Self-Heating Ability of Geopolymers Enhanced by Carbon Black Admixtures at Different Voltage Loads

Lukáš Fiala 1,* , Michaela Petříková 1, Wei-Ting Lin 2 , Luboš Podolka 3 and Robert Černý 1 1 Department of Materials Engineering and Chemistry, Faculty of Civil Engineering, Czech Technical University in Prague, Thákurova 7, 166 29 Prague 6, Czech Republic; michaela.petrikova@fsv.cvut.cz (M.P.); cernyr@fsv.cvut.cz (R.C.) 2 Department of Civil Engineering, College of Engineering, National Ilan University, No.1, Sec. 1, Shennong Rd., I-Lan 260, Taiwan; wtlin@niu.edu.tw 3 Department of Civil Engineering, Faculty of Technology, Institute of Technology and Business in České Budějovice, Okružní 517/10, 370 01 České Budějovice, Czech Republic; podolka@mail.vstecb.cz * Correspondence: fialal@fsv.cvut.cz; Tel.: +420-22435-7125


Introduction
Building materials with new functional properties that extend their usability in sophisticated applications, so-called multifunctional or smart materials, are currently in high demand by the construction industry. Studies dealing with their design, experimental determination of material properties, and testing of newly achieved abilities have been, and still are, mainly focused on cementitious composites. A comprehensive review dealing with a definition and classification of smart concretes and structures and possible applications introduced by Han et al. [1] showed that a variety of possible enhancements exist. Some of the new functional properties, such as self-sensing, self-heating, energy harvesting, or electromagnetic shielding/absorbing, are crucially dependent on electrical properties that are, in the case of common aluminosilicate-based building materials, often very poor. Therefore, some electrically conductive admixtures are necessary for the formation of a conductive net within the material matrix. Much effort has been devoted to studies dealing with influence of carbon-based and metallic admixtures on new functional properties of cementitious precursor. The differences in heat evolution of slag activated by NaOH, water glass, and a combination of NaOH and water glass were presented by Altan and Erdogan [23], and a faster initial reaction was observed in NaOH by isothermal calorimetric measurements by Haha et al. [24]. Mostly used alkaline activators are mixtures of sodium or potassium hydroxide (NaOH, KOH) with sodium or potassium water glass (n·SiO 2 ·Na 2 O, n·SiO 2 ·K 2 O) [25].
Multifunctional geopolymers can be designed in a similar way to multifunctional cementitious materials by the addition of electrically conductive admixtures [26]. However, the design of multifunctional geopolymers and their acquired abilities are not so well explored. Rovnaník et al. [27] compared self-sensing properties of alkali-activated slag mortars with Portland cement mortars and concluded that, due to some content of iron particles in slag, an applicable sensitivity is evident in practice, even without any electrically conductive admixture, whereas in the case of Portland cement, mortar self-sensing is detectable but not sufficient for practical applications. Another similar study performed by Rovnaník et al. [28] dealt with the self-sensing ability of a geopolymer mortar based on slag activated by water glass with GP admixture. They concluded that such materials exhibit a self-sensing ability but with a significant decrease in compressive strength. Concerning the self-heating ability, it was experimentally confirmed on small alkali-activated slag samples with a CB admixture in the amount of 8.89 wt. % at 32.1 and 41.5 V by Fiala et al. [29] Slag as a high-calcium precursor is a solid waste generated by the iron and steel industry. In 2014, slag was produced in the amount of 250 Mt within the 1.6 Gt of global steel production [30]. In 2013, the annual slag production of one of the leading producers, China, reached more than 100 million tons with just 29.5% utilization rate, which is very low in comparison to industrial countries. The utilization rate reaches 98.4% in Japan, 87.0% in Europe, and 84.4% in the United States. As of 2016, more than 300 million tons of accumulated steel slag has not been used effectively in China, which, taking into account large steel slag emissions, causes an important environmental problem for China [31]. Slag in granulated form is a precursor that can be relatively easily alkali-activated, and originating geopolymers can be used in the construction industry.
Within the research presented in this paper, granulated blast-furnace slag (GBFS) was used as a precursor for alkali activation by water glass, and CB admixtures were added in various dosages in order to enhance the effective electrical properties of the designed geopolymers that would be promising in terms of the self-heating ability. Subsequently, material properties involving basic physical, mechanical, thermal, and electrical properties were experimentally determined, and self-heating tests were conducted in order to verify the self-heating ability of such materials. It was observed that the self-heating ability of the tested materials started from a CB amount of 1.5 wt. %, and such material is able to generate heat at a DC voltage of 40 V leading to a small temperature increase. The best self-heating performance was observed for geopolymers with 2.25 wt. % of CB at 100 V, where the temperature increase was about 110 • C in approximately one hour.

