Construction and Building Materials 204 (2019) 458–467 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat Experimental methods to determine the feasibility of steel slags as supplementary cementitious materials Ying Wang a,b, Prannoy Suraneni a,⇑ a b Department of Civil, Architectural, and Environmental Engineering, University of Miami, Coral Gables, FL 33146, USA School of Civil and Resources Engineering, University of Science and Technology Beijing, 100083, China h i g h l i g h t s Hydraulic and pozzolanic behavior of basic oxygen furnace slag (BOFS), ladle furnace slag (LFS), and ground granulated blast furnace slag (GGBFS) are studied. Multiple tests are performed to determine the feasibility of these slags as SCMs. GGBFS shows the best performance; BOFS and LFS should not be used as SCMs as-is. The combined use of test methods is more accurate. a r t i c l e i n f o Article history: Received 12 November 2018 Received in revised form 11 January 2019 Accepted 27 January 2019 Keywords: Steel slag Supplementary cementitious materials Isothermal calorimetry Thermogravimetric analysis a b s t r a c t Limited applications for the large amounts of steel slags produced globally exist. In this study, the hydraulic and pozzolanic behavior of basic oxygen furnace slag (BOFS) and ladle furnace slag (LFS) are studied and contrasted with the behavior of ground granulated blast furnace slag (GGBFS). Pozzolanic test, isothermal calorimetry, thermogravimetric analysis, and compressive strength test are performed to determine the feasibility of these slags as supplementary cementitious materials (SCMs). The pozzolanic test is performed by measuring the heat release and calcium hydroxide consumption of slagcalcium hydroxide blends at 50 °C and pH 13.5. The other tests are performed on cementitious pastes with 20% slag replacement by mass at 23 °C and 50 °C. GGBFS shows the best performance among all slags in the tests, followed by BOFS, followed by LFS. It is suggested that BOFS and LFS should not be used as SCMs as-is. The poor performance of LFS is due to flash setting, likely caused due to the rapid reaction of C12A7, which causes hydration retardation and poor strength gain. The BOFS shows moderate performance as it has low hydraulic activity, likely explained by its low amorphous content and the presence of low amount of reactive crystalline phases. An important finding from this study is that in order to state whether materials could be used as SCMs or not, the combined use of pozzolanic test, isothermal calorimetry, and compressive strength test is more accurate than the use of only one test method. While the results obtained may here not be broadly generalizable to other materials due to variations in slag chemistry, the approach suggested may be generalizable to obtain insights into whether any powder may be used as SCM. Ó 2019 Elsevier Ltd. All rights reserved. 1. Introduction Slagging agents are commonly used in crude iron and crude steel manufacture to strip impurities from the iron ore in the blast furnace and from the crude iron and scrap steel in the steel furnace. The slagging agents combine with the impurities to form iron and steel slags, which are separated from the metals. The slags are ⇑ Corresponding author. E-mail address: [email protected] (P. Suraneni). https://doi.org/10.1016/j.conbuildmat.2019.01.196 0950-0618/Ó 2019 Elsevier Ltd. All rights reserved. cooled and processed to form aggregates or powders, which can potentially be used in the construction industry. Different slag types may be obtained depending on the type of furnace and the cooling method employed. Exact data are unavailable on slag production, but estimates indicate that worldwide iron slag and steel slag outputs were approximately 330 million tons and 200 million tons, respectively (2017 data) [1]. Numerous applications for aircooled iron slag exist, including as aggregates in concrete, asphaltic paving, fill; and as a feed for cement kilns [2–11]. Ground granulated blast furnace slag (GGBFS) is used as a partial replacement 459 Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 for Portland cement in concrete mixtures or in blended cements [12–14]. GGBFS is highly amorphous and is a hydraulic material. Concrete containing GGBFS usually has longer setting times and lower early strength but shows higher later strength, denser microstructure and better durability compared with Portland cement concrete [10]. In contrast to iron slags, uses for steel slags such as basic oxygen furnace slag (BOFS), electric arc furnace slag (EAFS), and ladle furnace slag (LFS), are somewhat more limited. These slags typically contain free lime (CaO) that can result in an expansion that must be dealt with through appropriate means to ensure its use in construction [2–22]. Steel slags may also contain free MgO, which is deleterious as it may hydrate, expand, and cause reduction in strength at later ages. BOFS can contain significant free CaO and free MgO, while these contents are typically lower for EAFS and LFS [23]. The use of such slags as supplementary cementitious materials (SCMs) requires a detailed understanding of the slag properties due to their special production processes and chemistry [6]. BOFS has been suggested to be somewhat reactive due to its composition and concrete containing BOFS has greater corrosion resistance than that containing conventional Portland cement [8,24–26]. LFS shows cementitious behavior, especially in the presence of alkaline activators [14,27], but when mixed with water, strength gain may be poor, due to hydration to crystalline aluminate phases, which may cause rapid setting in addition to conversion issues (similar conversion processes to those seen in calcium aluminate cements may result due to the presence of calcium aluminate phases) [27]. While steel slags may have crystalline calcium silicate phases, they may not necessarily be reactive [28]. The reactivity of EAFS has been reported to be rather weak [29]. Steel slags have also been used in masonry mortars and self-compacting concrete providing mixtures with adequate workability, shrinkage, strength, and durability [8]. As studies on steel slags are somewhat limited, better understanding and utilization of steel slags are of great significance in light of shortages in conventional SCMs and in reducing landfilled waste. In order to determine whether steel slags could be used as SCMs, information about their pozzolanic and hydraulic behavior must be obtained. This is typically done using the strength activity index test [15,22] or other similar tests that monitor strength gain [30–32], however, numerous factors apart from pozzolanic and hydraulic behavior affect strength gain [9]. For example, extremely finely ground limestone can pass a strength test, even though it is more appropriately classified as a filler rather than as a SCM. It is important