1. Introduction
Superabsorbent polymers (SAPs) are 3D polymer networks that can retain large amounts of liquids, up to a thousand times their own dry weight, forming a hydrogel without dissolving [
1]. This special feature makes SAPs very interesting to be used in concrete applications for different reasons [
1]: mitigating autogenous shrinkage [
2,
3,
4], modifying the rheology of fresh concrete [
2,
3], self-sealing and self-healing of cracked concrete [
5,
6,
7,
8,
9,
10,
11], increasing the freeze-thaw resistance of concrete [
12,
13] etc. Hence, they can help to increase the durability of concrete structures in different ways.
Various types of SAPs are currently available. On a chemical base, two extreme types can be distinguished: ionic (charged) and non-ionic (neutral) SAPs [
14]. Commercially available SAPs are mostly ionic polymers and show higher absorption capacity. Another distinction can be made based on the type of cross-links occurring between the polymer chains: covalent, ionic or hydrogen bonding. For application in mortar or concrete, covalent cross-links are preferred as they are stable and strong [
14]. These cross-links are the most essential part of the SAPs, as these bonds prevent the SAPs from dissolving. In the case of highly concentrated solutions of acrylic acid, self-crosslinking through hydrogen bonding can occur, although it has been found that these SAPs have a lower water absorption compared to polymers with the participation of the crosslinking agent and covalent bonds [
15].
In addition to the aforementioned different properties of SAPs, the following parameters can also be used to distinguish SAPs: type (synthetic, semi-synthetic, natural), cross-linking degree, polymerization technique (bulk, suspension), particle size, etc.
A decrease in swelling ratio is expected when the crosslinking density increases for SAPs submerged in deionized water [
16,
17] or in calcium rich solutions [
18]. This result is logical as the cross-links will impede the SAPs from swelling. However, a certain amount of crosslinking is always needed as otherwise the material would dissolve upon exposure to liquids.
SAPs with varying particle sizes have already been implemented in mortar or concrete: 100 µm [
19], 125–150 µm [
20], 125–250 µm [
3], 100–500 µm [
21], 106–425 µm [
16], 425–850 µm [
16] and even up to some mm [
22]. For mitigating autogenous shrinkage in cementitious materials, smaller SAP particles (d
50 around 100–150 µm) are used [
19,
20], whereas for self-sealing and self-healing larger particles, sizes up to 500 µm are often reported in literature [
5,
7,
22]. A smaller particle size is expected to have a smaller swelling capacity due to the smaller active circumference of the SAP compared to the bulk volume [
21], although other researchers state the opposite: smaller SAPs will have a larger swelling capacity due to their larger surface area [
22,
23,
24]. However, several researchers state that the particle size does not have an influence on the swelling capacity [
16,
25]. In this paper, SAPs with average dry particle size of 10, 40 and 100 µm were investigated. These particle sizes are smaller than the sizes studied in most published papers. However, it could be interesting to embed these SAPs in concrete, as their dimensions are in the same range as the other constituents in concrete like cement, fly ash or limestone.
When synthesizing the SAPs, the possibility and ease of upscaling should be kept in mind, as (large) concrete structures will require a lot of SAPs. In this view, topics such as water consumption, need for purification, synthesis time, polymerization technique, etc. will play a role in deciding which SAPs will be suitable for application in concrete. Moreover, it is important to know whether the solubles or the fraction of the non-cross-linked polymer present in the final SAP have a negative effect on the mortar properties. This parameter is often expressed as its complement called the gel fraction. Most SAPs used in literature have a high gel fraction meaning that the amount of material that is chemically incorporated in the 3D network is high. In literature, the reported gel fractions vary from >80% [
26], 95% [
27,
28] to nearly 100% [
29]. The gel fraction of SAPs can be increased by removing unreacted particles via dialysis with demineralised water. However, this procedure is quite time and water consuming. Zohuriaan-Mehr states that one of the functional features of an ideal SAP material is a low soluble content and residual monomer [
25]. In this study, it was investigated whether the presence (of a high number) of solubles (40% or more) has a negative effect on mortar properties. It could be interesting to know whether it is necessary to purify the SAPs and to remove the non-cross-linked soluble fraction or not, keeping the ease of upscaling in mind.
