Freeze–thaw resistance and sorptivity of fine‐grained alkali‐activated cement concrete

The paper investigates the freeze–thaw resistance and sorptivity behavior of fine‐grained alkali‐activated concrete cured at ambient temperature. A blended binder system containing fly ash, ground granulated blast furnace slag, and silica fume was used. A combination of sodium hydroxide and sodium silicate was used as an activator. The freeze–thaw resistance was evaluated based on mass loss (scaling), and the extent of internal damage was assessed by testing the ultrasonic time at different cycles. Initial and secondary sorptivity coefficients were calculated based on the cumulative water absorption values at different time intervals. Alkali content, sodium silicate to sodium hydroxide ratio, and water to binder ratio were investigated. The experimental results showed that water to binder ratio is the most significant parameter for the scaling; higher ratios result in higher scaling. In terms of internal damage, alkali content is the most significant. The increase of alkali increased the amount of internal damage in the concrete. The initial sorptivity coefficient increased with the water and alkali content and decreased with the silicate content. The secondary sorptivity coefficient showed no significant change with the investigated parameters.


| INTRODUCTION
The cement industry is facing significant challenges from different directions. On the one hand, the industry is under continued pressure to improve its production methods due to environmental concerns. On the other hand, the demand for cement is increasing. Depleting fossil fuel reserves and scarcity of raw materials further add to the problem the industry is facing. Improved production methods to reduce emissions are now high on the agenda. However, despite the progress made to improve cement production efficiency, about a ton of CO 2 is released for every ton of clinker produced, making the cement industry contribute about 8% of the global CO 2 emissions. 1,2 Partial or complete replacement of ordinary Portland cement (OPC) with a more environmentally friendly alternative cement has been the center of research for decades. A high ranking among these alternatives is alkali-activated cement (AAC). AAC mainly uses byproducts or secondary materials such as fly ash (FA), ground granulated blast furnace slag (GGBS), and silica fume (SF). Previous investigations showed promising results in different aspects, such as mechanical behavior improvements and fire and acid resistances. [3][4][5][6][7] Frost action by freezing and thawing cycles is a major issue in cold climates leading to the deterioration of concrete structures. Some researchers 8,9 reported the excellent freeze-thaw resistance of AAC, while others reported their poor performance. 10 Being a relatively new material, there are only limited studies on the freeze-thaw resistance of AAC concretes. Tekle et al. 11 investigated the freeze-thaw resistance of two blended binder AAC concretes and showed that the proper design of the concrete with lower water content is critical to achieving good freeze-thaw resistance. Gifford and Gillott 12 reported that sodium silicate-activated AACs have a better freeze-thaw resistance than those activated with sodium carbonate. Cai et al. 13 found that activator to slag ratio and slag content are the most prominent factors affecting the freeze-thaw resistance for slag-based AACs. Bilek et al. 14 showed that the freeze-thaw resistance of AAC depends on the alkali to silicate ratio of the activator. They also found that water curing negatively impacts freeze-thaw resistance compared to curing by covering with foils.
Sorptivity is a measure of a material's ability to absorb and transport water by capillary suction. It provides an engineering measure of microstructure and properties critical for durability. Despite various research on the characterization of sorptivity for OPC concrete, the sorptivity of AAC concrete is yet to be understood, particularly blended binder AAC systems. Albitar et al. 15 found that fly ash-based AAC concrete has higher cumulative water absorption and sorptivity coefficient than OPC. Another study on fly ash-based AAC observed the opposite. 16 This shows that in addition to the binder type, the sorptivity of AAC depends on other parameters.
There are different types of AACs depending on their source materials, activators, and curing methods, making it even more challenging to get a conclusive remark on their freeze-thaw resistance and sorptivity behavior. Hence, more research is warranted in the area. This study investigates a blended binder (FA, GGBS and SF) AAC fine-grained concrete activated with sodium silicate and sodium hydroxide solutions. Durability problems associated with freezing and thawing come in two primary forms: internal cracking and surface scaling. The internal cracking can be evaluated through the dynamic elastic modulus. The surface scaling is assessed by material mass loss on the concrete surface. In this study, both internal cracking and surface scaling are investigated. Sorptivity is evaluated by measuring the absorbed water at different time intervals. The alkaline solid to binder ratio, the water to binder ratio, and the sodium silicate to sodium hydroxide ratio were varied to investigate their effect on freeze-thaw resistance and sorptivity.

