State of art report
Condensed silica fume in concrete
iTY" Thomas Telford L London Delivered by ICEVirtualLibrary.com ...
51 downloads
524 Views
5MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
State of art report
Condensed silica fume in concrete
iTY" Thomas Telford L London Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:43:26
FIP C O M M I S S I O N O N C O N C R E T E Chairman:
T. W. Kirkbride, UK
Members P. Acker, France B. K. Bardhan-Roy, UK T. W. Bremner, Canada R. D. Browne, UK R. Calzona, Italy H. Daneng, China
Technical Secretary: Z. George, India S. Helland, Norway C. Jaegermann, Israel E. Lakatos, Hungary F. D. Lydon, UK J. Muhl, FRG
H. E. Gram, Sweden
P. Poitevin, France C. Souwerbren, The Netherlands H. Steinegger, FRG J. Strasky, Czechoslovakia W. Wilk, Switzerland
W O R K I N G G R O U P O N C O N D E N S E D SILICA F U M E IN C O N C R E T E Chairman: Members P. Acker, France
S. Helland, Norway
H. E. Gram, Sweden
E. J. Sellevold, Norway
Published by T h o m a s Telford Ltd, T h o m a s Telford H o u s e , 1 H e r o n Q u a y , L o n d o n E 1 4 9 X F First published 1988 British Library Cataloguing in Publication Data Condensed silica fume in concrete. 1. Concrete. Aggregates. Silica fume in concrete I. Federation Internationale de la Precontrainte II. Series 666'.893 ISBN: 0 7277 1373 6 © Federation Internationale de la Precontrainte, 1988 All rights, including translation, reserved. Except for fair copying, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise, without the prior writ ten permission of the Publications Manager, Publications Division, Thomas Telford Ltd, Thomas Telford House, 1 Heron Quay, London E14 9 X F . Although the Federation Internationale de la Precontrainte does its best to ensure that any information it may give is accurate, no liability or responsibility of any kind (including liability for negligence) is accepted in this respect by the Federation, its members, its servants or agents. Typeset in Great Britain by MHL Typesetting Ltd, Coventry. Printed and bound in Great Britain by Bell and Bain Ltd, Glasgow.
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:43:26
FOREWORD
The importance of c o n d e n s e d silica fume in concrete w a s discussed at the meetings o f the FIP C o m m i s s i o n on Concrete in Calgary ( 1 9 8 4 ) and L o n d o n ( 1 9 8 5 ) , w h e n it was decided to prepare a 'state o f the art' report based o n a world r e v i e w o f published material. At the N e w D e l h i m e e t i n g , in February 1 9 8 6 , it w a s agreed to u s e an exist ing paper by Erik J. S e l l e v o l d and Terje N i l s o n as a framework for the publication. The paper had originally been prepared for presentation at the International W o r k s h o p on C o n d e n s e d Silica F u m e in Concrete, organized by Canmet in June 1 9 8 7 . The C o m m i s s i o n appointed the f o l l o w i n g working party to adapt the document into an FIP publication: Steinar Helland (Chairman) (Selmer Furuholmen a/s, N o r w a y ) , Paul A c k e r (Laboratoire Central des Ponts et C h a u s s e e s , France), Hans Erik Gram (Swedish Cement and Concrete Research Institute), and Erik J. Sellevold ( N o r w e g i a n Building Research Institute). All 25 m e m b e r s o f the C o m m i s s i o n w e r e contacted by mail for their v i e w s on the original paper. The working party started by basing their work o n the broad response, which included a considerable amount o f n e w information, research w o r k , and sug gestions for the presentation. The main modifications to and extensions o f the original paper are (a) a considerably increased number o f illustrations to i m p r o v e readability (b) t w o n e w chapters, o n e o n health aspects and o n e o n national standards, c o d e s and recommendations (c) a number o f case studies to illustrate the practical use and application o f c o n densed silica fume (d) a total rewriting o f the chapter on fire resistance, based on n e w information (e) a general updating, incorporating new information published in the past two years. The major part o f the updating will be found in chapters 5 , 7 , and 8. In chapter 5 ('Hardening concrete') in particular, the problems with l o w curing temperatures have been highlighted. In chapters 7 and 8 ('Durability' and 'Corrosion') a considerable amount o f n e w information on frost resistance, alkali silica reaction, and corrosion has been included The draft for this report was discussed and approved by the Commission at the meeting in Stavanger, N o r w a y , during the F I P / N B S y m p o s i u m on H i g h Strength Concrete in June 1987. T.W. KIRKBRIDE Chairman FIP Commission on
Concrete
ACKNOWLEDGEMENT The work done by M r S. Helland and his working party, and in particular by Mr Erik J. S e l l e v o l d , the main author, w h o s e w o r k on the report w a s financed by a grant from The N o r w e g i a n Concrete A s s o c i a t i o n , is gratefully a c k n o w l e d g e d .
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:43:26
CONTENTS
1.
Introduction
1
2.
Health aspects
2
3.
Pozzolanic and filler effects
2
3.1
Reactivity and reaction products
2
3.2
Pore structure
4
3.3
Conclusions
5
4.
5.
6.
7.
Fresh Concrete
6
4.1
Water demand
6
4.2
Concrete colour
7
4.3
Cohesiveness and stability
7
4.4
Plastic shrinkage
7
4.5
Setting time
7
4.6
Conclusions
8
Hardening concrete
8
5.1
Strength development and temperature
8
5.2
Heat development
10
5.3
Conclusions
11
Hardened concrete: mechanical properties
11
6.1
Compressive strength
11
6.2
Tensile and flexural strength
12
6.3
Brittleness and f'-modulus
13
6.4
Fly ash-CSF combinations
13
6.5
Bond properties
14
6.6
Shrinkage
15
6.7
Creep
16
6.8
Fire resistance
17
6.9
Abrasion-erosion resistance
17
6.10 Conclusions
17
Hardened concrete: durability
19
7.1
Permeability
19
7.2
Frost resistance
20 Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:43:26
7.3
8.
9.
Chemical resistance
25
Corrosion of reinforcement
28
8.1
pH-values: pozzolanic reaction and carbonation
28
8.2
Chlorides
30
8.3
Rate of corrosion
31
8.4
Conclusions
31
National standards, codes, and recommendations
10. References
32
32
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:43:26
1.
INTRODUCTION
Condensed silica fume (CSF) is a by-product of the smelting process used to produce silicon metal and ferrosilicon alloys. Other names for CSF that can be found in the literature include microsilica, ferrosilicon dust, arc furnace silica, silica flue dust, amorphous silica, and volatized silica. Even 'very fine-grained siliceous fly ash' has been used. Besides CSF, microsilica seems to be the most commonly accepted name. The CSFs used in the investigations reviewed here all come from the produc tion of silicon metal, or of ferrosilicon alloy containing more than 75% silicon. Such CSFs share the following main characteristics: Si02 content in the range 85—98%, mean particle size in the range 0.1 —0.2 /xm, spherical par ticle shape with a number of primary agglomerates, and amorphous particles. Details concerning production, filter ing, and variations in physical and chemical characteristics are available elsewhere, and will not be covered in this report. Since CSF is a by-product of the production of silicon metal and ferrosilicon alloys, the produced quantity will be sensitive to fluctuations in the metal trade. Table 1 shows the estimated production of CSF in 1984 for some countries. No data are available for East European coun tries and the USSR. A number of other countries, e.g. Brazil and China, have plants producing silicon metal or ferrosilicon alloys; however, not much of the fume is filtered, and so CSF from these countries is not currently available to the concrete industry. It is impossible to give exact figures regarding pricing of CSF. However, as a general rule the price is higher than that of cement. For special CSF-based proprietary products the price may be as high as 15 times the price of cement. CSF for use in concrete is either in a 'natural' state, densified, or in slurry form mixed with 50% water by weight. General field experience and laboratory tests have shown remarkably litde difference in the properties of hardened concrete containing CSF with different characteristics or in different forms. This is in sharp con trast to general experience with other fly ashes. The type and form of CSF may significantly influence fresh con crete properties, however, and in particular the rheological properties. It is not presendy possible to relate such dif ferences to specific physical or chemical characteristics of the CSF. No distincfion is made here between types and forms of CSF, but it is implied that the CSFs used in the investigations reviewed share the broad characteristics oudined above. This state of the art report covers the properties of con crete containing CSF in the fresh state, during harden ing, and in the hardened state, with the emphasis on durability properties. The review is based on published reports, of which approximately 400 are available, the majority of Norwegian origin and written in Norwegian. Most of the reports contain original laboratory data, some are review articles covering limited topics, and a few are concerned with laboratory investigations of concrete from old structures. The first tests on CSF in concrete were made in the early
1950s at the Norwegian Institute of Technology. At the same time, CSF was included among a large number of additive-cement combinations to produce concretes for long-term exposure to the acidic, high sulphate content water in a tunnel segment in the Oslo alum shale region. Results of these tests were reported after 20 years of expo sure, and a final report after 30 years is now in preparation. The first documented use of CSF in structural concrete took place at the Fiskaa smelting plant, Norway, in 1971 — the concrete has since been investigated on several occa sions. Following the start of large-scale filtering in the mid-1970s, the use of CSF, both in practice and in laboratory investigation, was begun in several places. In Gothenburg, Sweden, a readymix plant used CSF exten sively — including its use in concrete for a large wharf. In Denmark and Norway, readymix plants also began pro duction, and systematic laboratory work started at the Norwegian Institute of Technology. In Iceland, work was started with a view to reducing the effects of alkaliaggregate reactions using CSF, resulting in the produc tion of a cement containing 7.5% CSF. Outside Scandinavia, reports began to appear at the end of the 1970s, particularly from Quebec, Canada, where the practical use of CSF in readymix concrete was started in 1981. Since then, research work and practical use of CSF in concrete has begun in many countries, and is spreading rapidly. When considering the properties of CSF concrete, it is important to keep in mind that CSF is used in two dif ferent ways: (a) as a cement replacement, in order to obtain reduc tion in the cement content — usually for economic reasons (b) As as addition to improve concrete properties — both in the fresh and hardened state. For normal low-grade structural concrete the required strength can be obtained with an extremely low cement content when CSF is used. The debate in Scandinavia regarding CSF in concrete has mainly been focused on the durability aspects of this approach. The true promise of CSF, however, lies in the approach aiming to design concrete for specific production processes and to achieve better durability, or to enable the production of ultra-highstrength concrete on a routine basis.
Table 1.
Estimated
Country
Norway United States France Australia South Africa Japan West Germany Canada Sweden
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:45:51
production
and use of CSF in 1984
Total quantity produced: Mt
Utilized in cement and concrete products: Mt
140 100 60 60 43 25 25 23 14
40 2 0 20 0 0 0 11 5
C S F IN
CONCRETE
Norwegian standards allow up to 10% dosage of CSF by weight of cement in concrete. Normal usage is generally less than this. The term 'CSF concrete', as used in this review, refers to dosages of 10% CSF or less, unless other wise specified. High CSF dosages combined with superplasticizers and extremely low water/cement (W/C)
2.
H E A L T H ASPECTS
It is well known that crystalline silica such as quartz may cause silicosis. The particle size is here thought to be critical, and threshold limit values are frequently given for respirable dust. The question arises as to whether CSF represents the same health hazard. CSF consists mainly of sub-micron particles, and is hence suspected to be respirable (particles of less than 5 /xm are respirable). However, this may depend on the strength of secondary agglomerates which are present, as well as local physiochemical conditions in the respiratory system. A large number of X-ray diffraction analyses sug gest that CSF is an amorphous material and should therefore be less dangerous than a crystalline material. Heggestad et al.^ have, at least partly, confirmed the amorphous structure of CSF by thermosonimetry. Jahr^ has written several review articles on the pos
3.
ratios are the basis for a new generation of concretes with extreme properties which are not discussed here. Nomenclature. In this review the W/C ratio is calculated on the basis of cement content only. The W/C + S ratio is used for CSF mixes, where S is the quantity of CSF. CSF dosage is given as a percentage of weight of cement.
sible health hazard in the handling of CSF. His conclu sion, which, among other factors, is based on a study of 865 workers in Norwegian and Swedish ferrosilicon plants, and on results from animal experiments by Gl0mme^ is that the tendency to cause lung changes is far less for CSF than for respirable quartz. The Norwegian authorities have established a threshold limit value (TLV) of 2 - 0 mg/m^ for respirable CSF. For comparison, alpha-quartz has a TLV of 0 - 2 mg/m^, whereas for inert dust the value is 5 - 0 mg/m^. In the USA, CSF has only recently been listed by the American Conference of Governmental Industrial Hygienists."^ Amorphous silica is given a TLV of 5 - 0 mg/m^ and quartz 0-1 mg/m^. In many applications CSF is handled as an aqueous slurry, reducing the dust problem virtually to zero.
P O Z Z O L A N I C A N D FILLER EFFECTS
CSF is both a reactive pozzolana and an effective filler. Both properties combine to explain the effects CSF has on the properties of cement-based products. 3.1 Reactivity and reaction products A number of reports have appeared on the pozzolanic reactivity ^ f CSF. The subject has been reviewed by Hjorth^ and R e g o u r d . ^ The pozzolanic reactivity of CSF in cement pastes has been demonstrated by measuring the amount of calcium hydroxide at different times in pastes with varying dosages of CSF. TGA-DTA and X-ray diffraction methods have been used. The results generally show high pozzolanic reactivity,^"although some studies have found me dium or low'"^-'^ reactivity. Other studies on pozzolanic reactivity include those by Traetteberg'^ and Chatterji et al^^ Fig. 1 shows calcium hydroxide contents for various CSF dosages determined for mature cement paste specimens. Extrapolation of the curve indicates that roughly 24% CSF will eliminate the calcium hydroxide. This figure varies in the literature, and depends on both the method used to determine the calcium hydroxide con tent and the composition of the cement. CSF has been found to have an accelerating effect on the hydration of white cement,'^ roughly equal to the effect of a fine calcium carbonate filler. Fig. 2 shows the calcium hydroxide contents against time for a reference paste and one containing 12% CSF. Up to about 2 days the CSF mix has a higher calcium hydroxide content than the reference mix, but then the curves cross, presumably
because calcium hydroxide is consumed faster by the poz zolanic reaction than it is generated by the cement hydration. Wu and Y o u n g s t u d i e d the reaction of CSF with tricalcium silicate and with calcium hydroxide. They con cluded that CSF accelerates the hydration of C3S. Halse et ai^^ found that C3S hydration was 'enhanced' rather than accelerated by CSF, while Traetteberg^^ concluded that when lignosulphonates were used, it led to a marked reduction in cement hydration over long periods of time, both with and without CSF. Traetteberg's data'^ have been re-evaluated by Markestad.^^ Cheng-yi and Feldman^^ studied the hydration of cement paste and mortar with varying amounts of CSF and ground quartz sand. Both additives were found to accelerate the cement hydration during the first period, but after 14 days the calcium hydroxide content was eliminated in pastes containing 30% CSF. Mixes with 10% CSF led to a reduction in calcium hydroxide content of about 8% by weight of cement, implying a cement/CSF (C/S) ratio of about 0 • 7 — a very low value. For mortar mixes'^ a similar calculation yields a C/S value of about 1, which is a value in closer correspondence with other observations. The pozzolanic reactivity has been investigated in mixes containing only CSF and calcium hydroxide. Buck and Burkes^' detected well-crystallized calcium silicate hydrate (CSH I) after 7 days of curing at 38°C. Grutzeck et al}'^'^^ observed a silica-rich gel on the CSF surface shortly after having mixed CSF in calcium hydroxide solu-
2
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:45:51
POZZOLANIC
tion or in cement paste. The gel later combined with calcium hydroxide to form CSH. Wu and Y o u n g f o u n d that CSF behaved similarly to synthetic silica fume, differing only in rates of reaction according to their specific surface areas. According to Wu and Young, three kinds of CSH are formed in C3S-CSF systems: (a) that formed directly from C3S hydration; (b) that formed from the reaction between calcium hydrox ide and silica fume (with a slighdy lower C/S ratio); and (c) that formed by reaction between CSH and CSF (with a low C/S ratio and high degree of polymerization). The latter will only form in mixes containing more CSF than is needed to consume all the calcium hydroxide, i.e. it is not relevant for normal CSF concrete. Kurbus et al}^ mixed CSF with calcium hydroxide at a water/solid ratio of 0 • 50, and cured the pastes at 55°C and 90°C. The reaction was very temperature-dependent; at 90°C, 6 8 - 9 5 % of added lime had reacted after 2 • 5 h (depending on initial quantities), while at 55°C only 2 5 - 5 5 % had reacted in 2 - 5 h. At 20°C the reaction has been found'^ to be very slow for a mix with C/S = 1 - 0 , but after 110 days there was no sign of calcium hydrox ide on TGA curves. C/S ratios in the range 0-9—1 - 3 have been reported for cement-CSF mixes.^"•^ Regourd^ found the C/S ratio to decrease with increasing dosage of CSF — a natural result, since a high CSF dosage implies a higher fraction of CSH formed by the pozzolanic reaction in the total amount of CSH. A consequence of the low C/S ratio in the CSH is an increased capacity to incorporate foreign ions such as alkalis and aluminium. The high capacity to incorporate alkalis has been demonstrated by analysis of pore water squeezed out of hardened cement-CSF pastes^^'^^ (Fig. 3). These factors can partly explain the resistance of CSF concrete to aggressive chemicals and alkali-aggregate reac tion expansions.^
A • O
0
4
8
30
A N D FILLER
EFFECTS
40
T i m e in w a t e r : d
Fig. 2. Non-evaporable water content (WJ and calcium hydroxide content for white cement pastes (W/C = 0-60) with and without 12% CSF added (calculated from TGA results). (From ref 10).
The microstructure of the binder phase in CSF concrete appears 'very dense and amorphous'.^ Diamond^'^ reports that in contrast to normal concrete, in a properly formulated CSF concrete the CSH gel particles cannot be 'visualized as individual particles, but rather as a massive, dense structure'. Calcium hydroxide appears as small local crystals, rather than large masses which act as flaws in normal pastes. According to Diamond^^ and Regourd,^ the dense paste structure essentially extends to the true aggregate boundary in a dense CSF concrete. This eliminates the normal porous region of about 40—50 jicm, rich in calcium hydroxide, which surrounds aggregate grains in normal concrete. It should be emphasized that these observations apply to high-CSF dosage, highstrength mixes. In CSF concrete of normal strength with moderate CSF dosage the changes in microstructure are less marked.
C u r e d for 5 2 d a y s C u r e d for 11 O d a y s C u r e d for 6 5 d a y s
12
16
Silica content: % of c e m e n t weight
Fig. 1. Calcium hydroxide contents (measured by thermal gravimetric analysis (TGA)) of mature pastes made with white Portland cement. The W/C ratio is constant (0 • 60), and various amounts of CSF added. (From ref 10).
Fig. 3. Concentration of (a) and (b) OH— in pore solu tions expressed at times indicated from cement paste and CSFbearing pastes, all at water/binder ratio of 0-50. (From ref 25).
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:47:26
C S F IN 80
CONCRETE
r 6 1 2 inert filler (after l l O d a y s )
10
100
1000
Pore radius: n m
Fig. 4. Mercury intrusion in mature (65-day-old) white cement pastes (W/C = 0-60). The two last digits in the identification numbers indicate the percentage of CSF added. The inert filler is a fine precipitated calcium carbonate. The C-S paste is a calcium hydroxide-CSF mix with C/S = 1-0. (From ref 10).
The amount of water bound in the CSH formed by the pozzolanic reaction was found to be the same as that con tained in the calcium hydroxide,'^ i.e. C/H = 1-0. Recalculation of the raw data by Meland^^ showed no in crease in bound water per gram of cement in pastes con taining CSF, in agreement with Sellevold et al.^^ D i a m o n d , h o w e v e r , found the bound water per gram of cement to increase in CSF-cement pastes in comparison with pure cement pastes. Cheng-yi and Feldman'' found for W/C + 5 = 0-45 that the non-evaporable water per gram of cement was 0 - 2 0 7 for a reference paste, 0-172 with 10% CSF, and 0 - 2 1 9 with 30% CSF. For W/C + 5 = 0-25 the numbers were: 0-144 in the reference paste, 0-138 with 10% CSF, and 0-163 with 30% CSF. Thus, the picture is not entirely clear at present. It is consistently observed, however, by calorimetry and by the develop ment of non-evaporable water content or calcium hydrox ide over time, that CSF accelerates the hydration of cement. In a series of paste mixes to which different amounts of CSF were added at a constant W/C ratio, it was found that the total volume porosity to water in the pastes was independent of the CSF dosage. This implies that the chemical shrinkage caused by the pozzolanic reaction is greater than that of the cement hydration; it was estimated to be 12 cm^/100 g CSF, compared with a value of about 5 cm^/100 g cement. As a consequence of this, CSF concrete cured without access to water will exper ience a higher degree of self-desiccation, and consequently a lower internal relative water vapour pressure. This has been confirmed by measurements^^ where a mix with W/C = 0 - 4 0 and 10% CSF had an internal relative humidity (RH) value of 70% after 6 months of sealed curing. 3.2 Pore structure Pore structure plays a major role in determining the permeability, and thereby durability properties, of cementbased products. Recent work by Mehta^^ and Manmohan and Mehta^' have demonstrated a relationship between pore structure, permeability, and durability for blended cements. The pore structure of cement-CSF pastes has
been studied by Sellevold et al. using water adsorption, mercury penetration, and freeze calorimetry. They con cluded that increasing CSF dosage at constant W/C ratio did not change the total porosity as measured by water adsorption, but led to a refinement of the pore structure, i.e. less of the pore space consisted of capillary pores, where water can freeze and mercury penetrate. For pastes where part of the cement was replaced by CSF on a 3:1 basis and the water content kept constant, the capillary porosity was unchanged, indicating that CSF was roughly three times as 'efficient' as cement in reducing capillary porosity. By comparing pore structure data for CSF pastes with pastes where an almost inert filler of equal fineness was used, it was concluded that most of the pore refine ment effect was caused by the pozzolanic activity of the CSF. Fig. 4 shows mercury penetration results for mature pastes. Traetteberg^^ measured mercury penetration in mortars with varying CSF contents, and concluded that CSF was very efficient in subdividing the pore space. Mehta and Gjorv^^ measured mercury penetration in cement pastes with W/C = 0 - 7 4 and in equivalent pastes where 30% of the cement volume was replaced by fly ash, CSF, or an equal volume of the two. The results showed that at 90 days the total penetration was equal for the con trol and the CSF pastes. For the control paste, however, more than 50% of the available pore space was large pores ( > 0 - l /xg), while the CSF paste contained only about 10% large pores. Fly ash also had a pore-refining effect, but far less than CSF. Cheng-yi and Feldman^^'^"*'^^ studied the porosity of pastes and equivalent mortars with 0, 10, and 30% CSF as replacement for cement in a 1:1 ratio. For pastes, the results of mercury intrusion agreed with others: increased CSF dosage leads to a finer pore structure. After the first
0' 10'
lO'*
10^ 10' Pore diameter: n m
10^
Fig. 5. Mercury intrusion and reintrusion into 90-day-old cement pastes and mortars (W/C + S = 0-45). Plotted from data in refs 34 and 35.
