Minimum SCM requirements in mixtures containing limestone cement to control thaumasite sulfate attack

Construction and Building Materials 84 (2015) 19–29

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Minimum SCM requirements in mixtures containing limestone cement to control thaumasite sulfate attack Sajjad Mirvalad ⇑, Michelle Nokken 1 Department of Building, Civil and Environmental Engineering, Faculty of Engineering and Computer Science, Concordia University, 1455 de Maisonneuve Blvd. West, EV-6.139, Montreal, Quebec, Canada H3G 1M8

h i g h l i g h t s  Canadian SCM recommendations for cold temperature sulfate resistance were evaluated.  Fly ash replacements exceeded the expansion limits, suggesting further research.  The best sulfate resistance was achieved when slag and fly ash were used together.  Mass change study on mortar samples in sulfate attack can complement expansion.

a r t i c l e

i n f o

Article history: Received 11 July 2014 Received in revised form 19 January 2015 Accepted 26 February 2015 Available online 21 March 2015 Keywords: Portland-limestone cement Thaumasite Ettringite Supplementary cementing materials Sulfate resistance Sustainability

a b s t r a c t Different supplementary cementing materials (SCMs) were used in Portland-limestone cement (PLC) blends in order to achieve sulfate resistance. The resistance was evaluated by studying mortar expansion and mass change in ettringite sulfate attack (ESA) and thaumasite sulfate attack (TSA). The results showed that the PLC blends containing the Canadian standard recommended amounts of SCMs were not resistant against TSA. However, by increasing the SCM amount, the desired resistance was achieved. Additional SCM types not mentioned in CSA were found as high sulfate resistant blended PLC; among those, the ternary blends of PLC, slag, and fly ash were found to be the most resistant. Additionally, for all blends, a very good correlation between the expansion and mass change results was found. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Portland-limestone cement (PLC) has been exploited for decades in the European countries and is progressively becoming widely used in all over the world. The principal reason for developing such a cementitious material is its economic benefits and interesting sustainability characteristics. The clinker content of Portland-limestone cement is lower than ordinary Portland cement as this type of cement is a blend of ground clinker, gypsum and limestone. Therefore, manufacture of Portland-limestone cement involves lower fuel consumption and lower greenhouse gas emissions (CO2 and NOx) in comparison to ordinary Portland cement. Considering the sustainability interests in world

⇑ Corresponding author. Tel.: +1 (514) 649 4292. E-mail addresses: [email protected] (S. Mirvalad), [email protected] (M. Nokken). 1 Tel.: +1 (514) 848 2424x7806. http://dx.doi.org/10.1016/j.conbuildmat.2015.02.074 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.

development, usage of Portland-limestone cement has drawn particular attention. The first standardization of limestone containing cement was by the European standard in 1987 when EN197 specified a type of cement called PKZ, which consisted of 15 ± 5% limestone and 85 ± 5% clinker [1]. Years later, in 2000, different blends of Portland cement with ground limestone were delineated by the European standard, EN197-1 [2]. This standard allowed replacement of Portland cement with limestone powder up to 35% by weight [2]. The latest version of this standard was published in 2011. In the North America, the concept of addition of ground limestone to Portland cement is rather new in comparison to Europe. In 2008, the Canadian standard, CSA A3001-08 [3], introduced a new category of cement (Portland-limestone cement) containing up to 15% ground limestone. This kind of Portlandlimestone cement, designated as GUL in CSA A3001-10 [4], is now practically prepared with 12–13% limestone filler in the cement plants in Canada [5]. In the United States, in 2012, ASTM C595 [6] introduced Portland-limestone cement with up to 15%

20

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29

replacement of clinker with limestone in the process of cement manufacture. One of the main concerns with practical usage of PLC is its poor performance in sulfate exposure, especially at low temperatures. The process of sulfate attack to PLC has been well studied by different researchers [7–12]. Considering these research outcomes, many standards have assigned limitations for the use of PLC when sulfate ions are present in the field. In the UK, the Thaumasite Expert Group (TEG) declares ‘‘those cements in which the amount of limestone filler can range from 6% to 35%, should not be permitted in conditions where sulfate concentrations in the groundwater are in excess of 0.4 g/L’’ [13]. In Canada, CSA A3000 (2008) [14] does not classify PLC as moderate or high sulfate resistant cement. Consequently, CSA A23.1 (2009) [15] has prohibited the use of Portland-limestone cement in sulfate environments. In the United States, ASTM C595 (2012) [6] as well as AASHTO M240 (2012) [16] have defined similar limitations for Portland-limestone cements. The concerns about the performance of PLC in a sulfate environment relate to a type of sulfate attack that is called thaumasite sulfate attack (TSA). This type of sulfate attack targets the main part of the hydrated cement matrix, the calcium silicate hydrate (CSH) gel, and turns it to a white non-cohesive substance called thaumasite CaSiO3
CaCO3
CaSO4
15H2O [17]. TSA is more severe than the other types of sulfate attack and can completely disintegrate the cement paste matrix ending in incoherence of hydrated cement paste and a whole concrete structure failure. Concrete attacked by severe thaumasite sulfate attack ‘‘can be crumbled by a hammer, or even by hand’’ [18]. Usually, the initial visual sign of TSA is formation of sub-parallel cracks filled with a white substance [13]. Severe TSA in hydrated cement paste preferably occurs in cold (<15 °C) conditions in presence of moisture, sulfate ions, carbonate ions, and active form of alumina, e.g. ettringite [13,19,20]. When the temperature rises above 15 °C, the rate of TSA considerably decreases [21]. Generally, thaumasite can form at temperatures from 15 °C to 25 °C but at a much slower rate of formation [19]. Normally, the predominant mode of sulfate attack to any type of Portland cement at ordinary temperatures (higher than 15 °C) is ettringite sulfate attack (ESA). In this process of deterioration, sulfate ions attack calcium hydroxide and aluminate hydrates inside the hydrated cement paste and form gypsum and ettringite, respectively. It should be noted that, during ESA, when carbonate ions are present in the system, it is likely that minor amounts of thaumasite form in the paste [19]. In cold conditions, when the temperature drops below 15 °C, the mode of sulfate attack tends to switch to thaumasite sulfate attack depending on the amount of carbonates available in the pore solution of hydrated cement. Portland-limestone cements are highly susceptible to TSA at low temperatures because of the carbonate content of limestone powder [21]. Generally, when there is not any external source of carbonate ions such as calcareous aggregates, formation of thaumasite in ordinary Portland cements or general use cements is negligible because of the lack of carbonate bearing components inside the system. Therefore, recent research mainly focuses on TSA in Portland-limestone cements. It has been found in different research that supplementary cementing materials (SCMs), can impede or retard thaumasite formation in carbonate-bearing concrete and mortar samples, and improve their performance against TSA [22–26]. The principal effect of SCMs is the refinement of the hydrated cement matrix by taking part in pozzolanic reactions and formation of secondary calcium silicate hydrate (CSH) gels as well as in some cases physically filling the voids due to their fineness. SCMs are specifically beneficial in reducing the problems occurring from TSA. Pozzolans, when added to cement, react with calcium hydroxide in the hydrated cement paste and lower the calcium to silicon ratio

in the CSH phase. Hydrated cement phases with lower amounts of calcium hydroxide are more resistant against TSA. Therefore, addition of SCMs is effectively advantageous in preparing sulfate resistant mixtures [27,28]. The positive effect of SCMs on the resistance of cements against sulfate attack compensates for the weakness of PLC against TSA. This observation has led the Canadian standard, CSA A3001-10, to recommend minimum amounts of SCMs to be added to Portland-limestone cement in order to achieve acceptable resistance. CSA A3001-10 recommends addition of 25% Type F fly ash, 40% slag, 15% metakaolin or a combination of 5% Type SF silica fume with 25% slag or 20% fly ash for moderate and high sulfate resistant blended PLCs. It should be noted that these recommendations are the minimum action and do not guarantee resistance against TSA. In fact, the standard has assigned expansion limitations for mortar bars that must be achieved in order to use blended PLC in sulfate exposure (Table 1). This paper evaluates and compares performance of PLC in combination with different SCMs in various replacement levels exposed to sodium sulfate solutions in both ESA and TSA. Considering the CSA A3001-10 expansion limitations, the present study evaluates the standard recommendations for addition of SCMs to PLC. As well, it introduces binary and ternary blends of PLC that, considering the standard limits, can be categorized as high sulfate resistant blended PLC (HSLb) and can be safely used in the field. Additionally, this paper discusses the change in the mass of mortar samples and its correlation with expansion.