| Experimental method and mix proportions
The study is part of a larger effort to create a concrete mix for use in textile-reinforced concrete. Textile  reinforcement requires fine-grained concretes due to its smaller openings; hence, no coarse aggregate was used. The mixtures are therefore termed fine-grained AAC (FGAAC). The constituents of the precursor are set at 55% FA, 40% GGBS, and 5% SF based on the authors' previous work. 20 A control mixture termed FGAAC-C was designed targeting a high-strength concrete. The mix parameters of FGAAC-C were then varied to understand their effect on the freeze-thaw resistance and sorptivity of the mixture. The sand to binder ratio was kept constant at 1.82. The parameters investigated were alkaline solids to binder (precursors) ratio (AS/B), sodium silicate to sodium hydroxide solids ratio (SS/SH), and total water to total solid binders ratio (TW/TS). SS and SH are only the solid portions of the alkaline liquid. TW is the amount of the free water and the water from the alkaline solution, while TS is the total amount of solid binder used, that is, precursors (B) and alkaline solids (AS). The mix proportion of each mix is shown in Table 2.

| Mechanical strength tests
The sodium silicate and sodium hydroxide solution were mixed in the required proportion at least 24 h before concrete mixing. AAC specimens were prepared by mixing the dry materials (sand and binder) in a mixer for about 2 min. Afterward, the prepared alkaline solution was mixed with the additional water, added slowly to the dry mixture, and mixed for about 4 min. The SP was added halfway during the wet mixing. The mixture was then placed into 40 mm Â 40 mm Â 160 mm prisms for compressive and flexural testing according to EN 1015-11. 21 Three prisms were produced for each mixture. After casting, the specimens were placed in the environmental control room (20 C temperature and 65% relative humidity). The specimens were demolded after 24 h and kept in the environmental control room until the test day.

