Salt Scaling Resistance of Variable w/c Ratio Air-Entrained Concretes Modified with Polycarboxylates as a Proper Consequence of Air Void System

The values of the air void parameters in hardened concrete (spacing factor L ≤ 0.200 mm and micro air content A300 ≥ 1.5%), determined on the basis of the Powers model, in concretes produced today do not always guarantee the frost resistance of the concrete, especially when in surface impact with the participation of de-icing agents. The literature indicates that the modified polycarboxylates used to liquefy concrete mixes are one of the factors involved in changing the air void system; therefore, the aim of the article was to determine the dependence of the air void parameters and the resistance to scaling of concretes liquefied to a constant consistency by the use of modified polycarboxylates in the spectrum of variability of the ratio w/c = 0.53 ÷ 0.30. In the research program, twelve concrete mixes were made with a constant proportion of aggregate and paste: six air-entrained—with a constant air content of 5.5 ± 0.5%—and six non-air-entrained. The air void parameters were determined in accordance with EN 480-11, while the resistance to scaling was determined in accordance with CEN/TS 12390-9 and assessed according to the criteria of SS 137244. The analysis of the test results showed that liquefaction with modified polycarboxylates did not affect the w/c limit values, enabling obtaining concretes resistant to scaling. They are, respectively, 0.35 in the non-air-entrained concretes and 0.50 in the air-entrained concretes with an air content of 5.5 ± 0.5. Moreover, the commonly used criterion for ensuring the frost resistance of air-entrained concretes, L ≤ 0.200 mm and A300 ≥ 1.5%, requires supplementing with the minimum value of the w/c ≤ 0.50.


Introduction
Scaling, which is a surface effect of frost with the participation of defrosting agents, is the dominant factor in the destruction of road and bridge concretes [1][2][3][4][5][6][7][8]. Defrosting agents used on road concrete pavements may cause several-millimeter splinters of hardened cement paste on the concrete surface; this is caused by an increase in osmotic pressure, microcracks due to thermal shocks, or chemical reactions [1,5,[8][9][10][11]. While freezing on a concrete surface, saline solution forms a two-material composite of ice and concrete. The temperature of this composite drops below the melting point of the solution. Then, the ice layer shrinks five times more than the concrete underlying it does [2,3,8]. As a result, the top layer of concrete breaks into many small sheets. Tensile stress develops in concrete along the perimeter of these patches, causing the defects to spread over the concrete surface, culminating in the removal of a thin layer of concrete. Fracture mechanics allows us to predict that the fracture propagation and the depth at which the fracture rotates parallel to the transition zone are dictated by the mechanical properties of the constituent materials. Therefore, if the properties of the concrete do not differ significantly across the surface, each freeze-thaw cycle causes a relatively constant amount of damage. The resistance to scaling of concrete is determined by the strength of the top layer. The estimated chipping stress is marginal in relation to the tensile strength of the cement binder (∼3 MPa) [2]. Any In the research program [36] on the scope of the influence of w/c on the AVS, it was shown that the AE of concrete mixtures at the level of 5.5 ± 0.5% allows the obtaining of the AVP required to ensure frost resistance [6,14]: L ≤ 0.200 mm and A 300 ≥ 1.50%. In addition, the AVP is described by mathematical relations with high correlation coefficients: total air content (A), L, α, and A 300 in both AC and NAC change with the w/c ratio. Lowering the w/c ratio value increases the viscosity of the concrete mixes, thus influencing the airentraining process; at a constant consistency, this requires a higher dose of SP, the side effect of which is an additional content of pores above 300 µm, which are unfavorable to frost. The w/c ratio does not affect the content of pores below 300 µm, which are favorable for frost resistance. The most favorable structure of air entrainment was obtained in concrete with w/c = 0.45 due to the optimal viscosity of the concrete mixture allowing for the emergence of large pores during compaction and preventing the coalescence of fine pores, as well as to the content of SP, which does not cause additional AE.
In the previously performed test program [33], at a constant value of the w/c ratio = 0.45, a very good resistance to scaling of all the AC was obtained, which made it impossible to determine the relationship between AVP and the resistance to scaling. Moreover, the research on the influence of w/c on the AVS in concrete [36] shows that as the w/c decreases and the demand for SP increases, the value of the concrete, unfavorable for frost resistance, increases. Therefore, the aim of the article was to determine the relationship between AVP and the resistance to scaling of concretes liquefied to a constant consistency with the use of modified polycarboxylates in a wide range of the variability of the w/c ratio. . The cement has a specific surface according to a Blaine of 4516 cm 2 /g and a density of 3.1 g/cm 3 . The physical and chemical properties of the cement specified by the manufacturer are presented in Table 1. Natural sand and a crushed basalt aggregate with a maximum grain of 16 mm and a sand point of 36.8% were used in the concrete mixes. The crumb pile was iterated to obtain an aggregate mix with a minimum sum of cavity and water demand. Based on our own analysis of the grain size distribution of individual aggregates, a grain size distribution curve was constructed for the crumb pile, composed by iteration. The graining curve of the aggregate used, presented in Figure 1, has been drawn into the boundary curves for the reference concrete in accordance with EN 480-1 [37]. The obtained graining curve shows insignificant contents of oversize and undersize. However, they were allowed in the test program due to the fulfillment of more important criteria in the scope of tightness and water demand.  Table 2 lists the properties of the air-entraining admixture (AEA) and the superplasticizer (SP), as specified by the manufacturers. The chemical compositions of the admixtures are protected by the manufacturer's patent. The percentage contents of the AEA and SP were calculated in relation to the cement weight. The amounts of AEA and SP were determined experimentally in order to obtain a constant air content (Vp) in the air-entrained mixes at the level of 5.5 ± 0.5% and the consistency class of S2 for all the concrete mixes (slump 50-90 mm) [35].

