DEDICATION
The genesis of this book can be traced back to July 2003, the date of the Carbon 2003 Conference. This presented Eduardo J. Bottani and myself with an excellent opportunity to introduce our project to lain Craig, the Commissioning Editor with whom we discussed the scope and contents of the book. Soon afterwards, the publication of the book was approved by Elsevier Science and we established initial contacts with some prospective authors. However, the unexpected death of Eduardo J. Bottani on October 24,2003, meant that I now had a dual responsibility, not only to the publisher and the authors who had already agreed to contribute chapters, but also to the memory of Eduardo, to whom I wish to dedicate this book. Needless to say, I accept full responsibility for any faults in the editing as Eduardo passed away before we began the work in earnest. Prof. Eduardo Jorge Bottani was born in Bahia Blanca, Argentina, on July 2, 1955. After completing his studies in Chemistry (speciality: Physical Chemistry) at the Universidad Nacional de La Plata (UNLP), he went on to gain a PhD in the same university and speciality in 1982. His thesis, carried out under the supervision of Prof. Lydia E. Cascarini de Torre, was entitled "Study of interactions in physically adsorbed molecules." Except for several periods ofstudy abroad, first at Louvain-Ia-Neuve, Belgium, and later at State College, PA, USA (with Prof. W.A. Steele), Eduardo J. Bottani spent most of his life in La Plata and his entire professional career at INIFTA, a renowned research institute devoted to Basic and Applied Physical Chemistry. It was here while a member of the Physical Adsorption Laboratory that he held several positions as a research scientist for CIC. Towards the end of his life he also collaborated with Profs A.J. Arvia and E.A. Castro, successive directors of the Institute, in organizational tasks. While carrying out his scientific research, he taught General, Inorganic and Physical Chemistry at UNLP (his Alma Mater which, I regret to say, returned him much less than he had given her). Later on, he was given the post of Visiting Professor at the Universidad Nacional del Litoral (UNL) in the city of Santa Fe, where, together with Prof. H.S. Odetti, he formed a flourishing research team. I first met Eduardo on October 12, 1992, when Lydia Cascarini and he welcomed me at Ezeiza Airport in Buenos Aires (the date is easy to remember as it was exactly 500 years to the day after a famous discovery known to all mankind). This was the beginning of a period of collaboration that has culminated in the production of this book. Since 1999, Eduardo J. Bottani spent one month a year in Oviedo, where we combined our efforts in investigating gas physisorption on different materials such as carbon blacks, fullerenes, and carbon v
vi
Dedication
nanotubes. Moreover, he and I shared an interest in art (especially architecture) and made many weekend visits to different monuments of our Spanish cultural heritage. At the time he passed away, Prof Bottani was Deputy Director of INIFTA and General Secretary of the Argentine Chemical Society (Asociaci6n Qu£mica Argentina, AQA). Shortly before, he had been appointed Editor-in-Chief of the Journal of the Atgentine Chemical Society, where he took up the challenge of converting the traditional Anales de la AQA into a truly international journal. He was also responsible for the organization of a number of Argentine Conferences on Chemistry. All of these facts about his life show that Eduardo J. Bottani was at the summit of his career at the time of his death, sadly yet another unfortunate case of a gifted person with a promising future brought to a premature end. Eduardo is survived by his wife Jovita Montoni de Bottani, daughter Estefania, and sons Eduardo and Ezequiel, to whom, together with his mother, Dona Elvira Gar6foli de Bottani and rest of the family, friends, and colleagues I offer this book as a token of my esteem. Juan M.D. Tasc6n
FOREWORD
It has been a great pleasure for me to write a foreword for this very interesting book on Adsorption by Carbons in response to an invitation I first received from the editors in August 2003 when the book was still only a project. This invitation has afforded me the opportunity and the privilege of uniting the efforts of an outstanding group of renowned scientists as authors of the different chapters, some of whom have been either my colleagues or disciples. Part 1 of the book is of an introductory nature. Thus, Chapter 1 provides an updated overview of adsorption by carbons, differentiates the features of adsorption on nonporous and porous carbons, and identifies the main obstacles still hindering the study of gas adsorption by porous carbons. Recent major developments and research needs in this field are also mentioned. Chapter 2 presents a survey of carbons in the context of adsorption and classifies them on the basis of structural criteria. This chapter might help nonspecialists in carbon to find materials that they can use for adsorption purposes. Part 2 of the book addresses the fundamental aspects of adsorption by carbons. The first few chapters deal with the energy aspects of gas adsorption. Thus, Chapter 3 focuses on thermodynamic quantities, with particular attention being paid to their significance and relation to experimental results. Cursory descriptions of the classical and statistical thermodynamic approaches are presented in the form of essential equations. Chapter 4 presents two types of simulation techniques (Monte Carlo and Molecular Dynamics) that help us to understand the behavior of molecules on surfaces of carbons and other materials. It is pointed out that the former method is more useful for studying transport properties, whereas the latter provides information about the thermodynamic properties at the adsorbent/adsorbate interface. New areas under development in this field are outlined. Chapter 5 focuses on models of porous carbons that are useful for predicting and explaining the behavior of adsorbed phases. The discussion encompasses both classical single-pore models (e.g., the well-known slit-shaped pore) as well as more realistic models that include factors such as connectivity or tortuosity. The connection between models and experimental techniques and the need for improving methods of obtaining experimental data are emphasized. Chapter 6 deals with adsorption hysteresis or, to be more precise, the origin of adsorption hysteresis. It addresses the peculiar characteristics of hysteresis in the case of carbon adsorbents, namely, structural lability and the presence of surface chemical structures (principally, oxygenated groups) that modify the energetics and wettability behavior of carbon surfaces. The surface heterogeneity of carbons is analyzed in Chapter 7 on the basis that practically all real adsorbents XVll
xviii
Foreword
are heterogeneous. The big differences within the ensemble of carbon materials (ranging from higWy homogeneous graphites to extremely heterogeneous activated carbons) are discussed and the factors that cause heterogeneity are analyzed. The wetting of solid surfaces by liquids is addressed next in Chapter 8. To analyze the wetting of carbon, the study was extended from graphite (by far the most widely studied carbon surface) to include other materials such as carbon nanotubes, the geometry of which greatly affects the wetting properties. The theory and simulation of gas adsorption on carbon nanotube bundles is addressed in Chapter 9, which focuses on the structural, dynamic, and thermal properties of gases adsorbed at sites such as grooves between pairs of nanotubes, interstitial channels, or inner spaces inside the individual tubes. This chapter serves as a background for Chapter 15, which belongs to Part 3 of the book. The section on fundamentals ends with Chapter 10, which follows two general approaches (the generalized Gaussian model and the bivariate model) to characterize the energy topography of carbon surfaces. The former model is attractive due to its simplicity and works well with substrates with a rough-surface adsorptive energy. The bivariate model is applicable to heterogeneous surfaces with two kinds of sites. The series of 10 chapters that constitute Part 3 of the book deals mainly with the use of adsorption as a means of characterizing carbons. Thus, the first three chapters in this section complement each other in the use of gas-solid or liquid-solid adsorption to characterize the porous texture and/or the surface chemistry of carbons. Porous texture characterization based on gas adsorption is addressed in Chapter 11 in a very comprehensive manner and includes a description of a number of classical and advanced tools (e.g., density functional theory and Monte Carlo simulations) for the characterization of porosity in carbons. Chapter 12 illustrates the use of adsorption at the liquid-solid interface as a means to characterize both pore texture and surface chemistry. The authors propose these methods (calorimetry, adsorption from solution) to characterize carbons for use in such processes as liquid purification or liquid-solid heterogeneous catalysis, for example. Next, the surface chemical characterization of carbons is comprehensively treated in Chapter 13, which discusses topics such as hydrophilicity and functional groups in carbon as well as the amphoteric characteristics and electrokinetic phenomena on carbon surfaces. The next block (Chapters 14-18) is principally devoted to the characterization of several types of carbon that deserve attention for their novelty. Thus, fullerenes, which are dealt with in Chapter 14, are studied as adsorbents with various objectives in mind: for the fundamental investigation of their surface energetics; as tools for the preconcentration and analysis of species in solution; or even as possible substrates for hydrogen storage. This last topic is still a "hot" one in the case of carbon nanotubes, which explains why two whole chapters (besides Chapter 9) have been dedicated to this type of material. One of them (Chapter 15) addresses hydrogen adsorption from a more theoretical point of view, while the other (Chapter 16) takes a look at the actual knowledge obtained from the experimental results published in the literature and therefore
Foreword
xix
offers the reader a complementary, more practical view. Chapter 15 identifies two problems that continue to obstruct progress in research, namely, the variability of the materials being studied by different teams and the lack of reliable computational methods for determining adsorption potentials or chemisorption interactions. Chapter 16 discusses the nature of different types of adsorption sites on nanotube surfaces and concludes that further research is needed to make such sites more easily identifiable, particularly high energy binding sites. The next two chapters deal mainly with the use of adsorption to characterize porous solids. In the case of activated carbon fibers (Chapter 17), methods to characterize microporosity, and particularly ultramicroporosity, by physical adsorption are of particular relevance for understanding the behavior of these adsorbents and extending the range of their applications. Moreover, in Chapter 18 the pore structure of ordered mesoporous carbons is shown to differ greatly from that of conventional activated carbons for which most of the available data treatment methods have been developed. Therefore, suitable procedures for correctly analyzing the pore structure of these novel carbons are proposed in this chapter. Two chapters related with the solid-liquid interface conclude Part 3. The electrochemical behavior of carbons is the subject of Chapter 19, which discusses the characteristics of carbon electrode materials with reference to concepts such as roughness or fractality, and also the electrochemical kinetics on carbon electrodes. Chapter 20 deals with the application of scanning probe microscopy to the study of inorganic and organic adsorbates deposited on highly oriented pyrolytic graphite (HOPG) at the submonolayer and monolayer level. The basal plane surface of HOPG is taken as a model system, thanks to its atomic-scale smoothness and low chemical reactivity. Part 4 of the book deals with the applications of adsorption in different fields of technology with explanations as to why carbons exhibit a particular behavior. First of all, gas-phase applications are addressed, either for the removal of pollutants (both volatile organic compounds and inorganic gases, Chapter 21) or for gas separation and storage (Chapter 22). These two chapters present comprehensive summaries of the surface science involved in these important processes and provide clues for selecting the right carbons to be used as adsorbents. Another application of carbons in the field of energy is that of electrochemical energy storage. Chapter 23 discusses the two most important current lines of investigation, namely, lithium batteries and supercapacitors, and relates their performance to carbon characteristics and identifies present research needs. Finally, Chapters 24-27 deal with the environmental applications of carbons as adsorbents for the removal of pollutants from aqueous solutions. These four chapters are highly complementary. Thus, Chapter 24, which addresses the problems associated with the removal of inorganic species, finds its "alter ego" in Chapter 25, which deals with the adsorption of organic solutes from dilute aqueous solutions. Both chapters provide insights into the fundamental reasons for the performance exhibited by a carbonaceous adsorbent. The global topic of water purification using carbons as adsorbents is addressed in Chapter 26, which
xx
Foreword
deals with the science and technology involved in the removal of a surprising variety of pollutants from water employing activated carbon, either in powdered or in granular form. Finally, Chapter 27 is somewhat complementary to Chapter 25 as it discusses the sorption of viscous organics (as opposed to dilute organic solutes). However, there is a significant difference in the type of adsorbent. Indeed, the removal of viscous organics (e.g., heavy oils and biomedical molecules) requires a macroporous adsorbent (e.g., exfoliated graphite) rather than a microporous one. From the preceding overview it is easy to imagine the enormous effort required of the editors and the authors of the different chapters to bring this important piece of work to a satisfactory conclusion. I believe that this will become a reference book for any person interested in the subject of adsorption and carbons. It will be useful not only to those beginning their study of activated carbons and related materials, but also to specialists wishing to further explore this interesting field of research. New discoveries are constantly being made in this area, leading to the solution of numerous problems, both of a theoretical nature and in the applied field ofmodern science and technology. Let me convey again my congratulations to the editors and my esteem for Prof Bottani of whom I have a fond memory. Finally, may I wish Prof Tascon further success in this interesting field of research. Prof Dr. Juan de Dios Lopez-Gonzalez Granada and Madrid, Spain
PREFACE
The essential aim of this work is to fill the gap that exists between the fields of adsorption and carbon materials, an area that, to our knowledge, has not been encompassed so far in one single book. Several books address the phenomenon of adsorption from both a fundamental and an applied perspective, while publications on the structure, properties, and applications of carbons, either general or restricted to specific types of materials, are increasingly common. There are, also, a number of works devoted to porosity in carbons or other solids. However, adsorption is involved in many areas other than porosity characterization. In short, the interplay between adsorption and carbon materials has not been addressed yet in one volume. There is a vacuum of knowledge between both fields that, if filled, could give birth to new concepts and ideas. Adsorption cannot occur without the active and mutual participation of the adsorbent and adsorbate. Indeed, the book is purposely entitled "Adsorption by Carbons" (rather than, for instance, "Adsorption on Carbons") to emphasize the dynamic character of adsorption and the active participation of the carbonaceous adsorbent, which not only provides adsorption sites but also attracts adsorbates to its surface. The book consists of four parts. Part 1 which is the shortest (two chapters), introduces the reader to the field of adsorption by carbons and to the realm of carbon materials. The following eight chapters address the fundamental aspects of adsorption by carbons through such topics as adsorption energetics, computer simulations, modeling, surface heterogeneity, and so on. Indeed, this second part ofthe book develops a series ofconcepts that contribute to a better understanding of what follows in the third section, in which adsorption is mainly envisaged as a tool to characterize carbon surfaces. The third part, consisting of 10 chapters, begins with a look at the adsorption methods used to study the porous texture and surface chemistry of carbons, and then follows with chapters devoted to several novel types of materials to conclude with a discussion on certain aspects of the electrochemistry of adsorption by carbons. Finally, the fourth part (seven chapters) deals with the most significant technological applications of adsorption by carbons, either at the gas-solid or at the liquid-solid interface, which have direct implications especially for the fields of environment and energy. Overall, the ensemble of 27 chapters tries to cover the subject of adsorption by carbons as comprehensively as possible. In my view, the main strength of the book derives from the stature of the contributing authors. I have always thought that Eduardo Bottani and I were being a little overambitious (perhaps even daring?) in inviting the best specialists in the various areas to contribute chapters. Some were our friends, others were XXI
xxii
Preface
just acquaintances we had met at conferences, and there are still a few whom I still do not know personally. But all of them share one thing in common, and that is that they are leaders in their respective fields. The important thing is that we succeeded in our invitation, and I am very happy with the result of this collective effort. The book has an undeniably international flavor, as it includes authors from 13 different countries. Apart from a possibly slight imbalance in favor of the New World (USA and Argentina), the geographical distribution of the authors is fairly representative of the places where adsorption by carbons is being investigated. I will purposely avoid citing authors' names here, as this would make the preface outstandingly long. Let me simply thank all the authors collectively for having accepted the invitation to produce their chapters, for the care they have taken in preparing them, and for their continued willingness to help me. My thanks are extended to the staff of Elsevier, particularly lain Craig, Commissioning Editor, Kristi Green, Administrative Editor, and Sunita Sundararajan, Project Manager, for their very professional assistance at many stages of book's preparation. Last but not least, my warmest thanks go to Prof Dr Juan de Dios LopezGonzalez for having accepted to write a foreword to this book. Neither Eduardo Bottani nor I are direct descendants of the scientific school that originally spread from Granada University under his guidance. The international reputation that the Spanish-speaking scientific community enjoys in the field of adsorption by carbons owes much to the efforts of that school. Interestingly, the first papers that came to the attention of Eduardo and myself at the beginning of our research careers in the late 1970s were those of Professor Lopez-Gonzalez and coworkers. Thus we shared the same introductory academic experience in two such distant places as far apart as La Plata and Madrid! I refer, of course, to physical distance, since few countries are so close to each other emotionally as are Argentina and Spain. Juan M.D. Tascon
LIST OF CONTRIBUTORS
Alejandro J. Arvia Instituto de Investigaciones Fisicoquimicas Te6ricas y Aplicadas (INIFTA) Universidad Nacional de La Plata-Consejo Nacional de Investigaciones Cientificas y Tecnicas La Plata, Argentina
[email protected] Teresa J. Bandosz Department of Chemistry City College of New York New York, NY, USA
[email protected] Fran~ois
Beguin Centre de Recherche sur la Matiere Divisee CNRS-U niversite Orleans Cedex, France
[email protected] Henry Bock Department of Chemical and Biomolecular Engineering North Carolina State University Raleigh, NC, USA
[email protected] Hans-Peter Boehm Department of Chemistry and Biochemistry University of Munich, Germany
[email protected] Mary J. Bojan Department of Chemistry The Pennsylvania State University University Park, PA, USA
[email protected] Agustin E. Bolzan Instituto de Investigaciones Fisicoquimicas Te6ricas y Aplicadas (INIFTA) UNLP-CIC-CONICET La Plata, Argentina
[email protected] (or)
[email protected] XX111
xxiv
List of Contributors
Eduardo J. Bottani Instituto de Investigaciones Fisicoquimicas Te6ricas y Aplicadas (INIFTA) UNLP-CIC-CONICET La Plata, Argentina M. Mercedes Calbi Department of Physics Southern Illinois University Carbondale, IL, USA
[email protected] Diego Cazorla-Amoros Departamento de Quimica Inorganica Universidad de Alicante Alicante, Spain
[email protected] Milton W. Cole Department of Physics The Pennsylvania State University University Park, PA, USA
[email protected] Hans Darmstadt Departement de genie chimique Universite Laval Quebec, Canada
[email protected] Renaud Denoyel MADIREL, CNRS-Universite de Provence Marseille, France
[email protected] DuongD. Do School of Engineering University of Queensland St Lucia, Qld, Australia
[email protected] HaD. Do School of Engineering University of Queensland St Lucia, Qld, Australia
[email protected]
List of Contributors
Catherine Faur-Brasquet Ecole des Mines de Nantes Nantes cedex, France
[email protected] Elzbieta Frackowiak Institute of Chemistry and Technical Electrochemistry Poznan University of Technology Poznan, Poland fracko@fct. put.poznan.pI Silvina M. Gatica Department of Physics The Pennsylvania State University University Park, PA, USA
[email protected] Keith E. Gubbins Department of Chemical and Biomolecular Engineering North Carolina State University Raleigh, NC, USA
[email protected] Michio Inagaki Faculty of Engineering Aichi Institute of Technology Yakusa, Toyota, Japan
[email protected] Norio Iwashita National Institute of Advanced Industrial Science and Technology Onogawa, Tsukuba, Japan
[email protected] Timur S. Jakubov Department of Applied Chemistry Royal Melbourne Institute of Technology Melbourne, Australia
[email protected] J. Karl Johnson Department of Chemical and Petroleum Engineering University of Pittsburgh Pittsburgh, PA, USA National Energy Technology Laboratory Pittsburgh, PA, USA
[email protected]
xxv
xxvi
List of Contributors
Feiyu Kang Department of Materials Science and Engineering Tsinghua University Beijing, China
[email protected] Pierre Le Cloirec Ecole des Mines de Nantes Nantes cedex, France
[email protected] Angel Linares-Solano Departamento de Quimica Inorganica Universidad de Alicante Alicante, Spain
[email protected] Amelia Martinez-Alonso Instituto Nacional del Carbon, CSIC Oviedo, Spain amelia@incar. csic. es Aldo D. Migone Department of Physics Southern Illinois University Carbondale, IL, USA
[email protected] Carlos Moreno-Castilla Departamento de Quirnica Inorganica Facultad de Ciencias Universidad de Granada Granada, Spain
[email protected] Gayle Newcombe Cooperative Research Centre for Water Quality and Treatment Salisbury, South Australia Australia
[email protected] James P. Olivier Micromeritics Instrument Corp., Inc. Norcross, GA, USA
[email protected]
List of Contributors
xxvii
Jorge Pikunic Department of Chemical and Biomolecular Engineering North Carolina State University Raleigh, NC, USA
[email protected] Antonio J. Ramirez-Pastor Laboratorio de Ciencias de Superficies y Medios Porosos Universidad Nacional de San Luis, CONICET San Luis, Argentina
[email protected] Jose L. Riccardo Laboratorio de Ciencias de Superficies y Medios Porosos Uniyersidad Nacional de San Luis, CONICET San Luis, Argentina
[email protected] Franfoise Rouquerol MADIREL, CNRS-Universite de Provence Marseille, France
[email protected] Jean Rouquerol MADIREL, CNRS-Universite de Provence Marseille, France
[email protected] Ryong Ryoo National Creative Research Initiative Center for Functional Nanomaterials and Department of Chemistry (School ofMolecular Science BK21, Korea Advanced Institute of Science and Technology Daejeon, Korea
[email protected] (or)
[email protected] Roberto C. Salvarezza Instituto de Investigaciones Fisicoquimicas Te6ricas y Aplicadas (INIFTA) La Plata, Argentina
[email protected] Kenneth S.W. Sing BruneI University Uxbridge, UK
[email protected]
xxviii
List of Contributors
Shivaji Sircar Department of Chemical Engineering Lehigh University Bethlehem, PA, USA
[email protected],
[email protected] William A. Steele Department of Chemistry The Pennsylvania State University University Park, PA, USA
[email protected] Fabian Suarez-Garcia Instituto Nacional del Carbon, CSIC Oviedo, Spain
[email protected] Juan M.D. Tascon Instituto Nacional del Carbon, CSIC Oviedo, Spain
[email protected] Fernando Teran Arce Instituto de Investigaciones Fisicoquimicas Teoricas y Aplicadas (INIFTA) Universidad Nacional de La Plata-Consejo Nacional de Investigaciones Cientificas y Tecnicas La Plata, Argentina
[email protected] Masahiro Toyoda Faculty of Engineering Oita University Oita, Japan
[email protected] Eugene A. Ustinov School of Engineering University of Queensland St Lucia, Qld, Australia
[email protected] Maria E. Vela Instituto de Investigaciones Fisicoquimicas Teoricas y Aplicadas (INIFTA) Universidad Nacional de La Plata-Consejo Nacional de Investigaciones Cientificas y T ecnicas La Plata, Argentina
[email protected]
List of Contributors
Giorgio Zgrablich Laboratorio de Ciencias de Superficies y Medios Porosos Universidad Nacional de San Luis CONICET, San Luis, Argentina Departamento de Quimica Universidad Aut6noma Metropolitana-Iztapalapa Mexico D.F., Mexico
[email protected]
Jose L. Zubimendi Instituto de Investigaciones Fisicoquimicas Te6ricas y Aplicadas (INIFTA) Universidad Nacional de La Plata-Consejo Nacional de Investigaciones Cientificas y Tecnicas La Plata, Argentina
xxix
ADSORPTION ON ACTIVATED CARBON FIBERS Angel Linares-Solano and Diego Cazorla-Amoros Deptamento de Qufmica Inorganica, Universidad de Alicante, Alicante, Spain
Contents 17.1 Introduction 17.2 Preparation of ACFs 17.3 Characterization of ACFs 17.4 Some Examples of ACF Applications 17.5 Conclusions Acknowledgments References
43 1 433 43 6 447 449 449 449
17.1 INTRODUCTION
Activated carbon fibers (ACFs) are porous carbons with a fiber shape and a well-defined porous structure that can be prepared with a high adsorption capacity. Although the ACFs are very promising materials, they have not still a market as important as the activated carbons (ACs) due to their difference in production costs. The main characteristics and advantages of the ACFs are as follows [1-3]. (i) They have both high apparent surface area and adsorption capacity. (ii) They have fiber shape with a small diameter (ranging between 10 and 40 f-Lm), which is a very important characteristic for new applications requiring higher packing density (i.e., gas storage) [3]. (iii) ACFs are light materials and can be easily woven into different fabrics (i.e., cloths, felts). Adsorption by Carbons ISBN: 978-0-08-044464-2
© 2008 Elsevier Ltd.
All rights reserved.
431
43 2
Chapter 17 Adsorption on Activated Carbon Fibers
(iv) The pore size distribution of the ACFs, being essentially microporous materials, is narrow and uniform although mesoporous ACFs can also be prepared. (v) The narrow diameter essentially eliminates mass transfer limitations, the adsorption-desorption rates being very rapid. Since the ACFs are fibrous materials that can be easily molded and woven, filters can be designed that do not have the settling and channeling problems of the conventional granular and powder ACs [1]. Due to their low hydrodynamic resistance, they can be used as thin cloths for the treatment of high flow of gases, very useful for control of gas-phase pollution [2]. The development of ACF and AC cloths is closely related to that of carbon fibers (CFs). This makes that the raw materials used for the preparation of ACFs be, chronologically, the same as for CFs. Thus, in 1966, viscose and acetate cloths were, like for CFs, the first materials used to obtain ACFs [4,5]. The low yield of the ACFs, and CFs, obtained from the above precursors, oriented the research towards the seek of other raw materials for the preparation of cheaper CFs and ACFs with a higher yield. In this way, ACFs were prepared from 1970 using lignin (with the brand of Kayacarbon ALF), polyvinylchloride [6] (i.e., Saran polymer, already used to obtain ACs) and phenolic precursors [7]. The high yield and the good mechanical properties of the ACFs obtained make these precursors very useful for this application. In fact, Economy and Lin [8] developed ACFs from a phenol formaldehyde precursor, which are commercialized since 1976 under the name of Novolak. In 1980, Kuray Chemical Co. Ltd commercialized ACFs from phenolic resin under the name ofKynol 1 . The preparation of PAN-based ACFs was initiated in 1976 by Toho Rayon. Ltd2 and the use of pitch to obtain ACFs started in 1985, and these ACFs were commercialized by Osaka Gas Co. Ltd (AD'ALL) 3 . Due to the low price of the pitch and high yield of the ACFs obtained, the manufacture of pitch-based ACFs has increased considerably, the pitch being one of the main precursors of ACFs nowadays [1]. The research on ACFs is not different from what is usual for other materials. It focuses on understanding the preparation process, the characterization of the materials, and the analysis of their performance in given applications. A literature search on this topic gives us more than 600 contributions, taking into account only the papers published in journals. The research on ACFs starts more than 30 years ago regarding their preparation (as it has been detailed above). However, most of the work done concentrates mainly in the last 20 years, and it is essentially focused on their characterization and applications. Since a detailed review of all these contributions is out of the scope of this chapter, we will only make reference to the most representative works done on the aspects that will be described anon. 1 Nippon Kynol Inc., Japanese Patent 7583, 1980. 2Toho Rayon Co. Ltd., Japanese Patent 30810, 1978. 3Renoves A series, Application of AD 'ALL (Activated Carbon Fibre), Product Catalogue.
17.2
Preparation of ACFs
433
Thus, this chapter on adsorption in ACFs will cover the following sections: (i) the fundamentals on the preparation routes, (ii) the characterization by gas adsorption and other complementary new techniques, and (iii) some examples of applications of ACFs.
17.2 PREPARATION OF
ACFs
Once the precursor (i.e., pitch, polymer) is transformed into a fiber shape by a suitable spinning process and is carbonized after a proper stabilization stage, the activation of the resulting CF is needed to increase its adsorption capacity. The starting points for the activation of the CF precursor are not different from those for conventional granular or powder ACs. To prepare ACFs, the precursor and the method of preparation need to be appropriately selected. These two factors have great importance as they determine the final porous structure of the ACF. For a given precursor, the main stage determining the porous structure is the method of activation. The objective of activation is both to increase the number of pores and to increase the size of the existing ones, so that the resulting porous carbon has a high adsorption capacity. The preparation of porous CFs can be achieved by any of the following three methods: (i) the pyrolysis of appropriate precursors, (ii) the controlled carbon gasification with a reactant gas (i.e., physical activation), and (iii) the so-called chemical activation. Apt examples on the first method can be found in the work done by Oya and co-workers [9, 10]. The polymer blending technique can be very useful to create porosity during a heat treatment in an inert atmosphere of a mixture of two polymers. In this method a pyrolyzing and a carbonizing polymer are blended and, after the heat treatment, pores are formed due to the volatilization of the pyrolyzing polymer. The preparation of ACFs by physical activation includes a controlled gasification of the CFs at temperatures between 800 and 1000°C with an oxidant gas, so that carbon atoms are removed selectively. The removal of the outer and less ordered carbon atoms leads to the creation of new micropores and/or the widening of their size, which results in an increase in their pore volume. Thus, for a given precursor, the pore size distribution in the ACF depends on the preparation conditions (mainly temperature, time, and gas flow), the activating agent used, and the presence ofcatalysts. Some representative examples of the influence of the experimental conditions mentioned earlier can be found elsewhere [2, 11-21]. In order to have an efficient activation process, the reaction must take place inside the CFs, at least, predominantly compared with the reaction occurring outside. If there exists only external reaction, the obtained material does not develop porosity. However, if the reaction occurs inside the fiber, there is porosity development: the higher the amount of carbon removed, the higher the porosity development.
Chapter 17 Adsorption on Activated Carbon Fibers
434
Carbon dioxide and steam are the activating agents most commonly used, whose reactions with carbon are endothermic. These gases react with the carbon atoms in the precursor according to the following reactions: C + CO 2
*+
2CO
LliH
= 159.0kJjmol
Although the activation with carbon dioxide or steam produces essentially microporous ACFs, strong differences have been found between these two activating agents regarding the porous texture and the mechanical properties of the ACFs [12, 13]. The addition of metals such as cobalt, silver, rare earth metals, or platinum either to the starting pitch (followed by spinning, stabilization, and carbonization) or to the CF, followed by gasification with steam, allows the preparation of ACFs with significant mesoporosity [14-19, 21]. The chemical activation process consists of mixing a carbonaceous precursor with a chemical activating agent, followed by a pyrolysis stage [22-25]. The material after this stage is richer in carbon content and presents a much ordered structure and, after the thermal treatment and the removal of the activating agent, has a well-developed porous structure. Different compounds can be used for the activation; among them, KOH, NaOH, H 3 P0 4 , and ZnCl2 have been reported in the literature [22-25]. The chemical activation presents advantages over the physical one that can be summarized as follows: (i) the chemical activation uses lower temperatures and pyrolysis time, (ii) it usually consists of one stage, (iii) the yields obtained are higher, (iv) it produces highly microporous ACs and (v) it is a suitable method for applying to materials with a high ash content [23-25]. On the other hand, the chemical activation presents disadvantages such as the need of a washing stage after the pyrolysis and the corrosiveness of the chemical agents used. Although the work done on physical activation of CFs is wide, the research on chemical activation of CFs is scarce and mainly corresponds to the use of lalkaline hydroxides as activating agents [26, 27]. The chemical activation must be done under well-controlled experimental conditions in order not to destroy the fiber shape. The resulting ACFs are essentially microporous materials (i.e., pore size below 2 nm) although differences exist depending on the activating agent used and the starting CF. In the activation by hydroxide, the main variables affecting the final porous texture are hydroxide/carbon ratio, heating rate, temperature, and time ofpyrolysis. Moreover, there are two additional parameters that have recently been r~ported [24,25]: nitrogen flow rate and the washing stage (washing with water or washing with hydrochloric acid), which have an important role in porosity development. Figures 17.1 and 17.2 contain N 2 adsorption isotherms of ACFs prepared by physical and chemical activations, as examples of the results that can be
17.2
Preparation of ACFs
435
60 - - - - - - - - - - - - - - - - - - - ,
50 40 ~ (5
E E
30 20 10 o.a..--------r----r--~-....-__._-__r_-,.______r-__l
o
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
P/PO
Figure 17.1 N 2 adsorption isotherms for chemically activated samples.
