Inﬂuence of Slag Cement on the Permeability of Concrete for Biological Shielding Structures

: Durability of concrete designed for radiation shielding structures is an important issue in nuclear power plant safety. An investigation of the permeability of concrete containing heavyweight aggregates and water-bearing aggregates was performed with respect to gaseous and liquid media. Mix design was developed using Portland and slag cement, crushed magnetite and serpentine aggregate. The use of slag cement in concrete containing magnetite and serpentine aggregates resulted in the substantial improvement of the compressive strength in comparison with Portland cement concrete. The application of slag cement was found to reduce the chloride ingress, regardless of the special aggregate use. The coe ﬃ cient of chloride migration was within the range 5 ÷ 8 × 10 − 12 m 2 / s and 17 ÷ 25 × 10 − 12 m 2 / s for slag cement concrete and Portland cement concrete, respectively. At the same time, the carbonation depth was increased twice for slag cement concrete in comparison to Portland cement concrete. However, the maximum carbonation depth after one year of exposure to 1% CO 2 was only 14 mm for slag cement concrete, and 7 mm for reference concrete. The total pore volume evaluated using mercury intrusion porosimetry was inﬂuenced by the type of special aggregate used. It was shown that concrete with various contents of magnetite aggregate and slag cement achieved the smallest total pore volume. While serpentine coarse aggregate caused an increase in total pore volume in comparison to concrete with magnetite aggregate.


Introduction
Nuclear power is a sustainable and carbon-free source of electricity. Increased interest to build new nuclear power plants has been observed due to growing environmental concerns related to fossil fuel energy sources. Additionally, the total cost of using nuclear energy in energy and electricity plans and policies is significantly lower than that of other energy sources, including renewable energy sources [1]. However, after the accident of the Fukushima Daiichi nuclear power plant in 2011, the public opinion trends demonstrated a lower degree of trust in nuclear energy technologies and facilities. In response, significant enhancements of key plant safety functions, in particular related to large-scale natural disasters, were introduced [2]. Nuclear safety-related concrete structures are constantly required to provide both structural and radiation shielding functions. For economic reasons the expected service life of principal concrete structures should be at least 60 years. Therefore, the problem of adequate long-term durability of concrete exposed to harsh environmental conditions becomes significant [3][4][5].
In [6], the environmental conditions were reviewed both for nuclear containment structures and biological shielding structures. The durability design is largely based on a suitable selection of concrete constituents and mix proportioning that is guided by the combined action of radiation, Binici [30] investigated the heavyweight concrete durability, including among others the resistance to chloride ions. He found that the resistance to the chloride penetration was significantly higher for concrete incorporating barite, both coarse and fine aggregate in comparison to reference concrete with limestone aggregate. The relationship between temperature and carbonation of concrete was found by Zhang et al. [31]. They revealed that the higher the temperature was, the larger the carbonation depth was noticed. At temperatures less than 300 • C, the rate of increase in the carbonation depth was relatively low, but above 450 • C, the carbonation depth increased sharply. At a temperature higher than 600 • C, the carbonation depth of ordinary concrete determined after 28 days increased more than three times (from 7 mm to 23 mm).
The aim of this investigation is to evaluate the impermeability of concrete containing special aggregates with respect to gaseous and liquid media. On the basis of characterization of concrete microstructure, its mechanical and physical properties and its suitability for application in structural elements of nuclear power plants are evaluated.

