https://doi.org/10.14311/APP.2022.33.0238 Acta Polytechnica CTU Proceedings 33:238–244, 2022 © 2022 The Author(s). Licensed under a CC-BY 4.0 licence Published by the Czech Technical University in Prague STABILITY OF CONCRETE CONTAINING BLAST-FURNACE SLAG FOLLOWING EXPOSURE TO CYCLIC ELEVATED TEMPERATURE Ahmad Iravania, ∗, Volkert Feldrappeb, Andreas Ehrenbergb, Steffen Andersa a University of Wuppertal, Institute of Structural Engineering, Department Construction Materials, Pauluskirch Str. 11, 42285 Wuppertal, Germany b FEhS - Building Material Institute, Bliersheimer Str. 62, 47229 Duisburg, Germany ∗ corresponding author: a.iravani@uni-wuppertal.de Abstract. Concrete is widely used in constructions such as industrial floors or airducts in steel- and casting industry where it is often exposed to long-term or cyclic elevated temperatures. For these applications, thermal stability of concrete is of vital importance. The strength reduction due to elevated temperatures depends on the temperature level and concrete composition. In this study, the effects of blast-furnace slag cement (CEM III/A) and basaltic aggregates were investigated at temperatures 250◦C to 700 ◦C in comparison to conventional Portland cement (CEM I) containing quarzitic aggregates. The concretes were cyclically exposed to high temperatures. Special attention was paid to mass loss, residual compressive and residual flexural strength depending on type of cement and aggregate as well as the number of thermal cycles. Mass loss and strength loss increased with increasing maximum temperature level, as expected. It was generally observed that concretes containing CEM III/A displayed significantly higher residual mechanical properties for almost all temperature levels. Concretes containing a combination of CEM III/Awith basaltic aggregates showed significantly higher stability at elevated temperatures compared to other concrete mixtures. It is further shown that apart from the maximum temperature the number of thermal cycles is important for the residual mechanical properties. Keywords: Blast furnace slag, concrete, cyclic elevated temperature, residual compressive strength, residual flexural strength. 1. Introduction Durability is one of the most important characteris- tics of concrete constructions, but it can become deci- sive for elevated temperature environments including cyclic heating and cooling. Applications are for in- stance concrete floors or airducts in steel and casting industry. The use of concretes with higher long-term thermal stability when exposed to continuous cycles of heating and can reduce the cost and down time for repair or replacement. The behaviour of concrete exposed to single high- temperature events such as fire has been extensively studied [1]. However, the stability of concrete sub- jected to long-term or cyclic exposure to elevated temperatures is still a research topic [2–4]. Physi- cal and chemical changes in concrete such as micro- cracking caused by different thermal extension of ag- gregates and cement stone or dehydration of calcium- silicate-hydrate phases in cement stone during are ir- reversible changes. The hot strength of concrete was compared to its residual strength within the course of cooling process and after reaching ambient tempera- ture. Results demonstrated significant differences in strength of the hot, cooling and cooled down con- crete [5, 6]. Frangi et al. [3] reported additional 10% reduction in the compressive strength compar- ing residual compressive strength and hot compres- sive strength, as long as temperature remains below 600 ◦C [3]. Lucio-Matin et al. [7] report an increasing number of micro-cracks up to 25 cycles when concrete subjected to 1, 25 and 75 cycles between 290 ◦C and 550 ◦C. The cracks initiated in the first heating cycle, with the number of cracks increasing up to the 25th cycle, but remained unchanged henceforth. These findings demonstrate a kind of thermal fatigue in con- crete. Due to the comparatively low compressive strength of the concrete used here, explosive spalling is not dealt with in this study. It is commonly known that the type of cement, the aggregates as well as the interaction of cement stone and aggregates affect the behaviour of con- crete at high temperatures. Concretes containing basaltic aggregate or blast-furnace slag as aggregate have shown better thermal stability compared to con- ventional quarzitic concrete [8–10]. Generally, the lower thermal expansion of e.g. basalt or slags re- duce thermal stresses and micro-cracking in the con- crete structure. As