Mix Proportion and Its Preparation
Twelve concrete mixes were made in order to unequivocally determine the relationship between the AVP and the resistance to scaling in the concretes liquefied with SP based on MP. Six air-entrained concretes (ACs), labeled AC1-AC6, and six non-air-entrained concretes (NACs), labeled NAC1-NAC6, were performed. In order to minimize the influence of factors shaping the AVS, the following constant parameters were adopted: the paste content of 30%; type, grain size, and aggregate content (sand P0/2 = 717 kg/m 3 and basalt grit G4/8 = 459 kg/m 3 and G8/16 = 688 kg/m 3 ); and cement type CEM I 42.5R, as well as the consistency of the concrete mixes (slump 50 ± 10 mm) and air content in the concrete mixes (Vp). In the research program, the composition of the paste was variable in terms of the w/c, content of cement, water, SP, and AEA. To isolate the variability of the scaling resistance, the w/c varying in the range of 0.53 ÷ 0.30 was used in the pavement concretes (Table 3). NAC1 non-air-entrained control concrete fulfilled the requirements for the reference concrete III according to EN 480-1 [37] and, similarly to the air-entrained concrete of the same AC1 formulation, did not contain SP. The remaining concrete mixes, both the air-entrained AC2-AC6 and the non-air-entrained NAC2-NAC6, were liquefied with SP based on MP. The basic recipes of the mixes with the same numbers 1 to 6 differed only as to the content of the air-entraining admixture (AEA). One AEA type, the base of which is a combination of natural resins and synthetic tensides, was used in all the air-  Table 2 lists the properties of the air-entraining admixture (AEA) and the superplasticizer (SP), as specified by the manufacturers. The chemical compositions of the admixtures are protected by the manufacturer's patent. The percentage contents of the AEA and SP were calculated in relation to the cement weight. The amounts of AEA and SP were determined experimentally in order to obtain a constant air content (Vp) in the air-entrained mixes at the level of 5.5 ± 0.5% and the consistency class of S2 for all the concrete mixes (slump 50-90 mm) [35].