to differentiate fillers and SCMs, because replacement levels are limited for fillers and fillers may not provide later-age contributions to durability that SCMs provide. Other pozzolanic tests include those that monitor calcium hydroxide consumption, such as Chappelle and Frattini tests [32,33], however, it is not clear if such methods are appropriate for slags, which are more hydraulic. A promising alternative to determine the pozzolanic and hydraulic behavior of SCMs is a test that measures the heat release and calcium hydroxide consumption of SCMs blended with calcium hydroxide in a high pH solution [33,34]. This test offers significant advantages over conventional pozzolanic tests as it provides information about both pozzolanic and hydraulic behavior for a wide variety of SCMs. In this study, the pozzolanic and hydraulic behavior of BOFS and LFS and their effects on the properties of cementitious pastes are quantified and compared with GGBFS. This is done using four methods: the aforementioned pozzolanic test, isothermal calorimetry, thermogravimetric analysis, and strength gain [33]. There are two main objectives in this study. The first objective is to better quantify BOFS and LFS for use as SCMs in concrete. The second, and more general objective is to develop an approach using a series of tests which may be used to determine whether any powder can be used as SCM. 2. Materials and methods 2.1. Materials Blast furnace slag (BFS), basic oxygen furnace steel slag (BOFS), flue gas desulfurization gypsum from Xingtai (Hebei province) and ladle furnace slag (LFS) from Liuzhou (Guangxi province) were used in this study. An ASTM Type I/II cement was also used in the study. The bulk oxide composition (as obtained from a calibrated X-ray fluorescence device (FluoroMax-4, Horiba)) and Blaine fineness of the cement, gypsum and the slags are shown in Table 1. All slags have considerable contents of Ca, Al, and Mg. The presence of free CaO and free MgO may cause expansion during the hydration process [23]. The LFS used here is considerably different from conventional LFS materials (although it should be pointed out that steel slags show considerable variation in their chemistries); it has relatively higher amounts of Al and lower amounts of Si. This is because Al bars were used to extract and purify the steel during the manufacture (information provided by the manufacturer). As the materials have similar fineness values, differences in reactivity due to differences in fineness are unlikely. Fig. 1 shows X-ray diffraction (XRD) patterns (Rigaku D/Max-RB diffractometer, Rigaku, Cu Ka anode) for the three slags. As quantitative XRD was not performed, evaluations about amorphous and crystalline nature have some uncertainty associated with them. The GGBFS is mostly amorphous (based on the amorphous hump and only one crystalline peak) but shows the presence of crystalline Fukalite (Ca4Si2O6(CO3)(OH)2). Due to its amorphous nature and its chemical composition with large amounts of CaO, SiO2, and Al2O3 (Table 1 and Fig. 1), it is expected that the GGBFS would behave as an SCM, consistent with literature [12–14]. The BOFS appears to be far less amorphous than the GGBFS, as an obvious amorphous hump is not detected [15]. It shows the presence of numerous crystalline phases, including b-C2S, c-C2S, C3S (monoclinic), CaO (free CaO), MgO (free MgO), and RO phase (a continuous solid solution composed of divalent metal oxides, such as FeO, MgO, MnO, CaO, etc.). The Fe2O3 content is high (25% by mass), although no significant amount of metallic iron was observed. As C3S and C2S (note that in cement chemistry notation C = CaO, S = SiO2, and A = Al2O3) react with water and result in strength gain in portland cement, the BOFS could potentially be used as an SCM (although classically SCMs are amorphous aluminosilicates and not crystalline) [9]. The LFS appears to have amorphous content lesser than GGBFS but greater than BOFS (semi-quantitatively based on the sizes of the amorphous humps [15]). Strong peaks of the crystalline phase Mayenite (Ca12Al14O33, C12A7) are detected in the XRD Table 1 Oxide composition (mass %) and Blaine fineness (m2/kg) of the raw materials. Material CaO SiO2 Al2O3 MgO Fe2O3 Na2O K2O SO3 Blaine fineness GGBFS BOFS LFS Cement Gypsum 46.54 41.03 55.86 64.33 48.13 29.78 16.15 0.73 20.46 2.62 12.18 4.31 36.91 4.68 0.96 6.00 3.64 3.40 1.17 1.48 1.22 27.03 1.16 3.94 0.48 0.45 0.20 0.00 0.16 0.08 0.41 0.17 0.01 0.35 0.21 0.36 0.16 0.40 2.47 44.03 420 400 400 399 373 460 Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 Fig. 1. X-ray diffraction (XRD) patterns for three slags. pattern, in addition to peaks of CaO and MgO. As it seems to have a moderate amorphous content, the LFS is also a potential SCM. C12A7 is a component of calcium aluminate cement (CAC) and other rapid-set cements; it reacts extremely rapidly with water in the absence of gypsum [27,35–37]. In the absence of gypsum, some have suggested that the initial hydration products of C12A7 will convert from metastable C2AH8 to stable C3AH6, causing an increase in porosity and a consequent decrease in strength [27]. BOFS and LFS show minor peaks of free CaO and free MgO, which may be deleterious and cause reduction in strength. The phases and amorphous contents of the BOFS and LFS are broadly consistent with literature [4,6,8,10,14,27]. 2.2. Experimental methods Reagent grade calcium hydroxide and potassium hydroxide were used to prepare the model systems for the pozzolanic test [33]. The model system consisted of a 3 to 1 mass ratio of calcium hydroxide and SCM mixed with solutions of 0.5 M potassium hydroxide, at a liquid-to-solid ratio of 0.9. A total of 40 g of material was mixed in a plastic container using a spatula for 4 min. Immediately after mixing, 6–7 g of material was transferred into a glass ampoule, which was then placed in an isothermal calorimeter (TAM Air, TA Instruments) that had been preconditioned at 50 °C. Due to temperature differentials, it typically took 45 min for the signal to stabilize, and this part of the signal was not recorded. For highly reactive SCMs, this may result in a slight underestimation of the total heat release. The heat flow from the samples was recorded for 240 h, after which the samples were removed from the calorimeter. Approximately 50 mg of material was taken from the sample and thermogravimetric analysis (Q50, TA Instruments) (TGA) was performed on it to determine calcium hydroxide contents using the tangential method [38]. The analysis was performed by heating from 23 to 600 °C at a rate of 10 °C/minute in a nitrogen purged atmosphere. The calcium hydroxide consumption is based on a comparison of the final and initial amounts of calcium hydroxide normalized to the mass of SCM in the model systems. Cementitious pastes were made using a 20% SCM mass replacement level and a water-to-cementitious