2. Experimental program 2.1. Materials and mixtures In order to evaluate the performance of Portland-limestone cement in a sulfate environment, two general use Portland-limestone cements (Type GUL according to CSA A3001-10), one provided by Ciment Québec, and the other one supplied by Holcim Canada (Joliette plant) were used in this study. The limestone contents from Ciment Québec and Holcim Canada were 10% and 13.5%, respectively. In this paper, the PLC from Ciment Québec is designated as ‘‘L’’, and the one from Joliette cement plant is titled as ‘‘GUL’’. In addition to PLC, different SCMs were obtained to evaluate their effect on the performance of PLC when added at different replacement levels. A grade 80 slag from Lafarge Stoney Creek plant in Ontario, a Type F fly ash provided by Ciment Québec from Belledune plant in New Brunswick, as well as metakaolin, silica fume (SF), and intermediate silica fume (SFI) all supplied by Ciment Québec were used. It should be mentioned that SFI is a new class of silica fume with lower content of SiO2 assigned by CSA A3001-08. While class SF silica fume must have at least 85% SiO2 content, SFI as an intermediate silica fume must contain higher than 75% SiO2 by mass. The SCMs were selected considering their availability in the region that would make their usage feasible for industrial purposes. The chemical compositions of the mentioned SCMs as well as the Portland-limestone cements used in this study are presented in Table 2.

Table 1 Expansion limitations for mortar samples studied according to test CSA A3004-C8 (CSA A3001-10).

Maximum expansion at 6 months (%) Maximum expansion at 18 months (%)

MSb and MSLba

HSb and HSLba

Reference

0.10

0.05c

CSA A3004-C8 Procedure A CSA A3004-C8 Procedure B

b

0.10

b

0.10

a MSb stands for moderate sulfate resistant blended hydraulic cement; HSb stands for high sulfate resistant blended hydraulic cement; MSLb stands for moderate sulfate resistant blended Portland limestone cement; HSLb stands for high sulfate resistant blended Portland limestone cement. b If the increase in expansion between 12 and 18 months exceeds 0.03%, the sulfate expansion at 24 months shall not exceed 0.10% in order for the cement to be deemed to have passed the sulfate resistance requirement. c If the expansion is greater than 0.05% at 6 months but less than 0.10% at 1 year, the cement shall be considered to have passed.

21

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29 Table 2 Chemical composition of the used materials in the research.

SiO2 Al2O3 Fe2O3 CaO MgO SO3 K2O Na2O TiO2 P2O5 SrO Mn2O3 Cr2O3 ZnO Loss on ignition C3S C2S C3A C4AF Equivalent alkalis Blaine fineness (m2/kg)

% % % % % % % % % % % % % % % % % % % %

‘‘L’’ 10% limestone Ciment Québec

‘‘GUL’’ 13.5% limestone Holcim Canada

Fly ash (Type F)

Silica fume

Silica fume I

Metakaolin

Slag (grade 80)

18.40 4.60 2.80 60.60 2.20 3.90 0.89 0.27 – – – – – – 4.60 61.00 6.80 7.30 8.60 0.85 399

19.00 4.10 2.68 60.40 1.80 3.05 0.83 0.20 0.15 – – – – – 7.46 70.00 1.67 6.30 8.16 0.75 468

47.34 18.80 8.58 5.36 2.32 0.75 1.96 1.28 0.94 0.22 0.17 0.10 0.05 0.03 – – – – – 2.57 490

83.75 1.81 1.06 1.58 0.17 0.05 0.88 0.34 0.09 0.08 0.00 0.12 0.01 0.10 9.09 – – – – 0.92 –

81.05 1.17 3.27 1.70 0.81 0.08 1.81 0.27 0.02 0.07 0.00 5.85 0.04 0.23 2.65 – – – – 1.46 –

61.45 29.23 1.20 2.35 0.36 0.19 1.92 0.19 0.68 0.03 0.01 0.08 0.01 0.00 2.31 – – – – 1.45 1690

38.6 9.90 0.67 37.90 9.30 3.00 0.52 0.17 0.42 – – – – – – – – – – 0.51 380

Table 3 The studied blends of Portland-limestone cement. Designation

PLC – Ciment Québec (L) %

Slag %

1 2 3

L-35 Slag L-40 Slag L-45 Slag

65 60 55

35 40 45

4 5 6

L-20 FA L-25 FA L-30 FA

80 75 70

7 8 9

L-3 SF L-5 SF L-8 SF

97 95 92

10 11 12 13

L-25 L-30 L-35 L-40

14 15 16

L-3 SFI L-5 SFI L-8 SFI

97 95 92

17 18 19

L-10 MK L-15 MK L-20 MK

90 85 80

20

L Control

100

21 22 23

GUL Control GUL-40 Slag GUL-25 FA

Slag-10 Slag-10 Slag-15 Slag-20

FA FA FA FA

65 60 50 40

Fly ash Type F %

Silica fume %

Silica fume I %

Metakaolin %

PLC – Holcim (GUL) %

20 25 30 3 5 8 25 30 35 40

10 10 15 20 3 5 8 10 15 20

40 25

In order to investigate the resistance of PLC when blended with the selected SCMs, 23 cement blends were devised as presented in Table 3. Blends number 1 through number 20 are prepared with the PLC containing 10% inter-ground limestone and blends number 21, 22, and 23 are made with the other PLC, which contains 13.5% inter-ground limestone. According to CSA A3001-10, PLC can only be considered in sulfate environment when it is in form of moderate or high sulfate resistant blended PLC. According to this standard, such cement shall conform to the assigned expansion limits (Table 1) and contain a recommended minimum of 25% Type F fly ash or 40% slag or 15% metakaolin or a combination of 5% Type SF silica fume with 25% slag or a combination of 5% Type SF silica fume with 20% Type F fly ash. The recommendations regarding the minimum SCM replacements are based on the limited research completed until the time of the standard publication. The PLC blends in this study are designed in order to evaluate the standard recommendations and to find new binary and ternary blends of PLC that can be categorized as high sulfate resistant blended PLC to be considered in later modifications of the standard.

