(NAS Colloquium) Geology, Mineralogy, and Human Welfare

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

i

COLLOQUIUM ON GEOLOGY, MINERALOGY, AND HUMAN WELFARE

NATIONAL ACADEMY OF SCIENCES WASHINGTON, D.C. 1999

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ii

NATIONAL ACADEMY OF SCIENCES

Colloquium Series In 1991, the Nationfal Academy of Sciences inaugurated a series of scientific colloquia, five or six of which are scheduled each year under the guidance of the NAS Council's Committee on Scientific Programs. Each colloquium addresses a scientific topic of broad and topical interest, cutting across two or more of the traditional disciplines. Typically two days long, colloquia are international in scope and bring together leading scientists in the field. Papers from colloquia are published in the Proceedings of the National Academy of Sciences (PNAS).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

GEOLOGY, MINERALOGY, AND HUMAN WELFARE

Geology, Mineralogy, and Human Welfare

A Colloquium sponsored by the National Academy of Sciences November 8-9, 1998 PROGRAM Sunday, November 8, 1998 AGRICULTURAL MINERALOGY/SOILS, SURFACES Garrison Sposito, University of California, Berkeley, Surface geochemistry of the clay minerals 9:40am 10:20am Paul M. Bertsch, University of Georgia/SREL, Characterization of complex mineral assemblages: implications for contaminant transport and environmental remediation 11:20am Samuel J. Traina, Stanford University, Contaminant bioavailability in soils, sediments and aquatic environments Gordon E. Brown, Jr., Stanford University, Mineral surfaces and bioavailability of heavy metals: A molecular-scale perspective 12:00pm AEROSOLS AND CLIMATE Joseph M. Prospero, University of Miami RSMAS, Long-range transport of mineral dust in the global atmosphere: Impact of African 2:00pm dust on the environment of the southeastern United States 2:40pm Peter R. Buseck, Arizona State University, Airborne minerals and related aerosol particles: Effects on climate and the environment OCEANS AND BIOMINERALOGY Miriam Kastner, Scripps Institution of Oceanography, Oceanic minerals and rocks, their origin, occurrence, and economic significance 3:40pm 4:20pm Keith A. Kvenvolden, U.S. Geological Survey, Potential effects of gas hydrates on human welfare Jillian F. Banfield, University of Wisconsin, Madison, Biological impact on mineral dissolution - application of the lichen model to 5:00pm understanding mineral weathering in rhizosphere

iii

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

GEOLOGY, MINERALOGY, AND HUMAN WELFARE

Monday, November 9, 1998 RADWASTE, MINING, AND ENVIRONMENTAL ISSUES 9:10am Rodney C. Ewing, University of Michigan, Nuclear waste forms of actinides 9:50am Robert B. Finkelman, U.S. Geological Survey, The health impacts of domestic coal use in China 10:50am Robert P. Nolan, City University of New York, A risk assessment for exposure to grunerite asbestos (amosite) in an iron ore mine 11:30am D. Kirk Nordstrom, U.S. Geological Survey, Negative pH, efflorescent mineralogy, and consequences for environmental restoration at the Iron Mountain Superfund site, California PURE AND APPLIED MINERALOGY 2:00pm David R. Pevear, Exxon Production Research Co., Illite and hydrocarbon exploration 2:40pm Jeffrey E. Post, Smithsonian Institution, National Museum of Natural History, Manganese oxide minerals: crystal structures & economic and environmental significance 3:40pm John D. Sherman, UOP Research Center, Synthetic zeolites and other microporous oxide molecular sieves 4:20pm Frederick A. Mumpton, SUNY - College, Natural zeolites - la rocca magica? J.V. Smith, University of Chicago, Biochemical evolution: III. Polymerization on organophilic silica-rich surfaces; crystal-chemical 5:00pm modeling; formation of first cells; geological clues

iv

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

TABLE OF CONTENTS

v

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA

Table of Contents

Geology, mineralogy, and human welfare Joseph V. Smith Characterization of complex mineral assemblages: Implications for contaminant transport and environmental remediation Paul M. Bertsch and John C. Seaman Surface geochemistry of the clay minerals Garrison Sposito, Neal T. Skipper, Rebecca Sutton, Sung-ho Park, Alan K. Soper, and Jeffery A. Greathouse Contaminant bioavailability in soils, sediments, and aquatic environments Samuel J. Traina and Valérie Laperche Airborne minerals and related aerosol particles: Effects on climate and the environment Peter R. Buseck and Mihály Pósfai Oceanic minerals: Their origin, nature of their environment, and significance Miriam Kastner Mineral surfaces and bioavailability of heavy metals: A molecular-scale perspective Gordon E. Brown, Jr., Andrea L. Foster, and John D. Ostergren Long-range transport of mineral dust in the global atmosphere: Impact of African dust on the environment of the southeastern United States Joseph M. Prospero Biological impact on mineral dissolution: Application of the lichen model to understanding mineral weathering in the rhizosphere Jillian F. Banfield, William W. Barker, Susan A. Welch, and Anne Taunton A risk assessment for exposure to grunerite asbestos (amosite) in an iron ore mine R. P. Nolan, A. M. Langer, and Richard Wilson Potential effects of gas hydrate on human welfare Keith A. Kvenvolden Health impacts of domestic coal use in China Robert B. Finkelman, Harvey E. Belkin, and Baoshan Zheng Nuclear waste forms for actinides Rodney C. Ewing Illite and hydrocarbon exploration David R. Pevear Manganese oxide minerals: Crystal structures and economic and environmental significance Jeffrey E. Post Negative pH, efflorescent mineralogy, and consequences for environmental restoration at the Iron Mountain Superfund site, California D. Kirk Nordstrom and Charles N. Alpers La roca magica: Uses of natural zeolites in agriculture and industry Frederick A. Mumpton Synthetic zeolites and other microporous oxide molecular sieves John D. Sherman Biochemical evolution III: Polymerization on organophilic silica-rich surfaces, crystal–chemical modeling, formation of first cells, and geological clues Joseph V. Smith, Frederick P. Arnold, Jr., Ian Parsons, and Martin R. Lee

3348–3349 3350–3357 3358–3364 3365–3371 3372–3379 3380–3387 3388–3395 3396–3403 3404–3411 3412–3419 3420–3426 3427–3431 3432–3439 3440–3446 3447–3454 3455–3462 3463–3470 3471–3478 3479–3485

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

vi

LIST OF ATTENDEES

List of Attendees Charles N. Alpers, US Geological Survey Phyl Amadi, Salt River Project David Applegate, Geotimes Ellen Avery Jillian F. Banfield, University of Wisconsin-Madison Paul M. Bertsch, University of Georgia, SREL Gordon E. Brown, Stanford University Peter R. Buseck, Arizona State University Fernando Camara, Arizona State University Winifred Caponigri, Holy Cross College Bill Casey, University of California, Davis Sing-Foong Cheah, University of California, Berkeley Ron Churchill, California Department of Conservation John Clinkenbeard, California Division of Mines and Geology Patricia Colville, James Hardie Building Products Michael L. Cummings, Department of Geology Frank Cynar Tammy Dickinson, National Research Council Katerina Edward W.G. Ernst, Stanford University Rodney C. Ewing, University of Michigan Robert B. Finkelman, U.S. Geological Survey Andrea Foster, Stanford University Julia Gaudinski, Univeristy of California, Irvine Tom Gihring, University of Wisconsin-Madison M. Charles Gilbert, University of Oklahoma Janet Gordon, Pasadena City College Priscilla C. Grew, University of Nebraska-Lincoln Shalini Gupta Noel Heim George R. Helz JoAnn Holloway, Land, Air and Water Resources Robert M. Housely Rick Humphreys, State Water Resouces Control Board Dawn Janney, Arizona State University Miriam Kastner, Scripps Institution of Oceanography Christopher Kim, Stanford University Stephan Kraemer, University of California, Berkeley Konrad B. Krauskopf, Stanford University Keith A. Kvenvolden, U.S. Geological Survey Jia Li, Arizona State University Gwen Loosmore, Lawrence Livermore National Laboratory Elizabeth Magno Juraj Majzlan, University of California, Davis Kevin Mandernack, Colorado School of Mines Carrie Masiello, University of California, Irvine Carleton B. Moore Frederick A. Mumpton, SUNY-College Paulina Mundkowski, University of Chicago Alexandra Navrotsky, University of California, Davis William Nesse, University of Northern Colorado Heino Nitsche, University of California, Berkeley Robert Nolan, City University of New York D. Kirk Nordstrom, U.S. Geological Survey Everest Tan Ong, University of Chicago John Ostergren, Stanford University Sung-Ho Park, University of California, Berkeley Jill D. Pasteris, Washington University Adina Paytan, University of California, San Diego Erich U. Petersen David R. Pevear, Exxon Production Research Co. Bernard Pipkin F.D. Pooley, Cardiff University Jeffrey E. Post, Smithsonian Institution Joseph M. Prospero, University of Miami, RSMAS Jim Ranville, Colorado School of Mines Malcom Ross, U.S. Geological Survey Donald Runnells, Shepherd Miller, Inc. Martin S. Rutstein, State University of New York John D. Sherman, UOP Research Center David K. Shuh, Lawrence Berkeley National Laboratory Fiorella Simoni, George Mason University J.V. Smith, University of Chicago Neville Smith, Lawrence Berkeley National Laboratory Garrison Sposito, University of California, Berkeley John O. Strong Rebecca Sutton, University of California, Berkeley Steve Sutton Alexis Templeton, Stanford University Samuel J. Traina, Stanford University Flavio Vasconcelos, Colorado School of Mines Glenn Waychunas, Lawrence Berkeley Laboratory Potter Wickware, Molecular Applications News Erika Williams, University of Michigan Howard G. Wilshire

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

GEOLOGY, MINERALOGY, AND HUMAN WELFARE

3348

Geology, mineralogy, and human welfare

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3348–3349, March 1999 Colloquium Paper This paper is the introduction to the following papers, which were presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. JOSEPH V. SMITH* PNAS is available online at www.pnas.org.

Department of Geophysical Sciences and Center for Advanced Radiation Sources, 5734 S. Ellis Avenue, The University of Chicago, Chicago, IL 60637 The complex sciences of geology and mineralogy couple the focused sciences of physics, chemistry, and biology to the “diffuse” disciplines of ecology and the environment. For this colloquium, 18 papers have been selected on matters related to human welfare, particularly health and both physical and mental wellbeing, to demonstrate the importance of new research plans and new instrumentation. Agricultural Mineralogy, Soils, and Surfaces. Emerging “chemical microscopes” using neutrons, synchrotron x-rays, and electrons allow physicochemical characterization of mineral surfaces and adsorbed molecules and ions in soils (1). Plant growth depends on subtle interactions between mineral surfaces, saline fluids, and microbes. Incorporation of “good” trace elements into food depends on the interaction of organic and inorganic components, as does that of toxic ones. One-quarter of the wheat and rice crops are lost to Mn-oxidizing bacteria. Soils become contaminated with mobile toxic elements, including Pb, Cd, Se, and As, which can affect plant growth and food safety (2, 3 and 4). Aerosols and Climate. Mineral dust blown from drying geological basins pervades the atmosphere, and falls to earth with both good results—loess soil in central Europe, China, and North America was important for early agriculture, and still supports large populations—and bad ones—air pollution causes lung problems (5). Chemical microscopes provide physicochemical analysis of tiny particles, particularly useful for distinguishing the natural and industrial components (6). Oceans. Minerals in the oceans range from several dozen types in bioorganisms to various zeolites grown from volcanic ash, sulfides in hot smokers in volcanic ridges, and precipitates in Mn-rich nodules (7). The reactions with sea-water depend on temperature and composition and ultimately can be related to climate and plate-tectonic processes. Highly touted as a major energy source for the future are the methanewater clathrate beds on cold ocean beds; however, current evidence is not promising for successful commercialization (8). Biomineralogy. Chemical microscopes coupled with biochemical techniques are opening up a rapidly expanding field of studies on microbes. Before these tools were developed, mineralogists could only speculate on how microbes concentrated useful elements (including uranium!) into ore bodies, how microbes interact with atmospheric gases to modify the climate, how deep-seated ones relate to the spatial distribution and chemical signatures of natural gas and oil in the continents, and so on. Microbes can make organic acids that accelerate mineral weathering to make soil minerals (good), and eat away outdoor statues (bad) (9). Honeycombed surfaces of weathered feldspars may have been the first home of primitive cells where they were protected from destruction by solar ultraviolet radiation. The internal hydrophobic silica-rich surface in nanometer-wide channels of a zeolite mineral formed from abundant volcanic ash (e.g., mutinaite = synthetic silicalite-ZSM-5) might have scavenged organic species from the proverbial water-rich “soup” and catalyzed assembly into primitive polymers that extruded like spaghetti to become tangled up to form the nucleus of a proto-cell (10). Radwaste, Mining, and Environmental Issues. Perhaps a billion people, including some in developed countries, currently ingest harmful amounts of toxic elements. A prime aim of this colloquium was careful evaluation of selected problems, and establishment of an international plan for coupling scientific and administrative skills for mitigation (11, 12, 13 and 14). Pure and Applied Mineralogy. Improvements in the chemistry of electric storage batteries are related to mineralogy (e.g., long-life Pb; high-energy Li, etc). Over 400 minerals contain Mn; we selected Mn-oxides for presentation because of advances in understanding their chemistry using the new chemical microscopes (15). Perhaps the most spectacular advances have been in the petroleum industries, to the great benefit of the consumer. Three-dimensionalseismic imaging, slant drilling, and other engineering advances are tripling the recovery of petroleum from geologic reservoirs and actually advancing the provable reserves (although most prognostications assert that supply will not meet demand some time before 2030). The value of mineral geochemistry was illustrated by use of subtle argon-age dating of clay minerals across a potential basin to predict its yield of oil (16). An even broader success story has been the invention of zeolite/molecular sieve adsorbent/catalysts and industrial development of myriad applications (17). Almost unknown to all but the zeolite chemists and engineers are everyday applications: the 3-fold increase in yield of gasoline from petroleum from tailored zeolite catalysts, also higher octane number and lower pollutants in automobile exhausts; clearer multipane windows from zeolite adsorbing dirty vapors; safer brakes in trucks and trains from pressure-swing zeolite adsorbent; longer life of refrigerators; and selective adsorbents in nuclear waste. Just moving forward into major use are natural zeolites whose low cost and high exchange capacity are leading to bulk applications in agriculture, gardening, and waste management (18). Geoscientists now have the “chemical microscopes” and other tools to study the scientific characteristics of toxic materials, and we can now study at the atomic level those interactions between the inorganic and organic worlds that have positive aspects for human welfare. However, the available funds are far too small to properly service the growing community of environmental geoscientists. Hence, the colloquium concluded with presentations on a concept for establishment of a new efficient program for instrumentation in the environmental sciences costing only $100 million to be complemented by a similar one for research, teaching, and public

*e-mail: [email protected].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

GEOLOGY, MINERALOGY, AND HUMAN WELFARE

3349

outreach over the initial five years at universities, colleges, and experimental stations. An additional $100 million is needed to bring together research professionals and students around the world to quantify the dangers to human health of toxic elements, including As, Se, and Pb; to devise plans for dissemination of information; and to evaluate ideas for remediation in the context of diplomatic, social science, and economic planning procedures. To conclude, particularly important for this colloquium is that the boundaries between subdisciplines are falling; that humans are now moving as much material on the Earth's surface as geological processes; that natural climatic changes are being modified willy-nilly by human activities; and that the recent increase in population and use of energy is beginning to slow down as fundamental limits are approached, but they may not slow down fast enough. Doubling of the population might be sustainable, but quadrupling almost certainly would lead to serious problems and possible catastrophes. Biological evolution, as seen in the context of geologic time, indicates that fortune goes with increasingly skilful use of resources of many types, not the maximum use of resources.

1. Sposito, G., Skipper, N. T., Sutton, R., Park, S.-h., Soper, A. K. & Greathouse, J. A. (1999) Proc. Natl. Acad. Sci. USA 96, 3358–3364. 2. Bertsch, P. M. & Seaman, J. C. (1999) Proc. Natl. Acad. Sci. USA 96, 3350–3357. 3. Traina, S. J. & Laperche, V. (1999) Proc. Natl. Acad. Sci. USA 96, 3365–3371. 4. Brown, G. E., Jr., Foster, A. L. & Ostergren, J. D. (1999) Proc. Natl. Acad. Sci. USA 96, 3388–3395. 5. Prospero, J. M. (1999) Proc. Natl. Acad. Sci. USA 96, 3396–3403. 6. Buseck, P. R. & Pósfai, M. (1999) Proc. Natl. Acad. Sci. USA 96, 3372–3379. 7. Kastner, M. (1999) Proc. Natl. Acad. Sci. USA 96, 3380–3387. 8. Kvenvolden, K. A. (1999) Proc. Natl. Acad. Sci. USA 96, 3420–3426. 9. Banfield, J. F., Barker, W. W., Welch, S. A. & Taunton, A. (1999) Proc. Natl. Acad. Sci. USA 96, 3404–3411. 10. Smith, J. V., Arnold, F. P., Jr., Parsons, I. & Lee, M. R. (1999) Proc. Natl. Acad. Sci. USA 96, 3479–3485. 11. Ewing, R. C. (1999) Proc. Natl. Acad. Sci. USA 96, 3432–3439. 12. Finkelman, R. B., Belkin, H. E. & Zheng, B. (1999) Proc. Natl. Acad. Sci. USA 96, 3427–3431. 13. Nolan, R. P., Langer, A. M. & Wilson, R. (1999) Proc. Natl. Acad. Sci. USA 96, 3412–3419. 14. Nordstrom, D. K. & Alpers, C. N. (1999) Proc. Natl. Acad. Sci. USA 96, 3455–3462. 15. Post, J. E. (1999) Proc. Natl. Acad. Sci. USA 96, 3447–3454. 16. Pevear, D. R. (1999) Proc. Natl. Acad. Sci. USA 96, 3440–3446. 17. Sherman, J. D. (1999) Proc. Natl. Acad. Sci. USA 96, 3471–3478. 18. Mumpton, F. A. (1999) Proc. Natl. Acad. Sci. USA 96, 3463–3470.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3350

Characterization of complex mineral assemblages: Implications for contaminant transport and environmental remediation

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3350–3357, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. PAUL M. BERTSCH* AND JOHN C. SEAMAN PNAS is available online at www.pnas.org.

Advanced Analytical Center for Environmental Sciences, Savannah River Ecology Laboratory, The University of Georgia, Drawer E, Aiken, SC 29802 ABSTRACT Surface reactive phases of soils and aquifers, comprised of phyllosilicate and metal oxohydroxide minerals along with humic substances, play a critical role in the regulation of contaminant fate and transport. Much of our knowledge concerning contaminant-mineral interactions at the molecular level, however, is derived from extensive experimentation on model mineral systems. Although these investigations have provided a foundation for understanding reactive surface functional groups on individual mineral phases, the information cannot be readily extrapolated to complex mineral assemblages in natural systems. Recent studies have elucidated the role of less abundant mineral and organic substrates as important surface chemical modifiers and have demonstrated complex coupling of reactivity between permanent-charge phyllosilicates and variable-charge Fe-oxohydroxide phases. Surface chemical modifiers were observed to control colloid generation and transport processes in surface and subsurface environments as well as the transport of solutes and ionic tracers. The surface charging mechanisms operative in the complex mineral assemblages cannot be predicted based on bulk mineralogy or by considering surface reactivity of less abundant mineral phases based on results from model systems. The fragile nature of mineral assemblages isolated from natural systems requires novel techniques and experimental approaches for investigating their surface chemistry and reactivity free of artifacts. A complete understanding of the surface chemistry of complex mineral assemblages is prerequisite to accurately assessing environmental and human health risks of contaminants or in designing environmentally sound, cost-effective chemical and biological remediation strategies. The transport and fate of contaminants in soils and groundwater are highly coupled to the nature and relative abundance of the reactive mineral phases. Clay and oxide minerals, along with humified organic matter, comprise the surface reactive phases that are the primary controllers of sorption processes in soils, thus serving as important regulators of contaminant transport. Major challenges in understanding the processes controlling contaminant behavior in the environment include the complexity of the soil and aquifer matrix and the enormous spatial scales over which these processes occur. Although it is well established that a fundamental understanding of molecular-level interactions is required to explain the underlying mechanisms controlling the fate and transport of solutes and contaminants in soils and subsurface environments, there has been limited success in translating molecular-level information to observations made at the larger scales. Although several explanations for this conundrum can be advanced, a prominent one is that much of our knowledge concerning the surface chemistry of clay and oxide minerals primarily is derived from experiments conducted on model mineral phases. These studies have established boundary conditions defining sorbate/mineral surface interactions and have identified the surface functional groups involved in surface complexation reactions, but they have produced little information that can be readily extrapolated to complex mineral assemblages typically present in heterogeneous soil and aquifer materials (1). Thus, utilization of bulk mineralogical data to represent predominant reactive phases in complex natural systems often has failed to reliably predict solute and contaminant behavior. Reactive Mineral Phases in Soils: An Historical View The pioneering work on reactive mineral phases in soils, which focused primarily on adsorption of group IA and IIA cations, has been comprehensively reviewed (2, 3) as has more recent work on specific sorption of metals and metalloids (1, 4). The concept that surface reactive phases in soils are colloidal and comprised of Al(OH)3, Fe(OH)3, and SiO2 hydrogels was proposed over a century ago (5). By the mid-1920s, a comprehensive understanding of the surface chemistry of Al, Fe, and Si colloids and their role in cation sorption was emerging, largely based on extensive investigations of Mattson (6, 7 and 8). Mattson viewed reactive phases in soils as mixtures of Al2O3, Fe2O3, and SiO2 colloids. Based on the observation that soils with a high SiO2/Al2O3 + Fe2O3 ratio had higher cation exchange capacities (CEC) and that soils with a low SiO2/Al2O3 + Fe2O3 ratios had high anion exchange capacities at low pH and higher CEC at high pH, Mattson concluded that the SiO2 colloids were primarily responsible for the CEC of a soil and that the Al2O3 and Fe2O3 colloids were amphoteric in nature. Mattson's compelling evidence for mixtures of positively and negatively charged colloids, based on careful cation/anion sorption experiments and electrophoresis, was largely disregarded as attention shifted to a new, rapidly emerging paradigm of reactive mineral phases predicated on the notion that soil clays were comprised primarily of crystalline phases. Two classic papers by Pauling (9, 10) figured prominently in this paradigm shift. Shortly thereafter, soil chemists applied x-ray diffraction to soil clays and discovered the existence of, and delineated the structures for, the major classes of phyllosilicate clays commonly found in soils (11, 12). Soon after it was demonstrated that most phyllosilicates in soils had a permanent negative charge resulting from substitution of lower valence cations in both the tetrahedral and octahedral layers (13). For decades the surface chemistry of reactive phases in soils would be interpreted primarily according to this paradigm, i.e., that predominant reactive phases in soil were

Abbreviations: pznc, point of zero net charge; EM, electron microscopy. *To whom reprint requests should be addressed.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3351

crystalline and comprised of negatively charged minerals of the phyllosilicate class. Three rather fortuitous circumstances solidified this view of reactive mineral phases. First, free Fe oxides and organic macromolecules were removed from soil clay fractions via pretreatment to improve x-ray diffraction patterns by minimizing background scatter and improving preferred orientation of the phyllosilicate clay minerals. Second, most of the active soil mineralogy groups emerging during this period were limited to geographical areas characterized by young circumneutral soils having clay fractions dominated by 2:1 phyllosilicates; albeit, on a worldwide basis these soils were more of an exception. Finally, much of the experimentation during this period continued to involve the adsorption/exchange of class IA and IIA cations, both of which are relatively weakly bound and present at relatively high concentrations (an exception is K+, whose chemistry is controlled by a unique combination of cation size, low hydration energy, and structural properties of micaceous minerals and their weathering products). Thus, much of the data generated under these conditions was consistent with the phyllosilicate model, and the distribution of phyllosilicates within a given soil clay fraction generally could be used to predict observed cation exchange behavior for this limited range of extensively studied soils. There continued to be prominent exceptions to this model that could be better interpreted according to a Mattson-like model of surface reactive phases (14, 15). Evidence for anion adsorption to soil clays having low SiO2/Al2O3 + Fe2O3 ratios and slightly acidic pH continued to appear. Evidence for positively charged regions (edge sites) on phyllosilicate clays in slightly acidic suspensions appeared during this time (16, 17, 18 and 19). This model also was used to interpret anion adsorption and complex flocculation/dispersion behavior of kaolinite suspensions (20). Clearly, the phyllosilicate model of reactive mineral phases based largely on 2:1 minerals in soils of circumneutral pH was limited in its extent of applicability. As mineralogical techniques improved and experimental approaches evolved, another very important body of literature on hybrid phyllosilicate-Al/Fe oxohydroxides emerged. The discovery (21) that 2:1 minerals in soils weathered from parent materials rich in mica schist were interlayered with nonexchangeable, positively charged hydroxo-Al polynuclear components stimulated a significant body of research that continues to this day and includes investigations on an important class of zeolite-like clay catalysts (22). Although this finding explained a number of properties related to the surface chemistry of many 2:1 soil clays, the research emphasis on the hydroxy-interlayered minerals largely focused on explaining the unique adsorption behavior of large weakly hydrated monovalent cations, such as K+, NH4+, and Cs+, with less emphasis on anion sorption. Only many years later would the role of this complex mineral assemblage in the specific sorption of transition metals be considered (22). Concurrent with these exciting developments, a new paradigm of surface reactive mineral phases was emerging. The structural aspects of important functional groups on oxide minerals were being unraveled (4, 23). The notion of surface structural hydroxyl groups having acid/base properties that could quantitatively explain the observed amphoteric behavior of oxides became firmly established (24). Thus, an accurate model of surface functional groups that could explain Mattson's original observations was emerging, and a number of studies on anion adsorption to metal oxide surfaces followed quickly as did spectroscopic evidence for the proposed reactive surface hydroxyls (4). It was now established that solutes could interact with charged metal oxide surfaces via electrostatic (outer sphere) reactions or through specific ligand exchange reactions with the surface functional groups (inner sphere). The conceptual model of surface complexation to describe nonspecific and specific adsorption of anions and cations was advanced shortly thereafter by the classic work of Schindler and Gamsjager (25) and Stumm et al. (26). The surface complexation model has remained the basic framework for research on metal and anion sorption to metal oxide surfaces to the present time (1, 4), and many studies have demonstrated the importance of metal oxides as resident phases for a variety of metals and metalloids (1, 27). Although extensive modeling efforts have demonstrated reasonable success for predicting metal and metalloid sorption to model monomineralic metal oxide phases, applications to natural systems have been less than satisfying (1). A major challenge in extending such results to complex mineral assemblages typically found in nature has been the identification and quantification of the primary reactive phase and associated surface functional groups. High surface area, low abundance metal oxohydroxide phases, and organic materials can be coassociated with more prominent mineral grains as armoring agents or as surface coatings. The term surface coating as used here does not imply the presence of a uniform gel-like phase as is often envisioned. Rather, it is used to describe domains of crystalline or noncrystalline components coassociated with well-defined mineral grains. The complex nature of the electrostatic and van der Waals interactions between finegrained crystalline and poorly ordered phases with mixed surface-charge properties has hampered the development of suitable models to represent surface reactive functional groups in mixed mineral assemblages. Adsorption studies using binary mixtures of model mineral phases have demonstrated remarkable complexity, with adsorption generally being very poorly predicted by considering a weighted sum of individual mineral components (1, 28). Mixed Mineral Assemblages in Natural Systems It is becoming increasingly clear that many natural mineral phases possess different surface chemical properties than their model mineral analogues. Zachara and others (29, 30 and 31) have provided compelling evidence suggesting that the small crystallite size of soil smectites enhances the importance of edge site aluminol (Al-OH) functional groups imparting an oxide-like behavior compared with the widely used Source Clay Repository, SWy-1 montmorillonite. Other studies have suggested that organic or metal oxide minerals may be the primary reactive phases in soils and sediments even at relatively low abundance (32, 33, 34, 35, 36 and 37). A major theme that emerges from these investigations is that surface modifiers in the form of organic/ metal oxohydroxide armoring agents or coatings, rather than bulk mineralogical composition per se, control the surface chemistry of reactive phases in soils and aquifers. For example, it has been demonstrated that organic constituents coassociated with variable charge minerals significantly alter the point of zero net charge (pznc, the pH at which the cation and anion exchange capacities are equal), shifting the pznc to significantly lower pH values (32, 33, 35, 36 and 37). Conversely, Fe and Al oxohydroxide phases coassociated with quartz, and permanent charge phyllosilicate minerals have been found to shift the pznc to higher pH values (38, 39 and 40). A number of recent studies have focused on the surface chemistry of mineral assemblages isolated from natural systems (33, 34 and 35, 37, 38, 40, 41 and 42). Characterizing natural mineral assemblages is challenging, because it has been virtually impossible to isolate them free of artifacts. In fact, the methods used to isolate and concentrate clay minerals involve dispersion of the clay fraction via treatment with harsh reagents designed to significantly alter surface charge properties and destroy complex mineral assemblages present in the original material. Recently, however, collection and examination of complex mineral assemblages has been achieved in a different context: that dealing with the transport of colloidal phases through porous media. The past decade has witnessed great

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3352

interest in the generation and transport of mineral colloidal phases through natural porous media (33, 34, 40, 42, 43, 44, 45, 46 and 47). Interest in this subject has paralleled evidence that colloidal minerals are important vectors for facilitating the transport of contaminants in certain environments (43, 47, 48, 49 and 50). Mobile colloids can be generated by a number of mechanisms, including precipitation of colloidal size phases, dissolution of cementation agents composed of fine-grained crystalline and poorly crystalline secondary minerals, and release from soil and aquifer materials via physicochemically controlled dispersion processes. Transport of the mineral colloids also depends on a number of factors, including fluid flow rate, electrostatic and van der Waals forces between colloids and between colloids and matrix minerals, and physical factors related to the relative size of the colloids and pores and pore throats. Recent evidence has indicated that mobile colloids comprised of minerals and complex mineral assemblages can be generated via dispersion processes and transported through many soils and groundwater systems with relatively minor changes in solute chemistry of the invading fluid, thereby avoiding the serious artifacts typically encountered in the isolation of complex mineral assemblages. Mineralogy and Surface Chemistry of Complex Mineral Assemblages Isolated From Soils and Aquifer Materials Although several studies have demonstrated the enhanced mobility of contaminants in the presence of mobile colloids, far fewer have focused on characterizing the mineralogical composition and surface charge properties of the mineral and organic-mineral assemblages comprising the mobile phase. Over the past decade, our investigations have focused on providing evidence for the facilitated transport of contaminants associated with mobile colloidal phases and in defining the mechanisms leading to the generation and transport of mobile colloidal phases and solutes (33, 34, 40, 46, 49, 51, 52 and 53). These studies have examined surface chemical controls on colloid generation and of colloid and solute migration in surface and subsurface highly weathered oxide-rich systems having similar bulk clay mineralogy. The samples examined are coarse textured (≥85% sand; <9% clay), have varying quantities of Fe-oxide and organic carbon, a predominance of exchangeable Al, low pH, low pore water ionic strength, and similar bulk clay mineralogies (Table 1). They are also poorly structured with little or no evidence for meso/macropore development, thus minimizing complications involving preferential flow. They both have silt and sand fractions composed entirely of quartz, thus minimizing artifacts resulting from dissolution of less stable minerals that contribute solutes and complicate solution chemistry. The properties of these highly weathered soils and aquifer materials have important commonalities with those widely distributed in humid tropics, which play an important role in global geo-chemical cycles (41). Table 1. Chemical and mineralogical characteristics of a surface soil (Orangeburg Series) and three subsurface sediments representative of the Tobacco Road (TR) formation from the Atlantic Coastal Plain Orangeburg series TR1 TR2 TR3 pH 4.61 5.37 5.33 5.18 Extractable Fe (mg kg−1) CDB* 15.0 73.5 111.0 359.4 3.1 1.4 1.2 2.3 NH4-oxalate k, HIV, gibb k, m, g k, m, g g, k, m Clay mineralogy k, kaolinite; HIV, hydroxy-interlayered vermiculite; m, mica (illite); gibb, gibbsite; g, goethite. *Citrate dithionite extraction.

Our results have demonstrated that colloids mobilized from surface soils have high negative electrophoretic mobilities (−2.5 to −3.5 µm cm s−1·V−1), inconsistent with mineral composition of the assemblages and pore water pH. The mobile colloids generally are enriched in kaolinite and Al- and Fe-oxohydroxide phases in the ≈ 200-nm size range and more dilute in quartz and hydroxy-interlayered vermiculite (≈700 nm to 1 µm) relative to the bulk clay mineralogy of the soil horizons where the colloids were generated. More colloids were generated in surface soils having slightly elevated pH (≥5.5) and low total electrolyte concentrations (<2 molc m−3). The mobile colloidal phase also was found to have high concentrations of surface-associated organic C, which explained the anomalous high negative electrophoretic mobilities (33, 34). As much as 50% of the total organic C in the surface soil horizons was associated with the water-dispersible clay, a fraction of the total clay thought to be an indicator of the potential for mineral colloids to become mobilized under physicochemical perturbation (34, 44). These results suggest that, consistent with previous investigations on model organic-clay mineral suspensions and on natural mineral assemblages, the surface charging mechanism of both the mobile and immobile reactive phases in these surface soils was dominated by organic-mineral interactions or surface coatings, which lower the pznc of the mineral assemblage compared with the model mineral phases of similar composition. Surprisingly, the generation and transport of mobile colloidal phases and the composition of these phases were extremely sensitive to minor changes in pore water chemistry and to a narrow range in pore water flow velocities (33, 34, 44). Thus, the generation and transport of colloids from managed surface soil systems, such as might be experienced in remediation of contaminated soils or the reclamation of highly disturbed sites, is an extremely dynamic and complex process. Similar to the surface soils, minor differences in solute chemistry were found to have a profound influence on colloidal generation and transport in the subsurface aquifer materials. In contrast to the results obtained from surface soils, those derived from the organic poor subsurface materials indicate a completely different surface charging mechanisms despite the similar mineralogy of the bulk clay fractions. Dynamic column transport experiments with simulated groundwater and various salt solutions (K+, Na+, Mg2+, and Ca2+; Cl− and SO42−) at concentrations ranging from 1 × 10−4 to 0.1 M were used to examine the processes regulating surface chemistry of the mixed mineral systems. Information gleaned from trends in colloid transport, effluent pH, and solute leaching histories, combined with characterization of the mobile colloidal phases indicate a complicated surface charging mechanism in these mixed mineral systems. The generation and transport of mobile colloids was found to be minimal in the presence of groundwater or Na+ solutions, but significantly enhanced on introduction of dilute Ca2+ solutions, a result at odds with the well-established empirical observations of dispersion and flocculation behavior of soil colloids (Fig. 1; refs. 40 and 46). Effluent pH was found to decrease between 1 and 1.5 units concurrent with the introduction of the dilute Ca2+ salt solution, which contrasts to the groundwater and Na+ salt systems, where pH remained relatively constant. Electrophoretic mobility measurements of the colloids generated in the Ca2+ transport studies reveal a net positive surface charge, strongly suggesting reversal of the net matrix mineral assemblage charge from slightly negative to strongly positive on the introduction of the dilute Ca2+ salt solutions. Additional evidence that the immobile reactive mineral assemblages possessed a net positive charge is the significant retention of anionic tracers observed in transport experiments (see below). A major feature of the mobile colloid elution profile is the concentration maximum at ≈2.5 pore volumes of CaCl2 injectate that was followed by a rapid decrease in colloid

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3353

concentration between ≈2.5 and 5 pore volumes because of destabilization of the suspended colloids as the ionic strength of the efluent approaches that of the influent. Evidence to support this explanation is provided in the column colloid transport histories, which reveal a second colloid concentration maximum when deionized water was introduced as the influent solution after 10 pore volumes of the Ca2+ solution.

FIG. 1. Influence of treatment solution (1 × 10−3 and 0.1 M NaCl; 5 × 10−4 M CaCl2; 5 × 10−4 M CaCl2, pH 3.00) on effluent turbidity (a measure of colloid concentration) (a), pH (b), and the electrophoretic mobility (c) of mobile colloids from sample TR1 (Table 1). Columns were leached with a given influent for 10 pore volumes at a Darcy velocity of ≈ 0.72 m d−1 followed by several pore volumes of deionized (DI) water. To explain the complex surface charging processes observed in this mixed mineral system, we propose a mechanism involving the strong coupling between surface exchange reactions on permanent negatively charged phyllosilicate minerals (kaolinite and mica) and subsequent protonation of the variable-charge Fe-oxohydroxide mineral components (Fig. 2). According to this model, Ca2+ undergoes exchange with native cations associated with the negatively charged phyllosilicate minerals, of which Al occupies ≈85% of the exchange phase. Evidence for ion exchange of Ca2+ with native cations is provided by examining the effluent discharge concentrations of major cations from the column, where the delay in Ca2+ transport is accompanied by the breakthrough of Mg2+ and trace levels of Na+ (data not shown). The model also suggests that the observed decrease in effluent pH is a result of exchanged Al undergoing hydrolysis reactions. Consistent with this model, we were unable to detect elevated Al in the effluents but could measure a decrease in cation exchangeable Al. Dilute Na+ influent solutions were ineffective at displacing native exchangeable cations associated with the phyllosilicate minerals, explaining the differences between the Na+ and Ca2+ systems in effluent pH and colloid transport histories. This mechanism was confirmed in transport experiments with more concentrated Na + solutions (0.1 M) where native cations, including Al, were displaced, as evidenced by elevated effluent cation concentration and depressed pH. Mobile colloids were not detected in these high Na+ systems, however, because of ionic strength destabilization. Consistent with the proposed mechanism, introduction of an acidified CaCl2 solution enhanced positive surface charge development as evidenced by the high mobile colloid yields throughout the leaching event, i.e., higher positive surface charging countered the ionic strength destabilization mechanism. Examination of the mobile colloids by transmission electron microscopy (EM) and selected area electron diffraction reveals that they are comprised primarily of aggregates of microcrystalline Al-substituted goethite along with complex mineral assemblages of goethite-armored kaolinite and crandallite (Ca Al3[PO4]2[OH]5·H2O) in the 200- to 300-nm size range (Fig. 3). Examination of mobile colloids by scanning EM reveals that the phyllosilicates and phosphate minerals present in the mobile phase are extensively armored in all instances and generally fall in the 100- to 300-nm size range. Examination of isolated bulk clay reveals that the kaolinite and crandallite are present in the bulk clay fraction in two major size populations; the 100- to 300-nm size class as observed in the mobile phase and the 700-nm to 1-µm size class, which is predominant and presumably more representative of the reactive minerals comprising the immobile matrix. The former size class also contains the micaceous minerals found in the bulk clay mineral fraction. Scanning EM images of the bulk clay suggest that even these larger mineral grains of kaolinite and micaceous minerals are partially armored along the negatively charged basal surfaces with goethite crystallites and

FIG. 2. Schematic representation of the coupled mechanisms controlling effluent pH and the generation of positive surface charge on the mobile colloids and the stationary matrix.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3354

FIG. 3. Transmission EM and selected-area electron diffraction pattern (Inset) for (A) crandallite and (B) kaolinite armored with finegrained goethite and (E) goethite aggregates generated during dynamic transport experiments with 5 × 10−4 M CaCl2 solutions. Transmission EM image of crandallite (C) and kaolinite (D) from bulk clay with Fe-oxides removed by dithionite extraction. Electrophoretic mobility behavior of (F) complex mineral aggregates generated during reactive transport experiments when repeatedly analyzed by laser doppler velocimetry: (a) initial mobility distribution, (b) third consecutive, (c) fifth consecutive, (d) sixth consecutive mobility distribution, respectively, and (e) typical mobility distribution observed after sample relaxation (t ≈ 5 min). (G) Average electrophoretic mobility of colloidal suspension as a function of consecutive analyses. (38).

that these assemblages often are coassociated with larger quartz grains as has been observed in at least one other natural subsurface system

Thus, in these oxide-rich subsurface systems, Feoxohydroxide surface modifiers increase the pznc of the complex mineral assemblages, resulting in surface reactivity

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3355

that is controlled by the development of a net positive surface charge. Based on electrophoretic mobility measurements of stable colloidal suspensions generated in the column transport experiments we conclude that the complex mineral assemblages observed in the EM images are not artifacts of sample isolation and preparation, i.e., surface armoring of negatively charged basal surfaces with Fe-oxohydroxide crystallites in the 10- to 30-nm size range is required to fully explain the observed high surface positive charge of the mineral aggregates. Further evidence for complex mineral aggregates comprised of a phyllosilicate/phosphate mineral core and a Fe-oxohydroxide veneer is provided on longer-term exposure of the assemblages to fluctuating electric fields. During consecutive electrophoretic mobility measurements we observed evidence that the complex mineral assemblages were disaggregating in the electric field, resulting in a suspension having two primary charged populations, one positively charged and one negatively charged (Fig. 3). The strong bias for enhanced scattering intensity observed for the negatively charged colloid population can be explained by the relative size and thus scattering by the 100–300 nm phyllosilicate and phosphate minerals dominates over the 10–30 nm goethite crystallites. When the disaggregated mineral mixture is allowed to reequilibrate in the absence of an electric field the system is found relax to the original state, i.e., the mixed mineral assemblage again is formed and the net positive charge reestablished. Other evidence of the rather fragile nature of the complex mineral assemblages in these natural systems was found in an attempt to conduct conventional batch flocculation/ dispersion and critical coagulation concentration experiments. On either air drying or physical disturbance of the sample, the surface chemistry of the mineral assemblages was found to be highly biased toward the permanently negatively charged mineral components in the mineral mixture, an observation consistent with several previous studies (56, 57). Even drying minimally disturbed columns via purging with Ar before conducting the dynamic transport experiments resulted in the complete loss of colloid generation and transport as observed for identically prepared columns that were maintained in the field moist condition. These observations have significant ramifications for how experiments are designed to examine surface chemical properties of mixed mineral assemblages. Furthermore, they provide some insight into the electrostatic forces involved in the stabilization of the complex mineral assemblages, although major challenges remain for developing methods for isolating, examining, and quantitatively describing the electrostatic/ van der Waals forces involved in their stabilization. Surface Chemistry of Mixed Mineral Assemblages and the Implication for Solute Transport Surface charge reversal of reactive mineral phases according to the proposed model should be manifested in the reactive transport behavior of anionic solutes. This phenomenon was examined by investigating Br− breakthrough behavior in additional dynamic transport experiments. Consistent with the proposed model based on colloid generation and transport behavior, Br− displayed significant retardation when referenced to the truly conservative 3H2O (Fig. 4). Also consistent with the proposed model was the greater retardation of Br− with increased abundance of Fe-oxohydroxide minerals. The fact that the surface sample displays enhanced Br− transport, compared with the 3H2O tracer, suggests anion exclusion, consistent with the proposed role of organic matter on surface charge characteristics of the mineralogically similar surface soil and, furthermore, demonstrates the sensitivity of dynamic transport compared with batch experiments for investigating subtle changes in surface charge characteristics of mixed mineral systems. Also the fact that Br− transport was more retarded in samples with higher pH (because of the greater abundance of Fe oxohydroxide minerals) provides additional evidence for the coupling of the surface reactions between the phyllosilicates and the Fe-oxohydroxides. The use of Mg2+ compared with K+ as the counter ion resulted in a greater decrease in effluent pH and greater Br− retardation, an observation also consistent with the proposed coupled reaction model (53). As with the colloid generation and transport behavior, disturbance of the samples for use in batch experiments or air drying before dynamic transport experiments greatly reduced or eliminated observed anion retardation (54). Finally, dynamic transport experiments conducted over a wide range of Br− concentrations revealed nonlinear sorption and at high concentrations of influent Br− salts (≥ 0.1 M) a nearly complete masking of the anion retardation behavior, with breakthrough appearing to be conservative (53, 54). This observation is critical because column transport experiments are commonly conducted to calibrate physical parameters for field scale tracer experiments. The concentrations used in these column experiments are typically 2–3 orders of magnitude higher than those used in the subsequent field scale tracer experiments. Thus, sorption reactions in the field typically would be misinterpreted as having physical significance (i.e., mixing, flow rate, permeability, regions of immobile/mobile water, stratification, etc.).

FIG. 4. Bromide breakthrough (A) and effluent pH (B) for 10−3 M KBr tracer solutions in columns packed with materials described in Table 1. Column pore volumes were calibrated based on tritium breakthrough (adapted from ref. 54). Surface Chemistry of Mixed Mineral Assemblages and the Implication for Contaminant Transport and Subsurface Remediation There are a number of important implications of this work to the modeling of contaminant transport and to the environmental remediation of contaminated aquifers. In oxide-rich,

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3356

organic-poor subsurface environments, typical of many aquifer systems, underestimation of contaminant transport distances can result from both an overestimation of contaminant sorption to reactive mineral phases that are assumed to be related to mineral abundance and assumed to posses a static surface chemistry, and by a misunderstanding of the primary mechanisms leading to the generation and transport of mineral colloids when predominantly negative charge surfaces are assumed to control the surface chemical behavior. For example, previous modeling efforts on coarse-textured highly weathered sediments similar to those studied here have considered quartz and kaolinite as the primary reactive phases and predicted limited metal and actinide mobility from an acidic plume (51). However, metal transport distances from the source plume were found to be significant and the primary mechanisms for this apparent enhanced transport were found to be a charge reversal on the matrix mineral phases leading to limited sorption reactions and the transport of trace levels of actinides and other metals specifically sorbed to colloids. Other studies also have demonstrated that the surface chemical properties of aquifer materials are dominated by high surface area phases of relatively low abundance (38). Greater emphasis must be placed on the identification and surface chemical characterization of complex mineral assemblages in natural systems to accurately define the mechanisms controlling solute and contaminant transport. An additional implication of these results relates to remediation of contaminated oxide-rich, organic-poor subsurface environments. Understanding the surface chemical controls of the reactive mineral phases has facilitated the development of an enhanced groundwater remediation technology, which is predicated on the selective mobilization through surface chemical manipulation of the highly reactive Feoxohydroxide phases and Fe-oxohydroxide armored minerals, which are the primary resident phase for both inorganic and organic contaminants (58). Conclusions Investigations of complex mineral assemblages in a highly weathered coarse-textured system have demonstrated that the surface chemistry of these assemblages is influenced by complex physicochemical interactions between natural organic constituents, phyllosilicate, and Feoxohydroxide phases. In surface soil environments containing as little as 1% organic matter, the surface chemistry was found to be controlled by organic constituents coating the phyllosilicate and Al- and Fe-oxohydroxide clay minerals, resulting in a much higher negative charge and a pznc shifted to much lower values than predicted based on bulk mineralogical composition. The surface charge modification by organic constituents was found to control the flocculation/dispersion processes of clay mineral assemblages in the surface soils, as well as the transport of mineral colloidal phases through soils. The complex mineral assemblages isolated from subsurface environments were found to be comprised of Fe-oxohydroxide phases partially or totally armoring the more abundant phyllosilicate minerals present. Unperturbed, these systems appear to be near the pznc; however, minor changes in solute chemistry can induce surface charge reversal (from slightly net negative to strongly net positive) through an elaborate coupled reaction between the permanent negatively charged and variable charged surfaces leading to the dispersion and transport of Fe-rich colloidal mineral phases. The surface charge reversal of these systems also is manifested by a significant retention of anions, such as Br −, which traditionally have been considered as conservative tracers in hydrological investigations of these systems. Defining the surface chemical properties of these complex mineral assemblages has a number of important implications for solute and contaminant transport. The development of robust predictive models describing solute and contaminant transport requires a thorough understanding of physical transport parameters (generally derived from solute tracer experiments) and solute/contaminant-mineral surface interactions. Most reactive transport modeling efforts define reactive mineral phases based on the relative abundance of the clay minerals present and surface chemistry defined by studies with model minerals. That the surface chemistry of reactive phases in aquifers may be controlled by minerals of relatively low abundance and more importantly, by complex physicochemical interactions occurring between individual components in complex mineral assemblages, suggests that we need less emphasis on studies of model minerals and more research on mineral assemblages isolated from natural systems. Clearly, the results of this and other investigations demonstrate that information based entirely on mineral abundance is insufficient for predicting the surface charge characteristics of natural mixed mineral systems. An important finding of these investigations is that the surface charge behavior defined in dynamic transport experiments could not be reproduced with conventional batch experiments. Both air drying or significant physical manipulation of the aquifer materials apparently results in a disruption of the complex mineral assemblages that appear to be primarily composed of a phyllosilicate core with a partial or total Feoxohydroxide veneer. Batch experiments were found to produce results strongly biased toward the permanent negative-charged components in the complex mineral assemblages and could not be used to predict solute transport or flocculation/dispersion behavior observed in both column and field scale transport experiments. These observations raise important questions concerning methods used for determining surface chemistry of complex mineral assemblages and pose significant challenges for designing isolation techniques so that complex mineral assemblages can be investigated in a meaningful way. It is clear that a detailed understanding of reactive mineral phases in soils and aquifers is necessary to accurately evaluate environmental and human health risks associated with contaminants and to design technologies for the protection or remediation of soil and water resources. This work was partially funded by Cooperative Agreement DE-F609–96SR18546 between the U.S. Department of Energy and The University of Georgia. 1. Davis, J. A. & Kent, D. B. (1990) in Reviews in Mineralogy, Mineral-Water Interface Geochemistry, eds. Hochella, M. F., Jr. & White, A. F. (Mineralogical Society of America, Washington, D.C.) Vol. 23, 177–260. 2. Kelley, W. P. (1948) Cation Exchange in Soils (Reinhold, New York). 3. Thomas, G. W. (1977) Soil Sci. Soc. Am. J. 41, 230–238. 4. Dzombak, D. A. & Morel, F. M. M. (1990) Surface Complexation Modeling, Hydrous Ferric Oxide (Wiley, New York). 5. Van Bemmelen, J. M. (1888) Landwirtsch. Vers. Stn. 35, 69–136. 6. Mattson, S. (1926) J. Agri. Res. 33, 553–567. 7. Mattson, S. (1927) Int. Congr. Soil Sci. Trans. 1st, 1927, 199–211. 8. Mattson, S. (1931) Soil Sci. 32, 343–365. 9. Pauling, L. (1930) Proc. Natl. Acad. Sci. USA 16, 123–129. 10. Pauling, L. (1930) Proc. Natl. Acad. Sci. USA 16, 578–582. 11. Hendricks, S. B. & Fry, W. H. (1930) Soil Sci. 29, 457–476. 12. Kelley, W. P., Dore, W. H. & Brown, S. M. (1931) Soil Sci. 31, 25–55. 13. Marshall, C. E. (1935) Z. Kristallogr. Mineral. 91, 433–449. 14. Schofield, R. K. (1939) Soils Fert. 2, 1–5. 15. Mehlich, A. (1952) Soil Sci. 73, 361–374. 16. van Olphen, H. (1950) Rec. Trav. Chim. 69, 1308–1312 (I), 1313–1322 (II). 17. van Olphen, H. (1951) Discuss. Faraday Soc. 11, 82–84. 18. Schofield, R. K. & Samson, H. R. (1953) Clay Miner. Bull. 2, 45–51.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CHARACTERIZATION OF COMPLEX MINERAL ASSEMBLAGES: IMPLICATIONS FOR CONTAMINANT TRANSPORT AND ENVIRONMENTAL REMEDIATION

3357

19. van Olphen, H. (1977) An Introduction to Clay Colloid Chemistry (Wiley, New York), 2nd Ed. 20. Schofield, R. K. & Samson, H. R. (1954) Discuss. Faraday Soc. 18, 135–145, 220. 21. Rich, C. I. & Obenshain, S. S. (1955) Soil Sci. Soc. Am. Proc. 19, 334–339. 22. Barnhisel, R. I. & Bertsch, P. M. (1989) Minerals in Soil Environments (Soil Science Soc. of America, Madison, WI), 2nd Ed., Soil Science Society of America Book Series, No. 1, pp. 729–788. 23. Parks, G. A. & DeBruyn, P. L. (1962) J. Phys. Chem. 66, 967–973. 24. Parks, G. A. (1965) Chem. Rev. 65, 177–198. 25. Schindler, P. W. & Gamsjager, H. (1972) Kolloid Z. Z. Polym. 250, 759–763. 26. Stumm, W., Huang, C. P. & Jenkins, S. R. (1970) Croat. Chem. Acta 42, 223–244. 27. Jenne, E. A. (1977) in Trace Element Sorption by Sediments and Soils-Sites and Processes, Symposium on Molybdenum in the Environment, eds. Chappell, W. & Peterson, K. (Dekker, New York )Vol. 2, pp. 425–553. 28. Hendershot, W. H. & Lavkulich, L. M. (1983) Soil Sci. Soc. Am. J. 47, 1252–1260. 29. Zachara, J. M., Smith, S. C., Resch, C. T. & Cowan, C. E. (1992) Soil Sci. Soc. Am. J. 56, 1074–1084. 30. Zachara, J. M., Smith, S. C., McKinley, J. P. & Resch, C. T. (1993) Soil Sci. Soc. Am. J. 57, 1491–1501. 31. Zachara, J. M. & Smith, S. C. (1994) Soil Sci. Soc. Am. J. 58, 762–769. 32. Tipping, E. (1981) Geochim. Cosmochim. Acta 45, 191–199. 33. Kaplan, D. I., Bertsch, P. M., Adriano, D. C. & Miller, W. P. (1993) Environ. Sci. Technol. 27, 1193–1200. 34. Kaplan, D. I., Bertsch, P. M. & Adriano, D. C. (1997) Soil Sci. Soc. Am. J. 61, 641–649. 35. Kretzschmar, R., Robarge, W. P. & Weed, S. B. (1993) Soil Sci. Soc. Am. J. 57, 1277–1283. 36. Kretzschmar, R., Hesterberg, D. & Sticher, H. (1997) Soil Sci. Soc. Am. J. 61, 101–108. 37. Heil, D. & Sposito, G. (1993) Soil Sci. Soc. Am. J. 57, 1246–1253. 38. Coston, J. A., Fuller, C. C. & Davis, J. A. (1995) Geochim. Cosmochim. Acta 59, 3535–3547. 39. Rengasamy, P. & Oades, J. M. (1977) Aust. J. Soil Res. 15, 235–242. 40. Seaman, J. C., Bertsch, P. M. & Strom, R. N. (1997) Environ. Sci. Technol. 31, 2782–2790. 41. Chorover, J. & Sposito, G. (1995) Geochim. Cosmochim. Acta 59, 875–884. 42. Kretzschmar, R., Robarge, W. P. & Amoozegar, A. (1995) Water Resourc. Res. 31, 435–445. 43. McCarthy, J. F. & Zachara, J. M. (1989) Environ. Sci. Technol. 23, 496–502. 44. Kaplan, D. I., Sumner, M. E., Bertsch, P. M. & Adriano, D. C. (1996) Soil Sci. Soc. Am. J. 60, 269–274. 45. Gschwend, P. M., Backhus, D. A., MacFarlane, J. K. & Page, A. L. (1990) J. Contam. Hydrol. 6, 307–320. 46. Seaman, J. C., Bertsch, P. M. & Miller, W. P. (1995) Environ. Sci. Technol. 29, 1808–1815. 47. Penrose, W. R., Polzer, W. L., Essington, E. H., Nelson, D. M. & Orlandini, K. A. (1990) Environ. Sci. Technol. 24, 228–234. 48. McCarthy, J. F. & Degueldre, C. (1993) in Environmental Particles—Part II: Sampling and Characterization of Particles of Aquatic Systems, eds. Buffle, J. & van Leeuwen, H. P. (Lewis, Ann Arbor, MI), pp. 247–315. 49. Kaplan, D. I., Hunter, D. B., Bertsch, P. M., Bajt, S. & Adriano, D. C. (1994) Environ. Sci. Technol. 28, 1186–1189. 50. Grolimund, D., Borkovec, M., Barmettler, K. & Sticher, H. (1996) Environ. Sci. Technol. 30, 3118–3123. 51. Kaplan, D. I., Bertsch, P. M., Adriano, D. C. & Orlandini, K. A. (1994) Radiochim. Acta 66/67, 181–187. 52. Kaplan, D. I., Bertsch, P. M. & Adriano, D. C. (1995) Ground Water 33, 708–717. 53. Seaman, J. C., Bertsch, P. M. & Miller, W. P. (1995) J. Contam. Hydrol. 20, 127–143. 54. Seaman, J. C., Bertsch, P. M., Korom, S. F. & Miller, W. P. (1996) Ground Water 34, 778–783. 55. Seaman, J. C. (1998) Soil Sci. Soc. Am. J. 62, 354–361. 56. McMahon, M. A. & Thomas, G. W. (1974) Soil Sci. Soc. Am. Proc. 38, 727–732. 57. Boggs, J. M. & Adams, E. E. (1992) Water Resourc. Res. 28, 3325–3336. 58. Seaman, J. C. & Bertsch, P. M. (1998) U.S. Patent 5,846, 434.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SURFACE GEOCHEMISTRY OF THE CLAY MINERALS

3358

Surface geochemistry of the clay minerals

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3358–3364, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. GARRISON SPOSITO*†, NEAL T. SKIPPER‡, REBECCA SUTTON*, SUNG-HO PARK*, ALAN K. SOPER§, AND JEFFERY A. GREATHOUSE¶ PNAS is available online at www.pnas.org.

ABSTRACT Clay minerals are layer type aluminosilicates that figure in terrestrial biogeochemical cycles, in the buffering capacity of the oceans, and in the containment of toxic waste materials. They are also used as lubricants in petroleum extraction and as industrial catalysts for the synthesis of many organic compounds. These applications derive fundamentally from the colloidal size and permanent structural charge of clay mineral particles, which endow them with significant surface reactivity. Unraveling the surface geochemistry of hydrated clay minerals is an abiding, if difficult, topic in earth sciences research. Recent experimental and computational studies that take advantage of new methodologies and basic insights derived from the study of concentrated ionic solutions have begun to clarify the structure of electrical double layers formed on hydrated clay mineral surfaces, particularly those in the interlayer region of swelling 2:1 layer type clay minerals. One emerging trend is that the coordination of interlayer cations with water molecules and clay mineral surface oxygens is governed largely by cation size and charge, similarly to a concentrated ionic solution, but the location of structural charge within a clay layer and the existence of hydrophobic patches on its surface provide important modulations. The larger the interlayer cation, the greater the influence of clay mineral structure and hydrophobicity on the configurations of adsorbed water molecules. This picture extends readily to hydrophobic molecules adsorbed within an interlayer region, with important implications for clay–hydrocarbon interactions and the design of catalysts for organic synthesis. The clay minerals are layer-type aluminosilicates, ubiquitous on our planet in geologic deposits, terrestrial weathering environments, and marine sediments (1, 2). Their name derives from the micrometer-sized particles into which they crystallize. This small particle size, in turn, endows these minerals with an important surface reactivity that plays a major role in the terrestrial biogeochemical cycling of metals, in the chemical homeostasis of the oceans, and in a broad variety of managed processes, including oil and gas production, industrial catalysis, pharmaceutical delivery, and radioactive waste disposal. Metal nutrients such as K+ or Ca2+ are retained in temperate-zone soils on negatively charged clay mineral surfaces but eventually can be released for consumption in the biosphere or for buffering these soils against excess acidity brought in by applied fertilizers or contaminated rainwater (3). Clay minerals precipitated from seawater in nearshore depositional environments can similarly influence the geochemical cycles of metal cations such as K+ (4), as well as the oceanic buffering of atmospheric CO2 on a global scale (5). In engineered settings, clay mineral swelling promoted by Na+ adsorption plays a significant role in petroleum extraction (6) and in the construction of environmental liners (7). Major impact on industrial organic synthesis comes in the many designed catalysts developed from clay minerals with adsorbed polymeric cations (8). Certain clay minerals are isostructural with mica (1, 2) but are not as well crystallized because of random isomorphic cation substitutions in their structure (9, 10). These cation substitutions lead to a negative net surface charge that induces an electrical double layer on clay mineral surfaces when they are exposed to aqueous electrolyte solutions (i.e., to natural waters). Water molecules can be intercalated between clay layers to create an interlayer ionic solution that causes swelling phenomena related to electrical double layer properties (3, 6, 11, 12). The structure of the double layer that forms in swelling clay mineral interlayers has been the object of much geochemical research (12), but only in recent years, aided especially by insights gained from studies of concentrated aqueous ionic solutions (13), has the powerful tandem mix of spectroscopy and molecular modeling been able to clarify matters. Two particularly effective innovations have been isotopic-difference neutron diffraction and Monte Carlo computer simulation (13). The present article is a brief account of our own recent efforts in applying these two innovations to the hydrated clay minerals, which are of widespread importance in terrestrial surface geochemistry. The emphasis here is on a molecular picture of clay hydrate interlayer structure, particularly the issues of how similar this structure is to that in concentrated ionic solutions and how different it is from the tetrahedral hydrogen-bonded network characteristic of liquid water in bulk. This kind of fundamental understanding is essential to improved modeling of global elemental cycles and better design of engineered clay materials (5, 6, 7 and 8). Crystal Structures of Clay Minerals. Clay minerals are stacked, polymeric sandwiches of tetrahedral and octahedral sheet structures (3, 9, 10, 11 and 12). They are classified first into “layer types,” differentiated by the number of tetrahedral and octahedral sheets that have combined, and then into “groups,” differentiated by the kinds of isomorphic cation substitution that have occurred (10). Layer types are sketched in Fig. 1A, and the groups are identified in Table 1. The 1:1 layer type consists of one tetrahedral sheet fused to an octahedral sheet. It is represented in Table 1 by the kaolinite group, whose generic chemical formula is [Sin1Al4−n1]Aln1 + n2−4Fe(III)n3O10(OH)8 · nH2O where cations enclosed in square brackets are located in the tetrahedral sheet, and the stoichiometric coefficients (ni, i = 1,2,3), which may be fractions, are constrained by charge neutrality and mass balance conditions given in Table 1. The octahedral sheet has two-thirds cation-site occupancy (dioctahedral sheet; full occupancy gives a trioctahedral sheet), and the structural hydration coefficient n = 0, except for 10-Å halloysite (1 Å = 0.1 nm), for which n = 4. Normally, there is no significant isomorphic substitution for Si in the tetrahedral sheet (n1 = 4) or for Al in the octahedral sheet (n2 = 4), and, therefore, no significant negative structural charge occurs in this dioctahedral clay mineral (9, 10).

*Earth Sciences Division, Mail Stop 90/1116, Ernest Orlando Lawrence Berkeley National Laboratory, University of California, Berkeley, CA 94720; ‡Department of Physics and Astronomy, University College, Gower Street, London WC1E 6BT, United Kingdom; §ISIS Facility, Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire OX11 0QX, United Kingdom; and ¶Department of Chemistry, University of the Incarnate Word, 4301 Broadway, San Antonio, TX 78209 †To whom reprint requests should be addressed at: Hilgard Hall #3110, University of California, Berkeley, CA 94720-3110. e-mail: gsposito@nature. berkeley.edu. Abbreviation: MC, Monte Carlo.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SURFACE GEOCHEMISTRY OF THE CLAY MINERALS

3359

FIG. 1. (A) Crystal structure of 1:1 and 2:1 layer type clay minerals, where X (shaded circles) is usually OH and M can be Al, Mg, Fe, etc. (B) Siloxane cavity in the basal plane of a tetrahedral sheet. The 2:1 layer type has two tetrahedral sheets fused to an octahedral sheet. Three clay mineral groups having this structure are illite, vermiculite, and smectite (9, 10). Their generic chemical formula is Cx[Sin1Al8−n1]Aln1+n2−8Fe(III)n3Fe(II)n4Mgn5Mn6O20(OH)4 where charge and mass balance constraints again appear in Table 1. The symbol Cx represents x moles of monovalent cation charge that balances negative structural charge created by isomorphic replacement of tetrahedral Si by Al; dioctahedral Al by Mg, Fe(II); or trioctahedral Mg by the unspecified structural cation M (for example, Li in the trioctahedral mineral hectorite). The layer charge x is thus the number of moles of excess electron charge per chemical formula unit that is produced by isomorphic cation substitutions (Table 1). It is balanced by cations adsorbed on or near the basal plane of a tetrahedral sheet. The three 2:1 groups are differentiated in two principal ways. As indicated in Table 1, the layer charge generally decreases in the order illite → vermiculite → smectite. The vermiculite group is further distinguished from the smectite group by a much greater degree of isomorphic substitution in the tetrahedral sheet (9, 10). Among dioctahedral smectites, those for which substitution of Al for Si in the tetrahedral sheet exceeds that of Fe(II) or Mg for Al in the octahedral sheet are called beidellite, and those for which the reverse is true are called montmorillonite (11). The emphasis in the present article will be on vermiculite and smectite clay minerals because of their importance in terrestrial weathering processes and designed industrial applications (5, 6, 7 and 8). Siloxane Surface Reactivity. The plane of oxygen ions bounding each side of a 2:1 clay mineral layer (i.e., the basal plane of a tetrahedral sheet) is called a siloxane surface (12). A reactive site associated with this surface is the hexagonal cavity formed by the bases of six corner-sharing Si tetrahedra (Fig. 1B). It has a diameter of ≈0.26 nm and is bordered by six oxygen ions, with a hydroxyl group rooted at the bottom in the octahedral sheet. In dioctahedral clay minerals, this hydroxyl group points toward the empty metal site in the octahedral sheet whereas, in trioctahedral clay minerals, it points perpendicularly to the siloxane surface (9). The reactivity of the siloxane surface depends on the nature of the local charge distribution in the clay layer (12). In the absence of nearby isomorphic cation substitutions that create negative charge, a siloxane surface functions only as a mild charge donor. If isomorphic substitution of Al by Fe(II) or Mg occurs in the octahedral sheet, the resulting excess negative charge makes it possible for the surface to form reasonably strong adsorption complexes with cations and water molecules. If isomorphic substitution of Si by Al occurs in the tetrahedral sheet, with the excess negative charge thereby localized much nearer to the periphery of the siloxane surface, much stronger adsorption complexes with cations and stronger hydrogen bonds to vicinal water molecules become possible. Charged sites also exist on the edges of clay mineral crystallites (12). Their role in clay mineral surface geochemistry ranges from critical to subordinate as the layer type shifts from 1:1 to 2:1 (9, 11, 12). Cation adsorption complexes can be classified as either inner-sphere or outer-sphere (3). An inner-sphere surface complex has no water molecule interposed between the surface functional group and the small cation or molecule it binds whereas an outer-sphere surface complex has at least one such interposed water molecule. Outer-sphere surface complexes thus comprise solvated adsorbed cations. Surface complexes involving metal cations are illustrated in Fig. 2 for a 2:1 layer type clay mineral. Ions bound in surface complexes are distinguished from those adsorbed in the diffuse portion of the electrical double layer (also illustrated in Fig. 2) because the former species remain immobilized on a siloxane surface over molecular time scales that are long when compared with, for example, the 4–10 ps required for a single diffusive step by a solvated ion in aqueous solution (13). The well known outer-sphere surface complex formed by bivalent metal cations (for example, Ca2+) in the interlayer region of montmorillonite (compare the left

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SURFACE GEOCHEMISTRY OF THE CLAY MINERALS

3360

side of Fig. 2) is immobile on the molecular time scale of ≈100 ps probed by electron spin resonance spectroscopy and by quasielastic neutron scattering (14, 15). These three types of clay mineral surface species—inner sphere complex, outer-sphere complex, and diffuse-layer— represent different modes of adsorption of aqueous cations that contribute to the formation of an electrical double layer on charged siloxane surfaces (3). Table 1. Some clay mineral groups Group Layer type Kaolinite 1:1 Illite 2:1 Vermiculite 2:1 Smectite 2:1

Layer charge, x <0.01 1.4−2.0 1.2−1.8 0.5−1.2

Structural charge balance* x = 4 − n1† x = 8 − n1 + n4 + n5 + n6‡ x = 8 − n1 + n4 + n5 + n6§ x = 8 − n1 + n4 + n5 + n6‡

Structural mass balance* ni = 8† ni = 12‡ ni = 12‡ ni = 12‡

*Based on an O10(OH)8 (1:1) or O20(OH)4 (2:1) unit cell formula. †See text for definitions of ni (i = 1, . . . , 6). ‡The formulas given are dioctahedral clay minerals with bivalent M in the octahedral sheet. For trioctahedral smectite and vermiculite with monovalent ni = 14. M in the octahedral sheet, x = 2(8 – n1) – n2 – n3 + n6 and

Interlayer Surface Structure by Neutron Diffraction. The surface species “cartooned” in Fig. 2 emerged conceptually from the results of in situ spectroscopic studies (14, 15). The consensus of these studies is that hydrated smectite and vermiculite interlayers are in many ways similar in structure to a concentrated ionic solution (14). Aqueous solution molecular structure, in turn, has been explored especially well over the same time-period by isotopic-difference neutron diffraction (13, 16). In this methodology, isotopic substitution of one or more diffracting atoms is performed (for example, substitution of hydrogen by deuterium in interlayer water) to create differences in coherent neutron scattering cross section that facilitate locating the atom accurately in relation to its diffracting neighbors (13, 16). Isotopic-difference neutron diffraction methods are reviewed concisely by Skipper et al. (17) for hydrated 2:1 clay minerals. Neutron diffraction studies of the interlayers in these systems have already provided valuable insight as to molecular structure in the electrical double layer formed on siloxane surfaces (17, 18 and 19).

FIG. 2. Cartoon of the three types of small cation adsorption by a 2:1 layer type clay mineral. The “Stern Layer” comprises only surface complexes, which can form in the interlayer region (left) as well as on single siloxane surfaces (right). Characteristic residence time scales of the three adsorbed species are compared at upper right to the time scales of in situ spectroscopic methods used to detect them. Fig. 3 is a visualization of the arrangement of water molecules around Ca2+ adsorbed in the interlayer region of the two-layer hydrate of trioctahedral vermiculite, based on the results of H/D isotopic-difference neutron diffraction experiments (18). Because the water protons could be identified separately from those in the clay mineral structure and were distinguishable from other interlayer atoms, it was possible to show rather clearly that the Ca2+ are octahedrally coordinated to their nearest-neighbor water molecules, as also occurs in concentrated, as opposed to dilute, aqueous solutions of CaCl2 (13, 18). Moreover, four of the solvating water molecules form hydrogen bonds with the siloxane surfaces, as expected for clay minerals with isomorphic substitutions only in the tetrahedral sheets (12). Because the two-layer hydrate of vermiculite has about eight water molecules per Ca2+ in the interlayer region

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SURFACE GEOCHEMISTRY OF THE CLAY MINERALS

3361

(18), two water molecules per Ca2+ must be nonsolvating. Skipper et al. (18) were able to find these water molecules tucked into the cavities of the siloxane surface. Other neutron diffraction studies of trioctahedral vermiculite hydrates have also reported this phenomenon (17, 18 and 19).

FIG. 3. Visualization of Ca2+ (large black sphere) in an octahedral solvation complex (water molecules, with smaller black spheres as O and red spheres as H) in the interlayer region of Ca-vermiculite. Portions of the opposing clay mineral layers are shown, with structural protons also indicated in red. Compare to the left side of the cartoon in Fig. 2. These structural inferences are in accord with—and improve on—the results of earlier neutron-scattering, electron spin resonance [facilitated by Cu(II) doping], and x-ray diffraction studies of the two-layer hydrate of Ca-vermiculite (20, 21 and 22), the last of which places Ca2+ precisely between a triad of surface O on one siloxane surface and a hexagonal cavity in the other. Neutron-scattering and electron spin resonance studies also confirm the immobilization of the interlayer Ca(H2O)62+ solvation complex on a 100-ps timescale, which is an order of magnitude longer than a diffusive time step for Ca2+ in aqueous solution (13). Interlayer Surface Structure by Computer Simulation. Monte Carlo (MC) computer simulations are well known as essential components of research on aqueous ionic solutions (13). The underlying paradigm in these simulations is to construct intermolecular potential functions that represent parametrically all of the known interactions in a system then devise a strategy for sampling the phase space of the interacting system to compute its chemical properties. In a typical MC simulation, the configuration space of the system is sampled randomly under the guidance of an algorithm based in equilibrium statistical mechanics (23). Convergence of the simulation is monitored by examining the stability of calculated system properties (for example, the layer spacing in the case of hydrated clay minerals) as sampling proceeds. The synergistic relationship between experiment and molecular modeling indigenous to the study of aqueous ionic solutions has not been possible in the study of hydrated clay minerals until very recently, directly after the appearance of fourth-generation supercomputers and the emergence of convenient parametric models for water–smectite and cation– smectite potential functions based on quantum mechanical insight (24, 25 and 26). These developments encouraged the undertaking of systematic simulation studies of low-charge smectite hydrates, particularly the ubiquitous Wyoming montmorillonites, which have both octahedral and tetrahedral charge sites (6, 27, 28, 29, 30, 31, 32, 33 and 34). Prototypical Wyoming montmorillonite corresponds to n1 = 7.75, n2 = 3.75, n3 = n4 = 0, n5 = 0.5, and n6 = 0 in the chemical formula given above for 2:1 layer type clay minerals, with x = 0.75 as the layer charge. The results of these simulations have generated a number of fundamental questions that require additional molecular-scale experimentation, thus sustaining the theory–experiment dialogue that defines fundamental geochemical research. The computer simulations summarized in the present article were performed by using the code MONTE (35), developed by N. T. Skipper and K. Refson, with phase-space sampling strategies as described by Skipper et al. (27, 28) and Chang et al. (25, 26). Fig. 4 and Fig. 5 are “snapshots” of equilibrium interlayer configurations based on MC simulations (34) of the two-layer hydrates of Naand K-Wyoming montmorillonite. The species shown in Fig. 4 is Na+ bound in an outer-sphere surface complex to an octahedral charge site of the clay mineral. This visualization, like Fig. 3, confirms the spectroscopy-inspired cartoon in Fig. 2. Fig. 4 includes only water molecules in a solvation shell confined to within 3.2 Å of the central Na+, the individual Na-H2O separations varying from 2.2 to 2.5 Å. These vicinal water molecules form a distorted octahedron, in agreement with Na–H2O distances and coordination numbers determined for concentrated NaCl solutions with the same H2O/Na ratio ≈ 10 (13). Fig. 5 illustrates K+ bound in an inner-sphere surface complex to an octahedral charge site on Wyoming montmorillonite. Portions of both siloxane surfaces are shown, and, in this case, water molecules both within and outside the primary solvation shell of K+ (K–O separations varying from 2.8 to 3.7 Å) are depicted. The coordination number of K+ with O is eight, but two are contributed by oxygen ions in the siloxane surface. Thus, K+ is in distorted cubic coordination with its neighboring O, consistent with geometric concepts based on the K+/O radius ratio (3). Strongly solvating monovalent cations like Na+ or Li+ have a tendency to form only inner-sphere surface complexes with tetrahedral charge sites and only outer-sphere surface complexes with octahedral charge sites on smectites (28, 33). This trend can be related to partial desolvation facilitated by the smaller distance of closest approach between an interlayer cation and a charge site that exists for a tetrahedral as opposed to an octahedral site, which necessarily lies deeper in the clay layer. For large cations like K+, however, the desolvation process is always facile because of a weak interaction with water molecules (13, 31), and inner-sphere surface complexation is not so dependent on the location of the charge site in the clay layer. Fig. 6 is a snapshot, based on MC simulation (33), of the interlayer configuration in Li-hectorite, a trioctahedral smectite having only octahedral charge sites. The view in the figure is along the clay layer c axis, with only one of the two opposing siloxane surfaces shown but with all cations and water molecules in a MC simulation cell depicted. This stable hectorite hydrate, which has been investigated extensively by a variety of spectroscopic techniques, has an average of three water molecules per Li+, which corresponds to a very concentrated ionic solution (≈18.5 molal). Fig. 6 conforms to the trend expected for small interlayer cations, in that only outer-sphere surface

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SURFACE GEOCHEMISTRY OF THE CLAY MINERALS

3362

FIG. 4. Visualization of Na+ bound in an outer-sphere surface complex in the interlayer region of Wyoming montmorillonite, based on MC simulation. A portion of the siloxane surface structure also is shown.

FIG. 5. Visualization of K+ bound in an inner-sphere surface complex in the interlayer region of Wyoming montmorillonite, based on MC simulation. Green lines extend from K+ (black sphere) to nearest-neighbor O in the surface complex. Dashed lines indicate hydrogen bonds between water molecules. Portions of the opposing two siloxane surfaces also are shown, with the beige sphere at the bottom of the figure (center) indicating a site of Al3+ substitution for Si4+.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SURFACE GEOCHEMISTRY OF THE CLAY MINERALS

3363

FIG. 6. Visualization of the interlayer configuration in Li(H2O)3− hectorite, based on MC simulation (33). The Li+ are bound in outersphere surface complexes with two water molecules. Other water molecules are keyed into the siloxane surface cavities.

FIG. 7. Visualization of a methane molecule adsorbed in the interlayer region of the three-layer hydrate of Na-montmorillonite, based on MC simulation. The typical 20-fold coordination between CH4 and O occurs, but with nearly half of the O being in the siloxane surface. complexes have formed, comprising just two solvating water molecules, in agreement with the average hydration number of 2.3 found in isotopic-difference neutron diffraction studies of very concentrated LiCl solutions (16). The nonsolvating water molecules have keyed themselves into the hexagonal cavities of the siloxane surface (one such molecule is at the center of the simulation cell), reminiscent of the situation in trioctahedral vermiculite (17, 18 and 19). Much experimental and theoretical information points to an inherent hydrophobicity of the siloxane surface [see, for example, the summary by Jaynes and Boyd (36)] were it not for the presence of layer charge. Pyrophyllite, the uncharged analog of montmorillonite, and talc, the uncharged analog of vermiculite, both have hydrophobic siloxane surfaces (36, 37). Studies of the effect of layer charge on the adsorption of both water and hydrocarbon molecules by smectites (36, 38) indeed show that surface hydrophobicity increases as the layer charge decreases. Recent molecular dynamics simulations of ion solvation and mobility in aqueous solution (39, 40) suggest that large cations like K+ tend to interact with water molecules not only through their positive charge but also through solvent cage formation, which is just what hydrocarbon molecules do (41, 42 and 43). This hydrophobic tendency may be the basis for K+ associating directly with clay mineral O (Fig. 5) instead of forming a well organized solvation complex near octahedral charge sites, as does Na+ in Fig. 4. Fig. 7 exposes the inherent hydrophobicity of the siloxane surface even more directly through a MC snapshot of a methane molecule adsorbed in the interlayer region of the three-layer hydrate of Na-montmoril

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SURFACE GEOCHEMISTRY OF THE CLAY MINERALS

3364

lonite (H2O/CH4 = 23). The well known 20-fold coordination cage induced by CH4 in bulk water (41, 42) was reproduced successfully by the CH4-O potential function used in the MC simulation (43). In the interlayer of a Na-montmorillonite hydrate, however, methane coordinates to eight clay mineral O and approximately a dozen water molecule O to form this cage in a highly distorted coordination structure. This kind of hydrophobic association, which may be favored for CH4 over a purely solvent-based arrangement of neighboring O atoms, could play an important role in the chemical evolution of organic molecules as mediated by clay minerals (44). The multifaceted nature of interactions within clay mineral interlayers leads necessarily to complexity in the structure of adsorbed water. This complexity is well illustrated by a consideration of water molecule orientations in the two-layer hydrate of Na-montmorillonite as revealed by MC simulation (29). Sodium–water molecule interactions in this system produce a local coordination structure like that in concentrated aqueous solutions of NaCl (13, 29), but Na+ interactions with tetrahedral charge sites are still strong enough to allow innersphere surface complex formation with oxygen ions in the siloxane surface. The configuration of water molecules differs between inner-sphere and outersphere surface complexes. When these two species are forced to cohabit within the constrained spatial domain that exists in an interlayer region, disorder in the water network is likely, with distorted H-bonds and an array of water dipole orientations taking on almost every possible direction (29). This disorder is enhanced by an evident attraction between water molecules and the cavities in the siloxane surface, which gives rise to nonsolvating water molecules keyed into these holes irrespective of the type of adsorbed cation (6, 18, 19, 29, 30 and 31, 33). The characteristics of adsorbed water on 2:1 clay minerals also reveal the competition between interlayer cations and clay mineral structure for intercalated water molecules, as well as that between hydrophilic and hydrophobic interactions. This competition produces a complex electrical double layer structure whose origins and behavior are beginning to be understood at a fundamental level as the basis for progress toward improved design in applications (7, 8) that provide palpable benefits for humankind. Without the unfailing support of Dr. John Maccini (U.S. National Science Foundation) and Dr. Sally Benson (Lawrence Berkeley National Laboratory), the simulation research described herein would not have been possible. The authors thank the National Energy Research Scientific Computing Center for allocations of time on its Cray supercomputers and Angela Zabel for excellent preparation of the typescript. The research reported in this paper was supported in part by National Science Foundation Grant EAR-9505629 and in part by the Director, Office of Energy Research, Office of Basic Energy Sciences, Geosciences Division of the U.S. Department of Energy under Contract DE-AC03-76SF00098. 1. Pauling, L. (1930) Proc. Natl. Acad. Sci. USA 16, 123–129, 578–582. 2. Hendricks, S. B. & Fry, W. H. (1930) Soil Sci. 29, 457–479. 3. Sposito, G. (1989) The Chemistry of Soils (Oxford Univ. Press, New York). 4. Michalopoulos, P. & Aller, R. C. (1995) Science 270, 614–617. 5. McCauley, S. E. & DePaolo, D. J. (1997) in Tectonic Uplift and Climate Change, ed. Ruddiman, W. F. (Plenum, New York), pp. 427–467. 6. Karaborni, S., Smit, B., Heidug, W., Urai, J. & van Oort, E. (1996) Science 271, 1102–1104. 7. Kajita, L. S. (1997) Clays Clay Miner. 45, 609–617. 8. Zielke, R. C., Pinnavaia, T. J. & Mortland, M. M. (1989) in Reactions and Movement of Organic Chemicals in Soils, ed. Sawhney, B. L. & Brown, K. (Soil Science Society of America, Madison), pp. 81–97. 9. Bailey, S. W., ed. (1988) Hydrous Phyllosilicates (Mineralogical Society of America, Washington, D.C.). 10. Moore, D. M. & Reynolds, R. C. (1997) X-Ray Diffraction and Identification and Analysis of Clay Minerals (Oxford Univ. Press, New York). 11. Borchardt, G. (1989) in Minerals in Soil Environments, ed. Dixon, J. B. & Weed, S. B. (Soil Science Society of America, Madison, WI), pp. 675–727. 12. Sposito, G. (1984) The Surface Chemistry of Soils (Oxford Univ. Press, New York). 13. Ohtaki, H. & Radnai, T. (1993) Chem. Rev. (Washington, D.C.) 93, 1157–1204. 14. Sposito, G. & Prost, R. (1982) Chem. Rev. (Washington, D.C.) 82, 553–573. 15. Sposito, G. (1993) Ion Exch. Solv. Extr. 11, 211–236. 16. Neilson, G. & Enderby, J. E. (1989) Adv. Inorg. Chem. 34, 195–217. 17. Skipper, N. T., Soper, A. K. & McConnell, J. D. C. (1991) J. Chem. Phys. 94, 5751–5760. 18. Skipper, N. T., Soper, A. K. & Smalley, M. V. (1994) J. Phys. Chem. 98, 942–945. 19. Skipper, N. T., Smalley, M. V., Williams, G. D., Soper, A. K. & Thompson, C. H. (1995) J. Phys. Chem. 99, 14201–14204. 20. Tuck, J. J., Hall, P. L., Hayes, M. H. B., Ross, D. K. & Hayter, J. B. (1985) J. Chem. Soc. Faraday Trans. 1 81, 833–846. 21. McBride, M. B. (1986) in Geochemical Processes at Mineral Surfaces, ed. Davis, J. A. & Hayes, K. F. (Am. Chem. Soc., Washington, D.C.), pp. 362– 388. 22. Slade, P. G., Stone, P. A. & Radoslovich, E. W. (1985) Clays Clay Miner. 33, 51–61. 23. Allen, M. P. & Tildesley, D. J. (1987) Computer Simulation of Liquids (Clarendon, Oxford). 24. Skipper, N. T., Refson, K. & McConnell, J. D. C. (1989) Clay Miner. 24, 411–425. 25. Skipper, N. T., Refson, K. & McConnell, J. D. C. (1991) J. Chem. Phys. 94, 7434–7445. 26. Skipper, N. T., Refson, K. & McConnell, J. D. C. (1993) in Geochemistry of Clay-Pore Fluid Interactions, eds. Manning, D. C., Hall, P. L. & Hughs, C. R. (Chapman & Hall, London), pp. 40–61. 27. Skipper, N. T., Chang, F.-R. C. & Sposito, G. (1995) Clays Clay Miner. 43, 285–293. 28. Skipper, N. T., Sposito, G. & Chang, F.-R. C. (1995) Clays Clay Miner. 43, 294–303. 29. Chang, F.-R. C., Skipper, N. T. & Sposito, G. (1995) Langmuir 11, 2734–2741. 30. Chang, F.-R. C., Skipper, N. T. & Sposito, G. (1997) Langmuir 13, 2074–2082. 31. Chang, F.-R. C., Skipper, N. T. & Sposito, G. (1998) Langmuir 14, 1201–1207. 32. Boek, E. S., Coveney, P. V. & Skipper, N. T. (1995) J. Am. Chem. Soc. 117, 12608–12617. 33. Greathouse, J. & Sposito, G. (1998) J. Phys. Chem. B 102, 2406–2414. 34. Sposito, G., Park, S.-H. & Sutton, R. (1999) Clays Clay Miner, in press. 35. Skipper, N. T. (1996) MONTE User's Manual (Department of Physics and Astronomy, University College, London). 36. Jaynes, W. F. & Boyd, S. A. (1991) Clays Clay Miner. 39, 428–436. 37. Bridgeman, C. H., Buckingham, A. D., Skipper, N. T. & Payne, M. C. (1996) Mol. Phys. 89, 879–888. 38. Sposito, G., Prost, R. & Gaultier, J.-P. (1983) Clays Clay Miner. 31, 9–16. 39. Lynden-Bell, R. M. & Rasaiah, J. C. (1997) J. Chem. Phys. 107, 1981–1991. 40. Koneshan, S., Rasaiah, J. C., Lynden-Bell, R. M. & Lee, S. H. (1998) J. Phys. Chem. B 102, 4193–4204. 41. Swaminathan, S., Harrison, S. W. & Beveridge, D. L. (1978) J. Am. Chem. Soc. 100, 5705–5712. 42. Swaminathan, S., Harrison, S. W. & Beveridge, D. L. (1979) J. Am. Chem. Soc. 101, 5832–5833. 43. Skipper, N. T., Bridgeman, C. H., Buckingham, A. D. & Mancera, R. L. (1996) Faraday Discuss. Chem. Soc. 103, 141–150. 44. Cairns-Smith, A. G. (1982) Genetic Takeover and the Mineral Origins of Life (Cambridge Univ. Press, New York).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CONTAMINANT BIOAVAILABILITY IN SOILS, SEDIMENTS, AND AQUATIC ENVIRONMENTS

3365

Contaminant bioavailability in soils, sediments, and aquatic environments

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3365–3371, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. SAMUEL J. TRAINA* AND VALÉRIE LAPERCHE PNAS is available online at www.pnas.org.

School of Natural Resources, The Ohio State University, 2021 Coffey Road, Columbus, OH 43210 ABSTRACT The aqueous concentrations of heavy metals in soils, sediments, and aquatic environments frequently are controlled by the dissolution and precipitation of discrete mineral phases. Contaminant uptake by organisms as well as contaminant transport in natural systems typically occurs through the solution phase. Thus, the thermodynamic solubility of contaminant-containing minerals in these environments can directly influence the chemical reactivity, transport, and ecotoxicity of their constituent ions. In many cases, Pbcontaminated soils and sediments contain the minerals anglesite (PbSO4), cerussite (PbCO3), and various lead oxides (e.g., litharge, PbO) as well as Pb2+ adsorbed to Fe and Mn (hydr)oxides. Whereas adsorbed Pb can be comparatively inert, the lead oxides, sulfates, and carbonates are all highly soluble in acidic to circumneutral environments, and soil Pb in these forms can pose a significant environmental risk. In contrast, the lead phosphates [e.g., pyromorphite, Pb5(PO4)3Cl] are much less soluble and geochemically stable over a wide pH range. Application of soluble or solid-phase phosphates (i.e., apatites) to contaminated soils and sediments induces the dissolution of the “native” Pb minerals, the desorption of Pb adsorbed by hydrous metal oxides, and the subsequent formation of pyromorphites in situ. This process results in decreases in the chemical lability and bioavailability of the Pb without its removal from the contaminated media. This and analogous approaches may be useful strategies for remediating contaminated soils and sediments. The Earth's surface is dominated by the elements O, H, Si, Al, Fe, Ca, Na, K, Mg, Ti, and P. As oxides, these elements account for ≈96% of the total mass of the continental crust (1). Many of the remaining elements in the periodic table (together with C) are essential for life. Many trace elements are toxic to a wide range of organisms when concentrated and some are toxic to most even at very low concentration. These latter contaminants are natural substances (with the exception of the transuranics) and life on Earth evolved in their presence. Human activity has altered the distribution and forms of these elements, locally increasing their relative toxicities and the frequency with which they are encountered by living organisms. At present, various elements are listed as priority pollutants by the U.S. Environmental Protection Agency. Their concentrations in soils, as well as in surface and ground water, typically are regulated based on total concentration. This system provides a convenient regulatory framework for establishing acceptable levels of contaminant metals and oxoanions in environmental media. However, the environmental science community recognizes that total concentration is not an accurate predictor of the bioavailability or chemical lability of a given contaminant in soils, sediments, or aquatic environments. Rather, the toxicity of a substance, be it an element, an ion, or a molecule typically is controlled by its chemical and physical state, or speciation (2, 3, 4, 5, 6, 7, 8, 9 and 10). Thus, regulations based on absolute concentration may be convenient, but their scientific validity may be in question. Knowledge of the link between chemical speciation and bioavailability is not new, nor did it originate with concerns of contaminant toxicity or environmental pollution. The fields of soil chemistry and soil fertility were, in part, created by the recognition that total soil concentration is a poor predictor of the bioavailability of essential nutrients required for plant growth and food production. This recognition has led to extensive research on the identity and form of nutrient elements in soils and fertilizers with emphasis on predicting their bioavailability. An example is the attention paid to the chemistry and mineralogy of P in soils and fertilizers (11, 12 and 13). Phosphorous is often a limiting nutrient, and supplementation of soil P through the addition of P-containing amendments has been practiced at least as early as 287 BC (13). The apatite mineral family is the most ubiquitous form of P in the Earth's crust, as well as the most geochemically stable one in neutral to alkaline environments (14). Although used extensively as fertilizer, its intrinsically low solubilities makes apatite a very poor choice. Instead, more soluble forms of P commonly are used as amendments to P-deficient agricultural soils. Thus, when serving as a nutrient, the total concentration of P in the agricultural amendment is not as important as the form or availability of the P in the amendment material. In addition to being an essential nutrient, P is an important environmental contaminant. Excessive influx of P into fresh water can lead to increases in primary productivity (photosynthesis) and accelerated sedimentation (15). This process, cultural eutrophication, can result in the growth of deleterious species of algae, depletions of available O2, and production of toxic metabolic products by a number of phytoplankton. Numerous studies of fresh-water systems have shown the inseparable link between the chemical form or species of P entering lakes and the potential for P-induced eutrophication (16). Indeed, bioassay measurements indicated that P present as apatites is much less bioavailable to planktonic algae than dissolved phosphate (16). Presumably, the lower bioavailability of apatites is a result of their low solubility in neutral and alkaline environments (as also in similar agricultural environments). That apatite as well as other solid-phase forms of P are less bioavailable than dissolved P strongly influenced the institutional controls implemented for the protection of the Great Lakes of North America. Initial efforts were focused on the removal of dissolved P from wastewater discharges followed by a later effort to reduce sediment loads from agriculture. The latter is dominated by particulate P (16).

*To whom reprint requests should be addressed. e-mail: [email protected]. Abbreviations: XRD, x-ray diffraction; SEM, scanning electron microscopy; AFM, atomic force microscopy.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CONTAMINANT BIOAVAILABILITY IN SOILS, SEDIMENTS, AND AQUATIC ENVIRONMENTS

3366

These examples illustrate two important points relevant to contaminant chemistry and ecotoxicology. First, a substance beneficial in one environment (e.g., bioavailable P in an agricultural soil) may be deleterious in another (e.g., bioavailable P in lake waters). Second, the specific form of a substance has a profound impact on its bioavailability with solid-phase forms (including sorbed species) generally controlling its bioavailability in natural environments. It is clear that dissolved substances are generally more labile and bioavailable than solids, but can one generalize about the relative bioavailabilities of two or more different solids, each containing the same element of interest? As discussed below, this question may be assessed by considering the solubility products of the solids. The relationship between solubility and bioavailability of contaminants has important ramifications for environmental risk assessment and environmental remediation. Adoption of an environmental impact or bioavailability mode of contaminant regulation can on be accomplished if a formal assessment indicates that some parameter other than total concentration (e.g., solubility) controls the bioavailability and ecotoxicology of that substance. Concomitantly, recognition that bioavailability can be tied to solubility rather than total concentration allows one to consider remediation strategies based on in situ reductions solubility of a contaminant, rather than its complete removal. Solubility and the Solubility Product The assertion that the bioavailability of a given element in soils or sediments is controlled by its solid-phase form rests on the assumption that uptake of a contaminant by a target organism happens through the solution phase. This is a safe assumption to make when one considers the uptake of an ion by plants. It may be less appropriate for the uptake of contaminants by fauna where ingestion or inhalation of particulate material can represent a significant mechanism of contaminant exposure. Nevertheless, with the possible exception of radiological damage, toxic responses to a contaminant generally require absorption by biological tissue. Even for ingestion, the relative bioavailabilities and toxicities of different mineral forms of a given element are subject to their relative solubility (17, 18). This phenomenon may in part be caused by the linkage between intrinsic solubility and relative dissolution rates, as discussed below. In any event, the equilibrium solubility of a given mineral and its dissolution kinetics profoundly affect the bioavailability and chemical lability of its constituent ions. For a given solid, MxLy, a general dissolution reaction is:

[1]

where My+(aq) and Lx−(aq) are the aqueous metal and ligand ions M and L, respectively. An equilibrium constant for this reaction is defined as:

[2]

where [ ] denote activities. A solubility product is then defined as:

[3]

If the solid MxLy is in its standard state then [MxLy] becomes unity and Ksp becomes: [4] Ksp = [My+] x [Lx−]y. For a fixed activity of Lx−, the solid with the smallest numerical value of Ksp will support the smallest equilibrium activity of My+. Commonly, the toxicity of metal M is directly proportional to the activity of the free metal ion, My+, regardless of whether My+ or some hydrolytic species, or complex ion-pair is the most toxic form of M. This relationship results from the direct relationship between the activity of My+ and all other species of dissolved M (including complex ion-pairs and hydrolytic species). Therefore, the least toxic solid form of M will have the smallest aqueous equilibrium activity of My+. Analogously, the most toxic solid will be that which supports the largest aqueous equilibrium activity of My+. This so-called solubility product model of contaminant availability has several limitations. First, organisms represent intrinsically dynamic systems, and their reaction with the surrounding environment is typically far from equilibrium. Second, adsorption reactions and mass transfer constraints may lower the aqueous activity of M below that supported by any known phase of M. The formation of solid solutions also may lower the aqueous activity of M below that supported by any known pure phase of M. Finally, the solubility product of MxLy is expressed in terms of its dissolution into its constitutive ions. It does not account for additional side reactions that could lead to consumption of L. Such side reactions could alter the apparent relative solubilities of different M-containing solids. With the exception of the first proviso (local equilibrium) the other limitations of the solubility product model can be addressed by substitution of the appropriate equilibrium constants to account for desorption from particulate surfaces, dissolution of solid solutions, and/or incongruent dissolution reactions. What then, can one do about the assumption of local equilibrium? The law of detailed balancing (19) indicates that to a first approximation the relative dissolution rates of a series of solids with identical specific surface areas, each containing metal M, will be inversely proportional to their solubility products, (when corrected for side reactions with the surrounding solution). Obviously, in natural systems solids of different chemical composition rarely have identical specific surface areas. Additionally, some solids may undergo surface controlled dissolution under the same conditions that promote diffusion controlled dissolution reactions for other solids containing M. Nevertheless, wide differences in mineral solubilities do produce differential dissolution rates among multiple solids, each containing the metal M. This assertion is true for virtually all aqueous solutions, including surface and ground waters and soil solutions, as well as gastrointestinal tracts. The net outcome is that the solids with greater solubilities generally can be characterized as having greater dissolution rates, resulting in greater bioavailabilities and chemical labilities of their constituent elements. Recognition that the solubility product serves as a relative constraint on the reactivity and potential toxicity of metals and oxoanions in surficial environments provides a strategy for environmental treatment. Often, it is not possible to remove toxic elements from contaminated soils and sediments. In these cases, inducing changes in the mineralogy of a contaminant (e.g., conversion of a metal-carbonate to a metalphosphate) may allow one to significantly lower its solubility and its corresponding ecotoxicity. The remainder of this paper will examine the feasibility of this approach with emphasis on the formation of stable Pb precipitates in contaminated soils. Pb-Phosphates The orthophosphate ion forms sparingly soluble solids with several toxic metals, including Cd, Zn, Pb, and several of the actinides. Much research has explored the utility of using phosphates to reduce the mobility and bioavailability of these metals in contaminated environments as well as in a number of waste forms. A full discussion of all of these metals and all of these applications is beyond the scope of this paper. For

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CONTAMINANT BIOAVAILABILITY IN SOILS, SEDIMENTS, AND AQUATIC ENVIRONMENTS

3367

brevity we focus on the formation and bioavailability of Pb-phosphates. The Pb-phosphates are some of the most insoluble Pb(II)-solids known to form under surficial geochemical conditions (Table 1). At standard state, the Pb-phosphates are at least 44 orders of magnitude less soluble than galena (PbS), anglesite (PbSO4), cerussite (PbCO3), litharge (PbO), and crocoite (PbCrO4), Pb-solids common to soils contaminated by mining and smelting activities and by paint (20, 21). Nriagu (22, 23 and 24), Santillan-Medrano and Jurinak (25), and Sauvé et al. (26) suggest that Pb-phosphate phases may control the solubility of Pb in noncalcareous soils. Indeed, in oxidized, noncalcareous environments, Pb-phosphates should form at the expense of other Pb solids if sufficient P is present. Natural Pb-phosphate minerals have been identified in soils impacted by the weathering of Pb ores (17, 20, 27, 28) as well as in roadside soils presumably contaminated by automobile emissions, and in urban soils (29). Often these natural Pb-phosphates are present as PbCa solid solutions (17, 27, 28 and 29) as predicted by Nriagu (24). However, essentially pure Pb-phosphates also have been identified in contaminated soils (20). In light of their intrinsically low solubilities and their natural occurrence in some contaminated soils, effort has been given to inducing the formation of Pb-phosphates in Pb-contaminated soils and soil materials through the addition of P. The treatment of Pb-contaminated soils with additions of highly soluble forms of orthophosphate as (Na2HPO4 or KH2PO4) can reduce the bioavailability of Pb as assessed by an in vitro gastrointestinal assay (30, 31) as well as induce the formation of Pb-phosphate particles (32). Unfortunately, treatment with highly soluble P increases the risk of offsite P migration (29, 33) and eutrophication of surrounding surface waters. An alternate approach is to use a lower solubility source of P such as apatite. These Ca-phosphates are prevalent as accessory minerals in igneous rocks and as low-temperature precipitates in soils and sedimentary environments. Apatite Chemistry The hexagonal (P63/m) crystals of the apatite group [Ca5(PO4)3X] comprise three dominant end-members, where X = OH in hydroxylapatite, F in fluorapatite, and Cl in chlorapatite. The solubility product constants of the most common hydroxyl and fluoro end members are 10−3.1 and 10−25, respectively. Apatites of geologic origin are dominated by the fluorapatites, exhibiting inhomogeneous solid solution with Cl, OH, and CO3 (11, 34, 35). The Ca in apatites resides in two distinct crystal sites (Fig. 1). The Ca(1) site is coordinated to nine oxygens. The Ca(2) site is comprised of CaO5X octahedron. In fluor- and hydroxylapatites an additional weak bond to O exists (0.15 valence units), resulting in CaO5X(O) polyhedra (34). Natural apatites exhibit extensive substitution with the incorporation of K, Na, Mn, Ni, Cu, Co, Zn, Sr, Ba, Cd, Sn, Y, and rare earth elements in the Ca sites (refs. 11, 34, and 36, and references therein). Substitution occurs at both Ca sites, depending on the ionic radius of the ion in question. Table 1. Solubility products of selected Pb minerals Mineral Litharge Anglesite Cerussite Pyromorphite Hydroxypyromorphite Fluoropyromorphite Bromopyromorphite Corkite Hindsalite Plumbogummite

Formula PbO PbSO4 PbCO3 Pb5(PO4)3Cl Pb5(PO4)3OH Pb5(PO4)3F Pb5(PO4)3Br PbFe3(PO4)(SO4)(OH)6 PbAl3(PO4)(SO4)(OH)6 PbAl3(PO4)2(OH)5·H2O

Log Ksp* 12.9 −7.7 −12.8 −84.4 −76.8 −71.6 −78.1 −112.6 −99.1 −99.3

*Data from ref. 20 and references cited therein.