Mix Proportion and Its Preparation
Twelve concrete mixes were made in order to unequivocally determine the relationship between the AVP and the resistance to scaling in the concretes liquefied with SP based on MP. Six air-entrained concretes (ACs), labeled AC1-AC6, and six non-air-entrained concretes (NACs), labeled NAC1-NAC6, were performed. In order to minimize the influence of factors shaping the AVS, the following constant parameters were adopted: the paste content of 30%; type, grain size, and aggregate content (sand P0/2 = 717 kg/m 3 and basalt grit G4/8 = 459 kg/m 3 and G8/16 = 688 kg/m 3 ); and cement type CEM I 42.5R, as well as the consistency of the concrete mixes (slump 50 ± 10 mm) and air content in the concrete mixes (Vp). In the research program, the composition of the paste was variable in terms of the w/c, content of cement, water, SP, and AEA. To isolate the variability of the scaling resistance, the w/c varying in the range of 0.53 ÷ 0.30 was used in the pavement concretes (Table 3). NAC1 non-air-entrained control concrete fulfilled the requirements for the reference concrete III according to EN 480-1 [37] and, similarly to the air-entrained concrete of the same AC1 formulation, did not contain SP. The remaining concrete mixes, both the air-entrained AC2-AC6 and the non-air-entrained NAC2-NAC6, were liquefied with SP based on MP. The basic recipes of the mixes with the same numbers 1 to 6 differed only as to the content of the air-entraining admixture (AEA). One AEA type, the base of which is a combination of natural resins and synthetic tensides, was used in all the air-entrained mixtures, AC1-AC6. The amount of AEA was regulated to obtain Vp at the level of 5.5 ± 0.5%. The mixes were made in a horizontal plan mixer with a volume of 100 dm 3 . The dry ingredients, the aggregates and the cement, were mixed for 0.5 min; then, part of the mixing water was added and mixed for another 0.5 min. The SP was added successively, along with 2 dm 3 of water and mixed for 1 minute. In the air-entrained mixes, as the last step, AEA was added with 2 dm 3 of water and then mixed for 2 min. The total mixing time for the NAC mixes was 2 min, and for the AC mixes, it was 4 min. The test samples were demolded after 24 h and stored in a chamber at 20 ± 2 • C and a humidity of 95 ± 5%, in accordance with EN 12390-3 [38].

Concrete Tests
The test scheme of the concrete mixes and hardened concretes is presented in Figure 2.

Concrete Mix Tests
The concrete mixes were tested within 10 minutes from the mixing of the components. The consistency, by the slump-test method, in accordance with EN 12350-2 [ Table 4.
Air Void Parameters (AVP) The AVPs were determined in accordance with EN 480-11 [43] for each series of hardened concretes in two samples with dimensions of 150 mm × 150 mm × 20 mm, using the automatic RapidAir 457 system. As a result of the analysis, our basic AVPs in hardened concrete were determined: A-total content of air, α-specific surface, L-spacing factor, A 300 -micro air content.

Results
The results of the concrete mixtures and the compressive strength (f c ) and density (D) of the hardened concretes after 28 days of curing are presented in Table 5.

The Results of Concrete Mix Tests
The slumps of all the mixtures were in the range of 50-90 mm and according to EN 206 [35], and they can be classified as consistency class S2 according to the adopted assumption ( Figure 3).  The densities (D) of the NAC mixtures ranged from 2380 to 2443 kg/m 3 , while the densities of the AC mixtures ranged from 2283 to 2352 kg/m 3 . The air contents (Vp) ranged from 0.8 to 3.0% in the NAC mixtures and from 5.0 to 6.0% in the AC mixtures. The Vp in the AC mixtures met the assumed criterion of 5.5 ± 0.5%. The obtained values confirm the adopted assumptions for the dosing of the SP and AEA admixtures, which enable the assumed consistency class and air content in the AC to be achieved.

The Results of Hardened Concretes Tests
The compressive strengths ranged from 48.0 to 82.8 Mpa in the AC and from 31.  The densities (D) of the NAC mixtures ranged from 2380 to 2443 kg/m 3 , while the densities of the AC mixtures ranged from 2283 to 2352 kg/m 3 . The air contents (Vp) ranged from 0.8 to 3.0% in the NAC mixtures and from 5.0 to 6.0% in the AC mixtures. The Vp in the AC mixtures met the assumed criterion of 5.5 ± 0.5%. The obtained values confirm the adopted assumptions for the dosing of the SP and AEA admixtures, which enable the assumed consistency class and air content in the AC to be achieved.