Materials and Methods
Blast furnace slag cement (CEM III/A 42.5N LH/HSR/NA) was investigated in the research. As a reference, also low heat, sulfate resistant and low alkali Portland cement CEM I 42.5N (OPC) was used. The specific surface was 4700 and 3800 cm 2 /g, respectively. CEM III was characterized by density 2.99 g/cm 3 and CEM I by 3.15 g/cm 3 . The water demand was 34% for CEM III and 28% for CEM I. Compressive and flexural strength after 28 days was 58.2 and 9.5 MPa and 52.6 and 8.1 MPa for CEM III and CEM I, respectively. The basic cement properties and the detailed chemical and physical properties of cement are given in [32].
The testing program covered concrete mixes with variable magnetite and serpentinite coarse aggregate. Amphibolite coarse aggregate was used as a reference in the concrete mix. A detailed composition of concrete mixes is presented in Table 1, and the properties of the fresh mix are presented in Table 2. As fine aggregate, in all mixes, a siliceous sand of 0-2 mm was used. The content of sand was 20%, except its content in reference concrete (30%). The magnetite to serpentine aggregate ratio was 1:2 (S-MS_2/1 and P-MS_2/1) or 2:1 (S-MS_1/2 and P-MS_1/2). The cement content was constant and amounted to 350 kg/m 3 and water to cement ratio 0.48. Water reducer based on modified phosphonate and high-range water reducer based on polycarboxylate and modified phosphonate were used.
Specimens 150 × 150 × 150 mm were used to test compressive strength and cylinders ø = 100 and h = 200 mm for chloride ion migration and as well as for porosity accessible for water. Specimens 100 × 100 × 100 mm were cast for CO 2 ingress and microstructure analysis. The specimens were conditioned in a climatic chamber at temperature of 20 ± 2 • C and relative humidity of 95% up to 28 days and next in RH = 55 ± 5% (room conditions) up to day of measurement.
The compressive strength was determined on three 150 × 150 × 150 mm specimens for each concrete mix after 7, 28 and 90 days of curing following the standard procedure PN-EN 12390-3:2019 [33].
The chloride ions penetration was tested according to Nordtest Method NT Build 492 [34] using the rapid chloride ion migration test. The value of non-steady-state chloride migration coefficient, D nssm , was determined for a non-steady flux of chloride ions under the action of the external electrical field after 56 and 112 days of curing the concrete specimens and calculated from the Fick's second law. A description of the applied test method is presented in [3]. Each result concerning compressive strength and well as the chloride migration coefficient is an average of three measurements.
The carbonation of concrete was investigated at a constant concentration of carbon dioxide (1%), at 22 • C and 60% relative humidity. The depth of carbonation was tested on one beam from each concrete mix on the freshly split surface with phenolphthalein solution after 0, 28, 56, 90, 180 and 360 days of exposure, according to PN-EN 13295:2005 [35]. The carbonation depth was determined on the average of 20 measurements tested after a prescribed exposure time.  Analysis of the concrete microstructure was performed on thin sections cut from concrete prisms. For this purpose, the fluorescent epoxy impregnated thin sections were prepared according to [24]. The concrete specimens were cut in smaller pieces (40 × 50 mm), which were then vacuum impregnated with a low viscous resin together with yellow fluorescent dye. The thin section subjected to further microscopic analysis was characterized by thicknesses of 20 ± 2 µm. Thin section analysis was performed using the optical polarizing microscope Olympus BX51 with a digital camera. The polarized light (PPL), crossed polarized light (XPL), and also with lambda plate and UV light were applied for thin sections analysis.
A mercury intrusion porosity (MIP) test was also executed to evaluate the differences in pore content and pore size distribution. MIP measurements were carried out using cores with diameter 9 mm and length 25 mm drilled from concrete specimens. Drilled specimens were crushed and then a cement mortar was separated to further measurements. All investigated specimens were characterized by the same volume and were smaller than 2 cm 3 , limited by the volume of the measurement container. The analyzed specimens were dried until a constant weight at 35 • C and then they were kept in tightly closed containers to avoid a humidity exchange with the environment. Quantachrome POREMASTER 60 mercury porosimeter, which was able to detect the pores up to 5 nm with the maximum pressure of 414 MPa, was used. The accuracy of measurement was 0.001 PSI, which corresponds to 5 nm resolution. The volume of intruded mercury was measured with accuracy 0.0001 cm 3 /g. A cement matrix without aggregate was separated from three cores from each concrete after 120 days of water curing.
Two cores from each concrete mix (100 ± 2 mm diameters and 50 ± 2 mm height) were used to assess the open porosity according to NF P18-459: 2010 [36]. They were vacuumed, saturated with tap water and weighed in three water saturation states: fully saturated and weighted with hydrostatic balance, fully saturated and weighted with an analytical balance, and dried to constant mass at 105 • C. Based on the obtained results, the porosity available for water was determined.

Compressive Strength
The results of the 7, 28 and 90 days of compressive strength for concrete made with various coarse aggregates and both CEM I and CEM III are presented in Figure 1.
Energies 2020, 13, x FOR PEER REVIEW 5 of 16 balance, fully saturated and weighted with an analytical balance, and dried to constant mass at 105 °C. Based on the obtained results, the porosity available for water was determined.

