3.1. Characteristics of Heat-Treated Serpentines
The temperature interval for forming an active metastable phase was in the range of 550–800 °C, according to the data of thermal analysis of serpentines (
Figure 1). The dependence of activity on temperature is extreme since it is determined by the completeness of two processes that affect this indicator differently. On DSC curves, these processes correspond to the endothermic effect of the crystal lattice destruction of the serpentine mineral and the exothermic effect of the formation of high-temperature phases. The superposition of two peaks on DSC curves of the antigorite sample was observed, i.e., the formation of high-temperature inactive compounds began even before the completion of the serpentine lattice destruction (
Figure 1a, curve 1). In contrast, the temperature of these processes in lizardite and chrysotile differed by about 100 degrees (
Figure 1b–d, curve 1), i.e., the formation of high-temperature phases began after the destruction of the initial serpentine minerals.
Breuil et al. [
21] demonstrated that magnesium leaching from samples correlates with the content of the active amorphous phase; thus, the acid-neutralizing capacity reflects the interaction activity of heat-treated serpentines with aqueous solutions. This process is a key factor in forming the magnesia-silicate phase [
22]. The acid-neutralizing capacity of samples obtained at different firing temperatures is demonstrated in
Table 2. Sample SAP (antigorite) exhibited maximum activity at a temperature of 700 °C. The acid-neutralizing capacity of SCH and SLH samples in the temperature range of 650–700 °C differed slightly. Further, the samples obtained at optimal temperatures (antigorite—700 °C, lizardite, and chrysotile—650 °C) were studied.
The reactivity of the samples obtained at the optimal temperature increased in the SAP-SLH-SCH series, which corresponded to the values of the activation energy for the dehydroxylation process. Chrysotile was the least stable serpentine mineral with an activation energy of 184 kJ/mol; this value for lizardite and antigorite was significantly greater—221 and 255 kJ/mol, respectively [
23].
Evolution in the serpentine mineral composition after heat treatment was monitored by DSC and XRD. Endo-effect significantly reduced the DSC curves of the heat-treated lizardite and chrysotile caused by dehydroxylation of the serpentine minerals. Therefore, high destruction degree of the original serpentine minerals was observed under optimal firing conditions (650 °C).
XRD patterns had broad humps from 17 to 43° (2Θ) in all heat-treated samples (
Figure 2, curve 2), suggesting the presence of an amorphous component [
15]. Heat-treated chrysotile (SCH) and lizardite (SLH, SLK) samples did not contain crystallized serpentine minerals. Simultaneously, the formation of magnesium silicates was observed; in particular, micro- to nanocrystal forsterite Mg
2SiO
4 and enstatite Mg
2Si
2O
6 in an amorphous phase were found, according to [
24].
The IR spectra patterns proved the forsterite presence in a heat-treated SCH sample. Bands at 874 and 991 cm
−1 corresponded to the vibrations of SiO
4 tetrahedra in the forsterite structure; the broadening of these bands indicated its amorphous state. Heat-treated SLH and SLK contained micro- to nanocrystal enstatite in an amorphous phase along with forsterite (
Figure 2c,d, curve 2). For these samples, the content of the amorphous phase was determined, and the content of crystalline serpentine minerals was also quantified (
Table 3). The crystallinity of serpentine concentrates went down after roasting at 650 °C. The content of the crystalline serpentine mineral has mostly decreased in the lizardite SLH from 75.2 to 4.8 wt.%, whereas in the SLK samples from 44.8 to 13.9 wt.%.
Antigorite better retains its original structure under the optimal firing temperature (700 °C). The main endo effect of the crystal lattice destruction of the serpentine mineral was preserved on the DSC curve (
Figure 1a, curve 2). Main basal reflexes of the initial mineral were observed for the heat-treated antigorite sample obtained at 700 °C (
Figure 2a, curve 2). Antigorite is more resistant to roasting due to its less defective structure when compared with other serpentine varieties [
25]. Relatively large separate microblocks in the antigorite structure cause its stability and difficulties of water diffusion from the crystal lattice. Diffusion of water vapor in the antigorite lattice in the direction perpendicular to the basal plane (axis
c) is absented. Diffusion along the
a-axis was limited, and only along the
b-axis did water molecules move freely [
23].