ratio (w/cm) of 0.40. This was done for the three slags and a control cement paste without any slag. A total of 40 g of material was hand mixed in a plastic container using a spatula for 4 min. These samples were hand mixed in order to better compare the results with the pozzolanic test. Immediately after mixing, 6–7 g of material was transferred into a glass calorimeter ampoule, which was then placed in an isothermal calorimeter (TAM Air, TA Instruments) that had been preconditioned at two temperatures 23 ± 0.05 °C and 50 ± 0.05 °C. The heat flow from the samples was recorded for 168 h, after which time the samples were removed from the calorimeter. For samples at 50 °C, approximately 50 mg of material was taken from the sample and thermogravimetric analysis (Q50, TA Instruments) was performed on it to determine calcium hydroxide contents using the tangential method [38]. The analysis was performed by heating from 23 to 600 °C, at a rate of 10 °C/minute in a nitrogen purged atmosphere. Compressive strength testing was also performed on the cementitious pastes. Pastes with the same mixture design as earlier were mixed using a paddle mixer. The mixing procedure is according to ASTM C305-14 [39]. These pastes were cast into 5 cm cubic molds and cured in a moist chamber with a relative humidity greater than 95% for one day. Cubes were then demoulded and then cured inside a solution of simulated pore solution at 23 ± 1 °C. The composition of the simulated pore solution was 0.2 M NaOH + 0.6 M KOH + 0.03 M Ca(OH)2 [40–42]. At 1, 7, 28, and 91 days, three cubes were tested in compression and presented strength results are averages of the three replicates. After samples were tested for compressive strength, small pieces from the core of the samples were ground, and thermogravimetric analysis (Q50, TA Instruments) was performed on it to determine calcium hydroxide contents in a manner similar to that outlined in the previous paragraph. Pastes containing LFS at w/cm 0.40 showed rapid setting behavior, perhaps due to the C12A7 reaction [37,43]. For compressive strength and thermogravimetric analysis, the LFS pastes were tested at a w/cm of 0.50 and hand mixed, similar to the hand mixing adopted for the pozzolanic testing. It was not possible to easily mix pastes with LFS at w/cm 0.40 or mix using a paddle mixer for compressive strength tests because of rapid setting, which did not allow for the casting of paste cubes. Therefore, results presented in this work for compressive strength and thermogravimetric analysis are for the LFS at w/cm 0.50. While the use of different w/cm values may present a possible issue, isothermal calorimetry and thermogravimetric analysis results of the LFS pastes at w/cm 0.40 and 0.50 were compared and showed little effect of the w/ cm on the paste behavior, as discussed later in the text. In order to investigate the rapid setting of the LFS, pastes with w/cm 0.40 were mixed with additional gypsum of 6% and 12%. Isothermal calorimetry till 7 days and thermogravimetric analysis (at 7 days) at 23 °C was carried out in a manner similar to that described earlier. A summary of the testing that was performed is shown in Table 2. 3. Results and discussion 3.1. Pozzolanic test Fig. 2 shows the heat release plotted against the calcium hydroxide consumption for the slags. In addition, SCM classification and corresponding regions in the curve for these classifications from previous results are also shown in the graph for comparison [34]. Both isothermal calorimetry and thermogravimetric analysis results had less than 5% coefficient of variability in replicate testing. The slags show relatively low calcium hydroxide consumption, ranging from 11.6 to 32.2 g/100 g SCM. The slags ordered by increasing order of calcium hydroxide consumption are LFS < BOFS < GGBFS. The heat release values range from 217.5 to 496.8 J/g SCM, and slags ordered by increasing order of 461 Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 Table 2 Summary of testing regime. Test Mixture design Materials Conditions Ages Pozzolanic test Slag: CH = 1:3, pH 13.5 All slags 50 °C, sealed cured Isothermal calorimetry Isothermal for 10 days, TGA at 10 days For 7 days 23 °C, sealed cured 50 °C, sealed cured 20% SCM, w/cm = 0.40; w/cm = 0.50 23 °C, cured in pore 1, 7, 28, 91 days for LFS solution 50 °C, sealed cured 7 days 23 °C, cured in pore 1, 7, 28, 91 days solution Additional testing performed on LFS is isothermal calorimetry till 7 days and TGA at 7 days at 23 °C for pastes with w/cm 0.40, w/cm 0.50, and w/cm 0.40 with 6% and 12% additional gypsum. Additional testing performed on BOFS is TGA at 7 and 28 days at 23 °C and 50 °C for BOFS pastes mixed with water at a water-to-solids ratio of 0.40. TGA Compressive strength Additional tests 20% SCM, w/cm = 0.40 Cement pastes with slags and control paste (no slag) Fig. 2. Pozzolanic test results for the tested slags. Classification results (empty circles and labels) from prior work [33,34] are also provided for comparison. heat release are BOFS < LFS < GGBFS. The calcium hydroxide consumption of the LFS is negative, meaning that LFS generates calcium hydroxide in the pozzolanic test, possibly because of the reaction of the free lime present in the slag. GGBFS has the highest calcium hydroxide consumption, which could be because of its higher amorphous SiO2 + Al2O3 content. Heat release for GGBFS and LFS are similar and significantly greater than BOFS. This may be because of the greater amounts of amorphous phases in the GGBFS and LFS compared to the BOFS [15]. While the BOFS contains reactive crystalline phases such as C3S, it may be that the more reactive phases are present in lower quantities, which can be confirmed using quantitative XRD. Based on the classification of SCM reactivity [33,34], all the slags are hydraulic materials; BFS and LFS are more reactive and BOFS is less reactive. 