FIG. 1. Projection of the hydroxylapatite structure down the c axis, as well as the two cation sites in hydroxylapatite. Structural data from ref. 52. Atom sizes are not to scale. Reaction of Dissolved Pb with Apatites Apatites have been examined extensively for use in the removal of toxic metal ions from wastewater and aquatic solutions (33, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50 and 51). Mechanisms of metal uptake vary with identity of the sorbate, the sorbent, and the solution conditions. From changes in solution composition, powder x-ray diffraction (XRD) patterns, and scanning electron micrographs, Ma et al. (33) described the reaction of dissolved Pb2+ with hydroxylapatite by the sequential dissolution and precipitation reactions: Ca5(PO4)3OH + 7H+⇔5Ca2+ + 3H2PO4− + H2O [5] 5Pb2+ + 3H2PO4− + H2O⇔Pb5(PO4)3OH + 7H+, [6] where Pb5(PO4)3OH is the mineral hydroxypyromorphite. The overall reaction: Ca5(PO4)OH + 5Pb2+⇔Pb5(PO4)3OH + 5Ca2+ [7] is exergonic with a standard state Gibbs energy change of −137.08 kJ·mol−1 (51). The pyromorphite group also includes the mineral pyromorphite [Pb5(PO4)3Cl], fluoropyromorphite [Pb5(PO4)3F], bromopyromorphite [Pb5(PO4)3Br], and various arsenate and vanadate analogs. The solubility of the pyromorphites decreases fluoropyromorphite > hydroxypyromorphite > bromopyromorphite > pyromorphite, with log Ksp of −72, −77, −78, and −84, respectively. The pyromorphites are isostructural with the apatites. Discrete products of Pb reactions with hydroxylapatite are detectable by powder XRD and by scanning electron microscopy (SEM) in model aqueous systems when the initial aqueous Pb concentrations are >5 mg·liter −1 and the initial pH ranges from 3 to 7 (33). Under these conditions, discrete Pb-bearing solids (identified by XRD as hydroxypyromorphite) form in less than 10 min. They have different crystal

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CONTAMINANT BIOAVAILABILITY IN SOILS, SEDIMENTS, AND AQUATIC ENVIRONMENTS

3368

habits than the original hydroxlyapatites and do not contain Ca (within the detection limits of energy dispersive x-ray analysis). Apparently they form according to the sequential reactions described in Eqs. 5 and 6 and not from ion substitution of Pb for the Ca in the apatite particles (33). Fourier transform IR spectroscopy and x-ray absorption fine structure spectroscopy indicate that the reaction of <1 mg Pb·liter−1 with hydroxylapatite still results in the formation of hydroxypyromorphite (52). Ex situ and in situ atomic force microscopy (AFM) studies of Pb-reacted hydroxylapatite indicate that when initial conditions are far from equilibrium (>1 mg Pb·liter−1, pH = 6), pyromorphite can nucleate homogeneously as a result of interactions between dissolved Pb and phosphate (50, 51). In situ, AFM measurements of Pb solutions reacting with hydroxylapatite particles, showed only “clean” apatite surfaces without coatings of pyromorphite crystals (51). Nevertheless, pyromorphite crystals were found in the outflow from the AFM liquid cell, suggesting that homogeneous nucleation had occurred. Additionally, the presence of pyromorphite needles atop the AFM cantilever (Fig. 2) was consistent with precipitation of the Pb-phosphates in solution and not on the surfaces of the hydroxylapatite (51). The initial composition of the solution phase influences the interactions of dissolved Pb with apatites. Ma et al. (42) examined the effects of NO3, Cl, F, SO4, and CO3 on the immobilization of aqueous Pb by hydroxylapatite. Pb concentrations were reduced from an initial 5–100 mg·liter−1 to <15 µg·liter−1 (the Environmental Protection Agency's drinking water limit for Pb) except at very high concentrations of CO3. Hydroxylapatite was transformed to hydroxypyromorphite in the presence of NO3, SO4, and CO3, to pyromorphite after reaction with PbCl2, and to fluoropyromorphite after reaction with PbF2. These reaction products were identified by XRD and SEM.

FIG. 2. Pyromorphite crystals formed from the reaction of dissolved Pb with hydroxylapatite. (A) An ex situ tapping mode AFM image of effluent from the AFM fluid cell. Scan size = 661 nm on a side. (B) SEM image of pyromorphite crystals deposited atop the AFM cantilever after reaction of dissolved Pb with hydroxylapatite in an AFM liquid cell. [Reproduced with permission from ref. 51 (Copyright 1998, Elsevier Science).] Ma et al. (53) explored the interactions of dissolved Pb with natural fluorapatites and carbonated fluorapatites. These solids varied in their capacity to remove aqueous Pb (from 39% to 100%). The fraction of Pb removed was not related to the initial surface areas of the apatite particles, but rather to their dissolution rates. Fluoropyromorphite and hydrocerussite were the principal Pb phases formed in these experiments. The exact composition of the products formed by the reaction of aqueous Pb with apatite depends on the solution pH. In the regime of pH 3.1–6.2, Chen et al. (49) found that mixtures of dissolved Pb and carbonate-containing fluorapatite reacted to form fluoropyromorphite by the coupled reactions: Ca5(PO4)3−x(CO3)xF1+x + 6H+⇔5Ca2+ + (3−x)H2PO4− + xH2CO30 + (1 + x)F− [8] 5Pb2+ + 3H2PO4− + F− Pb5(PO4)3F + 6H+. [9] At pH 6.6–6.8, a mixed hydroxylated fluorapatite forms. Ca5(PO4)3−x(CO3)xF1+x+(6 − x)H+⇔5Ca2+ + (3−x)H2PO4− + xHCO3− + (1 + x)F− [10] 5Pb2+ + 3H2PO4− + (F−,OH−)⇔Pb5(PO4)3(F,OH) + 6H+. [11] At circumneutral pH, hydrocerussite and carbonated hydroxyl fluoropyromorphite form by the reactions: Ca5(PO4)3−x(CO3)xF1+x + (6 − x)H+⇔5Ca2+ + (3 − x)H2PO4− + xHCO3− + (1 + x)F− [12] 3Pb2+ + 2HCO3− + 2H2O⇔Pb3(CO3)2(OH)2 + 4H+ [13] − − 2+ − − 5Pb + 3(H2PO4 ,HCO3 ) + (F ,OH )⇔Pb5(PO4,CO3)3(F,OH) + 6H+. [14] Finally, at pH 10.7–11.9, the reaction products consist of hydrocerrusite, hydroxypyromorphite, and lead oxide fluoride (49). Formation of the lead oxide fluoride is described as: 2Pb2+ + H2O+2F−⇔Pb2OF2 + 2H+. [15] Competition with other metal ions also influences the reactions of aqueous Pb with apatites. At initial solution concentrations of <20 mg·liter−1, dissolved Zn, Cd, Ni, Cu, Fe(II), and Al have no discernible effect on the immobilization of 20 mg Pb·liter−1, by hydroxylapatite. Additionally, significant quantities of these metals also are removed from solution (43). Cadmium and Zn also have been shown to have no influence on influence the uptake of Pb by carbonated fluorapatite (49). In these latter experiments, the initial concentrations of dissolved Cd, Zn, and Pb all were equimolar. When the concentration of competing metals far exceeds that of Pb, (M/Pb = 7:1, where M = Zn, Cd, Ni, Cu, Fe(II), or Al and dissolved Pb = 20 mg·liter−1) the interactions of Pb with hydroxylapatite can be significantly inhibited (43). Ma et al. (43) reported dissolved Cu was the most effective in inhibiting Pb immobilization by hydroxylapatite, followed by Fe(II), Cd, Zn, Al, and Ni. In all cases, hydroxypyromorphite

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CONTAMINANT BIOAVAILABILITY IN SOILS, SEDIMENTS, AND AQUATIC ENVIRONMENTS

3369

was the only reaction product detected by XRD besides hydroxylapatite. The intensities of hydroxypyromorphite XRD peaks decreased with increased concentrations of competing metals. Inhibition of hydroxypyromorphite formation was positively correlated with the solubility of known M-phosphates; however, no metal-phosphates other than hydroxylapatite could be detected with XRD or SEM (43). Additionally, geochemical calculations indicated that the solutions were all undersaturated with respect to known Zn-, Cd-, Ni-, Cu-, Fe(II)-, or Al-phosphate phases. The nature of this inhibition is not known, but it possible that these metal ions passivated the surfaces of the apatite, through the formation of sorbed, or surface-precipitated species. Reaction of Apatites with Sorbed and Solid-Phase Pb The pyromorphites are much less soluble than the other Pb-solids commonly present in terrestrial and aquatic environments (Table 1). Thus, it is expected that if sufficent soluble P is present, pyromophites will form at the expense of the “native” Pb-solids (including both adsorbed and precipitated Pb). Ma et al. (33) observed XRD-detectable hydroxypyromorphites after the reaction of hydroxylapatite with Pbsaturated cation exchange resins with hydroxylapatite (note, the initial concentration of dissolved Pb was <1 mg·liter−1). SEM showed that the spherical particles of the exchange resin were coated with hexagonal hydroxypyromorphite needles after they were reacted with apatite. Precipitates were not detectable on the surfaces of the reacted apatite particles. Zhang et al. (54) observed the formation of pyromorphite when hydroxylapatite particles were reacted with goethite suspensions containing sorbed Pb. In these experiments, the Pb-treated goethites and the apatite particles were separated by dialysis membranes. Hexagonal pyromorphite crystals formed on the inside surfaces of the dialysis membranes, indicating that nucleation on the apatite surface is not required for pyromorphite precipitation as was also suggested by Lower and coworkers (50, 51). Consistent with thermodynamic predictions hydroxypyromorphite forms at the expense of more soluble Pb-minerals when they are exposed to apatites. Laperche et al. (21) found XRD- and SEM-detectable hydroxypyromorphite when pure systems of PbO and cerussite each were reacted with hydroxylapatite at pH 5, 6, and 7. Decreased pH caused more rapid reaction rates with greater loss of the parent phases and increased formation of hydroxypyromorphite (Fig. 3). Apparently dissolution of the apatite and/or the original Pb-phases was rate limiting. In all of these cases, initial precipitation of pyromorphite or hydroxypyromorphite quickly reduced the concentration of dissolved Pb. Growth of the Pb-phosphate crystals required continued dissolution of the apatite particles, desorption of adsorbed Pb from the surfaces of the exchange resin and the goethite, and dissolution of the PbO and cerussite phases. The driving force for these dissolution and desorption reactions was formation of the Pb-phosphates. Reactions of Apatites with Pb-Contaminated Soils The interactions of apatites with soil Pb are similar to those observed in model solutions and laboratory mixtures of pure solids. When added to a soil slurry from an automobile battery cracking facility, hydroxylapatite caused a 99% reduction in dissolved Pb (from 3,370 to 36 µg·liter−1) over a 5-h period (33). Similarly, amendment of Pb-contaminated soils with natural fluorapatites reduced the leachability of Pb in soil columns (53) and decreased its extractability by neutral salts (MgCl2) and weak acids (Na-acetate) (55, 56). Laperche et al. (21) used direct physical methods to show apatite-induced formation of pyromorphites in soil slurries. When synthetic hydroxylapatite was added to a soil contaminated by paint residues, these investigators observed decreases in the intensity of the XRD peaks associated with the “native” Pb (cerussite) and appearance of peaks attributed to pyromorphite. The rate and magnitude of changes in the XRD peaks was greater at pH 5 than 7, presumably because of the more rapid dissolution of the cerussite and the hydroxylapatite.

FIG. 3. SEM images of the reaction products of hydroxylapatite and solid-phase forms of Pb (cerussite) at pH 5 (A) and pH 7 (B). [Reproduced with permission from ref. 21 (Copyright 1996, American Chemical Society).] Effects of Apatites on the Bioavailability of Pb in Contaminated Soils The conversion of soil or sedimentary Pb from highly reactive, chemically labile forms to less reactive solids should result in concomitant decreases in bioavailability. Chlopeka and Adriano (57) evaluated this hypothesis by adding natural apatite from North Carolina to a soil amended with varying amounts of Pb-containing flue dust. These materials then were used as potting mixes in glass-house experiments that produced a single crop of maize (Zea mays, var. Pioneer 3165) followed by single crop of barley (Hordeum vulgare, var. Boone). Apatite amendment resulted in decreases in extractable Pb from the soil as well as decreases in tissue Pb concentrations in both crops. Laperche et al. (58) investigated the use of apatite minerals to induce in situ formation of Pb-phosphates in contaminated soil and determined the impact of apatites on Pb uptake by plants. Subsamples of a Pb-contaminated soil (containing 37,026 mg Pb·kg−1 soil from paint residues) were mixed with sufficient quantities of either synthetic hydroxylapatite or natural fluorapatite to convert 33%, 66%, 100%, and 150% of the native soil Pb to either pyromorphite or hydroxypyromorphite. These materials then were planted with sudax grass (Sorghum bicolor L. Moench) a hybrid of sorgum (Sorghum vulgare L. Moench) and sudan grass (Sorghum vulgare var. sudanese). In all cases, hydroxylapatite amendments decreased the concentrations of Pb in the above-ground biomass (shoots)

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CONTAMINANT BIOAVAILABILITY IN SOILS, SEDIMENTS, AND AQUATIC ENVIRONMENTS

3370

(by 92–98%) relative to the unamended soil. Fluorapatite-induced reductions in the concentration of Pb in the shoots ranged from 87% to 96%. The effect of apatite amendments on the Pb concentration in the roots was much different. For both apatites, the lowest concentrations of Pb within the root tissue were associated with the smallest levels of apatite amendments. Increased levels of apatite addition corresponded to increases in the quantity of Pb associated with the roots. At the greatest levels of apatite amendment, the root Pb actually exceeded that found in the unamended soil. Examination of the root surfaces with SEM, energy dispersive x-ray analysis, and XRD analysis of root-associated particles indicated the presence of Ca-substituted pyromorphites on those plants grown in the apatite-treated soils (Fig. 4). Similar particles were not found on the root surfaces of sudax grown in the unamended soils; nor could similar particles be found in the bulk soil after reaction with apatite. Apparently, addition of apatite to the contaminated soil resulted in precipitation of pyromorphite particles on the exterior of the root surfaces. Local acidity within the rhizosphere may have enhanced the local dissolution of apatite grains, facilitating pyromorphite precipitation. Cotter-Howells and Caporn (32) also observed the precipitation of Ca-substituted pyromorphites on plant roots (Agrostis capillaris) grown on Pb-contaminated soils. In this case pyromorphite precipitation may have resulted from root-exudate phosphatase, causing increased local concentrations of phosphate in the rhizosphere. In any event, the formation of pyromorphite likely decreased the bioavailability of Pb. Soils, surface sediments, and surficial aquatic environments are open, dynamic systems best characterized as mixtures of meta-stable solids. It is safe to say that these systems never attain absolute thermodynamic equilibrium, and one can expect to find multiple forms of a given element present. Thus, the bioavailability and chemical lability of toxic elements in these systems are transient properties that are controlled by the reaction dynamics and the total quantity of the most reactive forms of these elements present. Ideally, treatment technologies would facilitate the complete conversion of a toxic element (e.g., Pb) from all pre-existing forms to the most geochemically stable phase; but as we know, kinetic constraints will prevent this from happening. Fortunately, decreases in the chemical reactivity of and bioavailability of a given element (e.g., Pb) can be accomplished by elimination of the most reactive forms (58). Conclusions Human activity on this planet has altered the distribution and form of various elements in the periodic table. Often these activities have converted many potentially toxic metals from nonreactive, geochemically stable solids into forms that are more soluble and bioavailable, increasing their effective toxicity. Fortunately, it is often possible to reverse this process, transforming reactive forms of toxic metals to less labile species through appropriate precipitation reactions. Although such an approach does not remove the element in question from the biosphere, it can significantly reduce its bioavailability. In many instances, this approach may be a practical alternative to more invasive methods of environmental restoration (e.g., excavation and removal of contaminated materials). In the event that more extensive treatment is chosen, geochemical stabilization may still be desirable as a rapid response treatment or to increase the chemical stability of excavated materials in landfill environments.

FIG. 4. SEM micrograph of sudax root grown in Pb-contaminated soil mixed with hydroxylapatite. Note the pyromorphite crystals on the root. This approach is not limited to Pb, nor is it only possible to induce precipitation of phosphates. Extensive research has been conducted on various treatment technologies designed to remove toxic ions from contaminated waters or to stabilize these elements in soil waste materials or contaminated soils and sediments (5). Many of these efforts involve the formation of geochemically stable solids through precipitation and/or adsorption reactions. In essence, these methods attempt to close the circle, converting labile forms of toxic elements into less reactive solids more consistent with long-term geochemical equilibrium.

1. Krauskopf, K. B. & Bird, D. K. (1995) Introduction to Geochemistry (McGraw-Hill, New York). 2. Hayes, K. F. & Traina, S. J. (1998) in Soil Chemistry and Ecosystem Health, ed. Huang, P. M. (Soil Sci. Soc. Am., Madison, WI), pp. 45–84. 3. Newman, M. C. & McIntosh, A. W. (1991) Metal Ecotoxicology Concepts and Application (Lewis, Chelsea, MI). 4. Adriano, D. C. (1992) Biogeochemistry of Trace Metal (Lewis, Chelsea, MI). 5. Iskandar, I. K. & Selim, H. M. (1992) Engineering Aspects of Metal-Waste Management (Lewis, Chelsea, MI). 6. Hamelink, J. L., Landrum, P. F., Bergman, H. L. & Benson, W. H. (1994) Bioavailability Physical, Chemical, and Biological Interactions (Lewis, Chelsea, MI). 7. Ernst, W. H. O. (1996) Appl. Geochem. 11, 163–167. 8. Davis, A., Sellstone, C., Clough, S., Barrick, R. & Yare, B. (1996) Appl. Geochem. 11, 409–423. 9. Davies, N. A., Taylor, M. G. & Simkiss, K. (1997) Environ. Pollut. 96, 179–184. 10. Peijnenburg, W. J. G. M., Posthuma, L., Eijsackers, H. J. P. & Allen, H. E. (1997) Ecotox. Environ. Saf. 37, 163–172. 11. Lindsay, W. L., Vlek, P. L. G. & Chien, S. H. (1989) in Minerals in Soil Environments, eds. Dixon, J. B. & Weed, S. B. (Soil Sci. Soc. Am., Madison, WI), 2nd Ed., pp. 1089–1130. 12. Troeh, F. R. & Thompson, L. M. (1993) Soils and Soil Fertility (Oxford Univ. Press, New York). 13. Tisdale, S. L., Nelson, W. L., Beaton, J. D. & Havlin, J. L. (1993) Soil Fertility and Fertilizers (Macmillan, New York). 14. Lindsay, W. L. (1979) Chemical Equilibria in Soils (Wiley, New York), pp. 162–209. 15. Laws, E. A. (1993) Aquatic Pollution: An Introductory Text (Wiley, New York). 16. Lee, G. F., Jones, R. A. & Rast, W. (1980) in Phosphorous Management Strategies for Lakes, eds. Loehr, R. C., Martin, C. S. & Rast, W. (Ann Arbor Science, Ann Arbor, MI), pp. 259–308. 17. Davis, A., Drexker, J. W., Ruby, M. V. & Nicholson, A. (1993) Environ. Sci. Technol. 27, 1415–1425. 18. Davis, A., Ruby, M. V., Bloom, M., Schoof, R., Freeman, G. & Bergstrom, P. D. (1996) Environ. Sci. Technol. 30, 392–399. 19. Lasaga, A. C. (1998) Kinetic Theory in the Earth Sciences (Princeton Univ. Press, Princeton). 20. Ruby, M. V., Davis, A. & Nicholson, A. (1994) Environ. Sci. Technol. 28, 646–654. 21. Laperche, V., Traina, S. J., Gaddam, P. G. & Logan, T. J. (1996) Environ. Sci. Technol. 30, 3321–3326. 22. Nriagu, J. O. (1972) Inorg. Chem. 11, 2499–2503. 23. Nriagu, J. O. (1973) Geochem. Cosmochim. Acta 37, 367–377. 24. Nriagu, J. O. (1974) Geochem. Cosmochim. Acta 37, 367–377.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

CONTAMINANT BIOAVAILABILITY IN SOILS, SEDIMENTS, AND AQUATIC ENVIRONMENTS

25. Santillan-Medrano, J. & Jurinak, J. J. (1975) Soil Sci. Soc. Am. Proc. 39, 851–858. 26. Sauvé, S., McBride, M. B. & Hendershot, W. H. (1997) Environ. Pollut. 2,149–155. 27. Cotter-Howells, J. D. & Thornton, I. (1991) Environ. Geochem. Health 12, 127–135. 28. Cotter-Howells, J. D., Champness, P. E., Charnock, J. M. & Patrrick, R. A. D. (1994) J. Soil Sci. 45, 393–402. 29. Cotter-Howells, J. D. (1996) Environ. Pollut. 1, 9–16. 30. Rabinowitz, M. B. (1993) Bull. Environ. Contam. Toxicol. 51, 438–444. 31. Berti, W. R. & Cunningham, S. C. (1997) Environ. Sci. Technol. 31, 1359–1364. 32. Cotter-Howells, J. & Caporn, S. (1996) Appl. Geochem. 11, 335–342. 33. Ma, Q.Y., Traina, S.J., Logan, T. J. & Ryan, J.A. (1993) Environ. Sci. Technol. 27, 1803–1810. 34. Hughes, J. M., Cameron, M. & Crowley, K. D. (1989) Am. Mineral. 74, 870–876. 35. Santos, R. V. & Clayton, R. N. (1995) Am. Mineral. 80, 336–344. 36. Rakovan, J. & Reeder, R. J. (1994) Am. Mineral. 79, 892–903. 37. Suzuki, T., Hatsushika, T. & Hayakawa, Y. (1981) J. Chem. Faraday Trans. 1 77, 1059–1062. 38. Suzuki, T., Hatsushika, T. & Miyake, M. (1982) J. Chem. Soc. Faraday Trans. 1 78, 3605–3611. 39. Suzuki, T., Ishigaki, K. & Miyake, M. (1984) J. Chem. Soc. Faraday Trans. 1 80, 3157–3165. 40. Takeuchi, Y., Suzuki, T. & Arai, H. J. (1988) J. Chem. Eng. Jpn 21, 98–100. 41. Takeuchi, & Arai, H. J. (1990) J. Chem. Eng. Jpn 23, 75–80. 42. Ma, Q. Y., Logan, T. J. & Traina, S. J. (1994) Environ. Sci. Technol. 28, 408–419. 43. Ma, Q. Y., Traina, S. J. & Logan, T. J. (1994) Environ. Sci. Technol. 28, 1219–1228. 44. Xu, Y. & Schwartz, F. W (1994) J. Contam. Hydrol. 15, 187–206. 45. Xu, Y., Schwartz, F. W. & Traina, S. J. (1994) Environ. Sci. Technol. 28, 1472–1480. 46. Jeanjean, J., Rouchard, J. C., Tran, L. & Fedoroff, M. (1995) J. Radioanal. Nucl. Chem. Lett. 201, 529–539. 47. Reichert, J. & Binner, J. G. P. (1996) J. Mater. Sci. 31, 1231–1241. 48. Chen, X., Wright, J. V., Conca, J. L. & Peurrung, L. M. (1997) Water Air Soil Pollut. 98, 57–78. 49. Chen, X., Wright, J. V., Conca, J. L. & Peurrung, L. M. (1997) Environ. Sci. Technol. 98, 624–631. 50. Lower, S. K., Maurice, P. A., Traina, S. J. & Carlson, E. H. (1998) Am. Mineral. 83, 147–158. 51. Lower, S. K., Maurice, P. A. & Traina, S. J. (1998) Geochim. Cosmochim. Acta 62, 1773–1780. 52. Laperche, V. & Traina, S. J. (1998) in Adsorption of Metals by Geomedia, ed. Jenne, E. A. (Academic, New York), pp. 255–276. 53. Ma, Q. Y., Logan, T. J. & Traina, S. J. (1995) Environ. Sci. Technol. 29, 1118–1126. 54. Zhang, P., Ryan, J. A. & Bryndzia, L. T. (1997) Environ. Sci. Technol. 31, 2673–2678. 55. Ma, L. Q. & Rao, G. N. (1997) J. Environ. Qual. 26, 788–794. 56. Ma, L. Q., Choate, A. L. & Rao, G. N. (1997) J. Environ. Qual. 26, 801–807. 57. Chlopeka, A. & Adriano, D. C. (1997) Sci. Total Environ. 207, 195–206. 58. Laperche, V., Logan, T. J., Gaddam, P. & Traina, S. J. (1997) Environ. Sci. Technol. 31, 2745–2753.

3371

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3372

Airborne minerals and related aerosol particles: Effects on climate and the environment

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3372–3379, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. PETER R. BUSECK* AND MIHÁLY PÓSFAI† PNAS is available online at www.pnas.org.

Departments of Geology and Chemistry/Biochemistry, Arizona State University, Tempe, AZ 85287 ABSTRACT Aerosol particles are ubiquitous in the troposphere and exert an important influence on global climate and the environment. They affect climate through scattering, transmission, and absorption of radiation as well as by acting as nuclei for cloud formation. A significant fraction of the aerosol particle burden consists of minerals, and most of the remainder—whether natural or anthropogenic—consists of materials that can be studied by the same methods as are used for fine-grained minerals. Our emphasis is on the study and character of the individual particles. Sulfate particles are the main cooling agents among aerosols; we found that in the remote oceanic atmosphere a significant fraction is aggregated with soot, a material that can diminish the cooling effect of sulfate. Our results suggest oxidization of SO2 may have occurred on soot surfaces, implying that even in the remote marine troposphere soot provided nuclei for heterogeneous sulfate formation. Sea salt is the dominant aerosol species (by mass) above the oceans. In addition to being important light scatterers and contributors to cloud condensation nuclei, sea-salt particles also provide large surface areas for heterogeneous atmospheric reactions. Minerals comprise the dominant mass fraction of the atmospheric aerosol burden. As all geologists know, they are a highly heterogeneous mixture. However, among atmospheric scientists they are commonly treated as a fairly uniform group, and one whose interaction with radiation is widely assumed to be unpredictable. Given their abundances, large total surface areas, and reactivities, their role in influencing climate will require increased attention as climate models are refined. There is widespread concern over the enhanced global warming that might result from the buildup of “greenhouse gases” in the atmosphere. The effects of aerosols (suspensions of solid or liquid particles in air) on Earth's radiation balance is less widely realized, and recognition of the role of airborne minerals has occurred only relatively recently. Climate is fundamentally influenced by Earth's energy budget, which depends on radiation received from the sun and energy radiated back to space. Incoming radiation is primarily in the visible range, whereas exiting radiation is largely in the IR. Greenhouse gases (H2O, CO2, CH4, N2O, etc.) absorb IR radiation and radiate it back to Earth's surface. Anthropogenic emissions of greenhouse gases cause increases in surface temperature (the “greenhouse effect”) and can have profound effects on climate and thus on societal welfare (1, 2). Aerosol particles also have a major influence on global climate and climate change; they can locally either intensify or moderate the effects of the greenhouse gases through the scattering or absorption of both incoming solar radiation and thermal radiation emitted from Earth's surface. Aerosols also act as cloud condensation nuclei (CCN) and thereby modify the radiative properties of clouds. The profound effects of atmospheric aerosols are surprising in view of their exceedingly low concentrations: the volumetric ratio of aerosol particles to atmospheric gases is between roughly 10−10 and 10−14 (3). The focus of this paper is on those particles, their compositions and structures and their effects on climate and, to a lesser extent, on the environment.‡ A growing awareness of the impact of particulate aerosols on climate, and the incompletely recognized but serious effects of anthropogenic aerosols, is summarized in several recent reviews (4, 5 and 6). One reason for the relatively slow recognition of the role of particulate aerosols is that their study has fallen to disparate groups of scientists. Radiative transfer and other physical properties tend to be handled by one group (largely meteorologists and physicists), whereas chemical effects such as acid rain are emphasized by different scientists (mainly chemists). Perhaps the least attention to date has been on the geochemistry and mineralogy of aerosol particles and the effects of speciation. Efforts to control greenhouse gases have been formalized by international treaty, e.g., the 1997 Kyoto Protocol on Climate Change. A comparable international effort to understand and control anthropogenic aerosol emissions has not (yet) occurred, at least in part because the extent to which they affect climate is not satisfactorily known. The incremental effects of anthropogenic increases in greenhouse gases are long lived (decades to centuries), whereas those of aerosols are shorter (weeks) (7). However, the sizes, compositions, and atmospheric lifetimes of particulate aerosols can vary spatially and temporally, and their strongest effects tend to be near their sources. If aerosols indeed offset climate responses to greenhouse gases, then the climate effects of greenhouse gases are even more substantial than has been recognized. What role do mineralogists and geochemists have in addressing these and related issues of fundamental importance for human society and welfare? The main difference between most aerosol particles and the materials that are routinely studied by mineralogists is that most terrestrial minerals are not as fine grained. There are major problems with studying fine-grained materials, and therefore atmospheric chemists have traditionally emphasized bulk analyses to determine aerosol types. However, it is the individual chemical species that affect the radiative balance and

Abbreviations: CCN, cloud condensation nuclei; TEM, transmission electron microscope; SEM, scanning electron microscope; MBL, marine boundary layer; FT, free troposphere; NSS, non-sea salt; AFM, atomic force microscopy; ACE, Aerosol Characterization Experiments; FeLINE, experiments in the equatorial Pacific; ASTEX/ IMAGE, experiments in the North Atlantic. *To whom reprint requests should be addressed. e-mail: [email protected]. †Present address: Department of Earth and Environmental Sciences, University of Veszprém, Veszprém, POB 158, H8201 Hungary. ‡Aerosol particles, to the extent they consist of nonanthropogenic homogeneous inorganic solids of more or less uniform composition and have ordered structures, fit generally accepted definitions of minerals. However, except for the title, in this manuscript we follow the usage common among atmospheric scientists and use minerals to refer to materials that once resided on Earth s land surface.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3373

climate as well as visibility and health. Paraphrasing a recent statement (8) and allowing for slight exaggeration, interpreting environmental and health effects of aerosols from bulk rather than individual-particle analyses is like interpreting mortality reports in a war zone from bulk airborne lead concentrations rather than from bullets. Our group has focused on the painstaking but necessary analysis of individual particles. High-spatial-resolution methods, using electron beams as the primary probes of both chemistry and structure, have been developed to study increasingly fine-grained minerals. We examine the inorganic and, in special cases, the organic fraction of aerosol particles with electron microprobe analyzers and scanning electron microscopes (SEMs) and transmission electron microscopes (TEMs). In this paper we provide a background to the above issues and indicate ways in which mineralogical experience and experimental techniques can provide uniquely useful information. We first review the broad problems and briefly describe the analytical techniques, then discuss some of our recent transmission electron microscopy results regarding sulfate, soot, sea salt, and mineral aerosols. Aerosols, Climate, and the Environment Nature of Airborne Minerals and Other Inorganic Aerosols. Andreae (9) estimated that the largest components of the global atmospheric aerosol are, in decreasing mass abundances: mineral aerosols—primarily from soil deflation but also with a minor component (<1%) from volcanoes (16.8 Tg), sea salt (3.6 Tg), natural and anthropogenic sulfates (3.3 Tg), products of biomass burning excluding soot (1.8 Tg) and of industrial sources including soot (1.4 Tg), natural and anthropogenic nonmethane hydrocarbons (1.3 Tg), natural and anthropogenic nitrates—largely from NOx (0.6 Tg), and biological debris (0.5 Tg). In general, a distinction is made between primary particles, which are injected into the troposphere, and secondary particles, which form within the troposphere. Sea salt from spray, desert dust, volcanic mineral emissions, and re-entrained road dust are examples of primary aerosols. In contrast, particles produced by condensation of gases result in secondary aerosols. Primary aerosol particles tend to be larger, dominating the “coarse” fraction, which is >1 µm in diameter and mostly mechanical in origin. The fine fraction is enriched in secondary particles, largely between 0.1 and 1 µm in diameter and mainly chemical in origin. The smaller size range is also called the “accumulation mode,” in distinction to the “coarse mode.” In addition, there is a “nucleation mode,” with particles smaller than 0.1 µm. Recently “nanoparticles” in the 3- to 10-nm range have also been distinguished (10). Human activities affect the amounts, types, and distributions of aerosols that enter the atmosphere. Anthropogenic particles are especially abundant in the submicrometer portion of the aerosol, and they provide a major uncertainty in estimating climatic effects (11, 12). Different particle types (mineral dust, sulfates, carbonaceous materials, sea salt, organic compounds) can occur in the same air mass. An important question that we commonly address in our individual-particle work is, to use the terminology of atmospheric chemistry, whether the particles are internally or externally mixed. Phases that are externally mixed occur within the same aerosol but in discrete separate particles. If they are internally mixed, then they occur within the same particles; inhomogeneous internal mixtures are much like minerals in a rock, whereas homogeneous internal mixtures are solutions. Changes from predominantly external to internal mixtures can occur when particles grow by coagulation with different species, a process that is especially common during entrainment into clouds. The differences between internal and external mixtures can significantly affect the optical properties and radiative efficiency of the aerosol and its ability to act as CCN. The nature and magnitude of these effects have received considerable attention (13, 14), but there are differences of opinion. For example, some authors (15, 16) indicate that scattering calculations for mixed aerosols are not significantly affected by assumptions regarding internal or external mixing, whereas others (17, 18) reach the opposite conclusion for mixtures of sulfate and soot. Clearly, the problems have not been resolved, and most current models do not yet consider the subtleties of internal mixtures. Atmospheric aerosol particles can also profoundly influence the environment. Dust, smoke, and haze locally impair visibility and health in both urban and rural regions. The harmful respiratory health effects of certain mineral and anthropogenic particles are well documented and have led to a plethora of federal rules and regulations (19, 20). Knowledge of the compositions and microphysical properties of aerosols is critical first for understanding and then for ameliorating some of these pernicious environmental and health effects. Effects of Airborne Minerals and Other Particles on Climate — Radiative Forcing. Forcing is the term used to describe changes imposed on the planetary energy balance. It is measured in watts per square meter (Wm−2). Aerosol radiative forcing, which refers to the effects of aerosols, is termed direct if it results from backscattering and absorption of radiation by the aerosol particles themselves, and indirect if it results from the influence of the particles on the optical properties, amounts, and lifetimes of clouds. Positive forcing results in a net warming at Earth's surface, and negative forcing results in a net cooling. The magnitude of the radiative effects of aerosol particles depends on their compositions, sizes and size distributions, abundances, hygroscopicities, surface properties, densities, and refractive indices (15, 21, 22 and 23). Some of these parameters are interdependent, and they can vary with locality, sources, and environmental variables such as intensity of sunlight and relative humidity. Also, inventories of concentrations of particle types, especially those with a broad range of spatial and temporal distributions such as from industrial, arid urban, and particular geological sources (deserts, volcanoes) need to be well known for specific particulate assemblages. Their vertical distributions and underlying surface albedos are also important (24). The composite effect is complex and will require extensive measurement before it is well understood. Sulfates are thought to be the most important scatterers of solar radiation on a global scale, producing a net cooling at Earth's surface, whereas soot tends to be a major absorber of the solar radiation and so has a net warming effect (4). The role of mineral dust is more ambiguous (9, 25). Particles tend to be most efficient in scattering radiation having wavelengths comparable to their physical sizes; submicron dust particles are efficient scatterers of the incoming sunlight and thus can have a cooling effect, especially near Earth's surface. On the other hand, mineral particles also absorb light and thus have a heating effect at the altitude at which they occur (cf. Mineral Dust). Because the outgoing radiation is in the IR, and silicate minerals have bands in which they absorb in the IR, they can act as “greenhouse particles.” The larger mineral aerosol particles tend to have shorter atmospheric lifetimes and to be most concentrated in the lower troposphere, near their source areas, and so produce localized effects. Cloud droplets form on aerosol particles as nuclei. The number, sizes, and compositions of such CCN have major influences on cloud formation. Hygroscopic materials such as sulfates and sea salts are especially efficient as nuclei; mineral dust and combustion products can also be effective, especially if they are wettable or acquire hygroscopic coatings (26). Increased numbers of CCN lead to more cloud droplets and concurrent decreases in droplet sizes (for given cloud water contents) (27, 28 and 29). Because of multiple scattering within the cloud, cloud albedo tends to increase with numbers and small sizes of hygroscopic aerosol particles, which results in increased cooling. In addition, clouds with more and smaller droplets are less prone to rain and drizzle formation and therefore persist longer, having more time to exert their cooling effect.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3374

Mineralogical Techniques Applied to Single Aerosol Particles Electron-Beam Analyses. The traditional method of studying the chemistry of aerosol particles is through bulk methods. By using such techniques, large numbers of particles are analyzed en masse. However, the results provide no information about individual aerosol species and the mixing state of the aerosol; knowing whether the aerosol particles are attached to one another or to other types of particles is critical for understanding processes that involve or modify the particles during atmospheric transport. Our approach has been to emphasize the individual particles because the radiative, environmental, and health effects of particles depend on their speciation rather than their averaged bulk compositions. Our primary instruments are the analytical SEM, electron microprobe analyzer (EMPA), and TEMs. The SEM and EMPA can operate in automated mode, running unattended around the clock and producing large numbers of analyses (30, 31, 32, 33, 34 and 35). Particle analysis using TEMs, which are the main instruments used for the results reported here, cannot be automated because of the complexity of the method (sample tilting in diffraction mode to obtain critical crystal orientation, combined with changes to imaging and analysis modes to obtain size, shape, and chemical information). Thus, the TEM does not readily provide the statistical depth produced by the SEM or EMPA, but it can be used to analyze far smaller particles (down to <1 nm vs. ≥0.1 µm) and, perhaps more importantly, it can provide crystallo-graphic, morphologic, size, and chemical information for individual particles. Many of the aerosol particles of greatest interest for influencing climate fall within a size range that is accessible only by using the TEM (36, 37, 38, 39, 40, 41, 42, 43 and 44). A general background for mineralogical transmision electron microscopy is given by ref. 45. Aerosol Time-of-Flight Mass Spectrometry. A recent instrumental development of considerable interest is the aerosol time-of-flight mass spectrometer (ATOFMS), which can be used for rapid measurements of both sizes and compositions of individual particles (46, 47 and 48). Particles pass through a laser beam, which blasts each particle into positively and negatively charged ions. These are then analyzed in the ATOFMS to determine the atomic or molecular weight of each ion, from which fragment compositions can be determined. Particle sizes are calculated from their velocities as they move through the ATOFMS. A result is that chemical and size information is obtained from individual particles in real time, although back calculations are required to infer parent species from their ion fragments. Samples. Because the oceans cover ≈70% of the globe, much attention has been given to the profound effects of the oceans on climate. A recent area of interest is the role of aerosol particles in the marine troposphere and the effects they have on climate. Marine aerosols are important because there is generally enough humidity to form clouds, but there can be a shortage of CCN; thus, the indirect climate effect of aerosols is far more pronounced over sea than land. Aerosols in such areas are also the most likely to represent unpolluted, “global background” conditions, and the contributions of natural and anthropogenic sources are easier to identify than in continental or polluted oceanic atmospheres. Two recent international research programs are shedding important new insights onto the role of aerosols in the troposphere. The Aerosol Characterization Experiments (ACE) took place in late 1995 in the Southern Ocean near Tasmania (ACE-1) and in 1997 in the North Atlantic (ACE-2). They produced integrated measurements from ships, airplanes, and ground stations; ACE-1 involved researchers from 45 institutions in North America, Europe, Australia, and Asia (49). In the ACE-1 campaign we used two platforms. At Cape Grim, Tasmania, we collected aerosol particles onto filters in a two-stage impactor (for SEM analysis) and directly onto TEM grids. We also used a one-stage impactor that was mounted on a C-130 aircraft. In this paper we use results from a Lagrangian experiment in which a tagged air mass was followed and the aerosol evolution within it studied over time. The airplane flew “stacked circles,” with each circle at a different altitude. From the ACE-2 experiment we use data obtained from particles collected onto TEM grids that were placed on filters. The sampling station was on a mountain top (at 2,600 m altitude) on Izaña, Canary Islands. We also include results from earlier experiments in the equatorial Pacific (FeLINE; ref. 41) and the North Atlantic (ASTEX/MAGE; refs. 42, 50). Open Problems, Issues, and Results Sulfates§ and Associated Soot and Organic Species. Sulfates are probably the main climate-cooling aerosols (51, 52 and 53). They scatter solar radiation and are effective as CCN; the result is negative forcing and thus cooling at Earth's surface. The radiative forcing of sulfate aerosol particles, especially in the Northern Hemisphere, is roughly equivalent in magnitude but opposite in sign to the combined forcing by the greenhouse gases (12, 51, 53, 54 and 55). If sulfates are internally mixed with other aerosol species, their hygroscopic behavior and optical properties can change dramatically (14, 17, 18) and result in diminished cooling. It is therefore important to determine the state of mixing of these particles. We studied sulfate particles in sample sets obtained from (i) the polluted marine boundary layer (MBL) near the Azores Islands (ASTEX/MAGE), (ii) the remote unpolluted Southern Ocean MBL and free troposphere (FT) (ACE-1), and (iii) the essentially clean North Atlantic FT at Izaña, Canary Islands (ACE-2). In distinction to the relatively small amount of primary sulfate that is in sea water and thus occurs in every sea-salt aerosol particle, the secondary sulfate aerosol, which is what interests us here, is called non-sea salt (NSS) sulfate. TEM Observations. The compositions of hydrated sulfate particles may change during sample processing or in the vacuum of a TEM. For such samples we must rely on indirect evidence, such as morphological features, to identify the compositions of

FIG. 1. TEM images of (NH4)2SO4. (a) The selected-area electron-diffraction pattern (upper left) confirms the identification. The arrow points to a soot aggregate. (Azores, North Atlantic, ASTEX/MAGE); (b) Rings of small (NH4)2SO4 crystals that formed as the sulfate particles dehydrated. The dimensions of the halos can be used to distinguish among particles that likely had different water contents while still airborne. (Southern Ocean, ACE-1.)

§Here we use “sulfate” for particles that consist purely or predominantly of sulfates and are formed in the atmosphere by either homogeneous or heterogeneous reactions involving SO2 gas. They have compositions ranging from H2SO4 to (NH4)2SO4. Sea-salt sulfate and the NSS sulfate formed by conversion of the original sea-salt particles are not considered in this section.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3375

the original atmospheric particles. Fig. 1a shows a typical particle that was collected above the North Atlantic Ocean when a polluted air mass of European origin was sampled. An energy-dispersive x-ray spectrum shows O, N, and S, and the selected-area electron-diffraction pattern was used to confirm its identity as crystalline (NH4)2SO4. Acidic particles are more hygroscopic than pure (NH4)2SO4 and so will contain more water and spread farther on the TEM grid than neutral species. The halos of discrete smaller particles that form (Fig. 1b) presumably result from the acidity of the original sulfate (37) and are likely related to the amount of adsorbed water. Different samples typically have characteristic distributions of sulfates with or without halos. For example, the images in Fig. 1a and Fig. 1b and Fig. 2a were obtained from different samples; we assume that the halo-free particle (Fig. 1a) was (NH4)2SO4 even while in the air. The sulfate with a single ring of smaller particles (Fig. 2a) was slightly acidic, whereas the particles with multiple rings (Fig. 1b) were more acidic. Other differences in appearance reflect the degree of decomposition produced by the electron beam of the TEM and contrast of the image. Internal Mixing of Sulfate and Soot. There is increasing evidence that carbonaceous aerosols can have major impacts on aerosol radiative forcing (56, 57). Soot both scatters and absorbs radiation at solar wavelengths and thus produces cooling at Earth's surface while warming the atmosphere around the particle. The net effect is a slight positive forcing. The resulting warming of the atmosphere can have particularly significant effects over highly reflective areas such as those covered with snow. When soot derived from fossil fuels is entered into models, it produces a positive forcing that can offset 40% or more of the negative global radiative forcing of sulfate (17). Organic particles can be the dominant CCN aerosols above tropical rain forests (58) and densely populated areas such as the East Coast (56, 57). Internal mixtures of sea salt, sulfate, and material that is presumably organic carbon occur widely (59). We found that many sulfate particles are internally mixed with soot in the marine troposphere (Fig. 1a and Fig. 2a). Soot/sulfate aggregates had been observed in polluted urban environments (40, 60, 61 and 62), but it is surprising that even in the remote oceanic atmosphere a significant fraction of sulfates contain soot inclusions (44). Up to ≈90% of the submicron ammonium sulfate particles contain soot aggregates in samples collected above the North Atlantic Ocean during a pollution episode (ASTEX/ MAGE). In the clean atmosphere above the Southern Ocean during the ACE-1 experiment, 11 to 46% of the sulfate particles contain soot. Soot/sulfate aggregates comprise about 20% of all sulfate particles in two samples that we studied from Izaña (ACE-2). Because we observed similar relative numbers of soot/ sulfate internal mixtures in the aerosol from two essentially clean but geographically distant locations (Southern Ocean and Izaña, North Atlantic), we believe it likely that soot/sulfate internal mixing is a globally important phenomenon and must be considered when the climate effects of sulfate aerosols are modeled (44). We observed variations in the compositions and microstructures of soot particles and in their associations with other species. (i) The polluted North Atlantic ASTEX/MAGE samples have soot associated with silica fly ash spherules that contain various metals; the association suggests that a coal-burning power plant was the likely source of the soot. (ii) Depending on the sample, between 20% and 50% of soot particles from both the Southern Ocean and Izaña contain significant K, which is typical of biomass burning (63, 64, 65 and 66). (iii) Only C and O are detectable in most soot from the Southern Ocean; such particles typically have “compact” microstructures, as described below. Because they contain no associated fly ash, K, V, or other metals, any of which would suggest alternate sources, we conclude they were likely emitted by aircraft.

FIG. 2. TEM images of an internal mixture of (NH4)2SO4 and soot. (a) The halo is similar to those in Fig. 1. The arrow points to a soot aggregate. (Southern Ocean, ACE-1); (b) High-resolution image of the arrowed tip of the soot aggregate in a. A degree of ordering is evident in the onion-like graphitic layers, seen edge on. (c) A large branching soot aggregate; such aggregates are typical of combustion processes (95). (Southern Ocean, ACE-1.) Most soot particles from the Southern Ocean atmosphere consist of only a few 10- to 50-nm globules. The (NH4)2SO4 particle in Fig. 2a contains such a typical small soot inclusion. Individual graphitic layers are wavy and subparallel, forming the poorly crystalline turbostratic structure characteristic of soot (Fig. 2b). Large branching soot aggregates (Fig. 2c) are less common than the small particles. Our observations of soot/sulfate particles are significant for several reasons. Internally mixed soot and sulfate raise the question of whether they became attached through (i) coagulation, (ii) processing in cloud droplets, or (iii) condensation and oxidization of SO2 on the soot surfaces. If mechanism (iii) is responsible for the observed aggregates, then soot particles provided nuclei for heterogeneous sulfate formation even in the remote marine troposphere, a pathway for sulfate formation that has not been considered in climate models. Further studies are needed to unravel the processes that bring atmospheric sulfate and soot particles together. A soot inclusion within a sulfate particle changes the optical properties of the sulfate. That particle, if soot-free, would have a cooling effect on the atmosphere; however, if it contained enough soot, it could emit IR radiation and thereby heat its immediate environment. The magnitude of the effect depends on the size of the inclusion. On the other hand, nucleation of sulfate particles on soot would result in additional CCN particles that exert increased indirect cooling. The presence of absorbing soot inclusions in clouds can have a major offset of the expected increase of cloud albedo. Clearly, the aggregation of soot with sulfate can cause changes in both the direct and indirect radiative forcing of sulfates and thus needs further study. Internal Mixing of Sulfate and Organic Species. For determining climate effects, it is important to know whether inorganic particles contain organic coatings. Such coatings can influence the hygroscopic behavior of particles by retarding water evaporation (67, 68) and increasing or decreasing water adsorption by inorganic aerosols (69). Many particles of ammonium sulfate, when sublimated with the electron beam in the TEM, leave visible residues on the grid. The residues in Fig. 3 include a soot aggregate and dark films that indicate only S in their energy-dispersive x-ray spectra. The most reasonable identification of the residue films is that they consist of organic compounds that coated the original sulfate particles. Indirect evidence for the probable presence of organic coatings on sulfates is provided by combining atomic force microscopy (AFM) and TEM images of the same sulfate particles (43). The AFM image (Fig. 4) was obtained under ambient conditions at a

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3376

relative humidity of 31%. The TEM image shows that the same particles are significantly smaller in the vacuum of the TEM (at 0% relative humidity) than during the AFM study. The amount of water lost when they were inserted into the TEM was calculated and was in excess of what could be expected for pure (NH4)2SO4. We believe that organic coatings on the particles are responsible for the observed anomaly in the hygroscopic behavior of these sulfates.

FIG. 3. TEM image of ammonium sulfate (a) before and (b) after it was sublimated by the electron beam. We believe the dark films in b are residues of organic material that coated the aerosol particle before sampling. The arrow marks a small soot particle. (Southern Ocean, ACE-1.) Sea Salt. Sea-salt aerosol particles are generated when rising bubbles burst at the surface of the ocean and water drops are ejected into the atmosphere (Fig. 5 a and Fig. 5 b). The droplets are typically 0.1 to 100 µm in diameter (70). Sea-salt particles can dominate light scattering by aerosols and comprise a significant fraction of CCN above some regions of the unpolluted oceans (48). The large mass of sea-salt aerosol means that it dominates the particulate surface area in the marine troposphere and, through heterogeneous surface reactions, plays an important role in the atmospheric cycles of Cl, S, and N (Fig. 5c). There is controversy about the significance of sea salt for aerosol radiative effects in the MBL. It had been thought that sea-salt particles are mostly too large to be efficient scatterers of solar radiation and that the smaller NSS sulfate is the main contributor to radiative effects in the MBL. However, recent data indicate that in some regions sea salt can be a major source of CCN and can even dominate NSS sulfate (21, 70). Individual-particle analysis can be used to determine sea salt vs. sulfate number concentrations and sizes to obtain a better understanding of the radiative importance of sea salt. During ACE-1 we found that in the windy Southern Ocean, sea salt is a constituent in almost all MBL particles; only less than 1% were pure sulfate (48), although in other areas sulfate is a major particle type (31, 71).

FIG. 4. (a) AFM and (b) TEM images of identical sulfate particles. Note the decrease in size caused by dehydration within the TEM. The amount of lost water is larger than expected and suggests increased hygroscopicity through organic coatings. (AFM image by Huifang Xu) (Azores, North Atlantic, ASTEX/MAGE.) FIG. 5. TEM images of sea salt. (a and b) Subhedral halite (NaCl) and euhedral sulfate crystals. The particle in b belongs to the smallest sea-salt particles that occur in the ACE-1 samples. (Southern Ocean, Cape Grim, ACE-1); (c) Halite particles in various stages of conversion to sulfate and nitrate. Grain A is partly converted, whereas C has been completely converted to nitrate and grains B to sulfates. (Azores, North Atlantic, ASTEX/MAGE.)

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3377

FIG. 6. Na-Cl-S plots showing changes in composition of aerosol particles with altitude (Upper) and time (Lower). Each of the upper triangles represents a sample from a single altitude (meters above sea level are indicated) and contains data from between 30 to 70 particles. The upper samples were collected during a series of flights on one day. The lower triangles contain composites of samples from several altitudes; the one on the left contains all particles indicated above, whereas the one on the right contains those collected 26 hr later. Each contains data from about 250 particles. Mg, K, and Ca are not included in the diagrams because their ratios to Na do not change (or are within our analytical error). (Southern Ocean, ACE-1.) Changes occur with altitude in the compositions of sea-salt particles collected above the Southern Ocean (Fig. 6), and the reaction pathway from A (fresh sea salt) to B (sulfate) can be traced. Unreacted particles dominate the samples collected within the MBL (at ≈30, 150, and 600 m); their S contents are consistent with the presence of sea-salt sulfate (i.e., sulfate that was in the seawater droplet when it was ejected into the atmosphere). There is more variation in particle compositions at 150 and 600 m than near the sea surface (at 30 m), but the differences are small. In addition to unreacted sea salt, S-rich particles occur at 1,200 and 2,100 m, above the thermal inversion. Several particles in the 2,100m sample were completely converted to sulfate. More reacted particles occur in the FT, i.e., at high altitudes, presumably because they spent more time in the atmosphere and had more time to react with SO2 than the particles in the MBL. Some particles in the FT may have been transported from remote locations where SO2 concentrations could have been higher than at the ACE-1 study area. Temporal changes in sea-salt compositions can also be observed from the ACE-1 samples. During Lagrangian experiment “B,” an air mass was tagged by balloons and sampled from flights on 3 consecutive days. Sea-salt compositions from samples collected during the first (R24) and last (R26) flights are summarized in the lower part of Fig. 6. The temporal changes resemble those with increasing altitude; reacted sulfate-rich particles formed during the 26 hr between the flights. Particle aging is noticeable in the FT samples, with more sulfate particles in the second sample set than in the first. The longer the sea-salt particles reside in the air, the longer they are exposed to SO2 and the more likely they are to convert to sulfates. Mineral Dust. A significant fraction of the atmospheric particulate burden consists of mineral dust; its injection rate into the atmosphere can vary temporally and spatially, e.g., as a result of dust storms, volcanic eruptions, and anthropogenic activities. The mineral-dust burden tends to be especially high near source regions. Important examples include semiarid and arid lands, areas where land use is changing, and, in general, in the tropical and subtropical belts (72, 73). Moreover, changes in climate as well as land use can profoundly affect the amount of mineral dust that enters the troposphere. Drought and increased desertification by human activities can dramatically increase the dust available for deflation, and the tropospheric dust burden will increase appreciably. The relatively short atmospheric lifetime of much such dust means its radiative forcing adjusts relatively rapidly to changes in emissions. Mineral aerosols have a dual forcing role, producing both warming and cooling effects. The details are not well known because the scattering characteristics of mineral aerosols are hard to determine (74). Duce (75) estimates a direct forcing of −0.75 Wm−2, which is roughly equivalent to that from sulfate from biogenic gases (−0.68 Wm−2). It has been suggested that the direct radiative forcing of mineral aerosols approaches that of anthropogenic sulfates (73) and that in the tropical and subtropical North Atlantic region mineral dust is the dominant lightscattering aerosol (72). Tegen et al. (76, 77) estimate that disturbed soils contribute roughly half of the total atmospheric dust, and that the negative shortwave forcing and the positive IR forcing approximately cancel at the top of the atmosphere, but that internally the energy is redistributed, leading to climate change. Because of the lack of sufficient data, all models make many assumptions about the optical and radiative properties of mineral dust and are subject to revision as improved data are acquired.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3378

FIG. 7. TEM images of mineral dust collected from the marine troposphere. (a) Internal mixture of presumably terrestrial silicate and anhydrite with sea salt (Azores, North Atlantic, ASTEX/MAGE); (b) smectite (clay) and quartz (Q). The small grain size of the clay is visible at the thin edge (the arrows mark hexagonal platelets). Selected-area electrondiffraction patterns of clay and quartz are at the upper left and lower right, respectively. (Canary Islands, North Atlantic, ACE-2); (c) TEM image of goethite, FeO(OH), collected 2,600 m above sea level. Fe-bearing minerals like this could be important nutrient sources in remote oceans. (Canary Islands, North Atlantic, ACE-2.) Anthropogenically generated dust such as arises from transportation and industry can result in massive injections of minerals into the atmosphere. Between 30 and 50% of the soil dust burden may result from human activities (78). The resulting climatic effect can equal or even exceed that from aerosols generated by burning of fossil fuels and locally can be comparable to that from clouds (79). Mineral-dust particles provide important reactive surfaces for atmospheric reactions. Mineral aerosols significantly affect the cycles of N, S, and atmospheric oxidants (80). An annual average of 40% (locally reaching 70%) of total sulfate is associated with mineral dust over Asia, the western United States, Australia, and North Africa, and large parts of the oceans have over 10% of the sulfate associated with mineral aerosol (81). The association of sulfates with the larger mineral particles means that the local cooling effects of the sulfate aerosol are diminished because of a decrease in the incremental mass scattering efficiency of the sulfate. The mineral particles, being partly coated with hygroscopic sulfates—sometimes as the result of cloud processing—can also become CCN. Such processing of mineral-dust particles over the eastern Mediterranean has converted them into giant CCN that influence precipitation and the concentration of ice crystals in convective clouds (82). Our TEM studies of individual mineral-dust particles aim at identifying major mineral species in the aerosol and determining whether they are aggregated with other aerosol species. Such information is useful for obtaining improved refractive indices and size and shape data for model calculations of the radiative forcing of dust. The observation of internal or external mixtures is important for determining which chemical reactions take place on mineral surfaces and how the original particles change during atmospheric transport. The compositions and aggregations of the mineral aerosols collected when a polluted air mass of European origin arrived at our sampling site near the Azores Islands reflect their transport over the ocean. An unidentified silicate is associated with euhedral anhydrite, CaSO4, and sea salt (Fig. 7a). Smec-tite (Fig. 7b) is the most common mineral, and anhydrite is also widespread (42). Because the aerosol was transported in an air mass that contains high concentrations of SO2 from European pollution sources (83), it is likely that original carbonate particles reacted with SO2 to form anhydrite. Such conversions were observed in Asian (84) and Saharan (85) dust plumes. Under high-dust conditions, mineral particles provide an important pathway for SO2 removal from the atmosphere (81). Internal mixtures of sea salt, sulfate, and mineral particles were also observed by Andreae et al. (86), who concluded that such aggregates likely formed in clouds. An interesting consequence of the transport of mineral aerosol particles to the oceans is that they can serve as sources of biological productivity. Large areas of the ocean, ranging from the tropical equatorial Pacific to the polar Antarctic and the Southern Ocean, contain fewer phytoplankton/zooplankton than expected from the abundance of nutrients in the sea water. A correlation exists between a relative lack of Fe and this underproductivity (87, 88). Iron in the oceans far from land is mainly provided by continental dust (89, 90 and 91), which can be transported across large oceanic expanses (92, 93 and 94). We have identified goethite (Fig. 7c) and other Fe minerals from the aerosol in the FT. It appears as if atmospheric transport of mineral dust is widespread and can have major effects on life at the bottom of the food chain in the large areas of the oceans that are far from the nutrients provided by river flow from the continents. Nutrients derived from mineral dust may be limiting factors on primary productivity. We thank Dr. J. R. Anderson for help with sample collection, general assistance, and many useful discussions. Helpful reviews and comments were provided by Drs. J. R. Anderson, M. O. Andreae, J. M. Prospero, and S. E. Schwartz. This research was funded through grants from the National Science Foundation. Electron microscopy was performed by using instruments of the Arizona State University Facility for High Resolution Electron Microscopy. 1. Farquhar, G. D. (1997) Science 278, 1411. 2. Mahlman, J. D. (1997) Science 278, 1416–1417. 3. Molina, M. J., Molina, L. T. & Kolb, C. E. (1996) Annu. Rev. Phys. Chem. 47, 327–367. 4. Charlson, R. J. & Heintzenberg, J., eds. (1995) Aerosol Forcing of Climate (Wiley, New York). 5. Houghton, J. T., Meiro Filho, L. G., Callander, B. A., Harris, N., Kattenberg, A. & Maskell, K., eds. (1996) Climate Change 1995: The Science of Climate Change, Contribution of Working Group I to the Second Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) (Cambridge Univ. Press, Cambridge, U. K.). 6. National Research Council, Panel on Aerosol Radiative Forcing and Climate Change (1996) A Plan for a Research Program on Aerosol Radiative Forcing and Climate Change (Natl. Acad. Press, Washington, DC ). 7. Schwartz, S. E. & Andreae, M. O. (1996) Science 272, 1121–1122. 8. Michaels, R. A. (1997) Science 278, 1696. 9. Andreae, M. O. (1995) in Future Climates of the World: A Modeling Perspective, World Survey of Climatology, ed. Henderson-Sellers, A. (Elsevier, Amsterdam), Vol. 16, pp. 341–392. 10. Weber, R. J., Marti, J. J., McMurry, P. H., Eisele, F. L., Tanner, D. J. & Jefferson, A. (1997) J. Geophys. Res. 102, 4375–4385. 11. Hegg, D., Larson, T. & Yuen, P. F. (1993) J. Geophys. Res. 98, 18435–18439. 12. Schwartz, S. E. (1996) J. Aerosol Sci. 3, 359–382. 13. Ackerman, T. P. & Toon, O. B. (1981) Appl. Opt. 20, 3661–3668. 14. Chylek, P., Videen, G., Ngo, D., Pinnick, R. G. & Klett, J. D. (1995) J. Geophys. Res. 100, 16325–16332. 15. Pilinis, C., Pandis, S. N. & Seinfeld, J. H. (1995) J. Geophys. Res. 100, 18739–18754.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

AIRBORNE MINERALS AND RELATED AEROSOL PARTICLES: EFFECTS ON CLIMATE AND THE ENVIRONMENT

3379

16. Tang, I. (1997) J. Geophys. Res. 102, 1883–1893. 17. Haywood, J. M. & Shine, K. P. (1995) Geophys. Res. Lett. 22, 603–606. 18. Haywood, J. M., Roberts, D. L., Slingo, A., Edwards, J. M. & Shine, K. P. (1997) J. Clim. 10, 1562–1577. 19. Abelson, P. H. (1998) Science 281, 1609. 20. National Research Council, Committee on Research Priorities for Airborne Particulate Matter (1998) Research Priorities for Airborne Particulate Matter: I. Immediate Priorities and a Long-Range Research Portfolio (Natl. Acad. Press, Washington, DC). 21. Pszenny A., Keene, W., O'Dowd, C., Smith, M. & Quinn, P. (1998) Intl. Global Atmospheric Chem. (IGAC) IGACtivities NewsLetter 11, 6–12. 22. Ogren, J. A. (1995) in Aerosol Forcing of Climate, eds. Charlson, R. J. & Heintzenberg, J. (Wiley, New York), pp. 215–226. 23. Lacis, A. A. & Mishchenko, M. I. (1995) in Aerosol Forcing of Climate, eds. Charlson, R. J. & Heintzenberg, J. (Wiley, New York), pp. 11–42. 24. Haywood, J. M. & Ramaswamy, V. (1998) J. Geophys. Res. 103, 6043–6058. 25. Andreae, M. O. (1996) Nature (London) 380, 389–390. 26. Kaufman, Y. J. & Fraser, R. S. (1997) Science 277, 1636–1639. 27. Twomey, S. (1977a) Atmospheric Aerosols (Elsevier, New York). 28. Twomey, S. (1977b) J. Atmos. Sci. 34, 1149–1152. 29. Baker, M. B. (1997) Science 276, 1072–1078. 30. Germani, M. S. & Buseck, P. R. (1991) Anal. Chem. 63, 2232–2237. 31. Anderson, J. R., Buseck, P. R. & Saucy, D. A. (1992) Atmos. Environ. 26A, 1747–1762. 32. De Bock, L. A., Van Malderen, H. & Van Grieken, R. E. (1994) Environ. Sci. Technol. 28, 1513–1520. 33. Katrinak, K. A., Anderson, J. R. & Buseck, P. R. (1995) Environ. Sci. Technol. 29, 321–329. 34. Buseck, P. R. & Anderson, J. R. (1998) in Advanced Mineralogy 3, ed. Marfunin, A. S. (Springer, Berlin), pp. 292–300. 35. Anderson, J. R. & Buseck, P. R. (1998) in Advanced Mineralogy 3, ed. Marfunin, A. S. (Springer, Berlin), pp. 301–312. 36. Sheridan, P. (1989) Atmos. Environ. 23, 2371–2386. 37. Sheridan, P. J., Schnell, R. C., Kahl, J. D., Boatman, J. F. & Garvey, D. M. (1993) Atmos. Environ. 27A, 1169–1183. 38. Parungo, F., Kopcewicz, B., Nagamoto, C., Schnell, R., Sheridan, P., Zhu, C. & Harris, J. (1992) J. Geophys. Res. 97, 15867–15882. 39. McInnes, L. M., Covert, D. S., Quinn, P. K. & Germani, M. S. (1994) J. Geophys. Res. 99, 8257–8268. 40. Katrinak, K. A., Rez, P. & Buseck, P. R. (1992) Environ. Sci. Technol. 26, 1967–1976. 41. Pósfai, M., Anderson, J. R., Buseck, P. R., Shattuck, T. W. & Tindale, N. W. (1994) Atmos. Environ. 28, 1747–1756. 42. Pósfai, M., Anderson, J. R., Buseck, P. R. & Sievering, H. (1995) J. Geophys. Res. 100, 23063–23074. 43. Pósfai, M., Xu, H., Anderson, J. R. & Buseck, P. R. (1998a) Geophys. Res. Lett. 25, 1907–1910. 44. Pósfai, M., Anderson, J. R., Buseck, P. R. & Sievering, H. (1998b) J. Geophys. Res. 103, in review. 45. Buseck, P. R. (1992) in Minerals and Reactions at the Atomic Scale: Transmission Electron Microscopy, ed. Buseck, P. R. (Min. Soc. Am., Washington, DC), pp. 1–36. 46. Prather, K. A., Nordmeyer, T. & Salt, K. (1994) Anal. Chem. 66, 1403–1407. 47. Gard, E. E., Kleeman, M. J., Gross, D. S., Hughes, L. S., Allen, J. O., Morrical, B. D., Fergenson, D. P., Dienes, T., Gälli, M. E., Johnson, R. J., et al. (1998) Science 279, 1184–1187. 48. Murphy, D. M., Anderson, J. R., Quinn, P. K., McInnes, L. M., Brechtel, F. J., Kreidenweis, S. M., Middlebrook, A. M., Pósfai, M., Thomson, D. S. & Buseck, P. R. (1998) Nature (London) 392, 62–65. 49. Bates, T. S., Huebert, B. J., Gras, J. L., Griffiths, F. B. & Durkee, P. A. (1998) J. Geophys. Res. 103, 16297–16318. 50. Huebert, B. J., Pszenny, A. & Blomquist, B. (1996) J. Geophys. Res. 101, 4319–4329. 51. Charlson, R. J., Langner, J., Rodhe, H., Leovy, C. B. & Warren, S. G. (1991) Tellus 43AB, 152–163. 52. Rivera-Carpio, C. A., Corrigan, C. E., Novakov, T., Penner, J. E., Rogers, C. F. & Chow, J. C. (1996) J. Geophys. Res. 101, 19483–19493. 53. Charlson, R. J., Schwartz, S. E., Hales, J. M., Cess, R. D., Coakley, J. A., Jr., Hansen, J. E. & Hofmann, D. J. (1992) Science 255, 423–430. 54. Kiehl, J. T. & Briegleb, B. P. (1993) Science 260, 311–314. 55. Taylor, K. E. & Penner, J. E. (1994) Nature (London) 369, 734–737. 56. Hegg, D. A., Livingston, J., Hobbs, P. V., Novakov, T. & Russell, P. (1997) J. Geophys. Res. 102, 25293–25303. 57. Hobbs, P. V. (1998) Intl. Global Atmospheric Chem. (IGAC) IGACtivities NewsLetter 11, 3–5. 58. Novakov, T. & Penner, J. E. (1993) Nature (London) 365, 823–826. 59. McInnes, L. M., Bergin, M. H., Ogren, J. A. & Schwartz, S. E. (1998) Geophys. Res. Lett. 25, 513–516. 60. Okada, K. (1985) Atmos. Environ. 19, 743–757. 61. Mészáros, A. & Mészáros, E. (1989) Aerosol Sci. Tech. 10, 337–342. 62. Parungo, F., Nagamoto, C., Zhou, M.-Y., Hansen, A. D. A. & Harris, J. (1994) Atmos. Environ. 28, 3251–3260. 63. Andreae, M. O. (1983) Science 220, 1148–1151. 64. Cachier, H., Ducret, J., Brémond, M.-P., Yoboué, V., Lacaux, J.-P., Gaudichet, A. & Baudet, J. (1991) in Global Biomass Burning, ed. Levine, J. S. (MIT Press, Cambridge, MA), pp. 174–180. 65. Artaxo, A., Gerab, F., Yamasoe, M. A. & Martins, J. V. (1994) J. Geophys. Res. 99, 22857–22868. 66. Turn, S. Q., Jenkins, B. M., Chow, J. C., Pritchett, L. C., Campbell, D., Cahill, T. & Whalen, S. A. (1997) J. Geophys. Res. 102, 3683–3699. 67. Gill, P. S., Graedel, T. E. & Weschler, C. J. (1983) Rev. Geophys. Space Phys. 21, 903–920. 68. Andrews, E. & Larson, S. M. (1993) Environ. Sci. Technol. 27, 857–865. 69. Saxena, P., Hildemann, L. M., McMurry, P. H. & Seinfeld, J. H. (1995) J. Geophys. Res. 100, 18755–18770. 70. O'Dowd, C. D., Smith, M. H., Consterdine, I. E. & Lowe, J. A. (1997) Atmos. Environ. 31, 73–80. 71. Anderson, J. R., Buseck, P. R., Patterson, T. L. & Arimoto, R. (1996) Atmos. Environ. 30, 319–338. 72. Li, X., Maring, H., Savoie, D., Voss, K. & Prospero, J. M. (1996) Nature (London) 380, 416–419. 73. Bergametti, G. & Dulac, F. (1998) Intl. Global Atmospheric Chem. (IGAC) IGACtivities NewsLetter 11, 13–17. 74. Penner, J. E., Charlson, R. J., Hales, J. M., Laulainen, N. S., Leifer, R., Novakov, T., Ogren, J., Radke, L. F., Schwartz, S. E. & Travis, L. (1994) Bull. Am. Meteorol. Soc. 75, 375–400. 75. Duce, R. A. (1995) in Aerosol Forcing of Climate, eds. Charlson, R. J. & Heintzenberg, J. (Wiley, New York), pp. 43–72. 76. Tegen, I., Lacis, A. A. & Fung, I. (1996) Nature (London) 380, 419–422. 77. Tegen, I., Hollrig, P., Chin, M., Fung, I., Jacob, D. & Penner, J. (1997) J. Geophys. Res. 102, 23895–23915. 78. Tegen, I. & Fung, I. (1995) J. Geophys. Res. 100, 18707–18726. 79. Sokolik, I. N. & Toon, O. B. (1996) Nature (London) 381, 681–683. 80. Andreae, M. O. & Crutzen, P. J. (1997) Science 276, 1052–1058. 81. Dentener, F. J., Carmichael, G. R., Zang, Y., Lelieveld, J. & Crutzen, P. J. (1996) J. Geophys. Res. 101, 22869–22889. 82. Levin, Z., Ganor, E. & Gladstein, V. (1996) J. Appl. Meteorol. 35, 1511–1523. 83. Sievering, H., Gorman, E., Ley, T., Pszenny, A., Springer-Young, M., Boatman, J., Kim, Y., Nagamoto, C. & Wellman, D. (1995) J. Geophys. Res. 100, 23075–23081. 84. Winchester, J. W. & Wang, M.-X. (1989) Tellus 41B, 323–337. 85. Glaccum, R. A. & Prospero, J. M. (1980) Marine Geol. 37, 295–321. 86. Andreae, M. O., Charlson, R. J., Bruynseels, F., Storms, H., van Grieken, R. E. & Maenhaut, W. (1986) Science 232, 1620–1623. 87. Duce, R. A. (1986) in The Role of Air-Sea Exchange in Geochemical Cycling, ed. Buat-Menard, P. (D. Reidel Pub. Co., Dordrecht), pp. 497–529. 88. Martin, J. H. & Gordon, R. M. (1988) Deep-Sea Res. 35, 177–196. 89. Duce, R. A. & Tindale, N. W. (1991) Limnol. Oceanogr. 36, 1715–1726. 90. Martin, J. H., Gordon, R. M. & Fitzwater, S. E. (1991) Limnol. Oceanogr. 36, 1793–1802. 91. Martin, J. H., Coale, K. H., Johnson, K. S., Fitzwater, S. E., Gordon, R. M., Tanner, S. J., Hunter, C. N., Elrod, V. A., Nowicki, J. L., Coley, T. L., et al. (1994) Nature (London) 371, 123–129. 92. Arimoto, R., Duce, R. A., Savoie, D. L. & Prospero, J. M. (1992) J. Atmos. Chem. 14, 439–457. 93. Prospero, J. M., Uematsu, M. D. & Savoie, L. (1989) in Chemical Oceanography, eds. Riley, J. P. & Chester, R. (Academic, New York), Vol. 10, pp. 187–218. 94. Prospero, J. M. (1999) Proc. Natl. Acad. Sci. USA 96, 3396–3403. 95. Katrinak, K. A., Rez, P., Perkes, P. R. & Buseck, P. R. (1993) Environ. Sci. Technol. 27, 539–547.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3380

Oceanic minerals: Their origin, nature of their environment, and significance

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3380–3387, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. MIRIAM KASTNER* PNAS is available online at www.pnas.org.

Scripps Institution of Oceanography, University of California–San Diego, La Jolla, CA 92093-0212 ABSTRACT The chemical and isotopic compositions of oceanic biogenic and authigenic minerals contain invaluable information on the evolution of seawater, hence on the history of interaction between tectonics, climate, ocean circulation, and the evolution of life. Important advances and greater understanding of (a) key minor and trace element cycles with various residence times, (b) isotopic sources and sinks and fractionation behaviors, and (c) potential diagenetic problems, as well as developments in high-precision instrumentation, recently have been achieved. These advances provided new compelling evidence that neither gradualism nor uniformitarianism can explain many of the new important discoveries obtained from the chemistry and isotopic compositions of oceanic minerals. Presently, the best-developed geochemical proxies in biogenic carbonates are 18O/16O and Sr/Ca ratios (possibly Mg/ Ca) for temperature; 87Sr/86Sr for input sources, Cd/Ca and Ba/Ca ratios for phosphate and alkalinity concentrations, respectively, thus also for ocean circulation; 13C/12C for ocean productivity; B isotopes for seawater pH;, U, Th isotopes, and 14C for dating; and Sr and Mn concentrations for diagenesis. The oceanic authigenic minerals most widely used for chemical paleoceanography are barite, evaporite sulfates, and hydrogenous ferromanganese nodules. Marine barite is an effective alternative monitor of seawater 87Sr/86Sr, especially where carbonates are diagenetically altered or absent. It also provides a high-resolution record of seawater sulfate S isotopes, (evaporite sulfates only carry an episodic record), with new insights on factors affecting the S and C cycles and atmospheric oxygen. High-resolution studies of Sr, Nd, and Pb isotopes of well-dated ferromanganese nodules contain invaluable records on climate driven changes in oceanic circulation. Before the first global oceanographic HMS Challenger expedition (1873–1876) and the publication by Murray and Renard (1) on deep-sea sediments that was based on the expedition, little was known about oceanic minerals, their origin, distribution, or relevance. The first recovery of long cores on the Albatross Swedish expedition (1947–1949) revolutionized our understanding of the Pleistocene-Holocene ice ages recorded in oceanic minerals. The enormous potential of these minerals, especially of their chemical and isotopic tracers, containing information on the chemical history of seawater, thus on the record of interaction between tectonics, climate, ocean circulation, and evolution of life was realized. Since, great advances have been made in our understanding of the chemical and isotopic signatures of marine minerals, and therefore, of their applicability for unraveling Earth system's evolution and operation. The history of the rates and nature of interactions of physical, chemical, and biological processes in the ocean, which reflect the interplay between Earth's internal and external environmental processes, is recorded only in oceanic minerals. Although detrital minerals also are discussed, first-order invaluable information on Earth system processes is recorded and stored in oceanic minerals; hence, in this synthesis the emphasis is on biogenic and anthigenic minerals. Ocean-atmosphere climate models depend on such data. In addition, the role of biology in oceanic authigenic mineral formation is briefly evaluated. The oceanic minerals francolite (carbonate-F-apatite), hydrothermal Fe-Cu-Zn sulfides, and possibly Mn-Fe oxyhydroxides also may have economic significance. Characterizing active processes of formation and understanding modes of occurrence of these minerals in the modern ocean are necessary for assessing their potential economic importance and for guiding exploration for similar deposits in older geologic terrains. The potential influence of the large amount of clathrate hydrate of methane in continental margins on global warming and on its status as a future energy resource is discussed by Kvenvolden (124) in this issue of the Proceedings. Sediment Types, Sources, and Distribution in the Ocean The sediments in the ocean, which consist of three major components of detrital, biogenic, and authigenic origins, contain direct and/or indirect evidence of chemical and material inputs to the ocean and recycling within it. Even though each of the detrital minerals in the marine sedimentary record contains significant environmental information (Table 1), the focus of this paper is on the oceanic minerals. Because oceanic minerals record the history of the ocean-atmosphere system and hydrothermal mineralization strongly impacts seawater composition, the formation and occurrence of hydrothermal sulfide minerals is summarized as well. Detrital Minerals. Aluminosilicate minerals, ultimately derived from the weathering and erosion of rocks on land, comprise the bulk of detrital sediments. Detrital sediments are transported by water, wind, or ice. Their distribution in the ocean thus depends on climate and weathering at the source, the geographic disposition of rivers, existence of glaciers, and prevailing winds. They are most abundant in continental margins but occur over the entire seafloor. Detrital minerals and their spatial and temporal distribution contain extremely important information on climate and tectonics (refs. 2, 3, 4, 5 and 6 and references therein) such as the changing patterns and intensities of winds (refs. 7, 8 and 9 and references therein), the nature and intensity of weathering and erosion, and detritus transport by rivers and ice. A prime example are pulses of massive detritus transport by icebergs from northeast Canada into the Labrador Sea and North Atlantic Ocean during glacial retreat. These episodes are known as Heinrichs events (i.e., ref. 10). Fresh water from the melting of these icebergs disrupted the oceanic conveyor circulation and Heinrichs events correspond in time with important climate changes; for example, one makes the end of the last glacial cycle.

Abbreviation: SST, sea surface temperature. *To whom reprint requests should be addressed. e-mail: mkastner@ ucsd.edu.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3381

Volcanic ejecta comprise another type of detrital sediment carried by water and wind into the ocean. Volcanic sediments provide important information on periods of intense island arc and submarine volcanic activities, which can be compared with records of submarine tectonics and with climate and biological productivity records (11, 12). About 4–6 × 104 tons/year of cosmogonic particles survive the Earth's atmosphere and enter the sedimentary record. They include recognizable magnetic spherules (13), which have distinct geochemical signatures, in particular, extreme 3He concentrations and 3He/4He ratios (14, 15 and 16), and high Ir and Os concentrations and distinct Os isotope ratios (17, 18 and 19). They also provide crucial information on meteorite impacts and impact catastrophes (i.e., refs. 18 and 20) and on the history of cosmogonic bombardment, which may be one of the controls on climatic fluctuations (21). Biogenic Minerals. Dissolved weathering products constitute most of the salt in the sea. The major- and minor-element chemistry of biogenically produced minerals, most importantly calcite, aragonite, and opal-A (amorphous silica) reflect the geologic processes and rates that control the chemistry of seawater, ocean circulation, and biologic evolution. The less reactive, long residence time elements such as Cl, Na, K, Mg, Ca, and Sr do not vary throughout the ocean. Others vary with depth and between oceans because of biological cycling and scavenging by particles (22). Plants and organisms preferentially extract some elements, primarily C, N, P, Ca, and Si, from seawater to form soft tissues and minerals. Some of these elements are consumed by other organisms or redissolve, thus they are internally recycled. The rest reach the seafloor mostly as calcite and opal-A. At the seafloor, the rather segregated distribution of calcite, formed by coccolithophores (phytoplankton) and foraminifera (zooplankton), and of opal-A formed by diatoms (phytoplankton) and radiolaria (zooplankton), are important indicators of the history of productivity and ocean circulation. The trace element chemistry and isotopic compositions of these skeletal components are the most powerful tools available to unravel the effects of the interplay between tectonic and surficial processes on the chemical history of seawater (i.e., refs. 23, 24, 25, 26, 27, 28, 29, 30, 31 and 32 and references therein).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3382

In addition to calcite and opal-A, aragonite is an important indicator of chemical paleoceanography. Corals form aragonitic skeletons. Because of their seasonal growth bands and formation near the sea surface, the isotopic and trace element compositions of coral heads are important as recorders at high resolution (decadal or longer) of late Quaternary sea level fluctuations, sea surface paleo-temperatures, and rainfall data (i.e., refs. 33, 34, 35 and 36). Another important mineral is celestite, which forms biogenically by acantharians (zooplankton); however, its high solubility prevents its preservation as fossils. Despite the absence of celestite from the geologic record, its importance lies in the influence on recycling of Sr and Ba in the uppermost km of the ocean, with important implications for chemical paleoceanographic interpretations of Sr/Ca and Ba/Ca ratios in corals and planktonic foraminifera. Fish-teeth apatite is highly useful for both stratigraphy and chemical paleoceanography, especially in red clay deep-sea sediments where biogenic calcite and opal-A are absent because of dissolution (37, 38). Magnetite biomineralization first was identified in chiton teeth (39), as it hardens their surface. The geologically important magnetite consists of the morphologically distinct single domain crystals formed by magnetotactic bacteria (39, 40). This magnetite is responsible for much of the stable magnetic remanence in many marine sediments, which is valuable in paleomagnetic studies. Organic matter-rich sediments signify periods of higher productivity and/or higher organic C preservation in low-oxygen waters, either in a more intense and expanded oxygen minimum zone or in low-oxygen bottom waters (i.e., ref. 41). The assumed link between organic C preservation and water oxygen content has been challenged (i.e., ref. 42) and presently is being tested. Nevertheless, the stable carbon isotopic composition of organic matter in marine sediments helps to identify periods of high productivity or of high terrestrial organic matter input from enhanced continental weathering. This information is essential for modeling the oceanic C cycle and atmospheric CO2 and O2 fluctuations (i.e., refs. 22, 43, and 44 and references therein). Prime examples of periods with widely distributed organic C-rich sediments are (a) the midCretaceous abnormally organic C-rich “black” sediments in all major ocean basins, concentrated primarily in Aptian (115–110 Ma) and Cenomanian/Turonian (≈90 Ma) (i.e., refs. 45 and 46 and references therein), known as the Crateceous Anoxic Events (CAE), and (b) the late Neogene Mediterranean sapropels (i.e., ref. 47 and references therein). The mid-Cretaceous organic C burial during the CAE has significant implications for the ocean-atmosphere system C imbalance, as seen in the C isotopic composition of marine carbonates (48), hence on atmospheric O2 content. Such “black” shales are potential reservoir rocks. Worldwide, Cretaceous strata are known for being prolific hydrocarbon producers (i.e., refs. 49, 50 and 51). Authigenic Minerals. Authigenic minerals form by in situ inorganic precipitation on the seafloor and within the sediment column. Barite is the only mineral so far reported to also form in the water column (52, 53). Some authigenic mineral reactions are bacterially mediated; the bacteria modify the immediate geochemical environment, inducing mineral formation. For Paleo-environmental studies, the most important marine authigenic minerals are barite, francolite (a carbonate fluor apatite), evaporites, especially anhydrite/gypsum and halite, and Mn-Feoxyhydroxide that occur as nodules or crusts. The concentrations and isotopic compositions of oceanic conservative components in these authigenic minerals that form at the seafloor are extremely informative for studies of the history of seawater chemistry. Nonconservative chemical components also reveal information about distinct water masses, thus about ocean circulation in the geologic past. In addition to diagenetic opal-CT and quartz that form from the dissolution of biogenic opal-A, other common authigenic alumino-silicate minerals are smectites and zeolites. Dolomite and pyrite are also widespread authigenic minerals. Authigenic alumino-silicates from both close to the sediment-seawater interface and the sediment column mainly form by replacement of precursor minerals or mineraloids, but also can form by direct precipitation from pore fluids. On short time scales (decadal to millenial) the total amount of oceanic authigenic silicate remineralization seems to exert an extremely small impact on the global balance between CO2 degassing rate and atmospheric CO2 removal rate by continental weathering of silicates (43, 54, 55). Research on the contribution of silicate remineralization to the global CO2 budget in the modern ocean presently is being actively pursued (56). The important marine clay minerals are smectites. The iron smectite, nontronite, is widespread in the Pacific ocean floor where hydrothermal activity is prevalent and sedimentation rates are slow <1 cm/kyr. It forms at the seafloor from hydrothermal Fe oxyhydroxides and opal-A (57). In addition, nontronite and saponite replace precursor volcanic components, especially volcanic glass. Diagenetic illitic clays are celadonite and glauconite. Celadonite is associated primarily with volcanic matter alteration. Glauconite forms in subtropical and tropical margins in shallow to intermediate water depths, near the sediment-water interface beneath high productivity upwelling regions. These areas have intermediate to slow sedimentation rates, and both ferric and ferrous Fe are simultaneously available in pore waters. The zeolite Phillipsite is common at the seafloor or at shallow burial (58, 59). It replaces volcanic material in slow sedimentation environments such as the South Pacific deep sea. Clinoptilolite is also common and forms within the sediment column where silica concentrations are elevated to stabilize it, mostly from opal-A dissolution (59, 60). Occasionally clinoptilolite forms pseudo-morphs after radiolaria tests (60). Analcime is significantly less abundant and typically forms in volcanic ash-rich sediments (59). Authigenic calcite is widespread in the ocean. It forms mostly from biogenic calcite recyrstalization and is geochemically distinct in its minor and trace element concentrations from its precursor. This recrystallization process ultimately transforms calcareous ooze into chalk and limestone. Dolomite forms where pore-fluid seawater is modified bacterially or by physical-chemical processes. In continental margins where pore fluids become suboxic to anoxic, dolomite forms both as a primary precipitate and by replacing precursor Ca and Ca-Mg carbonates (61, 62). It also occurs in carbonate platforms where dolomite replaces CaCO3 by the thermally driven convective flow of seawater that is being inorganically and bacterially modified along the flow path. The chemical and isotopic compositions, primarily of carbon and oxygen, reveal the origins of the various dolomites. Rhodochrosite and siderite are uncommon marine minerals. Manganese and Fe mostly substitute for Mg in the dolomite structure in suboxic to anoxic pore fluid environments where they are mobilized. In these environments, only in the presence of much detrital Mn and Fe, rhodochrosite and/or siderite may form during late diagenesis where the reactions have driven Mg/Ca ratios in the pore fluids to levels below dolomite stability. Pyrite forms where sulfate is bacterially reduced. In these environments Fe+3 is reduced and mobilized, reacting with the sulfide to form pyrite. Because of the large isotopic fractionation by the bacterial reduction of sulfate to sulfide (≈40%) the amount of pyrite formation modulates the seawater-sulfate sulfur isotopic composition and the atmospheric oxygen content, as discussed below. Greigite, another Fe sulfide, may form kinetically instead of pyrite, especially in the presence of Fe excess or in S-deficient anoxic pore waters. The variability in pyrite versus greigite formation reflects the pH and Fe/S ratio of the pore fluid, which originally was modified by sulfate-reducing bacteria. Greigite has distinct magnetic properties. The most important economic authigenic mineral is francolite, which forms phosphorite deposits when tectonic-oceanographic conditions are favorable (63, 64). Similar to C, about 80% of the global P burial occurs in continental margins despite the much greater area of the deep ocean. Understanding the origin of

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3383

francolite and phosphorites, used principally as fertilizer in detergents and other industrial applications, is essential for effective exploration for new phosphorite deposits. Oceanic phosphorites form more than 80% of the world phosphate, whereas guano deposits are considerably less economically important. Based on recent research, francolite forms in continental margins at slow to intermediate sedimentation rates and more efficiently in suboxic to oxic bottom waters where Fe-redox cycling of P prevails (64, 65 and 66). Its occurrence in this environment is characteristic of high sea-level stands during interglacial periods, where it is enriched by winnowing and redeposition during glacial periods (67). Repeated formation and enrichment of this mineral is necessary to form economic deposits. Secular variations in the occurrence of large phosphorite deposits in the geologic record reflect shifts in global regional patterns of P deposition that are controlled by tectonics and climate. The chemical and isotopic signatures in phosphorite francolite that forms at or close to the sediment-seawater interface are those of the bottom water oceanic environment. Because of slow formation and sedimento-logical reworking and mixing, the chemical paleoceanographic information stored in phosphorite francolite, although extremely valuable, is of moderate to low resolution. Submarine Hydrothermal Sulfide Deposits. Submarine hydrothermal massive sulfide deposits forming at divergent plate boundaries have been discovered only in the past ≈20 years. These deposits primarily consist of the sulfides chalcopyrite, sphalerite and/or wurtzite, pyrrhotite, pyrite and/or marcasite, and the associated minerals anhydrite, barite, and opal-A. At the seafloor in oxidizing low-temperature environments hydrothermal sulfides are unstable, they are oxidized to form Fe oxides and secondary sulfides such as covellite, digenite, and bornite, except if rapidly covered by new volcanic eruptions (i.e., refs. 68 and 69). As yet, observations mainly of surface seafloor sites have been conducted, and the magnitude and structures of these sulfide deposits have not yet been thoroughly characterized. The discovery of massive sulfide deposits has had a profound impact on our understanding of ore-forming processes, but their economic significance is still uncertain. At fast- and slow-spreading centers, seawater penetrates several km into oceanic basement (maximum depth is above the top of the magma chamber), where it is heated to 350–400°C and interacts with oceanic basement and sediments. This fluid ascends by buoyancy to form hydrothermal vents and sulfide deposits. The discharging fluids support rich chemosynthetic, previously unknown biological communities. The suggestion that the deep-sea hydrothermal vents environment was possibly the most conducive environment for the origin of life on Earth is controversial. About 30% of the oceanic heat budget is removed by hydrothermal circulation (70, 71). Chemical and isotopic exchange reactions between the heated seawater and oceanic basement have a profound influence on seawater chemistry (72, 73), and thus also effect the mineralogy, geochemistry, and physical properties of the altering oceanic basement (74, 75). Hot and acidic, pH 3–4, discharging hydrothermal fluids are strongly enriched, not only in metal complexes and 3He, but also in Li, K, Rb, Ca, Ba, and Si, and have no Mg and sulfate (72). Overall the fluxes of these elements out and into the midocean ridges are comparable to those from river fluxes. Precipitation of minerals occurs when the hot acid, sulfide, and metal-rich fluids mix with cold oxidizing alkaline seawater. The entire ocean volume circulates through oceanic basement at spreading centers in 5-7 million years. Especially the common newly formed hydrous silicate minerals, smectite, ordered mixed layer smectite/illite, epidote, chlorite, serpentine, and amphiboles exert important controls on the chemical and physical behaviors of the oceanic plate when it is subsequently subducted. In particular, as a result of mineral dehydration reactions, these minerals determine the amount and timing (temperature) of fluids released in the seismogeneic zone in relation to earthquake cycles and arc volcanism (76, 77 and 78). In 1994, on Ocean Drilling Program Leg 158 drilling into an actively forming large sulfide mound in the Trans-Atlantic Geo-traverse (TAG) area of the Mid-Atlantic Ridge, at ≈26°N, a modern analog of fossil ophiolite-hosted deposits such as those exposed on Cyprus or Oman, revealed that this sulfide deposit has been forming for at least 20,000 years by successive hydrothermal episodes, including brecciation and cementation by anhydrite and quartz of previous precipitates. Brecciation of these deposits is induced by anhydrite dissolution between mineralization episodes. The widespread occurrence of anhydrite indicates seawater entraintment and its heating to >150°C. Upon cessation of the hydrothermal activity the anhydrite dissolves and does not survive in the geologic record. Pyrite is abundant in the upper flow region together with the anhydrite and with quartz. The hottest hydrothermal chimneys (black smokers) are composed of mostly chalcopyrite and anhydrite; the cooler ones (white smokers) are dominated by sphalerite. Estimates are that 2.7 million tons of massive sulfides exist above the seafloor and at least 1.2 million tons beneath it in the fluid upflow zones at TAG (79). This is comparable to an average-sized Cyprus-type ophiolite sulfide deposit (80). The basalt below is chloritized. In sedimented and back arc spreading centers, barite is a common cementing mineral together with anhydrite. Examples are the Guaymas Basin, Juan de Fuca Ridge, and Mariana Trough. Unlike anhydrite, barite is more likely to survive burial and resist diagenesis and dissolution. Examples of Recent Advances in Studies of Oceanic Minerals and Implications for Chemical Paleoceanography and Global Change Changes in ocean chemistry respond to climate-producing variations in the exogenic cycle and are recorded in oceanic minerals, but the record is far from simple and bulk chemical analyses are inadequate for paleoceanographic studies. Having well-dated materials is essential for placing geochemical records in a firm chronology. Important advances and greater understandings of (a) key minor and trace element cycles, (b) isotopic sources and sinks and fractionation behaviors especially of the light stable isotopes (with isotope masses <40), (c) the potential diagenetic problems of mineral diagenesis, and (d) important developments in high-resolution analytical instrumentation recently have been achieved. These advances have clearly reaffirmed that neither gradualism nor uniformitarinism can explain many of the recent important discoveries in paleoceanography. Evidence for Rapid Climate Shifts. Recent results obtained from the oxygen isotopes of foraminifera tests from deep-sea cores indicate that the evolutions of seawater chemistry, ocean circulation, biology, and climate are punctuated by periods of rapid (decadal to millennial) changes. Among the most well-known and geochemically clearly marked punctuating events are the rapid cooling and warming events that lasted a few thousand to a few hundred years during the longer Pleistocene-Holocene glacial-interglacial cycles, for example, the younger Dryas (i.e., refs. 81, 82, 83 and 84). Similar results, but at even higher resolution and frequencies, have been obtained from ice core studies (i.e., refs. 85, 86 and 87). A second prime example is the abrupt and major extinction at the Cretaceous-Tertiary boundary when a bolide struck Earth (i.e., refs. 18, 19 and 20). Oceanic minerals hold the key to understanding how the records left by such events and the resulting feedbacks might be used to gain knowledge about how Earth works. These event studies leave us better prepared to predict rapid global warming and may allow us to avoid surprises in climatic response to anthropogenic perturbations. The Mediterranean Salinity Crisis. The notable finding of a thick evaporative formation of halite and anhydrite/gypsum underlain and overlain by deep-sea sediments, in the deep Mediterranean Sea (and also the Red Sea), at >3 km water depths, indicates that the Mediterranean Sea almost completely

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3384

dried up in the late Miocene Messinian stage 6–5 Ma when the Gibraltar isthmus tectonically separated it from the Atlantic Ocean. When the strait of Gibraltar barrier broke at the end of the Messinian, the Mediterranean Sea was abruptly filled with Atlantic Ocean water (within ≈100 years) (i.e., ref. 88). Minor and Trace Elements in Biogenic Calcite and Aragonite and Their Significance. To understand the unique geochemical records stored in oceanic minerals it is necessary to develop an in-depth understanding of biogeochemical processes in the modern oceans and establish new and redundant proxies for the most significant oceanographic processes. To be a reliable monitor, a mineral must inherit the chemical and isotopic composition of seawater at the time of its formation, must retain this signal intact, and remain inert to chemical exchange after burial. Another prerequisite for an effective paleoceanographic proxy is that the chemical behavior and oceanic distribution of the elements are rather well known. The prime mineral used for paleoceanographic studies is biogenic calcite. Foraminifera shells in core tops are composed of low Mg-calcite (<1–4 mol% MgCO3) that contains minor and trace element concentrations, most importantly, Mg, Sr, Ba, Cd, (and possibly Na). These elements substitute for Ca. Li, B, REE, U are important trace elements in biogenic calcite that act as oceanographic environmental indicators. It has been shown that, similar to paleothermometry using the oxygen isotope ratios of these tests (e.g., ref. 23), the concentrations of these tracers in the foraminifera tests, although not representative of the thermodynamic distribution coefficients, generally reliably and consistently reflect the seawater composition and the biogeochemical conditions of formation. The biogeochemical controls on the incorporation of these minor and trace elements in the biogenic calcite and aragonite, however, are poorly known. Over the past several years, through difficult and persistent studies, it has been demonstrated that minor and trace elements in biogenic calcite can be used as proxies for a variety of important chemical oceanographic properties. The pioneering work by Boyle (89, 90) on cleaning procedures for studies of Cd in planktonic and benthonic foraminifera as a proxy for seawater phosphate, and for using Cd/Ca ratios in foraminifera shells for ocean circulation studies opened the field for new developments of trace element proxies of very low concentrations in the CaCO3 lattice. Improvements in culturing foraminifera, and especially in analytical techniques for accurate chemical and isotopic analyses of trace components and isotope ratios in small purified samples, such as the multicollector inductively coupled plasma mass spectrometry (MC-ICPMS) or the ion probe, provide great promise for documenting variations in seawater chemistry and sea surface temperature (SST) at very high resolution. The application of such techniques also should help to distinguish between physical-chemical versus biogeochemical controls on the trace element distributions in the calcites and aragonites. Presently, the best developed and most used geochemical proxies in foraminifera calcite shells are: • Oxygen isotope ratios and Sr/Ca ratios. Both are primarily important as temperature indicators; the latter also indicates seawater Sr concentrations. Oxygen isotopes, however, are influenced by both temperature and ice volume (i.e., refs. 24, 91, and 92). • Sr isotope compositions hold information about fluctuations in the inputs of Sr into the ocean, through time, by climate, tectonics, weathering processes, and hydrothermal activity. It is very useful for stratigraphic correlation and dating (i.e., refs. 93, 94 and 95 and references therein). • Cd/Ca and Ba/Ca ratios can be used as indicators of phosphate and alkalinity concentrations, respectively (i.e., refs. 31, 89, and 90). The C cycle is directly linked to nutrient concentrations and productivity, for example to phosphate concentration, therefore both Cd and Ba are important for paleoceanography (22). • Carbon stable isotope ratios are used to study ocean productivity (i.e., refs. 26 and 44). • The B isotope ratios of foraminifera (11B and 10B) reflect variations in seawater pH, an important parameter influencing atmospheric CO2 (i.e., refs. 32 and 96), because the boric acid to borate ratio is strongly pH dependent, and the two dissolved B species have distinct B isotopic signatures. • U/Th isotopes can be used for dating (i.e., refs. 97 and 98). • Sr, and possibly Mn, concentrations are excellent indicators of diagenesis (i.e., refs. 99 and 100). This is extremely important for monitoring the long-term integrity of mineral species in chemical oceanographic studies. • Because Mg readily substitutes for Ca in the calcite structures and is a major conservative component of seawater, the Mg/Ca ratio in foraminifera shells may be realized as a valuable palaeotemperature indicator. This possibility is being pursued at several laboratories (i.e., ref. 101). Strontium, Ba, and Mg can substitute for Ca in aragonite. The large Sr and Ba are more abundant in aragonite, and the smaller Mg is more abundant in calcite. As a result, seasonal banded corals that grow at the sea surface, composed of aragonite, show great promise for preserving records of ocean chemistry and temperature over a few centuries with a seasonally to biweekly resolution. Questions like the “exact” timing of the last glaciation or the causes and feedbacks of the very rapid climate changes observed in deep sea and ice cores are important for our understanding of the dynamics of the large ice sheets, their effect on ocean circulation, climate, and SST anomalies in the tropics. The latter is of a key parameter in paleoclimate analysis, thus, for understanding the causes of past rapid climate fluctuations such as those responsible for the El Niño Southern Ocean phenomenon. For climate studies, U-Th dated bands of coral species that live in the upper 5–20 m of the ocean are being used. For the SST record, in addition to oxygen isotope ratios, a measure of both SST and ice volume, the same coral bands also are analyzed for Sr/Ca and Mg/Ca ratios. These ratios are thought to provide more reliable SST data than oxygen isotopes. Corals from the equatorial Atlantic, Pacific and Indian oceans thus have been analyzed for these components (i.e., refs. 36, 83, and 102, 103, 104, 105 and 106). The Sr/Ca and Mg/Ca ratios in the aragonite structure are controlled by the activity of Sr, Mg, and Ca in seawater and the Sr/Ca and Mg/Ca distribution coefficients between aragonite and seawater that also depend on temperature. At present it is assumed that the biological controls on these distribution coefficients are minimal and that the Sr/Ca and Mg/Ca ratios in seawater are conservative. However, especially for high-resolution temperature data based on Sr/Ca ratios, the effect of the precipitation of acantharians SrSO4 shells in the upper water column on these ratios in corals needs to be better constrained. The most important result obtained from recent studies of the δ18O values and Sr/Ca (and Mg/Ca) ratios in tropical corals is that SST fluctuation between glacial and interglacial periods in the tropics were considerably larger, by 2–5°C, than previously recorded by using the oxygen isotopic ratios of foraminifera. The resulting implications for climate modeling are profound. Aragonite is unstable at surface seawater and transforms to calcite. Therefore the above-discussed coral studies are primarily applicable for Holocene corals. Even in Holocene corals aragonite may undergo diagenesis. As yet very little is known about the behavior of trace elements during diagenesis of coral aragonite. The absence of calcite is the only reliable test available that indicates that coral aragonite is pristine; this, however, is inadequate. To being able to extend in time the unique and high-resolution climatic records stored in coral aragonite, several laboratories have begun to pursue this problem. Marine Barite as a Monitor of Seawater Sr and S Isotope Compositions and Ocean Productivity. Marine barite is an ubiquitous minor phase in Pelagic sediments (52, 53), particularly

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3385

underlying regions of high biological productivity where it reaches concentrations in the sediment of over 2 weight percent (carbonate free). The exact mechanism of barite formation is not known, although there are indications that it precipitates inorganically directly from seawater, in micro-environments containing decaying organic matter, acantharian shells, and siliceous frustules and tests (i.e., ref. 53). This authigenic barite forms micro-crystals or aggregates ranging from 0.5 to 5 mm that is very different in habit from hydrothermal barite (Fig. 1). The water column is slightly undersaturated with respect to barite (107), therefore, some barite must dissolve at greater depths in the water column. This is observed in dissolved Ba concentration-depth profiles in the ocean. In most sediments, especially those beneath high productivity zones, porewaters rapidly become saturated with respect to barite; hence, most of this barite is preserved in oxic marine sediments (108, 109). Sr isotope analyses of marine barite and its crystal habit suggest no significant diagenetic alteration of this mineral occurs after burial in oxic pelagic sediments (109). The presence of morphologically and geochemically unaltered marine barite in calcareous sediments with clear indications of carbonate diagenesis (for example, Deep Sea Drilling Project Leg 85 Site 574, at >250 m burial depth), further supports this suggestion. Moreover, the 226Ra activities of marine barite indicate that the mineral behaves as a closed system in oxic environments (110). Thus, marine barite is a highly suitable mineral for a range of chemical paleoceanographic studies, for example, for seawater Sr and S isotopic compositions and paleo-productivity. Minor amounts of particulate barite also form in submarine hydrothermal systems (111). Barite dissolves in sulfate-reducing sediments and reprecipitates at the oxic-anoxic boundaries of such sediments. These hydrothermal and diagenetic barites may have Sr isotope ratios that are different from contemporaneous sea water; they therefore are unsuitable for chemical paleoceanographic studies, but they are relatively easily recognized (Fig. 1). As discussed earlier, the strontium isotope ratios in sea water are influenced by climate, tectonics, weathering, and hydrothermal activity at ocean ridges. The evolution of this ratio through time, determined primarily by measuring the strontium isotope composition of marine carbonates, holds information about variations in these processes. It is also useful for stratigraphic correlation and age dating. Carbonates, however, are absent from some marine sediments, such as siliceous oozes and red clays, and can be significantly diagenetically altered in others, especially in Eocene and older sediments. Like calcite marine barite is an effective monitor of seawater 87Sr/86Sr. The microcrystals of marine barite separated from Holocene Pacific, Atlantic, and Indian ocean sediments all record the modern seawater 87Sr/86Sr value, and the 87Sr/86Sr values of barite from sediment samples spanning the past 35 million years all fall within the range of published data for carbonates over this time period (109). Global changes in climate and atmospheric chemistry also are intimately related to the sedimentary sulfur cycle, and seawater sulfate is one of the main reservoirs of the sulfur exogenic cycle. Hence, knowledge of the chemical and isotopic composition of sulfate is of great importance for understanding coupled geo-chemical cycles on Earth, especially of C, S, O, and P. Evidence for large-scale transfers of S between different sedimentary reservoirs is provided by the evaporite-based isotope record for oceanic sulfate (112, 113). Because the geological record of marine evaporites is episodic, with gaps of tens of million years and evaporites are susceptible to diagenesis, a continuous seawater sulfate sulfur isotope curve for the Cenozoic recently has been generated at ≈1 million year resolution (114). The new curve fills significant gaps in the evaporite-based data set. A comparison between the new sulfate sulfur isotope data and the carbon isotope record of marine carbonates does not reveal the clear systematic coupling between the S and C cycles assumed in the existing major global geochemical cycles models. The new data indicate that changes in pyrite sulfur and organic carbons burial rates did not balance each other over a few to tens of millions of years. This has important implications for modeling of the atmospheric oxygen concentration over Phanerozoic time.

