**1. Introduction**

Cements with granulated blast furnace slag are widely employed due to their lower carbon footprint, as a strategy for sustainable development in the field of construction. The use of ground granulated blast furnace slag (GGBFS), which presents an amorphous structure and shows pozzolan characteristics, in concrete as an additive has a positive effect on the properties of fresh and hardening concrete [1,2]. The use of GGBFS provides the important advantage of helping to avoid thermal cracks in concrete due to the low hydration process [2]. In fact, as previous results have shown, the hydration of GGBFS is slower than that of ordinary CEMI cement. Concrete with ground granulated blast furnace slag has a later setting time and a lower stiffness [3].

When equal amounts of cement and water binder (w/b) are used, concretes with slag content have a lower compressive strength at early ages and higher compressive strength at late ages than Portland cement [2]. Furthermore, with a specific compressive strength, slag concrete has a better mechanical performance in terms of tension than concrete made with Portland cement [2]. However, the study by Shumuye et al. [4] showed that the compressive strength of the concrete decreased as the slag content increased.

Environmental conditions and the temperature exposure during curing has a strong effect on concrete mechanical properties [3,5]. When the material is subjected to heating at higher temperatures up to 1000 ◦C, like during a fire, thermal damage occurs due to dehydration of the cement paste and the thermal mismatch of strains between the shrinking cement paste and expanding aggregates, which induces cracking [6,7]. Moreover, during the phase of cooling down to the ambient temperature, stresses induced by inversed thermal gradients result in the development of cracks within the cement paste that will affect permeability, and also may compromise the durability of the material after a fire [8,9]. The changes of concrete's mechanical properties at high temperatureshave been widely investigated [6–12], helping us to better understand the behavior of concrete structuresin a fire situation and to determine parameters influencing its behavior. The evolution of concrete mechanical properties in fire depends on the concrete composition: presence of mineral additions [4,13], w/c water cement ratio [14,15], the nature and type of aggregates [10–12]. Moreover, the concrete heating conditions: heating rate and maximum temperature of exposure play a major role in concrete strength evolution, as well as the testing procedure: hot tested concrete or tested after temperature exposure and cooling down to the ambient temperature [7,15]. Nevertheless, for material mechanical properties testing, a slow heating rate is recommended in order to ensure limitation of the thermal gradient inside the specimen. In the literature investigations the heating rates of 0.1–10 ◦C/min are employed. Nevertheless, the heating rates recommended by RILEM International Union of Laboratories and Experts in Construction Materials, Systems and Structures [16] depend on the specimen diameter and are from 0.5 to 2.0 ◦C/min for accidental conditions (fires).

The existing knowledge regarding the behavior of high performance concrete in a fire was recently reviewed by the RILEM Technical Committee HPB-227 [8], however, there are still no clear reports as to whether the properties of concretes subjected to high temperatures change in a similar or a very different way, depending on the cement type used.

According to Shumuye et al. [4], the addition of GGBFS seems to improve the resistance of concrete to fire conditions. It was highlighted that, when the exposure to fire temperature increased from 200 to 400 ◦C, the compressive strength increased for concrete with slag (70% ordinary Portland cement OPC and 30% slag cement, as well as 50/50 proportions). For the group of concretes with 30% OPC and 70% slag cement, the opposite behavior was observed. The concrete mix containing GGBFS usually has a lower thermal expansion coefficient than Portland cement. The 15% and 30% replacement of CEMI by GGBFS gives coefficients of thermal expansion of 22.7 <sup>×</sup> 10−6/ ◦C and 17.2 <sup>×</sup> 10−6/ ◦C, respectively, which is 99.2% and 75.5% of the value obtained for Portland cement paste [4]. However, a recent study by Asamoto et al. [17] highlighted that the reduction in the elastic modulus and increase in permeability of the concrete with GGBFS subjected to 65◦Cwere larger than those of concrete without slag. Indeed, astonishingly, this can be attributed to a larger thermal expansion coefficient and larger cement paste shrinkage with the slag, leading to the formation of microcracks around the aggregate.

Moreover, it can be concluded that the addition of aluminosilicate minerals like fly ash, ground granulated blast furnace slag (GGBFS), and silica fume (SF) can affect concrete behavior at high temperatures in a way that may produce spalling of heated concrete in material that is denser, and thus less permeable [18–20]. Lower permeability leads to moisture clog occurrence and increase of vapor pore pressures inside the heated concrete [20]. The moisture clog effect was explained and linked with the permeability decrease observed in temperature from 100 to 200 ◦C but this effect is observed when the permeability is tested at hot stage and not after cooling down when the residual values of permeability are determined, like in present study. An important finding on gas pore pressure development were provided by works of Kalifa et al. [19,20] and linked with the permeability.

Cases of fires that took place in engineering facilities (Gotthard tunnel, Chunnel tunnel, or Mont Blanc tunnel, for example) have caused numerous fatalities, but also significant financial losses. During these fires an important loss of concrete in tunnel linings was observed. The load-bearing capacity of the structural elements was reduced due to the explosive spalling. The spalling may take different forms, from small concrete pieces chipping, known as the popcorn effect, to explosive behavior when larger pieces of concrete are separated from the concrete element with great energy [21–24]. In all cases, concrete fire spalling leads to the exposure of steel reinforcement, which is sensitive to high temperatures [24,25]. So far, it has been confirmed that the type and composition of concrete, including the aggregate type, water cement ratio, pozzolanic mineral material, and moisture content of concrete, affect its behavior in fire conditions [8,13,19]. Research aimed at understanding the causes of the spalling phenomenon, as well as determination of material parameters affecting its intensity, has been carried out by experiments [21,22,24] and numerical analysis [21,25,26]. Thus, concrete spalling is one of the most interesting and complex phenomena occurring in concrete exposed to fire conditions. The RILEM Technical Committee 256-SPF: Spalling of concrete due to fire: Testing and modelling has been established, and is mainly dedicated to studying this specific behavior.