100 60 75

Three blends of ‘‘L’’ were set to the CSA A3001-10 recommendation for addition of slag, fly ash and metakaolin, (L-40 Slag, L-25 FA, L-15 MK). With the purpose of further refining the suggested limits of these SCMs, other blends were also set with 5% lower and higher than the recommended percentage levels. Silica fume is only mentioned in the standard at amount of 5% and in combination with slag or Type F fly ash. Therefore, blends of ‘‘L’’ were studied solely with silica fume with percentages of 3%, 5%, and 8% toward evaluating effectiveness of this SCM alone on impeding sulfate attack to PLC. The same percentages were also studied for intermediate silica fume so that the performance of these two types of silica fumes would be compared in parallel. Moreover, owing to the fact that both slag and fly ash are effective for the durability in sulfate attack [5,13,25,26,29–31], and the Canadian standard does not have any suggestions for ternary blends of PLC containing these two SCMs, four ternary blends of PLC with slag and fly ash were devised (blends number 10, 11, 12, and 13). As slag possesses both cementing and pozzolanic characteristics, while Type F fly ash only has pozzolanic characteristics, the percentage of slag in the blends was chosen higher than fly ash. CSA A3001-10 limits the

22

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29

amount of SCMs in ternary blends to 60%. Accordingly, a ternary blend was designed with 40% slag (the recommended percentage according to the standard) and 20% fly ash (blend number 13: L-40 Slag-20 FA). Three other ternary blends were also set by reducing amounts of slag and fly ash. With the purpose of verifying recommendations of the standard on addition of slag and fly ash that are the most used SCMs, binary blends of the other type of Portland-limestone cement (GUL) with 40% slag and 25% fly ash were also set. It should be mentioned that sample numbers 20 and 21 are control samples prepared with each studied cement.

during the process of sulfate attack, mortar samples gain weight [39–41]. This is attributed to formation of gypsum, ettringite and thaumasite. Therefore, mass change can help in studying the process of sulfate attack. In this study, the mass change was performed along with the expansion test. For the test, the excess water on the surface of the mortar bars was removed using paper towel, then the samples were weighed.

3. Results and discussion 2.2. Mixing and casting procedures

3.1. Expansion of mortar bars in sulfate attack The Canadian standard on expansion of mortar bars due to external sulfate attack, CSA A3004-C8 (2010) [32] was used in order to evaluate the sulfate resistance of PLC blends. This standard test method has two procedures for testing of sulfate attack. The first one, procedure A, is sulfate attack at standard temperature (23 °C). This procedure determines the resistance against formation of ettringite and gypsum as these are the dominant materials formed during the process of sulfate attack at temperatures above 15 °C [13]. The second one, procedure B, is performed at 5 °C, and evaluates the resistance against thaumasite sulfate attack. It should be mentioned that the most favorable temperature for formation of thaumasite is found around 5 °C [21,33]. Both procedures A and B are practically similar to ASTM C1012 [34], and the most integral difference is the defined ambient temperature. Standard mortar mixtures were prepared according to ASTM C305 [35] (similar to CSA A3004-C1 [36]) and ASTM C109 [37] (similar to CSA A3004-C2 [38]) with the blends of PLC presented in Table 3. Each mixture was proportioned as one part blended PLC and 2.75 parts standard graded sand, and the water to cement ratio was 0.485. Although the standard indicates that the water to cement ratio should be changed to give constant flow, it was decided to keep the w/c constant. The constant w/c allows better comparison among the mixtures as well as avoids several trial mixtures to achieve the flow required. The constant w/c approach has been used by researchers in the past [12], likely for the same reasons. Blends of PLC with SCMs were manually prepared in the laboratory. For each mix, the desired amounts of PLC and SCM were obtained and hand mixed in a bowl. Then, the prepared blended PLC, standard graded sand, and water were mixed in accordance to ASTM C109 and ASTM C305. Prepared mortar was cast into metal cube molds of 50  50  50 mm as well as metal prisms molds of 25  25  285 mm. The molds were then put in 100% humidity containers, placed in an oven with temperature of 35 ± 3 °C for 23 ± 0.5 h. Afterwards, the mortar samples were demoulded and placed in saturated limewater at 23 °C. 2.3. Measurement procedures At the time of demoulding of each set of mortar samples, the compressive strengths of two mortar cubes were tested according to ASTM C109 (or CSA A3004-C2) with the purpose of checking if the samples achieved the required average compressive strength of 20 ± 1 MPa. The compressive strength measurements were monitored daily until the average strength of mortar cubes reached 20 ± 1 MPa. As for the CSA A3004-C8 procedure A (ESA), once the companion mortar cubes attained the specified strength, the length of mortar bars was measured with a digital comparator with the accuracy of 0.002 mm, and the length was defined as the initial length. Six mortar bars prepared for each mixture were then placed into two plastic containers (three bars in each one) with close-fitting lids containing 50 g/L sodium sulfate solution prepared with de-ionized water and anhydrous sodium sulfate. It should be mentioned that the standard assigns such highly concentrated solution in order to accelerate the process of sulfate attack. The mortar bars were placed on plastic rods in the containers so that the solution would engulf their whole perimeter. The volume of the solution was limited to 4 ± 0.5 times of the volume of the submerged mortar bars. The containers were then kept in the lab at temperature of 23 ± 2 °C. In procedure B of the standard (TSA), when the companion mortar cubes reached the desired compressive strength, the mortar bars were first cooled to 5 ± 2 °C. To do so, the bars were immersed in a precooled container of tap water in a refrigerator with temperature of 5 ± 2 °C for 7 h. It should be mentioned that the water container was placed in the refrigerator at least 7 h prior to immersion of mortar bars. Subsequently, the initial length of mortar bars was measured, then they were immersed in 50 g/L sodium sulfate solution, which was previously prepared and precooled in the refrigerator for at least 7 h according to the CSA A3004-C8 standard. The same ratio of solution to the sample was conformed to as for the procedure A. Following placing mortar bars in sodium sulfate solution, continuous length change measurements with the same comparator was carried out on mortar bars at 1, 2, 3, 4, 8, 13, 15 weeks followed by 4, 6, 9, 12, 15, 18, 21, and 24 months. After each measurement, the sodium sulfate solution was renewed. The new solution was prepared at least 7 h earlier, and stored at 5 ± 2 °C or 23 ± 2 °C in order to reach the desired test temperature for ESA and TSA. There is not any standard procedure for studying the mass change of mortars in sulfate environment. However, during this study, it was decided to investigate the mass change of mortar bars along with their expansion. It has been observed that

3.1.1. Ettringite sulfate attack (23 °C) The expansion measurements were carried out according to CSA A3004-C8 on mortar bars, immersed in sodium sulfate solution, after 1, 2, 3, 4, 8, 13, 15, 17, 26, 39, 52, 65, 78, 91, and 104 weeks. Table 4 presents the average expansion at some of the times; each average represents the mean of 6 mortar bars. For all blends, expansion increased during the 2-year study. During the process of ESA, gypsum and ettringite form due to the reaction of sulfate ions with the hydration products of cement. Formation of gypsum and ettringite is known to be expansive. Therefore, it can result in expansion, surface deterioration and spalling of concrete and ultimately a total failure [42]. In this study, continuous presence of sodium sulfate in the system (as the solution was renewed after each measurement) along with presence of calcium hydroxide and calcium aluminate hydrates (results of cement hydration) was in favor of formation of gypsum and ettringite, which led to a continuous expansion. Formation of ettringite and gypsum in sulfate attack to Portland cement and Portlandlimestone cement has been well studied before [8]. With regards to the procedure A of CSA A3004-C8, there are 6month expansion limits of 0.10% and 0.05% for moderate sulfate resistant blended PLC (MSLb) and high sulfate resistant blended PLC (HSLb), respectively (CSA A3001-10). Considering the average expansions after 6 months, presented in Table 4, all tested blends of PLC other than ‘‘GUL Control’’ (sample No. 21), expanded less than the limit for HSLb. Nonetheless, the average expansion of ‘‘GUL Control’’ after 6 months was less than the limit for moderate sulfate resistance. It should be mentioned that this sample showed

Table 4 Expansion of mortar bars subjected to ettringite sulfate attack (23 °C) and tested according to CSA A3004-C8-A. Average expansion (%) 3 months 6 months 12 months 18 months 24 months 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

