Characterization of Multimodal Silicas Using TG / DTG / DTA , Q-TG , and DSC Methods

The formation of hierarchical, multimodal porosity materials with controlled shape and size of pores is the essential challenge in materials science. Properties of silica materials depend largely on different features: crystal structure, dispersity, surface composition, and porosity as well as the method of preparation and possible modification. In this paper, multimodal silicas obtained using different additives are presented. A-50 and A-380 aerosils and wide-porous SiO2 milled at 300 rpm were used as the additives in the sol stage at 20 ◦C, the sol–gel stage followed by hydrothermal modification (HTT) at 200 ◦C, or in the mechanochemical treatment (MChT) process. The characterizations were made by application of N2 adsorption/desorption, SEM imaging, quasi-isothermal thermogravimetry (Q-TG), dynamic thermogravimetry/derivative thermogravimetry/differential thermal analysis (TG/DTG/DTA), and cryoporometry differential scanning calorimetry (DSC) methods. Results showed that such a one-step preparation method is convenient and makes it possible to obtain multimodal silicas of differentiated porous structures and surface chemistry.


Introduction
Silica belongs to the oxide materials commonly used as efficient adsorbents, phases or phase carriers for gas or liquid chromatography, supports for active phases in catalysis, dispersion agents, and in numerous other applications [1,2].With so many applications, a very important challenge for materials science is the development of a SiO 2 preparation method with diverse porous structures and physicochemical properties.A special case is materials with an ordered structure and those characterized by multimodal porosity.Such materials are prepared using, for example, polystyrene latex spheres, a novel block of copolymers and templates (an ionic liquid surfactant) [3], or close-packed polystyrene beads, octadecyltrimethylammonium chloride (template), and tetraethoxysilane (TEOS) [4].However, a new, attractive method of obtaining such silicas can be the use of SiO 2 additive of various origins and structures in the selected stage of the traditional sol-gel procedure [5].The second component can be added in the sol or gel stage using specific processes, e.g., hydrothermal treatment (HTT) or mechanochemical treatment (MChT).
Due to the specific structure of dry materials, it is necessary to characterize both the shape and size of the pores as well as surface chemistry with full particulars.The term "silica" indicates a substance of stoichiometric composition, i.e., SiO 2 , but it also includes hydrated structures of SiO 2 •xH 2 O composition.This means that water is chemically bound with the SiO 2 surface in a nonstoichiometric amount.Such -OH groups also determine some important properties [6,7].Thus, it is necessary to elaborate a convenient method to characterize the porosity of multimodal silicas.Many modern research methods are used to assess the characteristics of silica materials.These include, among others, adsorption of different gases and liquids, thermal analysis under dynamic or quasi-isothermal conditions [8][9][10][11][12][13][14][15], FTIR and NMR spectroscopies [8,9,14,[16][17][18], ESR [8], SEM or TEM imaging [8,9], and many other methods.
The aim of the paper was to study the structural and thermal properties of multimodal silicas of different origins.The silicas were obtained using the addition of the second component (SiO 2 as well) in the sol stage at 20 • C, the sol-gel transition at 200 • C under hydrothermal conditions (HTT), or in the wet gel stage during MChT.Characterization was made by application of N 2 adsorption/desorption as well as SEM.Moreover, three different methods of studying the interactions between the SiO 2 surface and water were utilized: (1) quasi-isothermal desorption of water from the silica surface and the internal spaces (Q-TG), (2) thermogravimetry under dynamic conditions to characterize the amount and character of surface hydroxyl groups (thermogravimetry/derivative thermogravimetry/differential thermal analysis (TG/DTG/DTA)), and (3) differential scanning calorimetry (DSC) used as a cryoporometric method.
The introduction of additives into the reaction mixture was made in different ways: in the sol stage at 20 • C in air (samples 1, 2, and 3), the sol-gel transition at 200 • C in the autoclave (HTT, samples 4, 5, 6, and 7), or in the wet gel stage during MChT at 300 rpm (samples 8, 9, 10) or 500 rpm (sample 11).MChT was performed using the ball mill Pulverisette-7 (Fritsch, Germany).