FIG. 1. Scanning electron micros copy micrographs of (A) marine barite separated from late Miocene (9.6 Ma) equatorial Pacific calcareous sediment, Deep Sea Drilling Project, Site 575. (B) Hydrothermal barite from a submarine sulfide chimney at Juan de Fuca Ridge. Ocean productivity influences organic carbon supply to the sediment and its burial efficiency, thus affecting the partitioning of CO2 between the ocean and atmosphere and influencing climate. Large glacial to interglacial fluctuations in atmospheric CO2 concentrations have been observed in ice cores and related to variations in ocean productivity (85, 87). To discern the coupling between ocean circulation, productivity, and climate, it is important to be able to estimate past ocean productivity and to reconstruct its history from the record of marine sediments. Upwelling of CO2-rich waters in the equatorial Pacific provides the largest natural source of CO2 to the atmosphere. As indicated above, because of its relative stability in the oxic pelagic environment marine barite accumulation rate in sediments is suitable for reconstructing changes in ocean biological productivity. Fluctuations in barite accumulation rates down-core (shown in Fig. 2) indicate that during glacial periods of the past 45 × 104 years, the productivity in the central and eastern equatorial Pacific was about two times that during intervening interglacial periods (115). This result is consistent with other evidence that productivity was high in the eastern and central equatorial Pacific during the last glacial. Ocean Dynamics and Climate Records in Hydrogenous Ferromanganese Crusts and Nodules. In the modern ocean, the thermohaline “conveyor” system circulation (in ≈1,500 years) controls the transfer of heat and solutes between the ocean basins and thus modulates climate. Dense cold surface seawater, formed in the North Atlantic, is carried throughout the ocean basins at depth and is replaced by northward transport of warmer water. Because circulation patterns most likely have been different in the past, the resulting global environmental, thus geochemical changes, are stored in biogenic and authigenic marine minerals. Hydrogenous Mn-Fe oxyhydroxide nodules and crusts effectively scavenge and incorporate geochemically important trace elements from seawater, including Be and Th-U, making them suitable for age dating. Recent high-resolution studies especially of Sr, Nd, and Pb isotopes (i.e., refs. 116, 117, 118 and 119), of well-dated samples by 10Be/9Be isotope ratios (i.e., refs. 120, 121 and 122) using a high-precision laser ablation, multicollector inductively coupled plasma mass spectrometry have successfully recorded important oceanic circulation changes. For example, the closure of the Central American isthmus, which changed an E-W ocean circulation to a N-S pattern, had clear geochemical consequences with evidence stored in Mn-crusts and nodules (123). Because of the slow growth of these nodules (1 to <10 mm/million year) high-precision sampling of them permits much higher resolution studies of paleoclimatic fluctuations than previously possible. Especially because of the very short residence time of Pb, the significant input variations in Pb isotopes in these minerals (i.e., refs. 117 and 122), on time scales similar to oceanic oxygen isotopes in sediments, provide important insights on the history of coupling between climate and input of Pb into the ocean, thus on continental erosion, atmospheric CO2, and climate-driven changes in ocean circulation through time.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3386

FIG. 2. (A) Accumulation rate of barite (AR BaSO4) and productivity in the equatorial Pacific in the past 45 × 104 years; data from two piston cores VNTR01–08PC at 110°W (○) and TT013–72PC at 140°W (●). Modified from figure 3 in ref. 115. The Role of Biology in Oceanic Authigenic Minerals Formation. Questions about the role of biology in the formation of oceanic authigenic minerals have been raised. As yet, there is no evidence for a direct biological involvement in the formation of any of the known oceanic authigenic minerals. Bacterial processes induce the formation of the oceanic authigenic minerals francolite, pyrite, and/or greigite, and of the carbonates dolomite, siderite, and rhodochrosite, by modifying their immediate geo-chemical environment. Biology also controls the preservation of bacterially produced magnetite. For example, in organic-rich sediments with bacterial reduction of Fe3+ to Fe2+, single domain magnetite crystals are unstable and the remanent magnetization of these sediments is not preserved. In continental margin sediments beneath high productivity regions with moderate sedimentation rates and burial of reactive organic matter, bacterial nitrate, Fe3+, and sulfate reduction occur, and reactive Fe3+ is bacterially reduced to Fe2+, setting in motion the Fe-redox cycling of P. This increases the availability of P for francolite and eventually promotes phosphorite formation, especially where bottom waters are oxic (64, 65 and 66). Fe-redox cycling occurs in the uppermost sediment column close to the sediment-seawater interface. Below this zone, sulfide, produced by bacterial sulfate reduction, combines with the bacterially mobilized ferrous ion, which precipitates mostly as pyrite, but occasionally first as greigite, because of kinetic constraints. The first appearance of greigite occurs especially in areas of excess Fe and/or deficient S in sediments. Pore water pH, modulated by sulfate-reducing bacteria, also controls the formation of greigite versus pyrite. The sulfur isotope ratios of these sulfides indicate the important role of bacterial mediation in their formation. In addition to P- and S-bearing minerals, biology also affects the formation of dolomite. Dolomite, although the thermodynamically stable carbonate mineral in seawater, does not form in the pelagic environment because the activation energy required for the formation of this ordered mineral at the prevailing low temperatures in the ocean is very high. It does, however, form in modified seawater with higher than normal alkalinity and/or Mg/Ca ratios that help to overcome the activation energy barrier. Dissolved sulfate kinetically inhibits the precipitation of carbonate minerals, especially of dolomite. Both its removal by bacterial sulfate reduction and the simultaneous alkalinity production promote dolomite formation (61, 62). As a result, dolomite forms inorganically in the marine environment in evaporative sub-basins where its formation is kinetically enhanced by high Mg/Ca ratios and elevated temperatures. It also forms by bacterial mediation in continental margins where most of the buried organic matter in the ocean occurs and pore fluid alkalinity is high. Consequently, on many margins authigenic calcite forms and pore fluid Mg/Ca ratios increase to between 6 to >20, compared with the seawater ratio of 5.4. Subsequently, at even higher alkalinity and lower sulfate concentrations dolomite forms instead of calcite. Dolomite also forms from precursor carbonates in carbonate platforms where thermally driven convective fluid flow drives reactions between modified pore fluids by the above mentioned bacterially mediated reactions. Bacterial sulfate reduction and alkalinity production is also widespread in evaporative sub-basins and is an important mechanism for dolomite formation in these environments. An important consequence of the vertical sequence of bacterial redox reactions on other authigenic carbonate mineral formation is the continuous mobilization of available reactive Mn2+ and Fe2+ by iron-reducing bacteria. This can cause the formation of rhodochrosite, ankerite, and/or siderite, respectively, instead of dolomite, in pore fluids with lowered Mg/Ca (<1–2) from earlier carbonate and silicate diagenetic reactions. Because of the high Mg content in seawater siderite is less common as an authigenic oceanic mineral than in continental environments. In addition to carbonates, biological involvement also has been invoked in marine barite formation (i.e., refs. 52 and 53), to explain its formation in the ocean water column, which is undersaturated with respect to barite solubility (107); oxidation of organic S to sulfate in organic-rich micro environments was proposed (i.e., refs. 52, 53, and 108), but the biological involvement in marine barite formation is yet to be demonstrated unequivocally.

1. Murray, J. & Renard, A. F. (1891) Challenger Reports (Longmans, London); reprinted (1965) (Johnson, London). 2. Lisitzin, A. P. (1972) Sedimentation in the World: Ocean (Soc. Econ. Paleont. Mineral., Tulsa, OK), Spec. Pub., 17. 3. Berger, W. H. (1973) in Chemical Oceanography, eds. Riley, J. P. & Chester, R. (Academic, London), pp. 266–387. 4. Hay, W. W. (1996) Geol. Rundsch. 85, 409–437. 5. Raymo, M. E. & Ruddiman, W. F. (1992) Nature (London) 359, 117–122. 6. Edmond, J. M. (1992) Science 258, 1594–1597. 7. Olivarez, A. M., Owen, R. M. & Rea, D. K. (1991) Geochim. Cosmochim. Acta 55, 2147–2158. 8. Prospero, J. M. (1996) in Particle Flux in the Ocean, eds.Ittekkott V., Honjo, S. & Depetris, P. J. (Wiley, New York), pp. 133–151. 9. Rea, D. K., Hilde, S. & Leah, H. J. (1998) Paleoceanography 13, 215–224. 10. MacAyeal, D. R. (1993) Paleoceanography 8, 767–773. 11. Larson, R. L. (1991) Geology 19, 547–550. 12. Officer, C. B. & Drake, C. L. (1985) Science 227, 1161–1167. 13. Murrel, M. T., Davis, P. A., Nishizumi, K. & Millard, H. T. (1980) Geochim. Cosmochim. Acta 44, 2067–2074. 14. Ozima, M., Takayanagi, M., Zashu, S. & Amari, S. (1984) Nature (London) 311, 448–450. 15. Anderson, D. L. (1993) Science 261, 170–176. 16. Farley, K. A. (1995) Nature (London) 376, 153–156. 17. Barker, J. L. & Anders, E. (1968) Geochim. Cosmochim. Acta 32, 627–645. 18. Alvarez, L. W., Alvarez, W., Asaro, F. & Michel, H. V. (1980) Science 208, 1095–1108. 19. Luck, J. M. & Turekian, K. K. (1983) Science 222, 613–615. 20. Esser, B. K. & Turekian, K. K. (1988) Geochim. Cosmochim. Acta 52, 1383–1388. 21. Muller, R. A. & MacDonald, G. J. (1995) Nature (London) 377, 107–108. 22. Broecker, W. S. & Peng, T.-H. (1982) Tracers in the Sea (Eldigio, Palisades, NY). 23. Epstein, S., Buchsbaum, R., Lowenstam, H. A. & Urey, H. C. (1953) Geol. Soc. Amer. Bull. 64, 1315–1325. 24. Emiliani, C. (1978) Earth Planet. Sci. Lett. 37, 349–352. 25. Savin, S. M., Douglas, R. G. & Stehli, F. G. (1975) Geol. Soc. Amer. Bull. 86, 1499–1510. 26. Shackleton, N. J. & Opdyke, N. D. (1977) Nature (London) 270, 216–219. 27. Broecker, W. S. (1982) Geochim. Cosmochim. Acta 46, 1689–1705. 28. Graham, D. W., Bender, M. L., Williams, D. F. & Keigwin, L. D. (1982) Geochim. Cosmochim. Acta 456, 1281–1292. 29. Burke, W. H., Denison, R. E., Hetherington, E. A., Koepnick, R. B., Nelson, H. F. & Otto, J. B. et al. (1982) Geology 10, 516–519. 30. Boyle, E. A. (1988) Paleoceanography 3, 471–489. 31. Lea, D. W. & Boyle, E. A. (1989) Nature (London) 338, 751–753. 32. Spivack, A. J., You, C.-F. & Smith, H. J. (1993) Nature (London) 363, 149–151. 33. Emiliani, C., Hudson, J. H., Shinn, E. A. & George, R. Y. (1978) Science 202, 627–629. 34. Fairbanks, R. G. (1989) Nature (London) 342, 637–642. 35. Wellington, G. M. & Dunbar, R. B. (1995) Coral Reefs 14, 5–25.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

OCEANIC MINERALS: THEIR ORIGIN, NATURE OF THEIR ENVIRONMENT, AND SIGNIFICANCE

3387

36. De Villiers S., Nelson, B. K. & Chivas, A. R. (1995) Science 269, 1247–1249. 37. Doyle, P. S. & Riedel, W. R. (1979) Micropaleontology 25, 337–364. 38. Ingram, B. L., Coccioni, R., Montanari, A. & Richter, F. M. (1994) Science 264, 546–550. 39. Lowenstam, H. A. & Weiner, S. (1989) On Biomineralization (Oxford Univ. Press, Oxford). 40. Chang, S.-B. R. & Kirschvink, J. L. (1989) Annu. Rev. Earth Planet. Sci. 17, 169–195. 41. Emerson, S. & Hedges, J. I. (1988) Paleoceanography 3, 621–634. 42. Calvert, S. E. & Pedersen, T. F. (1992) in Productivity, Accumulation and Preservation of Organic Matter in Recent and Ancient Sediments, eds. Whelan, J. K. & Farrington, J. W. (Columbia Univ. Press, New York), pp. 231–263. 43. Sundquist, E. T. (1991) Q. Sci. Rev. 10, 283–296. 44. Shackleton, N. J. (1987) in Marine Petroleum Source Rocks, eds. Brooks, J. & Fleet, A. J. (Geol. Soc., London), p. 423–434. 45. Schlanger, S. O. & Cita, M. B., eds. (1982) Nature and Origin of Cretaceous Carbon-Rich Facies (Academic, London). 46. Jenkyns, H. C. (1980) J. Geol. Soc. (London) 137, 171–188. 47. Calvert, S. E. & Pedersen, T. F. (1993) Mar. Geol. 113, 67–88. 48. Scholle, P. A. & Arthur, M. A. (1980) Am. Assoc. Petrol. Geol. Bull. 64, 67–87. 49. Arthur, M. A. & Schlanger, J. O. (1979) Am. Assoc. Petrol. Geol. Bull. 63, 870–885. 50. Demaison, G. J. & Moore, G. T. (1980) Am. Assoc. Petrol. Geol. Bull. 64, 1179–1209. 51. Tissot, B. Demaison, G., Masson, P., Delteil, J. R. & Combaz, A. (1980) Am. Assoc. Petrol Geol. Bull. 64, 2051–2063. 52. Dehairs, F., Chesselet, R. & Jewab, J. (1980) Earth Planet. Sci. Lett. 49, 528–550. 53. Bishop, J. K. B. (1988) Nature (London) 332, 341–343. 54. Berner, R. A., Lasaga, A. C. & Garrels, R. M. (1983) Am. J. Sci. 283, 641–683. 55. Berner, R. A. & Caldeira, K. (1997) Geology 25, 955–956. 56. Michalopoulos, P. & Aller, R. C. (1995) Science 270, 614–617. 57. Alt, J. F. (1988) Mar. Geol. 81, 227–239. 58. Bernat, M., Bieri, R. H., Koide, M., Griffin, J. J. & Goldberg, E. D. (1970) Geochim. Cosmochim. Acta 34, 1053–1071. 59. Kastner, M. & Stonecipher, S. A. (1978) in Natural Zeolites, Occurrence, Properties, Use, eds. Sand, L. B. & Mumpton, F. A. (Pergamon, Oxford), pp. 199–220. 60. Kastner, M., (1981) in The Ocean Lithosphere: The Sea, ed. Emiliani, C. (Wiley, New York), Vol. 7, pp. 915–980. 61. Baker, P. A. & Kastner, M. (1981) Science 213, 214–216. 62. Burns, S. J. & Baker, P. A. (1987) J. Sed. Petrol. 57, 128–137. 63. Bentor, Y. K. (1980) Soc. Econ. Paleontol. Mineral Spec. Publ., 3–18. 64. Schuffert, J. D., Kastner, M. & Jahnke, R. A. (1998) Mar. Geol. 146, 21–31. 65. Ingall, E. & Jahnke, R. (1994) Geochim. Cosmochim. Acta 58, 2571–2575. 66. Van Cappellen, P. & Ingall, E. D. (1996) Science 271, 493–496. 67. Garrison, R. E., Kastner, M. & Kolodny, Y. (1987) in Cenozoic Basin Development of Coastal California, eds. Ingersoll, R. V. & Ernst, W. G. (Prentice– Hall, Englewood Cliffs, NJ), Vol. VI, pp. 348–381. 68. Haymon, R. M. & Kastner, M. (1981) Earth Planet. Sci. Lett. 53, 363–381. 69. Hannington, M. D., Jonnasson, I. R., Herzig, P. M. & Petersen, S. (1995) Geophys. Monogr. 91, 115–157. 70. Sclater, J. G., Parsons, B. & Jaupart, C. (1981) J. Geophys. Res. 86, 11522–11535. 71. Stein, C. A. & Stein, S. (1994) J. Geophys. Res. 99, 3081–3096. 72. Edmond, J. M., Measures, C., McDuff, R. E., Chan, L.-H., Collier, R., Grant, B., Gordon, L. I. & Corliss, J. B. (1979) Earth Planet. Sci. Lett. 46, 1–18. 73. East Pacific Rise Study Group (1981) Science 213, 31–40. 74. Muehlenbacks, K. & Clayton, R. N. (1972) Can. J. Earth. Sci. 9, 172–184. 75. Humphris, S. E. & Thompson, G. (1978) Geochim. Cosmochim. Acta 42, 107–125. 76. Zindler, A. & Hart, S. R. (1986) Annu. Rev. Earth Planet. Sci. 14, 115–118. 77. Tatsumi, Y. (1989) J. Geophys. Res. 94, 4697–4707. 78. Alt, J. C., Teagle, D. A., Brewer, T., Shanks, W. C. & Halliday, A. (1998) J.Geophys. Res. 103, 12365–12380. 79. Humphris, S. E., Herzig, P. M., Miller, D. J., Alt, J. C., Becker, K., Brown, D., Brügmann, G., Chiba, H., Fouquet, Y., Gemmell, J. B., et al. (1995) Nature (London) 377, 713–716. 80. Richards, J. G., Cann, J. R. & Jensenius, J. (1989) Econ. Geol. 84, 91–115. 81. Dansgaard, W., White, J. W. C. & Johnsen, S. J. (1989) Nature (London) 339, 532–534. 82. Broecker, W. S. (1992) Q. Res. 38, 135–138. 83. Fairbanks, R. G. (1990) Paleoceanography 5, 937–948. 84. Bard, B., Hamelin, B., Arnold, M., Montaggioni, L., Gabioch, G., Faure, G. & Rougerie, F. (1996) Nature (London) 382, 241–244. 85. Neftel, A., Oeschger, H., Schwander, J., Stauffer, B. & Zumbrunn, R. (1982) Nature (London) 295, 220–223. 86. Brook, E. J., Sowers, T. & Orchardo, J. (1996) Science 273, 1087–1093. 87. Severinghaus, J. P., Sowers, T., Brook, E. J., Alley, R. B. & Bender, M. L. (1998) Nature (London) 391, 141–146. 88. Hsü, K. J., Montadert, L., Bernoulli, D., Cita, M. B., Erickson, A., Garrison, R. E., Kidd, R. B. Mèlierés, F., Müller, C. & Wright, R. (1977) Nature (London) 267, 399–403. 89. Boyle, E. A. (1988) Paleoceanography 3, 471–489. 90. Boyle, E. A. (1992) Annu. Rev. Earth Planet. Sci. 20, 245–287. 91. Graham, D. W., Bender, M. L., Williams, D. F. & Keigwin, L. D. (1982) Geochim. Cosmochim. Acta 46, 1281–1292. 92. Delaney, M. L., Bé, A. W. H. & Boyle, E. A. (1985) Geochim. Cosmochim. Acta 49, 1327–1341. 93. Elderfield, H. (1986) Paleogeog. Paleoclimatol. Paleoecol. 57, 71–90. 94. DePaolo, D. J. & Ingram, B. L. (1985) Science 227, 938–941. 95. Hodell, D. A., Mueller, P. A., McKenzie, J. A. & Mead, G. A. (1989) Earth Planet. Sci. Lett. 92, 165–178. 96. Sanyal, A., Hemming, N. G., Hanson, G. N. & Broecker, W. S. (1995) Nature (London) 373, 234–236. 97. Edwards, R. L., Chen, J. H. & Wasserburg, G. J. (1986) Earth Planet. Sci. Lett. 81, 175–192. 98. Slowley, N. C., Henderson, G. M. & Curry, W. B. (1996) Nature (London) 383, 242–244. 99. Baker, P. A., Gieskes, J. M. & Elderfield, H. (1982) J. Sed. Petrol. 52, 71–82. 100. Boyle, E. A. (1983) Geochim. Cosmochim. Acta 47, 1815–1819. 101. Nurnberg, D., Binja, J. & Hemleben, C. (1996) Geochim. Cosmochim. Acta 60, 803–814. 102. Hart, S. R. & Cohen, A. L. (1996) Geochim. Cosmochim. Acta 60, 3075–3084. 103. Shen, G. T., Cole, J. E., Lea, D. W., Linn, L. J., McConnaughey, T. A. & Fairbanks, R. G. (1992) Paleoceanography 7, 563–588. 104. Beck, J. W., Edwards, R. L., Ito, E., Taylor, F. W., Recy, J., Rougerie, F., Joannot, P. & Henin, C. (1992) Science 257, 644–647. 105. McCulloch, M., Mortimer, G., Esat, T., Xianhua, L., Pillans, B. & Chappel, J. (1996) Earth Planet. Sci. Lett. 138, 169–178. 106. Mitsuguchi, T., Matsumoto, E., Abe, O., Uchida, T. & Isdale, P. J. (1996) Science 274, 961–963. 107. Church, T. M. & Wolgemuth, K. (1972) Earth Planet. Sci. Lett. 15, 35–44. 108. Dymond, J., Suess, E. & Lyle, M. (1992) Paleoceanography 7, 163–181. 109. Paytan, A., Kastner, M. Martin, E. E., Macdougall, J. D. & Herbert, T. (1993) Nature (London) 366, 445–449. 110. Paytan, A., Moore, W. & Kastner, M. (1996) Geochim. Cosmochim. Acta 60, 4313–4319. 111. Feely, R. A. (1987) J. Geophys. Res. 92, 11343–11363. 112. Holser, W. T. & Kaplan, I. R. (1966) Chem. Geol. 1, 93–135. 113. Claypool, G. E., Holser, W. T., Kaplan, I. R., Sakai, H. & Zak, I. (1980) Chem. Geol. 28, 199–260. 114. Paytan, A., Kastner, M., Campbell, D. & Thiemens, M. H. (1998) Science 282, 1459–1462. 115. Paytan, A., Kastner, M. & Chavez, F. D. (1996) Science 274, 1355–1357. 116. Albarede, F. & Goldstein, S. L. (1992) Geology 20, 761–763. 117. Christensen, J. N., Halliday, A. N., Godfrey, L. V., Hein, J. R. & Rea, D. K. (1997) Science 277, 913–918. 118. Abouchami, W., Goldstein, S. L., Galer, S. J. G., Eisenhauer, A. & Mangini, A. (1997) Geochim. Cosmochim. Acta 61, 3957–3974. 119. O'Nions, R. K., Frank, M. von Blanckenburg, F. & Ling H.-F. (1998) Earth Planet. Sci. Lett. 155, 15–28. 120. von Blanckenburg, F., O'Nions, R. K. & Hein, J. R. (1996) Geochim. Cosmochim. Acta 60, 4957–4963. 121. Ling, H.-F., Burton, K. W., O'Nions, R. K., Kamber, B. S., von Blanckenburg, F., Gibbs, A. J. & Hein, J. R. (1997) Earth Planet. Sci. Lett. 146, 1–12. 122. Frank, M. & O'Nions, R. K. (1998) Earth Planet. Sci. Lett. 158, 121–128. 123. Burton, K. W., Ling, H.-F. & O'Nions, R. K. (1997) Nature (London) 386, 382–385. 124. Kvenvolden, K. A. (1999) Proc. Natl. Acad. Sci. USA 96, 3420–3426.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3388

Mineral surfaces and bioavailability of heavy metals: A molecularscale perspective

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3388–3395, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. GORDON E. BROWN, JR.*†‡, ANDREA L. FOSTER*, AND JOHN D. OSTERGREN* PNAS is available online at www.pnas.org.

ABSTRACT There is a continual influx of heavy metal contaminants and pollutants into the biosphere from both natural and anthropogenic sources. A complex variety of abiotic and biotic processes affects their speciation and distribution, including adsorption onto and desorption from mineral surfaces, incorporation in precipitates or coprecipitates, release through the dissolution of minerals, and interactions with plants and microbes. Some of these processes can effectively isolate heavy metals from the biosphere, whereas others cause their release or transformation to different species that may be more (or less) bioavailable and/or toxic to organisms. Here we focus on abiotic adsorption and precipitation or coprecipitation processes involving the common heavy metal contaminant lead and the metalloids arsenic and selenium in mine tailings and contaminated soils. We have used extremely intense x-rays from synchrotron sources and a structure-sensitive method known as x-ray absorption fine structure (XAFS) spectroscopy to determine the molecularlevel speciation of these elements at concentrations of 50 to several thousand ppm in the contaminated environmental samples as well as in synthetic sorption samples. Our XAFS studies of As and Pb in the mine tailings show that up to 50% of these contaminants in the samples studied may be present as adsorbed species on mineral surfaces, which makes them potentially more bioavailable than when present in sparingly soluble solid phases. Our XAFS studies of Se(VI) sorption on Fe2+-containing sulfates show that this element undergoes redox reactions that transform it into less bioavailable and less toxic species. This type of information on molecular-level speciation of heavy metal and metalloid contaminants in various environmental settings is needed to prioritize remediation efforts and to assess their potential hazard to humans and other organisms. The earth's surface and near surface regions are dominated by interfaces among solids, liquids, and gases. Such interfaces, particularly those between natural solids and aqueous solutions, play an enormously important role in a number of geological and geochemical processes. For example, “water–rock” interactions over geologic time have been major contributors to both the rock cycle and the geochemical cycling of elements. In an environmental context, interfacial processes such as mineral dissolution, mineral precipitation, and the sorption and desorption of chemical species are responsible for the release and/or sequestration of heavy metals that may eventually become pollutants in soils and groundwater. The importance of interfacial processes is summed up well in the following quotation from Werner Stumm (1): “Almost all of the problems associated with understanding the processes that control the composition of our environment concern interfaces, above all the interfaces of water with naturally occurring solids.” Recognition of the importance of surface chemical reactions in geochemical and environmental contexts can be traced back to some of the pioneering work on soil–fluid interactions, particularly the papers by H. S. Thompson (2) and J. Thomas Way (3), who were early soil chemists. The paper by Way reported on the filtration of “liquid manure” through a loamy soil, which resulted in the manure being “deprived of color and smell.” Almost 150 years later, we have a more fundamental understanding of the sorption and cation exchange phenomena that account for Way's observation. We are also beginning to develop a molecular-scale understanding of the complexity and interplay of the chemical and biological processes that control element cycling in soils and sediments, some of which are shown in Fig. 1. Such processes range from dissolution of mineral particles in soils, which can release natural contaminants into pore waters, to the binding or sorption of metals (M) and organic ligands (L) to mineral surfaces, which can effectively immobilize contaminants and reduce their bioavailability. Precipitation is another common means of sequestering a heavy metal if the precipitated phase is relatively insoluble. Some heavy metal contaminants such as lead normally exist in minerals in one dominant oxidation state, whereas others such as arsenic and selenium can exist in several oxidation states and can undergo oxidation or reduction when they interact with mineral surfaces or organic compounds, which act as oxidants or reductants. Microorganisms and plants can have a profound influence on chemical reactions involving contaminants. For example, microorganisms often play a major role in the degradation of organic contaminants and in the oxidation and reduction of heavy metals. In the case of plants, the root–soil interfacial region, referred to as the rhizosphere (circled area in soil profile in Fig. 1) is an area of particularly intense chemical and biological activity where organic acids, sugars, and other organic compounds are exuded by live plant roots. The pH can be as much as 2 units lower, and microbial counts can be 10- to 50-fold higher at the root surface than in the bulk soil a few millimeters away. Thus, mineral weathering rates and the solubility of mineral elements and anthropogenic contaminants are generally greater in surface soils, where plant and microbial activity are higher than in deeper parts of the soil and geologic column. Because of this complexity, experimental and theoretical studies that probe the nature of these processes at a fundamental level are difficult to perform on natural samples, and the results are sometimes difficult to interpret because of the large number of interacting abiotic and biotic processes. In fact, it is often difficult to distinguish between abiotic and biotic transformation mechanisms involving heavy metals and metalloids, as both can affect the pathways of reactions in a synergistic fashion (see, e.g., ref. 4). One approach to this problem involves analytical and experimental studies of simplified model systems in which variables can be controlled to simulate processes in more complex natural systems. It is also essential to carry out parallel studies of the natural systems to place constraints on the variables and types of processes that control contaminant speciation and distribution and to develop testable hypotheses that can be addressed by appropriately chosen model systems. We have adopted this combined model system/natural system approach to help interpret analytical results on natural systems containing heavy metal and metalloid contaminants.

*Department of Geological and Environmental Sciences, Stanford University, Stanford, CA 94305-2115, and †Stanford Synchrotron Radiation Laboratory, Stanford Linear Accelerator Center, P.O. Box 4349, Stanford, CA 94309 ‡To whom reprint requests should be addressed. e-mail: [email protected].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3389

FIG. 1. Schematic illustration of a variety of molecular environmental science processes affecting contaminant elements in soils and groundwater. Ultimately, the impact of heavy metals and other environmental contaminants on humans and other organisms depends on their concentration levels, toxicity, and bioavailability, i.e., the extent to which they are absorbed by the blood or stored in internal organs. The toxicity and bioavailability of a heavy metal, in turn, depend in part on reactivity and solubility, which are determined by the speciation or chemical form of the element. The term “speciation,” as used here, refers to (i) the identity of the element, (ii) its oxidation state, (iii) its physical state [i.e., phase association; presence in a liquid, gaseous, or solid phase (amorphous or crystalline), colloidal particle, animal or plant cell, or biofilm; presence as a surface coating or thin film on a solid, as a sorption complex (monomeric or polymeric) on a solid, colloidal particle, or an organic substance; etc.], (iv) its empirical formula, and (v) its detailed molecular structure. For example, chromium in the Cr(VI) form occurs as a tetrahedral oxoanion in aqueous solution, is quite soluble and mobile in groundwater and surface waters, and is toxic to organisms (5, 6). In contrast, when in the Cr(III) form, chromium is typically bonded to six oxygens in an octahedral complex, is relatively insoluble and immobile, and poses little risk to organisms (5). As shown by this example, a fundamental understanding of the potential impact of a particular environmental contaminant on humans must include a molecular-level description of its speciation, as well as a number of other interrelated factors (7). This point is illustrated by the blood lead levels measured in humans in several sites, where lead levels in surface soils range from hundreds to thousands of ppm (Fig. 2). These data (8) show that significantly different levels of blood lead are found in human populations at the different sites and do not correlate directly with bulk Pb soil concentrations. One possible explanation for these differences is the presence of different species of lead at these sites, with lead at some sites being in a more bioavailable form. When present in crystalline solids such as galena or pyromorphite, lead should be less soluble, less mobile, and less bioavailable than when present as sorbed species on mineral surfaces, where it may be relatively easily removed into solution by a reduction in pH from neutral to acidic, such as might occur when a lead-contaminated soil sample is ingested by an organism (see, e.g., ref. 9). These differences are well illustrated by the results of in vivo swine studies using soil and mine waste material from Leadville, CO, which have shown that Pb bioavailability ranges from <5% for selected tailings materials, which are dominated by galena (PbS), to 45% for surface soils where the majority of Pb occurs as “Fe-Mn-Pb oxide” phase (s) (10).

FIG. 2. Blood lead levels in humans in several major lead mining districts (modified after ref. 12). Although the correlation between speciation of environmental contaminants and their bioavailability has been shown in recent studies (11, 12, 13, 14, 15, 16 and 17), there is relatively limited use of information on molecular-level speciation in setting maximum contaminant levels (see, e.g., ref. 18). For example, the current EPA limit for arsenic in drinking water is 50 ppb total arsenic without designation of the species of As present, which can occur as inorganic oxoanions [As(III)O33− and As(V)O43−], as organic arsenicals, or in other forms; the trivalent form of As, either in inorganic or organic compounds, is generally more toxic than pentavalent compounds of As (19). In the past, cleanup efforts for contaminated soils or mine tailings were often driven by total contaminant element concentration without sufficient attention to speciation or the relative toxicities or bioavailabilities of individual

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3390

species of a contaminant element (18). This practice is due in large part to the difficulty in deriving molecular speciation information for low concentration levels of a given contaminant and in applying this information to field-scale situations. Potential consequences of this approach are remediation efforts and major expenditures that may not be warranted because of the presence of a heavy metal or metalloid in chemical forms that are not readily bioavailable. In this paper, we summarize recent spectroscopic work on the molecular-scale speciation of arsenic, selenium, and lead in contaminated soils and mining wastes with the objective of relating speciation to potential bioavailability. We also briefly discuss some of the natural sources of these elements as well as general concepts needed to understand variations in their uptake or sorption on mineral surfaces in contact with bulk aqueous solutions. These heavy metals (Pb) and metalloids (As, Se) are among the most important environmental contaminants, affecting millions of people. For example, arsenic pollution in drinking water in Bangladesh is currently causing a health risk to over 70 million people (20). Arsenic also poses a potential risk in the Mother Lode District of California, where tailings from gold mining activities of the 1800s contain arsenic levels as high as 5,000 ppm (21, 22). Similarly, selenium contamination in the Central Valley of California threatens wildfowl populations as well as agricultural production in this region where Se concentrations in topsoils and efflorescences are as high as 50 ppm (23). Lead is arguably the most widespread elemental contaminant worldwide and has had a major impact on human welfare (24, 25). Lead has many diverse physiological and biochemical effects in animals and humans, particularly children, where it has been shown to affect intelligence quotients in children exposed to lead-containing paints and soils (26, 27). In the contaminated soil and mine tailings samples considered in this paper, As, Se, and Pb concentrations ranged from as low as 50 ppm to as high as several weight percent. In the model laboratory sorption samples considered here, the amount of heavy metal present on particle surfaces, measured in percentage of an ideal monolayer (ML), ranged from less than 10% ML to greater than 100% ML, or, in terms of concentration/unit area, from as low as 0.1 µmol/m2 to >20 µmol/m2. At these very low concentrations, sensitive element-specific probes capable of providing molecular-scale structure–composition information on heavy metal speciation are essential. Synchrotron-based x-ray absorption fine structure (XAFS) spectroscopy methods are well suited for this type of work because of the extremely high intensity of synchrotron x-rays and their energy tunability, which provide element sensitivity and element specificity, respectively. XAFS spectroscopy is capable of providing quantitative information on the geometry, composition, and mode of attachment (inner- vs. outer-sphere surface complexes vs. three-dimensional precipitates; monodentate vs. bidentate surface attachment, etc.) on specific elements at surface coverages as low as ≈0.05 ML by using existing synchrotron radiation sources and x-ray detectors (28, 29 and 30). The information derived from this work includes a molecular-scale description of the dominant type(s) of surface complexes or precipitates that form when these metals or metalloids partition from aqueous solution onto the mineral surface, a process generally referred to as sorption when the details of the adsorption mechanism are not fully known. We also present similar spectroscopic studies on the speciation of these heavy metals/ metalloids in several natural multiphase systems, including Secontaminated soils and sediments and As- and Pbcontaminated mining wastes. We show that significant fractions (up to 50 atom%) of As and Pb in our samples of mine tailings from the Mother Lode District of Central California and Leadville, CO, respectively, occur as sorbed species that are potentially bioavailable. We also discuss spectroscopic results on Se in contaminated soils that indicate redox transformations from more toxic to less toxic species. Natural Sources of Heavy Metal Contaminants and Pollutants. One common public misconception is that inorganic contaminants or pollutants (contaminants at sufficiently high concentrations to pose a hazard to exposed organisms) are introduced into soils and natural waters only through anthropogenic emissions. Because of their presence in minerals and the fact that minerals undergo dissolution caused by chemical weathering, heavy metals and metalloids are continually released into the environment at various rates and concentrations from natural materials. When concentrated in certain types of mineral deposits exposed at the earth's surface, they can be released at sufficient concentrations to be considered pollutants (e.g., arsenic released from the weathering of arsenical pyrite and arsenopyrite). When such deposits are mined, the release is typically enhanced because of the crushing of ore and gangue (waste) materials and the resulting increase in surface area of material exposed to weathering processes. In these cases, pollution can become a serious problem, especially if the oxidation of sulfide minerals, which are often present in heavy metal ore deposits, causes acid mine drainage, which can enhance the mobility of heavy metal cations (see ref. 31 for a case study). For example, the oxidation of pyrite (FeS2) produces ferric hydroxide solids, sulfate, and hydrogen ions by means of a reaction of the following type: FeS2(s) + 15/4 O2(g) + 7/2 H2O(1) → Fe(OH)3(s) + 2 SO4(aq) + 4 H+(aq) [1]

which results in decreased pH, enhanced dissolution of minerals, and enhanced release of cations that are adsorbed on mineral surfaces in the vicinity of the dissolving pyrite. Desorption of surface-bound Pb(II) and other cations from mineral surfaces under acidic solution conditions results in an increase in the concentration level of potentially bioavailable forms of heavy metal cations in surface waters and groundwater. The following four sections are intended to provide a brief overview of some of the minerals that contain arsenic, selenium, and lead as major, minor, or trace components. Some of the phases listed are potential candidates for secondary minerals in various types of base-metal deposits or tailings (e.g., Pb-Zn-Cu sulfide deposits); others are primary minerals. A relatively comprehensive list of primary and secondary minerals containing these elements can be found on the Internet at the address: http://un2sg4.unige.ch/athena/mineral/minppcl.html. Although some of the minerals listed below are rare and may seem unimportant, they could be present at relatively high concentrations in certain contaminated or polluted environments containing one or more of these heavy metals or metalloids. Natural Forms of Arsenic in the Environment. Arsenic is the 51st most abundant element in crustal rocks (average concentration = 1.8 ppm) and can occur in the 5+, 3+, 0, 1−, and 2− oxidation states in different geological environments (32). In general, reduced inorganic As found in sulfide minerals is relatively low in toxicity, but oxidized inorganic As(III) and As(V) compounds are significantly more toxic than many organoarsenicals. Moreover, As(III) compounds are two to three times more toxic than As(V) compounds (20). Both As(III) and As(V) form anionic species when in aqueous solution, adsorbed to mineral surfaces, or incorporated into precipitates. As(V) adsorbs more strongly to mineral surfaces than does As(III), thus is generally less mobile and potentially less bioavailable (33, 34). Arsenic occurs naturally as a major component in arsenates [containing As(V)O4 units] (e.g., adelite [CaMgAsO4(OH)], chalcophyllite [Cu18Al2(AsO4)3(SO4)3(OH)27·33H2O], clinoclase [Cu3(AsO4)(OH)3], duftite [CuPb(AsO4)(OH)], hoernesite [Mg3(AsO4)2·8H2O], scorodite [FeAsO4·2H2O]); arsenites [containing As(III)O3 units] (e.g., armangite [Mn26(As18O50) (OH)4CO3], ecdemite [Pb6As2O7Cl4], finnemanite [Pb5 (AsO3)3Cl], paulmooreite [Pb2As2O5], trigonite [Pb3Mn

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3391

(AsO3)2(AsO2(OH)], trippkeite [CuAs2O4]); elemental arsenic [As(0)], sulfides and arsenides (e.g., arsenopyrite [FeAsS], lautite [CuAsS], skutterudite [(Cu, Ni)As3], smaltite [(Co, Ni)As3−x]); sulfosalts (e.g., orpiment [As2S3], realgar [AsS], tennantite [(Cu, Fe)12As4S13]); or as a minor or trace component in minerals (e.g., in arsenical pyrite, where As can occur in arsenopyrite inclusions or as part of a solid solution, in jarosite [KFe3(SO4)2(OH)6], where As(V) substitutes for S(VI), or in hydrotalcite-like anionic structures, which have the general formula [M(II)1 2 x+ n− n− −xM(III)x(OH)2] (A x/n)·mH2O, where M(II) and M(III) are divalent and trivalent cations, respectively, and A represents anions such as CO3 −, SO 2−, CrO 2−, AsO 3−, and AsO 3−, which could occupy inner-layer positions to neutralize the positive layer charge (35). 4 4 3 4 Natural Forms of Selenium in the Environment. Selenium is the 66th most abundant crustal element (average concentration = 0.05 ppm) and occurs in the 6+, 4+, 0, and 2− oxidation states in different geologic settings depending on the Eh and pH values (32). In reduced forms, selenium is relatively insoluble and immobile, thus poses little danger to organisms. However, in oxidized forms, particularly Se(VI), selenium is mobile in aqueous solutions and poses a significant risk to organisms. Selenium is present as a major component in selenates [containing Se(VI)O4 units] (e.g., olsacherite [Pb2(SeO4)(SO4)], schmiederite [Pb2Cu(II)2(Se(IV)O3)Se(VI)O4(OH)4]); selenites [containing Se (IV)O3 units] (e.g., chalcomenite [CuSeO3·2H2O], hannebachite [Ca2(SeO3)2·H2O], molybdomenite [PbSeO3], sofiite [Zn2SeO3Cl2]); elemental selenium [Se(0)]; selenides (e.g., berzelianite [Cu2Se], umangite [Cu3Se2]); or as a minor to trace component in others (e.g., substituting for S in sulfides such as pyrite and in sulfates such as barite and jarosite). Natural Forms of Lead in the Environment. Lead is the 36th most abundant element in the earth's crust (average concentration = 13 ppm) and is generally present in the 2+ oxidation state in inorganic compounds and rarely present as elemental lead (32); it is not sensitive to oxidation or reduction over the normal range of Eh and pH values encountered in various geologic settings at or near the earth's surface. Lead can occur as a major element in a wide variety of minerals (e.g., anglesite [PbSO4], cerrusite [PbCO3], galena [PbS], pyromorphite [Pb5(PO4)3 (Cl, OH, F)]), and as a trace or minor element, especially by substitution for Ca2+ and K+ (e.g., Pb-bearing apatites and jarosites). Heavy Metal Uptake on Mineral Surfaces. Sorption on mineral surfaces is an important process that can bind and sequester heavy metals and other aqueous contaminant ions. Sorption can dramatically reduce the mobility of contaminants in groundwater and, in the case of redox-sensitive elements, result in their transformation into a less (or more) toxic species through reduction or oxidation reactions. The effectiveness of sorption reactions in binding an ion is determined by a number of variables, including (i) pH, (ii) the charge on the mineral surface as a function of pH, (iii) the type of sorption complex formed, (iv) competition between different ions for the same types of reactive surface sites, (v) the presence of organic and/or inorganic ligands that can inhibit or enhance sorption of a metal ion, and (vi) the presence of surface coatings such as biofilms that may block reactive sites and/or create new sorption sites. A discussion of each of these variables is beyond the scope of this paper. In this section we review the general modes of sorption of aqueous cations and anions on mineral surfaces and discuss the specific types of surface complexes formed by As, Se, and Pb, as revealed by x-ray absorption spectroscopy. Mineral surfaces in contact with aqueous solutions have a point of zero charge (pHpzc), which is the pH value (or small range of pH values) at which the surface is electrically neutral. Points of zero charge of silicate and oxide surfaces have been critically evaluated in several studies [e.g., (36, 37)] and range from 2–3 (SiO2) to ≈12 (MgO). Below the pHpzc value, the charge of a mineral surface is positive, indicating an excess of protons bonded to the surface. Above the pHpzc, the surface is negatively charged indicating an excess of OH− groups. At very high pH values, oxo ions (O2−) may occur on the surface of an oxide or silicate in contact with an aqueous solution. The pH of uptake or release of a heavy metal depends to a significant extent on whether the heavy metal occurs as a cation or anion in solution. For example, Pb(II) in aqueous solutions exists as a cation over a wide pH range, thus its affinity for a given mineral surface generally increases with increasing pH. In contrast, As(III), As(V), Se(IV), and Se(VI) behave as anions in aqueous solutions, thus are not strongly sorbed at high pH values where the mineral surface is negatively charged. The type of surface complex formed has an important effect on the mobility of a metal ion. Some metal ions bond directly to the mineral surface, losing waters of hydration and forming an inner-sphere complex (see Fig. 1). Such ions are relatively difficult to desorb except for large pH changes, thus are relatively immobile. However, if the metal ion forms a weakly bound outer-sphere complex in which the ion is surrounded by waters of hydration and no direct chemical bonds to the surface are formed, the metal can be easily desorbed when pH changes. An indication of the strength of binding of an aqueous cation or anion to a mineral surface can be obtained from the macroscopic uptake behavior as a function of ionic strength. For example, when an increase in the ionic strength of the background electrolyte (e.g., NaNO3) reduces the degree of uptake of an ion, the ion is less strongly bound to a mineral surface than when adsorbate uptake shows no ionic strength dependence. In the former case, the ion may form a weakly bound outer-sphere complex, whereas in the latter case, a strongly bound innersphere complex may be indicated. However, direct spectroscopic verification of the mode of sorption is required to verify conclusions drawn from macroscopic measurements alone. Among the ions examined here, Pb(II) forms strongly bound inner-sphere complexes on many mineral and metal oxide surfaces (38, 39 and 40), as do the oxoanions AsO43− (41, 42, 43, 44 and 45), AsO33− (46), and SeO32− (47). In contrast, the oxoanion SeO42− (47) forms more weakly bound complexes that are thought to be predominantly of the outer-sphere type, although there is still some controversy about the dominance of inner-sphere vs. outer-sphere complexes in the case of selenate sorption on iron (oxy)hydroxides (48). The mode of sorption of these species was determined by using XAFS spectroscopy. The strength of binding and stability of an inner-sphere complex will depend in part on how many bonds it forms with surface functional groups. For example, bidentate or tridentate surface complexes, attached by two and three bonds, respectively, will typically be more difficult to desorb than monodentate complexes (attached by a single bond), all other factors being equal. Another important factor in the effectiveness of sorption reactions in removing contaminant metal ions from solution is the possibility of multinuclear complex formation (49). In this case, more than one metal ion is involved in the surface complex. The progression from mononuclear to multinuclear surface complexes generally occurs with increasing sorbate metal concentration (29, 50), with precipitation occurring at a point where the solution becomes supersaturated with respect to the hydroxide of the metal-ion sorbate (50). When other ions are present, such as the anions carbonate, sulfate, or phosphate, additional complications can arise, including competition for surface sites, formation of solution complexes, or precipitation of solids made up of the sorbate ion and the anion. In addition to the types of adsorption described above, another potentially important sorption mechanism involves incorporation of certain types of metal ions into a solid coprecipitate made up of the sorbate ion and metal ions derived from dissolution of the mineral surface. Studies by several groups over the past few years have shown that this type of sorption product can form in aqueous systems containing sorbate ions like Co(II) (refs. 51, 52 and 53) and Ni(II) (54, 55) and sorbents like α-Al2O3, kaolinite [Al2Si2O5(OH)4], and pyrophyllite [Al2Si4O10(OH)2]. The result

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3392

ing solid phase can have a hydrotalcite-type structure (35) and may be relatively insoluble, and therefore decreases the metal ion bioavailability. A different mechanism of heavy metal sequestration by means of a coprecipitation mechanism is produced by amending Pb-contaminated soils with phosphate, resulting in the growth of the Pb-phosphate pyromorphite (56).

FIG. 3. As K-edge EXAFS data and fluorescent x-ray images for As-contaminated mine tailings from the Mother Lode District of California (derived from data reported in ref. 59). Fits to the EXAFS data are shown in red and molecular models of the As(V) environments in the model compounds used in the fits are also shown.

FIG. 4. Pb LIII-edge EXAFS data and fluorescent x-ray images for Pb-contaminated mine tailings from Leadville, CO (derived from data reported in ref. 62). EXAFS spectra of crystalline model compounds are shown in the center panel, and linear least-squares fits of EXAFS spectra of two mine tailings samples by using the model compound spectra are shown in the side panels. Bar plots of the abundance of different lead species are shown in the bottom panels on each side of the figure. EXPERIMENTAL APPROACH Details of the procedures we used in preparing model heavy metal sorption samples for XAFS spectroscopy studies can be found in Bargar et al. (38, 39). All sorbents were synthetic high-surface area powders (≈50 ≥ 100 m2/g) and are available commercially or can be synthesized. The identity and phase purity of the powdered sorbents were determined by powder x-ray diffraction, and in selected cases by transmission electron microscopy coupled with energy dispersive x-ray analysis. Surface area was determined for powdered sorbents by N2-Brunauer– Emmet– Teller (BET) measurements. Powdered sorbents were placed in 0.01M NaNO3 solutions and pH was allowed to equilibrate. In separate experiments, the heavy metal was added to each solution in contact with the sorbent at initial aqueous concentrations ranging from ≈0.1 to 0.01 M. pH was adjusted by addition of NaOH to achieve a desired percentage uptake of the heavy metal onto each sorbent, which was verified by graphite furnace atomic absorption (GFAA) spectrophotometry analysis of the supernatant. The resulting sorption densities ranged from ≈0.1 to >20 µmol/m2. Each sample was allowed to “equilibrate” at the final pH for 24–36 hr, then before GFAA analysis the sample was centrifuged and about 95% of the supernatant was removed. Except at the highest metal concentrations, the solutions were undersaturated with respect to the stable precipitate; thus, for example, no Pb(OH)2 or PbO precipitate phase was expected to

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3393

form in the case of the model lead sorption systems. In each case, the sorption sample in the form of a wet paste was loaded into a Teflon sample holder sealed with Mylar windows, and XAFS spectra were collected at the Stanford Synchrotron Radiation Laboratory (SSRL) on wiggler-magnet beamlines within 2 days of sample preparation. XAFS spectra for the model sorption samples were collected in the fluorescence-yield mode by using either a Stern–Heald type gas-filled ionization chamber detector (57) or a 13-element Ge detector under ambient conditions. The XAFS spectra of crystalline model compounds were collected in transmission mode under ambient conditions. In none of these experiments was evidence found for oxidation or reduction of the sorbate ion during or after XAFS data collection, as indicated by the lack of energy shifts in the edge positions, extended x-ray absorption fine structure (EXAFS)-derived metal-oxygen bond lengths, and x-ray photoelectron spectroscopy measurements. XAFS data analysis procedures used for the powdered sorption samples are described in Bargar et al. (38, 39) and O'Day et al. (58). XAFS spectroscopic analysis of the natural soil and mine waste samples containing As, Se, and Pb was carried out by using the experimental and data analysis procedures described in Foster et al. (59, 60), Pickering et al. (61), and Ostergren et al. (62), for As, Se, and Pb, respectively. These natural samples were not significantly modified from their conditions in the field, except for size separation in selected cases, in an attempt to preserve the same speciation of heavy metals present in the original samples. About 100 mg of each sample was placed in a Teflon holder and covered with thin Mylar tape, and XAFS data were collected by using fluorescence-yield detection (either a Stern–Heald detector or a multielement solid state detector) at ambient temperature (298 K) and pressure [1 atm (101.3 kPa)]. In addition to the normal EXAFS spectral fitting, which yields information on the identity and arrangement of first and second neighbors around the central absorber, linear least-squares fitting of the x-ray absorption near edge structure (XANES) spectra was conducted for the As- and Se-contaminated samples, and linear least-squares fitting of the XAFS spectra was conducted for the Pb-contaminated samples. By using this approach (59, 61, 62), the quantitative speciation of the heavy metal was determined by fitting the spectrum of the contaminated sample with spectra of model compounds, including both crystalline phases and model sorption samples with the heavy element sorbed at low surface coverage on different mineral or organic substrates. This approach works well when the spectral signatures of the different models are significantly different, which is often the case. Examples of this type of fitting are given in the sections below. The amount of surface-bound heavy metals in the contaminated soil and mine waste samples was determined in this manner. In addition to XAFS analysis of the contaminated samples, each sample was examined by powder x-ray diffraction, optical microscopy, and electron microprobe to determine the types and amounts of crystalline phases present (59, 62, 63). In addition, selected samples were examined by surface-sensitive x-ray photoelectron spectroscopy (62). RESULTS AND DISCUSSION As in Mine Tailings from the Mother Lode District, California. Almost 150 years of gold mining in the Mother Lode of California has resulted in significant concentrations of arsenic in mine tailings (up to 5,000 ppm), some of which have been used for housing developments in the Sierra Nevada Foothills of Central California. Because of the toxic effects of high concentrations of arsenic on humans and other organisms, there is concern about these tailings, and a number of studies are underway to determine the potential health hazard of these tailings to humans. We have used XAFS spectroscopy to determine the oxidation state, local coordination environment (to a radius of ≈7 Å around As), and the relative proportion of different As species in model compounds and three California mine wastes: a fully oxidized tailings (Ruth Mine), a partially oxidized tailings (Argonaut Mine), and a roasted sulfide ore (Spenceville Mine) (59). Analysis of the XANES spectra of these contaminated samples indicates that As(V) is the predominant oxidation state in the Ruth and Spenceville mine samples, but mixed oxidation states were observed in the Argonaut mine-waste. We obtained qualitative information about As(V) chemical speciation by fitting the XANES spectra of mine samples with a linear combination of component (model compound) spectra (Fig. 3). These analyses suggest the presence of As (V) species similar to those found in scorodite (FeAsO4·2H2O) and As(V) adsorbed on goethite (α-FeOOH) and gibbsite [γ-Al(OH)3]. Nonlinear least-squares fits of mine waste EXAFS spectra indicate variable As speciation in each of the three mine wastes. We conclude that ferric oxyhydroxides and aluminosilicates (probably clay) bind roughly equal portions of As(V) in the Ruth Mine sample. Our analysis suggests that tailings from the Argonaut Mine contain ≈20% reduced As bound in arsenopyrite (FeAsS) and arsenical pyrite (FeS2-xAsx) and ≈80% AS (V) in a ferric arsenate precipitate such as scorodite. Roasted sulfide ore of the Spenceville Mine contains As(V) substituted for sulfate in the crystal structure of jarosite [KFe3(SO4)2(OH)6], and sorbed to hematite surfaces. Determination of solid-phase As speciation by means of EXAFS spectroscopy is a valuable first step in the evaluation of As bioavailability, because the mobility and toxicity of As compounds vary with As oxidation state. As bound in crystalline or x-ray amorphous precipitates is generally considered to be less available for uptake by organisms than when sorbed to mineral surfaces. Se-Contaminated Soils in the Central Valley of California. Selenium occurs naturally in sediments and soils in many parts of the Western U.S. and is assumed to be incorporated in pyrites in marine sedimentary rocks such as shales. When such soils are irrigated for agricultural purposes, this indigenous element becomes soluble and is transported in agricultural drainage waters to ponds and reservoirs where it becomes concentrated in water-borne plants and animals (up to 3,000 ppm). One result of this concentration process was discovered in the early 1980s by government scientists at the Kesterson National Wildlife Refuge in Merced County, California. Wildlife, particularly waterfowl, died or were born deformed from consumption of high levels of selenium (64). Similar problems have since been documented at nine sites in eight Western states comprising some 1.5 million acres of farmland. This problem has major financial and health implications in the San Joaquin Valley of California, a vast area that produces a significant portion of the nation's vegetables, fruits, and other crops. If farmers are prevented from draining irrigation waters in this region, the rapid buildup of salts in the soil will quickly make production of these crops impossible. About 500,000 acres of farmland in the San Joaquin Valley—one quarter of the Valley's agricultural acreage—are at stake, with an annual crop production worth about $500 million (65). To provide molecular-level information on the chemical forms of selenium present in Se-contaminated soils from the Kesterson Reservoir area, we have carried out XAFS spectroscopy studies that showed that selenate and selenite are present in the top few cm of soil adjacent to the drainage ponds but are reduced to elemental selenium at lower soil levels (61, 66). In carefully controlled laboratory studies of soil columns to which selenatecontaining solutions were added, Tokunaga et al. (67, 68) found that selenate is rapidly converted into elemental selenium. The reduction can occur by means of both biotic (4, 69) and abiotic (70) pathways. However, when reoxidized to the selenate form during irrigation, selenium becomes highly mobile and is transported as an aqueous complex in drainage waters. Various solutions to this problem have been proposed, including bacterial reduction, immobilization, and removal of selenium from drainage waters or the use of drainage waters to irrigate land on which salt-resistant plants such as cotton, Eucalyptus trees, and atriplex are grown. None have been adopted and some skeptics doubt that

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3394

a viable solution, which satisfies environmental, financial, and political constraints, will be found. An eventual solution to this problem will require a detailed knowledge of the redox chemistry of selenium, its speciation in soils and groundwaters, and the effect of microbial organisms and inorganic reductants on its speciation. Pb in Mine Tailings from Leadville, CO. Pb is a ubiquitous environmental contaminant in soils because of the intensive use of Pb in batteries, paints, alloys, and solder, ammunition, gasoline additives, and other commercial products and the production of lead by mining and smelting activities. A recent study of the history of atmospheric lead deposition over the past 12,370 years, as measured in a peat bog in the Jura Mountains of Switzerland (71), has shown that the greatest lead flux (15.7 mg/m2/yr) occurred in 1979. This level is 1,570 times the natural background value of 0.01 mg/m2/yr. However, since the elimination of tetraethyl lead as a gasoline additive, beginning in the 1970's in the U.S. and some other countries, lead contamination levels have dropped significantly (25). They still exceed natural background levels by orders of magnitude in soils in many nonurban localities where soils have become polluted as a result of mining and smelting activities and in urban localities where paints and other anthropogenic sources of lead contaminate soils. The bioavailability of Pb is known to vary widely among different Pb species, and this fact is often cited to explain apparently dramatic variations in Pb bioavailability from site to site (e.g., ref. 10). Understanding the detailed relationship between speciation and bioavailability necessarily begins with a complete, accurate, and direct identification of Pb species in environmental media, such as soils and mine waste. Working toward this goal, researchers have recently begun applying synchrotron radiation-based x-ray absorption spectroscopic (XAS) techniques to determine the molecular-scale details of Pb speciation at contaminated sites (62, 63, 72, 73). Here we summarize the results of our work on Pb-bearing mine tailings from Leadville, CO (62). Using the unique advantages of XAS techniques, we emphasize the identification and characterization of poorly crystalline and/or finegrained species, such as sorption complexes and poorly crystalline coprecipitates, which are likely to control Pb bioavailability and mobility in natural systems. Bulk Pb concentrations range from 6,000 to 10,000 ppm in the two tailing piles we sampled at the Leadville site. These concentrations necessarily raise human health and environmental concerns, but bioavailability and chemical lability of Pb in these materials vary dramatically and show little correlation with bulk concentrations (10). Because these samples are heterogeneous multiphase mixtures (Fig. 4), a variety of complementary analytical methods were used, including powder x-ray diffraction, scanning electron microscopy, electron probe microanalysis, x-ray photoelectron spectroscopy, and synchrotron radiation-based x-ray absorption. By using this suite of techniques, and XAS techniques in particular, in conjunction with physical and chemical separation techniques, we were able to identify and characterize a number of species not amenable to detection by conventional microanalytical techniques. In particular, we found direct spectroscopic evidence for Pb adsorbed to mineral surfaces and variations in this surface-bound component with pH as would be predicted on the basis of simplified model system studies of adsorption processes. Least-squares fitting of EXAFS spectra shows that 50% of the total Pb in selected samples of the carbonate-buffered tailings with near-neutral pH occurs as adsorption complexes on iron (hydr)oxides, whereas Pb speciation in sulfide-rich low pH samples is dominated by Pb-bearing jarosites; we find no evidence for adsorbed Pb in these latter samples. Importantly, the dominant Pb species in each of the tailing piles could not be definitively identified without the molecular-scale information provided by EXAFS analysis. Because these species likely control Pb transport and bioavailability in these environments, our results clearly illustrate the need for molecular-scale characterization as basis for understanding the behavior of and health risks posed by Pb in natural environments. SUMMARY AND CONCLUSIONS The heavy metal Pb and the metalloids As and Se are among the most common environmental contaminants resulting from anthropogenic activities and the weathering of natural mineral deposits. These elements occur in a variety of chemical forms or species that can vary widely in solubility, mobility, toxicity, and bioavailability, depending on their speciation. In contaminated soils and mine tailings, for example, they can occur in primary minerals, secondary minerals formed by weathering of primary minerals in situ, solid precipitates formed by reactions of contaminant ions in groundwater with other aqueous ions, and adsorbed species. Although it is relatively easy to determine the types of solid phases present in a contaminated soil or mine tailings sample and the concentration levels of heavy metals and metalloids they contain by using a combination of x-ray diffraction and analytical methods, it is considerably more difficult to assess the importance of adsorbed heavy metal/ metalloid species, particularly at very low surface coverages. Adsorbed species may comprise a significant fraction of the heavy metal or metalloid present, and they are often the most bioavailable fraction. By using a combination of synchrotron-based XAFS spectroscopy and other analytical methods, the molecular-level speciation of As, Se, and Pb, including the types of sorbed species, was determined for selected mine tailings and contaminated soils. Sorbed As and Pb species were found to be major components with potentially high bioavailability in mine tailings from the Mother Lode District of California and Leadville, CO, respectively. Similar studies have shown that the most toxic and potentially bioavailable forms of selenium, Se(VI), are transformed rapidly to environmentally benign forms in contaminated soils through a combination of biotic and abiotic processes. A detailed knowledge of molecular-level speciation of heavy metal/metalloid contaminants and the bioavailability of the different species of a contaminant element is necessary for setting maximum contaminant limits. Knowledge of speciation is also required for efficient and costeffective remediation efforts. This point is well illustrated by cleanup efforts at the former uranium processing plant at Fernald, OH, which serves as the host for the Uranium in Soils Integrated Demonstration. This demonstration project used carbonate soil-washing procedures to remove hexavalent uranium. However, this conventional remediation procedure was not totally effective because of the presence of secondary phases containing U(IV) and U(VI), the latter being in the form of insoluble phosphates. These uranium species were detected by using a combination of XAFS, optical luminescence, Raman spectroscopy, scanning electron microscopy, and powder x-ray diffraction (74) and microXAFS spectroscopy (75). Changes in remediation procedures based on this type of speciation information could result in significant cost savings and more efficient cleanup of environmental contaminants, both of which should be major societal goals. The studies by our group presented in this paper were supported by the Department of Energy, Basic Energy Sciences, and the National Science Foundation. All of the synchrotron-based spectroscopic work was carried out at the Stanford Synchrotron Radiation Laboratory (SSRL), which is supported by the Department of Energy (Basic Energy Sciences and Office of Biology and Environmental Research) and the National Institutes of Health. We are grateful to the staff of SSRL for their technical support of this work. We also acknowledge the collaborations of T. Tokunaga and S. Myneni (Lawrence Berkeley National Laboratory) and I. Pickering (SSRL) in the studies of Secontaminated soils; G. Morin, F. Juillot, and G. Calas (University of Paris 7) in the studies of Pb-contaminated soils from northeastern France, and G. A. Parks and the late T. N. Tingle (Stanford University) in the studies of As- and Pb-contaminated mine tailings.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MINERAL SURFACES AND BIOAVAILABILITY OF HEAVY METALS: A MOLECULAR-SCALE PERSPECTIVE

3395

1. Stumm, W., Werhli, B. & Wieland, E. (1987) Croat. Chem. Acta 60, 429–456. 2. Thompson, H. S. (1850) R. Agric. Soc. Engl. J. 11,. 3. Way, J. T. (1850) R. Agric. Soc. Engl. J. 11, 313–379. 4. Oremland, R. S., Hollibaugh, J. T., Maest, A. S., Presser, T. S., Miller, L. G. & Culbertson, C. W. (1989) Appl. Environ. Microbiol. 55, 2333–2343. 5. Gauglhofer, J. & Bianchi, V. (1991) in Metals and Their Compounds in the Environment, Merian, E., ed. (VCH, Weinheim, Germany), pp. 853–878. 6. Babich, H. & Stotzky, G. (1983) in Aquatic Toxicology, Nriagu, J. O., ed. (Wiley, New York) pp. 1–46. 7. Chaney, R. L., Ryan, J. A. & Brown, S. L. (1997) Proc. Workshop on Environmentally Acceptable Endpoints: Chlorinated Organics, Energetics, and Heavy Metals, (American Academy of Environmental Engineers). 8. Gulson, B. L., Mizon, K. J., Law, A. J., Korsch, M. J., Davis, J. J. & Howarth, D. (1994) Econ. Geol. 89, 889–908. 9. Ruby, M. V., Davis, A., Schoof, R., Eberle, S. & Sellstone, C. M. (1996) Environ. Sci. Technol. 30, 422–430. 10. Casteel, S. W., Brown, L. D., Dunsmore, M. E., Weis, C. P., Henningsen, G. M., Hoffman, E., Brattin, W. J. & Hammon, T. L. (1997) Draft Report — Bioavailability of Lead in Soil and Mine Waste from the California Gulch NPL Site, Leadville, Colorado, (United States Environmental Protection Agency). 11. Davis, A., Drexler, J. W., Ruby, M. V. & Nicholson, A. (1993) Environ. Sci. Technol. 27, 1415–1425. 12. Davis, A., Ruby, M. V. & Bergstrom, P. D. (1992) Environ. Sci. Technol. 26, 461–468. 13. Davis, A., Ruby, M. V., Goad, P., Eberle, S. & Chryssoulis, S. (1997) Environ. Sci. Technol. 31, 37–44. 14. Dieter, M. P., Matthews, H. B., Jeffcoat, R. A. & Moseman, R. F. (1993) J. Toxicol. Environ. Health 39, 79–93. 15. Freeman, G. B., Johnson, J. D., Liao, S. C., Feder, P. I., Davis, A. O., Ruby, M. V., Schoof, R. A., Chaney, R. L. & Bergstrom, P. D. (1994) Toxicology 91, 151–163. 16. Gasser, U. G., Walker, W. J., Dahlgren, R. A., Borch, R. S. & Burau, R. G. (1996) Environ. Sci. Technol. 30, 761–769. 17. Ng, J. C., Kratzmann, S. M., Qi, L., Crawley, H., Chiswell, B. & Moore, M. R. (1998) Analyst 123, 889–892. 18. Davis, A., Ruby, M. V., Bloom, M., Schoof, R., Freeman, G. & Bergstrom, P. D. (1996) Environ. Sci. Technol. 30, 392–399. 19. Yamauchi, H. & Fowler, B. A. (1994) in Arsenic in the Environment, Nriagu, J. O., ed. (Wiley Interscience, New York), pp. 35–53. 20. ScienceScope (1998) Science 281, 1261. 21. Greenwald, J. (1995) in Time Magazine, p. 1. 22. Vogel, N. (1995) in San Jose Mercury News, p. 1. 23. Presser, T. S. (1994) in Selenium in the Environment, Frankenberger, W. T., Jr. & Benson, S. L., eds. (Dekker, New York), pp. 139–155. 24. Nriagu, J. O. (1998) Science 281, 1622–1623. 25. Nriagu, J. O., ed. (1978) The Biogeochemistry of Lead in the Environment (Elsevier, New York). 26. Ewers, U. & Schilpköter, H.-W. (1991) in Metals and Their Compounds in the Environment, Merian, E., ed. (VCH, Weinheim, Germany), pp. 971–1014. 27. Agency for Toxic Substances and Disease Registry (1988) The Nature and Extent of Lead Poisoning in Children in the United States: A Report to Congress. DHHS Doc. No. 99–2966 (U. S. Department of Health Human Services, Public Health Service, Atlanta, GA). 28. Brown, G. E., Jr. (1990) in Mineral-Water Interface Geochemistry, Reviews in Mineralogy, Vol. 23, Hochella, M. F., Jr. & White, A. F., eds. (Mineralogical Society of America, Washington, DC), pp. 309–363. 29. Brown, G. E., Jr., Parks, G. A. & O'Day, P. A. (1995) in Mineral Surfaces, Vaughan, D. J. & Pattrick, R. A. D., eds. (Chapman & Hall, London), pp. 129–183. 30. Brown, G. E., Jr., Parks, G. A., Bargar, J. R. & Towle, S. N. (1998) in Proceedings of the American Chemical Society Symposium on Kinetics and Mechanisms of Reactions at the Mineral/Water Interface, Sparks, D. L. & Grundl, T., eds. (Am. Chem. Soc., Washington, DC), in press. 31. Moore, J. N. & Luoma, S. N. (1990) Environ. Sci. Technol. 24, 1278–1285. 32. Greenwood, N. N. & Earnshaw, A. (1984) Chemistry of the Elements (Pergamon, Oxford). 33. Frost, R. R. & Griffin, R. A. (1977) Soil Sci. Soc. Am. J. 41, 53–57. 34. Korte, N. C. & Fernando, Q. (1991) Crit. Rev. Environ. Control 21, 1–39. 35. Cavani, F., Trifiro, F. & Vaccari, A. (1991) Catal. Today 11, 173–301. 36. Parks, G. A. (1965) Chem. Rev. (Washington, DC) 65, 177–198. 37. Sverjensky, D. A. (1994) Geochim. Cosmochim. Acta 58, 3123–3129. 38. Bargar, J. R., Brown, G. E., Jr. & Parks, G. A. (1997) Geochim. Cosmochim. Acta 61, 2617–2637. 39. Bargar, J. R., Brown, G. E., Jr. & Parks, G. A. (1997) Geochim. Cosmochim. Acta 61, 2639–2652. 40. Strawn, D. G., Scheidegger, A. M. & Sparks, D. L. (1998) Environ. Sci. Technol. 32, 2596–2601. 41. Waychunas, G. A., Rea, B. A., Fuller, C. C. & Davis, J. A. (1993) Geochim. Cosmochim. Acta 57, 2251–2269. 42. Manceau, A. (1995) Geochim. Cosmochim. Acta 59, 3647–3653. 43. Waychunas, G. A., Davis, J. A. & Fuller, C. C. (1995) Geochim. Cosmochim. Acta 59, 3655–3661. 44. Fendorf, S. E., Eick, M. J., Grossl, P. & Sparks, D. L. (1997) Environ. Sci. Technol. 31, 315–328. 45. Foster, A. L., Brown, G. E., Jr., Tingle, T. N. & Parks, G. A. (1998) Am. Mineral. 83, 553–568. 46. Manning, B. A., Fendorf, S. E. & Goldberg, S. (1998) Environ. Sci. Technol. 32, 2383–2388. 47. Hayes, K. F., Roe, A. L., Brown, G. E., Jr., Hodgson, K. O., Leckie, J. O. & Parks, G. A. (1987) Science 238, 783–786. 48. Manceau, A. & Charlet, L. (1994) J. Colloid Interface Sci. 168, 87–96. 49. Chisholm-Brause, C. J., O'Day, P. A., Brown, G. E., Jr. & Parks, G. A. (1990) Nature (London) 348, 528–531. 50. Chisholm-Brause, C. J., Brown, G. E., Jr. & Parks, G. A. (1991) in XAFS VI, Sixth International Conference on X-Ray Absorption Fine Structure, Hasnain, S. S., ed. (Ellis Horwood, New York), pp. 263–265. 51. d'Espinose de la Caillerie, J.-B., Kermarec, M. & Clause, O. (1995) J. Am. Chem. Soc. 117, 11471–11481. 52. Towle, S. N., Bargar, J. R., Brown, G. E., Jr. & Parks, G. A. (1997) J. Colloid Interface Sci. 187, 62–82. 53. Thompson, H. A., Parks, G. A. & Brown, G. E., Jr. (1999) Geochim. Cosmochim. Acta, in press. 54. Scheidegger, A. M., Lamble, G. M. & Sparks, D. L. (1997) J. Colloid Interface Sci. 186, 118–128. 55. Scheidegger, A. M., Strawn, D. G., Lamble, G. M. & Sparks, D. L. (1998) Geochim. Cosmochim. Acta 62, 2233–2245. 56. Traina, S. J. & Laperche, V. (1999) Proc. Natl. Acad. Sci. USA 96 3365–3371. 57. Lytle, F. W., Greegor, R. B., Sandstrom, D. R., Marques, E. C., Wong, J., Spiro, C. L., Huffman, G. P. & Huggins, F. E. (1984) Nucl. Instrum. Methods 226, 542–548. 58. O'Day, P. A., Chisholm-Brause, C. J., Towle, S. N., Parks, G. A. & Brown, G. E., Jr. (1996) Geochim. Cosmochim. Acta 60, 2515–2532. 59. Foster, A. L., Brown, G. E., Jr., Tingle, T. N. & Parks, G. A. (1998) Am. Mineral. 83, 553–568. 60. Foster, A. L., Brown, G. E., Jr. & Parks, G. A. (1998) Environ. Sci. Technol. 32, 1444–1452. 61. Pickering, I. J., Brown, G. E., Jr. & Tokunaga, T. (1995) Environ. Sci. Technol. 29, 2456–2459. 62. Ostergren, J. D., Brown, G. E., Jr., Parks, G. A. & Tingle, T. N. (1999) Environ. Sci. Technol., in press. 63. Morin, G., Juillot, F., Ostergren, J. D., Ildefonse, P., Calas, G. & Brown, G. E., Jr. (1999) Am. Mineral. 84, 420–434. 64. Ohlendorf, H. M. & Santolo, G. M. (1994) in Selenium in the Environment, Frankenberger, W. T., Jr. & Benson, S. L., eds. (Dekker, New York), pp. 69– 117. 65. Benson, M. (1993) San Jose Mercury News, July 4, p. 1A. 66. Tokunaga, T. K., Brown, G. E., Jr., Pickering, I. J., Sutton, S. R. & Bajt, S. (1997) Environ. Sci. Technol. 31, 1419–1425. 67. Tokunaga, T. K., Pickering, I. J. & Brown, G. E., Jr. (1996) Soil Sci. Soc. Am. J. 60, 781–790. 68. Tokunaga, T. K., Sutton, S. R., Bajt, S., Nuessle, P. & SheaMcCarthy, G. (1998) Environ. Sci. Technol. 32, 1092–1098. 69. Losi, M. E. & Frankenberger, W. T., Jr. (1998) J. Environ. Qual. 27, 836–843. 70. Myneni, S. C. B., Tokunaga, T. K. & Brown, G. E., Jr. (1997) Science 278, 1106–1109. 71. Shotyk, W., Weiss, D., Appleby, P. G., Cheburkin, A. K., Frei, R., Gloor, M., Kramers, J. D., Reese, S. & Van Der Knapp, W. O. (1998) Science 281, 1635–1640. 72. Cotter-Howells, J. D., Champness, P. E., Charnock, J. M. & Pattrick, R. A. D. (1994) Eur. J. Soil Sci. 45, 393–402. 73. Manceau, A., Boisset, M.-C., Sarret, G., Hazemann, J.-L., Mench, M., Cambier, P. & Prost, R. (1996) Environ. Sci. Technol. 30, 1540–1552. 74. Morris, D. E., Allen, P. G., Berg, J. M., Chisholm-Brause, C. J., Conradson, S. D., Donohoe, R. J., Hess, N. J., Musgrave, J. A. & Tait, C. D. (1996) Environ. Sci. Technol. 30, 2322–2331. 75. Bertsch, P. M., Hunter, D. B., Sutton, S. R., Bajt, S. & Rivers, M. L. (1994) Environ. Sci. Technol. 28, 980–984.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3396

Long-range transport of mineral dust in the global atmosphere: Impact of African dust on the environment of the southeastern United States Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3396–3403, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. JOSEPH M. PROSPERO* PNAS is available online at www.pnas.org.

University of Miami, Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami, FL 33149 ABSTRACT Soil dust is a major constituent of airborne particles in the global atmosphere. Dust plumes frequently cover huge areas of the earth; they are one of the most prominent and commonly visible features in satellite imagery. Dust is believed to play a role in many biogeochemical processes, but the importance of dust in these processes is not well understood because of the dearth of information about the global distribution of dust and its physical, chemical, and mineralogical properties. This paper describes some features of the large-scale distribution of dust and identifies some of the geological characteristics of important source areas. The transport of dust from North Africa is presented as an example of possible long-range dust effects, and the impact of African dust on environmental processes in the western North Atlantic and the southeastern United States is assessed. Dust transported over long distances usually has a mass median diameter <10 µm. Small wind-borne soil particles show signs of extensive weathering; consequently, the physical and chemical properties of the particles will greatly depend on the weathering history in the source region and on the subsequent modifications that occur during transit in the atmosphere (typically a period of a week or more). To fully understand the role of dust in the environment and in human health, mineralogists will have to work closely with scientists in other disciplines to characterize the properties of mineral particles as an ensemble and as individual particles especially with regard to surface characteristics. There is increased interest in the properties of small airborne particles (aerosols) because of the role that they play in many environmental processes. Much of this interest stems from the possible impact of aerosols on climate-related processes that involve radiation and clouds. For this reason, aerosol studies have focused on the chemical and physical properties of aerosols that relate to radiation and to hygroscopic behavior. For many years, research efforts have mainly focused on anthropogenic aerosols, especially sulfate aerosols. Sulfate receives most attention because humans have massively affected the global cycle of atmospheric sulfur mainly through emissions from combustion sources (the burning of fossil fuels, oil and coal) and because the physical and chemical properties of sulfate particles make them especially efficient for affecting radiation propagation in the atmosphere and cloud nucleation (1). Humans have lived with pollutant aerosols for a relatively short time in their history, since the beginning of the industrial age. In contrast, humans have lived with wind-borne mineral dust over the entire course of their history on earth. This long history can be clearly read from the record retained in ice and snow cores and in deep sea sediments. These show that the concentration of dust has varied over an extremely wide range through time and that dust activity can change very abruptly. Here, I focus on the present-day transport of mineral dust to better understand the processes that affect transport and the possible environmental effects. Soil dust transport has a number of implications for humans. At the most fundamental level, the history of human agriculture has been closely tied to loess deposits. Most loess soils are comprised of mineral particles that are largely derived from till and outwash at the front of glaciers (2). Winds generate dust clouds; the larger-sized particles (tens of micrometers in diameter) because of their high settling velocity are deposited relatively rapidly (generally within hundreds to a thousand kilometers of the source) to form deep blankets of soil. Such soils can be extremely fertile and easily tilled. They lie predominantly in the mid-latitudes, where weather tends to be favorable for agricultural pursuits (relatively mild temperatures, a long growing season, adequate rainfall). Massive loess deposits are found throughout Europe and Asia in regions that saw the development of many of the early civilizations. In the new world, the loess deposits of the Midwestern United States and in southern South America supported a highly productive agriculture that facilitated colonization and rapid economic development. The generation and transport of dust are processes that continue to this day, playing an important role in geochemical and geophysical processes, including the addition of nutrients to soils and to the oceans (1). More recently, there is a new focus on the effects of aerosols on human health. The Environmental Protection Agency (EPA) is required by the Clean Air Act to set standards for air quality at levels that protect public health. To this end, the EPA has established a new standard that focuses on particles <2.5 µm in diameter; only these small particles can efficiently penetrate into the lungs. It has long been known that, in certain industrial environments (e.g., mines, factories), there were clearly identifiable health effects associated with specific minerals (e.g., silica, asbestos). However, a substantial fraction of wind-borne soil dust is found in this “respirable” size range. Thus, we must consider the possible impact of ambient mineral dust on human health. Because the most effective particles from the standpoint of radiation, nucleation, and health are those with diameters of a few micrometers or less, they will have a relatively long lifetime in the atmosphere with respect to gravitational settling—in the absence of precipitation removal, on the order of

Abbreviations: EPA, Environmental Protection Agency; AOT, aerosol optical thickness; AVHRR, Advanced Very High Resolution Radiometer; TOMS, Total Ozone Mapping Spectrometer; AI, aerosol index. *To whom reprint requests should be addressed. e-mail: jprospero@ rsmas.miami.edu.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3397

several weeks. During this time, they can be transported thousands of kilometers by winds. To assess the impact of such particles on climate, biogeochemical process, and health, we must have a good understanding of the sources of the particles, the processes that affect dust mobilization and transport, and the consequent global distribution. I review various aspects of dust transport and show that mineral particles can have a clearly discernable impact on atmospheric properties at great distances from the source. I discuss the global scale distribution of dust and the factors that affect its mobilization and long-range transport, including geological and geomorphological factors. Then, I focus on dust transport over the North Atlantic, showing how dust transport varies on a temporal and spatial scale. As an example, I present data from South Florida that show that African dust comprises an important part of the ambient aerosol, that it plays an important role in air quality, and that it has implications regarding geochemical processes and human health. Finally, I summarize those aspects of long-range dust transport that are poorly characterized and warrant further study. The Global Scale Distribution of Mineral Dust Studies performed over the past few decades have clearly established that large amounts of soil dust are mobilized by winds, mostly in arid regions, and that substantial quantities can be carried great distances (reviewed in refs. 2, 3, 4, 5, 6 and 7). Many aspects of global scale mineral dust mobilization and transport are summarized in an excellent review by Duce (8). I (9) review dust transport to the global ocean. Guerzoni and Chester (10) present papers focusing on African dust transport with emphasis on the Mediterranean. The volume by Leinen and Sarnthein (11) contains papers on dust studies relevant to paleoclimatic interpretations. Satellites can give an excellent picture of the transport of dust and other aerosols on a global scale: e.g., Fig. 1 shows the global distribution of aerosol optical thickness (AOT) as estimated from the National Oceanic and Atmospheric Administration Advanced Very High Resolution Radiometer (AVHRR) (12). AOT is estimated from backscatter radiation measurements made at an effective wavelength of 0.63 µm; high values of AOT usually indicate high concentrations of suspended particles. Because the AOT algorithm requires that the underlying surface has a low and constant albedo, AOT measurements can only be made over oceans. Fig. 1 shows the mean AOT distributions for July. The most prominent features in the figure are the very large “plumes” of high values of AOT that extend westward from the coast of Africa and eastward from the Middle East. The plume emerging from the west coast of North Africa is unambiguously attributable to dust whereas that off the west coast of South Africa is attributed to biomass burning. The high values of AOT over the Arabian Sea are due to dust transported from the Middle East.

FIG. 1. AVHRR aerosol optical depth, mean for July, 1988 through July, 1989. (R. Husar, personal communication; see also ref. 12.) There are examples of other types of plumes that show somewhat elevated values of AOT but are not nearly so prominent, coherent, or persistent as the plumes attributed to dust and smoke. In particular, in July, pollutant aerosols over the North Atlantic appear as a plume that emerges from the east coast of North America and the west coast of Europe and also along the east coast of Asia. Nonetheless, the effects of pollutant aerosols (as interpreted from AVHRR) are modest in comparison to the dust and biomass burning aerosols. Unfortunately, AVHRR (and other satellites operating in the visible spectrum) cannot be readily used to obtain information about specific source areas because of the restriction of the algorithm to ocean surface retrievals. Recently, the Total Ozone Mapping Spectrometer (TOMS) instrument has been used for detecting absorbing aerosols based on the spectral contrast at 340 and 380 nm in the upwelling ultraviolet spectrum (13). TOMS is sensitive to UV-absorbing aerosols such as mineral dust, volcanic ash, and soot aerosol from fossil fuel combustion sources and biomass burning. Because the UV surface reflectivity is typically low and nearly constant over both land and water (14), TOMS can detect aerosols over continents as well as oceans (15). The UV spectral contrast can be used in a nonquantitative way as an aerosol index (AI). The

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3398

temporal and spatial variability of TOMS AI has been matched to specific types of absorbing aerosols and with known sources: for example, specific volcanic eruptions, forest fires, and large dust events (15, 16 and 17). Fig. 2 a and Fig. 2 b shows global distributions of the frequency of occurrence of relatively high TOMS AI values for the months of January and July, 1984 (15). A comparison of Fig. 2 with Fig. 1 shows that there is a good conformity between the plumes located over the oceans. An exact match is not expected because the data in Fig. 1 are mean values of AOT, whereas Fig. 2 shows the frequency of occurrence of elevated values of TOMS AI. Also, the AI is not a quantitative measure of aerosol concentration. TOMS is sensitive to a number of aerosol physical properties and, of most importance, to the altitude of the aerosols layer (13). Dust Sources and Regional Characteristics Fig. 2 also provides information about the distribution of dust (and smoke) over land, and, of most importance, it yields clues about the location of dust sources. The most striking feature is that the greatest number of sources and the most active ones are distributed across a band of arid regions that extends from the west coast of North Africa, across the Middle East, and into central Asia. [Because most dust sources have a strong seasonal activity pattern, they will not necessarily be visible in the two examples shown in Fig. 2. Many of the patterns discussed here are visible in the figures presented in Herman et al. (15), where they show maps identical to those in Fig. 2 for each month of 1984 and 1988.] In contrast, there are many arid regions, including deserts, that do not show any significant dust activity. It is notable that there are no prominent dust plumes associated with any of the southern hemisphere arid regions, including the deserts in Australia, the Kalahari and Namib deserts in southern Africa, and the Atacama and Patagonian deserts in South America. (Note also the absence of any evidence of persistent absorbing aerosol production and transport over the heavily industrialized and developed regions of North America and Europe.) This is not to say that there is no dust mobilization in these regions. Indeed, the daily TOMS AI product shows relatively large-scale dust episodes in regions where one might expect to see them—for example, in the southwestern United States (15)—but the frequency is relatively low and the AI values tend to be small.