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:47:26
POZZOLANIC
intrusion, the mercury was evaporated and a second in trusion performed. Hysteresis between the two curves in dicates discontinuity in the pore structure: the pore struc ture may be 'broken up" during the first intrusion. This hysteresis increased markedly with increasing CSF con tent. For mortars''' the effect is even more pronounced (Fig. 5J. Cheng-yi and Feldman related the effect to the reaction of CSF with calcium hydroxide, particularly the high calcium hydroxide concentrations around sand grains in mortars. They also demonstrated that CSF mixes have higher ratios between mortar and paste compressive strength than mixes without CSF, and attribute this to an improved bt)nd between sand and binder matrix. The present authors have also observed that CSF mortars have a higher fraction of pores with diameters above 100 nm relative to equivalent pastes than is the case for mortars and pastes without CSF. That a pozzolanic reaction between CSF and calcium hydroxide at the interface improves the bond appears to be natural, but we find it less natural that such an improved bond phase has a coarser and more unstable pore struc ture than is found in mixes with no CSF. The Technological Institute in Denmark has developed a method to estimate the capillary porosity of concrete by microscopic examination of thin sections impregnated with fluorescent epoxy. Applying this method to CSF con cretes from field and lab
3.3 Conclusions CSF is considered to be a very reactive pozzolana. When used in cement systems it prtxluces a CSH gel with a lower C/S ratio than the cement hydration; consequently, it has a high capacity to incorporate foreign ions, particularly alkalis. The nature of the hydration products of CSF, and their infiuence on cement hydration, are not entirely understtwd at present. CSF also has a definite filler effect that is believed to distribute the hydration products in a more homogenous fashion in the available space. These two factors have the combined effect of refining the pore structure when CSF is added to cement-based mixes. The refinement of the pore structure leads to reduced permeability, and is considered to be the main factor responsible for the influence CSF has on the mechanical and durability properties of concrete.
Case study A. Distillery in Bergen, Norway In 1982 a production plant for alcoholic beverages was constructed in Bergen, on the western coast of Norway. The total volume of concrete was 17 000 m \ The job requirements were (a) 35 MPa characteristic strength at 28 days (b) high early strength (c) the walls had to be poured without the help of tremies (d) a smooth surface on the walls without voids or honeycombing (e) very stable fresh concrete, without any tendency to segregation or bleeding — necessary because (he walls were up to 9 m in height, had two layers
A N D FILLHR
BFFECTS
of reinforcing bars, and had a thickness of only 30 cm (f) low cost. These requirements were met by the following mix: 110 kg ordinary Portland cement 110 kg rapid-hardening cement 22 kg CSF 2.5 kg lignosulphonate (40%) W/C + S = 0 - 7 0 slump = 15—20 cm. As the local price of the CSF was only 140% that of the cement, this recipe gave an inexpensive concrete, since the efficiency factor of the CSF with respect to strength was about 4. By using aggregate with continuous grading to avoid segregation and honeycombing and to introduce a surplus of fines by adding the CSF and the finely ground rapidhardening cement to avoid bleeding, the walls were poured without any disfiguring pores. The site was inspected after 18 months of exposure. The depth of carbonation was then in the range of 0 - 5 — 2 - 0 mm, indicating that even at these high W/C + S ratios it is possible to achieve a dense concrete. Source: ref. 37. Case study B. Slipformed fertilizer storage silos at Heroya, Norway Two silos for storing fertilizers were slipformed at Heroya in Norway in 1982. Each silo had a height of 28 m, a diameter of 27 m, and a wall thickness of 27 cm. The walls were heavily reinforced both with ordinary bars and post-tensioned cables. Due to the highly corrosive properties of calcium nitrate, a dense concrete of extremely good chemical resistance had to be used. Since the paste is the weakest part of the material, the client wanted to minimize the cement and water content. At the same time, the heavy reinforcement and the sensitive slipforming operation required a concrete of good workability. To meet these requirements the following mix was used: 260 40 112 19
kg kg 1 1
sulphate-resistant cement CSF total water content combination of melainine and lignosulphonate-based plasticizers 0.15 I air-entering agent slump value = 17 cm air content = 8% mean strength at 28 days = 65 MPa.
By introducing such high amounts of plasticizers, the yield shear strength of the fresh concrete was reduced without altering the viscosity to the same degree as usually occurs when water is added to the mix to increase the slump value. To get a normal workability and a nor mal viscosity in the concrete, entrained air was used. The slipforming operation was executed without any major problems, in spite of the high content of CSF and plasticizers. After 5 years of service there are no signs of deterioration in the structure, and concrete based on the above principles has since become the normal require ment in the Norwegian fertilizer industry. Source: ref. 38.
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:47:26
C S F IN
CONCRETE
4.
FRESH C O N C R E T E
The influence of CSF on workability properties of con crete depends somewhat on the type of concrete. The addi tion of CSF to a given mix will generally lead to a lower slump, and a more cohesive mix. The 'static' and 'dynamic' behaviour of CSF concrete do not relate in the same way as for normal concrete, i.e. the slump measure does not predict the response to vibration in the usual way. For practical purposes it is generally recommended that the slump should be 20—30 mm higher for a CSF con crete to obtain the same workability as that of normal con crete. This should be kept in mind when considering water demand tests based on equal slumps. 4.1 Water demand Systematic measurements of water demand to reach a given slump for different CSF dosages have been made by Johansen^^ and Loland and Hustad.'*^ For very lean concrete (cement 100 kg/m^), it was found that the water demand decreased as CSF was added,"^^ as has been reported by Aitcin et al.^^ for lean mixes with less than 10% CSF. In concrete with a cement content of more than 250 kg/m^, the water demand will increase when adding CSF or even replacing cement on a 1:1 basis, when no water-reducing agents are used. Sellevold and Radjy"^^ analysed the data from references 39 and 40, together with data by Dagestad"^^ on water demand in concrete mixed RP38 RP38FA SP30
10
20
30
50
40
C S F content: kg/m^ (a) L D w i t h all 3 cements
F r o m ref . 5
with three types of cement, different dosages of waterreducing agents, and 0%, 8% and 16% CSF by weight of cement. It was found that the increased water demand per kilo of CSF added was of the order of 1 litre. However, water-reducing agents had a much stronger effect in CSF concrete. A lignosulphonate dosage (dry) of 0 • 2—0 • 4 % by weight of cement was sufficient to equal the water demand of the control mixes and those with 10% CSF (Fig. 6). It was also found that the water-reducing agent should be dosed by weight of CSF to keep the water demand near that of the control concrete. A dosage of 4% dry lignosulphonate by weight of CSF is a good starting point for CSF dosages between 5% and 15% by weight of cement. The investigations cited above, as well as general Norwegian experience, show that lignosulphonatebased water-reducing agents are as efficient as, or more efficient than superplasticizers in reducing water demand in CSF concrete. Markestad'*^ also analysed Dagestad's raw data."^^ He concluded that the water reduction achieved by waterreducing admixtures in CSF concrete was directly pro portional to the admixture concentration in the mixing water, rather than to the amount of CSF. He proposed a mix design method based on an efficiency factor of a given admixture. Maage and Dahl"^^ measured water demand for con cretes with different types of cement and various CSF dosages. At a constant dosage of lignosulphonate, the water demand increased with increasing CSF dosage for all types of cement. Carette and Malhotra"^^ and Petersson et al. ^'^ have also measured water demand in CSF concrete. Their conclu sions are in general agreement with the discussion above, but not numerically identical. As already mentioned, the type of CSF and its specific interaction with all the con stituent materials in a given concrete determine the water demand of concrete. Systematic laboratory tests based on fixed mix proportions in a control concrete have little value. When CSF is used, the chosen mix design may not be the optimum. For instance, the great fineness of CSF usually permits a coarser grading curve for the other materials. Adsorption of water-reducing admixtures on CSF has been measured by Meland^^ and by Buil et al.
-1
2?
0)
li •13 + 1
/ / / / /
0
'
LD: Condensed naphthalene sulphonate P:Lignosulphonate
10
20
W a t e r - r e d u c i n g a d m i x t u r e : % of C S F w e i g h t
Case study C . Housing project in Tromse, Norway Ehiring the period 1980—82, a housing compound con sisting of 1100 flats was constructed in Tromso, nor thern Norway. The project consumed 27 000 m^ of concrete with a requited characteristic strength of 25 MPa. The site demanded a concrete that enabled them to strip the scaffolding of the 6 m span decks after 16 h, even during the winter period. For the prefabricated produc tion and the construction of columns, the demand varied between 16 h and 6 h (1 or 2 cycles/d). This resulted in the following mix:
(b)
Fig. 6. (a) The influence of CSF content on the water demand of mixes not containing water-reducing admixtures, (b) The in fluence of water-reducing admixture dosage on the water demand of CSF concrete. SP30 = ordinary Portland cement; RP38 = rapid-hardening Portland cement; RP38FA = RP38 with 20% fly ash. (From ref 42).
240 kg ordinary Portland cement 24 kg CSF 4 kg plasticizers W/C + 5 = 0 . 7 0 - 0 - 7 5 slump = 1 8 - 2 0 cm mean strength at 28 days = 32 MPa.
6
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:38:02
FRESH
Due to the cheap CSF, which was 75% of the price of cement while the efficiency factor for strength was between 3 and 4, the site obtained an inexpensive mix of high workability. To ensure high early strength, the main bulk of the concrete was heated by steam injection up to 40 —60°C. During the winter period, infra-red heating of the formwork was also adopted. At these high temperatures the pozzolanic reaction is strongly accelerated, resulting in a relative strength gain of the CSF concrete at an early age, equivalent to ordinary concrete at the same temperature. The elevated temperature of the mix resulted in a reduced retardation from the lignosulphonate. On the other hand, the risk of crack ing due to plastic shrinkage increased when this CSF concrete was used in deck slabs. As the concrete was batched on the site, giving short transportation time, there were no problems due to slump loss or quick setting. The depth of carbonation, measured after 3! years, was about 6 mm, indicating that the carbonation front will reach the reinforcement after a period of roughly 50 years. To improve the quality of the concrete cover, the Norwegian c«xie of practice was revised in 1986; the maximum W/C + S for an outdoor construction is now 0-60. Source: ref. 50.
4.2 Concrete colour CSF may vary considerably in colour, which in turn affects the colour of the fresh concrete. Generally, CSF darkens the concrete. Bellander'" determined the "degree of darkness', according to Swedish standards, of concretes made with a light and a dark CSF. The dark CSF pro duced a darker fresh concrete, but after 20 days of storage in laboratory air there was no measurable difference be tween the two CSF concretes and the control concrete. The initial difference is presumably eliminated by dryingcarbonation effects. 4.3 Cohesiveness and stability Increased cohesiveness is the most obvious difference between CSF concrete and normal concrete. One conse quence of this cohesiveness is reflected in dynamic workability tests such as 'Thaulow strokes' (a Norwegian test somewhat analogous to the flow table test) or a modified Vebe test where the time to obtain a slump of 250 mm on a Vebe table is recorded. Both Johansen"* and Maage and Dahl*' found that at equal slumps, con crete containing CSF required more energy input for a given flow (Fig. 7). In practice, this problem is overcome by using a higher slump for CSF concrete. The cohesiveness imparts stability to the mix, an effect of major imp
CONCRETE
rare, but in two such studies Johansen''' and Maage and Dahl'*'' found that CSF used as a cement replacement led to a large reduction in bleeding, even for mixes with no water-reducing agents and therefore increased water con tent. Bellander'' stated in a review article on practical experience in Sweden that CSF also improves the ability of concrete to be handled and transported without separa tion. In a comprehensive study of 25 concrete mixes,'**' the bleeding and separation tendencies were evaluated qualitatively. Again, the conclusion was that CSF greatly improved the stability of concrete, particularly in very lean mixes. 4.4 Plastic shrinkage Plastic shrinkage and cracking takes place when evaporation from a fresh concrete surface exceeds the rate of bleeding water from the concrete. The fact that bleeding can be practically eliminated in CSF concrete makes it vulnerable in this respect. Johansen^' made a systematic investigation of plastic shrinkage cracking, and concluded that the critical time is just around the time when the set ting of the concrete takes place. The effect of CSF was, as expected from practical experience, that the risk of cracking increased. The problem can be eliminated by applying a proper curing procedure to the concrete sur face. Experience shows that under conditions of fast evaporation (wind and sun), curing measures must be taken immediately after placing the concrete. 4.5 Setting time The setting characteristics of cements are measured on pastes with a standardized consistency, using a Vicat apparatus. Maage**^ measured setting times for cement pastes with 0 and 10% CSF. All pastes, except one con trol paste, were mixed with 0.4% dry lignosulphonate by weight of cement. The setting of each of the pastes with admixture was retarded relative to the control. However, the retardation effect of the lignosulphonate was much less in pastes containing CSF than in pure cement paste. This observation agrees with the common practical experience that CSF concrete can tolerate a higher lignosulphonate dosage than normal concrete without suffering unaccept able retardation. Isothermal calorimetry data on pastes with and without CSF and lignosulphonate are in agreement with this.*'
lOOt-
Q.
E
5
10 C S F c o n t e n t : % b y w e i g h t of c e m e n t
Fig. 7. Time of vibration to reach a slump value of 250 mm for concrete containing different types of cement and varying CSF content. {From ref. 45).
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:38:02
C S F IN
CONCRETE
Penetration measurement on mortar is a more practically relevant test than the Vicat test. Bellander^' found a tend ency to increased setting time with increasing CSF dosage for concretes of equal strength, while Petersson et al^^ found little influence of CSF dosage for concretes with constant cement plus CSF content. Neither of these in vestigations used water-reducing admixtures; they are therefore of little practical relevance, because almost all CSF concrete is produced with the aid of water-reducing agents. Much practical experience has been gained concerning the setting time of CSF concretes with admixtures, but no comprehensive laboratory investigations have been reported. As for CSF and control concretes of equal 28-day strength, the former will normally have a longer setting time and slower early strength development, since it has a higher W/C + S ratio. A commonly used method for practical control of the setting time and early strength development is to use a combination of lignosulphonate
5.
and superplasticizer, where the relative amounts are varied to achieve the desired result. 4.6 Conclusions The major effects of CSF on the workability of con crete are to increase the cohesiveness and stability of a mix. Consequently, bleeding and separation are greatly reduced. The increased cohesiveness means that a higher slump is needed to match the workability of a control concrete. CSF addition leads to increased water demand. Waterreducing agents or superplasticizers are necessary to realize the potential of CSF in practice. The use of such admix tures has a greater effect in CSF concrete than in control concrete. The lack of bleeding in CSF concrete makes it more vulnerable to plastic shrinkage cracking than ordinary con crete. Protective measures must be taken under conditions creating high evaporation rates from the concrete surface.
HARDENING CONCRETE
5.1 Strength development and temperature Many investigations have measured strength develop ment with respect to time for CSF concretes. Most of these are confined to compressive strength and curing at 2 0 ° C . In some reports heat curing has been used. In this section only compressive strength development and temperature effects are considered for moist-cured specimens. For moist curing near 20°C it has been established that the strength development for a CSF concrete is slower than for a control concrete of equal 28-day compressive strength. This difference increases with increasing CSF dosage and decreasing temperature. It does not mean, however, that high early strengths cannot be obtained using CSF. One-day strength as high as 100 MPa has been reported by Biirge^^ for a concrete quality that can hardly be made with Portland cement alone (Fig. 8). High early strength in CSF concrete without the use of heat curing or accelerators also means very high later strength. The percentage of the 28-day strength that is obtained the first day increases with a decreasing W/C ratio for normal con crete. The percentage for CSF concrete appears to be the same as for normal concrete with the same W/C ratio, but the CSF concrete characterized this way has a higher absolute strength at 28 days.
Fig. 8. Compressive strength of concrete with gap-graded quartz aggregate (binder 450 kg/m\ cement 351 kg/m^). (From ref. 56).
CSF contributes most to strength development at 20°C in the time range from 3 to 28 days. Empirical relation ships between short-time and 28-day strengths are therefore not applicable to CSF concrete, which has a much higher strength gain than normal concrete in this time range. A comprehensive investigation^^ of strength development of 84 mixes up to 90 days showed a general trend for the CSF mixes to lag most behind the control mixes between 1 and 7 days, at equal 28-day strengths. At very short curing times the filler effect of the very fine CSF particles presumably acts as an accelerator, and by adsorpdon of water-reducing agents offsets some of their retarding effects. For the period between 28 and 90 days the general tendency was for the CSF mixes to have a higher strength gain than the controls. This particularly applied to lean mixes. Johansen^^'^'^ measured strength up to 3 years and concluded that there was little effect of CSF on either the strength gain between 28 days and 1 year or between 1 and 3 years for water-stored specimens. Aitcin^^ reported on Canadian experience that, for concretes of equal 28-day strength, the CSF mixes will have somewhat lower long-term strengths. This is in contrast to the results cited above. Long-term water storage is of limited prac tical interest, however; results from dry storage will be discussed below. The problems of obtaining early strength for CSF con crete are partly overcome in practical use (readymix, concrete element production) by using warm concrete and applying insulation, or by heat curing. A number of systematic investigations on the effects of heat have also been carried out. The first, on concretes mixed at 25°C and 5 0 ° C , did not show higher 1-day and 3-day strengths relative to 28-day strengths for CSF concrete compared with c o n t r o l s . T h i s is possibly because the high initial temperature was only maintained over a short period of time. Sandvik^^ measured strength development of 25 MPa concrete with 0 and 10% CSF at 5, 20, and 40°C. The concretes were mixed at 20°C and placed immediately
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:38:02
HARDENING
at their curing temperatures; after 1 day they were trans ferred to water baths. The results showed the normal pattern at 20°C, and very slow development for CSF con crete at 5°C — even at 28 days it had only reached about two-thirds of the strength of the control concrete. However, at 40°C the 1-day strength of the CSF concrete was more than 50% higher than the strength of the con trol. This dramatic increase is not supported by later work. Sandvik^^ later reported strength development for 40 MPa concrete with 0 and 10% CSF, cured isothermally at 20°C and 4 0 ° C . Ordinary Portland cement (OPC) and rapid-hardening Portland cement (RHPC) gave basically the same results: at 20°C the CSF mixes lag behind up to 28 days, and at 4 0 ° C the CSF mixes are accelerated more than the control mixes, shifting the crossing point in the strength development to about 3 days. Skurdal^^ reported strength development at 20, 30, and 50°C for concrete with 0, 10, and 20% CSF, all with 28-day strengths at 20°C of about 30 and 65 MPa. Cur ing at 50°C reduced the 28-day strength of all the concretes by about 20% relative to that obtained by curing at 20°C. Roughly the same reduction was found for concretes cured for 1 day at 50°C and then for 27 days at 2 0 ° C . 30°C curing gave a strength reduction for the control concrete of about 10%, but no reduction was found for CSF mixes. The effect of temperature on early strength was similar to that found by Sandvik^^ 30°C was not sufficient to produce 1-day strengths in CSF concretes equal to that in the control, but at 50°C CSF mixes were about equal to controls of 30 MPa concrete, while they surpassed con trol strengths of 65 MPa mixes. Maage and Hammer^^ reported on a comprehensive investigation involving four cement types (OPC, 10 and 25% pulverised fuel ash (PFA) blends, and a 15% slag blend) with 0, 5, and 10% CSF. The mixes were made at 5 ° C , 20°C and 35°C and maintained at these temperatures in water for 28 days, after which they were stored at 20°C in water for up to one year. The com pressive strength was measured from 16 h. Mixes in three strength classes were made: 15, 25, and 45 MPa. For 20°C curing, the CSF had the same influence on strength development as described above, regardless of the cement type (Fig. 9). Figs 10 and 11 show relative strength development at 5°C with and without 10% CSF for the four cement types, and similar data for 35°C cur ing are shown in Figs 12 and 13. At 5°C the blended cements lag behind OPC up to 28 days as expected; with 10% CSF the lag increases — it looks as if the pozzolanic reactions have not contributed much to strength in the 28-day period. At 35°C the CSF mix is more strongly accelerated (in comparison with cur ing at 20°C) than the reference mixes, particularly be tween the first and seventh day. Helland^'^ reported the general experience in Norway of adding CSF to Portland and fly-ash-blended Portland cement. Assuming that the relative rate of reaction at dif ferent temperatures, 6, corresponds to the Arrhenius equation H = exp
E(d) R
1 293
1 213 + 6
16h
Id
Fig. 9. Compressive strength development of concrete watercured at 20° C, with various CSF dosages. Each curve represents mean values for 4 cement types; 100% represents 28-day strength for each mix type. (From ref 63).
28-day strength at20°C | ) 100
50 P30 MP3010%flyash MP30 25%flyash MP3015%slag
16h 1d
14d
3d
3
28d
Age
6 months
12
Fig. 10. Development of compressive strength in reference con crete cured in water at 5°Cfor 28 days, then at 20°C. 100%) represents 28-day strength at 20°Cfor each cement t\pe. (From ref 63).
5°C ^ 20^0 28-day strength at 20°C
I
100"
O P30 • MP3010%flyash • MP30 25%flyash AMP3015%slag
16h 1d
3d
7d
14d Age
where R = 8.3 J/moPC, the activation energy E(d) increases by about 10% as 10% of CSF is added. This is demonstrated in Fig. 14 for a Norwegian rapidhardening cement.
CONCRETE
28d
6
12 months
Fig. 11. Development of compressive strength in concrete con taining 10%) CSF and water-cured at 5 °C for 28 days, then at 20°C. 100%c represents 28-day strength at 20°Cfor each cement type (with 10%c CSF). (From ref 63). 9
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:52:00
C S F IN
CONCRETE
300r
-RHPCplus10%CSF -RHPC
5 4
°
I
3
1
2
200
"5 1 tr
0-5 ^
100 2 8 - d a y strength at 2 0 ' C
I •
P30 MP3010%flyash M P 3 0 2 5 % l l y ash MP3015%slag
—I 16h I d
3d
7d
14d
I
1 12 months
28d
Fi((. 12. Development of compressive strength in reference con crete cured in water at 35 °Cfor 28 days, then at 20°C. 100% represents 28-da\ strength at 20°C for each cement type. (From ref 63). 300r
200
100
O ' I J J L. 16h i d
3d
_1_ 7d
14d
28d
• A •
P30 MP3010%flyash MP3025%tlyash MP3015%slag
6
-I 12 m o n t h s
A9»
Fig. 13. Development of compressive strength in concrete con taining 10% CSF and water-cured at 35°C for 28 da\s. then at 20°C. lOOJc represents 28 day-strength at 20°C for each cement type. (From ref 63).
5.2 Heat development Isothermal heat development data already cited"*^ showed that total heat development after 2 days was reduced when cement was replaced 1:1 by CSF. However, after such a short period of time little of the pozzolanic reaction has taken place. Adiabatic measurements on concrete'"^ show that the heat development by the poz zolanic reaction is I - 2 times as high per gram of CSF as the cement hydration per gram of cement, assuming that the two chemical reactions do not influence each other in a CSF concrete. For high-strength concrete/water-binder ratio below 0 - 4 0 and 15% CSF, Helland** found negligible contribu tion of CSF to heat development, in spite of a large con-
025 10
20
30 40 Temperalure:°C
50
60
Fig. 14. Rate of reaction at different temperatures for a Norwegian RHPC with and without 10% CSF. The relation ship is valid up to about 407c of standard cube 28-day strength. (From ref 67)
tribution to strength. This observation is important in the practical use of CSF in high-strength concrete. Since the strength potential is several times higher per unit weight of CSF than for cement, and because the heat development under normal field conditions is slower than under adiabatic conditions, it is possible to design a lowheat concrete of a specified strength using CSF. This was shown by Rasmussen.*^ Skurdal,*"^ and Lessard et al.*^ 5.3 Conclusions CSF concrete is more sensitive to curing temperaure than OPC concrete. The main contribution of CSF to con crete strength development at 20 °C takes place from about 3 to 28 days after mixing. For a CSF and a control con crete of equal 28-day strength, the strength of the CSF concrete will be lower over the entire time period with 20°C curing. Curing at elevated temperatures has a greater accelerat ing effect on CSF concrete than on control concrete. Evidence indicates that a curing temperature of roughly 50°C is necessary for CSF concrete to equal the 1-day strength of an equivalent control mix. Curing at temperatures below 20°C retards strength development more for CSF concrete than for control con crete. In practical work, the early strength problem may be overcome by using warm concrete and insulation, or by applying heat. CSF makes it possible to design low-heat concrete over a wide range of strength levels.