During heating, the permeability usually progressively increases [27–29], exceptwhenthe permeability of concrete may decrease [30] due to the moisture clog effect. In this situation the water vapor pressure increases in the material's pore network, which may lead to spalling behavior. It is believed that the interaction of high temperature, an increase in water vapor pressure in the material pores, and the internal stress state is responsible for the occurrence of concrete spalling [19–26]. It seems that the key parameter governing the occurrence of spalling is its permeability. In denser and less permeable concretes the risk of spalling is higher. Researchers have shown that in fire conditions, concretes that are modified with the addition of mineral additives like silica fume and calcareous filler are prone to spalling behavior. As the spalling behavior of concrete is mainly governed by its permeability, researchers have been testing the influence of GGBFS addition on concrete permeability. Recently, Karahan [27] showed an increase of concrete transport properties after exposure to temperatures of 400 ◦C, accompanied by compressive strength reduction. Moreover, the conclusion of the authors indicated an optimum GGBFS/cement blend from the point of view of material behavior in a fire of 50–70% slag content as the cement replacement.

Hence, the results available do not reflect all the relevant aspects of this topic, and additional investigation is required. The literature results cannot be compared to each other due to the fact that the mixes differ. A research programme was therefore proposed which would allow for a clear comparison of the influence of cement type on the mechanical and physical properties of concrete at high temperatures. For this we performed various tests on identical concrete mixes, for which the only changing factor was the cement. Therefore, the main goal of this work is to present the comparison of the changes in mechanical and physical properties of concretes made with two different cement types; CEMI and CEMIII. For all four concretes, the composition of cement paste, as well as the volume of cement paste and mortar, remained the same. Thus, the study reflected solely the cement type effect of Portland cement versus slag cement on the mechanical performances and permeability of concretes made with two types of aggregates: crushed basalt (B) and riverbed gravel (RB). For all the concretes tested, the amount of all components (cement paste and mortar volume) and aggregate type and nature, as well as the particle size distribution, was identical, apart from the type of cement.

This research investigates the mechanical performances and permeability of concretes made with different cements, to compare their reference mass transport capacities, strength, and stiffness after high temperature exposure. The reference values of permeability enable one to assess their potential for spalling in fire conditions, as denser and less permeable materials are prone to this behavior. Furthermore, the evolution of permeability with heating temperature was investigated, as well as the compressive strength and splitting tensile strength. Moreover, the stress strain curves were determined, and the modulus of elasticity was determined. All residual mechanical performances (fcT, ftT, ET) were evaluated after heating to temperature T (◦C), which corresponds to the post-fire performance of concrete in situations where the assessment of material properties is required. In this specific situation, the residual permeability of concrete is also an issue because it governs all aspects of durability, and there may be a need for assessment when a decision must be made on the further use of concrete elements after a fire.

#### **2. Materials, Specimen Preparation, Curing, and Heating**

The concretes investigated in this research were manufactured with the following components: Portland cement CEMI 42.5R and CEMIII/A 42.5 N containing 53% GGBFS, quartz sand 0/2mm, and one of two types of coarse aggregate: (B)basalt or (RB) riverbed gravel.

Cements from Lafarge (Małogoszcz, Poland) were used for both the CEMI 42.5 R Portland cement and CEMIII/A 42.5 N slag cement. The chemical characteristics of these cements are provided in Table 1, the physical characteristics in Table 2, and the mechanical characteristics in Table 3.


**Table 1.** The chemical characteristics of CEM I and CEM III cements(oxide analysis, % by mass).

**Table 2.** Physical characteristics of CEMI and CEMIII cements.




Two types of aggregates were used in this research programme: gravel from Dunajec River (Dwudniaki, Poland) and crushed basalt.

In Table 4, the concrete mixes are presented. The cement paste volume was 300 dm3/m<sup>3</sup> and the mortar volume was 550 dm3/m3. The concretes are denominated as B CEMI, B CEMIII, RB CEMI, and RB CEMIII. Plasticizer (BASF BV 18 (My´slenice, Poland) and superplasticizer (BASF Glenium SKY 591 (My´slenice, Poland) were used and the water-cement ratio (w/c) of the concretes was equal to 0.3.


**Table 4.** Mix composition of the test concretes.

All concrete cubic and cylindrical specimens were cast in plastic molds and stored for 24 h. After preliminary 24 h curing, the molds were covered with plastic lids for 7 days to prevent water evaporation. Samples were stored in laboratory conditions at T = 20 ± 5 ◦C and relative humidity HR = 50% ± 5%. Cylindrical specimens dedicated to permeability measurements were cut into discs with a diameter of 150 mm and thickness of 50 mm at the age of 28 days. At 90 days, all specimens for mechanical performance testing and permeability were heated in an electric furnace to T = 200, 400, 600, 800, and 1000 ◦C. As recommended by RILEM [16], a heating rate of 0.5 ◦C/min was applied. A slow heating rate is applied for concrete mechanical behavior testing at high temperatures in order to ensure limitation of the thermal gradient inside the specimen. When the target temperature was reached it was maintained for three consecutive hours in order to obtain a homogenous temperature in the whole cross section of the specimen. Afterwards, all specimens were cooled down inside of the furnace chamber.

#### **3. Testing Procedures**