Textural Characteristics
The low-temperature (−195 • C) isotherms of nitrogen adsorption/desorption were measured using an ASAP 2010 adsorption analyzer (Micromeritics).The specific surface areas (S BET ) were calculated from the isotherms using the Brunauer-Emmett-Teller (BET) method.The total pore volume (V p ) was equal to that of liquid nitrogen adsorbed at p/p o ≈ 0.98 (p and p o denote the equilibrium and saturation pressure of nitrogen at −195.8 • C, respectively).Assuming the cylindrical shape of pores, the average pore radius was calculated as R = 2V p /S BET [19].The nitrogen desorption data were used to calculate the pore size distribution (PSD, differential f V (R)~dV p /dR and f S (R)~dS/dR) using the self-consistent regularization (SCR) procedure under non-negativity condition (f V (R) ≥ 0 at any pore radius R) at the fixed regularization parameter α = 0.01.A pore model was applied with cylindrical (C) pores and voids (V) between the spherical aggregates of primary nonporous nanoparticles, NPNP packed in random aggregates (self-consistent regularization (SCR) procedure applied to integral adsorption equations based on a complex model with slit-shaped and cylindrical pores and voids between spherical nonporous particles packed in random aggregates (SCV), SCV/SCR method) [20].The differential f S (R) functions were recalculated to the incremental pore-size distribution (IPSD v , dV/dR = f(R)) using the procedure described in Reference [21].The differential f S (R) functions were used to estimate the deviation (∆w) of the pore shape [21].The f V (R) functions were also used to calculate the contribution of micropores (V micro , S micro for R p < 1 nm), mesopores (V meso , S meso for 1 nm < R p < 25 nm), and macropores (V macro , S macro for R p > 25 nm) [10,22,23].The obtained results are summarized in Table 1.

Thermal Analysis
The thermal studies of the samples were carried out during heating in air using a derivatograph C (Paulik, Paulik & Erdey, MOM, Budapest).TG/DTG/DTA curves were registered.Ceramic crucibles were used.
Thermal analysis of tested silicas under dynamic conditions was performed in order to calculate the number of hydroxyl groups present on the surface.The initial water content in the samples depends on the modification method.Therefore, before the measurements, the samples were dried in oven at 200 • C to remove the physically bound water.Dry silicas (m s ~10 mg) were then heated up to 1200 • C (10 • C min −1 ).On the basis of mass losses in different temperature ranges, the contents of vicinal silanol groups (200-400 • C), germinal silanol groups (400-900 • C), and isolated silanol groups (900-1200 • C) on the surface of silicas were calculated.It should be noted that the former can form surface and structural hydrogen bonds, while the latter do not combine with each other by hydrogen bonds because the distances between them are too short.Simultaneously, internal (structural) -OH groups were removed when the temperature increased from 200 • C to 900 • C. As a result, only siloxane groups were present on the silica surface at a temperature higher than 1200 • C [24].Moreover, the total water content was calculated on the basis of mass loss in the range of 20-1200 • C.
The concentration of hydroxyl groups in relation to the mass of silica υ OH (mmole g −1 ) was calculated according to the following equation: where 18 is the molar mass of water, and ∆m denotes the sample mass loss (% w/w).It should be mentioned that two OH surface groups form one water molecule.However, the concentration of hydroxyl groups in relation to the specific surface area OH (σ OH [OH nm −2 ]) was determined from the de Boer equation [25]: where S denotes the specific surface area, and ∆m denotes the sample mass loss (% w/w).