Materials and Methods
A high-calcium precursor, GBFS SMŠ 380 (39.8% CaO), produced by Kotouč Štramberk Ltd. was activated by water glass Susil produced by Vodní sklo a.s. GBFS is an industrial byproduct of iron production with a fineness of 380 m 2 ·kg −1 (Blaine). The average grain size of the slag particles determined by laser granulometry was d 50 = 15.5 µm and d 90 = 38.3 µm. Activation was performed by sodium silicate solution (water glass Susil MP 2.0 with a molar ratio SiO 2 /Na 2 O = 2.07). Three normalized CEN fractions of fine quartz sand (PG1, PG2, PG3) produced by Filtrační písky, Ltd., that complied with theČSN EN 196-1 standard were used as a filler. The effective electrical properties of the composites were enhanced by CB VULCAN 7H. CB is essentially elemental carbon in the form of spherical particles and aggregated clusters of those particles with a high surface area and high electrical conductivity. The average grain size of CB particles was d 50 = 0.52 µm and d 90 = 17.6 µm. In Figure 1, the particle size distribution of CB VULCAN 7H determined by laser granulometry is presented. CB VULCAN 7H is mainly manufactured for tires and industrial rubber production [32]. Its production is performed by thermal cracking of heavy aromatic feedstock, such as oil, in a hot flame. Oil is injected into a furnace hot flame zone where hydrocarbons are cracked to carbon and hydrogen by means of quenching the flame by water. Because of the reasonably high electrical conductivity and high surface area of such a way of processing CB, it can be used to optimize the electrical properties of polymers [33] and inorganic building materials [34,35].
In Table 1, the compositions of geopolymers with an optimized amount of mixing water are given. Seven different mixtures, the reference geopolymer (CB 0), and geopolymers with CB in the amounts of 0.75 wt. %, 1.25 wt. %, 1.5 wt. %, 1.75 wt. %, 2 wt. %, and 2.25 wt. % were designed and prepared by the following procedure. First, suspensions (10% and 15%) were prepared by adding a given amount of CB powder into water with nonionic surfactant Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) in the form of 0.5% solution and stirred by homogenizer IKA ULTRA-TURRAX for 15 min. An initial amount of additional water was then added to the GBFS suspension and stirred by a mixer for several minutes. In order to eliminate foaming during mixing leading to the formation of large pores, 1% solution of siloxane-based air-detraining agent Lukosan S (Lučební závody, Kolín, Czech Republic) was added. Subsequently, three fractions of sand were added to the mixture and stirred again by a mixer for several minutes. Consistency of fresh mixtures was tested according to the ČSN EN 1015-3 standard Determination of Consistence of Fresh Mortar by Flow Table, which is mainly used for cementitious mortars but is used also for alkali-activated ones [36]. The water-to-slag ratio of the mixtures was adjusted so that the average base diameter was equal to 160 mm, which is within the plastic range (140-200 mm) closer to the dry consistency bounds. Final mixtures with optimized water-to-slag ratios were then poured into molds and covered by a plastic cover. After one day, solidified samples were demolded and placed into a water bath for 28 d. Before the experiments (except the experimental determination of mechanical properties), samples were dried in an oven and subsequently cooled down in desiccator with silica gel.   CB VULCAN 7H is mainly manufactured for tires and industrial rubber production [32]. Its production is performed by thermal cracking of heavy aromatic feedstock, such as oil, in a hot flame. Oil is injected into a furnace hot flame zone where hydrocarbons are cracked to carbon and hydrogen by means of quenching the flame by water. Because of the reasonably high electrical conductivity and high surface area of such a way of processing CB, it can be used to optimize the electrical properties of polymers [33] and inorganic building materials [34,35].