Case study D. Submerged concrete bridge at Karmoy, Norway To transport gas from the oil fields of the North Sea to the exposed western coast of Norway, a double pipeline was constructed in 1982—83. At the shore approach the pipelines were placed inside a concrete tun nel acting as an underwater bridge over the rocky sea bed. The tunnel has a length of 590 m and was cast as five separate prefabricated elements with lengths vary ing from 90 to 150 m and with a displacement of up to 7000 t. To reduce the environmental loadings from the waves, it was essential to reduce the duRgOMlMMi.i)/itie s U i ^
10 Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:52:00
MECHANICAL
ture. This resulted in the use of high-strength concrete and a dense reinforcement (330 k g / m \ including 80 kg/m^ post-tensioned cables). To ensure proper con creting, a highly workable mix was needed, the require ment being at least 24 cm slump. Thus the mix for the main elements was: 400 kg
SP-30-4A Portland cement (a high-strength cement) 32 kg CSF 4 kg naphthalene-based plasticizers 3 kg lignosulphonate-based plasticizers W/C + S = 0.38 slump = 2 4 - 2 6 cm
6.
PROPERTIES
mean strength at 28 days = 85 MPa mean strength at 90 days = 95 MPa. The Norwegian code of design now allows the use of C-105. In 1982, however, C-65 was the highest tabulated grade. For this reason the elements were designed in C-65. The requirement was met with a large margin, and the in situ strength today has passed 100 MPa. The CSF was used not only as an addition to ensure the specified compressive strength, but also to obtain the high specified slump value without getting problems with bleeding and segregation. The whole project, in cluding design and construction, was completed in 9 months. . Source: ref. 69.
H A R D E N E D C O N C R E T E : M E C H A N I C A L PROPERTIES
6.1 Compressive .strength The contribution of CSF to concrete strength can be expres.sed in at least two different ways. Loland and Hustad* ™ introduced the term 'efficiency factor' (K), defined by the equation
Sorensen'''' found K factors in the range from 2 to 5. increasing for richer mixes and decreasing with higher CSF contents. Strength results inconsistent with the general trend 100
ch
w C
+
KS)^
90
where the subscripts R and S indicate reference and CSF concrete respectively. The equation is based on the general assumption that, forgiven materials, each property of con crete is a function of the W/C ratio. By applying a K fac tor to the amount of CSF, it is converted into an equivalent amount of cement. The efficiency factor may be calculated for any property of concrete. It should be pointed out that A" is a 'marginal' quantity and therefore sensitive to the accuracy of the raw data. This was made clear by Mtxleer,^' who studied batch-to-batch variation for con crete with and without CSF, as well as by Maage and Hammer."' Another factor, the 'cement replacement factor' {K^ is defined as the difference in cement contents needed to produce equal strengths in a reference and a CSF con crete with the same slump, divided by the CSF content. takes into account different water demands in the two types of concrete; its numerical value therefore depends on the type and dosage of water-reducing agent used. This is why it is considered less 'fundamental' than the effi ciency factor, K. The relationship between the two fac tors was discussed by Sellevold and Radjy.""' Examples of based on experience with readymix concrete are given by Skrastins and Zoldners.^' Fig. 15 shows that the general shape of the strength against W/C ratio curve is maintained, but shifted to a higher level, when CSF is added to concrete. The majority of the papers published on CSF concrete contain data on compressive strength, but only a few con tain sufficient data to allow calculation of K factors. Laland^' analysed data from Johansen^^ and Loland and Hustad'" and found that A" = 2 for CSF mixes with 300 kg/m' or higher cement content, while for leaner mixes K = }>. Sellevold and Radjy"^ reported to be between 2 and 4, the highest values being for rich mixes, in contrast to Loland's findings." In line with general practical experience they also found K to be higher for 8% CSF dosage than for 16% by weight of cement.
\
80
\ 8%CSF-\
•
X
70' 6 % CSF
I
60
\
I"
\
\
\
20
\ R e f e r e n c e concrete
10
0 30
\
0 40
0 50
0 60
070
0 80
^
0 90
10
M
W/C
Fig. 15. 28-d(iy compressive strength versus W/C ratio for con crete with different CSF contents. Concretes with and without water-reducing admixtures are not differentiated. Cement is standard Portland cement (SP30): strength is 28-day cube strength for a lOOx lOOx ]00 mm cube. (From ref 42).
appear in the literature. For example, Carette and Malhotra** reported strengths for concretes with a W/C + S ratio of 0 - 4 0 where 0 - 3 0 % of the cement was replaced by OPC. Different amounts of a superplasticizer were used to maintain workability. The 28-day com pressive strength did not vary for mixes with CSF from 5 to 20%, in contrast to expectations based on Norwegian results. The strength potential of CSF was found to vary somewhat with cement type."' However, when the uncertainty in the efficiency factor is taken into account, it has been shown by analysis of results with four cement types (one OPC mix, two PFA blends and one slag blend) 11
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:52:00
C S F IN
CONCRETE
that such differences are probably not statistically significant.^^ Malhotra and Carette^^ have shown that low dosage of CSF may be used to compensate for strength reduction in blended cements with 50% slag, but only at ages above 14 days. There has been much speculation regarding the use of water-reducing agents in CSF concrete, and their role in dispersing the fine particles. From a practical point of view it is generally necessary to use them, since the strengthgiving potential of CSF may otherwise be cancelled out by increased water demand. However, there is little evidence that mixes without water-reducing agents pro duce reduced properties in CSF concrete when a com parison is made on an equal W/C + S b a s i s . O n e exception is for flexural strength, which will be discuss ed later. Microscopic examinations^^ have revealed dif ferent degrees of CSF dispersion in hardened concrete; however, such differences have not been related to dif ferences in concrete properties. The results cited above apply to water-cured concrete of 28 days or higher age. A number of papers have in vestigated the effects of early drying on strength proper ties. Since the major contribution of CSF to strength develops within 3—28 days, it is natural to suspect that CSF concrete is vulnerable to inadequate curing. Johansen^^'^^ and Sellevold and Radjy"^^ reported com pressive strength development for continuously watercured CSF concrete and parallel cube specimens exposed to laboratory air directly after demoulding. Concrete both with and without CSF suffers long-term strength loss due to the early drying, in the range of 10—20% reduction relative to water-cured companion specimens. There is little difference in this respect between the two types of concrete, except possibly that the strength loss is somewhat greater for CSF concretes of strengths less than 30 MPa than for equivalent controls. Maage and Hammer^^ found that dry curing from demoulding of 100 mm^ cubes made with OPC and 10% PFA-blended cement did not reduce the strength of the concrete relative to water-cured controls until after 6 months. With a 25% PFA blend and a 15% slag blend
§ 3
/
11 O
CO
C S F concrete
•
P C concrete
o
A
A
2
W
I ' c
B 2
0
1
2
3
4
5
6
7
8
9
-I 10
S q u a r e root of c o m p r e s s i v e s t r e n g t h : ( M P a ) V2
Fig. 16. Relationship between uniaxial tensile strength and compressive strength for a range of different concrete qualities. (From ref. 76).
the reduction appeared after about 14 days of dry curing. 10% CSF mixes increased the effect, and the reduction in strength started after 7 days curing for all four cement types, with the cements ranked as in the reference mixes. After one year the reduction for dry-cured reference mixes was 10—30% depending on the cement type, while the CSF mixes showed 20—45% reduction relative to watercured companions. The curing was exceptionally bad in these experiments, since the 100 mm^ cubes dried from all sides directly after exposure. A follow-up series run at three laboratories (not yet published) indicates that the strength losses due to bad curing are much reduced when the cubes are allowed to dry from one side only. Loland and Hustad^^ found that either type of concrete wet-cured 7 days before exposure to drying, suffered lit tle strength loss, as was found by Peterson et al.^'^ for concretes wet-cured 5 days before exposure. It is evident from these results that CSF concrete requires protection at early ages to realize its potential. 6.2 Tensile and flexural strength The relationship between compressive strength and ten sile, flexural, or splitting strength has been studied in a few cases. Loland and Hustad"^^ reported direct tension and flexure data for 25 mixes, continuously water-cured or exposed to drying after 7 days in water. Ages at the time of test were about 3 months for the water-cured samples and 1 year for those exposed to drying. Loland and Gjorv"^^ plotted the results as tensile strength versus square root of compressive strength for both curing con ditions (Fig. 16). Dry-stored specimens without CSF have a higher ratio of tensile to compressive strength than wetstored specimens. For CSF mixes the curing condition has no systematic influence, and the curve falls in between the two curves for the control mixes. Flexure results follow the same pattern as results for tensile strengths. Johansen^^'^^ reports both compressive and flexural strength data for concrete exposed to drying on demoulding. Ages at the time of test were 1 and 3 years. There was little overall change in either strength proper ties from 1 to 3 years, and no significant influence of CSF was found. A plot of flexural versus compressive strength showed that the same relationship was valid for mixes with 0 and 5% CSF, both with and without water-reducing agents. Mixes with 11% CSF and water-reducing agent also followed the same relationship, but the 11% mix without water-reducing agent and 25 % CSF mixes both with and without water-reducing agents had lowered ratios of flexural strength to compressive strength. Maage and Hammer^^ measured tensile strength for mixes with four cement types and various CSF dosages, both water-cured and exposed to drying after demoulding. Water-cured CSF mixes uniformly produced higher ten sile strengths, for given compressive strengths, than reference mixes with all types of cement. After 3 months of dry curing, the blended cement mixes performed significantly worse ( 1 5 - 3 5 % ) than their water-cured com panions, while the OPC mix only suffered a small reduc tion in tensile strength due to dry curing. For concrete with 10% CSF, OPC mixes also suffered a loss of about 25% in tensile strength, while blended cements lost 40—50% of tensile strength. Thus, the tensile strength of concrete made with blended cement, or any cement with CSF, is very sensitive to curing conditions. Fig. 17 shows that all the concrete mixes produce more or less equal curves for the two curing conditions in a plot of tensile
12
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
MECHANICAL
strength against compressive strength. (Note that the overall plot in Fig. 17 shows a general trend, but obscures the detailed results given above.) Fig. 17 is in contrast to the behaviour when dry curing is preceded by 7 days of water curing as discussed above (Fig. 16). The pic ture is consequently not entirely clear at present, but it indicates that the early curing history is quite important. Dispersion may also play a role, since mixes without plasticizers were more sensitive to bad curing than mixes containing plasticizers. This applies to concretes of equal compressive strength. 6.3 Brittleness and E-modulus Cement pastes with more than 10% CSF have been found to be very brittle.'^'^^ Principles of fracture mechanics have been applied to study brittleness or duc tility in CSF concrete. Loland and Hustad^^ compared ductility of 25 MPa concrete with 0 and 10% CSF by weight of cement. They concluded that the difference was very small, with a tendency to somewhat lower ductility in the CSF concrete. A more comprehensive study by Loland^^ (reviewed by Loland^^) covered many grades of concrete, as well as wet and dry curing. A plot of 'maxi mum fracture zone deformation' against tensile strength did not reveal any systematic differences between CSF and control concrete. The curing condition had an in fluence, but the major effect was, as commonly agreed, that increased strength led to increased brittleness. Tunnel segments of high strength concrete ( > 80 MPa) with CSF were produced to protect gas pipelines on the western shore of Norway. Details of the project, including stress-strain diagrams for the concrete, were reported by Einstabland et al The stress-strain curves were almost linear up to 0.2% strain, with a fracture strain of close to 0-3%. More complete stress-strain diagrams, using a stiff testing machine designed for high-strength and ultrahigh-strength ( > 100 MPa) concrete with CSF, has recently been reported by Helland et al}^ They found that as the strength increased, the fracture strain also in creased. In addition,the 'falling' part of the stress-strain curve increased in steepness (Fig. 18). This is in line with experience with normal concrete. A brittle concrete does
PROPERTIES
not necessarily mean a brittle construction; that depends on design and reinforcement. Helland et al.^^ also reported test data for over-reinforced beams, which did not disintegrate, but failed gradually — presumably because the compressive failure was local and spread progressively inwards in the cross-section. The high failure strain of ultra-high-strength CSF con crete is a consequence of the fact that the ^-modulus does not increase nearly as much as the compressive strength. Loland^^ gave ^-modulus versus compressive strength data for a large number of concretes, demonstrating this. He found no significant difference between CSF and con trol concretes. Sellevold et al. showed that the dynamic f'-modulus of cement pastes increased with increasing CSF addition. In concrete, however, the aggregates dominate the ^-modulus. Hence, increased stiffness of the binder phase is reflected only to a moderate degree in concrete. Larrard et al. investigated the fracture characteristics of normal-strength concrete (54 MPa), high-strength con crete (76 MPa), and very-high-strength concrete (105 MPa) containing 10% CSF. They concluded that although the latter two showed 'quasi-explosive' behaviour in normal compression testing, they did show 'significantly better fracture toughness than usual normal concrete', and consequently that 'it is absolutely possible to build duc tile structures with high-strength and very-high-strength concrete'. 6.4 Fly a s h - C S F combinations Several investigators have explored the possibility of using CSF in combination with fly ash.^^'^^-^^'^^"^^ The purpose has generally been to use the highly reactive CSF to compensate for the slow strength development associated with fly ash in concrete. Results obtained by Carette and Malhotra^^ confirm this possibility (Fig. 19). Work in this area has recently been carried out in Nor4 3 , 6 3 rpi^g blended cement is finely ground, so that strength development matches that of OPC. A wide range of properties of CSF concrete made with such a blended cement are being studied. Results do not indicate any
120
W a t e r cured
- 0 % CSF . 7-10% CSF - 15% CSF, basalt aggregate
100
80
if
% TO
(D ^
in
2
60
S
CO
Air cured
40
20
_L 10
20
30
40
50
60
C o m p r e s s i v e strength: M P a
Fig. 17. Relationship between compressive and tensile strengths for water-cured samples and samples exposed to dry ing in air from the point of demoulding. Four cement types, and 0 and 10% CSF dosages are included. (From ref. 63).
0-1
0-2
0-3
0-4
0-5
0-6
Strain: % o
Fig. 18. Stress-strain diagram for high strength concrete tested in compression. (From ref. 81). 13
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
C S F IN
CONCRETE %CSF '20 + 8 0 % control 15 + 8 5 % control 10 + 9 0 % control
50 r 40 h
\- 95% control
^Reference: 100% Portland cement Control: 70% Portland cement, 30% fly ash 28 Age: days
Fig. 19. Compressive strength versus age for concrete a W/C + S ratio of 0-50. (From ref 83).
with
significant difference in CSF effects in concretes contain ing OPC or a blended cement containing 10% fly ash. Results for blended cements containing 25 % fly ash show greater sensitivity to curing temperature and curing con dition, as already discussed above. 6.5 Bond properties Bond properties of concrete include bond to aggregates, bond to reinforcing steel, bond to fibres, and bond when fresh concrete is placed on old concrete. Johansen and Dahl^^ investigated concrete-to-concrete bond for various combinations of normal and CSF con crete. They concluded that CSF was 'an important aid material to obtain good bond at joints or for two-layer construction'. One contributing reason for this is that segregation is small in CSF concrete. Hence, a weak top layer as a result from a horizontal pour is avoided. Carles-Gibergues et al. have studied the zone form ed between aggregate and cement paste (the 'aureole').
This zone is a point of weakness, being available for crack propagation in concrete. They report that the characteristics of the aureole are altered in the presence of CSF. In particular, it reduces the thickness of the zone. Regourd^ observed an 'excellent cement paste-aggregate bond' in mortars with CSF. Part of the reason for an im proved bond phase is presumably the greater stability of CSF mixes. Biirge^^'^^ reported improved bond strength to reinforcing steel by the use of CSF both for high-strength lightweight concrete and for high-strength normal weight concrete. A project to investigate interface structure and bond strength (pull-out test) between concrete and reinforcing steel has been reported by Gjorv et ai^^ and Monteiro et al"^^ They tested concretes with varying strength levels and CSF dosages from 0 to 16% by weight of cement. They concluded that addition of CSF gave im proved pull-out strength (Fig. 20), particularly at high compressive strength levels of the concrete. The inter face zone was found to be 'more densified', i.e. reduced both in porosity and thickness. Bache^^ reported high bond strength to plastic fibres in a review of ultra-high-strength CSF mixes, as did Krenchel and Shah.^"^ Ramakrishnan and Srinivasan^^ studied steel fibre reinforced concrete and found that, in general, CSF improved the performance characteristics of steel fibrereinforced concrete. Scandinavia has recently seen the widespread use of steel fibre-reinforced shotcrete using the wet process, since this makes it possible to avoid wire mesh reinforcing nets. A
30 000 r
(b) 24 000
18 0 0 0
3 ?
12 0 0 0
6000
30 000 r (d) 24 000
0% CSF 8% CSF 16% CSF
-L. 3000
5000
7000
9000
11 0 0 0
C o m p r e s s i v e s t r e n g t h : p.s.i.
13 0 0 0
3000
5000
7000
9000
-L.
-I
11 0 0 0
13 000
C o m p r e s s i v e s t r e n g t h : p.s.i.
Fig. 20. Pull-out strength versus concrete compressive strength: (a) deformed bars (upper posi tion); (b) plain bars (upper position); (c) deformed bars (lower position); (d) plain bars (lower position). (From ref 91). 14
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
MECHANICAL
recent comprehensive study of field projects with labtiratory follow-up'**' concluded that CSF is an import ant ingredient in the mix. but due to lack of control in field situations the effects of CSF are difficult to quantify. The need for an accelerator in CSF mixes to maintain low rebound losses appears to have been reduced. Use of the methcxl in the USA was reported by Corwine."^' A comprehensive review of CSF shotcrete in Canada was recently given by Morgan et The use of CSF in carbon fibre-reinforced cements has been investigated by Ohama et al.'*'' They conclude that CSF plus a water-reducing agent arc efficient in dispers ing the carbon fibres, and that the CSF improves fibre anchorage. Bcrgstrom and Gram'"" used CSF to reduce the pH in the pore water of the cement matrix in order to improve the durability of alkali-sensitive fibres. They concluded that embrittlemcnt was delayed but not prevented for glass fibres. For sisal fibres they found that embrittlemcnt "can be avoided almost completely". Experience with CSF in asbestos-free fibre-reinforced cement was reported by Radjy et al. All these studies indicate that CSF can be used to improve bond properties of different kinds, as well as to improve durability properties. This is an important area that may increase the potential of cement-based binders. Case study E . Norwegian shotcrete In 1986 about 45 (XX) m' of shotcrete were produc ed in Norway, mainly for rock stabilization and tunnei lining. The majority of this was pr(xluced by the wet process method and applied by robots. By this method the concrete is mixed with water at the batching plant, and only an accelerator is added at the nozzle. 60—70% of the production was steel fibre-reinforced, and about 85% contained CSF. CSF is used to improve the pumpability, the strength (due both to better quality of the concrete from the batch ing plant and to the reduced need for accelerator), the permeability, and the bond to the fibres. A normal mix for wet process shotcrete is: 4 0 0 - 6 0 0 kg cement up to 10% CSF 2 - 8 litres plasticizers max size of the aggregate = 4 - 1 2 mm W/C + S = 0 - 4 0 - 0 - 6 0 slump value = 12 — 24 cm. The normal requirements for the applied shotcrete are grade C-25 or C-35. However, by the use of CSF and plasticizers it is also possible to produce high quality shotcrete with compressive strength up to 100 MPa. Source: refs 102. 103. 6.6 Shrinkage Traetteberg and Alstad"^^ measured shrinkage of pastes with W/C + S-ratioof 0 - 5 0 , with 0, 5, and 15% of the cement replaced by CSF. The pastes were cured for either 2 or 28 days before exposure to relative humidities from 12 to 75%. For samples exposed after 2 days curing, the results indicated greater shrinkage at increasing CSF con tents. This was particularly significant at RH values below 50%, i.e. in a range below that at which concrete shrinkage normally is assessed. The same trend was observed for pastes pre-cured for 28 days. The weight loss of the CSF
PROPERTIES
pastes was smaller than for the control pastes in the 'capillary range', i.e. at RH values above about 40%. These observations seem natural taking into account that the pastes were made with a constant W/C + S ratio: CSF pastes have finer pore structure and consequently retain more capillary condensed water at a given RH. At RH values below the capillary range, the weight loss and the shrinkage are controlled by the amount of CSH, which is greater for the CSF pastes. The shrinkage of concrete depends on the shrinkage of the cement paste, but also on the volume fraction of the aggregate. Several investigations have been made of shrinkage in concrete by standard methcxis, i.e. un restrained shrinkage of prismatic specimens at RH values of 50 - 60%. Johansen measured .shrinkage on con crete prisms exposed to 50% RH directly after demoulding and after 28 days of water curing. The W/C + S ratio varied from 0-37 to 1 -06, and the CSF dosage from 0 to 25% by weight of cement. The mortar fraction was also sieved from the fresh concrete, and parallel shrinkage measurements were performed on mortar prisms. The con cretes were produced with predeterminal cement plus CSF contents both with and without a water-reducing agent. The water demand to prtxluce the desired slump therefore varied widely, resulting in varying W/C + S ratios as well as aggregate volume fractions. It is therefore difficult to compare the measured shrinkage values directly. For samples precured for 28 days and with 0. 5. and 10% CSF the shrinkage values at 3 years varied little and were not directly related to either water-reducing agents, aggregate volume fraction, or W/C + S ratio. AT 25% CSF dosage the influence of water-reducing agents was evident in that they pri.xluced significantly higher aggregate volume frac tions and therefore lower shrinkage over the entire W/C -f- 5 range. This applied to both 1 and 28 days of precuring. In general, only 1 day of pre-curing resulted in higher shrinkage, particularly for CSF mixes and for con trol mixes with W/C ratios of less than 0 - 6 0 . Thus, for quality concrete (W/C + S < 0-60) no significant dif ference was found between control mixes and mixes containing 0, 5, and 10% CSF. The 25% CSF mixes pro duced higher shrinkages, particularly mixes with no waterreducing agents. Shrinkage data on mortars generally followed the same pattern, with a small tendency to increased shrinkage with increasing CSF dosage up to 10% and a clearly significant difference at 25% CSF dosage. Johansen applied a simple composite model to relate the shrinkage values for the concretes and for their equivalent mortars. He found that increasing CSF dosage generally led to a lower ratio between concrete and mor tar shrinkage, and concluded that this must be caused by a higher degree of stress relaxation in CSF concrete, since the shrinkage potential in the mortars was not fully reflected in concrete shrinkage. He also suggested that this might be caused by increased microcracking in CSF concrete. A s discussed in section 7.1.2, microscope studies of field CSF concretes have suggested a higher density of microcracks than in control concrete.^** Since the importance of such microcracking for concrete pro perties has not been established, the implications of this are unclear. Loland and Hustad™ measured shrinkage for 25 con crete mixes, with 0, 10, and 20% CSF by weight of cement and W/C + S ratios from 0-37 to 2 -11. The specimens were exposed to 60% RH after 7 days of moist curing. The shrinkage after 14 months was not significantly difLS
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
C S F IN
CONCRETE
ferent for concretes with W/C -\- S ratios over 0 - 6 0 . At lower values the mixes containing CSF had less shrinkage than the controls. Carette and Malhotra"^^ moist-cured the concrete for 28 days and subsequently measured shrinkage 84 days after production. Two series were tested. One used about 230 kg cement plus CSF, with CSF dosage from 0 to 43 %. No water-reducing agent was used; hence the W/C + S ratio varied from 0 - 6 4 to 0 - 8 4 with increasing CSF dosage. The other series had the same range of CSF dosage, but the cement plus CSF was kept constant at about 400 kg/m^ and the W/C -h S ratio at 0 - 4 0 by using dif ferent amounts of superplasticizer. For both series the pat tern was less shrinkage for the concretes containing CSF. The measuring time was short in these experiments, which may have distorted the results. For example, Maage'^^ has reported shrinkage with 0 and 10% CSF cured for 28 days prior to exposure, where shrinkage for the CSF mix was less than for the control of equal strength at times up to several months, but then tended to catch up (Fig. 21). When exposed immediately after demoulding, the CSF mixes showed somewhat greater shrinkage over the whole time span. Maage's experiments were performed both with OPC and a blended cement with 10% fly ash interground. The shrinkage properties of the two cements did not differ. Malhotra and Carette^^ measured shrinkage over up to 100 days in slag-cement concretes containing CSF, and found only marginal differences relative to controls. Petersson et al.^^ measured shrinkage after 7 days of wet curing for concretes with 0, 7, and 13% CSF con tent by weight of cement. The mixes were lean, with W/C -h S ratios from 0 - 8 0 to 0 - 9 4 and 28-day compressive strengths from 27 to 35 MPa. At short curing times the CSF mixes had higher shrinkage, but at curing times of over 80 days there was little difference from the control. Buil and Acker'^^ measured shrinkage of a control concrete {W/C = 0-44) and one containing 33% CSF (W/C -h 5 = 0-40). The CSF concrete had shrinkage strains roughly one-half as large as those of the control after about one year. For high-strength mortars with and without 40% CSF, Buil et al."^^ found slightly higher shrinkage after 90 days for the CSF mix (99 MPa) than for the control without superplasticizer (50 MPa), while a control (C3 MPa) with the same superplasticizer dosage as the CSF mix had almost twice the shrinkage. For a series of very-high-strength concretes with CSF,
Helland et al^^ measured shrinkage according to ASTM C157.^^^ They found that shrinkage values were low in comparison with normal concretes, and that shrinkage was linearly related to the water content of the concretes. Wolsiefer'^^ has also reported shrinkage for highstrength ( > 100 MPa) CSF concrete. For samples moistcured for 14 days prior to exposure (ASTM C257^^^), lower shrinkage was observed than for reference concrete, while exposure after 1 day led to higher shrinkage for CSF concrete. It appears from the existing evidence that at equal W/C -h S ratios, CSF paste has a higher shrinkage potential than a control paste. This effect is most evident at CSF dosages above 10% by weight or cement, as also seen in mortar tests. The effect of exposure to very early dry ing is less clear: pastes with or without CSF showed litde sensitivity, while mortars showed a clear tendency to in crease shrinkage. This tendency was of the same magnitude for mortars with 0, 5, or 10% CSF. Shrinkage data for concrete are even more difficult to assess, since the volume fraction of aggregates plays an important role, and in laboratory experiments this variable is not usually optimized. Minimum shrinkage is obtained when the binder phase volume fraction is minimum and its quality maximum (low W/C -h S ratio). Hence, the use of water-reducing agents is of importance in that it allows reduction in binder phase volume fraction. However, water-reducing agents by themselves lead to increased shrinkage, thus offsetting some or all of the advantage gained by the reduced water content. The results reviewed here indicate that concrete shrinkage is little influenced by CSF contents, at least up to 10% by weight of cement. For concrete exposed to dry ing very early, shrinkage is increased; this effect is most marked for lean CSF mixes {W/C -h 5 > 0 • 60) and high CSF contents ( > 10% by weight of cement). This con clusion seems natural, since early drying inhibits the pozzolanic reaction. CSF concrete dries out more slowly than control con crete of equal strength. This may partly explain why the higher shrinkage potential in cement paste containing CSF is not reflected in concrete shrinkage. It is known that slow drying (or large specimen size) leads to decreased shrinkage for a given concrete. Finally, it should be noted that crack sensitivity of con crete is not only related to shrinkage, but to the creep and stress relaxation properties as well. Data on shrinkage alone are therefore of limited value in connection with crack sensitivity.