cement paste shrinks during heat- ing, lower thermal expansion of the aggregates is sup- 238 https://doi.org/10.14311/APP.2022.33.0238 https://creativecommons.org/licenses/by/4.0/ https://www.cvut.cz/en vol. 33/2022 Blast-furnace Slag Concrete Sustainability Mix 1 Mix 2 Mix 3 Mix 4 Water 286 286 286 286 [kg/m3] CEM I 42.5R 550 550 − − [kg/m3] CEM III/A 42.5N − − 550 550 [kg/m3] Quartz 0/4 mm 1354 − 1354 − [kg/m3] Basalt 0/4 mm − 1534 − 1534 [kg/m3] Stabilizer (1% Cem.) 5.5 − 5.5 − [kg/m3] w/c-ratio 0.52 0.52 0.52 0.52 [−] Initial compressive strength at the age of 46 days 58.4 73.9 65.0 75.4 [MPa] Initial flexural strength at the age of 46 days 8.7 11.7 7.0 10.8 [MPa] Table 1. Overview of concrete mixtures. Concrete age Climat Remarks 1 day − Demoulding after casting 1 - 28 days Submerged in water Optimal condition for hydration 28 - 35 days 20 ◦C / 65 % r.h. Drying 35 - 38 days 50 ◦C in dry box Reducing water content 38 - 46 days 20 ◦C / 65 % r.h. Hygric homogenization 46 days − Initial compressive strength as reference 47 days Cyclic temperature exposure Starting high temperature exposure temperatures: 250 ◦C, 500 ◦C, 700 ◦C cycles: 1, 5 and 20 Table 2. Overview of conditioning process. posed to reduce micro-cracking. According to an- other studies, the residual compressive strength of concrete containing blast-furnace cement (CEM III) is significantly higher compared to concretes with Portland cement (CEM I), after single temperature exposure. Higher residual strengths are reported for concretes containing combinations of CEM III and basalt as aggregate following exposure to cyclic heat- ing [2, 4]. Furthermore, alkali-activated binders us- ing ground granulated blast-furnace slags have shown to improve resistance to high temperatures between 500 − 800◦C. This can be attributed to the ability of alkali-activated slags to chemically react forming calcium-silicate-hydrate phases (CHS) [10]. Therefore, the residual strength of concrete after exposure to cyclic heating/cooling provides essential questions for further investigations concerning the ef- fect of blast-furnace slag in cyclic high-temperature applications. It the aim of this study to quantitatively inves- tigate increases in residual compressive strengths if temperature resistant aggregates, namely basalt, as well optimized binders containing blast furnace slag are used in concrete subjected to thermal cycles at elevated temperatures. As reference concrete con- ventional concrete made of quarzitic aggregates and Ordinary Portland Cement are tested. Apart from the residual compressive strength, residual flexural strength and mass loss are considered. 2. Experimental program 2.1. Materials and curing of concrete mixes The experimental program is designed to show the effects of blast-furnace slag cement and basaltic ag- gregates on the residual strengths and mass loss of concrete. As reference concrete made of quarzitic aggregates and Ordinary Portland cement are used. The resulting matrix of concrete mixtures was fully varied (Table 1). In order to be able to compare the results w/c- ratio and lime content must kept constant. Mixture 1 is the reference containing Ordinary Portland Ce- ment (CEM I) and quarzitic aggregate. Mixture 2 is a combination of CEM I and basalt. Mixtures 3 and 4 have been developed to investigate the effect of ce- ment type, replacing CEM I by blast-furnace slag ce- ment (CEM III/A). Specimens were cast in standard prisms (40 × 40 × 160 mm3). The maximum grain size was limited to 4 mm. Both cements were supplied by the same manufacturer representing market-based products. Thus, the clinker in both cements is the same. 239 A. Iravani, V. Feldrappe, A. Ehrenberg, S. Anders Acta Polytechnica CTU Proceedings Figure 1. Sketch of the heating program, each complete cycle having a duration of 24 h. Figure 2. Residual compressive strength after 1 heating cycle (a), 5 heating cycles (b) and 20 heating cycles (c). Crosses (X) denote untestable mixtures due to complete degradation. In order to reach comparable moisture contents in all mixtures and to reduce the water content to a value usually reached after about 90 days (d) of cur- ing in air, pre-storage was scheduled (Table 2). After demoulding, specimens were stored in water for 28 d, followed by 7 d in standard climate (20 ◦C, 60 % rel- ative humidity). Afterwards concretes were dried at 50 ◦C for three days in order to reduce the physically bounded moisture. This pre-drying was implemented to reduce the impact of temperature-induced steam pressure build up on the strength of the