Compressive Strength
The results of the 7, 28 and 90 days of compressive strength for concrete made with various coarse aggregates and both CEM I and CEM III are presented in Figure 1. The highest value of compressive strength tested after 7, 28 and 90 days for concrete made with CEM I achieved reference concrete made with amphibolite aggregate, 51, 71 and 85 MPa, respectively. In concretes containing CEM I and special aggregate, the highest compressive strength, regardless of curing time, was achieved for concrete with magnetite aggregate: 45, 62 and 80 MPa after 7, 28 and 90 days. Concretes made with serpentine aggregate were characterized by lower values of compressive strength, although very close to the previous one. The change in magnetite and serpentine content (P-MS_1/2 and P-MS_2/1) did not affect the compressive strength values significantly. All concretes made with CEM I belonged to the strength class C55/67 (after 90 days).
In concrete containing CEM III cement, the beneficial influence of a special aggregate type on compressive strength was visible. The highest values of compressive strength were achieved by concretes containing a mixture of magnetite and serpentine aggregate, S-MS_1/2 and S-MS_2/1. After 90 days of curing, they were characterized by 85 and 91 MPa. A similar value was achieved by reference concrete, 87 MPa, which placed them in the compressive strength class C70/85. Slightly lower values of compressive strength were achieved for concretes made with pure magnetite aggregate, S-M-81 MPa (the compressive strength class C60/75).
In all analyzed series, the influence of CEM III in comparison to CEM I on the chloride ions migration coefficient was clearly visible. A significant increase in the resistance of concrete to the penetration of chloride ions was observed, after 56 days as well as after 120 days of curing. The use of slag cement was found to reduce the chloride migration coefficient after 120 days for even two categories of chloride ions migration resistance.
The reference concrete contained slag cement achieved the lowest values of Dnssm, 3.8 × 10 −12 m 2 /s after 56 days and 1.7 × 10 −12 m 2 /s after 120 days which classifies it into good and very good classes of The highest value of compressive strength tested after 7, 28 and 90 days for concrete made with CEM I achieved reference concrete made with amphibolite aggregate, 51, 71 and 85 MPa, respectively. In concretes containing CEM I and special aggregate, the highest compressive strength, regardless of curing time, was achieved for concrete with magnetite aggregate: 45, 62 and 80 MPa after 7, 28 and 90 days. Concretes made with serpentine aggregate were characterized by lower values of compressive strength, although very close to the previous one. The change in magnetite and serpentine content (P-MS_1/2 and P-MS_2/1) did not affect the compressive strength values significantly. All concretes made with CEM I belonged to the strength class C55/67 (after 90 days).
In concrete containing CEM III cement, the beneficial influence of a special aggregate type on compressive strength was visible. The highest values of compressive strength were achieved by concretes containing a mixture of magnetite and serpentine aggregate, S-MS_1/2 and S-MS_2/1. After 90 days of curing, they were characterized by 85 and 91 MPa. A similar value was achieved by reference concrete, 87 MPa, which placed them in the compressive strength class C70/85. Slightly lower values of compressive strength were achieved for concretes made with pure magnetite aggregate, S-M-81 MPa (the compressive strength class C60/75).
Energies 2020, 13, 4582 6 of 16 were characterized by high Dnssm coefficient, from 21.3 to 31.2 × 10 −12 m 2 /s, and after 120 days from 17.0 to 25.1 × 10 −12 m 2 /s. All the above concretes were assigned to class unacceptable, however among them, concrete with magnetite aggregate showed the best resistance to chloride ions penetration, and concrete with serpentine aggregate the worst. The chloride ion migration coefficient increased with increasing the content of serpentine aggregate. Reference concrete with amphibolite aggregate both after 56 and 120 days achieved acceptable class, respectively 13.5 and 9.1 × 10 −12 m 2 /s.

