PERFORMANCE OF CONCRETE CONTAINING T E R N A R Y BINDERS IN C H L O R I D E - L A D E N E N V I R O N M E N T S B J Magee Purdue University United States of America M R Jones R K Dhir University of Dundee United Kingdom A B S T R A C T . Based on the premise that if one additional binder material in concrete provides improved performance, then further benefits are likely when using two, this study was carried out to examine selected properties of ternary binder concrete (TBC). Various fresh concrete properties were considered as well as strength development and chloride resistance of TBC. In addition, the nature of the near surface pore structure of the concrete has been inferred from its initial surface absorption. Performance in all instances was compared to PC and binary PC/PFA concrete of equivalent 28-day cube strengths of 20, 40 and 60 N/mm2. In general, no practical dissimilarity existed between TBC, PC and PC/PFA mixtures in terms of fresh concrete properties. When designed for equivalent 28-day strength, TBC exhibited lower strengths than the controls. Equations enabling estimates of strength reductions are proposed. Improved long-term strength was achieved by TBC, with PC replacement levels of around 60% by mass identified as an optimum. Using accelerated electrochemical and penetration test methods, it has been shown that the chloride resistance of all the ternary binder concrete (TBC) is significantly higher than corresponding PC and PC/PFA mixes. This improved performance is attributed mainly to the improved microstructure of TBC mixes. Keywords: Ternary binder concrete, PFA, GBS, SF, Mix handling/placing/finishing characteristics, Strength development, Chloride ingress. D r B r y a n J Magee is currently a post-doctoral research associate at the School of Civil Engineering, Purdue University, Indiana, USA. Prior to this appointment, he held a similar position at the Department of Civil Engineering, University of Cape Town, South Africa. He received his BSc(Eng) and PhD degrees from the University of Dundee, Scotland. His research interests include cement and concrete technology, with emphasis on the use of supplementary cementitious materials and the issue of long-term durability D r R o d Jones is a chartered civil engineer and senior lecturer in the Concrete Technology Unit in the Department of Civil Engineering at the University of Dundee. His research focuses mainly on binder technology, concrete durability and repair and maintenance. Professor R a v i n d r a D h i r is the Director of the Concrete Technology Unit and Professor of Concrete Technology at the University of Dundee. He is a member of numerous national and international technical committees and has published extensively on many aspects of concrete technology, binder science, durability and construction methods. Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. 228 Magee, Jones, Dhir INTRODUCTION Premature corrosion damage of embedded reinforcement in concrete continues to present one of the most serious durability problems facing engineers. Although it is established [1-3] that improved resistance to corrosion damage is achievable through the use of materials such as pulverised-fuel ash (PFA), granulated blastfurnace slag (GBS) and silica fume (SF) in binary blends, case studies indicate that many existing structures continue to suffer . Encouragingly, recent research [5,6] indicates that ternary blended binders in concrete allow high levels of durability to be achieved, particularly with respect to chloride ingress. These findings assume particular significance with the adoption of the cement standard ENV 197-1 (1996), which presents European engineers with the opportunity to tailor concrete for high levels of durability via binder optimisation. Indeed, ENV 197-1 allows cocktails of three or more binder materials to be prescribed for any given concrete mix. The value of using ternary binder concrete (TBC) has already been realised in many countries, with its advantages exploited for some major infrastructure projects. For example, the Great Belt Link in Denmark was constructed using a PC/PFA/SF blend . In this case, TBC was used as an integral component of a multistage protection strategy designed to provide a service life of 100 years. this study was undertaken to examine both the practicality and performance of TBC. Although the main focus of the study was the chloride resistance of TBC, selected fresh properties and strength development were additionally considered. Despite ENV 197-1 identifying a total of 25 types of cement (of which Types II-N and V allow ternary blended binders) based around 8 