60 - . - - - - - - - - - - - - - - - - - - - - - ,
50
94% burn-off
40 ~ (5
E E
30
20
o.....--,....--~----r---_r__,r__~-___r__-_r___,r---"___l
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
P/PO
Figure 17.2 N 2 adsorption isotherms for samples activated by CO 2 •
obtained [27]. The isotherms are of type I in all the cases and very high adsorption capacities (BET [stands for Brunauer, Teller, and Teller] surface areas close to 3000 m 2 I g) and very high micropore volumes (close to 1 mll g) can be reached. Only at high porosity development, the knee of the isotherms becomes wide, indicating the presence of supermicroporosity (i.e., pore size between 0.7 and 2nm) and narrow mesoporosity (i.e., between 2 and 7nm). We can see that chemical activation allows to noticeably overpass the maximum adsorption capacity reached by physical activation.
Chapter 17 Adsorption on Activated Carbon Fibers
In summary, as it happens with granular and powder carbons, a proper selection of the CF and the activation method and experimental conditions permits the preparation of ACFs with a tailored pore size distribution, with the additional advantage of their fiber shape and small diameters that allow faster mass transfer rates when compared with conventional ACs.
17.3
CHARACTERIZATION OF ACFs
Because the ACFs are porous carbons that have no significant differences compared with other porous carbons, the techniques used for their characterization are almost the same. Since the porosity in carbons is the responsible for their adsorption properties, the analysis of the different types of pores (size and shape), as well as the pore size distribution, is very important to foresee the behavior of these porous solids in final applications. We can state that the complete characterization of the porous carbons is complex and needs a combination of techniques, due to the heterogeneity in the chemistry and structure of these materials. There exist several techniques for the analysis of the porous structure, from which we can underline physical adsorption of gases, mercury porosimetry, small-angle neutron and X-ray scattering (SANS and SAXS), transmission and scanning electron microscopy (TEM and SEM), scanning tunnel microscopy (STM), immersion calorimetry, etc. Regarding the evaluation of porosity of ACs, physical adsorption of gases is, undoubtedly, the most widely used technique [28]. Different adsorptives, such as N 2 , CO 2 , Ar, He, CH 4 , benzene, and nonane, can be used for this purpose [28-39]. Due to the considerable sensitivity of nitrogen adsorption isotherms to the pore structure in both microporous and mesoporous regimes and to its relative experimental simplicity of the pore structure, measurements of subcritical nitrogen adsorption at 77 K are the most used. However, the main disadvantage of N 2 adsorption at 77 K is that when it is used for the characterization of microporous solids, diffusional problems of the molecules inside the narrow microporosity (i.e., pore size below 0.7 nm) may occur [29, 33, 34, 37]. Moreover, there is an additional experimental difficulty in the adsorption of subcritical nitrogen because very low relative pressures (10- 8 -10- 4) are required to extend the range of porosity studied to the narrow microporosity, which requires more sophisticated and expensive adsorption equipments. To overcome these problems, the use of other adsorptives has been proposed [29-31, 33-35, 37-39]. It is important to note that gas adsorption is only sensitive to the open porosity, that is, the porosity that is accessible to the adsorptive used. Thus, there are pores that do not communicate with the surroundings; such pores are called "closed pores." The closed state depends on the probe size, particularly, in the case of gas adsorption. This closed porosity is not associated with the adsorption properties and permeability of the molecules but does affect the mechanical
17.3 Characterization of ACFs
437
properties and the density of the solid material, which is of special relevance to the case of ACFs. From the point of view of gas adsorption, which is the main objective of this chapter, there are no differences in the results obtained between the ACFs and the granular and powder ACs, except for the kinetics of gas adsorption due to the special pore structure of the ACF [2, 40, 41] and for the higher packing density that can be obtained with them due to their fiber shape [3]. In this way, although there are important differences in the pore structure and distribution of porosity among the ACFs and the conventional ACs (this aspect will be described in a next section), the adsorption isotherms are not sensitive to them and do not allow to distinguish the shape of the porous carbons (i.e., fiber, granular, powder, monolith). As an example of the above statement, Fig. 17.3 contains the N 2 adsorption isotherms for powder AC with different adsorption capacities [3]. These isotherms, compared with those in Figs. 17.1 and 17.2, clearly demonstrate that the adsorption isotherms do not permit neither to distinguish the ACF from the AC nor to deduce differences in the pore size distribution. However, the unique fiber shape and porous structure of the ACF are advantages that permit to deepen into the fundamentals of adsorption in microporous solids [31]. ACFs are essentially microporous materials [13, 31], with slit-shaped pores and a quite uniform pore size distribution [42, 43]. Thus, they have simpler structures than ordinary granulated ACs [31] and can be considered as model microporous carbon materials. For this reason, important contributions to the understanding of adsorption in microporous solids for the assessment ofpore size distribution have been made using ACF [31,33,34,39, 42-46], which merit to be reviewed.
50 , - - - - - - - - - - - - - - - - - - - - - - - - ,
40
~
30
C\I
Z
(5
~
20
10
O __--r----.----~-___r__-~-___r__-_r_-___r_-~-___l o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
P/Po
Figure 17.3 N 2 adsorption isotherms for granular and powder activated carbon samples.
Chapter 17 Adsorption on Activated Carbon Fibers
In this sense, we will discuss first the usefulness of ACFs to understand micropore characterization from gas adsorption and, after that, we will focus on the latest contributions to the research on the activation-pore structure relationship for ACFs. The knowledge on the latter aspect is essential for its implication on future development of applications for the ACFs.
17.3.1 Adsorption on the ACF and Its Usefulness to Understand Micropore Characterization Due to the small size of the micropores (i.e., size below 2 nm), adsorbate filling at low relative pressures may occur. The presence of micropores in ACFs, and in most of the ACs, causes that most of the adsorption takes place within them and, at least, 90 % of the total surface area corresponds to micropores. The adsorption in microporosity is not so well understood and simple to interpret as for adsorption in mesopores or nonporous solids, which has led to an important research effort since more than 50 years ago, trying to establish experimental methods and refined models useful to explain adsorption in micropores (i.e., assessment of micropore volume and micropore size distribution [28, 35, 36, 47-49]. When the pore size is similar to the size of the adsorbate molecule and the adsorption temperature is below the critical temperature, a number of peculiarities occur that should be emphasized [50]: (i) The equilibrium time for the adsorption may be extremely long, what may be a source of error for the evaluation of microporosity. For example, this occurs for N 2 at 77 K in samples with narrow microporosity (size below 0.7 nm). When the size of the molecule is similar to the size of the pore, the activation energies for passing the molecules through the pore entrance and for the diffusion of the molecules inside the micropores can make the kinetics of adsorption very slow and a temperature-dependent process. (ii) The interaction energy between a free surface of a solid and an adsorbate molecule is rather lower than in a micropore, as a consequence of the overlap of the adsorption potential from neighboring walls. This overlap leads to a strong adsorption of the gas by the micropore and, then, to an enhancement of the heat of adsorption. (iii) The adsorption process in the micropores occurs by a volume filling mechanism rather than a surface coverage mechanism. Then, the amount adsorbed for different adsorptives (expressed as volume of liquid) at a relative pressure near unity is very similar. (iv) Surface areas ofsome well-developed ACs are unrealistically high, compared with the calculated for 1 g of carbon in the form of an extended graphite layer plane, counting both sides (2630 m 2 / g).
The first point presented above, i.e., the diffusional limitations of N 2 adsorption at 77 K, is extremely important considering that it is the most widely used adsorptive for the characterization of the porous materials, which can lead to
17.3 Characterization of ACFs
439
misinterpretations. One material that without doubt shows the unsuitability of N 2 adsorption is the CMS used for the separation of N 2 /0 2 or CH 4 /C0 2 and hence with a microporosity in the size of 0.3-0.4 nm. As an example, CF without any activation treatment does not give N 2 adsorption at 77 K but exhibits a quite fast CO 2 adsorption at temperatures above 298 K and an increasing methane uptake with increasing adsorption temperature [51]. This reflects that it is an activated process extremely sensitive to the temperature and the molecule size (0.33 and 0.38 nm for CO 2 and CH 4 , respectively [52]). Some additional examples of CMSs are collected in the following data that include the micropore volume deduced from CO 2 adsorption, measured in our laboratory, for non-activated carbon fibers obtained from different precursors (i.e., carbon fibers from coal tar pitch and from petroleum pitch). It must be noted that the N 2 adsorption is negligible for all of them, but V(C0 2 ) (ml/g) =0.18, 0.25, 0.19, and 0.05. These examples show that their microporosity cannot be evaluated through N 2 adsorption at 77 K. Consequently, this strong limitation (that occurs not only on CMSs but also on ACFs with low activation degrees - see below) needs to be solved. Two approaches can be considered taking into account the two experimental variables mentioned above, i.e., molecule size and temperature: (i) the use of an smaller size adsorbate making the adsorption at the boiling temperature and (ii) the use ofan adsorbate at temperatures higher than the boiling point; at these conditions, the kinetic energy of the molecules is increased and they can overcome the activation energy for diffusion within the porosity. He adsorption at 4.2 K has been proposed [30, 31, 38] as a promising method for the accurate determination ofmicroporosity. A He atom is the smallest atom; it has a spherical shape and interacts weakly with any solid surface [30]. He adsorption requires lower equilibrium times and the amount adsorbed is higher than in the case of N 2 at 77 K. From this research the authors concluded that the micropore analysis by N 2 adsorption at 77 K is insufficient and may give misleading conclusions [30]. In spite of the interesting results obtained with He, the experimental conditions used (adsorption at 4.2 K) make this technique not available for routine characterization of microporous solids. CO 2 adsorption, either at 273 or at 298 K [29, 33-37, 39, 49, 50], is another useful alternative for the assessment of the narrow microporosity. In such case, though the critical dimension of the CO 2 molecule is similar to that of N 2 , the higher temperature of adsorption used for CO 2 results in a larger kinetic energy of the molecules, which are able to enter into the narrow porosity. In this way, CO 2 adsorption has demonstrated that CO 2 is an appropriate complementary adsorptive for the analysis of the microporosity [37]. In the following, the usefulness of CO 2 adsorption at 273 K to achieve a rather complete characterization of the porous texture of microporous carbons will be discussed. We will base our study on the results already published [33-35, 37] in which samples with different characteristics were used and CO 2 adsorption experiments at high pressures (up to 4 MPa) were performed. In this study, the ACFs with different contents of microporosity have been very useful. The use
440
Chapter 17 Adsorption on Activated Carbon Fibers
of high pressures also was of utmost importance as it permitted the comparison of both N 2 and CO 2 adsorptions at comparable relative pressures. Since N 2 adsorption is done at 77 K and CO 2 at 273 or 298 K, the experiments cannot be directly compared, which introduces strong concerns about the similarities and differences among both adsorptives. Thus, a better way to compare the two experiments is to plot the characteristic curves [33-35, 37]. These characteristic curves, obtained applying the Dubinin-Radushkevich (DR) equation [47] to the adsorption isotherms, are the plot of the logarithm of the volume ofliquid adsorbed versus the square of the adsorption potential corrected for the affinity coefficient ((3) of the adsorptive ((AI (3)2 = (R TIn ifo If) I (3)2, T being the temperature, f the fugacity, and 10 the saturation fugacity). Figure 17.4 contains the characteristic curve for N 2 adsorption at 77 K, calculated from an adsorption isotherm within the usual relative pressure range of 10-3 - 1, and the characteristic curve for CO 2 adsorption at 273 K, deduced from the isotherm done at subatmospheric pressures. The figure shows that both characteristic curves have a similar shape and could overlap, but there is a gap of adsorption potential that is not covered with these adsorption experiments. To cover this gap and, consequently, to demonstrate that the adsorption mechanism for both adsorptives is similar, there are two possibilities. The first one is to use high pressures for CO 2 trying to cover a higher range of relative fugacities, which will allow us to analyze the whole range of porosity and the peculiarities
(AI(3)2 (kJ/mol)2 200 400 600 800 1000 1200 1400 1600 1800 2000 0+---+-----+-----+-----1--+----+----+-----+------+------1
o
o
(Alf3)2
100
200
300
400
500
-0.5 -0.5 -1
-1.5
~
:5
~
E-1.5~~
-2 -2.5 -3 -3.5 -4
Figure 17.4 Characteristic curves for an activated carbon fiber (ACF) that includes the N 2 adsorption data at 77 K (relative pressure from 10-3 to 1) (0) and the CO 2 adsorption data at 273 K at subatmospheric pressures (.).
17.3 Characterization of ACFs
44 1
(AI(3)2 (kJ/mol)2
o
o
200
800
600
400
1000
o
1200
1400
1600
(A/j3)2 200 300
100
400
1800
2000
500
-0.5 -0.5 -1
-1.5
5:' of:
s:
c- -1.5 O"'''2!-~ ............. '"n, '- ...........
-2
-..........,.......
....... '
-2.5
-3
-2.5
............ .....................
..
"--"'"-,,
, -,
........-
................
-3.5
......... , ........
.
.........
.................
......IiI...
-4-&...-------------------------1 Figure 17.5 Characteristic curves for an activated carbon fiber (ACF) that includes the N 2 adsorption data at 77 K (relative pressure from 10-3 to 1) (0), the CO 2 adsorption data at 273 K at subatmospheric pressures (.) and the CO 2 adsorption isotherm at high pressures (Li).
of CO 2 adsorption at temperatures higher than the boiling temperature (but below the critical temperature). Figure 17.5 contains the same characteristic curve presented in Fig. 17.4, but now it includes the data obtained at high CO 2 adsorption pressures. The figure clearly shows that the N 2 characteristic curve superimposes over the CO 2 one. This indicates that there is a large similarity in the mechanism by which these two gases are adsorbed in the different ranges of porosity. This characteristic curve (shown as an example by many others [33-35,37, 50]) allows us to state that CO 2 adsorption at 273 K provides similar information as adsorption of N 2 at 77 K. The second possibility to cover the gap in the characteristic curves shown in Fig. 17.4 is to use lower relative pressures for N 2 adsorption (i.e., below 10-3 and up to 10-7 ). Thus, a series of ACFs with different degrees of activation, and consequently with different degrees of microporosity, will be discussed next. In the case of an ACF (see Fig. 17.6) with very narrow microporosity, the characteristic curve for N 2 at 77 K remains always below that for CO 2 , in agreement with the kinetically restricted adsorption for N 2 in these types ofsamples. We can see that the amount of N 2 adsorbed is lower than that of CO 2 in all the relative pressure range studied. In the case of an ACF with a higher degree of burn-off (see Fig. 17.7), the characteristic curve for N 2 superimposes over that for CO 2 up to a value of (AI (3)2 higher than about 400 (k]/mol) 2. From this value on, a large deviation downward occurs. This behavior has been observed with other microporous carbon materials and, in fact, it is a common problematic feature
44 2
Chapter 17 Adsorption on Activated Carbon Fibers
(AI(3)2 (kJ/mol)2 0
0
500
1000
1500
2000
2500
3000
-2
-4
5:' ~
-6
-8
-10
-12
Figure 17.6 Characteristic curves for sample CFC14: (.) CO 2 adsorption at subatmospheric pressures; (A) CO 2 at high pressures; (~) N 2 at 77 K. (Reproduced with permission from Ref. [33]. Copyright 1996 American Chemical Society.) (AI(3)2 (kJ/mol)2 500
1000
1500
2000
2500
3000
3500
-2
-4
> ~ -6 -8
-10
-12 .......- - - - - - - - - - - - - - - - - - - - - - -
Figure 17.7 Characteristic curves for sample CFSSO: (.) CO 2 adsorption at subatmospheric pressures; (A) CO 2 at high pressures; (~) N 2 at 77 K. (Reproduced with permission from Ref. [33]. Copyright 1996 American Chemical Society.)
of the characteristic curves for N 2 adsorption in microporous carbons and hence a limitation of its use in such low relative pressures (i.e., from 10- 3 to 10- 7 ). The limitations of N 2 to characterize materials with narrow microporosity are reinforced in the next example. Figure 17.8 [33] contains a magnification
443
17.3 Characterization of ACFs
(A/~)2 (kJ/mol)2
200
o
300
400
500
600
-2 .........--..
-6
-8..1..-------------------.. . . Figure 17.8 Characteristic curves for samples CFC14, CFC40, and CFSSO in the (A/~)2 range of 200-600 (k]/mol)2 for the N 2 adsorption isotherm at 77 K. (Reproduced with permission from Re£ [33]. Copyright 1996 American Chemical Society.)
of the characteristic curve for N 2 for ACFs with different degrees of burnoff The adsorption potential, (AI (3)2, at which the large deviation downward starts depends on the burn-off of the sample as it is clearly shown in Fig. 17.8. These deviations that happen at adsorption potentials (A/~)2 between 250 and 400 (k]/mol)2 (i.e., at relative pressures of N 2 lower than 10- 5 for the sample with a 50 % burn-off [BO] for CFS50- and lower than 10- 4 for the sample with a 14 % BO for CFC14) show that N 2 adsorption in the narrow microporosity is influenced by diffusional limitations. With increasing burn-of£ the porosity widens, the accessibility ofN2 increases, and hence the relative pressure at which N 2 can enter into the porosity decreases. It must be noted that the large deviation downward has also been related by other authors to a change in the adsorption mechanism at very low relative pressures [53]. Usually, many researchers calculate the pore size distribution from the adsorption isotherms applying a model, which accounts for the heterogeneity of the pore size distribution. Thus, the general adsorption isotherm [54] is used in which different pore size distribution functions and local adsorption isotherms can be used. If the DR or Dubinin-Astakhov (DA) equations are used and a Gaussian distribution of micropores is assumed, simple equations are obtained which may show, at least semiquantitatively, the pore size distribution. This can be a direct and simple way to compare the two adsorptives analyzed (i.e., N 2 and CO 2). If the pore size distributions obtained from the two adsorptives are similar, the commented similarities between CO 2 and N 2 should be reinforced. Figure 17.9 presents the pore size distributions obtained for the sample with the widest micropore distribution (sample CFS50, Fig. 17.7), applying the DubininStoeckli (DS) equation [28] to the CO 2 and N 2 data. It must be remembered that because the high-pressure CO 2 adsorption data provide information for the whole range of relative fugacities, this type of calculations and comparisons are
444
Chapter 17 Adsorption on Activated Carbon Fibers
o
2
3
4
L (nm)
Figure 17.9 Pore size distributions obtained for the sample CFSSO applying the DubininStoeckli (DS) equation to the CO 2 and N 2 adsorption data.
straightforward. Figure 17.9 shows that there is a good concordance between both distributions, confirming the validity of CO 2 to characterize the microporosity and its reliability at subatmospheric pressures to characterize the narrow microporosity. The mean pore size for the CO 2 distribution is somewhat smaller than for N 2 , which can be a consequence of the downward deviation observed for the characteristic curves for N 2. These results obtained with ACFs are important for their relevance in the use of N 2 adsorption in the characterization of porosity. To "measure" the narrow microporosity with N 2 at 77 K, low relative pressures (i.e., from 10-4 to 10-7 , i.e., high adsorption potentials) must be used. These low relative pressures need more sophisticated and expensive adsorption equipments and cannot be reached with conventional ones. Additionally, as a consequence ofthe diffusional limitations, N 2 adsorption at 77 K cannot be used to determine the micropore volume of the narrowest porosity. All this makes necessary the use of other adsorptives to analyze this range of porosity. The research done shows that He adsorption at 4.2 K [30, 31, 38] or CO 2 adsorption at 273 K at subatmospheric pressures [33-35, 37, 50] can be used for this purpose, although the second one is more convenient from an experimental point of view. In fact, because the adsorption temperature used for CO 2 adsorption is 273 K, the saturation pressure for this gas is high and, hence, the relative pressures are low (about 10-4). These low relative pressures can be easily reached with conventional equipments working up to 0.1 MPa, avoiding also the additional diffusional limitations that happen with N 2 •
17.3.2 Understanding the Activation-Pore Structure Relationship of ACFs: Effect of Activating Agent and Burn-Off Degree An additional issue of special relevance for the characterization of the ACFs is their fiber shape, since it introduces differences on the porous texture
17.3 Characterization of ACFs
445
compared with the conventional ACs. In fact, it is not only important to determine the pore volume and pore size distribution of the ACF, but also the distribution of the pores across the fiber diameter, which will be a result of the activation process (i.e., activation temperature, activating agent, and CF used). Although a remarkable effort has been done on the porosity analysis and structural characterization of the ACF [55-57], the research on pore distribution within the fibers is scarce and very recent [58-60]. We have dedicated an important effort to analyze the development ofporosity in isotropic pitch-based CFs [13, 58-63]. The previous studies showed that the ACFs had a different evolution of porous structure with burn-off depending on the activating gas used (steam and CO 2 ) [13, 61-63]. From the N 2 and CO 2 adsorption data at 77 and 273 K, respectively, and the measurements of tensile strength and fiber diameter [13], it was shown that CO 2 essentially develops narrow microporosity (size lower than 0.7 nm) and causes a steady decrease in the tensile strength with burn-o~ while the fiber diameter does not change significantly. In comparison, steam activation results in a wider pore size distribution and, after the initial stages of the activation, the tensile strength remains nearly constant and the fiber diameter decreases. From these results, we speculated that CO 2 activation deepens into the bulk of the fiber, whereas steam mainly focuses the activation in the external parts. The confirmation of this interpretation can only be carried out analyzing the porosity development across the fiber diameter, on a single CF, using a technique with a lateral resolution of some micrometers. SAXS can be a useful technique for this purpose, provided that a micrometersize beam with a sufficient intensity is available. It must be noted that SAXS technique gives information about the structure of matter on a micro- and mesoscopic scale; then, SAXS can be used to characterize the porosity of ACs in a size range from few angstrom to about 2000 A. The use of this technique for that purpose is becoming more important, as reflected in the considerable number of papers published on this subject [62, 64-73]. This growing use is due to the fact that, unlike gas adsorption techniques, the SAXS intensity profile is also sensitive to the shape and orientation of the scattering objects so that additional information can be obtained. The availability at ESRF of X-ray microbeams with sizes down to less than 2 f.,Lm (Microfocus Beamline, ID13), together with a position-resolved X-ray scattering method makes the technique named as f.,L-SAXS suitable for analyzing single fibers [74, 75]. Experiments demonstrating the successful use of this technique in fibrous materials were carried out, e.g. on cellulose [74, 75] and CFs [76]. In the work carried out with CFs [76] the internal structure of single CFs from different precursors (PAN-based fiber and mesophase-pitch based fibers) was investigated. In this way, we have applied for the first time the f.,L-SAXS technique to the characterization of a single ACF and we have shown the usefulness of the technique to measure a single ACF and to get positionresolved information on the pore distribution across the fiber diameter [58-60]. In the following, main conclusions obtained will be briefly reviewed.
446
Chapter 17 Adsorption on Activated Carbon Fibers
(a)
(b)
1.E-2
0.1
(c)
10 Intensity
100
1000
Figure 17.10 Two-dimensional scattering patterns corresponding to (from left to right): (a) CF, (b) CFC29, and (c) CFC50. (Reproduced with permission from Re£ [58]. Copyright 2002 Elsevier Science Ltd.)
Figure 17.10 presents the two-dimensional scattering patterns corresponding to the original CF and ACF with two degrees of activation (samples CFC29 and CFC50), once the background has been subtracted. Samples CFC29 and CFC50 are ACFs prepared by CO 2 activation up to 29 and 50 % burn-off: respectively. These two-dimensional scattering patterns have been plotted for a maximum value of scattering vector (q = 41T sin(8)jA.) of 10 nm- 1 and correspond to measurements performed in the center of the fibers. This figure shows that the original CF, which cannot be characterized by N 2 adsorption, has some scattering, which is low, as it corresponds to its low micropore volume determined by CO 2 adsorption at 273 K (V = 0.18 ml/g). The comparison of the scattering pattern of the original CF (Fig. 17.10(a)) and the ACFs (Fig. 17.10(b) and (c)) shows that the scattering intensity increases with the bum-off degree, which agrees with gas adsorption results. These measurements clearly show the usefulness of the technique for the characterization of single fibers, even though they can have low porosity (which is the case of the original CF). Once the suitability of f..L-SAXS technique for the characterization of the porous texture of a single CF is demonstrated, the next step is to use the microscopic position resolution of this technique for characterizing the porosity development across the fiber diameter for CO 2 -and steam-activated CFs. Thus, the second series of experiments carried out in this work consisted in obtaining the scattering measurements in different regions of the same fiber, across the fiber diameter. From the scattering results, the Porod invariant (PI), which is a parameter related with the porosity development, was estimated for each scattering measurement [77]. From these calculations, the pore distribution across the fiber diameter could be deduced. The results showed that the scattering profiles, as a
17.4 Some Examples of ACF Applications
447
function of the position of the fibers, are different for CO 2-and steam-activated materials. In the case of steam, the scattering from all the internal zones is very similar and much lower than from the external part of the fiber. On the other hand, in the case of CO 2, the porosity is much more developed in the center compared with the steam-activated CFs. All the J.1-SAXS results indicate that CO 2 activation produces a more extensive porosity development across the fiber diameter than steam that focuses in the external zone of the fibers, which agrees with the decrease in fiber diameter for steam activation compared with CO 2 [13]. This means that CO 2 molecules penetrate more easily than H 20 into the carbon matrix. These results, obtained due to the fiber shape, provide the first direct proof of the different behavior of CO 2 and steam as activating agents and of the different pore distribution that produces the activation with these gases.
17.4
SOME ExAMPLES OF ACF ApPLICATIONS
Since the applications of porous carbons are treated in detail in Part IV of this book, in this section we will provide a list of applications trying to emphasize the advantages of the ACF over the conventional AC. There are a large number of applications in which the performance of the ACF has been extensively analyzed. They range from conventional gas-and liquid-phase adsorption to antibacterial and energy storage uses. Anon, we will mention some of them:
(i) Gas-phase adsorption: The use of ACFs for gas-phase adsorption has been extensively studied since it is one of the main uses of ACs. The advantage of the ACF over the conventional AC is a consequence of their porous structure. As we have already mentioned, the microporous ACFs essentially contain microporosity that is readily available to the adsorbate, and the mass transfer limitations due to the diffusion within meso-and macroporosity are absent [1, 2], making the adsorption process faster. Moreover, the ACFs avoid the settling and channeling effects of the ACs [1,2]. Consequently, a large number of works on the use of ACFs for gas-phase adsorption can be found. As an example we can mention the adsorption of gases such as S02 [78,79], NO x [80, 81], and VOC [2,41]. (ii) Liquid-phase adsorption: Liquid-phase adsorption on ACFs has similar advantages as gas-phase adsorption and their use for drinking water treatment and removal of organic compounds is well documented [2, 40, 41]. Essentially, the main advantage is the faster adsorption rate when compared with the AC. (iii) Carbon molecular sieves (CMSs): CFs and ACFs have been studied for the preparation and use as CMSs for the separation of gases such as C0 2 /CH 4
448
(iv)
(v)
(vi)
(vii)
Chapter 17 Adsorption on Activated Carbon Fibers
and Nz/O z [51, 82-84]. It was shown that the narrow microporosity of the nonactivated CF can be useful for this application and that the porosity of the ACF can be tailored by cracking of hydrocarbons to develop good quality CMSs. Antibacterial use of ACFs: Related to water treatment, Oya and coworkers developed for the first time antibacterial ACF consisting on Ag-containing samples [15-18]. ACFs as catalyst and catalyst support: There are interesting examples on the use of ACFs as catalyst and catalyst support [85-88]. The use of ACFs for this purpose is a consequence of the application of carbon materials in catalysis for decades [89]. The versatility of the carbon materials (i.e., porosity and surface chemistry), together with the special properties of the ACF, makes these materials rather interesting, and the number of papers on this subject has increased considerably since 1997. An interesting example is the oxidation of SOz to H zS0 4 by Oz and in the presence of water [85, 88]. ACF as catalyst support of different metal catalysts was studied analyzing the influence of the catalyst dispersion and activity (i.e., Refs [87] and [88]). Gas storage (CH 4 and Hz): The storage of gases such as methane and hydrogen is of special interest for their use as fuel in conventional applications. They both have considerable advantages from an environmental point of view. The main limitation for the use of these gases is their storage in the onboard fuel tank with a sufficiently high energy density. One possibility is the storage of these gases in carbon materials, which is a subject of strong research effort. The research is much more advanced in the case of methane [3, 90]. The use of porous carbons for vehicular fuel and for large-scale transportation of methane is a consequence of the enhanced adsorption of this gas within the microporosity. The advantage of the ACFs over ACs is their essentially microporous character and higher packing density, although the larger price of the ACF is detrimental [3, 90]. In the case of Hz, the limitation is much stronger than with methane due to its very low boiling point. The work done on adsorption of Hz in ACFs and ACs, shows that the highest values of hydrogen adsorption are close to 1 wt % at 10 MPa [91], a rather low value compared with the targeted 6.5 wt % value. Supercapacitors: The use of ACFs (mesoporous and microporous ACFs) as supercapacitors has been reported elsewhere [92]. The interest of mesoporosity instead ofmicroporosity for this application in propylene carbonate electrolytes (i.e., non-aqueous electrolytes) is a consequence of the size of the ions that cannot enter into the narrowest microporosity. Then, all the available surface is not used for the double-layer formation resulting in a lower capacitance (EDLC) than expected from the apparent surface area of the porous carbon. Additionally, the role of the surface chemistry on the EDLC is very important for this application.
References
17.5
449
CONCLUSIONS
ACFs are fibrous microporous carbons that can easily be activated, rendering adsorbents with high adsorption capacities and high surface areas, as happens with classical ACs. Although their adsorption behavior does not differ from other forms of microporous carbons, their fiber shape allows to get better performance in some applications. For this reason, ACFs can be considered, in relation to the classical ACs, as a new and more powerful generation of microporous carbons. The microporosity of ACFs, as also happens with other forms of ACs, is responsible for most of their applications (i.e., volatile organic compound [VOC] removal, gas separation, methane and hydrogen storage). For some of the abovementioned applications, narrow microporosity (micropore < 0.7 nm) is a key factor to understand their behavior, to enhance their performance and applications and to improve their preparation process. Consequently, the microporosity and particularly the narrow microporosity need a correct characterization microporosity that has to be done by using physical adsorption. N 2 at 77 K is the adsorptive most widely used. However, for CFs and ACFs (with low degree of activations) that have CMS properties, N 2 at low relative pressures and low adsorption temperature (77 K) is worthless because it presents diffusional adsorption problems, and hence a lack of adsorption equilibrium, in narrow mlcropores. CO 2 at 273 K is a useful adsorptive that has an adsorption mechanism quite similar to N 2 (confirmed by the high-pressure adsorption data). Its use is strongly recommended as a complement to N 2 at 77 K to perform a correct characterization of the microporosity.
ACKNOWLEDGMENTS The authors thank MCYT (Proj. PPQ2003-03884) for financial support.
REFERENCES 1. Donnet, J.B. and Bansal, R.Ch. (1990). Carbon Fibers. Marcel Dekker. 2. Suzuki, M. (1994). Activated carbon-fiber - fundamentals and applications. Carbon,
32,577-86. 3. Lozano-Castello, D., Cazorla-Amoros, D., and Linares-Solano, A. (2002). Powdered activated carbons and activated carbon fibers for methane storage: a comparative study. Energy Fuels, 16, 1321-8.