The Results of Hardened Concretes Tests
The compressive strengths ranged from 48.0 to 82.8 Mpa in the AC and from 31.1 to 59.7 Mpa in the NAC. The compressive strength of both the NAC and the AC concretes increases with the reduction of the w/c value (Figure 4). In the NAC concretes, it is a linear  Table 6. The mass losses due to scaling decrease as the value of the w/c ratio decreases in both the AC and the NAC. The AC showed a much higher resistance to scaling than the NAC. NAC1-NAC4 are of unacceptable quality, while NAC5-NAC6 are of acceptable quality. AC1 is unacceptable; AC2 and AC3 are acceptable; and AC4 is good, while AC5 and AC6 are of very good quality.   Table 6. The mass losses due to scaling decrease as the value of the w/c ratio decreases in both the AC and the NAC. The AC showed a much higher resistance to scaling than the NAC. NAC1-NAC4 are of unacceptable quality, while NAC5-NAC6 are of acceptable quality. AC1 is unacceptable; AC2 and AC3 are acceptable; and AC4 is good, while AC5 and AC6 are of very good quality.

The Air Void Parameters
The AVP parameters, A, L, α, and A 300 , characterizing the air void system in hardened concretes, are presented in Table 7.

The Influence of w/c on the Scaling Resistance
Mass losses due to scaling ( Figure 5) appear after just 7 cycles in the non-air-entrained concretes, with w/c ≤ 0.5 (NAC1; NAC2). However, they do not exceed 0.1 kg/m 2 , and thus, the concretes are of very good quality [45]. After 14 cycles, the NAC1 concrete (w/c = 0.53) is of acceptable quality, and the NAC2 (w/c = 0.5) is of good quality. After 21 cycles, the non-air-entrained concretes with w/c ≤ 0.5 (NAC1; NAC2) are of unacceptable quality, and the NAC3 concrete is of acceptable quality. Between the 21st and the 28th cycle, the NAC1-3 concrete samples (w/c ≤ 0.45) disintegrate, i.e., their volume is destroyed [2,3,5,6]. After 28 cycles, the first flaking appears in the NAC4 concrete (w/c = 0.4). Between the 28th and the 35th cycle, the remaining non-air-entrained NAC4-6 concretes and the control air-entrained concrete AC1 (w/c = 0.53) are scaled off. Between the 35th and the 42nd cycle, the AC1 concrete is destroyed, further scaling occurs in the NAC4-6 concretes, and the first flaking appears in the AC2 and AC3 concretes. Between cycle 42 and 49, the NAC4 concrete becomes of unacceptable quality, while the NAC5-6 and AC2-3 concretes are of acceptable quality and remain in these categories until cycle 56. Only concretes AC4-6 (w/c ≤ 0.4) do not show any signs of flaking after 56 cycles and remain of very good quality.
The dependences of mass loss and w/c after 56 scaling cycles for both the AC and the NAC presented in Figure 6 indicate evident decreases in mass losses along with the reduction of the w/c value. The variability of the w/c ratio determined in the test program made it possible to obtain AC concretes with variable resistance to scaling from unacceptable to very good quality, which makes it possible to assess the influence of SP on the resistance to scaling. This type of analysis was not possible in [33] due to the obtained very good quality of all the air-entrained concretes. Of course, we are dealing here with a simultaneous interaction of three changing parameters, i.e., the w/c ratio, the cement content, and the increasing SP content as the w/c decreases. The limit value in the NAC of acceptable quality is w/c = 0.35, while in the AEC with an air content of 5.5 ± 0.5% it is w/c = 0.5. The obtained results confirm the literature data [1,5,7,8,46] with regard to the w/c limit values enabling the obtaining of concretes resistant to scaling; in addition, they confirm the same MP with which the new generation of concrete mixes is liquefied; although they change the air void structure [36], they do not affect the resistance to scaling. Thus, the problem of the decrease in resistance to scaling in the SCC concretes [28], as a result of additional air entrainment when liquefied with modified polycarboxylates, does not apply to pavement concretes, in which concrete mixes with consistency classes S2-S4 are standardly used [35]. The dependences of mass loss and w/c after 56 scaling cycles for both the AC and the NAC presented in Figure 6 indicate evident decreases in mass losses along with the reduction of the w/c value. The variability of the w/c ratio determined in the test program made it possible to obtain AC concretes with variable resistance to scaling from unacceptable to very good quality, which makes