Quasi-Isothermal Desorption of Water
Water thermodesorption under the quasi-isothermal conditions was performed under the conditions of low coverage of the surface with adsorbed water.The water vapor was adsorbed on the silica surface at 20 • C. A portion (ca. 1 g) of adsorbent was placed in a desiccator containing bidistilled water for 72 h.Then, the weighed portions (ca. 25 mg) of the adsorbents were placed in a spherical spiral platinum crucible.Al 2 O 3 was used as a reference material.The measurements were performed in the temperature range of 20-300 • C using a derivatograph C (Paulik, Paulik and Erdey, MOM, Hungary).
The experimental Q-TG data corresponded to the water mass loss with increasing temperature.When measurements are performed over a temperature range without sample mass changes, then the ∆m = f(T) dependence reflects the changes in concentration of the water adsorbed on the adsorbent surface under equilibrium.The procedure has been previously described in detail [11,12,[26][27][28].
The quasi-isothermal thermogravimetric analysis can be used to characterize energetic and structural heterogeneity of silica surfaces and allows to determine changes in the free energy of adsorbed water layer during their desorption as well as the dependence of the effective radius of a drop evaporated from the surface on the adsorbed water concentration.Additionally, during such investigations, the structure of water layers adsorbed on the silica surface was determined.The dependences dG/dM H2O = f(C H2O ) were examined under the conditions of low coverage of the adsorbent surface with the adsorbed water.Such analysis can show the formation of various types of water clusters on the heterogeneous surfaces.
On the basis of the obtained data, the total concentrations of water adsorbed on the surface of studied adsorbents (mg g −1 ) were calculated.The number of statistical monolayers "n" of water was determined from the following formula [28]: where TG max is the maximum mass loss at 300 • C (mg g −1 ), N is the Avogadro number, ω is the surface occupied by one adsorbed water molecule (assumed 0.09 nm 2 ), S BET is the specific surface area of the tested adsorbents (m 2 g −1 ), and M is the molar mass of water (18 g mol −1 ).

DSC Investigations
DSC examination of interactions between the silica surface and water was performed using a differential scanning calorimeter (PYRIS Diamond, Perkin Elmer Instruments, USA) at the constant freezing/melting rate of 2 • C min −1 .PYRIS Diamond DSC was calibrated using distilled water (melting temperature Tm = 0 • C) and the standard indium sample (Tm = 156.6 • C) using the standard calibration procedure.
Before the investigations, the samples were prepared according to the procedure used for Q-TG measurements.The saturated samples (about 5 mg) were sealed into aluminum pans.As a reference, an empty aluminum pan was used.Measurements were made in the helium atmosphere (inert gas).In order to prevent system defrosting, curtain gas (nitrogen) was used.The freezing/melting studies were carried out in the temperature range of +20 to −50 • C. The melting thermograms were used to determine the PSD curves and define the temperature of the maximum (T max ) and the beginning (T onset ) of the peaks as well as the thermal (∆H) characteristics of the outgoing phase transformations.
In cryoporometry DSC, the fact of the depression of liquids' freezing point because of confining effect in the pores can be observed.The degree of pore filling by media of different polarity depends on the surface nature of a tested material and is related to the porosity.In this study, it was assumed that the ice in the pores of R min size melts at T onset , while the temperature T max corresponds to the melting of water in the pores of R av .The correlation between the liquid freezing point depression and the cylindrical pore radius R p [nm] is described by the Gibbs-Thomson equation (GT) [26][27][28][29][30]: where T m and T m0 are the temperatures of melting of ice confined in the pores in the bound and bulk forms, respectively, and k GT is the constant (k GT = 32.33K•nm).The PSD dV/dR [cm 3 nm −1 g −1 ] was determined from the ice melting thermogram: where dq/dt, ρ, β, m, and ∆H(T) are the heat flow, the ice density, the rate of changes of temperature, the sample mass, and the ice melting enthalpy, respectively.The water density is expressed as follows: The ∆H [J g −1 ] dependence on the temperature was calculated from the following equation:

Scanning Electron Microscopy
The morphology and structure of the silica aggregates were analyzed using SEM with a DualBeam Quanta 3D FEG FEI apparatus under low vacuum conditions at the accelerating voltage of 30 kV and magnification up to 100,000×.The samples were fixed to the aluminum tables with a diameter of 10 mm and covered with the adhesive carbon film.