different additional materials, only those of immediate relevance to the UK were considered. Combinations of Portland cement (PC) to BS 12 (1991), PFA to BS EN 450 (1995), GBS to BS 6699 (1991) and SF were, therefore, considered within the scope of this work, used in combinations based on the requirements of ENV 197-1. EXPERIMENTAL MIX PROPORTIONS On a practical basis and in order to limit large numbers of possible binder combinations, additional materials were proportioned based on replacement levels similar to those used in typical binary mixes. Such binder proportions were also considered suitable due to their potential to provide practical mixes with high levels of chloride resistance [1-3]. Using a mix design method developed at the University of Dundee , TBC mix proportions were selected to give standard 28-day cube strengths of 20, 40 and 60 N/mm . The control mixes were PC and PC/30% PFA concrete of equal cube strengths. The mix proportions used are given in Table 1, which additionally includes a binder proportion summary where each binder material is represented as a percentage of the total binder content by mass. 2 Performance of each binder combination chosen was highlighted by maintaining fixed free water and coarse aggregate contents (185 1/m and 1200 kg/m respectively). Unit volumes of the mixes were subsequently achieved by varying the fine aggregate contents. It should be noted that with the free water content selected, 28-day cube strengths of 60 N/mm were not achievable with the PC/PFA/GBS blends. All mixes obtained a nominal slump of 75 mm without the need for a plasticising chemical admixture except for those using SF, where the very cohesive nature of the concrete required the use of a minimal dosage of superplasticiser (see Table 2 for details). All concrete specimens were standard water cured for 28-days and all tests were initiated at that time. 3 3 2 Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. Ternary Binder Concrete 229 [able 1 Summary of concrete mix proportions 28-DAY CUBE STRENGTH, N/mm MIX CODE CONST 1 IT I'NT M A T E R I A L S , kg/m' Binder 2 PC PFA PC control C1 C2 C 3 20 40 60 225 325 430 - PC/PFA C4 C 5 C6 20 40 60 PC/PFA/GBS TBC 1 TBC 2 TBC 3 TBC 4 TBC 5 TBC 6 TBC 7 GBS S F Fine Aggregate BINDER PROPORTION, % BY M A S S - 820 745 630 100/0/0 100/0/0 220 270 405 95 115 175 600 540 360 70/30/0 70/30/0 70/30/0 20 20 20 20 40 40 40 140 100 195 85 230 155 155 50 120 70 150 55 175 90 595 565 475 475 545 425 330 42.5/15/42.5 32.5/35/32.5 25/15/60 20/35/45 60/15/25 32.5/35/32.5 25/15/60 PC/PFA/SF TBC 8 TBC 9 TBC 10 TBC 11 TBC 12 20 20 40 40 60 180 140 250 220 265 20 140 30 220 265 20 30 30 45 65 725 580 615 410 355 80/10/10 45/45/10 80/10/10 45/45/10 45/45/10 PC/GBS/SF TBC 13 TBC 14 TBC 15 TBC 16 40 40 40 60 195 155 105 160 25 35 35 50 625 580 545 370 65/25/10 45/45/10 25/65/10 25/65/10 1 ooo/o/ control 140 100 195 195 100 155 370 85 155 245 370 FRESH PROPERTIES Mix Observations Owing to the unfamiliarity of TBC, qualitative observations for ease of compaction, cohesiveness, colour and finishing characteristics were initially documented, see Table 2. Clearly, TBC mixes exhibited no visual dissimilarity to good quality PC concrete. In fact, with regards to fresh properties, TBC was generally of higher quality than the PC and PC/PFA controls. Of practical significance, TBC was easily compacted, exhibited no visible bleeding and produced an excellent surface finish. Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. 230 Magee, Jones, Dhir Table 2 Fresh properties and visual observations of TBC MIX SLUMP, mm CODE PC SP DOSAGE AIR 1 C O N T E N T , O B S E R V A T I O N S OF MIX H A N D L I N G % CHARACTERISTICS control C 1 85 - 1.2 Normal m i x appearance. C2 75 - 1.1 Normal m i x appearance. C3 60 - 1.1 Normal m i x appearance. G o o d w o r k a b i l i t y and c o m p a c t i o n . Slight bleeding. G o o d w o r k a b i l i t y and c o m p a c t i o n . PC/PFA control C4 90 - 1.0 Normal m i x appearance. C5 75 - 1.2 Normal m i x appearance. C6 65 - 1.1 Normal m i x appearance. Very workable, easy compaction. G o o d c o h e s i o n and compaction. H a n d l i n g and c o m p a c t i o n g o o d . PC/PFA/GBS TBC 1 75 - 1.1 Normal m i x appearance, with g o o d TBC 2 65 - 1.3 Normal m i x appearance. TBC 3 80 - 1.2 Normal m i x appearance, no bleeding, good TBC 4 60 - 1.1 G o o d w o r k a b i l i t y , c o m p a c t i o n and TBC 5 60 - 1.1 N o r m a l m i x a p p e a r a n c e , quite stiff, n o TBC 6 50 - 1.0 Normal m i x appearance. TBC 7 50 - 1.3 Normal mix, g o o d workability, w o r k a b i l i t y , c o m p a c t i o n a n d finish G o o d workability, slight bleeding. c o h e s i o n and s u r f a c e finish. cohesion. Appeared normal. bleeding, good cohesion. c o m p a c t i o n , finish and c o h e s i o n . PC/PFA/SF TBC 8 65 0.7 (7.0) 0.9 TBC 9 60 0.9 (9.0) 0.8 A v e r a g e w o r k a b i l i t y but m i x e a s i l y c o m p a c t e d . D a r k in a p p e a r a n c e . Fairly harsh m i x , n o b l e e d i n g . E a s y c o n c r e t e c o m p a c t i o n . D a r k grey. T B C 10 60 0.9 (9.0) 0.6 M i x appeared normal. G o o d w o r k a b i l i t y , e a s y h a n d l i n g and workability. T B C 11 55 1.1 ( 1 7 . 