450
Chapter 17 Adsorption on Activated Carbon Fibers
4. Doying, E.G. (1965). US Patent 3,256,206. Union Carbide Corporation. 5. Rodgers, S.J., Udavcak, R.J., and Mausteller, J.W. (1965). Development of medium flow rate, high-purity, particulate, and iodine filters. Intern. Symp. Fission Prod. Release Transport Accident Conditions, Oak Ridge, Tenn., 1204-15. 6. Boucher, E.A., Cooper, R.N., and Everett, D.H. (1970). Preparation and structure of saran-carbon fibres. Carbon, 8, 597. 7. Arons, G.N. and Macnair, R.N. (1972). Activated carbon fiber and fabric achieved by pyrolysis and activation of phenolic precursors. Text. Res. J., 42, 60. 8. Lin, R.Y. and Economy, J. (1973). The preparation and properties of activated carbon fibers derived from phenolic Precursor, Appl. Polym. Symp., 21, 143-152. 9. Oya, A. (2003). Polymer blend technique for designing carbon materials. In Carbon Alloys (E. Yasuda et aI., eds). Carbon Alloys. Elsevier Science Publishers B.V., pp. 129-42. 10. Oya, A. and Kasahara, N. (2000). Preparation of thin carbon fibers from phenolformaldehyde polymer micro-beads dispersed in polyethylene matrix. Carbon, 38, 1141-4. 11. Ko, T.H., Chiranairadul, P., Lu, C.K., et ai. (1992). The effects of activation by carbon-dioxide on the mechanical-properties and structure of pan-based activated carbon-fibers. Carbon, 30, 647-55. 12. Ryu, S.K., Jin, H., Gondy, D., et ai. (1993). Activation of carbon fibres by steam and carbon dioxide. Carbon, 31, 841-2. 13. Alcaniz-Monge, J., Cazorla-Amoros, D., and Linares-Solano, A. (1994). Effect of the activating gas on tensile-strength and pore structure of pitch-based carbonfibers. Carbon, 32, 1277-83. 14. Iacocca, R.G. and Duquette, D.J. (1993). The catalytic effect of platinum on the oxidation of carbon fibres. J. Mater. Sci., 28, 1113-19. 15. Oya, A., Yoshida, S., Abe, Y., et al. (1993). Antibacterial activated carbon fiber derived from phenolic resin containing silver nitrate. Carbon, 31, 71-3. 16. Oya, A., Wakahara, T., and Yoshida, S. (1993). Preparation of pitch-based antibacterial activated carbon fiber. Carbon, 31, 1243-7. 17. Oya, A., Kimura, M., Sugo, T., et ai. (1994). Antibacterial activated carbon fiber derived from methyl methacrylate-grafted phenolic resin fiber. Carbon, 32, 107-10. 18. Oya, A., Yoshida, S., Alcafiiz-Monge, J., and Linares-Solano, A. (1996). Preparation and properties of an antibacterial activated carbon fiber containing mesopores. Carbon, 34, 53-7. 19. Tarnai, H., Ikeuchi, M., Kojima, S., and Yasuda, H. (1997). Extremely large mesoporous carbon fibers synthesized by the addition of rare earth metal complexes and their unique adsorption behaviors. Adv. Mater., 9, 55-8. 20. Blanco-Lopez, M.C., Martinez-Alonso, A., and Tascon J.M.D. (2000). N 2 and CO 2 adsorption on activated carbon fibres prepared from Nomex chars. Carbon, 38, 1177-82. 21. Wan, Y.Z., Wang, Y.L., Cheng, G.X., et ai. (2001). Preparation and characterization of activated carbon fiber supporting silver-loaded mesoporous molecular sieves. Carbon, 39, 1605-16. 22. Bansal, R.Ch., Donnet, J., and Stoeckli, F. (1998). Active Carbon. Marcel Dekker. 23. Illan-Gomez, M.J., Garcia-Garcia, A., Salinas-Martinez de Lecea, C., and LinaresSolano, A. (1996). Activated carbons from Spanish coals. 2. Chemical activation. Energy Fuels, 10, 1108-14.
References
45 1
24. Lozano-Castello, D., Lillo-Rodenas, M.A., Cazorla-Amoros, D., and LinaresSolano, A. (2001). Preparation of activated carbons from Spanish anthracite. I. Activation by KOH. Carbon, 39, 741-9. 25. Lillo-Rodenas, M.A., Lozano-Castello, D., Cazorla-Amoros, D., and LinaresSolano, A. (2001). Preparation of activated carbons from Spanish anthracite II. Activation by NaOH. Carbon, 39, 751-9. 26. Lu, A.H. and Zheng, J.T. (2001). Study of microstructure of high surface area polyacrylonitrile activated carbon fibers. J. Colloid Interface Sci., 236, 369-74. 27. Macii-Agull6, J.A., Moore, B.C., Cazorla-Amor6s, D., and Linares-Solano, A. (2004). Activation of coal tar pitch carbon fibres: physical activation vs chemical activation. Carbon., 42, 1367-70. 28. Rouquerol, F., Rouquerol,J., and Sing, K. (1999). Adsorption by Powders and Porous Solids. Principles, Methodology and Applications. Academic Press. 29. Rodriguez-Reinoso, F. and Linares-Solano, A. (1988). Microporous structure of activated carbons as revealed by adsorption methods. In Chemistry and Physics of Carbon, Vol. 28 (P.A. Thrower, ed.). Marcel Dekker, p. 1. 30. Kaneko, K., Setoyama, N., and Suzuki, T. (1994). Ultramicropore characterization by He adsorption. In Characterization of Porous Solids III (J. Rouquerol et aI., eds). Elsevier Science Publishers B.V, p. 593. 31. Setoyama, N., Ruike, M., Kasu, T., et al. (1993). Surface characterization of microporous solids with He adsorption and small angle X-ray scattering. Langmuir, 9,2612-17. 32. Sosin, K.A., Quinn, D.F., and MacDonald J.A.F. (1996). Changes in PSD of progressively activated carbons obtained from their supercritical methane isotherms. Carbon, 34, 1335-41. 33. Cazorla-Amor6s, D., Alcafiiz-Monge, J., and Linares-Solano, A. (1996). Characterization of activated carbon fibers by CO 2 adsorption. Langmuir, 12, 2820-4. 34. Cazorla-Amor6s, D., Alcafiiz-Monge, J., de la Casa-Lillo, M.A., and LinaresSolano, A. (1998). CO 2 as an adsorptive to characterize carbon molecular sieves and activated carbons. Langmuir, 14, 4589-96. 35. Lozano-Castello, D., Cazorla-Amoros, D., Linares-Solano, A., and Quinn, D.F. (2002). Micropore size distributions of activated carbons and carbon molecular sieves assessed by high-pressure methane and carbon dioxide adsorption isotherms. J. Phys. Chern. B, 106, 9372-9. 36. Sweatman, M.B. and Quirke, N. (2001). Characterization of porous materials by gas adsorption at ambient temperatures and high pressure. J. Phys. Chern. B, 105, 1403-11. 37. Lozano-Caste1l6, D., Cazorla-Amor6s, D., and Linares-Solano, A. (2004). Usefulness of CO 2 adsorption at 273 K for the characterization of porous carbons. Carbon., 42, 1233-42. 38. Setoyama, N., Kaneko, K., and Rodriguez-Reinoso, F. (1996). Ultramicropore characterization of microporous carbons by low temperature helium adsorption. J. Phys. Chern., 100, 10331-6. 39. Guillot, A., Stoeckli, F., and Bauguil, Y. (2000). The microporosity of activated carbon fibre KF1500 assessed by combined CO 2 adsorption and calorimetry techniques and by immersion calorimetry. Adsorp. Sci. Technol., 18, 1-14. 40. Shmidt, J.L., Pimenov, A.V., Lieberman, A.I., and Cheh, H.Y. (1997). Kinetics of adsorption with granular, powdered, and fibrous activated carbon. Sep. Sci. Technol., 32, 2105-14.
45 2
Chapter 17 Adsorption on Activated Carbon Fibers
41. Brasquet, C. and Le Cloirec, P. (1997). Adsorption onto activated carbon fibers: application to water and air treatments. Carbon, 35, 1307-13. 42. Kaneko, K., Shimizu, K., and Suzuki, T. (1992). Intrapore field-dependent micropore filling of supercritical N 2 in slit-shaped micropores.]. Chem. Phys., 97, 8705-11. 43. Matsumoto, A., Kaneko, K., and Ramsey,J.D.F. (1993). Neutron scattering investigations of the structure and adsorption properties of activated carbon fibers. Stud. Surf. Sci. Catal., 80, 405-12. 44. Kobayashi, N., Enoki, T., Ishii, C., et al. (1998). Gas adsorption effects on structural and electrical properties of activated carbon fibers.]. Chem. Phys., 109, 1983-90. 45. Li, Z., Kruk, M.,Jaroniec, M., and Ryu, S.-K. (1998). Characterization ofstructural and surface properties of activated carbon fibers.]. Colloid Interface Sci., 204, 151-6. 46. Alcaniz-Monge, J., Linares-Solano, A., and Rand, B. (2001). Water adsorption on activated carbons: study of water adsorption in micro- and mesopores.]. Phys. Chem. B, 105, 7998-8006. 47. Dubinin, M.M. (1960). The potential theory of adsorption of gases and vapors for adsorbents with energetically non uniform surfaces. Chem. Rev., 60, 235-41. 48. Dubinin, M.M. (1966). Porous structure and adsorption properties of active carbons. In Chemistry and Physics of Carbon (P.L. Walker, ed.). Marcel Dekker, 51-120. 49. Ravikovitch, P.I., Vishnyakov, A., Russo, R., and Neimark, A.V. (2000). Unified approach to pore size characterization of microporous carbonaceous materials from N 2 , Ar, and CO 2 adsorption isotherms. Langmuir, 16, 2311-20. 50. Linares-Solano, A., Salinas-Martinez de Lecea, C., Alcaniz-Monge,J., and CazorlaAmoros, D. (1998). Further advances in the characterization of microporous carbons by physical adsorption of gases. Tanso, 185, 316-25. 51. de la Casa-Lillo, M.A., Alcaniz-Monge, J., Raymundo-Pinero, E., et al. (1998). Molecular sieve properties of general-purpose carbon fibres. Carbon, 36, 1353-60. 52. Breck, D.W. (1974). Zeolite Molecular Sieves. John Wiley, p. 636. 53. Ehrburger-Dolle, E. (1997). Analysis of the derived curves of adsorption isotherms. Langmuir, 13, 1189-98. 54. McEnaney, B., Mays, T.J., and Causton, P.D. (1987). Heterogeneous adsorption on microporous carbons. Langmuir, 3, 695-9. 55. Donnet, J.B., Qin, R.Y., Park, S.J., et al. (1993). Scanning tunneling microscopy study of activated carbon fibers. J. Mater. Sci., 28, 2950-4. 56. Huang, Z.H., Kang, F.Y., Huang, W.L., et al. (2002). Pore structure and fractal characteristics of activated carbon fibers characterized by using HRTEM.]. Colloid Interface Sci., 249, 453-7. 57. Nakayama, A., Suzuki, K., Enoki, T., et al. (1996). Electronic and magnetic properties of activated carbon fibers. Bull. Chem. Soc. ]pn., 69, 333-9. 58. Lozano-Castello, D., Raymundo-Pinero, E., Cazorla-Amoros, D., et al. (2002). Characterization of pore distribution in activated carbon fibers by microbeam small angle X-ray scattering. Carbon, 40, 2727-35. 59. Lozano-Castello, D., Raymundo-Pinero, E., Cazorla-Amoros, D., et al. (2002). Microbeam small angle X-ray scattering (f,.LSAXS): a novel technique for the characterization of activated carbon fibers. Stud. Surf. Sci. Catal., 144, 51-8. 60. Lozano-Castello, D., Cazorla-Amoros, D., and Linares-Solano, A. (2003). Microporous solid characterization: use of classical and "new" techniques. Chem. Eng. Technol., 26, 852-7.
References
453
61. Alcafiiz-Monge, J., Cazorla-Amoros, D., and Linares-Solano, A. (1997). Production of activated carbons: use of CO 2 versus H 2 as activating agent. A reply to a letter from P.L. Walker, Jr. Carbon, 35, 1665-8. 62. Cazorla-Amoros, D., Salinas-Martinez de Lecea, C., and Alcafiiz-Monge, J. (1998). Characterization of activated carbon fibers by small angle X-ray scattering. Carbon, 36,309-12. 63. Lozano-Castello, D., Cazorla-Amoros, D., and Linares-Solano, A. (2000). Characterization of activated carbon fibers by positron annihilation lifetime spectroscopy (PALS). Stud. Surf. Sci. Catal., 128, 523-32. 64. Gardner, M.A., North, A.N., Dore, J.C., et al. (1994). Characterization of pore size in activated carbons by small-angle X-ray scattering. Stud. Surf. Sci. Catal., 87, 273-81. 65. Renouprez, A. and Avom,J. (1988). Characterization of active carbons from palmtree fibers using nitrogen adsorption and small angle X-ray scattering. Stud. Surf. Sci. Catal., 39, 49-54. 66. Foster, M.D. and Jensen, K.F. (1991). SAXS investigation of model carbon pore structure and its change with gasification. Carbon, 29, 271-82. 67. Guet,J.M., Lin, A., Linares-Solano, A., and Salinas-Martinez de Lecea, C. (1991). Characterization of activated carbon: an approach to the activation process by SAXS and optical microscopy. Stud. Surf. Sci. Catal., 62, 379-87. 68. Gupta, A. and Harrison, I.R. (1994). Small-angle X-ray scattering (SAXS) in carbonized phenolic resins. Carbon, 32, 953-60. 69. Nakagawa, T., Nishikawa, K., and Komaki, I. (1999). Change of surface fractal dimension for witbank coal with heat-treatment studied by small angle X-ray scattering. Carbon, 37, 520-2. 70. Li, X.K., Liu, L., Li, Z.H., Wu, D., and Shen, S.D. (2000). The characterization of ultrafine carbon powders by SAXS and Raman spectra. Carbon, 38, 623-41. 71. Diduszko, R., Swiatkowski, A., and Trznadel, B.J. (2000). On surface ofmicropores and fractal dimension ofactivated carbon determined on the basis ofadsorption and SAXS investigations. Carbon, 38, 1153-62. 72. Hall, P.J., Brown, S., Fernandez, J., and Calo, J.M. (2000). The effects of the electronic structure of micropores on the small angle scattering of X-rays and neutrons. Carbon, 38, 1257-9. 73. Cohaut, N., Blanche, C., Dumas, D., et al. (2000). Small-angle X-ray scattering (SAXS) in carbonized phenolic resins. Carbon, 38, 1391-400. 74. Milller, M., Czihak, C., Vogl, G., et al. (1998).Direct observation of microfibril arrangement in a single native cellulose fiber by microbeam small-angle X-ray scattering. Macromolecules, 31, 3953-7. 75. Miiller, M., Czihak, C., Burghammer, M., and Riekel, C. (2000). Combined X-ray microbeam small-angle scattering and fibre diffraction experiments on single native cellulose fibres. J. Appl. Crystallogr., 33, 817-19. 76. Paris, 0., Loidl, D., Peterlik, H., et al. (2000). The internal structure of single carbon fibers determined by simultaneous small-and wide-angle X-ray scattering. J. Appl. Crystallogr., 33, 695-9. 77. Antxustegi, M.M., Hall, P.J., and Calo, J.M. (1998). Development of porosity in Pittsburgh no. 8 coal char as investigated by contrast- matching small-angle neutron scattering and gas adsorption techniques. Energy Fuels, 12, 542-6.
°
454
Chapter 17 Adsorption on Activated Carbon Fibers
78. Raymundo-Pinero, E., Cazorla-Amoros, D., Salinas-Martinez de Lecea, C., and Linares-Solano, A. (2000). Factors controlling the S02 removal by porous carbons: relevance of the S02 oxidation step. Carbon, 38, 335-44. 79. Kisamori, S., Mochida, I., and Fujitsu, H. (1994). Roles of surface oxygen groups on poly(acrylonitrile)-based active-carbon fibers in S02 adsorption. Langmuir, 10, 1241-5. 80. Yang, C.M. and Kaneko, K. (2002). Nitrogen-doped activated carbon fiber as an applicant for NO adsorbent.]. Colloid Interface Sci., 255, 236-40. 81. Shirahama, N., Moon, S.H., Choi, K.H., et ai. (2002). Mechanistic study on adsorption and reduction ofN0 2 over activated carbon fibers. Carbon, 40, 2605-11. 82. Kawabuchi, Y., Kawano, S., and Mochida, I. (1996). Molecular sieving selectivity of active carbons and active carbon fibers improved by chemical vapour deposition of benzene. Carbon, 34, 711-17. 83. de la Casa-Lillo, M., Moore, B.C., Cazorla-Amoros, D., and Linares-Solano, A. (2002). Molecular sieve properties obtained by cracking of methane on activated carbon fibers. Carbon, 40, 2489-94. 84. Villar-Rodil, S., Martinez-Alonso, A., and Tascon, J.M.D. (2002). Carbon molecular sieves for air separation from Nomex aramid fibers.]. Colloid Interface Sci., 254, 414-6. 85. Mochida, I., Kuroda, K., Miyamoto, S., et ai. (1997). Remarkable catalytic activity of calcined pitch based activated carbon fiber for oxidative removal of S02 as aqueous H 2S0 4 , Energy Fuels, 11, 272-6. 86. Raymundo-Pinero, E., Cazorla-Amoros, D., and Linares-Solano, A. (2003). The role of different nitrogen functional groups on the removal of S02 from flue gases by N-doped activated carbon powders and fibres. Carbon, 41, 1925-32. 87. Macias-Perez, M.C., Salinas Martinez de Lecea, C., and Linares-Solano, A. (1997). Platinum supported on activated carbon cloths as catalyst for nitrobenzene hydrogenation. Appl. Catal. A Gen., 151,461-75. 88. de Miguel, S.R., Villella, J.I., Jablonski, E.L., et ai. (2002). Preparation of Pt catalysts supported on activated carbon felts (ACF). Appl. Catal. A Gen., 232, 237-46. 89. Radovic, L.R. and Rodriguez-Reinoso, F. (1997). Carbon materials in catalysis. Chern. Phys. Carbon, 25, 243-358. 90. Lozano-Castello, D., Alcaniz-Monge, J., de la Casa-Lillo, M., et al. (2002). Advances in the study of methane storage in porous carbonaceous materials. Fuel, 81, 1777-803. 91. de la Casa-Lillo, M., Lamari-Darkrim, F., Cazorla-Amoros, D., and Linares-Solano, A. (2002). Hydrogen storage in activated carbons and activated carbon fibers. ] Phys. Chern. B, 106, 10930-4. 92. Shiraishi, S. (2003). Polymer blend technique for designing carbon materials. In Carbon Alloys (E. Yasuda et aI., eds). Elsevier Science Publishers B.V., pp. 447-57.
OVERVIEW OF PHYSICAL ADSORPTION BY CARBONS Kenneth S.W. Sing Brunei University, Uxbridge, UK
Contents Introduction Physisorption on Nonporous Carbons 1.3 Physisorption by Porous Carbons 1.4 Concluding Remarks References
1.1
1.2
3 5 7 11 12
1.1 INTRODUCTION Physical adsorption (physisorption) phenomena were already well-known in the early years of the twentieth century, when various attempts were made to explain the underlying principles [1]. Some investigators followed Polanyi [2] in picturing the adsorbed state (the adsorbate) as a thick multimolecular film under the influence oflong-range forces emanating from the surface. Others were more strongly influenced by the monumental work of Langmuir [3] and considered that the adsorbate was normally in the form of a monolayer. A third approach, which was based on the application of the Kelvin equation, drew attention to the role of capillary condensation [4, 5]. The interest in physisorption was further strengthened in 1938 by the publication of the Brunauer-Emmett-Teller (BET) theory ofmultilayer adsorption [5, 6]. Over the past 50 years the BET theoretical model has been subjected to a considerable amount of criticism [7], but the BET method has remained the most popular procedure for determining the surface area ofadsorbents, catalysts, and various other porous and finely divided materials. Adsorption by Carbons
ISBN: 978-0-08-044464-2
© 2008 Elsevier Ltd.
All rights reserved.
3
4
Chapter 1 Overview of Physical Adsorption by Carbons
Langmuir had mentioned in his 1916 paper [3] that it was inconceivable that a well-defined monolayer could be formed in very narrow, molecular-sized pores of an activated carbon, but it was left to Dubinin [8, 9] to introduce the concept of "micropore filling." In his original theory of volume filling, Dubinin [8] adapted Polanyi's concept of the adsorption potential expressed as a temperatureinvariant characteristic curve. It was subsequently realized [7, 10, 11] that a physisorption isotherm of the typical "Langmuir" shape (now referred to as a Type I isotherm [10]), is generally associated with micropore filling and not monolayer coverage. It is also generally accepted that micropore filling is dependent on the overlap of adsorption forces [7] and should not be regarded as a form of capillary condensation since it does not involve the formation of a meniscus. Three groups of pores of different width, w, were defined by Dubinin [9]. The classification, which was adopted in a revised form by the IUPAC [10], is as follows: in micropores w<1"'V2 nm; in mesopores W 1"'V2-50 nm; in macropores W >1"'V50 nm. It also expedient [11] to subdivide the micropores into ultramicropores (w <1"'V1 nm) and supermicropores (w 1"'V1-2 nm). However, all these dimensions are somewhat arbitrary and imprecise because the stages of pore filling are dependent on the gas-solid system as well as the pore geometry [11]. Similarly, there is no precise definition of the currently popular term "nanopore," which is often applied to a pore in the supermicropore or narrow mesopore range. Highly porous carbons can be produced from a variety of natural and synthetic precursors [11, 12]. Relatively inexpensive activated carbons are useful adsorbents, but generally their surface and pore structures are exceedingly complex [11, 13]. However, it is now possible to prepare a number of more uniform carbonaceous adsorbents. For example, molecular sieve carbons (MSCs) are available with narrow distributions of ultramicropores, which exhibit welldefined molecular selectivity [11], and carbon nanotubes, aerogels, and membranes are also amongst the most interesting advanced materials for research and development [12, 14]. Carbon blacks with specific surface areas of up to 100 m 2 / g can be regarded as essentially nonporous [15] since they give reversible Type II isotherms in the IUPAC classification [10]. Early physisorption measurements on carbon blacks [1] were designed to test the validity of surface areas determined the BET method [6]. Carbon blacks were considered [5] to be especially suitable for this purpose because the discrete nature of their spheroidal particles allowed electron microscopy to be used for the evaluation of the particle size distribution. Certain well-characterized carbon blacks are still extremely useful as reference adsorbents [11, 16]. In its original state, the surface of a carbon black is energetically heterogeneous [15], but as discovered by Beebe et al. [17] the heterogeneity is considerably reduced by heat treatment in an inert atmosphere. As a result of this form of "graphitization," adsorption isotherms of the noble gases and lower hydrocarbons are transformed from Type II to Type VI in the IUPAC classification [10] - the latter having stepwise multilayer character [7, 11, 18]. By studying the energetics of adsorption of many polar and nonpolar molecules on highly I"'V
1.2
Physisorption on Nonporous Carbons
5
graphitized carbon blacks, Kiselev and his coworkers were able to confirm [19] that the gas-carbon interactions were essentially nonspecific (i.e., insensitive to the polar nature of the adsorptive molecule). In contrast, polar molecules (dipolar or quadrupolar) were found to undergo pronounced specific interactions with exposed hydroxyls or cationic sites on the surface of oxides and zeolites. In recent years various new experimental techniques and computerized procedures have become available for the determination and analysis of physisorption data. It is perhaps all the more surprising that some widely used methods are still based on old and over-simplified principles. This short overview is not intended to provide a critical survey of current research on gas adsorption by carbons and indeed the recent advances in research and development are amply described in the following chapters of this book. However, it is pertinent in this introductory overview to draw attention to some of the major obstacles, which have restricted progress and which now offer the opportunity for further research. The distinctive features of gas/carbon physisorption systems are also briefly compared with the adsorptive behavior of oxides and zeolites.
1.2 PHYSISORPTION ON NONPOROUS CARBONS As already indicated, a well-defined stepwise (Type VI) isotherm is obtained when a noble gas or lower hydrocarbon is adsorbed on a basal graphitic surface at an appropriate temperature [7, 11]. The regular steps can extend up to four or five molecular layers, but become less sharp with increased distance from the adsorbent surface. An increase in temperature also produces a progressive blurring of the layer-by-layer adsorption [7]. The appearance of such regular multilayer steps in isotherms on uniform surfaces supports the view that (a) the influence of the surface structure can extend well beyond the first adsorbed layer and (b) the multilayer steps are associated with a form oflocalized physisorption. Exfoliated graphite and graphitized carbon black are especially useful for studying monolayer structures and phase changes in the subrnonolayer region of physisorption isotherms [11, 18]. Small discontinuities may be due to a bimodal distribution of surface sites [18] or other abnormal effects [11], but on a uniform surface such substeps are associated with two-dimensional (2-D) phase changes [11]. These 2-D transitions on graphite can be clearly revealed and characterized by means ofadsorption energy [20] or heat capacity [21] measurements. There is no sign of any stepwise character in the reversible Type II isotherms of argon, krypton, and the lower hydrocarbons on ungraphitized carbons [11]. Similar Type II isotherms are also given by the adsorption of these nonpolar gases on most inorganic oxides - an exception being methane on the (100) face of MgO [11]. These results are of course consistent with the heterogeneous nature of ungraphitized carbon and oxide surfaces. Turning now to nitrogen adsorption at 77 K, we find that ungraphitized carbons, oxides and other nonporous adsorbents all give typical Type II isotherms 11
6
Chapter
1
Overview of Physical Adsorption by Carbons
and when they are normalized (or reduced to unit area) these isotherms are almost identical over a wide range of pipO. On the surface ofgraphitized carbons, however, nitrogen isotherms at 77 K are of "hybrid" shape. In the multilayer range above plpo rv 0.5 they have the characteristic Type II shape, but at lower plpo they are stepwise (a fairly sharp first layer step and a wavy second layer step). The fact that at 77 K the path of the nitrogen multilayer isotherm is rather insensitive to change in the adsorbent structure is of some practical importance. It is one reason why nitrogen adsorption is used for the assessment of the surface area and pore size distribution [7, 11]. When physisorption occurs on the heterogeneous surface of an ungraphitized carbon or an oxide, an increase in surface coverage is normally accompanied by a progressive decrease in the differential energy of adsorption [11, 21]. On the graphitized surface any initial decrease in adsorption energy is restricted to a region of very low surface coverage and thereafter the differential energy either remains almost constant or increases to a maximum at high surface coverage. In this respect, there is an interesting difference between the behavior of n-hexane and benzene [11]. In the case of hexane, the marked increase in adsorption energy is the result of fairly strong lateral interactions between the adsorbed molecules, whereas the almost constant adsorption energy given by benzene appears to indicate very little adsorbate-adsorbate interaction over a wide range of surface coverage [11, 21, 22]. Interpretation of the variation of differential adsorption energy with surface coverage may not always be straightforward. For example, a strong adsorbentadsorbate interaction may allow some compression of the monolayer resulting in repulsion rather than attraction between neighboring molecules [23]. Also, the almost constant differential adsorption energy may be due to compensation between the effect of surface heterogeneity and the attractive adsorbateadsorbate interactions [24]. In the case of adsorption on a uniform surface, Henry's law is to be expected at low surface coverage [11,22]. Unless the operational temperature is relatively high, the Henry's law region is normally restricted to a very small part of the isotherm (e.g., below plpo rvl x 10-4 ) and the deviations may be in either direction. From a fundamental standpoint, the virial analysis of low-coverage adsorption data is important since it is analogous to the treatment of imperfect gases and nonideal solutions [22]. Many different equations have been used to interpret monolayer-multilayer isotherms [7, 11, 18, 21, 22] (e.g., the equations associated with the names: Langmuir, Volmer, Hill-de Boer, Fowler-Guggenheim, Brunauer-EmmettTeller, and Frenkel-Halsey-Hill). Although these relations were originally based on adsorption models, they are generally applied to the experimental data in an empirical manner and they all have limitations of one sort or another [7, 10, 11]. The multilayer Frenkel-Halsey-Hill (FHH) equation is usually expressed in the following form: log
pO) k ( -p =n-r
(1.1)
1.3 Physisorption by Porous Carbons
7
where n is the amount adsorbed at the relative pressure pj pO and k and rare empirical constants [7, 22]. The FHH plots of log [log (p jpO)] vs log n are generally linear over a wide multilayer range (e.g., pjp°1'"'...l0.4-0.9). A characteristic value of r is obtained from the slope of the FHH plot. In the case of nitrogen adsorption at 77 K on nonporous carbons and silicas, the derived values of r are generally remarkably constant (r = 2.70), which confirms the almost "universal" character of the nitrogen multilayer [11]. Linear FHH multilayer plots are also given by other adsorptives such as butane, neopentane, methanol, dichloromethane, and benzene, but the corresponding values of r appear to be somewhat more variable [11, 25-28]. Recent fundamental studies of physisorption by nonporous carbons have involved the application of molecular simulation [23, 29], density functional theory, and lattice gas models [30]. Important aspects studied have included the way in which the molecular packing in each layer is likely to be affected by a systematic alteration of the lattice parameters and interaction energies and the resulting changes in the isotherm shape [21, 30].