3.2. Isothermal calorimetry Fig. 3 shows the heat flow (mW/g cementitious material) and cumulative heat release (J/g cementitious material) for the slags and the control cement at 23 ± 0.05 °C. The variation in ultimate heat release for replicate samples was less than 3%. The LFS shows a significant retardation of the hydration; its main hydration peak is at approximately 28 h, while the other materials have a main peak at approximately 9 h. The order of the heat release from greatest to least at 168 h is LFS > Cement > GGBFS > BOFS. While the LFS strongly retards early hydration, at later ages, it has the Fig. 3. The (a) heat flow and (b) cumulative heat release of cementitious pastes at 23 °C. greatest heat release. This is likely because of the large amount of the C12A7 in the LFS, which causes a sulfate imbalance [44– 48]. The reaction of the C12A7 releases a significant amount of aluminum ions, which may be adsorbed on calcium silicate surfaces, thereby retarding silicate dissolution and cement hydration [47,49–51]. These results suggest that BOFS could potentially be used as SCM, as its heat flow and heat release are in general similar 462 Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 to BFS (which is in turn similar to the cement). The use of GGBFS and BOFS reduces the total heat release somewhat and could potentially be used to solve the issue of thermal cracking due to the different thermal stresses in mass concrete structures [52]. GGBFS is more reactive than BOFS, therefore, the heat release of the GGBFS is closer to the cement than the BOFS. As the hydration continues, the heat flow of the GGBFS gradually exceeds that of the cement, so the heat release gradually approaches the cement. This is because of hydration and pozzolanic reactions of the GGBFS with alkali hydroxide and calcium hydroxide [13]. It is not obvious to identify SCMs using isothermal calorimetry using just heat flow and heat release, though more complicated measures including isolation of slag contribution to heat release may help in such identification [53]. Fig. 4 shows the heat flow (mW/g cementitious material) and heat release (J/g cementitious material) for the slags and the control cement at 50 ± 0.05 °C. The LFS shows a significant retardation of the hydration; its main hydration peak is at approximately 13 h, while the other materials have a main peak at approximately 3 h. The order of the heat release from greatest to least at 168 h is GGBFS > LFS > Cement > BOFS. Compared with results at 23 °C, all slags experienced significant thermal activation (compared to the cement), as the heat release of GGBFS and LFS exceeded that of the cement paste, and the heat release of BOFS was almost equal to the cement paste. At 50 °C, the hydration kinetics of all pastes Fig. 4. The (a) heat flow and (b) cumulative heat release of cementitious pastes at 50 °C. are accelerated, as reflected in lower time to reach the maximum heat release. The maximum heat release values of BF, BOFS, LFS, and cement paste at 50 °C are increased by 26%, 20%, 11% and 9%, respectively, when compared to values at 23 °C. The maximum heat flow of GGBFS, BOFS, and cement are approximately four times that at 23 °C, while the heat flow of LFS is two times that at 23 °C. 3.3. Thermogravimetric analysis (TGA) The calcium hydroxide evolution of the paste samples at 23 °C is shown in Fig. 5(a). In general, for tested samples, calcium hydroxide content had less than 3% coefficient of variability in replicate testing. For the cement paste, the calcium hydroxide content increases sharply from 1 to 7 days, and then somewhat levels off. GGBFS initially shows a greater calcium hydroxide content than the cement paste, likely due to filler effect [54–56]. At later ages, GGBFS shows lower calcium hydroxide amounts than cement, due to dilution and calcium hydroxide consumption by the slag [33,34]. BOFS shows slightly higher calcium hydroxide amounts than cement at early ages (due to the filler effect), and at later ages, shows a calcium hydroxide content between the GGBFS and the cement. This may be expected from the pozzolanic Fig. 5. (a) Calcium hydroxide content and (b) calcium hydroxide consumption of cementitious pastes. Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 test (Fig. 2), as the BOFS has very little calcium hydroxide consumption. The LFS initially shows very low calcium hydroxide content (3.96 g/100 g paste at 1 day). This is not due to calcium hydroxide consumption, but due to the retardation of cement hydration (Fig. 3). At later ages, LFS and BOFS show similar calcium hydroxide values, which is expected from the pozzolanic test (Fig. 2). The difference between the calcium hydroxide amounts in the cement pastes (no SCM) and the cementitious paste (with SCM) divided by the amount of the SCM gives the calcium hydroxide consumption for each SCM in cementitious pastes. When considering the contribution of the cement in generating calcium hydroxide amount in the cementitious paste, the filler effect should be taken into account [54–56]. Here, based on the previous results [51,57], a filler effect of 5% (an increase in the degree of hydration of cement by 5%) is assumed. Fig. 5(b) shows the calcium hydroxide consumption at different ages for the tested materials. GGBFS has an increasing calcium hydroxide consumption during the testing duration, although the rate reduces as time increases. The calcium hydroxide consumption of BOFS decreases and becomes more negative as time increases, indicating that BOFS is likely producing calcium hydroxide in cementitious pastes with SCM. This may be because of the reaction of the free lime in the BOFS (Fig. 1). The calcium hydroxide consumption of LFS is very large at 1 day, which is not an actual consumption, but due to retardation of the cement hydration. At later ages, the LFS has a negative calcium hydroxide consumption, suggesting that, similar to BOFS, it is producing calcium hydroxide under these conditions. As mentioned earlier, in this set of tests, the w/cm value for LFS pastes is 0.50, which is different from the other pastes, 0.40. Fig. 6(a) shows the effect of temperature on calcium hydroxide contents in the pastes at 7 days. It is noted that these samples are not under the same curing conditions as pastes at 23 °C are cured in saturated pore solutions whereas pastes at 50 °C are sealed cured. As the temperature increases, the calcium hydroxide content of the cement paste increases. However, the calcium hydroxide content of the BOFS and LFS increases only slightly. The calcium hydroxide of the GGBFS remains almost unchanged. The calcium hydroxide consumption, as defined earlier, is shown in Fig. 6(b). As the temperature increases, the calcium hydroxide consumption of the GGBFS increases, or, in other words, the GGBFS consumes more calcium hydroxide produced by the cement. Similar observations have been made in literature [52]. The BOFS shows slightly greater calcium hydroxide consumption as temperature increases. The exact reason for this is unclear and may be due to a complex interplay between the acceleration of cement and BOFS hydration as the temperature increases. As the temperature increases, the LFS shows significantly greater calcium hydroxide consumption, likely because the cement hydration is not as retarded at a higher temperature. 