0 ) 0.9 M i x stiff, w i t h l o w s l u m p . C o m p a c t i o n a c h i e v e d . V e r y dark c o l o u r . T B C 12 50 1.5(13.0) 0.9 M i x stiff, w i t h l o w s l u m p . C o m p a c t i o n a c h i e v e d . V e r y dark c o l o u r . PC/GBS/SF T B C 13 75 1.0(12.0) 0.8 Discolouration, average workability, mix c o m p a c t e d satisfactorily. T B C 14 70 1.0(10.0) 0.9 H a r s h , stiff m i x but g o o d c o m p a c t i o n . No bleeding, discolouration. T B C 15 70 1.1 ( 1 2 . 0 ) 0.7 T B C 16 65 1.4(16.0) 0.9 M i x appeared normal. G o o d workability, h a n d l i n g and c o m p a c t i o n . Stiff, harsh m i x . N o b l e e d i n g , h i g h cohesion. Slight discolouration. 1 Superplasticiser dosage as % of total binder content (and SF content) by mass Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. Ternary Binder Concrete 231 However, there was some variation in the colour of TBC when assessed at 28 days. Mixes rich in GBS and PFA tended to range from lighter shades of grey to typical shades. Alternatively, and due to the nature of the slurry used, mixes rich in SF ranged from darker shades of grey to almost black. Entrapped Air Contents As the potential for TBC to entrap air was uncertain, all concrete was tested in compliance with BS 1881: Part 106 (1983) after undergoing full compaction. As shown in Table 2, results indicate minor differences existing between the controls and PC/PFA/GBS mixes. In contrast, entrapped air contents obtained for all TBC mixes containing SF were on average 30% lower than the controls. Why the observed reduction in entrapped air occurs is unclear, although some hypotheses have been described for binary mixes. As superplasticisers are dispersants they tend to repel air that has become, in part, stabilised by orientation of the surfactant. For this reason, high workability concrete containing superplasticisers are often characterised by foam, caused by a rapid release of entrapped air . As all TBC mixes inclusive of SF contained superplasticiser, this theory may, in part, explain the reduced air contents recorded. STRENGTH DEVELOPMENT Specimens (100 mm cubes) were tested at intervals up to 180-days in accordance with BS 1881: Part 116 (1983). TBC mixes generally exhibited lower early strength development than PC and PC/PFA control concrete. Indeed, at 1 and 3-days, respective TBC strengths were on average 50% and 30% lower than PC concrete, and 25% and 15% lower than PC/PFA concrete. These strength differences decreased with time, however, and by 90 days the TBC strengths were on average 25% and 20% higher than the PC and PC/PFA concrete respectively. This level of strength improvement was maintained up to 180 days. Cube strengths obtained for TBC mixes at 1, 3, 90 and 180 days are given in Figure 1, expressed as a percentage of PC control results and plotted with respect to the percentage of PC replacement used for each mix. 20 100 60 40 170 6 H (a) Y = -[0.4X + 9.3] -20 'ex.. w PQ U U 150 A A -40 x oo -60 Y = - [ 0 . 7 X + 15.5] A* o \ o o O O 90-day A A A (b) A 180-day -80 130 o A 0 A b-,.o A <=>R o 110 \ x O1-day 0 A 3-day 90 -100 0 20 40 60 % O F PC REPLACEMENT, by mass Figure 1 Relationship between TBC strength and percentage of PC replacement used Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. 100 232 Magee, Jones, Dhir Figure 1(a) indicates that as PC replacement in the TBC mixes increased, strength development at 1 and 3 days decreased with respect to the PC control. By calculating the equations of the best-fit lines shown, it is possible to estimate early strength development of TBC. For example, consider a TBC mix with a PC replacement of X%. Strength at 1 and 3 days would be approximately [0.7 X + 15.5]% ± 9.0% (at a 90% confidence) and [0.4 X + 9.3]% ± 8.0% (at a 90% confidence) lower than a typical PC mix respectively. In contrast, the relationship between improved TBC strength and PC replacement at 90 and 180-days was not consistent. Indeed, Figure 1(b) indicates that an optimum PC replacement exists, in terms of ultimate strength development, of