1.3
PHYSISORPTION BY POROUS CARBONS
Activated carbons having BET areas well in excess of 1000 m 2 / g can easily be prepared by the carbonization and activation of a variety of precursors [11, 12]. By the carefully controlled carbonization and activation of uniform precursors (e.g., viscose rayon) under conditions of low burnoffit is possible to obtain ultramicroporous carbons having molecular sieve properties and giving high-affinity, reversible Type I isotherms [11] (designated Type la) with no detectable supermicroporosity or mesoporosity. Many activated carbons have complex nanoporous structures as revealed by the composite nature of their physisorption isotherms [11]. For example, hysteresis loops of Type H4 in the IUPAC classification [10] are usually obtained after extensive burnoff More uniform nanoporous and mesoporous carbons can be produced, however, by the application of special pretreatment and activation procedures [11, 12]. Certain carbon aerogels also give well-defined Type IV isotherms [14]. As already indicated, much of the pioneering work on micropore filling was carried out by Dubinin [8, 9]. The first version of the temperature-invariant "characteristic curve" for micropore filling was expressed in the exponential form (1.2)
where np is the micropore capacity, A is the "adsorption affinity," and E is a characteristic energy for the given system. From Eqn (1.2), we can easily
8
Chapter
1
Overview of Physical Adsorption by Carbons
arrive [7] at the well-known Dubinin-Radushkevich (DR) equation in its usual linear form
log(n)
= log (n
p) -
Dlog
2
(~)
(1.3)
where the constant D is related to E. Thus, the plot of log (n) against lo~ (po Ip) should be linear with slope D and intercept log (n p ). To obtain vp (the micropore volume) from np ' it is necessary to assume that the density of the adsorbate is the same as that of the liquid adsorptive at the operational temperature [11]. Ultramicroporous carbons generally give linear DR plots [7, 11] over wide ranges of pipO, but the extent of the linear region is much more restricted with most nanoporous carbons [11]. Also, DR plots on many nonporous and mesoporous carbons exhibit similar ranges of linearity [7]. Since D, E, and np are all empirical parameters, their significance is questionable. In consequence, it must be emphasized that the simple DR plot cannot always give a true assessment of the effective micropore volume [11]. The Dubinin-Astakhov (DA) equation, which was proposed as a more general characteristic curve equation [9], can be conveniently expressed in the form
.!!..- = exp np
[_(~) N] 13 Eo
(1.4)
where N is another empirical constant and E is replaced by f3E o (13 being a scaling factor). A limited amount of experimental evidence appears to confirm an inverse relation between Eo and the pore width, w, and also between Nand the extent of the pore size distribution (the extreme value of N = 3 being given by an ultramicroporous carbon molecular sieve). A complementary approach introduced by Stoeckli and his coworkers [31, 32] was based on the notion that the simple DR equation is invalid if the microporosity is heterogeneous. It was argued that the overall isotherm can be regarded as the sum of the DR isotherms corresponding to individual groups of micropores and that for a continuous distribution ofpore size the summation can be replaced by integration. A Gaussian pore size distribution was assumed and the integral transform was solved by the use of a mathematical device equivalent to an "error function." The mathematical elegance of Stoeckli's contribution is impressive, but it must be kept in mind that it does not allow for the effects of surface heterogeneity or changes in the various molecular interactions in controlling the different stages of pore filling [11]. In addition to the well-known monolayer equations already referred to (e.g., the Langmuir, Hill-de Boer and virial equations), there are others such as those associated with the names of Freundlich, Sips, and Toth, which have been applied to physisorption isotherms on microporous solids [11]. These relatively simple, empirical equations are useful for the interpolation and extrapolation of chemical engineering data, but they are not capable of throwing any new light
1.3 Physisorption by Porous Carbons
9
on the mechanisms of surface coverage or pore filling. As pointed out by various authors [7, 11, 18], the range of fit of a particular equation is not enough by itself to confirm the validity of the underlying theory. It follows that all derived quantities such as surface area and pore size require critical appraisal before their physical validity can be accepted [7, 10, 11]. Considerable interest has been shown in recent years in the presentation of physisorption isotherms in a generalized integral form. This kind of generalized treatment was first applied to the submonolayer region [21], but much of the current interest is centered on the analysis ofmicropore filling isotherms [33, 34]. In principle, the approach provides a means of constructing model isotherms by systematically combining hypothetical "local" isotherms and energy distribution functions. Thus, the generalized expression for the overall adsorption, n(p) can be written in the integral form
n(p) = no
f n*(p, E)f(E)dE
(1.5)
where no represents either the monolayer capacity or the micropore capacity, n*(p, E) is the local isotherm corresponding to the adsorption energy E, and f(E) is the distribution function for the adsorption energy. Although Eqn (1.4) can be derived from Eqn (1.5), it is not possible to arrive at an unambiguous interpretation of the energy distribution function without additional information since f (E) may be associated with surface heterogeneity and/or a micropore size distribution. Thus, for the filling of micropores an equivalent generalized expressIon IS
n(p)
= np
f n*(p, w)f(w)dw
(1.6)
where f (w) is now the micropore size distribution. Various empirical procedures, which can be used for the analysis ofphysisorption isotherms, are based on the comparison with standard adsorption data obtained on nonporous reference materials [7, 11]. These procedures have the advantage that they do not depend on any a priori assumptions concerning the mechanisms of adsorption or pore filling. The standard isotherm on the chemically similar reference adsorbent is plotted either as a multilayer thickness curve of t vs pipO or preferably in the reduced form of as vs pipO. The dimensionless quantity as is defined as nlns ' ns being the amount adsorbed at a preselected plpo (usually 0.4). The as plot is constructed by plotting n for the given nanoporous adsorbent against the corresponding as for the nonporous reference. Deviations from linearity can be attributed to the various stages of micropore or mesopore filling and, if certain conditions are satisfied, it is possible to derive approximate values of internal and external areas and ultramicropore and supermicropore volumes [11]. By analyzing the as plots of many adsorptives on various nanoporous carbons, Sing and his coworkers have obtained a considerable amount of experimental evidence [11, 35] to support the view [35] that the mechanisms of micropore filling are to a large extent dependent on the ratio of pore size:molecular
10
Chapter
1
Overview of Physical Adsorption by Carbons
diameter. If the pore width is no more than about two molecular diameters (i.e., if w<"'-'2o-) there is a significant overlap of adsorbent-adsorbate interactions [35,36] and the ultramicropores are filled at very low plpo. This has been termed [7, 35] "primary micropore filling" and is manifested in a pronounced distortion of the initial part of the isotherm [37] - and the corresponding as plot. The wider micropores (supermicropores) are filled over a range of higher plpo by a cooperative process [11, 35], after monolayer adsorption has taken place on the pore walls, but before the onset of capillary condensation [7]. Two major problems are encountered in the classical interpretation ofthe capillary condensation range ofa physisorption isotherm. The first is concerned with the validity of the Kelvin equation and the second with the existence of adsorption hysteresis. It is generally accepted that the Kelvin equation breaks down completely below a certain pipO when the condensate loses its stability [7, 11]. In addition to this complete failure of the Kelvin equation, it now seems likely [11, 14, 38] that the effective values of surface tension and molar volume require progressive adjustment as the pore width is reduced below, say, 10 nm. The presence of adsorption hysteresis cannot be reconciled with the laws of classical thermodynamics. It is evident that there are various forms of adsorption hysteresis [7], which require different explanations [7, 39, 40]. In the capillary condensation range, well-defined hysteresis loops are generally associated with delayed condensation or percolation [11] through pore networks or "ink bottles" [38, 39]. In the case of activated carbons, delayed condensation is likely to be the most important mechanism [11], but we cannot rule out the other effects. The computational procedures now used in the application of density functional theory and molecular simulation for the prediction and analysis of physisorption isotherms are based on the statistical mechanics of confined fluids [14]. These important advances are described in several chapters of this book and therefore the present introductory remarks are confined to a few general comments. Whichever computational procedure is adopted [39, 40], it is first necessary to define a 3-D model of the pore structure within a solid of known and uniform composition [14]. It has been customary to assume that the pores of different width are all of the same shape (e.g., slits in activated carbons). Further assumptions made by many investigators are that the filling or emptying of each group of pores can occur independently and reversibly, that the internal surface is uniform and that the solid-fluid and fluid-fluid interactions can be expressed in terms of standard potential functions [14]. The Monte Carlo method is generally considered [14, 39] to be the most convenient means of using computer simulation to generate adsorption equilibrium data by application of the Boltzmann distribution law in order to accept or reject the random movements of large numbers of molecules. Although the thermodynamic parameters may be readily changed, grand canonical Monte Carlo (GCMC) molecular simulation is usually preferred with the chemical potential being taken as the independent variable. Because of the present availability of commercial software, the nonlocal version of density functional theory (i.e., NLDFT) is now widely used for pore
1.4 Concluding Remarks
11
size analysis. With the aid ofNLDFT, it has been possible to generate a series of "individual pore isotherms" for nitrogen adsorption at 77 K on a wide range of graphitic slit-shaped pores [14, 34]. The NLDFT-derived isotherms have been found to be in excellent agreement with results obtained by GCMC simulation and also in broad agreement with the corresponding experimental data [14, 40]. The NLDFT calculations have shown inter alia the effects produced by changing the operational temperatures and the dimensions of the probe molecules. For example, although the molecular dimensions of nitrogen and carbon dioxide are similar, their adsorptive behavior was shown [40] to depend on the difference in operational temperatures (e.g., 77 K for nitrogen and 273 K for carbon dioxide). Indeed, this difference had already been successfully exploited by RodriguezReinoso and his coworkers [41] by using carbon dioxide at 273 K for probing the very narrow pores (of w "'-'0.35 nm) in activated carbons. The highly distinctive character of the adsorption of water vapor by porous carbons has long been of great interest [5, 7]. Some microporous carbons give well-defined Type V water isotherms, the steep riser often being associated with a fairly narrow hysteresis loop. The low affinity of the carbon surface for water is evidently due to the small number of hydrophilic sites and the relatively weak nonspecific interactions [11] (dispersion forces). However, the initial uptake of water at low p/ pO can be increased by surface oxidation, which leads to the formation of additional hydrophilic sites and the appearance of a small knee [7] (i.e., a change in isotherm shape from Type V to Type IV). The clustering of hydrogen-bonded water molecules finally leads to complete pore filling of narrow slit-shaped pores [33,42]. This is in contrast to the hydrophobic behavior of Silicalite [11, 42], which has an interconnected tubular pore structure. Thus, thin layers of hydrogen-bonded water molecules can be accommodated in slitshaped pores of "'-'0.5 nm, but this is not possible in tubular pores of similar width because it would involve too much distortion of the hydrogen-bonded water structure [42, 43].
1.4
CONCLUDING REMARKS
As already indicated, there are two related obstacles have impeded progress in the vast amount of research on the physisorption of gases by porous carbons. First, we have to contend with the complexity of the surface and pore structure ofmost carbon adsorbents; and secondly, we have to face the difficulties involved in the interpretation of physisorption isotherms. The simplest structural model for a microporous carbon is the slit-pore model, in which each pore is in the form of a parallel-sided and rigid graphitic slit. It is also assumed that the filling and emptying processes are not dependent on the spatial arrangement of the pores. It therefore follows that for a given gas/carbon system and temperature, the notional thermodynamic state of the adsorbate (e.g., its chemical potential and enthalpy) is solely dependent on the distribution of
Chapter
12
1
Overview of Physical Adsorption by Carbons
slit width. Obviously, this simple model cannot account for nonequilibrium effects such as delayed condensation or network percolation or for any significant variation of pore shape. Although the slit-pore model is still popular, it is evident [43, 44] that considerable refinement of the model is now possible. For example, allowance can be made for nonuniformity of the internal surface in the form of pit defects and corrugated layers [44]. Also included in the more sophisticated models [44] are randomly etched graphite surfaces and different assemblages of microcrystals of graphite or graphene platelets. By the statistical modeling of the topological features of pore networks [45] and the application of fractal analysis [46], it is also possible to identify various short-range patterns and different degrees of structural disorder. The rapid developments now taking place in the computer modeling of pore networks are of considerable interest to both theoreticians and experimentalists. We can expect further progress in this direction, but as pointed out by Levitz [45] it must be kept in mind that a particular statistical model may not be a unique representation of the real material. It is therefore equally important to extend the present limited range of well-characterized ordered and partially disordered carbons. The development of user-friendly equipment has facilitated the use of gas adsorption alongside other techniques for the characterization of highly active carbons of all types. In the near future, the limitations of the most popular procedures for the analysis of physisorption isotherms and the derivation of surface areas and pore size distributions (e.g., by the BET, DR, and Barrett-JoinerHalenda (BJH) methods) will inevitably remain a serious handicap. However, it should be kept in mind that it is always useful to closely examine the isotherm shape and apply an empirical method of analysis [11]. There is no immediate prospect of finding any suitable alternative to nitrogen for routine physisorption measurements, but in future nitrogen adsorption at 77 K should be regarded as only the first stage in the characterization of porous carbons. The next stage will depend on the nature of the carbon and its application (e.g., for gas storage or separation, water treatment, respiratory protection or catalysis). This is likely to involve equilibrium and dynamic adsorption measurements with selected probe molecules of different size, shape, and molecular structure. We can also anticipate an increased need for chemical engineering data, which will necessitate multicomponent adsorption measurements at high and low pressures.
REFERENCES 1. Deitz, V.R. (1944). Bibliography of Solid Adsorbents. National Bureau of Standards. 2. Polanyi, M. (1916). Adsorption of gases by a solid non-volatile adsorbent. Verb. Deut. Physik. Ges., 18, 55-80. 3. Langmuir, I. (1916). Constitution and fundamental properties of solids and liquids. I. Solids. J. Am. Chern. Soc., 38, 2221-95.
References
13
4. Foster, A.G. (1932). Sorption of condensable vapors by porous solids. I. Applicability of the capillary theory. Trans. Faraday Soc., 28, 645-57. 5. Brunauer, S. (1944). The Adsorption of Gases and Vapors. Oxford University Press. 6. Brunauer, S., Emmett, P.H., and Teller, E. (1938). Adsorption of gases in multimolecular layers.]. Am. Chern. Soc., 60, 309-19. 7. Gregg, S.J. and Sing, K.S.W. (1982). Adsorption, Suiface Area and Porosity. Academic. 8. Dubinin, M.M. (1955). A study of the porous structure of active carbons using a variety of methods. Q. Rev. Chern. Soc., 9, 101-14. 9. Dubinin, M.M. (1966). Porous structure and adsorption properties of active carbons. In Chemistry and Physics of Carbon (P.L. Walker, ed.). Marcel Dekker, pp. 51-119. 10. Sing, K.S.W., Everett, D.H., Haul, R.A.W., et al. (1985). Reporting physisorption data for gas/solid systems. Pure Appl. Chern., 57, 603-19. 11. Rouquerol, F., Rouquerol, ]., and Sing, K. (1999). Adsorption by Powders and Porous Solids. Academic. 12. Rodriguez-Reinoso, F. (2002). Production and applications of activated carbons. In Handbook of Porous Solids (F. Schuth, K.S.W. Sing and ]. Weitkamp, eds). Wiley/VCH, pp. 1766-827. 13. McEnaney, B. (2002). Properties of activated carbons. In Handbook of Porous Solids (F. Schuth, K.S.W. Sing, and]. Weitkamp, eds). Wiley/VCH, pp. 1828-63. 14. Pikunic, ]., Lastoskie, C.M., and Gubbins, K.E. Molecular modeling of adsorption from the gas phase. In Handbook of Porous Solids (F. Schuth, K.S.W. Sing, and ]. Weitkamp, eds). Wiley/VCH, pp. 182-236. 15. Sing, K.S.W. (1994). Physisorption of gases by carbon blacks. Carbon, 32,1311-7. 16. Badalyan, A. and Pendleton, P. (2003). Analysis of uncertainties in manometric gas adsorption measurements. Langmuir, 19, 7919-28. 17. Beebe, R.A., Biscoe,]., Smith, W.R., and Wendell, C.B. (1947). Heats ofadsorption on carbon black.]. Am. Chern. Soc., 69, 95-101. 18. Young, D.M. and Crowell, A.D. (1962). Physical Adsorption of Gases. Butterworths. 19. Kiselev, A.V. (1965). Non-specific and specific interactions ofmolecules ofdifferent electronic structures with solid surfaces. Disc. Faraday Soc., 40, 205-18. 20. Rouquerol,]., Rouquerol, F., and Grillet, Y. (1989). Energetical aspects ofnitrogen and argon adsorption. Pure Appl. Chern., 61, 1933-36. 21. Rudzinski, W. and Everett, D.H. (1992). Adsorption of Gases on Heterogeneous Suifaces. Academic. 22. Steele, W.A. (1974). The Interaction of Gases with Solid Surfaces. Pergamon. 23. Aronovich, G.L. and Donohue, M.D. (2001). Surface compression in adsorption systems. Colloids Surf. A, 187, 95-108. 24. Vernov, A. and Steele, W.A. (1993). Computer simulations ofbenzene adsorbed on graphite. In Fundamentals ofAdsorption IV (M. Suzuki, ed.). Kodansha, pp. 695-701. 25. Carrott, P.J.M., Ribeiro Carrott, M.M.L., Cansado, I.P.P., and Nabais, ].M.V. (2000). Reference data for the adsorption of benzene on carbon materials. Carbon, 38,465-74. 26. Carrott, P.J.M., Ribeiro Carrott, M.M.L., and Cansado, I.P.P. (2001). Reference data for the adsorption of methanol on carbon materials. Carbon, 39, 193-200. 27. Carrott, P.J.M., Ribeiro Carrott, M.M.L., and Cansado, I.P.P. (2001). Reference data for the adsorption ofdichloromethane on carbon materials. Carbon, 39, 465-72. 28. Carrott, P.J.M. and Sing, K.S.W. (1989). Multilayer adsorption of nitrogen and alkanes by non-porous carbons and silicas. Pure Appl. Chern., 61, 1836-40.
14
Chapter 1 Overview of Physical Adsorption by Carbons
29. Sweatman, M.B. and Quirke, N. (2001). Characterization of porous materials by gas adsorption: comparison of nitrogen at 77 K and carbon dioxide at 298 K on activated carbon. Langmuir, 17, 5011-20. 30. Do, D.D. and Do, H.D. (2002). Analysis of adsorption data of graphitized thermal black with DFT-lattice gas theory. Adsorption, 8, 309-24. 31. Stoeckli, H.F., Houriet,J.P., Perret, A., and Huber, U. (1979). The adsorption of simple gases by strongly activated carbons in relation to heterogeneity. In Characterization of Porous Solids (S.J. Gregg, K.S.W. Sing, and H.F. Stoeckli, eds). Society of Chemical Industry, pp. 31-9. 32. Stoeckli, H.F. Characterization of microporous ca~bons by adsorption and immersion techniques. In Porosity in Carbons O.W. Patrick, ed.). Edward Arnold, pp.67-92. 33. Jorge, M. and Seaton, N.A. (2002). Characterization of the surface chemistry of activated carbon by molecular simulation of water adsorption. In Characterization of Porous Solids VI (F. Rodriguez-Reinoso, B. McEnaney, J. Rouquerol, and K. Unger, eds). Elsevier, pp. 131-8. 34. Sweatman, M.B. and Quirke, N. (2001). Modeling gas adsorption in slit-pores using Monte-Carlo simulation. Mol. Simul., 27, 295-321. 35. Carrott, P.J.M. and Sing, K.S.W. (1988). Assessment of microporosity. In Characterization of Porous Solids I (K.K. Unger, J. Rouquerol, K.S.W. Sing, and H. Kral, eds). Elsevier, pp. 77-100. 36. Everett, D.H. and Powl, J.C. (1976). Adsorption in slit-like and cylindrical micropores in the Henry's law region. J. Chem. Soc. Faraday Trans. I, 72, 619-36. 37. Kenny, M., Sing, K.S.W., and Theocharis, C. (1993). The use of high-resolution adsorption measurements for the study ofporous solids. In Fundamentals ofAdsorption IV (M. Suzuki, ed.). Kodansha, pp. 323-32. 38. Ravikovitch, P.1. and Neimark, A.V. (2002). Density functional theory of adsorption in spherical cavities. Langmuir, 18, 1550-60. 39. Vishnyakov, A. and Neimark, A.V. (2003). Monte Carlo simulation test of pore blocking effects. Langmuir, 19, 3240-7. 40. Neimark, A.V. and Ravikovitch, P.I. (2000). Density functional theory of adsorption hysteresis and nanopore characterization. In Characterization of Porous Solids V (K.K. Unger, G. Kreysa, and J.P. Baselt, eds). Elsevier, pp. 51-60. 41. Rodriguez-Reinoso, F., Garrido, J., Martin-Martinez, J.M., et al. (1989). The combined use of different approaches in the characterization of microporous carbons. Carbon, 27, 23-32. 42. Carrott, P.J.M., Kenny, M.B., Roberts, R.A., et al. (1991). The adsorption ofwater vapor by microporous solids. In Characterization of Porous Solids II (F. RodriguezReinoso, J. Rouquerol, K.S.W. Sing, and K.K. Unger, eds). Elsevier, pp. 685-92. 43. Kaneko, K. (2000). Specific intermolecular structures of gases confined in carbon nanospace. Carbon, 38, 287-303. 44. Bandosz, T.J., Biggs, M.J., Gubbins, K.E., et al. (2003). Molecular models of porous carbons. In Chemistry and Physics of Carbons, Vol. 28 (L.R. Radovic, ed.). Marcel Dekker, pp. 41-228. 45. Levitz, P.E. (2002). Statistical modeling of pore networks. In Handbook of Porous Solids (F. Schuth, K.S.W. Sing, andJ. Weitkamp, eds). Wiley/VCH, pp. 37-80. 46. Neimark, A.V. (2002) Fractal analysis. In Handbook of Porous Solids (F. Schuth, K.S.W. Sing, andJ. Weitkamp, eds). Wiley/VCH, pp. 81-105.
OVERVIEW OF CARBON MATERIALS IN RELATION TO ADSORPTION Juan M.D. Tasc6n Instituto Nacional del Carbon, CSIC, Oviedo, Spain
Contents Introduction Structures of Elemental Carbon: Carbon Allotropes and Polytypes 2.3 The Sp2 Carbon Forms: Graphitic, Graphitizable, and Nongraphitizable Carbons 2.4 Structural Characterization of Carbon Materials: The Basic Structural Units and Their Stacking and Orientation Degrees 2.5 Conclusions Acknowledgments References 2.1
15
2.2
17 21
24 42 43 43
2.1 INTRODUCTION The importance of adsorption for different carbon materials and, conversely, the contribution of each type of carbon to the field of adsorption is very different. This reflects the wide variability in properties of solid carbons [1, 2], which makes their surface properties important in very different fields and for different reasons. Thus, graphite, due to its relatively simple structure, has often been used as a model material to simulate the adsorption of different molecules on its surface, or to carry out adsorption measurements on a well-controlled surface. Likewise, carbon blacks, particularly those thermally treated ("graphitized"), have often been used as reference nonporous adsorbents, as they only exhibit an external surface. The absence of open porosity and high chemical inertia are attributes that make glass-like carbon a material frequently used in Adsorption by Carbons ISBN: 978-0-08-044464-2
© 2008 Elsevier Ltd. All rights reserved.
15
16
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
electrodes for electrochemical applications; for this reason, glass-like carbon has special relevance for liquid-solid adsorption in the context of electrochemistry. Furthermore, the adsorption behavior (for both gases and liquids) of highly porous carbons such as activated carbons is of great interest because of their relevance for technologies such as pollution control, heterogeneous catalysis, or gas and electric energy storage to mention just a few applications [3]. It is not surprising therefore that a great number of adsorption studies on carbons have been carried out using these materials as adsorbents. A corollary to this is the vast abundance of adsorption studies addressing the characterization of (micro)porosity in activated carbons. On the opposite side of the spectrum of surface area development, carbon fibers (CFs) [4, 5], used principally as reinforcement in composites, are virtually nonporous and exhibit a surface area that is only slightly greater than their geometrical area. Since the surface characteristics of carbon fibers are important in fields such as interfacial fiber-matrix behavior in composites, researchers in this area have adapted their methodologies, or developed new ones, to study adsorption on these low-surface area solids. This explains, for instance, the widespread application of inverse gas chromatography to this class of carbons [6]. Newer fibrous forms of carbon such as carbon nanofibers (CNFs) and carbon nanotubes (CNTs) are also being studied for their adsorption properties, although in this case the aim is to investigate not only their interaction with matrices in composites, but also their possible use as media for gas storage. The examples given above show that, although only a few types of carbons (particularly activated carbons) appear to dominate the scene where carbonaceous solids and adsorption are concerned, the carbon materials involved belong to very different classes, from highly crystalline to highly disordered. Carbon fibers (more specifically, those produced from pitches) reproduce within themselves the wide variety of carbon properties, depending chiefly on the nature of the precursor and preparation conditions. Thus, in order to provide an overview of carbons in relation to adsorption, it is necessary to be comprehensive and consider all possible classes of carbon materials, though paying particular attention to those that are most relevant to adsorption. This approach is the one adopted this chapter, which mainly attempts to classify the carbons, compare their 'properties, and introduce them to those interested in their adsorption behavior. The classification criteria used here are chiefly associated with crystalline structure (particularly, nanotexture), which we believe to be the property that offers the most thorough classification of carbon materials. Whenever available, structural models (rather than e.g. micrographs) will be used here to schematically illustrate the structural features of the various classes of carbonaceous solids. For more extensive or detailed information on carbon materials, the reader is referred to a number of comprehensive publications [1, 2, 7-15], some of which are particularly focused on structural aspects [11-14]. In this chapter, we will try to follow the terminology norms for carbon as a solid developed by the International Committee for Carbon Characterization and Terminology, and later adopted by the IUPAC [16]. Some
in
2.2
Structures of Elemental Carbon: Carbon Allotropes and Polytypes
17
previous reviews [7, 8] also include detailed information on the terminology employed for these materials. For the adsorption terminology (especially in connection with the characterization of porous solids, pore size classification, etc.), we will follow the corresponding IUPAC norms [17,18].
2.2 STRUCTURES OF ELEMENTAL (ARBON: (ARBON ALLOTROPES AND POLYTYPES
Carbon has an atomic number of 6 and its ground-state electron configuration is [He]2s22p 2. Like some other p-block elements (e.g., P, S, Sn), solid elemental carbon exhibits the phenomenon of allotropy (i.e., occurrence of the element in different forms that vary in structure and equation of state). Although other approaches (e.g., LCAO) have been used [7, 13] to justify the occurrence of different structures for this element in solid form, the hybridization concept provides a very intuitive explanation and is very useful for classifying carbon-based solids. In this approach, one of the s electrons can be promoted and hybridized with different numbers of p orbitals to give rise to three types of hybrid orbitals: Sp3, Sp2, and sp. These are at the origin of the three basic carbon structures, respectively, diamond, graphite, and carbynes. Table 2.1 summarizes the relationship existing between the type ofchemical bond, the stereochemistry, and the corresponding structures for carbon in solid form. In the diamond structure, carbon atoms are present in Sp3 hybridization, with a tetrahedral stereochemistry and a face-centered cubic structure that is shown in Fig. 2.1. Besides natural diamond, synthetic diamond has been produced since General Electric first announced its successful high-pressure synthesis in 1955. Sintered polycrystalline diamond, different types ofdiamond films, and diamondlike carbon are other types ofdiamond-related synthetic materials, some ofwhich are noncrystalline [13, 19]; these solids have their own terminology [10,20]. Unlike other carbonaceous solids, diamond has a rather limited and specific relevance to adsorption. Indeed, ever since the publication of a pioneering work Table
2.1 Relationship between hybridization, stereochemistry, and structures of solid carbon
Sp3
Tetrahedral
Sp2
Planar trigonal
sp
Linear
Diamond Lonsdaleitea Hexagonal graphite Rhombohedral graphitea Carbynes
aThese can be regarded as polytypes rather than different structures.
18
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
(a)
(c)
Figure 2.1 Relationship between the graphite (a), lonsdaleite (b), and diamond (c) structures: c is the interlayer distance. (Reprinted from Re£ [23] with permission from Elsevier.)
in 1961 when an industrial diamond powder was first used as an adsorbate for nitrogen and argon [21], adsorption work on diamonds has mainly tended to confine itself to well-defined surfaces [22]. A second type of solid made up of carbon atoms in Sp3 hybridization is lonsdaleite. This structure consists of tetrahedra of carbon atoms assembled into a wurtzite-type hexagonal network. Lonsdaleite, a rare mineral, has been found in meteorites (e.g., the famous metallic meteorite Canyon Diablo, Coconino Co., Arizona, USA). Lonsdaleite can also be prepared by subjecting graphite to static pressure [19]. Figure 2.1 shows an analogy existing between lonsdaleite, diamond, and graphite structures, where the cross-hatched hexagons of carbon atoms adopt boat, armchair, and planar configurations, respectively [23]. As is well known, fullerenes are the only molecular solids in the realm of carbon materials. Their structure consists of polyhedra formed by the assembling of a series of pentagonal and hexagonal rings of carbon atoms. The hybridization state of fullerenes is intermediate between Sp3 and Sp2 (the Sp3 character deriving from the curvature in the polyhedron), this state being [24] Sp2.28 for the molecule of C 60 or buckminsterfullerene, the archetype of this carbon structure. Something similar occurs with carbon nanotubes, where curvature is present in the cylinder itself and in the two fullerene hemispheres that are thought to close up its ends. More information on this point is given below in this section, when the ternary diagram (CSp3, C Sp2, Csp) of carbonaceous solids is discussed. The review by Bandosz et al. [15] contains much information on a number of fullerene-related materials such as exo- and endohedral fullerenes, heterofullerenes (e.g., C s9 N, C 69 N), and fullerites (i.e., molecular solids formed by the aggregation of fullerenes at moderate temperatures and pressures; they can in turn condense into polymerized fullerites or accommodate intercalate species to form fullerite intercalation compounds). Graphite is the next allotrope of carbon. This consists of sheets of Sp2 carbon atoms with a planar trigonal stereochemistry that form a flat, condensed
2.2
Structures of Elemental Carbon: Carbon Allotropes and Polytypes
19
a
a A
A
c
A
B
B
A
A
A
Simple hexagonal graphite
Bernal graphite
Rhombohedral graphite
Figure 2.2
Graphite structures: simple hexagonal, Bernal hexagonal, and rhombohedral: a is the distance between nearest-neighbours, and c is the interlayer distance. A, B, C denote the stacking sequence ofgraphenes. (Reprinted from Ref. [25] with permission from Elsevier.)
system of hexagonal rings. In turn, these sheets, called graphenes, are bound to each other in the z-axis direction by a very weak metallic-type bond (similar in strength to van der Waals forces). Two possible types of graphene stacking exist (Fig. 2.2): hexagonal, Bernal type (ABAB ... sequence), and rhornbohedral (ABCABC ... sequence). The so-called simple hexagonal structure (AA ... sequence) [25], also depicted in Fig. 2.2, is only hypothetical. The rhombohedral structure is irreversibly transformed into the hexagonal one at temperatures above 1600 K. For this reason, it is not found in synthetic graphites obtained by thermal treatment. In addition, attempts to prepare pure rhombohedral graphite have so far failed. Due to the minor importance of rhombohedral graphite, from this point on, the terms "hexagonal graphite" and "Bernal graphite" will be referred to simply as graphite. In fact, this simplification has already been explicitly accepted by the IUPAC terminology norms [16]. Carbynes are white solids made up of carbon atoms with sp hybridization. Two main possibilities [23] exist for this linear structure: polyine (-C:=C-C:=C-) and polycumulene (==C==C==C==C==); theoretical predictions and experimental evidence also point to the existence of cyclic carbynes [10]. According to a classification based on the type of bond present (Table 2.1), and also according to chronology, carbynes (rather than fullerenes) should be considered as the third allotropic form of carbon. Our scarce knowledge of carbynes and doubts about whether they really exist in pure form [10, 15, 26] are factors that have contributed to this erroneous interpretation. Moreover, lonsdaleite and rhombohedral graphite should be considered as polytypes rather than new structures with a different equation of state [27]. Strictly speaking, therefore, they should not be regarded as true allotropes of carbon (Table 2.1). A series of impure varieties termed (X- and ~-carbynes, chaoite, carbon VI, and carbons VIII-XIII have been reported. Some of these solids have a natural origin (e.g., chaoite was discovered in the crater of the meteorite Ries, Bavaria, Germany), whereas others have been synthesized through reactions such as the oxidative dehydropolycondensation of acetylene [28]. This type of structure has been detected in macroscopic amounts in the presence of heteroatoms, e.g., in polyvinyl cyanides, H-(C=C)n-CN (where n = 2 - 4). Attempts to prepare
Chapter
20
2
Overview of Carbon Materials in Relation to Adsorption
large quantities of carbynes led a decade ago to the discovery [29] of carbolites, a class of solids with a low mass density that exhibits different structures depending on the gas present when they were synthesized. The fact that the structure of carbolite is complementary to that of graphite and a series of intriguing physical properties make this esoteric (according to Dresselhaus [30]) carbon phase appealing. Phase diagrams for carbon typically depict the zones where diamond, lonsdaleite, and graphite exist (plus those for the liquid and vapor states) without considering the rest of the solid phases. Some diagrams that are very rich in detail, such as the one shown in Fig. 2.3, depict several zones of formation of metastable phases or spontaneous transformations (depicted as A up to ]), this knowledge having been obtained principally from efforts to convert graphite into diamond [31]. However, no information is provided on other carbon phases. Thus, authors such as Dresselhaus [30] and Donnet [32] have suggested the need to define precisely the areas of stability of novel carbon forms and include them in such diagrams. For instance, Delhaes [13] has suggested a tentative area for carbyne existence that encompasses ~4000-5000 K, and zero to a few gigapascal. In this regard, the ternary diagram shown in Fig. 2.4 represents a significant step forward as it shows the different carbon materials as a function of the type of hybridization [27]. The three corners in the triangle correspond to the Sp3, Sp2, and sp "pure" states, and along the three sides are carbon forms with intermediate hybridization states spn. Thus, along the Sp2_Sp3 side are the carbon nanotubes (near the Sp2 corner) and fullerenes (2 < n < 3, where n increases with the decreasing number of carbon atoms in the fullerene molecule, i.e., with an increasing degree of curvature). The two other sides of the triangle include
50 I I
30
r
DIAMOND
20
II
E
40
C? Q. ~
:; (/)
Q)
a:
I
,
~
Q)
(/)
G
~H
,F
, \
, .....