Structural Analysis
Figures 1-3 present the low-temperature nitrogen adsorption/desorption isotherms as well as the PSD curves for the silicas that were prepared during the addition of various dispersed silicas (aerosils or milled silica gels) in the sol stage (Figure 1) or as a result of mechanochemical (Figure 2) or hydrothermal (Figure 3) modifications.Table 1 includes the designation and structural parameters of the obtained materials.The course of analyzed curves indicated that each of the modifications affected the porous structure parameters of the materials.The shape of the initial material isotherms (S0, type 1 according to the International Union of Pure and Applied Chemistry (IUPAC) classification [31], a lack of hysteresis loop) with the addition of A-50 (S1) and A-380 (S2) in the sol stage caused appearance of hysteresis loops (loop H1) on the isotherms of the modified samples (type IV, Figure 1a), indicating the creation of cylindrical pores of differentiated radii dominating R dom .The addition of 30% fumed silica A-50 (S1) led to the formation of pores of R av = 1.85 nm (R dom ~2.3 nm and about 24 nm radii).However, the addition of 10% A-50 and 10% A-380 (S2) resulted in the formation of pores of R av = 1.25 nm (R dom ~1 and 3 nm (S2, Figure 1)).This differentiation was due to the differences in structure and porosity of the additive (A-50: S BET = 50 m 2 g −1 , particle diameter D g = 44 nm; A-380: S BET = 370 m 2 g −1 , D g = 6 nm).The addition of 20% SiO 2-MChT to the structure of the initial material in the sol stage (S3) resulted in the formation of a type of hysteresis loop with a complex profile, indicating the presence of pores of different sizes (Figure 1a inset).This was also reflected in the PSD curves calculated by SCR.The PSD curves indicated the presence of pores of the radii R dom ~2.1 and 5 nm (Figure 1b, Table 1).Intensive changes of the obtained silicas were clearly visible when comparing their structures with those of additives A-50 and A-380.This was particularly evident on the PSD V curves (Figure 1b inset).; %S meso = (S meso /S BET )*100%; S macro [m 2 g −1 ]; %S macro = (S macro /S BET )*100%; V micro [cm 3 g −1 ]; %V micro = (V micro /V p )*100%; V meso [cm 3 g −1 ]; %V meso = (V meso /V p )*100%; V macro [cm 3 g −1 ]; %V macro = (V macro /V p )*100%; V p [cm 3 g −1 ]; R av [nm]; R dom (N 2 ) [nm]; R dom (differential scanning calorimetry (DSC)) [nm]; ∆w-The deviation of the pore shape.Figure 2 presents the nitrogen adsorption/desorption isotherms as well as the PSD curves for the materials obtained with the differentiated silica additives (A-50 (S8), A-380 (S9), A-50 + A-380 mixture (S10), and SiO 2-MChT (S11)) in the gel stage followed by mechanochemical milling at 300 or 500 rpm.In all cases, intensive changes of the structure were observed compared to the initial material (S0) due to the dispersed silica addition.The addition of both A-50 (sample S8), A-380 (S9) and application of their mixture (A-50 + A-380, sample S10) caused the obtained materials to be characterized by type IV isotherms (Figure 2a) according to the IUPAC classification [31].However, their hysteresis loops had a complex shape, indicating the differentiated shapes of the pores contained in these materials.These were mostly spherical pores with numerous narrowings and open bottle-shaped terminations.The reason for this was a kind of treatment during which the mechanical deformation of structure could take place.A faintly developed maximum at R dom ~1.5 nm, 6 nm, and ~50 nm appeared on the PSD curve obtained for sample S8.The shapes of the other curves (S9 and S10) also indicated a complex, multimodal character of porous structure, with maxima at 1.5 and 12 nm (S9) and 1.5 and 4.2 nm (S10).The most differentiated silica material of multimodal structure was obtained during milling with the addition of SiO 2-MChT at 500 rpm (S11).The increased mill rotation caused the obtained silica to be characterized by the complex hysteresis loop (Figure 2a), and the PSD curve possessed three distinct maxima, indicating the presence of pores with the radii R dom ~0.8, 2.5, and 7.7 nm (Figure 2b, Table 1).
Figure 3 presents the isotherms and PSD curve obtained for the materials prepared using the differentiated silica additives (A-50 (S4), A-380 (S5), A-50 + A-380 mixture (S6), and SiO 2-MChT (S7)) in the sol stage followed by hydrothermal treatment at 200 • C. The shape of the analyzed isotherms indicated that this kind of modification had the most significant effect on the redevelopment of the silica surface structure.The obtained type IV isotherms (Figure 3a) were characterized by well-developed hysteresis loops, indicating the creation of mesoporous structure.The obtained materials possessed the pores of unified sizes about ~1 to ~6 nm.Only sample S7 possessed pores of R dom ~12 nm (Figure 3b).As can be seen from the analysis of the data in Table 1, all materials had a slightly developed structure of macropores.Although the maxima can be seen in the region R av > 50 nm on the PSDv curves (Figure 1b, Figure 2b, and Figure 3b), they are slightly developed.
The parameters included in Table 1 indicate that the kind of additives (A-50, S-380, SiO 2-MChT ) significantly affected the development of specific surface area S BET .The materials obtained with the addition of A-380 were characterized by the largest surface areas with S BET quantities ranging from ~380 to ~487 m 2 g −1 , whereas the lowest S BET values were observed for the silicas with the addition of wide-porous, ground SiO 2-MChT .However, the kind of modification also significantly affected the surface development and porous structure.The addition of the second component (A-50 or A-380) in the sol stage or using the mechanochemical modification enabled the preparation of micro/mesoporous structure materials.The contribution of both types of pores (S1 and S2) was ~50% (%S and %V).Using SiO 2-MChT as the additive led to preparation of materials with significant dominance of mesopores (76.8% − %S meso , 88.6% − %V meso , Table 1), regardless of the modification method.
Hydrothermal modification proved to be the most effective.The obtained materials (S4, S5, S6, and S7) were mainly mesoporous (%S meso ~72.4-96.6%,%V meso ~90.6-97.7%,Table 1) of radii (R p ) and total pore volume (V p ) larger than for the materials prepared by other methods (addition in the gel or MChT stage).This was due to the essence of hydrothermal modification during which dissolution of small silica molecules and their deposition on larger molecules take place.Such mass transport results in significant redevelopment of the structure [10,32,33].
Modifications of porous structure resulting in differentiation of pore shapes and sizes, as evidenced in broad bands on the PSD curves (Figure 1b, Figure 2b, Figure 3b), caused some deviation in the applied CV/SCR model.This was clearly seen as the values ∆w (Table 1) indicated a wide range of deviation from the assumed (cylindrical) model of pores.