around 60% by mass. Beyond this level, it is likely that Ca(OH)2 available for pozzolanic reactions may be limited. It should be recognised, however, that as all concrete specimens were standard water cured until the required test age, continued pozzolanic activity was not prevented. ACCELERATED CHLORIDE RESISTANCE TESTS E l e c t r o c h e m i c a l C h l o r i d e Ingress Predicting a potentially high chloride ingress resistance of the TBC mixes, an accelerated 14day electrochemical chloride transmission test  using 12V DC potential difference (PD) was initially adopted. PD indices were subsequently calculated using Fick's First Law and these are compared for all mixes in Figure 2. Figure 2 Comparison of PD indices obtained for control and TBC mixes In comparison to both the PC and PC/PFA control mixes, the use of ternary blended binders resulted in a dramatic reduction in PD index. Considering the 20 N/mm mixes for example, PD indices obtained for TBC were on average 90% and 85% lower than those obtained for the control PC and PC/PFA concrete respectively. 2 Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. Ternary Binder Concrete 233 The degree of improved performance decreased slightly with an increase of concrete strength to 40 N/mm and corresponding reductions for these TBC mixes were on average 89% and 82%. At the conclusion of the 14-day test period no chloride transmission was measured for either of the 60 N/mm TBC mixes considered. 2 2 Comparing the PD indices obtained it is evident that only minor differences existed (results ranged from between 0.30 to 0.90), irrespective of the concrete strength or the binder combination used. These variations are, in fact, within the accuracy of the test method and may, therefore, not be significant . This trend most likely reflects equally very high levels of chloride resistance for all TBC. It is interesting that as the level of PC replacement ranged from between 20 to 80% by mass, the results from this test method suggest that no optimum PC replacement exists in order to achieve maximum chloride resistance. Chloride P e n e t r a t i o n Owing to the rapid nature of the electrochemical chloride transmission test and the resulting likelihood of limited chloride binding , a restricted series (i.e. mixes TBC 13-16 omitted) of chloride immersion testing was additionally carried out. All test specimens (75 x 75 x 300 mm prisms) were immersed in 5 M. NaCl solution and sealed on 5 faces to give uniaxial chloride penetration. After an immersion period of 28, 56 and 91 days, powder samples were obtained by incremental drilling at depths of 0-5 mm, 6-15 mm and 16-30 mm from the ascast surface. X-ray fluorescence was subsequently used to analyse these samples for their total chloride content. Using results obtained after 91 days of immersion, chloride profiles were constructed as shown in Figure 3. Figure 3 Chloride profiles constructed from 91 day accelerated chloride penetration results Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. 234 Magee, Jones, Dhir The significant advantage of using TBC mixes is, once again, clearly apparent. For example, considering the 16-30 mm results collectively over the range of 28 day strengths, the TBC mixes had on average 80% lower chloride concentration than the PC/PFA controls. As would be expected due to shorter exposure periods, increasingly lower chloride levels, in comparison to Figure 3, were obtained for all mixes at 56 and 28 days. However, profiles constructed from results at each of these ages produced very similar trends to those shown in Figure 3, with TBC outperforming the controls in all instances. CHLORIDE DIFFUSION COEFFICIENTS Owing to the accelerated nature of the chloride tests considered, the results obtained (PD indices and percentages of total chlorides by mass of concrete for the electrochemical and penetration tests respectively) are not comparable and cannot easily be used to evaluate concrete's resistance to chloride ingress. For this reason, coefficients of chloride diffusion were estimated or calculated from the results obtained from each test. Curves have been reported in the literature , correlating values of PD index and coefficients of chloride diffusion for PC and PC/30% PFA concrete. As no established relationship exist between coefficients of diffusion and PD indices for TBC, the curves given for the PC/30%) PFA concrete were adopted for both the PFA control and TBC mixes. Using these curves, coefficients of chloride diffusion have been estimated as shown in Table 3. It is recognised that further research is merited in this area, to establish relationships between