| Freeze-thaw tests
For the freeze-thaw specimens, the concrete was cast into 150 mm cube molds. Two polytetrafluoroethylene (PTFE) plates were placed firmly in the cube mold, one at each side of the mold, as per the requirement of EN 12390-9. 22 The test surface is the concrete surface cast against the PTFE plate. The test specimens were demolded after 24 h and kept in the environmental room covered with plastic until 21 days. The test specimens were then sawn to the standard size, that is, 150 mm Â 150 mm Â 70 mm (length Â width Â height). The lateral surfaces of the test specimen were cleaned and treated with a primer and sealed with aluminum foil butyl rubber tape. The test specimens were left in the environmental control room for 7 days. Five specimens were prepared for each of the investigated parameters. The degree of saturation of concrete at the beginning of the freezing and thawing test is critical to the freeze-thaw resistance of the concrete. 23 Deterioration increases with the degree of saturation. EN 12390-9 22 recommends 7 days of presaturation of the dried concrete before the freezing and thawing test starts. As per the recommendation, after curing in the environmental control room, the test specimens were placed in the test containers on 5 mm high spacers with the test surface at the bottom. Deionized water was then poured into the container up to a height of 10 mm without wetting the top surfaces of the specimens. The  specimens were left in the container for 7 days for saturation with the lids closed at a temperature of 20 C. The increase in the specimen weight was then determined at the end of the saturation period.
After saturation, the test specimens were treated with an ultrasonic bath to clean any loose particles on the test surface. The specimens were then exposed to cycles of freezing and thawing. A cycle of freezing and thawing lasts for 12 h. The temperature is reduced from +20 C to À20 C in 4 h. After keeping the temperature at À20 C for 3 h, the temperature is increased back to +20 C in 4 h and kept at this temperature for an hour. Figure 1 shows the measured temperature during the freezing and thawing cycles. Schleibinger Geräte Teubert & Greim GmbH CDF/CIF-freeze-thaw testing equipment was used for the freeze-thaw test.
Surface scaling, moisture uptake and ultrasonic measurements were carried out at the beginning of the freeze-thaw test (0 freeze-thaw cycles) and after 4, 10, 16, 24, 38, and 56 cycles. Figure 2 shows the test surface, the sealing, and the ultrasonic testing of the freeze-thaw specimens. Vikasonic ultrasonic testing machine from Schleibinger Geräte Teubert & Greim GmbH was used for the ultrasonic measurement.
The total scaling is calculated by using Equation (1): where m n is the total quantity of scaled material in relation to the test surface at each measurement, g/m 2 , μ s is the mass of the scaled material at each measurement. The total scaling is the sum of all measurements up to the nth cycle. A is the size of the test surface in m 2 calculated on the basis of the linear dimensions. The moisture uptake is calculated using Equation (2): where ΔW n is the moisture uptake of each test specimen at each measurement, in % by mass, W 0 is the reference mass of each test specimen without the mass of the sealing material after dry storage, in g, W 1 is the mass of each test specimen including the mass of the sealing material before the beginning of presaturation, in g and W n is the mass of each test specimen at each measurement, in g. The ultrasonic transit time is calculated based on Equation (3). The change in transit velocity τ n after freeze-thaw cycles is calculated separately for each specimen using Equation (4). The internal damage is assessed using the relative dynamic modulus of elasticity (R u,n ) which is calculated from the ultrasonic transit time according to Equation (5).
where t c is the transit time in the test liquid in μs, l c is the transit length of the test liquid, V c is the velocity of the ultrasonic signal in the test liquid. It is taken as 1,490 m/s. τ n is the relative transit velocity, t cs is the total transit time after presaturation in μs, prior to the first freeze-thaw cycle, t n is the total transit time after n freeze-thaw cycles in μs. R u,n is the relative dynamic modulus of elasticity. The lower the value, the higher the internal damage.

| Sorptivity tests
The rate of water absorption determines the durability performance of concrete. The specimens for the sorptivity test were prepared according to ASTM C1585. 24 The specimens were obtained from molded cylinders with a diameter of 100 mm and a height of 200 mm. The cylinders were 5 months old at the time of preparation. Three specimens were obtained from the cylinder for each mix group by cutting it to a length of 50 mm per ASTM C1585. The specimens were placed in an environmental chamber at a temperature of 50 C and relative humidity of 80% for 3 days. The specimens were then placed in a sealable container and stored for 15 days at 23 C. Then, sealing was applied on all the surfaces except the bottom. The specimens were kept in water up to a depth of F I G U R E 1 Measured temperature during freezing and thawing cycles 2-3 mm above the base of the specimen. Figure 3 shows the water absorption test in progress. After that, the specimens were weighed at different time intervals up to 22 days to measure the amount of water absorbed. Care was taken to clean all the surface water before weighing. The initial and secondary sorptivity were calculated from the cumulative water absorption and the square root of the time lapse. Table 3 shows the average results for the 28-day compressive and flexural strength, moisture uptake, total scaling, and internal damage at different freeze-thaw cycles. The minimum compressive strength was obtained for FGAAC-TW/TS mix. At 36.9 MPa, this mix satisfies the minimum concrete strength recommended (25 MPa) before concrete is exposed to freezing and thawing cycles. 23,25 The control mix has a much higher strength,