FIG. 2. Statistics on the occurrence of high values of TOMS aerosol index. The shaded area shows the number of days (lightest shading, 5–10 days; heaviest shading, 25–30 days) when moderate-to-high concentrations of absorbing aerosol were detected by the TOMS/ NIMBUS-7 satellite. (a) January, 1984. (b) July, 1984. (Figures adapted from ref. 15, figure 4.) The other striking feature in Fig. 2 is that some regions stand out as persistent dust sources. Although many source regions are obscured by dense dust plumes (for example, in West Africa, where high concentrations of dust are present almost continuously), many areas are clearly visible as well defined geometric patterns that persist from year to year. These features can often be linked to specific geographical locations. For example, in Fig. 2 a and Fig. 2 b, a large circular pattern is located at the eastern end of the North African dust plume. This area coincides with the Chad basin. Between 5,000 and 10,000 yr ago, when the climate was more humid, Lake Chad was much larger (Mega-Chad); its surface was 320 m above sea level compared with 200 m today (18). The floor of the former lake now forms a monotonous desert plain with dunes, shallow wadis, and salt flats (18). The soils in this region are readily deflated and dust storms are quite frequent throughout the region (19). In addition to anecdotal evidence, visibility reports (20) confirm the extremely high frequency of dust storms and low visibility in this region, and they show the same seasonal pattern that is seen in TOMS. The dense Harmattan dust clouds that affect the coastal regions of the Gulf of Guinea during the winter have long been attributed to dust transported largely from the Chad basin, especially in the northern region, in the Bodele Depression (19). The TOMS AI product does indeed show that this region is persistently active and is the source of the consistently high AI values; furthermore, an inspection of daily TOMS AI images shows that, within the Chad basin, the highest AI values are obtained in the region of the Bodele depression. Other active areas are visible in Fig. 2 a and Fig. 2 b. In northwest North Africa, there is an active spot in Tunisia and northern Algeria, on the south side of the Sahara Atlas mountains. Two large chotts (salt lakes) are located in this region, which is part of a large depression that was formerly an arm of the sea and that extends 400 km westward from the Gulf of Gabes. The chotts, which receive runoff from the mountains, are covered with water only in the lowest areas, except after periods of heavy rains. Thus, large areas of chott sediments are frequently exposed and subject to deflation. The dust sources observed by TOMS are most likely associated with these features. There is a persistently active region in northeastern Libya and western Egypt. In West Africa, persistent dust features appear on the western flanks of the Ahaggar mountains. In Fig. 2b, there is a prominent area in northeastern Sudan between 30°E and the Red Sea and flanked to the southwest by the Ethiopian Highlands. Other well defined source patterns are visible on the Arabian Peninsula extending up the Tigris-Euphrates region. A persistent feature is seen in the Aral Sea region. Finally, there is a well defined feature in western China, north of the Indian subcontinent; this is located in the Tarim Basin between the Tibetan Plateau and the Tian Shan Mountains.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3399

Many of the dust sources visible in TOMS have a sharply defined geometry that appears to be determined by characteristics of the topography. There has been considerable discussion and debate for many years about the types of terrains and environments that can serve as a source of fine-grained dust. Pye (ref. 2, pp. 65–73) identifies (among others) wadi sediments, lake and playa sediments, alluvial fans, and alluvial floodplain sediments. Although field investigations have validated these source types at specific sites, it has been difficult to extrapolate to larger scales. For example, Herrmann et al. (19) summarize the reported locations of dust activity in North Africa; these would appear to suggest that, in effect, almost all of North Africa is a source of dust. This observation may be literally true at ground level. But TOMS suggests that, from the standpoint of large-scale dust events and long-range transport, some sources are much more active and effective than others. The source terrains identified by the TOMS satellite appear to be generally consistent with those listed in Pye (2). These implied source areas have a number of features in common. They are located in arid regions. Many prominent sources are found in well known ancient sedimentary basins (e.g., Mega-Chad, Tigris-Euphrates) (18). A common characteristic of the TOMS sources is that they generally lie in topographical lows that currently receive runoff from surrounding highlands. The association of the TOMS-identified dust sources with such topographical features suggests a mechanism that could explain the activity of these sources. Precipitation in the highlands weathers rocks and soils. Fine particles are carried downstream to the basin and are deposited in river channels and wadis; in the dry season, the deposits become exposed, dry out, crack, and flake. When the wind velocity increases, the disrupted soil surface is easily deflated, and clouds of fine-grained dust are carried away (21). Although many sources can be identified in TOMS, there are large regions where the sources are obscured by persistent dust clouds during much of the year—in West Africa and the Arabian Peninsula, for example. In some cases, TOMS shows active sources that cannot be readily associated with topographical lows nor with readily identifiable sources of runoff. Nonetheless, it is reasonable to assume that the dominant dust sources in such regions will most likely have the same environmental attributes as the visible sources cited above and those listed in Pye (2). On the other hand, other types of hypothesized sources are not validated. For example, Pye (2) specifically mentions stony deserts and uses the Gobi as an example. TOMS does not show the Gobi as a major dust source (see below). Furthermore, Fig.2 shows that many regions that would seem to provide favorable conditions for dust activity are, in fact, free of significant sources. This suggests that there are other factors involved. In Australia, for example, the absence of significant dust activity might be attributable to character of the soil particles. Kiefert et al. (22) compare Saharan and Australian dusts. Australian suspended dusts display particle size modes between 8 and 12 µm that are surprisingly uniform in time and space. In contrast (and in agreement with the discussion above) dusts collected in Mali (West Africa) were much finer, with a mode at 2–3 µm. Australian soils and sediments tend to be highly aggregated, with large quantities of clay pellets. On the basis of the TOMS product, we conclude that aridity is a necessary, but not sufficient, characteristic of dust sources. It is also informative to compare the distribution of dust sources with the distribution of loess deposits (e.g., ref. 2). In general, the geographical distribution of dust sources is distinctly different from that of loess deposits. This leads to the conclusion that loessal soils do not seem to consistently supply substantial amounts of dust to the atmosphere under present day conditions. There are certainly some loess deposits, especially those in the People's Republic of China, that appear to be the source of major dust events (23). Although these may have a widespread and significant impact on air quality and geochemical processes, they are sporadic events and they are limited to the spring months. Also, dust activity may be aggravated to a large extent by human activity, especially agriculture. As pointed out above, in China, the Tarim Basin is the most persistent feature in the TOMS AI. In contrast, the Gobi Desert, which lies to the northeast of the Tarim Basin, does not appear to be a major source of long-range dust, contrary to common belief. This should not be too surprising because the Gobi is predominantly a stony desert, with only 5% covered with sand dunes; indeed, the name “Gobi” in Chinese means “gravelly, pebbly plain.” Processes in Dust Source Areas In order for mineral dusts to be carried great distances, it is necessary that a substantial fraction of the deflated dust has a size under ≈10 µm in diameter. The limitation of the source dust to this size range has implications regarding the weathering processes by which particles of this size are produced and also about the consequent physical, chemical, and mineralogical properties of “long-range” dust. The association of dust sources with topographical lows is consistent with the abundance of fine-grained soil material. We would expect that the soil particles will be highly weathered; the particle surface can be physically abused, and it may have acquired surface coatings (e.g., iron oxide deposits, adhering flecks of very fine clay particles). In the case of clay minerals, the lattice structure may be distorted through removal or substitution of elements. In short, we should not expect to see fresh mineral entities. Contrary to general belief, sand dunes are not usually good sources of fine particles. This point is emphasized by the observation made above that the TOMS sources are associated with topographical lows that suggest that wet processes are important. This is not to say that substantial amounts of dust are not (or could not be) generated from sand dunes but, rather, that other types of sources are clearly important. The size distribution of the deflated dust is a strong function of many factors, including the physical properties of the soil matrix (e.g., the size distribution of soil particles, soil moisture, cohesiveness), the condition of the surface (e.g., whether it has been disturbed, the degree of protection afforded by vegetation), and the characteristics of the wind field above the surface (21). Studies of dust mobilization in soils from arid regions show that particles <10 µm in diameter can be released in large numbers. For example, measurements in Mali, West Africa, show a bimodal distribution with one mode at 44 µm in diameter and another at 5 µm; the large particle mode is attributed to relatively localized sources, and the 5-µm mode is attributed to long-distance transport (24). Measurements in dust storms in Tadzhjikistan show a substantial mass mode in the size range <10 µm; at a distance of 100 km downwind of a dust source, the dust mass peak was in the range 2–5 µm in diameter (7). In a study of windblown dust from agricultural fields in the Pacific Northwest, Clairborn et al. (25) compared size distributions during windy conditions with those during nonwindy conditions. They found that concentrations in the size range of 1–10 µm in diameter increased by as much as a factor of 5. A substantial fraction of the wind-borne dust particles are in the size range <1.0 µm in diameter. These are released from the soil by a “sandblasting” process (26). SEM studies show that the larger (1–10 µm) airborne particles are often coated with clay-like platelets (see also ref. 1). The impaction of wind-driven large particles dislodges the fine particles. Studies in wind tunnels using different types of soils (27) show that the concentration of submicrometer particles increases sharply when the wind velocity attains a threshold value that imparts

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3400

enough momentum to impacting particles to effect the dislodgment of the fine particles. Given the fact that the suspended dust has such a small mass median diameter, we would expect that the composition will be largely determined by the composition of the small particles in the soil and that clay minerals will be strongly represented. Schütz (28 and 29) reviews the literature on the mineralogical composition of African dusts. He shows that the dominant clay minerals are illite and kaolinite with additional amounts of smectite, montmorillonite and chlorite. Quartz and calcite are also important. As one might expect, there are regional differences in the composition that reflect the regional differences in geology and weathering processes. For example, kaolinite is more prominent in dust from the low latitudes of West Africa, and illite is more prominent in dust from the northern regions, toward the Mediterranean coast; this latitudinal distribution is consistent with the idea that kaolinite is a favored weathering product in warm, humid environments (e.g., the tropics), whereas illite is a product of midlatitude weathering conditions. This latitudinal distribution of illite is observed in dust collected over the world oceans (3). Schütz also shows differences in the composition of eastern North African dusts versus western dusts. Nonetheless, the dusts derived from sources within relatively large regions have a relatively uniform mineralogical composition. Caquineau et al. (30) show that, in wind-borne dusts over the eastern tropical Atlantic, the illite/kaolinite ratios fall into consistent patterns; in contrast, the amounts of quartz can vary widely and show no regional pattern. Although some regional mineralogical differences are noted, in general, these differences are not great. As a consequence, it is difficult to identify unique characteristics that would allow for a straightforward and unambiguous attribution to specific source regions (31 and 32). The elemental composition of North African dust is also rather homogeneous. If one normalizes composition data to Al or Fe (29), one finds that, for the vast majority of elements, the ratios are quite close to those of average crustal abundances (i.e., generally within a factor of about 0.5 to 2–5). Much higher ratios are found for some elements that are known to have substantial anthropogenic sources (e.g., Zn, As, Sb, etc.); also, the size distribution of these elements is shifted strongly to the submicrometer fraction, as one might expect for pollution sources (33) in contrast to the soil-derived elements, which peak in the supramicrometer size range. The uniformity in dust composition reflects the fact that the mobilized fraction of the soil is a highly weathered product that has been derived from number of sources in the region. Dust-sized particles in soils have gone through repeated cycles of mobilization and deposition, mixing with materials from other regional sources. Thus, dust that travels across the Atlantic is the homogenized product of a long chain of geological, weathering, and meteorological processes. African Dust Transport to Miami: An Example of Possible Air Quality Issues General Characteristics of Miami Aerosols. Thus far, I have discussed the general character of dust sources and some of the characteristics of the aerosols generated in those regions. Here, I report on the properties of dust after it has been transported across the North Atlantic, a distance of at least 5,000 km from the sources in North Africa. I focus primarily on measurements made in Miami, Florida, where continuous daily aerosol measurements have been made at a coastal site since 1974 (34). This data set may well be one of the longest records of daily aerosol composition measurements in the U.S. Data for the years 1989–1996 (Fig. 3) show a clear seasonal periodicity, with the maximum dust concentrations in June, July, and August. This temporal pattern is consistent with that seen in the satellite products, AVHRR (ref. 12; Fig. 1) and TOMS (ref. 15; Fig. 2).

FIG. 3. Daily mineral dust concentrations in Miami, 1989 to 1996. Measurements are made at a coastal site during onshore wind conditions. Dust episodes usually extend over several days or more; given the persistence of the trade-wind flow, this suggests that the scale of the dust events is on the order of several hundred to 1,000 km; this scale is consistent with satellite depictions of aerosol distributions over the western North Atlantic and the Caribbean (Fig. 1 and Fig. 2) (12, 15). Thus, measurements of aerosol concentration and composition in Miami should be representative of a very large region. The seasonal pattern of dust concentrations in Miami is similar to that at Barbados, West Indies (13° 15′ N, 59° 30′ W), where the University of Miami aerosol group has carried out a continuous sampling program since 1965 (35, 36 and 37). Fig. 4 shows the Barbados and Miami dust record for the period 1989 to 1996 (34). The major differences between the records from the two sites are that the dust concentrations are consistently higher at Barbados and that the dust transport season on Barbados is longer than that in Miami, where transport starts later in the year and ends earlier. As a result, the annual mean concentration at Barbados is ≈2.5 times that in Miami (9). The mineralogical composition of dust collected in the western Atlantic is identical to that collected off the coast of Africa (38); the dominant constituents are clay minerals and quartz. Long-Term Record and Relation to Climate. Summer dust transport has been a persistent feature throughout the 23 years of measurements in Miami, 1974 to 1996, as shown by the monthly mean dust concentrations in Fig. 5 (34). Nonetheless, there are very substantial variations in dust concentration over this period. Concentrations were consistently high during the period 1983–1987. The early 1980s was a time of severe drought in North Africa. Previous work has shown that summer dust concentrations measured in the trade winds at Barbados are anticorrelated with rainfall in the sub-Saharan (Sahel) region of North Africa (35, 39). In this regard, note that the highest monthly mean dust concentrations in Miami were obtained in 1983, the year after the onset of one of the most intense El Niño events in recent history; a similar sharp increase in dust also was observed on Barbados (35). The longer-term variability of dust transport could be linked in a complex way to other climate variables, such as the North Atlantic Oscillation (40).

FIG. 4. Monthly mean mineral dust concentrations at Barbados and Miami for the period 1989 to 1996. (Note the difference in the scales for Barbados and Miami.)

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3401

FIG. 5. Monthly mean dust concentrations measured in Miami, Florida, for the period 1974–1996. Size Distributions. In the discussion of dust properties in North Africa, it was noted that there was a clearly identifiable mode in the mass distribution in the size range below 10 µm. Measurements over the western North Atlantic show that the dominant size in the mass distribution is in the range of several micrometers and that there is very little mass above ≈10 µm; thus, the larger particles have been deposited during transit. Studies on Barbados (41) and Miami (42) during African dust events show that about one-third to one-half of the dust mass was less than 2.0–2.5 µm in aerodynamic diameter. In general, the mass median diameter of mineral dust over the oceans is typically 2–3 µm (see ref. 8 for a review of the literature on the size distribution of mineral dust over the oceans). Distribution of African Dust in the Eastern United States. The temporal and geographical extent of African dust transport to the United States is nicely depicted by Perry et al. (43), who studied PM 2.5 soil dust particle concentrations (based on the measured concentration of Al, Ca, Fe, Si, and Ti in twice-weekly daily samples) in a network of approximately 70 sites located in national parks and wilderness areas during the period 1992–1995. They observed that the highest individual PM 2.5 soil concentrations were associated with sites in the eastern United States during the summer, not in the arid southwest, as one might expect. Furthermore, there was a large-scale coherence in the temporal variability of the high PM 2.5 values, suggesting that they were associated with a large-scale forcing process. The elemental composition of the samples in these large-scale PM 2.5 events was distinctly different from other types of soil-dominated samples; furthermore, the composition of the samples at sites in the eastern United States during these events was identical to samples collected in the Virgin Islands. These observations are consistent with the hypothesis that the high PM 2.5 episodes were associated with incursions of African dust. Indeed, the progress of some of these dust incursions could be followed in the data as they moved from the Caribbean and Gulf of Mexico into the southern states and across the northeast United States. Independently, in a retrospective study of data from a field program held in central Illinois in the summer of 1979, Gatz and Prospero (44) noted the occurrence of unusually high concentrations of Si, Al, and other crustal elements carried by winds from the Gulf of Mexico. Concurrent mineral dust measurements at Miami show that there was a strong influx of North African dust at the same time. High concentrations of African dust are also routinely observed on Bermuda during the summer (45 and 46). Geological and Geochemical Implications of Dust Deposition. Geologically and geochemically significant amounts of dust are deposited in precipitation. In Miami, rain collected during intense summer dust events can have a turbid appearance. When filtered, some rains yield a “cake” of red-brown mud (47) which, when dried and rubbed with the fingers, produces an extremely fine, rouge-like powder. In contrast, during winter rain events, there is very little sediment, and the sediment is often gray-colored and gritty, characteristics that are associated with particles derived from pollution sources and from local soils that often have a high content of calcium carbonate. Measurements of dust deposition in precipitation in Miami over a 1-yr period (1982–1983) yielded a total Al deposition rate of 0.10 g·m−2·yr−1 (47), which is equivalent to a mineral dust deposition rate (8% Al) of ≈1.25 g·m−2·yr−1. This deposition rate is comparable to the long-term accumulation rate of aeolian minerals in the deep sea sediments of the tropical North Atlantic (47). When placed in the context of the meteorology and climatology of the region, it is clear that dust events are large-scale phenomena that can affect a very large region. This conclusion is supported by other studies. Landing et al. (48) measured the concentration of soil-related species (Al, Fe) in a network of sampling stations in Florida from the panhandle to the Florida Keys. Data show a well defined summer maximum at all sites, a seasonality that is consistent with the summer dust maximum observed at the University of Miami site (47). Concentrations were remarkably uniform during the summer months throughout the state. The deposition rate of Al in precipitation measured by Landing et al. (48) at five sites distributed over the length of Florida in 1993–1994 ranged from 0.062 to 0.148 g·m−2·yr−1 (equivalent to dust deposition rates of 0.78 to 1.9 g·m−2·yr−1). Almost all of the deposition of soil dust takes place during the summer months, which is also the rainy season in Florida. Dust deposition appears to be high throughout the western North Atlantic (49). Bermuda has an extensive cover of fine-grained red clayey paleosols. Herwitz et al. (50) found that the Zr/Y Zr/La ratios of these soils closely resembled that of the <2-µm fraction of African dust; contributions from two other hypothesized sources on North America (Great Plains loess and Mississippi River Valley loess) could not be detected. Muhs et al. (51) concluded on the basis of elemental composition that the soils on Barbados, Jamaica, the Florida Keys, and the Bahamas that African dust was the most important contributor to the soils on all of the islands. Mineral Dust and the EPA PM 2.5 Standard for Suspended Particles. There is renewed interest in aerosol properties because of evidence that the exposure of humans to high concentrations of airborne particulate matter can have a detrimental effect on health (52). Much evidence is based on epidemiological studies of death rates and respiratory-related hospital admissions (e.g., aggravated asthma, severe respiratory symptoms, and chronic bronchitis). Regions that have relatively high mean concentrations of particles tend to have higher rates of admissions than those with lower mean levels. Short-term exposures also appear to have an impact, as reflected in sharply increased admissions during severe pollution episodes. Until recently, the EPA standard for sus

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3402

pended particulate matter focused on particles having a diameter of 10 µm or smaller. In July, 1997, the EPA established a new standard for particles 2.5 µm in diameter or smaller (henceforth, the PM 2.5 standard) (52). EPA's new PM 2.5 standard specifies an annual mean of 15 µg·m −3 and a 24-hour mean of 65 µg·m−3. However, as shown above, African dust is the dominant aerosol constituent in south Florida during the summer. African air masses would, in effect, bring into the region very high PM 2.5 “background” aerosol concentrations. During dust events, the concentration of dust, coupled with particles from local emissions, could conceivably yield aerosol concentrations that exceed the EPA's recently implemented PM 2.5 standard. Thus, to account for the impact dust on air quality, it will be necessary to develop a set of diagnostic indicators for African dust, such as the mineral composition of the dust (38), its elemental composition (43, 44, 45 and 46, 50, 54), or its morphological characteristics (1). As previously stated, African dust has a remarkably uniform mineralogical and elemental composition. In contrast, locally or regionally derived dusts in the eastern United States appear to have distinctly different characteristics and a much more variable composition, as suggested by the studies of elemental composition (9, 43 and 44, 48). Mineral Particles and Health. The EPA standards for airborne particulate materials are largely based on epidemiological studies of specific at-risk communities in regions that experience relatively high concentrations of anthropogenic particles and other airborne pollutants. Although these studies suggest that increased illness and morbidity are associated with increased concentrations of airborne particles, it has not been possible to unambiguously identify specific cause–effect relationships with specific components in aerosols nor to validate mechanisms by which low concentrations of particles could cause cardiopulmonary toxicity (55). Amdur (56) and Utell and Samet (55) review various hypotheses regarding mechanisms; they present evidence that suggests that aerosols coated with first row transition metals are especially efficient in producing an inflammatory response in the lungs. In particular, they suggest that surface-complexed iron can generate hydroxyl radicals in lung tissue and that these radicals have acute lung toxicity. In this regard, it should be noted that African dust particles collected on Barbados have a total Fe content of 3.4% (57). The dust particles are heavily coated with iron, which accounts for the characteristic red-brown color of filters collected during dust events. Under acid conditions, 6.2% of the total Fe content of the aerosol was readily soluble (57); thus, it might be expected that a substantial fraction of the Fe on dust could be readily mobilized in the lung once the particles are deposited on lung tissue. The ease with which trace species are desorbed or extracted from mineral particles will strongly depend on the specific properties of the mineral particles and their weathering history. Consequently, we might expect that the health effects of dust particles could vary greatly from region to region. Thus, any strategy to address the health issues of mineral dust particles will have to incorporate studies of the properties of the mineral particles themselves. Conclusions The presence of high concentrations of mineral dust over such large areas of the Earth has implications in many areas of science: meteorology, climate, and biogeochemical processes. To properly assess the role of dust in these processes, we must have a better understanding of the properties of the airborne soil particles. The mass median diameter of long-range dust is generally <10 µm, typically about several micrometers. The soil particles in this size fraction usually show signs of severe chemical and physical weathering. The particle surface may be coated with oxides or salts, other particles (e.g., clays) could be attached to the surface, or the particle itself might be made up of an agglomeration of smaller particles. Consequently, particles will have very complex physical and chemical properties that cannot be elucidated by simply studying the bulk properties of the dust; individual particle analysis will be required. Dust generation is a highly nonlinear process that is very sensitive to climate change. Indeed, the geological record shows that dust mobilization has varied tremendously through time. Recent research suggests that mineral dust also plays a significant role in climate forcing (see, for example, refs. 59, 60, 61, 62 and 63). Thus, there could be feedback between the climate– dust generation processes and the climateforcing effects of mobilized dust. To understand the role of dust in climate, it will be necessary to develop models that can both characterize the radiative effects of dust and predict the location and output rates of dust sources as a function of climate. The modeling of dust sources (58, 59, 60 and 61) is perhaps the most difficult task facing climate modelers. To accomplish this goal, it will be necessary for geologists and mineralogists to work closely together to study the processes that affect dust mobilization and also to relate the mineral properties of dust to the source terrains. Finally, there is a renewed interest in the health effects of fine particles as reflected in the new EPA regulations regarding suspended particles <2.5 µm in diameter. In many regions of the world, mineral dust is the dominant aerosol constituent, and, consequently, dust could constitute a widespread health threat. Although some types of mineral particles (e.g., asbestos, silica dust) are unambiguously linked to health issues, the health effects of ambient soil dust are unknown. A number of hypotheses focus on the chemical and physical properties of the particle surface, but these remain unproven. Nonetheless, it is clear that the investigation of health effects will require the participation of mineralogists who can characterize the surface properties of individual particles. This work was carried out as a part of the Atmosphere/Ocean Chemistry Experiment supported by the National Science Foundation, Grants ATM-94-14812, ATM-94-14808, and ATM-94-14846. 1. Buseck, P. R. & Pósfai, M. (1999) Proc. Natl. Acad. Sci. USA 96, 3372–3379. 2. Pye, K. (1987) Aeolian Dust and Dust Deposits (Academic, London). 3. Prospero, J. M. (1981) in The Sea, The Oceanic Lithosphere, ed. Emiliani, C. (Wiley Interscience, New York), Vol. 7, pp. 801–974. 4. Prospero, J. M., Uematsu, M. & Savoie, D. L. (1989) in Chemical Oceanography, eds. Riley, J. P., Chester, R. & Duce, R. A. (Academic, London), Vol. 10, pp. 187–218. 5. Middleton, N. J. (1990) in Techniques for Desert Reclamation, ed. Goudie, A. S. (Wiley, New York), pp. 87–108. 6. Goudie, A. S. & Middleton, N. J. (1992) Climatic Change 20, 197–225. 7. Golitsyn, G. & Gillette, D. A. (1993) Atmos. Environ. 27A, 2467–2470. 8. Duce, R. A. (1995) in Dahlem Workshop on Aerosol Forcing of Climate, eds. Charlson, R. J. & Heintzenberg, J. (Berlin), pp. 43–72. 9. Prospero, J. M. (1996) in Particle Flux in the Ocean, eds. Ittekkott, V., Honjo, S. & Depetris, P. J. (Wiley, New York), pp. 19–52. 10. Guerzoni, S. & Chester, R., eds. (1996) The Impact of Desert Dust Across the Mediterranean (Kluwer, Dordrecht, The Netherlands ). 11. Leinen, M. & Sarnthein, M., eds. (1989) Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global Atmospheric Transport (Kluwer, Boston). 12. Husar, R., Prospero, J. M. & Stowe, L. L. (1997) J. Geophys. Res. 102, 16889–16909. 13. Torres, O., Bhartia, P. K., Herman, J. R., Ahmad, Z. & Gleason, J. (1998) J. Geophys. Res. 103, 17099–17110. 14. Herman, J. R. & Celarier, E. A. (1997) J. Geophys. Res. 102, 28003–28012. 15. Herman, J. R., Bhartia, P. K., Torres, O., Hsu, C., Seftor, C. & Celarier, E. (1997) J. Geophys. Res 102, 16911–16922.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LONG-RANGE TRANSPORT OF MINERAL DUST IN THE GLOBAL ATMOSPHERE: IMPACT OF AFRICAN DUST ON THE ENVIRONMENT OF THE SOUTHEASTERN UNITED STATES

3403

16. Seftor, C. J., Hsu, N. C., Herman, J. R., Bhartia, P. K., Torres, O., Rose, W. I., Schneider, D. J. & Krotkov, N. (1997) J. Geophys. Res. 102, 16749. 17. Chiapello, I, Prospero, J. M., Herman, J. R & Hsu, N. C. (1999) J. Geophys. Res., in press. 18. Bridges, E. M. (1990) World Geomorphology (Cambridge Univ. Press, Cambridge, U.K.). 19. Herrmann, L., Stahr, K. & Jahn, R. (1999) Contrib. Atmos. Phys., in press. 20. Mbourou, G. N., Bertrand, J. J. & Nicholson, S. E. (1997) J. Appl. Meteorol. 36 (7), 868–882. 21. Gillette, D. A. & Passi, R. (1988) J. Geophys. Res. 93, 14233–14242. 22. Kiefert, L., McTainsh, G. H. & Nickling, W. G. (1996) in The Impact of Desert Dust Across the Mediterranean, eds. Guerzoni, S. & Chester, R. (Kluwer, Dordrecht, The Netherlands), pp. 183–190. 23. Prospero, J. M., Uematsu, M. & Savoie, D. L. (1989) in Chemical Oceanography, ed. Riley, J. P. (Academic, New York), Vol. 10, pp. 187–218. 24. Gillies, J. A., Nickling, W. G. & McTainsh, G. H. (1996) Atmos. Environ. 7, 1081–1090. 25. Clairborn, C., Lamb, B., Miller, A., Beseda, J., Clode, B., Vaughan, J., Kang, L. & Newvine, C. (1998) J. Geophys. Res. 103, 19753–19767. 26. Gomes, L., Bergametti, G., Coude-Gaussen, G. & Rognon, P. (1990) J. Geophys. Res. 95, 13927–13935. 27. Alfaro, S., Gaudichet, A., Gomes, L. & Maille M. (1998) Geophys. Res. Lett. 25 (7), 991–994. 28. Schütz, L. (1997) in Proceedings of the Alfred-Wegener-Conference: Sediment and Aerosol, eds. von Hoyningen-Huene, W. & Tetzlaff, G. (Leipzig, Germany), pp. 9–13. 29. Schütz, L. (1989) in Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global Atmospheric Transport, eds. Leinen, M. & Sarnthein, M. (Kluwer, Dordrecht, The Netherlands), pp. 359–384. 30. Caquineau, S., Gaudichet, A., Gomes, L., Magonthier, M. C. & Chatenet, B. (1998) Geophys. Res. Lett. 25, 983–986. 31. Molinaroli, E. (1996) in The Impact of Desert Dust Across the Mediterranean, eds. Guerzoni, S. & Chester, R. (Kluwer, Dordrecht, The Netherlands), pp. 153–162. 32. Herrmann, L., Jahn, R. & Stahr, K. (1996) in The Impact of Desert Dust Across the Mediterranean, eds. Guerzoni, S. & Chester, R. (Kluwer, Dordrecht, The Netherlands), pp. 173–172. 33. Gullu, G. H., Olmez, I. & Tuncel, G. (1996) in The Impact of Desert Dust Across the Mediterranean, eds. Guerzoni, S. & Chester, R. (Kluwer, Dordrecht, The Netherlands ). 34. Prospero, J. M. (1999) J. Geophys. Res., in press. 35. Prospero, J. M. & Nees, R. T. (1986) Nature (London) 320, 735–738. 36. Savoie, D. L., Prospero, J. M. & Saltzman, E. S. (1989) J. Geophys. Res. 94, 5069–5080. 37. Li, X., Maring, H., Savoie, D., Voss, K. & Prospero, J. M. (1996) Nature (London) 380, 416–419. 38. Glaccum, R. A. & Prospero, J. M. (1980) Mar. Geol. 37, 295–321. 39. Prospero, J. M., Savoie, D. L., Arimoto, R. & Huang, F. (1993) Eos Trans. AGU 74, 146. 40. Moulin, C., Lambert, C. E., Dulac, F. & Dayan, U. (1997) Nature (London) 387, 691–694. 41. Li-Jones, X. & Prospero, J. M. (1998) J. Geophys. Res. 103, 16073–16083. 42. Hardy, K. A., Akselsson, R., Nelson, J. W. & Winchester, J. W. (1976) Environ. Sci. Technol. 10, 176–182. 43. Perry, K. D., Cahill, C. A., Eldred, R. A., Dutcher, D. D. & Gill, T. E. (1997) J. Geophys. Res. 102, 11225–11238. 44. Gatz, D. F. & Prospero, J. M. (1996) Atmos. Environ. 30, 3789–3799. 45. Arimoto, R., Duce, R. A., Savoie, D. L. & Prospero, J. M. (1992) J. Atmos. Chem. 14, 439–457. 46. Arimoto, R., Duce, R. A. Ray, B. J., Ellis, W. G., Jr., Cullen, J. D. & Merrill, J. T. (1995) J. Geophys. Res. 100, 1199–1214. 47. Prospero, J. M., Nees, R. T. & Uematsu, M. (1987) J. Geophys. Res. 92, 14723–14731. 48. Landing, W. M., Perry J. J., Jr., Guentzel, J. L., Gill, G. A. & Pollman, C. D. (1995) Water Air Soil Pollut. 80, 343–352. 49. Prospero, J. M., Barrett, K. Church, T., Dentener, F., Duce, R. A., Galloway, J. N., Levy, H., II, Moody, J. & Quinn, P (1996) Biogeochemistry 35, 27–73. 50. Herwitz, S. R., Muhs, D. R., Prospero, J. M. Mahan, S. & Vaughn, B. (1996) J. Geophys. Res. 101, 23389–23400. 51. Muhs, D. R., Bush, C. A, Stewart, K. C., Rowland, T. R. & Crittenden, R. C. (1990) Quaternary Res. 33, 157–177. 52. Wilson, R. & Spengler, J., eds. (1996) Particles in Our Air: Concentrations and Health Effects (Harvard Univ. Press, Cambridge, MA). 53. 62 Federal Register 138 (1997), Appendix K, 2.4a. 54. Glaccum, R. A. (1978) M.S. thesis (Univ. of Miami, Miami, FL). 55. Utell, M. & Samet, J. (1996) in Particles in Our Air: Concentrations and Health Effects, eds. Wilson, R. & Spengler, J. (Harvard Univ. Press, Cambridge, MA), pp. 169–188. 56. Amdur, M. (1996) in Particles in Our Air: Concentrations and Health Effects, eds. Wilson, R. & Spengler, J. (Harvard Univ. Press, Cambridge, MA), pp. 85–121. 57. Zhu, X. R., Prospero, J. M. & Millero, F. J. (1997) J. Geophys. Res. 102, 21297–21306. 58. Marticorena, B., Bergametti, G., Aumont, B., Callot, Y., N'Doume, C. & Legrand, M. (1997) J. Appl. Meteorol. 36 (7), 868–882. 59. Tegen, I. & Fung, I. (1995) J. Geophys. Res. 100, 18707–18726. 60. Tegen, I., Lacis, A. A. & Fung, I. (1996) Nature (London) 380, 419–422. 61. Tegen, I., Hollrig, P., Chin, M., Fung, I., Jacob, D. & Penner, J. (1997) J. Geophys. Res. 102, 23895–23916. 62. Sokolik, I. N. & Toon, O. B., (1996) Nature (London) 381, 681–683. 63. Lacis, A. A. & Mishchenko, M. I. (1995) in Dahlem Workshop on Aerosol Forcing of Climate, eds. Charlson, R. J. & Heintzenberg, J. (Berlin), pp.11–42.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3404

Biological impact on mineral dissolution: Application of the lichen model to understanding mineral weathering in the rhizosphere

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3404–3411, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. JILLIAN F. BANFIELD*, WILLIAM W. BARKER, SUSAN A. WELCH, AND ANNE TAUNTON PNAS is available online at www.pnas.org.

Department of Geology and Geophysics, University of Wisconsin, Madison, WI ABSTRACT Microorganisms modify rates and mechanisms of chemical and physical weathering and clay growth, thus playing fundamental roles in soil and sediment formation. Because processes in soils are inherently complex and difficult to study, we employ a model based on the lichen– mineral system to identify the fundamental interactions. Fixed carbon released by the photosynthetic symbiont stimulates growth of fungi and other microorganisms. These microorganisms directly or indirectly induce mineral disaggregation, hydration, dissolution, and secondary mineral formation. Model polysaccharides were used to investigate direct mediation of mineral surface reactions by extracellular polymers. Polysaccharides can suppress or enhance rates of chemical weathering by up to three orders of magnitude, depending on the pH, mineral surface structure and composition, and organic functional groups. Mg, Mn, Fe, Al, and Si are redistributed into clays that strongly adsorb ions. Microbes contribute to dissolution of insoluble secondary phosphates, possibly via release of organic acids. These reactions significantly impact soil fertility. Below fungi– mineral interfaces, mineral surfaces are exposed to dissolved metabolic byproducts. Through this indirect process, microorganisms can accelerate mineral dissolution, leading to enhanced porosity and permeability and colonization by microbial communities. Mineral Weathering, Microbes, and Geochemical Cycles The Importance of Mineral Weathering. Rocks at the Earth's surface typically formed at high temperature and pressure. Exposure of the minerals to oxygenated solutions initiates chemical and physical reactions, resulting in mineral dissolution and crystallization of new phases, such as clays, that are more stable at Earth's surface conditions (Fig. 1). Nano-crystalline products contribute abundant reactive surface area and thus can impact bioavailability of beneficial and toxic elements (Fig. 2). Weathering affects the compositions of ground water, river and lake water, and ultimately, of oceans (1). Resistant primary and secondary minerals are redistributed to form sediments and soils. Thus, weathering leads to a major geochemical fractionation near the Earth's surface. Such reactions have occurred throughout geological time, shaping the compositions of the mantle, crust, hydrosphere, and atmosphere. Mineral weathering also directly impacts humans, affecting water quality, agriculture, architectural stability, landscape evolution, integrity of repositories for high level nuclear waste, and distribution of mineral resources. Microbial Distributions in Natural Environments. Microbes inhabit diverse environments at, and near, the Earth's surface. Their potential to cause geochemical change is immense. Viable cells exist in extreme environments, from subzero temperatures in Antarctica (2, 3, 4 and 5) to above boiling temperatures in hot springs and hydrothermal vents (6). Microbes are essentially ubiquitous in sediments, and metabolically active cells have been discovered in rocks buried several kilometers below the Earth's surface (6, 7, 8, 9, 10, 11, 12, 13 and 14). Of the three domains of life, the majority of microorganisms reported to date from soils and sediments are bacteria (11, 15, 16, 17, 18, 19, 20 and 21). Archaea, minor members of many microbial populations, are especially important under more extreme conditions, such as those encountered in saline lakes (22) and very hot aqueous environments (22, 23 and 24). Eukaryotes typically occur in more moderate environments, although some fungi and protists thrive at very low pH (refs. 25 and 27; K. J. Edwards, T. M. Gihring, and J.F.B., unpublished data). Recent surveys of subsurface aquifers show large microbial populations, ≈105– 108 cells/cm3 (15, 28, 29 and 30). Cell concentrations ranging from 103 to 109 cells/cm3 have been reported from soils, sediments, and natural waters (15, 31, 32 and 33). Greater than 108 cells/cm2 of surface area occur on metal sulfide minerals (25, 26 and 27). Why Do Microbes Interact with Minerals? Some biologically essential elements are readily available from natural waters (e.g., Ca, Si, carbonate needed for structural fabrication) (34), but a subset must be actively scavenged (e.g., Fe, K, P). All organisms need Fe, but the solubility of Fe in natural oxygenated near-surface waters is low (35). Some organisms synthesize Fe-specific complexing agents to improve Fe bio-availability (35, 36, 37, 38, 39, 40 and 41). Other microbes utilize compounds that act as electron shuttles to improve accessibility to redox sites within minerals. Enzyme cofactors such as Mo, Cu, Zn, Mg, Fe, Cr, and Ni can be derived by dissolution of sulfide minerals and ferromagnesian silicates. Phosphorus, required for construction of DNA, RNA, ADP, ATP, phospholipids, and polyphosphates, is available by dissolution of minerals such as apatite [Ca5(PO4)3(F, Cl, OH)] and other typically less soluble secondary phosphates (42, 43 and 44). A subset of organisms (lithotrophs) associate closely with minerals because they derive their metabolic energy from inorganic substrates (e.g., Mn+2, Fe +2, S, NH +, and H ) (7, 45, 46, 47 and 48). Other microbes (heterotrophs) utilize organic material originating from photosynthetic or 4 2 lithotrophic microorganisms, and compounds such as O2, NO3−, Fe+3, and SO4−2 serve as electron acceptors (45, 46). Some interactions between cells, organic products, and minerals liberate ions from surfaces. This may be incidental (indirect consequences of growth) or under cellular control. Deciphering these interactions constitutes an important challenge for the future. Regardless of whether microbe–mineral interactions are directed or otherwise, microbial stimulation of mineral dissolution directly affects the fertility of agricultural and other soils. Microbial Controls on Mineral Weathering Reactions. Most geochemical research on mineral weathering has focused

*To whom reprint requests should be addressed. e-mail: jill@geology. wisc.edu.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3405

FIG. 1. A granite weathering profile that passes from almost fresh rock (below) into soil. Characterization of the mineralogy and microbiology of samples from throughout this profile reveals chemical and physical changes that accompany soil formation. on inorganic aspects (49). These studies provide valuable information on chemical dissolution rates, mechanisms, and products. Results reveal the order of reactivity of silicate minerals and show how rates depend on temperature, pH, and mineral and solution compositions. The absolute values of rates remain in considerable dispute. Conversely, soil scientists and agronomists have long recognized the fundamental importance of microorganisms in soil development (50). Understanding of near-surface systems requires integrated mineralogical, geochemical, and biological analysis. A variety of organic metabolic products can dramatically suppress or greatly accelerate rates of dissolution and secondary mineral formation (51, 52, 53, 54, 55 and 56). Energy generation by catalysis of redox reactions greatly affects element speciation (45, 46). The form of nitrogen is dramatically modified by microbial nitrogen fixation. Organic detritus greatly modifies the water retention capability and physical properties (porosity and permeability) of sediments and soils (57). Conversely, availability of metal oxides, sulfides, and hydroxides able to support lithotrophic growth, redox state, and pH dramatically controls microbial populations. Clay precipitation reactions may provide proton sources or sinks, and a variety of silicate minerals may serve as sources of metal reductive and oxidative power. In this paper, we discuss physical, chemical, and biochemical aspects of mineral dissolution and clay formation. Our approach is to analyze the lichen–mineral microcosm to identify the key factors responsible for microbe–mineral interactions and then to combine mineral and microbial characterization and experimentation to quantify the impact of these factors. Incipient Weathering and Microbial Colonization. In the initial stages of alteration, only rock surfaces exposed to air and water are colonized by biofilms. Because of space restrictions, incipient weathering at distance from the biofilm-mineral interface is predominantly inorganic. Inorganic reactions are restricted to sites of fluid access, typically at grain boundaries or in proximity to defects, and to reactions involving readily exchangeable ions (such as interlayer sites in layer silicates) (58, 59 and 60). Transmission electron microscope investigations show that incipient silicate weathering involves surface hydration, surface recrystallization, and release of ions to solution (60, 61). The scales of elemental redistribution are small, and primary mineral chemical gradients are preserved (62). Surface energy minimization leads to secondary minerals (clays and nanocrystalline oxyhydroxides) in highly specific orientations with respect to the parent mineral (Fig. 3). Thus, mineral surfaces are often coated by clays that largely fill space created by dissolution (63), leaving little room for microbial occupancy. Only after rock density is significantly reduced by dissolution does sufficient (micron-scale) space develop for microbial colonization to proceed. Dissolution of reactive phases (e.g., reduced minerals such as sulfides, olivine, and soluble solids such as glass) may facilitate microbial colonization. Metal sulfides, minor components of most crustal rocks (e.g., granites, basalts, metamorphic rocks), may be the first minerals to undergo chemical weathering, leading to generation of acid by the reaction FeS2(aqueous) + 15/4O2(gas) + 7/2H2O(liquid) → Fe(OH)3(solid) + 2H2SO4(aqueous) The rate of bacterial oxidation of ferrous iron released from pyrite surfaces is up to one million times faster than the inorganic oxidation rate at low pH (64). Because Fe3+ is the predominant pyrite oxidant at low pH, acid generation rates in natural environments are largely determined by microbial activity (65, 66). If sulfide mineral weathering in rocks is microbially mediated, sulfide dissolution can lead directly to colonization by lithotrophic species. Sulfuric acid generated by sulfide dissolution accelerates dissolution of surrounding silicates, increasing porosity available for microbial access and colonization. What Can We Learn from the Lichen–Mineral Microcosm? A wide variety of microorganisms colonize mineral surfaces, among the most familiar of which are lichens. Although

FIG. 2. High-resolution transmission electron microscope (HRTEM) image of nanocrystalline material produced by chemical weathering.

FIG. 3. High-resolution transmission electron microscope (HRTEM) image showing clays reaction products formed during early weathering of pyroxene develop in highly specific orientations with respect to the parent structure.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3406

FIG. 4. Optical micrograph showing a cross section through the lichen–feldspar interface. Photosynthetic microorganisms (pm) exist within the upper levels of a mass of fungal hyphae (h). A fungal fruiting body (a) is present in this image. Fungal hyphae contribute to physical weathering by penetrating feldspar cleavages and grain boundaries to expose the interior of crystals to microbial colonization. classically described as symbiotic associations between photosynthetic microorganisms and fungi (67, 68), lichens are actually extremely complex microbial communities. A mass of fungal hyphae, or thallus, composes the majority of any lichen. Photosynthetic microorganisms lie just beneath the upper surface. Although these are typically green algae, other photosynthetic microbes such as diatoms and cyanobacteria occur. This upper zone is a region of carbon transfer, in the form of sugars, between the photosynthetic assemblage and the fungal partner. Lower light levels preclude photosynthesis deeper within the lichen thallus, but other prokaryotes reside among the fungal hyphae. Little is known regarding the biodiversity of this nonphotosynthetic assemblage, and the role these organisms play in nutrient transfer from substratum to fungus is a fertile area for research. Lichens accelerate the degradation of minerals by physical and chemical methods and are ideal microcosms in which to study microbially mediated mineral weathering. Their extent is easily defined and sampled, and a limited assemblage of minerals grading from fresh to weathered is usually present. Ideally, mineral surfaces of known age (e.g., tombstones, building surfaces, or quarry faces) are studied, permitting fieldbased weathering rate determinations. Fungal hyphae, perhaps acting in concert with physical weathering mechanisms such as freeze-thaw, penetrate mineral cleavages and grain boundaries, leading to accumulation of substratum-derived mineral fragments within the lower thallus (Fig. 4). The intact organomineral interface from lichen-encrusted boulders in a 90-year-old rock quarry (52) was characterized by transmission electron microscopy. Mineral surfaces in microbially colonized regions are coated in complex mixtures of high molecular weight polymers, clays, and oxyhydroxides. Cells are attached to mineral surface and mineral weathering proceeds via polymer-mediated dissolution/transport/ recrystallization. Chemical weathering is accelerated relative to uncolonized surfaces.

FIG. 5. Zone model cartoon illustrating mineral weathering occurring in zones that are impacted by microbes to different degrees and in different ways. Zone 4 includes unweathered rock and rock incipiently weathered by inorganic reactions. Zone 3 is where reactions are accelerated by dissolved organic molecules (predominantly acids) but cells are in direct contact with reacting mineral surfaces. Zone 2 is the area of direct contact between microbes, organic products, including polymers, and mineral surfaces. Zone 1 is where photosynthetic members of the symbiosis generate fixed carbon and where crystalline lichen acids precipitate. A zone model for microbially mediated mineral weathering has been developed (Fig. 5) based on correlation of different styles of silicate mineral weathering with pore size-controlled microbial distributions (53). In brief, Zone 1 consists of the upper lichen thallus and is devoid of weathering of substratum-

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3407

FIG. 6. This energy-filtered transmission electron microscope (EFTEM) zero loss image reveals that complex mixtures of organic polymers (p) and clay minerals (c) exist at the lichen–microbe interface. derived mineral particles. Zone 2 is a region of extreme mineral weathering, characterized by direct contact between cells, extracellular polymers, associated compounds, and mineral surfaces. Complex, nanometer-scale mixture of clay products, organic polymers, and primary minerals are common (Fig. 6). In some cases, mineral formation on cell surfaces preserves relict cell shapes. Unlike essentially inorganic weathering, the secondary mineral orientations are less controlled by the primary mineral structure and are more determined by the distribution of polymers onto which they nucleate. Zone 3 weathering reactions, although accelerated by microbial products, are not mediated by direct microbial contact. Orientations of secondary minerals are more commonly determined by orientations of the primary minerals, as for essentially inorganic reactions. Unweathered minerals and minerals undergoing early, predominantly inorganic reactions comprise Zone 4 (Fig. 5). Application of Insights from the Lichen–Mineral Interface to Soils Microbial Populations in Soils. Microbial populations increase in abundance and diversity as rock is weathered and transformed to soil. In parallel, the microstructure and chemical complexity of the system increases, making high-resolution studies of processes occurring in soils extremely difficult. For this reason, the lichen–rock interface is a preferable system for study. Insights from the lichen system then must be tested for their applicability to soils. A diverse community, comprised of extremely high numbers of symbiotic and nonsymbiotic microorganisms, inhabit the soil zone around the roots of vascular plants. Although microbial concentrations in soils are very high, ≈109 cells/cm3, only ≈1–10% of the total, are alive (29). Thus, viability may ultimately prove to be a more important parameter than diversity or overall cell counts for mineral weathering studies. The assemblage of plant roots and their associated micro-flora and fauna in intimate contact with soil particles is termed the rhizosphere. Because many studies reveal the importance of microbe–mineral interactions to plant nutrition (69), it is critical to understand the rates and mechanisms of mineral transformations occurring in the rhizosphere. Which Microorganisms Are Involved? Microbial populations in soils are very diverse. One gram of soil can contain anywhere from hundreds to many thousands of species of microorganisms (16, 17, 18, 19, 20 and 21, 70, 71). Furthermore, the level of diversity from agricultural soils reported in molecular biological studies is astonishing. Identification of individual microbial cells in natural samples minerals remained challenging until recently. Traditional microbial identification techniques relied on culturing and isolation of microbial cells from a natural sample and then characterization of microbial isolates via a series of biochemical tests. This approach gives very biased population data because, although a large fraction of the visible cells are viable, only 0.01–10% of microorganisms are culturable by using current methods. Although these techniques yield some information on microbial populations and diversity, all information on the spatial distribution of organisms is lost. Recent breakthroughs in molecular biological approaches have provided powerful approaches to study the distribution and diversity of microbial cells in situ (71, 72, 73). The basis of the molecular approach to microbial population analysis is to extract, amplify, purify, and sequence DNA, typically the ribosomal DNA, from a natural microbial population and then to identify individual organisms by comparison between the obtained sequences and sequences from known organisms. Regardless of whether an organism has been identified previously, differences between its sequence and those of known organisms can be used to place the species onto a phylogenetic tree (71, 72 and 73). Once the sequence has been obtained, a DNA-specific, fluorescently labeled probe can be used to identify individual cells in environmental samples or in laboratory cultures (25, 26 and 27, 71, 73, 74). These probes are designed so that they will bind only to ribosomal RNA of organisms with complementary sequence (71, 73). Under appropriate experimental conditions, the degree of probe specificity (e.g., binds to all bacteria versus binds to Thiobacillus caldus) is determined by choice of probe sequence (e.g., an oligonucleotide common to all bacteria versus one common to only the target species). Microscopic visualization using multiple probes allows direct in situ characterization of the microbial population at the species (or higher) level. Combined DNA sequencing and probing has been used successfully in a number of natural environments (25, 26 and 27, 73, 74). In some cases, it has been possible to quantify the proportions of bacteria, archaea, and eukaryotes and to correlate abundances with geochemical conditions. This has provided new insights into interconnections between physical, chemical, and biological components of natural systems. Analysis of DNA from environmental samples and characterization of environmental samples using probes constructed with DNA sequences of interest are powerful new approaches for geochemical studies. These tools have not yet been deployed widely but should allow future researchers to evaluate interrelationships between microbial ecology and geochemical and physical parameters (25). Ultimately, these data can be correlated with metabolic information to infer how and why microbes interact with mineral surfaces in natural environments. DNA has been extracted and sequenced from several hundred soil microorganisms (16, 17, 18, 19, 20 and 21, 70). Most of the sequences were from bacteria, but many did not correspond to any of the known species. Several could not be classified into any of the larger known bacteria groups. Recently, archaea have been consistently reported from the rhizosphere. The distribution and importance of these organisms remain unclear. Experimental Quantification of Processes Occurring in Lichen Zones: Insights for Rhizosphere Processes. Plant roots excrete a combination of sugars, organic acids, and amino acids. It is believed that plant roots exude carbohydrates to encourage and sustain the associated microbial community. This large-scale carbon transfer from photosynthesizer to a nonphotosynthetic microbial community is analogous to Zone 1 in Barker and Banfield's (53) microbial weathering model (Fig. 5).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3408

FIG. 7. Secondary phosphate minerals formed on the surface of apatite during the early stages of weathering. These insoluble phases bind P in a relatively nonbioavailable form. The plant–microbe symbiosis is not the only similarity between the rhizosphere and lichen communities. Within soils, some microorganisms live in solution, but most are attached to mineral surfaces where they can directly affect mineral reactions (75, 76). Attached cells, largely fungi and bacteria, may impact mineral dissolution, precipitation, and clay hydration reactions in a manner analogous to that seen in Zone 2, at the lichen–mineral interface (Fig. 5). Physical restrictions placed on microbial distribution by pore size also apply to rhizosphere environments. Consequently, Zone 3 (Fig. 5) reactions are expected in soils as well. Although only the outer surface of a given mineral grain may be colonized, interior surfaces may show enhanced dissolution because of organic acids of plant and microbial origin. Elevated carbonic acid levels may result from fungal and bacterial degradation of organic matter (77, 78). Several to several hundred micromolar concentrations of oxalate, acetate, lactate, formate, pyruvate, propionate, malate, succinate, citrate, isocitrate, and aconitate have been detected in rhizosphere soil (79, 80 and 81). The fertility of numerous soils is limited by the abundance of bioavailable phosphate. In incipiently weathered rocks, microbial growth also may be phosphorus-limited because the abundance of organophosphates is low. Phosphorus limitation can develop during weathering because of either phosphorus adsorption onto secondary iron oxyhydroxide surfaces or binding of phosphorus into insoluble secondary minerals. During the initial stages of weathering, apatite is replaced by chemically or microbially precipitated secondary phosphate minerals containing iron (strengite), aluminum (variscite), and lanthanides (e.g., rhabdophane and florencite; Fig. 7). Secondary lanthanide phosphates persist until at least one-third of the initial mineral constituents are removed by dissolution (82). These phosphates are completely solubilized in the soil zone (83). DNA staining, in combination with high-resolution scanning electron microscope observations and unpublished experimental data, suggest that solubilization of secondary phosphates results from microbial colonization (Fig. 8). Microbes apparently respond to phosphate limitation by locating themselves at sites of inorganic phosphorus release or by excretion of organic acids or other complexing compounds to actively dissolve the otherwise insoluble secondary phosphate minerals (5, 43, 84). Further work is needed to determine which organic compounds are important. The effectiveness of likely compounds should be tested experimentally. Analysis of organic constituents in phosphorus-stressed cultures could be used to verify the microbial response. Identification of the mechanisms involved and microorganisms responsible for release of phosphorus from resistant secondary phosphates may provide new strategies for enhancement of productivity of low fertility soils (e.g., how to enhance the effectiveness of phosphate fertilizers and to manage soil fertility through manipulation of microbial populations).

FIG. 8. Microorganisms colonize surfaces of secondary minerals in pits formed by apatite dissolution. Microbial processes mobilize phosphorus from insoluble secondary phases.

FIG. 9. (Left) Fe released from biotite to solution in three experiments, two of which used bacterial cultures (B0428 and B0665) and one that was a control experiment. Significant enhancement of biotite dissolution rate is observed. (Right) Results of a feldspar dissolution experiment (pH 4.0) using undifferentiated and medium molecular weight polysaccharides demonstrate enhanced feldspar dissolution by two to three orders of magnitude under some conditions. Dissolution of inorganic phosphates may localize microbial activity and stimulate weathering of adjacent minerals. For example, it has been shown (85) that microbially mediated feldspar dissolution rates increase when feldspars contain apatite inclusions. Because feldspar dissolution rates are strongly pH-dependent, it is probable that the results can be attributed to release of acidic microbial byproducts. Mineral dissolution studies with cultures of bacteria and fungi show a dramatic increase in dissolution rates of feldspar, biotite (Fig. 9a), quartz, apatite, and other minerals (54, 86, 87, 88, 89, 90, 91, 92 and 93). In experiments with bacteria, mica, and feldspar, there is a direct correlation between microbial organic ligand production and increased release of Si, Al, and Fe (88, 89 and 90). Although these experiments replicate natural systems to some extent, they are inherently complex because of the combination of acidity effects, ion-complexing effects, and growth media effects. Thus, it is difficult to distinguish the contributing factors. Lichen–mineral interface characterization studies suggest that microbial colonization can dramatically impact both the rates and the mechanisms of silicate mineral weathering reactions via direct and indirect processes. To compare the effectiveness of organic–mineral interactions that depend on contact between microbes and mineral surfaces with those interactions that only involve soluble compounds, it is necessary to develop appropriately simplified experimental models. Acid production is the most basic mechanism by which microbes affect weathering reactions. We infer that this is largely responsible for enhanced reactivity in Zone 3 (Fig. 5).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3409

Solutions of organic acids in concentrations comparable to or slightly higher than ground water show increase in dissolution rates of less than one order of magnitude (56, 57). These relatively small effects may hide much larger responses in natural systems in which local microenvironments may be characterized by very high acid concentrations because of cell proximity. It is possible to demonstrate directly that microbes can cause low pH microenvironments at mineral surfaces. The pH values can be measured by using microelectrodes (94) or, in even smaller volumes, by using confocal-based techniques. For example, pH values of 3–4 were detected in proximity to bacterial cells within cleavages in biotite when the bulk solution pH was 7.0 (88). As acidity increases, below pH = 5, the rates of silicate mineral dissolution increase by a factor of anH+ (95). Lowering pH to 3–4 corresponds to a 10- to 1,000-fold increase in dissolution rate. The mechanism of the reaction changes as well. Typically, elements such as Fe and Al are relatively insoluble at neutral pH and form secondary Fe-hydroxides and vermiculite or smectite-like silicate clays. However, as acidity increases, Fe and Al solubility and mobility increase, probably leading to formation of different secondary minerals, such as the aluminosilicates kaolinite and halloysite. In addition to inorganic acid production, microbes also can catalyze mineral weathering rates by production of organic ligands. Ligands can complex with ions on the mineral surface and can weaken metal–oxygen bonds. Alternatively, ligands indirectly affect reactions by forming complexes with ions in solution, thereby decreasing solution saturation state. Experimental studies using relatively dilute solutions of compounds such as oxalic acid, citric acid, pyruvate, α-ketoglutarate, acetate, propionate, lactate, etc. have shown rate enhancements for silicate dissolution of up to one order of magnitude (96, 97, 98, 99, 100 and 101). The effect is somewhat similar to that of acid production because organic ligands affect silicate mineral dissolution stoichiometry by complexing with, and increasing the solubility of, less soluble major ions such as Al and Fe (96, 102, 103). The transformation of biotite to vermiculite with the release of the interlayer K is perhaps one of the most important biologically mediated geochemical reactions occurring in the rhizosphere (104, 105 and 106). Uptake of K+ by microbial cells and plant roots lowers solution saturation state, thereby indirectly promoting the weathering. However, plants and associated ectomycorrhizal populations have more aggressive methods for weathering sheet silicates. The extensive transformation of biotite to vermiculite has been attributed to root-induced pH decreases in the rhizosphere and acid dissolution of the mica structure (104). In a similar experimental study (93), plant roots and ectomycorrhizal fungi actively produced oxalate and increased biotite weathering when they were stressed for K and Mg. Plant roots and associated microbial populations also physically disrupt sheet silicates, exposing new surface area for chemical alteration (107). It has been shown that microorganisms produce intermediate molecular weight organic compounds that are important in iron transport to the cell surface (36, 37). Under conditions of Fe limitation, microbes produce siderophores that have a very high affinity for Fe+3. Siderophores form very strong stable bidentate complexes with octahedrally coordinated Fe+3 as well as with other major elements in minerals, e.g., Al and Si. Several hundred siderophores from microbial cultures and natural environments have been isolated and identified, most of which have either hydroxamate or catecholate functional groups. Estimates of siderophore concentrations in soil microenvironments range from 10 to 2,000 mg/ liter (38, 39). Mineral weathering experiments with naturally occurring siderophores show that these compounds can accelerate the rate of Feoxide and silicate mineral dissolution by about one order of magnitude and thereby dramatically impact Fe cycling in soils (35, 38, 39, 41). In addition to producing acids and low and intermediate molecular weight organic ligands, microbes also produce high molecular weight polymers. Observations of naturally weathered minerals associated with lichens show that, in the direct zone, these polymers coat surfaces. Polymers can affect mineral weathering reactions by several processes (52, 53). Slime layers are ≈99% water, so they can increase the contact time between water and the mineral surface. Polymers also increase the diffusion of ions away from the mineral surface. However, these extracellular polymers may have more direct chemical effects on mineral weathering as well. The effect of high molecular weight organic molecules on mineral dissolution rates was quantified by using dissolution experiments involving model extracellular polymers. The model polysaccharides (alginates) varied in the ratio of constituent mannuronic and guluronic acids. Mannuronic and guluronic acids differ in their effect because of different concentrations and orientations of carboxylic acids functional groups. Like low molecular weight organic acids, polysaccharides can increase the extent of mineral weathering (Fig. 9b), presumably by complexing with ions in solution, thereby lowering solution saturation state (51, 108). Microbially induced metal binding and mineralization is important in the rhizosphere as well as at the lichen–mineral interface. Cell surfaces and associated polymers provide effective nucleation sites for secondary silicate mineral precipitation. Some rocks contain high abundances of minerals enriched in potentially toxic elements (e.g., As, Cd, Cu, U). Microorganisms react to high metal concentrations in a variety of ways. One biochemical response to toxic substances is production of extracellular polymers that bind and effectively immobilize the compound and, in some cases, biomineralization. For example, microbes immobilize U by intra- and extracellular precipitation of secondary minerals (109, 110). Alternatively, some microbes exposed to arsenic employ mobile genetic structures (plasmids) to manufacture proteins that (although counterintuitive) reduce the less toxic As5+ to more toxic As3+ and secrete it from their cells (111). In other cases, microbes can completely volatilize the toxic element (e.g., by formation of methyl-mercury compounds) (111). Thus, microbial responses to toxic metals can dramatically change metal abundances and elemental speciation. Applications of Microbe-Mineral Interaction Studies for the 21st Century. Among the many challenges for the 21st century are questions concerning how (i) to maintain agricultural soil fertility to ensure food production for a still-growing world population; (ii) to prevent further environmental damage; and (iii) to invent cost-effective ways to remediate existing contamination of soil, sediments, and water. Improved knowledge of how microorganisms interact with their environments and contribute to geochemical transformations will be critical to these endeavors. We have established above that microbes are important agents of physical and chemical change in natural systems. In fact, global-scale models verify that chemical weathering reactions directly impact climate. It is widely accepted that, early in Earth's history, the atmosphere was dominated by CO2 and that oxygen concentrations only increased with the evolution of efficient photosynthetic microbial populations. Early inorganic rock weathering resulted in accumulation of Ca and Mg in ocean waters, leading to precipitation of Ca, Mg-carbonates and draw down of CO2. The rates of chemical weathering of Ca-silicates, Mg-silicates, and Ca, Mg-silicates determine the rate of supply of Ca and Mg to oceans and thus affect the magnitude of this critical feedback mechanism (112, 113, 114, 115 and 116). Mineral dissolution rates are important inputs into global climate models (112, 113, 114 and 115). However, the impact of microbial processes on weathering reactions over geological time are

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3410

unclear, and the relevant rates remain in dispute. If microorganisms significantly affect mineral dissolution rates, then the evolution of the biosphere (dominated by microorganisms for the majority of time), atmosphere, hydrosphere, and lithosphere have been closely coupled (117). Much can be learned about the long term response of the Earth system through quantifying the microbial–mineralogical feedback mechanisms. Thus, a challenge for the next century will be to develop more rigorous and comprehensive models for climate change. Climate change implies altered patterns of rainfall and temperature and thus, modification of the processes occurring in surficial materials, including soils. Prediction of the consequence of climate change requires a rather comprehensive understanding of the interrelationships between mineralogy, soil chemistry, microbial populations, rainfall, and temperature. Such understanding will benefit from investigations of soil-forming mineral reactions occurring under diverse climatic and hydrological conditions. Analyses must include detailed characterization of primary minerals and their weathering products, a census of microorganisms and determination of their patterns of distribution, and contextural information about soil physics and chemistry. It should be possible to increase the rates of soil development and optimize methods used to maintain or increase soil fertility through knowledge about microbial populations and how microbes cooperate to affect mineral dissolution, degradation of organic compounds, immobilization of ions, precipitation of minerals, and change in solution chemistry. For example, identification of microbial species capable of solubilizing secondary phosphates may provide new insights for improved fertilizer efficiency (e.g., simultaneous addition of microorganisms or microbial compounds with phosphate fertilizers). Understanding of microbial metal tolerance and of metal resistance strategies should lead to new approaches to environmental cleanup. Understanding how microbes affect the fate of contaminants in the natural environment will be critical for waste treatment and site remediation. An example of current regulatory interest is the long term, near-site containment of nuclides released from geological repositories. Challenges such as this demand sophisticated models that include all factors, inorganic and biological. Interdisciplinary study of microbe-mineral-solution interactions are required for model development. As we learn about the remaining 99–99.9% of organisms we currently know essentially nothing about, new species with new capabilities will be discovered. These may be vitally important sources of new compounds of tremendous technological importance. The discovery of hyperthermophilic microorganisms by Thomas Brock (University of Wisconsin, Madison, WI) (26) led to the identification of a thermally stable enzyme, taq polymerase, which has revolutionized the PCR step of DNA analysis. It is difficult to predict what the future will hold, but the discovery of enzymes capable of carrying out redox reactions at silicate and oxide mineral surfaces as well as in solutions may be of some importance for enhancement of soil fertility, economic metal extraction, and materials science. The global-scale balance between inorganic and microbial process is unclear, and the quantitative effects of microbes on geochemical cycles are essentially a mystery. At present, we are identifying specific processes that may be important and are quantifying them via laboratory studies. Ecological aspects of microbially dominated systems, including the symbioses and competition, largely remain a challenge for the future. New understanding of Earth's surface processes should come from detailed and quantitative studies that utilize current molecular biological and biochemical approaches in concert with high-resolution mineralogical and geochemical analyses. The authors acknowledge the editorial assistance of Dr. J. V. Smith (University of Chicago) and thank Dr. R. A. Eggleton (Australian National University), K. Edwards (University of Wisconsin, Madison, WI), T. Gihring (University of Wisconsin, Madison, WI), and Dr. P. Bond (University of Wisconsin, Madison, WI) for their contributions to this work. This research was supported by Grants EAR-9317082 and EAR-9706382 from the National Science Foundation and Grant DE-FG02-93ER14328 from the Department of Energy.

1. Banfield, J. F. & Nealson, K. H (1997) Rev. Mineral. 35, 1. 2. Vincent, W. F. & James, M. R. (1996) Biodivers. Conserv. 5, 1451. 3. Vishniac, H. S. (1996) Biodivers. Conserv. 5, 1356. 4. Broady, P. A. (1996) Biodivers. Conserv. 5, 1307. 5. Johnston, C. G & Vestal, J. R. (1993) Microb. Ecol. 25, 305. 6. Kerr, R. A. (1997) Science 276, 703. 7. Stevens, T. O. & McKinley, J. P. (1995) Science 270, 450. 8. Stevens, T. O., McKinley, J. P. & Fredrickson, J. K. (1993) Microb. Ecol. 25, 35. 9. Onstott, T. C., Tseng, H. Y., Phelps, T. J. & Colwell, F. S. (1995) Abstr. Programs Geol. Soc. Am. 27, 6, 312 (abstr.). 10. Phelps, T. J., Raione, E. G., White D. C. & Fliermans, C. B. (1989) Geomicrobiol. J. 7, 79. 11. Balkwill, D., Boone, D. R., Colwell, F. S., Griffin, T., Kieft, T. L., Lehman, R. M., McKinley, J. P., Nierzwicki-Bauer, S., Onstott, T. C., Tseng, H. Y., et al. (1994) Eos Trans. Am. Geophys. Union 75, 34. 12. Ghiorse, W. C. (1997) Science 275, 789. 13. Fredrickson, J. K. & Onstott, T. C. (1996) Sci. Am. 275, 68. 14. Fyfe, W. S. (1996) Science 273, 448. 15. Balkwill, D. J. (1989) Geomicrobiol. J. 7, 33. 16. Borneman, J., Skroch, P. W., O'Sullivan, K. M., Palus, J. A., Rumjanek, N. G., Jansen, J. L., Nienhuis, J. & Triplett, E. W. (1996) Appl. Environ. Microbiol. 62, 1935. 17. Borneman, J. & Triplett, E. W. (1997) Appl. Environ. Microbiol. 63, 2647. 18. Kennedy, A. C. & Gewin, V. L. (1997) Soil Sci. 162, 607. 19. Torsvik, V., Sorheim, R. & Goksoyr, J. (1996) J. Ind. Microbiol. Biotechnol. 17, 170. 20. Rheims, H., Rainey, F. A. & Stackebrandt, E. (1996) J. Ind. Microbiol. Biotechnol. 17, 159. 21. Franzmann, P. D. (1996) Biodivers. Conserv. 5, 1295. 22. Zimmer, C. (1995) Discover 16, 30. 23. Barns, S. M., Fundyga, R. E. & Jeffries, M. W. (1994) Proc. Natl. Acad. Sci. USA 91, 1609. 24. Barns, S. M. & Nierzwicki-Bauer, S. A. (1997) Rev. Mineral. 35, 35. 25. Edwards, K. J., Goebel, B. M., Rodgers, T. M., Schrenk, M. O., Gihring, T. M., Cardona, M. M., Hu, B., McGuire, M. M., Hamers R. J., Pace, N. R., et al. (1999) Geomicrobiol. J., in press. 26. Brock, T. D. (1978) Thermophilic Microorganisms and Life at High Temperatures. (Springer, New York). 27. Edwards, K. J., Schrenk, M. O., Hamers, R. J. & Banfield, J. F. (1998) Am. Mineral. 83, 1444. 28. Sinclair, J. L. & Ghiorse, W. C. (1989) Geomicrobiol. J. 7, 15. 29. Bone, T. L. & Balkwill, D. L. (1988) Microb. Ecol. 16, 49. 30. Fredrickson, J. K., Garland, T. R., Hicks, R. J., Thomas, J. M., Li, S. W. & McFadden, K. M. (1989) Geomicrobiol. J. 7, 53. 31. Albrechtsen, H.-J. & Winding, A. (1992) Microb. Ecol. 23, 303. 32. Kampfer, P., Steiof, M. & Dott, W. (1991) Microb. Ecol. 21, 227. 33. Hicks, R. J. & Fredrickson, J. K. (1989) Geomicrobiol. J. 7, 67. 34. de Vrind-de Jong, E. W. & de Vrind, J. P. M. (1997) Rev. Mineral. 35, 267. 35. Stone, A. T. (1997) Rev. Mineral. 35, 309. 36. Neilands, J. B. (1981) Annu. Rev. Biochem. 50, 715. 37. Neilands, J. B. (1982) Annu. Rev. Microbiol. 36, 285. 38. Hersman, L., Maurice, P. & Sposito, G. (1996) Chem. Geol. 132, 25. 39. Hersman, L., Lloyd, T. & Sposito, G. (1995) Geochim. Cosmochim. Acta 59, 3327. 40. Ochs, M., Brunner, I., Stumm. W. & Cosovic, B. (1993) Water Air Soil Pollution 68, 213. 41. Watteau, F. & Berthelin, J. (1994) J. Soil Biol. 30, 1. 42. Jansson, M. (1987) Microb. Ecol. 14, 81. 43. Leyval, C. & Berthelin J. (1989) Plant Soil 117, 103. 44. Cumming, J. R. (1993) Tree Physiol. 13, 173.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOLOGICAL IMPACT ON MINERAL DISSOLUTION: APPLICATION OF THE LICHEN MODEL TO UNDERSTANDING MINERAL WEATHERING IN THE RHIZOSPHERE

3411

45. Chapelle, F. H. (1993) Ground Water Microbiology and Geochemistry (Wiley, New York). 46. Lovely, D. R. & Chapelle F. H. (1995) Rev. Geophys. 33, 365. 47. Ehrlich, H. L. (1996) Chem. Geol. 132, 1. 48. Berthelin, J. (1983) in Microbial Geochemistry, ed. Krumbein, W. E. (Blackwell Scientific, Oxford), p. 223. 49. White A. F. & Brantley S. L. (1995) Rev. Mineral. 31, 1. 50. Huang, P. M. & Schnitzer, M., eds (1986) Interactions of Soil Minerals with Natural Organics and Microbes. (Soil Science Society of America, Madison, WI). 51. Welch, S. A., Barker, W. W. & Banfield, J. F. (1999) Geochim. Cosmochim. Acta, in press. 52. Barker, W. W. & Banfield, J. F. (1996) Chem. Geol. 132, 55. 53. Barker, W. W. & Banfield, J. F. (1998) Geomicrobiol. J. 15, 223. 54. Ullman, W. J., Kirchman, D. L., Welch, S. A. & Vandevivere, P. (1996) Chem. Geol. 132, 11. 55. Fortin, D., Ferris, F. G. & Beveridge, T. J. (1997) Rev. Mineral. 35, 161. 56. Drever, J. I. (1994) Geochim. Cosmochim. Acta 58, 2325. 57. Drever, J. I. & Vance, G. F. (1994) in The Role of Organic Acids in Geological Processes, eds. Lewan, M. D. & Pittman, E. D. (Springer, New York), p. 138. 58. Banfield, J. F. & Eggleton, R. A. (1988) Clays Clay Miner. 36, 47. 59. Banfield, J. F. & Eggleton, R. A. (1990) Clays Clay Miner. 38, 77. 60. Hochella, M. & Banfield J. F. (1995) Rev. Mineral. 30, 353. 61. Banfield, J. F., Veblen, D. R. & Jones, B. F. (1990) Contrib. Mineral. Petrol. 106, 110. 62. Banfield J. F. & Barker W. W. (1994) Geochim. Cosmochim. Acta 58, 1419. 63. Banfield, J. F., Jones, B. J & Veblen, D. R. (1991) Geoochim. Cosmochim. Acta 55, 2781. 64. Singer, P. C. & Stumm, W. (1970) Science 167, 1121. 65. Norstrom, D. K. & Southam G. (1997) Rev. Mineral. 35, 360. 66. Nordstrom, D. K. & Alpers, N. (1999) Rev. Econ. Geol. 6, in press. 67. Hale, M. E. (1983) The Biology of Lichens (Edward Arnold, London), p. 180. 68. Nash, T. H. (1996) Lichen Biology (Cambridge Univ. Press, New York), p. 303. 69. Jungk, A. O. (1996) in Plant Roots: The Hidden Half (Dekker, New York), p. 529. 70. Bintrim, S. B., Donohue, T. J., Handelsman, J., Roberts, G. P. & Goodman, R. M. (1997) Proc. Natl. Acad. Sci. USA 94, 277. 71. Amann, R., Snaider, J. & Wagner, M. (1996) J. Bacteriol. 178, 3496. 72. Pace, R. N., Stahl, D. A., Lane, D. J. & Olsen, G. J. (1986) Adv. Microbiol. Ecol. 9, 1. 73. Amman, R. I., Ludwig, W. & Schleifer, K.-H. (1995) Microbiol. Rev. 59, 143. 74. Schrenk, M. O., Edwards, K. J., Goodman, R. M., Hamers, R. J. & Banfield, J. F. (1998) Science 279, 1519. 75. Holm, P. E., Nielsen, P. H., Albrechtsen, H.-J. & Christensen, T. A. (1992) Appl. Environ. Microbiol. 58, 3020. 76. Hazen, T. C., Jimenez, L. & de Victoria, G. L. (1991) Microb. Ecol. 22, 293. 77. Chapelle, F. H., Zelibor, J. L., Jr., Grimes, D. Y. & Knobel L. L. (1987) Water Resources Res. 23, 1625. 78. Chapelle, F. H. & Lovely, D. R. (1990) Appl. Environ. Microbiol. 56, 1865. 79. Fox, T. R. & Comerford, N. B. (1990) Soil Sci. Soc. Am. J. 54, 1139. 80. Shen, Y., Strîm, L., Jînsson, J. & Tyler, G. (1996) Soil Biol. Biochem. 28, 1163. 81. Grierson, P. F. (1992) Plant Soil 144, 259. 82. Banfield, J. F. & Eggleton, R. A. (1989) Clays Clay Miner. 37, 113. 83. Taunton, A. E., Welch S. A., Santelli, C. M., Fournelle, J. & Banfield, J. F. (1998) in Geological Society of America Meeting: Abstracts with Programs (Geological Society of America, Boulder, CO), A304. 84. Illmer, P. A., Barbato A. & Schinner, F. (1995) Soil Biol. Biochem. 27, 265. 85. Rodgers, J. R., Bennett, P. C. & Hiebert, F. K. (1998) Geological Society of America Meeting: Abstracts with Programs (Geological Society of America, Boulder, CO), A305. 86. Barker, W. W., Welch, S. A. & Banfield, J. F. (1997) Rev. Mineral. 35, 391. 87. Berthelin, J. & Belgy, G. (1979) Geoderma 21, 297. 88. Barker, W. W., Welch, S. A., Chu, S. & Banfield, J. F. (1998) Am. Mineral. 83, 1551. 89. Welch, S. A. & Ullman W. J (1999) Chem. Geol., in press. 90. Vandevivere, P., Welch, S. A., Ullman, W. J. & Kirchman, D. L. (1994) Microb. Ecol. 27, 241. 91. Thorseth, I. H., Furnes, H. & Tumyr, O. (1995) Chem. Geol. 119, 139. 92. Callot, G., Maurette, M., Pottier, L. & Dubois, A. (1987) Nature (London) 328, 147. 93. Paris, F., Bottom, B. & Lapeyrie, F. (1996) Plant Soil 179, 141. 94. Parasuraman, C. S. (1995) Dissertation (Univ. of Delaware, Newark). 95. Blum, A. & Lasaga, A. C. (1988) Nature (London) 31, 431. 96. Welch, S. A. & Ullman, W. J. (1993) Geochim. Cosmochim. Acta 57, 2725. 97. Welch, S. A. & Ullman, W. J. (1996) Geochim. Cosmochim. Acta 60, 2939. 98. Stillings, L. L., Drever, J. I., Brantley, S. L., Sun, Y. & Oxburgh, R. (1996) Chem. Geol. 132, 79. 99. Bennett, P. C., Melcer, M. E., Siegel, D. I. & Hassett, J. P. (1988) Geochim. Cosmochim. Acta 52, 1521. 100. Hansley, P. L. (1987) J. Sediment. Petrol. 57, 666. 101. Manley, E. P. & Evans, L. J. (1986) Soil Sci. 141, 106. 102. Wogelius, R. A. & Walther, J. V. (1991) Geochim. Cosmochim. Acta 55, 943. 103. Antweiler, R. C. & Drever, J. I. (1983) Geochim. Cosmochim. Acta 47, 623. 104. Hinsinger, P., Elsass, F., Jaillard, B. & Robert, M. (1993) J. Soil Sci. 44, 535. 105. Hinsinger, P. & Jaillard, B. (1993) J. Soil Sci. 44, 525. 106. Berthelin, J. & Leyval, C. (1982) Plant Soil 68, 369. 107. April, R. & Keller, D. (1990) Biogeochemistry 9, 11. 108. Welch, S. A. & Vandevivere, P. (1995) Geomicrobiol. J. 12, 227. 109. Macaskie, L. E., Yong, P., Doyle T. C., Roig, M. G., Diaz, M. & Manzano, T. (1997) Biotechnol. Bioeng. 53, 100. 110. Jeong, B. Y., Hawes, C., Bonthrone, K. M. & Macaskie, L. E. (1997) Microbiology 143, 2497. 111. Silver, S. (1997) Rev. Mineral. 35, 345. 112. Walker, C. G., Hays, P. B. & Kasting, J. F. (1981) J. Geophys. Res. 86, 9776. 113. Berner, R. A., Lasaga, A. C. & Garrels, R. M. (1983) Am. J. Sci.283, 641. 114. Volk, T. (1987) Am. J. Sci. 287, 763. 115. Brady, P. V. (1991) J. Geophys. Res. 96, 18101. 116. Brady, P. V. & Carroll, S. A. (1994) Geochim. Cosmochim. Acta 58, 1853. 117. Schwartzman, D. W. & Volk, T. (1991) Paleogeogr. Paleoclimatol. Paleoecol. 90, 357.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

3412

A risk assessment for exposure to grunerite asbestos (amosite) in an iron ore mine

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3412–3419, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. R. P. NOLAN*†, A. M. LANGER*, AND RICHARD WILSON‡ PNAS is available online at www.pnas.org.

ABSTRACT The potential for health risks to humans exposed to the asbestos minerals continues to be a public health concern. Although the production and use of the commercial amphibole asbestos minerals—grunerite (amosite) and riebeckite (crocidolite)— have been almost completely eliminated from world commerce, special opportunities for potentially significant exposures remain. Commercially viable deposits of grunerite asbestos are very rare, but it can occur as a gangue mineral in a limited part of a mine otherwise thought asbestos-free. This report describes such a situation, in which a very localized seam of grunerite asbestos was identified in an iron ore mine. The geological occurrence of the seam in the ore body is described, as well as the mineralogical character of the grunerite asbestos. The most relevant epidemiological studies of workers exposed to grunerite asbestos are used to gauge the hazards associated with the inhalation of this fibrous mineral. Both analytical transmission electron microscopy and phase-contrast optical microscopy were used to quantify the fibers present in the air during mining in the area with outcroppings of grunerite asbestos. Analytical transmission electron microscopy and continuous-scan x-ray diffraction were used to determine the type of asbestos fiber present. Knowing the level of the miner's exposures, we carried out a risk assessment by using a model developed for the Environmental Protection Agency. We evaluate the potential for any risk to health in miners that might arise after the release of grunerite asbestos from a seam in an iron ore mine. None of the analytical criteria required for the mineral's identification were ambiguous (the objects studied were asbestos fibers, not cleavage fragments). A geological survey of the asbestos seam indicated localization in a relatively small area of the mine. No asbestos of any other variety was detected in the blast pattern and drill core samples. To evaluate the potential for asbestos exposure, an air sampling program that included area and personal samples was initiated. Both types of samples were analyzed by phase-contrast optical microscopy and analytical transmission electron microscopy (ATEM). The risk assessment calculations were referenced to the fibers ≥5 µm long, with fiber counts obtained by phase-contrast optical microscopy using standard National Institute of Occupational Safety and Health–Mine Safety and Health Administration (MSHA) methods. The grunerite asbestos identified in the iron ore mine is a known human carcinogen and merits special attention, although its presence in the mine appears to be an anomaly. The best evidence for the pathogenicity of grunerite asbestos has come from epidemiological studies of workers in factories where predominantly this fiber type was used. The mortality studies of lung cancer, mesothelioma, and asbestosis among grunerite asbestos exposed workers are reviewed. In addition, lung content analysis using ATEM was used to characterize the fiber concentrations found in lung tissues of individuals who developed asbestos-related diseases after exposure. The results of the air sampling program are used to calculate the mine work required to inhale a similar number of fibers as that found in the lungs of mesothelioma cases. The exposures measured in the iron ore mine are several factors of ten lower than the occupational exposures that occurred in the studied groups. Unlike the comparisons of lung content described above that assumes a threshold, the Environmental Protection Agency (EPA) model assumes a linear dose-response, where each exposure is associated with an incremental increase in risk. Brief Review of the Occupational Health Effects Associated with Asbestos Exposure. The earliest reports on the health effects of exposure to asbestos occurred among individuals who were exposed predominately to chrysotile asbestos (1). The first case in the English literature of asbestos-related pulmonary fibrosis described as asbestosis was reported in 1927 and occurred in a chrysotile textile worker. Although the first medical indications of any effect of asbestos on health was reported in 1906 in France and the United Kingdom, it (as with other diseases, like silicosis) was frequently complicated by the presence of tuberculosis. However, by 1938, asbestosis was generally accepted by industry and government health units as an occupational disease with distinct clinical, radiological, lung function, and pathological characteristics. Case reports of lung cancer accompanying asbestosis first began to appear in the literature during the 1930s. The evidence associating these diseases was greatly strengthened by the information Merewether provided for the 1947 Report of the Chief Inspector of Factories (England). He reviewed the accumulated data from 1923–1946 and found a 13.2% prevalence of lung cancer among the 235 autopsies of individuals know to have died with asbestosis, compared with 1.3% in 6,884 cases of silicosis. A high prevalence of lung cancer was found among other autopsy series of asbestosis cases, such as Wyer (1949), where 14.8% lung cancer was found among 115 asbestosis deaths (1), although at a meeting in Zagreb in 1953, Merewether (2) expressed doubt about the relationship between asbestosis and cancer of the lung, perhaps because of the limitations of an autopsy series. In 1955, Sir Richard Doll published a comprehensive epidemiological survey of employees of chrysotile asbestos textile plant in Rochdale, England (3). Individuals employed for 20 or more years experienced lung cancer ≈14 times more frequently than the general population (11 cases observed/0.8 expected). The results became available at the same time that

Abbreviations: ATEM, transmission electron microscopy; SMR, standardized mortality ratio; OSHA, Occupational Safety and Health Administration; MSHA, Mine Safety and Health Administration. *Environmental Sciences Laboratory, Brooklyn College of The City University of New York, 2900 Bedford Avenue, Brooklyn, NY 11210; and ‡Harvard University, 9 Oxford Street Rear, Cambridge, MA 02138 †To whom reprint requests should be addressed. e-mail: [email protected].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

3413

the association between lung cancer and cigarette smoking was being established. Defining the increase in the risk of developing lung cancer when an individual's exposure to chrysotile asbestos is insufficient to produce asbestosis is mostly theoretical. Changes in the diagnostic criteria of asbestosis have further complicated the matter. In 1960, Wagner et al. (4) reported 33 cases of a malignant tumor known as mesothelioma, which he attributed to crocidolite exposure. The discovery focused attention on the question of asbestos fiber type and disease. This rare tumor was the last of the three major asbestosrelated diseases to be identified. The potency of chrysotile to induce this tumor in humans remains a subject of considerable controversy. It also is clear that exposure to crocidolite asbestos, actinolite-tremolite asbestos, and grunerite asbestos produce considerably higher incidence of this disease, sometimes even after exposures that are considered quite low. The patterns of mesothelioma depending on asbestos fiber type are strikingly different in that a high mortality for mesothelioma is never found among individuals exposed only to chrysotile asbestos (5), although from time to time, individuals present with pleural mesothelioma and high concentrations of chrysotile are found to be present in the pulmonary tissue by lung content analysis (6). Geological Survey of the Area of the Mine Containing Grunerite Asbestos. The grunerite asbestos is confined to quartz–ankerite– grunerite veins of the host rock. These veins contain medium- to coarse-grained quartz, ankerite, stilpnomelane, and grunerite fiber distributed throughout a specific bench face (Fig. 1). The veins range up to 3 feet thick. The major veins occur within a magnetite–chert–silicate unit at the contact of the host rock and metadiabase sill units. The larger veins generally conform to the compositional banding of the host rock, but smaller veins commonly cut across the structure. Long-fibered asbestos mineral development is restricted to the thicker conformable veins. Grunerite asbestos is developed within the quartz–ankerite–stilpnomelane veins and along its contact with the host rock and sills. The veins were deformed structurally, exhibiting signs of shearing, brecciation, faulting, and folding. Minor quartz–carbonate veins occur, which lack asbestos-like minerals. The grunerite asbestos is discontinuous along the strike of the veins. Locally, recrystallization or replacement within the host rock has resulted in relatively coarse-grained acicular amphibole. The coarse-grained amphiboles are most notable in the silicate layers, but occur occasionally within the magnetite–chert bands, particularly near grunerite asbestos. Fibrous amphiboles occur irregularly in cross-cutting and concordant vein-like structures over a gradational zone from the host wall rock, with fairly coarser grained amphiboles, to quartz-ankerite– stilpnomelane–grunerite veins. The coarse grunerite asbestos occurs discretely within, and immediately adjacent to, the quartz–ankerite– stilpnomelane veins (Fig. 2). Strongly sheared horizons in the host rock close to the veins have formed platy, bladed, and fibrous mineral habits, only some of which are asbestiform. At several places along the strike of the quartz–ankerite–stilpnomelane–grunerite veins, the host rock has been tightly folded immediately adjacent to the vein (several inches on both sides). Essentially no deformation is observed just inches away from tight folding.

FIG. 1. The grunerite asbestos occurred along the wall and on top of the lower bench on the left. Banded, vuggy, quartz–fluorite–pyrite–chalcopyrite veins occur locally (most notably at the extreme southern end of the mapped bench) possibly in association with the quartz–ankerite–stilpnomelane–grunerite veins. The mineralogy and appearance of the sulfide veins indicate a different generation of development, but no clear cross-cutting relationships were observed. Minor quartz–magnetite–pyrite–chalcopyrite veins and veinlets occur. Analysis of Bulk Samples. Three bulk samples, selected from highly fibrous seams, were analyzed by polarized light microscopy, continuous-scan x-ray diffraction, and ATEM. In the United States, MSHA and the Occupational Safety and Health Administration (OSHA) regulate six minerals under the asbestos standard (Table 1). Five are amphiboles. These minerals have diverse elemental compositions (7). Each of the named minerals can exist in three different morphological forms or habits (8) that have been shown to effect their biological potential (9). In the asbestos habit, the fiber occurs as parallel fibrils, which form polyfilamentous bundles. It is this habit that is believed to cause cancer, and only this asbestos habit is regulated by MSHA and OSHA. The two other habits are nonasbestiform, occurring as splintery fiber, and massive anhedral nodules. When crushed, however, the nonasbestiform amphiboles may form elongated cleavage fragments that morphologically resemble fibers. Difficulties arise when cleavage fragments occur in association with amphibole asbestos. Two of the asbestos minerals (cummigtonite–grunerite and tremolite–actinolite) form a solid solution series in which Fe2+ and Mg2+ substitute. Although actinolite, grunerite, and tremolite do occur in nature as asbestos minerals, an occurrence of cummingtonite asbestos has not been reported.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

3414

FIG. 2. The coarse grunerite asbestos vein occur within, and immediately adjacent to the quartz–ankerite–stilpnomelane veins. All three of the highly fibrous samples were analyzed by polarized light microscopy, continuous-scan x-ray diffraction, and ATEM. None of the analytical criteria required for the mineral's identification are ambiguous (10). The asbestos seam is localized to a relatively small area of the mine. No other asbestos fiber type was detected in 24 blast pattern and drill core samples collected to evaluate the depth to which the seam extends. Evaluation of Air Samples from the Mine. To evaluate the potential for asbestos exposure by inhalation, an air sampling program (including both area and personal samples) was initiated. The personal samples were job classification-specific and sufficient in number to evaluate the range of exposures that would occur during mining of the ore. Of the 179 personal air samples collected, the mean concentration was 0.05 fiber per ml (all fiber ≥5 µm), and the highest exposure was 0.39 fiber per ml (all fiber ≥5 µm) (Table 2). None exceeded the MSHA asbestos standard (2 fiber per ml) (all fiber ≥5 µm) or action level, although 13.4% did exceed the current OSHA asbestos standard of 0.1 fiber per ml (all fiber ≥5 µm) (Table 3). Comparison of Epidemiological Studies of Workers Exposed to Iron Ore Dust and Those Exposed to Asbestos Dust. The four epidemiological studies described cover mortality. Such studies of causes of death, are used to determine whether a cohort (a group of individuals defined by exposure to some agent) dies more frequently from a particular disease than would otherwise be expected (based on rates in the reference population, e.g., everyone in the U.S.A.). Diseases such as lung cancer occur with a natural background. Cigarette smoking elevates the expected background death rate, and cancer incidence may be further increased by exposure to certain environmental agents. The assumption is made that the fraction of people that smoke is the same in the exposed as the control group. Epidemiological cohort studies allow for the determination of association between exposure to some agent and an increase in the occurrence of a specific disease. The standardized mortality ratio (SMR) is the number of deaths observed of a specific disease in the cohort divided by the number of deaths from that cause expected for the reference population, multiplied by 100. As the years of exposure increases, the SMR should also rise because of the increase in dose. Table 1. Mineralogy of the six minerals regulated under the asbestos standard in the United States. Amosite is occasionally referred to as cummingtonite–grunerite asbestos Mineral name Mineral group Chemical formula Commercial name Amosite Grunerite Amphibole (Fe2+, Mg)7[Si8O22] (OH)2‡ Anthophyllite asbestos Anthophyllite* Amphibole (Mg, Fe2+)7[Si8O22] (OH)2 Chrysotile Chrysotile Serpentine Mg3[Si2O5] (OH)4 Crocidolite Riebeckite Amphibole Na2Fe3+2(Fe2+, Mg)3[Si8O22] (OH)2 † Tremolite asbestos Tremolite* Amphibole Ca2Ma5[Si8O22] (OH)2 Actinolite asbestos Actinolite*† Amphibole Ca2(Mg, Fe2+)5[Si8O22] (OH)2 *These minerals do not have separate names for their asbestos analogs. Mineralogists now refer to amosite as grunerite asbestos and cocidolite as riebeckite asbestos, although the commercial names persist in the literature. †Tremolite-actinolite also form a solid-solution series between a calcium–magnesium-end member (tremolite) and a calcium–iron magnesium-end member (actinolite). ‡For amosite (grunerite asbestos) the Fe2+ is present in at least 5 of the 7 available x structural sites. Table 2. Results of the analysis of three hundred and twenty-six air samples collected while mining a grunerite asbestos (amosite) seen by the NIOSH-7400 methods No. of samples Concentration of fibers per ml† (all fibers ≥5 µm in the air) Range of fiber concentration Air sample Area During mining 137 0.02 ± 0.02 0.001–0.20 During blasting 10 0.01 ± 0.01 0.002–0.02 Total 147 Personal Drilling 110 0.06 ± 0.05 0.001–0.23 Shovel 22 0.06 ± 0.09 0.008–0.39 Production truck 23 0.04 ± 0.05 0.005–0.24 Track dozer 20 0.05 ± 0.04 0.005–0.17 Blast 2 0.03 ± 0.03 0.013–<0.05 Unidentified Sample 2 0.05 ± 0.03 0.028–0.07 Total 179 0.05 ± 0.05 0.001–0.39 Total samples analyzed 326 0.04 ± 0.05 0.001–0.39 Field 11 5.2‡ Laboratory 7 2.7‡ Unspecified 2 1.6‡ Not analyzed 3 349 Total samples taken †Values given as arithmetic mean ± SD—the fiber ‡Controls, values are fibers per mm2 of filter area.

concentrations a log normal.

A cohort of 17,800 asbestos insulation workers in the United States and Canada was followed from January 1, 1967 until the end of 1986 (11, 12). At the end of 1986, after almost 302,000 person-years of observation, 4,951 deaths occurred, while only 3,453 deaths were expected. The increased incidence of lung cancer accounted for >50% of the excess deaths (Table 4). The SMR (100 × observed/expected cases) for lung cancer was 435, whereas 8.6% and 9.3% of the deaths were caused by asbestosis and mesothelioma, respectively. Although the insulators were exposed to all of the commercial asbestos fiber types, the major fiber type retained in the worker's lung tissue was grunerite asbestos (12).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

Table 3. United States regulations concerning occupational exposure to asbestos fiber Regulatory agency Standard for an 8-hour time-weighted average, Excursion level, fibers per ml asbestos fibers per ml MSHA 2.0 ≤10‡ OSHA* 0.2 None Allowed OSHA† 0.1 None Allowed

3415

Action level, asbestos fibers per ml 1.0 0.1 0.05

All data refer to all fibers ≥5 µm. *Department of Labor Reg. 1986, 29CFR 1910-1926. Effective July 21, 1986. †Department of Labor Reg. 1994, 29CFR 1910, Effective October 11, 1994. ‡In a 15-min period.

Vermiculite Ore Containing Tremolite Asbestos. The mineral vermiculite has the generalized chemical formula (Mg, Ca)0.35(Mg, Fe, Al)3 (Al, Si)4O10(OH)2nH2O. On heating, the mineral loses water rapidly and expands to form a lightweight aggregate used for various purposes, e.g., insulation, soil conditioning, and filter medium. Various amphibole minerals associated with vermiculite have been the focus of health concerns, rather than vermiculite itself. The health effects among the miners and millers in Libby, Montana exposed to vermiculite containing tremolite asbestos have been studied by two groups of investigators (13, 14, 15, 16 and 17). Each investigation was designed as a mortality study and a cross-sectional chest radiographic survey. Slightly different criteria were used to define each cohort: the McDonald study (13, 14) contained 406 men with 165 deaths, and the Amandus study (15, 16 and 17) contained 575 men with 161 deaths. Both research groups used historical air samples to estimate exposure indices for the cohort members. The dust levels in the past were made with a device called a midget impinger, and the unit of concentration of dust was expressed in millions of particles per cubic foot (mppcf) of air. Conversion factors have been used to change the mppcf unit to an approximate number of fibers per milliliter of air (fibers per ml ≥5 µm), the units used in modern risk assessment (13, 15, 18). The exposure in the mill before the installation of dust control equipment in 1964, was estimated to be 400 and 168 fibers per ml (all fiber ≥5 µm), respectively. Dust levels between 1965 and the closure of the mill in 1974 were estimated by McDonald et al. and Amandus et al. to ≈20 and ≈33 fibers per ml (all fiber ≥5 µm), respectively. These were the highest exposures measured except for 20% higher dust levels during floor sweeping. McDonald and colleagues calculated the SMR for total mortality as 117, with 23 lung cancers observed against 9.4 expected (SMR = 245) and 4 mesotheliomas (2.4%). The SMR for the total mortality in the Amandus cohort was 110, with 20 lung cancers where ≈9 cases were expected (SMR = 223) and 2 mesotheliomas (1.2%). The lung cancer SMR for >20 years since first exposure for all exposure levels were 242 and 279 for the McDonald and Amandus cohorts, respectively. Both cohorts had an SMR of 250 for nonmalignant respiratory disease. Table 4. Deaths from lung cancer asbestosis and mesothelioma* among 17,800 asbestos insulation workers in the United States and Canada (1967– 1986)* Person-years Lung Cancer SMR Asbestosis Mesothelioma Years from onset of exposure E O O/E O No. per 100,000 per yr O No. per 100,000 per yr <15 61,655 3.9 9 232 1 1.6 0 0 15–19 52,709 11.6 37 318 14 26.6 5 9.5 20–24 57,595 27.5 95 346 31 53.8 18 31.8 25–29 50,518 46.6 183 393 52 102.9 73 144.5 30–34 37,165 57.4 281 490 59 158.8 105 287.5 35–39 20,340 46.8 239 511 84 413.0 91 447.4 40–44 10,200 30.8 155 503 80 784.3 59 578.5 45–49 5,256 18.8 75 399 33 627.8 58 1103.4 >50 6,151 25.4 94 370 73 1,186.8 49 796.6 301,593 268.7 1168 435 427 141.6 458 151.9 Total *Best evidence: Causes of death categorized after review of best available information (autopsy, surgical, and clinical). E, expected; O, observed.

Two Cohort of Minnesota Iron Ore Workers. Taconite is a term used particularly in the Lake Superior region of Minnesota for certain iron-containing rocks from the Biwabik Iron Formation. A high-grade ore concentrate is obtained from commercial-grade taconite that contains enough magnetite (Fe3O4) to be economically processed by fine grinding and wet-magnetic separation. Taconite is a hard, dense, fine-grained metamorphic rock that contains substantial quartz (20–50%) and magnetite (10–36%) and various mineral constituents, including hematite, carbonates, amphiboles (principally of the cummingtonite–grunerite series, although actinolite and hornblende also occur), greenalite, chamosite, minnesotaite, and stilpnomelane. Reserve Mining Company. Analysis of mortality data obtained on men who were employed from 1952–1976 has been reported (19). The study was initiated by concerns in the early 1970s that asbestos was released into the air and dumped into lake water during processing of the taconite rock (20, 21). It was inferred that this dust posed a risk to the miners as well as to the general public. Silver Bay and Duluth obtained their drinking water from Lake Superior, into which the pulverized waste rock (or tailings) from the pellet plant was deposited at Silver Bay. The U.S. Department of Justice considered this a potential health hazard. The Department alleged that the amphibole in the waste rock (cummingtonite–grunerite) was asbestos and the exposures would cause gastrointestinal cancer through ingestion and lung cancer from inhalation of the water- and airborne fibers (although they had done no calculation of this). The Reserve cohort consisted of 5,751 men, of which 907 had worked for the company for >20 years and 298 were deceased. The men had been exposed to respirable dust concentrations from 0.02 to 2.75 mg/M3, the modal range being 0.2–0.6 mg/M3. The fibrous particulate content of the dust was occasionally >0.5 fibers per ml (all fibers ≥5 µm), but usually the concentration was much lower. The observed and expected deaths and SMR for all men who had worked one year or longer from 1952–1975 are given in Table 5. There was no relationship between the mortality observed and lifetime exposure to silica dust (that was as high as 1,000 mg/M3 × years). There was no suggestion that deaths from cancer increased after 10 or 20 years of latency. No deaths from mesothelioma or asbestosis were reported.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

Table 5. Selected causes of mortality for men who worked one year or longer for the Reserve Mining Company Deaths Cause of death ICD* Expected All causes 000–E999 343.7 Cardiovascular disease 402, 404, 410–429 123.8 Cancers All 140–209 63.4 Respiratory 160–163 17.9 Digestive 150–159 17.6 Urinary 188–189 3.0 Genital 180–187 3.3 Selected nonmalignant respiratory diseases 470–474, 480–486, 490, 491, 493, 510–519. 6.8 All external causes E800–E998 72.8 E810–E823 31.2 Motor vehicle accidents

3416

Observed 298 112

SMR† 87 90

58 15 20 3 3 4 76 38

92 84 114 101 91 59 104 122

Source: Higgins et al. (1983) *International Classification of Causes of Death, 8th Revision. †Standardized mortality ratio, based on white male mortality in Minnesota, 1952–1976.