specimens. Finally, the specimens were stored again in standard climate for 7 d to ensure a homogenous moisture dis- tribution. At the time of the first cycle of thermal exposure, specimens had an age of 45 − 48 d. Mass loss, residual flexural strength and residual compres- sive strength were measured after exposure to 1 and 5 heating cycles with the maximum temperature of 250 ◦C, as well as 1, 5 and 20 heating cycles at max- imum temperatures of 500 ◦C and 700 ◦C. 2.2. Heating programme / temperature regime Heating started at ambient temperature and in- creased with a heating rate of 5 K/min until the de- sired maximum temperature 250 ◦C, 500 ◦C or 700 ◦C was reached in a conventional tempering furnace. Temperature was kept constant for 4 h to ensure the homoňgeneity of the temperature in the specimen and to allow chemical reactions to take place. The cool- ing process was free but slow cooling for about 16 h until the temperature reached about 50 − 60 ◦C. Mea- surements and strength tests were always carried out after cooling down in ambient condition. Figure 2 shows a sketch of the cyclic heating regime. Regard- less the temperature level, one cycle took 24 h. There- fore, specimens which were heated up to 250 ◦C, were stored longer at ambient temperature before the next cycle compared to specimens heated to 700 ◦C. The specimens were examined in terms of water release (mass loss) immediately before testing the residual flexural and residual compressive strength. 3. Results and discussion 3.1. Residual compressive strength Generally, residual compressive strength depends on the maximum temperature during heating. It is demonstrated that the type of cement and aggregate significantly affect the residual compressive strength. Furthermore, the number of cycles affect the residual strengths significantly. The residual strengths after heating is expressed relative to the initial strength of concrete prior to thermal treatment at the concrete age of 45 − 48 d, as described in section 2.1. 3.1.1. Effect of maximum temperature and number of heating cycles Figure 2a (left) displays the residual compressive strength for all mixtures after a single cycle with max- imum temperatures of 250 ◦C, 500 ◦C and 700 ◦C. The known increasing reduction of residual strength [1, 5] with an increasing temperature level is confirmed. As expected, basaltic aggregates increase the residual strength compared to quartz at temperature levels of 500 ◦C and higher. 240 vol. 33/2022 Blast-furnace Slag Concrete Sustainability Figure 3. Residual compressive strength for concrete containing CEM I and basalt (a) and quartz (b). Figure 4. Combined effects of type of cement and aggregate on the residual compressive strength after cyclic heating with maximum temperatures of 500 ◦C and 700 ◦C. Figure 2b and 2c display the residual compres- sive strength after 5 and 20 thermal cycles respec- tively. An ongoing reduction of residual compres- sive strength can be seen after 5 heating cycles for all tested mixtures (Figure 2b). Similar to previous findings, the reduction of strength is significant for maximum temperatures of 500 ◦C and 700 ◦C. In ad- dition to the dehydration and degradation of the ce- ment paste, the increase in thermal expansion of ag- gregates, especially quartz, leads to the formation of micro-cracks resulting in higher degradation. Crack formation is further increased by alternating expan- sion and shrinkage while subjected to thermal cy- cles. Accordingly, stability of concrete significantly depends on the maximum exposed temperature as well as the number of exposed heating/cooling cy- cles. Similar results have been obtained within the current work, which are in accordance with findings e.g. of Alonso et al [5]. In Figures 3b and 3c it is furthermore obvious that none of the mixtures containing CEM I was testable after 20ăthermal cycles due complete degradation to powder. Mixtures containing CEM III/A Maintained considerable residual compressive strengths even af- ter 20 thermal cycles. 