Thermoporometry DSC
The analysis of thermograms of ice melting in the pores of studied materials pointed out the multistage course of this process.The characteristic temperatures of water freezing and ice melting in sample S1 determined from the thermograms DSC are presented in Table 2.As can be seen from the data in Table 2, the water in the S1 material froze up to −37.93 • C, although the freezing point of volumetric water is 0 • C. Such reduction in T onset was observed for all samples.The reduction in ice freezing temperature in relation to the standard temperature (0 • C) was the result of strong interactions of water with the surface of narrow pores of the studied materials.The multistage phase transformations indicated differentiated porosity of silica materials, which was also confirmed by the PSD V curves determined by cryoporometry DSC (PSD V-DSC ; Figure 4) obtained from the ice melting thermograms.Figure 4 presents the PSD V -DSC curves for the exemplary samples S2, S3, and S11.The clearly seen maxima on the observed curves were used for the determination of R dom-DSC (Table 1).The analysis of the obtained parameters proved the multimodal character of the obtained silica, and the values were in good agreement with those obtained from the N 2 adsorption data, particularly in the area of mesopores (Figure 1b, Figure 2b, Figure 3b).However, water freezing in the macropores (R > 50 nm) formed monolayers at a distance from the silica surface, which belonged to the clusters filling the pores in bulk water.Therefore, there were no peaks in the range R > 50 nm on the PSD V -DSC curves, which could be observed for the chosen PSD V curves for some materials (Figure 1b, Figure 2b, Figure 3b).
Table 3 presents the values of changes in water freezing and ice melting enthalpies in the studied materials.One can observe that the obtained values were much lower than the standard value of ice melting (333.7 J g −1 ).This was the result of strong interactions of water with the surface of silicas containing numerous hydroxyl groups.As can be seen, the highest values ∆H f (the closest to the standard value ∆H • ) were determined for the sample characterized by the pores of largest R av (S11).This confirmed the earlier observations about weak interactions of water molecules in wide pores owing to which they preserve properties of volumetric water to some extent.