Minnesota Taconite Miners. A second epidemiological study of Minnesota taconite workers employed at the Erie and Minntac mines was reported (22). The study cohort, followed from 1947–1988 with a minimum observation period of 30 years for all participants, was made up of 3,341 men, of which 1,058 were deceased. Dusts in the two mines are reported as containing 28–40% and 20% quartz at Erie and Minntac mine, respectively. Concentrations of fibrous particulates were nearly always <2 fibers per ml (all fibers ≥5 µm). These fibrous particulates included elongate cleavage fragments and are assumed to be similar to those objects reported at Reserve Mining. The total number of deaths was significantly fewer than expected, SMR = 83 (based on U.S. male rates) and 91 (based on Minnesota male rates). SMR for all cancer (including lung cancer), diseases of the circulatory system, and nonmalignant respiratory disease were fewer than expected when compared with both reference groups (Table 6). There was one reported case of mesothelioma in a 62-year-old worker whose exposure to taconite had begun only 11 years before his death. Although latency periods as short as 15 years have been reported among insulation workers, mesothelioma generally occurs following a long latency period of 25 years or more (23). This person had previously been employed in the railroad industry, as a locomotive fireman and engineer, an occupational environment where both amosite and crocidolite asbestos insulation was used and opportunity for exposure existed (12). It is unlikely that this particular taconite exposure contributed to the appearance of mesothelioma. Analysis of the mortality data, with a minimum latency period of 30 years, provided no evidence to support any association between exposure to quartz or elongated cleavage fragments of amphibole with lung cancer, nonmalignant respiratory disease, or any other specific disease. Comparison of Occupational Cohorts Exposed to Iron Ore and Asbestos. The American and Canadian asbestos insulation workers are generally thought to have had exposure to the three principal commercial asbestos fiber types—grunerite asbestos, crocidolite, and chrysotile (12). The tremolite asbestos in the vermiculite at Libby. Montana has never been extensively used in commerce in the United States. The vermiculite workers are an example of the effect of amphibole asbestos at concentrations of ≈1% in the ore. The mortality experience of the two asbestos-exposed groups are distinctly similar. Each shows an elevated risk of lung cancer, mesothelioma, and asbestosis (a nonmalignant respiratory disease). Of the 1,058 deaths reported in the most recent study of Minnesota taconite workers, one would have expected about 250 lung cancer (23.6%) and about 98 mesotheliomas (9.3%) if their mortality experience was similar to American and Canadian insulators (11). Instead, the actual number of lung cancer and mesotheliomas (Table 6) was 65 (6.1%) and 1 (0.09%), respectively. Actually 32 fewer lung cancer occurred than the 97 expected (SMR = 67) using the rates for U.S. white males. The one mesothelioma that did occur had a latency of ≈11 years in taconite mining. In the large insulation cohort (17,800 workers), no mesothelioma was reported with a latency <15 years, indicating the present case was unlikely to be related to his taconite dust exposure (11, 23). The mortality experience of the iron ore workers is, in fact, overall less than expected, indicating they are healthier than the general population. This healthy workers effect is commonly observed among many employed groups. Epidemiological and Lung Content Analysis of Grunerite Asbestos-Exposed Workers. Before the United States entering Table 6. Deaths by major causes (1948–1988) in taconite miners and millers exposed for 3 months or more before 1959 Deaths Cause of death (ICD, 7th Revision, 1955) Expected Observed All causes (001–998) 1,272.5 1,058 All malignant neoplasms (140–205) 267.7 232 Digestive organs and peritoneum (150–159) 70.5 66 Stomach (151) 12.0 11 Large intestine (153) 23.9 26 Respiratory system (160–164) 97.0 65 Bronchus, tracheas, lung (162–163) 92.2 62 Kidney (180) 6.8 12 Lymphopoietic (200–205) 25.8 29 All diseases of circulatory system (400–468) 575.1 477 Arteriosclerotic heart disease (420) 481.8 368 Cirrhosis of liver (581) 35.5 24 Nonmalignant respiratory disease (470–527) 77.2 55 All external causes of death (800–998) 112.3 114 All accidents (800–962) 74.4 79 Motor vehicle accidents (810–835) 33.4 32 Suicide (963, 970–979) 27.3 32 Cause unknown 19 Number of workers 3,431 Number of person-years 10,055 Deaths per 1,000 person-years 10.5 +1.8% Adjustment of cause-specific SMRs for missing Certificates

Source: Cooper et al. (22)

SMR 83 87 94 92 109 67 67 177 112 83 76 68 71 102 106 96 117

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

Table 7. Fiber count (given in millions of fibers per gram of dried lung tissue) by type of pathology Pathology Total Grunerite Asbestos (Amosite) Mean SD Mean SD Lung Cancer 1,483 2,568 1,433 2,590 Mesothelioma 1,035 1,039 1,000 1,013 358 490 297 463 Other

3417

Number of Cases 14 5 24

Source: Gibbs et al. (25). World War II, a grunerite asbestos factory was established in Paterson, New Jersey to supply the U.S. Navy with asbestos insulation for the pipes, boilers, and turbines in ships. From 1941–1945, 933 men were recruited to work in this plant, which operated until November 1954. Of these, 820 men formed a cohort and provided a unique group of individuals with an intense short-term exposure and a long-term follow-up (24). Among these individuals, no mesotheliomas occurred with less than a 6-month exposure history or a latency of <20 years. Although the concentration of asbestos fibers in the air of the Paterson plant was never determined, few occupational health experts would estimate the exposure at <30 fibers per ml (all fibers ≥5 µm). Therefore, 6 months of work at the plant is equivalent to 15 fibers per ml × years. The mean fiber levels in the iron ore mine are 0.05 fibers per ml. Therefore, it would require about 300 years of exposure in the iron ore mine to reach the 15 fiber per ml × years level. For the workers in the Paterson plant the concentration of grunerite asbestos present in the lung tissue of any individual with an asbestosrelated disease has not been reported. However, in a report about workers in a British grunerite asbestos factory, lung tissue taken at autopsy from 14 lung cancer and 5 mesothelioma cases were examined for fiber levels (25). The mineral fibers were separated from the lung tissue and analyzed by using ATEM. Although the factory principally used grunerite asbestos, a small amount of chrysotile had also been used. Of the 43 cases in which sufficient tissue was available for fiber analysis, grunerite asbestos was present at a 20-fold higher concentration than the three other commercial asbestos fiber types. In both the lung cancer and mesothelioma cases, ≈97% of the total fiber burden was grunerite asbestos (Table 7). The mean fiber concentration was about 1.483 × 109 and 1.035 × 109 fibers per gram of dry lung tissue for lung cancer and mesothelioma, respectively. The mean fiber concentration was ≈45% higher in the lung cancer cases than in the mesothelioma cases. Assuming the total dry weight of an average pair of human lungs to be ≈150 gm, the mean total concentration of fiber in the five mesothelioma cases would be 1.5 × 1011 fibers (25). The mean fiber concentration in the air of the iron mine was 0.05 fibers per ml (all fibers ≥ 5µm). The fiber number in the lung tissue represents fibers of all lengths, whereas the air data is only for those ≥5 µm. The 0.05 fibers per ml (all fibers ≥5 µm) represents an index of the fibers present in the air. The fibers <5 µm and ≥5 µm but too thin to be visible by phase-contrast microscopy were not counted. One method to approximate the total number of fibers per ml is to interpolate from data where the total size distribution of grunerite asbestos has been reported, as at the Penge Mine in the Republic of South Africa (26). Using the length and diameter data from Penge and assuming 0.05 fibers per ml represents the fibers ≥5 µm in lengths and ≥0.25 µm in diameter, a multiplication factory of 6.2 was interpolated. The total fiber concentration in the iron mine is therefore assumed to be 0.05 fibers per ml × 6.2, or 0.33 fibers per ml (all fibers). A second method is to add the fiber counts of 11 air samples from the mine analyzed by phase-contrast optical microscopy and ATEM to estimate total exposure. When the two values were added, the mean exposure was 1.18 ± 0.57 fibers per ml (all fibers). The exposure is 3.6-fold greater than that estimated by using the size distribution of grunerite asbestos in the Penge mining environment, although the mean exposure for the 11 air samples was 0.08 ± 0.05 fibers per ml (all fibers ≥5 µm), which exceeds the average of the 179 personal air samples of 0.05 ± 0.05 fibers per ml (all fibers ≥5 µm). All of the grunerite asbestos fibers counted by ATEM were <5 µm long. To inhale a concentration of fibers similar to the concentration in the lung tissue of the mesothelioma cases (1.5 × 1011 fibers) would require inhaling 4.7 × 1011 ml of air in the iron ore mine, assuming an exposure of 0.33 fibers per ml. For the purpose of this model, we pessimistically assume no clearance, although the lung has mechanisms to clear inhaled particles that can be very effective. Assuming on average an individual inhales 10,000 ml of air per minute, this is 600,000 ml per hour, or 4,800,000 ml per 8-hour shift. This seems a very large number, but it would require ≈98,000 days in the iron ore mine with an exposure of 0.33 fibers per ml (at 1.18 fibers per ml exposure, it would require 27,000 days) just to inhale a similar number of fibers to that found in the only series of lung content analysis of grunerite asbestosrelated mesotheliomas. The range is 75–265 years of daily 8-hour shifts of exposure to inhale a similar number of fibers to that found in the lung tissue of the factory mesothelioma cases. Table 8. Risk of death in a lifetime for some selected environmental exposures Activity Heavy cigarette smoking All causes of death Lung cancer only Total U.S. motor vehicle accidents All deaths Pedestrian deaths U.S. air pollution (calculated deaths from assumed correlation) Frequent airline passenger, 200,000 miles for 35 years Accident Cosmic ray cancers U.S. natural radiation background at sea level (cancers) excluding radon gas U.S. home deaths All Falls (mostly over age 65) Drowning deaths (nontransport causes) Diagnostic x-ray in USA (cancer) Person living with a smoker (cancer) Person in brick building, added natural radiation One transcontinental round-trip flight per year Accident Cosmic rays Upper level of risk EPA claims to regulate Falling meteorite Drinking water with 100 mg/ml choloroform (EPA level) Eating 1.1 kg charcoal-broiled steak per week (cancer only) World Health Organization (1974) acceptable risk for drinking water Struck by falling airplane (average over entire U.S.)* Smoking three cigarettes in a lifetime (all deaths)* Lower level of risk EPA claims to regulate* Lightning*

Lifetime risk per 100,000 35,000 9,000 23,000 1,200 100 2,000 400 300 200 600 200 80 200 100 70 15 15 15 5 5 2 1 0.4 0.3 0.1 0.1–0.15

*Those activities with lifetime similar to the lung cancer and mesothelioma risk calculated for the iron ore miners. All other activities listed pose a higher risk.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

Table 9. Lifetime risk for 1-year continuous exposure per 100,000 people for 0.001 fibers per ml Age on onset of exposure Average* Lung Cancer Risk Mesothelioma Risk Nonsmoker Smoker 30 0.31 0.06 0.62 0.21 45 0.25 0.05 0.50 0.03 50

3418

Total Cancer Risk Nonsmoker 0.27 0.1 0.08

Smoker 0.83 0.6 0.53

*From Table 6-3 EPA (1986) for men.

Risk Assessment from Mining in the Iron Ore Mine. In the past, workers were exposed to aerosols containing high concentrations of asbestos fibers. To obtain a quantitative risk estimate from the low exposures, we used a model developed for the Environmental Protection Agency to quantify the risk of asbestos-related disease (27). This model is developed to fit the type of data described above, the exposures during mining of the iron ore are orders of magnitude lower than the occupational exposures which occurred in the cohorts used to parameterize the dose component in the equations of the risk models. Nonetheless, the high exposure-response relationships of the past were used to interpolate the risk to the current low exposures encountered in the iron ore mine in linear (proportional) relationships. We know of no scientist who has argued that this linear dose-response model underestimates the risk. The risk assessment model requires that the concentrations of asbestos fibers in the air be determined. Risk assessment is based on counting all fibers ≥5 µm in length in the occupational environment by phase-contrast microscopy, at ≈×500 magnification (Table 2). Risk estimates were considered for the following two scenarios: (i) A bench containing approximately 1 million tons of rock was removed in 22 days. Assuming the average employee is 45 years old, what is the lifetime risk for lung cancer and mesothelioma? No air sampling was done at that site, and it is uncertain whether any asbestos exposure took place. Assume the fiber levels are similar to those given in Table 2. (ii) Approximately 30 days of drilling remain to be done on the bench containing the seam of grunerite asbestos (28 days in the sill and two days in the waste iron formation). Assuming the sill contains no asbestos (so far none has been found), what would be the lifetime risk to the drillers for lung cancer and mesothelioma assuming they are 45 years old? Table 6-3 from the EPA risk model (27) was used. This table is for an exposure to a concentration over a long time. It can be used for a 2or 22-day exposure if it is assumed that the exposure integrated over time is the relevant parameter. (i) There is a linear dose-response relationship. Any proposed biological mechanism of which we are aware involves the exposure integrated over time. (ii) If the peak exposure is the parameter of concern, the risk is proportional to the frequency of peak exposures. The integrated exposure is also proportional to the total time of possible exposure and goes down with time. The average lung cancer risk among smokers and nonsmokers was reported by the EPA. The risk number found in the EPA Table 6-3 is the average for smokers and nonsmokers, but the actual lung cancer risk from asbestos exposure is five times less for nonsmokers and double for smokers. Because mesotheliomas are assumed not to be related to smoking, the number applies to both smokers and nonsmokers. Exposure. The average of the exposures monitored is appropriate for calculating the risk to a worker not otherwise identified. The mean airborne concentration of 179 personal air samples was 0.05 fibers per ml (all fibers ≥5 µm) (Table 2). This value assumes all the fibers were asbestos and that each person was continuously exposed (8-hour time-weighted average) over a 22-day period. The EPA calculated for continuous exposure over different periods of time, and therefore the iron ore mining exposure is converted to be equal to the exposure average over 1 year, <E>. <E> = 22/365 × 8/24 × 0.05 = 0.01 fibers per ml (all fibers ≥5 µm). The life-time risk can be read directly from Table 6-3 (27) at 30 and 50 years of age at onset of exposure (45 years of age is interpolated) (Table 9). Scenario I. The total cancer risk for the individual exposure beginning at 45 years of age is 0.1 and 0.6 in 100,000 for nonsmokers and smokers, respectively (see Table 8 for comparison with selected different lifestyles and environmental exposures). This assumes a linear doseresponse. If all of the cancer risk is assumed to be lung cancer, it is equivalent to smoking 2 or 12 cigarettes in a lifetime for 0.1 and 0.6 in 100,000 people respectively. The risk for someone smoking one cigarette is 0.05 per 100,000 people (or, smoking 2 cigarettes is associated with a lung cancer risk of 1 in 1 million). Scenario II. In this scenario, there will be a 2-day exposure (not the 22-day of Scenario I), so the risk becomes 2/22 or 1/11 of the risk of Scenario I (0.1 in 1,000,000 for nonsmokers, and 0.6 in 1,000,000 for smokers) (Table 10). These are risks accumulated in a lifetime. Note also that according to the assumption pertaining to the risk calculation; each new exposure adds to this risk independent of the past risk. Of course, if asbestosis is a precondition for lung cancer, there exists a lung cancer threshold (28, 29). Although new exposures can add to past ones, they only increase the risk where the total exposure exceeds the threshold. That the EPA model overestimates the risk of lung cancer is widely believed (30). Although the above is a best estimate, an important consideration is how much larger could the risk be to that individual. An examination of Table 2 indicates the extreme exposure level of 0.39 fibers per ml (all fibers ≥5 µm) was seven times larger than the mean 0.05 fibers per ml (all fibers ≥5 µm). This suggests the most extreme risk is seven times greater than given above. These risks are put into perspective in Table 8. We thank Mr. Paul Nordstrom for providing the survey of the bench containing grunerite (amosite) asbestos. We acknowledge support from a Higher Education Advanced Technology grant from the State of New York and Cleveland-Cliffs, Inc. Table 10. Lifetime risk for two mining scenarios in the iron ore mine compared to selected relative risks* Activity Lifetime risk per 100,000 Iron Ore Mining Scenario I Nonsmokers 0.1 Smokers 0.6 Scenario II Nonsmokers 0.01 Smokers 0.06 Lung cancer heavy cigarette smoking 9,000 US motor vehicles, all deaths 1,200 Drowning deaths, nontransport caused 80 Upper limit of risk EPA claims to regulate 15 *See Table 8 for additional comparison.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

A RISK ASSESSMENT FOR EXPOSURE TO GRUNERITE ASBESTOS (AMOSITE) IN AN IRON ORE MINE

3419

1. Murray, R. (1990) Br. J. Ind. Med. 47, 361–365. 2. Merewether, E. R. A. (1954) Arch. Hyg. Rada. 4, 365–382. 3. Doll, R. (1955) Br. J. Ind. Med. 12, 81–86. 4. Wagner, J. C., Sleggs, C. A. & Marchand, P. (1960) Br. J. Ind. Med. 17, 260–271. 5. McDonald, J. C. & McDonald, A. D. (1996) Eur. Respir. J. 9, 1932–1942. 6. Nolan, R. P., Langer, A. M. & Addison, J. (1994) Environ. Health Pespect. 102, Suppl. 5, 245–250. 7. Veblen D. R. & Wylie A. G. (1993) in Health Effects of Mineral Dusts, eds. Guthrie, G. D. & Mossman, B. T. (Mineralog. Soc. Am., Washington, D.C.), pp. 61–137. 8. Langer, A. M., Nolan, R. P. & Addison, J. (1991) in Mechanisms in Fibre Carcinogenesis, eds. Brown, R. C., Hoskins, J. A. & Johnson, N. F. (Plenum, New York), pp. 253–267. 9. Nolan, R. P., Langer, A. M., Oechsle, G. W., Addison J. & Colflesh, D. E. (1991). in Mechanisms in Fibre Carcinogenesis, eds. Brown, R. C., Hoskins, J. A. & Johnson, N. F. (Plenum, New York), pp. 231–251. 10. Ross M., Kuntze R. A. & Clifton R. A. (1984) Special Technical Publication 834 (Am. Soc. Testing Mat., Philadelphia), pp. 139–147. 11. Selikoff, I. J. & Seidman, H. (1991) Ann. N.Y. Acad. Sci. 647, 1–14. 12. Langer, A. M. & Nolan, R. P. (1998) Monaldi Arch. Chest Dis. 53, 168–180. 13. McDonald, J. C., McDonald, A. D., Armstrong, B. & Sebastien, P. (1986) Br. J. Ind. Med. 43, 436–444. 14. McDonald, J. C., Sebastien, P. & Armstrong, B. (1986) Br. J. Ind. Med. 43, 445–449. 15. Amandus, H. E., Wheeler, R., Jankovic, J. & Tucker, J. (1987) Am. J. Ind. Med. 11, 1–14. 16. Amandus, H. E. & Wheeler, R. (1987) Am. J. Ind. Med. 11, 15–26. 17. Amandus, H. E., Althouse, R., Morgan, W. K. C., Sargent, N. & Jones, R. (1987) Am. J. Ind. Med. 11, 27–37. 18. Health Effects Institute-Asbestos Research (1991) Asbestos in Public and Commercial Buildings: A Literature Review and Synthesis of Current Knowledge (Health Effects Inst., Cambridge, MA ). 19. Higgins, I. T. T., Glassman, J. H., Oh, M. S. & Cornell, R. G. (1983) Am. J. Epidemiol. 118, 710–719. 20. Schaumberg, F. D. (1976) Judgement Reserved (Reston Publishing Reston, VA), pp. 1–265. 21. Langer, A. M., Maggiore, C. M., Nicholson, W. J., Rohl, A. H., Rubin, I. B. & Selikoff, I. J. (1979) Ann. N.Y. Acad. Sci. 330, 349–372. 22. Cooper, W. C., Wong, O., Trent, L. S. & Harris, F. (1992) J. Occup. Med. 34, 1173–1180. 23. Liddell, D. (1988) Proceedings of the Symposium on Health Aspects of Exposure to Asbestos in Building (Harvard Univ. Press, Cambridge, MA) pp. 47– 68. 24. Seidman, H., Selikoff, I. J. & Hammond, E. C. (1979) Ann. N.Y. Acad. Sci. 330, 61–89. 25. Gibbs, A. R., Gardner, M. J., Pooley, F. D., Griffiths, D. M., Blight, B. & Wagner, J. C. (1994) Environ. Health Persp. 104, Suppl. 5, 261–263. 26. Pooley, F. D. & Clark, N. J. (1980) in Biological Effects of Mineral Fibres, ed. Wagner, J. C. (Int. Agency Res. Cancer, Lyon, France), Vol. 1, pp. 79–86. 27. U.S. Environmental Protection Agency (1986) Airborne Asbestos Health Assessment Update. EPA/ 600/8.84/003F, pp. 198. 28. Weiss, W. (1999) Chest 115, 536–549. 29. Hughes, J. M. & Weill, H. (1991) Br. J. Ind. Med. 48, 229–233. 30. Camus, M., Siemiatycki, J. & Meek, B. (1998) N. Engl. J. Med. 338, 1565–1571.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

POTENTIAL EFFECTS OF GAS HYDRATE ON HUMAN WELFARE

3420

Potential effects of gas hydrate on human welfare

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3420–3426, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. KEITH A. KVENVOLDEN* PNAS is available online at www.pnas.org.

U.S. Geological Survey, 345 Middlefield Road, MS999, Menlo Park, CA 94025 ABSTRACT For almost 30 years, serious interest has been directed toward natural gas hydrate, a crystalline solid composed of water and methane, as a potential (i) energy resource, (ii) factor in global climate change, and (iii) submarine geohazard. Although each of these issues can affect human welfare, only (iii) is considered to be of immediate importance. Assessments of gas hydrate as an energy resource have often been overly optimistic, based in part on its very high methane content and on its worldwide occurrence in continental margins. Although these attributes are attractive, geologic settings, reservoir properties, and phase-equilibria considerations diminish the energy resource potential of natural gas hydrate. The possible role of gas hydrate in global climate change has been often overstated. Although methane is a “greenhouse” gas in the atmosphere, much methane from dissociated gas hydrate may never reach the atmosphere, but rather may be converted to carbon dioxide and sequestered by the hydrosphere/biosphere before reaching the atmosphere. Thus, methane from gas hydrate may have little opportunity to affect global climate change. However, submarine geohazards (such as sediment instabilities and slope failures on local and regional scales, leading to debris flows, slumps, slides, and possible tsunamis) caused by gas-hydrate dissociation are of immediate and increasing importance as humankind moves to exploit seabed resources in ever-deepening waters of coastal oceans. The vulnerability of gas hydrate to temperature and sea level changes enhances the instability of deep-water oceanic sediments, and thus human activities and installations in this setting can be affected. The potential effects of gas hydrate on human welfare are not understood with certainty, but enough information has been collected and enough knowledge gained over the past 30 years to make preliminary assessments possible. To make these assessments, however, some geoscience background is necessary. Definition. Naturally occurring gas hydrate is a solid, ice-like substance, composed of rigid cages of water molecules that enclose molecules of gas, mainly methane. Chemically, this substance is a water clathrate of methane, where “water clathrate” refers to the rigid cage structure of hydrogen-bonded water molecules, but is commonly called “methane hydrate” or, in general terms, “gas hydrate.” The maximum amount of methane is fixed by the geometry of the clathrate. In an ideally saturated methane hydrate, the molar ratio of methane to water is 1:5.75, that is equal to a volumetric ratio at standard conditions of methane gas to water of 216:1 or a volumetric ratio of methane gas to solid hydrate of 164:1 (1). Occurrence. Gas-hydrate deposits occur under specific conditions of pressure and temperature, where the supply of methane is sufficient to initiate the formation of, and to stabilize, the hydrate (clathrate) structure (2). These conditions exist on earth in the upper 2,000 m of sediments in two regions: (i) continental, including continental shelves, at high latitudes in polar regions, where surface temperatures are very cold (<0°C) and (ii) submarine continental slopes and rises, where not only is bottom water cold (≈0°C) but also pressures are high (>3 MPa). Thus, in polar regions, gas hydrate is found where temperatures are cold enough for onshore and offshore permafrost to be present. In offshore sediment of outer continental and insular oceanic margins, gas hydrate is found at water depths >300–500 m, depending on bottom-water temperatures. The presence of gas-hydrate deposits in these oceanic margins has been inferred mainly from the appearance on marine seismic profiles of an anomalous reflection (Fig. 1) that coincides with the predicted boundary (based on assumed pressure/temperature considerations) of the base of the gashydrate stability zone (Fig. 2). This reflection is commonly called a bottom-simulating reflection (BSR) because it approximately mimics sea floor topography. BSRs have been mapped at depths below the sea floor, ranging from near the sea floor to ≈1,100 m (5); the upper limit of the gas-hydrate zone in outer continental margin sediment is ordinarily the sea floor. Gas-hydrate samples have been recovered at 27 oceanic continental margin locations (6), providing direct confirmation of gas-hydrate occurrence. The worldwide locations of known and inferred gas hydrate are shown in Fig. 3. Estimates of Methane Content. Chersky and Makogon (8) proposed that the amount of methane in naturally occurring gas hydrate is potentially “enormous,” but the estimated amounts were highly speculative because of incomplete knowledge of gas-hydrate occurrence. The Potential Gas Committee (9) summarized the early estimates for the world: methane in gas-hydrate deposits ranging from 3.1 × 1015 to 7,600 × 1015 m3 for oceanic sediments and from 0.014 × 1015 to 34 × 1015 m3 for permafrost regions. Because oceanic gas hydrate apparently contains significantly more methane, it is emphasized in global estimations of the methane content of gas hydrate. The upper limit estimates above are from Dobrynin et al. (10) and appear to be overly optimistic. They are “rough estimates” based on permafrost coverage and zones of gashydrate stability in oceanic sediments without apparent regard for distributions of sedimentary basins or sources of methane. Estimates made during the period from 1980 to 1990 of the amounts of methane in oceanic sediments were summarized by Kvenvolden (11). During this decade, an increased understanding of gas-hydrate occurrence has generally resulted in estimates within the lower ranges of previous ones (Fig. 4). For example, Kvenvolden (7) estimated the methane content of global gas-hydrate occurrence at 21 × 1015 m3. The calculated amount of gas hydrate in the outer continental margin of the Arctic Basin of 1.1 × 1015 m3 (12) was extrapolated to outer continental margins of the remainder of the world by multiplying by 20 because the length of the Arctic margin is about 5% of the total length of continental margins worldwide.

Abbreviations: GCM, general circulation model; BSR, bottom-simulating reflection; LPTM, latest Paleocene thermal maximum. * To whom reprint requests should be addressed. e-mail: kk@octopus. wr.usgs.gov.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

POTENTIAL EFFECTS OF GAS HYDRATE ON HUMAN WELFARE

3421

FIG. 1. A 12-fold multichannel seismic reflection profile from the crest and eastern flank of the Blake Outer Ridge. The strong BSR is inferred to represent the base of the gas hydrate stability zone. Modified from ref. 3. By using a completely independent approach in which the organic carbon content of sediment was used as one guide, MacDonald (13) estimated the amount of methane in worldwide gas-hydrate deposits at 21 × 1015 m3. This estimate neglects any gas hydrate in sediment at water depth over 3,000 m. That the estimates of Kvenvolden (7) and MacDonald (13) are equal at 21 × 1015 m3 is a coincidence, but the convergence of ideas has made this value the consensus estimate. This consensus value is bounded at an upper limit of 40 × 1015 m3 determined by Claypool (14) and a lower limit of 10 × 1015 m3 determined by Makogon (cited in ref. 2) in the same decade. In the 1990s, Gornitz and Fung (15) and Harvey and Huang (16) used plausible ranges of relevant variables to provide estimates of possible methane release under various general circulation model (GCM) predictions of global climate change. The issue here, however, is not global climate change, but rather the amount of methane in gas hydrate, which serves as the basis for considerations of global climate change. Both approaches included 1° × 1° grid resolution of some of the variables. In the case developed by Gornitz and Fung (15), estimates of the amount of hydrate methane were made from pressure–temperature phase relations and a plausible range of thermal gradients, sediment porosities, and pore fillings, based on two theories of gas-hydrate formation: (i) in situ microbial gas generation and (ii) pore fluid expulsion models. A proxy for organic carbon (the ultimate source of methane) was cleverly obtained by using data from a coastal zone color scanner, which measures oceanic photosynthetic pigments and is indirectly related to primary production. The calculated ranges of amounts of methane in oceanic gas hydrate are between 26 × 1015 and 139 × 1015 m3, with the most likely values at the lower end of the range, i.e., 26 × 1015 m3 for the in situ microbial gas generation model and 115 × 1015 m3 for the pore fluid expulsion model. These estimates lie within the range of early values but are higher than the consensus estimate of 21 × 1015 m3 (7, 13).

FIG. 2. Phase diagram showing boundary between free methane gas (no pattern) and methane hydrate (pattern) for a pure water and pure methane system. The addition of salts, such as NaCl, to water shifts the curve to the left. Adding CO2, H2S, C2H6, or C3H8 to methane (CH4) shifts the boundary to the right, thus reducing the pressure for gas-hydrate stability at a given temperature. Depth scale assumes lithostatic and hydrostatic pressure gradients of 10.1 kPa·m−1. Redrawn after Katz et al. (4). Harvey and Huang (16) used data on ocean depths, temperature at the sediment-water interface, ocean sediment thermal conductivity, and the geothermal heat flux on a 1° × 1° global grid to estimate the volume of the zone of gas-hydrate stability in oceanic sediments. They calculated the total methane content in gas hydrate in this volume of sediment to be 23 × 1015, 46 × 1015, and 91 × 1015 m3, depending on the assumptions regarding the pore fractions occupied by gas hydrate. They selected the intermediate value (46 × 1015 m3) as the best estimate. Other global estimates of hydrate methane were made during this same time period and resulted in values smaller than the consensus estimate of 21 × 1015 m3. Holbrook et al. (17), using mainly seismic information from the Blake Ridge (Ocean Drilling Program Leg 164), concluded that global estimates of methane in gas hydrate are too high by as much as a factor of 3, which calculates to about 7 × 1015 m3. This conclusion was discussed by Dickens et al. (18), who used mainly direct measurements of in situ methane abundances stored in gas hydrate and free gas in sediment from the same area, i.e., the Blake Ridge. They concluded that global estimates of methane in gas hydrate in the range from ≈2 × 1015 to 20 × 1015 m3 are acceptable, based on current knowledge. Other low estimates of the methane content of worldwide gas hydrate include 10 × 1015 m3 by Makogon (cited in ref. 2), a value which has now been revised to 15 × 1015 m3 (19) and 1 × 1015 m3 by Ginsburg and Soloviev (20), who challenge all larger estimates. Table 1 summarizes the critical worldwide estimates of hydrate methane, and Fig. 4 shows how the magnitude of these estimates has changed over about two decades. It is evident that by 1990, the range of estimates had been greatly constrained from estimates available in 1980. However, since 1990, the range of estimates has expanded slightly, but the consensus value of 21 × 1015 m3 remains about midway between the extremes. It is quite likely that the global amount of hydrate methane is considerably less than 1017 m3 but probably is greater than 1015 m3, with the actual value in the lower or intermediate part of the range. Gas Hydrate and Human Welfare. Chemists have known about gas hydrate since the early part of the 19th century (2). The petroleum industry became aware of this substance in the 1930s when gas-hydrate formation was discovered to be the cause of pipeline blockage during transmission of natural gas (23). In the 1960s, naturally occurring gas hydrate was found in the Siberian Messoyakha gas field (19), and in the 1970s it was recognized that gas hydrate occurs naturally not only in polar continental regions but also in shallow sediment under deep water of the oceanic outer continental margins (24, 25). Since a review by Kvenvolden and McMenamin (5) of the geological occurrences of natural gas hydrate, it has become increasingly evident that naturally occurring gas hydrate is a significant component of the shallow geosphere and has been postulated by Kvenvolden (26) to be of societal relevance in a least three ways: resource, climate, and hazard. Now the author believes that only the hazard aspect is of immediate importance in considerations of human welfare.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

POTENTIAL EFFECTS OF GAS HYDRATE ON HUMAN WELFARE

3422

FIG. 3. The Earth showing locations of known and inferred gas-hydrate deposits in oceanic sediment of outer continental margins and in permafrost (continental) regions. Modified from ref. 7. Potential Energy Resource. Two factors make gas hydrate attractive as a potential energy resource: (i) the enormous amount of methane that is apparently sequestered at shallow sediment depths within 2,000 m of the surface of the Earth and (ii) the wide geographical distribution of gas hydrate (Fig. 4). According to MacDonald (27), the energy density (volume of methane at standard conditions per volume of sediment) of methane hydrate is 10-fold greater than the energy density of other unconventional sources of gas, such as coal beds, tight sands, black shales, and deep aquifers, and 2- to 5-fold greater than the energy density of conventional natural gas.

FIG. 4. Changing magnitude of estimates of the methane content (×1015 m3) of worldwide gas-hydrate deposits from 1980 to 1998. Given these attractive factors, it is reasonable to conclude that natural gas hydrate could serve as a future energy resource, as suggested recently by Collett and Kuuskraa (28). However, there are some negative factors that suggest that overly optimistic assessments are being made of gas hydrate as a future energy resource. One such factor is the overall petroleum geology of the gas-hydrate deposits. As Levorsen (29) pointed out years ago, an essential element in any oil or gas reservoir is permeability. Without permeability, there can be little gas accumulation, nor can the accumulated gas be produced by drilling, for it cannot move into production wells quickly enough. Low-permeability sediments (mainly clay and claystone) are the principal lithologies in gas hydrate-bearing sections of the Blake Ridge, offshore from the southeastern United States (30). Dillon et al. (31) recognized the low permeability of the sediments in this region and suggested that faults act both as permeability barriers to gas and as conduits for gas. The idea of trying to produce gas from such a system Table 1. Summary of world estimates of methane content ×1015 m3 of oceanic gas hydrate Methane estimate × 1015 m3 Best estimate High, Low Value Before 1988 3.1 — — 5–25 7,600 — After 1988 40 — 10 — 21 — 21 — 26 26–140 46 23–91 1 — 7 — 15 — — 2–20 Estimates are rounded to two significant figures.

Reference McIver (21) Trofimuk et al. (22) Dobrynin et al. (10) Kvenvolden and Claypool (14) Makogon (cited in ref. 2) Kvenvolden (7) MacDonald (13) Gornitz & Fung (15) Harvey & Huang (16) Ginsburg & Soloviev (20) Holbrook et al. (17) Makogon (19) Dickens et al. (18)

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

POTENTIAL EFFECTS OF GAS HYDRATE ON HUMAN WELFARE

3423

of clay and claystone does not appear to be particularly attractive, especially with current technology. This one example certainly does not apply to all regions of oceanic gashydrate occurrence, but it is the best known example discovered and described thus far and warns of the potential petroleum geologic problems that may be found elsewhere. Although there is a large amount of methane in naturally occurring gas hydrate, there is a constraint on this amount that is often not appreciated. Hunt (32) pointed out that the gas-hydrate enrichment factor decreases with depth. That is, the gas-hydrate reservoir can hold, ideally at standard condition, “about six times as much gas as free gas held in the same space.” However, this enrichment factor decreases with depth because free gas compresses considerably with depth and gas hydrate does not. In the example given by Hunt (32), the enrichment factor decreased from 6 to 1.25 in the depth range from 274 to 1,219 m. Thus, with increasing depth, the presence of gas hydrate can be a disadvantage in that a given volume of gas hydrate will contain less gas than could be present if the gas were in a free state. Interest in gas hydrates by the gas industry has waxed and waned over the past 30 years. At present, however, interest appears to be increasing, and certainly nations such as Japan and India, with immediate energy needs, have undertaken major efforts to investigate gas hydrate as an energy source (28). Successful recovery of methane from gas hydrate will require the attention and infrastructure of the gas industry. In fact, the location of this infrastructure is likely to dictate where, if ever, gas hydrate is to be first produced commercially. Although naturally occurring gas hydrate was recognized in the 1960s, the gas industry has been slow to develop methodologies to recover methane from this substance. This slowness is due, in part, to generally abundant gas supplies and a lack of economic incentives leading to recent assessments such as those of Rogner (33), who wrote that “in the foreseeable future, there will be little need for the development of gas hydrates.” Various schemes have been considered at an academic level (34), but application of these schemes and successful results have not been documented. Development of the Messoyakha gas field in Western Siberia during the past 30 years (19) is often cited as an example of successful gashydrate exploitation (35), but even this single example is now questioned (36) because the observed gas hydrate may be secondary, a result of production of conventional gas from this field. Therefore, all of these discouraging factors, such as low permeability sediments, decreasing enrichment factor with depth, lack of sustained gas-industry interest, current limited gas-industry infrastructure at most gas-hydrate locations, and no good field example yet of successful methane production from gas hydrate, diminish the potential for gas hydrate becoming a significant energy resource. As far as human welfare is concerned, methane hydrate as an energy resource is not of immediate interest. Any attempts at full scale production will probably not happen until well into the 21st century (2). Global Climate Change. Methane is an important trace component of the atmosphere, having a concentration of about 6.9 × 1012 m3. This concentration was increasing at a rate of 0.9% yr−1, (37, 38 and 39) until a few years ago, when this rate decreased (40). Because methane is radiatively active, it is a “greenhouse” gas that has a global warming potential 20 times greater than an equivalent weight of carbon dioxide when integrated over 100 years (41). The earth's atmosphere has a wide variety of sources and sinks for methane (42), including methane hydrate. Methane hydrate exists in metastable equilibrium and is affected by changes in pressure and temperature that occur mainly with changes in sea level. The amount of methane that is present in gas hydrate onshore and offshore is perhaps 3,000 times the amount in the present atmosphere; an instantaneous release of methane from this source could have an impact on atmospheric composition and thus on the radiative properties of the atmosphere that affect global climate (13). There are obstacles, however, to methane from gas hydrate ever reaching the atmosphere. Instead of instantaneous release, much methane associated with gas hydrate could vent slowly over geologic time, providing opportunities for its oxidation to carbon dioxide by microbial and chemical processes; the oceans could then act as a sink for the produced carbon dioxide. If any methane did reach the atmosphere, it should react with hydroxyl radicals (42) in about 10 years, unless the supply of radicals is overwhelmed. Thus, the role for methane hydrate in global climate change undoubtedly depends on the rate of methane release, and that rate is largely unknown. Evidence for slow gas release from oceanic sediments is provided by the widespread occurrence of pockmarks on the ocean floor (43) and direct observations of slow release of methane from sea-floor gas hydrate and associated vents in the Gulf of Mexico (44). In contrast, a possible example of rapid release is given by a 700-km2 collapse depression on the crest of the Blake Ridge (45). The amount of methane released from this depression during its formation might have been limited, however, by rehydration of escaping methane caused by a combination of hydrostatic pressure and cold bottom waters reaching the floor of the newly formed feature. Two ideas have been proposed to explain the possible role of methane hydrate in global climate change, but neither of these ideas considered the powerful effect of oxidation on the methane released from gas hydrate. Nisbet (46) suggested that methane from continental gas hydrates contributed to the rapid rise in atmospheric methane at the end of the last major glaciation about 13,500 years ago. In this scenario, polar continental gas-hydrate deposits are destabilized by pressure reduction of melting ice sheets, causing temperature increases because of the released methane. The resulting warming provides a strong positive feedback that amplifies methane emissions and ultimately helps to end the ice age. A different scenario was proposed by Paull et al. (47). They suggested that outer continental margin gas-hydrate deposits release methane during a falling sea level, that is, during global cooling. The resulting decrease in pressure causes this gas hydrate to dissociate. The released methane enhances global warming and triggers deglaciation. Thus, methane derived from outer continental margin gas-hydrate deposits in this scenario is believed to be an important factor in limiting the extent of glaciation during a glacial cycle. As interesting as both of these scenarios are, they are both quite speculative, because it is not even known how gas hydrates behave in the present climate regime. Kvenvolden (48) suggested that gas-hydrate deposits of the polar continental shelves are presently most vulnerable to climate change. These areally extensive shelves, formerly exposed to very cold surface temperatures (−10° to −20°C) have been and are being transgressed by a much warmer polar ocean (≈0°C). The polar shelf surface, therefore, has experienced a +10°C or more change in temperature over at least the past 10,000 years. If this suggestion is correct, then escape of methane from assumed gas-hydrate deposits of polar continental shelves should be observable. To test this idea, methane concentrations in water overlying the Beaufort Sea continental shelf of Alaska have been measured when ice is present and absent during 1990–1995 (Fig. 5). Preliminary results (49) showed that methane concentrations in the Beaufort Sea under the winter ice canopy were greater than when the ice is absent. Continued studies (50, 51) confirmed these observations and showed that the average difference in methane concentrations between ice-covered and ice-free conditions is ≈15 nM; however, carbon isotopic determinations indicated that most of the methane came from coastal microbial processes and not from gas

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

POTENTIAL EFFECTS OF GAS HYDRATE ON HUMAN WELFARE

3424

FIG. 5. Map of a portion of the Beaufort Sea continental shelf offshore from Alaska showing station locations and concentrations of methane (in nM) in water samples collected when ice was present and absent from 1993 through 1995. Based on data from refs. 50 and 51. hydrate. Nevertheless, the Beaufort Sea continental shelf of Alaska appears to be a seasonal source of a very minor contribution of methane to the atmosphere. Although this test failed to show significant release of methane from assumed gas-hydrate occurrence, the test remains inconclusive until similar research is carried out on the broad continental shelves of the Siberian Arctic Ocean. Large excursions in the carbon isotopic record of carbonate in oceanic sediments from the latest Paleocene thermal maximum (LPTM) have been attributed to massive dissociation of gas hydrate (52, 53). Although this idea is attractive in explaining the isotopic record, the processes and consequences of this idea need further examination. With an increase in temperature or a decrease in pressure, gas hydrate dissociates. Slow dissociation of massive amounts of gas hydrate could take place, with the methane becoming trapped. If this trap is breached then the methane release could be very rapid, causing a “blast of gas” as suggested by Dickens et al. (53). Evidence for trapping of gas beneath gas hydrate is provided by the worldwide occurrence of BSRs marking the interface between gas hydrate above and free gas below. As discussed previously, it is the rate of release of trapped methane or the size of the “blast” that is unknown. Also unknown is the effectiveness of the water column in oxidizing the methane as it passes through toward the atmosphere. The amount of methane that can actually reach the atmosphere to affect global climate change is uncertain; however, the oxidized methane, now present as carbon dioxide in the water of the ocean, may cause change in the carbon isotopic record of sediments as observed during the LPTM (52, 53). Excursions in the atmospheric concentration of methane during the Holocene have been noted in ice cores from Antarctica (54). Consideration has been given to the role that gas hydrate might have played in producing these excursions (55), and the conclusion was reached, based on ice cores from both Antarctica and Greenland, that gas hydrate was a relatively small source of methane during at least the later Holocene. The potential impact of methane hydrate on future global warming has been evaluated in a GCM study (16). This study concluded that even for worst case scenarios, the impact on future global warming caused by gas hydrate will be small; the uncertainty in future global warming due to gas-hydrate destabiliztion is smaller than the uncertainty due to warming caused by fossil fuel use. Therefore, the potential role for gas hydrate in global climate change is diminished because of the possibility of rapid oxidation of the released methane to carbon dioxide, thus enhancing the solubility of the methane carbon in the ocean water. Scenarios of past global climate change caused by methane released from gas hydrate are all very speculative; one test in the Arctic of methane release during the current climate cycle failed to show much methane from gas hydrate; one GCM study demonstrated little impact of gas hydrate on future global warming; and records of excursions of methane concentrations during the Paleocene and Holocene, although possibly caused by gas-hydrate dissociation, do not provide compelling evidence that this methane actually affected global climate. With respect to human welfare, it is not immediately important whether gas hydrate has a role in global climate change. To be on the safe side, however, it would be best in the future not to perturb natural gas-hydrate stability by continuing societal practices that enhance global warming. Geologic Hazard. Gas hydrate as a geohazard has been considered in detail previously (56). Before gas hydrate forms in usual geologic settings, vast quantities of methane and water are free to migrate within the interstitial pore spaces of consolidating sediment. During gashydrate formation, methane and water become immobilized as a solid, restricting pore space and retarding the migration of fluids. Solid water (rather than liquid water) occupies the pore spaces, and the sedimentological processes of consolidation and mineral cementation are greatly inhibited, although gas hydrate itself can act as metastable cementation (bonding) agent. The permeability of the sediment to gases and liquids decreases as more gas hydrate forms. Eventually, gas hydrate may occupy much of the pore space within the zone of gas-hydrate stability. Continued sedimentation leads to deeper burial of the gas hydrate. Finally, the gas hydrate will be buried so deeply that temperatures at the base of the stability zone will be reached at which the gas hydrate is no longer stable. The solid gas/water mixture (i.e., the gas hydrate) will become a liquid gas/water mixture. Thus, the basal zone of the gas hydrate becomes underconsolidated and possibly overpressured because of the newly

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

POTENTIAL EFFECTS OF GAS HYDRATE ON HUMAN WELFARE

3425

FIG. 6. Diagram showing the effects of changes in sea level on submarine gas hydrate and the resulting failures and gas release. Adapted from McIver (57). released gas, leading to a zone of weakness (low shear strength, where failure could be triggered by gravitational loading or seismic disturbances), and submarine landslides result (57). The same conditions that cause gas-hydrate dissociation during continued sedimentation can also be brought about by the lowering of sea level or by an increase in bottom-water temperatures. These processes change the in situ pressure or temperature regime. In adjusting to the new pressure/ temperature conditions, the gas hydrates dissociate, producing an enhanced fluidized layer at the base of the gas-hydrate zone. Submarine slope failure can follow, giving rise to debris flows, slumps, slides, and collapse depressions such as described by Dillon et al. (45). Failure would be accompanied by the release of methane gas at least into the water column, but much of the methane is likely to be oxidized unless the gas release is catastrophic. A scenario illustrating submarine slope failure is shown in Fig. 6. The possible connection between gas-hydrate boundaries and submarine slide and slump surfaces was first recognized by McIver (58), and several possible examples were described later. These examples include surficial slides and slumps on the continental slope and rise of West Africa (59), slumps on the U.S. Atlantic continental slope (60), large submarine slides on the Norwegian continental margin (61, 62), sediment blocks on the sea floor in fjords of British Columbia (63), and massive bedding-plane slides and rotational slumps on the Alaskan Beaufort Sea continental margin (64). Periodic Pleistocene eustatic sea level transgressions and regressions provide mechanisms to account for the waxing and waning of submarine gas hydrate (64). For example, during the last Pleistocene regression, sea level lowered approximately 100 m between about 28,000 and 17,000 years before present, resulting in a reduction of total stress acting on the seafloor. The reduction in the total pressure initiates dissociation at the base of the gas hydrate, releasing excess methane and water. Failure follows on moderate slopes unless the increased fluid pressures can be adequately vented. On the Beaufort Sea continental slope is a zone of massive slides and slumps that coincides with a region of sediment inferred, from seismic reflection studies, to contain gas hydrate. Fluctuations in global climate, reflected in Pleistocene sea-level lowerings, likely caused these submarine slides and perhaps caused other slides on other continental margins where gas hydrate is present (64). These submarine disruptions of the seafloor, caused by gas-hydrate dissociation, impact human welfare if humanmade structures are located in regions of potential failure. As humankind expands its interest in the seafloor at increasing water depth, such as in the petroleum industry's search for oil and gas, stability of the seafloor becomes increasingly important for any engineering structures. The potential vulnerability of engineering structures to gas-hydrate dissociation in oceanic sediments has been recently recognized and described (65, 66). Risks to drilling and production through gas hydrate-bearing sediment have been addressed by Yakushev and Collett (67) for gas hydrate in Arctic regions, and these same concerns, such as casing collapse, gas leakage outside the conductor casing, and gas blowouts, will be applicable to gas hydrate of deep oceanic regions, only the problems will likely be more severe, but not necessarily as severe as suggested by Bagirov and Lerche (68), who discuss possible gas-hydrate hazards in the Caspian Sea. The geohazard aspects of gas hydrate provide an additional constraint on exploiting oceanic gas hydrate as a future energy resource. CONCLUSIONS The amount of methane sequestered in gas hydrate is undoubtedly very large, probably >1015 m3, but considerable <1017 m3. How this methane can or will affect human welfare is not yet defined. There is much current interest in gas hydrate as a potential (i) energy resource, (ii) factor in global climate change, and (iii) submarine geohazard. Of these three issues, only the third is considered here to be important at the present time for human welfare. It is argued that gas hydrate as a future energy resource has received overly optimistic assessments because of inadequate evaluation of the reservoir qualities of the geologic settings in which oceanic gas hydrate is found. Vacillating interest by the gas industry and the absence of good industrial examples of successful methane production from gas hydrate diminish enthusiasm for gas hydrate as a potential energy resource. It is also argued that gas hydrate is likely not a major factor in global climate change in that much of the methane that could be released from the dissociation of gas hydrate is probably oxidized to carbon dioxide, which dissolves in the water; most hydrate methane never reaches the atmosphere where it could function as a powerful “greenhouse” gas. Scenarios of climate change caused by methane release from gas hydrate in the geologic past are all speculative, and one test of methane release from gas hydrate in the current climate cycle failed. Gas hydrate may be an agent in global change (for example, altering the isotopic record of oceanic sediments), but not necessarily in global climate change. The evidence seems clear that gas hydrate is a geohazard, particularly in the oceans. As a geohazard, gas hydrate will affect human welfare as humankind moves to exploit the seafloor at ever increasing water depths. Human activities and installations in regions of gas-hydrate occurrence must take into account the presence of gas hydrate and deal with the consequences of gas-hydrate dissociation. 1. Davidson, D. W., El-Defrawy, M. D., Fulgem, M. O. & Judge, A. S. (1978) in Proceedings of the 3rd International Conference on Permafrost, National Research Council of Canada, Vol. 1, pp. 938–943. 2. Sloan, E. D. (1998) Clathrate Hydrates of Natural Gas (Dekker, New York), 2nd Ed. 3. Shipley, T. H., Houston, M. H., Buffler, R. T., Shaub, F. J., McMillen, K. J., Ladd, J. W. & Worzel, J. L. (1979) Am. Assoc. Pet. Geol. Bull. 63, 2204– 2213. 4. Katz, D. L., Cornell, R., Kobayashi, R., Poettmann, F. H., Vary, J. A., Elenblass, J. R. & Weinaug, C. G. (1959) Handbook of Natural Gas Engineering (McGraw–Hill, New York). 5. Kvenvolden, K. A. & McMenamin, M. A. (1980) Geol. Surv. Circ. (U. S.) 825, 1–11. 6. Booth, J. S., Rowe, M. M. & Fischer, K. M. (1996) Open-File Rep.–U. S. Geol. Surv. Rep. 96-272, pp. 1–17.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

POTENTIAL EFFECTS OF GAS HYDRATE ON HUMAN WELFARE

3426

7. Kvenvolden, K. A. (1988) Chem. Geol. 71, 41–51. 8. Chersky, N. & Makogon, Yu. F. (1970) Oil Gas Int. 10, 82–84. 9. Potential Gas Committee (1981) Potential Supply of Natural Gas in the United States (as of December 31, 1980) (Potential Gas Agency, Golden, CO). 10. Dobrynin, V. M., Korotajev, Yu. P. & Plyuschev, D. V. (1981) in Long-Term Energy Resources, eds. Meyer, R. F. & Olson, J. C. (Pitman, Boston), pp. 727–729. 11. Kvenvolden, K. A. (1993) U. S. Geol. Surv. Prof. Pap. 1570, 555–561. 12. Kvenvolden, K. A. & Grantz, A. (1990) in The Arctic Ocean Region, eds. Grantz, A., Johnson, L. & Sweeney, J. F. (Geol. Soc. Am., Boulder, CO) Vol. 50, pp. 539–549. 13. MacDonald, G. T. (1990) Clim. Change 16, 247–281. 14. Kvenvolden, K. A. & Claypool, G. E. (1988) Open-File Rep.–U. S. Geol Surv. Rep. 88–216. 15. Gornitz, V. & Fung, I. (1994) Global Biogeochem. Cycles 8, 335–347. 16. Harvey, L. D. D. & Huang, Z. (1995) J. Geophys. Res. 100, 2905–2926. 17. Holbrook, W. S., Hoskins, H., Wood, W. T., Stephen, R. A., Lizarralde, D. & Leg 164 Science Party (1996) Science 273, 1840–1843. 18. Dickens, G. R., Paull, C. K., Wallace, P. & the ODP Leg 164 Scientific Party (1997) Nature (London) 385, 426–428. 19. Makogon, Yu. F. (1997) Hydrates of Hydrocarbons (PennWell, Tulsa, OK). 20. Ginsburg, G. D. & Soloviev, V. A. (1995) 27th Annual Offshore Technology Conference, (Offshore Technology Conference, Richardson, TX), OTC 7693, Vol. 1, pp. 513–518. 21. McIver, R. D. (1981) in Long-Term Energy Resources, eds. Meyer, R. F. & Olson, J. D (Pitman, Boston), pp. 713–726. 22. Trofimuk, A. A., Chersky, N. V. & Tsarev, V. P. (1977) in Future Supply of Nature-Made Petroleum and Gas, ed. Meyer, R. F. (Pergamon, New York), pp. 919–926. 23. Hammerschmidt, E. G. (1934) Ind. Eng. Chem. 26, 851–855. 24. Stoll, R.D., Ewing, J. & Bryan, G. M. (1971) J. Geophys. Res. 76, 2090–2094. 25. Claypool, G. E. & Kaplan, I. R. (1974) in Natural Gases in Marine Sediments, ed. Kaplan, I. R. (Plenum, New York), pp. 99–139. 26. Kvenvolden, K. A. (1993) Rev. Geophys. 31, 173–187. 27. MacDonald, G. T. (1990) Annu. Rev. Energy 15, 53–83. 28. Collett, T. S. & Kuuskraa, V. A. (1998) Oil Gas J. 96, (19), 90–95. 29. Levorsen, A. I. (1967) Geology of Petroleum (Freeman, San Francisco), 2nd Ed. 30. Paull, C. K., Matsumoto, R., Wallace, P. J., et al. (1996) Proceedings ODP, Initial Reports (Ocean Drilling Program, College Station, TX), Vol. 164. 31. Dillon, W. P., Holbrook, W. S., Drury, R., Gettrust, J., Hutchinson, D., Booth, J. & Taylor, M. (1997) 29th Offshore Technology Conference, (Offshore Technology Conference, Richardson, TX) OTC 8296, pp. 201–209. 32. Hunt, J. M. (1979) Pet. Geochem. Geol. (Freeman, San Francisco ). 33. Rogner, H.-H. (1997) Annu. Rev. Energy Environ. 22, 217–262. 34. Holder, G. D., Kamath, V. A. & Godbole, S. P. (1984) Annu. Rev. Energy 9, 427–445. 35. Collett, T. S. (1993) U. S. Geol. Surv. Prof. Pap. 1570, 299–311. 36. Collett, T. S. & Ginsburg, G. D. (1998) Int. J. Offshore Polar Eng. 8, 22–29. 37. Khalil, M. A. K. & Rasmussen, R. A. (1983) J. Geophys. Res. C Oceans Atmos. 88, 5232–5144. 38. Steele, L. P., Fraser, P. J., Rasmussen, R. A., Khalil, M. A. K., Conway, T. H., Crawford, A. J., Gammon, R. H., Masarie, K. A. & Thoning, K. W. (1987) J. Atmos. Chem. 5, 125–171. 39. Blake, D. R. & Rowland, F. S. (1988) Science 240, 634–637. 40. Steele, L. P., Dlugokencky, E. J., Sang, P. M., Tans, P. P., Martin, R. C. & Masarie, K. A. (1992) Nature (London) 358, 313–316. 41. Shine, K. P., Derwent, R. G., Wuebbles, D. J. & Morcritte, J.-J. (1990) in Climate Change, The IPCC Scientific Assessment, eds. Houghton, J. T., Jenkins, G. J. & Ephraums, J. J. (Cambridge Univ. Press, New York), pp. 41–68. 42. Cicerone, R. J. & Oremland, R. S. (1998) Global Biogeochem. Cycles 2, 299–327. 43. Hovland, M. & Judd, A. G. (1988) Seabed Pockmarks and Seepages (Graham & Trotman, London). 44. MacDonald, I. R., Guinasso, N. L., Sassen, R., Brooks, J. M., Lee, L. & Scott, K. T. (1994) Geology 22, 699–702. 45. Dillon, W. P., Danforth, W. W., Hutchinson, D. R., Drury, R. M., Taylor, H. H. & Booth, J. S. (1998) in Gas Hydrates: Relevance to World Margin Stability and Climate Change, eds. Henriet, J.-P. & Mieneat (J. Geol. Soc., London) Special Publications, Vol. 137, pp. 293–302. 46. Nisbet, E. G. (1990) Can. J. Earth Sci. 27, 148–157. 47. Paull, C. K., Ussler, W. & Dillon, W. P. (1991) Geophys. Res. Lett. 18, 432–434. 48. Kvenvolden, K. A. (1988) Global Biogeochem. Cycles 2, 221–229. 49. Kvenvolden, K. A., Lilley, M. D., Lorenson, T. D., Barnes, P. W. & McLaughlin, E. (1993) Geophys. Res. Lett. 20, 2459–2462. 50. Lorenson, T. D. & Kvenvolden, K. A. (1995) Open-File Rep.–U. S. Geol. Surv., Rep. 95-70, pp. 1–84. 51. Lorenson, T. D. & Kvenvolden, K. A. (1997) Open-File Rep.–U. S. Geol. Surv., Rep. 97-54, pp. 1–77. 52. Dickens, G. R., O'Neil, J. R., Rea, D. K. & Owen, R. M. (1995) Paleoceanography 10, 965–971. 53. Dickens, G. R., Castillo, Maria M. & Walker, J. C. G. (1997) Geology 25, 259–262. 54. Chappellaz, J., Barnola, J. M., Raynaud, D., Korotkevich, Y. S. & Lorius, C. (1990) Nature (London) 345, 127–131. 55. Blunier, T., Chappellaz, J., Schwander, J., Stauffer, B. & Raynaud, D. (1995) Nature (London) 374, 46–49. 56. Kvenvolden, K. A. (1994) Ann. N.Y. Acad. Sci. 715, 232–246. 57. McIver, R. D. (1982) Am. Assoc. Pet. Geol. Bull. 66, 789–792. 58. McIver, R. D. (1977) Abstracts with Programs, Geological Society of America 9, 1089–1090 (abstr.). 59. Summerhayes, C. P., Bornhold, B. D. & Embley, R. W. (1979) Mar. Geol. 31, 265–277. 60. Carpenter, G. (1981) Geo-Mar. Lett. 1, 29–32. 61. Jansen, E., Befring, S., Bugge, T., Eidvin, T., Holtedahl, H. & Sejrup, H. P. (1987) Mar. Geol. 78, 77–107. 62. Bugge, T., Befring, S., Belderson, R. H., Eidvin, T., Jansen, E., Kenyon, H., Hotedahl, H. & Sejrup, H. P. (1987) Geo-Mar. Lett. 7, 191–198. 63. Bornhold, B. D. & Prior, D. B. (1989) Geo-Mar. Lett. 9, 135–144. 64. Kayen, R. E. & Lee, H. G. (1991) Mar. Geotechnol. 10, 125–141. 65. Borowski, W. S. & Paull, C. K. (1997) 29th Offshore Technology Conference, (Offshore Technology Conference, Richardson, TX), Vol. 1, OTC 8297, pp. 211–216. 66. Chaouch, A., and Briaud, J. L. (1997) 29th Offshore Technology Conference, (Offshore Technology Conference, Richardson, TX), Vol. 1, OTC 8298, pp. 217–224. 67. Yakushev, V. S. & Collett, T. S. (1992) Proceedings of the Second International Offshore and Polar Engineering Conference, eds. Chang, V. S., et al. (Int. Soc. Offshore Polar Eng., Golden, CO), Vol. 1, pp. 669–673. 68. Bagirov, E. & Lerche, I. (1997) Oil Gas J. 95, 99–102.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

HEALTH IMPACTS OF DOMESTIC COAL USE IN CHINA

3427

Health impacts of domestic coal use in China

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3427–3431, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. ROBERT B. FINKELMAN*†, HARVEY E. BELKIN*, AND BAOSHAN ZHENG‡ PNAS is available online at www.pnas.org.

ABSTRACT Domestic coal combustion has had profound adverse effects on the health of millions of people worldwide. In China alone several hundred million people commonly burn raw coal in unvented stoves that permeate their homes with high levels of toxic metals and organic compounds. At least 3,000 people in Guizhou Province in southwest China are suffering from severe arsenic poisoning. The primary source of the arsenic appears to be consumption of chili peppers dried over fires fueled with high-arsenic coal. Coal samples in the region were found to contain up to 35,000 ppm arsenic. Chili peppers dried over high-arsenic coal fires adsorb 500 ppm arsenic on average. More than 10 million people in Guizhou Province and surrounding areas suffer from dental and skeletal fluorosis. The excess fluorine is caused by eating corn dried over burning briquettes made from high-fluorine coals and high-fluorine clay binders. Polycyclic aromatic hydrocarbons formed during coal combustion are believed to cause or contribute to the high incidence of esophageal and lung cancers in parts of China. Domestic coal combustion also has caused selenium poisoning and possibly mercury poisoning. Better knowledge of coal quality parameters may help to reduce some of these health problems. For example, information on concentrations and distributions of potentially toxic elements in coal may help delineate areas of a coal deposit to be avoided. Information on the modes of occurrence of these elements and the textural relations of the minerals and macerals in coal may help predict the behavior of the potentially toxic components during coal combustion. The U.S. Environmental Protection Agency (EPA) recently issued a report to Congress on the health impacts of 189 potentially hazardous air pollutants (HAPs) emitted from coal-burning electric utility generators (1). In this report the EPA concludes that, with the exception of mercury, there is no compelling evidence to indicate that trace element emissions cause human health problems. The absence of detectable health problems is, in part, caused by the fact that the coals burned in the U.S. generally contain low to modest concentrations of HAP elements and that many coal-burning utilities use sophisticated pollution control systems that efficiently reduce the emissions of HAPs (2). Such is not the case in many developing countries, especially in homes where coal is used for heating and cooking. Domestic use of coal can present serious human health problems because the coals generally are mined locally with little regard to their composition, and the coals are commonly burned in poorly vented or unvented stoves, directly exposing residents to the emissions. This paper briefly describes health problems believed to be caused by, or exacerbated by, trace elements or organic compounds emitted during domestic combustion of coal. Although the examples used to illustrate these problems are taken from China, people in many other developing and undeveloped countries use coal in a similar way and may suffer from similar health problems. China is the world's largest coal producer and coal consumer. Coal production has increased steadily during the past 25 years to nearly 1.4 billion metric tons in 1996 (5). In contrast to most developed countries, such as the U.S., where domestic coal use constitutes a small fraction of 1% of coal consumption, a substantial portion of China's coal is used for domestic energy needs. Smith and Liu (3) estimate that worldwide 330 million people rely on coal for domestic energy needs and as many as 2.5–3 billion people are using even poorer-quality biomass fuels. However, Florig (4) estimates that more than 75% of China's primary energy needs are supplied by domestic coal. Coal stoves and small coal boilers provide more than 50% of the energy for urban households and 22% of rural households rely on coal (4). About 70 percent of the population in China resides in rural areas, thus Florig's data would indicate that about 400 million people in China rely on coal for their domestic energy needs. Health Problems Caused by Trace Element Emissions Wood had long been the primary source of energy in southwest China, but by the early part of this century the forests were largely denuded and the residents were forced to seek other sources of fuel. In southwest Guizhou Province (Fig. 1), surface exposures of coal are plentiful, and coal quickly became the primary fuel for domestic use. Unfortunately, some of these coals have undergone mineralization, causing their enrichment in potentially toxic trace elements such as arsenic, fluorine, mercury, antimony, and thallium. Burning the mineralized coals in unvented stoves volatilizes the toxic elements and exposes the local population to the toxic elements in the emissions. The situation is exacerbated by the practice of drying crops directly over the coal fires. In the autumn it is commonly cool and damp in the higher elevations of Guizhou Province. It is common practice for the residents of this region to dry their corn and chili peppers directly over the burning coals. Arsenic. Chronic arsenic poisoning, which affects at least 3,000 people in Guizhou Province, has been described by Zheng and others (6). Those affected exhibit typical symptoms of arsenic poisoning, including hyperpigmentation (flushed appearance, freckles), hyperkeratosis (scaly lesions on the skin, generally concentrated on the hands and feet), Bowen's disease (dark, horny, precancerous lesions of the skin: Fig. 2), and squamous cell carcinoma. Zheng and others (6) have shown that chili peppers dried over open coal-burning stoves may be a principal vector for the arsenic poisoning. Fresh chili peppers have less than 1 ppm arsenic. In contrast, chili peppers dried over high-arsenic coal fires can have more than 500 ppm arsenic. Significant amounts

*U.S. Geological Survey, Mail Stop 956, Reston, VA 20192; and ‡Institute of Geochemistry, Guiyang, Guizhou Province, People's Republic of China 55002 Abbreviations: SEM, scanning electron microscope; EDS, energy-dispersive x-ray analyzer. †To whom reprint requests should be addressed. e-mail: [email protected].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

HEALTH IMPACTS OF DOMESTIC COAL USE IN CHINA

3428

FIG. 1. Map of China indicating provinces in which there are reported health problems caused by domestic coal combustion. of arsenic also may come from other tainted foods, ingestion of dust (samples of kitchen dust contained as much as 3,000 ppm arsenic), and from inhalation of indoor air polluted by arsenic derived from coal combustion. The arsenic content of drinking water samples was below the Environmental Protection Agency's drinking water standard (7) of 50 ppb and does not appear to be an important factor. Detailed chemical and mineralogical characterization of the arsenic-bearing coal samples from this region recently was conducted by Belkin and coworkers (8, 9 and 10). They analyzed about 25 coal samples that they had collected from several locations within Guizhou Province. Instrumental neutron activation analyses of the coal indicate arsenic concentrations as high as 35,000 ppm. The magnitude of this concentration can be seen by comparison with U.S. coals. The mean concentration for arsenic in nearly 10,000 U.S coal samples is approximately 22 ppm, with a maximum value of about 2,000 ppm (11). Belkin and coworkers (8, 9 and 10) examined polished blocks of the coal by using a scanning electron microscope (SEM) equipped with energy-dispersive x-ray analyzer (EDS) and an electron microprobe (EMP). They observed a wide variety of As-bearing mineral phases in the coal samples. Pyrite is the most common sulfide, occurring as framboids, euhedral crystals, and irregular shapes. The range of As in pyrite determined by EMP analyses is from the detection limit (≈100 ppm) in unaltered framboids to about 4.5 weight percent in grains adjacent to arsenopyrite crystals. Arsenopyrite occurs in a variety of habits, including large 150- to 250-µm crystals, narrow, 1- to 5-µm veins, and small crystals (Fig. 3A). Selenium contents in arsenopyrite are fairly uniform (0.15 to 0.2 weight percent). A third As-bearing sulfide, composed of As, Pb, and S, is present rarely. Another group of As-bearing minerals contains arsenic in the 5+ valence state as arsenate commonly substituting for the phosphate group. An unidentified As-bearing iron phosphate, usually associated with banded iron oxide (Fig. 3B), as veins or masses has a P/As ratio on the order of 4. Jarosite [K2Fe63+(SO4)4(OH)12] was present as an alteration product of sulfides or as mixtures with iron oxide and commonly contained a few weight percent As. An additional As-rich phase was observed only as scattered micron-sized grains that contained only Fe and As (±O), as identified by SEM-EDS. The atomic ratio of Fe/As in this mineral is about 1 and the mineral may be scorodite, FeAsO4·2H2O. Some coals display evidence of movement of epigenetic fluids and show primary phases and their alteration products. One sample examined by Belkin and coworkers (8, 9 and 10) showed a complete range of diagenetic, low-As framboids that progresses through various stages to framboid pseudomorphs composed totally of As-bearing iron oxide (SEM-EDS and reflected light indicates hematite). Veins of jarosite, arsenopyrite, and iron oxide are common in some samples. Three samples from the same location had As concentrations in excess of 3 weight percent and were mineralogically unusual. Although they contain small grains and veins of arsenopyrite and As-bearing pyrite, the concentration of these phases is completely inadequate to account for the As abundance on a whole coal basis. However, in SEM back-scattered electron images, a distinct banding characterized by differing

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

HEALTH IMPACTS OF DOMESTIC COAL USE IN CHINA

3429

FIG. 2. Extensive scaly lesions (hyperkeratosis) are evident on the chest of a resident of this region. The dark spot over the left breast was diagnosed as Bowen's disease. image brightness is easily observed (Fig. 3C). Some of this banding forms box-like arrangements, but in all cases the bands appear to have sharp edges. The bands range from a few to tens and a few hundreds of µm in thickness. SEM-EDS results show that these bright bands are highly enriched in As (Fig. 3D). They also contain S, Fe, and traces of Al and Si (µm-sized clay particles). In fact, there is a relationship between the EDS count intensity for As and apparent brightness of the SEM image. Semiquantitative analysis by SEM-EDS demonstrate that the bright bands contain As at levels ≈3 weight percent. Fe concentrations in the bands are low but always present at levels from 0.2 to 0.4 weight percent, S is the only other major element found. By using an SEM, no discrete As-bearing phases could be resolved in these bands at 50,000 times magnification. Thin fragments of one sample were examined by an advanced field-emission transmission electron microscope. No discrete As-bearing phase could be observed by using this instrument at magnifications of 1 million times. Thus, finely dispersed arsenopyrite, As-bearing pyrite, or any other As-phase can be ruled out as the source of the As. To define the nature of bonding in the arsenic-bearing phases, a reconnaissance study of two high-arsenic samples was conducted by using high-energy x-rays from a synchrotron source (12). Collection of diffraction spectrum intensity across the XANES (x-ray absorption near-edge structure) and EXAFS (extended x-ray absorption fine structure) regions of an absorption spectrum can provide three-dimensional information on the electronic state and chemical coordination for each crystallographic site of the chosen element. Results from this work demonstrate that ≈100% of the As in one sample is AsO43− and that about 75% of the As in the other sample is AsO43− with the balance (25%) as sulfide-bound As. Thus, for the two coals examined, the preponderance of the As is in the 5+ valence state. An interesting and potentially important relationship exists between arsenic and gold. Southwest Guizhou Province contains numerous gold mines and deposits that have been identified as Carlin-type gold deposits (13). Carlin-type deposits are characterized by very fine-grained gold (generally less than 1 µm) and are associated with arsenic, antimony, mercury, and thallium. In Carlin-type deposits of the Great Basin region of Nevada, the sedimentary host rock is usually rich in organic matter. The Guizhou Province deposits also contain thallium and mercury (14, 15 and 16). Ashley and others (13) concluded that the Chinese and Great Basin deposits formed from low-salinity fluids at relatively low temperatures. Li and Peters (17), using fluid inclusion data, indicate that the Guizhou fluids ranged in temperature from 150°C to 240°C, comparable to the U.S. occurrences. It is geologically reasonable to assume that the introduction of arsenic into the coal strata is related to gold mineralization, although the exact mechanism is uncertain. Mineralogical characterization of the coals from Guizhou Province may help elucidate the geologic process that created the high-arsenic coals and the relationship of the high-arsenic coals to the gold. Knowledge of these processes and relationships may help determine the regional distribution of these environmentally dangerous coals. Information on the arsenic mineralogy also may help us to anticipate the behavior of arsenic during coal combustion. Preliminary characterization of residual ash in coal-burning stoves indicates high retention of arsenic. Mineralogical characterization in conjunction with combustion tests may determine whether one or more of the arsenic-bearing phases is primarily responsible for adsorption of arsenic on the chili peppers. Fluorine. The health problems caused by fluorine volatilized during domestic coal use are far more extensive than those caused by arsenic (Fig. 1). More than 10 million people in Guizhou Province and surrounding areas suffer from various forms of fluorosis (18, 19), and it also has been reported from 13 other provinces, autonomous regions, and municipalities in China (20). Typical symptoms of fluorosis include mottling of tooth enamel (dental flurosis) and various forms of skeletal fluorosis, including osteosclerosis, limited movement of the joints, and outward manifestations such as knock knees, bow legs, and spinal curvature. Fluorosis combined with nutritional deficiencies in children can result in severe bone deformation (Fig. 4). The etiology of fluorosis is similar to that of arsenism in that the disease is derived from foods dried over coal-burning stoves. Zheng and Huang (18) have demonstrated that adsorption of fluorine by corn dried over unvented ovens burning high (>200 ppm) fluorine coal is the probable cause of the extensive dental and skeletal fluorosis in southwest China. The mode of occurrence of fluorine in the coal is unknown. The problem is compounded by the use of clay as a binder for making briquettes. The clay used is a high-fluorine (mean value of 903 ppm) residue formed by intense leaching of a limestone substrate. Ando and others (20) determined fluorine contents of coals from two mines (559 and 802 ppm) and in the associated soils (592 and 669 ppm). They estimated that 97% of the fluoride exposure came from food consumption and 2% from direct inhalation. Zhang and Cao (19) report mean fluorine levels in coals from 11 regions in China to range from 203 to 1,513 ppm with a maximum value of 3,762 ppm. Mercury. There is also considerable concern about the health effects of mercury and the proportion of anthropogenic mercury in the environment (21). So far, there is no direct

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

HEALTH IMPACTS OF DOMESTIC COAL USE IN CHINA

3430

FIG. 3. (A) SEM back-scattered electron image of polished coal showing a pyrite grain (P) and adjacent arsenopyrite crystals (A). (B) SEM back-scattered electron image of polished coal showing the deposition of a banded complex of iron oxide and As-bearing iron phosphates. (C) SEM back-scattered electron image of polished block of arsenic-rich coal. Dark areas are coal, bright areas are mainly pyrite, milky area is coal containing organically bound arsenate. Fluids moving through the fracture in the coal appears to have removed arsenic from the organic matrix. (D) X-ray map depicting the distribution of arsenic in the coal. Red areas are high concentrations, and blue areas are low concentrations. Compare distribution of arsenic to the outline of the milky area in C. A–C originally appeared in ref. 9 and are republished with the permission of the Pittsburgh Coal Conference. evidence of health problems caused by mercury released from coal but there are circumstances where poisoning from mercury released from coal combustion may be occurring. Zhou and Liu (22) reported on chronic thallium poisoning in Guizhou Province, China, where the source of the thallium poisoning appears to be from vegetables grown on a mercury/thallium-rich mining slag. Most symptoms, such as hair loss, are typical of thallium poisoning. However, loss of vision in several patients from this region was considered to be unique (R. Dart, personal communication). Mineralogical analysis of the coal being used in the homes of people having visual impairment revealed abundant mercury minerals. Chemical analysis of a coal sample being used in Guizhou Province, China, indicates a mercury concentration of 55 ppm, which is about 200 times the average mercury concentration in U.S. coals. Selenium. Zheng and others (23) report nearly 500 cases of human selenosis in southwest China that are attributed to the use of seleniumrich carbonaceous shales known locally as

FIG. 4. Bone deformation caused by nutritional deficiency combined with exposure to high levels of fluorine from domestic coal combustion.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

HEALTH IMPACTS OF DOMESTIC COAL USE IN CHINA

3431

“stone coal.” The stone coals have as much as 8,390 ppm selenium. This selenosis is attributed to the practice of using combustion ash as a soil amendment. This process introduced large amounts of selenium into the soil and resulted in selenium uptake by crops. Symptoms of selenium poisoning include hair and nail loss. Organic Compounds. Esophageal cancer is a common fatal cancer and the fourth-leading cause of cancer death in China. Parts of Henan Province in north-central China (Fig. 1) have some of the highest rates of esophageal cancer in the world, with annual age-adjusted mortality rates of up to 169 per 100,000 and cumulative death rates of over 20% by age 75 for both sexes (24). Many studies have been carried out on the high incidence of esophageal and lung cancers in China, but the dominant causative agents of the cancer remain unclear. Polycyclic aromatic hydrocarbons (PAHs) released during unrented coal combustion in homes in China, have been cited as the primary cause for the highly elevated incidence of lung cancer (25). The PAH levels in homes burning “smoky” coal are so high that the resulting lung cancer mortality rate is five times the national average of China (26). Conclusions A better knowledge of coal quality parameters may help to minimize some of the health problems caused by domestic coal use. Information on the concentrations and distributions of potentially toxic elements in coal may assist people dependent on local coal sources to avoid those areas of a coal deposit having undesirably high concentrations of toxic compounds. Information on the modes of occurrence of potentially toxic elements and the textural relations of the minerals and macerals in which they occur may help us to anticipate the behavior of the potentially toxic components during coal cleaning, combustion, weathering, and leaching. Coal characterization offers geoscientists opportunities to directly contribute improved public health.

1. U.S. Environmental Protection Agency (1998) Study of Hazardous Air Pollutant Emissions from Electric Utility Steam-Generating Units (Government Printing Office, Washington, DC) EPA-453/ R-98-004A. 2. U.S. Environmental Protection Agency (1996) Study of Hazardous Air Pollutant Emissions from Electric Utility Steam-Generating Units: Interim Final Report (Government Printing Office, Washington, DC), EPA-453/R-96–013a. 3. Smith, K. R. & Liu, Y. (1994) in Epidemiology of Lung Cancer, ed. Samet, J. M. (Dekker, New York), pp. 151–184. 4. Florig, H. K. (1997) Environ. Sci. Technol. News 31, 274A–279A. 5. Coleman, L. L. (1998) International Coal: 1997 Edition (The National Mining Association, Washington, DC). 6. Zheng, B., Yu, X., Zhand, J. & Zhou, D. (1996) 30th Intl. Geologic Congress Abstr. 3, 410. 7. U.S. Environmental Protection Agency (1973) Water Quality Criteria 1972 (Government Printing Office, Washington, DC), EPA-R3–73033. 8. Belkin, H. E., Zheng, B. & Finkelman, R. B. (1997) Fourth International Symposium on Environmental Geochemistry (U.S. Geological Survey, Reston, VA), U.S. Geological Survey Open-File Report 97–496, p. 10. 9. Belkin, H. E., Zheng, B., Zhou, D. & Finkelman, R. B. (1997) Fourteenth Annual International Pittsburgh Coal Conference, CD-ROM (University of Pittsburgh). 10. Belkin, H. E., Warwick, P., Zheng, B., Zhou, D. & Finkelman, R. B. (1998) Fifteenth Annual International Pittsburgh Coal Conference, CD-ROM (University of Pittsburgh). 11. Bragg, L. J., Oman, J. K., Tewalt, S. J., Oman, C. L., Rega, N. H., Washington, P. M. & Finkelman, R. B. (1997) U. S. Geological Survey Coal Quality (COALQUAL), Version 2.0 (U.S. Geological Survey, Reston, VA), U. S. Geological Survey Open-File Report 97 134 (CD-ROM). 12. Huffman, G. P., Huggins, F. E., Shah, N. & Zhao, J. (1994) Fuel Processing Technol. 39, 47 62. 13. Ashley, R. P., Cunningham, C. G., Bostick, N. H., Dean, W. E. & Chou, I.-M. (1991) Ore Geol. Rev. 6, 131 151. 14. Qian, H., Chen, W. & Hu, Y. (1995) Geol J. Universities 1, 45 52 (in Chinese; English abstract). 15. Chen, D., Wang, H. & Ren, D. (1996) Acta Mineralogia Sinica 16, 307 314 (in Chinese; English abstract). 16. He, M. & Wu, X. (1997) Chinese Z Geochem. 16, 75 79. 17. Li, Z. & Peters, S. G. (1996) in Geological Society of America Program for 1996 Annual Meeting (Geological Society of America, Boulder, CO), p. A153 (abstr.). 18. Zheng, B. & Huang, R. (1989) in Developments in Geoscience, Contributions to 28th International Geologic Congress: 1989, eds. Washington, DC (Science Press, Beijing, China), pp. 171 176. 19. Zhang, Y. & Cao, S. R. (1996) Fluoride 29, 207 211. 20. Ando, M., Tadano, M., Asanuma, S., Matsushima, S., Wanatabe, T., Kondo, T., Sakuai, S., Ji, R., Liang, C. & Cao, S. (1998) Environ. Health Perspect. 106, 239 244. 21. U.S. Environmental Protection Agency (1998) Mercury Study Report to Congress: White Paper (Government Printing Office, Washington, DC) EPA-453/R-98-004B. 22. Zhou, D. & Liu, D. (1985) J. Environ. Health 48, 14 18. 23. Zheng, B., Hong, Y., Zhao, W., Zhou, H., Xia, W., Su, H., Mao, D., Yan, L. & Thornton, I. (1992) Chinese Sci. Bull. 37, 1725 1729. 24. Crowley, S. S., Orem, W. H., Roth, M. J., Finkelman, R. B., Scroggs, E. A. & Willett, J. (1998) Fifteenth Annual Meeting of the Society for Organic Petrology, Abstracts and Program (The Society for Organic Petrology), Vol. 15, pp. 99 97. 25. Mumford, J. L., He, X. Z., Chapman, R. S., Cao, S. R., Harris, D. B., Li, X. M., Xian, W. Z., Jiang, C. W., Xu, J. C., Chuang, J. C., et al. (1987) China Sci. 235, 217 220. 26. Mumford, J. L., Li, X., Hu, F., Lu, X. B. & Chaung, J. C. (1995) Carcinogenesis 16, 3031 3036.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3432

Nuclear waste forms for actinides

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3432–3439, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. RODNEY C. EWING* PNAS is available online at www.pnas.org.

Department of Nuclear Engineering and Radiological Sciences and Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109-2104 ABSTRACT The disposition of actinides, most recently 239Pu from dismantled nuclear weapons, requires effective containment of waste generated by the nuclear fuel cycle. Because actinides (e.g., 239Pu and 237Np) are long-lived, they have a major impact on risk assessments of geologic repositories. Thus, demonstrable, long-term chemical and mechanical durability are essential properties of waste forms for the immobilization of actinides. Mineralogic and geologic studies provide excellent candidate phases for immobilization and a unique database that cannot be duplicated by a purely materials science approach. The “mineralogic approach” is illustrated by a discussion of zircon as a phase for the immobilization of excess weapons plutonium. The disposition of “waste” generated by the nuclear fuel cycle is one of the most pressing, and potentially costly, environmental problems for the 21st century, a heritage from the atomic age of the 20th century. Proposed strategies are complicated, not only by the large volumes and activities of waste, but by the political and public policy issues associated with the long times considered for containment and disposal (104-106 years). Furthermore, the waste includes fissile material, e.g., 239Pu, of high energy content. Three primary sources of actinide-bearing waste in the United States are as follows. High-level waste (HLW) resulting from reprocessing to reclaim fissile materials for weapons production. Approximately 380,000 m3 (100 million gallons) of HLW have a total radioactivity of 960 million curies (1 Ci = 37 GBq) (1). The greatest volumes (340,000 m3) are stored in tanks at Hanford, WA, and Savannah River, SC. Over 99 percent of the present activity is from nonactinide radionuclides with half-lives <50 years (reprocessing has removed much of the actinide content); however, after 500 years, the total activity will be substantially reduced, and the primary radionuclides will be 238Pu, 131Sm, and 241Am. After 50,000 years, most of the activity will be associated with longer-lived radionuclides, such as 239Pu and 240Pu. Also related to reprocessing are much lower activity waste contaminated with transuranic elements, TRU waste. These are defined as containing 100 nanocuries of α-emitting transuranic isotopes, with half-lives >20 years, per gram of waste. Over 60,000 m3 are stored retrievably at Department of Energy sites, destined for disposal at the Waste Isolation Pilot Plant in New Mexico (2). The estimated cost of remediation and restoration actives in the Department of Energy complex during the next decades is in the order of 200 billion dollars (3). Used or spent nuclear fuel resulting from commercial power generation. Just over 20 percent of the electricity generated in the United States is produced by nuclear power reactors. In 1995, 32,200 metric tons of spent fuel with a total activity of 30,200 MCi were stored by the electric utilities at 70 sites (either in pools or in dry storage systems) (2, 4). By 2020, the projected inventory will be 77,100 metric tons of heavy metal (MTHM) with a total activity of 34,600 MCi. Although the volume of the spent fuel is only a few percent of the volume of HLW, >95% of the total activity (defense-related plus commercially generated waste) is associated with the commercially generated spent nuclear fuel (3). At present, none of the spent fuel will be reprocessed, and all is destined for direct disposal in a geologic repository. The dismantlement of nuclear weapons. Under the first and second Strategic Arms Reduction treaties, as well as unilateral pledges made by both the United States and Russia, several thousand nuclear weapons will be dismantled. Initially, this will result in an estimated 100 metric tons of weapons plutonium that will require long-term disposition. The disposition strategy should not only protect the public and the environment but must also ensure that the plutonium is not readily recoverable for use in weapons (5). Present U.S. strategy calls for “burning” the Pu as a mixed-oxide fuel in existing or modified reactors followed by direct disposal with commercially generated spent fuel in a geologic repository (6, 7). A smaller portion of the Pu (tens of metric tons) is destined for immobilization into a durable solid followed by geologic disposal. The present program has an anticipated cost of two billion dollars. World-wide, since the first creation of milligram quantities of plutonium by Glenn Seaborg in 1941, the global inventory of plutonium has reached 1,350 tons and continues to increase by ≈70 tons/year (Table 1). This commercially generated plutonium is m two forms: (i) incorporated in spent fuel destined for direct geologic disposal (>600 metric tons of plutonium is in the spent fuel in the U.S.); and (ii) plutonium separated by reprocessing of commercial fuel, which is estimated to reach 300 tons by the year 2000. This is greater than the amount of plutonium presently in nuclear weapons (9). Considering that the bare critical mass for weapons grade plutonium is 15 kg of metal (this number is substantially reduced in the presence of a neutron reflector), safe-guarding this plutonium is essential. In fact, the need for safeguards to protect against the diversion of separated plutonium applies equally to all grades of plutonium (10). The peaceful use of nuclear energy will inevitably require a strategy for the disposition and disposal of actinides. Why Are Actinides Important? Although there are a number of fission product radionudides of high activity (137Cs and 90Sr) and long half-life (99Tc, 200,000 years; 129I, 1.6 × 107 years) in spent nuclear fuel, actinides and their daughter products account for most of the radiotoxicity of nuclear waste after the first 500 years of disposal (Fig. 1). After several hundred years, radiotoxicity is dominated by 239Pu (half-life = 24, 100 years) and 237Np (half-life = 2,000,000 years). Thus, a major part of the long-term risk is directly related to the fate of these two actinides in the geosphere (natural, crustal concentrations of Pu are on the order of 10−11 ppm for 239Pu). Plutonium has several important and unique properties: (i) 239Pu is fissile; (ii) 239Pu with a half-life of 24,100 years decays to

Abbreviations: HLW, high-level waste; Ma, million years ago; MeV, million electronvolts; dpa, displacements per atom; REE, rare-earth element. *To whom reprint requests should be addressed. e-mail: rodewing@ umich.edu.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3433

Table 1. Estimated global plutonium inventory (metric tons) at the end of 1996 (8) In spent fuel USA Japan France Germany Total In operating reactors Separated by civilian reprocessing Military inventories Former USSR USA France China Israel, India, Pakistan Total Estimated world inventory, +70 metric tons/year production

650 90 70 50 860 80 150 140 100 6 5 1 252 1,350

235U, another fissile radionuclide with a half-life of 700 million years (the bare critical mass of weapons grade uranium is ≈50 kg); and (iii) Pu has four oxidation states (3+, 4+, 5+, and 6+) in natural water–rock systems. Although crystalline PuO2 has a low solubility, Pu may exist as PuO2+ or PuO22+ aqueous species, with the former predominating in oxidized natural waters. Additionally, in the 3+, 5+, and 6+ oxidation states, Pu forms strong carbonate complexes (pH >5) (12). Actual plutonium concentrations in solution are further complicated by the possibility of disproportionation among oxidation states, α-radiolysis of water to produce oxidants, such as H2O2, α-decay-induced amorphization of the solid that increases leach rates, and the formation of intrinsic actinide or actinide-bearing colloids that can increase actinide concentrations in ground waters by several orders of magnitude. Thus, the geochemistry of Pu has the full array of dissolution, transport, and precipitation mechanisms that are typical in geologic systems, in addition to radiation effects; as with other multivalent elements (e.g., insoluble U4+ vs. mobile U6+), this can lead to either dispersal or concentration in the geosphere. Thus, it is essential to evaluate the longterm behavior of Pu either as it exists in spent nuclear fuel or is immobilized in solid waste forms. The purpose of this paper is to illustrate the unique contributions that mineralogy and geochemistry can make in the design and selection of durable waste forms for the long term disposal of plutonium.