In Table 1, the compositions of geopolymers with an optimized amount of mixing water are given. Seven different mixtures, the reference geopolymer (CB 0), and geopolymers with CB in the amounts of 0.75 wt. %, 1.25 wt. %, 1.5 wt. %, 1.75 wt. %, 2 wt. %, and 2.25 wt. % were designed and prepared by the following procedure. First, suspensions (10% and 15%) were prepared by adding a given amount of CB powder into water with nonionic surfactant Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) in the form of 0.5% solution and stirred by homogenizer IKA ULTRA-TURRAX for 15 min. An initial amount of additional water was then added to the GBFS suspension and stirred by a mixer for several minutes. In order to eliminate foaming during mixing leading to the formation of large pores, 1% solution of siloxane-based air-detraining agent Lukosan S (Lučební závody, Kolín, Czech Republic) was added. Subsequently, three fractions of sand were added to the mixture and stirred again by a mixer for several minutes. Consistency of fresh mixtures was tested according to theČSN EN 1015-3 standard Determination of Consistence of Fresh Mortar by Flow Table, which is mainly used for cementitious mortars but is used also for alkali-activated ones [36]. The water-to-slag ratio of the mixtures was adjusted so that the average base diameter was equal to 160 mm, which is within the plastic range (140-200 mm) closer to the dry consistency bounds. Final mixtures with optimized water-to-slag ratios were then poured into molds and covered by a plastic cover. After one day, solidified samples were demolded and placed into a water bath for 28 d. Before the experiments (except the experimental determination of mechanical properties), samples were dried in an oven and subsequently cooled down in desiccator with silica gel. The bulk density was determined on dry samples with dimensions of 50 × 50 × 50 mm 3 by means of the gravimetric method. The matrix density was determined on samples with dimensions of 50 × 50 × 50 mm 3 by means of the vacuum saturation method. With respect to the bulk density and the matrix density, the total open porosity Ψ (%) was calculated by the following equation where ρ v (kg·m −3 ) is the bulk density, and ρ mat (kg·m −3 ) is the matrix density. Mechanical properties were determined on three samples with dimensions of 40 × 40 × 160 mm 3 according to the Czech StandardČSN EN 196-1. Samples cured for 28 d were first tested in terms of flexural strength by a three-point bending test. The length between the supports was 100 mm, and the loading rate was 0.15 mm/min. The compressive strength was then determined on six halves of the prisms originating from the previous flexural strength tests.
Thermal properties were determined on samples with dimensions 70 × 70 × 70 mm 3 by a commercial device ISOMET 2114 (Applied Precision, Ltd.) attached by a surface probe by means of the transient heat-pulse method. Such measurements were based on an analysis of the temperature response to the generated heat flow pulses. Heat flow was induced by electric heating using a resistor heater placed in the probe having direct thermal contact with the surface of the sample. First, the probe was heated up, and, subsequently, temperature decrease was monitored after the heater was turned off. With the known geometry of the probe and a decrease of the temperature due to dissipation of the heat in the sample, thermal properties were identified.
Electrical properties were measured in a two-probe arrangement in DC regime. First, the samples with dimensions 50 × 50 × 50 mm 3 were attached to electrodes where two opposite lateral sides were painted with a conductive carbon paint. Then, copper tape was pasted onto the first conductive layer in order to gain good contact of the samples with a power supply and wattmeter (self-heating experiment) and multimeter (electrical properties).
Resistance R (Ω) of the dried samples was measured by a Fluke 8846A 6 1 2 digit precise multimeter, and the electrical conductivity σ (S·m −1 ) was calculated with respect to the shape ratio of the samples (electrodes: 50 × 50 mm 2 , distance between electrodes: 50 mm) by the following equation where R (Ω) is the resistance of the sample, l (m) is the distance between electrodes, and S (m 2 ) is the area of electrodes. Self-heating experiments were performed on samples with dimensions of 50 × 50 × 50 mm 3 . Dried samples with electrodes attached in the same way as for determining the electrical properties were connected to a GW Instek GPR-11H30D voltage power supply and loaded by one or two voltage levels ( Figure 2). Mortar samples CB 0 and CB 0.75 were loaded by 40 V to prove the no self-heating ability that was expected because of the electrical conductivity measurements. CB 1.25 mortar and mortars with higher amounts of CB were tested at two voltage levels, 40 and 100 V. Electrical power was monitored by a GW Instek GPM-8213 wattmeter. Ambient laboratory temperature and temperatures of samples were monitored by Pt-100 probes supported by Ahlborn ALMEMO 8690-9A datalogger in the central position of the bottom sides of the samples perpendicular to the attached electrodes.