it possible to assess the influence of SP on the resistance to scaling. This type of analysis was not possible in [33] due to the obtained very good quality of all the air-entrained concretes. Of course, we are dealing here with a simultaneous interaction of three changing parameters, i.e., the w/c ratio, the cement content, and the increasing SP content as the w/c decreases. The limit value in the NAC of acceptable quality is w/c = 0.35, while in the AEC with an air content of 5.5 ± 0.5% it is w/c = 0.5. The obtained results confirm the literature data [1,5,7,8,46] with regard to the w/c limit values enabling the obtaining of concretes resistant to scaling; in addition, they confirm the same MP with which the new generation of concrete mixes is liquefied; although they change the air void structure [36], they do not affect the resistance to scaling. Thus, the problem of the decrease in resistance to scaling in the SCC concretes [28], as a result of additional air entrainment when liquefied with modified polycarboxylates, does not apply to pavement concretes, in which concrete mixes with consistency classes S2-S4 are standardly used [35]. The dependences of the mass losses after 56 scaling cycles and the compressive strength of the concrete after 28 days in the NAC, as presented in Figure 7, show an evident increase in resistance to scaling with increasing strength. However, in the AC, due to the dominant influence of air entrainment on both the scaling resistance and the decrease in compressive strength resulting from it, this relation has not been confirmed in the tests. The dependences of the mass losses after 56 scaling cycles and the compressive strength of the concrete after 28 days in the NAC, as presented in Figure 7, show an evident increase in resistance to scaling with increasing strength. However, in the AC, due to the dominant influence of air entrainment on both the scaling resistance and the decrease in compressive strength resulting from it, this relation has not been confirmed in the tests. The dependences of the mass losses after 56 scaling cycles and the compressive strength of the concrete after 28 days in the NAC, as presented in Figure 7, show an evident increase in resistance to scaling with increasing strength. However, in the AC, due to the dominant influence of air entrainment on both the scaling resistance and the decrease in compressive strength resulting from it, this relation has not been confirmed in the tests.            The total air contents in the NAC both at the stage of concrete mixes, Vp- Table 5, and the hardened concretes, A- Table 7 and Figure 8, increased from 0.8 to 2.4%, as compared to the control concrete NAC1, which was the only one that was not liquefied with SP. Therefore, Ref. [33] the side effect in the form of additional air entrainment after adding SP based on modified polycarboxylates is confirmed. A synergistic air-entraining effect of AEA and SP is observed in the AC. As shown in Figure 8 presenting the A and w/c relation, the resistance to scaling in the AC at a constant air-entrainment level significantly depends on the w/c ratio. For concretes with w/c = 0.53, the air entrainment at the level of 7.4% is not sufficient to ensure resistance to 56 scaling cycles. It should be noted that in order to obtain a constant consistency, in both the NAC and the AC concretes, the SP dose was increased, which weakened the top layer of concrete and, consequently, reduced the resistance to scaling [33].  The total air contents in the NAC both at the stage of concrete mixes, Vp- Table 5, and the hardened concretes, A- Table 7 and Figure 8, increased from 0.8 to 2.4%, as compared to the control concrete NAC1, which was the only one that was not liquefied with SP. Therefore, Ref. [33] the side effect in the form of additional air entrainment after adding SP based on modified polycarboxylates is confirmed. A synergistic air-entraining effect of AEA and SP is observed in the AC. As shown in Figure 8 presenting the A and w/c relation, the resistance to scaling in the AC at a constant air-entrainment level significantly depends on the w/c ratio. For concretes with w/c = 0.53, the air entrainment at the level of 7.4% is not sufficient to ensure resistance to 56 scaling cycles. It should be noted that in order to obtain a constant consistency, in both the NAC and the AC concretes, the SP dose was increased, which weakened the top layer of concrete and, consequently, reduced the resistance to scaling [33].