L-35 Slag L-40 Slag L-45 Slag L-20 FA L-25 FA L-30 FA L-3 SF L-5 SF L-8 SF L-25 Slag-10 L-30 Slag-10 L-35 Slag-15 L-40 Slag-20 L-3 SFI L-5 SFI L-8 SFI L-10 MK L-15 MK L-20 MK L Control GUL Control GUL-40 Slag GUL-25 FA

FA FA FA FA

0.019 0.020 0.025 0.032 0.034 0.019 0.033 0.024 0.021 0.026 0.022 0.017 0.015 0.033 0.028 0.026 0.023 0.023 0.019 0.019 0.030 0.013 0.032

0.031 0.035 0.029 0.036 0.038 0.023 0.040 0.034 0.032 0.033 0.030 0.023 0.021 0.050 0.040 0.035 0.030 0.032 0.024 0.047 0.097 0.029 0.035

0.041 0.045 0.043 0.050 0.051 0.036 0.053 0.048 0.043 0.039 0.035 0.029 0.026 0.080 0.054 0.044 0.053 0.042 0.029 0.146 0.288 0.042 0.050

0.050 0.056 0.045 0.055 0.057 0.041 0.075 0.057 0.051 0.043 0.039 0.029 0.029 0.153 0.082 0.052 0.063 0.049 0.036 0.324 0.511 0.048 0.056

0.053 0.061 0.052 0.061 0.062 0.046 0.100 0.063 0.057 0.047 0.042 0.032 0.029 0.275 0.135 0.060 0.083 0.056 0.036 0.569 0.799 0.050 0.056

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29

expansion of 0.288% after a year (Table 4). Therefore, considering CSA A3001-10, its expansion was not in the range for high sulfate resistant blended cements. Based on the results summarized in Table 4, the control samples (samples No. 20 and 21) revealed the highest expansions amongst all the studied blends after 2 years immersion in sodium sulfate solution. Although both PLCs (L and GUL) had nearly similar C3A contents, the ‘‘GUL Control’’ sample expanded more than ‘‘L Control’’. This can be attributed to the higher limestone content in ‘‘GUL’’ (13.5%) in comparison to ‘‘L’’ (10%). The higher limestone content means a lower cementing property, which weakens the hydrated cement matrix. This signifies the weakness of these cements against ESA when compared to the other blends of PLC with SCMs. According to Table 4, SCMs have reduced expansions in ESA. When PLC is partially replaced with SCMs, depending on the type of the SCM (from the viewpoint of aluminate content), the amount of C3A in the cementitious system may be reduced; thus, the possibility of formation of ettringite can be decreased. In addition, this replacement reduces the amount of calcium hydroxide in the hydrated cement paste by reducing the amount of C2S and C3S in the system as well as developing pozzolanic reactions in the later stages of hydration. The reduction in the amount of calcium hydroxide in the hydrated cement matrix leads to reduction in the possibility of formation of gypsum and ettringite. Additionally, since SCMs take part in pozzolanic reactions by reacting with calcium hydroxide in the hydrated cement matrix, secondary CSH gels are formed that reinforce the hydrated cement matrix. Evidently, the pozzolanic reactions reduce the porosity and permeability of the hydrated cement matrix. From this point of view, the SCMs are also effective in improving resistance against sulfate attack by reducing the chance of ingress of sulfate ions inside the hydrated cement matrix. Considering Table 4, it can be inferred that generally at the age of two years, the samples prepared with SCMs have expanded less than the control samples. Also, it could be found that for each type of SCM, in most cases, the expansion is decreased when the amount of SCM is increased in the blend. This observation denotes the effectiveness of SCMs on impeding ESA. Also, it can be observed that the ternary PLC blend containing 40% slag and 20% fly ash had the lowest expansion after two years of ESA. This blend had the highest amount of SCM among all the blends. Comparing expansion of this blend with expansion of the binary blends of ‘‘L’’ cement with slag and fly ash, the synergy of slag and fly ash in a ternary blend of PLC can be inferred. Normally, slag reduces the workability of mortar while fly ash increases this property. Hence, when slag and fly ash are simultaneously used in a blend of cement, the prepared mortar can benefit from the pozzolanic and cementitious properties of slag as well as pozzolanic properties of fly ash (Type F) while it possesses a good flow that helps proper compaction. Such good workability was observed in the prepared ternary mortars during this laboratory study. Expansion in mortar bars can result in formation of cracks and distortion (curving), which ultimately results in total failure of the samples. After the two-year study, noticeable cracks and distortion were found in mortar bars prepared with the two control blends of PLC. Cracks were mostly formed on the edges and at the end of those mortar bars. After similar period of time, the blends containing SCMs were totally sound, and only very small longitudinal cracks were seen on ‘‘L-3 SFI’’ sample. The mentioned difference in the samples’ visual conditions signifies the positive effect of the SCMs on mortars in ESA. In Fig. 1, there is a close look at the extent of cracks in ‘‘L Control’’ sample. The cracks were extensive, especially near the ends of the mortar bars. According to the close visual study, ESA did not cause disintegration of the mortar bars.

23

Fig. 1. Formation of cracks on the edges and at the end of the ‘‘L Control’’ mortar bars due to the two-year ESA.

3.1.2. Thaumasite sulfate attack (5 °C) The average expansion of mortar bars tested in TSA as per CSA A3004-C8 is depicted in Table 5. According to the standard procedure, the length change measurements were performed at ages of 1, 2, 3, 4, 8, 13, 15, 17, 26, 39, 52, 65, 78, 91, and 104 weeks, but in order to simply summarize the results, solely length changes at ages of 3, 6, 12, 18, and 24 months are reviewed in the table. All samples showed increasing expansions during the TSA test. The expansion along with the disintegration caused by severe TSA was followed by total failure in some of the samples as mentioned in the table. Due to the temperature, thaumasite is thermodynamically favored in these tests [43]. The expansions are due to TSA in which thaumasite is formed predominantly, but ettringite and gypsum are also developed. Formation of all these products is expansive, and results in deterioration. Formation of thaumasite, ettringite, and gypsum causing expansion has been reported in similar research [12]. Considering Tables 4 and 5, the expansions due to TSA were found higher than ESA. Since formation of thaumasite is associated with destruction of the CSH structure in the hydrated cement matrix, it has caused severe disintegration and failure in many of the studied blends during the 2-year study. In Table 5, the time to failure of the blends is also presented. The failure of a sample is considered the time in which the sample is broken or deteriorated in a way that performing the length change test is impossible. Although failure could have occurred at any time since the previous measurement, the time of failure is defined when a scheduled measurement was no longer possible. The failed samples were mostly disintegrated, fractured, or lost the cohesion to the measurement pins. The first to fail was blend No. 21 (GUL Control), which contained 13.5% ground limestone. The ‘‘GUL Control’’ mortar bars failed after 17 weeks of immersion in sodium sulfate solution at 5 °C. The other control sample, prepared with blend No. 20 (L Control) failed at 42 weeks. This type of cement contained 10% ground limestone. As the carbonate source is vital for thaumasite formation, the main reason for ‘‘GUL Control’’ to fail sooner than ‘‘L Control’’ is the higher limestone content. Compared to the ‘‘GUL Control’’ sample, addition of 40% slag or 25% fly ash to the PLC in blends No. 22 and 23, respectively, improved the performance in TSA by reducing the expansion and delaying the failure time. Results for the blends No. 1 through 19, which contain the other PLC (L), presented in Table 5, show that all blends containing Type F fly ash, as well as ‘‘L-3 SF’’, ‘‘L-3 SFI’’, ‘‘L-5 SFI’’, ‘‘L-10 MK’’, and ‘‘L-15 MK’’ deteriorated to failure during this study. It was notable that those mixtures containing slag were generally intact until 24 months. According to the table, generally, increasing amounts of SCMs decreased expansion thus improved the resistance against TSA. As mentioned previously, the CSA A3001-10 limits the expansion of mortar bars, tested in accordance with CSA A3004-C8-B, to 0.10% at 18 months in order to achieve moderate or high sulfate resistant blended PLC. With respect to this mandate and Table 5,