Depth of Carbonation
The results of the carbonation depth of concrete measured until 360 days are presented in Figure  3. As it was expected, the depth of carbonation increased, along with the age of exposure. None of the tested concretes except reference one with Portland cement and amphibolite aggregate (P-A) did not show a tendency to achieving the maximum depth of carbonation. The effect of cement type on the carbonation of concrete was clearly visible. The lowest value of the carbonation depth was achieved by concrete with Portland cement, and the highest results for carbonation depth were observed in concretes containing the slag cement. Among the concretes with slag cement and special aggregate, the highest depth of carbonation was achieved by concrete S-M with magnetite aggregate, 13.5 mm after one year of exposure in 1% CO2. Similar but slightly lower results were achieved by concrete with magnetite + serpentine S-MS_1/2-13.4 mm. In concrete with slag cement, the increase of serpentine aggregate caused an increase in carbonation depth, while in concrete with Portland cement the trend was opposite, the more serpentine aggregate the smaller the carbonation depth.
It is also worth noting that the progress of carbonation in concretes with slag cement and special aggregates occurred much faster than in the case of concrete with Portland cement. After 28 days of exposure in carbon dioxide the values of the carbonation depth were three times higher for concrete made with magnetite aggregate and more than twenty times higher for concrete made with serpentine aggregate compared to those with Portland cement. In all analyzed series, the influence of CEM III in comparison to CEM I on the chloride ions migration coefficient was clearly visible. A significant increase in the resistance of concrete to the penetration of chloride ions was observed, after 56 days as well as after 120 days of curing. The use of slag cement was found to reduce the chloride migration coefficient after 120 days for even two categories of chloride ions migration resistance.
The reference concrete contained slag cement achieved the lowest values of D nssm , 3.8 × 10 −12 m 2 /s after 56 days and 1.7 × 10 −12 m 2 /s after 120 days which classifies it into good and very good classes of the concrete resistance to chloride ions penetration. Concrete S-MS_2/1 achieved the lowest value of D nssm in the group of concretes with slag cement and special aggregate. The nonsteady chloride migration coefficient was relatively stable and achieved 5.8 × 10 −12 m 2 /s after 56 days and 5.4 × 10 −12 m 2 /s after 120 days (good class). After 120 days of curing, concretes with magnetite (S-M) and magnetite/serpentine with a ratio of 2:1 (S-MS_1/2) were characterized by similar values of D nssm and revealed good resistance to chloride ingress, respectively 7.0 × 10 −12 m 2 /s and 7.7 × 10 −12 m 2 /s.
In concretes containing Portland cement, differences between 56 and 120 days of curing are relevant for the chloride migration coefficient. After 56 days, all concrete containing special aggregate were characterized by high D nssm coefficient, from 21.3 to 31.2 × 10 −12 m 2 /s, and after 120 days from 17.0 to 25.1 × 10 −12 m 2 /s. All the above concretes were assigned to class unacceptable, however among them, concrete with magnetite aggregate showed the best resistance to chloride ions penetration, and concrete with serpentine aggregate the worst. The chloride ion migration coefficient increased with increasing the content of serpentine aggregate. Reference concrete with amphibolite aggregate both after 56 and 120 days achieved acceptable class, respectively 13.5 and 9.1 × 10 −12 m 2 /s.

Depth of Carbonation
The results of the carbonation depth of concrete measured until 360 days are presented in Figure 3. As it was expected, the depth of carbonation increased, along with the age of exposure. None of the tested concretes except reference one with Portland cement and amphibolite aggregate (P-A) did not show a tendency to achieving the maximum depth of carbonation. The effect of cement type on the carbonation of concrete was clearly visible. The lowest value of the carbonation depth was achieved by concrete with Portland cement, and the highest results for carbonation depth were observed in concretes containing the slag cement. Among the concretes with slag cement and special aggregate, the highest depth of carbonation was achieved by concrete S-M with magnetite aggregate, 13.5 mm after one year of exposure in 1% CO 2 . Similar but slightly lower results were achieved by concrete with magnetite + serpentine S-MS_1/2-13.4 mm. In concrete with slag cement, the increase of serpentine

Concrete Microstructure
All concretes were analyzed on thin sections in transmitted light to evaluate their microstructure. Digital images of thin sections (25 × 45 mm) consisting of aggregate grains and surrounding cement matrix with air voids filled by resin are presented in Figure 4. The separate images (1.5 × 2 mm) were automatically acquired and assembled into one image, which was then subjected to further analysis.
Uniform distribution of the coarse aggregate was visible in all concretes containing slag cement and special aggregates. Concrete made with magnetite aggregate only was characterized by a higher content of smaller air voids (max. 0.5 mm) than other concretes. Figure 4 shows the dependence, the more serpentine aggregate and the less magnetite aggregate in concrete with slag cement, the lower content of smaller air voids and the higher content of the individual, larger entrapped air. It is also worth noting that the progress of carbonation in concretes with slag cement and special aggregates occurred much faster than in the case of concrete with Portland cement. After 28 days of exposure in carbon dioxide the values of the carbonation depth were three times higher for concrete made with magnetite aggregate and more than twenty times higher for concrete made with serpentine aggregate compared to those with Portland cement.

Concrete Microstructure
All concretes were analyzed on thin sections in transmitted light to evaluate their microstructure. Digital images of thin sections (25 × 45 mm) consisting of aggregate grains and surrounding cement matrix with air voids filled by resin are presented in Figure 4. The separate images (1.5 × 2 mm) were automatically acquired and assembled into one image, which was then subjected to further analysis.

Concrete Microstructure
All concretes were analyzed on thin sections in transmitted light to evaluate their microstructure. Digital images of thin sections (25 × 45 mm) consisting of aggregate grains and surrounding cement matrix with air voids filled by resin are presented in Figure 4. The separate images (1.5 × 2 mm) were automatically acquired and assembled into one image, which was then subjected to further analysis.
Uniform distribution of the coarse aggregate was visible in all concretes containing slag cement and special aggregates. Concrete made with magnetite aggregate only was characterized by a higher content of smaller air voids (max. 0.5 mm) than other concretes. Figure 4 shows the dependence, the more serpentine aggregate and the less magnetite aggregate in concrete with slag cement, the lower content of smaller air voids and the higher content of the individual, larger entrapped air.

