Luminescent Hybrid Material Based on Boron Organic Phosphor and Silica Aerogel Matrix

A new luminescent hybrid material based on silica aerogel and a boron-containing coordination compound with 8-hydroxyquinoline was created, and its physicochemical and spectral-luminescent characteristics were studied. A simple scheme for the synthesis of a hybrid luminescent material was developed. Simultaneously with the synthesis of the aerogel, the formation of a boron-containing phosphor was carried out using an isopropanol solution of boric acid and 8-hydroxyquinoline. Using in situ luminescent measurements, the mechanisms of the formation of boron-based luminescent complexes in isopropanol and tetrahydrofuran media were established. Both hydrophilic and hydrophobic silica aerogels were tested as matrices for the hybrid material. The formation of a thin layer of a boron-containing coordination luminescent compound on the highly developed surface of the SiO2 aerogel made it possible to strongly stabilize the aerogel structure and noticeably increase the thermal stability of the synthesized hybrid material.


Introduction
It is common knowledge that the most effective luminescence occurs on materials' surfaces. Therefore, a material with a high specific surface can work most efficiently as a luminescent material if the matrix is transparent for its own radiation. New luminescent materials with unique properties are of great interest for the development of novel devices, for example, lighting panels for buildings instead of common lighting devices. When combining the lighting properties with low thermal conductivity, one can obtain a unique material for an outer thermal insulation for the generation of new building constructions.
This future material could be made on the base of a hybrid material that includes an aerogel, such as an inorganic matrix and an organic phosphor as a lighting source. Such properties are fully possessed by SiO 2 aerogel.
Silica aerogel is a solid material with a low density from 0.003 to 0.3 g/cm 3 , a high specific surface area up to 1500 m 2 /g, a mean pore size from 10 to 20 nm [1], an extremely low thermal conductivity of about 0.014 W/mK [2], and a high optical transparency in the visible spectral range of about 90%. Due to the cell-net structure, a silica aerogel is an efficient light-scattering inorganic matrix, and its porosity could provide the encapsulation of a large amount of substances, including efficient photoluminophors. Since luminescence is a surface phenomenon, an aerogel with a high specific surface area is an ideal substrate for phosphor formation and bonding at the molecular level to the aerogel structure through covalent bonds or van der Waal bonds. This is the general concept of an aerogel-based luminescent hybrid material.

Synthesis of BLCs and Photoluminescence Measurements in Solutions
We used two solvents to conduct the synthesis of BLC, namely, IPA and tetrahydrofuran (TGF). The dissolution of boric acid (H 3 BO 3 ) and 8-Hq in the media was carried out at a fixed temperature (45, 50, or 60 • C) in quartz-glass vessels with a preliminary heated solvent in an ultrasonic bath (1.3/2 TTC 35 kHz, Sapphire Ltd., Moscow, Russia) for 60 min. Then, 0.007 g of H 3 BO 3 and 0.044 g of 8-Hq were dissolved in 0.5 mL of IPA or TGF for different experiments. Then, as-prepared solutions were transferred into a quartz-glass cuvette placed into a 1 cm thermostabilized cuvette holder of an Ocean Optics spectrophotometer. The temperature was supported with 0.1 C accuracy by a BT 12-2 liquid thermostat (TERMEX Ltd., Moscow, Russia). Photoluminescent (PL) spectra were recorded at 365 nm excitation using a QE65000 spectrophotometer (Ocean Optics, Delray Beach, FL, USA). The PL spectral signal was collected for 20 s. Then, we had an exposure for a certain time at a fixed temperature and repeated the PL spectral recording.

One-Step Heterophase Synthesis of BLCs
A heterophase synthesis of solid BLC was conducted using extra-dried boron oxide wires and 8-Hq as raw components. Crystalline P 2 O 5 was placed outside of the reaction zone and was used as a water-absorbent agent. The general heterophase reaction

One-Step Heterophase Synthesis of BLCs
A heterophase synthesis of solid BLC was conducted using extra-dried boron oxide wires and 8-Hq as raw components. Crystalline P2O5 was placed outside of the reaction zone and was used as a water-absorbent agent. The general heterophase reaction could be presented by the formal chemical reaction (1) of tris(8-hydroxyquiniline) boron (Bq3) synthesis: (1) According to [10], the structure of amorphous boron oxide is represented by boron in the triple coordination (flat triangles), which is polymerized into [BO2]3n boroxol rings. The proportion of boron polymerized in boroxol rings exceeds 70%. However, one should take into account that the exact B-base product could differ from Bq3 because the reaction proceeded on the surface of B2O3 wires and stopped because of diffusion limitations.
The apparatus scheme for the direct synthesis was described in detail in Ref. [11].