| Sorptivity
Sorptivity is used as a parameter to assess the durability of concrete as it characterizes the capability to absorb and transmit water in a nonsaturated state. The water absorption in relation to the square root of time is plotted in Figure 4. The process of water absorption can be divided into two stages: initial absorption, characterized by a rapid rate of absorption, and secondary absorption, characterized by a steady rate of absorption. To calculate the sorption coefficient for each of these stages, obtained data were fitted to Equation (6).
where I is the water volume absorbed per unit cross-sectional area, S is the sorptivity coefficient, t is the time and S 0 is a correction term added to account for surface effects at the time sample is placed in contact with water. Figures 5 and 6 show the initial and secondary sorptivity coefficients obtained from the curve fitting of the cumulative water absorption in relation to the square root of time for each mix. Data points up to 6 h were used for the regression analysis of the initial sorptivity behavior as per ASTM 1585. For the secondary absorption behavior, however, the final data points between 5 and 22 days were used.  The initial sorptivity coefficient was much higher than the secondary coefficient for all the mixes. The initial is in the range of 4:3 Â 10 À3 mm=s 1=2 to 9:1 Â 10 À3 mm=s 1=2 while the final is about 3:5 Â 10 À4 mm=s 1=2 on average. The initial water absorption is mainly controlled by capillary pores. 27,28 After the initial rapid absorption is completed, water continues to transport into the gel pores, which is primarily controlled by diffusion. 29 The FGAAC-TW/TS showed a significantly higher absorbed water content and initial sorptivity coefficient. FGAAC-TW/TS has a high water to binder ratio; 0.51 TW/TS compared to 0.34 for the control. Similar to conventional OPC pastes, the accessible porosity of AAC pastes increases with the increase of the water to binder ratio. 30 Hence, FGAAC-TW/TS has higher pore connectivity than the other mixes resulting in more absorbed water and a higher rate of water absorption.
FGAAC-AS/B showed a higher water absorption value (Figure 4). It also showed a slightly higher initial sorptivity coefficient than the control ( Figure 5). This is due to the higher AS/B ratio, which also resulted in a lower compressive strength compared to the control mix. The increase in the alkali content above an optimum amount reduces the concrete's strength due to the formation of calcium hydroxide around the GGBS surface, preventing the formation of calcium silicate hydrate gel. 31,32 FGAAC-SS/SH showed a lower coefficient of sorptivity than the control mix. FGAAC-SS/SH showed higher compressive strength than the control mix, contributing to its lower water absorption. In FGAAC-SS/SH, the SS/SH ratio F I G U R E 5 Curve fitting for the initial sorption F I G U R E 6 Curve fitting for the secondary sorption was increased from 1.75 to 2.5. The higher SS/SH ratio means more silicate in the alkaline solution. Bignozzi et al. 33 reported that increasing the proportion of sodium silicate solution generally encourages gel formation, resulting in higher strength. The increased gel formation lowers the porosity of the concrete, consequently lowering the water absorption. No significant variation was observed in the secondary sorptivity coefficient between the mixes. Figure 7 shows the specimens' surfaces at 4 and 56 freeze-thaw cycles. Figure 8 shows the mean scaling values at each of the cycles. The concrete surface was barely damaged after four freeze-thaw cycles except for the FGAAC-TW/TS specimen, which showed some noticeable surface damage. At the 56th cycle, the FGAAC-TW/TS specimen was severely damaged with a significant amount of scaling as can be observed in Figures 7 and 8. This is due to the high water to binder ratio of the mixture. For non-air-entrained concretes, a water to binder ratio of less than 0.3 is usually needed to reach an acceptable durability level. 34 FGAAC-TW/TS has a TW/TS (which can be taken as equivalent to water to cement ratio) ratio of 0.51, leading to high capillary pores and consequently lower freeze-thaw resistance. According to the BAW code of Practice 35 the mean value of the surface scaling should be less than 1,000 g/m 2 after 28 freeze-thaw cycles. This criterion is shown in Figure 8. All mixes except the FGAAC-TW/TS satisfy this acceptance requirement.