Slag cement
Portland cement Uniform distribution of the coarse aggregate was visible in all concretes containing slag cement and special aggregates. Concrete made with magnetite aggregate only was characterized by a higher content of smaller air voids (max. 0.5 mm) than other concretes. Figure 4 shows the dependence, the more serpentine aggregate and the less magnetite aggregate in concrete with slag cement, the lower content of smaller air voids and the higher content of the individual, larger entrapped air.
The microscopic analysis of thin sections revealed that the zone around the serpentine aggregate differed from the zones surrounding the other aggregates (Figures 4 and 5). The zones of increased porosity were clearly visible on thin sections observed in plane-polarized light (PPL) as well as in UV light. Detailed microscopic analysis showed that the porous zone around the serpentinite aggregate was higher than in the magnetite or amphibolite aggregate. The cement type slightly affects the zone of discontinuity in concrete with serpentine aggregate. The width of that zone was from about 50 to 80 µm in concrete with Portland cement and from about 50 to 100 µm in concrete with slag cement. The microscopic analysis of thin sections revealed that the zone around the serpentine aggregate differed from the zones surrounding the other aggregates (Figures 4 and 5). The zones of increased porosity were clearly visible on thin sections observed in plane-polarized light (PPL) as well as in UV light. Detailed microscopic analysis showed that the porous zone around the serpentinite aggregate was higher than in the magnetite or amphibolite aggregate. The cement type slightly affects the zone of discontinuity in concrete with serpentine aggregate. The width of that zone was from about 50 to 80 μm in concrete with Portland cement and from about 50 to 100 μm in concrete with slag cement.

Mercury Intrusion Porosity Tests
Results of MIP measurement on cement paste taken from analyzed concretes with two types of cement are presented in Figures 6 and 7. The total volume of pores was similar, within the range 0.023-0.030 cm 3 /g for concrete with slag cement and 0.024-0.047 cm 3 /g for concrete with Portland cement.
The total pore volume for concrete made with slag cement and various magnetite aggregate content was similar, regardless of the amount of this aggregate. It was 0.022 cm 3 /g for concrete with magnetite aggregate (S-M), 0.023 cm 3 /g for magnetite/serpentinite ratio 1:2 (S-MS_1/2) and 0.029 cm 3 /g for magnetite/serpentinite ratio 2:1 (S-MS_2/1). The content of total pore volume for reference concrete (B19) was 0.030 cm 3 /g (Figure 6b).
A much more pronounced effect of the aggregate on the total pore volume content was visible in concretes containing Portland cement. The lowest value of the total pore volume achieved concrete

Mercury Intrusion Porosity Tests
Results of MIP measurement on cement paste taken from analyzed concretes with two types of cement are presented in Figures 6 and 7. The total volume of pores was similar, within the range 0.023-0.030 cm 3 /g for concrete with slag cement and 0.024-0.047 cm 3 /g for concrete with Portland cement.
Regardless of the cement type, the increase in magnetite aggregate content in concrete influenced the decrease in the total pore volume. A decrease in the total pore volume measured by MIP was observed with an increase in the volume of magnetite aggregate.  The total pore volume for concrete made with slag cement and various magnetite aggregate content was similar, regardless of the amount of this aggregate. It was 0.022 cm 3 /g for concrete with magnetite aggregate (S-M), 0.023 cm 3 /g for magnetite/serpentinite ratio 1:2 (S-MS_1/2) and 0.029 cm 3 /g for magnetite/serpentinite ratio 2:1 (S-MS_2/1). The content of total pore volume for reference concrete (B19) was 0.030 cm 3 /g (Figure 6b).
A much more pronounced effect of the aggregate on the total pore volume content was visible in concretes containing Portland cement. The lowest value of the total pore volume achieved concrete with magnetite aggregate (P-M) 0.024 cm 3 /g, lower than reference concrete (P-A) 0.029 cm 3 /g, subsequently concrete with magnetite/serpentine in ratio 1 to 2 (P-MS_1/2) and 2 to 1 (P-MS_2/1), respectively 0.039 cm 3 /g and 0.045 cm 3 /g (Figure 6a).