24

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29

Table 5 Expansion and time to failure of mortar bars subjected to thaumasite sulfate attack (5 °C) and tested according to CSA A3004-C8-B. Average expansion (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

L-35 Slag L-40 Slag L-45 Slag L-20 FA L-25 FA L-30 FA L-3 SF L-5 SF L-8 SF L-25 Slag-10 L-30 Slag-10 L-35 Slag-15 L-40 Slag-20 L-3 SFI L-5 SFI L-8 SFI L-10 MK L-15 MK L-20 MK L Control GUL Control GUL-40 Slag GUL-25 FA

FA FA FA FA

3 months

6 months

12 months

18 months

24 months

Time to failure (weeks)

0.010 0.013 0.019 0.036 0.036 0.025 0.030 0.025 0.026 0.025 0.023 0.020 0.019 0.034 0.030 0.026 0.026 0.024 0.024 0.015 0.376 0.007 0.175

0.025 0.027 0.027 0.086 0.089 0.047 0.040 0.037 0.036 0.034 0.031 0.028 0.024 0.047 0.040 0.038 0.032 0.036 0.033 0.038 Failed 0.025 Failed

0.047 0.041 0.039 Failed 0.196 Failed 0.063 0.047 0.046 0.048 0.044 0.038 0.035 Failed 0.071 0.043 0.078 0.104 0.043 Failed Failed 0.111 Failed

0.173 0.085 0.045 Failed Failed Failed 0.202 0.056 0.046 0.064 0.084 0.044 0.037 Failed 0.329 0.068 0.144 Failed 0.068 Failed Failed 0.435 Failed

0.518 0.219 0.055 Failed Failed Failed Failed 0.091 0.051 0.110 0.178 0.048 0.038 Failed Failed 0.255 Failed Failed 0.108 Failed Failed Failed Failed

No failure No failure No failure 52 65 52 104 No failure No failure No failure No failure No failure No failure 52 91 No failure 91 78 No failure 42 17 91 26

blends; ‘‘L-40 Slag’’, ‘‘L-45 Slag’’, ‘‘L-5 SF’’, ‘‘L-8 SF’’, ‘‘L-25 Slag-10 FA’’, ‘‘L-30 Slag-10 FA’’, ‘‘L-35 Slag-15 FA’’, ‘‘L-40 Slag-20 FA’’, ‘‘L-8 SFI’’, and ‘‘L-20 MK’’ expanded below the limit for sulfate resistant blended PLC. The CSA A3001-10 standard also limits the 24-month average expansion to 0.10% for mortar bars if their increase in expansion between 12 and 18 months exceeds 0.03%. If such limit is achieved, the cement can be classified as HSLb. Considering the expansion data in Table 5, among the blends expanded less than 0.10% after 18 months, expansions of ‘‘L-40 Slag’’ and ‘‘L-30 Slag10 FA’’ between 12 months and 18 months have exceeded 0.03%. As well, these two blends have expanded more than 0.10% after 24 months TSA, which means they cannot be considered as high sulfate resistant blended PLC. As seen in Table 5 the ternary PLC blend containing 40% slag and 20% fly ash showed the lowest expansion in TSA during this study. Therefore, from the point of view of expansion, this sample resisted the best. Partially replacing PLC with SCMs is beneficial from different aspects. Reactions related to formation of thaumasite are dependent to presence of carbonates. When PLC is partially replaced with SCM, the limestone content of the blended cement is reduced; therefore, the extent of formation of thaumasite is reduced. Moreover, by addition of specific SCMs, the C3A content of the blended PLC is lowered. So, formation of ettringite, which is an initiator for formation of thaumasite, is reduced [13]. When formation of ettringite is decreased, the expansion and crack formation due to its development is decreased; accordingly, the resistance against TSA is improved, as the cement matrix is less affected and less permeable to ingress of sulfate ions that can result in formation of thaumasite inside the hydrated cement paste. Additionally, SCMs reinforce the hydrated cement matrix by reacting with calcium hydroxide and forming secondary CSH gels. In such case, the silicon rich formed CSH matrix is more resistant against sulfate attack [27,28]. This reinforcement improves the resistance against disintegrations occurred by formation of thaumasite. Besides, pozzolanic reactions and formation of secondary CSH gels reduce the porosity of the hydrated cement paste and increase its density. This results in a lower permeable concrete that can better resist against TSA. Considering the results presented in Table 5, the CSA A3001-10 recommended minimum amounts of SCM to be considered in

moderate and high sulfate resistant PLC were not sufficient to achieve the desired resistance. Of course, when the SCM content was increased over the standard recommendations, the required resistance was attained for slag and metakaolin containing blends, but none of the PLC blends containing the studied Type F fly ash achieved the standard limits for expansion. This can be due to the fact that typically fly ash is more chemically variable compared to slag and metakaolin. Considering these results and the fact that the standard recommendations are rather new, more experimental study on various PLC and SCMs is required to draw accurate recommendations for addition of SCMs to PLC with the purpose of utilization in sulfate environments. As mentioned earlier, some of the PLC blends qualified as HSLb. The two-year average expansion of these blends in TSA is presented in Fig. 2. While all the presented blends are HSLb, there are noticeable differences in their expansions. Accordingly, considering expansion in CSA A3004-C8-B test as a criterion for resistance in TSA, the noted blends have different resistances. ‘‘L-40 Slag-20 FA’’, which contained the highest amount of SCM, had the best performance against TSA. The worst performance was related to ‘‘L-8 SFI’’. It is clear that slag, when used in high percentages, had a key role on improving resistance in TSA. Silica fume also was quite effective on performance improvement of PLC, the intermediate silica fume as expected was not as effective as silica fume. According to the figure, ‘‘L-8 SFI’’, ‘‘L-25 Slag-10 FA’’, and ‘‘L-20 MK’’ have passed the CSA A3001-10 expansion limit at some point after 18 months. Yet, they qualified as high sulfate resistant PLC as their expansion between 12 and 18 months was less than 0.03%. As clearly seen in Fig. 2, the rate of the length change for each blend changes as a function of time, and at older ages, the graphs of some of the blends have shown an increased rate in length change. Such observation denotes that the mortar bars are getting close to the failure. The differences in the length change rate of the blends at particular ages also emphasize the differences in resistance against TSA. ‘‘L-40 Slag-20 FA’’ that had the lowest expansion in two years also showed the lowest length change rate at the last interval of expansion test amongst the other high sulfate resistant blended PLCs. During this experimental study, it was clearly observed that the failure due to TSA was very extensive, and in addition to formation

25

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29

0.120 8% SFI 25% Slag+10%fly ash

CSA A3001-10 limit 0.100 20% MK

0.080

Average expansion (%)

5% SF

0.060 45% Slag

8% SF 35% Slag+15%fly ash 0.040 40% Slag+20%fly ash

0.020 L-25 Slag-10 FA-5 deg.

L-35 Slag-15 FA-5 deg.

L-40 Slag-20 FA-5 deg.

L-45 Slag-5 deg.

L-5 SF-5 deg.

L-8 SF-5 deg.

L-20 MK-5 deg.

L-8 SFI-5 deg.