The degree of serpentine activation was calculated as the ratio between experimental and theoretical acid-neutralizing capacity (%) (
Table 4). Theoretical acid-neutralizing capacity was calculated from the magnesium content in heat-treated samples. The degree of activation was 92% for chrysotile, 72% for lizardite, and 38% for antigorite. The lower the activation energy of dihydroxylation, the higher the degree of activation.
The study of serpentine thermolysis supervised by N.O. Zulumyan has shown that the efficiency of serpentine amorphization is primarily associated with the conditions of their formation in the Earth’s crust [
26]. The degree of silica extraction upon acid treatment in the current study was done in accordance with the method proposed by N.O. Zulumyan since it is the most objective index of the transformation degree of serpentine minerals into meta-serpentine [
27]. Heat-treated serpentine leaching by HCl (8 wt.%) led to the migration of (SiO
4)
4− anions from the serpentine silicate layer to the solution as orthosilicic acid [
25]. The silicon-containing phase, which was not leached by HCl (8 wt.%), had low activity and participated in the formation of a magnesia-silicate binder only under conditions of the alkaline component (MgO) high concentrations [
22]. Silicon was leached to a greater extent from lizardite (34%) and to a lesser extent from antigorite (28%) and chrysotile (26%). This indicator was not related to the activation energy of the dehydroxylation reaction of serpentine minerals and may reflect the impurities effect on the silicate components’ solubility. The silica solubility in the magnesia-silicate system with aluminum was higher [
28], and with calcium was lower [
29] compared to a magnesia–silicate system without impurities.
Amorphous silica, diagnosed by a reflex with a maximum of 22° (2Θ), was found in the leaching residues of all heat-treated serpentines (
Figure 2a–c, curve 4). Additional silicate phases were not diagnosed in the SCH sample; simultaneously, initial serpentine, forsterite, and enstatite were found in the SAP sample; forsterite and enstatite were determined in the SLH sample.
The degree of magnesium leaching by HCl (8 wt.%) was near 100% for chrysotile and 82–84% for lizardite and antigorite. This confirms the absence of magnesium silicates in the chrysotile leaching residue. The magnesium silicate products of the initial serpentines’ destruction in residues of antigorite and lizardite were more resistant to the reaction with hydrochloric acid.
The Mg/Si index in the reacting system is an important indicator affecting the composition and properties of the magnesia-silicate binding agent [
22]. This ratio varied from 3.9 to 5.1 in the active phase, which indicated Mg excess, i.e., the mineral formation during the interaction of heat-treated minerals with water proceeded in a system with high alkalinity. The pH values of aqueous suspensions of serpentines (heat-treated/hydrated heat-treated) increased in the series SAP (10.23/9.40)—SLK (10.34/9.55)—SCH (10.4/10.06).
3.3. The Hydration of Heat-Treated Serpentines
In contrast with the initial heat-treated samples, an additional endothermic effect at a temperature of 110–120 °C corresponding to the removal of physically bound water was revealed on the thermograms of hydrated heat-treated samples obtained after hardening for 28 days under wet conditions (
Figure 1, curve 3). At a temperature of 350–600 °C, the indefinite wide peak of the endothermic effect of water removal was observed; the peak is attributed to the dehydration process of both brucite Mg(OH)
2 and magnesium silicate hydrate (M-S-H) phases of variable composition [
34,
35]. M-S-H phases were identified by XRD reflections in 2Θ 22, 35, and 60° areas [
22].
Wide reflections corresponded to the decomposition products of serpentine minerals; forsterite and/or enstatite were observed in X-ray patterns of heat-treated serpentines in the areas of 2Θ 36–38° and 59–63°. The same compounds were retained in hydrated samples. According to [
35,
36,
37], these phases can be referred to as M-S-H phases, which have binding properties. The mechanism of the M-S-H phase formation during the interaction of heat-treated serpentine (lizardite) with an aqueous solution is described as sequential reactions: silica and magnesium dissolution; silica polymerization and precipitation on the surface of reacting particles; magnesium sorption from the solution to form a magnesium silicate phase [
38].