Porosity Accessible to Water
The results of the porosity accessible to water are presented in Figure 8. The lowest values of porosity were achieved by the reference concrete, with Portland cement 11.0% and with slag cement 11.4%. Use of special aggregates, both magnetite and serpentine increased open porosity for all cement types compared to reference concrete with amphibolite aggregate. The values of the porosity accessible to water were similar for concrete made with Portland cement and special aggregate, about 12.5 ± 0.3%. However, they differed when slag cement was used. In this group of concretes, the lowest value of porosity achieved concrete S-MS_1/2, with a magnetite/serpentine ratio of 1:2-12.8%, and the highest concrete S-MS_2/1 with a magnetite to serpentine ratio of 2:1-14.4%. Pore size distribution analyzed revealed the largest number of pores within the range below 300 nm. It influenced the total pore volume (Figure 7). Concretes containing slag cement (S-M, S-MS_1/2, S-MS_2/1) were characterized by the lowest volume of capillary pores >300 nm (Figure 7b). The volume was in the range of 0.005-0.007 cm 3 /g, which constituted 30-45% less than the volume of capillary pores for reference concrete (S-A). Concretes with Portland cement were characterized by the highest volume of capillary pores (>300 nm) with range 0.007-0.011 cm 3 /g (P-M, P-MS_1/2, P-MS_2/1), which was 20-110% more than reference concrete (P-A) with amphibolite aggregate (Figure 7a).
Regardless of the cement type, the increase in magnetite aggregate content in concrete influenced the decrease in the total pore volume. A decrease in the total pore volume measured by MIP was observed with an increase in the volume of magnetite aggregate.

Porosity Accessible to Water
The results of the porosity accessible to water are presented in Figure 8. The lowest values of porosity were achieved by the reference concrete, with Portland cement 11.0% and with slag cement 11.4%. Use of special aggregates, both magnetite and serpentine increased open porosity for all cement types compared to reference concrete with amphibolite aggregate. The values of the porosity accessible to water were similar for concrete made with Portland cement and special aggregate, about 12.5 ± 0.3%. However, they differed when slag cement was used. In this group of concretes, the lowest value of porosity achieved concrete S-MS_1/2, with a magnetite/serpentine ratio of 1:2-12.8%, and the highest concrete S-MS_2/1 with a magnetite to serpentine ratio of 2:1-14.4%.
Energies 2020, 13, x FOR PEER REVIEW 11 of 16 Figure 8. Porosity accessible to water measured after 120 days of curing.

Discussion
All obtained results regarding the permeability of the analyzed concretes: chloride migration, carbonation, pore size distribution and porosity accessible for water, as well as the results of compressive strength and microstructure are compatible with each other and complementary.
It is known from the literature that hydration of slag cement is slower than ordinary Portland cement. Concrete made with CEM III/A cement is characterized by delayed setting, [37]. The research conducted by Hager et al. [38] revealed that slag cement influenced the compressive strength of 90day concrete. In their research concrete with slag cement was characterized by higher compressive strength on average by 10 MPa for both basalt and river bed aggregate concrete in comparison to Portland cement. Kubissa et al. [27] investigated magnetite aggregate and blended cement with ground granulated blast furnace slag. They showed that the use of CEM II/B-S cement resulted in a significant increase in concrete compressive strength at the age of about 2 years, up to 45%. The influence of slag cement on the long-term compressive strength of the investigated concretes confirmed partly earlier results. The influence of CEM III/A cement in reference concrete with amphibolite aggregate was negligible. The 90 days compressive strength for reference concrete was similar, 85 MPa for Portland cement and 87 MPa for slag cement. After the same maturity period, the values of compressive strength for concrete made with only magnetite, regardless of the cement used, were lower than reference concrete but similar. It was 80.2 and 80.8 MPa for magnetite concrete with Portland and slag cement. However, the effect of the type of cement on the compressive strength of concrete containing a mixture of magnetite and serpentine aggregate was visible. Higher compressive strength determined after 90 days of curing were obtained when using slag cement, 84.6 MPa for S-MS_1/2 and 90.6 MPa for S-MS_2/1. When using Portland cement, concretes containing magnetite and serpentine in proportions of 1:2 and 2:1 were characterized by similar compressive strength, 71.9 and 70.7 MPa.
The influence of special aggregate on concrete compressive strength was investigated by [39,40]. Horszczaruk et al. [39] revealed that the use of magnetite aggregate instead of gravel and river sand improved mechanical properties, compressive (about 10 MPa) and splitting (about 1.2 MPa) strength of concrete made with CEM I and 10% of silica fume. Petrounias et al. [40] analyzed three groups of aggregate: serpentine, diabase-gabbro and albitite. The compressive strength of concrete with CEM II 32.5N cement showed the lowest value for concrete with serpentine aggregate. The conducted research showed that concrete made with magnetite aggregate achieved higher compressive strength