D
LIQ.
t
10
! 0 0
1000
3000 4000 2000 Temperature (K)
5000
6000
Figure 2.3 Phase diagram for carbon. The letters indicated correspond to various transformations between graphite, diamond, and lonsdaleite. (Reprinted from Ref. [31] with permission from Elsevier.)
2.3 The Sp2 Carbon Forms: Graphitic, Graphitizable, and Nongraphitizable Carbons
21
Diamond lonsdaleite
~Ollapse Carbyno(polyyne)diamonds mLCp (m=3,0
C20 , P/H=oo C21 , P/H=3 . C22 , P/H =2 Fullerenes
\
~
'Amorphous
'~carb.o. n.\~. .
/.
Hypothetical D-G hybrids
C82 , P/H =0.6
C73 , P/H =0.5
Condensation
Polycyclic , / networks
Carbyne
Monocyclic ...-... Graphynes ~ Layer-chain rings, carbons mLCp cyclo[N]carbons (m = 2, 0 < P < 1)
Graphite
Figure 2.4 Ternary diagram (Csp 3, C Sp 2, Csp) of carbonaceous solids. (Reprinted from Ref. [27] with permission from Elsevier.)
materials such as the less-known "graphynes" and monocyclic cyclo [N] carbons (1 < n < 2), and the hypothetic "superdiamonds" (1 < n < 3, n =f. 2). There are similar diagrams for more specific types of carbon materials - for instance, the ternary diagram (CSp 3, C Sp 2, H), which can be used to classify noncrystalline carbons such as amorphous carbon (a-C) films, hydrogenated carbon (a-C:H) films, or diamond-like carbons (DLC), to mention just a few [10, 13].
2.3
THE Sp2 CARBON FORMS: GRAPHITIC,
GRAPHITIZABLE, AND NONGRAPHITIZABLE CARBONS
Carbonaceous solids that are inside the triangle in Fig. 2.4 obviously contain mixtures of the three hybridization states. Japanese researchers [33, 34] recently coined the term "carbon alloys" to designate materials typically made up of carbon atoms in multicomponent systems, where the components undergo physical or chemical interactions with each other. This very broad concept extends to mixtures of different phases (as occurs, for instance, in carbon-carbon composites), and also to materials containing heteroatoms. The interesting point in the present context is that, according to the above definition, carbon atoms with different hybrid orbitals are considered as different components.
22
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
The ensemble of carbon materials containing the three types of hybridization includes many materials of practical interest, which are often referred to as "engineering carbons." In what follows, we will group all of these (Sp3+Sp2+Sp) solids under the common term "carbon forms." If one excludes some specific cases such as that of "diamond-like" amorphous carbon, these materials mainly contain carbon atoms in Sp2 hybridization. According to Marsh [35], "Although diamond has a greater emotional appeal than graphite as an allotrope of carbon structure, it is the graphite lattice which dominates discussions of structure within carbons." Indeed, it has become commonplace to consider the majority of carbon forms within the framework of the (hexagonal, Bernal type) graphite lattice, which is taken as reference for the highest degree of structural perfection. Hiittinger [36] has highlighted the potential of this anisotropic lattice as a common framework for many high-performance carbon materials, while Chung [37] has reviewed the special physical and chemical properties on which many applications of graphite and related solids are based. Both authors have identified oxidation sensitivity as one of the main shortcomings for practical applications, especially at high temperatures. The IUPAC [16] has adopted as a criterion for considering a material as graphite (or, more precisely, as a graphitic carbon), the requisite that the graphenes therein must be arranged parallel to each other in a tridimensional crystalline network. Conversely, one should not use the term graphitic carbon/material for solids (nongraphitic carbons), which do not exhibit a longrange tridimensional graphitic structure, even if they have been heat-treated at temperatures > 2500 K, in the so-called graphitization process. Nongraphitic carbons can in their turn be subdivided into graphitizable and nongraphitizable. As early as 1951, Franklin [38] established a classification into graphitizable and nongraphitizable carbons based on whether they were transformed or not into graphite by thermal treatment at 2273 K, and proposed the structural models that are reproduced in Fig. 2.5. According to Franklin's concept, graphitizable solids contain a series of units oriented in an approximately parallel fashion to each other (pregraphitic arrangement). In nongraphitizable
~~!IJb
~,~~~\\\~
~~C~~ ~~~~-'~~ Figure 2.5 Franklin's original schemes for graphitizable (lift) and nongraphitizable (right) carbons. (Reproduced from Re£ [38] with permission from the Royal Society.)
2.3 The Sp2 Carbon Forms: Graphitic, Graphitizable, and Nongraphitizable Carbons
23
solids, however, these units are arranged randomly and are cross-linked by disordered carbon species. The precise nature of these cross-links was not specified. It was assumed that on heating at high temperatures, these cross-links may be broken, but the activation energy needed for the motion of whole crystallites is high, and graphite is not formed. This model has been (and is still) widely accepted by the carbon community, despite its being considered as oversimplified, since in actual fact it is known that there is a continuum with no gap between extreme graphitizable and nongraphitizable situations [39]. Later on, the IUPAC [16] established a definition of nongraphitizable (also termed nongraphitizing) materials as "those non-graphitic solids that are not transformed into graphitic carbon by a simple thermal treatment at temperatures up to 3300 K and atmospheric pressure or lower" (i.e., through the graphitization process). This definition is empiric and somewhat arbitrary. The most representative type of nongraphitizable carbons is char, defined as a product of carbonization ofnatural or synthetic organic matter that has not passed through a fluid stage during carbonization. Other typically nongraphitizable carbon forms are carbon black, soot (plus aciniform carbon, an unsuccessful term that was coined to encompass the two previous types), glass-like (or "glassy") carbon, activated (also termed "active") carbon, wood charcoal, and many types of coal (bituminous coking coal and some anthracites being exceptions). The so-called graphitizable (also termed graphitizing) materials constitute a group that is complementary to the previous one. The IUPAC has simply defined them as those nongraphitic carbons that upon heat treatment are converted to graphitic carbon. The most representative type ofgraphitizable material is coke, which has been defined as a solid rich in nongraphitic carbon produced by the pyrolysis of an organic material that has passed, at least in part, through a liquid or liquid-crystalline stage during the carbonization process. There are a number of coke types, according to the precursor (e.g., petroleum coke, pitch coke, coal coke), the heat treatment temperature (e.g., green coke, calcined coke), or a combination of these and the production process (e.g., fluid coke, delayed coke, sponge coke, needle coke). Pyrolytic carbon, carbonaceous mesophase, and many types of pitch and carbon fibers (excluding, for instance, isotropic pitches and carbon fibers derived from them) are typical examples of graphitizable carbon forms. The difference between the definitions for char and coke derives from the formation of a mesophase or liquid crystal during the pyrolysis of certain types of precursors of carbons with a high degree of structural perfection [40]. Carbonaceous mesophase is formed from polyaromatic molecules with a laminar structure that are grouped together in parallel stacking by means of van der Waals forces (mesogens). Mesogens therefore adopt a pregraphitic arrangement that will transform into graphitic carbon when they are heated at higher temperatures. As an illustration of the thermal behavior of a typically graphitizable material, Fig. 2.6 shows a succession of optical micrographs for a pitch that generates mesophase during pyrolysis [40]. The initially formed mesogens group into anisotropic mesocarbon microspheres, which become separated from the
24
Chapter
Mesogens
2
Overview of Carbon Materials in Relation to Adsorption
Mesophase
Coalesced mesophase
Coke
Figure 2.6 Scheme of the carbonization process of a pitch (temperature increases from left to right). (Reproduced from Re£ [40] with permission from Taylor and Francis Group.)
fluid isotropic phase. As the temperature increases, polymerization and condensation reactions take place, the microspheres grow and coalesce, and viscosity increases until solidification takes place in the form of anisotropic coke. There are some exceptions to the rule that graphitizable materials must pass through a fluid stage to transform into graphite. On the one hand, it is well known that sucrose and other simple organic compounds give rise to nongraphitizable carbons, whereas they melt at low temperatures before the beginning of carbonization. However, this is not a significant exception since this involves just melting (""'459 K for sucrose) rather than carbonaceous mesophase formation and the like. Another, more important, exception is that of highly ordered graphitic carbon that can be produced by the carbonization-graphitization of polyimides, which takes place through the formation of a system of flat pores without any mesophase formation [41]. Moreover, the final product of this transformation has been shown to be one of the most perfect types of synthetic graphite ever produced, comparable in structural order to highly oriented pyrolytic graphite [1, 2]. Therefore, the above classification into graphitizable and nongraphitizable solids, which is phenomenological and somewhat arbitrary, is not strictly speaking true, and can only be considered as an approximate rule for classifying materials with a predominant Sp2 character into two main classes depending on their ability to transform into graphite.
2.4
STRUCTURAL CHARACTERIZATION OF CARBON
MATERIALS: THE BASIC STRUCTURAL UNITS AND THEIR STACKING AND ORIENTATION DEGREES
As Fig. 2.6 reveals, optical microscopy with polarized light makes it possible to distinguish various intermediate products formed from a precursor of graphitic carbon at different stages of pyrolysis. The observed morphology is the so-called optical texture, which enables carbon solids to be classified according to their degree of anisotropy. Besides optical microscopy, other techniques
2.4 Structural Characterization of Carbon Materials
25
with a higher degree of resolution, principally transmission electron microscopy (TEM), have significantly contributed to the characterization of the structure of carbon materials. To return to the case above, this technique provided direct evidence for the formation of flat pores and their sudden transformation into graphitic flat sheets at 2823 K upon the graphitization of Kapton polyimide. In this process, plane stacking changed from turbostratic (i.e., disordered in the c direction, see Section 2.4.3) to graphitic [41]. Oberlin and her group [42,43] have been widely recognized for their extensive studies of carbonaceous materials by TEM, mainly by combining selected area diffraction with dark-field and light-field imaging. One of the important concepts they established is that of basic structural unit (BSU) , defined as a planar aromatic structure of less than 10-20 rings and between two and four layers. A related concept is that of local molecular orientation (LMO), defined as an array ofBSUs with a near-common orientation. The arrangement ofBSUs, with a higher or lower degree ofLMO (in general, the regions of LMO are similar in size to the individual BSUs for carbons prepared at temperatures between 873 and 1273 K [39]), has increased our understanding of the mechanism ofgraphitization considerably. Thus, a high degree of cross-linking between the BSUs in nongraphitizable carbons leads to only a slight growth in the extent and order of LMO. On the other hand, lack of cross-linking in graphitizable carbons will allow the BSU elements to rearrange and coalesce with thermal treatment. The BSU concept also helps to explain the presence of intrinsic pores in certain carbon materials [44]. Figure 2.7 sketches the possible arrangements ofBSUs in carbons [42]. These can be reduced to two symmetries: spherical or cylindrical. All possible textures derive from these two basic arrangements if one considers the variable radii of curvature of lamellae; thus, an infinite radius of curvature gives rise to flat lamellae. We will come back to this figure when discussing the types of orientation found in carbonaceous materials (Sections 2.4.1 to 2.4.3). X-ray diffraction (XRD), generally considered the "ideal" technique for the structural characterization of materials, not only allows the structures of different carbon allotropes and polytypes to be distinguished from each other, but also enables the degree to which the structure of a given carbon form departs from the ideal graphite structure to be determined. By applying the Debye-Scherrer equation (adapted to carbon materials by Warren) to the (002) and (10) peaks, it is possible to calculate, respectively, the height (Lc) and width (La) of the graphitelike crystallites in these solids. Emmerich [45] has provided a useful comparison of the trends of variation of L c and Lc with heat treatment temperature for graphitizable and nongraphitizable carbons. The recent efforts of a group of Japanese researchers to normalize the application of XRD to carbon materials also deserves special mention [46]. Raman spectroscopy is another useful technique for classifying carbon materials according to structural criteria. Thus, the Raman spectra ofmost carbon forms contain, at least, two main bands termed G band (after graphite, r-v1575 cm- i ) and D band (after defects, r-v1355 cm- i ) [47]. Figure 2.8 illustrates the evolution of these two bands and the second-order spectrum (2600-3200 cm- i , G' band
26
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
-
1nm (a)
(d)
I.
Fiber. axiS
(f)
Figure 2.7 Sketches ofthe possible arrangementofbasic structural units (BSU)s. (a and b), true spherical symmetry (a, concentric texture; b, radial texture); (c-e), statistical spherical symmetry (c, crumpled sheets ofpaper; d, lamellar structure with infinite radius ofcurvature or infinite local molecular orientation (LMO); e, lamellar structure with a long-range statistical orientation); (f and g), cylindrical symmetry (f, true cylindrical symmetry; g, statistical cylindrical symmetry). (Reproduced from Re£ [42] with permission from Taylor and Francis Group.)
at '"'-'2700 cm- 1) for carbon solids with a different degree of structural order (decreasing from top to bottom in this figure) [48]. One can clearly see the decrease in intensity and widening of the D band, the parallel increase in the G band, and the disappearance of the G' band when moving progressively from
27
2.4 Structural Characterization of Carbon Materials
GC2
.' .....
GA2
OCA
_~ .. W.' _
~'
j ".,.•.... ._~
OVA
FC1
CM3 f
1000 1200 1400 1600 1800
$
•
2600 2800 3000 3200
Figure 2.8 Raman spectra of various carbon materials with an approximately decreasing degree ofstructural order in the following sequence: natural graphite (GC2), high surface area graphite (GA2), calcined needle coke (QCA), green needle coke (QVA), isotropic carbon fiber (FC1), and subbituminous coal (CM3). (Reprinted from Ref. [48] with permission from Elsevier.)
graphites to higWy disordered materials such as coals. The intensity ratio of the G and D bands as well as their wavenumbers and widths are useful quantitative criteria for comparing the degree of structural order of different carbon forms. The information provided by XRD is not equivalent to that obtained from Raman spectroscopy. Rather it is complementary. More specifically, XRD reflects the degree of graphitization in terms of the stacking of the basic constituents, whereas Raman spectroscopy reflects their degree of orientation. This is illustrated in Fig. 2.9, which shows the variation of two parameters derived from each of these techniques (respectively, the graphite-like interlayer spacing dOO2 ' and the [ID/(ID+ Ie)] intensity ratio) for a wide and varied set of carbons (including 45 different samples of graphitic, graphitizable, and nongraphitizable solids) [49]. For low values of both parameters, a horizontal line is obtained where data corresponding to various types of graphite accumulate. For these materials, d002 approaches the ideal value of 0.3354 nm. However, the graphite
28
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
0.41
• •
0.39
•
E
c
......... C\I
0
0.37
0
u
0.35
0.33 0
20
40
60
80
100
'01(10+ 'G) (%)
Figure 2.9
Comparison of structural parameters deduced from X-ray diffraction and Raman spectroscopy for a wide set of carbon materials. (Ref. [49] - reproduced by permission of the Royal Society of Chemistry.)
samples studied correspond to solids with different degrees of BSU orientation; therefore, the Raman intensity ratio varies from 0 to f"V40 %. A transition zone (line of intermediate slope) is followed by a zone with a wide dispersion of points that corresponds to the most disordered materials studied. For these materials, the [ID/(ID+ Ie)] ratio oscillates between 60 and 80 %, whereas d002 increases (vertical straight line). Here, the solids have their BSUs oriented at random and do not significantly differ from each other as examined by Raman spectroscopy. However, by means ofXRD they can be differentiated from the degree of stacking of graphenes inside their basic units (d002 varies from f"V0.35 to >0.40 nm). The importance ofthe degree oforientation ofthe basic constituents ofcarbon materials is highlighted in Fig. 2.10, proposed as early as 1985 by Inagaki [50], which depicts the main types of orientation found in carbons. This scheme has many points in common with the one shown in Fig. 2.7. In the following subsections we will use these two schemes as a basis for describing the main classes of nanotexture (defined here as arrangement of BSUs) found in carbonaceous solids, which afford a useful classification for understanding the behavior of carbons both in general and as adsorbents. Although, strictly speaking, lamellar and random orientations are special cases of spherical symmetry [42], we will treat them as separate types of texture to facilitate the classification of carbon solids into a larger variety of types. 2.4.1 Planar Orientation
This type of orientation is characteristic of graphite, of which some very perfect forms, such as kish graphite (precipitated from molten iron at high temperatures), highly oriented· pyrolytic graphite (HOPG), polyimide-derived
29
2.4 Structural Characterization of Carbon Materials
ORIENTED TEXTURE
RANDOM TEXTURE
PLANAR ORIENTATION
:I:~;e~
WS?
~ RANDOM ORIENTATION
f~~t1
~~COaXial
~~fer~nce AXIAL ORIENTATION ~~IS ~ ("'- ~ Radial Concentric
Reference point
POINT ORIENTATION
~
~ Radial -~---....
-----... - - - - -.... ~
Degree of orientation Degree of graphitization Heat treatment temperature
Figure 2.10 Types of orientation of the basic constituents of carbon materials. (Reprinted from Re£ [1] with permission from Elsevier.)
graphite films, or natural graphite flakes (e.g., from Ticonderoga [New York, USA], Madagascar, or Sri Lanka) represent paradigmatic cases of a high degree of structural order in the realm of carbon materials. As indicated in Section 2.1, some of the major connections of graphite with adsorption work have to do with carrying out measurements on a well-controlled surface, or, in the case of theoretical studies, to use its structure as a model for simulating the adsorption of different molecules on its surface. Moreover, HOPG is a material of choice for techniques such as scanning tunneling and atomic force microscopies and, as such, has often been used to directly visualize large molecules adsorbed on its surface. It is well known that various atoms, ions, and even molecules can be received into the graphite interlayer space between hexagonal layers of carbon atoms. In this way, intercalation compounds with rather unique properties are formed [1, 51]. When graphite intercalation compounds such as graphite hydrogen sulfate are subjected to very rapid heating (or flash heating), a type ofgraphite made up of flakes called exfoliated graphite is produced. Exfoliated graphite has a worm- or accordion-like shape with numerous macropores [44] (the macropores being the spaces left between successive flakes) and has a close relevance to adsorption, either for fundamental work (as it has a highly homogeneous surface) or connected with applications, either per se (e.g., in the removal of oil spills from sea water) [52], or as monolithic supports to deposit other adsorbent or catalyst phases [53]. A less widely known type of graphite is "high surface
30
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
area graphite," produced by extensive milling [54] and used as catalyst support or adsorbent [55]; it exhibits surface areas ranging between 100 and 300 m 2 / g. Various intermediate situations between perfect planar orientation and random orientation are found in solids such as pyrolytic carbons and cokes (i.e., typically graphitizable materials). Heat treatment temperature and the type ofprecursor (in the case ofcokes) are the main factors that determine the actual arrangement, i.e., tending toward either planar or random orientation. Unlike these solids, natural microcrystalline graphite (also termed "amorphous" graphite in certain contexts) should be classified as a randomly oriented material. This type ofgraphite is made up of very small domains that are oriented at random toward each other. Inside each domain, graphenes are stacked as in the most perfect types of graphite. The empty circle at the intersection of the two dashed lines in Fig. 2.9 corresponds to a natural microcrystalline graphite (from Mexico), which, according to the Raman band intensity ratio [ID/(ID+ Ie) = 67.5 %], is among the most randomly oriented materials studied by Cuesta et al. [49], whereas on the basis of XRD (d002 = 0.335 nm) it presents a high degree of graphitization. Likewise, synthetic isotropic high-density graphite, produced by cold isostatic pressing for use in a number of applications such as the structural parts of nuclear fission reactors [1], must be considered as randomly oriented from the point of view of nanotexture. On the other hand, graphite foams [56, 57], which are produced from mesophase pitches, are materials made up of an interconnected network of long graphite filaments and thus they exhibit isotropic material properties, though only on the macroscopic scale. On the nanoscale, however, they exhibit high Lc and La values, indicative of the presence of large graphitic domains. 2.4.2 Axial Orientation
Axial orientation is typical of materials with a tubular geometry (indeed, a fibrous morphology of carbon materials is possible, thanks to this axial orientation scheme). Axial orientation encompasses two extreme characteristic types: coaxial and radial. One of the best examples of coaxial orientation of BSUs is that of multiwall carbon nanotubes (MWCNTs), which consist of a series of concentric cylinders. Figure 2.11 shows one of the many examples of TEM micrographs of MWCNTs from the literature [58]. The parallelism of graphenes (in the so-called Russian doll arrangement) can be clearly seen and follows approximately the model that is depicted in an inset in the same figure. However, MWCNTs generally contain defects such as pentagonal and heptagonal rings, the presence of which produces bending and/or capping of individual nanotubes to yield structures such as carbon nanobamboos [59] and nanocones [60] (also termed nanohorns). The latter type of material has been quite frequently studied as adsorbent by Iijima, Kaneko and coworkers [61]. As in the case of fullerenes, readers interested in further information on these and other even more exotic nanotube-related solids (e.g., nanotubes with helical and toroidal structures) are referred to the review by 13artdosz et al. [15]. Adsorption by carbon nanotubes (especially the single-wall ones, SWCNTs) constitutes the
2.4 Structural Characterization of Carbon Materials
31
Figure 2.11 TEM image ofa concentric multiwall nanotube. The scale bar is 4 nm. The inset in the upper left comer illustrates the concentric arrangement of graphenes. (Reproduced from Re£ [58] with kind permission of Springer Science and Business Media.)
subject of Chapters 9, 15, and 16, and thus, we feel that the relevance to adsorption of these materials does not need to be dwelt upon any more at this point. Closely related to MWCNTs are CNFs, which can also be produced by the decomposition of hydrocarbon gas in the presence of transition-metal catalyst particles (catalytic chemical vapor deposition, CCVD) [62]. CNFs are highly graphitizable [63] as their main constituent is a layer ofpyrolytic carbon grown on an initially hollow graphitic filament generated on a catalyst particle. Some CNFs exhibit coaxial orientation, there being some doubt as to the distinction between them and MWCNTs [64]. CNFs can exhibit other characteristic nanotextures. One is the so-called platelet type, formed by platelets that are stacked on top of each other perpendicularly to the fiber axis. Other types are the herringbone and stacked-cup ones, which are depicted in Fig. 2.12; these two types of CNFs can be grown by the same method (chemical vapor deposition, CVD) , and they exhibit similar TEM images but differ in nanotexture and, hence, in their properties and potential applications [65]. As in the case of CNTs, CNFs are being actively investigated as adsorbents (if necessary, after activation), mainly in connection with gas storage.
32
Chapter
(a)
(b)
2
Overview of Carbon Materials in Relation to Adsorption
(c)
(d)
Figure 2.12 (a) and (c) Atomic models of stacked-cup and herringbone carbon nanofibers and (b) and (d) their respective simulated TEM images. (Reprinted from Ref. [65] with permission from Elsevier.)
The other extreme type of axial orientation consists of a radial alignment of carbon layers. This is typically found in ultrahigh-modulus, mesophase-pitchbased carbon fibers, which exhibit a radial arrangement in the cross-section of the fibers. Figure 2.13 shows two SEM micrographs of this type of fibers, where wedge-arranged straight layers are clearly visible, especially in the lower micrograph [66]. According to Mochida and coworkers [67,68], melt spinning (which is the basis of the preferred process for producing these fibers) is the most important step for determining the structures of mesophase-pitch-based CFs. During this process, microdomains (already present in the liquid crystal mesophase pitch) are aligned parallel to the fiber axis, forming fibrils. The fibrils are made up of a number of pleat units aligned parallel to the fiber axis. In turn, the pleats are composed ofgraphitic units. The radial open wedge probably forms by anisotropic shrinkage along the circumference of the fibers in the outer area at high temperature. Edie [69] has summarized the types of traverse textures that occur in mesophase-pitch-based CFs. The corresponding schemes are reproduced in Fig. 2.14, which highlights the striking variety of nanotextures that can be obtained from this type of precursor. It is worth mentioning at this point that the term" graphite fibers," still overused in certain contexts, is only justified when the material has a three-dimensional graphitic order. This term should therefore be limited to a few highly ordered mesophase-pitch-based CFs, which are truly graphitic, the general term "carbon fibers" being applied to the rest of the CFs. Less ordered, polyacrylonitrile (PAN)-based high-modulus and high-strength CFs have an intermediate, statistically cylindrical symmetry, which can be
2.4 Structural Characterization of Carbon Materials
33
Figure 2.13 SEM micrographs of mesophase-pitch-based carbon fibers. (Reprinted from Re£ [66] with permission from Elsevier.)
Radial
Onion-skin
Flat-layer
Radial-folded
Random
Line-origin
Figure 2.14 Traverse textures of mesophase-pitch-based carbon fibers. (Reprinted from Ref. [69] with permission from Elsevier.)
34
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
generated by introducing an elongation parallel to the fiber axis into the socalled crumpled sheets of paper model (Fig. 2.7, c and g) (see below) [42]. More detailed accounts on the nanotexture of these important carbon forms can be found elsewhere [70]. As is well known, the main field of application of CFs is as a reinforcement of polymeric matrices in composite materIals for aeronautic, aerospace, and other advanced uses. Improvement of the degree of adhesion between CFs and binders in these composites depends to a large extent on the surface properties of the fibers. This has been the motive of much research work on the surface properties of CFs, including many adsorption studies, which have been reviewed by various authors [4, 5,71]. Much less ordered than PAN-based high-strength CFs are the isotropic CFs. They are produced by the carbonization of isotropic pitch fibers (or other fibrous precursors such as phenolic resins or cellulose, including rayon), without any attempt to obtain a preferred orientation of the polyaromatic molecules in the fiber direction. Consequently, they have a random nanotexture and belong to the "low modulus" class of CFs [16]. Rather than being used for high-performance reinforcement purposes, they find their application as thermal insulators for furnaces or as reinforcements for cement [1]. Another important use of isotropic CFs is as a feedstock for the production of activated carbon fibers, a material dealt with in Section 2.4.4.
2.4.3 Point Orientation Like axial orientation, point orientation, which is characteristic of materials with a spherical geometry, falls into two extreme characteristic types. One is concentric, found, for instance, in carbon blacks [72] and carbon nanoonions [73], both of which consist of successive concentric spherical layers of BSUs. Figure 2.15 is a TEM image [74] of several carbon nano-onions (also termed multilayered fullerenes), which clearly illustrates the phenomenon of point concentric orientation. Many similar (but generally not so clear-cut) micrographs have been published for submicrometric, spherical carbon black particles as examined by TEM. A "venerable, and much reproduced" [75] model for the nanotexture of this material [76] can be seen in Fig. 2.7a. Carbon blacks have many connections with adsorption work since the inert nature of their surface and virtual absence of porosity (while exhibiting [15] external surface areas of up to 150 m 2 / g, thanks to their small particle size) have made them suitable as reference nonporous carbon materials, especially those that have been thermally treated at high temperatures. However, it is necessary to be cautious on this point since it is common in the literature to refer to such solids as "graphitized carbon blacks." It is true that, upon heat treatment, the outermost layers of certain carbon blacks may become polyhedral through the wiping-out of defects at boundaries between BSUs [77]. Correspondingly, graphitization begins to occur there, but the size of the flat regions is known never to surpass about one-third of the particle diameter [42]. Therefore, carbon blacks are generally considered nongraphitizable as a whole [15].
2.4 Structural Characterization of Carbon Materials
35
Figure 2.15 High-resolution TEM micrograph of carbon nano-onions. (Reprinted from Re£ [74] with permission from Elsevier.)
Carbon black is produced industrially in the form of different products (e.g., furnace black, thermal black, channel black, lampblack, acetylene black) with specific properties. In addition to the relevance of carbon black for basic research on adsorption, or as a reference solid, applications of this material in fields such as elastomer reinforcement, as modifier of certain properties of plastics (UV protection, electrical conductance, color), or as xerographic toners make its surface and interfacial properties extremely important. Soot is a randomly formed particulate material similar in nature to carbon black. The main (pragmatic, rather than conceptual) difference between these two carbon forms is that soot is generally formed as an unwanted by-product of incomplete combustion of pyrolysis, whereas carbon black is produced under strictly controlled conditions. Bansal and Donnet [78] have reviewed various possible mechanisms for the formation of soot and carbon black. Soot can retain a number of tars and resins on its surface. There is therefore some interest in studying the adsorption of polyaromatic hydrocarbons in soots, especially those of environmental significance such as diesel soot. Fullerene soot (also known as fullerene black, as in fact it is produced under controlled conditions) is a material generated in various fullerene synthesis processes by condensation of carbon species from the gas phase [79]. The subject of soot versus fullerene formation led the discoverers of buckminsterfullerene to propose a new mechanism for soot formation, known as the icospiral growth mechanism [80], which was refined later [81]. In this model, growth takes place in the form of a spiraling structure which, depending on the availability of a number of pentagonal rings, will either close up (fullerene) or curl around itself like a nautilus shell (soot). This is a field to which adsorption work
Chapter
E C
1J ,
2
Overview of Carbon Materials in Relation to Adsorption
0
-->..
A _-+ ----.... G-- ~__~ _.....~....
IISS-lllJI
A
o
B
11 11 1111'1 II.
c
m:n.-:::::z:ma E
F
Figure 2.16 Schematic model of the alignment of aromatic layers in carbon spherules with radial point orientation of basic structural units (BSUs). Several types of sections and the corresponding cross-sections of the spherules are also indicated. (Reprinted from Ref. [82] with permission from Elsevier.)
could contribute interesting results in order to clarify (i) the still eXIstIng controversy about the possible connection between fullerenes and soot/carbon black formation and (ii) the structure of soot/carbon black particles, which remains unresolved [75] despite the fact that these materials are among the oldest known carbonaceous solids. The radial arrangement is the other extreme type of point orientation. One of the carbon materials exhibiting a nanotexture very close to this model is the carbon spherule formed from mixtures of polyethylene and polyvinyl chloride by carbonization under pressure (30 MPa) [82]. A similar, although less "perfect" situation corresponds to mesophase spheres (Fig. 2.6), which are close to the radial point orientation near their surface. However, in their centers the orientation of the layers is not radial [83]. Figure 2.16 shows a model for the nanotexture of carbon spherules, where small aromatic layers are assumed to align approximately on circular conical surfaces with a common vertex at or near the center of the spherule. This proposed model was supported by an electron microdiffraction study of various micro-areas of sections of the spherules and by scanning electron microscopy, which evidenced the formation of a definite system of cracks on heat treatment to 3073 K [82].