Thermal Analysis
The investigated samples were subjected to thermogravimetric investigations under dynamic conditions to determine the content of physically adsorbed water and the amount of surface hydroxyl groups.In Figure 5, the exemplary thermograms of water desorption for the materials after HTT modification are presented.As one can see, the water desorption from the surface occurred in two main stages.The first step, up to 200 • C, was caused by dehydration (removal of the physically bound water).The second step, from 200 to 1200 • C, resulted from condensation of different types of silanols.The hydroxyl groups could be present on the surface (vicinal, germinal, or isolated hydroxyl ones) or inside the silica skeleton (intraglobular, structural hydroxyl groups).In the temperature range of 200-900 • C, intraglobular water was removed from the silica structure.
As can be seen from Figure 5a, the kind of additive affected the content of bound physical water (∆m in the temperature range of 20-200 • C).The smallest mass loss in this range was observed for the material with the addition of milled silica gel that was subjected to hydrothermal treatment (S7).This was associated with the smallest specific surface area with the simultaneous contribution of micropores (%S micro = 27.6%,Table 1) in which a strong retention of water molecules was observed.In the further range, the mass loss rate was comparable.Moreover, differentiation of preparation methods had a significant effect on the water content in the prepared materials.Figure 5b presents the dependence of water mass loss (TG%) on the way of modification using A-50 as the second component.Of all tested materials, the smallest amount of physically bound water was observed in the HTT-modified material.
Table 4 presents the quantitative results of determination of hydroxyl surface groups for the studied materials.As one can see, the materials obtained after the hydrothermal treatment were characterized by a relatively high content of isolated hydroxyl groups removed from the silica surface in the temperature range of 900-1200 • C (Table 4, S4, S5 and S6).From the analysis of the data, the total content of hydroxyl groups (σ OH OH nm −2 ) was close to the value for completely hydrated silica (σ OH = 4.6 − 4.9 OH nm −2 ) only for the hydrothermally modified silica [7,24,25].The values of the materials prepared in the sol stage and MChT processes were larger than 4.6 OH nm −2 , which was due to the fact that silica contains not only OH surface groups (geminal, vicinal, isolated) but also intraglobular water bound inside the skeleton.Its content depends on the way of modification.However, it is not considered as the form of hydroxyl surface groups.

Quasi-Isothermal Analysis
The prepared silicas were also subjected to thermal analysis under quasi-isothermal (Q-TG) conditions.In Figure 6, the Q-TG curves are presented for the exemplary samples.As one can see, the desorption of water from the surface of the studied materials proceeded in a wide temperature range in three stages, which is illustrated in detail in Figure 7.This figure represents the exemplary curve of surface free energy ∆G(C H2O ) changes in the whole range of tested water surface concentrations.At a large extent of hydration (Figure 7, C total ), evaporation of the water physically bound with the surface took place with the increasing temperature.Evacuation of water from the pores was not observed in this stage.Here the mass loss was small, being ~5-15% (Figure 6) of the initial concentration of absorbed water.
The area confined with the curve ∆G(C H2O ) between the points C total and C p determined the energy needed for evaporation of physically bound water.With further temperature increase (from the point C p , Figure 7), water desorption from the pores started.Removal of such type of water occurred over a relatively wide temperature range (from ~80 to ~160 • C, Figure 6).This was due to the fact that the boiling temperature of water in the pores increases with the decreasing pore diameter.Further increase in temperature was the result of the beginning of thermodesorption of a thin monolayer of water molecules bound directly with the adsorbent surface (Figure 7, A).Based on the data obtained by means of Q-TG, the dependences of the size of effective radius of the evaporating drop on the adsorbed water concentration (Figure 8a,b) and of dG/dM H2O =f(C H2O ) obtained for the low coverage of water for the tested adsorbents (Figure 8b) were determined.As proved earlier, all studied materials were characterized by the multimodal structure of pores, which was also reflected on the Q-TG curves.
The shapes of exemplary curves (S1, Figure 8a and S9, Figure 8b) indicate that in the initial stage of water thermodesorption (low temperature, large C H2O ), the effective radius of the evaporating drop was small.However, with the increase in temperature, it increased to the value determining the radius of the pores dominating R dom .A further increase in temperature resulted in reduction of the adsorbed water concentration and the size of effective radius of the evaporating drop until the next maximum was reached, which was the evidence that the materials possessed pores of differentiated sizes.The shape of the observed profiles depended on the R dom value.Such courses of curves were observed for all studied materials.Figure 8c,d presents the chosen dG/dM dependences in a function of statistical monolayers of water (n) under the conditions of low coverage with water, where dG/dM indicates the change of system free energy during water evaporation from the surface for S1 (Figure 8c) and S9 (Figure 8d) samples.It shows that on the curves obtained for the investigated silicas (also those not shown here) for which the degree of surface coverage was below 0.12 nm, there were multiple maxima.These maxima showed the formation of various types of numerous water clusters on the heterogeneous surfaces of silicas.