3.1.2. Effect of blast-furnace cement and basaltic aggregates The effect of basaltic and quarzitic aggregates on residual compressive strength are depicted in Fig- ure 3a and 3.b depending on the maximum tempera- ture. A comparison shows the higher thermal stabil- ity of the basalt-containing concretes subjected to 1 and 5 thermal cycles at both temperatures of 500 and 700 ◦C. The residual strength of concrete containing CEM I and basalt is about 55% and 30% after ex- posure to 5 heating cycles at 500 and 700 ◦C, respec- tively; whereas in the concrete containing quarzitic aggregate, the residual strength is lower than 20% in both cases. This better performance of the basaltic aggregates is attributed to the lower thermal expan- sion of basalt. Nevertheless, both mixtures exhibited complete degradation before 20 thermal cycles were reached. In order to characterize the effects of different types of cements, the residual compressive strengths are compared of mixtures with CEM I and CEM III/A are displayed in Figure 4. A pronounced decrease in compressive strength is observed for all mixtures be- fore 5 thermal cycles. Generally, mixtures containing CEM III/A show higher residual strengths compared to CEM I. The stability of CEM III/A containing mixtures is particularly visible after 5 thermal cy- cles, where CEM III/A obviously reach a maximum 241 A. Iravani, V. Feldrappe, A. Ehrenberg, S. Anders Acta Polytechnica CTU Proceedings Figure 5. Residual flexural strength of the concretes after 1 cycle (a), 5 cycles (b) and 20 cycles (c). Figure 6. Mass loss induced by cyclic heating for all tested concretes. degradation. It has reached a stable structure with- out further degradation. The authors currently have no a comprehensive explanation at hand. One expla- nation is that the powdered blast-furnace slag, which is itself stable against temperature as basalt, changes development of cracks and micro-cracks or the ther- mally induced shrinkage of the cement paste. An- other approach is a possible re-hydration tendency of CEM III/A due to the blast-furnace slag and mois- ture uptake in the air between temperature cycles. This re-hydration could result in a chemical closure of micro-cracks [11, 12]. 3.2. Residual flexural strength Apart from the residual compressive strength, the residual flexural strength was measured for all mix- tures, temperature levels and numbers of cycles. Fig- ure 6 displays the results after exposure to 1 (Fig- ure 6.a), 5 (Figure 6.b) and 20 (Figure 6.c) thermal cycles. A significant reduction of the residual flexural strength occurs after any thermal treatment. The re- duction is determined both by the temperature level as well as the number of thermal cycles. The loss of residual flexural strengths is generally higher com- pared to the reduction of the compressive strength. Figure 5.b and 5.c show significant stability of con- cretes containing CEM III/A after 5 and 20 cycles. It is worth mentioning that compared to mixtures with quartz, flexural strengths of mixtures with basalt are higher both, before (see Table 1) and after heating to temperatures 500 ◦C and 700 ◦C. This can be ex- plained by to better mechanical properties of basalt. 3.3. Mass loss after temperature exposure Figure 6 depicts the percentage of mass loss in all con- crete mixtures and temperature regimes. The mass loss increases with the maximum temperature. The increase of mass loss is attributed first to the release of 242 vol. 33/2022 Blast-furnace Slag Concrete Sustainability physically bound water, followed by beginning of de- hydration of CSH-phases and the subsequent release of further chemically bound water at higher temper- atures. This corresponds to the findings presented in section 3.1.1, where the maximum temperature sig- nificantly affected the residual compressive strength for all mixtures. As observed, and described in [12], mass does not change significantly after the first cycle and up to the 5th cycle of heating at temperatures of 500 and 700 ◦C for almost for all mixtures. Exposure up to 20 thermal cycles at 500 and 700 ◦C, indicate a slight increase of mass loss for concretes containing CEM III/A. The pronounced mass loss in the first cycle results in shrinkage of the cement stone and subsequent for- mation of a cracks. This is the reason for high degra- dation in residual compressive and flexural strengths in the first cycle. This hypothesis is in accordance with findings of Lucio-Martin et al. [7]. The effect of cement type on the concrete mass loss is evident from Figures 6b and 6d, where the higher mass loss for mix- tures containing CEM III/A compared to CEM I is observed. This might further imply that the bonding of water to the cement paste differs with the type of cement. From the results presented here it could be expected, that water is easier releasable in concrete containing CEM III/A cements. 