2.4.4 Random Orientation Positioned directly opposite to planar orientation in Fig. 2.10 is random orientation, which is typical of highly disordered carbon materials such as chars, activated carbons, wood charcoals, or low-rank coals. The basic constituents of these materials are randomly intermingled, many of the spaces between them forming either opened or closed pores. "Pure" random orientation is found in
2.4 Structural Characterization of Carbon Materials
37
carbon materials just after the carbonization of some precursor polymers, and also in glass-like carbon [2]. The latter carbon form is nongraphitizable and exhibits uncommon properties such as impermeability to gases and extremely low chemical reactivity. One of the structural models proposed to account for the behavior of glass-like carbon is the "ribbon" model of Jenkins and Kawamura [84, 85], illustrated in Fig. 2.17. This assumes that the molecular orientation in the polymer precursor material is retained to some extent after carbonization. According to this model the "fibrils" in the polymer become curved and twisted ribbons of graphite-like carbon. Since the basic constituents of glass-like carbon cannot be directly imaged by TEM due to their very small size, the modeling of their nanotexture has often been based on a TEM examination of high-temperature-treated samples. Thus, the "shell" model, proposed by Shiraishi [86] for heat-treated glass-like carbon, involves cage-like components enclosing closed pores [87]. The hexagonal layers are locally oriented in a concentric scheme. This model is considered [2] to be realistic as it agrees with the presence of closed porosity in this material and, hence, gas impermeability. More recently, Harris [75, 88] proposed a new structural model that is applicable to nongraphitizable carbons in general, based on fullerene-like elements. This model is based on the examination ofTEM images, where fullerene-related particles have been observed by the authors in nongraphitizable carbons, in most of the cases treated at high temperatures. According to this DI0del, these carbons consist of discrete fragments of curved carbon sheets, in which pentagonal and heptagonal rings are dispersed at random throughout networks of hexagons. Harris has argued that this model, schematized in Fig. 2.18, is applicable to glass-like carbons prepared at temperatures of around 1273 K, since this structure probably has a low reactivity and permeability to gases compared with the "ribbon" model, particularly if there is a high proportion of completely closed particles. Glass-like carbons treated at higher temperatures are probably formed by larger basic building blocks that resemble incomplete giant fullerenes.
j • • • Lc
t
Figure 2.17 Jenkins-Kawamura "ribbon" model for glass-like carbon. (Reproduced from Ref. [85] with permission from the Royal Society.)
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
Figure 2.18 Harris model for the structure of nongraphitizable carbons based on fullerenelike elements. (Reproduced from Ref. [88]. Copyright Maney Publishing.)
The next carbon forms that we will consider within the random orientation framework are activated carbons (ACs) (as well as the chars from which they are prepared). ACs (and highly porous carbons in general) have a disordered structure, which is not, however, amorphous, as has been sometimes claimed. In ACs, there is a certain two-dimensional order, but no three-dimensional graphitic ordering. The disorder in the c direction is termed "turbostratic." This concept was coined to describe a graphite-like structure with a random orientation of layer planes along the a axis and a rotation of layer planes along the c axis, so that the interlayer spacing is 0.344 nm (cf 0.335 nm in a graphite single crystal). Early models used to describe the structure of ACs included the Franklin model itself (Fig. 2.5), and a ribbon-like structure [89] somewhat similar to the Jenkins and Kawamura model for glass-like carbon (Fig. 2.17). Interestingly, these models were based on results obtained with polyvinylidene chloride (PVDC) (or Saran) char, a typical nongraphitizable material (unlike the wellknown graphitizability of polyvinyl chloride [PVC] char). Various arguments have been used to criticize these and other models based on the occurrence of Sp3 carbon besides Sp2 carbon in nongraphitizable materials [10, 75]. Oberlin and coworkers [42,90] (Fig. 2.7c) used a "crumpled sheets of paper" model to describe the nanotexture of porous carbons. Figure 2.19 depicts this model and includes an enlarged sketch of pores in sucrose-based carbons [91]. In Fig. 2.19a, each zone of molecular orientation is represented by a shaded area inside of which each individual stack is shown either as a rectangular unit or as two parallel segments depending on whether they are observed in perspective
39
2.4 Structural Characterization of Carbon Materials
(a)
(b)
Figure 2.19 Sketches of crumpled sheet nanotexture of a sucrose-based high-temperature char. (a) Enlarged representation of pores and (b) model of crumpled sheets of paper. (Reprinted from Re£ [91] with permission from Elsevier.)
or edge-on. In sucrose-based carbons, which are typically nongraphitizable, the BSUs are locally oriented in very small regions, so that the LMO is smaller than 10 nm, and the LMOs are distributed at random in a given fragment. According to Rouzaud and Oberlin [91], during the release of heteroatoms that takes place when the sucrose is carbonized, the diameter ofLMOs remains practically constant, whereas their thickness decreases. Due to this, numerous small pores develop with an irregular entangled shape following the crumpled sheets of paper model [39]. Harris and coworkers [92, 93] have proposed that their model that is based on fullerene-like elements developed for nongraphitizable carbons in general (Fig. 2.18) may be applicable to PVDC chars and sucrose chars heat-treated to 2373-2873 K, where they detected closed nanoparticles that were usually faceted and often hexagonal or pentagonal in shape. The authors argue that such particles derive from fullerene-related structures present in the original, freshly prepared carbons (when examined by TEM, the low-temperature carbons were usually featureless and did not reveal a detailed structure, although occasionally some very small closed particles were found there) [88]. The model based on these findings (Fig. 2.18) predicts a micropore size of the order of 0.5-1.0 nm, which is similar to the pore sizes observed in many (ultra)microporous chars or carbons activated to low burn-off The empty spaces left between the fullerene-like fragments in Fig. 2.18 are not slit-shaped, but Harris [75] argues that micropores in carbons may have random shapes, as has been found either experimentally or by theoretical calculations. Harris [75] has also pointed out that his model based on fullerene-like elements has some connections with the so-called random schwarzite structure [94], which is based in turn on the ordered schwarzite structure [95] (the term schwarzite was coined after the German mathematician H.A. Schwarz, who first investigated the periodic minimal surfaces). The key feature of schwarzite is the occurrence
Chapter
4°
(a)
2
Overview of Carbon Materials in Relation to Adsorption
(b)
Figure
2.20 Possible structures for ordered schwarzite (a) and random schwarzite (b). (Reprinted with permission from Ref. [94]. © 1992 The American Physical Society.)
of negative curvature (in the Gaussian sense of the word) due to the presence of heptagonal rings. In an ordered schwarzite there is only negative curvature (Fig. 2.20a), whereas random (amorphous) schwarzites (Fig. 2.20b) contain pentagons (besides heptagons) and therefore combine concave and convex surfaces to yield solids that contain pores of molecular dimensions. Whatever the significance of the similarity between Harris' and schwarzite models (the random schwarzite shown in Fig. 2.20b should contain pores in the order of 0.5-1.0 nm, as in the model in Fig. 2.18), it is certain that schwarzites, if they occurred, would bring a new flavor to adsorption studies on carbons in both fundamental and applied fields. Their structure is continuous, with no edges or dangling bonds and would act as a kind of carbon sponge [96] with homogeneous pores of (perhaps tunable) molecular dimensions. Studies on the simulation of gas adsorption on schwarzites are reviewed in Chapter 14. However, despite many attempts to synthesize schwarzites [15], these efforts seem to have been fruitless to date, at least in the case of the ordered ones. Something similar occurs with certain hypothetical combinations (Fig. 2.4) of Sp3, Sp2, and sp hybridized atoms (graphdiynes, fullerenediynes), which could form potentially porous, molecular carbon solids [96, 97]. To obtain these, active research programs based on synthetic chemistry routes have been launched [15, 98]. To return to activated carbons, these can occur in very different physical forms: granular (or particulate), powdered, fibrous, or even membrane (the latter can be either unsupported, or, more commonly, supported). These basic physical forms can be combined with binders and extruded to form pellets, monoliths, or even paper. All of these materials, which are very frequently used as adsorbents, differ significantly in shape, but not in their intrinsic nanotextural features. All of them are isotropic and have their BSUs randomly oriented. Although they are specifically treated in Chapter 17, activated carbon fibers (ACFs) and derived cloths and felts deserve special mention here due to their uniform pore size distribution (PSD) and small and uniform fiber diameter, which confer on them, respectively, both size selectivity and rapid
2.4 Structural Characterization of Carbon Materials
41
adsorption/desorption kinetics. As announced at the end of Section 2.4.2, ACFs are produced from nongraphitizable (and hence isotropic) fibrous precursors [99, 100], and therefore, their nanotexture is also random. Indeed, anisotropic carbon fibers cannot be activated physically, and perhaps only chemical activation with a strong alkali may render them porous. The uniformity in pore size of ACFs is generally attributed to their small diameter (rv 10 f.Lm), so that many pores are directly open to the outer surface, porosity development proceeding by pore deepening rather than the creation of new pores. In some outstanding cases of pore size homogeneity such as polyaramid-derived ACFs, a certain "memory effect" seems to exist, whereby a highly ordered structure in the precursor yields a denser, less defective char than those prepared from other precursors [101, 102]. Carbon molecular sieves (CMSs) are porous carbons with an even narrower distribution of pore size (or, more precisely, of pore entrance size), which have the ability to differentiate between molecules of similar dimensions on the basis of adsorption kinetics. A key step in producing CMSs is the deposition of carbon on pore entrances [103], which is carried out usually by the CVD of hydrocarbons. Needless to say, CMSs have a nanotexture that is as randomly oriented as that of the parent ACs or ACFs on which the CVD was carried out. A type of carbon with a characteristic bimodal PSD is carbon gel [104]. The porosity of carbon gels consists of mesopores located between the constituent particles (or "nodules" [105]) and micropores located within these particles [15]. The various existing types of carbon gels (carbon xerogels, aerogels of cryogels) differ from each other in the type of drying (evaporative, supercritical, or freezedrying, respectively) to which their organic precursor gels (e.g., resorcinolformaldehyde gels) were subjected. Carbon gels retain to a large extent the mesoscopic structure of their parent organic gels [15]. They are therefore nongraphitizable and isotropic. Their nanotexture is random, as one might expect for carbons derived from the polymers used to produce the parent organic gels. We will not discuss here models for pores in carbons, as this topic is treated in Chapter 5, and elsewhere in specialist [15] or general reviews [106, 107]. For similar reasons, we will not discuss porosity control [44, 108] in detail. However, porous carbons prepared by the template technique, especially the ordered ones, deserve special attention. Ordered mesoporous carbons have been known to scientists since 1989 when two Korean groups independently reported their synthesis using mesoporous silicas as templates [109, 110]. Further achievements have been described in more recent reports [111, 112]. One might have expected that the nanotexture of these materials would merely reflect the nature of the precursor used, namely phenol-formaldehyde [109] or sucrose [110] in the two first ordered mesoporous carbon syntheses (as is well known, these two precursors would have yielded randomly oriented, isotropic carbon had they been pyrolyzed/activated under more conventional conditions). However, the mesopore walls in some ordered mesoporous carbons exhibited a graphite-like, polyaromatic character [113, 114], as described in Chapter 18. This information was obtained by nitrogen adsorption at low relative pressures, as in classical
Chapter
42
2
Overview of Carbon Materials in Relation to Adsorption
studies of highly homogeneous exfoliated graphite surfaces [115-117], and were in agreement with XPS results. Ryoos group has gone a step further and prepared a family of mesoporous carbons composed of graphitic framework structures consisting of discoid graphene sheets. This was the first case of porous carbons having a structural regularity on both the meso- and atomic scale [118]. The range of mesopores achievable by the template method is limited by the structure of the available templates. To overcome this limitation, Li and Jaroniec [119] developed a method, whereby, in contrast with the so-called colloidal templating technique (which involves a fluid-type carbon precursor), the so-called colloid-imprinted carbons are prepared by imprinting solid particles ofa mesophase pitch (used as carbon precursor) with colloidal silica particles. The primary pore structure ofthe resulting mesoporous carbons (which have spherical mesopores in the ""'-'6-60 nm range) can be preserved upon graphitization. In addition, the latter treatment made the resulting materials more energetically homogeneous and attractive for chromatographic separations [120]. The case of zeolite-templated, ordered microporous carbons (ZTCs) appears even more intriguing. Kyotani and coworkers [121, 122] were the first to succeed in synthesizing this type of material by carbonizing a precursor (acrylonitrile, furfuryl alcohol) polymerized in zeolite nanochannels, followed by the deposition of carbon from propylene by CVD. Some of the ZTCs obtained had [123] a BET surface area as high as 4100 m 2 / gl, (the surface area value decreased to 3730m2 /g 1 when calculated by the "subtracting pore method" [124]) a finely tailored micropore size and a long-range periodicity originating from the parent zeolite. The material was observed by scanning tunneling microscopy [125] to consist of carbon clusters about 1 nm in size. The porosity therefore is probably made up of the voids that separate the clusters. These clusters must have formed inside, and adapted themselves to the supercages of the zeolite Y template, and would appear to consist of nanographenes with a curved topology. In connection with the latter point, we believe that the recent discovery [126] of a method to isolate individual graphenes may pave the way for unexpected findings in the field of carbons in general and in that of adsorption by carbons in particular. This first example of a truly bidimensional material that is also the thinnest conceivable object will shed light on whether or not graphenes of different sizes are flat or curved, continuous or discontinuous, and whether they may be made to curl into, e.g., random, or even ordered, schwarzites. We are undoubtedly at the dawn of an exciting new era of carbon science.
2.5
CONCLUSIONS
The well-known basic structures of diamond and graphite are offset by others (carbynes) which, although they are scarce and even their existence is doubted, contribute to our understanding of structural trends in carbonaceous solids as a function of the type of chemical bond present.
References
43
Carbon forms with major Sp2 hybridization may be phenomenologically classified into graphitizable and nongraphitizable. The criterion established to differentiate these two classes is too empirical, and there are important exceptions (e.g., polyimide-derived carbons) to this rule. Nevertheless, these two concepts are useful for purposes of classification. Structural criteria provided by techniques such as X-ray diffraction, Raman spectroscopy, and, especially, transmission electron microscopy help to establish a rational classification of the wide range of carbon materials. The degrees of graphitization and orientation of the basic constituents justify the properties of very different types of carbon solids. The planar, axial, point, and random types of orientation of the BSUs, plus the situations in between, give rise to a wide variety of solids that have different relations with adsorption. Although random nanotexture is particularly relevant as it gives rise to a well-developed porosity, other nanotextures are also important in relation to other aspects of adsorption by carbons. Models developed to describe the molecular structure of carbons are subject to continuous improvements due to the need to obtain accurate descriptions of features evidenced by experimental results.
ACKNOWLEDGMENTS Financial support from the Spanish CSIC is gratefully acknowledged.
REFERENCES 1. Inagaki, M. (2000). New Carbons. Control of Structure and Functions. Elsevier. 2. Inagaki, M. and Kang, F. (2006). Carbon Materials Science and Engineering - From Fundamentals to Applications. Tsinghua University Press. 3. Bansal, R.C. and Goyal, M. (2005). Activated Carbon Adsorption. Taylor and Francis. 4. McKee, D.W. and Mimeault, V.J. (1973). Surface properties of carbon fibers. Chem. Phys. Carbon, 8, 151-243. 5. Ehrburger, P. (1990). Surface properties of carbon fibres. In Carbon Fibers, Filaments and Composites Figueiredo, C.A. Bernardo, R.T.K. Baker, and K.J. Hiittinger, eds). Kluwer Academic Publishers, pp. 147-161. 6. Bottani, E.J. and Tasc6n, J.M.D. (2004). Energetics of physical adsorption of gases and vapors on carbons. Chem. Phys. Carbon, 29, 209-423. 7. Edwards, LA.S. (1989). Structure in carbons and carbon forms. In Introduction to Carbon Science (H. Marsh, ed.). Butterworths, pp. 1-36. 8. Marsh, H. (1997). Carbon materials: an overview of carbon artifacts. In Introduction to Carbon Technologies (H. Marsh, E.A. Heintz, and F. Rodriguez-Reinoso, eds). Publicaciones de la Universidad de Alicante, Chapter 1, pp. 1-34.
a.L.
44
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
9. Schlagl, R. (1997). Carbons. In Handbook ofHeterogeneous Catalysis, Vol. 1 (G. Ertl, H. Knazinger, andJ. Weitkamp, eds). Wiley-VCH, pp. 138-91. 10. McEnaney, B. (1999). Structure and bonding in carbon materials. In Carbon Materials for Advanced Technologies (T.D. Burchell, ed.). Pergamon, Chapter 1, pp. 1-33. 11. Bourrat, X. (2000). Structure in carbons and carbon artefacts. In Sciences of Carbon Materials (H. Marsh and F. Rodriguez-Reinoso, eds). Publicaciones de la Universidad de Alicante, Chapter 1, pp. 1-97. 12. Fiala,J. (2000). New structures in carbon-based materials. In Sciences of Carbon Materials (H. Marsh and F. Rodriguez-Reinoso, eds). Publicaciones de la Universidad de Alicante, Chapter 2, pp. 99-115. 13. Delhaes, P. (2001). Polymorphism in carbons and parent materials. In Design and Control of Structure of Advanced Carbon Materials for Enhanced Peiformance (B. Rand, S.P. Appleyard, and M.F. Yardim, eds). Kluwer, pp. 3-27. 14. Bourrat, X. (2001). Characterisation of carbon structure. In Design and Control of Structure of Advanced Carbon Materials for Enhanced Peiformance (B. Rand, S.P. Appleyard, and M.F. Yardim, eds). Kluwer, pp. 43-64. 15. Bandosz, T., Biggs, M.J., Gubbins, K.E., et a1. (2003). Molecular models for porous carbons. Chem. Phys. Carbon, 28, 41-228. 16. Fitzer, E., Kachling, K.-H., Boehm, H.-P., and Marsh, H. (1995). Recommended terminology for the description of carbon as a solid. Pure Appl. Chem., 67, 473-506. 17. Sing, K.S.W., Everett, D.H., Haul, R.A.W., et a1. (1985). Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity. Pure Appl. Chem., 57, 603-19. 18. Rouquerol, J., Avnir, D., Fairbridge, C.W., et al. (1994). Recommendations for the characterization of porous solids. Pure Appl. Chem., 66, 1739-58. 19. Wilks, J. and Wilks, E. (1994). Properties and Applications of Diamond. ButterworthHeinemann. 20. Gielisse, P.J. (1998). Mechanical properties of diamond, diamond films, diamondlike carbon and like-diamond materials. In Handbook of Industrial Diamonds and Diamond Films (M.A. Prelas, G. Popovici, and L.K. Bigelow, eds). Marcel Dekker, Chapter 3, pp. 49-88. 21. Lopez-Gonzalez, J.D., Deitz, V.R., and Carpenter, F.G. (1961). Adsorption of nitrogen and argon on mineralogical graphite and diamond at 77 and 90 o K.J. Phys. Chem., 65,1112-19. 22. D'Evelyn, M.P. (1998). Surface properties of diamond. In Handbook of Industrial Diamonds and Diamond Films (M.A. Prelas, G. Popovici, and L.K. Bigelow, eds). Marcel Dekker, Chapter 4, pp. 89-146. 23. Heimann, R.B., Kleiman, J., and Salansky, N.M. (1984). Structural aspects and conformation of linear carbon polytypes (carbynes). Carbon, 22, 147-55. 24. Haddon, R.C., Palmer, R.E., Kroto, H.W., and Sermon, P.A. (1993). The fullerenes: powerful carbon-based electron acceptors. Phil. Trans. R. Soc. Lond. Sere A - Phys. Sci. Eng., 343, 53-62. 25. Charlier, J.-C., Gonze, X., and Michenaud, J.-P. (1994). First-principles study of the stacking effect on the electronic properties of graphite(s). Carbon, 32, 289-99. 26. Kavan, L. and Kastner, J. (1994). Carbyne forms of carbon: continuation of the story. Carbon, 32, 1533-6. 27. Heimann, R.B., Evsyukov, S.E., and Koga, Y. (1997). Carbon allotropes: a suggested classification scheme based on valence orbital hybridization. Carbon, 35, 1654-8.
References
45
28. Kudryavtsev, Yu. P., Evsyukov, S., Guseva, M., et al. (1997). Carbyne - a linear chainlike carbon allotrope. Chem. Phys. Carbon, 25, 1-69. 29. Tanuma, S.1. and Palnichenko, A. (1995). Synthesis of low density carbon crystal "carbolite" by quenching of carbon gas.]. Mater. Res., 10, 1120-5. 30. Dresselhaus, M.S. (1977). Future directions in carbon science. Annu. Rev. Mater. Sci., 27, 1-34. 31. Bundy, F.P., Bassett, W.A., Weathers, M.S., et al. (1996). The pressuretemperature phase and transformation diagram for carbon; updated through 1994. Camon, 34, 141-53. 32. Donnet, J.B. (2006). Les allotropes du carbone: une grande famille. Actualite Chimique, 295-296, 115-18. 33. Tanabe, Y. and Yasuda, E. (2000). Carbon alloys. Carbon, 38, 329-34. 34. Yasuda, E., Inagaki, M., Kaneko, K., et al. (eds) (2003). Carbon Alloys: New Concepts to Develop Carbon Science and Technology. Elsevier. 35. Marsh, H. (1986). Structure in carbons. In Carbon and Coal Gasification Figueiredo and J.A. Moulijn, eds). Martinus Nijho£I: pp. 27-56. 36. Hiittinger, K.J. (1990). The potential of the graphite lattice. Adv. Mater., 2, 349-55. 37. Chung, D.D.L. (2002). Review graphite.]. Mater. Sci., 37, 1475-89. 38. Franklin, R.E. (1951). Crystallite growth in graphitiZing and non-graphitizing carbons. Proc. R. Soc. Lond. A, 209, 196-218. 39. Oberlin, A. (1984). Carbonization and graphitization. Carbon, 22, 521-41. 40. Granda, M., Santamaria, R., and Menendez, R. (2003). Coal tar pitch: composition and pyrolysis behavior. Chem. Phys. Carbon, 28, 263-330. 41. Inagaki, M., Takeichi, T., Hishiyama, Y., and Oberlin, A. (1999). High quality graphite films produced from aromatic polyimides. Chem. Phys. Carbon, 26, 245-333. 42. Oberlin, A. (1989). High-resolution TEM studies of carbonization and graphitization. Chem. Phys. Carbon, 22, 1-143. 43. Oberlin, A., Bonnamy, S., and Rouxhet, P.G. (1999). Colloidal and supramolecular aspects of carbon. Chern. Phys. Carbon, 26, 1-148. 44. Inagaki, M. and Tascon, J.M.D. (2006). Pore formation and control in carbon materials. In Activated Carbon Surfaces in Environmental Remediation (T.J. Bandosz, ed.). Elsevier, pp. 49-105. 45. Emmerich, F.G. (1995). Evolution with heat treatment of crystallinity in carbons. Carbon, 33, 1709-15. 46. Iwashita, N., Park, C.R., Fujimoto, H., et al. (2004). Specification for a standard procedure of X-ray diffraction measurements on carbon materials. Carbon, 42, 701-14. 47. Tuinstra, F. and Koenig, J.L. (1970). Raman spectrum of graphite. J. Chem. Phys., 53, 1126-30. 48. Cuesta, A., Dhamelincourt, P., Laureyns, J., et al. (1994). Raman microprobe studies on carbon materials. Carbon, 32, 1523-32. 49. Cuesta, A., Dhamelincourt, P., Laureyns,J., et al. (1998). Comparative performance of X-ray diffraction and Raman microprobe techniques for the study of carbon materials. J. Mater. Chem., 8, 2875-9. 50. Inagaki, M. (1985). Microstructure of carbon materials (in Japanese). Tanso, 122, 114-21.
a.L.
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
51. Akuzawa, N. (2003). Intercalation compounds. In Carbon Alloys. New Concepts to Develop Carbon Science and Technology (E. Yasuda, M. Inagaki, K. Kaneko, et aI., eds). Elsevier, pp. 99-108. 52. Inagaki, M., Kang, F., and Toyoda, M. (2004). Exfoliation of graphite via intercalation compounds. Chern. Phys. Carbon, 29, 1-69. 53. Celzard, A., Mareche,J.F., and Furdin, G. (2005). Modelling of exfoliated graphite. Prog. Mater. Sci., 50, 93-179. 54. Seibold, K.J. and Fischer, F.G. (1973). Properties of high surface area graphite. Carbon, 11, 683. 55. Castillejos-Lopez, E., Bachiller-Baeza, B., Nevskaia, D.M., et al. (2006). Interactions between toluene and aniline and graphite surfaces. Carbon, 44, 3130-3. 56. Klett, J., Hardy, R., Romine, E., et al. (2000). High-thermal-conductivity, mesophase-pitch-derived carbon foams: effect of precursor on structure and properties. Carbon, 38, 953-73. 57. Gallego, N.C. and Klett, J.W. (2003). Carbon foams for thermal management. Carbon, 41, 1461-6. 58. Bhushan, B. (ed.) (2004). Springer Handbook of Nanotechnology. Springer, p. 43. 59. Lee, C.J., Park, J.H. and Park, J. (2000). Synthesis of bamboo-shaped multiwalled carbon nanotubes using thermal chemical vapor deposition. Chern. Phys. Lett., 323, 560-5. 60. Iijima, S., Yudasaka, M., Yamada, R., et al. (1999). Nano-aggregates of singlewalled graphitic nano-horns. Chern. Phys. Lett., 309, 165-70. 61. Yang, C.M., Noguchi, H., Murata, K., et al. (2005). Highly ultramicroporous single-walled carbon nanohorn assemblies. Adv. Mater., 17, 866-86. 62. Endo, M., Kim, Y.A., Matusita, T., and Hayashi, T. (2001). From vapor-grown carbon fibers (VGCFs) to carbon nanotubes. In Carbon Filaments and Nanotubes: Common Origins, Differing Applications? (L.P. Biro, C.A. Bernardo, G.G. Tibbetts, and Ph. Lambin, eds). Kluwer Academic Publishers, pp. 51-61. 63. Paredes, J.I., Burghard, M., Martinez-Alonso, A., and Tascon, J.M.D. (2005). Graphitization of carbon nanofibers: visualizing the structural evolution on the nanometer and atomic scales by scanning tunneling microscopy. Appl. Phys. A, 80, 675-82. 64. Biro, L.P., Bernardo, C.A., Tibbetts, G.G., and Lambin, Ph. (2001). Carbon Filaments and Nanotubes: Common Origins, Differing Applications? Kluwer Academic Publishers, pp. 343-5. 65. Kim, Y.-A., Hayashi, T., Naokawa, S., et al. (2005). Comparative study of herringbone and stacked-cup carbon nanofibers. Carbon, 43, 3005-8. 66. Cho, T., Lee, Y.S., Rao, A.M., et al. (2003). Structure of carbon fiber obtained from nanotube-reinforced mesophase pitch. Carbon, 41, 1419-24. 67. Yoon, S.-H., Korai, Y., and Mochida, I. (1996). Axial nano-scale microstructures in graphitized fibers inherited from liquid crystal mesophase pitch. Carbon, 34, 941-56. 68. Mochida, I., Yoon, S.-H., Takano, N., et al. (1996). Microstructure of mesophase pitch-based carbon fiber and its control. Carbon, 34, 941-56. 69. Edie D.D. (1998). The effect of processing on the structure and properties of carbon fibers. Carbon, 36, 345-62. 70. Oberlin, A., Bonnamy, S., and Lafdi, K. (1998). Structure and texture of carbon fibers. In Carbon Fibers, 3rd edn a.-B. Donnet, T.K. Wang, J.e.M. Peng, and S. Rebouillat, eds). Marcel Dekker, pp. 85-159.
References
47
71. Wang, T.K., Donnet, J.-B., Peng, J.C.M., and Rebouillat, S. (1998). Surface properties of carbon fibers. In Carbon Fibers, 3rd edn G.-B. Donnet, T.K. Wang, J.C.M. Peng, and S. Rebouillat, eds). Marcel Dekker, pp. 231-309. 72. Donnet, J.-B., Bansal, R.C., and Wang, M.-J. (eds) (1993). Carbon Black Science and Technology, 2nd edn. Marcel Dekker. 73. Ugarte, D. (1992). Curling and closure of graphitic networks under electron-beam irradiation. Nature, 359, 707-9. 74. Roy, D., Chhowalla, M., Wang, H., et al. (2003). Characterisation of carbon nano-onions using Raman spectroscopy. Chern. Phys. Lett., 373, 52-6. 75. Harris, P.J.F. (2003). Impact of the discovery of fullerenes on carbon science. Chern. Phys. Carbon, 28, 1-39. 76. Heidenreich, R.D., Hess, W.M., and Ban, L.L. (1968). A test object and criteria for high resolution electron microscopy. J. Appl. Cryst., 1, 1-19. 77. Hess, W.M. and Herd, C.R. (1993). Microstructure, morphology and general physical properties. In Carbon Black Science and Technology G.-B. Donnet, R.C. Bansal, and M.-J. Wang, eds). Marcel Dekker, Chapter 3, pp. 89-173. 78. Bansal, R.C. and Donnet, J.-B. (1993). Mechanism of carbon black formation. In Carbon Black Science and Technology G.-B. Donnet, R.C. Bansal, and M.-J. Wang, eds). Marcel Dekker, Chapter 2, pp. 67-88. 79. Dresselhaus, M.S., Dresselhaus, G., and Eklund, P.C. (1996). Science of Fullerenes and Carbon Nanotubes. Academic Press, Chapter 3, pp. 110-42. 80. Zhang, Q.L., O'Brien, S.C., Heath, J.R., et al. (1986). Reactivity of large carbon clusters - spheroidal carbon shells and their possible relevance to the formation and morphology of soot. J. Phys. Chern., 90, 525-8. 81. Kroto, H.W. and McKay, K. (1988). The formation of quasi-icosahedral spiral shell carbon particles. Nature, 331, 328-31. 82. Hishiyama, Y., Yoshida, A., and Inagaki, M. (1982). Microstructure of carbon spherules. Carbon, 20, 79-84. 83. Brooks, J.D. and Taylor, G.H. (1969). The formation of some graphitized carbons. Chern. Phys. Carbon, 4, 243-86. 84. Jenkins, G.M. and Kawamura, K. (1971). Structure of glassy carbon. Nature, 231, 175-6. 85. Jenkins, G.M., Kawamura, K., and Ban, L.L. (1972). Formation and structure of polymeric carbons. Proc. R. Soc. Lond. A, 327, 501-17. 86. Shiraishi, M. (1984). Introduction to Carbon Materials (in Japanese). Carbon Society ofJapan, p. 29. 87. Yoshida, A., Kaburagi, Y., and Hishiyama, Y. (1991). Microtexture and magnetoresistance of glass-like carbons. Carbon, 29, 1107-11. 88. Harris, P.J.F. (1997). Structure of non-graphitising carbons. Intern. Mater. Rev., 42, 206-18. 89. Ban, L.L., Crawford, D., and Marsh, H. (1975). Lattice-resolution electronmicroscopy in structural studies of non-graphitizing carbons from polyvinylidene chloride (PVDC). J. Appl. Cryst., 8, 415-20. 90. Oberlin, A., Villey, M., and Combaz, A. (1980). Influence of elemental composition on carbonization. Pyrolysis of kerosene shale and kuckersite. Carbon, 18, 347-53. 91. Rouzaud, J.N. and Oberlin, A. (1989). Structure, microtexture, and optical properties of anthracene and saccharose-based carbons. Carbon, 27, 517-29.