0.000 0

8

16

24

32

40

48

56

64

72

80

88

96

104

Time (weeks) Fig. 2. Two-year average expansion of high sulfate resistant mortar bars in TSA, tested according to CSA A3004-C8.

of cracks and mortar bars deformation, degradation and disintegration occurred. The aggressiveness of TSA to the plain PLC (GUL) can be seen in Fig. 3. After only 4 months of immersion of mortar bars in sodium sulfate solution at 5 °C, visible cracks were found on the edges and at the end sections of the samples as well as noticeable deformation. This was followed by decomposition of hydrated cement paste in later ages. As presented in Fig. 3, seven months later the mortar bars were completely crumbled and disintegrated to a non-cohesive mush found on the base of the container. Such failure was also seen for the other control sample, and it is also reported in previous research [23,26]. The visual observations confirmed that deterioration due to TSA was quite extensive compared to ESA, and it was different from the aspect of type of failure since in addition to deformation and cracks, TSA resulted in decomposition of mortars.

Blends of PLC with SCMs visually exhibited greatly improved resistance to TSA in comparison to the control samples of plain PLC. Interestingly for each SCM, the visual condition improved with increase in the amount of SCM in the blend. As an example, the mortar bars prepared with silica fume and exposed to TSA for two years are presented in Fig. 4. The increase in the silica fume content clearly improved the performance of mortar bars. ‘‘L-3 SF’’ showed clear damage, surface loss and crack on edges and the end after two years. ‘‘L-5 SF’’ was in better condition; it had some cracks on the edges. ‘‘L-8 SF’’ did not show any visual sign of deterioration after two years of TSA. Among all the 23 sets of mortar bars in TSA, after two years, ‘‘L-8 SF’’, ‘‘L-45 Slag’’, ‘‘L-35 Slag-15 FA’’, and ‘‘L-40 Slag-20 FA’’ had the best performance considering visual condition. This was in accordance with the expansion results as seen in Table 5.

26

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29 Table 6 Mass changes of mortar bars – ettringite sulfate attack (23 °C). Average mass change (%) 3 months 6 months 12 months 18 months 24 months

4 months TSA

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

11 months TSA

Fig. 3. Effect of thaumasite sulfate attack on plain Portland-limestone cement (GUL).

3.2. Mass changes of mortar bars in sulfate attack

L-35 Slag L-40 Slag L-45 Slag L-20 FA L-25 FA L-30 FA L-3 SF L-5 SF L-8 SF L-25 Slag-10 L-30 Slag-10 L-35 Slag-15 L-40 Slag-20 L-3 SFI L-5 SFI L-8 SFI L-10 MK L-15 MK L-20 MK L Control GUL Control GUL-40 Slag GUL-25 FA

FA FA FA FA

0.4 0.5 0.4 0.7 0.7 0.7 0.8 0.7 1.0 0.4 0.5 0.4 0.3 1.0 0.7 0.6 0.7 0.7 0.7 0.6 0.9 0.3 0.8

0.5 0.5 0.5 0.9 0.9 0.9 1.0 0.9 1.1 0.5 0.6 0.5 0.3 1.4 0.9 0.8 1.0 0.8 0.7 0.9 1.4 0.4 1.0

0.7 0.7 0.6 1.2 1.2 1.1 1.4 1.3 1.4 0.8 0.7 0.5 0.6 1.8 1.4 1.2 1.2 1.0 0.9 1.5 1.9 0.6 1.3

0.8 0.8 0.7 1.4 1.3 1.2 1.7 1.6 1.8 0.9 0.9 0.7 0.8 2.1 1.7 1.6 1.5 1.2 1.0 2.1 2.6 0.6 1.5

0.9 1.0 0.8 1.5 1.4 1.2 1.9 1.8 1.9 1.1 1.0 0.8 1.1 2.5 2.0 1.8 1.7 1.3 1.0 2.9 3.5 0.7 1.5

Basically, when sulfate attack reactions take place, the resultants fill the voids inside the hydrated cement paste, which results in mass increase. Formation of ettringite, thaumasite and gypsum inside pores leads to expansion and crack propagation and eventual deterioration. Accordingly, mass increase can stop at certain points near failure and change to mass decrease due to disintegration. Lower increase in mass is a sign of resistance against formation of ettringite, thaumasite, and gypsum. Therefore, mass increase is a sign of the progress of sulfate attack, and it can help to monitor and compare the resistance of different samples against formation of ettringite and thaumasite. Considering this fact, the mass changes of mortar bars in ESA and TSA is studied. The results are presented as follows.

the lowest mass gain. This indicated that these blends were more resistant against ESA in comparison to the other blends. Considering the table, it is clear that the control blends had the highest mass gains due to ESA. Accordingly, addition of the studied SCMs improved the resistance of the PLC blends against ESA. It is also inferred from Table 6 that generally for each SCM, an increase in the amount of the SCM leads to a decrease in mass gain. In other words, increase in the amount of SCM in the PLC blends has mainly improved resistance against ESA.

3.2.1. Ettringite sulfate attack (23 °C) The mass changes of mortar bars immersed in sodium sulfate solution at 23 °C after 3, 6, 12, 18 and 24 months are presented in Table 6. The continuous mass increase, which is found in all samples, is a sign of formation of gypsum and ettringite. According to Table 6, after 24 months the binary blends containing slag and the ternary ones containing slag and fly ash demonstrated

3.2.2. Thaumasite sulfate attack (5 °C) The mass changes of mortar bars immersed in sodium sulfate solution at 5 °C at ages of 3, 6, 12, 18, and 24 months are summarized in Table 7. As gypsum, ettringite, and thaumasite are formed, a mass increase is detected in all mortar samples. However, since mortar is softened and disintegrated during TSA, especially when

L-3 SF

L-5 SF

L-8 SF

Fig. 4. Mortar bars containing silica fume after 24 months of thaumasite sulfate attack.

27

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29 Table 7 Mass changes of mortar bars – thaumasite sulfate attack (5 °C). Average mass change (%) 3 months 6 months 12 months 18 months 24 months 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

L-35 Slag L-40 Slag L-45 Slag L-20 FA L-25 FA L-30 FA L-3 SF L-5 SF L-8 SF L-25 Slag-10 L-30 Slag-10 L-35 Slag-15 L-40 Slag-20 L-3 SFI L-5 SFI L-8 SFI L-10 MK L-15 MK L-20 MK L Control GUL Control GUL-40 Slag GUL-25 FA

FA FA FA FA

0.5 0.4 0.6 1.0 0.9 0.9 1.2 1.1 0.9 0.6 0.6 0.5 0.3 1.2 0.7 0.9 1.0 1.1 1.0 0.9 1.8 0.4 1.2

1.0 0.6 0.7 1.6 1.7 1.6 1.6 1.4 1.1 1.0 1.0 0.7 0.5 1.8 1.2 1.3 1.6 1.9 1.3 1.7 Failed 0.8 Failed

1.4 0.9 1.0 Failed 3.3 Failed 2.5 2.1 1.5 1.5 1.4 0.9 0.7 Failed 2.3 2.1 5.8 3.0 1.7 Failed Failed 1.3 Failed

2.3 1.2 1.1 Failed Failed Failed 1.3 2.8 1.9 2.3 1.8 1.1 0.8 Failed 3.1 3.1 22.7 Failed 1.8 Failed Failed Failed Failed

3.6 1.6 1.3 Failed Failed Failed 4.5 3.8 2.2 2.8 2.3 1.4 0.9 Failed Failed 1.3 Failed Failed 2.0 Failed Failed Failed Failed

TSA. As mentioned earlier, these blends also showed the lowest mass increases in ESA. It is also inferred from Table 7 that in general, increase in the content of each SCM in the PLC blend has led to a decrease in the mass gain, which indirectly shows improvement in performance of samples in TSA.