Basic Physical Properties
In Table 2, the basic physical properties were represented by the bulk density, the matrix density, and the total open porosity. The bulk density was highest for the reference mortar CB 0 (2111 kg•m −3 ) and decreased systematically with increasing CB amounts down to 1720 kg•m −3 (CB 2.25). The matrix density was in the range of 2562-2602 kg•m −3 and did not exhibit a significant influence on the amount of CB admixture. The total open porosity was calculated from the bulk density and the matrix density by Equation (1). It was lowest for CB 0 (17.6%) and exhibited an increasing tendency with increasing amounts of CB up to 33.4% (CB 2.25). Compared to the reference mortar, the addition of 0.75 wt. % of CB resulted in an increase in porosity of 7.6%, whereas addition of 2.25% wt. % of CB almost doubled the porosity (increase of 15.8%).

Mechanical Properties
In

Basic Physical Properties
In Table 2, the basic physical properties were represented by the bulk density, the matrix density, and the total open porosity. The bulk density was highest for the reference mortar CB 0 (2111 kg·m −3 ) and decreased systematically with increasing CB amounts down to 1720 kg·m −3 (CB 2.25). The matrix density was in the range of 2562-2602 kg·m −3 and did not exhibit a significant influence on the amount of CB admixture. The total open porosity was calculated from the bulk density and the matrix density by Equation (1). It was lowest for CB 0 (17.6%) and exhibited an increasing tendency with increasing amounts of CB up to 33.4% (CB 2.25). Compared to the reference mortar, the addition of 0.75 wt. % of CB resulted in an increase in porosity of 7.6%, whereas addition of 2.25% wt. % of CB almost doubled the porosity (increase of 15.8%).

Mechanical Properties
In

Thermal and Electrical Properties
In Table 4, thermal properties were represented by the thermal conductivity, and the specific heat capacity and electrical properties were represented by the electrical conductivity. The highest thermal conductivity was observed for CB 0 (λ = 1.71 W·m −1 ·K −1 ). The decreasing tendency of the thermal conductivity with an increasing amount of CB admixture corresponds with the basic physical properties (the bulk density was directly proportional, and the total open porosity was inversely proportional).  The electrical conductivity increased significantly with an increasing amount of CB because of the formations of conductive paths within the geopolymer matrix. Such an increase is a very important assumption for self-heating ability of the tested mortars. The initial value of the electrically non-conductive mortar CB 0 (8 × 10 −7 Ω·m) improved to the highest value (1.3 × 10 −1 Ω·m) observed for CB 2.25, which was an increase of about 7 orders of magnitude. With respect to the measured data, it was expected that the percolation threshold for the self-heating ability would be at 1.5 wt. % of CB. The electrical conductivity of CB 1.5 mortar was 4000 times higher than that of the reference mortar (3.2 × 10 −3 Ω·m).