Hydrophilic Boron-Containing Aerogel Synthesis
The process and apparatus for silica gel fabrication was described in detail in Ref. [3]. At the first stage of the fabrication of a boron-containing luminescent aerogel, the gel was placed in a four-fold volume of a solution of H3BO3 in IPA and left on a shaker for 5-7 h. The procedure for introducing the boron-containing component was repeated three times. The obtained gel was placed in a high-pressure apparatus. Then, the 8-Hq powder preparation was added inside the apparatus in the amount that doubled the stoichiometric amount according to reaction (2). H3BO3 s + 3·(8-Hq) solute = Bq3 s + 3·H2O solute (2) The apparatus was sealed, heated to 40° C, and supplied with CO2 up to 120 bar. According to the study of the process of the formation of boron-containing luminescent preparations in IPA solution (see Section 3.1), it was established that we did not need more than 30 min to finish the synthesis reaction in solution. This means that the rate-limiting stage of the BLC formation process inside the gel is the diffusion process. It was found that exposure for 1 h was enough to complete the reaction of BLC synthesis in the gel volume. After that, the supercritical drying of the gels was carried out, and a boron-containing luminescent aerogel (BoronLightSil) was obtained.
The total boron content in BoronLightSil samples, according to an ICP-MS analysis (Table S1, Section 2.4.3), turned out to be directly proportional to the boric acid concentration in the IPA solution during the gel impregnation procedure (Figure 1). This indicates that all boron oxide introduced into the aerogel was successfully absorbed in the aerogel porous structure and formed the hybrid material. (1) According to [10], the structure of amorphous boron oxide is represented by boron in the triple coordination (flat triangles), which is polymerized into [BO 2 ] 3n boroxol rings. The proportion of boron polymerized in boroxol rings exceeds 70%. However, one should take into account that the exact B-base product could differ from Bq 3 because the reaction proceeded on the surface of B 2 O 3 wires and stopped because of diffusion limitations.
The apparatus scheme for the direct synthesis was described in detail in Ref. [11].

Hydrophilic Boron-Containing Aerogel Synthesis
The process and apparatus for silica gel fabrication was described in detail in Ref. [3]. At the first stage of the fabrication of a boron-containing luminescent aerogel, the gel was placed in a four-fold volume of a solution of H 3 BO 3 in IPA and left on a shaker for 5-7 h. The procedure for introducing the boron-containing component was repeated three times. The obtained gel was placed in a high-pressure apparatus. Then, the 8-Hq powder preparation was added inside the apparatus in the amount that doubled the stoichiometric amount according to reaction (2).
The apparatus was sealed, heated to 40 • C, and supplied with CO 2 up to 120 bar. According to the study of the process of the formation of boron-containing luminescent preparations in IPA solution (see Section 3.1), it was established that we did not need more than 30 min to finish the synthesis reaction in solution. This means that the rate-limiting stage of the BLC formation process inside the gel is the diffusion process. It was found that exposure for 1 h was enough to complete the reaction of BLC synthesis in the gel volume. After that, the supercritical drying of the gels was carried out, and a boron-containing luminescent aerogel (BoronLightSil) was obtained.
The total boron content in BoronLightSil samples, according to an ICP-MS analysis (Table S1, Section 2.4.3), turned out to be directly proportional to the boric acid concentration in the IPA solution during the gel impregnation procedure (Figure 1). This indicates that all boron oxide introduced into the aerogel was successfully absorbed in the aerogel porous structure and formed the hybrid material.