3.3. Comparing average expansion and average mass changes of mortar bars

it is close to the total failure, a mass loss can be found in some of the samples. This phenomenon was observed for ‘‘L-3 SF’’, ‘‘L-8 SFI’’, and ‘‘L-10 MK’’. Other samples were either not approaching failure, or had failed before the next measurement could be done. The lowest mass gain after 24 months according to Table 7 was 0.9% related to ‘‘L-40 Slag-20 FA’’. Therefore, from the point of view of mass changes, this sample performed the best against TSA. The mass gain of this sample after 24 months was equal of the mass gain of the control sample (sample No. 20) after only 3 months of TSA. It should be noted that ‘‘L-40 Slag-20 FA’’ also had the lowest expansion in TSA. Overall, comparing the mass gain results of the studied blends after 24 months, in addition to ‘‘L-40 Slag-20 FA’’, which had the best performance, the samples ‘‘L-45 Slag’’, ‘‘L-35 Slag-15 FA’’, and ‘‘L-8 SFI’’ had relatively low mass increases. Considering the mass change results in TSA at all ages, the binary PLC blends containing slag or ternary blends containing slag and fly ash showed the lowest mass increase. This demonstrates effectiveness of the mentioned blends on slowing the process of

During this study of periodic changes in mass and length of mortar bars, it was observed that, in general, the mass increased with the expansion. According to the results, mortar bars had expansion during both processes of sulfate attack. In fact, the cause for both expansion and mass gain is gypsum, ettringite, and thaumasite formation inside the hydrated cement paste; accordingly, there can be a correlation between the average expansion and the average mass gain of mortar samples. In Fig. 5, the average expansion of all blends in ESA is plotted over the average mass gain according to the two-year periodic results (368 data points). As seen in the figure, there is a good polynomial correlation with order of two between expansion and mass change results of all mortar prisms. Fig. 6 demonstrates similar relationship for mortar prisms in TSA. The number of data points is 303 (lower than ESA) because some of the samples deteriorated during the two-year study, and the negative mass changes (losses), observed for some of the samples near failure, are not presented. Generally, the increase in length has been associated with the increase in mass. However, there is not a good correlation as for ESA between the results. The main reason for such observation lies in the nature of TSA that is more aggressive than ESA and in the considerable diversity in performance of the studied blends. In TSA, in addition to expansion in the deterioration process, CSH removal occurs. During the course of two years TSA, many mortar samples started to deteriorate and lost the total integrity. This resulted in a discontinued mass gain, while the expansion was still occurring. Moreover, as explained and presented in the results section, there was a significant difference in the performance of mortar samples containing different types of SCMs. Such difference was even found in the samples with different contents of a single type of SCM. Accordingly, the correlation between the whole expansion and mass change data was not as good as the ESA. However, when the average expansion and the average mass change were studied separately for each blend, it was found that there was a very good polynomial correlation with the order of two between the two

0.800 0.700

Average expansion (%)

R² = 0.85 0.600 0.500 0.400 0.300 0.200 0.100 0.000 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Average mass change (%) Fig. 5. Average expansions versus average mass changes of mortar bars at different ages during 2-year ettringite sulfate attack (23 °C).

28

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29

0.800 0.700

Average expansion (%)

0.600 0.500 0.400

R² = 0.45 0.300 0.200 0.100 0.000 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Average mass change (%) Fig. 6. Average expansions versus average mass changes of mortar bars at different ages during 2-year thaumasite sulfate attack (5 °C).

Table 8 Correlation between the average expansion and the average mass change of each studied blend of PLC in ettringite sulfate attack and thaumasite sulfate attack (polynomial correlation with order of two). Sample

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

L-35 Slag L-40 Slag L-45 Slag L-20 FA L-25 FA L-30 FA L-3 SF L-5 SF L-8 SF L-25 Slag-10 L-30 Slag-10 L-35 Slag-15 L-40 Slag-20 L-3 SFI L-5 SFI L-8 SFI L-10 MK L-15 MK L-20 MK L Control GUL Control GUL-40 Slag GUL-25 FA

Correlation coefficient (R2)

FA FA FA FA

ESA

TSA

0.86 0.79 0.88 0.97 0.96 0.94 0.99 0.99 0.95 0.88 0.85 0.79 0.84 0.97 0.96 0.95 0.98 0.98 0.92 1.00 0.99 0.85 0.98

1.00 0.96 0.93 0.98 0.96 0.98 0.96 0.95 0.95 0.95 0.94 0.92 0.93 0.93 0.83 0.82 0.86 0.92 0.91 0.95 1.00 1.00 0.96

results in both ESA and TSA as seen in Table 8. The polynomial correlation found between the two results confirms that the mass change study can help to observe the deterioration in sulfate attack and can reinforce the expansion results. 4. Conclusions 1. All mortar bars prepared with the studied blends of PLC showed continuous expansion when immersed in sodium sulfate solution at 23 °C (ESA) and 5 °C (TSA) due to the formation of gypsum, ettringite, and thaumasite. Considering the expansion and mass change results as well as the visual inspections, TSA was found significantly more aggressive than ESA. These conclusions are expected and are well-documented in the literature. 2. An increase in the content of SCM in the PLC blend resulted in an improved resistance against ESA as well as TSA (less expansion and mass gain in mortar bars). Therefore, the studied SCMs

improved the resistance of mortar samples to ESA and TSA when added to PLC cement as well as when increased in their amount of addition. The present study showed that the CSA A3001-10 recommended minimum amounts of slag, metakaolin and fly ash were not sufficient for achieving a high sulfate resistant blended PLC. When the SCM content of PLC blends was increased over the standard recommendations for slag and metakaolin, the required resistance was reached, but none of the blends containing Type F fly ash expanded below the standard limits. However, fly ash composition is highly variable and expansion is expected to vary with different replacement levels; accordingly, more work in this area is suggested, especially on Type C fly ash that possesses both pozzolanic and cementing properties. 3. The mortar bars containing the ternary blends (PLC + slag + fly ash) had an improved performance from the viewpoint of expansion in ESA and TSA compared to the binary blends of PLC – slag and PLC – fly ash. The advantage of the ternary blends compared to the binary ones was the higher overall SCM content of ‘‘L-35 Slag-15 FA’’ and ‘‘L-40 Slag-20 FA’’ and the synergy of slag and fly ash. 4. Addition of certain amounts of SF or SFI to PLC was found quite effective on achieving high sulfate resistance. Solely addition of SF or SFI to PLC in order to achieve sulfate resistance has not been previously considered. 5. Investigating mass changes along with expansions is helpful in better understanding the performance of mortars in sulfate attack. In this study, the mass gain in the ESA as well as TSA conformed to the expansions in the two-year study; the samples with higher expansions had higher mass gains. There was a very good polynomial correlation with the order of two between the expansion and mass change of mortar samples of each PLC blend. 6. Overall, the findings of this research indicate that although the performance of PLC in TSA is extremely weak, addition of specific types of SCMs in adequate amounts may improve its resistance against TSA as high sulfate resistant blended Portland-limestone cement. In addition, it is found that current recommendations regarding the minimum required replacement levels of SCMs are too low, especially for fly ash as this study did not find performance of Type F fly ash satisfactory in TSA cases. Finally, the present research, with regards to the Canadian Standard, offers 8% silica fume, 45% slag or 20% metakaolin in binary blends of PLC as well as including 40% slag + 20% Type F fly ash or 35% slag + 15% Type F fly ash in a ternary blend of PLC in order to achieve high sulfate resistant blended PLC.