Assessment of Resistance to Scaling in the Light of AVP
The specific surfaces of all the NACs are below the recommended value of 24 mm −1 [7] for the frost resistance, and the AC is above the recommended value. In both the AC and the NAC, the specific surface decreases with a decrease in w/c (Figure 9), despite the fact that in the NAC the air content increases with a decrease in w/c. Additionally, according to the literature data [7,[14][15][16], the frost resistance should increase, at a constant air content, with a relatively finer structure of air entrainment in the concrete. It should be emphasized here that the decrease in the specific surface area of the pores with the decrease in w/c results from the increasing content of SP based on MP, which causes a disturbance in the AVS. At a constant air content, the content of the preferred pores (entrained air) decreases, while the that of the unfavorable pores (entrapped air) increases [36]. However, it has no effect on the resistance to scaling, for which, as shown by the test results, the strength of the top layer is more important and is mainly determined by the value of the w/c ratio.
All the NAC1-6 concretes have L > 0.292 mm, while all the AC1-6 concretes have L ≤ 0.200 mm ( Figure 10). That is, all the ACs at the level of 5.5 ÷ 0.5% meet the commonly accepted criterion for frost-resistant concretes, and yet, their resistance to scaling [7,9], with similar values of L = (0.138 ÷ 0.165) in the AC2-6 concretes, increases with the lowering of the w/c value. Attention should be paid to concrete AC1, which, despite obtaining the value of L = 0.101 mm, showed unacceptable quality after 56 scaling cycles.
All the AC1-6 air-entrained concretes had micropore contents higher than recommended for frost resistance [7], exceeding 1.5%. In the NAC, the values did not exceed 1.5%. However, in the case of the AC, trends similar to those for the parameters L and α can be observed ( Figure 11). Despite the fulfilment of the criterion, AC1 concrete with w/c = 0.53 is of unacceptable quality. Moreover, with a comparable value of A 300 = (2.35 ÷ 3.05), the resistance to scaling increases with the reduction of the w/c value. The obtained AVP values clearly indicate the influence of the modified polycarboxylates used to liquefy concrete mixes on the air void system. Nevertheless, by disturbing the air void system and deteriorating its parameters, they do not affect the resistance to scaling. Thus, in the light of the obtained results, it is necessary to look at the evaluation of the resistance to scaling in a new way, taking into consideration both the AVP and the w/c ratio.
The AVP obtained for AC1 concrete with w/c = 0.53, L = 0.101 mm, and A 300 = 4.28%, despite meeting the commonly used criterion for frost resistance assessment, L ≤ 0.200 mm and A 300 ≥ 1.5% [6,7,9], is concrete of unacceptable quality in the light of the requirements of SS 137244 [45]. Therefore, the above requirement, frequently [28,29] used spontaneously to assess frost resistance in the light of the obtained results, indicates that it is not spontaneously sufficient to ensure concrete scaling resistance. It should be noted that the AC1 concrete was not liquefied with SP; so, we are not dealing here with the side effect of air-entraining MP. The obtained test results confirm the inability to determine the unambiguous relationship between AVP and the resistance to scaling. In air-entrained concrete, it is necessary to supplement these requirements of L ≤ 0.200 mm and A 300 ≥ 1.5% with the minimum value of w/c ≤ 0.5. In addition, it should be emphasized that the requirement applies only to concretes made with Portland cement without additives and admixtures other than AEA and SP based on modified polycarboxylates.

Conclusions
The conducted research program, the aim of which was to determine the relationship between AVP and the resistance to scaling of concretes liquefied to a constant consistency with the use of MP in a wide range of variability of the w/c ratio, allowed for the formulation of the following conclusions:

•
The w/c variability in the range of 0.53 ÷ 0.30 determined in the test program made it possible to obtain AC with variable resistance to scaling from unacceptable to very good. • Liquefaction with MP did not affect the w/c limit values, enabling the obtaining of concretes resistant to scaling. They are, respectively, 0.35 in NAC and 0.50 in AC of Vp = 5.5 ± 0.5.

•
Resistance to scaling in concrete modified with polycarboxylates grows with a decrease in w/c. A more significant effect can be observed in NAC, because air entrainment is an equally important factor in AC.

•
MPs changed the AVS of AC; however, they did not affect their resistance to scaling. As a result of an increase in the SP content, at a constant air content, with a decrease in w/c in both the NAC and the AC, the entrained-air content decreases against the unfavorable entrapped air.

•
The commonly used criterion for ensuring the frost resistance of AC in the range of AVP, L ≤ 0.200 mm and A 300 ≥ 1.5%, requires supplementing with the minimum value of the w/c ≤ 0.50.