002
a X LU
002
004
i ^
C
0-06
^ 0081-
- 0 % C S F , c u r e d in w a t e r — 0 — 0 % C S F , c u r e d i n air — • — 1 0 % C S F , c u r e d in w a t e r - - • - - 1 0 % C S F , c u r e d in air
7d
28d Time
3
6 months
Fig. 21. Development of shrinkage for samples cured in water and in air containing P30 cement (OPC) and MP30 cement (90% OPC + 7 0 % PFA), both with and without 10%c CSF. (From ref 105).
6.7 Creep Few reports are available on creep of CSF concrete. Buil and Acker^^^ measured drying creep and creep in a sealed environment for a control concrete {W/C = 0 - 4 4 , 53 MPa) and one with 33% CSF {W/C -h 5 = 0-40, 76 MPa). Unloaded sealed samples suffered substantial strains, either caused by self-desiccation effects or in complete sealing. The creep of the sealed specimens was roughly equal for the two types of concrete, while the CSF mix had substantially less total creep deformation under drying conditions than the control. Wolsiefer^^^ measured creep of CSF concrete. Highstrength mixes were loaded at 12 h and at 28 days. Up to 4 months, the CSF concretes showed less creep than has been reported for normal high-strength concrete. Creep of high-strength concrete (112 MPa) containing
16
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
MECHANICAL
10% CSF was measured by Penttala,'^^ who found that the concrete drying creep values were some 80% of the CEB values, but that the creep takes place much faster than the CEB formula predicts. Close to 80% of the total deformation after 400 days in water before being loaded in atmospheres of 60% and 80% RH. Tazawa and Yonekura'*' also measured creep of CSF concretes with W/C ratios from 0 - 3 0 to 0 - 6 5 and 30% of the cement replaced by CSF, i.e. a high CSF dosage of about 43% by weight of cement. The strengths varied from 40 to 75 MPa — quite low for such compositions. Specimen were loaded after 28 days' curing in water and exposed to two conditions: storage in water, and drying at 50% RH. For water storge ('basic creep') the specific creep did not depend much on concrete strength level, and CSF concretes had slightly higher specific creep than controls. Drying creep decreased strongly with increas ing compressive strength, and CSF mixes had almost doubled the specific creep values compared with the con trols over the whole strength range. It should be noted that all of the results given here are for either very high strengths or very high CSF dosages. They may therefore not be representative of more nor mal CSF structural concrete. 6.8 Fire resistance Aaneland''^ exposed one face of concrete elements to ISO (wood fire) and HC (hydrocarbon) fire loads. Four concretes were tested, one of which contained 9% CSF by weight of cement. 28-day cube strengths for the four concretes were 32—35 MPa. The elements were tested after 12 weeks. All elements fulfilled the requirements as to temperature on the unexposed side. Spalling on the exposed side was also evaluated. The CSF concrete had a somewhat higher spalling, but the data were too limited to allow general conclusions to be drawn. Pedersen'^^ exposed 100 mm cylinders of highstrength CSF concrete (W/C + S = 0 -16 and 20% CSF) to slow heating (1 °C/min). Several of the specimens sud denly disintegrated at temperatures near 300°C. It should be kept in mind that this mix is extreme and by no means representative for CSF concrete. However, certain characteristics of CSF concrete are important in connec tion with the spalling problem. CSF concrete dries out very slowly (see section 7.1.1), while its thermal conduc tivity probably is as high or higher than for normal concrete. These factors indicate that high vapour pressure can be built up internally and cause spalling. Hertz^^"^ has discussed the question in connection with dense CSF concrete. These Danish results have motivated further studies of fire tests of high strength concrete. Maage and Rueslatten'^^ heated 100 mm cubes in a furnace to 1150°C over a 60 min period. Three concretes were tested: one with 15% CSF and mature strength of 138 MPa, and two reference concretes of about 80 MPa strength. All three concretes were tested in two moisture conditions: after continuous water storage, and after several weeks drying at 105°C. The test specimens were inspected during the heating at 570°C and 700°C. After the end of the tests the remaining strength was measured. At 570°C only one sample showed spalling or cracking, at 700°C more cracks were visible, and at 1125°C all samples were cracked. The remaining strength was about 8% of the original for all three concretes. One mix with 7% CSF (124 MPa at 19 months), exposed over a long
PROPERTIES
time to indoor climate, performed similarly to the others by visual inspection. The main point in this small test series is that no explosive effects were observed in spite of extreme variations in moisture states. Shirley et al^^^ produced 100 mm thick slabs from five mixes: one normal strength control (50 MPa), two highstrength controls (70 and 90 MPa), and two CSF mixes (70 and 120 MPa) with 10% and 15% CSF by weight of cement. The slabs were moist-cured for 7 days and then dried at about 30°C and 2 0 - 3 0 % RH until testing at ages of 8 0 - 1 3 0 days. The RH value at slab mid-depth was reduced to the range of 7 7 - 8 4 % at testing. Fire testing was carried out according to ASTM El 19,^'^ where one face of the slab is exposed to a rise in temperature to 600°C in about 10 minutes and to about 1100°C after 4 hours. The conclusion was that the tested concretes 'revealed no significant difference in behaviour'. There was no explosive behaviour, and none of the concretes showed even minor spalling on the exposed surface. As part of a large project on offshore concrete, Jensen et al.^^^ reported on 100 mm thick, prestressed concrete DT elements exposed to hydrocarbon fires reaching 1100°C in 30 min. Four elements of each of three con crete mixes were tested: control concrete (49 MPa), CSF concrete (10% CSF, 53 MPa) and a lightweight aggregate concrete (58 MPa). The prestressing cables were cut after 3 days, and from then on the elements were stored out side without protection for about 6 months (summer and autumn). From that point on, the curing varied as follows: four elements were stored outside under cover for 3 weeks before testing; four elements were stored inside at 20°C for 4 weeks before testing, and four elements were stored inside for 2 weeks, fire insulation was applied, and then they were stored inside for another 5 weeks before testing. The moisture content was determined before fire testing by chipping off pieces from the element sides and drying at 105°C. The moisture content varied from 3 - 4 % to 6-4% by weight of dry concrete; the lowest values were for CSF concrete and the highest for the control mix. Moisture content for neither concrete was very dependent on storage conditions. No measurements were made of the relative humidity in the concretes. The results of the fire tests were unsatisfactory for all uninsulated elements; the elements stored outside started to spall after about 5 minutes, and the prestressing wires were exposed after 8-5 minutes. The test was stopped after 31 min because one of the elements developed a hole right through. The development for the elements stored inside was similar, but somewhat delayed in time. Two of the insulated elements were undamaged, but on another the insulation scaled off and the behaviour was similar to that of the unin sulated elements. Jensen et al.^^^ also reported on two further test series: one on small test specimens and slabs (100 x 1000 X 1000 mm), and one on prestressed and reinforced beams, both loaded and unloaded during fire testing. The first of these series resulted in considerable damage to all specimens, particularly to those in a 'moist' condition. No significant difference was found for normal density concrete with compressive strength from 40 to 80 MPa, while lightweight concrete suffered the most pronounced damage. No CSF mix was included. The beams of the last test series largely confirmed the results on the DT beams, except that the CSF mix exper ienced most spalling. In general, prestressing resulted in increased risk of spalling. Specimens were tested at the 17
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
C S F IN
CONCRETE
ages of 13 weeks and 60 weeks; the latter were allowed to dry indoors for a year before testing. The risk of spalling after drying was low for both beam types with normal density concrete, but the risk was still apparent for lightweight concrete. The main conclusions of the work are that spalling and damage occur earlier than expected, and that prestress and increasing moisture content increase the severity of the effects. CSF concrete appears more sensitive to spalling. Lightweight concrete is more sensitive to spalling than normal density concrete, contrary to common expectation. The remaining strength of CSF concrete after heating to SSO^'C has been investigated by Dahl."^ Increasing the dosage of CSF resulted in somewhat reduced temperature resistance, while water storage after heat exposure led to large strength gains for mixes with 0 and 10% CSF. Mixes with 20% CSF experienced lower strength gains. W i l l i a m s o n e x p o s e d concrete blocks (approximate dimensions 150 X 260 X 500 mm) to the standard ASTM El 19'^^ time-temperature fire. Two reference mixes (30 and 80 MPa) and two CSF mixes (50 and 100 MPa) were tested. The relative humidity in the concrete was monitored during the storage period of about 6 months; it reached a level of 71 —74% RH at the time of testing. Williamson
90 h
Q>
6.9 Abrasion-erosion resistance Very low W/C ratio ( < 0 - 2 5 ) and high CSF dosage ( > 20%) mortars are known to be highly resistant to abra sion and wear. This property is currently used in a variety of applications. For concrete little published data are available. Holland'^^ has reported on a repair project using high-strength CSF concrete on the Kinzua dam still ing basin. Laboratory results on abrasion-erosion were very promising (Fig. 22), and repairs were carried out in 1983 after full-scale test pours were made. A followup report after one year's service'^^ states that 'the con crete appears to perform as intended'.
40 L
(a) (b) (c) (d) (e) (f) (g)
found no external spalling or internal damage to any of the specimens, and concluded that CSF concrete 'appears to give the same or better performance than comparable conventional concrete when exposed to the standard ASTM El 19 time-temperature fire exposure'. The reports cited here appear to have rather contradic tory results. The Danish test"^ showed an 'explosive' effect for a very dense mix {W/C 4- 5 - 0 • 16, 20% CSF) at a very low heating rate, while refs 115 — 120 indicate no such effects at much faster heating rates, a variety of moisture conditions, and concrete compositions to W/C + S = 0 - 2 2 , 15% CSF. It would be useful to perform experiments to determine if such a sharp limit on binder composition with regard to the explosive effect indeed exists. Except for the results in ref 115 it appears, as expected, that increasing moisture content leads to increased damage. The influence of concrete strength level and CSF content is less clear. None of the reports indicates that increased strength leads to increased damage by fire exposure. The role of CSF is only found to be detrimental in connection with spalling of prestressed beams.^'^ The major Norwegian study^'^ showed clearly that HC fire exposure in connection with prestressed or loaded con crete elements in a moist state is a serious problem at the moderate strength level of 5 0 - 6 0 MPa, regardless of mix composition.
Fibre-reinforced c o n c r e t e from K i n z u a stilling b a s i n . Conventional concrete. Pennsylvania limestone aggregate. C o n v e n t i o n a l c o n c r e t e . Virginia d i a b a s e a g g r e g a t e . C o n v e n t i o n a l c o n c r e t e . Mississippi chert a g g r e g a t e . A v e r a g e of C S F c o n c r e t e s p e c i m e n s p r e p a r e d during actual construction. C S F c o n c r e t e . Virginia d i a b a s e a g g r e g a t e . C S F concrete. Pennsylvania limestone aggregate.
Fig. 22. Abrasion-erosion test data for various concretes in the laboratory. (From ref. 121).
tested
6.10 Conclusions The contribution of CSF to any property of hardened concrete may be expressed in terms of an efficiency fac tor, K. For compressive strength K is in the range of 2—5, meaning that 2 to 5 kg of cement may be replaced by 1 kg of CSF in a given concrete without impairing the com pressive strength. This applies provided the water con tent is kept constant and the CSF dosage is less than about 20% by weight of cement. is a 'marginal' quantity, and cannot normally be determined accurately. CSF makes it possible to produce high-strength concrete (over 100 MPa) on a routine basis. Tensile and flexural strengths of CSF concrete are related to compressive strength in a manner similar to that of normal concrete. However, if CSF concrete is exposed to drying after only one day of curing in the mould, the tensile and flexural strengths are reduced more than for control concrete. The brittleness of normal concrete increases with in creasing strength level. CSF concrete appears to follow the same pattern as normal concrete in this respect. Various studies indicate that CSF can be used to improve any bond property of concrete; to aggregate, to reinforc ing steel, to various fibres, or to old concrete.
18
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
DURABILITY
Cement paste and mortar containing CSF appears to have a larger shrinkage potential than controls at equal W/C + 5 ratios. This increased potential has not been found to be reflected in increased concrete shrinkage. Fire exposure tests indicate that spalling is a potentially
7.
serious problem for high-strength concrete in general, and it is particularly sensitive to the moisture content of the concrete and to the state of stress. High-stress CSF concrete has a potential for high abrasion-erosion and wear resistance.
HARDENED CONCRETE: DURABILITY
A substantial number of investigations have been con cerned with various durability aspects of concrete con taining CSF. Review papers in English on physical and chemical durability have been published by Hjorth^ and Gjorv'^^ and on corrosion protection of reinforcement by Vennesland and Gjorv.^^"^ A general agreement exists that, apart from the chemical composition, the permeability of a given concrete is a good indicator of its durability both to physical and chemical aggression. Data on the permeability will therefore be discussed first in this section. 7.1 Permeability The term permeability may be applied to gas, vapour, and liquid transport through a porous material, although the term diffusion is more generally used when describ ing gas or vapour. Two types of experimental techniques have been used with CSF concrete and cement paste: observation of weight loss versus drying time, and water transport under a pressure gradient. Oxygen and chloride diffusion in water-saturated concrete have also been measured, but will be discussed in connection with cor rosion of reinforcement (section 8). 7.7.7 Drying experiments Sellevold et al. measured drying rates for 3 mm thick discs of cement-CSF pastes containing 0—20% of CSF by weight of cement. One series, in which the W/C ratio was kept constant and CSF used as an additive, gave the result that the relative diffusion coefficient was reduced from 1 - 0 at 0% CSF to 0-25 at 20% CSF. The reduction was non-linear, i.e. the effect per unit weight of CSF added decreased with increasing total CSF dosage. A parallel series, in which the water content in the paste was kept constant and the cement replaced by CSF in a ratio of 3:1, gave equal diffusion coefficients for 0, 8, and 16% CSF contents by weight of cement. This result indicates that with respect to water diffusion during drying the efficiency factor (K) of the CSF in cement paste is about 3, close to the value of generally found for compressive strength. Sorensen^"^ calculated relative diffusion coefficients from drying experiments on concrete discs containing 0, 8, and 16% CSF by weight of cement. The efficiency fac tor for the CSF with respect to drying calculated from these results was in the range 6—8, i.e. higher than K for compressive strength. These two investigations considered together show that the CSF has a greater effect in concrete than in cement paste, indicating that the boundary phase between aggre gate and paste is improved more by the use of CSF than the cement paste phase itself This indication has implica tions for any other property of concrete which depends on the quality of the interface, for example bond and water permeability.
7.7.2 Water permeability The first permeability tests on CSF concrete were car ried out in the 1960s by Markestad.'^^ Using a lean mix (W/C = 0 - 8 9 ) , and replacing 20% of the cement with CSF, the concrete was found to be 'completely im permeable' to water under 7 atm pressure for 15 days (60 mm thick disc specimens). In 1975, water penetra tion tests according to Swedish standards were carried out on concrete with 0, 10, and 20% CSF replacement of the cement on a 1:1 basis. This resulted in higher strengths for the CSF mixes, and about one-half as much water penetration in the 10% CSF mix as in the reference mix. Use of 20% CSF reduced the penetration further, but by substantially less than twice the reduction produc ed by 10% CSF. Note that no water-reducing agent was used in either of these investigations. As part of a comprehensive investigation, the water permeability of 25 concretes was measured by Hustad and Loland.^^^ The concretes contained 0, 10, and 20% CSF by weight of cement and were made both with and without water-reducing agents. The cement content ranged from 100 to 500 kg/m . When comparing permeability coef ficients on the basis of the same 28-day compressive strength, the CSF concretes were somewhat less permeable at strength levels up to 30—40 MPa. At higher strength levels the accuracy of the test method used was not high enough to permit comparisons; all mixes were essentially 'watertight'. The permeability coefficients for the strong concretes were in the range 1 0 t o 10"'^ m/s. Sandvik^^ has reported permeability coefficients for concretes of fixed composition except for 5, 10, and 20% replacement of cement by CSF on a 1:1 basis. The reference mix had 300 kg/m^ of cement and a 28-day cube strength of 31 MPa. The CSF mixes had 28-day cube strengths of 36, 43, and 44 MPa respectively. No waterreducing agents were used. The results do not permit com parison at equal strength levels. The mix containing 5% CSF produced a reduction in the permeability coefficient from 3 X 10"^' to 6 X lO"'^ m/s; at higher CSF replacement levels the coefficient was too low to be measured (i.e. below 1 0 " m / s ) (Fig. 23). Skurdal^^ measured permeability coefficients of a reference concrete and one containing 10% CSF by weight of cement. Both mixes had 28-day cube strengths of about 32 MPa at 2 0 ° C curing. Parallel samples were cured at 30°C and 5 0 ° C ; the former resulted in slightly lower strengths at 29 days, while the latter led to reductions of about 20% for both mixes. The permeability coefficients of the CSF mixes were consistently lower than for the reference mixes (20°C curing: 7 - 2 x 10"'^ and 0-5 x 10-'^ m/s; 30°C curing: 27 X 10"'^ and 0 - 8 x 1 0 " ' ' m/s; 50°C curing: 90 X 10'^^ and 74 X 10~'^ m/s). Heat curing of both types of concrete led to increased permeability. 19
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 14:55:23
C S F IN
CONCRETE
10 '
S t o r e d In w a t e r D r i e d for 1 4 daysat40'C prior to
testing
300/0 285/15 270/30 M I X proportions. c e m e n t / C S F : k g / i n '
240/60
Fig. 23. Hydraulic conductivity (permcahility) of different con crete compositions given different storage conditions. (From ref. 65).
Note that the values for the permeability coefficients in the two latter investigations do not match precisely for comparable concretes, in spite of the fact that the same experimental apparatus was used. This is a common phenomenon for penneability measurements, and emphas izes the need for caution when, for example, attempting to base evaluations of the effects of CSF on results from different test series. As part of a project to investigate the condition of field concretes with and without CSF. Maage''^" measured the water permeability of discs cut from drilled cores in various structures. The age of the structures varied from 3 to 9 years. The results are shown in Fig. 24. The abscissa is {W/C + 35) in order to compare concretes of similar compressive strengths. The trend is that CSF concretes have lower permeabilities at equal strengths. Microscopic examination of thin sections from the concretes was made by the Technological Institute in Denmark, which con cluded that CSF mixes in general were more dense, and that they contained more microcracks. The micrtKracks apparently have no influence on the permeability, nor on the measured stress-strain curves in compression.
1
10 "
10 ' 0-6 0-7 W / C (controls) W/C + 3S (CSF) Without C S F With C S F
10
Fig. 24. Permeability of samples drilled from siruciures, against W/C ratio for controls and W/C + iS ratio for CSF concrete. (From ref. 128).
In a comprehensive test series of CSF with blended cements already described,''' permeability was also measured.'-** The results were not entirely consistent, but on average CSF mixes had somewhat lower permeabil ity than control mixes at equal strength. CSF appeared to have least effect with a blended cement containing 25% PFA. The concretes were water-cured for an initial pericxl of 28 days and then exposed to 6 months of air-drying prior to testing. It has already been stated (section 7.2.2) that drying-rewetting treatment "opens" the pore structure of cement paste, and in particular of CSF-cement pastes.'" Johansen"" found that the permeability increased more for CSF concrete than for controls after such a moisture history. Mindess and Gray'" reported on a permeability study of CSF and control mixes, testing both pastes and cementaggregate composites. They concluded that CSF appears to decrease the permeability. However, the presence of aggregate was not found to have much influence on permeability. This is possibly because the aggregate-paste interface was tew small compared with that of normal con crete, and the great uncertainty associated with pemieability measurements therefore prevented the detection of any interface effects. 7.1.3 Conclusions The available data indicate that CSF in concrete reduces the permeability more than it improves the compressive strength, i.e. the efficiency factor is greater with respect to permeability than with respect to compressive strength. This appears to be particularly evident for low dosage levels of CSF, and at low concrete strength levels. Com parison of cement paste and concrete results indicates that it is particularly the aggregate-paste interface which is im proved by CSF. 7.2 Frost resistance The need for more durable concrete, in particular con crete with improved resistance to freeze/thaw exposure in the presence of salts, has motivated a number of in vestigations of CSF concrete. The investigations include studies of air pore system characteristics, basic studies of ice formation and pore structure, and freeze/thaw testing with and without deicing salts. 7.2./ Air entrainment As a part of a large study of air-entraining agents in concrete, Okkenhaug and Gjorv'^- and Okkenhaug'" have studied the effects of mixer type, mixing time, aggre gate grading, and air contents on the stability of the air content during handling of the concrete. They also studied air pore system characteristics for concrete both with and without CSF. Their conclusion was that a desired total air content could be obtained relatively easily in both types of concrete, by adjusting the dosage of air-entraining agent. Compared with control concrete, a higher dosage is needed in a CSF concrete without a water-reducing agent, but with a water-reducing agent the difference is decreased. O k k e n h a u g a l s o concluded that for a given content the air pore characteristics in CSF concrete are more favourable, and the air content is more stable with respect to vibration of the concrete, particularly in mixes with both water-reducing agents and air entrainment. Carette and Malhotra''*' found that for a low W/C + S ratio (0-40) the replacement of cement by CSF in a 1:1 ratio led to an increased need for an air-entraining agent.