Self-Heating Ability
In Figures 3-14, self-heating experiments are presented. Each self-heating experiment involved the determination of time dependencies of the ambient temperature, the temperature of the sample, and the power (dependent on power supply voltage and the passing current). Self-heating experiments conducted at 40 V on CB 0, CB 0.75, and CB 1.25 (Figures 3-5) and at 100 V on CB 1.25 ( Figure 6) confirmed the expectations obtained by measurements of electrical properties that such materials did not exhibit a self-heating ability. The passing current, which is directly proportional to the measured          A low self-heating ability was observed for CB 1.5 at 40 V, where Δt ≈ 2 °C and P ≈ 0.21 W (Figure 7). An increased passing current induced by an increase of the voltage level from 40 to 100 V resulted in a temperature increase of Δt ≈ 9.5 °C with the corresponding power P ≈ 1.18 W (Figure 8).   A low self-heating ability was observed for CB 1.5 at 40 V, where ∆t ≈ 2 • C and P ≈ 0.21 W (Figure 7). An increased passing current induced by an increase of the voltage level from 40 to 100 V resulted in a temperature increase of ∆t ≈ 9.5 • C with the corresponding power P ≈ 1.18 W (Figure 8). A low self-heating ability was observed for CB 1.5 at 40 V, where Δt ≈ 2 °C and P ≈ 0.21 W (Figure 7). An increased passing current induced by an increase of the voltage level from 40 to 100 V resulted in a temperature increase of Δt ≈ 9.5 °C with the corresponding power P ≈ 1.18 W (Figure 8).   A low self-heating ability was observed for CB 1.5 at 40 V, where Δt ≈ 2 °C and P ≈ 0.21 W (Figure 7). An increased passing current induced by an increase of the voltage level from 40 to 100 V resulted in a temperature increase of Δt ≈ 9.5 °C with the corresponding power P ≈ 1.18 W (Figure 8).  exhibited a good self-heating performance at 40 V, where ∆t ≈ 10 • C and P ≈ 1.25 W (Figure 11), and at 100 V, where ∆t ≈ 50 • C and P ≈ 7.41 W were achieved ( Figure 12).
Energies 2019, 12, x FOR PEER REVIEW 10 of 15 CB 1.75 exhibited a slightly better self-heating performance than the geopolymer mortars with a lower amount of CB. At 40 V, Δt ≈ 3 °C and P ≈ 0.51 W were achieved (Figure 9). At 100 V, further increases in the temperature Δt ≈ 22 °C and the power P ≈ 3.45 W were observed ( Figure 10). CB 2 exhibited a good self-heating performance at 40 V, where Δt ≈ 10 °C and P ≈ 1.25 W (Figure 11), and at 100 V, where Δt ≈ 50 °C and P ≈ 7.41 W were achieved ( Figure 12).    Energies 2019, 12, x FOR PEER REVIEW 10 of 15 CB 1.75 exhibited a slightly better self-heating performance than the geopolymer mortars with a lower amount of CB. At 40 V, Δt ≈ 3 °C and P ≈ 0.51 W were achieved (Figure 9). At 100 V, further increases in the temperature Δt ≈ 22 °C and the power P ≈ 3.45 W were observed ( Figure 10). CB 2 exhibited a good self-heating performance at 40 V, where Δt ≈ 10 °C and P ≈ 1.25 W (Figure 11), and at 100 V, where Δt ≈ 50 °C and P ≈ 7.41 W were achieved ( Figure 12).   Energies 2019, 12, x FOR PEER REVIEW 10 of 15 CB 1.75 exhibited a slightly better self-heating performance than the geopolymer mortars with a lower amount of CB. At 40 V, Δt ≈ 3 °C and P ≈ 0.51 W were achieved (Figure 9). At 100 V, further increases in the temperature Δt ≈ 22 °C and the power P ≈ 3.45 W were observed ( Figure 10). CB 2 exhibited a good self-heating performance at 40 V, where Δt ≈ 10 °C and P ≈ 1.25 W (Figure 11), and at 100 V, where Δt ≈ 50 °C and P ≈ 7.41 W were achieved ( Figure 12).         In Figure 15a, the maximal values of power from the conducted self-heating experiments loaded by 40 and 100 V are summarized. Geopolymer mortars with an amount of CB admixture starting at 1.5 wt. % were able to generate heat, leading to a temperature increase at both voltage levels. Comparing the heating power at 40 and 100 V of the mortars able to evolve heat, the highest increase was achieved for CB 2.25 (7.2 times higher power at 100 V than at 40 V), whereas the lowest non-zero increase was for CB 1.5 (5.6 times higher power at 100 V than at 40 V). In Figure 15b, the maximal achieved temperatures of the mortars during the self-heating experiments and the compressive strengths are presented. Quantities exhibited the opposite trend, where higher temperatures were achieved with an increasing amount of CB, but mechanical properties deteriorated. In Figure 15a, the maximal values of power from the conducted self-heating experiments loaded by 40 and 100 V are summarized. Geopolymer mortars with an amount of CB admixture starting at 1.5 wt. % were able to generate heat, leading to a temperature increase at both voltage levels. Comparing the heating power at 40 and 100 V of the mortars able to evolve heat, the highest increase was achieved for CB 2.25 (7.2 times higher power at 100 V than at 40 V), whereas the lowest non-zero increase was for CB 1.5 (5.6 times higher power at 100 V than at 40 V). In Figure 15b, the maximal achieved temperatures of the mortars during the self-heating experiments and the compressive strengths are presented. Quantities exhibited the opposite trend, where higher temperatures were achieved with an increasing amount of CB, but mechanical properties deteriorated.