coefficients of diffusion and PD indices for various TBC mixture combinations. Alternatively, coefficients of chloride diffusion were calculated from the chloride profiles shown in Figure 3, using Fick's Second Law. Clearly from Table 3, and most likely reflecting the assumptions made when estimating values from the electrochemical test data, variations existed in coefficients of chloride diffusion obtained from the two accelerated test methods. The results again illustrate, however, the greatly increased performance of TBC in comparison to the controls. Nomograms enabling estimates of service life for a given set of conditions have been established  and indicate that poor, good, very good and excellent performance is likely for coefficients of chloride diffusion in the ranges >10.0, 5.0-10.0, 1.05.0 and <1.0 x 10" cm /sec respectively. From this study, TBC mixes indicate the likelihood of very good and excellent performance in all cases, whereas only 60 N/mm PC/PFA concrete achieves very good performance. The PC control concrete results indicate the probability of poor performance at strengths up to 40 N/mm . The 60 N/mm PC concrete achieves only good performance status. 9 2 2 2 2 DISCUSSION OF TEST RESULTS Fresh Concrete From the qualitative analysis carried out, the use of ternary blended binders in concrete presents no immediate technical difficulties. Indeed, the compaction, cohesiveness, and finishing characteristics of TBC were all favourable. However, the colour variability noted has aesthetic implications, for example, in structures where selected elements use TBC with conventional or binary mixes in close proximity. TBC that is almost black in colour is not likely to be practicable in many cases and probable application for such mixes is limited to underground works. For structures where aesthetics are critical, suitable materials and material combinations should be chosen and trial mixed. Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. Ternary Binder Concrete 235 Table 3 Coefficients of chloride diffusion calculated or estimated from accelerated tests 2 MIX CODE 9 COEFFICIENT OF CHLORIDE DIFFUSION, cm x 10" Calculated from accelerated Estimated from accelerated chloride penetration test electrochemical chloride test 1 2 PC control C1 C2 C3 PC/PFA 65.0 25.0 9.5 95.0 40.0 11.0 15.0 8.5 5.5 35.0 17.0 11.0 2.0 1.0 1.0 1.7 0.5 1.0 1.1 4.0 4.0 5.0 5.5 3.5 3.5 2.0 2.1 3.0 0.5 0.4 0.4 3.5 5.5 4.0 2.0 0 control C4 C5 C6 PC/PFA/GBS TBC 1 TBC 2 TBC 3 TBC 4 TBC 5 TBC 6 TBC 7 PC/PFA/SF TBC 8 TBC 9 TBC 10 TBC 11 TBC 12 1 Calculated using Fick's Second Law. Estimated using established relationships between PD-index and coefficient of chloride diffusion, with relationship for PC/30%PFA assumed to be valid for TBC mixes. 2 S t r e n g t h Development The inherent low strength development of TBC mixes at early ages may have to be accounted for in practice, particularly if designing for applied pre-stressing forces, formwork pressures and stripping times etc. In addition to estimating early strength reductions using the equations proposed, it is likely that this drawback might be minimised by using accelerating admixtures, raised curing temperatures or by reducing the level of PC replacement. In adopting the latter course of action, however, the producer may sacrifice longer-term concrete performance, particularly with respect to durability. Clearly, the results indicate an optimum PC replacement level with respect to the longer-term strength development of TBC. It is likely that optimum performance in terms of durability may also be achieved using this PC replacement level (around 60% by mass). Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. 236 Magee, Jones, Dhir In many cases, however, achieving maximum strength development may not be the main design criterion and factors such as cost, heat evolution etc. may be the controlling parameter. This may result in selecting a TBC mix for a minimum total binder content or a minimum PC content. Chloride Resistance In terms of chloride resistance and due to either physical and/or chemical influences , the improved TBC performance is striking, with 20 N/mm TBC mixes outperforming 60 N/mm controls in many cases. Physically, pozzolanic reactions associated with additional materials are likely to refine the microstructure of concrete . The results recorded after 10 minutes (IS AT-10 results) indicated a large reduction of TBC surface absorption in comparison to the controls. ISAT-10 values of 75.0, 55.0 and 40.0 ml/m /sec x 10' were obtained for the 20, 40 and 60 N/mm PC concrete respectively, with reductions of 14%, 18% and 50% noted for the corresponding PC/PFA mixes. In contrast, average reductions in ISAT-10 values for the TBC mixes were 63%, 67%) and 80% at each strength grade respectively. This trend essentially mirrors that obtained for chloride resistance, as given in Figure 2. 