20 Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
DURABILITY
The mixes also contained a superplasticizer. The air pore characteristics were measured on the hardened concrete, but in contrast to Okkenhaug and Gjorv,^^^ who used 8% CSF by weight of cement, the results did not indicate any consistent improvement. A dose of 5% CSF did improve both the spacing factor and the specific surface, but higher CSF dosage led to less favourable values than for the con trol concrete. Virtanen'^"^ found that CSF concrete needed a higher dosage of air-entraining agent than a control mix to reach a given air content, but the dosage was less than that required for concrete containing fly ash and slag. A plot of spacing factors against air content for the various mixes produced identical curves. In a further test series, Virtanen^^^ investigated high quality concrete (40—50 MPa) for bridge edge beams, containing from 0 to 16% CSF, and different air-entraining agent dosages. For a given dosage of air-entraining agents an increase in CSF content resulted in reduced air con tent in the fresh mix. No measurements were made of the air pore system in the hardened concrete. Lehtonen*^^'^^ studied concrete of W/C + 5 = 0-45 with 0, 5, 10, and 15% CSF, and found that a higher dosage of air-entraining agent was needed to produce equal amounts of air with increasing CSF dosage. In agreement with O k k e n h a u g , h e found that the air content in CSF concrete was more stable with respect to vibration. Such benefits are apparently not produced automatically when CSF is used; for example, both Aitcin and Vezina'^^ and Rasmussen'^^ reported somewhat greater loss of air from CSF concrete during handling and plac ing than for concrete without CSF. Christensen and Jensen^"^^ reported a comprehensive quality control programme on concrete (W/C -h S = 0-35, 15% CSF) for bridge edge beams. The requirements were a minimum of 3 - 5% air, specific surface of air pores of more than 30 mm~\ and a maximum spacing factor of 0-20 mm. It was found that 28% of the concrete tested did not fulfil the air requirement, 10% failed the require ment regarding specific surface, and all tests gave satisfac tory spacing factors. The authors concluded that require ments for the air pore system in hardened concrete should be in terms of the most direcdy measured values: air con tent relative to paste volume, and the specific surface area of the pore system. Mathematically derived quantities, such as the spacing factor, are based on a number of assumptions usually not justified. As part of a large project concerning the effects of CSF in blended cements, air entraining of CSF mixes has been investigated by Maage and Dahl."^^ Pure Portland cements were used, as well as 10 and 25% fly-ash blends and a 15% blast-furnace slag blend. CSF dosage was 0, 5, and 10%. The mixes were designed to produce equal 28-day strength at two strength levels: 25 and 45 MPa. An effort was made to produce realistic mix proportions, as opposed to simple replacement of cement by a certain factor. The results showed that CSF had remarkably littie effect on the need for air-entraining agent. The response of the concrete to handling was simulated using a drum mixer and vibration treatment. CSF also had little effect on the stability of the air content. Seen as a whole, the evidence is that proper air entrain ment in CSF concrete can be obtained as easily as in ordinary concrete. The air pore structure will not automatically be improved, but will depend on all the fac tors involved in the production of concrete.
The stability of the air pore system to handling and vibration of the concrete may be improved by CSF, but this again depends on the mix design. The large varia tions experienced in air entrainment of fly-ash concretes are not found in concrete with CSF. 7.2.2 Moisture conditions The moisture history and condition of a specimen is of particular importance to its frost resistance. It has been established by Fagerlund^^' that a given concrete possesses a 'critical degree of saturation' (CDS) above which it is susceptible to rapid deterioration under freeze/thaw conditions; at moisture contents below this level, deterioration is very slow. The CDS is considered to be a material property; in order to assess the frost resistance of a concrete in a given environment it is necessary to know what the moisture content will be in practice. This is normally done by capillary suction experi ments, where pre-dried specimens are exposed to water and the weight gain with time is measured. The weight normally increases very quickly in the first day or two until it levels off, after which the increase is very slow and presumably represents filling of air voids. The method allows the prediction of a service life, equivalent to the time needed to reach the critical degree of saturation. Work by Vuorinen^"^^ has resulted in a Finnish National Standard'"^^ in which the 'pore protection fac tor' is defined as the ratio between the unfilled pore space after capillary suction and the total pore space available to water (determined by pressure saturation at 150 atm). Empirical data suggest that the pore protection factor should be at least 0 • 25 for concrete exposed to severe conditions. It is implicit in these methods that the concretes are exposed to drying and re wetting treatments. Sellevold and Bager''^'* have established by low temperature calorimetry that even mild drying (58% RH) and re wet ting alters the pattern of ice formation in cement paste and mortar quite dramatically. Freezing in virgin specimens of W/C less than 0 - 5 0 is gradual over a wide temperature range. However, after drying (at elevated temperatures) and re wetting, more ice forms, and this takes place in a very concentrated manner as the ice first nucleates. Thus the drying and rewetting implicit in the Finnish and the CDS methods are not unrealistic in prac tical terms, and the effects of moisture condition should be taken into account in any practical frost-resistance testing. Calorimetric determination of ice formation in hardened cement pastes with various amounts of CSF were made by Sellevold et al. They found that a control paste with W/C = 0 - 6 0 had a large 'primary' freezing peak near 0 ° C , indicating, as expected, bad frost resistance. Addi tion of 8% or more CSF resulted in an absence of the primary freezing peak — the first freezing now taking place around — 2 0 ° C . This behaviour was interpreted as a result of altered pore structure in the paste. However, as shown in Fig. 25, when pastes without primary freez ing peaks were dried gently (at 58% RH) and resaturated with water, a distinct primary freezing peak appeared — as had also been found to be the case for control pastes. Experiments have shown^"^^ that pastes with extremely low W/C ratios and high CSF contents do not give any primary freezing peaks even after drying and resaturation treatment. Thus, it is possible to produce CSF concrete which is
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
C S F IN
CONCRETE
Cool
Is
10h \
yC-S
0-8
0-6 h
0-4h
-50
-40
-30
-20
-10
C a l o r i m e t e r block t e m p e r a t u r e : °C (a)
50
40
30
^
2 0
10
-14
-12
-10
-8
-6
-4
-2
C a l o r i m e t e r block t e m p e r a t u r e : °C (b)
Fig. 25. Low temperature calorimetry (heat flow) for watersaturated cement paste samples. The last two digits in the identification numbers represent CSF dosage. All samples have W/C = 0-60. Areas under the peaks give the heat of fusion, proportional to amount of ice formed, (a) Virgin water-saturated pastes; (b) dried-re saturated pastes. (From ref. 10).
frost resistant without air entrainment, provided that the W/C ratio is low enough and the CSF content high enough. An estimate of the required mix proportion would be a W/C ratio of approximately 0 - 3 0 with a CSF dosage of 10% or more by weight of cement. Industrial trial pro duction of such a concrete is under way, by the Danish Road Authority among others. Two recent programmes have investigated the moisture state of CSF concrete in connection with frost-resistance tests. Virtanen^^^ determined the 'protective pore ratio' for a series of concretes with different air content and CSF dosage of 0, 4, 8, and 16%. For a given air content, the protective pore ratio increased with increasing CSF dosage. For example, at 6% air the ratio increased from 0-25 in the reference concrete to 0 • 70 for the one con taining 16% CSF. Thus, the absorption of water is much slower in CSF, and possibly not only air voids remain empty after a fixed time period of water suction. Lehtonen^^^ provided more information by presenting plots of water uptake against time. For concretes with W/C + = 0 - 4 5 and 0 and 10% CSF contents, the critical degree of saturation was not very different (about 0 • 83) in spite of variations in air content. The water suction behaviour was quite different, however. The reference concrete quickly reached a plateau, while the CSF con crete showed a much more gradual water absorption. These results indicate major differences between CSF and reference concrete in their drying and wetting behaviour which probably are important in practical conditions, but are not normally taken into account in frost and durabil ity tests of the type described below. 7.2,3 Frost resistance testing In the first paper published on CSF concrete, in 1952, Bernhardt*"^^ included results on frost resistance. No admixtures were used, which resulted in an extremely high cement content (730 kg/m^) to produce concrete of high slump and a W/C ratio of 0 - 4 0 . CSF dosage varied from 10 to 30%, and the W/C ratio from 0 - 4 0 to 0-96. The test procedure consisted of freezing in air and thawing in sea water; frost resistance was assessed in terms of weight loss as a function of frost cycles. Concretes with CSF performed significantly better than the controls, par ticularly for concretes well cured prior to exposure to freeze/thaw. This is in line with later findings proving that CSF concrete requires more time to reach its potential. Traetteberg^^ tested frost resistance of 15-day-old mor tars with various air contents. The CSF dosage was be tween 0 and 25 % and the W/C + S ratio in the range of 0 • 48 to 0 • 83. The test procedure involved freezing in air and thawing in water; damage was evaluated in terms of residual length change and decrease in dynamic Emodulus. All CSF mortars with W/C + 5 ratios of 0-60 or less showed excellent frost resistance in terms of residual length change and dynamic £'-modulus, while the controls required a W/C ratio of 0 - 4 0 to be resistant. However, strength tests made after exposure of both a control mortar and one with 5% CSF (both with airentraining and water-reducing agent) showed practically zero bending strength. For the CSF mortar, this is in direct contrast to the other 'indirect' measures of frost resistance. The observed contradiction between direct and indirect measures of frost resistance for samples with water reducers and air-entraining agents was not consistent for other mixes. Later work has not revealed any similar pattem, and hence one must conclude that Traetteberg's results
22
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
DURABILITY
are anomalous. The main conclusions of the paper were that CSF improves the frost resistance in mortars, and is less dependent on a proper air void system. Traetteberg also measured pore size distribution of the mortars by mer cury intrusion, and concluded that the positive effects of CSF on frost resistance are related to alterations in the pore structure. Cheng-Yi and Feldman^^'"^^ also tested frost resistance according to ASTM C666 (B)'^^ of mortars with W/C + S ratios of 0-45 and 0 - 6 0 , containing 0, 10, and 30% CSF. At the high W/C + S ratio, adding CSF improved the performance. At low W/C + 5, 10% CSF led to improved resistance; however, 30% CSF resulted in bad frost resistance despite the greater strength of the mix. None of the mortars was air-entrained. The good results for CSF mortars were attributed to an increase in the pore volume in the range 0-35—20 /xm, which the authors believed act as air pores because of their ink-bottle shape. However, no data were given on the degree of saturation during the tests. The authors suggested that the observa tion showing that the strongest mix gave the lowest frost resistance was caused by a combination of low permeabili ty and a higher CSF dosage than needed for complete poz zolanic reaction. As part of a comprehensive study, the frost resistance of 25 concrete mixes with and without CSF has been in vestigated.'^' The method used was that of ASTM C666;'^^ in addition to measurement of dynamic Emodulus, the compressive strength of the concretes after freeze/thaw treatment was determined. In an effort to pro vide an accelerated comparison between the concretes, none of them were air entrained. The results showed that the efficiency factor of the CSF with respect to frost resistance is greater than with respect to compressive strength. Carette and Malhotra'*^ also tested the frost resistance (ASTM C666'^^) of concrete containing 0 to 30% CSF, all having W/C + 5 = 0 - 4 0 . Various dosages of a superplasticizer were used to obtain workable concretes, and all concretes were air entrained. The results were that increasing CSF content led to decreased frost resistance, particularly for high CSF dosages. Air void characteristics were determined for the concretes, but they do not explain these apparently anomalous results. However, in terms of the general experience with CSF concrete, the com pressive strengths of the freeze/thaw series of concretes were also anomalous in that cement replacement levels from 5 to 20% CSF gave identical 28-day strengths. Nor mally, increased replacement levels should lead to substan tial strength increases. There is no obvious explanation for these differences. Aitcin and Vezina'^^ report on freeze/thaw resistance (using ASTM C666'^^) of reference concrete and con crete where cement was replaced 3:1 by CSF. CSF dosage was about 8%. Both mixes were air entrained with the same average spacing factor. The authors concluded that the 'CSF concrete was far superior to the plain concrete'. Yamato et al.,^^^ using the method of ASTM C666 (A),'^^ tested concretes with W/C + 5 ratios of 0 - 2 5 , 0 - 3 5 , 0 - 4 5 , and 0 - 5 5 , and CSF dosages of 0, 5, 10, 20, and 30%. Except for the highest W/C + S ratio, none of the mixes was air entrained. Freeze/thaw testing was initiated after 28 days of curing in water. None of the samples, including the air-entrained ones, had satisfac tory air void systems. The test results showed that all mixes with W/C + 5 = 0-25 had durability factors above 90,
decreasing somewhat with increasing CSF dosage. For higher W/C + S ratios none of the concretes performed satisfactorily, but the general trend was that the poorest performance was for 20 and 30% CSF dosages. This is in line with the results in refs 46 and 35, which indicate that high ( 2 0 - 3 0 % ) CSF dosage at W/C + S ratios in the range 0 - 3 5 - 0 - 5 5 is detrimental to frost resistance, while lower dosages generally appear to be beneficial over a wide range of W/C -h 5" ratios. The resistance of CSF concrete to freeze/thaw treatment with deicing salts has been reported by Sorensen.'^^ The test was performed according to RILEM Recommenda tion CDC 2,'^'' on virgin specimens as well as on com panion specimens pre-dried at 4 5 ° C for 14 days. The results showed that the pre-drying treatment had a detrimental effect, particularly on the control mixes. Air entrainment was in general very beneficial, but a concrete containing 10% CSF with W/C + S = 0 - 3 8 was resist ant without any air entrainment. This isolated result should not be understood as final, since mixes of equivalent com position have later proved less frost-resistant. Lehtonen'^^'^'' performed salt scaling experiments on air-entrained concretes with 0—15% of the cement replaced by CSF. All concretes had a W/C + S ratio of 0 - 4 5 . The results showed clear improvement by increas ing CSF dosage. Beyond the RILEM Standard 28 cycles, the trend for CSF concrete was an increased rate of scal ing. However, at 56 cycles the scaling of all CSF con cretes was well below that of the control mixes. Virtanen'^'' investigated the influence of CSF, fly ash, and slag replacement for cement on frost resistance. All mixes were both with and without air entrainment, and designed to give approximately equal compressive strength (30—40 MPa). Pore structure and air void characteristics were determined, as well as parameters from frost resistance and salt scaling experiments. Virtanen concluded that 'Condensed silica fume addition improves freeze/thaw resistance compared with cement concretes having the same strength and air content'. A further series of experiments'^^ focused on develop ing highly durable concrete for edge beams designed for road bridges. The mixes were based on a cement content of about 400 kg/m^ with 1:1 replacement of cement with CSF, and CSF dosages of 0, 4, 8, and 16%. Both superplasticizers and air-entraining agents were used. The 28-day compressive strengths were in the range 42—56 MPa. The freeze/thaw testing consisted of alter nately dipping 100 mm cubes in water at 20°C (16 h) and then freezing in a salt solution at — 15°C (8 h). The Fin nish standard requires that the volume must not decrease by more than 5% after 25 cycles. The result in all cases was that the scaling was reduced with increasing CSF con tent — in spite of decreased air content with increasing CSF dosage (Fig. 26). Rasmussen'^^ reported on a large long-term field exposure test programme on air-entrained concretes with several types of cement and varying CSF dosages. The interim report is on laboratory salt-frost testing (similar to ISO 4846'^^) on samples similar to those exposed in the field. The mixes were divided into three classes accord ing to cement content; in CSF mixes 10% dosage was used and cement was replaced by CSF in a ratio of 3:1. The poorest performance was for mixes with blended cements (25 % fly ash) and CSF (mixes with cement content of 200 and 260 kg/m^). This is probably because pre-curing consisted only of 14 days in water then 14 days in air; 23
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
C S F IN
CONCRETE
N o . of f r e e z e / t h a w c y c l e s (a)
CSF:% (b)
Fig. 26. Volume decrease: (a) as a function of freeze/thaw cycles during salt-frost testing; (b) as a Junction of CSF con tent after 100 cycles, for concretes with varying air-entraining admixture (AEA) content. (From ref. 135).
this is not enough time for the fly ash to develop its poten tial. The richer mixes (cement content 300—400 kg/m^) all performed well. One disturbing aspect of the results is the lack of repeatability of the freeze/thaw testing. In a second test, much less scaling was recorded for all samples relative to the first, presumably identical, test. The internal rank ing between the different mixes in the series was the same, however. This highlights the reproducibility problems associated with present standardized test methods. In their present stage of development they can apparently, at their best, be expected to produce a ranking order among dif ferent samples tested at the same time. Testing against absolute criteria is obviously hazardous. The Technological Institute, Department of Building Technology, Copenhagen, Denmark, has investigated drilled cores from a number of structures with CSF con crete as part of a quality control programme. Several methods were used to assess frost resistance, including a dilation method (equivalent to ASTM C671 '^^) and salt scaling tests. Christensen^^ concluded that 'CSF gener ally improves the frost durability of concrete as measured by frost resistance testing'. Recent work has further complicated the picture regard ing the role of CSF in the frost resistance of concrete. Petersson'^^ measured salt scaling of mortar mixes without air entrainment with water/binder-ratio of 0 • 35 (Fig. 27). A reference mix showed gradual scaling as expected, while a mix containing 19% CSF by weight of cement gave essentially no scaling up to 110 cycles, after which time the specimens disintegrated rapidly. Similar behaviour was observed for a mix with a water/binderratio of 0 • 54 and a CSF dosage of 11 % after 60 cycles. Petersson described two typical stages in salt-frost attack: the first scaling starts in the first few cycles and then levels off for good quality reference concrete; a second stage starts later, and may be caused by an increase in the degree of saturation over a critical level. The second stage appears to be more typical for CSF concrete, and leads to much more severe deterioration. Petersson^^^ also compared the effect of air-entraining agents alone with their effect when combined with waterreducing agents. At equal total air contents, the salt scal ing was much higher for the admixture combination, presumably because of a coarser air void system. The con crete did not contain CSF, but admixture combination is a factor to be aware of in connection with CSF concrete because it normally contains higher amounts of waterreducing agents. Several recent Canadian reports give apparently con flicting results. Hooton^^^ tested non-air-entrained con crete with water/binder-ratio of 0-35 with 0—20% cement replacement by CSF, according to ASTM C666 (A).'^^ All the CSF mixes gave superior performance up to 900 cycles. The control mix failed after 58 cycles. P i g e o n s u m m e d up extensive work at Laval Univer sity, both measuring salt scaling (ASTM C672'^^) and resistance to internal cracking (ASTM C666^^^). The approach was based on determining a critical spacing fac tor for a given mix, by testing a number of parallel mixes differing only in air void spacing factors. The resistance to internal cracking was found to be reduced for mixes containing 10% CSF and water/binder-ratios of 0-3 and 0 - 5 , relative to controls. The same conclusion was drawn for the salt scaling resistance of CSF concrete containing up to 10% CSF by weight of cement. The nature of the
24
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
DURABILITY
100 200 N u m b e r of f r e e z e / t h a w c y c l e s
Fig. 27. Scaling resistance curves (loss of mass) for concretes without (a) and with (b and c) CSF as functions of the number of freeze-thaw cycles. (From ref 157).
scaling was noted to be different for the two types of con crete: control mixes scale in discrete spots, while CSF mixes scale uniformly over the entire surface area. Aitcin and Pigeon'^' reported on field performance of CSF concrete exposed to de-icing salts. They concluded that CSF concretes perform as well as normal concretes, in spite of generally higher W/C ratios, provided that the air void spacing factor is less than 0-20—0-25 mm. This can be obtained with 'no special difficulty'. Bilodeau and C a r e t t e t e s t e d mixes with 0 and 8% CSF with water/binder-ratios from 0 - 4 0 to 0 - 6 4 , both for internal cracking (ASTM C666'^^) and salt scaling (ASTM C672'^^). All concretes were air-entrained with spacing factors of below 0 - 2 0 mm. The concretes with water/binder ratios of less than 0 • 60 performed well, with a tendency for CSF mixes to be more susceptible to scal ing. A number of tests were carried out, with curing and moisture treatment differing from the standard ASTM pro cedures. It was clear that such treatment has significant influence on the test results — requiring further investiga tion, according to the authors. The possibility of producing durable concrete without air entrainment has been investigated by Malhotra et aO^^ Water/binder ratios from 0 - 2 5 to 0 - 3 6 with CSF doses of 0—20% were used, both with and without air entrainment. Testing was according to ASTM C666.'^^ In direct contrast to the results by Hooton^^^ cited above, all mixes without air entrainment failed the test, with durability factors below 12. Air-entrained concretes con taining 10 and 20% CSF also failed to complete 300 cycles, probably because of air void spacing factors well over 0 - 2 0 mm. 7.2.4 Conclusions As is evident from the above, the situation today is unclear. Even when nominally identical procedures are applied to non-air-entrained, high-strength concrete, the results are contradictory.^^^'^^^ It is likely that for highstrength concrete other factors enter the picture, such as dispersion of the CSF, homogeneity of the binder phase, existence of microcracks, and bond to aggregates. These factors are not normally investigated, but they may be im
portant for high-strength concrete performance. There is evidence of reduced frost resistance relative to the high strength level, particularly for high-strength concrete with high CSF dosages ( > 15% by weight of cement). On the other hand, there is also evidence that low W/C ratio con crete with CSF has such fine pore structure that no ice forms above — 20°C, hence producing frost-resistant con crete without air entrainment. The question is open, and its solution may require other methods than W/C ratio, CSF dosage and compressive strength to characterize con crete structure. For normal-strength air-entrained concrete there is a difference between Canadian and Scandinavian laboratory results. Scandinavian results indicate improved frost resistance with the use of CSF, wheras the Canadian results indicate the opposite. Part of this conflict in results may be caused by the test methods used. The Canadian results are based on strict adherence to ASTM procedures, whereas in Scandinavia a variety of methods have been applied. It is clear that care must be taken in applying stan dard methods to CSF concrete. Curing time and moisture history are important to field performance, but these fac tors are fixed in the ASTM procedures. For example, as part of a round-robin test,^^"^ a con crete with a W/C ratio of 0 • 45 and about 5 % air content was found to give minimal salt scaling. The same con crete disintegrated quickly when it was dried at 50°C and placed in water one week prior to testing. On the other hand, a concrete with water/binder ratio of 0 - 3 5 , 7% CSF and 6% air appeared to perform well after such treatment. This illustrates the need to develop new test procedures that take into account factors known to be important to field performance. Experience of field performance of CSF concrete to date has not indicated any particular problem with frost resistance. At the present time it is clear that a proper air entrain ment system is necessary to protect normal-strength con crete, with or without CSF, from frost attack. 7.3 Chemical resistance 7.3.1 Leaching and efflorescence Efflorescence frequently occurs on concrete surfaces exposed to wetting-drying or to percolation of water through the concrete. The main cause is usually leaching of calcium hydroxide, which carbonates on the surface. Efflorescence is mainly an aesthetic problem, but if exten sive leaching of lime takes place in the concrete the porosi ty is increased, with decreased strength and durability as a consequence. Samuelsson'^^ measured the leaching of lime from mortar surfaces with 0, 5, and 10% CSF addition at fixed W/C ratios. Different pre-curing times and relative humidities were tested. He concluded that the most effi cient preventive measure was to pre-cure the concrete at 8 0 - 9 5 % RH for several days prior to exposure to leaching water. This allows hydration to proceed and carbonation to take place in the pore system. The effect of CSF was to reduce leaching, providing that the pre-curing period was at least 4 days in a moist atmosphere. Curing periods beyond 4 days were not tested. One might expect more effect from CSF addition if time is allowed for the poz zolanic reaction to go further, thereby reducing the free lime content and providing a finer pore structure. For concrete submerged in water, the leaching of lime is a major weakening factor. Carlsen and Vennesland'^^ made cement pastes with sulphate-resistant and rapid25
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
C S F IN
CONCRETE
hardening Portland cements, and 0, 5, and 10% CSF. The W/C -f- S ratio was kept constant at 0 - 5 0 . Mature 12-7 mm^ cube specimens were exposed to sea water for 540 days. All specimens showed strength reductions of about 50 %. The authors gave no explanation for this sur prising result. 7.3.2 Sulphate resistance The first published report on CSF in concrete^"^^ con tained data on sulphate resistance of concrete stored in a 10% sodium sulphate solution. Bernhardt concluded that the sulphate resistance was improved when 10—15% of the cement was replaced by CSF, but he stressed that the duration of the tests was too short to allow firm conclu sions to be drawn. In 1952 a large number of concrete specimens were placed in a tunnel in Oslo's alum shale region for longterm tests. The groundwater contains up to 4 g/1 SO3 and the pH varies from neutral to 2 -5. Two reports have been published so far, one'^^ on the 12-year results, and one*^^ on the 20-year results. The test specimens were produced with a variety of cements and additives, in cluding one set where 15% of ordinary Portland cement was replaced by CSF. All mixes had W/C ratios of about 0 - 5 0 , except for the CSF mix which had a higher water demand, and therefore a W/C + S ratio of 0 • 62. Damage was assessed by measuring volume reductions. After 20 years of exposure the most resistant mixes were those with sulphate-resistant cements and the CSF mix, which all per formed approximately equally (Fig. 28). Reasons for the good performance of the CSF mix in clude: the refined pore structure, and therefore reduced transport rate of harmful ions,^^ the lower calcium hydroxide content, and the increased amount of aluminium incorporated in the CSF, reducing the amount of alumina available for ettringite production.^ Inspection of the test samples after 30 years' exposure confirmed the results cited above. Laboratory investigations confirm the data from the field tests. Mather^^^ measured expansion in a sulphate solu-
10 Exposure: years
Fig. 28 Volume reduction of100 X 100 X 400 mm concrete prisms stored for 20 years in acidic sulphate-rich water in the Oslo alum-shale region. (From ref. 5, adapted from ref. 168).