Chapter
2
Overview of Carbon Materials in Relation to Adsorption
92. Harris, P.J.F. and Tsang, S.C. (1997). High-resolution electron microscopy studies of non-graphitizing carbons. Phil. Mag. A, 76, 667-77. 93. Harris, P.J.F., Burian, A., and Duber, S. (2000). High-resolution electron microscopy of a microporous carbon. Phil. Mag. Lett., 80, 381-6. 94. Townsend, S.J., Lenosky, T.J., Muller, D.A., et al. (1992). Negatively curved graphitic sheet model of amorphous carbon. Phys. Rev. Lett., 69, 921-4. 95. Mackay, A.L. and Terrones, H. (1991). Diamond from graphite. Nature, 352, 762. 96. Benedek, G., Vahedi-Tafreshi, H., Barborini, E., et al. (2003). The structure of negatively curved spongy carbon. Diamond Relat. Mater., 12, 768-73. 97. Diederich, F. and Rubin, Y. (1992). Synthetic approaches toward molecular and polymeric carbon allotropes. Angew. Chem. Int. Ed., 31,1101-23. 98. Bunz, U.H.F., Rubin, Y., and Tobe, Y. (1999). Polyethynylated cyclic p-systems: scaffoldings for novel two and three-dimensional carbon networks. Chem. Soc. Rev., 28, 107-19. 99. Mays, T.J. (1999). Active carbon fibers. In Carbon Materials for Advanced Technologies (T.D. Burchell, ed.). Pergamon, Chapter 3, pp. 95-118. 100. Menendez-Diaz, A. and Martin-Gu1l6n, I. (2006). Types of carbon adsorbents and their production. In Activated Carbon Surfaces in Environmental Remediation (T .J. Bandosz, ed.) . Elsevier, pp. 1-47. 101. Freeman,J.J., Gimblett, F.G.R., Hayes, R.A., et al. (1991). Adsorptive properties of activated carbons prepared from Kevlar*. In Characterization of Porous Solids II (F. Rodriguez-Reinoso, J. Rouquerol, K.S.W. Sing, and K.K. Unger, eds). Elsevier, pp. 319-28. 102. Villar-Rodil, S., Suarez-Garcia, F., Paredes, J.I., et al. (2005).Activated carbon materials of uniform porosity from polyaramid fibers. Chem. Mater., 17,5893-908. 103. Yang, R.T. (2003). Adsorbents. Fundamentals and Applications. Wiley-Interscience, Chapter 5, pp. 79-130. 104. Pekala, R.W. and Kong, F.M. (1989). Resorcinol-formaldehyde aerogels and their carbonized derivatives. Abstr. Pap. Am. Chem. Soc., 197, 113. 105. Job, N., Thery, A., Pirard, R., et al. (2005). Carbon aerogels, cryogels and xerogels: influence of the drying method on the textural properties of porous carbon materials. Carbon, 43, 2481-94. 106. Byrne, J.F. and Marsh, H. (1995). Introductory overview. In Porosity in Carbons: Characterization and Applications O.W. Patrick, ed.). Edward Arnold, pp. 1-48. 107. McEnaney, B. (2002). Properties of activated carbons. In Handbook of Porous Solids, Vol. 3 (F. Schuth, K.S.W. Sing, and J. Weitkamp, eds). Wiley-VCH, pp. 1828-63. 108. Kyotani, T. (2000). Control of pore structure in carbon. Carbon, 38, 269-86. 109. Ryoo, R., Joo, S.H., and Jun, S. (1999). Synthesis of highly ordered carbon molecular sieves via template-mediated structural transformation. J. Phys. Chem. B, 103, 7743-6. 110. Lee, J., Yoon, S., Hyeon, T., et al. (1999). Synthesis of a new mesoporous carbon and its application to electrochemical double-layer capacitors. Chem. Commun., 2177-8. 111. Lee, J., Han, S., and Hyeon, T. (2004). Synthesis of new nanoporous carbon materials using nanostructured silica materials as templates. J. Mater. Chem., 14, 478-86.
References
49
112. Ryoo, R. and Joo, S.H. (2004). Nanostructured carbon materials synthesized from mesoporous silica crystals by replication. Stud. Surf. Sci. Catal., 148, 241-60. 113. Darmstadt, H., Roy, C., Kaliaguine, S., et al. (2002). Surface chemistry of ordered mesoporous carbons. Carbon, 40, 2673-83. 114. Darmstadt, H., Roy, C., Kaliaguine, S., et al. (2003). Pore structure and graphitic surface nature of ordered mesoporous carbons probed by low-pressure nitrogen adsorption. Microp. Mesop. Mater., 60, 139-49. 115. Dupont-Pavlovsky, N. (1965). Adsorption d'azote et d'argon sur des gels de silice dans Ie domaine des basses pressions d'equilibre. Bull. Soc. Chim. Fr., 1269. 116. Duval, X. and Thorny, A. (1975). Interpretation of krypton adsorption-isotherms on exfoliated graphite. Carbon, 13, 242-3. 117. Thorny, A. and Duval, X. (1994). Stepwise isotherms and phase transitions in physisorbed films. Surf. Sci., 299-300, 415-25. 118. Kim, T.-W., Park, I.S., and Ryoo, R. (2003). A synthetic route to ordered mesoporous carbon materials with graphitic pore walls. Angew. Chern., 42, 4375-9. 119. Li, Z. and Jaroniec, M. (2001). Colloidal imprinting: a novel approach to the synthesis of mesoporous carbons. J. Am. Chern. Soc., 123, 9208-9. 120. Li, Z. andJaroniec, M. (2004). Colloid-imprinted carbons as stationary phases for reversed-phase liquid chromatography. Anal. Chern., 76, 5479-85. 121. Ma, Z., Kyotani, T., Liu, Z., et al. (2001). Very high surface area microporous carbon with a three-dimensional nano-array structure: synthesis and its molecular structure. Chern. Mater., 13, 4413-15. 122. Ma, Z., Kyotani, T., and Tomita, A. (2002). Synthesis methods for preparing microporous carbons with a structural regularity ofzeolite Y. Carbon, 40, 2367-74. 123. Matsuoka, K., Yamagishi, Y., Yamazaki, T., et al. (2005). Extremely high microporosity and sharp pore size distribution of a large surface area carbon prepared in the nanochannels of zeolite Y. Carbon, 43, 876-9. 124. Kaneko, K., Ishii, C., Ruike, M., and Kuwabara, H. (1992). Origin of superhigh surface area and microcrystalline graphitic structures of activated carbons. Carbon, 30, 1075-88. 125. Paredes, J.I., Martinez-Alonso, A., Yamazaki, T., et al. (2005). Structural investigation of zeolite-templated, ordered microporous carbon by scanning tunneling microscopy and Raman spectroscopy. Langmuir, 21, 8817-23. 126. Novoselov, K.N., Geim, A.K., Morozov, S.V., et al. (2004). Electric field effect in atomically thin carbon films. Science, 306, 666-9.
ENERGETICS OF GAS ADSORPTION BY CARBONS: THERMODYNAMIC QUANTITIES Eduardo J. BottanP and Juan M.D. Tasc6n 2 Ilnstituto de Investigaciones Fisicoqufmicas Te6ricas y Aplicadas (lNIFTA), UNLP-CIC-CONICET, La Plata, Argentina; 21nstituto Nacional del Carb6n, CSIC, Oviedo, Spain
Contents 3.1 Introduction 3.2 Classical Thermodynamics 3.3 Statistical Mechanics 3.4 Thermodynamic Quantities and Experimental Results 3.5 Conclusions Acknowledgment References
3.1
53 54 59 66 71
71 72
INTRODUCTION
This chapter is concerned with the energetics of gas adsorption on carbons, or more specifically the thermodynamic quantities involved when carbon materials are employed as the adsorbents. Out of these solids, activated carbons, due to their exceptional surface area development and consequent technological implications, are the carbonaceous adsorbents that have attracted the attention of most publications. A more comprehensive account of the energetic aspects of adsorption on carbons has been published elsewhere [1]. The behavior of molecules adsorbed on the surface of a solid depends on the properties of both the adsorbent surface and the adsorbate itself Among the most relevant characteristics of the adsorbent are its chemical nature, its Adsorption by Carbons ISBN: 978-0-08-044464-2
© 2008 Elsevier Ltd. All rights reserved.
53
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
54
topography, and the presence of impurities. For the adsorbate, it is important to consider molecular size, shape, and its electronic configuration. It has also been demonstrated that temperature or, more precisely, the thermal energy of the adsorbate and lateral interactions are factors from which a balance is established that defines the thermodynamic properties of the adsorbed phase. Since it is the interactions that determine the behavior of the adsorbed phase, the problem reduces itself to one single concept: energy. This justifies the relevance of the thermodynamic aspects of adsorption that constitute the subject analyzed here. Relevant previous publications on adsorption energetics include, besides the classical text by Gregg and Sing [2], a more recent book by Rouquerol et al. [3] on adsorption by powders and porous solids. This book covers thermodynamic aspects of adsorption at the gas-solid and liquid-solid interfaces, and an entire chapter is devoted to adsorption on activated carbons. In addition, two books by Bansal et al. [4, 5] review in commendable detail the literature on adsorption by activated carbons. The task of reviewing the thermodynamics of gas adsorption by carbons is complicated by several factors. One of them is the variability of carbon materials discussed in Chapter 2. Another is the large amount of published literature, produced over a long period and generated under very different environments and thus potentially very different conditions. This has prompted us to focus our analysis on recent work, assuming that older studies may be outdated by new ones thanks to advances in instrumentation and to the logical progress of SCIence. This chapter includes cursory descriptions of the classical and statistical thermodynamic approaches to gas adsorption in the form of a summary of the main equations. These aspects provide the basic knowledge necessary to design, understand, and interpret the experiments and the data derived from them. We complete this with a brief description of the basic principles and applications that constitute the bridge between theory and experiments. In every section, a certain logical order of presentation is maintained. We group the results obtained on homogeneous surfaces followed by those obtained from heterogeneous solids. Each group begins with adsorption of simple gases and is followed by other adsorbates of increasing complexity. Finally, papers dealing with adsorption of gas mixtures are discussed.
3.2
CLASSICAL THERMODYNAMICS
Several formalisms have been developed leading to what may be called practical thermodynamics. These treatments include the analog of solution thermodynamics, where the adsorbent and the adsorbate are considered as components in a two-phase equilibrium [6]. Another way to study the system is to use the surface excess approach, whereby the properties of the adsorbed phase are determined in terms of the properties of the real two-phase multicomponent
3.2 Classical Thermodynamics
55
system and the properties of the same system without an interface [7]. This method could be considered as an extension of solution thermodynamics [8]. Its main shortcoming is that specification of the reference system is sometimes problematic. The most preferred approach for studying the thermodynamics of adsorption on solids considers the adsorbed phase as a distinct phase located on the surface of the solid, which is considered to be inert. Here, the concept of inertness of the adsorbent presupposes that no chemical reactions between it and the adsorbate are possible, and that the structure of the solid is rigid. Thus, in this formalism, the properties of the adsorbent and the gas phase are not explicitly included in the calculation. According to the law of conservation of energy, the total energy U is a constant provided that the system is isolated and its volume remains constant. Thus, oUS,V,n
=0
(3.1)
where V is the volume, n is the amount of substance, and S is the entropy. Equation (3.1) can be applied to the gas alone if the solid adsorbent is considered inert. Following the treatment developed by Hill in his classical papers on the thermodynamics of small systems [9], it is possible to divide the adsorption space into small elements. These elements are sufficiently large to enable the characterization of the gas contained in them, and small enough to allow the thermodynamic properties inside them to be considered as local ones. Since the first law of thermodynamics must be valid for each of the space elements, Eqn (3.1) takes the form (3.2) where the superscript identifies the space element and J.L is the chemical potential. Equation (3.2) represents the energy changes due to the reversible transfer of thermal energy between different space elements, the reversible work due to volume changes, and the energy due to the reversible transfer of gas molecules across the boundary of the space element. A more useful form of Eqn (3.2) can be derived through the concept of spreading pressure, by assuming that the force per unit area perpendicular to the surface of the solid is different from the force per unit area on the plane parallel to the surface. The resulting equation is dU(a)
= TdS(a) -
c/J(a)dA(a) -
pd V(a)
+ j.Ldn(a)
(3.3)
Equation (3.3), written for the adsorbed phase, can be used to derive expressions relating to experimental quantities like the amount adsorbed at a given temperature and pressure, such as
{s(a) - S(g)} d T = {V(a) -
V(g)}
dp + ~d
(3.4)
56
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
where p is the vapor pressure of the adsorbed layer, and where the tildes indicate a mean molar quantity. From Eqn (3.4), it is possible to obtain ClausiusClapeyron-type expressions. Moreover, ifthe spreading pressure is kept constant,
sea) v(a) -
S(g)
(3.5)
V(g)
or, if a constant temperature is maintained,
(d¢) = dp
[v(a) _
V(g)]
n(a)
A
T
(3.6)
Several enthalpies can be obtained to describe the adsorption equilibrium. The first is the so-called equilibrium enthalpy, qeq:
_q = eq
[;(a)
+pv(a) + cPA n(a)
-
[;(g)
-pV(g)
(3.7)
Assuming that the gas phase is ideal and that the density of the adsorbed phase is close to the density of the bulk liquid it is possible to arrive at (3.8a) or dlnp ) ( dT cf>
(3.8b)
Equation (3.8) shows that the equilibrium enthalpy can be calculated from experimental data, i.e., if the dependence of the isotherm on temperature at constant spreading pressure is known. However the use ofEqns (3.8a) and (3.8b) is cumbersome because it requires the previous calculation of the spreading pressure instead of the use of raw experimental data, i.e., n(a) vs p. Furthermore, the molar quantities appearing in Eqn (3.7) are not the natural variables for adsorption systems. In preference to molar quantities, partial molar entropy and internal energy are generally used; these quantities measure the changes in these properties when an infinitesimal change occurs in the number of adsorbed moles at constant temperature, pressure, and area. To be able to relate these quantities to experimental measurements, differentiation of the chemical potential of the adsorbed phase (in Eqn (3.7)) and rearrangement finally leads to (3.9) This expression gives the definition ofthe enthalpy known as isosteric enthalpy of adsorption. Equation (3.9) can be simplified to
T (V(g) -
(3.10) v(a))
3.2 Classical Thermodynamics
57
Again, as in the case of Eqn (3.8), if the adsorbed phase is assumed to be liquid-like, Eqn (3.10) reduces to
(
dlnp ) dT n,A
(3.11)
There is at least one other enthalpy related to the experimental data. This enthalpy is obtained in a calorimetric experiment under adiabatic conditions. The experiment consists of adding gas, in a reversible manner, to the calorimeter containing the adsorbent. An alternative process could be considered as a way to simplify the problem. Instead of adding gas to the system, imagine that the adsorbed molecules are transferred from the gas phase to the adsorbed phase by the action of a piston that changes the gas phase volume by an amount d V(g) . Assuming that the area of the adsorbate is unchanged during the process, that the adsorbed phase is liquid-like, and that the gas phase is ideal, it is possible to derive the expression _ qad -
qst
+ V (g)
(
dp ) d (a) n
(3.12) ad
This relationship shows that it is possible to calculate the isosteric enthalpy of adsorption from calorimetric experiments. In summary, it has been shown how the enthalpies of adsorption are obtained either calorimetrically or from the dependence of the isotherms on temperature. Although the definitions given above for the different enthalpies of adsorption are rigorous, it is necessary to show that they exhibit the same properties as the enthalpy of vaporization, i.e., the heat necessary to vaporize one mole ofliquid at constant pressure and using a reversible and isothermal process. Moreover, only the isosteric enthalpy is related in a simple way to the heat required in a process that is reversible and isothermic. Suppose that moles of adsorbate are transferred to the gas phase at constant temperature and pressure. According to the first law of thermodynamics:
an
(3.13) If the area of the solid is kept constant, this expression leads to
(oQ) A
ou(a)) = (On(a)
On(a) + T,p,A
OV(g)) +p ( - on(g) On(g)
(OU(g))
-On(g)
on(g)
+ p (ov(a)) -On(a)
on(a) T,p,A
(3.14)
Using the definitions of mean molar and partial molar quantities, Eqn (3.14) becomes (3.15)
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
58
Since Bn(g)
= -on(a) = On, Eqn (3.15)
can be written as (3.16)
According to Eqn (3.9) this quantity is equal to the isosteric enthalpy, qst' which turns out to be equal to the heat per mole evolved in the reversible transfer of an infinitesimal amount of adsorbate from the adsorbed phase to the gas phase at constant temperature, pressure, and area. Up to this point the essential equations have been presented. Now, it is possible to analyze the work carried out in connection with the classical thermodynamic approach. The first systematic study of a thermodynamic adsorption quantity was perhaps the work done by de Boer and coworkers [10] on the determination, interpretation and significance of the enthalpy and entropy of adsorption. Their papers analyzed almost all aspects of the experimental determination of the entropy and how to interpret the values obtained in terms of two extreme models, i.e., those of mobile and localized adsorption, which today have lost much of their usefulness. To catalog the behavior of the adsorbed film as localized or mobile is a very simplistic solution and it has been demonstrated [9] that in most cases the adsorbed film is neither completely localized nor completely mobile. This approach also is somehow outdated because numerical simulations provide a better microscopic interpretation of the system's behavior. Fomkin et al. [11] have reported a slightly different treatment in which they use Eqn (3.11) to calculate the isosteric heat of adsorption of perfluoropropane adsorbed on PAC (powdered activated carbon) microporous carbon. Agarwal et al. [12] determined the entropy of the adsorbed phase for methane, ethane, ethylene, propane, carbon dioxide, and nitrogen adsorbed at high pressures on activated carbon. Bottani et al. [13] also employed the classical approach to obtain the entropy of the adsorbed phase for N 2 and CO 2 adsorbed on "graphitized" carbon blacks. The authors discussed several problems regarding the precision of the obtained values using Eqn (3.11) or equivalent equations, and how they could be employed to characterize the surface of carbonaceous materials. More recently, Sircar et al. [14] employed the Gibbsian Surface Excess model to describe the multicomponent adsorption of gas mixtures. They also showed that this model for multicomponent adsorption could define unambiguously the isosteric heats of adsorption for the components of a gas mixture. These variables can be experimentally determined using multicomponent differential calorimetry and then be used to describe the nonisothermal behavior of practical adsorbents. Mezzasalma [15] employed a condition of maximum irreversible entropy production in the framework of a variational procedure where the isotherm equations are represented by a convergent sequence of ordinary functions. Milewska-Duda et al. [16] employed the thermodynamic approach described above to derive an adsorption isotherm, similar to the BET equation, which
3.3 Statistical Mechanics
59
can describe adsorption in microporous structures provided that restrictions for pore capacity are taken into account. Asnin et al. [17] demonstrated that the classical thermodynamic approach does not contradict the molecular statistical theory and that it yields equations that are more general. Based on Eqn (3.4) and similar expressions for the internal energy, they analyzed the particular case of the Freundlich adsorption isotherm. With data obtained from Kr adsorption on high-modulus carbon fibers, Drzal et al. [18] determined the isosteric heat of adsorption and the entropy of the adsorbed phase and demonstrated that such fibers, which undergo a high-temperature graphitization treatment, possess a very homogeneous surface very similar to that of the basal plane of graphite. Sircar [19] has presented a thermodynamic treatment ofgas mixture adsorption on heterogeneous adsorbents with particular emphasis on the estimation of the isosteric heat of adsorption. He stated that the isosteric heat of adsorption on an energetically heterogeneous adsorbent could vary substantially depending on the fractional loadings of the adsorbates, which, in turn, depend on the equilibrium gas-phase pressure, temperature, and composition. Myers [20] has developed thermodynamic equations for adsorption of multicomponent gas mixtures on microporous adsorbents based on the principles of solution thermodynamics. He argued that the conventional spreading pressure and surface variables, which describe bidimensional films, must be abandoned for adsorption in micropores, in which spreading pressure cannot be measured experimentally or calculated from intermolecular forces. Li et al. [21] recently reviewed the progress made in predicting the equilibria of multicomponent mixture adsorption. They discussed the problems encountered in applying theories developed for subcritical mixtures to supercritical gases. In a recent paper, Chiang et al. [22] reported values of the free energy, enthalpy, and entropy of adsorption of volatile organic compounds (exemplified by benzene and methylethylketone) on seven samples of activated carbon. The starting point for their development was Eqn (3.11) for the isosteric heat of adsorption. Linders et al. [23] determined adsorption heats from the adsorption equilibrium constant and found that these values agree quite well with those obtained from uptake experiments using the integrated form ofEqn (3.11). They analyzed the experimental data obtained for n-butane adsorbed on two commercial activated carbons (Kureha and Sorbonorit B3) and for hexafluoropropylene adsorbed on activated carbon.
3.3
STATISTICAL MECHANICS
The statistical mechanics formalism is probably the most efficient way to connect molecular models with experimental data. We present here a brief summary of the most important equations used for numerical simulations. Of all the statistical ensembles that can be employed, the canonical and grand canonical
60
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
are the most popular. We also restrict our treatment to classical statistical thermodynamics. Thus no quantum effects are taken into account. The probability that molecules 1, 2, 3, ... , N are in the volume elements dr l' dr 2' • • • , dr N located at r l' r 2' •• • , r N is given by the Boltzmann expression (3.17)
where
zt)
is the normalization factor given by (3.18)
Here V is the volume available to the gas molecules and U(r 1 , r 2 , ••• , r N ) is the potential energy of the N molecules. Thus U can be taken as the sum of two terms, gas-solid (Ugs ) and gas-gas (Ugg ) interaction energies: N
L
U(r1,r2, ... ,rN)=LUgs(r;)+
Ugg(rij)
(3.19)
l~i<j~N
i=l
It must be pointed out that this expression implicitly contains terms that depend on the orientation of the molecule with respect to the surface and the orientation of a given molecule with respect to its neighbors when the molecules are nonspherical. Equation (3.19) assumes that the potentials are additive and pairwise; since it does not include three-body or higher terms, this must be considered as an effective potential [8]. A system that is constrained to have a constant number of molecules, volume, and temperature constitutes a canonical ensemble. The thermodynamic properties of the system can be calculated from the corresponding partition function, Q(N, V, 7) [24]. For the adsorbed phase the partition function can be written as z(a) N Q( N " V T) -- N!A3N
(3.20)
where A is a factor that includes the kinetic properties of the molecule. Equation (3.20) shows that the partition function can be written as the product of a configurational factor and a kinetic or nonconfigurational factor, A. This greatly simplifies the application of this approach to the theory of physical adsorption [8]. The main assumption implicitly contained in Eqn (3.20) is that the structural properties of the molecules are independent of the intermolecular interactions in all the important configurations of the system. Even though there is evidence to suggest that this is not strictly true, it is possible to derive the appropriate expressions to calculate the extent of the changes in those properties
61
3.3 Statistical Mechanics
when the molecule is adsorbed. The thermodynamic properties of the system can be calculated from the following expressions:
(3.21)
A=-kTlnQ alnQ )
U=-k a(l/T) (
Q) aln-kT ( av(a)
(3.22) N,v(a),A
(3.23)
P-
N,T,A
Q
A,. (aln o/-kT - -)
aA
(3.24) NT vCa)
where A is the Helmholtz free energy, U is the total energy, p is the pressure, and
(3.25)
where V* is the volume of the gas in its standard state and thus equal to NkT. Using Eqns (3.20) and (3.22) give -U(a)
== k [ alnZt) ]
a(l/T)
_ vCa) A NCa)
3
kN(a) aInA a(l/T)
(3.26)
Equations (3.20) and (3.25), for the gas phase, gives
- U(g)
3
==
kN(g) -alnA -a( 1/ T)
(3.27)
From Eqns (3.26) and (3.27)
au(a) [ a 2 lnZt) ] U = aN(a) = -k aNa(l/T)
alnA 3
-(a)
vCa) ,A,NCa)
+k a(l/T)
(3.28)
and u(g)
fJ(g)
=-
N(g)
alnA 3
= k--a(l/T)
(3.29)
62
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
Since the isosteric heat is given by
qst =
_U(a)
+ [;(g) + kT
(3.30)
substituting Eqns (3.28) and (3.29) into (3.30) yield
qst
=k[
a2lnZt) ]
+ kT
(/)
aNa 1 T
(3.31)
VCa),A,NCa)
In the canonical ensemble approach, the adsorbed phase is treated as a separate phase with a known volume and containing a fixed number of molecules at constant temperature. This approach is somewhat unrealistic even though the results obtained can be successfully correlated with experimental data [8]. If we consider that the gas-solid interactions induce a smooth gradient in the density of the gas as the surface is approached, another formalism is necessary. The solution is obtained by adopting the grand canonical ensemble, in which the fixed variables are the volume, temperature and chemical potential. The unknown quantities would be, e.g., the number of molecules, energy, and pressure. The chemical potential of the adsorbed phase, once the equilibrium condition has been achieved, is equal to the chemical potential of the gas phase, which is determined from the density at a point far from the surface. The amount adsorbed, N(a), can be defined as the difference between the total number of molecules in the system, N, and N*, the number of molecules in a hypothetical system of equal volume but with no gas-solid interactions. The grand partition function for the system with gas-solid interactions is given by
a= L
N~O
Q N (V, T) exp
(J-LN)
(3.32)
kT
where QN is the canonical partition function and J-L is the chemical potential. Assuming ideal behavior for the gas, and using Eqn (3.20), the following set of equations can be derived for the thermodynamic properties of the system: alnS) N- ( N ) ( alna T, V,A alnS )
U=-k ( a(l/T)
(3.33)
(3.34) JL,V,A
4>A+pV = kT In
a
(3.35)
As indicated at the beginninmg of this section, the statistical mechanics approach is employed to connect molecular models with experimental data. It is used as well to give a theoretical basis to empirical models or to test
3.3 Statistical Mechanics
other theoretical models. For example, an isotherm equation was derived by Chen and Yang [25] for the adsorption of gases and vapors on microporous and mesoporous solids based on statistical mechanics principles. The empirical Dubinin-Astakhov (DA) and Dubinin-Radushkevich (DR) equations were then shown to be approximated forms of such an isotherm equation. Another example is the isosteric heat of adsorption of simple fluids on flat surfaces derived from a two-dimensional equation of state. Good agreement was found between the calculated values and the experimental results taken from the literature [26] for many adsorbate-adsorbent systems. The comparison was satisfactory for a wide variety of adsorbate-adsorbent systems, which was taken as an indication that molecules confined within micropores may be treated as two-dimensional fluids. Garbacz et al. [27] also derived an expression for the isosteric heat of a partially mobile monolayer of a single gas on a heterogeneous adsorbent surface. They optimized the model parameters in order to describe several sets of experimental data, especially the ones containing "graphitized" carbon black as an adsorbent. Floess and VanLishout [28] calculated the adsorption energy for different surfaces and pore configurations summing the Lennard-Jones potential for the gas-solid interaction of a molecule with a graphite finite-size basal plane surface. They showed that restricted diffusion occurs only for a small range of pore sizes between c. 0.64 and 0.58 nm. In larger pores, the adsorbate is mobile, whereas pores smaller than 0.58 nm are assumed to be largely inaccessible. Murata and Kaneko [29] proposed a new equation of the absolute adsorption isotherm [30] for a supercritical gas in order to describe the adsorption of methane on activated carbons. The environmental aspects of supercritical gases confined in nanospaces have been reviewed by Kaneko [31]. The model assumes that the adsorption in micropores is slightly enhanced compared with that on a flat surface. The authors supported this assumption on the basis of comparison plots of experimental data obtained on activated carbons and flat surfaces like nonporous carbon black. Several aspects of the adsorption of self-associating molecules in microporous structures have been developed by Talu and Meunier [32]. Their approach is similar to that of the chemical interpretation of nonideality of vapor and liquid phases. The theory leads to type V isotherm behavior and can explain the transition between types I and V. The data at one temperature are represented by three parameters: Henry's law constant, saturation capacity, and reaction constant for "cluster" formation in the micropores. To describe a set ofisotherms obtained over a certain temperature interval, the theory can be used with five temperature-independent parameters to determine the entire phase behavior including the heat of adsorption. Besides obtaining a good agreement with experimental data, they found that the dimerization enthalpy of water in the micropores is lower than that in the vapor phase. Rudzinski and Panczyk [33] have recently reviewed the classical theories of adsorption and desorption kinetics and concluded that models based on the absolute rate theory were challenged by new theories linking the rate of adsorptiondesorption with the chemical potentials of bulk and adsorbed molecules. Of the
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
latter theories, the so-called statistical rate theory deserved the most advanced theoretical development. This model is based on both quantum mechanics and statistical thermodynamics. In particular, it uses as starting point both the Langmuir isotherm and the integral adsorption equation. The authors concluded that the parameters obtained with the statistical rate theory reproduced very well the behavior of both the kinetic and equilibrium aspects of isothermal adsorption measurements, and the parameters had a fully physical meaning. The absolute rate theory can fit the kinetic adsorption measurements at constant temperature (kinetic adsorption isotherms) using many sets of parameters, but some of the parameters determined always lacked physical meaning. Pan et al. [34] used the nonlocal density functional theory (DFT) [35] and the three-process Langmuir model (TPLM) [36] to predict the adsorption heats of propane and butane on carbon and compared these results with experimental data determined from isotherms measured on BAX-activated carbon (Westvaco) in the 297-333 K temperature interval. Both models agreed in showing that the adsorption heat for butane was c. 10 kJ/mol higher than that of propane at the same loading. The satisfactory agreement found prompted the authors to propose the use of the DFT method as it requires only one experimental isotherm in contrast with the numerous isotherms required by the classic technique. Following the same approach, Pan et al. [37] also predicted the adsorption heats of three binary gas mixtures (C0 2-C 2 H 4 , CH 4-C 2 H 6 , and CH 4-C 3 H g ) on homogeneous (BPL-6 carbon) and heterogeneous carbons (Westvaco BAX) at 350 K (here, homo/heterogeneity refers to the pore sizes). The isosteric heats showed complex behavior for the nonideal systems (the CH 4 -C 2 H 6 and CH 4-C 3 H g gas mixtures). Adsorbent heterogeneity played an important role in determining the behavior of the isosteric heats compared with the pure states. The authors attributed these differences to effects caused by the presence of the other (opposing) component of the mixture and to differences in the intermolecular forces between the adsorbate molecules. Myers [38] proposed an equation of state for the adsorption of mixtures of gases in porous materials. Even though the examples shown in this paper concerned adsorption on zeolites, the results can be employed to explain adsorption on porous carbonaceous materials. The author introduced the desorption quantities (enthalpy, entropy, and free energy), which have the advantage of being linearly correlated with the desorption properties of the pure components of the mixture. Bhatia [39] studied the transport of adsorbates in microporous random networks in the presence of an arbitrary nonlinear local isotherm. The transport model was developed by means of a correlated random walk theory, assuming pore mouth equilibrium at an intersection in the network and a local chemical potential gradient driving force. The author tested this model with experimental data of CO 2 adsorption on Carbolac measured by Carman and Raal [40]. He concluded that the experimental data are best predicted when adsorbate mobility, based on the chemical potential gradient, is taken to have an activation energy equal to the isosteric heat of adsorption at low coverage, obtained from the Henry's law region. He also concluded that the choice of the local isotherm
3.3 Statistical Mechanics
65
affects the predictions for the dependence of the diffusivity on coverage, even among isotherms that fit the equilibrium data equally well. Shekhovtsova and Fomkin [41] developed a discrete site model to describe the adsorption of methane on microporous adsorbents. The model was tested with adsorption on zeolite NaX and an activated carbon. A sharp decrease in the heats of adsorption was observed at high adsorbed amounts even in the supercritical temperature range. A multilayer adsorption theory was developed by Wang and Hwang [42] to describe the behavior of several adsorbates on activated carbons. The adsorbates employed included several alkanes, hydrogen sulfide, and carbon monoxide. The isosteric heats of adsorption for all gases were determined using the Clausius-Clapeyron equation. Cerofolini and Rudzinski [43] have reviewed the theoretical principles of single gas and mixture adsorption on heterogeneous surfaces. Their review is chronologically arranged from the earliest to the latest approaches. In the same book, Tovbin [44] reported the application of lattice-gas models to explain mixed-gas adsorption equilibria on heterogeneous surfaces; he also discussed [45] the kinetic aspects of adsorption-desorption on flat heterogeneous surfaces. The book [46] also contains other papers on different aspects of adsorption for the reader interested in surface diffusion processes. It is generally accepted that the adsorption of binary and ternary vapor mixtures by activated carbon beds can be successfully described by combining the Dubinin equation with the theory of Myers and Prausnitz [47]. A major advantage of this approach lies in the simplicity of the parameters required to describe the adsorption of vapor mixtures over a wide range of pressures and temperatures. Moreover, the combination of this method with computer simulation models developed by Ladugie et al. [48] extends the theory to the case of dynamic adsorption by activated carbons. Wintgens et al. [49] employed this approach to investigate the adsorption of vapor mixtures on activated carbons. The components were immiscible in the liquid state. They concluded that it
is possible to predict the adsorption of vapor mixtures corresponding to miscible and immiscible liquids if the variation of the characteristic energy of one component is taken into account. Riccardo and coworkers [50, 51] reported the results of a statistical thermodynamic approach to study linear adsorbates on heterogeneous surfaces based on Eqns (3.33)-(3.35). In the first paper, they dealt with low dimensional systems (e.g., carbon nanotubes, pores ofmolecular dimensions, corners in steps found on flat surfaces). In the second paper, they presented an improved solution for multilayer adsorption; they compared their results with the standard BET formalism and found that monolayer capacities could be up to 1.5 times larger than the one from the BET model. They argued that their model is simple and easy to apply in practice and leads to new values of surface area and adsorption heats. These advantages are a consequence of correctly assessing the configurational entropy of the adsorbed phase. Rzysko et al. [52] presented a theoretical description of adsorption in a templated porous material. Their method of solution uses expansions of size-dependent correlation functions into Fourier series. They tested
66
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
their method for model systems composed of hard spheres, or more specifically a hard sphere fluid in a polydisperse hard sphere disordered quench matrix.