Discussion
All obtained results regarding the permeability of the analyzed concretes: chloride migration, carbonation, pore size distribution and porosity accessible for water, as well as the results of compressive strength and microstructure are compatible with each other and complementary.
It is known from the literature that hydration of slag cement is slower than ordinary Portland cement. Concrete made with CEM III/A cement is characterized by delayed setting, Lura, P., et al. [37]. The research conducted by Hager et al. [38] revealed that slag cement influenced the compressive strength of 90-day concrete. In their research concrete with slag cement was characterized by higher compressive strength on average by 10 MPa for both basalt and river bed aggregate concrete in comparison to Portland cement. Kubissa et al. [27] investigated magnetite aggregate and blended cement with ground granulated blast furnace slag. They showed that the use of CEM II/B-S cement resulted in a significant increase in concrete compressive strength at the age of about 2 years, up to 45%. The influence of slag cement on the long-term compressive strength of the investigated concretes confirmed partly earlier results. The influence of CEM III/A cement in reference concrete with amphibolite aggregate was negligible. The 90 days compressive strength for reference concrete was similar, 85 MPa for Portland cement and 87 MPa for slag cement. After the same maturity period, the values of compressive strength for concrete made with only magnetite, regardless of the cement used, were lower than reference concrete but similar. It was 80.2 and 80.8 MPa for magnetite concrete with Portland and slag cement. However, the effect of the type of cement on the compressive strength of concrete containing a mixture of magnetite and serpentine aggregate was visible. Higher compressive strength determined after 90 days of curing were obtained when using slag cement, 84.6 MPa for S-MS_1/2 and 90.6 MPa for S-MS_2/1. When using Portland cement, concretes containing magnetite and serpentine in proportions of 1:2 and 2:1 were characterized by similar compressive strength, 71.9 and 70.7 MPa.
The influence of special aggregate on concrete compressive strength was investigated by [39,40]. Horszczaruk et al. [39] revealed that the use of magnetite aggregate instead of gravel and river sand improved mechanical properties, compressive (about 10 MPa) and splitting (about 1.2 MPa) strength of concrete made with CEM I and 10% of silica fume. Petrounias et al. [40] analyzed three groups of aggregate: serpentine, diabase-gabbro and albitite. The compressive strength of concrete with CEM II 32.5N cement showed the lowest value for concrete with serpentine aggregate. The conducted research showed that concrete made with magnetite aggregate achieved higher compressive strength after 90 days in comparison to concrete made with serpentine aggregate. The difference was small and amounted to 5 MPa for concrete with Portland cement and 3 MPa for concrete with slag cement. The obtained results of compressive strength with serpentine aggregate significantly exceed the values given in the report, Almenas, K., et al. [41]. In the Lithuanian NPP, at Ignalina, compressive strength of serpentine concrete which was used as the shielding concrete achieved 40 ÷ 62.5 MPa.
Plotting the porosity accessible to water versus the compressive strength of concrete reveals some degree of correlation for two types of cement. The tendency to slightly decrease the porosity available for water with the increase in the compressive strength of concrete with Portland cement and a much stronger relationship for slag cement was demonstrated. A similar relationship between the compressive strength and the water absorption at 28 days of concrete made with CEM I, CEM II/B-S and CEM III/A cements achieved Giergiczny et al. [42].
The relationship of the porosity accessible to water and both, the depth of carbonation and nonsteady chloride migration coefficient is presented in Figures 9 and 10. As in the case of compressive strength, the relationship of the porosity accessible for water and the resistance of concrete to carbonation is definitely more pronounced for concretes with slag cement. Higher CO 2 penetration depth values corresponded to a higher available porosity accessible for water values for concrete containing slag cement. This relationship is confirmed in the literature, Hager, I., et al. [38,43]. Concrete with slag cement was characterized by lower Cembureau permeability [44] for natural aggregate both crushed and uncrushed (gravel) in comparison to Portland cement, Hager, I., et al. [38]. The application of heavyweight and hydrous aggregates in concrete was investigated by Jaskulski et al. [43]. values given in the report, [41]. In the Lithuanian NPP, at Ignalina, compressive strength of serpentine concrete which was used as the shielding concrete achieved 40 ÷ 62.5 MPa.
Plotting the porosity accessible to water versus the compressive strength of concrete reveals some degree of correlation for two types of cement. The tendency to slightly decrease the porosity available for water with the increase in the compressive strength of concrete with Portland cement and a much stronger relationship for slag cement was demonstrated. A similar relationship between the compressive strength and the water absorption at 28 days of concrete made with CEM I, CEM II/B-S and CEM III/A cements achieved Giergiczny et al. [42].
The relationship of the porosity accessible to water and both, the depth of carbonation and nonsteady chloride migration coefficient is presented in Figures 9 and 10. As in the case of compressive strength, the relationship of the porosity accessible for water and the resistance of concrete to carbonation is definitely more pronounced for concretes with slag cement. Higher CO2 penetration depth values corresponded to a higher available porosity accessible for water values for concrete containing slag cement. This relationship is confirmed in the literature, [38,43]. Concrete with slag cement was characterized by lower Cembureau permeability [44] for natural aggregate both crushed and uncrushed (gravel) in comparison to Portland cement, [38]. The application of heavyweight and hydrous aggregates in concrete was investigated by Jaskulski et al. [43]. They revealed that concretes with magnetite and serpentine aggregate achieved an open porosity value from 12.2 to 13.3%. The increase in the open porosity was 1.2-2.3% in comparison with concrete containing amphibolite aggregate.  The total pore volume evaluated using mercury intrusion porosimetry was influenced by the content of magnetite aggregate in concrete containing slag cement. In Figure 11, the relationship between the relative content of magnetite aggregate and the total pore volume is presented. The total pore volume is linearly decreasing with increasing content of magnetite aggregate and a simultaneous Energies 2020, 13, 4582 13 of 16 decrease in the content of serpentine aggregate. The porosity is systematically lower for slag cement concrete. Such results demonstrate the beneficial effects of using slag cement and various content of crushed magnetite aggregate in radiation shielding concrete of low permeability.
Energies 2020, 13, x FOR PEER REVIEW 13 of 16 Figure 10. Relationship between the porosity accessible to water and non-steady-state chloride migration coefficient after 120 days of concrete curing.
The total pore volume evaluated using mercury intrusion porosimetry was influenced by the content of magnetite aggregate in concrete containing slag cement. In Figure 11, the relationship between the relative content of magnetite aggregate and the total pore volume is presented. The total pore volume is linearly decreasing with increasing content of magnetite aggregate and a simultaneous decrease in the content of serpentine aggregate. The porosity is systematically lower for slag cement concrete. Such results demonstrate the beneficial effects of using slag cement and various content of crushed magnetite aggregate in radiation shielding concrete of low permeability. Figure 11. The relationship between the relative content of magnetite aggregate and the total pore volume. Figure 10. Relationship between the porosity accessible to water and non-steady-state chloride migration coefficient after 120 days of concrete curing.
Energies 2020, 13, x FOR PEER REVIEW 13 of 16 Figure 10. Relationship between the porosity accessible to water and non-steady-state chloride migration coefficient after 120 days of concrete curing.
The total pore volume evaluated using mercury intrusion porosimetry was influenced by the content of magnetite aggregate in concrete containing slag cement. In Figure 11, the relationship between the relative content of magnetite aggregate and the total pore volume is presented. The total pore volume is linearly decreasing with increasing content of magnetite aggregate and a simultaneous decrease in the content of serpentine aggregate. The porosity is systematically lower for slag cement concrete. Such results demonstrate the beneficial effects of using slag cement and various content of crushed magnetite aggregate in radiation shielding concrete of low permeability. Figure 11. The relationship between the relative content of magnetite aggregate and the total pore volume. Figure 11. The relationship between the relative content of magnetite aggregate and the total pore volume.