Hydrophobic Boron-Containing Aerogel Synthesis
To obtain hydrophobic silica gels, the surfaces of previously obtained gels were subjected to modification. As a result of the modification, hydroxyl groups on the surfaces were replaced by nonpolar groups (alkyl groups) that block the adsorption of water and impart hydrophobic properties to the material. TMCS was used as a hydrophobizing agent. Since TMCS can chemically react with IPA, which had been previously used as a solvent, for the first reaction we changed the solvent to hexane in gel processing by placing the gels in four times the volume of IPA-hexane solutions with different concentrations of hexane (from 30 to 100 wt%) for 3-5 h in each solution. After changing the solvent, the gels were placed in a four times volume of a certain concentration (X wt%) of a TMCS-hexane solution and kept at 60 • C for 24 h in an oven.

Hydrophobic Boron-Containing Aerogel Synthesis
To obtain hydrophobic silica gels, the surfaces of previously obtained ge jected to modification. As a result of the modification, hydroxyl groups on were replaced by nonpolar groups (alkyl groups) that block the adsorption impart hydrophobic properties to the material. TMCS was used as a hyd agent. Since TMCS can chemically react with IPA, which had been previou solvent, for the first reaction we changed the solvent to hexane in gel processin the gels in four times the volume of IPA-hexane solutions with different conc hexane (from 30 to 100 wt%) for 3-5 h in each solution. After changing the gels were placed in a four times volume of a certain concentration (X wt%) hexane solution and kept at 60 °C for 24 h in an oven.
After hydrophobization, the samples were washed in IPA to remov chemicals. The introduction of the boron-containing component and boron c thesis via supercritical drying were carried out as described in Section 2.3.1. Table 1 shows the notation of the hydrophilic and hydrophobic BoronLig and synthesis conditions. An analysis of the IR spectra showed that the hydrophobization result After hydrophobization, the samples were washed in IPA to remove unreacted chemicals. The introduction of the boron-containing component and boron complex synthesis via supercritical drying were carried out as described in Section 2.3.1. Table 1 shows the notation of the hydrophilic and hydrophobic BoronLightSil samples and synthesis conditions. An analysis of the IR spectra showed that the hydrophobization resulted in a sufficient decrease in OH group content ( Figure 2).  BoronLightSil samples were prepared in the form of cylindrical monoliths. The samples were tested to establish the macroscopic and structural characteristics, i.e., the bulk density (ρ bulk ) with a standard deviation of 0.005 g/cm 3 ; the linear shrinkage (L) with the standard deviation <0.2%; the specific surface area (S BET ); the total pore volume (V total ), and the BJH pore volume (V BJH ). The techniques for characteristic determination were described in detail in Ref. [3]. The specific surface area (S BET ) was determined by the Brunauer-Emmett-Teller (BET) method. The pore size distribution, mean pore diameter (D), and pore volume (VBJH) were determined by the Barrett-Joyner-Halenda (BJH) method. An AccuPyc 1340 instrument (Micromeritics Instrument Corp., Norcross, GA, USA) was used to measure the skeletal density (ρsk) of BoronLightSil by helium pycnometry.

Luminescence Measurements
All spectra, except those for solutions (see Section 3.1), were recorded at room temperature.
A Fluorolog-3 FL3-22 spectrophotometer (Horiba Jobin Yvon, Portland, OR, USA) with double grating excitation and emission monochromators was used for the photoluminescence (PL) measurements of solid samples and aerogels in the wavelength range of 375-725 nm with 0.1 nm steps under the excitation of a Xenon 450W Ushio UXL-450S/O lamp.
The PL spectra and decay kinetics were treated by OriginPro 8 SR4 software (OriginLab Corp., Northampton, MA, USA), taking into account the original HORIBA software for instrumental corrections [12]. Photoluminescence quantum yield was determined using a G8 integration sphere (GMP SA, Renens, Switzerland).
The calculations of the chromaticity coordinates of the luminescence, according to the CIE RGB colorimetric system of HM films, were carried out using the addition curves.