463 Fig.6. (a) Calcium hydroxide content and (b) calcium hydroxide consumption of cementitious pastes at 7 days. 3.4. Strength gain Fig. 7 shows the compressive strength of the different pastes cured at room temperature. In general, compressive strength results had less than 5% coefficient of variability in replicate testing. At 1 day, cement paste had the highest compressive strength (21.2 MPa). The strength of GGBFS was slightly lower than the cement paste (18.4 MPa), and the strength of the BOFS was lower than the GGBFS (14.2 MPa). The strength of the LFS paste was extremely low (2.2 MPa). As hydration continued, the strength of GGBFS and cement paste samples continued to increase, and the GGBFS strength exceeded the cement paste strength at 7 days. BOFS showed a considerable increase in strength till 7 days. The strength of the BOFS paste also increased at later ages, but ultimate values were significantly lower than cement paste and GGBFS. Fig. 7. The compressive strength evolution of cementitious pastes. Ultimate strength (91 days) values of GGBFS (69.6 MPa) were approximately 10% greater than the cement paste (63.2 MPa) and 30% greater than the BOFS (48.5 MPa). The LFS pastes increased in strength from 1 to 7 days, and the strength increased only slowly 464 Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 after 7 days. However, the strength of these pastes was always significantly below 15 MPa [15]. The strength gain of the GGBFS is explained due to its hydraulic nature, due to the high amorphous content (which is likely a calcium aluminosilicate glass). This phase dissolves and reacts with the high pH pore solution and with calcium hydroxide to form C-S-H and other hydration products, thereby increasing the compressive strength [4,13,15,22,31]. The results of XRD show that the amorphous content in the BOFS is lower than in the GGBFS. The BOFS seems to have lower hydraulic nature than GGBFS, either because of a lower amorphous content or because of a lower content of reactive crystalline materials; this low hydraulic activity is confirmed by the pozzolanic test (Fig. 2) [22]. If the BOFS is unreactive or only slightly reactive, then the strength of the paste should be approximately 20% lower than the control cement paste. However, at 28 and 91 days, the reduction is between 23 and 30%, suggesting that another mechanism may also be responsible for the poor strength gain behavior of the BOFS paste. This could be due to the reaction of free CaO and free MgO, which are expansive and cause strength loss. TGA performed on BOFS paste mixed with water at a liquid–solid ratio of 0.40 at 7 and 28 days showed small amounts of calcium hydroxide at 50 °C but not at 23 °C. Based on this, a maximum of 2.6% of free CaO in the BOFS was estimated, which could additionally have contributed to the poor performance of the BOFS paste. No magnesium hydroxide was detected, suggesting that the free MgO content is negligible. A number of studies have shown that certain BOFS may also form a coating that is different in composition from the interior of the slag particle, which may hinder reactivity and the strength gain of BOFS pastes, which may be an additional contributor to the poor performance [58,59]. The poor strength gain behavior of the LFS is likely due to the presence of large amounts of C12A7, which causes flash set, sulfate imbalance, and retarded cement hydration [27,35–37,60]. Potential conversion of C2AH8 to C3AH6, which is known in LFS-water mixtures, may also contribute to the strength loss [27]. It seems that under these conditions, the strength gain does not ultimately recover. The relative strength gain of the pastes with slag is calculated as the ratio of the compressive strength of the cementitious pastes and the compressive strength of the cement paste [15,22]. Some have suggested a relationship between relative strength gain and slag basicity [15,22]. The latter is calculated as the sum of the mass fractions of CaO and MgO divided by the sum of the mass fraction of SiO2 and Al2O3. In this study, an obvious relationship between relative strength gain and basicity was not observed, which disagrees with previous research [13,15,22]. This discrepancy is explained by the fact that different slags have different amounts of amorphous phases and here only bulk chemical composition is known. In addition, some of the tested slags cause sulfate imbalance, which changes the strength gain behavior. As the strength gain behavior of the BOFS and LFS is poor, it is not suggested to use these materials as-is as SCMs. Fig. 8 shows the relationship between the compressive strength of the pastes and the calcium hydroxide consumption at each corresponding age for pastes ages cured at 23 °C. While there is significant scatter in the data, the paste strength gains show a roughly positive correlation with the calcium hydroxide consumption, that is, the more calcium hydroxide the slags consume, the greater the strength increases, although the relationship is not linear. In this analysis, the LFS data point to the far right with artificially high calcium hydroxide consumption is excluded. This result can be explained by reaction of amorphous phases in the slag with calcium hydroxide resulting in C-S-H gel, which increases the paste strength. The results obtained agree with literature on pozzolanic action of other SCMs [31,32,61] and reduction of calcium hydroxide in cementitious pastes with fly ash [62]. Fig. 8. Relationship between compressive strength and calcium hydroxide consumption. Relationships between strength and other parameters may also be considered. Fig. 9(a) shows the relationship between heat release from pozzolanic test and compressive strength of cementitious pastes, at ages of 1 and 7 days. As the heat release increases, the compressive strength of all samples increases; however, all materials show different relationships, suggesting that a unified relationship between heat release in the pozzolanic test and compressive strength is not possible for the tested materials. Or, in other words, the pozzolanic test heat release is not an accurate indicator of the compressive strength, when considering all slags tested here. Others have shown that the heat release in pozzolanic tests may be correlated with later age strength gain [30–32] (for example at 91 days, as the pozzolanic testing is accelerated). However, it is clear when comparing the GGBFS and LFS strength gains that such a relationship is not possible in our case. Fig. 9(b) shows the relationship between heat release from isothermal calorimetry and compressive strength of cementitious pastes, at ages of 1 and 7 days. The relationship seen for LFS is extremely different from the relationships seen for BOFS, GGBFS, and cement paste. If the point to the far right is excluded, then a roughly linear relationship between strength gain and heat release could be concluded, however, the scatter in the data suggests that the one may not be able to predict compressive strength development only by using heat release for these slags [31]. 