FIG. 1. Relative radiotoxicity on inhalation of spent nuclear fuel with a burnup of 38 megawatt days/kg U. The radiotoxicity values are relative to the radiotoxicity (horizontal line) of the quantity of uranium ore that was originally mined to produce the fuel (eight tons of natural uranium yields one ton of enriched uranium, 3.5% 235U) (11). Durable Waste Forms for Plutonium Although the development of waste forms for plutonium poses special problems and requirements for long-term durability, weapons Pu presents special opportunities: (i) As compared with high-level waste, the volumes are relatively small. For example, if Pu is immobilized in a typical waste form with a waste loading of 10 wt %, the 100 metric tons of weapons Pu can be immobilized in a volume of several hundred cubic meters. (ii) Weapons plutonium is remarkably pure, consisting of a Pu-Ga alloy (0.5–2% Ga) coated with a corrosion-resistant layer, generally nickel. This high purity provides a materials engineer with a wide range of potential processing techniques and the possibility of the production of phase pure waste forms at prescribed waste loading levels. The absence of highly active fission products, such as 137Cs and 90Sr, that are the primary source of ionizing radiation makes handling the material tractable, utilizing technologies comparable to those used to fabricate mixed-oxide fuels, (iii) Although the half-life of 239Pu (24,100 years) is much longer than that of the much higher activity fission products, a substantial amount of decay occurs over relatively short geologic time scales (e.g., containment of Pu for 10 half-lives requires on the order of 200,000 years). Thus, immobilization over the time required for substantial radioactive decay is short relative to the durability of some geologic materials, measured in many millions of years. The fact that 239Pu decays to 235U poses additional challenges for durability of the waste form, but the greater critical mass of 235U combined with the possibility of dilution of 235U by 238U can provide additional barriers to the potential for criticality. The essential question to mineralogists is, Are there naturally occurring, actinide-bearing phases of demonstrable chemical and physical durability that can be used for the long-term immobilization and disposal of weapons plutonium? This question is posed against a backdrop of 50 years of research and development of radioactive waste forms, mainly borosilicate glass and spent fuel. The first proposal for a waste form was made 45 years ago by Hatch (13), who suggested clays for the fixation of radionuclides into crystalline phases. Roy and McCarthy (14) were guided in their selection of appropriate phases by reference to minerals that are resistant to both geochemical alteration and radiation effects, and systematic studies were already in progress by the early 1980s (15). The most enduring proposal for a waste form was made by Ringwood et al. (16, 17) with the development of the Synroc concept, a Ti-based polyphase assemblage. The Synroc phase assemblage presently provides two of the candidate phases, pyrochlore and zirconolite, for Pu-immobilization (Table 2). These alternative concepts for waste forms led to a peak of research and development activity, mainly for the immobilization of the HLW in the tanks at the Savannah River site, during the late 1970s and early 1980s. This was a highly contentious period (18, 19 and 20) when a wide variety of “alternative” waste forms competed with borosilicate glass as the “reference” waste form (21). Although vitrification was finally selected for the Savannah River HLW (the Defense Waste Processing Facility began vitrifying waste in 1996), work on alternative waste forms has continued to the present day and provides the present basis for the discussion of waste forms for weapons plutonium (22). Rather than selecting a specific waste form in preference to others for Pu-immobilization, I want to illustrate what “mineralogical thinking” can bring to the table of such a discussion. The critical aspects of the question are the determination of “demonstrable, long-term physical and chemical durability.” Durability refers to a wide variety of properties: mechanical strength, thermodynamic stability, slow kinetics for corrosion processes, or retention of trace elements because of low diffusivity (in this case, actinides and neutron absorbers, such as Gd and Hf). The qualitative, geologic answer to such a question is obvious: e.g., the heavy minerals that survive weathering, erosion, transport, and deposition (sometimes many cycles) and persist as placer deposits in stream beds. The classic paper on heavy detrital

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3434

Table 2. Actinide-bearing phases that are candidates presently under consideration for plutonium immobilization Mineral Ideal formula GeoRef citations, 1998 Pyrochlore† (Ca,REE)Ti2O7 494 62 Zirconolite CaZrTi2O7 Apatite Ca4−xREE6+x(SiO4)6−y(PO4)y(O,F)2 5,098 Zircon ZrSiO4 8,055 1,711 Monazite CePO4 ZrO2 244 Baddelyite‡

Durable heavy mineral* Yes Yes Yes

*Heavy minerals identified and discussed in monograph on heavy mineral occurrences (23). Heavy minerals that were noted, but which are not included in this table, are gadolinite, allanite, thorite, titanite, and xenotime. Thorite (ThSiO4) and xenotime (YPO4) are isostructural with zircon; gadolinite, allanite, and titanite are either relatively rare or less durable than the minerals listed in the table. †Presently receiving the greatest attention within the U.S. program for the disposition of fissile materials. ‡Cubic and tetragonal polymorphs of ZrO have been considered as waste form phases; however, naturally occurring ZrO is most commonly 2 2 monoclinic baddelyite. Although not cited in ref. 23, it is included in this list because of its known durability and because it is considered both as a waste form and as inert matrix fuel for burning plutonium (9).

minerals by Hutton (23) provides a detailed list of heavy minerals (Table 2). Note the absence of potentially important actinide-bearing phases: pyrochlore, zirconolite, and baddeleyite, because of either their relative rarity or lower durability. The second issue is how much geologic and mineralogic data exist for these phases. Table 2 gives the number of citations for these phases as taken from GeoRef (American Geological Institute). The result is not surprising; those phases that are important to geochronology (zircon, apatite, and monazite) account for a major portion (95%) of the published literature. These studies are of two types: (i) laboratory studies to determine the ability of the minerals to retain isotopic signatures as a result of their physical and chemical durability in a variety of geologic environments; and (ii) agedating studies that essentially confirm the results of the laboratory studies in actual, long-term geologic environments. Both types of studies provide the essential data required for waste form design and selection. Present research on waste forms for Pu-immobilization includes a relatively short list of phases (Table 2). In the U.S., most of the effort within the Materials Disposition program of the Department of Energy focuses on immobilization in a ceramic (24), particularly cubic pyrochlore and its monoclinic derivative, zirconolite, because there are considerable data on these materials as waste form phases (22). However, the Department of Energy evaluation and selection process (24) is very different from the mineralogic approach presented in this paper. To illustrate the mineralogic approach, I review relevant work on zircon, drawn mainly from the mineralogic and geochemical literature, to demonstrate the utility of the mineralogic approach, as well as the extensive amounts of data that are already available in the literature. Although the structures of zircon (25) and actinide orthosilicates (26, 27) have been known for 30 years, I focus on studies applicable to the analysis of zircon as an actinide host phase. Zircon as an Actinide Waste Form Zircon (I41/amd; Z = 4) occurs in nature with uranium and thorium concentrations typically up to 5,000 ppm, but reaching 10 wt %. Zircon is an extremely durable mineral (28), often found as a heavy mineral in stream sediments, that, after transport over great distances, shows limited chemical alteration or physical abrasion (23, 29). The widespread distribution of zircon in the continental crust, its tendency to concentrate trace elements (lanthanides and actinides), its use in age dating, and its resistance to chemical and physical degradation (30, 31, 32 and 33) have made zircon probably the most useful accessory mineral in geologic studies. Zircon has been identified as an actinide-bearing phase in polyphase ceramic waste forms (34). Zircon also occurs in the Chernobyl lavas as an important actinide-bearing phase (6, 7, 8, 9, 10, 11 and 12 atomic percent uranium) (35). The propensity to incorporate actinides and its durability have lead to the suggestion that zircon be used to immobilize actinides (36, 37 and 38). Based on the ability of natural zircon to retain Pb, Gentry et al. (39) suggested that materials, like zircon, could effectively retain radioactive waste. This early suggestion was prescient but did not evaluate the extent of solid-solution of actinides in zircon and did not consider the much greater radiation damage that would occur in such a radioactive waste form (40). Structure. The zircon structure consists of triangular dodecahedral ZrO8 groups that form edge-sharing chains parallel to the a axis and SiO4 tetrahedral monomers that form edge-sharing chains with alternating ZrO8 groups parallel to the c axis (25, 26). U and Th replace the Zr in low concentrations; however, compositions of ASiO4, in which A4+ = Zr, Hf, Th, Pa, U, Np, Pu, and Am, have been synthesized (27). The regular increase in the unit cell volume with the increasing ionic radius of the A-site cation confirms the homologous topologies of these structures (Fig. 2). Four of these compositions, hafnon (HfSiO4), zircon, coffinite (USiO4), and thorite (ThSiO4), occur naturally. Structure refinements (27) and structural analyses (41, 42) suggest complete miscibility between ZrSiO4 and HfSiO4, but there are miscibility gaps on the ZrSiO4–USiO4–ThSiO4 joins (43). Zircon with 9.2 atom percent plutonium (8.1% Pu238; 1.1% Pu239) substituting for Zr has been synthesized (44). This is equal to a waste loading of 10 wt % Pu, but the maximum extent of the solubility of Pu in zircon has not been determined. That a pure, endmember composition, PuSiO4, has been synthesized (27) suggests extensive substitution of Pu for Zr is possible (42). The zircon structure is part of a larger class of ABO4 structure types (silicates and phosphates) and is closely related to the

FIG. 2. Variation in unit cell volume of actinide silicates, ASiO4, with the zircon structure as a function of ionic radius (VIIIA4+). Data are from Keller (27) after Speer (41).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3435

structure of monazite, CePO4, (another durable phase commonly used in geologic age-dating). Numerous A-site compositions (La, Pr, Nd, Sm, Eu, Gd, Tb, Tm, Yb, Lu, Sc, and Y) with the silicate and phosphate zircon and monazite structure-types have been synthesized, and their structures have been refined. These synthesized materials have provided the basis for detailed optical-absorption spectroscopy and electronic paramagnetic resonance studies of U, Pu, Cm, Np, and Gd incorporation into the zircon-structure type (45, 46). Geochronology. Recent progress in the utilization of zircon in age dating has come from the use of the sensitive high-resolution ion microprobe (SHRIMP), a method that allows the measurement of isotopic ratios on areas as small as 20–30 µm, thus providing age dates on separate zones within single crystals of zircon. Detrital zircons in a quartzite at Mount Narryer, Western Australia have been dated at 4,100– 4,300 million years ago (Ma), the oldest terrestrial minerals yet found (47). In Western Australia, the Jack Hills contain slightly younger (3,900– 4,270 Ma) detrital zircons (48). The zircons in Australia are individual, recycled grains in a younger (3,500 Ma) sequence of metamorphosed sedimentary rocks, as are similarly dated zircons from the Sino-Korean craton in northeast China (≥3,800 Ma) (49). The oldest so-called intact crust is found in the early Archean (3,800– 3,960 Ma) granitoids in northwestern Canada (50) and western Greenland (51). The zircons formed at the same time as the first crustal rocks on Earth. The oldest zircons in the solar system are found as rare inclusions in meteorites and were dated at 4,560 Ma (52). There are many hundreds of papers that can be cited to illustrate the use of zircon in dating very old rocks, but of greater importance are that these field studies establish the effects of geologic processes on the durability of zircon either under surface conditions of sedimentary transport (29) or at much higher temperatures, e.g., granulite facies metamorphism (53), extremely high metamorphic pressures (54, 55, 56 and 57), or meteorite impact (58). The most recent and dramatic example of the extraordinary ability of zircon to retain its U-Pb systematics is from studies of zircons from the Chicxulub impact structure of the Yucatan Peninsula (58, 59 and 60). The shocked zircons were exhumed from the Chicxulub basement rock during meteorite impact and dispersed in the fine dust of the impact cloud. Discordancies in the U-Pb systematics (e.g., Pb loss) are proportional to the extent of impact-induced shock textures, and isotopic resetting is consistent with partial lead loss at the time of impact (65 Ma), thus providing convincing support for the meteorite impact origin of the Chicxulub crater and its being the source of ejected material found at the Cretaceous-Tertiary (K-T) boundary in North America. Uranium–lead dating studies of shocked zircons in the fine-grained ejecta deposited in areas as wide apart as Colorado, Saskatchewan, and Haiti at the Cretaceous-Tertiary boundary have a predominant age of 545 Ma, in agreement with dates for shocked zircons from the Chicxulub crater (59 and 60). Only a material of remarkable durability, such as zircon, could preserve the isotopic signature of an event of such extreme conditions (60). Alteration and Disturbed U/Pb Systematics. Any process that can disturb the U/Pb isotopic systematics of zircon crystals has an important effect on whether the radiometrically determined dates are concordant or discordant; thus, there is an extensive literature that describes this type of alteration (30, 31 and 32, 61). Discordant ages are common, mainly because of Pb-loss and less frequently because of Uloss. The discordant ages are usually attributed to episodic loss of Pb or U during thermal events (61) and are enhanced by (i) physical degradation caused by microfracturing that is the result of the volume expansion associated with α-decay damage that results in an increase of surface area (57, 62, 63); and (ii) chemical alteration caused by radiation-induced amorphization that creates damaged, aperiod domains in which Pb-diffusion is enhanced and for which bulk leach rates are increased. Thus, the discordant ages are often the result of the accumulation of highly damaged regions resulting from α-decay damage; however, improved U/Pb dates may be obtained by removing the radiationdamaged regions either by physical abrasion or etching techniques (64). Differential etching experiments (using 48% HF) have shown by scanning electron microscopy that the removal of radiation damaged zones improved the concordance of U/Pb age dates in the remaining, unaltered material (32). Additionally, under extreme geologic conditions, e.g., zircons subjected to deformation in shear zones and altered by hydrothermal solutions, disturbed U/Pb systematics are clearly documented. Detailed experimental studies (65) have shown that experimentally induced U/Pb isotopic discordance in zircon is a complex function of zircon durability and the annealing of the radiation damage. Recent studies have investigated the incorporation of Pb into zircon (66) and the mechanisms by which Pb may be lost (67). The loss of Pb by diffusion is an extremely slow process (at 1,000°C, the diffusivity is 10−25 m2·s−1), and the low diffusivities confirm Pb closure temperatures in excess of 900°C and imply that U/Th/Pb isotopic ages are unlikely to be reset by thermal events alone except under the most extreme geologic conditions, i.e., partial melting or granulite-grade metamorphism (67). In summary, there are abundant data from geochronologic studies to show that natural zircons can quantitatively retain Pb for billions of years in the absence of episodic, thermally induced Pb loss (40). The minor alteration of zircon over long periods of time and under rather extreme conditions stands in contrast to the observations of other potential ceramic nuclear waste forms such as pyrochlore (68) for which the data of the type summarized above are simply not presently available. Diffusion of Tetravalent Cations and Rare Earth Elements. Even though diffusion rates for Pb are low, the diffusivities of U and Th at 1,100°C are orders of magnitude lower (69), suggesting complete containment of actinides for billions of years. The principal process by which this might change is α-decay radiation damage (40). Elements may preferentially segregate into aperiodic, damaged regions that have inherently higher solubilities than crystalline ceramics (70), and the annealing kinetics of recrystallization of damaged domains may substantially change the response of a material to the damage accumulation process (71). The development of a waste form for weapons 239Pu requires a knowledge of diffusion rates of the actinides, as well as elements with high neutron-capture cross-sections to control criticality (e.g., Hf and Gd). Natural zircons can contain up to several thousand parts per million rare earths, including up to 500 ppm of Gd, and zircon exhibits nearly complete solid solution with hafnon, HfSiO4. Thus, neutron absorbing nuclides may be incorporated into the zircon structure; however, the neutron absorber must remain within atomic scale proximity of the fissile radionuclide (239Pu or 235U). Recent determinations of diffusion rates for tetravalent cations (Th, U and Hf) give diffusivities in the order of 10 −22-10−20 m2·s−1 in the temperature range of 1,400–1650°C (72). Based on similarities in ionic radii, Pu4+ is expected to have a similar diffusivity, although slightly faster because of its smaller ionic radius, which means that it is essentially immobile under all but the most extreme geologic conditions. This is why fine-scale chemical zoning and isotopic signatures are preserved in the inherited cores of zircon crystals that have experienced protracted thermal events in their past history. In contrast, the rare-earth elements (REE) diffuse at rates 4–5 orders of magnitude faster than the tetravalent cations (73). The diffusion rates vary among the REEs in a systematic manner as a function of ionic radius; thus, for Gd, the estimated activation energy and diffusivity (1,000°C) are 189 kcal/mol and 3.2 × 10−26 m2·s−1, respectively. Again, the closure temperature of zircon for REEs is quite high (>1,000°C) for all but the smallest grains. Other phases that have been suggested as waste form phases, e.g., titanite and apatite, have closure temperatures many hundreds of degrees lower than those of zircon.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3436

Dissolution Studies. Studies of natural zircons under extreme laboratory conditions have confirmed the loss of U, Th, and Pb (30, 31, 65). The lead loss can be the result of grain boundary or volume diffusion (in which there is no dissolution of the zircon), or the bulk dissolution of zircon. However, at lower temperatures (<80°C) and near neutral pH values, i.e., conditions more pertinent to nuclear waste disposal, zircon is extremely insoluble. There are much less leach data in the literature than necessary for a full evaluation of zircon as a waste form. Crystalline zircon is stable to such an extent that the equilibrium concentrations of Zr and Si are in the order of 10−9 moles/liter (0.1 parts per billion) at 25° C (74). Dissolution of amorphous zircon followed a first order reaction based on Si concentrations. Zr concentrations remained <0.05 ppm, the instrument detection limit, because of precipitation of ZrO2 and ZrSiO4 (74). The leach rate of zircon increases with α-decay damage on the order of one to two orders of magnitude (75). Comparing the dissolution rate for a metamict zircon (74), r = 10−7 moles/m2/day at 80°C, pH 5, with the long-term rate of a nuclear waste borosilicate glass, r = 10−5 moles/m2/day (which equals 10−3 g/m2/day) in water at 98°C (76), shows that the dissolution rate of amorphous zircon is still considerably lower than that of glass in stagnant, silica-saturated solutions. In an open system (e.g., moving groundwater for which solubility limits are not reached), the leach rate for zircon (74) used in this comparison does not increase; however, the leach rate of borosilicate glass may increase by three orders of magnitude until reaching the final rate of reaction (21). Thus, one of the main advantages of zircon may be its high durability in an open system in which groundwaters are present because this allows considerably greater flexibility in disposal strategy (e.g., deep borehole). For this reason, it is essential to determine the maximum forward rate of the dissolution reaction. This is particularly challenging because the dissolution rate for zircon is low, and precipitation of zirconia removes Zr from the leaching solution. Recently, a high-temperature Soxhlet extractor was designed to measure the forward rate of dissolution of zircon in the range of 120–250°C. The measured rates were 4.1 × 10−4 g/m2/day at 250°C, 1.7 × 10−4 g/m2/day at 200°C, and 7.1 × 10−5 g/m2/day at 120°C. The rate extrapolated to 90°C is 4.6 × 10−5 g/m2/day; therefore, in an open system, in the absence of a solubilitylimiting phase, a 100-µm crystal of zircon would require 150,000 years for complete dissolution (77, 78). Under static conditions, the dissolution rate is substantially reduced. The experimental results are consistent with the high chemical durability of zircon in a wide range of geologic environments. Physical Properties. The mechanical properties of zircon change with increasing α-decay event dose (79) and as a result of implantation by Pb ions (540 keV) up to fluences of 3.3 × 1011 to 5 × 1015 ions/cm2 [a dose range that spans the crystalline-to-amorphous transition (80)]. The α-decay-induced softening leads to a decrease in hardness (40%) and the bulk elastic modulus (70%) (81), but there is an increase in fracture toughness probably caused by crack-tip blunting by the aperiodic domains (79). The ion beam irradiation results in softening (70%) and a decrease in modulus (42%). The principal effect of the changes in mechanical properties is the formation of a pronounced fracture system (62) caused by differential volume expansion in zones of different α-decay doses. A model has been developed that describes the radial and concentric fracture sets that are characteristic of zircon: principally, a function of the degree of damage (e.g., the amorphous fraction), the zone thicknesses, and the confining pressure (82). This type of analysis is required for evaluating the development of microfractures as a function of radiation damage, particularly for disposal in a deep borehole. Radiation Damage. Radiation damage resulting from the α-decay of 239Pu and its daughter products (e.g., 235U) has an important effect on the physical and chemical durability of actinide-bearing zircons. Depending on the waste loading, significant doses (>1018 α-decay events/g) accumulate, and crystalline phases become aperiodic in relatively short periods of time (103 years) (Fig. 3) (83, 84). Note that natural zircons can accumulate relatively high α-decay doses (horizontal line in Fig. 3), and this allows the comparison of data resulting from accelerated irradiation techniques (either actinide doping with short lived α-decay nuclides such as 238Pu or 244Cm that have reached α-decay doses of 1019 α-decay event/g or ion beam irradiation experiments). In an α-decay event, the α-particle dissipates most of its energy [4.5–5.8 million electronvolts (MeV) for actinides] (1 eV = 1.602 × 10–19J) by ionization processes over a range of 16–22 µm but undergoes enough elastic collisions along its path to produce several hundred isolated atomic displacements. The largest number of displacements occurs near the end of the α-particle range. The more massive but lower-energy αrecoil (86 keV 235U recoil from decay of 239Pu) dissipates nearly all of its energy in elastic collisions over a very short range, 30–40 nm, causing ≈1,000 atomic displacements. The density of energy deposited into the cascade is high (up to 1 eV/atom) and occurs over an extremely short time (<10−12 s). Thus, α single α-decay event generates ≈1,400 atomic displacements, which is significantly greater than the 0.1 displacements generated per β-decay event. Clearly, α-decay from incorporated actinides will have a profound effect on the structure of a crystalline solid. The cumulative effect of dose will be time- and temperature-dependent because of annearing and recrystallization of damaged areas. Studies of radiation effects in zircon have a long history (86, 87, 88, 89, 90 and 91). Zircon undergoes a radiation-induced transformation from the periodic-to-aperiodic state (metamict state) at doses over the range of 1018-1019 α-decay events/g [which equals 0.2–0.6 displacements per atom (dpa)] with a density decrease and a corresponding volume expansion of 18%. Previous studies include analysis of natural zircons that have accumulated α-decay event damage up to doses of nearly 0.7 dpa over 550 million years (63, 87, 88), 238Pu-doped zircons (half-life = 87.7 years) up to doses of 0.7 dpa in 6.5 years (44, 91, 92), and heavy ion beam irradiation using 2 MeV He+, 0.8 MeV Ne+, 1.5 MeV Ar+, 0.7–1.5 MeV Kr+, and 1.5 MeV Xe+ up to doses of 0.2–2.3 dpa in times of <1 hour (91, 93). All three types of damage experiments include detailed studies of annealing kinetics. The natural zircon studies (e.g., of samples that are 550 million years old) and the experimental results for ion beam irradiation experiments of <1-hour duration have a range of dose rates >108. In situ ion beam irradiation combined with high resolution transmission electron microscopy have demonstrated that the damage microstructures and the ingrowth of damage with increasing dose can be simulated by

FIG. 3. Cumulative α-decay dose as a function of waste form storage time for a ceramic, e.g., zircon, containing 1 and 10 wt % loading of (as in ref. 85). The dashed horizontal line indicates maximum doses reached in natural zircon from decay of uranium and thorium and daughter products. 239Pu

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3437

using heavy ion irradiation (Fig. 4). Thus, there is a firm basis for predicting the microstructure of the radiation-damaged zircon as a function of dose, temperature, and type of radiation (91, 93, 94). Systematic studies have been completed for monazite- and zircon-structure orthophosphates of a wide variety of A-site end-member compositions (95, 96) and orthosilicates (97, 98 and 99). Zircon, hafnon (HfSiO4), thorite (tetragonal ThSiO4), and huttonite (monoclinic ThSiO4) become amorphous in a two-stage process with increasing temperature when irradiated with 800 keV Kr+ or Xe+ ions in the temperature range of 20–1,100 K. The temperature above which amorphization does not occur (i.e., the temperature at which the rate of simultaneous annealing is equal to the rate of damage accumulation) increased in the order huttonite, zircon, hafnon, and thorite. When irradiated with heavy ions, all four orthosilicates may decompose into crystalline oxides: ZrO2, HfO2, or ThO2 plus amorphous SiO2 (98). Ion beam irradiation studies of synthetic and natural zircons and monazites (with impurities) revealed that impurities lower the dose required for amorphization and correspondingly increase the temperature above which a material cannot be amorphized. This suggests that impurities increase susceptibility to amorphization and inhibit annealing, particularly when coupled charge balance substitutions are required (96, 99). Details of structural rearrangements during damage accumulation and annealing have been obtained by extended x-ray absorption fine structure spectroscopy (EXAFS) studies of metamict zircons (89, 100). Metamictization is accompanied by major atomic reorganization: loss of well defined medium range order, disruption of the immediate environment of Zr, a decrease of the average Zr-coordination number, and tilting and distortion of SiO4 polyhedral. A two-stage thermal annealing process was observed that, at lower temperatures (400–500°C), resulted in the formation of minor Zr-rich domains (100). Finally, transmutation effects (239Pu decays to 235U) are important in crystalline materials because they may lead to phase instability. In the case of zircon, the solubility of U in zircon is only known approximately (4 ± 2 mole percent) (43); thus, for higher concentrations of U, one may expect the formation of USiO4, also with the zircon structure, and UO2. There are few crystalline ceramics for which such a wide variety of data on radiation damage are available. Based on these data, for a waste loading of 10 wt % of 239Pu under ambient conditions, the zircon will reach the saturation dose of damage (1.2 × 1019

FIG. 4. Comparison of radiation damage in natural (A–C) and 1.5 MeV Kr+-irradiated synthetic (D–F) zircon. (A) 5 × 1013 α-decay events/ mg (0.003 dpa). (B) 1.8 × 1015 α/mg (0.091 dpa). (C) 6.4 × 1015 α/mg (0.32 dpa). (D) 5 × 1013 Kr+/cm2 (0.057 dpa). (E) 1.5 × 1014 Kr+/cm2 (0.17 dpa). (F) 3 × 1014 Kr+/cm2 (0.34 dpa). Complete amorphization in both was observed after 0.5–0.55 dpa. Figure courtesy of L. M. Wang (91).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3438

α-decay events/g or 0.8 dpa) in <2,000 years (Fig. 3); thus, the properties of zircon must be considered in light of its radiation-damaged, aperiodic state. Based on the extensive database, it is possible to model the damage accumulation (e.g., percent amorphous fraction) as a function of dose and temperature for zircon and compare the results to those of other phases (e.g., apatite) (85). Fig. 5 is a plot of the critical amorphization dose for zircon as compared with a silicate apatite (data for both phases are based on ion beam irradiation experiments). Above 225°C, zircon with a 10 wt % loading of 239Pu will remain in the crystalline state because of thermal annealing. In contrast, silicate apatite anneals readily and will remain in the crystalline state at ambient temperatures. Such models are very sensitive to the activation energies for thermal annealing (101), and more recent work has suggested that temperatures as high as 400°C will be required to maintain the crystallinity of zircon (98). The efficacy of such an approach is that modeled results can be confirmed by comparison to naturally occurring zircons of known thermal history, and the modeled results are in good agreement with the amorphous fraction accumulation determined for natural zircons (unless there has been episodic thermal annealing) (101). The same type of analysis allows one to calculate minimum storage temperatures required to maintain the crystallinity of any actinide-bearing waste form (Fig. 6). At present, sufficient data for this type of analysis are only available for zircon. Stored energy values have been carefully determined (102) by transposed temperature drop calorimetry over the range of the periodic-toaperiodic transition on a suite of zircons from Sri Lanka (550 million years old). The energy released during annealing varies sigmoidally as a function of α-decay event dose reaching a saturation value of 322 ± 16 J/g at doses >5 × 1018 α-decay events/g. This is greater than values typical of nuclear waste glasses, which are generally <150 J/g and saturate at a dose of 1018 α-decay events/g; however, sudden release of this energy is not anticipated to cause a significant rise in temperature for either the glass or zircon. The magnitude of the enthalpy of annealing suggests that the radiation damage is pervasive on the scale of fractions of nanometers, perhaps leading to the formation of microdomains of amorphous SiO2-rich and ZrO2-rich regions in the metamict state. This suggestion is consistent with observations made by secondary ion mass spectrometry (SIMS) and high resolution transmission electron microscopy (HRTEM) (70) and extended x-ray absorption fine structure spectroscopy (EX-AFS) of annealed zircons (100). SUMMARY There can be little doubt that a phase such as zircon provides a demonstrable case for its long-term chemical and mechanical durability as an actinide waste form. The database for zircon is unique and extensive, providing ample opportunity to confirm experimental and modeled results against actual behavior in a wide range of geologic environments. If the issue is one of long-term verification of materials performance, no other comparable database exists for an actinide-bearing phase, as there is for zircon.

FIG. 5. Critical amorphization dose vs. storage temperature for silicate apatite and zircon containing 10 wt % 239Pu (85). FIG. 6. Minimum storage temperature of zircon containing 10 and 2 wt % 239Pu, respectively, vs. storage time to ensure complete crystallinity of zircon. Such modeled calculations are very sensitive to the activation energy. The dark solid line is for an activation energy of 3.6 eV; the lower, thinner line if for an activation energy of 2.1 eV (101). Are there other mineralogic candidates? Most definitely yes, and prominent among them are monazite and the polymorphs of zirconia. I believe that we have only just scratched the surface of mineralogic applications to issues related to nuclear waste disposal. Are there other mineralogic applications? Much can be learned about the alteration of UO2 in spent nuclear fuel by studying the alteration products of uraninite, UO2 + x (102, 103). I have benefited greatly from collaborations with Werner Lutze, Lumin Wang, and Bill Weber, and most importantly with students, Gregory Lumpkin and Al Meldrum, in the development of the ideas presented in this paper. Lynn Boatner's enthusiasm for this work has kept me in the game. I thank John Hanchar for very useful discussions of diffusion in zircon. This work has been sustained by support by Basic Energy Research Sciences of the Department of Energy (DE-FG02-97ER45656).

1. Linking Legacies Connecting the Cold War Nuclear Weapons Production Processes to Their Environmental Consequences (1997) Department of Energy, Office of Environmental Management 0319, Washington, DC. 2. Ahearne, J. F. (1997) Phys. Today 50, 24–29. 3. Crowley, K. D. (1997) Phys. Today 50, 32–39. 4. Richardson, J. A. (1997) Proc. Inst. Mech. Eng. Part A 211, 381–392. 5. Committee on International Security and Arms Control, National Academy of Sciences (1994) Management and Disposition of Excess Weapons Plutonium (National Academy Press, Washington, DC). 6. Panelon Reactor-Related Options for the Disposition of Excess Weapons Plutonium, National Research Council (1995) Management and Disposition of Excess Weapons Plutonium: Reactor-Related Options (National Academy Press, Washington, DC). 7. von Hippel, F. N. (1998) Nature (London) 394, 415–416. 8. Stoll, W. (1998) Mat. Res. Soc. Bull. 23, 6–16. 9. Oversby, V. M., McPheeters, C. C., Degueldre, C. & Paratte, J. M. (1997) J. Nucl. Mat. 245, 17–26. 10. Mark, J. C. (1993) Sci. Global Security 4, 111–128. 11. Hedin, A. (1997) SKB Technical Report 97–13 (Swedish Nuclear Fuel and Waste Management Co., Stockholm). 12. Langmuir, D. (1997) Aqueous Environmental Geochemistry (Prentice–Hall, Englewood Cliffs, NJ). 13. Hatch, L. P. (1953) Am. Sci. 41, 410–421. 14. Roy, R. (1979) in Scientific Basis for Nuclear Waste Management, ed. McCarthy, G. J. (Plenum, New York), Vol. 1, pp. 1–20. 15. Haaker, R. F. & Ewing, R. C. (1981) Naturally Occurring Crystalline Phases: Analogues for Radioactive Waste Forms (Pacific Northwest Laboratory, Richland, WA). 16. Ringwood, A. E. (1978) Safe Disposal of High Level Nuclear Reactor Wastes: A New Strategy (Australian National Univ. Press, Canberra, Australia ). 17. Ringwood, A. E., Kesson, S. E., Ware, N. G., Hibberson, W. & Major, A. (1979) Nature (London) 278, 219–223.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NUCLEAR WASTE FORMS FOR ACTINIDES

3439

18. Kerr, R. A. (1979) Science 204, 289–291. 19. Carter, L. J. (1979) Science 205, 287–289. 20. Garmon, L. (1981) Sci. News 120, 396–399. 21. Lutze, W. & Ewing, R. C., eds. (1988) Radioactive Waste Forms for the Future (North–Holland, Amsterdam). 22. Ewing, R. C., Weber, W. J. & Lutze, W. (1995) in Disposal of Weapon Plutonium Approaches and Prospects, eds. Merz, E. R. & Walter, C. E. (Kluwer, Dordrecht, The Netherlands), pp. 65–83. 23. Hutton, C. O. (1950) Geol. Soc. Am. Bull. 61, 635–716. 24. Cochran, S. G., Dunlop, W. H., Edmunds, T. A., MacLean, L. M. & Gould, T. H. (1997) Fissile Material Disposition Program Final Immobilization Form Assessment and Recommendation (Lawrence Livermore National Laboratory, Livermore, CA). 25. Robinson, K., Gibbs, G. V. & Ribbe, P. H. (1971) Am. Mineral. 56, 782–790. 26. Taylor, M. & Ewing, R. C. (1978) Acta Crystallogr. B 34, 1074–1079. 27. von Keller, C. (1963) Nukleonik 5, 41–48. 28. Hanchar, J. M. & Miller, C. F. (1993) Chem. Geol. 110, 1–13. 29. Riggs, N. R., Lehman, T. M., Gehrels, G. E. & Dickinson, W. R. (1996) Science 273, 97–100. 30. Pidgeon, R. T., O'Neil, J. R. & Silver, L. T. (1966) Science 154, 1538–1540. 31. Craig, H. (1968) Science 159, 447. 32. Krogh, T. E. & Davis, G. L. (1975) Year Book Carnegie Inst. Washington 74, 619–623. 33. Suzuki, K. (1987) Geochem. J. 21, 173–182. 34. Harker, A. B. & Flintoff, J. F. (1990) J. Am. Ceramic Soc. 73, 1901–1906. 35. Anderson, E. B., Burakov, B. E. & Pazukhin, E. M. (1993) Radiochim. Acta 60, 149–151. 36. Burakov, B. E. (1993) Proc. SAFE WASTE 93 2, 19–28. 37. Anderson, E. B., Burakov, B. E. & Vasiliev, V. G. (1993) Proc. SAFE WASTE 93, 2, 29–33. 38. Ewing, R. C., Lutze, W. & Weber, W. J. (1995) J. Mat. Res. 10, 243–246. 39. Gentry, R. V., Sworski, T. J., McKown, H. S., Smith, D. H., Eby, R. E. & Christie, W. H. (1982) Science 216, 296–298. 40. Ludwig, K. R., Zartman, R. E. & Goldich, S. S. (1984) Science 223, 835. 41. Speer, J. A. (1982) in Reviews in Mineralogy, Orthosilicates, ed. Ribbe, P. H. (Mineralogical Soc. Am., Washington, DC), pp. 67–135. 42. Speer, J. A. & Cooper, B. J. (1982) Am. Mineral. 67, 804–808. 43. Mumpton, F. A. & Roy, R. (1961) Geochim. Cosmochim. Acta 21, 217–238. 44. Weber, W. J. (1991) Radiat. Effects Defects Solids 115, 341–349. 45. Poirot, I. S., Kot, W. K., Edelstein, N. M., Abraham, M. M., Finch, C. B. & Boatner, L. A. (1989) Phys. Rev. B 39, 6388–6394. 46. Poirot, I., Kot, W., Shalimoff, G, Edelstein, N., Abraham, M. M., Finch, C. B. & Boatner, L. A. (1988) Phys. Rev. B 37, 3255–3264. 47. Froude, D. O., Ireland, T. R., Kinny, P. D., Williams, I. S., Compston, W., Williams, I. R & Myers, J. S. (1983) Nature (London) 304, 616–618. 48. Maas, R., Kinny, P. D., Williams, I. S., Froude, D. O. & Compston, W. (1992) Geochim. Cosmochim. Acta 56, 1281–1300. 49. Liu, D. Y., Nutman, A. P., Compston, W., Wu, J. S. & Shen, Q. H. (1992) Geology 20, 339–342. 50. Bowring, S. A., Williams, I. S. & Compston, W. (1989) Geology 17, 971–975. 51. Nutman, A. P., Friend, C. R., Kinny, P. D. & McGregor, V. R. (1993) Geology 21, 415–418. 52. Ireland, T. R. & Wlotzka, F. (1992) Earth Planet. Sci. Lett. 109, 1–10. 53. Roberts, M. P. & Finger, F. (1997) Geology 25, 319–322. 54. Deutsch, A. & Schärer, U. (1990) Geochim. Cosmochim. Acta 54, 3427–3434. 55. Claoué-Long, J. C., Sobolev, N. V., Shatsky, V. S. & Sobolev, A. V. (1991) Geology 19, 710–713. 56. Wayne, D. M. & Sinha, A. K. (1988) Contrib. Mineral. Petrol. 98, 109–121. 57. Wayne, D. M. & Sinha, A. K. (1992) J. Geol. 100, 593–603. 58. Bohor, B. F., Betterton, W. J. & Krogh, T. E. (1993) Earth Planet. Sci. Lett. 119, 419–424. 59. Krogh, T. E., Kamo, S. L., Sharpton, V. L., Marin, L. E. & Hildebrand, A. R. (1993) Nature (London) 366, 731–734. 60. Kamo, S. L. & Krogh, T. E. (1995) Geology 23, 281–284. 61. Davis, G.L., Hart, S. R. & Tilton, G.R. (1968) Earth Planet. Sci. Lett. 5, 27–34. 62. Chakoumakos, B. C., Murakami, T., Lumpkin, G. R. & Ewing, R. C. (1987) Science 236, 1556–1559. 63. Holland, H. D. & Gottfried, D. (1955) Acta Crystallogr. 8, 291–300. 64. Krogh, T. (1995) Geotimes Nov., 20–22. 65. Sinha, K. A., Wayne, D. M. & Hewitt, D. A. (1992) Geochim. Cosmochim. Acta 56, 3551–3560. 66. Watson, E. B., Cherniak, D. J., Hanchar, J. M., Harrison, T. M. & Wark, D. A. (1997) Chem. Geol. 141, 19–31. 67. Lee, J. K. W. (1993) in Defects and Processes in the Solid State: Geoscience Applications, The McLaren Volume, eds. Boland, J. N. & FitzGerald, J. D. (Elsevier Science, Amsterdam), pp. 423–446. 68. Lumpkin G. R. & Ewing R. C. (1996) Am. Mineral. 81, 1237–1248. 69. Cherniak, D. J. & Watson, E. B. (1998) Proc. Annu. Meet. Geol. Soc. Am. (abstr.). 70. McLaren, A. C., FitzGerald, J. D. & Williams, I. S. (1994) Geochim. Cosmochim. Acta 58, 993–1005. 71. Cherniak, D. J., Lanford, W. A. & Ryerson, F. J. (1991) Geochim. Cosmochim. Acta 55, 1663–1673. 72. Cherniak, D. J., Hanchar, J. M. & Watson, E. B. (1997) Contrib. Mineral. Petrol. 127, 383–390. 73. Cherniak, D. J., Hanchar, J. M. & Watson, E. B. (1997) Chem. Geol 134, 289–301. 74. Tole, M. P. (1985) Geochim. Cosmochim. Acta 49, 453–458. 75. Ewing, R. C., Haaker, R. F. & Lutze, W. (1982) in Scientific Basis for Nuclear Waste Management V, ed. Lutze, W. (Materials Research Soc., Pittsburgh), pp. 389–397. 76. Werme, L. O., Björner, I. K., Bart, G., Zwicky, H.-U., Grambow, B., Lutze, W., Ewing, R. C. & Magrabi, C. (1990) J. Mat. Res. 5, 1130–1146. 77. Helean, K. B. (1997) M.S. thesis (Univ. of New Mexico, Albuquerque). 78. Helean, K. B., Lutze, W. & Ewing, R. C. (1999) Proc. 100th Annu. Meet. Am. Ceramic Soc., in press. 79. Chakoumakos, B. C., Oliver, W. C., Lumpkin, G. R. & Ewing, R. C. (1991) Radiat. Effects Defects Solids 118, 393–403. 80. Oliver, W. C., McCallum, J. C., Chakoumakos, B. C. & Boatner, L. A. (1994) Radiat. Effects Defects Solids 132, 131–141. 81. Ozkan, H. (1976) J. Appl. Phys. 47, 4772–4779. 82. Lee, J. K. W. & Tromp, J. (1995) J. Geophys. Res. 100, 17,753–17,770. 83. Ewing, R. C., Weber, W. J. & Clinard, F. W., Jr. (1995) Prog. Nucl. Energy 29, 63–127. 84. Weber, W. J., Ewing, R. C., Catlow, C. R. A., Diaz de la Rubia, T., Hobbs, L. W., Kinoshita, C., Matzke, H., Motta, A. T., Nastasi, M., Salje, E. K. H., et al. (1998) J. Mat. Res. 13, 1434–1484. 85. Weber, W. J., Ewing, R. C. & Meldrum, A. (1997) J. Nucl. Mat. 250, 147–155. 86. Ewing, R. C. (1994) Nucl. Instr. Methods Phys. Res. B 91, 22–29. 87. Murakami, T., Chakoumakos, B. C., Ewing, R. C., Lumpkin, G. R. & Weber, W. J. (1991) Am. Mineral. 76, 1510–1532. 88. Nasdala, L., Pidgeon, R. T. & Wolf, D. (1996) Geochim. Cosmochim. Acta 60, 1091–1097. 89. Farges, F. & Calas, G. (1991) Am. Mineral 76, 60–73. 90. Woodhead, J. A., Rossman, G. R. & Silver, L. T. (1991) Am. Mineral. 76, 74–82. 91. Weber, W. J., Ewing, R. C. & Wang, L. M. (1994) J. Mat. Res. 9, 688–698. 92. Weber, W. J. & Maupin, G, D. (1988) Nucl. Instr. Methods Phys. Res. B 32, 512–515. 93. Wang, L. M., Ewing, R. C., Weber, W. J. & Eby, R. K. (1993) in Beam-Solid Interactions: Fundamentals and Applications, eds. Nastasi, M., Harriott, L. R., Herbots, N. & Averback, R. S. (Materials Research Soc., Pittsburgh), pp. 451–456. 94. Ewing, R. C. (1993) in Scientific Basis for Nuclear Waste Management XVI, eds. Interrante, C. G. & Pabalan, R. T. (Materials Research Soc., Pittsburgh) pp. 559–568. 95. Meldrum, A., Boatner, L. A., Wang, L. M. & Ewing, R. C. (1997) Nucl. Instr. Methods Phys. Res. B 127/128, 160–165. 96. Meldrum, A., Boatner, L. A. & Ewing, R. C. (1997) Phys. Rev. B 56, 13805–13814. 97. Meldrum, A., Boatner, L. A., Zinlde, S. J., Wang, S. Wang, L. M. & Ewing, R. C. (1999) Can. Mineral., in press. 98. Meldrum, A., Zinkle, S. J., Boatner, L. A. & Ewing, R. C. (1998) Nature (London) 395, 56–58. 99. Meldrum, A., Zinkle, S. J., Boatner, L. A. & Ewing, R. C. (1999) Phys. Rev. B, in press. 100. Farges, F. (1994) Phys. Chem. Miner. 20, 504–514. 101. Meldrum, A., Boatner, L. A., Weber, W. J. & Ewing, R. C. (1998) Geochim. Cosmochim. Acta 62, 2509–2520. 102. Ellsworth, S., Navrotsky, A. & Ewing, R. C. (1994) Phys. Chem. Miner. 21, 140–149. 103. Janeczek, J., Ewing, R. C., Oversby, V. M. & Werme, L. (1996) J. Nucl. Mat. 238, 121–130. 104. Burns, P. C., Finch, R. J., Hawthorne, F. C., Miller, M. L. & Ewing, R. C. (1997) J. Nucl. Mat. 249, 199–206.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ILLITE AND HYDROCARBON EXPLORATION

3440

Illite and hydrocarbon exploration

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3440–3446, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. DAVID R. PEVEAR PNAS is available online at www.pnas.org.

Exxon Production Research Co., P.O. Box 2189, Houston, TX 77252-2189 ABSTRACT Illite is a general term for the dioctahedral mica-like clay mineral common in sedimentary rocks, especially shales. Illite is of interest to the petroleum industry because it can provide a K-Ar isotope date that constrains the timing of basin heating events. It is critical to establish that hydrocarbon formation and migration occurred after the formation of the trap (anticline, etc.) that is to hold the oil. Illite also may precipitate in the pores of sandstone reservoirs, impeding fluid flow. Illite in shales is a mixture of detrital mica and its weathering products with diagenetic illite formed by reaction with pore fluids during burial. K-Ar ages are apparent ages of mixtures of detrital and diagenetic end members, and what we need are the ages of the end members themselves. This paper describes a methodology, based on mineralogy and crystallography, for interpreting the K-Ar ages from illites in sedimentary rocks and for estimating the ages of the end members. Illite is a general term for the dioctahedral mica-like clay mineral common in sedimentary rocks, especially shales (1, 2). Although it has a strict mineralogical definition (3), the name illite is often loosely used for any clay mineral with a 1-nm repeat in the x-ray powder diffraction data (4). Because shale is abundant at the earth's surface, its typical clay mineral, illite, impacts human welfare in several ways. In the petroleum industry, illite is of interest for two reasons: (i) It can provide an isotope date constraining basin heating events, and (ii) it may precipitate in the pores of sandstone reservoirs, impeding fluid flow. Because it is a potassium aluminum phyllosilicate, its time of formation can be determined by using K-Ar isotope dating. Illite holds Ar tightly because of the difficulty of migration (diffusion) through the crystal structure layers (5) at low temperatures. Of particular concern in resource exploration is the timing of hydrocarbon (HC) generation. When were the organic-rich source rock shales heated to ≈100°C, cracking the solid organic matter to oil and gas? It is critical to establish that HC formation and migration occurred after the formation of the trap (anticline, etc.) that is to hold the oil. We have long been able to find traps by using seismic methods, but we seldom are able to predict the presence of HC without expensive drilling. If integrated geologic evaluation of outcrops or nearby wells can show HC generation after trap formation, the risk of drilling a dry hole is reduced. Because illite forms in shales in response to heating in the same temperature range as oil formation (6), its K-Ar age is useful indeed. It has been recognized for some time (7) that illite in shales is a mixture of detrital mica and its weathering products with diagenetic illite precipitated from pore fluids during burial. Two important lines of evidence support this conclusion. First, grain size vs. mineralogy relations show a mixture of 2M1 and 1M (including 1Md) polytypes, with 1M increasingly abundant in the finer size fractions (7). Polytypes are a variety of polymorph distinguished by various repeating stacking arrangements of identical layers (3). 1M means one layer, monoclinic, etc. The 2M1 polytype certainly is expected (8) for the large detrital micas eroded from slates, schists, and phyllites. As we shall see, diagenetic illite that grows in bentonites and sandstones is exclusively 1M, which suggests that similar material mixed with 2M1 muscovite in shales is also diagenetic. Secondly, grain size vs. K-Ar age relations in shales invariably show age decreasing with grain size: The coarse fractions are typically older than the depositional (stratigraphic) age of the shale whereas the fine fractions are younger (9). The foregoing shows that illite in shales is a mixture of detrital and diagenetic components, with the latter more abundant in the fine fractions. But it also identifies the principal problem with practical use of K-Ar dating of illite in shales: The ages of bulk mixtures of detrital and diagenetic end members are rather meaningless, and what we need are the separate ages of the end members themselves. I describe a methodology, based on mineralogy and crystallography, for interpreting the K-Ar ages from illites in sedimentary rocks and for estimating the ages of the end members. Illite in Sedimentary Rocks One cannot discuss illite without touching the subject of mixed-layer illite/smectite (I/S), a mineral in which unit cell scale layers of illite and smectite are shuffled like a deck of cards. Clay mineralogists typically disaggregate a sample and prepare one or more grain size fractions as oriented aggregates (10) on a slide for x-ray powder diffraction (XRD) with a focusing diffractometer. Because the particles orient with 00l parallel to the slide, only the 00l reflections appear in the data. Illite has a series of 00l reflections based on a 1-nm periodicity; smectite, with interlayer water, has a 1.4-nm periodicity that can vary with humidity or treatment with organics. XRD patterns (00l series) for I/S typically are nonperiodic (nonintegral; they do not obey Bragg's Law) and do not look like a physical mixture of illite and smectite. They are interpreted (6) to result from a single diffraction from a faulted layer structure composed of two types of unit cells. There is a mature technology (10) for quantifying and modeling XRD data from mixed-layer clay minerals. I/S is common in shales; indeed, much of the illite in shales may be in the form of I/S. The percent of illite in I/S typically increases with depth and temperature in most of the world's sedimentary basins and with geologic age (6). This has been interpreted (or inferred) to indicate a progressive solid state or layer-by-layer transformation of smectite to illite in which the initial structure of the smectite is inherited by the illite (11). More recently, Nadeau (6, 10, 12) has introduced the dual concepts of fundamental particles and interparticle diffraction to explain mixedlayer clays. In this view, thin (2- to 10-unit

Abbreviations: HC, hydrocarbon; I/S, illite/smectite; XRD, x-ray powder diffraction; AFM, atomic force microscopy; IAA, Illite Age Analysis; my, million years.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ILLITE AND HYDROCARBON EXPLORATION

3441

cells) illite crystals precipitate in shales whereas smectite, feldspars, and other minerals dissolve. The diffraction effects of I/S result from coherent (in 00l) scattering amongst thin face-to-face illite crystals with hydrated interfaces that behave like smectite (are turbostratic). As crystals grow thicker, the number of interfaces decreases, which is seen in the XRD data as a decrease in smectite component of I/S. The observation of thin ideomorphic crystals of 1M illite with 1-nm surface growth steps in sandstones and shales (13) supports Nadeau's ideas. The subject of I/S remains controversial, but here I assume that increase in illite content of I/S with burial depth simply represents the growth of progressively thicker illite crystals. To extract useful chronologic information from K-Ar dating of illite, I have found the concept of grain-size vs. age spectra (size–age spectra) useful (Fig. 1a). A sample is routinely divided into three clay-size fractions: coarse (C = 0.2–2.0 µm), medium (M = 0.02–0.2 µm), and fine (F = <0.02 µm), and, for each, a routine K-Ar age is obtained. Using the <2-µm fraction generally excludes feldspar, so that the only Kbearing phases are illite and micas. Plotting these as simple bar graphs has revealed three major spectra shapes for sedimentary rocks: inclined, flat, and benched. These are typical of shales, K-bentonites, and sandstones, respectively. An inclined spectrum (Fig. 1a) is typical for shales, which are deposited with a wide initial size range of detrital micas. Usually the C fraction is older than the depositional age, but this depends on the proportion of diagenetic illite. The F fraction is typically younger than the depositional age because of the dominance of diagenetic illite. Importantly, as pointed out 35 years ago by Hower et al. (9), there is no way to use these dates, except as crude limits. All fractions appear to be physical mixtures, and we do not know the proportions. The mixture of old and young illite in shales can for some samples give K-Ar ages fortuitously close to depositional age (9). Note that K-Ar data from shales cannot be successfully interpreted by using the isochron method because shales are mixtures of things that formed at different times. They do, however, often give nice-looking, linear, but useless, “mixochrons.” Bentonites (stratigraphic definition) are an uncommon class of shale bed consisting of air-fall glassy volcanic ash altered to smectite (3). Kbentonites (3) are those that have undergone subsequent diagenesis to illite or I/S. They are of great value to illite studies because they do not contain detrital dioctahedral micas, only diagenetic illite. The size–age spectrum of a K-bentonite is typically flat (Fig. 1b); i.e., all size fractions have the same K-Ar age, younger than depositional age. Bentonites give the mean diagenetic age directly. If bentonites were common in the stratigraphic record, we could forget about trying to get meaningful ages from ordinary shales. They are useful for our dating problem because they give us an idea of what the pristine diagenetic illite is like. Mineralogic studies of K-bentonites are numerous, and XRD shows the illite and I/S to be entirely 1M polytype with moderate amounts of 120° rotational disorder (14, 15). 2M1 muscovite is never found as a diagenetic phase in K-bentonites of sedimentary basins. This is good news because it gives us a possible way to differentiate and quantify the diagenetic and detrital components in shales.

FIG. 1. (a) Size–age spectrum for shale. The sample is divided into three clay-size fractions: coarse (C = 0.2–2.0 µm), medium (M = 0.02–0.2 µm), and fine (F = <0.02 µm). An inclined spectrum is typical for shales, which are deposited with a wide initial size range of detrital micas. Usually, the C fraction is older than the depositional age, but this depends on the proportion of detrital mica. The F fraction is typically younger than the depositional age because of the dominance of diagenetic illite. (b) Size–age spectrum for a K-Bentonite is flat; i.e., all size fractions have the same K-Ar age, younger than depositional age. Bentonites give the diagenetic age directly because they do not contain detrital illite. Atomic force microscopy (AFM) shows the K-bentonite illite crystals to be only a few nanometers thick (Fig. 2), with a predominance of 1-nm growth steps. The former is confirmed by XRD studies of the 00l reflections (16); the latter agrees with their 1M polytype. The extraordinary thinness likely explains the abundance of diagenetic illite in the fine fractions of shales. Sandstones with a shale-like depositional matrix or abundant lithic grains have size–age spectra similar to shales and will not be discussed further. Clean sandstones consist only of sand-sized grains of quartz, feldspars, mica, etc., and lack depositional clay. They are deposited in a high-energy environment (like a beach) in which the fines are winnowed away. During diagenesis, feldspars and other rock constituents may react with pore fluids to precipitate illite or other diagenetic clays; hence, the fine material in these sandstones tends to be mostly diagenetic, and more so than for shales. A typical sandstone size–age spectrum (Fig. 3) is bench-shaped; i.e., the C fraction is older than depositional age whereas the M and F fractions have the same age, younger than depositional age. This flattening out in the finer fractions permits us to conclude that fine detrital mica is absent in these fractions and that we have measured the mean age of illite formation. Unfortunately, diagenetic illite is not so universally abundant in sandstones as it is in shales, and not all sandstones are clean sandstones. There are many studies of pore filling illites, both mineralogic and K-Ar dating (2, 6, 10). The abundant literature is primarily due to the negative effect illite has on permeability

FIG. 2. AFM deflection image of illite crystals from the Tioga K-bentonite. Scale is in nanometers. Individual growth steps are 1 nm high; the largest crystal is 7 nm thick. The image was made in air, contact mode, on a Digital Instruments (Santa Barbara, CA) Multi-Mode Nannoscope IIIa.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ILLITE AND HYDROCARBON EXPLORATION

3442

FIG. 3. Size–age spectrum of sandstone. The spectrum is typically bench-shaped; i.e., the C fraction is older than depositional age whereas the M and F fractions have the same age, younger than depositional age. The flattening out in the finer fractions indicates that fine detrital mica is absent in these fractions and that we have measured the mean age of illite formation. Symbols are same as in Fig. 1. of sandstone petroleum reservoirs. The illites are typically ideomorphic with a pronounced fibrous (lath) habit (long axis is crystallographic a axis) making them interesting subjects for microscopy (Fig. 4). They are often called “hairy illite” in the petroleum industry. The crystals are ideomorphic because they precipitate unconstrained from fluid in a relatively large pore. They are all 1M polytype, with a minor 120° rotational disorder. As in K-bentonites, they are thin (2–10 nm), with 1-nm growth steps and some evidence of spiral growth. Samples composed of especially thin crystals are I/S by XRD. There is no evidence for a smectite precursor. Individual laths may be intergrown at 120° to produce star-like aggregates or twins (Fig. 5). The twinning (a rotation of 120° with respect to the mirror plane containing the empty octahedral site) is after the “common mica twin law” (8) and likely accounts for much of the rotational disorder seen in the XRD data. The preceding has established that thin diagenetic illite crystals grow in sedimentary rocks and that they have distinct mineralogical features, such as I/S XRD effects and 1M polytype, that distinguish them from 2M1 muscovite. Much of our knowledge of disordered illite polytypes and I/S comes from the use of the programsNEWMOD (10) andWILDFIRE (14), which permit easy calculation of the complete powder XRD patterns of clay minerals. These programs form the basis for “unmixing” the mixtures we have been discussing. In the process of matching calculated to experimental data on polytypes and disorder in illite, some generalizations have emerged. Bentonites and fibrous (sandstone) illites are similar in many respects (1M with some 120° rotational disorder) but differ in that the cis-vacant form (15, 17) is more common in bentonites and the trans-vacant form (the traditional 1M structure) is more typical of fibers [discussion of nomenclature (14)].

FIG. 4. Scanning electron micrograph of pore-filling fibrous illite in a sandstone. FIG. 5. (A) AFM deflection image of sandstone illite. Laths are intergrown at 120° in a star-like aggregate or twin after the common mica twin law (a rotation of 120° with respect to the mirror plane containing the empty octahedral site) (8). Granular materials adhering to illite (especially on the right) are salts precipitated during sample preparation. The scale is in micrometers; the crystal is ≈1 µm long. This and subsequent images were made in air, contact mode, on a Universal AFM (ThermoMicroscopes, Sunnyvale, CA). (B) Close-up of the center in A. Lines show measurements of step height made on the height image (not shown). Note interlaced growth of 1-nm (10-Å) growth steps. Individual laths have a thickness of 6–8 nm. By powder XRD, this sample is 1M, with a minor 120° rotational disorder. Only the center will contribute to the disorder; the projecting laths (A) will not. The scale is in angstroms. Shales are different in that most shale illites (excluding the 2M1 component) show nearly maximum rotational disorder, including both 120° and 60° rotations (14) and are therefore the 1Md polytype (8). This means that each successive 1-nm layer is unrelated to the layer below it except that the hexagonal oxygen rings align to accommodate K. On the basis of AFM morphological observations, bentonite and sandstone illites

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ILLITE AND HYDROCARBON EXPLORATION

3443

grow primarily by spiral or step mechanisms whereas shale illites grow by nucleation (birth and spreading). Illites in shales (Fig. 6) show many small 1-nm-thick nuclei on the 00l of a larger substrate that may be detrital mica. These appear to be randomly placed epitaxial growths. Continued similar growth would create a 1Md illite. Bentonite and fiberous illites have nearly featureless 00l faces with one or more parallel growth steps. The contrasting mechanisms (growth vs. nucleation) are roughly in accord with the early discussion on the origin of polytypes (8). Transmission electron microscopy paints an apparently somewhat different view of shale illite (18), but it is not clear to me how much of that difference is related to the method of investigation (transmission electron microscopy vs. XRD). For example, the requirements for coherency are likely more stringent for XRD than for transmission electron microscopy. The predominance of 2M1 polytype in ion-milled whole-rock samples (18) is possibly due to detrital muscovite; at least, that is what shale K-Ar data (older than depositional age) suggest. Further discussion is beyond the scope of this review, but the questions raised by the transmission electron microscopy work on illite offer exciting directions for future research. Illite Age Analysis Returning to the shale size–age spectrum (Fig. 1a), it is obvious that a simple way to estimate the ages of the detrital and diagenetic end members is to quantitatively determine (by XRD) the proportions of the end members in each of the three size fractions, plot the points (normalized to 100%) as apparent K-Ar age vs. percent of detrital illite, and linearly extrapolate to 0 and 100% detrital to get the end member ages (Fig. 7). I call this Illite Age Analysis (IAA), and it is the subject of an Exxon patent (19). The extrapolated “diagenetic age” is the mean (integrated) age of the time interval over which illite grew. This could be a nearly instantaneous event in the case of illite formed in response to an igneous intrusion, or a 50-million-year (my) interval of burial in a sedimentary basin. Similarly, the “detrital age” is the mean age of the coarse micas, which may themselves be a mixture. Ideally, the detrital age corresponds to the mean time of uplift and cooling of the source terrain below the so-called blocking temperature for muscovite (250–300°C), below which Ar no longer diffuses out of the structure (20).

FIG. 6. AFM deflection image of a shale illite crystal. The surface is covered with small, 1-nm-thick growths or nuclei, possibly on the 00l of a larger substrate that may be detrital mica. These appear to be randomly placed epitaxial growths. Continued similar growth would create a 1Md illite. XRD shows 60% 1Md, with the rest 2M1. The XRD pattern for this sample is in Fig. 9b (C). The scale is in angstroms. FIG. 7. IAA plot of a shale sample. To estimate the ages of the detrital and diagenetic end members, we quantitatively determine (XRD) the proportions of the end members in each of three size fractions, plot the points (normalized to 100%) as apparent K-Ar age vs. percent of detrital illite, and linearly extrapolate to 0 and 100% detrital to get the end member ages. Lower diagram is XRD pattern (oriented aggregate, Cu radiation) showing discrete illite (detrital) and diagenetic I/S. Some distinctly questionable assumptions are made in using this method. First, can we treat the complex mixture that is shale as a twocomponent system with respect to illite? For example, what if there is detrital (recycled) 1M illite? Where we have had an independent test, such as a convenient bentonite interbedded with shale (21), or a date on large micas physically separated from the rock, the method works. Diagenetic illite is likely more easily weathered because of fine grain size; it may not survive as detritus. What if illite grew during two heating events 50 my apart? As we will see, for calibrating the thermal history of basins, only the integrated age is important. Certainly, two separate ages could not be extrapolated from IAA data alone. Could Ar leak out of the tiny illite crystals so the age would be too young? Illite formed by contact metamorphism gives the same age as the pluton, showing illite to be retentive of Ar (22). Small crystals often have fewer defects than large ones, and defects may control Ar loss (atom hopping vs. migration down tubes and cracks). Also, if a crystal is disrupted so it loses Ar, it will likely also lose K from the same region because it is in contact with a Na-rich pore fluid, in which case the K-Ar age will be unaffected. As long as the samples have not been heated above the generally accepted 250°C muscovite-blocking temperature, thermal argon diffusion is unlikely, but we really have few data on illite itself. Fortunately, drill holes in most sedimentary basins seldom get close to 200°C. How do we know that the relation in Fig. 7 is linear? It is not, really, but if the K content of both end members is similar, it is close enough. This is suggested by the observation that most of our many data sets fit a straight line rather well.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ILLITE AND HYDROCARBON EXPLORATION

3444

In practice, there are two approaches to quantify the end members using XRD. The first uses the 00l peaks and assumes the diagenetic illite is in I/S, and the detrital end member is discrete mica. These two can be distinguished on an XRD pattern (Fig. 8a). Quantification is the critical step and the source of most of the uncertainty in the IAA method. We calculate, from first principles, XRD patterns to match the experimental pattern (Fig. 8b). The basic method is that ofNEWMOD (10), but the actual calculation and matching are controlled by a genetic algorithm (23). From the range of calculations that have a good fit, we estimate an uncertainty for each point on the IAA plot and use a Monte Carlo method to project these uncertainties into the extrapolated end member ages. For samples that have data points mostly at one end or the other of the IAA plot, the uncertainty in estimating the age at the opposite end can be quite large. The second method uses polytypes (1Md and 2M1; see Fig. 9a). Because shales contain large amounts of rotationally disordered illite with a few, broad XRD peaks (Fig. 9b), anything resembling a real quantitative analysis was not easily done until WILDFIRE became available (14). Using an approach similar to the first, for each fraction, a calculated XRD pattern is optimized to the experimental data (Fig. 10). The polytype method is especially useful for samples lacking I/S, in which the peaks for illite and mica are superimposed. A similar routine was applied to a Paleozoic shale from Illinois (24).

FIG. 8. (a) Calculated XRD patterns (oriented aggregate, Cu radiation) for discrete illite and I/S made withNEWMOD. Patterns like these are added to match an experimental pattern. Blocks on the left show basic structural 2:1 layers. (b) This illustrates how an experimental XRD pattern (oriented aggregate, Cu radiation) is matched, and thus quantified, by a calculated mixture of discrete illite and I/S. Calculated pattern at the top is for 40% discrete (detrital) illite. FIG. 9. (a) XRD patterns (assuming random powder sample, Cu radiation) calculated with WILDFIRE of nondisordered 1M and 2M1 polytypes of dioctahedral mica. Note the distinguishing peaks in the central part of the patterns. At the left is the a-b projection of the unit cell showing the stacking sequences that characterize each polytype. (b) XRD patterns for the shale shown in Fig. 6. Size fractions are same as Fig. 1. K-Ar ages are C = 151, M = 110, and F = 78 my. The F pattern is typical for maximum disordered 1Md illite. The C pattern shows modulations for 2M1, and the model indicates 40% 2M1. Random powder mount, Cu radiation. The IAA technique permits only estimation of the component ages. Precision, calculated as above, averages ≈±15% of the estimated value (e.g., 20 ± 3 my) based on our experience, and can be larger where the diagenetic age is <10 my. Accuracy is unknown, but where tested (21) is almost as good as precision. Certainly, the diagenetic age from IAA is a much better choice for calibrating basin thermal history than a whole rock K-Ar age from shale or the age of an arbitrary fine fraction. The 40/39Ar dating technique has been used as an

FIG. 10. Experimental XRD pattern (Upper) and a calculated match (Lower) of a sample containing 15% 2M1 and the rest moderately disordered 1M.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ILLITE AND HYDROCARBON EXPLORATION

3445

alternative for separating diagenetic from detrital ages in mixtures (25). Although at present this is not as effective as IAA, continued progress on methodology and diffusion models may ultimately make this the method of choice. Applications Models are the key to using illite in basin thermal history calibration. The petroleum industry typically uses burial history, based on the stratigraphy (age and depth) in a well, to estimate thermal history (26). Other geologic data are used to estimate the amount of sediment eroded from unconformities, timing of uplift events, and basal heat flow. A computational model, which includes compaction, estimated rock thermal conductivity, and radiogenic heat generation, computes the temperature of each sedimentary layer through time as it is buried. The modeled results, such as present-day thermal gradient, are then compared with measured well temperatures, which define the real thermal gradient, and the model is adjusted to fit the measured data. Once the model is calibrated to data, kinetic expressions for thermal generation of oil and gas can be applied to the thermal history to give timing of HC generation. Unfortunately, present-day conditions may not tell us much about when a particular shale bed generated oil tens of millions of years ago. Present thermal gradient may not be a guide to past conditions. We need paleothermometers, rock properties that tell us about past thermal events, to properly constrain the model. The most widely used paleothermometer is based on vitrinite reflectance (%R), the increase in reflectivity of a coaly material found in rocks as a function of time and especially temperature (26). The thermal history is applied to a kinetic expression for %R, and model %R values are obtained; these are compared with measured values from rocks obtained from the well, and the model is adjusted to give a reasonable fit. But there is a problem: %R really gives only the maximum temperature; it tells us nothing about when that temperature was reached, and that is when the HCs were generated. A downhole increase in shale diagenetic illite (%I in I/S) is observed in many basins of the world (6, 20, 26), and this relation has been used as a paleothermometer in the same way as %R (27). We use experimental kinetics developed by Exxon (27); see ref. 28 for a comparison of several published kinetic expressions. The measured values of %I in I/S are compared with the modeled curve (Fig. 11), and the model is adjusted to optimize the fit. This alone does not give us much more than %R, but the age of the diagenetic illite can also be easily modeled by using the kinetic expression and can be calibrated with the measured diagenetic age from IAA. This gives us a powerful piece of chronologic information that is independent of assumptions about burial history. Note that a modeled age, like the IAA diagenetic age, will be an integrated age—the mean age of an illite-forming time interval.

FIG. 11. Plot of decimal fraction of illite in I/S from shales in a typical well. Individual points are sample measurements; the line is calculated from a burial history by using an experimental kinetic expression (23). The integration of paleothermometers is shown on a schematic thermal history in Fig. 12. Illite data constrain the burial or heating phase of a basin's thermal history, %R records maximum temperature, and apatite fission track analysis constrains timing of uplift and cooling. Discussion of the last technique is beyond the scope of this review. A diagrammatic example application is given in Fig. 13. The cross section shows petroleum source rocks separated from a structural trap by an unconformity at A. Did the source rocks mature (get heated) before or after deposition of the upper units containing the trap? If they produced oil before time A, then it is much less likely that they will be able to act as source for the trap. The question is one of amount of missing (eroded) section at A. If there was a large amount of uplift and erosion, then the source rocks could have been deep (hot) enough before A to produce oil. To solve the problem, samples are obtained from the source shales from outcrops or nearby wells (Fig. 13A, 1) and IAA (Fig. 13B, 2) is done to get the diagenetic age (Fig. 13B, 3). The thermal history plot (Fig. 13B, 4) is now anchored by a real date (IAA), which constrains the source heating and HC yields to post A time (Fig. 13B, 5). This indicates that HC supply to the trap will not be a risk factor for this prospect. IAA is especially useful in areas of complex structure, like fold and thrust belts, in which thermal history is not just a function of simple burial. Variants of the methods described have been used to successfully date normal and thrust faults (time of trap formation) and to predict growth of permeability-reducing illite in reservoir sandstones (29, 13). Growth of illite in shales, as in sandstones, appears to be a pore-filling process, but the pores are smaller and flatter. Shale permeability, like that of sandstones, is likely reduced by illite growth. This could improve the quality of a shale seal above a trap or could otherwise effect the mechanical properties of the shale. Clay-rich fault gouge typically has a flat size–age spectrum or an inclined spectrum with all ages younger than depositional age (29). It appears that upper crustal faulting (low temperature) can reset the illite K-Ar clock, but the mechanism is unclear. Heating seems unlikely, as % R indicates low temperatures. Crystal growth under conditions of deformation and unique fluid chemistry are likely involved. It is not clear that deformation alone can cause total Ar loss from illite. Fault gouge illites are an area of evolving research.

FIG. 12. Thermal history schematic showing integration of paleothermometers. Illite data constrain the burial or heating phase of a basin's thermal history, %R records maximum temperature, and apatite fission track analysis constrains timing of uplift and cooling.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

ILLITE AND HYDROCARBON EXPLORATION

3446

FIG. 13. (A) Cross section showing petroleum source rocks separated from a structural trap by an unconformity at A. Did the source rocks get heated before or after deposition of the upper units containing the trap? If they produced oil before time A, then it is much less likely that they will be able to act as source for the trap. The question is one of amount of missing (eroded) section at A. If there was a large amount of uplift and erosion, then the source rocks could have been deep (hot) enough before A to produce oil. (B) To solve the problem in A, samples are obtained from the source shales (1), and IAA (2) is done to get the diagenetic age (3). The thermal history plot (4) is now anchored by a real date (IAA), which constrains the source heating, and HC yields to post-A time (5). Sharp peaks on 5 show model generation of oil and gas, respectively. HC supply to the trap will not be a risk factor for this prospect. Conclusions Illite is a common mineral in sedimentary rocks, especially shales. Careful mineralogical analysis using new techniques developed by the clay mineral research community permits the extraction of quantitative information on the time and temperature of diagenetic illite formation. In hydrocarbon exploration, these data are used to calibrate the heating history of sedimentary basins to ascertain that oil or gas generation from source shales postdated trap formation. If generation preceded trap formation, the oil or gas would presumably have leaked off, and the well should not be drilled. Application of the mineralogical work reported here will decrease the risk of drilling a dry hole, reducing not only the expense but also any disturbance that might be caused by drilling. Further, because thermal conditions partly control the likelihood of the trap being filled with gas vs. oil, the illite work helps us find the particular type of HC we are looking for. Application to fault dating is useful not only to estimate HC trap timing but also may have potential in evaluating earthquake hazards. The price of oil and gas has remained low because of the combined effects of open competition and applied technology. Although threedimensional seismic is often featured by the media, many less dramatic advances also contribute to improving the efficiency of exploration and production. Because the earth is made of minerals, it is not surprising that mineralogy plays an essential role. I thank Exxon Production Research Co. for a productive research environment and thank several Exxon people: P. J. Houser for the AFM work, D. W. Webb and T. C. Phillips for XRD, and R. F. Ylagan for the polytype work and useful advice. Without the work and friendship of R. C. Reynolds, none of this would have been possible.

1. Grim, R. E., Bray, R. H. & Bradley, W. F. (1937) Am. Mineral. 22, 813–829. 2. Srodon, J. & Eberl, D. D. (1984) Rev. Mineral. 13, 495–546. 3. Jackson, J. A., ed. (1997) Glossary of Geology (American Geological Institute, Alexandria, VA). 4. Newman, A. C. D. & Brown, G. (1987) in Chemistry of Clays and Clay Minerals: Mineralogical Society Monograph No. 6, ed. Newman, A. C. D. (Longman, New York), pp. 1–128. 5. Brindley, G. W. & Brown, G., eds. (1980) Crystal Structures of Clay Minerals and Their X-Ray Identification: Mineralogical Society Monograph No. 6 (Mineralogical Society, London). 6. Eslinger, E. & Pevear, D. (1988) Clay Minerals for Petroleum Geologists and Engineers (Society of Economic Paleontolgists and Mineralogists, Tulsa, OK). 7. Bailey, S. W., Hurley, P. M., Fairbairn, H. W. & Pinson, W. H. (1962) Geol. Soc. Am. Bull. 73, 1167–1170. 8. Smith, J. V. & Yoder, H. S. (1956) Mineral. Mag. 31, 209–235. 9. Hower, J., Hurley, P. M., Pinson, W. H. & Fairbairn, H. W. (1963) Geochim. Cosmochim. Acta 27, 405–410. 10. Moore, D. M. & Reynolds, R. C., Jr. (1997) X-Ray Diffraction and the Identification and Analysis of Clay Minerals (Oxford Univ. Press, Oxford). 11. Hower, J. (1981) in Clays and the Resource Geologist, ed. Longstaffe, F. J. (Mineralogical Association of Canada, Toronto), pp. 60–80. 12. Nadeau, P. H. (1984) Science 225, 923–925. 13. Nagy, K. L. (1994) in Scanning Probe Microscopy of Clay Minerals, ed. Nagy, K. L. (Clay Minerals Society, Boulder, CO), pp. 204–239. 14. Reynolds, R. C., Jr (1993) in Computer Applications to X-Ray Powder Diffraction Analysis of Clay Minerals, eds. Reynolds, R. C. & Walker, J. R. (Clay Minerals Society, Boulder, CO), pp. 43–78. 15. Drits, V. A. & McCarty, D. K. (1996) Am. Mineral. 81, 852–863. 16. Reynolds, R. C., Jr. (1992) Clays Clay Miner. 40, 387–396. 17. Tsipursky, S. I. & Drits, V. A. (1984) Clay Miner. 19, 177–193. 18. Dong, H. & Peacor, D. R. (1996) Clays Clay Miner. 44, 257–275. 19. Pevear, D. R. (1994) U. S. Patent 5,288,695. 20. Clauer, N. & Chaudhuri, S. (1995) Clays in Crustal Environments (Springer, Berlin). 21. Pevear, D. R. (1992) in Water–Rock Interaction, eds. Kharaka, Y. K. & Maest, A. S. (A. A. Balkema, Rotterdam, the Netherlands), pp. 1251–1254. 22. Aronson, J. L. & Lee, M. (1986) Clays Clay Miner. 34, 483–487. 23. Pevear, D. R. & Schuette, J. F. (1993) in Computer Applications to X-Ray Powder Diffraction Analysis of Clay Minerals, eds. Reynolds, R. C. & Walker, J. R. (Clay Minerals Society, Boulder, CO), pp. 19–42. 24. Grathoff, G. H. & Moore, D. M. (1996) Clays Clay Miner. 44, 835–842. 25. Onstott, T. C., Mueller, C., Vrolijk, P. J. & Pevear, D. R. (1997) Geochim. Cosmochim. Acta 61, 3851–3861. 26. Robert, P. (1988) Organic Metamorphism and Geothermal History (D. Reidel, Dordrecht, Holland). 27. Huang, W.-L., Longo, J. M. & Pevear, D. R. (1993) Clays Clay Miner. 41, 162–177. 28. Elliot, W. C. & Matisoff, G. (1996) Clays Clay Miner. 44, 77–87. 29. Pevear, D. R., Vrolijk, P.J. & Longstaffe, F. J. (1997) in Geofiuids II '97, eds. Hendry, J. P. Carey, P. F., Parnell, J., Ruffell, A. H. & Worden, R. H. (Queen's Univ. Press, Belfast, U.K.), pp. 42–45.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3447

Manganese oxide minerals: Crystal structures and economic and environmental significance

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3447–3454, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. JEFFREY E. POST PNAS is available online at www.pnas.org.

Department of Mineral Sciences, Smithsonian Institution, Washington, DC 20560-0119 ABSTRACT Manganese oxide minerals have been used for thousands of years—by the ancients for pigments and to clarify glass, and today as ores of Mn metal, catalysts, and battery material. More than 30 Mn oxide minerals occur in a wide variety of geological settings. They are major components of Mn nodules that pave huge areas of the ocean floor and bottoms of many fresh-water lakes. Mn oxide minerals are ubiquitous in soils and sediments and participate in a variety of chemical reactions that affect groundwater and bulk soil composition. Their typical occurrence as fine-grained mixtures makes it difficult to study their atomic structures and crystal chemistries. In recent years, however, investigations using transmission electron microscopy and powder x-ray and neutron diffraction methods have provided important new insights into the structures and properties of these materials. The crystal structures for todorokite and birnessite, two of the more common Mn oxide minerals in terrestrial deposits and ocean nodules, were determined by using powder x-ray diffraction data and the Rietveld refinement method. Because of the large tunnels in todorokite and related structures there is considerable interest in the use of these materials and synthetic analogues as catalysts and cation exchange agents. Birnessite-group minerals have layer structures and readily undergo oxidation reduction and cation-exchange reactions and play a major role in controlling groundwater chemistry. Manganese (Mn) is the 10th most abundant element in the Earth's crust and second only to iron as the most common heavy metal; on average crustal rocks contain about 0.1% Mn (1). Geochemically, Mn behaves like Mg, Fe, Ni, and Co and tends to partition into minerals that form in the early stages of magmatic crystallization. Significant quantities of Mn persist, however, in melts and can be plentiful in late-stage deposits such as pegmatites (2). Mn is readily depleted from igneous and metamorphic rocks by interactions with surface water and groundwater and is highly mobile, as Mn(II), in acidic aqueous systems (2). Near the Earth's surface, Mn is easily oxidized, giving rise to more than 30 known Mn oxide/hydroxide minerals. These oxides are the major players in the story of the mineralogy and geochemistry of Mn in the upper crust and the major sources of industrial Mn. Most people's introduction to Mn oxides is the messy, black innards of a dry-cell battery. But in fact, Mn oxide/hydroxide (referred to generally as Mn oxide) minerals are found in a wide variety of geological settings and are nearly ubiquitous in soils and sediments. They occur as fine-grained aggregates, veins, marine and fresh-water nodules and concretions, crusts, dendrites, and coatings on other mineral particles and rock surfaces (e.g., desert varnish). Because Mn oxides commonly form at the interface between the lithosphere and hydrosphere, atmosphere and/or biosphere, they can provide environmentally relevant insights into certain types of interactions between these systems and potentially serve as long-term monitors of changes within a system. As ores, Mn oxides have been exploited since ancient times. In particular, pyrolusite (MnO2) was prized as a pigment and for its ability to remove the green tint imparted by iron to glass (3). By the mid-19th century Mn was an essential component in steel making, as a deoxidizer and desulfurizer and for making hard-steel alloys. Mn oxides are the predominant ore minerals in most of today's commercially important Mn deposits, commonly formed by weathering of Mn-rich carbonates or silicates, either by in situ oxidation or by dissolution followed by migration and reprecipitation (4). Approximately 80–90% of the current world production of Mn ore is consumed by the steel industry; on average, steel contains about 0.6 weight percent Mn but may be 10% or more in high-strength steels (5). Other uses include production of special Al alloys, Mn chemicals, catalysts, water-purifying agents, additives to livestock feed, plant fertilizers, colorant for bricks, and in batteries. Natural Mn oxide (primarily nsutite) is used as the cathodic material in zinc-carbon dry-cell batteries. In recent years, however, alkaline batteries, that use synthetic, electrolytic Mn oxide, have increasingly dominated the market (6). Ocean Mn Nodules The most extensive deposition of Mn oxides today occurs in the oceans as nodules, microconcretions, coatings, and crusts (7). Marine Mn nodules were first discovered in 1873 during the voyage of the HMS Challenger (8). Since then, Mn nodules have been found at almost all depths and latitudes in all of the world's oceans and seas (7); it has been estimated, for example, that they cover about 10–30% of the deep Pacific floor (9). Ocean Mn nodules typically are brown-black and subspherical-botroyoidal and consist of concentric layers of primarily Mn and iron oxide minerals. Other minerals commonly found in the nodules include: clay minerals, quartz, apatite, biotite, and feldspars (10). Most Mn nodules have formed around central nuclei that may be carbonate mineral fragments, pumice shards, animal remains, coral fragments, etc. (11). The nodules range from 0.5 to 25 cm in diameter, with an ocean-wide average of about 4 cm (11). Marine Mn oxide crusts and nodules concentrate at the sediment-water interface (12) but locally are distributed to a depth of 3 or 4 m (13). The nodules are most abundant in oxygenated environments with low sedimentation rates and reach their greatest concentration in deep-water at or below the calcium carbonate compensation depth (11). Accumulation rates range from 0.3 to 1,000 mm/yr in near-shore environments to about 1 cm/million yr in the deep ocean (14, 15). The source of the Mn is thought to be continental runoff and hydrothermal and volcanic activity at midocean spreading centers (16, 17).

Abbreviation: TEM, transmission electron microscopy.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3448

Research on the complex mineralogy of the Fe and Mn oxides in ocean Mn nodules has been hampered by the fact that the minerals typically occur as thin layers of fine-grained, poorly crystalline mixtures. Previous studies of the mineralogy of ocean nodules concluded that the dominant Mn oxide phases are birnessite (7 Å manganate), todorokite (10 Å manganate), and δ-MnO2 or vernadite (18). Both birnessite and todorokite commonly are found in the same nodule, but birnessite tends to predominate in nodules from topographic highs such as seamounts and ridges, and todorokite is more common in slightly more reducing near-shore and abyssal environments (19, 20). How ocean Mn nodules grow is a subject of intensive research and some debate. Nodules apparently grow principally by direct precipitation of Mn from seawater, but the types of reactions that occur in the water and at the precipitation surface are poorly known (2, 17). It also has been suggested that some Mn and Fe is supplied by upward diffusion through underlying reducing sediments (2). One scenario suggests that Mn oxide phases in ocean nodule form by catalytic oxidation and adsorption of Mn(II) on suitable substrates, such as mineral and rock fragments and fine-grained MnO2 and Fe(OH)3. Once initiated nodule formation is self-perpetuating because Fe and Mn are autocatalytically precipitated on the surface (2). Indeed Mn oxides, and Mn nodules themselves, have been recommended as oxidation catalysts for automobile exhaust systems (21) and for the reduction of nitric acid pollutants (22). It also has been proposed that in some environments bacteria might be the dominant catalysts for Mn oxide precipitation (7). For example, Mn-oxidizing and Mn-reducing bacteria isolated from deep-sea nodules have been shown to increase experimental deposition of Mn onto pulverized nodules (23). More recent studies (summarized in ref. 24) indicate that microorganisms can accelerate the rate of Mn(II) oxidation by up to five orders of magnitude over abiotic oxidation, and thus are likely responsible for much natural Mn(II) oxidation. Ocean nodules are of potential commercial interest because in addition to Mn they also contain significant amounts (several tenths to more than one weight percent) of Cu, Ni, Co, and other strategic metals (e.g., ref. 25). Laboratory experiments have shown that the sorption capacity of freshly precipitated Mn oxides is extremely high for a variety of metal cations (26, 27 and 28). Thus in seawater, adsorption by Mn oxide deposits may be the most important mechanism for controlling the concentration of heavy metals (27). The relatively slow accretion rates for deep-sea nodules provide ample opportunity for adsorption of heavy metals and is consistent with observations that more rapidly growing nodules tend to have lower trace metal concentrations (17). It is unclear how the heavy metals are bound in the nodule Mn oxide minerals, or even in which phases they are concentrated. Experiments have demonstrated that adsorption of heavy metals by hydrous Mn oxides is accompanied by release of protons (H+), suggesting that the cations are bound into the Mn oxides' atomic structures (27). Additional insights into the nature of the heavy metals likely await a more detailed understanding of the atomic structures and crystal chemistries of Mn oxide minerals in ocean nodules. Mn Oxide Minerals and the Environment The unusually high adsorption capacities and scavenging capabilities of Mn oxide/hydroxide minerals provides one of the primary controls of heavy metals and other trace elements in soils and aquatic sediments (28, 29). Understanding such controls is important for maintaining and improving fertility of soil, mitigating health affects in humans and animals, and for treatment of water for consumption and industrial use. Because Mn oxide minerals commonly occur as coatings and fine-grained aggregates with large surface areas, they exert chemical influences far out of proportion to their concentrations (28). The presence of only tiny amounts (e.g., a fraction of a weight percent of soil or sediment) of Mn oxide minerals might be adequate to control distribution of heavy metals between earth materials and associated aqueous systems (28). Additionally, Mn oxides can act as important adsorbents of phosphate in natural waters and surface sediments (30). Two useful applications of the scavenging ability of Mn oxide minerals are as geochemical exploration tools (25, 31, 32) and purification agents for drinking water (33). Recent studies also indicate that Mn oxide minerals in soils and stream sediments and as coatings on stream pebbles and boulders might serve as natural traps for heavy metals in contaminated waters from mines and other industrial operations (32, 34, 35). Similarly, Mn oxide absorbers effectively recover Ra, Pb, and Po from seawater (36), and it has been shown that the geochemical distribution of several naturally occurring radionuclides (234Th, 228Th, 228Ra, and 226Ra) is controlled by Fe and Mn oxides (37, 38). Hydrous oxides of Mn occur in most soils as discrete particles and as coatings on other mineral grains. Mn is highly mobile in acid, organic soils of the temperate and subarctic zones, but in the more alkaline tropical soils Mn might concentrate with residual laterites (2). It has been noted that frequently observed influences of pH, organic matter, lime, and phosphate on heavy metal availability in soils are understood principally in terms of their influence on the chemistries of hydrous oxides of Mn and Fe (28). The major Mn minerals reported in soils are lithiophorite, hollandite, and birnessite (39, 40); it is more typically the case, however, that because the Mn oxides are fine-grained and poorly crystalline (commonly referred to as amorphous) that no attempt is made to assign mineral designations. Mn oxides in soils and sediments readily participate in a wide variety of oxidation-reduction and cation-exchange reactions. They exhibit large surface areas and can be very chemically active. Birnessite directly oxidizes Se(IV) to Se(VI) via a surface mechanism (41), Cr(III) to Cr (VI) (42), and As(III) to As(V) (43). Certain Mn oxide minerals easily oxidize arsenate (III), the more toxic form of inorganic As, to arsenate (V), which can more effectively be removed from drinking water by existing water treatment procedures (44). Mn oxide minerals such as birnessite and todorokite readily undergo cation-exchange reactions (45), and studies have shown that the cation exchange capacity of Mn dioxide at pH 8.3 (about that of seawater) exceeds that of montmorillanite (46). Clearly, these kinds of reactions profoundly affect the chemistries of soils and associated aqueous solutions. Mn Oxide Minerals What accounts for the complexity and impressive variety of Mn oxide minerals? Mn occurs in natural systems in three different oxidation states: +2, +3, and +4, giving rise to a range of multivalent phases. Mn oxides also display a remarkable diversity of atomic architectures, many of which easily accommodate a wide assortment of other metal cations. Finally, Mn is abundant in most geological systems and forms minerals under a wide range of chemical and temperature conditions, and through biological interactions. Most Mn oxide minerals are brown-black and typically occur as intimately intermixed, fine-grained, poorly crystalline masses or coatings. Not surprisingly, identifying the particular mineral(s) in a Mn oxide specimen can pose quite a challenge. Hence many scientists report simply “Mn oxide,” rather than a particular mineral phase. Geologists have attempted to avoid the problem by simply referring to all soft (i.e., it blackens your fingers), brown-black, fine-grained specimens that were assumed to be Mn oxides as “wad.” Similarly, hard (does not

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3449

blacken your fingers), gray-black, botroyoidal, massive specimens were called “psilomelane.” Recent studies have shown that most so-called psilomelane specimens are predominantly the mineral romanechite (Ba.66Mn5O10·1.34H2O). There is no comparable correlation between specimens labeled wad and any particular Mn oxide mineral. Even today identification of the minerals in many Mn oxide samples is not straightforward. In general, powder x-ray diffraction is diagnostic for well-crystallized, monophasic samples. Unfortunately, the crystal structures, and consequently, the powder diffraction patterns are similar for many of the Mn oxide minerals. In many cases, it is necessary to supplement powder x-ray diffraction studies with other techniques, such as transmission electron microscopy (TEM), IR spectroscopy, and electron microprobe analysis. Despite the fact that Mn oxides have been extensively studied for the past several decades, the details of many of their atomic structures are poorly understood, and there are several phases for which even the basic crystal structures are not known. This paucity of crystallographic data has greatly hindered research on the fundamental geochemical behaviors of common Mn oxide minerals. In most cases the major limiting factor is the lack of crystals suitable for single-crystal x-ray or neutron diffraction experiments. In the past few years, however, other techniques such as TEM (high-resolution imaging and electron diffraction) and Rietveld refinements using powder x-ray and neutron diffraction data have provided important new insights into the atomic structures of Mn oxide minerals. The Rietveld method (see review in ref. 47) has made it possible to partially solve and refine structures from data collected from even relatively poorly crystalline samples (e.g., refs. 48 and 49). Other methods that have contributed to the understanding of Mn-oxide atomic structures include: IR spectroscopy, extended x-ray absorption finestructure spectroscopy, and thermogravimetric analysis. The recent application of charge-coupled device imaging plates to single-crystal x-ray diffraction experiments, particularly at synchrotron sources (50), should open possibilities for detailed studies of extremely tiny crystals of certain Mn oxide phases that were too small for use in conventional experiments. The basic building block for most of the Mn oxide atomic structures is the MnO6 octahedron. These octahedra can be assembled by sharing edges and/or corners into a large variety of different structural arrangements, most of which fall into one of two major groups: (i) chain, or tunnel, structures and (ii) layer structures. The tunnel Mn oxides are constructed of single, double, or triple chains of edge-sharing MnO6 octahedra, and the chains share corners with each other to produce frameworks that have tunnels with square or rectangular cross sections. The larger tunnels are partially filled with water molecules and/or cations. The layer Mn oxides, sometimes referred to as phyllomanganates, consist of stacks of sheets, or layers, of edge-sharing MnO6 octahedra. The interlayer regions can host water molecules and a wide range of cations. One of the complexities of Mn oxide crystal chemistry is the multiple valence states exhibited by Mn, commonly even in a single mineral. It is reasonably straightforward to measure the average Mn oxidation state for a mineral, but it is considerably more difficult to determine the proportions of Mn(IV), Mn(III), and/or Mn(II). In some cases, chemical analyses with bulk oxidation state measurements unambiguously indicate the correct Mn valence state, e.g., pyrolusite [Mn(IV)O2] or manganosite [Mn(II)O]. For other minerals, detailed crystal structure studies were able to reveal the valence state information. Manganite (MnOOH), for example, can be charge-balanced assuming all Mn(III) or a half-and-half mixture of Mn(II) and Mn(IV). Crystal structure refinements revealed that the Mn in manganite is in a highly distorted octahedral site, characteristic of the Jahn-Teller effects displayed by Mn(III). The Mn-O bond distances determined by crystal structure refinements also can provide insights into the Mn valence state(s), e.g., structure studies of hollandite minerals (51), romanechite (52), and todorokite (48) indicate that the reduced form of Mn in these minerals is Mn(III). Studies of Mn oxidation states also have been performed by using x-ray spectroscopy (e.g., ref. 53) and x-ray absorption near-edge structure spectroscopy (54). Summarized below are descriptions of the atomic structures and other information for some of the most important Mn oxide minerals (for a more in-depth presentation see ref. 55). The minerals and their chemical formulae are listed in Table 1. Mn Oxide Minerals with Tunnel Structures Pyrolusite MnO2. There are three known mineral polymorphs of MnO2; pyrolusite is the most stable and abundant, the others are ramsdellite and nsutite. In pyrolusite (β-MnO2), single chains of edge-sharing Mn(IV)O6 octahedra share corners with neighboring chains to form a framework structure containing tunnels with square cross sections that are one octahedron by one octahedron (1X1) on a side (Fig. 1). The structure is analogous to that of rutile (TiO2). The chain structure manifests itself in pyrolusite's typically acicular crystal morphology. The tunnels in pyrolusite are too small to accommodate other chemical species, and chemical analyses indicate that the composition deviates at most only slightly from pure MnO2. Pyrolusite commonly occurs as low-temperature hydrothermal deposits or as replacements after other Mn oxide minerals, particularly ramsdellite and manganite. It has long been assumed that Mn oxide dendrites, and other coatings, commonly found on rock surfaces are pyrolusite, but IR spectroscopy studies (70) have revealed that most such deposits are birnessite and/or romanechite, and in no case was pyrolusite identified. Ramsdellite MnO2. In the ramsdellite structure the Mn(IV)O6 octahedra are linked into double chains, each of which consists of two adjacent single chains that share octahedral edges. The double chains, in turn, link corners with each other to form a framework having tunnels with rectangular-shaped cross sections that are 1X2 octahedra on a side (57) (Fig. 1). The tunnels are generally empty but chemical analyses commonly reveal minor amounts of water, Na, and Ca that presumably are located in the channels. Ramsdellite is a relatively rare mineral, usually occurring in low-temperature hydrothermal deposits and commonly associated with, and Table 1. Important Mn oxide minerals Mineral Chemical formula Pyrolusite MnO2 Ramsdellite MnO2 Nsutite Mn(O,OH)2 Hollandite Bax(Mn4+,Mn3+)8O16 Cryptomelane Kx(Mn4+,Mn3+)8O16 Manjiroite Nax(Mn4+,Mn3+)8O16 Coronadite Pbx(Mn4+,Mn3+)8O16 Romanechite Ba.66(Mn4+,Mn3+)5O10·1.34H2O Todorokite (Ca,Na,K)x(Mn4+,Mn3+)6O12·3.5H2O Mn3+)O6(OH)6 Lithiophorite LiAl2( Chalcophanite ZnMn3O7·3H2O Birnessite (Na,Ca)Mn7O14·2.8H2O Vernadite MnO2·nH2O Manganite MnOOH Groutite MnOOH Feitknechtite MnOOH Hausmannite Mn2+ O4 Bixbyite Mn2O3 Pyrochroite Mn(OH)2 MnO Manganosite

Reference 56 57 58 51 51 59 60 52 48 61 62 49 63 64 65 66 67 68 66 69

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3450

FIG. 1. Polyhedral representations of the crystal structures of (A) pyrolusite, (B) ramsdellite, (C) hollandite, (D) romanechite, and (E) todorokite, looking approximately parallel to the Mn octahedral chains. probably altering to, pyrolusite. Ramsdellite is isostructural with goethite (FeOOH) and gibbsite (AlOOH). Nsutite MnO2. Nsutite (γ-MnO2) is an important cathodic material for use in dry-cell batteries. It is named after the large deposits of the mineral near Nsuta, Ghana. Although classified as a mineral, nsutite is actually an intergrowth between pyrolusite and ramsdellite. TEM images reveal a disordered structure consisting of regions of a ramsdellite-like phase and areas that appear to be ordered intergrowths of pyrolusite and ramsdellite (71). The generally accepted structure model for nsutite is an alternating intergrowth of ramsdellite and pyrolusite (Fig. 1), and therefore, most so-called nsutite samples are, in fact, mixtures of ramsdellite and nsutite. TEM studies also revealed some tunnels larger than 1X1 (pyrolusite) and 1X2 (ramsdellite), e.g., 1X3 and 3X3, as well as numerous defects and grain boundaries (71). All of these complexities undoubtedly affect the chemical and electrical properties of the material. Chemical analyses of nsutite typically show minor amounts of Na, Ca, Mg, K, Zn, Ni, Fe, Al, and Si, and about 2–4 weight percent water (58). These species probably are accommodated in the larger tunnels or along grain boundaries, and charge balance is maintained by substituting Mn(III) for some of the Mn(IV). Nsutite has been found in ore deposits worldwide (58), and one occurrence has been reported in ocean Mn nodules (72). It is a secondary, replacement mineral that commonly forms from oxidation of Mn carbonate minerals (58).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3451

Hollandite Group R.8–1.5[Mn (IV),Mn(III)]8O16R = Ba, Pb, K or Na. As in ramsdellite, the hollandite structure is constructed of double chains of edge-sharing MnO6 octahedra, but they are linked in such a way as to form tunnels with square cross sections, measuring two octahedra on a side (Fig. 1). The tunnels are partially filled with large uni- or divalent cations and, in some cases, water molecules. The charges on the tunnel cations are balanced by substitution of lower valence cations [e.g., Mn(III), Fe(III), Al(III), etc.] for some of the Mn(IV). The different minerals in the hollandite group are defined on the basis of the predominant tunnel cation: hollandite (Ba), cryptomelane (K), coronadite (Pb), and manjiroite (Na). Natural specimens having end-member compositions are unusual, and chemical analyses show a wide range of tunnel cation compositions. Hollandite minerals commonly occur intermixed and, in some cases, grade from one to another along a single crystal. They can be major phases in the oxidized zones of Mn deposits and important ores. Consistent with their chain structure, they typically are found as fibrous crystals, usually in compact botroyoidal masses. Less commonly, hollandite minerals form as prismatic crystals in hydrothermal vein deposits. In recent years there has been considerable interest in the hollandite minerals and in the hollandite structure-type in general, both for potential applications as solid ionic conductors (73) and for immobilizing certain radioactive cations as part of a waste storage system (74). Also it has been shown that at high pressures feldspar minerals transform to a hollandite-like structure (75), making this, perhaps, an important structure type in the lower crust and upper mantle. Romanechite Ba.66Mn(IV)3.68Mn(III)1.32O10·1.34H2O. The romanechite structure is constructed of double and triple chains of edgesharing MnO6 octahedra that link to form large tunnels with rectangular cross sections, measuring two by three octahedra (Fig. 1). The tunnels are filled with Ba cations and water molecules in a 1:2 ratio, and the charges on the tunnel cations are balanced by substitution of Mn(III) for some of the Mn(IV). Single-crystal x-ray diffraction studies indicate that the trivalent Mn concentrates on the octahedral sites that are at the edges of the triple chains. Romanechite typically occurs as botroyoidal masses in oxidized zones of Mn-rich deposits. Single crystals, at least those large enough for x-ray diffraction studies, are known only from Schneeberg, Germany. Cross sections of botroyoidal samples typically show very fine concentric layering (layers are tens to hundreds of microns thick). Electron microprobe analyses reveal minor fluctuations in composition among the different layers, mostly in concentration of Ba relative to Na, K, Ca, and Sr. In general, chemical analyses of romanechite deviate only slightly from the ideal formula. TEM studies have shown that romanechite and hollandite commonly intergrow on a very fine scale (76). The two structures interconnect via the common double octahedral chain. Upon heating above about 550°C, romanechite transforms to hollandite (77). Todorokite (Ca, Na, K).3-.5[Mn(IV), Mn(III), Mg]6O12·3– 4.5H2O. Todorokite is one of the major Mn minerals identified in ocean Mn nodules (10 Å manganate), and the likely host phase for strategic metals such as Ni, Co, etc. It is also a major mineral in the oxidized zones of many terrestrial Mn deposits. For many years the crystal structure of todorokite was a subject of considerable conjecture and controversy. Todorokite occurs with apparent platy or fibrous morphologies, supporting arguments for a tunnel- or layer-type structure (18). High-resolution TEM images (77) confirmed that todorokite has a tunnel structure constructed of triple chains of MnO6 octahedra. The triple chains share corners with each other to form large tunnels with square cross sections that measure three octahedra on a side (Fig. 1). TEM images also revealed intergrowths with todorokite having tunnels measuring 3X4, 3X5, and up to 3X9 octahedra in cross section (78, 79). The crystal structure of todorokite recently was refined by using the Rietveld method and powder x-ray diffraction data, revealing for the first time the major water and cation positions in the tunnels (48). Lower valence cations such as Mn(III), Ni(II), and Mg(II), which substitute for Mn(IV) to offset charges on the tunnel cations, appear to be concentrated into the sites at the edges of the triple chains, as in romanechite. Chemical analyses of todorokite show considerable variation in tunnel cation composition (80), and samples from ocean nodules have up to several weight percent Ni, Co, and/or Cu (81). Because of todorokite's large zeolite-like tunnels, there has been considerable interest in recent years in producing synthetic analogues for possible use as catalysts or molecular sieves (82). Todorokite typically occurs in Mn deposits as an alteration product of primary ores such a braunite. It also seems to be a important phase in many Mn coatings, dendrites, and varnishes (70). In the case of ocean nodules, the mechanism of todorokite formation is not well understood, but some experiments suggest that biological processes might play an important role (83). It has been speculated that nodular todorokite alters from a precursor buserite-like phase (48). Recently, studies have shown that todorokite can be synthesized starting with a Mgrich birnessite-like phase (45). MnOOH Minerals. There are three natural polymorphs of MnOOH; manganite is the most stable and abundant, the other two are feitknechtite and groutite. The manganite (γ-MnOOH) crystal structure is similar to that of pyrolusite but all of the Mn is trivalent and one-half of the O atoms are replaced by hydroxyl anions. The Mn(III) octahedra are quite distorted because of Jahn-Teller effects. Manganite typically occurs in hydrothermal vein deposits as acicular or prismatic crystals, or as an alteration product of other Mn-bearing minerals. In air manganite alters at 300°C to pyrolusite (84), and many crystals that appear to be manganite, in fact, are pseudomorphs, having been replaced by pyrolusite. Groutite (α-MnOOH) is isostructural with ramsdellite, but, as in manganite, with all Mn(III) and one-half of the O anions replaced by hydroxyl anions. Groutite is not a common mineral, but sometimes is intimately mixed with pyrolusite, to which it is probably altering. In 1945 a hydrous Mn oxide was synthesized that yielded an x-ray diffraction pattern identical to that of hausmannite with the exception of a strong extra reflection (d = 4.62 Å) and a general weakening of the remaining reflections, and it was given the name hydrohausmannite (85). Later a mineral was described from Franklin, NJ that gave an identical x-ray diffraction pattern to that of hydrohausmannite (86). Eventually it was determined that the original hydrohausmannite actually was a mixture of two phases: hausmannite and β-MnOOH (87). Electron micrographs showed that the β-MnOOH crystallized as hexagonal plates. It was assumed that all so-called hydrohausmannite mineral specimens were also mixtures, and the name feitknechtite was proposed for naturally occurring β-MnOOH (66). Feitknechtite is known only in very fine-grained mixtures, and consequently, its crystal structure has not yet been determined. Mn Oxide Minerals with Layer Structures Lithiophorite LiAl2[Mn(IV)2Mn(III)]O6(OH)6. The lithiophorite structure consists of a stack of sheets of MnO6 octahedra alternating with sheets of Al(OH)6 octahedra in which one-third of the octahedral sites is vacant (Fig. 2). In the ideal formula, Li cations fill the vacant sites in the Al layer, and charge balance is maintained by substitution of an equal number of Mn(III) for Mn(IV) cations (61). The layers are cross-linked by H bonds between hydroxyl H on the Al/Li layer

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3452

FIG. 2. Polyhedral representation of (A) lithiophorite showing alternately stacked layers of MnO6 (blue) and (Al, Li)(OH)6 (red) octahedra, (B) chalcophanite with Zn cations (green octahedra) occupying positions above and below vacancies in the Mn octahedral layers, and (C) Narich birnessite-like phase showing disordered H2O/Na sites (yellow) sandwiched between the Mn octahedral sheets. and O atoms in the Mn sheet. Lithiophorite commonly is found in weathered zones of Mn deposits and in certain acid soils, but also has been reported from low-temperature hydrothermal veins (88). It typically occurs in finely crystalline masses, but in Post-masburg, South Africa is found as large (1–2 cm) hexagonal plates. Chemical analyses show that the Li content of lithiophorite ranges from 0.2 to 3 weight percent and that transition metals such as Ni, Cu, and Co commonly substitute into the structure (89, 90). Extended x-ray absorption fine-structure spectroscopy suggest that Ni and Cu concentrate in the Al-OH sheets and Co in the Mn layers (91). Materials known as asbolanes are important components of certain Mn ore deposits and are thought to have a crystal structures similar to that of lithiophorite but with Al replaced by transition metal cations (92). Chalcophanite ZnMn3O7·3H2O. Chalcophanite is a common weathering product in many Mn-bearing base metal deposits. Its structure consists of sheets of edge-sharing Mn(IV)O6 octahedra that alternate with layers of Zn cations and water molecules (Fig. 2). One of seven octahedral sites in the Mn layer is vacant, and the Zn cations are above and below the vacancies (62). The water molecules form a hexagonal close-packed layer with one of seven molecules absent. Minerals recently have been described with the same structure as chalcophanite but with Mg (93) and Ni (94) instead of the Zn cations. The crystal structure of chalcophanite has been of interest because it is similar to, and, therefore, can serve as a model for, that of the more abundant and environmentally important birnessite, which has not been found in crystals suitable for detailed structural studies. Birnessite Group [Na,Ca,Mn(II)]Mn7O14·2.8H2O. Birnessite was first described as a natural phase from Birness, Scotland (95), and since then it has been recognized that birnessite and birnessite-like minerals occur in a wide variety of geological settings. As was mentioned above, it is a major phase in many soils and an important component in desert varnishes and other coatings and in ocean Mn nodules. It is also commonly found as an alteration product in Mn-rich ore deposits. It readily participates in oxidation-reduction and cation-exchange reactions and therefore might play a significant role in soil and groundwater chemistry. All known natural birnessite samples are fine-grained and relatively poorly crystalline. Consequently, it has not been possible to perform detailed studies of the crystal structures of these materials. The crystal structures of synthetic Na-, K-, and Mg-birnessite-like phases, however, recently were determined by using TEM and powder x-ray diffraction (49). The study confirmed that the basic structural unit is a sheet of MnO6 octahedra and revealed that the interlayer cations and water molecules occupied different positions in each of the three phases (Fig. 2). Powder x-ray diffraction patterns of the minerals ranceite and takanelite suggest that they are isostructural with birnessite, but with the dominant interlayer cations being Ca and Mn(II), respectively. The reported cation-exchange capacity of a synthetic Nabirnessite-like phase is 240 meq/100 g, with a preference for Ni and Ba over Ca and Mg (45). “Buserite.” When Na-birnessite is prepared synthetically, the initial precipitate yields a powder x-ray diffraction pattern similar to that of birnessite but with a 10 Å interlayer spacing, which upon drying collapses to the typical 7 Å birnessite spacing. The collapse presumably involves the loss of a water layer and is irreversible. The natural phase that is the presumed analogue of the synthetic 10 Å material has been called buserite (not an approved mineral name) (96, 97 and 98) and might be a common component of ocean Mn nodules before they dry out (99, 100). Cations such as Ni(II), Mg(II), Ca(II), and Co(II) tend to stabilize the buserite structure against collapse (101, 102). Vernadite MnO2·nH2O. Vernadite is a fine-grained poorly crystalline natural Mn oxide phase characterized by a powder x-ray diffraction pattern with broad XRD lines at 2.46, 1.42, and rarely at 2.2 A. Vernadite appears to be analogous to the synthetic phase δ-MnO2. Chemical analyses of vernadite samples commonly show minor amounts of K, Mg, Ca, Ba, and Fe (63) and 15–25 weight percent water. The crystal structure of vernadite is not known, but it has been proposed that vernadite is a variety of birnessite that is disordered in the layer-stacking direction (103, 104), thereby accounting for the absence of a basal reflection in the x-ray diffraction pattern. The lack of a basal reflection also could be the result of individual vernadite crystallites that are extremely thin plates, perhaps less than 100 A thick, such that there is no Bragg diffraction arising from the stacking direction. (66, 103). Vernadite is found in the oxidized zone of Mn ore deposits and might be a major phase in ocean Mn nodules and other Mn oxide crusts and coating. Other Mn Oxide Minerals Hausmannite [Mn(II)Mn(III)2O4] has a spinel-like structure with Mn(II) in the tetrahedral and Mn(III) in the octahedral

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3453

sites. It and Bixbyite [(Mn,Fe)2O3] typically are found in hydrothermal or metamorphic deposits. The amount of Fe that can be accommodated into the bixbyite structure is a function of temperature (105), and therefore the mineral is an important geothermometer in some ore deposits. The crystal structure of pyrochroite [Mn(OH)2] consists of stacked sheets of Mn(II)(OH)6 octahedra, and manganosite (MnO) is isostructural with halite. Both minerals are relatively rare, typically occurring in low-temperature hydrothermal veins in Mn-rich deposits. Summary There are more than 30 Mn oxide/hydroxide minerals, and many of them occur abundantly in a wide variety of geological settings. In addition to being important as ores of Mn metal, they also play an active role in the environmental geochemistry at the Earth's surface. Mn oxides are ubiquitous in soils and sediments, and because they are highly chemically active and strong scavengers of heavy metals, they exert considerable influences on the compositions and chemical behaviors of the sediments and soils and associated aqueous systems. The ability to successfully model and predict the chemical and thermodynamic properties of Mn oxide minerals and to prepare synthetic analogues depends to a large degree on a detailed understanding of their crystal structures. Unfortunately, many Mn oxide minerals occur only as fine-grained, poorly crystalline aggregates and coatings, making crystal structure studies extremely challenging. In recent years, however, an arsenal of new techniques, such as TEM, Rietveld refinements using powder diffraction data, extended x-ray absorption fine-structure spectroscopy, and single-crystal studies using charge-coupled device detectors and synchrotron sources, have slowly started and are continuing to unravel many of the inner secrets of Mn oxide minerals. Given the already considerable interest in Mn oxide minerals by geologists, soil scientists, microbiologists, chemical and environmental engineers, ceramicists, etc., the future for this group of characteristically dark minerals looks very bright.