| Surface scaling
As can be observed from Figures 7 and 8, the control mixture (FGAAC-C) performed better than any of the mixtures. The control mixture has lower AS/B, SS/SH and TW/TS ratios compared to the other mixes. As explained in Section 3.2, for a higher than optimum AS/B ratio, a decrease in the ratio results in lower amount of calcium hydroxide formation. This, in turn, increases the compressive strength and durability of the concrete. Hence, the lower scaling of FGAAC-C compared to FGAAC-AS/B is due to its lower calcium hydroxide formation. FGAAC-SS/SH has the highest compressive strength. This is due to the higher silicate, which encourages gel formation. 33 As can be observed from Table 3, FGAAC-SS/SH showed the lowest scaling up to the 24th cycle. However, after the 24th cycle, despite its highest 28 days strength, FGAAC-SS/SH did not show the lowest scaling. Previous studies have shown that sodium silicate-activated AACs have more surface cracks caused by shrinkage and lower internal adhesion than sodium F I G U R E 7 Freeze-thaw specimens' surface after 4 and 56 cycles F I G U R E 8 Mean scaling hydroxide-activated AACs. 32,36 Similarly, the higher silicate content of FGAAC-SS/SH mixes may have resulted in a higher shrinkage and cracks than FGAAC-C mixes, resulting in higher scaling. However, further investigation is warranted to understand this parameter better.

| Internal damage
The relative dynamic elastic modulus (R u,n ) can be used to measure the internal damage of concrete. This measures the deterioration of concrete samples even when there is no visible damage. According to the BAW code of Practice, 35 concrete is damaged if the relative dynamic elastic modulus calculated according to Equation (5) is less than 0.75. The relevant criterion for evaluating internal damage is the number of freeze-thaw cycles before the damage criterion is reached. As per the BAW Code of Practice, the acceptable number is 28 freeze-thaw cycles. Figure 9 shows the mean internal damage in terms of relative dynamic elastic modulus curves for the concrete mixtures. Similar to the scaling, the control mix performed better than any of the mixes, while the FGAAC-SS/SH mix showed similar behavior. The higher TW/TS ratio in FGAAC-TW/TS mixture increased the internal damage. However, FGAAC-TW/TS is no longer the worst-performing mix regarding internal damage. FGAAC-AS/B showed the worst internal damage despite its lower scaling. However, according to the acceptance criterion, it is still acceptable as the relative dynamic elastic modulus at 28 freeze-thaw cycles is higher than 0.75. FGAAC-AS/B's relative dynamic elastic modulus reached 0.75 at about 55 cycles.
The higher AS/B ratio of FGAAC-AS/B may have facilitated the formation of more calcium hydroxide in the matrix resulting in higher porosity and damage with the ingress of water. Fu et al. 9 also observed the better performance of low AS/B ratio mixes in terms of internal damage. Cai et al. 13 observed the higher significance of alkaline solution to slag ratio compared to slag content and sand ratio. A lower solution to slag ratio led to better performance. This is an important finding showing that a limit is required on the amount of alkaline content in the mixture for optimum mechanical and durability behaviors. Tekle et al. 26 studied 0.14, 0.18, and 0.22 AS/B ratios. The maximum 28-day compressive strength was obtained at 0.18.

| CONCLUSIONS
Freeze-thaw resistance and sorptivity of blended binder alkali-activated concrete were investigated. Alkaline solid to binder ratio (AS/B), sodium silicate to sodium hydroxide ratio (SS/SH), and total water to total solid (TW/TS) ratio were taken as parameters. The sorptivity, surface scaling, and internal damage were investigated. Based on the experimental results, the following conclusions can be drawn: • The initial sorptivity coefficient decreased with the SS/SH ratio and increased with the AS/B ratio. It increased significantly with the TW/TS ratio. • The secondary sorptivity coefficient showed no significant change with the mix design variation. • In the investigated parameter ranges, lower AS/B, SS/SH, and TW/TS ratios are desirable for better scaling resistance of the AAC concrete. • An increase in the AS/B ratio resulted in the worst internal damage.
The study showed that from the investigated parameters, lower water to binder ratio and an optimum amount of alkali to binder ratio are the most critical parameters for the durability of AAC concrete.

ACKNOWLEDGMENTS
The authors gratefully acknowledge the financial support of the Alexander von Humboldt Foundation (1206836-AUS-HFST-P).

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.