3.5. Effect of gypsum on LFS pastes The LFS pastes showed rapid set, likely due to C12A7 hydration, which resulted in a retardation of the cement hydration and a drastic reduction in the strength gain [43,50,51]. In order to investigate a potential solution, additional gypsum was added to LFS pastes at 23 °C [45,47,63]. In addition, the effect of w/cm on the hydration of LFS pastes was also investigated at w/cm 0.40 and 0.50. TGA was carried out on these pastes at the end of the testing duration (at 7 days) to investigate changes in the hydrate assemblage due to the addition of gypsum. Fig. 10 shows heat flow and heat release for the LFS with varying w/cm and gypsum additions. The w/cm does not have a strong effect on the hydration kinetics, although it does increase the cumulative heat release. The addition of gypsum results in a reduction of the peak time, with the addition of 12% gypsum reducing the peak time by approximately 20 h. As the gypsum content Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 Fig. 9. Relationship between heat release from (a) Pozzolanic test and (b) Isothermal calorimetry and compressive strength of cementitious pastes. increases, the aluminate (sulfate depletion) peak is delayed and the silicate peak is accelerated, which is a classic sign of sulfate imbalance [35,45,47]. The heat flow curve suggests that the peak that is being seen in the LFS pastes without gypsum is the silicate peak, which is delayed because of the early occurrence of the aluminate peak [44–46,48]. Without gypsum, C12A7 reacts almost immediately upon mixing, so the peak is not detected. With 12% gypsum addition, the aluminate and silicate peaks are almost overlapping, and suggest that the sulfate imbalance induced due to the C12A7 may be resolved by the addition of 12% or greater gypsum amounts. Compressive strength testing of the mixture with 12% gypsum at 1 day did not show a significant improvement (<1 MPa difference in strength), suggesting that even greater amounts of gypsum may be needed to address the sulfate imbalance. However, for actual applications, it may not always be feasible in terms of cost or convenience to add such a large amount of gypsum. Fig. 11 shows the derivative weight curve from the TGA. The peak around 140 °C belongs to calcium aluminate hydrates, which is reduced as the gypsum addition increases [44,64]. Peaks due to ettringite are also observed at approximately 100 °C in the pastes with gypsum but not in the pastes without gypsum. Other changes in the response or changes due to w/cm seem minor. 465 Fig. 10. The (a) heat flow and (b) cumulative heat release of LFS pastes with additional gypsum. Fig. 11. Derivative weight curve of LFS pastes with additional gypsum using TGA. 466 Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 Table 3 Comparison of the findings from different methods regarding the feasibility of a material as SCM. Test method GGBFS BOFS LFS Pozzolanic test Isothermal calorimetry SCM Not clear SCM, less reactive Not clear TGA Compressive strength SCM SCM Not clear SCM, less reactive SCM Retardation of hydration Not clear Not an SCM 3.6. Comparison of test methods Table 3 shows a comparison of the implications that the different test methods provide regarding the feasibility of a material as SCM. All methods provide some information about whether a material can be used as SCM or not. By combining the results, it is concluded that BOFS and LFS should not be used as SCMs as-is. While the BOFS could potentially be used as an SCM, its reactivity seems to be quite low, and the use of 20% BOFS replacement leads to a reduction in strength of between 23 and 30%. The following may be noted about the different test methods: 1. The pozzolanic test provides information about hydraulic and pozzolanic behavior of tested materials; however, interpretation may be complex for materials such as LFS. 2. It is not obvious to identify SCMs from isothermal calorimetry; however, the test reveals materials which have a negative effect on the hydration (such as LFS). 3. TGA identifies materials as SCMs provided the calcium hydroxide consumption is large enough, however, this may not be apparent at early ages. 4. Compressive strength testing identifies materials as SCMs; however, it may also misidentify filler materials as SCMs if they are fine enough. Using only one method to determine the feasibility of materials as SCMs may lead to misleading results, as also suggested by Fig. 9. Therefore, an important finding from this study is that an accurate manner in which materials may be identified as SCMs would be the combined use of test methods. This may be done by using the pozzolanic test, isothermal calorimetry, and compressive test results. The pozzolanic test provides information about whether the materials are pozzolanic or hydraulic and the extent of pozzolanic or hydraulic behavior [33]. If the interpretation is not obvious or materials appear less reactive, then a compressive strength test should be run to ensure that adequate compressive strength is achieved. The isothermal calorimetry is an additional check to ensure that the material does not have a strong negative effect on the cement hydration. The results from these tests may be used together with compositional information (XRF and XRD) and fineness information (Blaine fineness or laser diffraction) to study a wide variety of materials for their feasibility as SCMs. 4. Conclusions A comprehensive testing regime was carried out to evaluate whether steel slags (BOFS and LFS) could be used as SCMs. The main conclusions from the study are: 1. The pozzolanic test suggests that all tested slags are hydraulic; GGBFS and LFS are more reactive than BOFS. 2. Isothermal calorimetry shows that LFS strongly retards cement hydration, likely due to C12A7 reaction. This leads to sulfate imbalance which may be corrected through the use of additional gypsum. 3. TGA shows that GGBFS consumes calcium hydroxide. The other slags do not consume calcium hydroxide or produce calcium hydroxide. 4. Compressive strength testing reveals that GGBFS shows the best performance, exceeding the strength of the cement paste at later ages. The BOFS shows a moderate strength gain, however, ultimate strengths are considerably lower than that with cement paste. LFS shows negligible strengths, due to C12A7 reaction. 5. The results from this testing suggest that LFS and BOFS should not be used as SCMs as-is; however, modifications such as the addition of gypsum may be beneficial. 6. The combined use of pozzolanic test, isothermal calorimetry, and compressive strength test may be used to accurately identify whether materials may be used as SCMs. Conflicts of interest The authors have no conflicts of interest to declare. Acknowledgements Sivakumar Ramanathan, Nima Hosseinzadeh Nanehkaran, and Michael Croly (all at the University of Miami) are thanked for help with performing experiments detailed here. References [1] U.S. Geological Survey, Mineral commodity summaries 2018, 2018, doi: 10.3133/70194932. [2] V.S. Devi, B.K. Gnanavel, Properties of concrete manufactured using steel slag, Procedia. Eng. 97 (2014) 95–104, https://doi.org/10.1016/j. proeng.2014.12.229. [3] F. Faleschini, M.A. Fernández-Ruíz, M.A. Zanini, K. Brunelli, C. Pellegrino, E. Hernández-Montes, High performance concrete with electric arc furnace slag as aggregate: mechanical and durability properties, Constr. Build. Mater. 