S. Mirvalad, M. Nokken / Construction and Building Materials 84 (2015) 19–29

Acknowledgments The authors wish to acknowledge the financial support of CRIB (Centre de recherche sur les infrastructures en béton), a Strategic Cluster of FQRNT (Fonds de recherche du Québec – Nature et technologies). As well, the donation of materials by Ciment Québec and Holcim Canada are gratefully recognized. References [1] Schmidt M. Cement with interground additives – capabilities and environmental relief: II. ZKG Int., Edition B 1992;45(6):296–301. [2] European Committee for Standardization EN 197-1 (2000), ‘‘Cement – part 1’’, composition, specifications and uniformity criteria for common cements, Brussels, Belgium. [3] CSA A3001. Cementitious materials for use in concrete. Toronto, Ontario: Canadian Standards Association; 2008. [4] CSA A3001. Cementitious materials for use in concrete. Toronto, Ontario: Canadian Standards Association; 2010. [5] Hooton, R., Ramezanianpour, A., & Schutz, U. (2010). Decreasing the clinker component in cementing materials: performance of Portland-limestone cements in concrete in combination with supplementary cementing materials. Concrete sustainability conference, National Ready Mixed Concrete Association, Tempe, AZ, USA. [6] ASTM Standard C595. Standard specification for blended hydraulic cements. West Conshohocken, PA: ASTM International; 2012. www.astm.org. [7] Matthews JD. Performance of limestone filler cement concrete. In: Dhir RK, Jones MR, editors. Euro-cements – impact of ENV 197 on concrete construction. London: E&FN Spon; 1994. p. 113–47. [8] Gonzalez M, Irassar E. Effect of limestone filler on the sulfate resistance of low C3A Portland cement. Cem Concr Res 1998;28(11):1655–67. [9] Barker A, Hobbs D. Performance of Portland-limestone cements in mortar prisms immersed in sulfate solutions at 5 °C. Cement Concr Compos 1999;21(2):129–37. [10] Torres S, Sharp J, Swamy R, Lynsdale C, Huntley S. Long term durability of Portland-limestone cement mortars exposed to magnesium sulfate attack. Cement Concr Compos 2003;25(8):947–54. [11] Senhadji Y, Mouli M, Khelafi H, Benosman AS. Sulfate attack of Algerian cement-based material with crushed limestone filler cured at different temperatures. Turk. J. Eng. Environ. Sci. 2010;34:131–43. [12] Ramezanianpour AM, Hooton RD. Thaumasite sulfate attack in Portland and Portland-limestone cement mortars exposed to sulfate solution. Constr Build Mater 2013;40:162–73. [13] Crammond N. The thaumasite form of sulfate attack in the UK. Cement Concr Compos 2003;25(8):809–18. [14] CSA A3000. Cementitious materials compendium. Toronto, Ontario: Canadian Standards Association; 2008. [15] CSA A23.1. Concrete materials and methods of concrete construction. Toronto, Ontario: Canadian Standards Association; 2009. [16] AASHTO M240. Standard specification for blended hydraulic cement. Washington, DC: American Association of State and Highway Transportation Officials; 2012. [17] Hooton R, Thomas MDA. The use of limestone in Portland cements: effect on thaumasite form of sulfate attack. Portland Cement Association; 2002. 2658. [18] Shi C, Wang D, Behnood A. Review of thaumasite sulfate attack on cement mortar and concrete. J Mater Civ Eng 2012;24(12):1450–60. [19] Report of the Thaumasite Expert Group. The thaumasite form of sulfate attack: risks, diagnosis, remedial works and guidance on new construction. Department of the Environment, Transport and the Regions; 1999.

29

[20] Sharp JH. Surely we know all about cement–don’t we? Adv Appl Ceram 2006;105(4):162–74. [21] Bensted J. Thaumasite—background and nature in deterioration of cements, mortars and concretes. Cement Concr Compos 1999;21(2):117–21. [22] Tsivilis S, Kakali G, Skaropoulou A, Sharp J, Swamy R. Use of mineral admixtures to prevent thaumasite formation in limestone cement mortar. Cement Concr Compos 2003;25(8):969–76. [23] Higgins D, Crammond N. Resistance of concrete containing GGBS to the thaumasite form of sulfate attack. Cement Concr Compos 2003;25(8):921–9. [24] Mulenga D, Stark J, Nobst P. Thaumasite formation in concrete and mortars containing fly ash. Cement Concr Compos 2003;25(8):907–12. [25] Skaropoulou A, Tsivilis S, Kakali G, Sharp J, Swamy R. Thaumasite form of sulfate attack in limestone cement mortars: a study on long term efficiency of mineral admixtures. Constr Build Mater 2009;23(6):2338–45. [26] Ramezanianpour AM, Hooton RD. Sulfate resistance of Portland-limestone cements in combination with supplementary cementitious materials. Mater Struct 2013;46(7):1061–73. [27] Bellmann F, Stark J. Prevention of thaumasite formation in concrete exposed to sulphate attack. Cem Concr Res 2007;37(8):1215–22. [28] Bellmann F, Stark J. The role of calcium hydroxide in the formation of thaumasite. Cem Concr Res 2008;38(10):1154–61. [29] Thomas M, Rogers C, Bleszynski R. Occurrences of thaumasite in laboratory and field concrete. Cement Concr Compos 2003;25(8):1045–50. [30] Zhang F, Ma B, Wu S, Zhou J. Effect of fly ash on TSA resistance of cement-based material. J. Wuhan Univ. Technol.-Mater. Sci. Ed. 2011;26(3):561–6. [31] Qu G, Zhang A. Influence of temperature on the resistance to sulfate attack of limestone filler concrete. Rev. Româna˘ De Materiale/Rom. J. Mater. 2012;42(4):381–6 (in English). [32] CSA A3004-C8. Test method for determination of expansion of blended hydraulic cement mortar bars due to external sulphate attack. Toronto, Ontario: Canadian Standards Association; 2010. [33] Skalny J, Marchand J, Odler I. Sulfate attack on concrete. London: Spon Press; 2002. [34] ASTM Standard C1012. Standard test method for length change of hydrauliccement mortars exposed to a sulfate solution. West Conshohocken, PA: ASTM International; 2004. www.astm.org. [35] ASTM Standard C305. Standard practice for mechanical mixing of hydraulic cement pastes and mortars of plastic consistency. West Conshohocken, PA: ASTM International; 2006. www.astm.org. [36] CSA A3004-C1. Standard practice for mechanical mixing of hydraulic cement mortars and test method for determination of flow. Toronto, Ontario: Canadian Standards Association; 2008. [37] ASTM Standard C109. Standard test method for compressive strength of hydraulic cement mortars (using 2-in. or [50-mm] cube specimens). West Conshohocken, PA: ASTM International; 2008. www.astm.org. [38] CSA A3004-C2. Test method for determination of compressive strengths. Toronto, Ontario: Canadian Standards Association; 2008. [39] Aköz F, Türker F, Koral S, Yüzer N. Effects of sodium sulfate concentration on the sulfate resistance of mortars with and without silica fume. Cem Concr Res 1995;25(6):1360–8. [40] Santhanam M, Cohen MD, Olek J. Effects of gypsum formation on the performance of cement mortars during external sulfate attack. Cem Concr Res 2003;33(3):325–32. [41] Sotiriadis K, Nikolopoulou E, Tsivilis S. Sulfate resistance of limestone cement concrete exposed to combined chloride and sulfate environment at low temperature. Cement Concr Compos 2012;34(8):903–10. [42] Santhanam M, Cohen MD, Olek J. Sulfate attack research – whither now? Cem Concr Res 2001;31(6):845–51. [43] Schmidt T, Lothenbach B, Romer M, Scrivener K, Rentsch D, Figi R. A thermodynamic and experimental study of the conditions of thaumasite formation. Cem Concr Res 2008;38(3):337–49.