In this paper, SAPs with different cross-linking degrees, particle sizes and amount of solubles are investigated. By varying these parameters, insight can be gained on the influence of each of these parameters on SAP properties such as the swelling capacity. In a next step, the SAPs can be implemented in mortar to assess their influence on mortar properties like workability, compressive strength or hydration kinetics. Based on these results, the ‘ideal’ SAP with tuneable properties for a specific concrete application can be selected.
3. Results & Discussion
3.1. Influence of the Cross-Linking Degree
SAPs with five different amounts of cross-linker (mol%) were produced, namely 0.15 mol%, 0.38 mol%, 0.575 mol%, 0.775 mol% and 1.0 mol%. The measured amounts of solubles for these SAPs were respectively 28%, 44%, 18%, 36% and 19% and were, compared to other literature [
27,
28,
29,
30], quite high. The higher the amount of cross-linker, the lower the experimental swelling capacity obtained through a filtration test, as indicated in
Figure 1. The SAP with the highest amount of cross-linker (CS_1.0_40 with 1.0 mol% cross-linker) had a swelling capacity in demineralised water of 66 ± 4 g/g SAP while the SAP with the lowest degree of cross-linker (CS_0.15_40 with 0.15 mol% cross-linker) had a four times higher swelling capacity in demineralised water (270 ± 17 g/g SAP). This result was logical as the crosslinks will impede the SAPs from swelling. The same trend is seen for the swelling in cement filtrate solution and this is in accordance with findings reported earlier in literature [
16,
17,
18]. The results of all SAPs show that their swelling capacities in cement filtrate solution are much lower compared to their swelling capacity in demineralized water. Cations like K
+, Na
+, Mg
2+ or Ca
2+ present in cement filtrate solution, give rise to a so-called charge screening effect of the negatively charged polymer chains, resulting in a lower repulsion of chains and a lowered fluid absorption and less swelling of the SAP particles. Furthermore, the presence of divalent cations like Ca
2+ gives rise to an additional reduction of the swelling properties as these cations form strong complexes with the sulfonate groups and can therefore act as cross-linkers [
16,
33,
38,
39].
When looking at the calculated values of the swelling capacity that take into account the amount of solubles (
Figure 1), it can be seen that two mixtures do not follow the expected trend, namely CS_0.38_40 and CS_0.775_40. Both SAPs show a higher calculated swelling capacity than expected, due to their high amount of solubles of 44% and 36% for CS_0.38_40 and CS_0.775_40, respectively.
In a case where no additional water is added to the mixture, the workability would be negatively affected as the SAPs would absorb part of the mixing water. Therefore, it was decided to add an amount of additional water of 1.5 times the experimentally obtained swelling capacity in cement filtrate from the filtration test, based on previous tests using similar SAPs. The initial flow of the reference mixture was 205 mm and decreases after 120 min to 168 mm. The initial flow and flow over time of the reference and the five SAP-containing mixtures can be seen in
Figure 2.
With the aforementioned amount of additional water, two mixtures containing SAPs, namely CS_0.575_40 and CS_1.0_40 show similar workability in time as the reference. Their reduction in flow over the first two hours was limited and somewhat smaller than what was noticed for the reference mixture without SAPs. The SAP containing mixtures CS_0.15_40 and CS_0.38_40 however, showed a larger initial flow, meaning that the amount of additional water was not fully absorbed by the SAPs immediately. These two SAPs contained the smallest amount of cross-linker, respectively 0.15 mol% and 0.38 mol% and showed the largest deviation in swelling capacity, as can be seen in
Figure 1. For mixture CS_0.775_40 the initial flow is somewhat lower than for the reference mixture (17.5 cm instead of 20.5 cm). For this mixture, the amount of additional water was underestimated, and the SAPs absorbed part of the mixing water, resulting in a lower initial flow. However, when looking at the flow values after 120 min, it can be seen that all the mixtures have a similar flow in the range 18–22 cm. The reference mixture without SAPs has even the lowest flow of 16.75 cm after 120 min. Differences in swelling kinetics between the different SAPs could be a reason for the delayed water uptake by the SAPs and the varying workability over time.