tion of mortar prisms where 30% of the cement was replaced by an equal volume of various pozzolans. Three types of cements were used, and the CSF in combination with all the three cements proved to be the most efficient pozzolana in preventing expansion. Carlsen and Vennesland'^^ measured strength reduc tions of cement pastes after exposure to a sodium sulphate solution. They found good performance for a paste made with sulphate-resistant cement and one made with rapidhardening Portland cement containing 5% CSF, while the control with RHPC was destroyed quite quickly. Popovic et al.^^^ used OPC, a blended cement with 20% slag, and a blended cement with 15% natural poz zolana to produce control mortars and mortars where 15% CSF was added alone, and together with a superplasticizer to compensate for the increased water demand in CSF mixes. After 28 days of water curing the small prisms (25 X 25 X 160 mm) were exposed to a 10% ammonium sulphate solution. They concluded that sulphate corrosion is predominant in this solution, and that CSF prevents this type of corrosion as well as acid corrosion. The SteineggerKoch test in sodium sulphate solution 'confirmed this con clusion completely, and the fact that ordinary and blended Portland cement with admixture of silica fume exhibit bet ter durability than special sulphate-resisting cement'. It is interesting to note that in these results CSF was equal ly as effective with and without superplasticizers; i.e. at different total porosities CSF inhibits ammonium sulphate corrosion, which implies that chemical effects of CSF are more important than reduced permeability in this regard. M e h t a c o m p a r e d the resistance to chemical aggressives of concrete with OPC, latex-modified concrete, and a mix containing 15% CSF. The W/C -h S ratio of the three mixes was about 0 - 3 3 . Curing was in air for the latex mix; the two others were wet-cured for 7 days. After one week all mixes were air-cured for 6 weeks prior to exposure to the aggressives. Two sulphate solutions were used: 5% ammonium sulphate and 5% sodium sulphate. The failure criteria was the amount of time samples needed to suffer a 25% weight loss. Ammonium sulphate was equally destructive to the control and the CSF mix, while the latex mix suffered less weight loss. Mehta attributed this to the ability of ammonium sulphate to decompose CSH, while the latex coating delays the decomposition. Note that this result is in contrast to the one found by Popovic et al.^^^ where the CSF mix performed better than the control in a more concentrated solution. None of the three mixes decomposed in the sodium sulphate solu tion, a result M e h t a a t t r i b u t e d to the low W/C ratios. Hooton^^^ tested mortars with a W/C ratio of 0-49 and 10 and 20% cement replacement by CSF, according to ASTM C1012:''^^ immersion in a 5% sodium sulphate solufion. The control failed quickly, while the CSF mor tars expanded less than a mortar made with sulphateresistant cement (Fig. 29). 7.3.3 Alkali-aggregate reactions It is well known that reactive pozzolans can be used to control the expansions associated with the alkaliaggregate reaction. Pore-water analysis of CSF cement paste^^'^^ demonstrated the ability of CSF to reduce the alkali concentrations in the pore water quite rapidly, thus making it unavailable for the slower reaction with reac tive silica in the aggregates. Asgeirsson and Gudmundsson^^"* used CSF with high alkali Icelandic cements and reactive sands in mortar bar
26
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
DURABILITY 0-24-
0-20-
0-30
r
0-20
h
OPC
0-15 P r o p o s e d A S T M limits . . T y p e II
TypeV.
010
005
'OPC + 2 0 % C S F + \ superplasticizer \—-^-"'^''^ c
o
90
180
270
365 X LU
T i m e in 5 % sodium sulphate solution: d a y s
Fig. 29. ASTM CI 012 sulphate resistance of CSF mortar (From ref 158).
bars.
tests, and demonstrated the abihty ot CSF to reduce expan sions (Fig. 30). Olafsson'^^ reported further Icelandic experiences on blended cements made with CSF. The presence of salts (from sea-dredged aggregate) was generally found to in crease the alkali-aggregate expansions. CSF contents of 5 and 15% were found to reduce the expansions, both with and without the presence of salts. A characteristic of the expansions in the specimens containing CSF was that they took place over a long period of time. Short-term tests (14 days) are therefore not suitable to evaluate the effect of CSF. Expansions were found to decrease with decreas ing lime/silica ratios in the blended cements. This obser vation has been used to determine the necessary CSF dosage to prevent harmful expansions. All Icelandic cements are currently blends containing 7-5% CSF. Field experience with Icelandic CSF-blended cement has recently been reported by Sveinbjorsson.'^^ A total of 200 houses constructed in the period 1979—1986 were selected at random. Visual inspection was performed and cores drilled for microscopic examination before and after one month of storage in 'a container with favourable con ditions for alkali-aggregate reactions'. None of the samples was considered to have symptoms of alkali-aggregate damage, and the conclusion was that the use of CSF had been successful in preventing destructive alkali-aggregate expansion in concrete. Oberholster and Westra'^^ tested a number of mineral admixtures for their efficiency in reducing alkali-aggregate expansion in mortars containing high alkali cements. CSF was found to be the most efficient. The reactivity of several Scandinavian sands was tested by means of various methods in a Nordtest project. The prime purpose of the work was to compare test methods; however, a number of mortars clearly demonstrated the ability of CSF to reduce expansion. Perry and Gillott'^^ reported work on CSF and alkalisilica reactions. They measured mortar bar expansions at 23°C, 38°C and 51°C by using a pessimum amount of reactive opal aggregate and blended cements. They found that CSF reduced the total expansions considerably, but also found, as did Olafsson,'^^ that the expansions were delayed for CSF mixes. A CSF dosage of 5% increased expansion, and Perry and Gillott concluded that probably as much as 20% of the cement needs to be replaced by CSF to suppress effectively the expansion with their very reactive aggregates. A new aspect of their work was the
O-IOh
Time: months
Fig. 30. Expansion of mortar prisms made with high-alkali cement, reactive sand, and three CSF dosages. (From ref. 5, based on data from ref. 174).
finding that the use of a superplasticizer dramatically in creased expansion relative to identical mixes without superplasticizer. This was particularly evident for CSF mixes, and could not be explained by the alkalis con tributed by the superplasticizer. They also concluded that CSF was inefficient in suppressing expansion due to alkalicarbonate reaction, although some reduction was observed. Gillott'^^ has recently reviewed the effects of CSF on alkali-aggregate reactions. Nilsson and Peterson'^' studied the relationship be tween moisture state and 'pop-outs' caused by alkali-silica reactions. The use of CSF as an inhibitor was also tested, and found efficient in preventing pop-outs at a 5 % dosage level. A CSF dosage of 10 or 15% was needed to pre vent expansion. CSF from two different sources was used both by Perry and Gillott'^^ and by Nilsson and Peterson.'^' It is noteworthy that in both cases one source proved significantly more effective than the other. The effect of CSF as an inhibitor against pop-outs has also been investigated by P e t e r s o n . H e was unable to draw any firm conclusions from his results regarding popouts, but found that fine aggregates were not protected by as much as 10% CSF dosage. The investigation by Popovic et al.^^^ included a study of the efficiency of CSF in preventing alkali-silica expan sion in concrete. Using 23% CSF dosage by weight of a cement containing 0-8% total alkalis and Pyrex glass aggregate (ASTM 441'^^), no expansion was observed. Aitcin and Regourd'^"* reported on a field test followup after 3 years on a series of concretes with very reac tive aggregates and alkali-rich cement. Lean mixes with high CSF dosages ( 2 0 - 4 0 % ) showed no trace of gel formation, while traces were found in richer mixes with 15% CSF dosage. The alkali-aggregate reaction was con sidered to be 'under control' in all mixes. 27
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
C S F IN
CONCRETE
7.3.4 Other chemicals Improved resistance of concrete to a large number of chemically aggressive agents, including nitrates, chlorides, sulphates, and acids, is reported in a technical brochure from Norcem Cement. The concrete described generally has a higher CSF dosage than for normal CSF concrete, as the concrete is intended for special applications. Popovic et al^'^^ found that 15% CSF was efficient in 'preventing' corrosion of OPC mortars in 5% sulphuric acid solution and 5 % ammonium nitrate fertilizer solu tion. 'Preventing' in this connection means strength decrease relative to a control exposed to the same aggres sives — not relative to a control in water. CSF also greatly improved the resistance to a 10% ammonium nitrate solu tion. The performance of CSF with OPC, a 20% slag blend, and a 15% natural pozzolan blend was comparable, and reproducible. Mehta'^^ compared the response of a control concrete, a latex modified concrete, and a 15% CSF concrete to exposure to 1 % hydrochloric acid solution, 5% acetic acid solution, 1% lactic acid solution, and 1% sulphuric acid solution respectively. All concretes had W/C ratios of about 0-33. The results showed that the CSF mix generally had the best resistance to chemical attack. Feldman and C h e n g - y i m e a s u r e d the resistance of mortars to a 4% MgCl2 solution for 150—170 days, followed by exposure to a mixed chloride solution of 27-5% CaCl2, 3-9% MgCls, 1-2% NaCl, and 2 - 1 %
8.
NaHC03. The mortars had W/C + S ratios of 0-45 and 0 - 6 0 with 0, 10, and 30% dosage of CSF. Curing times before exposure were 7 and 28 days. The investigation included measurements of mercury intrusion, stiffness, calcium hydroxide contents, and non-evaporable water contents before and after exposure. The authors concluded that CSF and long curing time prior to exposure substan tially increased the resistance of the mortars. They attrib uted this effect mainly to inherent lowered permeability, but also to the reactions between excess CSF, low calcium CSH, unhydrated cement, and the salt solution which results in reduced total pore volume, and thereby reduced permeability. 7.3.5 Conclusions Many investigations as well as practical experience have indicated that a major potential advantage of CSF in con crete is to improve chemical resistance. Sulphate resistance and protection against alkali-aggregate reactions are two areas of particular promise. Recent reports indicate that protection may be increased against a variety of chemical aggressives. The reasons for the generally good performance of CSF concrete in chemically-aggressive environments include: (a) refined pore structure, and therefore reduced transfer rates of harmful ions; (b) reduced content of calcium hydroxide; and (c) lower C/S ratio of the reaction pro ducts, which increases the capacity to incorporate foreign ions such as aluminium or alkalis in the lattice.
CORROSION OF REINFORCEMENT
The corrosion process of reinforcement steel in con crete may be divided into two stages: the initiation stage and the propagation stage. Steel in concrete is normally in a passive state with respect to corrosion, because of the high pH value in the pore water. The passive iron oxide layer that normally protects the steel is destroyed when the pH value is reduced below about 10—11, or when chloride ions reach the steel surface even at higher pH values. The rate of corrosion, once the passive iron oxide layer is destroyed, depends on the presence of moisture and oxygen, and on the electrical resistivity of the concrete. 8.1 pH values: pozzolanic reaction and carbonation Scandinavian cements have a relatively high alkali con tent, which results in very high pH values for concrete pore water ( > 13.5). CSF reduces the pore water alkali content much more than the 'dilution' effect when it partly replaces cement on a 1:1 basis. Page and Vennesland^^ found pH values of pore water in mature pastes to be about 13-9, 13-4, 12-9, and 12-0 for 0, 10, 20, and 30% ce ment replacement levels respectively (Fig. 31). Thus, 30% replacement of cement by CSF is needed before the pH value drops below that of a saturated calcium hydroxide solution — approximately 12-5. Diamond^^ determined a pH value of 12-2 after 145 days' hydration for a paste with 30% of the cement replaced by CSF. In later work D i a m o n d r e p o r t e d pore water analyses made 4 and 24 hours after mixing of an equivalent paste, as well as a reference. There are some changes not caused by the dilution effects, but the main conclusion is that CSF has 'litde measureable ef
fect on the major dissolved component species' during the first 24 hours. More time is obviously needed for pro ducing the effects observed after long periods of time. Glasser and Marr^^^ mixed mortars with a low and a high alkali cement {W/C = 0-60) as well as blends where 15% of the cement was replaced by CSF. Pore water was squeezed out and analysed. The pH stabilized after 3 months; for the high alkali cement the values were 13-77 for the reference mix and 13 • 18 for the CSF mix, in line with earlier r e s u l t s . T h e calcium hydroxide content was reduced by more than one-half. CSF had much more effect than any of the fly ashes or natural pozzolana tested, but the reactivity with calcium hydroxide was comparable to a Degussa flame hydrolysis silica. CSF reduced the alkali contents to roughly one-third relative to the control OPC; the soluble sulphate was also markedly reduced to less than one-half, but the chloride concentration was increased somewhat in contrast to all the other mineral additives. The authors stated that it is the low C/S ratio of the CSH formed with CSF present which is responsible for the alkali depledon, either by sorption or by stabilizing alkali-rich compounds. Gautefall and Vennesland'^^ report pore water analysis for cement pastes made with OPC, fly ash blends (10 and 25%) and a slag blend (15%), all with 0 - 1 5 % CSF. The results regarding the ability of CSF to reduce pH values and alkali pore water concentrations are in general agree ment with results already cited, although the values of the reductions depend somewhat on the type of cement. The results cited here for high dosages of CSF are con sistent with thermogravimetric data which indicate that
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:00:50
C O R R O S I O N OF
in excess of 24% CSF by weight of cement is needed to consume all the calcium hydroxide.'^ We conclude that CSF does reduce the pH of concrete pore water for all cement types, but at normal dosage of CSF the reduction is far from sufficient to destroy the passivity of the steel. When concrete carbonates, the pH is reduced enough to depassivate the steel. Carbonation takes place both at the expense of the calcium hydroxide and the calcium oxide in the CSH phase. The fact that CSF concrete contains less calcium hydroxide than does control concrete does not therefore automatically mean that CSF concrete is more vulnerable to carbonation. Other factors also in fluence the rate of carbonation: curing history, moisture state, and CO2 diffusion rates. Because of the complex ity of the problem we consider direct measures of car bonation depths to be more relevant than theoretical predictions based, for example, on amounts of calcium hydroxide. Meland and Traetteberg'^^ measured calcium hydrox ide and calcium carbonate contents in 3—4 mm thick cement paste discs with W/C + 5 = 0 • 50 and 0 and 15 % of the cement replaced by CSF, after various times of exposure to 50% and 100% RH air with 1 and 3% CO2 content. The CSF paste was moist-cured for 21 days prior to exposure, and the control paste was moist-cured for 7 days. The calcium hydroxide contents of the uncarbonated pastes agreed reasonably well with data from Sellevold et al.,^^ taking into account the different cements used. However, other results were of a novel nature: in line with common experience, the control paste carbonated most at 50% RH, but the CSF paste carbonated much more at 100% RH than at 50% RH. Also, between 1 and 6 months substantial carbonation took place in the CSF paste, but the calcium hydroxide content hardly changed. A similar tendency was found for the control paste. It should be noted that the CO2 pressure was much higher than under natural conditions in these tests. We cannot draw any firm conclusions from these experiments, but they point out the need for more basic knowledge on carbonation, and on the influence of reactive pozzolans on cement hydration and the hydration products. Carbonation of concrete is usually measured by the phenolphthalein test on fracture surfaces perpendicular to the surface exposed to carbonation. Johansen, Vennesland and Gjorv,'^'^ and Vennesland'^' made car bonation measurements over periods of up to 3 years for a large number of concrete mixes with and without CSF. Vennesland's results showed that 10% CSF dosage had no influence on the carbonation depth for specimens cured for 7 days in water prior to exposure to laboratory air, when carbonation depth was plotted against the 28-day compressive strength of water-cured specimens. Johansen exposed the specimens direcfly upon demoulding. His results after 3 years show two clear effects: firstly, lack of moist-curing at least doubles the carbonation depth for both types of concrete; and secondly, CSF concrete is even more sensitive to lack of proper curing than normal con crete. Thus, as is well known, proper curing of concrete is essential to avoid excessive carbonation. Johansen'^^ measured carbonation depths on the same mixes after 6 years of exposure, both for specimens exposed directly after demoulding, and after 27 days of water curing (Fig. 32). Bad curing sfill has a major effect on carbonation depths; for mixes of higher strength than 40 MPa at 28 days, carbonation was roughly doubled relative to water-cured companions. The factor was smaller
REINFORCEMENT
i4r
13
12 •
0% CSh
O 10% CSF •
20% CSF 30% CSF
A I
I
10
20
30
40
50
60
70
80
90
Curing time: days
Fig. 31. Influence of CSF content on pH values of pore water squeezed from cement pastes (OPC, W/C + S = 0-50.) (From ref 26).
40 • O X A
X
o
30
O
0%CSF 5% CSF 1 0 % CSF 2 0 % CSF
X
20 h
o 1 0 h
E E
i
0
(a)
30
r X
•
20 h
o
1 0 h
X
o
20
40
60
2 8 - d a y c o m p r e s s i v e strength: M P a (b)
Fig. 32. Carbonation depths of concretes after 6 years of exposure at 20°C and 50% RH. Precuring prior to exposure: (a) 1 day in mould; (b) 1 day in mould, then 27 days in water. (From ref 192).
for weaker mixes, but the carbonation depths were greater. For example, badly-cured 25 MPa control concrete showed a depth of about 32 mm, while a well-cured com panion showed about 22 mm. The effect of CSF (5, 10, and 20%) was small at 5 % dosage, but the trend was clear for both types of exposure when comparison was made on an equal 28-day strength basis: increased CSF dosage led to increased carbonation depths. In order to obtain data from a variety of field condi tions, Maage and Sellevold'^^ took samples from 16 29
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:06:14
C S F IN
CONCRETE
buildings with CSF concrete and 11 controls. The age of the buildings varied from 40 to 80 months, and the con crete in all cases had a nominal characteristic strength of 25 MPa. The measured carbonation depths were adjusted to correspond to an age of 5 years. The result was that there was no significant difference between the mean values of carbonation for the two types of concrete (9—10 mm). The CSF concrete, however, showed much greater variation in carbonation depth than the control con crete. It seems likely that this reflects the greater sensitivity of CSF concrete to early curing conditions, which cer tainly vary considerably between building sites. These field data, as well as controlled laboratory results, are given by Skjolsvold.'^^ CSF concrete of higher quality from structures has also been investigated. Christensen-^^ found no influence of CSF on carbonation depth. Drilled cores from a pier in Gothenburg, Sweden, have been investigated by M a a g e . N o n e of five concretes (three with CSF) with core strengths from 44 to 68 MPa showed carbonation depths above 2 mm after about 7 years of service. The same was true for two 12-year-old concretes, one with CSF (60 MPa) and one control (50 MPa), from a silo cover at Fiskaa Verk, Kristiansand, Norway. Carbonation of concrete in the Middle East for a variety of curing and exposure conditions has been reported by de Fontenay.'^^ Some mixes with 5% CSF were included (Fig. 33). He concluded that CSF used as an addition to improve durability has no influence on the rate of carbonation. Based on the above, it seems as if CSF will increase carbonation at dosages above approximately 5% in badlycured medium-to-low grade concrete. There is no field evidence indicating that carbonation is a problem in high quality concrete, either with or without CSF. Laboratory tests over up to 6 years indicate that for equal strength, CSF concrete will carbonate somewhat deeper than con trol concrete regardless of strength class or exposure condition. 8.2 Chlorides Chlorides may be mixed into fresh concrete (as acceler ator or from aggregates) or may penetrate hardened con crete from external sources (such as deicing salts or sea water). In either case, only a proportion of the chloride ions will be available, and thus presumably aggressive in the pore water. The capacity of a concrete to bind
W/C+S = 0 - 7 8
WIC = 0 - 4 6
x-^'v lA//C = 0 - 5 0
I V / C + S = 0-51 SRPC SRPC+CSF
12 18 T i m e of e x p o s u r e : m o n t h s
24
Fig. 33. Carbonation depths in concrete containing sulphateresistant cement, both with and without CSF. (From ref. 195).
chlorides is therefore of interest, as well as the resistance it offers to the penetration of chlorides. Page and Vennesland^^ added chlorides to cement-CSF pastes, and later squeezed out pore water for analysis. Their results showed that as the CSF dosage increased, the fraction of the chlorides available in the pore water also increased. They suggested this effect to be caused by the lowered pH value, which increases the solubility and thereby reduces the quantity of Friedel's salt. The latter, a calcium chloro-aluminate-hydrate, is thought to be the dominant factor in binding chloride ions during cement hydration. The lower pH of CSF concrete also indicates that the threshold concentration of chlorides in the pore water, which will induce steel depassivation, is reduced. The hypothesis that there is a critical ratio of chloride to hydroxyl concentrations in concrete pore water with respect to inducing depassivation of steel was followed up in further work by Page and Havdahl,'^^ by electro chemical monitoring of the corrosion behaviour of steel electrodes in CSF-cement pastes of similar compositions to those studied e a r l i e r . T h e y concluded that the free chloride/hydroxyl ratio alone was not a reliable index for comparing the corrosiveness of hardened pastes with vary ing CSF dosages. Repassivation took place in some pastes after some initial corrosion had occurred, presumably because of insufficient mobility of chloride ions in the dense pore structure of CSF cement paste. The results thus indicate that acceptable chloride limits are lower for CSF concrete than for reference OPC concrete, but further work is required on exposure for longer periods of time, together with data on chloride diffusion. Monteiro et al. '^^ studied the steel-cement paste inter face for a reference mix and for one where 16% of the cement had been replaced by CSF. A 2% dosage of calcium chloride was added to the fresh mixes. After storage in a fog room for 180 days, the pastes were dried. This led to separation of the steel from the matrix. The interfacial zone was examined by SEM. Both mixes showed an interfacial film of large calcium hydroxide crystals, but only the CSF mix showed visual evidence of corrosion. The authors attributed this to a higher chloride-hydroxyl ratio in the pore water for the CSF mix. However, in light of the work by Page and Havdahl'^^ discussed above, it is not known if repassivation had occur red. Note that the chloride dosage was twice as high in ref 197 as in ref 196. Penetration of chlorides into CSF concrete from sea water has been studied by Fisher et al. With CSF used as additive, they found that it considerably reduced the diffusion coefficient of chlorides in concrete. For exam ple, a concrete with 8% CSF added had a chloride diffu sion coefficient of 1 • 1 x 10"^ cm^/s while an equivalent mix without CSF had a value of 1-5 X 10-^ cm^/s. Chloride penetration in concrete has been measured with somewhat inconclusive results.'^' Penetration from sea water into pastes was found to decrease markedly when 5 - 1 5 % of the cement was replaced by CSF'^^ (Fig. 34). The diffusion of chlorides through discs of hardened cement pastes has been measured by Gautefall.^^ W/C + S ratios of 0 - 5 , 0 - 7 , and 0 - 9 were used with CSF dosages of 5, 10, and 15%. OPC and a blended cement with 10% fly ash were used. The diffusion coefficients for blended cement pastes were 30—50% lower than for OPC pastes. The replacement of cement by CSF led to
30
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:06:14
C O R R O S I O N OF
marked reductions in the diffusion coefficient, particularly for OPC pastes. For example, 10% CSF led to reductions of 68—84% in the diffusion coefficients of OPC pastes. Byfors,^^' using the same technique, also found that replacement of cement by CSF in cement pastes led to substantial reduction in the chloride diffusivity. Based on available data, she also made a theoretical estimate of initiation time for diffusion-controlled chloride-initiated reinforcement corrosion, and concluded that both CSF and fly ash extended this time. For concrete bridge decks, the US Federal Highway Administration has described a rapid chloride permeability test. Christensen et al applied the test to compare a high-strength CSF concrete with a latex-modified concrete and a low-slump dense concrete. The CSF obtained a 'very low' permeability rating, while the two other mixes were rated as 'moderate'. Chloride penetration into concrete with 5—30% cement replaced by CSF has been measured by Marusin,^^^: penetration was reduced at all CSF dosage levels. 8.3 Rate of corrosion Vennesland and Gjorv'^"* and Vennesland'^' measured oxygen diffusion and electrical resistivity of concrete mixes with and without CSF. The oxygen difftision did not show any systematic dependence on CSF dosage or waterreducing agent content for water-saturated specimens. The electrical resistivity of CSF concretes increased significandy to exceed resistivity of controls, presumably because of both lower ion concentrations and the refined pore structure. Electrical resistivities of dry and wet-cured concretes made with four cement types and CSF dosages from 0 to 15% have been reported by Gautefall and V e n n e s l a n d . T h e dry-cured samples all had such a high resistance that no corrosion due to macrocell effects was possible. Wet-cured concrete generally showed only small decreases in resistivity with increasing W/C + S ratio for a given cement type and CSF dosage. CSF pro duced a substantial increase in resistance for all cement types (Fig. 35). Fisher et al. '^^ measured rates of corrosion in control concrete and in concrete with CSF addition. They found that the corrosion rate was about the same in the two con cretes for equal total chloride concentrations. They also found the resistivity of the concrete to be considerably increased on addition of CSF.