4. Conclusions From the generally known effects of an increasing re- duction in residual strengths with increasing temper- ature, the following conclusions can be drawn: • The residual compressive and residual flexural strength decrease strongest in the first thermal cycle, especially at temperatures of 500 ◦C and higher. The further decrease depends on the type of cement and type of aggregate. The performance gets better if either CEM III/A is chosen as cement or basalt as aggregate. The combination of both shows best results in terms of residual strengths. • At higher temperature levels, concretes containing basaltic aggregates perform better than quarzitic aggregates. • Blast-furnace slag cement (CEM III/A) can signif- icantly improve the residual compressive and flex- ural strengths, particularly after exposure to 20 thermal cycles and in combination with thermally stable aggregates such as basalt. • The mass loss increases with an increasing max- imum temperatures, but is not significantly im- pacted by the number of thermal cycles. It was furthermore observed that the mass loss of CEM III/A is generally about 1 % by mass higher com- pared to CEM I cement. 4.1. Residual flexural strength Apart from the residual compressive strength, the residual flexural strength was measured for all mix- tures, temperature levels and numbers of cycles. Figur 5 displays the results after exposure to 1 (Fig- ure 5a), 5 (Figure 5b) and 20 (Figure 5c) thermal cy- cles. Acknowledgements The IGF project 20318N of the Research Association "VDEh-Gesellschaft zur Förderung der Eisenforschung mbH" is funded by the AiF within the programs for spon- sorship by Industrial Joint Research (IGF) of the Federal Ministry of Economic Affairs and Energy (BMWI) based on an enactment of the German Bundestag. We further acknowledge our industrial partners for their project sup- port. References [1] U. Schneider. Verhalten von Beton bei hohen Temperaturen, Technische Universitaät Braunschweig, 1982.https://publikationsserver.tu-braunschwei g.de/servlets/MCRFileNodeServlet/dbbs_derivate _00043616/3444-1790.pdf. [2] S. Anders. Comparison of residual strengths of concretes with quarzitic, limestone and slag sand constituents after cyclic high-temperature exposure. Concrete Repair, Rehabilitation and Retrofitting, CRC Press, p. 223-4, 2015. [3] A. Frangi, C. Tesar, M. Fontana. Tragwiderstand von Betonbauteilen nach dem Brand. Bauphysik 28(3):170- 83, 2006. https://doi.org/10.1002/bapi.200610017. [4] H. Y. Wang. The effects of elevated temperature on cement paste containing GGBFS. Cement and Concrete Composites 30(10):992-9, 2008. https: //doi.org/10.1016/j.cemconcomp.2007.12.003. [5] K. D. Hertz. Concrete strength for fire safety design. Magazine of Concrete Research 57(8):445-53, 2005. https://doi.org/10.1680/macr.2005.57.8.445. [6] E. Klingsch, A. Frangi, M. Fontana. Experimental Analysis of Concrete Strength at High Temperatures and after Cooling. Acta Polytechnica 49(1), 2009. https://doi.org/10.14311/1087. [7] T. Lucio-Martin T, J. Puentes, M. C. Alonso. Effect of geometry in concrete spalling risk subjected to high temperatures for thermal inertia studies, Proceedings of the 6th International Workshop on Concrete Spalling due to Fire Exposure, p. 71-80, 2019. [8] K. Endell. Versuche über Längen-und Gefügeänderung von Betonzuschlagstoffen und Zementmörteln unter Einwirkung von Temperaturen bis 1200C: ausgeführt im Laboratorium der Dozentur für bauwissenschaftliche Technologie an der Technischen Hochschule zu Berlin in den Jahren 1928/29: Ernst; 1929. [9] Interner Abschlussbericht für die Verwendung eines hitzebeständigen Betons aus Hochofenstückschlacke mit Hochofenzement für Fundamente eines Trockenkühlers Internal test report FEhS Institut für Baustoff-Forschung (Duisburg, Germany). 243 https://publikationsserver.tu-braunschweig.de/servlets/MCRFileNodeServlet/dbbs_derivate_00043616/3444-1790.pdf https://doi.org/10.1002/bapi.200610017 https://doi.org/10.1016/j.cemconcomp.2007.12.003 https://doi.org/10.1680/macr.2005.57.8.445 https://doi.org/10.14311/1087 A. Iravani, V. Feldrappe, A. Ehrenberg, S. Anders Acta Polytechnica CTU Proceedings [10] Interner Abschlussbericht für Kokerei Dünkirchen Beton mit Hochofenstückschlacke Internal test report FEhS Institut für Baustoff-Forschung (Duisburg, Germany). [11] C. Alonso, L. Fernandez. Dehydration and rehydration processes of cement paste exposed to high temperature environments. Journal of Materials Science 39(9):3015-24, 2004. https: //doi.org/10.1023/b:Jmsc.0000025827.65956.18. [12] A. Iravani and S. Anders, Effects of cement type, aggregate type and concrete age on the mass loss of concrete exposed to elevated temperature, 5th International Workshop on Concrete Spalling due to Fire Exposure in, pp.63-72, 2017. 244 https://doi.org/10.1023/b:Jmsc.0000025827.65956.18