From these results, it can be seen that the correct amount of additional water is not that straightforward to determine. The correct amount of additional water, in order to obtain the same workability as the reference mixture, should be determined for each SAP separately by a trial-and-error procedure, starting from the swelling capacity in cement filtrate solution obtained from a filtration test. Not only the initial flow should be taken into account, but also the flow after some time, for example after 120 min, should be taken into account as different SAPs will have different swelling kinetics. Although this is a time-consuming method, it is very important to add the correct amount of additional water as an over-or underestimation of this amount will change the W/C ratio of the mixture, which will have a significant influence on the test results.
Figure 3 shows the heat production rate as a function of time. It can be seen that the hydration of the mixture containing SAPs and with the correct amount of additional water and thus approximately the same W/C ratio as the reference (namely CS_0.575_40) is slightly retarded with 2.5 h compared to the reference mixture as indicated by the small shift to the right and by the reduction in maximum heat flow values. The SAPs with the lowest swelling capacity, i.e., CS_1.0_40 with 1.0 mol% cross-linker, show a negligible effect on cement hydration. A retardation in setting time caused by the addition of SAPs was already reported by several authors [
40,
41,
42,
43]. However, the mixtures with SAP that show a slower absorption of the additional water, i.e., CS_0.15_40 and CS_0.38_40, show a much larger delay in the setting time of 6.5 h compared to the reference, as the actual W/C ratio of these mixtures is higher (and unwanted) compared to the reference [
44]. Therefore, the delay in setting time for these mixtures is caused by a combination of the higher W/C ratio and the presence of the SAPs. A premature release of stored water may also have caused this.
The compressive strength of the different mixtures after 3, 7 and 28 days is depicted in
Figure 4, together with the experimentally obtained swelling capacity in cement filtrate of the unpurified SAPs. Two main conclusions can be made based on these results:
- (1)
The higher the swelling capacity (i.e., the lower the amount of cross-linker), the lower the compressive strength. This result is logical as the swollen SAPs will create macro pores in the matrix, negatively affecting the compressive strength compared to the reference [
1,
2,
3,
45]. After 28 days, the compressive strength of SAP CS_0.15_40 (i.e., with the lowest amount of cross-linker) was 67% lower than the value noticed for the reference. SAP CS_1.0_40 (i.e., with the highest amount of cross-linker) had a compressive strength at that age which was 23% lower than what was noticed for the reference.
- (2)
The difference in strength with the reference decreases as a function of the mortar age. For example, for SAP CS_1.0_40 this difference decreased from −35% after 3 days to −29% after 7 days and to −23% after 28 days compared to the reference. This trend was observed for all the SAPs with different cross-linking degrees.
For the flexural strength at 3, 7 and 28 days similar trends were found. After 28 days, the flexural strength of SAP CS_0.15_40 (i.e., with the lowest amount of cross-linker) was 4.5 ± 0.2 MPa, which was 52% lower than the value noticed for the reference of 9.5 ± 0.5 MPa. SAP CS_1.0_40 (i.e., with the highest amount of cross-linker), and had a flexural strength at that age which was 26% lower than what was noticed for the reference. Also, the difference in flexural strength with the reference decreased as a function of the mortar age.
3.2. Influence of Particle Size
In a second test series, the influence of different particle sizes was investigated. For this purpose, the SAPs from Series 1 resulting in the lowest strength reduction and with the smallest effect on the hydration rate were selected, namely CS_0.775 (0.775 mol% cross-linker) and CS_1.0 (1.0 mol% cross-linker). These SAPs were ground to particle sizes with d
50 of 10, 40 and 100 µm. The particle size distributions of the SAP CS_1.0 with mean particle size 10, 40 and 100 µm are depicted in
Figure 5.