101 (2015) 113–121, https://doi.org/10.1016/j.conbuildmat.2015.10.022. [4] Y. Jiang, T.-C. Ling, C. Shi, S.-Y. Pan, Characteristics of steel slags and their use in cement and concrete—A review, Resour. Conserv. Recycl. 136 (2018) 187–197, https://doi.org/10.1016/j.resconrec.2018.04.023. [5] T.-H. Lu, Y.-L. Chen, P.-H. Shih, J.-E. Chang, Use of basic oxygen furnace slag fines in the production of cementitious mortars and the effects on mortar expansion, Constr. Build. Mater. 167 (2018) 768–774, https://doi.org/10.1016/ j.conbuildmat.2018.02.102. [6] J. Setién, D. Hernández, J.J. González, Characterization of ladle furnace basic slag for use as a construction material, Constr. Build. Mater. 23 (5) (2009) 1788–1794, https://doi.org/10.1016/j.conbuildmat.2008.10.003. [7] H. Yi, G. Xu, H. Cheng, J. Wang, Y. Wan, H. Chen, An overview of utilization of steel slag, Procedia. Environ. Sci. 16 (2012) 791–801, https://doi.org/10.1016/j. proenv.2012.10.108. _ Yüksel, A review of steel slag usage in construction industry for sustainable [8] I. development, Environ. Dev. Sustain. 19 (2) (2017) 369–384, https://doi.org/ 10.1007/s10668-016-9759-x. [9] T. Zhang, Q. Yu, J. Wei, J. Li, P. Zhang, Preparation of high performance blended cements and reclamation of iron concentrate from basic oxygen furnace steel slag, Resour. Conserv. Recycl. 56 (1) (2011) 48–55, https://doi.org/10.1016/j. resconrec.2011.09.003. [10] C. Shi, Steel slag—Its production, processing, characteristics, and cementitious properties, J. Mater. Civil. Eng. 16 (3) (2004) 230–236, https://doi.org/10.1061/ (ASCE)0899-1561(2004) 16:3(230). [11] T. Koh, S. Hwang, Field evaluation and durability analysis of an eco-friendly prestressed concrete sleeper, J. Mater. Civil. Eng. 27 (7) (2014) B4014009, https://doi.org/10.1061/(ASCE)MT.1943-5533.0001109. [12] F.J. Hogan, J.W. Meusel, Evaluation for durability and strength development of a ground granulated blast furnace slag, Cem. Concr. Aggr. 3 (1) (1981) 40–52, https://doi.org/10.1520/CCA10201J. _ Durmusß, Utilization and efficiency of ground [13] E. Özbay, M. Erdemir, H.I. granulated blast furnace slag on concrete properties—A review, Constr. Build. Mater. 105 (2016) 423–434, https://doi.org/10.1016/ j.conbuildmat.2015.12.153. [14] C. Shi, Characteristics and cementitious properties of ladle slag fines from steel production, Cem. Concr. Res. 32 (3) (2002) 459–462, https://doi.org/10.1016/ S0008-8846(01)00707-4. [15] H. Guo, S. Yin, Q. Yu, X. Yang, H. Huang, Y. Yang, F. Gao, Iron recovery and active residue production from basic oxygen furnace (BOF) slag for supplementary cementitious materials, Resour. Conserv. Recycl. 129 (2018) 209–218, https://doi.org/10.1016/j.resconrec.2017.10.027. Y. Wang, P. Suraneni / Construction and Building Materials 204 (2019) 458–467 [16] T. Koh, S.-W. Moon, H. Jung, Y. Jeong, S. Pyo, A feasibility study on the application of basic oxygen furnace (BOF) steel slag for railway ballast material, Sustainability 10 (2) (2018) 284, https://doi.org/ 10.3390/su10020284. [17] Y.-X. Li, Y.-M. Chen, J.-X. Wei, X.-Y. He, H.-T. Zhang, W.-S. Zhang, A study on the relationship between porosity of the cement paste with mineral additives and compressive strength of mortar based on this paste, Cem. Concr. Res. 36 (9) (2006) 1740–1743, https://doi.org/10.1016/j.cemconres.2004.07.007. [18] J.M. Manso, M. Losañez, J.A. Polanco, J.J. Gonzalez, Ladle furnace slag in construction, J. Mater. Civil. Eng. 17 (5) (2005) 513–518, https://doi.org/ 10.1061/(ASCE)0899-1561(2005) 17:5(513). [19] Á. Rodriguez, J.M. Manso, Á. Aragón, J.J. Gonzalez, Strength and workability of masonry mortars manufactured with ladle furnace slag, Resour. Conserv. Recy. 53 (11) (2009) 645–651, https://doi.org/10.1016/j.resconrec.2009.04.015. [20] R. Snellings, Assessing, understanding and unlocking supplementary cementitious materials, RILEM Tech. Lett. 1 (2016) 50–55, https://doi.org/ 10.21809/rilemtechlett.2016.12. [21] G. Wang, Y. Wang, Z. Gao, Use of steel slag as a granular material: Volume expansion prediction and usability criteria, J. Hazard. Mater. 184 (1–3) (2010) 555–560, https://doi.org/10.1016/j.jhazmat.2010.08.071. [22] J. Zhao, P. Yan, D. Wang, Research on mineral characteristics of converter steel slag and its comprehensive utilization of internal and external recycle, J. Clean. Prod. 156 (2017) 50–61, https://doi.org/10.1016/j.jclepro.2017.04.029. [23] A.S. Brand, J.R. Roesler, Steel furnace slag aggregate expansion and hardened concrete properties, Cem. Concr. Compos. 60 (2015) 1–9, https://doi.org/ 10.1016/j.cemconcomp.2015.04.006. [24] E. Belhadj, C. Diliberto, A. Lecomte, Properties of hydraulic paste of basic oxygen furnace slag, Cem. Concr. Compos. 45 (2014) 15–21, https://doi.org/ 10.1016/j.cemconcomp.2013.09.016. [25] P.Y. Mahieux, J.E. Aubert, G. Escadeillas, M. Measson, Quantification of hydraulic phase contained in a basic oxygen furnace slag, J. Mater. Civil. Eng. 26 (4) (2013) 593–598, https://doi.org/10.1061/(ASCE)MT.19435533.0000867. [26] A. Santamaría, A. Orbe, J.T. San Joe, J.J. González, A study on the durability of structural concrete incorporating electric steelmaking slags, Constr. Build. Mater. 161 (2018) 94–111, https://doi.org/10.1007/s41024-018-0037-1. [27] E. Adesanya, H. Sreenivasan, A.M. Kantola, V.-V. Telkki, K. Ohenoja, P. Kinnunen, M. Illikainen, Ladle slag cement—Characterization of hydration and conversion, Constr. Build. Mater. 193 (2018) 128–134, https://doi.org/ 10.1016/j.conbuildmat.2018.10.179. [28] J.J. Emery, Slag utilization in pavement construction, ASTM Special Publication 774 (1982) 95–118, https://doi.org/10.1520/STP32459S. [29] M.F. Rojas, M.I.S. de Rojas, Chemical assessment of the electric arc furnace slag as construction material: expansive compounds, Cem. Concr. Res. 34 (10) (2004) 1881–1888, https://doi.org/10.1016/j.cemconres.2004.01.029. [30] X. Li et al., Reactivity tests for supplementary cementitious materials: RILEM TC 267-TRM phase 1, Mater. Struct. 51 (6) (2018) 151, https://doi.org/10.1617/ s11527-018-1269-x. [31] R. Snellings, K.L. Scrivener, Rapid screening tests for supplementary cementitious materials: past and future, Mater. Struct. 49 (8) (2016) 3265– 3279, https://doi.org/10.1617/s11527-015-0718-z. [32] F. Avet, R. Snellings, A.A. Diaz, M.B. Haha, K. Scrivener, Development of a new rapid, relevant and reliable (R3) test method to evaluate the pozzolanic reactivity of calcined kaolinitic clays, Cem. Concr. Res. 85 (2016) 1–11, https:// doi.org/10.1016/j.cemconres.2016.02.015. [33] P. Suraneni, J. Weiss, Examining the pozzolanicity of supplementary cementitious materials using isothermal calorimetry and thermogravimetric analysis, Cem. Concr. Compos. 