1. Turekian, K. K. & Wedepohl, K. L. (1961) Geol. Soc. Am. Bull. 72, 175–192. 2. Crerar, D. A., Cormick, R. K. & Barnes, H. L. (1980) in Geology and Geochemistry of Manganese, eds. Varentsov, I. M. & Grasselly, Gy. (E. Schweizerbart'sche Verlagsbuchhandlung, Stuttgart), Vol. 1, pp. 293–334. 3. Varentsov, I. M. (1996) Manganese Ores of Supergene Zone: Geochemistry of Formation (Kluwer, Boston). 4. Roy, S. (1997) in Manganese Mineralization: Geochemistry and Mineralogy of Terrestrial and Marine Deposits, eds. Nicholson, K., Hein, J. R., Bühn, B. & Dasgupta, S. (Geological Society, London), Spec. Pub. 119, pp. 5–27. 5. Lankford, W. T., Jr., Samways, N. L., Craven, R. F. & McGannon, H. E., eds. (1984) The Making, Shaping, and Treating of Steel (Assoc. of Iron and Steel Engineers, Pittsburgh, PA). 6. U.S. Geological Survey (1998) Mineral Industry Surveys: Manganese (U.S. Geological Survey, Reston, VA). 7. Crerar, D. A. & Barnes, H. L. (1974) Geochim. Cosmochim. Acta 38, 279–300. 8. Murray, J. & Renard, A. F. (1891) Report on the Scientific Results of the Exploration Voyage of H.M.S. Challenger (Neill and Co., Longon). 9. Menard, H. W. & Shipek, C. J. (1958) Nature (London) 182, 1156–1158. 10. Riley, J. P. & Sinhaseni, P. (1958) J. Mar. Res. 17, 466–482. 11. Mero, J. L. (1977) in Marine Manganese Nodules, ed. Glasby, G. P. (Elsevier, Amsterdam), pp. 327–355. 12. Horn, D. R., Horn, B. M. & Delach, M. N. (1972) in Ferromanganese Deposits on the Ocean Floor, ed. Horn, D. R. (Lamont-Doherty Observatory, Columbia Univ., New York), pp. 9–17. 13. Cronan, D. S. & Tooms, J. S. (1967) Deep-Sea Res. 14, 117–119. 14. Ku, T. L. & Glasby, G. P. (1972) Geochim. Cosmochim. Acta 36, 699–703. 15. Crecelius, E. A., Carpenter, R. & Merrill, R. T. (1973) Earth Planet. Sci. Lett. 17, 391–396. 16. Rona, P. A., Harrison, R. N., Bassinger, B. G., Scott, R. B. & Nalwalk, A. J. (1976) Geol. Soc. Am. Bull. 87, 661–674. 17. Hein, J. R., Koschinsky, A., Halbach, P., Manheim, F. T., Bau, M., Kang, J. & Lubick, N. (1997) in Manganese Mineralization: Geochemistry and Mineralogy of Terrestrial and Marine Deposits, eds. Nicholson, K., Hein, J. R., Bühn, B. & Dasgupta, S. (Geological Society, London), Spec. Pub. 119, pp. 123–138. 18. Burns, R. G. & Burns, V. M. (1977) in Marine Manganese Nodules, ed. Glasby, G. P. (Elsevier, Amsterdam), pp. 185–248. 19. Barnes, S. S. (1967) Science 157, 63–65. 20. Glasby, G. P. (1972) Mar. Geol. 13, 57–72. 21. Weisz, P. B. (1968) J. Catal. 10, 407–408. 22. Wu, S. & Chu, C. (1972) Atmos. Environ. 6, 309–317. 23. Ehrlich, H. L. (1972) in Ferromanganese Deposits on the Ocean Floor, ed. Horn, D. R. (Lamont-Doherty Observatory, Columbia Univ., New York), pp. 63–70. 24. Tebo, B. M., Ghiorse, W. C., van Waasbergen, L. G., Siering, P. L. & Caspi, R. (1997) in Geomicrobiology: Interactions Between Microbes and Minerals, eds. Banfield, J. F. & Nealson, K. H. (Mineral. Soc. Am., Washington, DC), Rev. Min. 35, pp. 225–266. 25. Chao T. T. & Theobald, P. K. (1976) Econ. Geol. 71, 1560–1569. 26. Nicholson, K. & Eley, M. (1997) in Manganese Mineralization: Geochemistry and Mineralogy of Terrestrial and Marine Deposits, eds. Nicholson, K., Hein, J. R., Bühn, B. & Dasgupta, S. (Geological Society, London), Spec. Pub. 119, pp. 309–326. 27. Murray, J. W. (1975) Geochim. Cosmochim. Acta 39, 505–519. 28. Jenne, E. A. (1968) in Trace Inorganics in Water (Am. Chem. Soc., Washington, DC), ACS Advances in Chemistry Series, Vol. 73, pp. 337–387. 29. Young, L. B. & Harvey, H. (1992) Geochim. Cosmochim. Acta 56, 1175–1186. 30. Yao, W. & Millero, F. J. (1996) Environ. Sci. Technol. 30, 536–541. 31. Carpenter, R. H., Robinson, G. D. & Hayes, W. B. (1978) J. Geochem. Explor. 10, 75–89. 32. Whitney, P. R. (1975) J. Geochem. Explor. 4, 251–263. 33. Prasad, V. S. & Chaudhuri, M. (1995) Aqua 44, 80–82. 34. Lind, C. J. & Hem, J. H. (1993) Appl. Geochem. 8, 67–80. 35. Prusty, B. G., Sahu, K. C. & Godgul, G. (1994) Chem. Geol. 112, 275–291. 36. Towler, P. H., Smith, J. D. & Dixon, D. R. (1996) Anal. Chim. Acta 328, 53–59. 37. Todd, J. F., Elsinger, R. J. & Moore, W. S. (1988) Mar. Chem. 23, 393–415. 38. Wei, C. L. & Murray, J. W. (1991) Deep-Sea Res. 38, 855–873. 39. Taylor, R. M., McKenzie, R. M. & Norrish, K. (1964) Austr. J. Soil Res. 2, 235–248. 40. Golden, D. C., Dixon, J. B. & Kanehiro, Y. (1993) Austr. J. Soil Res. 31, 51–66. 41. Scott, M. J. & Morgan, J. J. (1996) Environ. Sci. Technol 18, 807–815. 42. Manceau, A. & Charlet, L. (1992) J. Colloid Interface Sci. 148, 425–442. 43. Huang, P. M. (1991) in Rates of Soil Chemistry Processes, eds. Sparks, D. L. & Suarez, D. L. (Soil Sci. Soc. Am., Madison, WI), pp. 191–230. 44. Driehaus, W., Seith, R. & Jekel, M. (1995) Water Res. 29, 297–305. 45. Golden, D. C., Chen, C. C. & Dixon, J. B. (1986) Science 231, 717–719. 46. Morgan, J. J. & Stumm, W. (1965) Adv. Water Pollut. Res. 1, 103–118. 47. Post, J. E. & Bish, D. L. (1989) in Modern Powder Diffraction, eds. Bish, D. L. & Post, J. E. (Mineral. Soc. of Am., Washington, DC), Rev. in Mineralogy 20, pp. 277–308. 48. Post, J. E. & Bish, D. L. (1988) Am. Mineral. 73, 861–869. 49. Post, J. E. & Veblen, D. R. (1990) Am. Mineral. 75, 477–489. 50. Pluth, J. J., Smith, J. V., Pushcharovsky, D. Y., Semenov, E. I., Bram, A., Riekel, C., Weber, H. P. & Broach, R. W. (1997) Proc. Natl. Acad. Sci. USA 94, 12263–12267. 51. Post, J. E., Von Dreele, R. B. & Buseck, P. R. (1982) Acta Crystallogr. B 38, 1056–1065.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

MANGANESE OXIDE MINERALS: CRYSTAL STRUCTURES AND ECONOMIC AND ENVIRONMENTAL SIGNIFICANCE

3454

52. Turner, S. & Post, J. E. (1988) Am. Mineral. 73, 1155–1161. 53. Yanchuk, E. A. (1977) Lithol. Miner. Res. 12, 733–737. 54. Schulze, D. G., Sutton, S. R. & Bajt, S. (1995) Soil Sci. Soc. Am. J. 59, 1540–1548. 55. Post, J. E. (1992) in Biomineralization, Processes of Iron and Manganese, eds. Skinner, H. C. W. & Fitzpatrick, R. W. (Catena Verlag, CremlingenDestedt, Germany), pp. 51–73. 56. Baur, W. H. (1976) Acta Crystallogr. B 32, 2200–2204. 57. Byström, A. M. (1949) Acta Chem. Scand. 3, 163–173. 58. Zwicker, W. K., Meijer, W. O. J. G. & Jaffe, H. W. (1962) Am. Mineral. 47, 246–266. 59. Nambu, M. & Tanida, K. (1967) J. Jpn. Assoc. Mineral. Petrol. Econ. Geol. 58, 39–54. 60. Post, J. E. & Bish, D. L. (1989) Am. Mineral. 74, 913–917. 61. Post, J. E. & Appleman, D. E. (1994) Am. Mineral. 79, 370–374. 62. Post, J. E. & Appleman, D. E. (1988) Am. Mineral. 73, 1401–1404. 63. Chukhrov, F. V., Gorshkov, A. I., Rudnitskaya, E. J., Berezovskaya, V. V. & Sivtsov, A. V. (1978) Izv. Akad. Nauk SSSR, Ser. Khim. 6, 5–19. 64. Dachs, H. (1963) Zeit. Kristall 118, 303–326. 65. Glasser, L. S. D. & Ingram, L. (1968) Acta Crystallogr. B 24, 1233–1236. 66. Bricker, O. (1965) Am. Mineral 50, 1296–1354. 67. Satomi, K. (1961) J. Phys. Soc. Jpn. 16, 258–265. 68. Geller, S. (1971) Acta Crystallogr. B 27, 821–828. 69. Sasaki, S., Fujino, K., Takeuchi, Y. & Sadanaga, R. (1980) Acta Crystallogr. A 36, 904–915. 70. Potter, R. M. & Rossman, G. R. (1979) Am. Mineral. 64, 1219–1226. 71. Turner, S. (1982) Ph.D thesis (Arizona State Univ., Tempe). 72. Manheim, F. T. (1965) in Symposia on Marine Chemistry, eds. Schink, D. R. & Corless, J. T. (Occasional Publications, Univ. of Rhode Island, Kingston), Vol. 3, 217–276. 73. Beyeler, H. U. (1976) Phys. Rev. Lett. 37, 1557–1560. 74. Ringwood, A. E., Kesson, S. E., Ware, N. G., Hibberson, W. & Major, A. (1979) Nature (London) 278, 219–223. 75. Ringwood, A. E. & Reid, A. F. (1967) Acta Crystallogr. 23, 1093–1099. 76. Turner, S. & Buseck, P. R. (1979) Science 203, 456–458. 77. Fleischer, M. & Richmond, W. E. (1943) Econ. Geol. 38, 269–286. 78. Turner, S. & Buseck, P. R. (1981) Science 212, 1024–1027. 79. Chukhrov, F. V., Gorshkov, A. I., Sivtsov, A. V. & Beresovskaya, V. V. (1978) Izv. Akad. Nauk SSSR, Ser. Khim. 12, 86–95. 80. Ostwald, J. (1986) Mineral. Mag. 50, 336–340. 81. Burns, V. M. & Burns, R. G. (1978) Earth Planet. Sci Lett. 39, 341–348. 82. Shen, Y. F., Zerger, R. P., Deguzman, R. N., Suib, S. L., McCurdy, L., Potter, D. I. & O'Young, C. L. (1993) Science 260, 511–515. 83. Mandernack, K. W., Post, J. E. & Tebo, B. M. (1995) Geochim. Cosmochim. Acta 59, 4393–4408. 84. Dasgupta, D. R. (1965) Mineral. Mag. 35, 131–139. 85. Feitknecht, W. & Marti, W. (1945) Helv. Chim. Acta 28, 129–148. 86. Frondel, C. (1953) Am. Mineral. 38, 761–769. 87. Feitknecht, W., Brunner, P. & Oswald, H. R. (1962) Z. Anorg. Allg. Chem. 316, 154–160. 88. De Villiers, J. E. (1983) Econ. Geol. 78, 1108–1118. 89. Ostwald, J. (1984) Mineral. Mag. 48, 383–388. 90. Ostwald, J. (1984) Ore Geol. Rev. 4, 3–45. 91. Manceau, A., Llorca, S. & Calas, G. (1987) Geochim. Cosmochim. Acta 51, 105–113. 92. Chukhrov, F. V., Gorshkov, A. I., Vitovskaya, V. I., Drits, V. A., Sivtsov, A. V. & Rudnitskaya, Y. S. (1982) Int. Geol. Rev. 24, 598–604. 93. Yan, G., Zhang, S., Zhao, M., Ding, J. & Li, D. (1992) Acta Mineral. Sinica 121, 69–77. 94. Grice, J. D., Gartrell, B., Gault, R. A. & Van Velthuizen, J. (1994) Can. Mineral. 32, 333–337. 95. Jones, L. J. P. & Milne, A. A. (1956) Mineral. Mag. 31, 283–288. 96. Giovanoli, R., Feitknecht, W. & Fischer, F. (1971) Helv. Chim. Acta 54, 1112–1124. 97. Burns, R. G., Burns, V. M. & Stockman, H. W. (1983) Am. Mineral. 68, 972–980. 98. Giovanoli, R. (1985) Am. Mineral. 70, 202–204. 99. Arrhenius, G. O. & Tsai, A. G. (1981) Scripps Inst. Oceanogr. 81–28, 1–19. 100. Ostwald, J. & Dubrawski, J. V. (1987) Neues Jahr. Mineral. Monatsh. 157, 19–34. 101. Giovanoli, R. & Bürki, P. (1975) Chimia 29, 114–117. 102. Paterson, E., Clark, D. R., Russell, J. D. & Swaffield, R. (1986) Clay Minerals 21, 957–964. 103. Giovanoli, R. & Arrhenius, G. (1988) in The Manganese Nodule Belt of the Pacific Ocean, eds. Halbach, P., Friedrich, G. & von Stackelberg, U. (Ferdinand Enke Verlag, Stuttgart), pp. 20–37. 104. Giovanoli, R. (1980) Miner. Deposita 15, 251–253. 105. Mason, B. (1944) Am. Mineral. 29, 66–69.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3455

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3455–3462, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA.

Negative pH, efflorescent mineralogy, and consequences for environmental restoration at the Iron Mountain Superfund site, California D. KIRK NORDSTROM*† AND CHARLES N. ALPERS‡ PNAS is available online at www.pnas.org.

ABSTRACT The Richmond Mine of the Iron Mountain copper deposit contains some of the most acid mine waters ever reported. Values of pH have been measured as low as −3.6, combined metal concentrations as high as 200 g/liter, and sulfate concentrations as high as 760 g/liter. Copious quantities of soluble metal sulfate salts such as melanterite, chalcanthite, coquimbite, rhomboclase, voltaite, copiapite, and halotrichite have been identified, and some of these are forming from negative-pH mine waters. Geochemical calculations show that, under a mine-plugging remediation scenario, these salts would dissolve and the resultant 600,000-m3 mine pool would have a pH of 1 or less and contain several grams of dissolved metals per liter, much like the current portal effluent water. In the absence of plugging or other at-source control, current weathering rates indicate that the portal effluent will continue for approximately 3,000 years. Other remedial actions have greatly reduced metal loads into downstream drainages and the Sacramento River, primarily by capturing the major acidic discharges and routing them to a lime neutralization plant. Incorporation of geochemical modeling and mineralogical expertise into the decision-making process for remediation can save time, save money, and reduce the likelihood of deleterious consequences. Mining and Water Quality Mining of metallic sulfide ore deposits (primarily for Ag, Au, Cu, Pb, and Zn) produces acid mine waters with high concentrations of metals that have harmful consequences for aquatic life and the environment. Deaths of fish, rodents, livestock, and crops have resulted from mining activities and have been noted since the days of the Greek and Roman civilizations. Mining and mineral processing have always created health risks for miners and other workers. In addition, mining wastes have often threatened the health of nearby residents by exposure to emissions of sulfur dioxide and oxides of As, Cd, Pb, and Zn from smelter stacks and flues, metal-contaminated soils, and waters and aquatic life with high concentrations of metals. As with most forms of resource extraction, human health risks accompany mineral exploitation. In 1985, the U.S. Environmental Protection Agency (EPA) estimated that 50 billion tons (45 × 1012 kg; 1 ton = 907 kg) of mining and mineral processing wastes had been generated in the United States and about 1 billion tons would continue to be generated each year (1). More recently, the EPA has described 66 “damage cases” at their web site (www.epa.gov, search for Mining and Mineral Processing Wastes, accessed Sept. 9, 1998) in which environmental injuries from mining activities in the U.S. are detailed. Government records indicate that many millions, perhaps billions, of fish have been killed from mining activities in the U.S. during this century (2). Incidents of arsenic poisoning in residents of Thailand result from arsenic contamination of the shallow groundwaters because of weathering of mine wastes (3). A mine flood disaster in Spain occurred in April 1998 in which about 6 million m3 of acid water and sulfide tailings escaped from a breached impoundment and covered about 6,500 acres of farmland and river banks along a 70-km reach of the Guadiamar River with fine-grained sulfides (details at www.csic.es). Numerous rivers, estuaries, and reservoirs throughout the world have been used as dumping grounds for the large volumes of waste produced during mineral extraction and processing. Mineral processing, in addition to fossil fuel and metal utilization, has increased the concentration of selected metals and nonmetals in the atmosphere. The emissions of As, Cd, Cu, Pb, Sb, and Zn from anthropogenic sources are all greater than emissions from natural sources, sometimes several times higher (4, 5). Acid mine drainage is produced primarily by the oxidation of the common iron disulfide mineral pyrite. Pyrite oxidation is a complex process that proceeds rapidly when this mineral and other sulfides are exposed to air. A simplified representation of this chemical process is given by the reaction of pyrite with air and water, +2 + 2H+(aq) [1] FeS2(s) + 7/2O2(g) + H2O(1) → in which the product is a solution of ferrous sulfate and sulfuric acid. The dissolved ferrous iron continues to oxidize and hydrolyze when the mine water is no longer in contact with pyrite surfaces, Fe2+(aq) + 1/4O2(g) + 5/2H2O(1) → Fe(OH)3(S) + 2H+(aq) [2] producing additional acidity. Iron- and sulfur-oxidizing bacteria, especially Thiobacillus ferrooxidans, are known to catalyze these reactions at low pH, increasing reaction rates by several orders of magnitude (6). These processes occur naturally and, indeed, natural acidic drainage is well known from many locations (7). Mining has the overall effect of dramatically increasing the oxidation rates by providing greater accessibility of air through mine workings, waste rock, and tailings, by creating greater surface area exposure through blasting, grinding, and crushing, and by concentrating sulfides in tailings. The overall rates of sulfide oxidation and metal

*United States Geological Survey, 3215 Marine Street, Boulder, CO 80303-1066; and ‡United States Geological Survey, Placer Hall, 6000 J Street, Sacramento, CA 95819-6129 Abbreviation: EPA, Environmental Protection Agency. †To whom reprint requests should be addressed. e-mail: [email protected].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3456

release in areas affected by mining are estimated to be orders of magnitude faster than natural rates. Another process, sometimes overlooked, plays an important role in the environmental consequences of mining: the formation of soluble, efflorescent salts. Acid ferrous sulfate solutions often become so enriched through rapid pyrite oxidation and evaporation that soluble salts form. These often appear as white, blue-green, yellow to orange or red efflorescent coatings on surfaces of waste rock, tailings, and in underground or open-pit mines. Acidity and metals, formerly contained in the acid mine water, are stored in the salts, which can quickly be dissolved by a rising groundwater table or be dissolved when exposed to rain and flowing surface waters, and then infiltrate to groundwaters. The Iron Mountain Mine Superfund site is an extreme example of how the formation of soluble efflorescent minerals can make certain remediation alternatives much more risky and potentially disastrous than might otherwise be imagined. Iron Mountain Iron Mountain is located in Shasta County, California, approximately 14 km northwest of the town of Redding (Fig. 1), in the southern part of the Klamath Mountains. “Iron Mountain Mine” is really a group of mines within Iron Mountain that include Old Mine, No. 8 Mine, Confidence-Complex, Brick Flat Open Pit Mine, Mattie Mine, Richmond and Richmond Extension Mines, and Hornet Mine. Ag, Au, Cu, Fe, Zn, and pyrite (for sulfuric acid production) were recovered at various times beginning in the early 1860s and ended with the termination of open-pit mining in 1962. Iron Mountain was once the largest producer of Cu in the state of California, and now it produces some of the most acidic waters in the world. Prior to the late 1980s when major remediation efforts began, approximately 2,500 tons of pyrite weathered every year from one mine alone (the Richmond Mine) and water containing about 300 tons per year of dissolved Cd, Cu, and Zn drained from the site into the Sacramento River. During periods of high runoff, sudden surges of acid mine waters into the Sacramento River have caused massive fish kills, which state and federal agencies have investigated since 1939. More than 20 fish-kill events have occurred in Sacramento River receiving waters since 1963, with at least 47,000 trout killed during a single week in 1967 (8). Furthermore, the town of Redding (with approximately 100,000 residents) receives its drinking water from the Sacramento River, downstream from the Iron Mountain site. Large, metal-rich sediment deposits containing toxic porewaters have built up in Keswick Reservoir, where the Spring Creek drainage from Iron Mountain empties into the Sacramento River. A brief history of mining, water management, and environmental action at Iron Mountain is outlined in Table 1.

FIG. 1. Location of Iron Mountain Mine, California (adapted from ref. 15). The mineral deposits are primarily massive sulfide lenses as much as 60 m thick containing up to 95% pyrite, variable amounts of chalcopyrite and sphalerite, and averaging about 1% Cu and about 2% Zn. Some disseminated sulfides occur along the south side of the mountain. The deposits at Iron Mountain and elsewhere in the West Shasta mining district are Devonian in age and have been classified as Kuroko type, having been formed in an island arc setting in a marine environment (9). The country rock is the Balaklala Rhyolite, a keratophyric rhyolite that has undergone regional metamorphism during episodes of accretion of oceanic crust to the continent. The brittle, fractured nature of the altered volcanic bedrock gives rise to a hydrologic conditions dominated by fracture-flow at Iron Mountain. The mineral composition of the rhyolite is albite, sericite, quartz, kaolinite, epidote, chlorite, and minor calcite; consequently it has little buffering capacity. Kinkel and others (10), Reed (11), and South and Taylor (12) have documented the chemical and isotopic compositions of ore, gangue, and country-rock minerals in the West Shasta mining district. Weathering of massive sulfide deposits at and near the surface has given rise to large gossan outcrops, enriched in Ag and Au. The 10 million tons of gossan in place prior to mining is the residue from at least 15 million tons of massive sulfide that weathered naturally. Table 1. Brief chronology of Iron Mountain mining and envrionmental activities Year Activity 1860s Discovery of massive gossan outcropping 1879 Silver discovered in gossan and mining begins 1897 Mountain Copper Co. acquires property and underground mining begins 1902 U.S. Forest Reserve sues company for vegetation damage from smelting activities 1907 Smelting ends and ore is transported to Martinez, CA, for processing 1928 California Fish and Game Commission files complaint regarding tailings dam 1939 State initiates water quality and fish toxicity studies 1943 Shasta Dam, upstream from Iron Mountain outflows, is completed 1950 Keswick Dam, downstream from Iron Mountain outflows, is completed 1955–1962 Open-pit mining of pyrite at Brick Flat for sulfuric acid production 1963 Spring Creek Debris Dam is completed, regulating outflow of acid mine waters to the Sacramento River 1967 Stauffer Chemical Co. acquires property 1976 Iron Mountain Mines, Inc., acquires property 1976–1982 State of California fines company for unacceptable releases of metals 1983 Iron Mountain listed on National Priorities List (NPL) for EPA Superfund, ranking as the third-largest polluter in the State of California Four Records of Decision by EPA have instituted several remedial activities that include partial capping, surface-water diversions, 1986–1998 tailings removal, and lime neutralization of the most acidic, metal-rich flows, reducing copper and zinc loads by 80–90%

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3457

A total of 7.5 million tons of sulfide ore was mined at Iron Mountain, and remaining reserves are estimated at approximately 15 million tons (13), so the overall size was at least 37.5 million tons prior to weathering. Preliminary paleomagnetic data on iron oxides in the gossan show portions with reversed polarity, indicating the gossan began forming at least 780,000 years ago. Secondary enrichment in the upper zones of the massive sulfides resulted in high concentrations of Cu (5–10%) and Ag (about 1 oz/ton). This enrichment took place at or near the water table during gossan formation. Three main massive sulfide ore bodies, the Brick Flat, the Richmond, and the Hornet, include most of the oxidizing sulfides causing the current water-quality problems. These ore bodies are thought to be parts of a single massive sulfide body about 0.8 km long, over 60 m wide, and over 60 m thick that was offset by two normal faults (Fig. 2). All three of these bodies have been mined, and the consequences include large changes in the hydrogeology, resulting in highly contaminated waters from tunnels, tailings, and waste rock piles. Acid Effluent from the Richmond Mine Conditions at Iron Mountain are nearly optimal for the production of acid mine waters, and this mine drainage is some of the most acidic and metal-rich reported anywhere in the world (14, 15). In the Richmond Mine, about 8 million tons of massive sulfide remain (13). At current weathering rates it would take about 3,200 years for the pyrite in the Richmond ore body to fully oxidize. The massive sulfide deposit is about 95% pyrite and is excavated by tunnels, shafts, raises, and stopes which allow rapid transport of oxygen by air advection. The sulfides are at or above the water table so that moisture and oxygen have ready access. Airflow is driven by the high heat output from pyrite oxidation. About 1,500 kJ of heat is released per mole (120 g) of pyrite. Air enters the main tunnel, heats up in the mine, then travels up through raises and shafts to the surface. The average flux of acid mine drainage from the Richmond portal indicates that about 2,400 mol of pyrite is oxidized every hour, producing about 1 kW of power or almost 9,000 kW per year. Water temperatures as high as 47°C have been measured underground, and the amorphous silica geothermometer (16, 17) would suggest temperatures of at least 50°C in the subsurface. In the early days of mining at Iron Mountain, fires were frequent during underground excavation, and temperatures of 430°F (221°C) were recorded at the ore surface (18). A considerable amount of historical data exist for effluent composition and discharge from the Richmond Mine because it is the largest single source of dissolved metals (both in terms of concentration and in terms of flux) in the Iron Mountain district. The Richmond ore body was discovered about 1915 but it was not mined on a large scale until the late 1930s and the war years (1940–1945). Regular monitoring of the Richmond Mine effluent by the California Regional Water Quality Control Board in cooperation with the EPA began in 1983. A summary of the data for discharge, pH, and Cu and Zn concentrations for 1983–1991 is shown in Table 2. Further compilation and details of Richmond portal effluent composition and discharge can be found in Alpers et al. (19).

FIG. 2. Cross-section of Iron Mountain (adapted from ref. 15). Table 2. Richmond Mine portal effluent characteristics, 1983–1991 Discharge, liter/s pH Zinc, mg/liter Copper, mg/liter

Mean 4.4 0.8 1,600 250

Range 0.5–50 0.02–1.5 700–2,600 120–650

Data are from ref. 19.

The variability in the Richmond effluent with time can be seen quite clearly for the 1986–1987 monitoring period. Fig. 3 covers the time period of late November 1986 to April 1987 and shows the rainfall (at Shasta Dam), and the consequent changes in Richmond Mine discharge and copper and zinc concentrations. One explanation for the large increase in copper concentrations is the dissolution of underground soluble salts from the flushing effect of meteoric recharge (see below). An observed increase in temperature with increased discharge may be the result of the dilution of concentrated sulfuric acid, the dissolution of soluble salts, and increased pyrite oxidation. One of the obvious options for remediation of the Richmond Mine was to plug it. Many mines have been plugged, but the consequences have not been consistently favorable. The EPA wanted to know what the consequences of plugging the Richmond Mine might be; for example, what would the composition of the resultant mine pool be? There was, however, no basis on which to speculate without some idea of the underground conditions. Hence, one of the activities of the Second Remedial Investigation Phase (1986–1992) under the Superfund Program was an underground survey of the Richmond tunnel and part of the mine workings. Prior to underground renovations in 1989–1990, the last underground tour, to the best of our knowledge, was in 1955 (Don White, U.S. Geological Survey, personal communication, 1989). The last mining had occurred in the late 1940s. Other than an occasional inspection by a company employee, there had been no recorded observation of the underground workings for 35–40 years. After underground renovations, entry was safe, and on September 10–12, 1990, water and mineral samples were

FIG. 3. Variations in rainfall, discharge, and copper and zinc concentrations for the Richmond portal effluent, 1986–1987 (adapted from ref. 19).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3458

collected. They revealed extremely acidic seeps with pH values as low as −3.6 and total dissolved solids concentration of more than 900 g/liter. The chemical compositions of five of the most acidic waters found underground in the Richmond Mine during 1990–1991 are shown in Table 3. These concentrations are the highest ever recorded for As, Cd, Fe, and SO4 and nearly the highest for Cu and Zn in groundwater. The high subsurface temperatures have induced considerable evaporation, which, in addition to pyrite oxidation, has caused the high concentrations of dissolved metals and sulfate. The reporting of negative pH values has been controversial, and for several good reasons. The conventional definition of pH based on the former National Bureau of Standards criteria and defined buffer systems limits the range of definable and measurable pH values to that of 1 to 13. Outside this range, the concept and measurement of pH are difficult at best. Furthermore, a new definition of pH must be used that is consistent with the conventional definition, different buffers must be used, and electrode performance and interferences must be determined. The most acceptable model for activity coefficients at present for defining pH below 1.0 is the Pitzer ion-interaction approach (20, 21). Acid mine waters are solutions of sulfuric acid, so the Pitzer model applied to sulfuric acid (22, 23) could serve as a definition for pH. Standardized sulfuric acid solutions would then serve as buffer solutions for calibration and the remaining question is the performance of standard glass membrane electrodes under these extreme conditions. Several Orion Ross glass membrane electrodes and a Sargent–Welch glass membrane electrode all performed well and could be calibrated up to a sulfuric acid concentration of about 8 molal. Another difficulty facing the definition of pH below 0.0 is scaling of individual ion activity coefficients. There is no generally accepted procedure for defining individual ion activity coefficients without some arbitrary assumptions. Two common methods with the Pitzer approach include “unscaled” Pitzer equations, and “MacInnes scaled,” using the MacInnes assumption (24). The MacInnes assumption is simpler, more flexible for a wide range of complex chemical compositions, and is more consistent with conventional speciation models applied to natural waters (24). It could be argued that the MacInnes assumption becomes less defensible at high concentrations where the unscaled approach should be more appropriate, but there is no obvious justification for using one approach over the other and the choice remains arbitrary. In the present investigations, the MacInnes scaling was used primarily because geochemists who have applied the Pitzer method to the interpretation of brines and saline waters find the MacInnes assumption more consistent with conventional practice. If the unscaled approach is used, the resultant pH values begin to differ significantly from MacInnes scaling for sulfuric acid solutions with pH values below −0.5. For example, at a sulfuric acid concentration of about 5.0 molal a scaled pH would be −2, whereas the unscaled pH would be notably higher, about −1.2. Some of these negative-pH mine waters were in apparent equilibrium with prominent soluble salts. Table 3. Compositions of five extremely acid mine water samples from the Richmond Mine Concentration of element in sample, mg/liter Element 90WA103 34.8°C 90WA109 38°C pH 90WA110A 42°C pH pH 0.48 −0.7 −2.5 Aluminum 2,210 6,680 1,420 Antimony 4.0 16 29 Arsenic(III) 8.14 38 32 Arsenic (total) 56.4 154 340 Barium 0.068 0.1 0.2 Beryllium 0.026 0.1 0.2 Boron 1.5 2.5 17 Cadmium 15.9 48.3 211 Calcium 183 330 279 Chromium 0.12 0.75 0.6 Cobalt 1.3 15.5 5.3 Copper 290 2,340 4,760 Iron(II) 18,100 79,700 34,500 Iron (total) 20,300 86,200 111,000 Lead 3.6 3.8 11.9 Magnesium 821 1,450 437 Manganese 17.1 42 23 Molybdenum 0.59 1.0 4.2 Nickel 0.66 2.9 3.7 Potassium 261 1,170 194 Selenium 0.42 2.1 4.2 Silicon (as SiO2) 170 34 35 Silver 0.16 0.65 2.4 Sodium 251 939 416 Strontium 0.25 0.49 0.90 Sulfur (as SO4) 118,000 360,000 760,000 Thallium 0.44 0.15 0.39 Tin 1.6 15 41 Titanium 5.9 125 1.0 Vanadium 2.9 11 15 Zinc 2,010 7,650 23,500 Melanterite Rhomboclase, römerite Associated mineral(s) A dash indicates no determination was made.

90WA110C 46°C pH −3.6 — — — — — — — — — — — — 9,790 16,300 — — — — — — — — — — — — — — — — Rhomboclase

91WA111 28°C — 6,470 15 74 850 <0.1 <0.1 — 370 443 2.6 3.6 9,800 — 68,100 8.3 2,560 119 2.3 6.3 11.1 <2.8 — 0.70 44 — — 1.6 — — 28 49,300

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3459

FIG. 4. Growth of cuprian melanterite in a manway of the Richmond Mine with stalactite dripping pH = −0.7 water into plastic beaker. (Photo by D.K.N. and C.N.A.) For example, a stalactite of zincian-cuprian melanterite had water dripping from the tip that had a pH of −0.7 (Table 3 and Fig. 4). Enormous quantities of highly soluble iron sulfate salts were found as efflorescences and precipitates, coating walls, ceilings, and floors of the mine and growing out of muck piles in colorful assemblages. Identification of these soluble salts made it possible to estimate what the composition of a mine pool formed by mine plugging might be. Soluble Salts and Consequences of the Mine-Plugging Scenario Ten soluble iron sulfate salts plus gypsum and chalcanthite were identified in the Richmond Mine. These minerals and their idealized formulae are listed in Table 4, with the iron salts in approximate sequence downward from the early formed to the later formed. Rhomboclase was found as stalactites and stalagmites (Fig. 5), and clusters of coquimbite, römerite, copiapite, and voltaite crystals were common throughout the mine (Fig. 6). Rhomboclase was rarely found without voltaite crystals. As long as an acid mine water is in contact with pyrite, the dissolved iron will remain in the ferrous state because of the strong reducing capacity of the pyrite. Rapidly flowing mine water will still maintain a high proportion of ferrous iron because the oxidation rate is often slow enough relative to the flow rate of the water. Consistent with this expectation, the only iron sulfate salts containing exclusively ferrous iron, melanterite, rozenite, and szomolnokite, are found close to pyrite sources and associated with more rapidly flowing waters. Ferric-bearing minerals are found to form in more stagnant conditions and can be considered to be hydrologic “dead-ends,” where much of the FeII has had time to oxidize to FeIII. Additional evidence for this mineralogical evolution is the observation that melanterite is the first-formed mineral when typical acid mine water is allowed to evaporate under ambient conditions and rhomboclase and voltaite are the last formed (25). Table 4. Idealized formulae of sulfate minerals found in the Richmond Mine Mineral Idealized formula Melanterite FeIISO4·7H2O Rozenite FeIISO4·4H2O Szomolnokite FeIISO4·H2O Copiapite Römerite Coquimbite Kornelite Rhomboclase

(H3O)FeIII(SO4)2·3H2O

Voltaite Halotrichite–bilinite Gypsum Chalcanthite

FeII(Al,FeIII)2(SO4)4·22H2O CaSO4·2H2O CuSO4·5H2O

A copper–zinc partitioning study of melanterite demonstrates that melanterite prefers copper over zinc (15). The consequences of this partitioning are that portal effluents will tend to have higher ratios of Zn/Cu during the dry season when melanterite is forming underground and lower Zn/Cu ratios in the wet season when these salts are dissolved and flushed from the mine workings. This trend is seen in the historical data on the Richmond Mine effluent (15). Dissolution of these soluble, iron sulfate salts (with variable amounts of copper, zinc, cadmium, and

FIG. 5. Stalagmite of rhomboclase (white) and coquimbite (purple) in the Richmond Mine. (Photo by C.N.A. and D.K.N.) FIG. 6. Cluster of coquimbite, voltaite, and copiapite from the Richmond Mine. (Photo by G. Robinson, Canadian Museum of Nature, Ottawa.)

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3460

aluminum substituting for the iron) can generate acidic solutions with high concentrations of dissolved metals. During the rising limb of a stream discharge in central Virginia after the onset of rain, Dagenhart (26) showed that rapid increases in the concentrations of Cu, Zn, Fe, and Al resulted from the dissolution of efflorescent salts found on upstream tailings and waste rock piles. This phenomenon must be common at mine waste sites and is likely to be an important cause of fish kills associated with periods of high runoff, especially after prolonged dry periods. Now we consider the consequences of dissolution of the enormous quantity of salts in the Richmond Mine in a mine-plugging scenario. The chemical composition of the mine pool created by plugging the Richmond Mine can be estimated by allowing these salts to dissolve in a volume of water equivalent to the void space created by the underground workings. The exact proportion of the different type of salts is not known, but the results of the calculations are not particularly sensitive to this factor. The amount of salts stored underground is a more critical factor, and so that was considered a variable. Computations were made by inputting the mineral compositions to thePHREEQE program (ref. 27, now superseded byPHREEQC, ref. 28) for a range of salt volumes.PHREEQE can calculate the speciation and chemical equilibrium for mass transfer processes such as precipitation, dissolution, oxidation–reduction reactions, ion exchange, and gas addition or removal (29). The results are shown in Fig. 7, where the resultant pH in the mine pool is plotted against the volume of added salts under two scenarios: active infiltration (actively injecting clean water) and passive infiltration (letting the groundwater naturally fill the void spaces). The latter scenario gives a worse picture because passive infiltration would allow more pyrite oxidation and the buildup of more acid waters. The salts probably occupy about 1% of the volume of the mine workings based on visual inspection from the limited subsurface survey. As can be seen in Fig. 7, however, an error in this value makes little difference. The consequences are that a mine pool of about 600,000 m3 with a pH at or below 1, with many grams of dissolved metals per liter (much like the current portal effluent), would likely form at or near the top of the groundwater table, in a rock with almost no neutralization capacity, and in which the hydrologic flow is governed by fractures, excavations, and drill-holes. Thus, plugging presents a remediation scenario that has a high degree of risk with potentially dangerous results. It has been common engineering practice to plug abandoned or inactive mines without monitoring, modeling, or even considering the physical and chemical consequences. Major leaks or failures at plugs, widespread and disseminated seeps of enriched acid mine waters, and increases in subsurface head pressures of more than 100 m have occurred. For some mine sites, plugging may ultimately prove to be successful, but more careful planning and peer review are essential to lessen the probability of disastrous results.

FIG. 7.PHREEQE simulation of water composition for mine pool after plugging the Richmond Mine. Regulatory Investigations and Remediation Several investigations and regulatory actions at Iron Mountain have been initiated by California State agencies over the last few decades. These are too lengthy to summarize here. Since the original listing of Iron Mountain on the National Priorities List in 1983, the EPA has authorized four Records of Decision (RODs) and has considered numerous options for remediation. A condensed version of the main remedial alternatives is as follows: • • • • • • • • • • •

No action Surface-water diversion Lime neutralization Capping (partial or complete capping of the mountain to prevent infiltration) Enlargement of Spring Creek Debris Dam (acid water storage and release structure) Ground-water interception Air sealing Mine plugging On-site leaching and solution extraction Continued mining under environmentally safe conditions Combined alternatives

Surface-water diversions have been installed to divert clean headwater streams around contaminated areas. The waters that are the largest sources of metal loadings have been captured and diverted to a lime neutralization plant. In the late 1980s, an emergency lime neutralization plant with a capacity of about 60 gallons per minute (gpm; 1 gallon = 3.8 × 10−3 m3) was installed to handle the worst flows from the Richmond and Hornet portals. By December of 1992, this plant had been expanded to handle 140 gpm, but was operated only 4 months per year during highest flows. In July of 1994 a new plant with a capacity of about 1,400 gpm began operation at the Minnesota Flats tailings site. In 1996 it was upgraded to 2,000 gpm, and high-density sludge treatment was added. Now it accepts drainage from Slickrock Creek (pumped from Old Mine and No. 8 Mine workings) as well as the Richmond and Hornet Mine portal effluents. The decision to build the larger treatment plant and to treat discharges from the Lawson tunnel (Hornet Mine) was also influenced by geochemical modeling. Opinion was divided as to whether the flow of acid mine water from the Lawson tunnel originates from the Richmond Mine by spillage or leakage or whether the Hornet ore body produces its own contaminant effluent. An ore chute and a raise that connected the two mines were identified from the old mine maps (Fig. 2). Because of its proximity to the surface and the collapsed nature of the mine workings, it was generally agreed that the Hornet Mine itself could not be effectively plugged. However, consultants proposed that plugging the Richmond Mine would stop or greatly reduce the flow from the Lawson tunnel. There was also reason to believe that during the intervening years since mining ceased, cave-ins and other ground failures had largely cut off direct connections from the Richmond to the Hornet. Alpers and others (19) studied the historical data on rainfall–discharge relationships between the two mines, Zn/Cu ratios as a signature of reactions within each mine site, and mass balance calculations for the two portal effluents. The most definitive method of determining the possible influence of the Richmond Mine water on the Lawson tunnel effluent was a mass balance approach. Using the known water compositions discharging from each mine and knowing the composition of the minerals that are reacting to form the effluent

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3461

waters, it is possible to calculate the mass amounts of minerals dissolved or precipitated to produce these waters by using the BALANCE program (30). Mineral reaction signatures were developed for each mine effluent separately, and then Richmond effluent was mixed with clean ground water and allowed to precipitate and dissolve additional minerals to determine if it was possible to derive the Lawson effluent from the Richmond. No version of this mass balance model produced a water that matched the Lawson effluent. Next, Richmond effluent was also mixed with Lawson effluent, and geochemical reaction was allowed, to see how much effluent each mine could be contributing to the Lawson. The model results indicated that not more than about 2% of the Richmond effluent could be present in the Lawson effluent. Therefore, the Hornet Mine is producing its own effluent independently of the Richmond Mine. Even if the Richmond Mine were successfully plugged, water from the Hornet would continue to be a significant problem and it would have to be treated. The fourth Record of Decision, issued in September of 1997, selected the construction of a dam on Slickrock Creek. This structure will capture the largest remaining loads of Cu and Zn and divert them to the neutralization plant for treatment. The remaining remediation is now focused on Boulder Creek, lower Spring Creek, Spring Creek Reservoir, and the metal-enriched sediments that formed in Keswick Reservoir from the neutralization of acid mine waters for nearly 50 years. The EPA and the potentially responsible parties remain in legal contention over the appropriate final remediation approaches to be used at Iron Mountain and the costs. Both the U.S. Government and the potentially responsible parties have funded a considerable number of investigations, remediation efforts, legal fees, and oversight management. The loads of copper, zinc, and cadmium into the Sacramento River have been reduced by 80–90%, and further remediation is in progress or being planned. The main challenge that remains is how to find a permanent (and passive) treatment solution in light of the fact that the mine drainage will continue for approximately 3,000 years unless the sulfide ore is mined out. Conclusion Prevention and control of contamination at mine sites is a challenging task, and remediation of large inactive mine sites such as Iron Mountain has proven to be extraordinarily difficult, complex, and expensive, not to mention litigious. The physical and chemical nature of the site makes it difficult to assess the effectiveness of remediation and the relative risks and costs of various alternatives and their contingencies. There are no easy solutions to these types of environmental problems, but several important points can be made about cleanup of mine waste sites on the basis of our experiences at Iron Mountain. First, there is tremendous value to having a technical advisory team of multidisciplinary professionals, without an obvious conflict of interest, to advise the regulatory agencies, to review data, and to make recommendations. Mine sites and their contaminants are complex functions of the geology, hydrology, geochemistry, pedology, meteorology, microbiology, and mining and mineral processing history, and their remediation is subject to considerations of economic limitations, available technology, and potential land use. Furthermore, the risks of failed remediation or no action are often poorly known. Assessing such risks involves toxicology, epidemiology, wildlife biology, and dealing with public perception. To ignore professionals in these areas, who can contribute both to the wisest choice of remediation strategies and to public awareness and education is to invite mistakes. Second, the effectiveness of a remedial alternative usually cannot be easily quantified or predicted. Hence, we must admit that remediation is experimental. Research is required to effect the best and most appropriate remediation available at a given time for a given site. Both longterm and short-term remediations are needed. For the short term, we need to fill in the knowledge gaps, especially as they pertain to a particular site. For the long term, we need to continue to develop better remediation techniques and mining and processing techniques that can utilize mine wastes and mineral deposits of lower grade. Mineralogical and geochemical knowledge make it possible to foresee the potential consequences of a remedial option and to plan a remediation strategy. The results of long-term research by the U.S. Geological Survey provided technical tools (computer programs for geochemical modeling and procedures for measuring pH) that could be used to answer important questions regarding remediation scenarios. Third, it would seem prudent to proceed on mine waste cleanup in a phased, iterative approach. Our natural inclination is to identify the worst part of a hazardous waste site and attempt to clean it up. For Iron Mountain, there is no single remedial solution that would clean up 90% of the problem on a permanent and maintenance-free basis (with the exception of completely mining the mountain). There are, however, several options (most of which have been exercised) that are low risk and low cost and should reduce the discharge of acid mine waters. These options can be instituted while deliberations and research continue to find the long-term solution. Fourth, mine waste sites commonly contain low-grade resources that are potentially mineable—it requires the right technology to make resource recovery economic. In an age of increasing recycling, recycling strategies should be applied to mine sites. Many mine wastes have already undergone further metals extraction and others could be stockpiled or tested for new uses. Additional research into metal recovery from acidic solutions could also provide economic incentive to recycling metals from mine drainage waste streams. Finally, Iron Mountain has been an extraordinary and extreme environment in which to study and document the processes of acid mine water production and efflorescent mineral formation, the value of which goes far beyond just the immediate remediation needs. The processes and properties found at Iron Mountain are probably commonplace at metal sulfide mine and mineral processing sites, but usually on a smaller scale. We now have some direct observations of the composition of water that produces efflorescent minerals. We have some idea of the consequences of efflorescent mineral dissolution when a mine is plugged. We can estimate the geochemical consequences of various remediation scenarios for mine sites with better confidence. Unraveling the dynamic processes that affect water–mineral interactions is often critical to solving hazardous waste problems in the hydrogeologic environment. We are grateful to personnel of Region 9 of the U.S. EPA, especially Rick Sugarek, for their continued support of our investigations on this project and to personnel of CH2M Hill for their help and assistance in our efforts to answer technically challenging questions. We thank the California Regional Water Quality Control Board in Redding and all the state agencies that have worked on Iron Mountain for their cooperation and support. Roger Ashley and Katie Walton-Day (U.S. Geological Survey) and James Hanley and Carol Russell (EPA) provided helpful reviews. We also acknowledge Rick Sugarek for making helpful suggestions on the manuscript. 1. USEPA (U.S. Environmental Protection Agency) (1985) Wastes from Extraction and Beneficiation of Metallic Ores, Phosphate Rock, Asbestos, Overburden from Uranium Mining, and Oil Shale: Report to Congress EPA/530-SW-85–033. 2. Nordstrom, D. K. & Alpers, C. N. (1999) in Environmental Geochemistry of Mineral Deposits, eds. Plumlee, G. S. & Logsdon, M. J., Reviews in Economic Geology (Soc. Econ. Geol., Littleton, CO), in press.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

NEGATIVE PH, EFFLORESCENT MINERALOGY, AND CONSEQUENCES FOR ENVIRONMENTAL RESTORATION AT THE IRON MOUNTAIN SUPERFUND SITE, CALIFORNIA

3462

3. Choprapawon, C. (1998) Abstracts of International Conference on Arsenic Pollution of Ground Water in Bangladesh: Causes, Effects, and Remedies; Dhaka, Feb. 8–12, 1998, pp. 77–78. 4. Church, T., Arimoto, R., Barrie, L. A., Dehairs, F., Dulac, F., Jickells, T. D., Mart, L., Sturges, W. T. & Zollar, W. H. (1990) in The Long-Range Atmospheric Transport of Natural and Contaminant Substances, ed. Knap, A. H. (Kluwer, Dordrecht, the Netherlands), pp. 37–58. 5. Buat-Ménard, P. (1993) in Global Atmospheric Chemical Change, eds. Hewitt, C. N. & Sturges, W. T. (Elsevier Science, Amsterdam), pp. 271–311. 6. Nordstrom, D. K. & Southam, G. (1997) in Geomicrobiology: Interactions between Microbes and Minerals, eds. Banfield, J. F. & Nealson, K. H., Reviews in Mineralogy (Mineral Soc. Am., Washington, DC), Vol. 35, pp. 361–390. 7. Runnells, D. D., Shepard, T. A. & Angino, E. E. (1992) Environ. Sci. Technol. 26, 2316–2322. 8. Nordstrom, D. K., Jenne, E. A. & Averett, R. C. (1977) Heavy metal discharges into Shasta Lake and Keswick Reservoir on the Sacramento River, California—A Reconnaissance During Low Flow, U.S. Geological Survey Open-File Report 76-49. 9. Albers, J. P. & Bain, J. H. C. (1985) Econ. Geol. 80, 2072–2091. 10. Kinkel, A. R., Hall, W. E. & Albers, J. P. (1956) Geology and base-metal deposits of the West Shasta copper-zinc district, Shasta County, California, U.S. Geological Survey Professional Paper 285. 11. Reed, M. H. (1984) Econ. Geol. 79, 1299–1318. 12. South, B. C. & Taylor, B. E. (1985) Econ. Geol. 80, 2177–2195. 13. Kaiser Engineering (1990) Iron Mountain Mine Property, West Shasta District, California—Technical Review. Prepared for Iron Mountain Mines, November 1981. 14. Nordstrom, D. K. (1977) Ph.D. Dissertation (Stanford Univ., Stanford, CA). 15. Alpers, C. N., Nordstrom, D. K. & Thompson, J. M. (1994) in Environmental Geochemistry of Sulfide Oxidation, eds. Alpers, C. N. & Blowes, D. W., American Chemical Society Symposium Series (Am. Chem. Soc., Washington, DC), Vol. 550, pp. 324–344. 16. Fournier, R. O. & Rowe, J. J. (1966) Am. J. Sci. 264, 685–697. 17. Fournier, R. O. (1985) in Geology and Geochemistry of Epithermal Systems, eds. Berger, B. R. & Bethke, P. M., Reviews in Economic Geology (Soc. Econ. Geol., Littleton, CO), Vol. 2, pp. 45–61. 18. Wright, L. T. (1906) Eng. Min. J. 81, 171–172. 19. Alpers, C. N., Nordstrom, D. K. & Burchard, J. M. (1992) Compilation and interpretation of water-quality and discharge data for acidic mine waters at Iron Mountain, Shasta County, California 1940–91, U.S. Geological Survey Water-Resources Investigations Report 91–4160. 20. Pitzer, K. S. (1973) J. Phys. Chem. 77, 268–277. 21. Pitzer, K. S. (1991) in Activity Coefficients in Electrolyte Solutions, ed. Pitzer, K. S. (CRC Press, Boca Raton, FL), 2nd Ed., pp. 75–153. 22. Pitzer, K. S., Roy, R. N. & Silvester, L. F. (1977) J. Am. Chem. Soc. 99, 4930–4936. 23. Clegg, S. L., Rard, J. A. & Pitzer, K. S. (1994) J. Chem. Soc. Faraday Trans. 90, 1875–1894. 24. Plummer, L. N., Parkhurst, D. L., Fleming, G. W. & Dunkle, S. A. (1988) A computer program incorporating Pitzer's equations for calculation of geochemical reactions in brines, U.S. Geological Survey Water-Resources Investigations Report 88-4153. 25. Buurman, P. (1975) Geologie en Mijnvouw 54, 101–105. 26. Dagenhart, T. V., Jr. (1980) M.S. thesis (Univ. of Virginia, Charlottesville). 27. Parkhurst, D. L., Thorstenson, D. C. & Plummer, L. N. (1980) PHREEQE—A computer program for geochemical calculations, U.S. Geological Survey Water-Resources Investigations Report 80-96. 28. Parkhurst, D. L. (1995) User's guide to PHREEQC—A computer program for speciation, reaction-path, advective-transport, and inverse geochemical calculations, U.S. Geological Survey Water-Resources Investigations Report 95-4227. 29. Alpers, C. N. & Nordstrom, D. K. (1999) in Environmental Geochemistry of Mineral Deposits, eds. Plumlee, G. S. & Logsdon, M. J., Reviews in Economic Geology (Soc. Econ. Geol., Littleton, CO), in press. 30. Parkhurst, D. L., Plummer, L. N. & Thorstenson, D. C. (1982) BALANCE—A computer program for calculating mass transfer for geochemical reactions in ground water, U.S. Geological Survey Water-Resources Investigations Report 82-14.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3463

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3463–3470, March 1999 Colloquium Paper

La roca magica: Uses of natural zeolites in agriculture and industry

This paper was presented at National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA. FREDERICK A. MUMPTON* Edit Inc., P.O. Box 591, Clarkson, NY 14430 PNAS is available online at www.pnas.org.

ABSTRACT For nearly 200 years since their discovery in 1756, geologists considered the zeolite minerals to occur as fairly large crystals in the vugs and cavities of basalts and other traprock formations. Here, they were prized by mineral collectors, but their small abundance and polymineralic nature defied commercial exploitation. As the synthetic zeolite (molecular sieve) business began to take hold in the late 1950s, huge beds of zeolite-rich sediments, formed by the alteration of volcanic ash (glass) in lake and marine waters, were discovered in the western United States and elsewhere in the world. These beds were found to contain as much as 95% of a single zeolite; they were generally flat-lying and easily mined by surface methods. The properties of these low-cost natural materials mimicked those of many of their synthetic counterparts, and considerable effort has made since that time to develop applications for them based on their unique adsorption, cation-exchange, dehydration–rehydration, and catalytic properties. Natural zeolites (i.e., those found in volcanogenic sedimentary rocks) have been and are being used as building stone, as lightweight aggregate and pozzolans in cements and concretes, as filler in paper, in the take-up of Cs and Sr from nuclear waste and fallout, as soil amendments in agronomy and horticulture, in the removal of ammonia from municipal, industrial, and agricultural waste and drinking waters, as energy exchangers in solar refrigerators, as dietary supplements in animal diets, as consumer deodorizers, in pet litters, in taking up ammonia from animal manures, and as ammonia filters in kidney-dialysis units. From their use in construction during Roman times, to their role as hydroponic (zeoponic) substrate for growing plants on space missions, to their recent success in the healing of cuts and wounds, natural zeolites are now considered to be full-fledged mineral commodities, the use of which promise to expand even more in the future. The discovery of natural zeolites 40 years ago as large, widespread, mineable, near-monomineralic deposits in tuffaceous sedimentary rocks in the western United States and other countries opened another chapter in the book of useful industrial minerals whose exciting surface and structural properties have been exploited in industrial, agricultural, environmental, and biological technology. Like talc, diatomite, wollastonite, chrysotile, vermiculite, and bentonite, zeolite minerals possess attractive adsorption, cation-exchange, dehydration–rehydration, and catalysis properties, which contribute directly to their use in pozzolanic cement; as lightweight aggregates; in the drying of acid-gases; in the separation of oxygen from air; in the removal of NH3 from drinking water and municipal wastewater; in the extraction of Cs and Sr from nuclear wastes and the mitigation of radioactive fallout; as dietary supplements to improve animal production; as soil amendments to improve cation-exchange capacities (CEC) and water sorption capacities; as soilless zeoponic substrate for greenhouses and space missions; in the deodorization of animal litter, barns, ash trays, refrigerators, and athletic footwear; in the removal of ammoniacal nitrogen from saline hemodialysis solutions; and as bactericides, insecticides, and antacids for people and animals. This multitude of uses of natural zeolites has prompted newspapers in Cuba, where large deposits have been discovered, to refer to zeolites as the magic rock, hence the title of this paper. The present paper reviews the critical properties of natural zeolites and important uses in pollution control, the handling and storage of nuclear wastes, agriculture, and biotechnology. The paper also pleads for greater involvement by mineral scientists in the surface, colloidal, and biochemical investigations that are needed in the future development of zeolite applications. PROPERTIES A zeolite is a crystalline, hydrated aluminosilicate of alkali and alkaline earth cations having an infinite, open, three-dimensional structure. It is further able to lose and gain water reversibly and to exchange extraframework cations, both without change of crystal structure. The large structural cavities and the entry channels leading into them contain water molecules, which form hydration spheres around exchangeable cations. On removal of water by heating at 350–400°C, small molecules can pass through entry channels, but larger molecules are excluded— the so called “molecular sieve” property of crystalline zeolites. The uniform size and shape of the rings of oxygen in zeolites contrasts with the relatively wide range of pore sizes in silica gel, activated alumina, and activated carbon, and the Langmuir shape of their adsorption isotherms allows zeolites to remove the last trace of a particular gas from a system (e.g., H2O from refrigerator Freon lines). Furthermore, zeolites adsorb polar molecules with high selectivity. Thus, polar CO2 is adsorbed preferentially by certain zeolites, allowing impure methane or natural gas streams to be upgraded. The quadrupole moment of N2 contributes to its selective adsorption by zeolites from air, thereby producing O2enriched products. The adsorption selectivity for H2O, however, is greater than for any other molecule, leading to uses in drying and solar heating and cooling. The weakly bonded extraframework cations can be removed or exchanged readily by washing with a strong solution of another cation. The CEC of a zeolite is basically a function of the amount of Al that substitutes for Si in the framework tetrahedra; the greater the Al content, the more extraframework cations needed to balance the charge. Natural zeolites have CECs from 2 to 4 milliequivalents/g (meq/g), about twice the CEC of bentonite clay. Unlike most noncrystalline ion

*To whom reprint requests should be addressed. e-mail: [email protected]. Abbreviations: CEC, cation-exchange capacity; meq, milliequivalent.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3464

FIG. 1. Scanning electron micrograph of plates of clinoptilolite from Castle Creek, ID [Reproduced with permission from ref. 3 (Copyright 1976, The Clay Minerals Society)]. exchangers, e.g., organic resins and inorganic aluminosilicate gels (mislabeled in the trade as “zeolites”), the framework of a crystalline zeolite dictates its selectivity toward competing ions. The hydration spheres of high field-strength cations prevent their close approach to the seat of charge in the framework; hence, cations of low field strength are generally more tightly held and selectively exchanged by the zeolite than other ions. Clinoptilolite has a relatively small CEC (≈2.25 meq/g), but its cation selectivity is Cs > Rb > K > NH4 > Ba > Sr > Na > Ca > Fe > Al > Mg > Li. This preference for larger cations, including NH4+, was exploited for removing NH4-N from municipal sewage effluent and has been extended to agricultural and aquacultural applications (1, 2). Clinoptilolite and natural chabazite have also been used to extract Cs and Sr from nuclear wastes and fallout. Most zeolites in volcanogenic sedimentary rocks were formed by the dissolution of volcanic glass (ash) and later precipitation of micrometer-size crystals, which mimic the shape and morphology of their basalt counterparts (Fig. 1; ref. 3). Sedimentary zeolitic tuffs are generally soft, friable, and lightweight and commonly contain 50–95% of a single zeolite; however, several zeolites may coexist, along with unreacted volcanic glass, quartz, K-feldspar, montmorillonite, calcite, gypsum, and cristobalite/tridymite. Applications of natural zeolites make use of one or more of the following properties: (i) cation exchange, (ii) adsorption and related molecular-sieving, (iii) catalytic, (iv) dehydration and rehydration, and (v) biological reactivity. Extrinsic properties of the rock (e.g., siliceous composition, color, porosity, attrition resistance, and bulk density) are also important in many applications. Thus, the ideal zeolitic tuff for both cation-exchange and adsorption applications should be mechanically strong to resist abrasion and disintegration, highly porous to allow solutions and gases to diffuse readily in and out of the rock, and soft enough to be easily crushed. Obviously, the greater the content of a desired zeolite, the better a certain tuff will perform, ceteris paribus. (See Table 1 for more information on the properties of zeolites.) Table 1. Representative formulae and selected physical properties of important zeolites* Zeolite Representative unit-cell formula Void volume, % Channel dimensions, Å Analcime Na10(Al16Si32O96)·16H2O 18 2.6 Chabazite (Na2Ca)6(Al12Si24O72)·40H2O 47 3.7 × 4.2 34 3.9 × 5.4 Clinoptilolite (Na3K3)(Al6Si30O72)·24H2O 35 3.6 × 5.2 Erionite (NaCa0.5K)9(Al9Si27O72)·27H2O 47 7.4 Faujasite (Na58)(Al58Si134O384)·240H2O 28 4.3 × 5.5 Ferrierite (Na2Mg2)(Al6Si30O72)·18H2O Heulandite (Ca4)(Al8Si28O72)·24H2O 39 4.0 × 5.5 4.4 × 7.2 4.1 × 4.7 34 4.6 × 6.3 Laumonitte (Ca4)(Al8Si16O48)·16H2O 28 2.9 × 5.7 Mordenite (Na8)(Al8Si40O96)·24H2O 6.7 × 7.0 31 4.2 × 4.4 Phillipsite (NaK)5(Al5Si11O32)·20H2O 2.8 × 4.8 3.3 47 4.2 Linde A (Na12)(Al12Si12O48)·27H2O (Na86)(Al86Si106O384)·264H2O 50 7.4 Linde X

Thermal stability (relative) High High High High High High Low

CEC, meq/g† 4.54 3.84 2.16 3.12 3.39 2.33 2.91

Low High

4.25 2.29

Medium

3.31

High High

5.48 4.73

*Modified from refs. 103 and 104. Void volume determined from water content. †Calculated from unit-cell formula.

APPLICATIONS Construction Dimension Stone. Devitrified volcanic ash and altered tuff have been used for 2,000 years as lightweight dimension stone. Only since the 1950s, however, has their zeolitic nature been recognized. Their low bulk density, high porosity, and homogeneous, close-knit texture have contributed to their being easily sawed or cut into inexpensive building blocks. For example, many Zapotec buildings near Oaxaca, Mexico, were constructed of blocks of massive, clinoptilolite tuff (4), which is still used for public buildings in the region. The easily cut and fabricated chabazite- and phillipsite-rich tuffo giallo napolitano in central Italy has also been used since Roman times in construction, and the entire city of Naples seems to be built out of it (Fig. 2). Numerous cathedrals and public buildings in central Europe were built from zeolitic tuff quarried in the Laacher See area of Germany. Early ranch houses (Fig. 3) in the American West were built with blocks of locally quarried erionite; they were cool and did not crumble in the arid climate. Similar structures made of zeolitic tuff blocks have been noted near almost every zeolitic tuff deposit in Europe and Japan (5). Cement and Concrete. The most important pozzolanic raw material used by the ancient Romans was obtained from the tuffo napolitano giallo near Pozzuoli, Italy (6, 7). Similar materials have been used in cement production throughout Europe. The high silica content of the zeolites neutralizes excess lime produced by setting concrete, much like finely powdered pumice or fly ash. In the U.S., nearly $1 million was saved in 1912 during the construction of the 240-mile-long Los Angeles aqueduct by replacing ≤25% of the required portland cement with an inexpensive clinoptilolite-rich tuff mined near Tehachapi, CA (8, 9).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3465

FIG. 2. Castel Nuovo (Naples, Italy) constructed of tuffo giallo napolitano [Reproduced with permission from ref. 105 (Copyright 1995, International Committee on Natural Zeolites)]. Lightweight Aggregate. Much like perlite and other volcanic glasses are frothed into low-density pellets for use as lightweight aggregate in concrete, zeolitic tuff can be “popped” by calcining at elevated temperature. Clinoptilolite from Slovenia and Serbia yields excellent aggregates of this type on firing to 1,200–1,400°C. Densities of ≥0.8 g/cm3 and porosities of ≤65% have been reported for expanded clinoptilolite products (10). These temperatures are somewhat higher than those needed to expand perlite, but the products are stronger (11). The Russian Sibeerfoam product is expanded zeolitic tuff and is used as lightweight insulating material (12). In Cuba, mortars for ferrocement boats and lightweight aggregate for hollow prestressed concrete slabs contain indigenous clinoptilolite (13, 14). The mortars have compressive strengths of ≤55.0 MPa; the ferrocement boats can withstand marine environments. Water and Wastewater Treatment Municipal Wastewater. Large-scale cation-exchange processes using natural zeolites were first developed by Ames (1) and Mercer et al. (2), who demonstrated the effectiveness of clinoptilolite for extracting NH4+ from municipal and agricultural waste streams. The clinoptilolite exchange process at the Tahoe–Truckee (Truckee, CA) sewage treatment plant removes >97% of the NH4+ from tertiary effluent (15). Hundreds of papers have dealt with wastewater treatment by natural zeolites. Adding powdered clinoptilolite to sewage before aeration increased O2-consumption and sedimentation, resulting in a sludge that can be more easily dewatered and, hence, used as a fertilizer (16). Nitrification of sludge is accelerated by the use of clinoptilolite, which selectively exchanges NH4+ from wastewater and provides an ideal growth medium for nitrifying bacteria, which then oxidize NH4+ to nitrate (17, 18 and 19). Liberti et al. (19) described a nutrient-removal process called RIM-NUT that uses the selective exchange by clinoptilolite and an organic resin to remove N2 and P from sewage effluent.

FIG. 3. Abandoned ranch house in Jersey Valley, NV, constructed of quarried blocks of erionite-rich tuff [Reproduced with permission from ref. 5 (Copyright 1973, Industrial Minerals)]. FIG. 4. Clinoptilolite-filled columns at a Denver, CO, water-purification plant [Reproduced with permission from ref. 106 (Copyright 1997, AIMAT)]. Drinking Water. In the late 1970s, a 1-MGD (million gallons per day) water-reuse process that used clinoptilolite cation-exchange columns went on stream in Denver, CO, (Fig. 4) and brought the NH4+ content of sewage effluent down to potable standards (<1 ppm; refs. 20, 21 and 22). Based on Sims and coworkers' (23, 24) earlier finding that nitrification of sewage sludge was enhanced by the presence of clinoptilolite, a clinoptilolite-amended slow-sand filtration process for drinking water for the city of Logan, UT, was evaluated. By adding a layer of crushed zeolite, the filtration rate tripled, with no deleterious effects. At Buki Island, upstream from Budapest, clinoptilolite filtration reduced the NH3 content of drinking water from 15–22 ppm to <2 ppm (25, 26). Clinoptilolite beds are used regularly to upgrade river water to potable standards at Ryazan and other localities in Russia and at Uzhgorod, Ukraine (27, 28).

FIG. 5. Methane-purification pressure-swing adsorption unit, NRG Company, Palos Verde Landfill, Los Angeles, CA [Reproduced with permission from ref. 106 (Copyright 1997, AIMAT)].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3466

The selectivity of several natural zeolites for Pb2+ suggests an inexpensive means of removing lead from drinking water. Adsorption and Catalysis Two principal uses of synthetic molecular sieves are the purification of gaseous hydrocarbons and the preparation of catalysts for petroleum refining. In general, natural zeolites do not compete with their synthetic counterparts in adsorption or catalytic applications because of their inherent lower adsorption capacities and, to some extent, to the presence of traces of Fe and other catalyst “poisons.” Most natural materials have smaller pore openings than the synthetics. Despite the low cost of the natural materials (a few cents per kilogram), the economics of hardware construction, activation, and regeneration favor the more expensive synthetics, even at $2.00/kg, for most adsorption applications. By using certain natural zeolites, however, researchers have made headway in the drying and purification of acid gases. Mordenite and chabazite, for example, can withstand the rigors of continuous cycling in acid environments and have been used to remove water and carbon dioxide from sour natural gas. Union Carbide Corporation (now UOP Corporation, Tarrytown, NY) marketed an AW-500 product (natural chabazite-rich tuff from Bowie, AZ) for removing HCl from reformed H2 streams (pH < 2), H2O from Cl2, and CO2 from stack gas emissions (29). NRG Corporation (Los Angeles, CA; ref. 30) used a pressure-swing adsorption process with Bowie chabazite to remove polar H2O, H2S, and CO2 from low-BTU (British thermal unit) natural gas and developed a zeolite-adsorption process for purifying methane produced by decaying garbage in a Los Angeles landfill (Fig. 5). A pressure-swing adsorption process using natural mordenite was developed in Japan to produce high-grade O2 from air (T. Tamura, unpublished work; refs. 31 and 32). Dominé and Häy (33) showed that the quadrupole moment of nitrogen is apparently responsible for its adsorption by a dehydrated zeolite in preference to oxygen, resulting in a distinct separation of the two gases for a finite length of time. Similar processes use synthetic CaA zeolite to produce O2 in sewage-treatment plants in several countries. In Japan, small zeolite adsorption units generate O2-enriched air for hospitals, in fish breeding and transportation, and in poorly ventilated restaurants. Modifying the surface of clinoptilolite with long-chain quaternary amines allowed it to adsorb benzene, toluene, and xylene in the presence of water, a process that shows promise in the clean up of gasoline and other petroleum spills (34, 35 and 36). These hydrophilic products can be treated further with additional amine to produce anion exchangers capable of taking up chromate, arsenate, selenate, and other metal oxyanions from aqueous solutions. Applications in catalysis include (i) a selective-forming catalyst developed by Mobil Corporation using natural erionite-clinoptilolite (37); (ii) a hydrocarbon conversion catalyst for the disproportionation of toluene to benzene and xylene, employing a hydrogen-exchanged natural mordenite (38); (iii) a catalyst using cation-exchanged clinoptilolite from Tokaj, Hungary, for the hydromethylation of toluene (39); and (iv) clinoptilolite catalysts for the isomerization of n-butene, the dehydration of methanol to dimethyl ether, and the hydration of acetylene to acetaldehyde (40). Nuclear Waste and Fallout Nuclear Waste. Early experiments were aimed at concentrating 137Cs and 90Sr from low-level waste streams of nuclear reactors and leaking repositories on clinoptilolite (41, 42 and 43). The “saturated” zeolite was transformed into concrete, glass, or ceramic bodies and stored indefinitely. Natural zeolites have superior selectivity for certain radionuclides (e.g., 90Sr, 137Cs 60Co, 45Ca, and 51Cr) compared with organoresins and are cheaper and much more resistant to nuclear degradation. Dozens of papers have demonstrated the ability of several natural zeolites to take up these and other radionuclides (44, 45, 46 and 47). A mixture of synthetic zeolite A and natural chabazite from Bowie, AZ, was used to take up Sr and Cs, respectively, from contaminated waters at Three Mile Island, PA (48). Clinoptilolite currently is used to remove Sr and Cs from low-level effluents from a nuclear power plant before they are released to the Irish Sea at Sellafield, U.K. (49), and to capture these isotopes from leaking repository containers at West Valley, NY (50). Nuclear Fallout. The same selectivities for Cs and Sr by zeolites permit treatment of radioactive fallout from nuclear tests and accidents. The addition of clinoptilolite to soils contaminated with 90Sr markedly reduced the strontium uptake by plants (51), and the presence of clinoptilolite inhibited the uptake of Cs in contaminated Bikini Atoll soils (52). Several zeolite processes have been developed to counteract the fallout from the 1986 Chernobyl disaster. Shenbar and Johanson (53) found that 137Cs in soils was not taken up by plants after treating the soil with a zeolite, and Forberg et al. (54) showed that a zeolite supplement to the diets of Swedish reindeer accelerated the excretion of 137Cs ingested with food contaminated by Chernobyl fallout. Zeolites added to soils reduced the uptake of 137Cs by pasture plants in the vicinity of Chernobyl (55), and dietary zeolite reduced sorption of radiocesium by sheep fed fallout-contaminated rations in Scotland (56). In Bulgaria, zeolite pills and cookies were prepared for human consumption to counteract Chernobyl fallout (57). The zeolite apparently exchanges 137Cs and 90Sr in the gastro-intestinal tract and is excreted by normal processes, thereby minimizing assimilation into the body. Agriculture Animal Nutrition and Health. Since 1965, studies in Japan using ≤10% clinoptilolite and mordenite as dietary supplements for swine and poultry showed that test animals generally grew faster than control groups, with simultaneous decrease in the amount and cost of the feed. Young and mature pigs fed rations containing 5% clinoptilolite gained 16% more weight than animals fed a normal diet (58, 59). The animals' excrement was less odoriferous because of the take up of NH4+ by the zeolite, and the number and severity of intestinal diseases decreased. To reduce the toxic effect of high NH4+ in ruminal fluids when nonprotein-nitrogens, such as urea and biuret, are added to animal diets, White and Ohlrogge (60) introduced both synthetic and natural zeolites into the rumen of test animals. NH4+ formed by the enzyme decomposition of the nonprotein-nitrogen was exchanged immediately by the zeolite and held for several hours until released by Na+ entering the rumen in saliva. This gradual release of the excess nitrogen allowed rumen organisms to synthesize cellular protein for assimilation into the animals' digestive systems. Several hundred studies of the effect of zeolites in animal diets have been made in the U.S. and elsewhere (61). The results are mixed, but in countries in which the level of animal productivity is not as high as in the U.S. and the sanitary conditions of feed lots and production facilities are much worse, substantial increases in productivity and mortality have been achieved (62). Weight gain may be caused by the zeolite acting as an ammonium reservoir in the gastrointestinal tract, thereby allowing the animal to use ingested nitrogen more efficiently. The prevention or minimization of scours and other intestinal diseases, however, is more baffling. An NH4+-containing zeolite may support the growth of nitrogen-loving bacteria that contribute to the health of the animals; the zeolite may take up deleterious heavy metals, or it may simply regulate pH in the gut system, resulting in fewer or less severe stomach ailments.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3467

FIG. 6. Tomatoes grown zeoponically in Havana, Cuba [Reproduced with permission from ref. 106 (Copyright 1997, AIMAT)]. These reactions await serious physiological and biochemical examination. Natural zeolites and some clay minerals have proven to be effective in protecting animals against mycotoxins (63, 64). The apparent ability of clinoptilolite and other zeolites to absorb aflatoxins that contaminate animal feeds has resulted in measurable improvements in the health of swine, sheep, and chickens (65, 66 and 67). Here also, more work is needed to verify and understand the mechanisms of such reactions. Agronomy and Horticulture. Natural zeolites are used extensively in Japan as amendments for sandy, clay-poor soils (68). The pronounced selectivity of clinoptilolite NH4+ and K+ also was exploited in Japan in slow-release chemical fertilizers. By using clinoptilolite-rich tuff as a soil conditioner, significant increases in the yields of wheat (13–15%), eggplant (19–55%), apples (13–38%), and carrots (63%) were reported when 4–8 tonne/acre zeolite was used (69). The addition of clinoptilolite increased barley yields (70); it also increased the yields of potatoes, barley, clover, and wheat after adding 15 tonne/ hectare to Ukrainian sandy loams (71). Clinoptilolite amended to a potting medium for chrysanthemums behaved like a slow-release K-fertilizer, yielding the same growth for the plants as daily irrigation with Hoagland's solution (72). The addition of NH4+-exchanged clinoptilolite in greenhouse experiments resulted in 59% and 53% increase in root weight of radishes in medium- and light-clay soils, respectively (73). A 10% addition of clinoptilolite to sand used in the construction of golf-course greens substantially reduced NO3-leaching and increased fertilizer-N uptake by creeping bent-grass, without disturbing the drainage, compaction, or “playability” of the greens (74, 75 and 76). In Italy, natural zeolites have been used as dusting agents to kill aphids afflicting fruit trees (J. L. Gonzalez, personal communication). The mechanism of this reaction is not known; the zeolite could be acting as a desiccant, although it is saturated almost completely with water before use, or its highly alkaline character in water could simply kill individual insects that come in contact with it. The use of clinoptilolite as the principal constituent of artificial soil was developed in Bulgaria in the late 1970s. The growth of plants in synthetic soils consisting of zeolites with or without peat, vermiculite, and the like has been termed zeoponics. A nutrient-treated zeoponic substrate used for growing crops and the rooting of cuttings in greenhouses produced greater development of root systems and larger yields of strawberries, tomatoes, and peppers, without further fertilization (77). Tomatoes and cucumbers are grown commercially outdoors in Cuba (Fig. 6) by using zeoponic substrate (78), and vegetables currently are supplied to Moscow in the winter from greenhouses that use zeoponic synthetic soils. By using a treated Bulgarian clinoptilolite product, cabbage and radishes have been grown aboard the Russian space station Mir (79). Considerable attention has been paid to zeoponic mixtures of NH4- or K-exchanged natural zeolites and sparingly

FIG. 7. Wheat grown zeoponically for use in space flights and stations, Johnson Space Center, Houston, TX [Reproduced with permission from ref. 106 (Copyright 1997, AIMAT)]. Photograph by D. W. Ming.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3468

FIG. 8. Zeolite deodorization products from Itaya, Japan [Reproduced with permission from ref. 106 (Copyright 1997, AIMAT)]. soluble phosphate minerals (e.g., apatite; refs. 80 and 81). The small amount of Ca in the soil solution in equilibrium with the apatite exchanges onto the zeolite, thereby disturbing the equilibrium and forcing more Ca into solution. The apatite is ultimately destroyed, releasing P to the solution, and the zeolite gives up its exchanged cations (e.g., K+ or NH4+). Taking the zeolite–apatite reaction one step further, the National Aeronautics and Space Administration prepared a substrate consisting of a specially cation-exchanged clinoptilolite and a synthetic apatite containing essential trace nutrients for use as a plant-growth medium in Shuttle flights (Fig. 7). This formulation may well be the preferred substrate for vegetable production aboard all future space missions (82) and in commercial green houses. Aquaculture. Natural zeolites can play three roles in aquaculture: (i) to remove ammonium from hatchery, transport, and aquarium waters; (ii) to generate oxygen for aeration systems in aquaria and transport; and (iii) to supplement fish rations. Zeolite cation exchange removes NH4+ from recirculating hatchery waters produced by the decomposition of excrement and/or unused food, much as NH4+ is removed from municipal sewage effluent (83, 84 and 85). Phillipsite from the Neapolitan Yellow Tuff was used to remove NH4+ from the effluent from saline water in shrimp-culture tanks (86). Although zeolites are not as successful with saline-water as they are with freshwater effluents, these results suggest limited applications in seawater systems. The NH3 content of tanktruck waters was reduced by such a system during the transport of live fish (87, 88 and 89), and Jungle Laboratories (Comfort, TX) developed a technique of adding natural zeolites to plastic bags containing tropical fish to take up NH3 generated during transport. In the U.S., at least four brands of granular clinoptilolite are currently on the market for use in filters in home aquaria, and fist-size blocks of zeolitic tuff have been sold as decorative (and ammonium-removing) additions to hobbyists' fish tanks. The physiological similarity of fish and poultry suggests that the results achieved by feeding natural zeolites to chickens might be duplicated in fish at a considerable savings in cost. Several authors have reported that small additions of clinoptilolite (2%) in the rations of trout resulted in ≤10% improvement in biomass production, with no apparent ill effects on the fish (90, 91 and 92). Animal-Waste Treatment. Natural zeolites are potentially capable of (i) reducing the malodor and increasing the nitrogen retentivity of animal wastes, (ii) controlling the moisture content for ease of handling of excrement, and (iii) purifying the methane gas produced by the anaerobic digestion of manure. Several hundred tonnes of clinoptilolite is used each year in Japanese chicken houses; it is either mixed with the droppings or packed in boxes suspended from the ceilings (K. Torii, unpublished work). A zeolite-filled air scrubber was used to improve poultry-house environments by extracting NH3 from the air without the loss of heat that accompanies ventilation in colder weather (93). Granular clinoptilolite added to a cattle feedlot (2.44 kg/m2) significantly reduced NH3 evolution and odor compared with untreated areas (94). As more and more animals and poultry are raised close to the markets to reduce transportation costs, the need to decrease air pollution in areas of large population becomes ever more apparent.

FIG. 9. Footwear and garbage-can zeolite deodorization products [Reproduced with permission from ref. 106 (Copyright 1997, AIMAT)].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3469

Other Applications Consumer Products. Numerous natural zeolite-containing consumer products have come on the market in the U.S., Japan, Hungary, Cuba, and Germany, chiefly as deodorizing agents and as pet litters to take up water and odor-causing NH3 from animal urine (Fig. 8). A horsestall refreshener consisting solely of crushed clinoptilolite also has been sold in the U.S., and similar materials have been used for years at major livestock shows to decrease odors in barns and stalls. These applications depend mainly on the ability of clinoptilolite to exchange NH4+ from aqueous solution, thereby preventing the release of NH3 into the atmosphere. Other natural zeolite deodorization products have been marketed to remove malodors from shoes, boots, and athletic footwear (e.g., Stinky Pinkys and Odor Zappers) and from garbage cans, refrigerators, and walk-in coolers (ref. 95 and 96; Fig. 9). The mechanism by which these products have achieved success is not well understood, but it is probably a surface-adsorption phenomena on the zeolite. In Russia, an unusual product (Zeo-Light) has been developed, consisting of zeolite-filled pillows and pads for funeral caskets to eliminate malodors. Medical Applications. Natural phillipsite and certain synthetic zeolites were found to be effective filter media to remove NH4+ from the dialysate of kidney patients during hemodialysis, thereby allowing the cleansed saline solution to be used repeatedly in recirculating home and portable dialysis systems (97, 98). Zeolites, especially the natural varieties, are substantially less expensive than the zirconium phosphate ion exchanger currently used. In Cuba, inexpensive, indigenous natural zeolites are being studied as buffers to reduce stomach acidity and to treat stomach ulcers (99). External application of zeolite powder has been found to be effective in the treatment of athlete's foot (100) and to decrease the healing time of wounds and surgical incisions (101, 102). Although no systematic study has been made, anecdotal information from three mining operations in the United States indicates that the cuts and scrapes of mine and mill workers exposed to on-the-job zeolite dust heal remarkably quickly. In Cuba, it is common practice to dust the cuts of horses and cows with clinoptilolite to hasten the healing process. The bactericidal reactivity of natural zeolites is an untouched field. ROLE OF MINERAL SCIENTISTS Mineral scientists have played leading roles in the development of applications for natural zeolites, e.g., Ames (41) studied the cationexchange properties of clinoptilolite and its use in treating municipal sewage and nuclear waste waters; Stojanovic (10) and Bush (11) studied the expansion properties of clinoptilolite; Filizova (57) developed zeolite pills and cookies to counteract Chernobyl fallout; White and Ohlrogge (60) developed rations for ruminants fed nonprotein-nitrogen-supplemented feeds; and U.S. Geological Survey (80) and National Aeronautics and Space Administration (82) mineral scientists used zeoponic substrate in space vehicles and greenhouse soils. Although they brought to the table a broad base of scientific knowledge, most were less than expert in agriculture or animal science or in surface chemistry, colloid chemistry, or biochemistry, and they often had to prod or persuade agricultural, chemical, biological, and engineering colleagues to apply their expertise to these fascinating problems. In the future, mineral scientists investigating the use of natural zeolites (or of any of the industrial minerals mentioned above) must become more proficient in surface chemistry, cation exchange, biological reactivity, and colloid chemistry if they are to continue to contribute to the development of zeolite applications. In turn, the chemist, agronomist, animal scientist, biochemist, and engineer must become more knowledgeable about the mineral sciences. Multidisciplinary, cooperative efforts are essential if we are to understand, for example, how a zeolite functions in an animal's gastrointestinal system, how a zeolite gives rise to greater growth or healthier animals, the specific mechanism of surface adsorption of amine-treated zeolites with other organics in aqueous solution, or how untreated zeolites adsorb offending odors from footwear and other enclosed spaces. The adsorption, cation-exchange, dehydration, catalytic, and biotechnical properties of zeolitic materials must therefore become a major part of a mineral scientist's portfolio if he or she is to become a full partner in future investigations. 1. Ames, L. L., Jr. (1967) in Proceedings of the 13th Pacific Northwest Industrial Waste Conference (Washington State Univ., Pullman, WA), pp. 135–152. 2. Mercer, B. W., Ames, L. L., Jr., Touhill, C. J., Van Slyke, W. J. & Dean, R. B. (1970) J. Water Pollut. Control Fed. 42, R95–R107. 3. Mumpton, F. A. & Ormsby, W. C. (1976) Clays Clay Miner. 24, 1–23. 4. Mumpton, F. A. (1973) Am. Miner. 68, 287–289. 5. Mumpton, F. A. (1973) Ind. Miner. (London) 73, 30–45. 6. Norin, E. (1955) Geol. Rundsch. 43, 526–534. 7. Sersale, R. (1958) Rend. Accad. Sci. Fis. Mat. Naples 25, 181–207. 8. Mielenz, R. C., Green, K. T. & Schlette, N. C. (1951) Econ. Geol. 46, 311–328. 9. Drury, F. W. (1954) Calif. Miner. Inf. Serv. 7(10), 1–6. 10. Stojanovic, D. (1972) Proc. Serbian Geol. Soc. for 1968–70, 9–20. 11. Bush, A. L. (1974) Minutes: 25th Annual Meeting (Perlite Institute, Colorado Springs, CO), p. 534 (abstr.). 12. Belitsky, I. A., Gorbunov, A. V., Kazantseva, L. K. & Fursenko, B. A. (1995) Russian Patent RU 2,033,982. 13. Gayoso Blanco, R. A., De Jongh Caula, E. & Gil Izquierdo, C. (1993) in Zeolites '91: Memoirs of the 3rd International Conference on the Occurrence, Properties & Utilization of Natural Zeolites, Havana, 1991, eds. Rodriguez Fuentes, G. & Andres Gonzales, J. (Int. Conf. Center, Havana, Cuba), pp. 203–207. 14. Gayoso, R. & Gil, C. (1994) in Proceedings of the 5th International Symposium on Ferrocement, eds. Nedwell, P. J. & Swamy, R. N. (Spon, London), pp. 141–150. 15. Butterfield, O. R. & Borgerding, J. (1981) Tahoe–Truckee Sanitation Agency Internal Report (Tahoe–Truckee Sanitation Agency, Truckee, CA). 16. Kalló, D. (1995) in Natural Zeolites '93: Occurrence, Properties, Use, eds. Ming, D. W. & Mumpton, F. A. (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 341–350. 17. Sims, R. C. (1972) Environ. Sci. Eng. Notes 9, 2–4. 18. Sims, R. C. & Little, L. W. (1973) Environ. Lett. 4, 27–34. 19. Liberti, L., Lopez, A., Amicarelli, V. & Boghetich, G. (1995) in Natural Zeolites '93: Occurrence, Properties, Use, eds. Ming, D. W. & Mumpton, F. A. (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 351–362. 20. CH2M-Hill (1975) Report for Board of Water Commissioners, Denver, Colorado (Cavvalis, OR). 21. Heaton, R. (1972) Water Wastes Eng. June, 32–34. 22. Rothberg, M. R., Work, S. W. & Lauer, W. C. (1981) in Municipal Wastewater for Agriculture, eds. D'itri, F. M., Aguirre Martinez, J. & Athie Lambarri, R. R. (Academic, New York), pp. 87–102. 23. McNair, D. R., Sims, R. C. & Grenney, W. J. (1986) Proc. Annu. Conf. Am. Water Works Assoc. (American Water Works Association, New York). 24. McNair, D. R., Sims, R. C., Sorensen, D. L. & Hulbert, M. (1987) J. Am. Water Works Assoc. 79(12), 74–81. 25. Hlavay, J. (1986) Hidrol. Kozl. 66, 348–355. 26. Hlavay, J., Inczedy, J., Földi-Polyak, K. & Zimonyi, M. (1988) in Occurrence, Properties & Utilization of Natural Zeolites, eds. Kalló, D. & Sherry, H. S. (Akademiai Kiado, Budapest), pp. 483–490. 27. Tarasevich, Y. I. (1984) in SlovZeo '84: Conference on the Study and Use of Natural Zeolites, Vysoke Tatry, Czechoslovakia, Part 2 (Czech. Sci. Tech. Soc., Kosice, Slovakia), pp. 76–81. 28. Tarasevich, Y. I. (1993) in Program & Abstracts: Zeolite '93: 4th International Conference on the Occurrence, Properties & Utilization of Natural Zeolites, Boise, Idaho, (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 199–201 (abstr.). 29. Union Carbide Corporation (1962) Linde Molecular Sieve Bulletin F-1617 (Union Carbide). 30. NRG Company (1975) Brochure, NRG NuFuel Company (NRG, Los Angeles).

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

LA ROCA MAGICA: USES OF NATURAL ZEOLITES IN AGRICULTURE AND INDUSTRY

3470

31. Tamura, T. (1971) British Patent 1,258,417. 32. Tamura, T. (1972) German Patent 2,214,820. 33. Dominé, D. & Häy, L. (1968) in Molecular Sieves (Soc. Chem. Ind., London), pp. 204–216. 34. Gao, F., Cadena, F. & Peters, R. W. (1991) in Proceedings of the 45th Purdue Industrial Waste Conference (Lewis, Chelsea, MI), pp. 509–516. 35. Cadena, F. & Cazares, E. (1995) in Natural Zeolites '93: Occurrence, Properties, Use, eds. Ming, D. W. & Mumpton, F. A. (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 309–324. 36. Bowman, R. S., Haggerty, G. M., Huddleston, R. G., Neel, D. & Flynn, M. M. (1995) ACS Symp. Ser. 594, 54–64. 37. Chen, N. Y. (1971) U.S. Patent 3,630,066. 38. Ohtani, S., Iwamura, T., Sando, K. & Matsumura, K. (1972) Japanese Patent 72046, 667. 39. Papp, J., Kalló, D. & Schay, G. (1971) J. Catal. 23, 168–182. 40. Kalló, D. (1988) in Occurrence, Properties & Utilization of Natural Zeolites, eds. Kalló, D. & Sherry, H. S. (Akademiai Kiado, Budapest), pp. 601–624. 41. Ames, L. L., Jr. (1959) U.S. At. Energy Comm. Unclassified Report (AEC, Washington, DC) Publ. No. HY-62607. 42. Mercer, B. W., Ames, L. L., Jr., & Smith, R. W. (1970) Nucl. Appl. Technol. ECL-152, 62–69. 43. Wilding, M. W. & Rhodes, D. W. (1965) U.S. At. Energy Comm. Document (AEC, Washington, DC) Publ. No. IDO-14657. 44. Daiev, C., Delchev, G., Zhelyazkov, V., Gradev, G. & Simov, S. (1970) in International Atomic Energy Agency, Vienna, Symposium on the Management of Low- & Intermediate-Level Radioactive Wastes (Int. At. Energy Agency, Vienna), pp. 739–746. 45. IAEA (1972) Tech. Rep. Ser. IAEA 136(68), 97–98. 46. Dyer, A. & Keir, D. (1984) Zeolites 4, 215–221. 47. Robinson, S. M., Kent, T. E. & Arnold, W. D. (1995) in Natural Zeolites '93: Occurrence, Properties, Use, eds. Ming. D. W. & Mumpton, F. A. (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 579–586. 48. Hofstetter, J. K. & Hite, G. H. (1983) Sep. Sci. Technol. 18, 1747–1764. 49. British Nuclear Technology (1987) British Nuclear Technology Paper 9 (Risley, Warrington, U.K.). 50. Grant, D. C., Skirba, M. C. & Saha, A. K. (1987) Environ. Prog. 6(2), 104–109. 51. Nishita, H. & Haug, R. M. (1972) Soil Sci. 114, 149–157. 52. Robinson, W. L. & Stone, G. R. (1988) Bikini Atoll Rehabilitation Committee Summary Report No. 6, (BARC, Berkeley, CA), Appendix A, A1–A48. 53. Shenbar, M. A. & Johanson, K. J. (1992) Sci. Total Environ. 113, 287–295. 54. Forberg, S., Jones, B. & Westermark, T. (1989) Sci. Total Environ. 79, 37–41. 55. Firsakova, S. K., Grebenchchikova, N. V., Timofeev, S. F. & Novik, A. A. (1992) Dokl. Vses. Akad. Skh. Nauk ira. V. I. Lenina (3), 25–27. 56. Phillippo, M., Gvozdanovic, S., Gvozdanovic, D., Chesters, J. K., Paterson, E. & Mills, C. F. (1988) Vet. Rec. 122, 560–563. 57. Filizova, L. (1993) in Program & Abstracts: Zeolite '93: 4th International Conference on the Occurrence, Properties, and Utilization of Natural Zeolites, Boise, Idaho (Int. Comm. Natl. Zeolites, Brockport, NY), pp. 88–90 (abstr.). 59. Onagi, T. (1966) Rep. Yamagata Stock Raising Inst. 11–22. 59. Kondo, N. & Wagai, B. (1968) Yotonkai 1–4. 60. White, J. L. & Ohlrogge, A. J. (1974) Canadian Patent 939,186. 61. Pond, W. G. (1995) in Natural Zeolites '93: Occurrence, Properties, Use, eds. Ming, D. W. & Mumpton, F. A. (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 449–457. 62. Mumpton, F. A. & Fishman, R. H. (1977) J. Anim. Sci. 45, 1188–1203. 63. Shell, T. C., Lindemann, M. D., Kornegay, E. T. & Blodgett, D. J. (1992) Va. Polytech. Inst. Anim. Sci. Rept. 10, 6–8. 64. Dvorak, M. (1989) Vet. Med. (Prague) 34, 307–316. 65. Galabov, I., Hristov, B. & Simeonov, K. (1991) in Program & Abstracts, Zeolites '91, (Int. Conf. Center, Havana, Cuba), p. 147 (abstr.). 66. Lon-wo, E., Zaldivar, V. & Margolles, E. (1993) Cuban J. Agric. Sci. 27, 199–204. 67. Kovac, G., Bartko, P., Vrzgula, L., Reichel, R., Nistiar, F., Mojzis, J. & Seidel, H. (1995) in Natural Zeolites '93: Occurrence, Properties, Use, eds. Ming, D. W. & Mumpton, F. A. (Int. Comm. Nat. Zeolites, Brockport, NY). pp. 459–466. 68. Minato, H. (1968) Koatsu Gasu 5, 536–547. 69. Torii, K. (1978) in Natural Zeolites: Occurrence, Properties, Use, eds. Sand, L. B. & Mumpton, F. A. (Pergamon, Elmsford, NY) 441–450. 70. Van Bo, N. (1988) Soviet Agric. Sci. (12), 62–64. 71. Mazur, G. A., Medvid, G. K. & Grigorta, T. I. (1984) Pochvovedenie (10), 73–78. 72. Hershey, D. R., Paul, J. L. & Carson, R. M. (1980) HortScience 15, 87–89. 73. Lewis, M. D., Moore, F. D., III, & Goldsberry, K. L. (1984) in Zeo-Agriculture: Use of Natural Zeolites in Agriculture & Aquaculture, eds. Pond, W. G. & Mumpton, F. A. (Westview, Boulder, CO), pp. 105–112. 74. Petrovic, A. M. (1990) Golf Course Manage. 58(11), 92–93. 75. Petrovic, A. M. (1993) in Program & Abstracts: Zeolite '93: 4th International Conference on Occurrence, Properties & Utilization of Natural Zeolites, Boise, Idaho (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 162–164 (abstr.). 76. Huang, Z. T. (1992) Ph.D. thesis (Cornell Univ., Ithaca, NY). 77. Petrov, G. S., Petkov, I. A., Etropolski, H. I., Dimitrov, D. N., Popov, N. N. & Uzunov, A. I. (1982) U.S. Patent 4,337,078. 78. Rodriguez Fuentes, G. (1993) in Zeolites '91: Memoirs of the 3rd International Conference on the Occurrence, Properties & Utilization of Natural Zeolites, Havana, 1991, Part 1, eds. Rodriguez Fuentes, G. & Andres Gonzales, J. (Int. Conf. Center, Havana, Cuba), pp. 25–33. 79. Ivanova, T. N., Bercovich, Y. A., Mashinsky, A. L. & Meleshko, G. I. (1993) Acta Astronaut. 29, 639–644. 80. Lai, T. M. & Eberl, D. D. (1986) Zeolites 6, 129–132. 81. Garcia Hernandez, J. E., Notario del Pino, J. S., Arteaga Padron, I. J. & Gonzalez Martin, M. M. (1992) Agrochimica 32, 1–11. 82. Ming, D. W., Barta, D. J., Golden, D. C., Galindo, C., Jr., & Henninger, D. L. (1995) in Natural Zeolites '93: Occurrence, Properties, Use, eds. Ming, D. W. & Mumpton, F. A. (Int. Comm. Nat. Zeolites, Brockport, NY), pp. 505–513. 83. Sims, R. C. (1977) M.S. thesis (Washington State Univ., Pullman, WA ). 84. Bruin, W. J., Nightingale, J. W. & Mumaw, L. (1980) Seattle Aquarium Tech. Rep. 9, 1–17. 85. Smith, C. F., Piper, R. G. & Tisher, H. R. (1981) U.S. Fish Wildl. Serv. Bozeman Inf. Leaflet 20, 1–17. 86. Ciambelli, P., Corbo, P., Lumare, F. & Porcelli, C. (1984) in Zeo-Agriculture: Use of Natural Zeolites in Agriculture & Aquaculture, eds. Pond, W. G. & Mumpton, F. A. (Westview, Boulder, CO), pp. 251–258. 87. Amend, D. F., Croy, R. R., Goven, R. A., Johnson, K. A. & McCarthy, D. H. (1982) Trans. Am. Fish. Soc. 111, 603–611. 88. Clary, S. D. (1978) Salmonid 2, 12–14. 89. Burkstaller, J. (1968) M.S. thesis (Mew Mexico State Univ., LasCrucas, NM). 90. Leonard, D. W. (1979) Trans. Soc. Min. Eng. AIME 79, 380–400, preprint. 91. Misaila, C., Misaila, E. R., Marton, A. & Bucar, N. (1990) Lucrarile S. C. R. Piscicola-Iasi 1, 223–230. 92. Buzmakov, G. T. & Arsenov, O. A. (1992) Rybn. Khoz. (Moscow) (6), 2–7. 93. Koelliker, J. K., Miner, J. R., Hellickson, M. L. & Nakaue, H. S. (1980) Trans. Am. Soc. Agric. Eng. 23, 157–161. 94. Miner, J. R. & Stroh, R. C. (1976) Trans. Am. Soc. Agriv. Eng. 19, 553–558. 95. Schwarz, J. & Wagner, C. (1994) Sci. Tech. Froid 1, 541–550. 96. Bermas, E. M. (1995) International Patent Appl. WO 95 15, 187. 97. Andersson, S., Grenthe, I. & Jonsson, E. (1975) German Patent 2,512,212. 98. Ash, S. R. (1986) U.S. Patent 4,581,141. 99. de Armas, M., Fernandez, G., Perez de Alejo, L. & Rodriguez, B. D. (1991) in Program & Abstracts: Zeolites '91 (Int. Conf. Center, Havana, Cuba), p. 188 (abstr.). 100. Lopez, D. Z. (1991) in Program & Abstracts: Zeolites '91, (Int. Conf. Nat. Zeolites, Havana, Cuba), p. 187 (abstr.). 101. Maeda, K. (1989) Eur. Patent Appl. EP 298, 726. 102. Ikeganmi, K. & Koide, M. (1993) Japanese Patent JP 05,285,209. 103. Breck, D. W. (1974) Zeolite Molecular Sieves (Wiley, New York). 104. Meier, W. M. & Olson, D. H. (1971) Adv. Chem. Ser. 101, 155–170. 105. Aiello, R. (1995) in Natural Zeolites '93, eds. Ming, D. W. & Mumpton, F. A. (Int. Comm. on Nat. Zeolites, Brockport, NY), pp. 589–602. 106. Mumpton, F. A. (1997) in Proceedings of the 3rd National Congress of the Association of Italian Materials Engineers, ed. Colella, C. (AIMAT, Naples, Italy), pp. xxxi-lxiv.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3471

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3471–3478, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA.