Conclusions
The following conclusions can be drawn:

•
Use of slag cement CEM III/A in concrete containing magnetite and serpentine aggregates resulted in the substantial improvement of the compressive strength in comparison with Portland cement concrete. The attained strength class was C70/85 and C55/67, respectively.

•
The application of slag cement was found to reduce the chloride ingress, regardless of the special aggregate use. In regard to Portland cement concrete, the coefficient of chloride migration D nssm decreased from 17 ÷ 25 × 10 −12 m 2 /s to 5 ÷ 8 × 10 −12 m 2 /s at the age of 120 days.

•
The effect of cement type on the concrete carbonation depth was clearly demonstrated: the lowest carbonation depth (4.5 ÷ 7.0 mm) was achieved for Portland cement concrete and the highest (11.0 ÷ 13.9 mm) for slag cement concrete.

•
The type of cement did not influence the uniformity of coarse aggregate distribution and microstructural homogeneity of hardened cement paste in concrete containing special aggregates.

•
The total pore volume evaluated using mercury intrusion porosimetry was influenced by the type of special aggregate. A linearly decreasing relationship was found between the relative content of magnetite aggregate and the total pore volume in concrete containing slag cement. The smallest total pore volume (0.024 cm 3 /g) was found for concrete with the highest content of magnetite aggregate and both cement types. Serpentine coarse aggregates caused an increase in total pore volume in comparison to concrete with magnetite aggregate.

•
The use of magnetite and serpentine aggregates resulted in an increase in the concrete porosity accessible to water by 1 to 3%, in concrete with Portland and slag cement, respectively.