SEM
The SEM images confirmed the properties of the silicas under investigation.Figure 9 shows the changes in the surface morphology of the samples after the mechanochemical modification and the addition of A-50 (S8), A-380 (S9), A-50 + A-380 mixture (S10), and SiO 2-MChT (S11).
In Figure 9a, one can see single globules of A-50 concentrated into bigger structures due to the mechanochemical treatment.They are packed quite loosely.However, the addition of A-380 (Figure 9b) resulted in the obtained materials being more compact, and the individual globules were hardly visible.The differences were seen well because the preparation was made in the gel stage, and the additives (A-50 and A-380) were characterized by great differences in the globule diameter (particle sizes: 44 and 6 nm, respectively).The morphology of the S10 (Figure 9c) sample was different.Between the smooth patches (just like in S5, after the A-380 addition), there were clusters of tiny A-50 globules (as for S8, after the addition of A-50).The structure of this material was heterogeneous.The surface of the S11 (Figure 9d) sample seemed to be the most uniform.It was obvious that the MChT caused the particle size of the ground material to be unified.Figure 10 presents the images of sample surfaces after hydrothermal treatment.It is known that HTT causes dissolution of silica small globules and their deposition on the surface of greater ones, which can also be seen in Figure 10.After the A-50 addition (Figure 9a, particle size D g = 44 nm), the silica particles were larger than in the case of the A-380 addition (Figure 10b, particle size D g = 6 nm).Moreover, the particles were combined and had irregular shapes.However, the addition of initially milled silica (SiO 2-MChT ) caused the structure of the obtained silica to be relatively homogeneous.

Figure 4 .
Figure 4. PSD V curves determined by cryoporometry DSC for the exemplary silica samples.

Figure 5 .
Figure 5. Thermograms of water desorption under quasi-isothermal thermogravimetry (Q-TG) conditions for samples with different additives modified by hydrothermal modification (HTT) (a) and using A-50 (as the second component) prepared in different ways (b).

Figure 6 .
Figure 6.The course of mass changes based on the Q-TG analysis for the chosen samples.

Figure 7 .
Figure 7.The dependence of ∆G value changes on the concentration of water adsorbed on the S1 sample.

Figure 8 .
Figure 8. Dependences of the effective radius of the evaporating drop on the concentration of adsorbed water for (a,b) as well as the dG/dM dependences under the conditions of small filling of the surface with adsorbed water for the silicas S1 (a,c) and S9 (b,d).

Table 1 .
The designation and structural parameters of silicas.

Table 2 .
Characteristic temperatures of the stages of DSC cooling and heating of water registered for S1 sample.onset , T end -Temperatures at the beginning and the end of the phase transition in • C, respectively; T max /T min -Temperature at which the transformation occurs fastest, • C; ∆H-Changes of free enthalpy of phase transition in the cooling or heating stage, J g −1 . T

Table 3 .
Changes of free enthalpy of the phase transition during freezing and melting of the ice in the tested materials.

Table 4 .
Quantitative results of thermal analysis of the samples under investigation.