Synthetic zeolites and other microporous oxide molecular sieves

JOHN D. SHERMAN PNAS is available online at www.pnas.org.

Research and Development Department, UOP LLC, 25 East Algonquin Road, Des Plaines, IL 60017-5017 ABSTRACT Use of synthetic zeolites and other microporous oxides since 1950 has improved insulated windows, automobile airconditioning, refrigerators, air brakes on trucks, laundry detergents, etc. Their large internal pore volumes, molecular-size pores, regularity of crystal structures, and the diverse framework chemical compositions allow “tailoring” of structure and properties. Thus, highly active and selective catalysts as well as adsorbents and ion exchangers with high capacities and selectivities were developed. In the petroleum refining and petrochemical industries, zeolites have made possible cheaper and lead-free gasoline, higher performance and lower-cost synthetic fibers and plastics, and many improvements in process efficiency and quality and in performance. Zeolites also help protect the environment by improving energy efficiency, reducing automobile exhaust and other emissions, cleaning up hazardous wastes (including the Three Mile Island nuclear power plant and other radioactive wastes), and, as specially tailored desiccants, facilitating the substitution of new refrigerants for the ozone-depleting chlorofluorocarbons banned by the Montreal Protocol. Relationships of Synthetic Zeolites to Natural Zeolites and Other Minerals. Only 6 of the >63 natural zeolites commonly occur in large beds: analcime (ANA),* chabazite (CHA), clinoptilolite (HEU), erionite (ERI), mordenite (MOR), and phillipsite (PHI) (1); ferrierite (FER) occurs in a few large beds. Each of the seven also has been synthesized, but only mordenite and ferrierite are manufactured in large quantity. Significantly, synthetic mordenite has large pores whereas natural mordenite has small pores (2). Besides mordenite and ferrierite, the principal synthetic (aluminosilicate) zeolites in commercial use are Linde Type A (LTA), Linde Types X and Y (Al-rich and Si-rich FAU), Silicalite-1 and ZSM-5 (MFI), and Linde Type B (zeolite P) (GIS). Other commercially available synthetic zeolites include Beta (BEA), Linde Type F (EDI), Linde Type L (LTL), Linde Type W (MER), and SSZ-32 (MTT). All are aluminosilicates or pure silica analogs. Recently, new nonaluminosilicate, synthetic molecular sieves became available commercially. They include aluminophosphates (family of AlPO4 structures); silicoaluminophosphates (SAPO family); various metal-substituted aluminophosphates [MeAPO family, such as CoAPO-50 (AFY)]; and other microporous framework structures, such as crystalline silicotitanates.† Most current commercial applications use aluminosilicate zeolites or their modified forms. Undoubtedly, commercial uses both for zeolites and other molecular sieves will continue to grow. Development of Synthetic Zeolites and Other Microporous Oxides. The first zeolite mineral (stilbite) was described in Sweden by Baron Cronstedt in 1756 (3, 4 and 5). Highlights of the history of adsorption studies of zeolites were reviewed by Breck (6). By 1926, the adsorption characteristics of chabazite were attributed to tiny pores (<5 Å in diameter) that allowed small molecules to enter but excluded larger ones: hence, the term “molecular sieve” (7). By 1945, Barrer classified zeolite minerals into three classes depending on the size of the molecules adsorbable rapidly, slowly, or not appreciably at room temperature or above (8, 9). However, zeolites did not find any significant commercial use until synthetic zeolites were discovered and developed (large, mineable deposits of natural zeolites were not discovered until the late 1950s). Barrer's 1948 synthesis of small-port mordenite at high temperatures and pressures heralded the era of synthetic zeolites (10). From 1949 through the early 1950s, the commercially significant zeolites A, X, and Y were discovered by Milton and Breck at the Tonawanda, New York, laboratories of the Linde Air Products Division of Union Carbide Corporation. These zeolites were synthesized from readily available raw materials at much lower temperature and pressure than used earlier. Many of the new synthetic zeolites had larger pore size than most of the known natural zeolites, allowing applications involving larger molecules. In addition, many had larger pore volume, giving higher capacity. In 1953, Linde Type A zeolite became the first synthetic zeolite to be commercialized as an adsorbent to remove oxygen impurity from argon at a Union Carbide plant (11). Synthetic zeolites were introduced by Union Carbide as a new class of industrial adsorbents in 1954 and as hydrocarbon-conversion catalysts in 1959. New zeolites and new uses appeared steadily through the 1960s. An explosion of new molecular sieve structures and compositions occurred in the 1980s and 1990s from the aluminosilicate zeolites to the microporous silica polymorphs to the microporous aluminophosphate-based polymorphs and metallo-silicate compositions (12). Molecular sieves now serve the petroleum refining, petrochemical, and chemical process industries as selective catalysts, adsorbents, and ion exchangers. Many zeolites can be synthesized with SiO2 higher or lower than in nature for the same framework type. Higher SiO2 generally gives greater hydrothermal stability, stronger-acid catalytic activity, and greater hydrophobicity as adsorbents. Conversely, lower SiO2 gives greater cation exchange capacity and higher adsorbance for polar molecules. Controlling the synthesis process optimizes a zeolite for different applications.

*The three-letter International Zeolite Association Structure Commission code for the framework topology of each zeolite is given in parentheses at the first mention of that zeolite in this paper (the full list is at http://www-iza-sc.csb.yale.edu/IZA-SC/). †The molecular sieves of AlPO4s, SAPOs, MeAPOs, etc., were discovered in the 1980s by scientists in the Tarrytown, NY laboratories of Union Carbide Corporation's Catalysts, Adsorbents and Process Systems (CAPS) group. In 1988, Union Carbide Corporation's Catalysts, Adsorbents and Process Systems and the Process Division of UOP of AlliedSignal merged to form a partnership company, called UOP, which is jointly owned by AlliedSignal and Union Carbide. UOP LLC has continued to develop both the materials and their applications. Abbreviations: SAPO, silicoaluminophosphate; 8R, eight-ring; EB, ethylbenzene; PSA, pressure swing adsorption; VSA, vacuum swing adsorption; tpd, U.S. tons of O2 per day; 3D, three-dimensional.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3472

Many synthetic zeolites have framework topologies not found to date among the natural zeolites. The natural zeolite faujasite has the same framework (FAU) and similar framework composition to the Type Y synthetic zeolite but is rare in nature. Where both natural and synthetic forms of the same zeolite are available in commercial quantity, the variable phase purity of the natural zeolite and the chemical impurities, which are costly to remove, can make the synthetic zeolite more attractive for specific applications. Conversely, where uniformity and purity are not important, the cheapness of a natural zeolite may favor its use. Hence, natural and synthetic zeolites seldom compete for the same applications. Structure and Properties of Synthetic Molecular Sieves. Zeolites have the chemical formula M2/nOAl2O3·xSiO2·yH2O, where the charge-balancing nonframework cation M has valence n, x is 2.0 or more, and y is the moles of water in the voids. The Al and Si tetrahedral atoms, or T-atoms, form a three-dimensional (3D) framework of AlO4 and SiO4 tetrahedra linked together by shared oxygen ions. Although an SiO4 tetrahedra is charge-balanced, an AlO4 tetrahedra has a negative charge balanced by a positive charge on M. Related pure SiO2 frameworks, such as silicalite-1 (MFI), are charge-balanced and do not need non-framework cations. Variants involve Ge substitution for Si in the framework or involve substitution of Fe, Co, Mn, Zn, Ti, or Mg for Al. In the related aluminophosphates (AlPO4), each negatively charged AlO4 tetrahedron is balanced by a positively charged PO4 tetrahedron, and nonframework cations are not needed. Still other variants include the silicoaluminophosphate (SAPO) structures in which Si substitutes some P in the AlPO4, framework; each added Si needs a nonframework cation to balance the charge on the framework. The pore geometry and volume in a specific microporous oxide are determined by the specific topology of the particular 3D framework. The lower the T-atom density per volume of the zeolite crystal, the higher the void fraction inside the crystal. The void fraction is 50% for NaX and 47% for NaA. The size of the largest pore in a zeolite is determined by the number of oxygen ions rimming the pore and its shape; e.g. a planar, circular eight-ring (8R) pore rimmed by eight oxygen ions has a diameter of 4.1 Å, as in Linde Type A zeolite, whereas the elliptical 8R pore of NaP zeolite (GIS) is 4.5 × 3.1 Å. Applications in separation and purification processes often used the ability of zeolites and other molecular sieves to exclude molecules too large to enter the pores and admit smaller ones. Similarly, shape-selective catalysis takes advantage of the ability of the pores to favor the admission of smaller reactant molecules, the release of smaller reaction product molecules, or the restriction of the size of transition-state complexes inside the micropores of the zeolite (13). Petroleum Refining Processes for the Production of Fuels Catalytic Cracking. The prime goal in petroleum refining is efficient conversion of crude oil into high-quality fuel components. Desired fuel fractions in order of increasing molecular weight are gasoline, aviation jet fuel, and diesel fuel. Gasoil and asphalt, with even higher molecular weight, are most often further processed by thermal cracking, catalytic cracking (to make gasoline) (14), and catalytic hydrocracking (to make jet fuel). A lower-boiling fraction, light straight-run naphtha, rich in pentanes and hexanes and some butane, is further processed by catalytic hydroisomerization. Strong acid catalytic activity of X and Y zeolites was discovered in 1957 by Rabo and was related to their crystallinity (15). This discovery laid the basis for zeolites in cracking, hydrocracking, and isomerization of hydrocarbons. From the early 1960s on, use of synthetic zeotites in catalysis and in related adsorption separation processes has dramatically transformed petroleum refining by vastly increasing the yield of high-quality fuels and reducing capital and operating costs, energy requirements, and adverse environmental impact. Zeolites also played a major role in allowing the efficient reformulation of gasoline to the present lead-free gasoline. In modern petroleum refineries in the world, gasoil and other heavier fractions from the crude oil fractionation unit are fed to fluid catalytic cracking units, which use small, fluidizable catalyst particles containing Type Y zeolite or other zeolites, or to hydrocracking units, which use fixed beds of larger catalyst particles also containing zeolites. The fluid catalytic cracking and hydrocracking units convert highermolecular-weight hydrocarbons to lighter ones suitable for gasoline, light fuel oils, olefins, and other uses. In both fluid catalytic cracking and hydrocracking, zeolite catalysts provide vastly superior combinations of strong acid catalytic sites, uniformity of pore structure, and stability, all of which provide improved selectivity, yield, durability, and cost over nonzeolite alternatives. In addition, these zeolites have provided much higher yield of gasoline and other high-quality fuels per barrel of crude oil, significantly reducing crude oil imports to the U.S. (>400 million barrels a year) and to other countries. Hydrocracking. The early 1960s saw increasing demand for high-octane gasoline for the high-compression-ratio engines in new highperformance cars. Demand also grew for diesel fuel for diesel-electric locomotives and low-freeze-point jet fuel. These needs were met by rapid growth in hydrocracking of the more-refractory crude fractions that were not converted to gasoline and lighter products in the catalytic cracking units. This growth was accompanied by the pioneering development by Roland Hanford at Union Oil, now Unocal, of new, zeolitebased hydrocracking catalysts with dramatically improved activity and selectivity. Hydrocracking grew rapidly in the 1960s and 1970s inside and then later outside the U.S. Worldwide hydrocracking capacity should grow from ≈2.5 million barrels per day in 1990 to ≈3.5 million in 2000 (16). In hydrocracking, hydrocarbon molecules and hydrogen gas pass over the zeolite catalyst, which converts higher-molecular-weight petroleum fractions to lower-molecular-weight fuels (17). For example, UOP's Unicracking process (developed jointly by the Molecular Sieve Department of Union Carbide, now part of UOP, and Unocal) uses base- or noble-metal hydrogenation-activity promoters impregnated on combinations of zeolite- and amorphous-aluminosilicates for cracking activity (18). The specific metals chosen and the proportions of the metals, zeolite, and nonzeolite aluminosilicates are optimized for the feedstock and desired product balance. The Isocracking process of Chevron also uses hydrocracking catalysts, some containing zeolites to increase the cracking function of these dual-function catalysts (19). The zeolites most frequently used in commercial hydro-cracking catalysts are partially dealuminated and low-sodium, or high-silica, Type Y zeolites in hydrogen or rare-earth forms. Other zeolites and mixtures of zeolites also are used. The zeolites often are imbedded in a highsurface-area amorphous matrix, which serves as a binder. The metals can reside inside the zeolite and on the amorphous matrix. Catalytic Dewaxing. Catalytic dewaxing yields various grades of lube oils and fuel components suitable for extreme winter conditions. Paraffinic (waxy) components, which precipitate out at low temperatures, are removed. In the UOP Catalytic Dewaxing process, the first stage saturates olefins and desulfurizes and denitrifies the feed via hydrotreating (20). In the second stage, a dual-function, non-noble-metal zeolite catalyst selectively adsorbs and then selectively hydrocracks the normal and near-normal long-chain paraffins to form shorter-chain (nonwaxy) molecules. Alternatively, as in the recently commercialized Chevron Isodewaxing process, the dewaxing results from isomerizing the linear paraffins to branched paraffins by using a SAPO-11 molecular sieve catalyst containing platinum (21, 22). Light Paraffin Hydroisomerization. Lead was added to gasoline to increase its octane number, especially for vehicles, introduced in the early 1960s, that had modern high-compression-

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3473

ratio, high-performance engines. The subsequent U.S.-legislated reduction of lead in gasoline required increased use of the catalytic hydroisomerization of the light straight-run naphtha fraction mentioned earlier. Some versions of UOP's hydroisomerization processes use highly active zeolite-based, Pt-containing hydroisomerization catalysts, such as UOP I-7, which contains modified synthetic (large-port) mordenite. In the presence of hydrogen at moderate conditions, such catalysts optimize isomerization and minimize hydrocracking (23). Linear paraffins in the feed convert to branched paraffins with higher octane number. The SudChemie HYSOPAR catalyst used in CEPSA's CKS ISOM process also uses a zeolite for the hydroisomerization of light naphtha.‡ To further increase octane level, products from a hydroisomerization unit can be sent to the Molex process, where the remaining loweroctane, linear paraffins are separated from the other compounds by using a zeolite adsorbent and a liquid desorbent; the Molex process is an example of UOP's Sorbex simulated-moving-bed technology (24). The extracted linear paraffins are recycled to the hydroisomerization unit, and the remaining higher-octane fraction is recovered for gasoline blending. The combination of the hydroisomerization and Molex processes boosts the research octane number of a typical feed from 68–70 to 89–92. Alternatively, if a refinery can use the linear paraffins, it need not recycle them to a hydroisomerization unit. For example, the paraffins may be added to the feed of an ethylene steam cracker, thus increasing the efficiency of the cracker and leading to lower energy consumption and a purer product. Linear paraffins also are used as intermediaries in some food processing. UOP's once-through zeolitic isomerization process (formerly known as the Shell Hysomer process) also uses a strongly acidic zeolite with a noble metal to hydroisomerize the light naphtha (25). Refiners with idle catalytic reformers or hydrotreaters can convert this equipment to use this process. To achieve higher octane levels, UOP's TIP total isomerization process uses the once-through isomerization process combined with UOP's IsoSiv process, which uses size-selective zeolite adsorption of the unreacted linear paraffins so that they can be recycled and converted to extinction (32). Both the TIP and IsoSiv processes originally were developed at Union Carbide's Molecular Sieve Department, now part of UOP. Petrochemicals Processing for Aromatics Production and Derivatives Ethylbenzene Synthesis. Styrene monomer, made by dehydrogenating ethylbenzene (EB), is the basic chemical for all polystyrene products. Ethylbenzene is made by using various catalysts to alkylate benzene with ethylene. Until 1980, nearly all EB was produced by liquidphase alkylation reactions using aluminum chloride catalyst. In 1980, vapor-phase alkylation using a heterogeneous catalyst was introduced to eliminate many problems of waste disposal and special metallurgy involving aluminum chloride. In 1990, liquid-phase zeolitic technology began to replace the Mobil/Badger vapor-phase process, based on ZSM-5 zeolite. The 1990 Lummus/UOP ethylbenzene liquid-phase process, using highly stable, poison-resistant zeolite catalysts manufactured by UOP, operates at low benzene-to-olefin ratio and high selectivity to EB.§ The UOC-4120 catalyst from UOP, used initially for both alkylation of benzene with ethylene and transalkylation of polyethylbenzenes and benzene to produce more EB, operated successfully for 7 years, with >5 million metric tons of capacity installed or ordered. The extremely low xylene content of the EB product permits the production of the highestpurity styrene monomer and lowers the costs in the styrene production unit. Current designs use the EBZ-100 catalyst for transalkylation and the EBZ-500 catalyst for alkylation. The new Mobil/Badger-Raytheon EBMax process commercialized in 1995 uses an MCM-22 (MWW) zeolite catalyst for liquid-phase alkylation; Mobil's TRANS-1 modified MFI catalyst for vapor-phase transalkylation of polyethylbenzene and cracking of C6 and C7 naphthenes; and TRANS-4 catalyst for liquid-phase transalkylation of polyethylbenzenes (26). Cumene Synthesis. More than 95% of the 7 million metric tons per year of cumene is used worldwide as the principal chemical for production of phenol and its acetone byproduct. The phenol yields phenolic resins, bisphenol-A, caprolactam, and other products. Phenolic resins are used extensively to bond plywood and composition board. Both phenol and acetone are used increasingly in the production of polymers such as epoxy, polycarbonate resins, and nylon-6. Most cumene is made by alkylating benzene with propylene over an acid catalyst, mostly solid phosphoric acid and minor AlCl3. Recent awareness of the negative environmental impact of spent-catalyst disposal has spurred a search for more benign alternatives. The world's leading technology (90% open market) for producing cumene is the UOP Catalytic Condensation process, which uses inexpensive solid phosphoric acid catalyst. Its high-purity cumene product has set the standard. However, side reactions over the solid phosphoric acid catalyst result in a 4–5% loss in cumene yield. The UOP Q-Max process, commercialized in 1996, uses the new, environmentally benign QZ-2000 zeolite catalyst for direct alkylation of benzene with propylene and incorporates a second step (transalkylation) to react the diisopropylbenzene, a byproduct of the first step, with benzene to form additional cumene.¶ It produces a higher-quality cumene product (>99.97% purity) at overall cumene yield of 99.7% and lower investment cost. The Mobil-Badger process using aluminum chloride catalyst also yields cumene. Recently, Mobil/Raytheon also developed a zeolite catalyst based on the relatively new zeolite MCM-22. Similarly, Dow and Kellogg developed “3D-DM” catalysts based on dealuminated forms of mordenite (27): Enichem, a catalyst based on zeolite Beta, and CD-Tech/Lummus, a catalytic distillation zeolite catalyst. para-Xylene Production from Mixed C8 Aromatics. Polyester fibers have revolutionized clothing. Many people in the U.S. take for granted wash-and-wear and permanent press clothing, and ironing of shirts is out of fashion. Demand has grown in developing countries because of the great comfort of fabrics made from cotton–polyester fibers blended in any proportion for any climate, as well as low cost, excellent durability, and ease of washing with little water and detergent (M. M. Sharma, personal communication). The worldwide annual production of 12–15 million tons p-xylene is expected to rise to ≈17–18 million tons in 10 years. Most p-xylene is used to make purified terephthalic acid, which is reacted with ethylene glycol to make the poly-(ethylene terephthalate), the basis of polyester fibers. The p-xylene is separated from mixed C8 aromatics (containing o-, m-, and p-xylenes and EB) by using either crystallization or adsorption processes. Since 1971, the new UOP's Parex adsorption separation process captured 60% of the worldwide p-xylene production. Another large and growing use for p-xylene is in the manufacture of poly(ethylene terephthalate) for bottles recyclable and environmentally benign. The Parex process uses the Sorbex technology mentioned earlier (28). Its critical ingredient is a special, p-xylene-selective adsorbent. Ionexchanged forms of synthetic FAU zeolite are

‡Floyd, F. M., Gilbert, M. F., Pascual, M. P. & Kohler, E., Middle East Petrotech 98: Second Middle East Refining and Petrochemicals Conference and Exhibition, September 14–16, 1998, Bahrain. §Woode, G. B., Zarchy, A. S., Morita, M. & Shinohara, K., Sud-Chemie Group 1998 International Styrene Symposium, June 14–18, 1998, Sapporo, Hokkaido, Japan. ¶Jeanneret, J., Greer, D., Ho, P., McGehee, J. & Shakir, H., 22nd Annual DeWitt Petrochemical Review, March 18–20, 1997, Houston, TX.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3474

used with desorbent liquids to recover >97% p-xylene at >99.9% purity from a raffinate containing EB and o- and m-xylenes. Xylene Isomerization. The raffinate from the Parex unit can go to an Isomar unit (29), licensed by UOP, for isomerization to a nearequilibrium mixture of xylenes, which are recycled to the Parex unit. The Isomar unit itself also uses UOP zeolite acid catalysts, such as the Ptbearing I-9 catalyst, which converts EB to xylenes, and the I-100 catalyst, which dealkylates EB to benzene. Both provide efficient EB conversion with excellent xylene retention. Disproportionation of Toluene and Transalkylation of Toluene and Trimethylbenzenes. Recent strong demand for p-xylene has begun to exceed the supply of mixed xylenes. Incorporating the Tatoray process (originated with Toray Industries in Japan and further developed and licensed by UOP) into the aromatics complex in a refinery can more than double the yield of p-xylene from a naphtha feedstock (30). The zeolite-based TA-4 catalyst has two principal functions, disproportionation of toluene into the more-valuable benzene and mixed xylenes, and transalkylation of toluene and trimethylbenzenes to mixed xylenes. The mixed xylenes then are added to the Parex unit to produce more p-xylene. p-Xylene Synthesis from Toluene. Xylenes can be produced by the zeolite-catalyzed disproportionation of toluene alone. Mobil developed the MTPX (Mobil toluene to para-xylene) process for internal use, and the licensed MSTDP (Mobil selective toluene disproportionation) process, based on ZSM-5 (MFI) “product shape-selective” zeolite catalysts (31). The toluene disproportionation generates mixed xylenes inside the catalyst, but the overall relative yield of p-xylene is greater than the thermodynamic equilibrium allows because the pxylene diffuses more rapidly out of the zeolite than do the o- and m-xylenes. UOP's recent PX-Plus process also uses a zeolite catalyst for pxylene synthesis by shape-selective disproportionation of toluene. Aromatics from Light Hydrocarbons. UOP's Cyclar process converts low-value LPG (propane, butanes) or light feedstocks containing olefins and paraffins to high-value, easily transportable, petrochemical-grade liquid aromatic products, particularly BTX (benzene, toluene, and xylenes). It uses a single galliummodified zeolite catalyst developed by BP and UOP in conjunction with UOP's CCR continuous catalytic regeneration system (32, 33). Acidic sites on the zeolite catalyze dehydrogenation, oligomerization, and cyclization. The shape-selectivity of the zeolite cavities helps promote the cyclization reactions and limits the size of the rings (34). M-Forming. Catalytic reforming produces a high-octane liquid reformate product rich in aromatics and hydrogen gas; light hydrocarbon gases, such as LPG; and C6 to C9 paraffins. Mobil's M-Forming process selectively hydrocracks linear and singly branched paraffins in gasoline reformate fractions to LPG by size-selective catalysis by using medium-pore ZSM-5 zeolite (35). Olefins produced from paraffin cracking alkylate the aromatics and also form some aromatics by oligomerization. Other Aromatics Produced by Sorbex Separations. Some other applications of the Sorbex zeolite-based simulated-moving-bed technology (36) are MX the Sorbex process, m-xylene from EB and o- and p-xylenes; the Cymex process, m- and/or p-xylene from a mixture of cymene isomers; and the Cresex process, m- and/or p-cresol from mixtures of cresol and xylenol. Petrochemicals Processing for Olefins Production Light Olefin Production by Methanol-to-Olefins (MTO) Process. Only ≈110 of the ≈2,500 billion cubic meters of natural gas produced annually is wasted (burned in flares). About 103 billion cubic meters per year of natural gas are processed to make liquefied natural gas, but at high cost. Alternatively, natural gas can be converted first to syngas (CO and H2) and then to the more-valuable, easily shipped methanol. However, the methanol market is too small for the available natural gas. The new UOP/HYDRO MTO process provides the means to efficiently convert methanol to even more valuable light olefins (ethylene, propylene, and butenes), which have large, commodity-type petrochemical markets: Ethylene and propylene represent the largest, together accounting for 120 million MTA, and growing (37). This process uses the product-shape-selective UOP MTO-100 catalyst based on a unique molecular sieve. During the late 1980s, Norsk Hydro, assisted by Sintef, started independent work, and UOP and Hydro agreed on joint development of the process, now available from UOP for commercial licensing. Norsk Hydro is running a large (0.75 metric tons/day) UOP/ HYDRO MTO demonstration plant in Porsgrunn, Norway. Olefin Isomerization. The 1990 Clean Air Act increased the demand for blendable ethers in motor fuels and created a demand for isobutene to make methyl tertiary butyl ether and for isopentene to make tertiary amyl methyl ether. In anticipation, the UOP I-500 catalyst, based on a SAPO structure, and two new processes were developed: Butesom for isobutene isomerization and Pentesom for pentene isomerization (38, 39, 40 and 41). In both processes, coke progressively accumulates on the catalyst and is periodically removed by a simple carbon burn-off in the reactor. The Lyondell IsoPlus process (42, 43) uses a ferrierite (FER) zeolite for the isomerization of olefins to isoolefins, and a Mobil patent (44) describes using a medium-pore zeolite catalyst (for example, ZSM-5) for similar applications. Oxygenates Removal Unit. Zeolite adsorbents are used in a UOP oxygenate removal unit down to >1 ppm total of trace oxygenates (e.g., DME, methanol, and methyl tertiary butyl ether) from C4 streams. Depending on the flow scheme, the C4 stream generally goes to a motor fuel alkylation (sulfuric acid or hydrofluoric acid) process or is recycled to a dehydrogenationetherification complex, which has a UOP Oleflex unit and a methyl tertiary butyl ether unit. The advantages of the oxygenate removal unit is that it minimizes the acid consumption otherwise associated with these oxygenates, thus minimizing the acid neutralization wastes, a significant environmental benefit (B. V. Vora, personal communication). In dehydrogenation, the oxygenate removal unit improves catalyst stability and lowers costs of methyl tertiary butyl ether production. Petrochemicals Processing for Detergents Production Linear Paraffins for Biodegradable Detergents. Petroleum derivatives account for most of the total surfactant production and household detergents. During the 1940s and 1950s, sodium dodecylbenzene sulfonate was the most widely used synthetic detergent. However, the dodecyl paraffin side group on the benzene ring is highly branched and not easily biodegraded. In the early 1960s, environmental concerns led to development of linear alkylbenzene sulfonate (LAS) detergents, which are both biodegradable and cost-effective. The key to the manufacture of the linear paraffins required to make linear alkylbenzene (LAB) and, hence, LAS is the use of size-selective synthetic zeolites that adsorb linear paraffins but exclude branched paraffins, naphthenes, and aromatics from mixtures spanning a range of boiling points, as in kerosene (C12 to C18). Although anticipated by the work of McBain (7) and Barrer (8, 9,&10), such a class separation of molecules spanning a range of boiling points was virtually impossible before development of the synthetic molecular sieves by Union Carbide in the 1950s. Two different processes, one vapor phase and the other liquid phase, are used. The vapor-phase IsoSiv process was developed at Union Carbide originally for octane improvement. To produce linear paraffins for detergents, kerosene feed, pretreated to acceptable quality and the desired carbon number range, passes at elevated temperature and just over atmospheric pressure through a bed of zeolite adsorbent that adsorbs just the linear paraffins. Just enough hexane vapor follows the kerosene feed to displace the nonadsorbed feed and isomeric hydrocarbons from the void spaces in the adsorber vessel. The effluent from this step is

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3475

combined with the adsorption effluent stream. The linear paraffins adsorbed in the zeolite are desorbed by purging the bed in the opposite direction with hexane. The hexane in the effluent is separated by distillation and is recycled. The remaining linear paraffins comprise the desired product. The liquid-phase Molex process, mentioned earlier, is most often used to produce plasticizers (C6-C10), LABs (C10-C15), and detergent alcohols (C13-C22+, but usually heavier than C16). To make linear paraffins for LAB, increased linearity and low aromatics content are desired. The new high-purity Molex process has improved product purity to 99.7% and reduced aromatics content to 0.05 wt %. In addition, the new OP ADS-34 zeolite adsorbent provides improved long-term separation performance in the Molex process. The linear paraffins made in the Molex process can be sent to UOP's Pacol and DeFine processes (45) for catalytic conversion to monoolefins. These pass to a UOP Detergent Alkylate process (46), which uses a hydrofluoric acid catalyst, or to a Detal process (offered for license in 1995 ), which uses a more environmentally friendly solid, heterogeneous catalyst to produce LAB from the monoolefins plus benzene. Unreacted linear paraffins are recycled to the Pacol and DeFine units and are converted to extinction. These catalytic processes use nonzeolite catalysts. Linear Olefins for Detergent Alcohols. As discussed previously, LAB accounts for half of the detergent intermediate market. Detergent alcohols made from linear olefms are another quarter. Detergent alcohols are made from C10-C15 alpha olefins derived from ethylene or from C10-C15 internal olefins derived from lower-cost kerosene feed. Linear olefins of improved purity are increasingly sought. Linear paraffins from the new high-purity Molex zeolite adsorption process are sent to the Pacol and DeFine processes to convert them to a mixture of monoolefins plus unreacted linear paraffins. The mixture is fed to UOP's Olex process, which uses a zeolite adsorbent in a Sorbex simulated-moving-bed process to separate the linear olefins and the unreacted paraffins, which are recycled back to the Pacol and DeFine units to extinction. The Olex process now uses the new UOP ADS-32 zeolite adsorbent, which provides improved capacity and rates. The linear olefins product has improved product purity (reduced aromatics and diolefins) as a result of improvements in the zeolite adsorbents and in the nonzeolite catalysts and operating conditions in the Molex-PacolDeFine-Olex sequence of processes. Separation and Purification Process Applications Molecular sieve adsorbents are used in many other separation and purification applications: (i) petroleum refining processes, used to remove CO2, chlorides and mercury from a variety of streams; to dry and purify liquids and gases in diverse applications; to treat alkylation unit feed to reduce acid consumption, regenerator use, and corrosion, and to treat refinery hydrogen to prevent corrosion in downstream equipment; to dry and desulfurize refined products; and to dry and purify feed and recycle hydrogen in isomerization units; (ii) petrochemicals, used to dry hydrocarbon liquids, cracked gas, and hydrogen; to dry and purify natural gas liquids, ethane and propane feedstocks in ethylene and polymer plants; and in ethylene, propylene, butadiene, butylenes, amylenes, and various other comonomers and solvents; (iii) natural gas treating, used to dry and desulfurize natural gas to protect transmission pipelines and to remove undesirable impurities from home cooking and heating gas and to desulfurize ethane, propane, and butane and for H2O and CO2 removal before cryogenic processing; (iv) industrial gas production and purification, used to remove H2O and CO2 from air before liquefaction and separation by cryogenic distillation, for pressure swing adsorption (PSA) separation of air, and in PSA purification of hydrogen by using zeolites and other adsorbents, such as activated carbon; (v) specialty and fine chemicals and pharmaceuticals, used for drying; for removal of impurities, including odors; and for other applications in the manufacture of specialty and fine chemicals and pharmaceuticals. Common features of these uses are now summarized. Preprocessing of Gases before Cryogenic Separations. Deep drying and CO2 removal are required before cryogenic liquefaction and subsequent separation processing to prevent formation of ice and dry ice, which would plug up the cryogenic processing equipment. Several synthetic zeolites exhibit great affinity for polar compounds such as H2O and CO2 and have high adsorption capacity at ambient temperature. They are used extensively in processing natural gas to make liquefied natural gas or to recover hydrocarbon liquids or helium; in processing air to make O2, N2, and Ar in cryogenic air-separation plants; and in treating ethylene and other olefins formed in ethylene steam-cracking plants before separation in cryogenic distillation separation units. These pre-treatments work to perfection: Passing ambient air over such zeolite adsorbents at room temperature makes the air drier (−60°C dew point or lower) than in the coldest part of Alaska in the depth of winter. Because the adsorbed impurities are strongly held on the zeolite adsorbents, they are regenerated for subsequent reuse in thermal-swing processes that pass a hot regeneration gas over the spent zeolite to heat it and to carry away the adsorbed compounds. The zeolite then is cooled to ambient temperature and is used to treat more gas. Removal of Impurities from Gases and Liquids Down to Low Levels. Because zeolites bind strongly to polar compounds, including hydrogen sulfide, mercaptans, organic chlorides, CO, and to mercury, they can purify many streams in petroleum refineries, petrochemical plants, natural gas production plants, and chemical plants. In refineries, zeolite adsorbents remove impurities detrimental to downstream processing, including catalyst poisons (e.g., oxygenates and sulfur), corrosive agents, and chloride compound byproducts from processes using chloride catalyst promoters (e.g., catalytic reformers). In natural gas production, zeolite adsorbents are used to dry the gas to prevent freezing and corrosion in pipelines, to remove sulfur compounds from the gas or LPG fractions to prevent corrosion in burners, and to remove compounds that are obnoxious or toxic (such as the odoriferous hydrogen sulfide and mercaptans in natural gas that form sulfur dioxide pollutants when burned for home cooking and heating). Worldwide, > 1,000 units process tens of billions of cubic feet of natural gas daily. Zeolites are used in the preparation of very high-purity fluids for special uses: e.g., gases used in the manufacture of electronics or gases and liquids used in modern analytical laboratory instruments. Air Separation by Pressure Swing Adsorption (PSA) Processes. The following draws primarily from reviews (47, 48, 49, 50, 51,&52) plus the author's personal reflections from direct involvement in zeolite adsorbent development over the last 30 years. Many zeolites adsorb N2 more strongly than O2 [the possible use of zeolites in air separation was indeed the principal impetus for the pioneering work of Milton (11)]. Also, because zeolites adsorb more of both N2 and O2 from air with increasing pressure, air can be separated by using a PSA process. The air is passed at an elevated pressure through a bed of zeolite particles that adsorb the N2 more strongly and hold it on the bed but allow O2 to pass through the bed. Then, the adsorbed N2 is discharged from the feed end of the bed as the pressure in the bed is lowered. Many variations on the process cycle were developed to improve efficiency and capital and operating costs. The PSA and vacuum-swing adsorption (VSA) processes use zeolite adsorbents to produce O2 of 90–94% purity (the balance is primarily argon). The O2 is used, for example, in the manu

Imai, T., Kocal, J. A. & Vora, B. V., Second Tokyo Conference on Advances in Catalytic Science Technology, August 21–26, 1994, Tokyo.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3476

facture of steel, glass, pulp, paper (in delignification and bleaching), and chemicals and in nonferrous metal recovery, waste incineration, and bioremediation. Zeolites provide benefits in energy efficiency, process efficiency, improved processing rates and product quality, and environmental impact. Over the last 25 years, improvements in the PSA and VSA O2 processes were driven by the development of zeolite adsorbents with improved N2 capacity and selectivity. Zeolites such as NaX and CaA made possible the development of the first economical PSA O2 process at a relatively small scale [up to ≈15 U.S. tons of O2 per day (tpd)] in the early- to mid-1970s. Second-generation adsorbents, such as CaX (53), and third-generation adsorbents, such as LiX (54), LiCaX or LiSrX (55, 56), and MgA (57), together with improved (vacuum PSA or VSA) processes have dramatically reduced both capital and operating costs. Of the third-generation Type X zeolites, only LiX has been used commercially as of 1997.** From the mid-1980s to the mid-1990s, these improvements provided a 5-fold reduction in adsorbent inventory and a nearly 2-fold reduction in power requirements. The commercial viability of a simple two-bed VSA system expanded to well over 100 tpd, and that of an even simpler one-bed system expanded to well over 40 tpd, allowing use of these noncryogenic systems in many applications formerly served by the cryogenic distillation of air. For the delivery of O2 of 90–94% purity, single-bed units are more economical because of lower capital costs (although with higher energy costs) than liquid O2 delivery in the 4–57 tpd range. Two-bed units have lower energy costs (but higher capital costs) and are more economical than either liquid O2 delivery or on-site cryogenic plants in the 57–235 tpd range.†† In 1994, VSA plus PSA O2 production was estimated to be 4–4.5% of the world demand for O2, the fourth largest chemical at 39 billion pounds in 1995 (51). In 1996, PSA/ VSA O2 production was estimated to be >3,500 tpd in the USA and >10,000 tpd (>265,000 Nm3O2 per hour) worldwide (47). Assuming a value of $20 per ton of O2, this production corresponds to a total market value of more than $75 million per year, and growing. Manufacturing Industries and Consumer Products Applications Small Oxygen Concentrators for Medical Use (Medox). In the U.S., a dozen companies manufacture small-scale PSA oxygen concentrators for patients with emphysema and chronic obstructive pulmonary disease. As with the large-scale PSA O2 units, these small PSA concentrators use zeolites to produce 90 and 95% pure oxygen, the balance mainly argon and nitrogen. They can dramatically improve the quality of life. The PSA units are engineered to be small (about the size of a small end table), readily transportable, (weighing ≈40 pounds), quiet, and reliable. They use 3–7 pounds of zeolite adsorbent to produce between 3 and 6 liters per minute of oxygen. Users are freed from needing highpressure cylinders of oxygen delivered or stored in their homes. Use of these concentrators has grown substantially over the last 20 years (S. R. Dunne, personal communication). In 1996, home medical equipment reimbursements (from U.S. Medicare) associated with oxygen concentrators totaled $1.1 billion (58). Most are PSA units, the rest being primarily membrane units, which produce much lower oxygen concentration. Automotive Air-Conditioning and Stationary Refrigerant Drying. Zeolite desiccants remove water and acids formed by breakdown of the refrigerant mixed fluid thus protecting the system from freeze-up and corrosion. UOP supplied zeolite 4AXH-5 desiccant for automotive use and 4A-XH-6 for stationary refrigerant drying. These desiccants dominated the refrigerant industry for use with refrigerants R-12 and R-22 and the associated mineral oil lubricants. However, the 1987 Montreal Protocol heralded their demise because the very long life of fugitive R-12 emissions in the atmosphere became linked to ozone depletion and global warming. The leading contender to replace R-12 was R-134a refrigerant, but R-134a was found to be unstable in the presence of the 4AXH-5 desiccant, leading to acids, sludge, deterioration of the desiccant, and possible failure of the refrigerant system. A UOP team developed the new XH7 zeolite desiccant, which is compatible with the new R134a lubricant systems, to meet the critical legislated deadlines. The deadlines for original automotive equipment manufacture and fleet testing were set at 3 years before system production, an extremely short development time for such a complex application. Another new zeolite desiccant, XH9, was developed a couple of years later; in addition to automobiles, it is widely used in refrigerators (home refrigerators, supermarket freezers, and display cases) and stationary air conditioners. Because XH7 and XH9 desiccants are also compatible with systems using R12 refrigerant and mineral oil lubricants, dryers using the new desiccants can be prefit into R12 systems before total conversion to R134a systems. The XH7 desiccant today holds almost all of the automotive air-conditioning market formerly held by the 4AXH-5 desiccant before the advent of the new refrigerants. Consumers benefited because the availability of current systems was not disrupted and any danger to the environment was alleviated: hence, the American Chemical Society Heroes of Chemistry Award in March 1998 to the UOP team (A. P. Cohen, S. L. Correll, P. K. Coughlin, and J. E. Hurst). Worldwide, millions of pounds of zeolite desiccants are installed in air conditioning units in passenger cars and light trucks. For stationary systems, most refrigerators in the U.S. and many elsewhere use zeolite desiccants to dry and remove acids from the refrigerants. The new desiccants, together with hermetically sealed systems with internal pumps, have extended the service life of refrigeration units by at least two to three times. Air Brake Dryers for Heavy Trucks and Locomotives. Most goods produced worldwide are moved to market by heavy trucks and locomotives whose brake systems are actuated mostly by clean dry air at high pressure. Air brake systems are engineered to be fail-safe: When the air supply system fails, the brakes engage and prevent the trucks or locomotives from moving. A key element of the air supply systems of a truck is a PSA dryer, typically using a single packed bed of ≈3 pounds of molecular sieve to dry the compressed gas; locomotives require more absorbent. The air compressor on a truck runs for 1–3 min at a time. The compressed gas is dried and passed to a reservoir that in turn supplies air pressure to the brakes to prevent the unintended actuation of the brakes while the truck is moving. When the reservoir is full, a signal shuts off the compressor; the dryer is depressurized, and a little dry air is bled back through the dryer to partially regenerate the bed of molecular sieve (S. R. Dunne, personal communication). These dryers have significantly improved the reliability and safety of braking systems for large trucks and locomotives. Insulated Glass Windows. Most insulated glass produced worldwide is manufactured with desiccant contained in channels (or in matrices) that separate the panes of double-, triple-, or quadruple-paned windows. The desiccant scavenges moisture and other trace compounds, such as solvents or plasticizers, that may evolve during manufacturing. Although the sealants used for the manufacture of insulated glass windows are excellent, a finite amount of moisture still leaks into the windows over time. Desiccants, primarily zeolites, prevent fogging, mists, or formation of dew between the windowpanes because they lower the dew point of the gases inside the windows to levels far below the lowest expected surface temperature of the glass. Insulated glass provides aesthetic features, improved human comfort, and energy savings that make them a truly economical and beneficial addition to both commercial buildings and homes (S. R. Dunne, personal communication). Residential and nonresidential dualpane (insulating) windows and patio doors containing zeolite adsorbents have a total window area of ≈46 billion square feet worldwide. The estimated present energy savings in heating during winter and cooling during summer from the use of these insulating windows is equivalent to 450 million barrels of oil per year.

**Notaro, F., Schaub, H. R. & Kingsley, J. P., Second Joint China/U.S. Chemical Engineering Conference, May 22, 1997. ††Notaro, F., Schaub, H. R. & Kingsley, J. P., Second Joint China/U.S. Chemical Engineering Conference, May 22, 1997.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3477

Environmental Protection Applications In addition to the benefits already listed, many other applications provide environmental benefits. Builders for Phosphate-Free Laundry Detergents. In the 1960s, growing public awareness of eutrophication of natural waters led to efforts to reduce the inflow of plant nutrients, especially phosphate and ammonia or nitrate. Dead algae sinks to the bottom of a pond or lake, where it depletes the oxygen in the water. Too much growth of algae depletes oxygen so much that fish die. As a result, many states, particularly those bordering the Great Lakes, banned phosphate in laundry detergents. The prime function of phosphate “builders” in laundry detergent powders is removal of the hardness ions Ca2+ and Mg2+ in the wash water by complexing. Zeolite ion exchangers in powder form also can provide this service by removing Ca2+ and Mg2+ ions from the solution and replacing them with soft ions such as Na+. Zeolite NaA was known to have high selectivity and capacity for calcium, and its application as a builder in heavy-duty powder detergents was developed in the 1970s, primarily by scientists at Henkel (59, 60, ‡‡) in Germany and Procter and Gamble (61, 62) in the U.S. Round NaA zeolite particles, a few micrometers across, are small enough to pass through the openings in the weave of the fibers in clothing and are not filtered out to form encrustations on the cloth. Recently, zeolite P (GIS), as maximum aluminum P or MAP, was developed by scientists at Unilever (and Crossfield) (63) as an alternative builder for the same applications, and debate on the relative merits of NaA and NaP zeolites continues (64). Today, the conversion of the USA detergent market to zero-phosphate formulas is virtually complete. In Europe, one-third of the powder detergents are zeolite based, and Canada is ≈50% converted. Latin America and many of the Pacific region countries continue to use phosphates (65). In 1987, the Kao Corporation in Japan introduced Attack, a compact powder detergent that has higher bulk density and higher surfactant level and needs lower dosage. Use of compact powders in the Japanese laundry market grew to >90%. In the U.S., from 1990 to 1994, the use of compact powders grew from 2% to >90%. All compact powders in the U.S., Europe, and Japan have no phosphates. Zeolites used as builders in compact powders serve as particle-formation aids. This use of zeolites has facilitated changes in the process of detergent manufacture from spray drying and to alternative processes such as agglomeration. This shift, in turn, has led to increased use of zeolites in laundry detergent powders. Automotive Emissions Control. New stable zeolites have been used successfully in diverse automotive emissions control problems. A very serious challenge in automotive emissions control today is control of nitrogen oxides (NOx) emitted from lean-burn diesel engines. Many catalyst developers and academic researchers are using zeolites with a wide array of added base and precious metals as catalysts to enable hydrocarbon storage, NOx reduction, and oxidation of both hydrocarbons and CO. Today, zeolites are used commercially for enhanced hydrocarbon oxidation in conventional diesel engines and NOx reduction. Four-way catalysts, which provide NOx reduction, HC oxidation, CO oxidation, and particulate control, are being developed. For gasoline-fueled vehicles, the most-serious problem is the cold-start period. Over 90% of the hydrocarbon emission by a car during a cold start occurs within the first 3 min of engine operation. A hydrocarbon trap must contain an adsorbent that captures most of the hydrocarbons during this period. Once captured, most of the hydrocarbons must be held by the adsorber until the catalytic converter has heated up enough to be capable of oxidizing them. Then the adsorber must release the hydrocarbons to be oxidized, rendering them harmless to the atmosphere. The adsorber also must be mechanically, thermally, and hydrothermally durable enough to withstand the harsh environment of the exhaust gas stream. Especially since 1995, major advances have been achieved in the development of hydrocarbon traps. Improved hydrothermal stability of the molecular sieve adsorbent, improved chemistry of the wash coating, and the addition of a noble metal catalyst directly onto the adsorber brick are technical milestones that enable successful implementation of hydrocarbon traps in emissions control (S. R. Dunne, personal communication).§§ Several major automakers have demonstrated excellent emissions reduction and system durability. General Motors achieved 50,000-mileaged converter performance that surpasses Environmental Protection Agency requirements for use on a car designated as a low-emissions vehicle: i.e., a vehicle must emit nonmethane hydrocarbons at a weighted rate <0.075 grams per mile in the U.S. federally mandated test protocol. Mercedes Benz has achieved emissions that beat the ultralow-emissions vehicle standards of <0.04 grams per mile in the same testing (S. R. Dunne, personal communication). The total U.S. car and light truck production rate is ≈12 million vehicles per year. Over the next 10 years, emissions control technologies of all kinds will be implemented to keep the auto manufacturers in compliance with the law. Europe has a larger vehicle production rate and more eligible vehicles. Radioactive Waste Management. A special use category that is small in both the quantity of both molecular sieves and the gases and liquids processed is radioactive waste management, in which zeolites and other new molecular sieve ion exchangers (66, 67), adsorbents, and catalysts have been used for >25 years. Although only small quantities are used, the past and future environmental benefits are large indeed. UOP's IONSIV zeolite ion exchangers were used for the radioactive waste cleanup at Three Mile Island; the West Valley commercial nuclear fuel reprocessing site; and the Hanford, Savannah River, Oak Ridge, and other U.S. Department of Energy nuclear waste storage sites (67). Most recently, and of special interest, is the current use of the new IONSIV IE-911 crystalline silicotitanate (CST) ion exchangers for the cleanup of the radioactive wastes in the Melton Valley tanks at Oak Ridge (68) and the planned use elsewhere in similar and other applications. The effectiveness of CST was discovered (69) by researchers at Sandia National Laboratories and Texas A & M University, and its further product and manufacturing process development and commercial manufacture was carried out in 1994–1995 by UOP under a Cooperative Research and Development Agreement with Sandia (70, 71). In 1996, this work earned an R&D 100 Award for the Sandia, Texas A & M, and UOP researchers (72). Other Smaller Environmental Applications. Many smaller applications of zeolites in catalysis and adsorption, although important and beneficial to mankind, are not discussed here because of space limitations. Noteworthy are the growing uses of zeolites and other molecular sieves as adsorbents in separations of volatile organic compounds and other pollutants, in desiccant cooling and dehumidification, and as ion exchangers for pollution abatement and toxic waste management. Conclusions Likely Future Applications. Use of zeolites as catalysts in the manufacture of some fine chemicals should expand. New zeolite and other microporous oxide catalysts should be developed with improved selectivity and new functionalities, perhaps for strong base and oxidation catalysis, chiral synthesis, and possibly, membrane reactors. Desiccant cooling and dehumidification, and sorption heat pumps, may achieve serious success. New ion exchange applications of new microporous oxides also may be expected. Experience has taught that the availability of new materials normally precedes by many years the discovery of all of their useful properties and the conception and development of new uses. Impacts of Molecular Sieves on Human Welfare. From these numbers, Roth estimated the value of fuels and chemicals produced using catalysts in 1989 to be $891 billion per year, or 17% of the U.S. gross national product, and judged the corresponding worldwide product values for fuels and chemicals to be $2.4 trillion per year (73). Of course, not all catalysts are based on zeolites, but for petroleum-based fuels and petrochemicals, most catalysts are now zeolitic. Thus, the impact of zeolites in these areas is clearly great. Likewise, the use of zeolites in catalytic converters to reduce undesirable emissions from vehicles also represents a significant present market (with large growth potential) and substantial benefits to mankind in pollution abatement. As described earlier, synthetic zeolites play critical roles in the production of fuels, petrochemicals, and other products essential to modern societies; in pollution avoidance or abatement; in energy efficiency; and in the efficient use of natural raw materials. They also contribute to the quality and performance of the ultimate products because of the greater purity and uniformity of the intermediates made by using zeolites.

‡‡Schwuger, M. J. & Smolka, H. G., 49th National Colloid Symposium, June 1975, Clarkson College, Potsdam, NY. §§Dunne, S. R. & Taqvi, S. M., AIChE Annual Meeting, Session on Environmental Catalysis, November 1997, Los Angeles,

CA.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

SYNTHETIC ZEOLITES AND OTHER MICROPOROUS OXIDE MOLECULAR SIEVES

3478

The many benefits achieved from the applications of zeolites and other molecular sieves are the fruits of the basic investments made decades ago, and into the present, in the research areas of mineralogy, geology, geochemistry, structure and properties of natural zeolites, exploratory materials synthesis, materials characterization methods and their application, and exploratory research on structure, property, and functionalities. The basic concepts and understanding from these efforts, coupled with creative consideration of how the properties and functionalities so discovered might be of service to solve the needs of mankind, continue to create new benefits. We all contribute by standing on the shoulders of giants. In the field of synthetic zeolites and related materials, and their applications, my personal giants include R. M. Barrer, R. M. Milton, D. W. Breck, E. M. Flanigen, J. A. Rabo, L. B. Sand, P. E. Pickert, C. G. Gerhold, D. B. Broughton, G. T. Kerr, G. H. Kuehl, J. J. Collins, G. E. Keller, W. M. Meier, and J. V. Smith. In various ways, the prior work of each has had specific impact on my own work over the years. I also gratefully acknowledge the contributions of my many friends and colleagues at UOP, especially P. T. Barger, J. C. Bricker, A. P. Cohen, N. A. Cusher, S. R. Dunne, G. J. Gajda, S. H. Hobbs, J. A. Johnson, D. C. Kaminsky, P. J. Kuchar, R. L. Patton, M. W. Schoonover, B. V. Vora, S. T. Wilson, and C. M. Yon, the thoughtful secretarial support of Sharon Lambert, and the skilful editorial support of Sandy Weiss. I thank J. V. Smith and the National Academy of Science for organizing this colloquium and inviting this contribution and thank UOP for supporting this endeavor and many others. Finally, I thank my wife, Carol, for patience and understanding.

1. Sheppard, R. A. (1971) Adv. Chem. Ser. 101, 279–310. 2. Sand, L. B. (1968) in Molecular Sieves (Soc. of Chem. Ind., London), pp. 71–77. 3. Cronstedt, A. (1756) Akad. Handl. (Stockholm) 18, 120 4. Cronstedt, A. (1758) Mineral (Stockholm), 102. 5. Schlenker, J. L. & Kuehl, G. H. (1993) Proc. 9th Internatl. Zeolite Conf. 1, 3. 6. Breck, D. W. (1974) Zeolite Molecular Sieves, Structure, Chemistry and Use (Wiley, New York). 7. McBain, J. W. (1932) The Sorption of Gases and Vapors by Solids (Routledge, London). 8. Barrer, R. M. (1945) J. Soc. Chem. Ind. (London) 64, 130. 9. Barrer, R. M. (1949) Q. Rev. 3, 293. 10. Barrer, R. M. (1948) J. Chem. Soc. 2158. 11. Milton (1968) in Molecular Sieves (Soc. Of Chem. Ind., London), pp. 199–203. 12. Flanigen, E. M. (1991) Stud. Surf. Sci. Catal. 58, 13–34. 13. Chen, N. Y., Garwood, W. E. & Dwyer, F. G. (1989) Shape Selective Catalysis in Industrial Applications (Dekker, New York), pp. 1–301. 14. Hemmler, C. L. & Tajbl, D. G. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 3.55. 15. Rabo, J. A., Pickert, P. E., Stamires, D. N. & Boyle, J. E. (1960) Proceedings of the Second International Congress on Catalysis (Editions Technol., Paris), p. 2055. 16. Schcrzer, J. & Gruia, A. J. (1996) Hydrocracking Science and Technology (Dekker, New York). 17. Bridge, A. G. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), pp. 14.15 and 14.33. 18. Reno, M. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 7.41. 19. Bridge, A. G. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 7.21. 20. Genis, O. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 8.49. 21. Miller, S. J. (1994) in Studies in Surface Science and Catalysis, eds. Weitkamp, J., et al., (Elsevier Science, Amsterdam), Vol. 84C, pp. 2319–2326. 22. Miller, S. J. (1994) Microporous Materials 2, 439–450. 23. Cusher, N. A. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 9.15. 24. Sohn, S. W. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 10.75. 25. Cusher, N. A. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 9.29. 26. (April, 1997) Hydrocarbon Asia, 56. 27. Wallace, J. W. & Gimpel, H. E. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 1.15. 28. Jeanneret, J. J. (1997) in Handbook of Petroleum Refining Processes. ed. Myers, R. A. (McGraw–Hill, New York), p. 2.45. 29. Jeanneret, J. J. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 2.37. 30. Jeanneret, J. J., (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 2.55. 31. Olson, D. H. & Haag, W. O. (1984) ACS Symp. Ser. 248, 275. 32. Jeanneret, J. J. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 2.27. 33. Barlow, M. T. (1988) U.S. Patent 4,761, 511. 34. Gosling, C. D., Wilcher, F. P., Sullivan, L. & Mountford, R. A. (1991) Hydrocarbon Process. 35. Chen, N. Y., Garwood, W. E. & Heck, R. H. (1987) Ind. Eng. Chem. Res. 26, 706–711. 36. Jeanneret, J. J. & Mowry, J. R. (1997) in Handbook of Petroleum Refining Processes, ed. Myers, R. A. (McGraw–Hill, New York), p. 10.45. 37. Vora, B. V., Marker, T. L., Barger, P. T., Nilsen, H. R., Kvisle, S. & Fuglerud, T. (1997) Stud. Surf. Sci. Catal. 107, 87–98. 38. Gajda, Gregory J. (1992) U.S. Patent 5,132, 484. 39. Gajda, G. J. & Barger, P. T. (1993) U.S. Patent 5,191,146. 40. Ozmen, S. M., Abrevaya, H., Barger, P., Bentham, M. & Kojima, M. (1993) Fuel Reformulation 3, 54–59. 41. Davis, S. (1997) p. 10.45 in Handbook of Petroleum Refining Processes. Ibid. 42. Wise, J. B. & Powers, D. (1994) ACS Symp. Ser. 552, 273–285. 43. Suzuki, S. (1994) Idemitsu Giho 37, 297–300. 44. Chin, A. A., Krambeck, F. J., Wong, S. S. & Yurchak, S. (1992) U.S. Patent 5,166, 455. 45. Vora, B. V. (1985) U.S. Patent 4,523, 045. 46. Vora, B. V. (1985) U.S. Patent 4,523,048. 47. Notaro, F. (1996) Sci. Tech. Froid 215–233. 48. Kumar, R. (1996). Sep. Sci. Technol. 31, 877–893. 49. Armor, J. N. (1995) Adv. Chem. Ser. 245, 321–334. 50. Ruthven, D. M., Farooq, S. & Knaebel, K. S., (1994) Pressure Swing Adsorption (VCH, New York). 51. Reiss, G. (1994) Gas Sep. Purif. 8, 95–99. 52. Campbell, M. J., Lagree. D. A. & Smolarek, J. (1993) AIChE Symp. Ser. 294, 104–108. 53. Coe, C. G. & Kuznicki, S. M. (1984) U.S. Patent 4,481, 018. 54. Chao, C. C. (1989) U.S. Patent 4,859,217. 55. Chao, C. C., Sherman, J. D., Mullhaupt, J. T. & Bolinger, C. M. (1992) U.S. Patent 5,174,979. 56. Coe., C. G., Kirner, J. F., Perantozzi, R. & White, T. R. (1992) U.S. Patent 5,182,813. 57. Coe, C. G., MacDougall, J. E. & Weigel, S. J. (1994) U.S. Patent 5,354,360. 58. Schworm, K. & Gruenwald, E. (1997) Home Care (July),35–42. 59. Berth, P., Jakobi, G., Schmadel, E., Schwuger, M. J. & Krauch, C. H. (1975) Angew. Hem. Intern. Edit. 14, 94. 60. Smolka, H. G. & Schwuger, M. J. (1978) in Natural Zeolites: Occurrence, Properties, Use, eds. Sand, L. B. & Mumpton, F. A. (Pergamon, New York), pp. 487–493. 61. Savitsky, A. C. (March, 1977). Soap Cosmet. Chem. Spec. 53, 29. 62. Wier, B. H., Grosse, R. J. & Cilley, W. A. (1982) Environ. Sci. Technol. 16, 617. 63. Adam, C. J., Araya, A., Carr, S. W., Chapple, A. P., Franklin, K. R., Graham, P., Minihan, A. R., Osinga, T. J. & Stuart, J. A. (1997) in Progress in Zeolite and Microporous Materials (Elsevier Science, Amsterdam), Vol. 105. 64. Borgstedt, E. v. R., Sherry, H. S. & Slobogin, J. P. (1997) in Progress in Zeolite and Microporous Materials (Elsevier Science, Amsterdam), Vol. 105. 65. Showell, M. S. (1998) Surfactant Sci. Ser. 71, 1–19. 66. Roddy, J. W. (1981). Survey: Utilization of Zeolites for the Removal of Radioactivity from Liquid Waste Streams. (National Technical Information Service, Springfield, VA). 67. Sherman, J. D. (1984) NATO ASI Ser. Ser. E 80, 583–623. 68. Lee, D. D., Walker, J. F. Jr., Taylor, P. A. & Hendrickson, D. W. (1997) Environ. Prog. 16, 251–262. 69. Anthony, R. G., Dosch, R. G., Gu, D. & Philip, C. V. (1994) Ind. Eng. Chem. Res. 33, 2702–2705. 70. Braun, R., Dangieri. T. J., Fennelly, D. J., Sherman, J. D., Schwerin, W. C., Willis, R. R., Bray, L. A., Brown, G. N., Brown, N. E., Miller, J. E., et al. (1996) Int. Top. Meet. Nucl. Hazardous Waste Management 96, 204–213. 71. Dosch, R.G., Brown, N. E., Stephens, H. P. & Anthony, R. G. (1993) Technol. Programs Radioact. Waste Management Environ. Restoration 2, 1751– 1754. 72. Anon. (1996) R&D (Cahners) 38, 36. 73. Roth, J. F. (1991). CHEMTECH (June) 357–362.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOCHEMICAL EVOLUTION III: POLYMERIZATION ON ORGANOPHILIC SILICA-RICH SURFACES, CRYSTAL–CHEMICAL MODELING, FORMATION OF FIRST CELLS, AND GEOLOGICAL CLUES

3479

Proc. Natl. Acad. Sci. USA Vol. 96, pp. 3479–3485, March 1999 Colloquium Paper This paper was presented at the National Academy of Sciences colloquium “Geology, Mineralogy, and Human Welfare,” held November 8–9, 1998 at the Arnold and Mabel Beckman Center in Irvine, CA.

Biochemical evolution III: Polymerization on organophilic silicarich surfaces, crystal–chemical modeling, formation of first cells, and geological clues (biological evolution/silica/feldspar/zeolite/first cell walls)

JOSEPH V. SMITH*†, FREDERICK P. ARNOLD, JR.‡, IAN PARSONS§, AND MARTIN R. LEE§ PNAS is available online at www.pnas.org.

ABSTRACT Catalysis at organophilic silica-rich surfaces of zeolites and feldspars might generate replicating biopolymers from simple chemicals supplied by meteorites, volcanic gases, and other geological sources. Crystal–chemical modeling yielded packings for amino acids neatly encapsulated in 10-ring channels of the molecular sieve silicalite-ZSM-5-(mutinaite). Calculation of binding and activation energies for catalytic assembly into polymers is progressing for a chemical composition with one catalytic AI–OH site per 25 neutral Si tetrahedral sites. Internal channel intersections and external terminations provide special stereochemical features suitable for complex organic species. Polymer migration along nano/ micrometer channels of ancient weathered feldspars, plus exploitation of phosphorus and various transition metals in entrapped apatite and other microminerals, might have generated complexes of replicating catalytic biomolecules, leading to primitive cellular organisms. The first cell wall might have been an internal mineral surface, from which the cell developed a protective biological cap emerging into a nutrient-rich “soup.” Ultimately, the biological cap might have expanded into a complete cell wall, allowing mobility and colonization of energy-rich challenging environments. Electron microscopy of honeycomb channels inside weathered feldspars of the Shap granite (northwest England) has revealed modern bacteria, perhaps indicative of Archean ones. All known early rocks were metamorphosed too highly during geologic time to permit simple survival of large-pore zeolites, honeycombed feldspar, and encapsulated species. Possible microscopic clues to the proposed mineral adsorbents/catalysts are discussed for planning of systematic study of black cherts from weakly metamorphosed Archaean sediments. Introduction and Summary of Biochemical Evolution: Part I. Darwin/Oparin/Haldane/Watson/Crick biological evolution provides a plausible framework for integrating the patchy paleontological record with the complex biochemical zoo of the present Earth (literature review: ref. 1). But how could the first replicating and energy-supplying molecules have been assembled from simpler materials that were undoubtedly available on the early protocontinents? Bernal preferred “life” to begin by catalytic assembly on the surface of a mineral, but all pre-1998 attempts using clays and other minerals to assemble an integrated scheme of physicochemical processes had significant weaknesses. Catalysis of organic compounds dispersed in aqueous “soup” requires a mechanism for concentrating the organic species next to each other on a catalytic substrate. Biochemically significant polymers, such as polypeptides and RNAs, must be protected from photochemical destruction by solar radiation and must not be overly heated. A stable cell wall is needed to protect the first primitive organism. Part I (1) pointed out that certain inorganic materials have internal surfaces that are both organophilic and catalytic, allowing efficient capture of organic species for catalytic assembly into polymers in a protective environment. These physicochemical features are related to the state of the art for zeolite catalysts in the chemical industry, the observed properties of zeolite, feldspar (2), and silica minerals, and a plausible framework for the accretion and early history of the Earth's crust and atmosphere (1). Various materials from the zeolite, feldspar, and silica mineral groups were listed as having surfaces with the capacity to adsorb organic species preferentially over water molecules and catalyze them into polymers. We focus here on mutinaite, a zeolite mineral recently discovered in Antarctica, which is the natural analog of the ZSM-5/ silicalite series of synthetic microporous materials (Note: Microporous does not imply that the pores are of micrometer size; indeed the pores in zeolites are generally less than a nanometer across). This type of molecular sieve is based on a tetrahedral framework containing a threedimensional channel system spanned by rings of 10 oxygen atoms (Fig. 1 Upper Left and Upper Right). The silica-rich end-member of the ZSM-5 series, silicalite, is very organophilic, and Al-substituted synthetic relatives catalyze organic reactions at Al–OH regions. Silicalite provides a useful basis for modeling adsorption/catalytic processes that would apply in principle, but not in detail, to other materials in paper I (1). Nonexperts in computer modeling of crystal structures might note the conventions in Fig. 1 Upper Left and Upper Right. Threedimensional imaging must be idealized and truncated. Fig. 1 Upper Left displays 10 oxygen atoms as spheres half the conventional atomic radius of 1.4 Å. All other atoms are shown merely by the intersection of spokes. Each tetrahedrally coordinated (T) atom lies at the intersection of four yellow spokes and each O atom at the intersection of two maroon spokes. Fig. 1 Upper Right shows all the O atoms as half-size spheres, and the Si and Al types, respectively, of T atoms as yellow and pink spheres joined by thin grey spokes. Only four 10-rings are shown lying in the wall of the channel, and you, the reader, must imagine the channel extending up and down to the surface of the crystal where some adjustment of chemical bonding is needed. Fig. 1 Upper Right is deliberately tilted slightly with respect to Fig. 1 Upper Left. The conventions for amino acids are given in the Fig. 1 Upper Right legend.

*Department of Geophysical Sciences and Center for Advanced Radiation Sources, 5734 South Ellis Avenue, The University of Chicago, Chicago, IL 60637; ‡Advanced Research Systems, 5640 South Ellis Avenue, The University of Chicago, Chicago, IL 60637; and §Department of Geology and Geophysics, University of Edinburgh, Edinburgh EH9 3JW, United Kingdom †To whom reprint requests should be addressed. e-mail: [email protected].

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOCHEMICAL EVOLUTION III: POLYMERIZATION ON ORGANOPHILIC SILICA-RICH SURFACES, CRYSTAL–CHEMICAL MODELING, FORMATION OF FIRST CELLS, AND GEOLOGICAL CLUES

3480

FIG. 1. Computer graphics of part of the atomic framework of silicalite/ZSM-5 with amino acids encapsulated in energetically favored positions. (Upper Left) Tetrahedral framework of silicalite/ZSM-5 showing a 10-ring channel down the y-axis. Most of the figure consists of spokes linking atom positions. One 10-ring of O atoms is shown by spheres displayed at half the formal atomic radii. See text for explanation: oxygen atoms, maroon spheres and intersection of maroon spokes; tetrahedral (T) atoms, intersection of yellow spokes. Only four 10-rings are shown, whereas a perfect crystal would have an infinite number defining the channel. Also shown are the tilted five-rings. Ten-ring channels present in the plane of the paper are difficult to see without rotation on the video display. (Upper Right) Four glycine molecules in the zwitterion configuration encapsulated in silicalite/ZSM-5. Glycine consists of a central C atom bonded to two H, one carboxyl COO−, and one amine NH3+. The cluster of three molecules near the middle of the near-vertical channel has been optimized to interact mutually by way of hydrogen bonding and to be suspended by van der Waals bonding from the O atoms of the 10-ring channel. The fourth molecule is oriented along a horizontal 10-ring channel. All the framework O atoms are represented by half-size maroon spheres. The tetrahedrally coordinated atoms are represented by small spheres differentiated by color: Si, yellow, Al, pink. The glycine molecule is represented by a stick model with conventional color code: O, red; C, grey; N, blue; H, white. The orientation of the channel system is rotated slightly from that in Upper Left. (Lower Left) Three glycine molecules within a 10-ring channel of silicalite, viewed down the y- axis. Coloring as in Upper Left; silicalite/ZSM-5 framework shown as tetrahedra (Si, Al) and balls (O); glycine shown as tubes. Note the alignment of the amino acids parallel to the channel and restricted lateral positions within the channel. (Lower Right) Two of the glycine molecules from Lower Left, viewed along the z-axis. Note the ‘head-to-tail' alignment of the carboxylate group of an amino acid with the amino group of the next amino acid. Once again, the positional constraints on the amino acids in the channel, as well as their parallel alignment with the channel, are emphasized. Introduction to New Unpublished Studies. This third part integrates the current state of research on biochemical evolu

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOCHEMICAL EVOLUTION III: POLYMERIZATION ON ORGANOPHILIC SILICA-RICH SURFACES, CRYSTAL–CHEMICAL MODELING, FORMATION OF FIRST CELLS, AND GEOLOGICAL CLUES

3481

tion that will be presented at the Colloquium on Geology, Mineralogy, and Human Welfare. [The second part on weathered honeycombed feldspars and associated bacteria in a modern granite was under preparation as this third part was being completed and will be published (57) in the regular part of the Proceedings before this paper. Its key contents are given briefly in this paper and are illustrated in Fig. 3.] We begin with crystal–chemical modeling of amino acids inside 10-ring channels of the chosen zeolite, silicalite (Fig. 1 Upper Right, Lower Left, and Lower Right and Fig. 2 Upper and Lower). The channel walls are electrically neutral except for arbitrary replacement of 4% of the silicon–oxygen tetrahedra by aluminum-oxygen-hydroxyl catalytic centers. This 1 in 25 replacement yields nice graphics and has no particular scientific significance. Moreover, this ratio can be varied to increase or decrease the spacing of catalytic centers along the 10-ring channels and to vary the electrical forces on adsorbed molecules and the repeat distances of the polymers generated by catalytic condensation. We end with new ideas for generating primitive protocells inside the honeycombed weathered surfaces of feldspars [Part II (57)]. Fig. 3 shows scanning-electron micrographs of the crystallographically controlled channels in feldspars from the Shap granite, northwest England, together with associated modern bacteria, as models for speculation on the development of the first primitive cells. Now to details. Preliminary Simulations of Encapsulation of Amino Acids in Silicalite/ZSM-5 and Catalytic Generation of Biopolymers. The crystal–chemical reviews in refs. 1 and 2 are updated by papers in the following areas: • crystal chemistry of high-silica materials: Fourier Transform-Raman studies of single-component and binary adsorption in silicalite-1 (3); vapor adsorption in thin silicalite-1 films studied by spectroscopic ellipsometry (4); adsorption/desorption of n-alkanes on silicalite crystals (5); adsorption equilibria of Cl to C4 alkanes, CO2, and SF6 on silicalite (6); adsorption of linear and branched alkanes in zeolite silicalite-1 (7); combined quantum mechanical/molecular mechanics ab initio modeling, demonstrating that the most stable Brønsted sites occur in high-silica zeolites (8); simulation of adsorption and diffusion of hydrocarbons in silicalite, demonstrating that a linear hydrocarbon moves more freely than a branched one, whose CH group becomes locked at a channel intersection (9); adsorption isotherms of linear alkanes in ferrierite— smaller ones, C1–C5, fill the entire pore system, whereas C6 and C7 fit only in a 10-ring channel unless forced into an eight-ring channel by pressure (10, 11); nuclear magnetic resonance of 1H in water adsorbed on silicalite (12); heterogeneity of Brønsted acid sites in Al-substituted faujasites (13); nuclear magnetic resonance of 17O in silica, albite glasses, and stilbite (14, 15); simulation of alkane adsorption in aluminophosphate-5, and calorimetry of alkane absorption in high-silica zeolites (16, 17); hydrophobic properties of all-silica beta zeolite (18); structural location of sorbed pnitroaniline in silicalite/MFI molecular sieves from x-ray powder diffraction and 29Si Magic Angle Spinning–NMR (19); nature, structure, and composition of hydrocarbon species obtained by oligomerization of ethylene on acidic H-ZSM-5 molecular sieve (20): • surface chemistry of various minerals: coordination models for simple surfaces of oxide and silicate minerals (21); the role of intragranular microtextures and microstructures in chemical and mechanical weathering, direct comparisons of experimentally and naturally weathered alkali feldspars (22); • synthesis: RNA-catalyzed nucleotide synthesis of a pyrimidine (23); conversion of amino acids into peptides at 373 K and pH 7–10 on (Ni, Fe)S surfaces (24), synthesis of glycylglycine dipeptide in the presence of kaolin clay and zeolites of Linde Type A, faujasite, and beta types (25); thermodynamic calculation of amino acid synthesis in hot water and application to hydrothermal vents (black smokers) on ocean floor (26); polymerization of various amino acids on hydroxylamine and illite mica, with increasing adsorption affinity of oligomers longer than 7-mer (27, 28,& 29).] Fig. 1 Upper Right, Lower Left, and Lower Right and Fig. 2 Upper and Lower illustrate the current state of chemical modeling of amino acids encapsulated in a silicalite containing one Al substitution for 25 tetrahedral Si. Simulations were carried out using the Sorption module within the MSI/Cerius 2 program system (issued by Pharmacopeia, Princeton, NJ). The Consistent Valence Force Field was used for all atoms. Fig. 1 Lower Left and Lower Right and Fig. 2 Upper illustrate the results of packing simulations based on Monte Carlo techniques for glycine and histidine molecules encapsulated within the 10-ring channels of silicalite. It is possible at low pressure to pack 28 glycine molecules per unit cell or eight histidine. Fig. 1 Lower Left, and its rotated version in Fig. 1 Lower Right, demonstrate how the restriction of lateral motion by the channel system, coupled with charge effects at the amino and carboxyl ends of the amino acids, assists in orienting them correctly for the production of polypeptides. It can also be seen from these illustrations that one of the chief difficulties of the standard model for the formation of life, that of achieving sufficient concentration of reactants while excluding or minimizing environmental degradation, is overcome. Not only are the growing biopolymers protected from outside interference and concentrated in the channels, but the limited degrees of freedom in molecular movements assist in orienting them optimally for polymerization. Simple molecular mechanics simulations within the SPARTAN computer package (Wavefunction, Irvine, CA) of various model complexes are shown in Fig. 2 Upper and Lower.Fig. 2 Upper (stereoview) illustrates an adenine hydrogen bound to the hydroxyl site of the 10ring channel, with the carboxyl end of a glycine residue bound to the amino group of the adenine. This complex is correctly oriented for protonation of the hydroxyl group of the amino acid, followed by the elimination of water and the formation of an amide bond between the base and the amino acid. Such a reaction would be facilitated by a second metal site in the region, and provides one possibility for a precursor to the autocatalytic biopolymers of the ‘pre-RNA' world. Electrostatic potential calculations on the system, using the PM3 semiempirical Hamiltonian within the MOPAC molecular orbital package, indicate that this orientation is favorable for the proposed reaction. Fig. 2 Lower illustrates one possibility of the bonding of an amino acid to the hydroxyl site of a zeolite-type material. It is obvious that the nitrogen functionalities of the histidine ring could form hydrogen bonds. Furthermore, it is also possible to bind the hydroxyl group of the carboxyl functionality to the acid site within the zeolite framework, then dehydrate and form a covalent bond between the surface and the amino acid. This process is analogous to the functionalized glass or plastic beads used in commercial DNA or protein synthesis, where the polymer chain grows away from the supported terminus. Reaction with hydronium ion, presuming mildly acidic media, would enable cleavage of the chain and would release the peptide into solution. In passing, it should also be noted that the amino terminus of the amino acid, which is facing away from the viewer, is oriented along the axis of the channel and hence, in analogy to Fig. 1 Lower Right, is oriented optimally for further reaction. Speculations on Biochemical Evolution Currently Under Evaluation. These illustrative results give confidence for speculations that a microporous aluminum-substituted silica material with mainly hydrophobic channels and widely spaced Al–OH catalytic centers might act as a sausage machine for production of biopolymers that became assembled into

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOCHEMICAL EVOLUTION III: POLYMERIZATION ON ORGANOPHILIC SILICA-RICH SURFACES, CRYSTAL–CHEMICAL MODELING, FORMATION OF FIRST CELLS, AND GEOLOGICAL CLUES

3482

FIG. 2. Computer graphics of part of the atomic framework of silicalite/ZSM-5 with amino acids encapsulated in energetically favored positions. (Upper) Stereopair of 10-ring channel of silicalite, showing a hydrogen-bound adenine–glycine complex itself hydrogen-bound to an acid site of the framework. Some framework atoms removed for clarity. This stereopair illustrates the ability of the zeolitic material both to accommodate large biopolymer precursors and to provide sites at which reactions may occur. (Some students initially have difficulty viewing stereopairs. If you have this problem, try putting a bright spot at the red atoms at the extreme top right and bottom left. Your brain should then be able to make your eyes swivel to achieve stereo with the left eye seeing the left dot. This contrasts with the cross-eye technique used in some biological modeling.) (Lower) Histidine and water molecules in silicalite. Histidine was chosen for modeling because its imidazole ring can switch electronic states readily to catalyze the making and breaking of bonds, as well as to provide several potential sites for binding and reactivity. The histidine molecules are shown by a ball-and-stick arrangement and are colored as in Upper. The water molecules are represented by a bent bicolor rod with two white ends representing H, and the red center represents O. The framework is represented by tetrahedra whose shared vertices are at O positions. A further reason for choosing histidine is its prevalence as a metal-binding site in modern proteins, undoubtedly an important function in the prebiotic world.

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOCHEMICAL EVOLUTION III: POLYMERIZATION ON ORGANOPHILIC SILICA-RICH SURFACES, CRYSTAL–CHEMICAL MODELING, FORMATION OF FIRST CELLS, AND GEOLOGICAL CLUES

3483

protocells in protected honeycombs in weathered feldspars. Various new matters are being evaluated currently from the viewpoint of physical and chemical processes and are modeled in detail for later publication: • First, channel intersections may prove important for stereochemical control of larger functional groups, especially at the end of a biopolymer. The intersection of an internal channel with the outer surface should be even more important and indeed might be considered as an anchor for a polymer that projects outwards into a ‘soup.' • Second, after outward migration from an internal channel, the first biopolymers would begin to coil up like a snake and in certain places, such as a tapered tube in a honeycombed feldspar approximately 5 to 100 nanometers across, would begin to interact closely with the aluminosilicate surface. • Third, various biopolymers of different types might begin to interact and begin the evolution toward a protobacterium. • Particularly important would be the first generations of persistent energy-generating species containing phosphorus and electrontransferable transition metals. Very important is that K-rich feldspars from granites contain micrometer inclusions of the calcium— phosphate–hydroxide/halide mineral apatite and transition-metal oxides, including ilmenite, spinel, and hematite, which might well be the primary reservoirs of these key elements. • At some stage, a protocell lining inside an aluminosilicate tube might develop a bilipid lining that would extend into a cap, ultimately allowing detachment from the silicate and free motion through a soup. Again, one might imagine a sausage machine popping off a free cell as the remaining protocell reconstituted itself ready for generation of the next free cell. Schematic graphics are being envisaged along with ideas for chemical bonding schemes. • All these processes would involve subtle effects related to diurnal and annual temperature cycles and wet/dry cycling driven both by solar radiation and lunar tides that would change the spatial distribution of chemical forces across mineral surfaces. Ideas are not developed enough so far to warrant further description here. Mineralogical Observations and Speculations: Electron Microscopy of Honeycombed Weathered Feldspars—Bacteria and Protocells. Turning now to mineralogical information, Fig. 3 contains four scanning-electron micrographs of the crystallographically controlled honeycomb weathering on a modern surface of a K-feldspar from the Shap granite (30). Particularly important are the micrometer-scale sausage shapes interpreted as electron scattering from bacteria, somewhat shrunken from interaction with the electron beam. Many have no particular orientation with respect to the feldspar, but the bacterium in Lower Right is interpreted as sitting neatly in a crevice. The near correspondence between the segmentation of the proposed bacterium and the spacing of the feldspar honeycomb is intriguing. Perhaps it may ultimately be possible to quantify the original chemical linkages between the inorganic substrate and the unshrunken bacterium and to use them for modeling the above ideas. Coupling the catalytic production of polymers at the nanometer scale with bacteria at the micrometer scale is plausible in the geologic context, but requires many flights of imagination and a lot of faith. On the present Earth, volcanic glass transforms to zeolites in continental basins and ocean floors; the zeolites become metamorphosed to feldspars; and the whole mineralogical assemblage becomes converted over geologic time into granitic metamorphic rocks. Hence, we are comfortable in proposing that zeolites and feldspars would have coexisted on the early Earth, for which only the resultant granitic metamorphic rocks have been seen so far. Hence we can suggest for discussion purposes that a zeolite/silica/ altered feldspar sausage machine fed a range of biological polymers into feldspar honeycombs. As discussed above, intermingling of polymers, generation of P-bearing energy-transporting species from apatite, and hydrogen-bond coupling between organic species and silicarich walls, would have generated primitive protocells. To conclude the evolution into the first organisms, a cap between the dangerous outer regime of ‘soup' and the inner protected world might have expanded to completely enclose the protocell so that it could swim into the future. Part IV, under preparation, will show graphics illustrating the scientific factors underlying these flights of fancy about cell formation. Conclusion. We conclude with matters of specific geological import. From the humanistic viewpoint, it would be extremely significant if the early forms of life had left behind some physicochemical evidence of their existence. The current carbon-isotope evidence of bulk samples is indicative of some kind of early biological evolution, but has no particular import for the atomic-level ideas presented above. A review of the geological evidence indicates that at least most and perhaps all of the early Archean rocks have been metamorphosed to a high enough level that all volcanic rocks have recrystallized. There can be little doubt that volcanoes would have been pumping out ash containing crystals of Kfeldspar and silica minerals. By analogy with modern conditions, much of the ash would have been converted into zeolite beds, and there might well have been zoned beds of zeolite minerals interacting with salty lakes sloshed by tides and impacts. Some K-feldspar and zeolite crystals would have been exposed to an acidic rain, and honeycombed and grooved faces should have occurred (1, 2). Primitive molecules would certainly have been available dispersed in ‘soups', as envisaged by many writers (1). Here are some ideas for testing whether minerals produced by metamorphic recrystallization of earlier igneous origin might have retained some specific signature indicative of subtle biological processes involving feldspar, zeolite and silica minerals: • Ancient cherts (silica–hydroxyl-rich aggregates) range in color at least from black to brown, red, and orange-yellow. At least some of the color variation must result from transition metals, especially Fe and Mn, at different redox states. Might some carbonaceous species have survived in the black cherts? If so, would careful analysis reveal organic breakdown products specific to primary biocatalytic precursors? • Because early organisms would have needed P and various transition metals, would their absence or low abundance in the metamorphosed siliceous rocks be indicative of early biological scavenging by organisms that escaped into the ‘soup'? • Particularly challenging, because of the possibility of complete failure, would be a hunt for x-ray diffraction evidence of surviving Sirich molecular sieves. Silicalite and other large-pore zeolites have strong low-angle diffractions that would stand out in lowbackground patterns obtained with synchrotron x-rays, even at a concentration below 1%. Since the review in ref. 1, the following geological/biological publications have been added:

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOCHEMICAL EVOLUTION III: POLYMERIZATION ON ORGANOPHILIC SILICA-RICH SURFACES, CRYSTAL–CHEMICAL MODELING, FORMATION OF FIRST CELLS, AND GEOLOGICAL CLUES

3484

FIG. 3. Four scanning-electron micrographs of weathered feldspar from the Shap granite. (Upper Left) Resin cast of honeycomb texture. The cast is somewhat flexible, so that some of the etched dislocations appear to be curved, although they were almost straight in the original feldspar, which has been dissolved away in HF. (Upper Right) A near-planar surface close to bar601 with a trace of etched dislocations running horizontally and vertically across the image with ellipsoidal bacteria, some in strings like sausages hanging in a U.K. butcher's shop. (Lower Left) More deeply weathered surface showing occasional traces of etched dislocations, with sausage-shaped bacteria. (Lower Right) Detail of honeycomb on the 001 surface of a feldspar honeycomb. The holes are etch pits formed on paired outcrops of dislocations that formed on exsolution lamellae. The bacterium, although perhaps partly shrunken by the instrument vacuum, is segmented on a scale remarkably similar to the spacing of the etch pits. Details of the feldspar weathering are given in Part II (57). • stardust and meteorites: Photochemical evolution of interstellar/precometary organic material (31); silica-rich micrometer objects in a carbonaceous chondrite (32); • planetary impact processes: Survival of amino acid in large comet impacts (33); • early geologic events on Earth: Nitrogen fixation by volcanic lightning (34); redox state of upper-mantle peridotites under the ancient cratons, and possible equilibrium of diamonds with methane-nitrogen-rich fluids (35); new revised Pb-ages of Greenland gneisses at 3.65–3.70 instead of earlier 3.85 gigayear-before-present (36); interpretation of geologic evidence in favor of plate-tectonic processes in the Archean era (37); interpretation of Archean magmatism and deformation in nonplate tectonics terms (38); details of Precambrian clastic sedimentation that partly match and partly differ from recent processes (39); evidence from mature quartz arenites in various Archean shields of stable continental crust containing quartz-rich granitoid rocks (40); microbiological evidence for Fe(III) reduction to Fe(II) on early Earth, and support for earlier idea that Fe(III) was a more likely electron acceptor than S in microbial metabolism (41), birth of the Earth's atmosphere, and the behavior and fate of its major elements (42); • bacteria, cell walls, various matters: Text on bacterial biogeochemistry, with final chapter on origins and evolution of biogeochemical cycles/prebiotic Earth and mineral cycles/ theoretical perspectives on the origins of life (Oparin-Haldane theory, Cairns–Smith ideas on clays and life, pyrite, and the origins of life, “thioester world”) (43); intracellular bacteria in protozoa (44); plant cell wall proteins (45); gene molecular sequences of Archea and details of thermophiles and cold-dwelling types (46); hydrogen consumption by methanogens on the early Earth (47); genome sequences from a dozen bacteria and a yeast fit with a three-kingdom world (48); ‘Eukaryotes are suggested to have arisen through symbiotic association of an anaerobic strictly hydrogen-dependent strictly autotrophic archaebacterium (the host) with a eubacterium (the symbiont) that was able to respire, but generated molecular hydrogen as a waste product of anaerobic heterotrophic metabolism,' (49); bacteria in sediments (50); • chert: The following papers about the siliceous nodules known as chert and about related siliceous materials should be useful in thinking about how to characterize ancient chert: Evidence of volcanic origin of chert in the Permo-Triassic Sydney Basin (51); growth of chalcedony by assembly of short linear polymers with silica monomers (52); growth fault control of ≈3.5 Gybp Early Archaean cherts, barite mounds, and chert-barite veins, North Pole Dome, Eastern Pilbara, Western Australia, carbonaceous aggre

About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution.

BIOCHEMICAL EVOLUTION III: POLYMERIZATION ON ORGANOPHILIC SILICA-RICH SURFACES, CRYSTAL–CHEMICAL MODELING, FORMATION OF FIRST CELLS, AND GEOLOGICAL CLUES

3485

gates in grey chert (53, 54); transformation of black to white chert (55); classic Rhynie chert locality with evidence for a low-energy lacustrine environment with periodic desiccation on exposed mud flats (56). To conclude: From the viewpoint of geology, mineralogy, and human welfare, it is quite obvious that major questions on biochemical evolution remain unanswered, but might become accessible to quantitative study with the new analytical tools developed over the past few decades. Surface biogeochemistry is a subject whose time has come. F.P.A. thanks Advanced Research Systems for computer facilities. I.P. and M.R.L. thank the U.K. National Environment Research Council for a grant. J.V.S. thanks many scientists at UOP for allowing him to participate in their pioneering work on organophilic silicic molecular sieves and wishes to acknowledge the pioneering independent parallel studies by scientists from Mobil Corporation on the ZSM-5 zeolite series.

1. Smith, J. V. (1998) Proc. Natl. Acad. Sci. USA 95, 3370–3375. 2. Smith, J. V. (1998) Proc. Natl. Acad. Sci. USA 95, 3366–3369. 3. Ashtekar, S., Hastings, J. J. & Gladden, L. G. (1998) J. Chem. Soc. Faraday Trans. 94, 1157–1161. 4. Bjorklund, R. B., Hedlund, J., Sterte, J. & Arwin, H. (1998) J. Phys. Chem. B 102, 2245–2250. 5. Millot, B., Methivier, A. & Jobic, H. (1998) J. Phys. Chem. B 102, 3210–3215. 6. Sun, M. S., Shah, D. B., Xu, H. H. & Talu, O. (1998) J. Phys. Chem. B 102, 1466–1473. 7. Vlugt, T. J. H., Zhu, W., Kapteijn, F., Moulijn, J. A., Smit, B. & Krishna, R. (1998) J. Am. Chem. Soc. 120, 5599–5600. 8. Braendle, M. & Sauer, J. (1998) J. Am. Chem. Soc. 120, 1556–1570. 9. Smit, B., Loyens, L. D. J. C. & Verbist, G. L. M. M. (1997) J. Chem. Soc. Faraday Trans. 106, 93–104. 10. van Well, W. J. M., Cottin, X., de Haan, J. W., Smit, B., Nivarathy, G., Lercher, J. A., van Hooff, J. H. C. & van Santen, R. A. (1998) J. Phys. Chem. B 102, 3945–3951. 11. van Well, W. J. M., Cottin, X., Smit, B., van Hooff, J. H. C. & van Santen, R. A. (1998) J. Phys. Chem. B 102, 3952–3958. 12. Turov, V. V., Brei, V. V., Khomenko, K. N. & Leboda, R. (1998) Microporous Mesoporous Mater. 23, 189–196. 13. Maekawa, H., Saito, T. & Yokokawa, T. (1998) J. Phys. Chem. B 102, 7523–7529. 14. Sierka, M., Eichler, U., Datka, J. & Sauer, J. (1998) J. Phys. Chem. B 102, 6397–6404. 15. Xu, Z. & Stebbins, J. F. (1998) Geochim. Cosmochim. Acta 62, 1803–1809. 16. Maris, T., Vlugt, T. J. H. & Smit, B. (1998) J. Phys. Chem. B 102, 7183–7189. 17. Savitz, S., Siperstein, F., Gorte, R. J. & Myers, A. L. (1998) J. Phys. Chem. B 102, 6865–6872. 18. Stelzer, J., Paulus, M., Hunger, M. & Weitkamp, J. (1998) Microporous Mesoporous Mater. 22, 1–8. 19. Mentzen, B. F. & Lefebvre, F. (1998) J. Chim. Phys. Biol. 95, 1052–1067. 20. Stepanov, A. G., Luzgin, M. V., Romannikov, V. N., Sidelnikov, V. N. & Paukshtis, E. A. (1998) J. Catal. 178, 466–477. 21. Koretsky, C. M., Sverjensky, D. A. & Sahai, N. (1998) Am. J. Sci. 298, 349–438. 22. Lee, M. R., Hodson, M. E. & Parsons, I. (1998) Geochim. Cosmochim. Acta 62, 2771–2788. 23. Unrau, P. J. & Bartel, D. P. (1998) Nature (London) 260, 260–263. 24. Huber, C. & Wächtershäuser, G. (1998) Science 281, 670–672. 25. Zamaraev, K. I., Salganik, R. I., Romannikov, V. N., Vlasov, V. A. & Khramtsov, V. V. (1995) Doklady Chem. (Transl. of Dokl. Akad. Nauk.) 340, 56– 58. 26. Amend, J. P. & Shock, E. L. (1998) Science 281, 1659–1662. 27. Orgel, L. E. (1998) Origins Life Evol. Biosphere 28, 227–234. 28. Hill, A. R. Jr., Böhler, C & Orgel, L. E. (1998) Origins Life Evol. Biosphere 28, 235–242. 29. Liu, R. & Orgel, L. E. (1998) Origins Life Evol. Biosphere 28, 245–257. 30. Lee, M. R. & Parsons, I. (1998) J. Sediment. Res. Sect. A 68, 198–211. 31. Allamandola, L. J., Bernstein, M. P. & Sandford, S. A. (1997) in Astronomical and Biochemical Origins and the Search for Life in the Universe, eds. Cosmovici, C. B., Bowyer, S. & Wertheimer, D. (Editrice Composipori, Milan). 32. Nazarov, M. A., Kurat, G. & Brandstätter, F. (1998) Meteorites Planet. Sci. 33, A115. 33. Pierazzo, E. & Chyba, C. F. (1998) Meteorites Planet. Sci. 33, A122–123. 34. Navarro-Gonzaléz, R., Molina, M. J. & Molina, L. T. (1998) Geophys. Res. Lett. 25, 3123–3126. 35. Simakov, S. K. (1998) Geochim. Cosmochim. Acta 62, 1811–1820. 36. Kamber, B. S. & Moorbath, S. (1998) Chem. Geol. 150, 19–41. 37. de Wit, M. J. (1998) Precambrian Res. 91, 181–226. 38. Hamilton, W. B. (1998) Precambrian Res. 91, 143–179. 39. Eriksson, P. G., Condie, K. C., Tirsgaard, H., Mueller, W. U., Altermann, W., Miall, A. D., Aspler, L. B., Catuneanu, O. & Chiarenzelli, J. R. (1998). Sediment. Geol. 120, 5–53. 40. Donaldson, J. A. & de kemp, E. A. (1998) Sediment. Geol. 120, 153–176. 41. Vargas, M., Kashefi, K., Blunt-Harris, E. L. & Lovley, D. R. (1998) Nature (London) 395, 65–67. 42. Javoy, M. (1998) Chem. Geol. 147, 11–25. 43. Fenchel, T., King, G. & Blackburn, H. (1998) Bacterial Biogeochemistry: The Ecophysiology of Mineral Cycling (Academic ,London). 44. Görtz, H. & Brigge, T. (1998) Naturwissenschaften 85, 359–368. 45. Cassab, G. I. (1998) Annu. Rev. Plant Physiol. Plant Mol. Biol. 49, 281–309. 46. DeLong, E. (1998) Science 280, 542–543. 47. Kral, T. A., Brink, K. M., Miller, S. L. & McKay, C. B. (1998) Origins Life Evol. Biosphere 28, 311–319. 48. Doolittle, R. F. (1998) Nature (London) 392, 339–342. 49. Martin, W. & Müller, M. (1998) Nature (London) 392, 37–41. 50. Nealson, K. H. (1997) Annu. Rev. Earth Planet. Sci. 25, 403–443. 51. Dutta, P. K. (1998) Sediment. Geol. 117, 123–132. 52. Heaney, P. J. (1993) Contrib. Mineral. Petrol. 115, 66–74. 53. Nijman, W., de Bruijne, K. H. & Valkering, M. E. (1998) Precambrian Res. 88, 25–52. 54. Sugitani, K. (1992) Precambrian Res. 57, 21–47. 55. Shanmugam, G. & JB Higgins, J. B. (1988) Am. Assoc. Pet. Geol. Bull. 72, 523–525. 56. Trewin, N. H. & Rice, C. M. (1992) Scott. J. Geol. 28, 37–47. 57. Parsons, I., Lee, M. R. & Smith, J. V. (1998) Proc. Natl. Acad. Sci. USA 95, 15173–15176.