Similar tests as for test Series 1 were performed. The results are summarized in
Table 3 and
Table 4. The results of the SAP CS_1.0 with different particle sizes of 10, 40 and 100 µm were not significantly different (α = 0.05) concerning the swelling capacity in both demineralised water and cement filtrate solution (
Table 3) and mortar compressive strength after 28 days (
Table 4). The swelling capacity in demineralised water for CS_0.775_40 was found to be significantly different from the swelling capacity of SAPs with particle sizes of 10 µm and 100 µm (
Table 3). It must be noted that the swelling capacity of the SAPs with particle sizes of 40 µm was not measured at the same time as for the particle sizes of 10 and 100 µm. As a result, differences in temperature, solution or other conditions could have had an influence on the results. In cement filtrate solution, the swelling capacity was not significantly different (α = 0.05) between the studied particle sizes for SAP CS_0.775 (
Table 3). This is in accordance with findings in literature [
16,
25]. Also, the mortar compressive strength at 28 days was not significantly different between the studied particle sizes for this SAP (
Table 4).
From these results it can be concluded that a change in particle size in the range of 10–100 µm does not have a significant influence on the studied parameters.
In
Table 4, the importance of determining the correct amount of additional water based on both the initial flow and the flow in time can be seen. The initial flow of mixtures CS_0.775_40 and CS_1.0_40 was remarkably lower than the flow of the other SAP-containing mixtures. However, the same remark as earlier must be made that the tests with SAPs with particle size of 40 µm were not conducted at the same time as the tests on SAPs with particle sizes of 10 and 100 µm. Although several mixes showed a much larger initial flow compared to the reference, this difference disappears after 120 min and all the mixtures had a similar flow in the range of 18–19 cm.
3.3. Influence of Solubles
With the ease of upscaling in mind, the need to purify the SAPs and remove the non-cross-linked soluble fraction was further investigated. For this purpose, three more SAP types were synthesized by ChemStream with different amounts of solubles: 100%, 40% and 10% solubles. The d50 of these SAPs were respectively 20 µm, 100 µm, 30 µm. From the previous paragraph, it was proven that the particle size did not have a significant influence on the results.
From
Table 5 and
Table 6 it can be seen that the lower the amount of solubles, the higher the swelling capacity and as a consequence the lower the compressive strength. Although the SAP CS_SOL_100% was made without any cross-linker, this SAP showed a very limited swelling capacity in demineralized water and cement filtrate solution. The Ca
2+ ions present in cement filtrate solution can form strong complexes with the sulfonate groups of the SAPs and can therefore act as cross-linkers, resulting in a limited swelling capacity, even in the absence of cross-linker.
However, the presence of solubles itself has no effect on the studied mortar properties as CS_SOL_100% shows similar results compared to the reference mixture, see
Table 6. The only effect of the amount of solubles can be seen when investigating the hydration kinetics, as the graph for CS_SOL_100% has shifted to the right, meaning an increase of the setting time, see
Figure 6. The peak in the heat production rate shifts from 840 min (i.e., 14 h) for the reference mixture to almost the double amount of time of 1635 min (around 27 h) for the mixture CS_SOL_100%.
It can be concluded that extra purification of the SAPs is not needed as the presence of solubles in unpurified SAPs does not negatively affect the studied mortar properties. It was therefore confirmed that omitting the extra purification step for the SAPs used in Series 1 and Series 2 had no negative effect.
However, from an economical point of view, it could be interesting to purify the SAPs in order to have the most efficient SAP addition in the concrete. In the case of unpurified SAPs, part of the added SAP will not be active and will thus not have any benefit in the concrete. In case of purification of the SAPs, the water and energy consumption and drying time must be optimized in order to still be economical interesting.
The possibility, ease and cost-effectiveness of purifying the SAPs should be considered when upscaling the SAP production.