2 2 2 2 2 It has been reported  that some aluminate phases in concrete may play a role in the binding of chloride ions. Although the bulk AI2O3 content may be a relatively poor estimate of the actual amount of chloride binding sites available, it is at least generally indicative. Despite variations in total binder content existing between mixes, calculations indicated that the quantity of AI2O3 (per kg/m ) in the majority of TBC mixes was not significantly different than that in the PC/30% PFA controls and indeed in many cases was lower, in particular for the SF mixes. This suggests that overall it is probably the improved microstructure of TBC that has the dominant effect on increased chloride resistance. 3 CONCLUSIONS 1. In terms of the fresh properties examined, no practical difference existed between TBC, PC and PC/PFA mixes. TBC was, however, very cohesive but easily compacted and produced a high quality surface finish. The use of superplasticiser was found necessary in all TBC mixes containing SF. TBC containing PC/PFA/GBS entrapped similar quantities of air to PC and PC/PFA control mixes. All TBC containing SF entrapped on average 30% less air than control mixes. TBC had a wide range of colours, ranging from light grey to almost black, depending on the combinations of binder materials used. 2. When designed for equivalent 28-day strength, TBC exhibited on average 50% and 30% lower strength than PC control concrete and 1 and 3 days. These differences reduced with time, however, and the TBC strength results were on average 25% higher than the PC control at 90 and 180 days. For optimum long-term strength development, PC replacement levels of around 60% by mass were identified for TBC. 3. The results obtained from both the electrochemical ingress and penetration tests indicated that the chloride resistance of TBC was significantly better than both the PC and PC/PFA control mixes. The initial surface absorption of TBC was reduced in comparison to the controls. Since there were only minor differences in bulk AI2O3 content, this suggests that the chloride resistance of TBC is mainly the result of microstructural refinement. No optimum replacement levels for PC or combinations of binder materials for chloride resistance were found and it is recommended, therefore, that mix proportions are chosen to maximise strength development (i.e. by selecting a PC replacement of around 60%). Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved. Ternary Binder Concrete 237 REFERENCES 1. DHIR, R K, and MATTHEWS, J D. Durability of PFA Concrete. The Use of PFA in Construction, Ed. RK Dhir and MR Jones, pub E&FN Spon, London, 1992. 2. BIJEN, J. Blast Furnace Slag Cement for Durable Structures. Stichting BetonPrisma, 1996, 62pp. 3. ZHANG, M H, and GJ0RV, O E. Cement and Concrete Research. Vol. 23, no. 411, 1985. 4. WALLBANK, E J. The performance of concrete in Bridges: A survey of 200 highway bridges. Department of Transport, April 1989, 96pp. 5. JONES, M R, DHIR, R K, and MAGEE, B J. Concrete Containing Ternary Blended Binders: Carbonation & Chloride Resistance. Cemt & Concrete Research, Vol.27, No.6,pp.825-831, 1997. 6. MAGEE, B J, and ALEXANDER, M G. Durability performance of concrete containing condensed silica fume. Submitted for publication. 7. VINCENSTEN, L J, and HENRIKSON, K R. The Great Belt Link - Built to last. Concrete International, Vol.14, No.25, 1992, pp.30-33. 8. JONES, M R, MAGEE, B J, and DHIR, R K. Concrete Containing Ternary Blended Binders: Fresh Properties, Strength Development and Mix Design. In preparation 9. HEWLITT P C. Private communication with Dundee University, 1998. 10. DHIR, R K, JONES, M R, AHMED, H E H. and SENEVIRATNE, A M G. Rapid estimation of chloride diffusion in concrete. Mag of Concrete Research. Vol.42, No. 177, 1990, pp. 177-185. 11. ALONSO, C, and ANDRADE, C. Corrosion of steel reinforcement in carbonated mortar containing chlorides. Advances in Concrete Research, Vol.1, No.3, pp. 155-163, 1988. 12. DHIR, R K, JONES, M R, and AHMED, H E H. Concrete durability: estimation of chloride concentration during design life. Mag. of Concrete Research, Vol. 43, No. 154, 1991,pp.579-590. 13. ROY, D M. Hydration of blended cements containing slag, fly ash or silica fume. Proceedings Institution of Concrete Technology. Coventry, UK, 1987. Downloaded by [ Griffith University] on [25/10/17]. Copyright © ICE Publishing, all rights reserved.