Discussion
With respect to the data presented in Table 2 and 3, decreases in the bulk density, compressive strength, and flexural strength and an increase in the total open porosity of the designed geopolymer mortars were observed with an increasing amount of CB admixture. This mainly was due to the high surface area of particles and aggregated clusters of particles of such ECA filler ( Figure 1). The average grain size of CB was d50 = 0.52 μm and d90 = 17.6 μm, which was significantly lower compared to the grain size of the slag binder (d50 = 15.5 μm and d90 = 38.3 μm). Because of the high surface area of CB, more mixing water was needed (Table 1), which resulted in an increase in the total open porosity. However, an increased amount of water was necessary for the preparation of mixtures with plastic consistencies with an average base diameter equal to 160 mm according to the ČSN EN 1015-3 standard Determination of Consistence of Fresh Mortar by Flow Table. The highest decrease in bulk density and the highest increase in total open porosity observed between the reference mortar (CB 0) and the mortar with the highest amount of CB (CB 2.25) were about 18.5% and 90%.
Mechanical properties of the mortars with higher CB dosages were negatively influenced by an increased total open porosity. The compressive strength of CB 0 (83.45 MPa) was significantly higher compared to that of the mortars with the self-heating ability. The decrease was significant (CB 1.5 about 76% compared to CB 0), but remained at a good level (20.38 MPa up to 1.5 wt. % of CB admixture). Geopolymer mortars with an amount of CB higher than 1.75 wt. % exhibited low compressive strength, equal to 7.31 MPa, which was about a 91% decrease compared to the reference mortar. However, it should be noted that the compressive strength of cementitious mortars widely used in practice with aggregates up to 2 mm is usually up to 10 MPa. In the case of the flexural strength, the highest value was observed for CB 0 (8.26 MPa) and decreased with an increasing amount of CB, but not as much as in the case of the compressive strength (CB 2.25 compared to CB 0,

Discussion
With respect to the data presented in Tables 2 and 3, decreases in the bulk density, compressive strength, and flexural strength and an increase in the total open porosity of the designed geopolymer mortars were observed with an increasing amount of CB admixture. This mainly was due to the high surface area of particles and aggregated clusters of particles of such ECA filler ( Figure 1). The average grain size of CB was d 50 = 0.52 µm and d 90 = 17.6 µm, which was significantly lower compared to the grain size of the slag binder (d 50 = 15.5 µm and d 90 = 38.3 µm). Because of the high surface area of CB, more mixing water was needed (Table 1), which resulted in an increase in the total open porosity. However, an increased amount of water was necessary for the preparation of mixtures with plastic consistencies with an average base diameter equal to 160 mm according to theČSN EN 1015-3 standard Determination of Consistence of Fresh Mortar by Flow Table. The highest decrease in bulk density and the highest increase in total open porosity observed between the reference mortar (CB 0) and the mortar with the highest amount of CB (CB 2.25) were about 18.5% and 90%.
Mechanical properties of the mortars with higher CB dosages were negatively influenced by an increased total open porosity. The compressive strength of CB 0 (83.45 MPa) was significantly higher compared to that of the mortars with the self-heating ability. The decrease was significant (CB 1.5 about 76% compared to CB 0), but remained at a good level (20.38 MPa up to 1.5 wt. % of CB admixture). Geopolymer mortars with an amount of CB higher than 1.75 wt. % exhibited low compressive strength, equal to 7.31 MPa, which was about a 91% decrease compared to the reference mortar. However, it should be noted that the compressive strength of cementitious mortars widely used in practice with aggregates up to 2 mm is usually up to 10 MPa. In the case of the flexural strength, the highest value was observed for CB 0 (8.26 MPa) and decreased with an increasing amount of CB, but not as much as in the case of the compressive strength (CB 2.25 compared to CB 0, about a 71% decrease). Taking into consideration cement mortars widely used in practice with flexural strengths up to 2.5 MPa, all the designed geopolymer mortars are comparable.