The hydration products can be identified using differential scanning calorimetry (DSC). Based on the results presented in [
35,
39], the mass losses were determined in the temperature ranges (°C): 20–350 (dehydration, phase D), 350–600 (dehydroxylation of magnesium silicate binder and binder precursor, phase B), 600–900 (dehydroxylation of serpentine minerals, phase S). This indicator was determined for initial, heat-treated, and hydrated heat-treated (after 28 days of hardening) samples.
Table 5 shows data on the phase content calculated for a dehydrated sample, i.e., excluding phase D. The general trend for all hydrated heat-treated samples was a decrease in the phase S content compared to initial samples and an increase in phase B compared to heat-treated samples. The number of phases and their ratio related to the degree of activation were calculated by the acid-neutralizing capacity. The greater the degree of activation, the less is phase S; the more phase B is, the greater the relative content of phase B is, calculated as phase B/(phase B + phase S).
3.4. Compressive Strength of Binder—Molded Hydrated Heat-Treated Serpentines
The binder formation during the interaction of heat-treated serpentines with water allows one to obtain materials with certain strengths. The compressive strength characteristics of the samples appear in
Table 6. Upon hardening within 7 days, the strengths of the SAP and SCH samples were almost the same (2.2 MPa), whereas this indicator for the SLH and SLK samples was significantly lower (0.5 MPa). Upon hardening for over a year, the chrysotile sample SCH had the highest strength (about 8 MPa), whereas the strength of antigorite SAP amounted to 3 MPa. Although the lizardite strength SLH increased to 1.1 MPa after a year of hardening, this indicator remains lower than in other varieties of serpentine. The strength of SLK did not increase over the entire hardening period and remained at 0.5 MPa.
The results presented in
Table 6 show the absence of a relation between compressive strength and the content of precursor of magnesium silicate binding agent in the heat-treated serpentines. The content of the heat-treated serpentine active phase increased in the row SAP–SLH–SCH, whereas the strength of the SLH samples was less compared with the SAP samples. Thus, the sample strength was affected not only by the active phase content but also by factors primarily attributed to the structure of serpentine minerals. The layered structure of microcrystallites, which is pretty common for lizardite, is the most likely reason for this discrepancy [
40].
Microscopic studies assist in explaining the results of the compressive strength. Unlike the loose microstructure of lizardite (
Figure 4a), the particles of the antigorite constitute a uniform, dense material (
Figure 4b). The surface texture of the particles of the initial chrysotile was characterized both by a porcelaneous shell-like fracture (as in antigorite) and by the presence of microcrystallinity (as in lizardite) (
Figure 4c).
New visible phases formed after the interaction between heat-treated serpentine minerals and water (
Figure 4b,d,e). The particles of antigorite and chrysotile made a uniformly compacted homogeneous structure in the newly formed binding material (
Figure 4d,f), while the lizardite sample represented a conglomerate of chaotic oriented plates, significantly reducing the strength of the resulting material (
Figure 4b).
The strength of the hardening products was affected by the crystalline structure of serpentine minerals as well as by their origin, namely their chemical composition, content, and impurity profile. For instance, the variations in the strength of the SLH and SLK samples in which lizardite is the main serpentine component were observed. After hydration, the crystallinity reduced in the SLH sample and increased slightly in the SLK sample (
Table 3). XRD revealed the recrystallization of a serpentine mineral in the SLK sample (
Figure 2d, curve 3). A decrease in the amorphous phase from 55.5 in heat-treated SLK to 48.5% in the hydrated heat-treated SLK sample indicated the processes of reverse crystallization of the binder precursor during the formation of binding material, which led to the reduction of the material strength. In addition to the difference between lizardite samples in terms of the content of the serpentine component and its behavior during firing and hydration, the mineral composition of impurities was also noteworthy. Due to its origin, the Kovdor serpentine sample differed from Khabozero one since it contained vermiculite, which negatively affected the strength of target materials.