REINFORCEMENT
Preece et al.^^^ studied the electrochemical behaviour of steel in dense CSF cement mortars. They found that the mortar provided a high degree of corrosion protec tion to the steel, and attributed this to: (a) a very low W/C -h S ratio, resulting in a very fine pore structure which apparently limits the access of water; and (b) very high electrical resistivity, which limits the galvanic current even in the absence of passivity of the steel. Biirge^^^ has reported that the use of a very dense polymer-modified cement-CSF mortar with corrosion in hibitors added is effective as a protective coating for steel in corrosive environments. Cracks in concrete may result in the formation of con centrated anodic areas at the exposed steel, and conse quently high corrosion rates. The ability of concrete to 'self-heal' the cracks is important, to avoid the corrosion problem. Gautefall and Vennesland^^^ investigated the self-healing capacity of concretes with and without CSF in sea water with access to oxygen, simulating a splashzone situation. They found no significant difference be tween the two types of concrete. 8.4 Conclusions As discussed above, the individual factors controlling the corrosion of steel in concrete are known and may be investigated. The present evidence indicates, for concretes of equal strength, that: (a) in regard to chloride-induced corrosion, the use of CSF will extend the initiation time; and (b) in regard to carbonation-induced corrosion of lowto-medium grade concrete, CSF may shorten the initia tion fime. For quality concrete with compressive strength exceeding about 40 MPa, carbonation is not generally regarded as a problem. However, in practical situations it is the combination of these factors which governs the risk of corrosion, and information on the individual factors is not sufficient to allow a direct prediction of the corrosion protection offered by different concrete mixes. Present evidence suggests that the use of CSF as an addition to improve concrete durability will also improve the concrete's ability to pro tect embedded steel from corrosion. It should also be noted that the primary factors govern ing corrosion protection of steel in concrete are probably the quality of the initial curing and that the concrete cover over the reinforcing steel is sufficient. o A
A 0 •
20 50r
V
SP30 SP30 + 1 0 % C S F MP30(10%fly ash) M P 3 0 ( 1 0 % fly ash) + 5% CSF MP30 (10% fly a s h ) + 10% CSF MP30 (25% fly a s h ) MP30 (25% fly a s h ) + 5% CSF MP30 (25% fly a s h ) + 10% CSF
E 40 CSF
!io
^^30 >
o
' 20
0
5
10
15
20
25
5
10
15
20
Penetration depth: m m
Penetration depth: m m
(a)
(b)
Fig. 34 Chloride (OPC; (a) W/C + S were exposed to sea as a percentage by
25
penetration into hardened cement paste = 0-50; (b) W/C + S = 0- 70). Specimens water for 6 months. Chloride content given weight of paste. (From ref J 99).
lOh
0
0-5
0-6
0-7
0-8
0-9
10
M
1-2
1-3
W/C + S
Fig. 35. Electrical resistivity of concrete with different cement types, W/C + S ratios and CSF dosages. Mature water-cured specimens. (From ref 189). 31
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:06:14
CSF IN C O N C R E T E
9.
NATIONAL STANDARDS, CODES, AND RECOMMENDATIONS
Arabian Gulf, Rasheeduzzafar et al. Proposal for a code of practice to ensure durability of concrete construction in the Arabian Gulf environment. King Fahd University, Saudi Arabia.
Norway. NS 3098. Portland cements, specification of properties, sampling and delivery. NS 3420. Specification texts for building and construc tion., NBR, Oslo.
Canada. CAN/CSA-A23.5-M86. Supplementary ing materials.
cement
Sweden. Statens PlanverkPFS 1985: 2. Mineral additions to concrete, approval rules. Statens Planverk, Stockholm.
Denmark. DS 4 U . Code of practice for the structural use of concrete. The Danish Academy of Technical Sciences Hefte no. 25. Basic Concrete Specification. Copenhagen, May 1986.
UK. British Board of Agrement. Agrement Certificate No. 85/1568. USA. ACI 266-3 Detroit.
Draft.
American Concrete Institute,
Finland. Suomen Betoniyhdistys r.y. Betoninormit 1987, RakMK BY, by 15, Helsinki.
10.
REFERENCES
1. Heggestad et al. Investigation of Elkem microsilica by t h e r m o s o n i m e t r y . Thermochimia Acta, 1984, 72, 202-212. 2. Jahr J. Possible health hazard from different types of amorphous silicas — suggested threshold limit values. Health effects of synthetic silica particulates. Edited by D a n n o m D . American Society for Testing of Materials, 04 732000-17. 3. Glomme J. Risikoen for stovlungesykdom i ferrosilisiumsmelteverk. Del II: Dyreeksperimentelle undersokelser. Arbeidsforskningsinstituttene, O s l o , N o r w a y , 1964. (In Norwegian). 4. American Conference of Governmental Industrial Hygenists. Threshold limit values for chemical substances in the work environment adopted by ACGIHfor 1984—85. A C G I H , Cincinatti, 1984, 3 9 . 5. Hjorth L. Microsilica in Concrete. Nordic Concrete Federation, Oslo, 1982, Publication N o . 1. 6. Regourd M . Condensed silica fume. Edited by Aitcin P . C . Universite de Sherbrooke, Q u e b e c , 1983, 2 0 - 2 4 . 7. Regourd M . Microstructure of high strength cement paste systems. Mat. Res. Soc. Symp. Proc, 4 2 , 1985, 3 - 1 7 . 8. Regourd M . et al. Hydraulic reactivity of various pozzolanas. Proc. 5th Int. Symp. on Concrete TechnoL, Monterrey, M a r . 1 9 8 1 . 9. Regourd M . et al. Microstructure of field concretes con taining silica fume. Proc. 4th Int. Conf on Cement Microscopy. Las Vegas, M a r . 1 9 8 3 . 10. Sellevold E.}. et al. Silica fume — cement pastes: hydra tion and pore structure. N o r w e g i a n Institute of Technology, T r o n d h e i m , 1982, Report B M L 8 2 . 6 1 0 , 19-50. 11. Huang Cheng-yi and Feldman R . F . Hydration reactions in Portland cement — silica fume blends. Cement and Concrete Res., 1985, 15, 5 8 5 - 5 9 2 . 12. Huang Cheng-yi and Feldman R . F . Influence of silica fume on the microstructural development in cement mor tars. Cement and Concrete Res., 1985, 15, 2 8 5 - 2 9 4 . 13. Traetteberg A. Silica fume as a pozzolanic material. // Cemento, 1978, 7 5 , N o . 3 , 3 6 9 - 3 7 6 . 14. Scheetz B . E . et al. Effect of composition of additives upon microstructures of hydrated cement composites. Proc. 3rd Int. Conf. on Cement Microscopy, Houston, Mar. 1981, 3 0 7 - 3 1 8 . 15. Chatterji N . et al. Pozzolanic activity of byproduct silicafume from ferro-silicon production. Cement and Concrete Res., 1982, 12, 7 8 1 - 7 8 4 . 16. Traetteberg A. Silikastov i fabrikkbetong. Vurdering av hydratasjonsforlop. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1979, Report STF65
A 7 9 0 1 4 . (In Norwegian). 17. Chatterji S. et al. Pozzolanic property of natural and syn thetic pozzolanas: a comparative study. Publication SP-79. American Concrete Institute, 1983, I, 2 2 1 - 2 3 3 . 18. Zhao-qi W u and Young J. F . T h e hydration of tricalcium silicate in the presence of colloidal silica. J. Mat. Sci., 1984, 19, 3 4 7 7 - 3 4 8 6 . 19. Halse Y. et al. Calorimetry and microscopy of fly ash and silica fume cement blends. Brit. Ceramic Proc, 1984, 35, 4 0 3 - 4 1 7 . 20. Markestad A. Silikastov i fabrikkbetong. Vurdering av hydratasjonsforlop. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1980, Report STF65 A 8 0 0 6 5 . (In Norwegian). 2 1 . Buck A . D . and Burkes J . P . Characterization and reac tivity of silica fume. U S Army Engineer Waterways Experiment Station, Structures Laboratory, Vicksburg, 1981, Miscellaneous paper SL-81-13. 22. Grutzeck M . W . et al. Mechanism of hydration of con densed silica fume in calcium hydroxide solutions. Publication SP-79. American Concrete Institute, 1983, II, 6 4 3 - 6 6 4 . 2 3 . Grutzeck M . W . et al. Mechanism of hydration of Portland cement composites containing ferrosilicon dust. Proc. 4th Int. Conf on Cement Microscopy, Las Vegas, M a r . 1982, 1 9 3 - 2 0 1 . 24. Kurbus B . et al. Reactivity of S i 0 2 fume from ferro silicon production with C a ( 0 H ) 2 under hydrothermal conditions. Cement and Concrete Res., 1985, 15, 134-140. 2 5 . Diamond S. Effect of microsilica (silica fume) on pore solution chemistry of cement pastes. J. Am. Ceramic Soc., 1983, 6 6 , M a y , N o . 5. 26. Page C . L . and Vennesland 0 . Pore solution composition and chloride binding capacity of silica-fume cement pastes. Mater, and Structs, 1983, 16, N o . 9 1 , 1 9 - 2 5 . 2 7 . Diamond S. Scientific basis for the use of microsilica in concrete. Presented at First S e m . on Elkem Microsilica Technology, Elkem Chemicals, Sao Paulo, Sept. 1984. 2 8 . Meland I. Hydratasjonsvarme, hydratasjonsforlop og reaktivitet i flygeaskesement og PC sement med- og uten silikastov. F C B / S I N T E F , N o r w e g i a n Institute of Technology, Trondheim, 1983, Report STF65 A83022. (In Norwegian). 29. Nilsson L . O . Desorption isotherms for silica-fume/ cement mortars. Institute of Moisture Issues in Material and Building Technology, Trelleborg, Sweden, 1984, Report IF 8 4 3 1 . 30. Mehta P.K. Sulphate resistance of blended Portland cements containing pozzolans and granulated blastfurnace
32
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:07:53
REFERENCES
31.
32.
33.
34.
35.
36.
37. 38. 39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
slag. Proc. 5th Int. Symp. on Concrete TechnoL, Monter rey, Mar. 1981. Manmohan D. and Mehta P.K. Influence of pozzolanic, slag, and chemical admixtures on pore size distribution and permeability of hardened cement pastes. Cement, Concrete, and Aggregates, 1981, 3, N o . 1, Summer, 63-67. Traetteberg A. Frost action in mortar of blended cement with silica dust. ASTM STP 691. American Society for Testing and Materials, 1980, 5 3 6 - 5 4 8 . Mehta P.K. and Gjorv O . E . Properties of Portland cement concrete containing flyash and condensed silica fume. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1982, Report STF65 A82030. Feldman R . F . and Huang Cheng-yi. Properties of Portland cement — silica fume pastes. I. Porosity and sur face properties. Cement and Concrete Res., 1985, 15, 765-774. Huang Cheng-yi and Feldman R . F . Dependence of frost resistance on the pore structure of mortar containing silica fume. Adjournal, 1985, S e p t . - O c t . , Technical Paper, Title N o . 8 2 - 6 8 , 7 4 0 - 7 4 3 . Christensen P. Afprovning af kvalitetskarakterisering af beton tilsat Silica. Teknologisk Institut, Byggeteknik, Copenhagen, 1982. (In Danish). Helland, S. Silica-tilsetningsstoff til betong. Nordisk Betong, 1986, No. 3—4. (In Norwegian). Helland S. Slipforming of concrete with very low water content. Concrete, 1984, 18, N o . 12, D e c , 1 9 - 2 2 . Johansen R. Silicastov i fabrikksbetong. Langtidseffekter. F C B / S I N T E F , Norwegian Institute of Technology, T r o n d h e i m , 1 9 7 9 , R e p o r t S T F 6 5 F 7 9 0 1 9 . (In Norwegian). Loland K.E. and Hustad T. Report 1: fresh concrete and methods of data analysis. FCB/SINTEF, Norwegian Insti tute of Technology, Trondheim, 1981, Report STF65 A81031. Aitcin P . C . et al. The use of condensed silica fume in concrete. Proc. Ann. Meet. Materials Research Soc, Boston, M a s s . , Nov. 1981, 3 1 6 - 3 2 5 . Sellevold E.J. and Radjy F . F . Condensed silica fume (microsilica) in concrete: water demand and strength development. Publication SP-79. American Concrete Institute, 1983, II, 6 7 7 - 6 9 4 . Dagestad G. Undersokelse av samvirke mellom sement SP 30 og RP 38, tilsatt silika, flygeaske og tilsetningsmidlene P 40 og Lomar D. Oslo, 1983, N O T E B Y Report no. 19844. (In Norwegian). Markestad S.A. A study of the combined influence of silica fume and water reducing admixture on water demand and strength of concrete. Materials and Struc tures, 1986, 19, N o . 109, J a n . - F e b . Maage M . and Dahl P . A . Modifisert Portlandsement. Delrapport 2. Betsongsammensetning og fersk betongs egenskaper. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1985, Report STF65 A85022. (In Norwegian). Carette G.G. and Malhotra V . M . Mechanical properties, durability and drying shrinkage of Portland cement con crete incorporating silica fume. Cement, Concrete, and Aggregates, 1983, 5, N o . 1, Summer, 3 - 1 3 . Petersson P . E . et al. Flygaska och kiselstoft som tillsatsmaterial i bruk och betong. Statens Provningsanstalt. Boras, Sweden, 1983, SP-RAPP 1983:11. (In Swedish). Meland I. Adsorption of plasticizer by condensed silica fume. FCB/SINTEF, Norwegian Institute of Technology, Trondheim, 1985, Report S T F 6 5 A 8 5 0 3 5 . Buil M. et al. High strength mortars containing condensed silica fume. Cement and Concrete Res., 1984, 14, 693-704. Helland S. Silica-tilsetningsstoff til betong. Nordisk
51.
52. 53.
54.
55.
56. 57.
58. 59.
60.
61.
62.
63.
64. 65.
66.
67.
68.
69.
70.
71.
72.
Betong, 1986, N o . 3 - 4 , 1 5 - 1 9 . (In Norwegian). Bellander U. Experiences with silica fume in ready mixed concrete. Norwegian Institute of Technology, Trondheim, 1982, Report B M L 82.610, 2 5 7 - 2 6 4 . (In Swedish). Helland S. Slipforming of concrete with low water con tent. Concrete, 1984, 18, D e c , N o . 12, 1 9 - 2 1 . J o h a n s e n R. Risstendens ved plastisk svinn. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1980, S T F 6 5 A 8 0 0 1 6 . (In Norwegian). M a a g e M . Silikastovs vannbehov, storkning av pastaprover. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1982, Report STF65 F 8 2 0 1 8 . (In Norwegian). Meland I. Influence of condensed silica fume and fly ash on the heat evolution in cement pastes. Publication SP-79. American Concrete Institute, 1983, II, 6 6 5 - 6 7 6 . Biirge T . A . 1400 psi in 24 hours. Concrete Int.: Des. and Constr., 1983, 5, N o . 9, Sept., 3 6 - 4 1 . Johansen R. Report 6: long-term effects. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1981, Report S T F 6 5 A 8 1 0 3 1 . Aitcin P . C . Condensed silica fume. Edited by Aitcin P . C . Universite de Sherbrooke, Quebec, 1983. Johansen R. Silikastov i fabrikksbetong. Varmblandet betong. F C B / S I N T E F , N o r w e g i a n I n s t i t u t e of Technology, Trondheim, 1979, Report STF65 F 7 9 0 3 5 . (In Norwegian). Sandvik M . Fasthetsutvikling for silicabetong ved ulike temperaturnivd. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1981, Report S T F 6 5 F81016. (In Norwegian). Sandvik M . Effekten av den ferske betongens temperatur pa fasthetsutviklingen. N o r w e g i a n I n s t i t u t e of Technology, Trondheim, 1982, Report B M L 8 2 . 3 0 3 . (In Norwegian). Skurdal S. Egenskapsutvikling for silicabetong ved forskjellige herdetemperaturer. Norwegian Institute of Technology, Trondheim, 1982, Report B M L 82.416. (In Norwegian). Maage M . and Hammer T.A. Modifisert Portlandsement. Delrapport 3. Fasthetsutvikling og E-modul. F C B / S I N T E F , Norwegian Institute of Technology, T r o n d h e i m , 1 9 8 5 , R e p o r t S T F 6 5 A 8 5 0 4 1 . (In Norwegian). Helland S. Silica — dlsetningsstoff til betong. Nordisk Betong, 1986, N o . 3 - 4 , 1 5 - 1 9 . (In Norwegian). Sandvik M . Silicabetong: herdevarme, egenskapsutvik ling. F C B / S I N T E F , Norwegian Institute of Technology, T r o n d h e i m , 1 9 8 3 , R e p o r t S T F 6 5 A 8 3 0 6 3 . (In Norwegian). Helland S. Temperature and strength development in con crete with W/C ratio less than 0 - 4 0 . Proc. Utilization of High Strength Concrete, Stavanger, June 1987. Available TAPIR, N-7034, Trondheim-NTH, Norway. Rasmussen T . H . Low heat concrete containing PFA, silica powder and superplasticizer. Aalborg Portland, Aalborg, 1981, C B L Internal Report No. 2 8 . (In Danish). Lessard S. et al. Development of a low heat of hydration blended cement. Publication SP-79. American Concrete Institute, 1983, II, 7 4 7 - 7 6 3 . Einstabland T. High strength concrete — underwater con crete bridge for shore approach at Karmoy. Nordisk Betong, 1986, N o . 1 - 2 , 5 5 - 5 8 . Loland K . E . and Hustad T. Report 2: mechanical prop erties. FCB/SINTEF, Norwegian Institute of Technology, Trondheim, 1981, Report STF65 A 8 1 0 3 1 . Modeer M . Silica fume as replacement for some of the cement in ordinary concrete with standard Portland cement; compressive cube strength tests. Norwegian Insti tute of Technology, Trondheim, 1982, Report B M L 82.610, 9 1 - 1 0 3 . (In Norwegian). Skrastins J.I. and Zoldners N . G . Ready mixed concrete 33
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:08:55
CSF IN C O N C R E T E
73.
74.
75.
76. 77.
78.
79.
80.
81. 82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
incorporating condensed silica fume. Publication SP-79. American Concrete Institute, 1983, 11, 8 1 3 - 8 2 9 . Loland K . E . Fasthets- og deformasjonsegenskaper i herdnet tilstand — herdebetingelser. Presented at the seminar Bruk av silika i betong, Norsk Sivilingeniorers Forening, Oslo, Oct. 1983. (In Norwegian). Sorensen E . V . Concrete with condensed silica fume. A preliminary study of strength and permeability. Norwegian Institute of Technology, Trondheim, 1982, Report B M L 8 2 . 6 1 0 , 1 8 9 - 2 0 2 . (In Danish). Malhotra V . M . and Carette G . G . Mechanical properties of concrete incorporating both fly ash and condensed silica fume. Proc. ACI/RILEM Symp. Technology of Concrete when Pozzolans, Slags and Chemical Admixtures Are Used, Monterrey, M a r . 1985, 3 9 5 - 4 1 4 . Loland K . E . and Gjorv O . E . Silikabetong. Nordisk Betong, 1 9 8 1 , N o . 6, 1 - 6 . (In Norwegian). Justesen C . F . Performance of densit injection grout for prestressing tendons. Technical Advisory Service, Aalborg Portland, Aalborg, Denmark, A p r . 1981. Loland K . E . and Hustad T . Lastrespons av C25-betong med og uten silikatilsetning. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1980, Report STF65 A 8 0 0 4 8 . (In Norwegian). Loland K . E . Matematisk modellering av betongens d e f o r m a s j o n s - og b r u d d e g e n s k a p e r basert pa skademekaniske prinsipper — anvendelse pa betong med og uten tilsetning av silika. Norwegian Institute of Technology, Trondheim, Report B M L 8 1 . 1 0 1 . (In Norwegian). Einstabland T. et al. Ilandforing av gassledninger til Kar m o y . Nordisk Betong, 1 9 8 3 , N o . 1, 5 - 1 1 . (In Norwegian). Helland S. et al. Hoyfast betong. Presented at Norsk Betongdag, Trondheim, Oct. 1983. (In Norwegian). Larrard F . et al. Fracture toughness of high strength con crete. Proc. Utilization of High Strength Concrete, Stavanger, June 1987, 2 1 5 - 2 2 3 . Carette G . G . and Malhotra V . M . Early age strength development of concrete incorporating fly ash and con densed silica fume. Publication SP-79. American Con crete Institute, 1983, II, 7 6 5 - 7 8 4 . Regourd M . et al Use of condensed silica fume as filler in blended cements. Publication SP-79, American Con crete Institute, 1983, II, 8 4 7 - 8 6 5 . Tendal O . and Sanne P. Effekt av silica og flygeaske pa betongens fasthetsutvikling. Norwegian Institute of Technology, Trondheim, 1982, Report B M L 8 2 . 5 0 1 . (In Norwegian). Carles-Gibergues A. et al. U s e of fly ash and silica for manufacturing rapid hardening cements. Proc. Colloque International Util. Sous-Produits et Dechets dans le Genie Civil (C.R.), Paris, France, 1980. (In French). Johansen R. and Dahl P . A . Silikabetongens heftegenskaper. Presented at the seminar Bruk av silika i betong, Norske Sivilingeniorers Forening, Oslo, Oct. 1983. (In Norwegian). Carles-Gibergues A. et al. Contact zone between cement paste and aggregate. Proc. Int. Conf. on Bond in Con crete. Edited by Bartos P. Applied Science Publishers, London, 1982, 2 4 - 3 3 . Biirge T . A . High strength lightweight concrete with con densed silica fume. Publication SP-79. American Con crete Institute, 1983, II, 7 3 1 - 7 4 5 . Biirge T . A . Densified cement matrix improves bond with reinforcing steel. Proc. Int. Conf. on Bond in Concrete. Edited by Bartos P. Applied Science Publishers, London, 1982, 2 7 3 - 2 8 1 . Gjorv O.E. et al. Effect of condensed silica fume on the steel-concrete bond. Norwegian Institute of Technology, Trondheim, 1986, Report B M L 8 6 . 2 0 1 . Monteiro P.J. et al. Effect of condensed silica fume on
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106. 107. 108.