83 (2017) 273–278, https://doi.org/10.1016/j. cemconcomp.2017.07.009. [34] P. Suraneni, A. Hajibabaee, S. Ramanathan, Y. Wang, J. Weiss, New insights from pozzolanic testing of supplementary cementitious materials, Cem. Concr. Compos. (2018). Under review. [35] D. Damidot, A. Rettel. Effect of gypsum on CA and C12A7 hydration at room temperature. In: Proceedings of the 11th International Congress on the Chemistry of Cement (ICCC), Durban, South Africa, 2003, pp. 1855–1865. [36] R.N. Edmonds, A.J. Majumdar, The hydration of 12CaO. 7A12O3 at different temperatures, Cem. Concr. Res. 18 (3) (1988) 473–478, https://doi.org/ 10.1016/0008-8846(88)90082-8. [37] A.J. Majumdar, B. Singh, R.N. Edmonds, Hydration of mixtures of C12A7 and granulated blastfurnace slag, Cem. Concr. Res. 19 (6) (1989) 848–856, https:// doi.org/10.1016/0008-8846(89)90097-5. [38] T. Kim, J. Olek, Effects of sample preparation and interpretation of thermogravimetric curves on calcium hydroxide in hydrated pastes and mortars, Transp. Res. Rec. 2290 (1) (2012) 10–18, https://doi.org/10.3141/ 2290-02. [39] ASTM, C305 Standard practice for mechanical mixing of hydraulic cement pastes and mortars of plastic consistency, ASTM International (1999). [40] G. Blanco, A. Bautista, H. Takenouti, EIS study of passivation of austenitic and duplex stainless steels reinforcements in simulated pore solutions, Cem. 467 Concr. Compos. 28 (3) (2006) 212–219, https://doi.org/10.1016/j. cemconcomp.2006.01.012. [41] M.M. Mennucci, E.P. Banczek, P.R.P. Rodrigues, I. Costa, Evaluation of benzotriazole as corrosion inhibitor for carbon steel in simulated pore solution, Cem. Concr. Compos. 31 (6) (2009) 418–424, https://doi.org/ 10.1016/j.cemconcomp.2009.04.005. [42] M. Saremi, E. Mahallati, A study on chloride-induced depassivation of mild steel in simulated concrete pore solution, Cem. Concr. Res. 32 (12) (2002) 1915–1921, https://doi.org/10.1016/S0008-8846(02)00895-5. [43] H.-G. Park, S.-K. Sung, C.-G. Park, J.-P. Won, Influence of a C12A7 mineral-based accelerator on the strength and durability of shotcrete, Cem. Concr. Res. 38 (3) (2008) 379–385, https://doi.org/10.1016/j.cemconres.2007.09.016. [44] M. Antoni, J. Rossen, F. Martirena, K. Scrivener, Cement substitution by a combination of metakaolin and limestone, Cem. Concr. Res. 42 (12) (2012) 1579–1589, https://doi.org/10.1016/j.cemconres.2012.09.006. [45] J.W. Bullard et al., Mechanisms of cement hydration, Cem. Concr. Res. 41 (12) (2011) 1208–1223, https://doi.org/10.1016/j.cemconres.2010.09.011. [46] M. Gawlicki, W. Nocuń-Wczelik, Ł. Ba˛k, Calorimetry in the studies of cement hydration: setting and hardening of portland cement–calcium aluminate cement mixtures, J. Therm. Anal. Calorim. 100 (2) (2009) 571–576, https://doi. org/10.1007/s10973-009-0158-5. [47] L. Nicoleau, E. Schreiner, A. Nonat, Ion-specific effects influencing the dissolution of tricalcium silicate, Cem. Concr. Res. 59 (2014) 118–138, https://doi.org/10.1016/j.cemconres.2014.02.006. [48] A. Quennoz, K.L. Scrivener, Interactions between alite and C3A-gypsum hydrations in model cements, Cem. Concr. Res. 44 (2013) 46–54, https://doi. org/10.1016/j.cemconres.2012.10.018. [49] R. Snellings, Solution-controlled dissolution of supplementary cementitious material glasses at pH 13: the effect of solution composition on glass dissolution rates, J. Am. Ceram. Soc. 96 (8) (2013) 2467–2475, https://doi. org/10.1111/jace.12480. [50] P. Suraneni, R.J. Flatt, Use of micro-reactors to obtain new insights into the factors influencing tricalcium silicate dissolution, Cem. Concr. Res. 78 (2015) 208–215, https://doi.org/10.1016/j.cemconres.2015.07.011. [51] P. Suraneni, M. Palacios, R.J. Flatt, New insights into the hydration of slag in alkaline media using a micro-reactor approach, Cem. Concr. Res. 79 (2016) 209–216, https://doi.org/10.1016/j.cemconres.2015.09.015. [52] H. Moon, S. Ramanathan, P. Suraneni, C.-S. Shon, C.-J. Lee, C.-W. Chung, Revisiting the effect of slag in reducing heat of hydration in concrete in comparison to other supplementary cementitious materials, Materials 11 (10) (2018) 1847, https://doi.org/10.3390/ma11101847. [53] V. Kocaba, E. Gallucci, K.L. Scrivener, Methods for determination of degree of reaction of slag in blended cement pastes, Cem. Concr. Res. 42 (3) (2012) 511– 525, https://doi.org/10.1016/j.cemconres.2011.11.010. [54] W.A. Gutteridge, J.A. Dalziel, Filler cement: the effect of the secondary component on the hydration of portland cement: Part I. A fine nonhydraulic filler, Cem. Concr. Res. 20 (5) (1990) 778–782. [55] T. Oey, A. Kumar, J.W. Bullard, N. Neithalath, G. Sant, The filler effect: The influence of filler content and surface area on cementitious reaction rates, J. Am. Ceram. Soc. 96 (6) (2013) 1978–1990, https://doi.org/10.1111/jace.12264. [56] E. Berodier, K. Scrivener, Understanding the filler effect on the nucleation and growth of C-S-H, J. Am. Ceram. Soc. 97 (12) (2014) 3764–3773, https://doi.org/ 10.1111/jace.13177. [57] P. Suraneni, V.J. Azad, O.B. Isgor, J. Weiss, Role of supplementary cementitious material type in the mitigation of calcium oxychloride formation in cementitious pastes, J. Mater. Civil. Eng. 30 (10) (2018) 04018248, https:// doi.org/10.1061/(ASCE)MT.1943-5533.0002425. [58] A. Coomarasamy, T.L. Walzak, Effects of moisture on surface chemistry of steel slags and steel slag-asphalt paving mixes, Transp. Res. Rec. 1492 (1995) 85–95. [59] K. Mitsunori, K. Torii, S. Hasaba, N. Nicho, K. Oda, Applicability of basic oxygen furnace slag as a concrete aggregate, ACI Special Publication 79 (1983) 1123– 1142. [60] W. Fajun, M.W. Grutzeck, D.M. Roy, The retarding effects of fly ash upon the hydration of cement pastes: The first 24 hours, Cem. Concr. Res. 15 (1) (1985) 174–184, https://doi.org/10.1016/0008-8846(85)90024-9. [61] G.C. Isaia, C. Geraldo, A.L.G. Gastaldi, R. Moraes, Physical and pozzolanic action of mineral additions on the mechanical strength of high-performance concrete, Cem. Concr. Compos. 25 (1) (2003) 69–76, https://doi.org/10.1016/ S0958-9465(01)00057-9. [62] B.K. Marsh, R.L. Day, D.G. Bonnery, Strength gain and calcium hydroxide depletion in hardened cement pastes containing fly ash, Mag. Concrete. Res. 38 (134) (1986) 23–29, https://doi.org/10.1680/macr.1986.38.134.23. [63] X. Li, Stabilization mechanism of municipal solid waste incineration fly ash and characterization of active mineral hydration products Doctoral dissertation, Donghua University, Shanghai, 2015. [64] W. Prince, M. Espagne, P.-C. Aïtcin, Ettringite formation: a crucial step in cement superplasticizer compatibility, Cem. Concr. Res. 33 (5) (2003) 635– 641, https://doi.org/10.1016/S0008-8846(02)01042-6.
0
Puede agregar este documento a su colección de estudio (s)
Iniciar sesión Disponible sólo para usuarios autorizadosPuede agregar este documento a su lista guardada
Iniciar sesión Disponible sólo para usuarios autorizados(Para quejas, use otra forma )