3-4
THERMODYNAMIC QUANTITIES
AND EXPERIMENTAL RESULTS
The connection between experimental results and thermodynamic quantities gives rise to different problems depending on the experimental design employed to obtain the data. In what follows we will analyze the relationship between the various heats of adsorption and the experimental data obtained under different conditions. To obtain enthalpy variations directly from experiments, it is necessary to carry out these experiments at constant pressure (or spreading pressure), but in general experiments are not carried out under these conditions. This is the reason why Letoquart et al. [53] recommend expressing the experimental data in terms of internal energy. In this way, writing the necessary expressions to calculate all the thermodynamic quantities is straightforward. Another problem arises from the fact that most thermodynamic equations have been deduced for a closed system, which is often not the experimental situation [53]. As usual, we consider the solid adsorbent to be inert, in the sense that its internal energy is constant at constant temperature and its total surface area is kept constant. For the gas phase, we assume ideal behavior. In any experimental setup that constitutes an open system, what is measured is the heat exchange with the surrounding media, which includes the exchange of matter and a certain amount of work. In this case, one is tempted to introduce thermodynamic potentials, i.e., to introduce entropy. Nevertheless, we feel it more prudent to find a way to relate the experimental data with the internal energy without introducing any additional hypothesis. The evaluation of internal energy change allows the energy conservation law to be used in two ways. We assume that our system contains c components distributed among 'P phases. During the adsorption process the system receives some work (W), heat (Q), and a certain amount of moles of the ith component (LlNi ) from the exterior, each one having a known molar internal energy, B ie . Under these conditions the change in internal energy is given by i=c
~E
= W+Q+" L...J B·
Ie
~N.I
(3.36)
i=l
Alternatively, if the ith component has a molar internal energy Bij when it is in the jth phase (where Nij represents the number of moles of the ith component present in the jth phase), the internal energy change can be written as j=cp i=c
LlE
= LLLl (NijBij) i=l i=l
(3.37)
3.4 Thermodynamic Quantities and Experimental Results
Using Eqns (3.36) and (3.37) it is easy to obtain the expression that gives the experimentally determined thermal effect, Q, in terms of the internal energies. The result is j=({) i=c
i=c
"" d (N..lJ e lJ..) - "" Q = "" ~~ ~ e·Ie dN. - W I
j=l i=l
(3.38)
i=l
Consider an infinitesimal transformation that takes the system from one equilibrium state to another one that is close to it. To calculate the amount of heat involved in the transformation it is necessary to derive Eqn (3.38). Thus (3.39) where we have used the fact that dN = dNa + dNg . If we also use the fact that the gas admitted into the adsorption cell is at the same temperature (thus its internal energy does not change), it is possible to write Eqn (3.39) as follows: (3.40) N ow it is necessary to calculate the work involved, d W. Given that the gas is ideal, it follows that:
dW=PoV
(3.41)
This equation is easily transformed into (3.42) Considering that the total volume of the system is constant, Vg , the change in the number of gas moles can be expressed in terms of the pressure. Thus Eqn (3.42) leads to
dW = RTdNa + VgdP
(3.43)
Substituting into Eqn (3.40): (3.44) This equation can be written in an equivalent form to show that it corresponds to a differential heat of adsorption. (3.45) This is in fact an isothermal heat of adsorption. The last term, which depends on the experimental setup, is obtained from the calibration of the equipment
68
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
and the slope of the isotherm. We must now consider the fact that Eqn (3.45) is valid if the gas is reversibly admitted into the adsorption cell, a condition that is seldom obtained. We must now develop the expressions corresponding to a finite addition of gas. The heat experimentally measured corresponds to a change in the number of adsorbed moles, flNa = N z - N l . Equation (3.45) must be integrated, giving
The result is
Q = (B az -
Bg ) N aZ - (B al - Bg ) N al -
RT (Naz - N al )
-
Vg (Pz - Pl )
(3.47)
The first and second terms of this equation are the integral heats of adsorption, Qint. Thus,
Q
= Qint_ Qint-RTdN z 1 a
Vg (PZ -P) 1
(3.48a)
We have shown that the experimentally measured heat exchange is directly connected to the integral heat of adsorption at constant temperature. The experimental data needed to perform the calculation are Q, the number of moles adsorbed at two pressures, these pressures and the volume of the adsorption cell (the so-called dead space). In a completely analogous way, it is possible to obtain an expression corresponding to the adiabatic heat of adsorption. The final result is
(3.48b) Now we will discuss a series of papers using the classical thermodynamic, the statistical thermodynamic, direct determination of the heat of adsorption, or a combination of these methods. The theory behind each paper has been outlined in the previous sections as well as in the present one. Della Gatta [54] has considered the energy changes occurring during adsorption in gas-solid and solution-solid systems in connection with the measurement of the adsorption enthalpies. He focused on the description of heat-flow microcalorimeters and calibration techniques. Piper et al. [55] have published a good description of an adiabatic calorimeter. Other types of calorimeters are the conduction calorimeter (the temperature of the sample is made equal to the surrounding temperature by simple conduction), compensation calorimeters (the sample temperature is made equal to the surroundings by means of power compensation), and the "isoperibol" calorimeter (this is a conventional temperature rise calorimeter, also known as Thomsen or Berthelot calorimeter) [3].
3.4 Thermodynamic Quantities and Experimental Results
Groszek [56, 57] studied the adsorption of simple gases (C0 2, CH 4 , S02' 02' He, and N 2) on microporous carbons using flow adsorption microcalorimetry. Shen and Bulow [58] demonstrated that the isosteric adsorption technique (Eqns (3.11) or (3.31)) is a useful and effective tool to obtain highly accurate thermodynamic data for microporous adsorption systems like the heat of adsorption given by Eqn (3.47). They studied the adsorption of CO 2 and N 2-0 2 mixtures on a super-activated, almost entirely microporous, carbon (M-30, from Osaka Gas) and three faujasite-type zeolites. They also estimated the energetic heterogeneity of the solids due to specific interactions between the adsorbate and the solid. Salem et al. [59] determined several thermodynamic functions from the excess and absolute adsorption isotherms for N 2, Ar, and methane on an activated carbon (AS, used for respiratory protection) and on a 13X zeolite. They concluded that for a consistent thermodynamic interpretation ofhigh-pressure excess adsorption data it is necessary to consider the specific adsorption quantities (the ones that can be measured directly). However, such quantities depend very much on the initial values of adsorption, i.e., the region of Henry's law. If this region is not included in the experimental data, the integral molar and specific adsorption quantities are estimated with a systematic error of unknown magnitude. Another conclusion they reached is that a molecular interpretation of high-pressure adsorption data can be achieved only by using the absolute isosteric and differential molar adsorption quantities. In the case of microporous solids, this is relatively easy to do; however, it is a very difficult task in the case of macroporous and nonporous solids. Rychlicki and Terzyk [60] determined the heat of adsorption and the integral molar entropy of methane adsorbed on microporous carbons with different degrees of oxidation. They found that the oxidation of the micropore surface affected the adsorption of methane because of the existence of an endothermic effect during adsorption probably due to specific adsorbate-adsorbent interactions. A similar effect was found in the case of CCl4 adsorption [61, 62]. These authors also studied the adsorption of methanol, ethanol and CCl4 on a series of microporous carbons [63] and compared the data obtained with the adsorption of the same adsorbates on a "graphitized" carbon black. They concluded that at low coverage, association of the studied gases in the micropores does not occur. The adsorption of gas mixtures has been extensively studied. For example, Wendland et al. [64] applied the Born-Green-Yvon approach using a coarse grained density to study the adsorption of subcritical Lennard-Jones fluids. In a subsequent paper, they tested their equations with simulated adsorption isotherms of several model mixtures [65]. They compared the adsorption of model gases with an equal molecular size but different adsorption potentials. They discussed the structure of the adsorbed phase, adsorption isotherms, and selectivity curves. Based on the vacancy solution theory [66], Nguyen and Do [67] developed a new technique for predicting the multicomponent adsorption equilibria of supercritical fluids in microporous carbons. They concluded that the degree of adsorption enhancement, due to the proximity of the pore
70
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
walls, is different for different adsorbates and that it increases with a decrease in pore size. They found good agreement between their model and experimental data. Do and Do [68] have recently reviewed the adsorption of supercritical fluids in porous and nonporous carbons. Ustinov and Do [69] proposed a model for binary mixture adsorption accounting for energetic heterogeneity and intermolecular interactions based on statistical thermodynamics. This model is able to describe the molecular rearrangement of a mixture in a nonuniform adsorption field inside a cavity. The Helmholtz free energy obtained has upper and lower limits, which define a permissible range in which all possible solutions may occur. One limit corresponds to a completely chaotic distribution of molecules within a cavity, and the other to a maximum ordered molecular structure. Their model can also describe the negative deviations from Raoult's law exhibited by N 2-0 2 mixtures. Bakaev and Steele [70] developed a lattice gas model of mixed adsorption on heterogeneous surfaces. The model considers an ideal mixture since it is assumed that the difference between the energies of the components on each adsorption site is the same. They showed that mixing at constant spreading pressure in this case is equivalent to that at constant coverage. The interaction between adsorbed molecules induces deviations from ideality, but the excess chemical potentials of the components calculated in the random mixing approximation depend linearly upon mole fractions, unlike the regular bulk solutions, which for the same approximation have a quadratic dependence on mole fractions. Experimental separation factors (selectivities) for methane-ethane mixtures on activated carbon [71] were compared with those calculated according to the lattice gas model of mixed adsorption [70] or by means of the ideal adsorption solution theory (lAST). The lattice gas model under discussion provided only rough evaluations of the selectivities in mixed adsorption but, unlike the lAST predictions, showed no dependence on the composition of the gas phase or its total pressure. According to the authors, the main limitation of the lattice gas model is that of the Langmuir model as a whole, namely, the requirement that the adsorption capacities of all the components in a mixture be equal. More recently, Ustinov and coworkers [72, 73] developed a thermodynamic approach based on an equation of state to model the gas adsorption equilibrium over a wide range of pressure. Their model is based on the Bender equation of state, which is a virial-like equation with temperature dependent parameters based on the Benedict-Webb-Rubin equation of state [74]. They employed the model [75, 76] to describe supercritical gas adsorption on activated carbon (Norit R 1) at high temperature, and extended this treatment to subcritical fluid adsorption taking into account the phase transition in elements of the adsorption volume. They argued that parameters such as pore volume and skeleton density can be determined directly from adsorption measurements, while the conventional approach of He expansion at room temperature can lead to erroneous results due to the adsorption of He in narrow micropores of activated carbon. More recently, Li et al. [77] have reviewed the thermodynamic basis ofa novel model ofthe combined cycle ofa solar-powered adsorption-ejection refrigeration
Acknowledgment
71
system. They described the adsorption isotherms using a virial-like equation and the heats of adsorption were calculated with expressions derived from Eqn (3.48a).
3.5
CONCLUSIONS
The determination ofadsorption thermodynamic quantities such as adsorption heats can now be performed through direct or indirect methods with a great degree ofaccuracy. The foundations of gas-solid interface calorimetry have been well established by combining adsorption microcalorimetry with adsorption in quasi-equilibrium. The experimental results reported so far, obtained from different calorimetries, concur with the values calculated from adsorption isotherms. There is a more or less generalized agreement that the isosteric adsorption heat is strongly affected by the microstructure of the adsorbent, particularly in the case of porous solids. This magnitude is better suited for structural analysis than other thermodynamic quantities. The use of the Clausius-Clapeyron equation to determine the isosteric adsorption heat has several limitations both theoretical and experimental, that are well known. Gas mixture adsorption is a field that is still waiting for a better theory to explain the experimental data. The Ideal Adsorbed Solution Theory cannot explain all the facts and needs to be replaced by a new model that includes nonideal effects, and adsorbent surface heterogeneity in particular. This field is acquiring increasing relevance because of its technological implications. Adsorption calorimetry, based on the use of different adsorbates, is now employed to probe the effects of various types of surface modification treatments on the surface chemistry of solids. The technique is employed in particular to investigate and characterize activated carbons. Recent studies show good agreement between the results from this technique and immersion calorimetry. In summary, the determination of thermodynamic quantities is almost a standard routine connected to the majority of experimental and simulation techniques. Thermodynamic models for adsorption can now be validated using both experimental and computer simulation results. New developments are regularly reported concerning new adsorption equations or models to explain experimental data of a very different nature.
ACKNOWLEDGMENT
Financial support from the Spanish CSIC is gratefully acknowledged.
72
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
REFERENCES 1. Bottani, EJ. and Tasc6n, J.M.D. (2004). Energetics of physical adsorption of gases and vapors on carbons. Chem. Phys. Carbon, 29, 209-423. 2. Gregg, SJ. and Sing, K.S.W. (1982). Adsorption, Surface Area and Porosity, 2nd edn. Academic Press. 3. Rouquerol, F., Rouquerol,J., and Sing, K. (1999). Adsorption by Powders and Porous Solids. Principles, Methodology and Applications. Academic Press. 4. Bansal, R.C., Donnet, J.B., and Stoeckli, F. (1988). Active Carbon. Marcel Dekker. 5. Bansal, R.C. and Goyal, M. (2005). Activated Carbon Adsorption. Taylor and Francis, Chapter 2, pp. 67-143. 6. Myers, A.L. and Monson, P.A. (2002). Adsorption in porous materials at high pressure: theory and experiment. Langmuir, 18, 10261-73. 7. T6th, J. (2003). On thermodynamic inconsistency of isotherm equations: Gibbs' thermodynamics.]. Colloid Interface Sci., 262, 25-31. 8. Steele, W.A. (1974). The Interaction of Gases with Solid Surfaces. Pergamon Press. 9. Hill, T.L. (1994). Thermodynamics of Small Systems. Dover Publications. 10. de Boer, J.H. (1953). The Dynamic Character of Adsorption. Clarendon Press. 11. Fomkin, A.A., Nikiforov, Yu.V., Sinitsyn, V.A., and Savel'ev, E.G. (2002). Adsorption ofperfluoropropane on the PAC microporous carbon adsorbent. Russ. Chem. Bull., Int. Ed., 51, 2161-4. 12. Agarwal, R.K., Amankwah, K.A.G., and Schwarz,J.A. (1990). Analysis ofadsorption entropies of high pressure gas adsorption data on activated carbon. Carbon, 28, 169-74. 13. Bottani, E.G., Llanos,J.L., and Cascarini de Torre, L.E. (1989). Entropy determination from physisorption isotherms: a useful tool for carbon surface characterization. Carbon, 27, 531-6. 14. Sircar, S., Mohr, R., Ristic, C., and Rao, M.B. (1999). Isosteric heat of adsorption: theory and experiment.] Phys. Chem. B, 103, 6539-46. 15. Mezzasalma, S.A. (1999). Hysteresis and isotherm equations in gas-solid adsorption from maximum entropy production.]. Phys. Chem. B, 103,7542-50. 16. Milewska-Duda, J., Duda, J.T., Jodlowski, G., and Kwiatkowski, M. (2000). A model for multilayer adsorption of small molecules in microporous materials. Langmuir, 16, 7294-303. 17. Asnin, L.D., Fedorov, A.A., and Cheekryshkin, Yu.S. (2000). Thermodynamic quantities of adsorption described by Freundlich isotherm. Russ. Chem. Bull., 49, 178-80. 18. Drzal, L.T., Mescher, J.A., and Hall, D.L. (1979). The surface composition and energetics of Type HM graphite fibers. Carbon, 17, 375-82. 19. Sircar, S. and Cao, D.V. (2002). Heat of adsorption. Chem. Eng. Technol., 25, 945-48. 20. Myers, A.L. (2002). Thermodynamics of adsorption in porous materials. AIChE ]., 48, 145-60. 21. Li, M., Zhou, L., Wu, Q., and Zhou, Y.P. (2002). Progress in predicting the equilibria of multi-component gas/solid adsorption. Prog. Chem., 14, 93-7.
References
73
22. Chiang, H.L., Huang, C.P., and Chiang, P.C. (2002). The adsorption of benzene and methylethylketone onto activated carbon: thermodynamic aspects. Chemosphere, 46, 143-52. 23. Linders, M.J.G., van den Broeke, L.J.P., Nijhuis T.A., et al. (2001). Modelling sorption and diffusion in activated carbon: a novel low pressure pulse-response technique. Carbon, 39, 2113-30. 24. Hill, T.L. (1960). An Introduction to Statistical Thermodynamics. Addison-Wesley. 25. Chen, S.G. and Yang, R.T. (1994). Theoretical basis for the potential theory adsorption isotherms. The Dubinin-Radushkevich and Dubinin-Astakhov equations. Langmuir, 10, 4244-9. 26. Valenzuela, D.P. and Myers, A.L. (1989). Adsorption Equilibrium Data Handbook. Prentice Hall. 27. Garbacz, ].K., Kowalczyk-Dembinska, H., and Dabrowski, A. (1996). Partially mobile adsorption of single gases on graphitized carbon black surfaces. Adsorpt. Sci. Technol., 14, 69-76. 28. Floess, ].K. and VanListhout, Y. (1992). Calculation of adsorption energies in carbon micropores. Carbon, 30, 967-73. 29. Murata, K. and Kaneko, K. (2001). The general equation of supercritical gas adsorption isotherm. J. Phys. Chem. B, 105, 8498-503. 30. Murata, K., Miyawaki, ]., and Kaneko, K. (2002). A simple determination method of the absolute adsorbed amount for high pressure gas adsorption. Carbon, 40, 425-8. 31. Kaneko, K. (1999). Control of supercritical gases with solid nanospaceenvironmental aspects. In Adsorption and its Applications in Industry and Enviromental Protection, Studies in Surface Science and Catalysis, Elsevier, Vol. 120B (A. Dabrowski, ed.). pp. 635-756. 32. Talu, O. and Meunier, F. (1996). Adsorption ofassociating molecules in micropores and application to water on carbon. A IChE J., 42, 809-19. 33. Rudzinski, W. and Panczyk, T. (2002). Remarks on the current state of adsorption kinetic theories for heterogeneous solid surfaces: a comparison of the ART and SRT approaches. Langmuir, 18, 439-49. 34. Pan, H.H., Ritter, ].A., and Balbuena, P.B. (1998). Isosteric heats of adsorption on carbon predicted by density functional theory. Ind. Eng. Chem. Res., 37, 1159-66. 35. Lastoskie, C.M., Quirke, N., and Gubbins, K.E. (1997). Structure of porous adsorbents: analysis using density functional theory and molecular simulation. In Equilibria and Dynamics of Gas Adsorption on Heterogeneous Solid Surfaces, Studies in Surface Science and Catalysis, Elsevier, Vol. 104 (W. Rudzinski, W.A. Steele and G. Zgrablich, eds). pp. 745-75. 36. Drago, R.S., Burns, D.S., and Lafrenz, T.]. (1996). A new adsorption model for analyzing gas-solid equilibria in porous materials. J. Phys. Chem., 100, 1718-24. 37. Pan, H., Ritter,].A., and Balbuena, P.B. (1999). Binary isosteric heats of adsorption predicted from density functional theory. Langmuir, 15, 4570-78. 38. Myers, A.L. (2003). Equation of state for adsorption of gases and their mixtures in porous materials. Adsorption, 9, 9-16. 39. Bhatia, S.K. (1994). Transport of adsorbates in microporous solids: arbitrary isotherm. Proc. R. Soc. Lond. A, 446, 15-37. 40. Carman, P.C. and Raal, F.A. (1951). Diffusion and flow of gases and vapours through micropores. 3. Surface diffusion coefficients and activation energies. Proc. R. Soc. Lond. A, 209, 38-58.
74
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
41. Shekhovtsova, L.G. and Fomkin, A.A. (1990). Discrete site model for methane adsorption on microporous adsorbents. Bull. Acad. Sci. USSR Div. Chem. Sci., 39, 867-70. 42. Wang, C.H. and Hwang, B.J. (2000). Characterizing the adsorption behaviors of various adsorbates on activated carbons via the multi-layer theory. J. Chin. Inst. Chem. Eng., 31, 333-8. 43. Cerofolini, G.F. and Rudzinski, W. (1997). Theoretical principles of single- and mixed-gas adsorption equilibria on heterogeneous solid surfaces. In Equilibria and Dynamics of Gas Adsorption on Heterogeneous Solid Suifaces, Studies in Suiface Science and Catalysis, Elsevier, Vol. 104 (W. Rudzinski, W.A. Steele and G. Zgrablich, eds). pp. 1-103. 44. Tovbin, Y.K. (1997). Application of lattice-gas models to describe mixed-gas adsorption equilibria on heterogeneous solid surfaces. In Equilibria and Dynamics of Gas Adsorption on Heterogeneous Solid Suifaces, Studies in Suiface Science and Catalysis, Elsevier, Vol. 104 (W. Rudzinski, W.A. Steele and G. Zgrablich, eds). pp. 105-52. 45. Tovbin, Y.K. (1997). Theory of adsorption-desorption kinetics non flat heterogeneous surfaces. In Equilibria and Dynamics of Gas Adsorption on Heterogeneous Solid Suifaces, Studies in Suiface Science and Catalysis, Elseveir, Vol. 104 (W. Rudzinski, W.A. Steele and G. Zgrablich, eds). pp. 201-84. 46. Rudzinski, W., Steele, W.A., and Zgrablich, G., eds. (1997). Equilibria and Dynamics of Gas Adsorption on Heterogeneous Solid Suifaces, Studies in Suiface Science and Catalysis, Elsevier, Vol. 104. 47. Myers, A.L. and Prausnitz, J .M. (1965). Thermodynamics of mixed-gas adsorption. AIChEJ., 11, 121-7. 48. Ladugie, P., Lavanchy, A., and Touzani, R. (1992). Study of a simple model for chemical adsorption by active carbon beds. Math. Eng. Ind., 3, 247-64. 49. Wintgens, D., Lavanchy, A., and Stoeckli, F. (1999). Static adsorption, by activated carbons, of vapour mixtures corresponding to immiscible liquids. Adsorpt. Sci. Technol., 17,761-9. 50. Ramirez-Pastor, A.J., Pereyra, V.D., and Riccardo, J.L. (1999). Statistical thermodynamics of linear adsorbates in low dimensions: application to adsorption on heterogeneous surfaces. Langmuir, 15, 5707-12. 51. Riccardo, J.L., Ramirez-Pastor, A.J., and Roma, F. (2002). Multilayer adsorption with multisite occupancy: an improved isotherm for surface characterization. Langmuir, 18, 2130-4. 52. Rzysko, W., Sokolowski, S., and Pizio, 0. (2002). Theory of adsorption in a polydisperse templated porous material: hard sphere systems. J. Chem. Phys., 116, 4286-92. 53. Letoquart, C., Rouquerol, F., and Rouquerol, J. (1973). Heats of adsorption.l. Heats of physisorption derived in terms of internal energy from experimental data. J. Chim. Phys., 70, 559-73. 54. Della Gatta, G. (1985). Direct determination of adsorption heats. Thermochim. Acta, 96,349-63. 55. Piper, J., Morrison, J.A., Peters, C., and Ozaki, Y. (1983). Heats and entropies of adsorption ofN 2 on grafoil at 79.3K.J. Chem. Soc. Faraday Trans. I, 79, 2863-74. 56. Groszek, A.J. (1997). Heats of adsorption and desorption of CO 2, CH 4 , S02' O 2 and N 2 on microporous carbons. Carbon, 35, 1399-405.
References
75
57. Groszek, AJ., Avraham, I., Danon, A., and Koresh,].E. (2002). Interaction of0 2 , N 2 and He at room temperature with carbon molecular sieves sensed by adsorption measurements. Colloid Surf. A, 208, 65-70. 58. Shen, D. and Bulow, M. (1998). Isosteric study of sorption thermodynamics of single gases and multi-component mixtures on microporous materials. Microporous Mesoporous Mater., 22, 237-49. 59. Salem, M.M.K., Braeuer, P., Szombathely, M.V., et al. (1998). Thermodynamics of high pressure adsorption of argon, nitrogen, and methane on microporous adsorbents. Langmuir, 14, 3376-89. 60. Rychlicki, G. and Terzyk. A.P. (1995). Energetics of methane adsorption on microporous activated carbons.]. Therm. Anal., 45,1183-7. 61. Rychlicki, G., Terzyk, A.P., and Szymanski, G.S. (1993). Thermodynamic properties of CCl4 adsorbed on microporous activated carbon. Polish]. Chern., 67, 1121-30. 62. Rychlicki, G., Terzyk, A.P., and Zawadzki, ]. (1994). The influence of carbon surface oxidation on thermodynamic properties of carbon-tetrachloride adsorbed at different temperatures. Polish]. Chern., 68, 557-63. 63. Rychlicki, G., Terzyk, A.P., and Zawadzki,]. (1993). Low-coverage adsorption of methanol, ethanol and carbon-tetrachloride on homo and heterogeneous surface differential heat and integral molar entropy. Polish]. Chern., 67, 2019-28. 64. Wendland, M., Salzmann, S., Heinbuch, D., and Fischer,]. (1989). Born-GreenYvon results for adsorption ofa simple fluid on plane walls. Mol. Phys., 67,161-72. 65. Wendland, M., Heinbuch, D., and Fischer,]. (1989). Adsorption of simple gas mixtures on a plane wall: Born-Green-Yvon for structure, adsorption isotherms and selectivity. Fluid Phase Equil., 48, 259-77. 66. Ding, L.P. and Bhatia, S.K. (2001). Application of heterogeneous vacancy solution theory to characterization of microporous solids. Carbon, 39, 2215-29. 67. Nguyen, C. and Do, D.D. (2001). Multicomponent supercritical adsorption in microporous activated carbon materials. Langmuir, 17, 1552-7. 68. Do, D.D. and Do, H.D. (2003). Adsorption of supercritical fluids in non-porous and porous carbons: analysis of adsorbed phase volume and density. Carbon, 41, 1777-91. 69. Dstinov, E.A. and Do, D.D. (2002). Mixed gas equilibrium adsorption on zeolites and energetic heterogenity of adsorption volume. Langmuir, 18, 3567-77. 70. Bakaev, V.A. and Steele, W.A. (1996). Adsorbed mixtures on a heterogeneous surface. The lattice gas model. Langmuir, 12, 6119-26. 71. Richter, E., Schutz, W., and Myers, A.L. (1989). Effect of adsorption equation on prediction of multicomponent adsorption equilibria by the ideal adsorbed solution theory. Chern. Eng. Sci., 44, 1609-16. 72. Dstinov, E.A., Vashchenko, L.A., and Polyakov, N.S. (2001). Statistical model of equilibrium adsorption of non-ideal mixtures on zeolites. Russ. Chern. Bull., 50, 220-7. 73. Dstinov, E.A., Do, D.D., Herbst, A., et al. (2002). Modeling of gas adsorption equilibrium over a wide range of pressure: a thermodynamic approach based on equation of state.]. Colloid Interface Sci., 250, 49-62. 74. Benedict, M., Webb, G.B., and Rubin, L.C. (1940). An empirical equation for thermodynamic properties of light hydrocarbons and their mixtures I. Methane, ethane, propane and n-butane.]. Chern. Phys., 8, 334-45.
Chapter 3 Energetics of Gas Adsorption by Carbons: Thermodynamic Quantities
75. PoIt, A., Platzer, B., and Maurer, G. (1992). The Bender equation of state for 14 polyatomic pure substances. Chern. Tech. (Leipzig), 44, 216-24. 76. Platzer, B. and Maurer, G. (1993). Application of a generalized Bender equation of state to the description of vapour-liquid equilibria in binary systems. Fluid Phase Equilibria, 84, 79-110. 77. Li, C.H., Wang, R.Z., and Lu, Y.Z. (2002). Investigation of a novel combined cycle of solar powered adsorption-ej ection refrigeration system. Renewable Energy, 26~ 611-22.
MONTE CARLO AND MOLECULAR DYNAMICS Mary
J.
Bojan and William A. Steele
Department of Chemistry, The Pennsylvania State University, University Park, PA, USA
Contents 4.1 Introduction 4.2 Overview of Computer Simulations 4.3 Conclusions References
4.1
77 78
97 98
INTRODUCTION
Computer simulations have become an indispensable tool in the efforts to forge links between the behavior of molecules on the atomic level and experimentally measurable properties. This has been especially true in the area of surface science in general and specifically in studies of carbon surfaces. Due to the abundance of experimental data for gases interacting with carbon surfaces, and because carbon surfaces are relatively simple, many of the pioneering simulations of surface phenomena involved studies of gases interacting with homogeneous (graphite) and heterogeneous carbon surfaces. Pioneering studies include the Monte Carlo (MC) simulations of Ar interacting with graphite by Rowley et al. [1-3] and the molecular dynamics (MD) simulations of N 2 on graphite by Talbot et al. [4-6]. As computers became more powerful, the expense of computing dropped, making it possible to simulate systems of ever-increasing size and complexity. At the same time, the complexity and cost of experimental work has made simulations even more indispensable since simulation results often suggest interesting directions for experimentation while they continue to aid in the interpretation of the results. The goal of MC and MD simulations is to calculate properties of a particular macroscopic thermodynamic system from the configurations (positions and Adsorption by Carbons ISBN: 978-0-08-044464-2
© 2008 Elsevier Ltd. All rights reserved.
77
Chapter 4 Monte Carlo and Molecular Dynamics
velocities) of many microscopic states. The techniques differ in their approach to generating the configurations. With MC a completely stochastic (probabilistic) algorithm is used, while in MD, the method is deterministic. In this context an ensemble is a collection of a large number of systems (made up of particles) constructed to be a replica on a macroscopic level of the thermodynamic system of interest. In MD, the configurations in a particular thermodynamic ensemble are generated by solving the equations of motion for the collection of classical particles (atoms or molecules) moving in a potential field created by other particles or by a surface. In MC, new configurations are generated by making