109. 110.
111.
112.
113.
114.
the steel-cement paste transition zone. Norwegian Institute of Technology, Trondheim, Report B M L 86.205. Bache H . H . Densified cement/ultrafine particle-based m a t e r i a l s . P r e s e n t e d at t h e 2nd Int. Conf. on Superplasticizers in Concrete, Ottawa, June 1981. Krenchel H . and Shah S. Applications of polypropylene fibres in Scandinavia. Concrete Int.: Des. and Constr., 1985, 2 , N o . 3 , M a r . , 3 2 - 3 4 . Ramakrishnan V. and Srinivasan V. Performance characteristics of fibre reinforced condensed silica fume concrete, 1983, Publication SP-79 American Concrete Institute, II, 7 9 7 - 8 1 2 . Opsahl O . A . A study of a wet-process shotcreting method. Norwegian Institute of Technology, Trondheim, 1985, Report B M L 8 5 . 1 0 1 . Cor wine J . W . Steel fiber-reinforced shotcrete — applica tion and procedure. The Construction Specifier, 1984, July, 4 4 - 4 8 . Morgan D . R . et al. Evaluation of silica fume shotcrete. Proc. Int. Workshop on Condensed Silica Fume in Con crete Edited by Malhatra M . , C A N M E T , Montreal, May 1987. Ohama Y. et al. Properties of carbon fibre reinforced cement with silica liime. Concrete Int.: Des. and Constr., 1985, 7, N o . 3 , M a r . , 5 8 - 6 2 . Bergstrom S.G. and Gram H . E . Durability of alkalisensitive fibres in concrete. Int. J. Cem. Compos, and Lightwt Cone, 1984, 6, N o . 2 , M a y , 7 5 - 8 0 . Radjy ¥.¥. etal. Use of microsilica additives in asbestosfree fiber reinforced cements. Presented at 3rd Int. Symp on Developments in FRC Composites, Sheffield, July 1986. Opsahl O . A . Norsk sproytebetong i internasjonalt perspektiv. Nordisk Betong, 1987, N o . 1, 9 - 1 5 . (In Norwegian). Opsahl O.K. A study of a wet-process shotcreting method. Dr. ing Thesis, Norwegian Institute of Technology, Trondheim, 1985. Traetteberg A. and Alstad R. Volumstabilitet i blandingssementer med rdjernslagg og silikastov. F C B / S I N T E F , Norwegian Institute of Technology, T r o n d h e i m , 1 9 8 1 , Report S T F 65 A 8 1 0 3 4 . (In Norwegian). Maage M . Modifisert Portland cement. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1983, Report S T F 65 A 8 3 0 6 3 . (In Norwegian). Buil M . and Acker P. Creep of silica fume concrete. Cement and Concrete Res., 1985, 15, 4 6 3 - 4 6 6 . American Society for Testing and Materials. ASTM Philadelphia, C 1 5 7 . Wolsiefer J. Ultra high strength field placeable concrete in the range 10,000 to 18,000 psi (69 to 124 M P a ) . Presented at the 1982 Ann. Con v. American Concrete Institute, Atlanta, Jan. 1982. A S T M , Philadelphia, C257. Penttala V . Mechanical properties of high strength con cretes based on different binder combinations. Proc. Utilization of High Strength Concrete, Stavanger, June 1987. Tazawa E. and Yonekura A. Drying shrinkage and creep of concrete with condensed silica fume. Publication SP-91, American Concrete Institute, 1986, II, 9 0 3 - 9 2 1 . A a n e l a n d K . A . Brannteknisk provning. Bygningskonstruksjoner av betong og stdl. Division of Building Materials, Norwegian Institute of Technology, Trond heim, 1979. (In Norwegian). Pedersen S. Beregningsmetoder for varmepdvirkede betongkonstruktioner. Institute of Building Design, Technical University of Denmark, Lyngby, 1981. (In Danish). Hertz K. Heat-induced explosion of dense concretes. Insti tute of Building Design, Technical University of Den-
34 Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:08:55
REFERENCES
mark, Lyngby, 1984, Report N o . 166. 115. Maage M . and R u e s l a t t e n H . Trykkfasthet og blceredannelse pd brannpdkjent hoyfastbetong. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, N o r w a y , M a r . , 1987, Report STF65 A87006. (In Norwegian). 116. Shirley S.T. et al. Fire endurance of high strength con crete slabs. Concrete Int. in press. 117. American Society for Testing and Materials, A S T M , Philadelphia. E l 19. 118. Jensen J.J. et al. Offshore concrete structures exposed to hydrocarbon fire. Proc. Concrete for Hazard Protec tion. Edinburgh, Sept. 1987. Concrete Society, A C I , C E B , CIB and R I L E M for F I P . 119. Dahl P . A . Silicabetong ved hoye temperaturer. F C B / S I N T E F , Norwegian Institute of Technology, T r o n d h e i m , 1 9 8 2 , R e p o r t S T F 6 5 A 8 1 0 4 1 . (In Norwegian). 120. Williamson R.B. and Rashed A.I. A Comparison of ASTM El 19 fire endurance exposure of two EMS AC concretes with similar conventional concretes. Fire Research Laboratory, University of California, Berkeley, July 1987. 121. Holland T . C . Abrasion-erosion evaluation of concrete mixtures for stilling basin repairs, Kinzua Dam, Penn sylvania. US Army Engineer Waterways Experiment Sta tion, Structures Laboratory, Vicksburg, 1983, Miscellaneous Paper SL-83-16. 122. Holland T.C. etal. Use of silica-fume concrete to repair abrasion-erosion damage in the Kinzua D a m Stilling Basin. Publication SP-91. American Concrete Institute, 1986, II, 8 4 1 - 8 6 3 . 123. Gjorv O . E . Durability of concrete containing condensed silica fume. Publication SP-79. American Concrete Insdtute, 1983, II, 6 9 5 - 7 0 8 . 124. Vennesland 0 . and Gjorv O . E . Silica concrete — pro tection against corrosion of embedded steel. Publication SP-79, American Concrete Institute, 1983, II, 7 1 9 - 7 2 9 . 125. Markestad S.A. An investigation of concrete in regard to permeability problems and factors influencing the results of permeability tests. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1977, Report STF65 A77027. 126. Johansson L. Tilsats av kiseldioxidstoft i betong. Swedish Cement and Concrete Research Institute at the Institute of Tchnology, Stockholm, 1975, CBI Report N o . 7587. (In Swedish). 127. Hustad T. and Loland K . E . Report 4: permeability. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1981, Report STF65 A 8 1 0 3 1 . 128. Maage M . Effect of microsilica on the durability of con crete structures. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1984, Report S T F 6 5 A84019. 129. Hammer T.A. and Maage M . Modifisert Portlandsement. Delrapport 4: permeabilitet. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1985, Report STF65 A85040. (In Norwegian). 130. Johansen R. NS 3420 — Innforing av miljoklasser. Presented at Norsk Betongdag, Haugesund, Oct. 1984. (In Norwegian). 131. Mindess S. and Gray R.J. Effect of silica fume additions on the permeability of hydrated Portland cement paste and the cement-aggregate interface. Proc. ACI/RILEM Symp. on Technology of concrete when pozzolans, slags and chemical admixtures are used, Monterrey, March 1985, 121-141. 132. Okkenhaug K. and Gjorv O . E . Influence of condensed silica fume on the air-void system in concrete. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1982, Report S T F 6 5 A82044. 133. Okkenhaug K. Silikastovets innvirkning pd lufiens stabilitet i betong med L-stoffog med L-stoff i kombinas-
134.
135.
136.
137.
138.
139. 140.
141.
142.
143. 144.
145.
146.
147.
148.
149.
150. 151.
152.
153.
154. 155.
jon med P-stoff. Swedish Cement and Concrete Research Institute at the Institute of Technology, Stockholm, 1983, CBI Report 2 : 8 3 , 1 0 1 - 1 1 5 . (In Norwegian). Virtanen J. Freeze-thaw resistance of concrete contain ing blast furnace slag, fly ash or condensed silica fume. Publication SP-79. American Concrete Institute, 1983, II, 9 2 3 - 9 4 2 . Virtanen J. Mineral by-products and freeze-thaw resistance of concrete. Publikation nr. 22:85. Dansk Betonforening, Copenhagen, 1985, 2 3 1 - 2 5 4 . Lehtonen V . Frostbestandighet hos betong med flygaska eller kiselstoft. Cement- och Betong-instituttet, Informationsdagen, Stockholm, 1983. (In Swedish). Lehtonen V. T h e influence of pozzolanic admixtures on the frost resistance of hardened concrete. Publikation nr. 22:85. Dansk Betonforening, C o p e n h a g e n , 1985, 217-230. Aitcin P . C . and Vezina D . Resistance to freezing and thawing of silica fume concrete. Cement, Concrete, and Aggregates, 1984, 6, N o . 1, 3 8 - 4 4 . Rasmussen T . H . Long term durability of concrete. Nor dic Concrete Federation, Oslo, 1985, Publication N o . 4. Christensen P. and Jensen J.H. Luftiblanding i silicabeton til Farobroens kantelementer. Publikation nr. 22:85. Dansk Betonforening, Copenhagen, 1985, 1 8 1 - 1 9 1 . (In Danish). Fagerlund G. Kritisk vattenmettnad. Publikation nr. 22:85. D a n s k Betonforening, C o p e n h a g e n , 1985, 7 9 - 1 0 8 . (In Swedish). Vuorinen J. O m skyddporforhallandet hos betong. Publikation nr 22:85. Dansk Betonforening, Copenhagen, 1985, 2 8 3 - 2 9 2 . (In Swedish). Suomen Betoniyhdistys. SFS 4 4 7 5 . Finnish Standard. SFS 4 4 7 5 , SB, Helsinki. Sellevold E.J. and Bager D . H . Some implications of calorimetric ice formation results for frost resistance testing of cement products. Publikation nr. 22:85. Dansk Betonforening, Copenhagen, 1985, 4 7 - 7 4 . Preece C M . et al. Chloride ion diffusion in low porosity silica cement paste. Norwegian Institute of Technology, Trondheim, 1982, Report B M L 82.610, 5 1 - 5 8 . Vejdirektoratet, Broafdelingen. Betons holdbarhet Rap port nr. 4: Ryd-broen, Forsok med silicabeton. Copenhagen, D e n m a r k , 1982. (In Danish). Cowi Consult, Radgivende Ingeniorer. Forsog med kantelementer af silicabeton, 1981. Vejdirektoratet, Copenhagen, 1982. (In Danish). Bernhardt C I . S i O 2 - s t 0 v som sementtilsetning. Betongen i dag, 1952, 17, N o . 2, A p r . , 2 9 - 5 3 . (In Norwegian). Feldman R . F . and Huang Cheng-yi. Microstructural properties of blended cement mortars and their relation to durability. Proc. RILEM Seminar on Durability of Con crete Structures under Normal Outdoor Exposure. Han nover, M a r . 1984. American Society for Tesdng and Materials. A S T M , Philadelphia C 6 6 6 . Opsahl O . A . Report 5: frost resistance. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1981, STF65 A 8 1 0 3 1 . Yamato et al. Strength and durability of concrete incor porating condensed silica fume. Proc. 2nd Int. Conf. on Use of Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, Madrid, Apr. 1986, American Concrete In stitute, Publication S P - 9 1 , 1986, II, 1 0 9 5 - 1 1 1 9 . Sorensen E . V . Freezing and thawing resistance of con densed silica fume (microsilica) concrete exposed to deic ing chemicals. Publication SP-79. American Concrete Institute, 1983, II, 7 0 9 - 7 1 8 . R I L E M Recommendation C D C 2 , R I L E M , Paris. Intemadonal Organization for Standardization, ISO 4846, Geneva. 35
Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:08:55
C S F IN C O N C R E T E
156. American Society for Testing and Materials. A S T M , Philadelphia, C 6 7 1 . 157. Petersson P . E . T h e influence of silica fiime on the saltfrost resistance of concrete. Presented at Int. Seminar, Chalmers University of Technology, Gothenburg, A p r . 1986. 158. Hooton R . D . Some aspects of durability with condensed silica fume in pastes, mortars and concretes. Proc. Int. Workshop on Condensed Silica Fume in Concrete. Edited by Malhotra M . C A N M E T , Montreal, M a y 1987. 159. Pigeon M . Freeze-thaw durability and de-icing salt scal ing resistance of condensed silica fume concrete. Proc. Int. Workshop on Condensed Silica Fume in Concrete. Edited by Malhotra M . C A N M E T , Montreal, May 1987. 160. American Society for Testing and Materials. A S T M C 6 7 2 , Philadelphia. 161. Aitcin P - C . and Pigeon M . Performance of condensed silica fume concrete used in pavements and sidewalks. Durability of Building Materials, 1986, 3, 3 5 3 - 3 6 8 . 162. Bilodeau A . and Carette G . G . Resistance of condensed silica fume concrete to the combined action of freezing and thawing cycling and de-icing salts. Proc. Int. Workshop on Condensed Silica Fume in Concrete. Edited by Malhotra M . C A N M E T , Montreal, M a y 1987. 163. Malhotra V . M . et al. Mechanical properties and freez ing and thawing resistance of high strength concrete incor porating silica fiime. Proc. Int. Workshop on Condensed Silica Fume in Concrete. C A N M E T , Montreal, M a y 1987. 164. Sellevold E.J. Unpublished results of Scandinavian Round-Robin-Tests, N o r w e g i a n Building Research Institute, Oslo, N o r w a y , 1987. 165. Samuelsson P . T h e influence of silica fume on the risk of efflorescence on concrete surfaces. Norwegian Institute of Technology, Trondheim, 1982, Report B M L 82.610, 2 3 5 - 2 4 4 . (In Swedish). 166. Carlsen R. and Vennesland 0 . Sementers sulfat- og sjovannsbestandighet. FCB/SINTEF. Norwegian Institute of Technology, Trondheim, 1982, Report STF65 F82010. (In Norwegian). 167. Fiskaa O . et al. Betong i Alunskifer. Norwegian Geotechnical Institute, Oslo, 1 9 7 1 , Publication no. 86. (In Norwegian). 168. F i s k a a O . M . Betong i A l k u n s k i f e r . Norwegian Geotechnical Institute, Oslo, 1973, Publication no. 101. (In Norwegian). 169. Unpublished results, Norwegian Building Research Insti tute, Oslo. 170. Mather K. Factors affecting sulphate resistance of mor tars. Proc. 7th Int. Conf. on Chemistry of Cements, 1980, 4, 5 8 0 - 5 8 5 . 171. Popovic K. et al. Improvement of mortar and concrete durability by the use of condensed silica fume. Durabil ity of Building Materials, 1984, 2 , 1 7 1 - 1 8 6 . 172. Metha P.K. Chemical attack of low water cement ratio concretes containing latex or silica fume as admixtures. Proc. ACI/RILEM Symp. on Technology of Concrete when Pozzolans, Slags and Chemicals Admixtures Are Used. Monterrey, M a r . 1985, 3 2 5 - 3 4 0 . 173. American Society for Testing and Materials. A S T M C 1 0 1 2 , Philadelphia. 174. Asgeirsson H . and Gudmundsson G. Pozzolanic activity of silica dust. Cement and Concrete Res., 1979, 9, 249-252. 175. Olafsson H . Effect of silica fume on the alkali-silica reac tivity of cement. Norwegian Institute of Technology, Trondheim, 1982, Report B M L 82.610, 1 4 1 - 1 4 9 . (In Norwegian). 176. Sveinbjornsson S. Alkali-aggregate damages in concrete with microsilica-field study. Icelandic Building Research Institute, Reykjavik, 1987. 177. Oberholster R . E . and Westra W . B . T h e effectiveness of
178.
179.
180.
181.
182.
183. 184.
185.
186.
187.
188.
189.
190.
191.
192. 193.
194.
195.
196.
197.
mineral admixtures in reducing expansion due to alkali aggregate reaction with Malmesbury group aggregates. Presented at Conf on Alkali-Aggregate Reaction in Con crete, Cape T o w n , M a r . , 1981. Olafsson H . and Thaulow N . Alkali-silica reactivity of sands. Technological Institute, Copenhagen, 1981, Nord test Project 173-79. Perry C. and Gillott J . E . T h e feasibility of using silica ftime to control concrete expansion due to alkali-aggregate reaction. Durability of Building Materials, 1985, 3, 133-146. Gillott J . E . T h e influence of silica fume on alkaliaggregate reactions. Proc. Int. Workshop on Condensed Silica Fume in Concrete. C A N M E T , Montreal, 1987. Nilsson L . O . and Peterson O . Alkali-silica reactions in Scania, Sweden. Division of Building Materials, Lund I n s t i t u t e of T e c h n o l o g y , L u n d , 1 9 8 3 . R e p o r t TVBM-3014. Peterson O . Pop-out formation. A laboratory study with a Swedish opaline gravel. Proc. 6th Int. Conf. on Alkalis in Concrete. Copenhagen, 1983, 2 9 1 - 2 9 8 . American Society for Testing and Materials. ASTM C 4 4 1 , Philadelphia. Aitcin P . C . and Regourd M . T h e use of condensed silica liime to control alkali-silica reaction — a field case study. Cement and Concrete Res., 1985, 15, 7 1 1 - 7 1 9 . Norcem Cement Division. Corrocem: anti-corrosion additive for concrete. Technical Bulletin from Norcem Cement Division, Concrete Technology Dept., Slemmestad, N o r w a y , 1981. Feldman R . F . and Huang Cheng-yi. Resistance of mor tars containing silica fume to attack by a solution con taining chlorides. Cement and Concrete Res., 1985, 15, 411-420. Diamond S. Influence of chemical admixtures on mix water-pore solution compositions and on early hydration. Proc. ACI/RILEM Symp. on Technology of Concrete When Pozzolans, Slags, and Chemical Admixtures are Used. Monterrey, M a r . 1985, 1 - 1 8 . Glasser F . P . and Marr J. The effect of mineral additives on the composition of cement pore fluids. Brit. Ceramic Proc, 1984, 35, 4 1 9 - 4 2 9 . Gautefall O . and Vennesland 0 . Modifisert Portlandsement. Delrapport 5: Elektrisk motstand og pH-nivd. F C B / S I N T E F , Norwegian Institute of Technology, T r o n d h e i m , 1 9 8 5 , R e p o r t S T F 6 5 A 8 5 0 4 2 . (In Norwegian). Meland I. and Traetteberg A . Karbonatisering i sement med slagg eller silikastov. F C B / S I N T E F , Norwegian Institute of Technology, Trondheim, 1981, Report STF65 A 8 1 0 3 3 . (In Norwegian). Vennesland 0 . Report 3: Corrosion properties. F B C / S I N T E F , T h e Norwegian Institute of Technology, Trondheim, 1981, Report STF65 A 8 1 0 3 1 . Private communication with R. Johansen, FCB/SINTEF, Norwegian Institute of Technology, Trondheim, 1986. Maage M . and Sellevold E.J. Feltmalinger av kar bonatisering i betong med og uten silikastov. Nordisk Betong, 1985, N o . 1, 2 1 - 2 2 . (In Norwegian). Skjolsvold O . Carbonation depths of concrete with and without condensed silica fume. Publication SP-91. American Concrete Institute, 1986, II, 1 0 3 1 - 1 0 4 8 . de Fontenay C. le Sage. A study of the effect of concrete admixtures, concrete composition and exposure condi tions on carbonation in Bahrain. Presented at 1st Int. Conf. on Deterioration and Repair of Reinforced Concrete in the Arabian Gulf, Bahrain, Oct. 1985. Page C . L . and Havdahl J. Electrochemical monitoring of corrosion of steel in microsilica cement pastes. Mater, and Structs, 1985, 18, N o . 103, J a n . - F e b . , 4 1 - 4 7 . Monteiro P . J . M . et al. Microstructure of the steel-cement paste interface in the presence of chloride. Cement and
36 Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:08:55
REFERENCES
Concrete Res., 1985, 15, 7 8 1 - 7 8 4 . 198. Fisher K . P . et al. Corrosion of steel in concrete: some fundamental aspects of concrete with added silica. Norwegian Geotechnical Institute, Oslo, 1982, Report no. 51304-06. 199. Gautefall O . Modifisert Portland sement. F B C / S I N T E F , Norwegian Institute of Technology, Trondheim, 1984, Arbeidsnotat, Prosjekt Nr. 651357.00, Notat N r . 6. (In Norwegian). 200. Gautefall O . The effect of condensed silica fume on the diffusion of chlorides through hardened cement paste. Proc. 2nd Int. Conf. on the Use of Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, Madrid, American Concrete Institute, Publication S P - 9 1 , 1986, II, 9 9 1 - 9 9 9 . 2 0 1 . Byfors K. Influence of silica fume and flyash on chloride diffusion and pH values in cement paste. Cement and Concrete Res., mi, 17, 1 1 5 - 1 3 0 .
2 0 2 . Christensen D . W . et al. Rockbond: a new microsilica concrete bridge deck overlay material. Proc. Int. Bridge Conf, Pittsburgh, June 1984, 1 5 1 - 1 6 0 . 2 0 3 . Marusin S.L. Chloride ion penetration in conventional concrete and concrete containing condensed silica fume. Publication SP-91. American Concrete Institute, 1986, II, 1 1 1 9 - 1 1 3 3 . 2 0 4 . Preece CM. etal. Electrochemical behaviour of steel in dense silica-cement mortar. Publication SP-79. American Concrete Institute, 1983, II, 7 8 5 - 7 9 6 . 2 0 5 . Biirge T . A . Densified silica-cement coating as an effec tive corrosion protection. Corrosion of Reinforcement in Concrete Construction. Society of Chemical Industry, London, 1983, 3 3 3 - 3 4 2 . 2 0 6 . Gautefall O . and Vennesland 0 . Effects of cracks on the corrosion of embedded steel in silica-concrete compared to ordinary concrete. Publication No. 2, Nordic Concrete Federation, Oslo, 1 9 8 3 , 1 7 - 2 8 .
37 Delivered by ICEVirtualLibrary.com to: IP: 129.132.211.123 On: Sun, 26 Jun 2011 15:08:55