ICP-MS Analysis
To analyze the chemical purity of the initial preparations and synthesized products, we used inductively coupled plasma mass spectrometry with the preliminary transfer of a solid sample to the liquid phase. For dissolving preparations, we used extra pure water (Aqua-Max-Ultra 370 Series, Young Lin Instruments Co. Ltd., Gyeonggi-do, 14042, Korea) with a specific resistance of 18 MΩ·cm, high-purity nitric (HNO 3 ) acid (7N7) purified by a Berghoff BSB-939-IR surface distillation system (BERGHOFF GmbH&Co., Wenden, Germany), and high-purity sulfuric acid (H 2 SO 4 ) (8N Ultrapur, Sigma Aldrich, Burlington, MA, USA) in a SPEEDWAVE-FOUR microwave decomposition system (Berghof GmbH, Koenigsee, Germany) equipped with DAP-100 PTFE autoclaves (BERGHOFF GmbH&Co.).
Analytical measurements were carried out on a NexION 300D inductively coupled plasma mass spectrometer (ICP-MS) (PerkinElmer Inc., Boston, MA, USA). We used the TotalQuant method [13] for the determination of 65 chemical elements' concentrations. The standard solutions (PerkinElmer Inc.) were used for calibration.

TEM Analysis
Micrographs of the samples were obtained using JEM 2100F and JEM 2100M electron transmission microscopes (maximum accelerating electron voltage-200 kV, chromatic and spherical aberration correctors, and X-ray energy-dispersive analyzer). The sample preparation for electron microscopy consisted of the deposition of small BLC strips (thickness less than 100 µm) on copper and carbon grids, without the deposition of additional materials. Moderate sample drift was observed only at maximum magnifications (2 M or more).

Boron-Based Luminescent Complexes
When we studied the solution synthesis of BLC. We found that in 30 min the maxima of the PL wavelengths (λ max PL ) reached constant values, which indicated that the synthesis reaction of BLC came to an end ( Figure 3). transmission microscopes (maximum accelerating electron voltage-200 kV, chromatic and spherical aberration correctors, and X-ray energy-dispersive analyzer). The sample preparation for electron microscopy consisted of the deposition of small BLC strips (thickness less than 100 µ m) on copper and carbon grids, without the deposition of additional materials. Moderate sample drift was observed only at maximum magnifications (2 M or more).

Boron-Based Luminescent Complexes
When we studied the solution synthesis of BLC. We found that in 30 min the maxima of the PL wavelengths ( ) reached constant values, which indicated that the synthesis reaction of BLC came to an end ( Figure 3). The most interesting result was obtained for the analysis of the PL spectra recorded at different reaction temperatures. It turned out that during the reaction two different boron-based luminescent complexes were formed. The complexes differed in the value. The transition from one temperature to another led to a redistribution of the concentrations of these complexes, as evidenced by the redistribution of the intensities of the PL bands for . A decomposition of the PL spectra in energy coordinates ( Figure 4) made it possible to determine that one complex (type I) was characterized by = 535 nm, while the second complex (type II) had = 427 nm. The most interesting result was obtained for the analysis of the PL spectra recorded at different reaction temperatures. It turned out that during the reaction two different boronbased luminescent complexes were formed. The complexes differed in the λ max PL value. The transition from one temperature to another led to a redistribution of the concentrations of these complexes, as evidenced by the redistribution of the intensities of the PL bands for λ max PL . A decomposition of the PL spectra in energy coordinates ( Figure 4) made it possible to determine that one complex (type I) was characterized by λ max PL = 535 nm, while the second complex (type II) had λ max PL = 427 nm.  It is well-known that 8-Hq complexes with Al, Ga, and In have a number of polymorphic modifications that are characterized by differences in structure, and their values differ by a maximum of 50 nm [14][15][16]. In our case, the difference in was almost 100 nm, which can hardly be explained by the formation of different polymorphic modifications. Taking into account the complex polymer formations based on boron oxide groups in the liquid, it can be assumed that at different temperatures different degrees of coordination of the boron complex with 8-Hq arise. It is known that a decrease in the mass of the coordination metal and the amount of ligands for metal complexes with 8-Hq generally leads to a decrease in [17]. Thus, in the row Alq3Caq2Liq for structurally It is well-known that 8-Hq complexes with Al, Ga, and In have a number of polymorphic modifications that are characterized by differences in structure, and their λ max PL values differ by a maximum of 50 nm [14][15][16]. In our case, the difference in λ max PL was almost 100 nm, which can hardly be explained by the formation of different polymorphic modifications. Taking into account the complex polymer formations based on boron oxide groups in the liquid, it can be assumed that at different temperatures different degrees of coordination of the boron complex with 8-Hq arise. It is known that a decrease in the mass of the coordination metal and the amount of ligands for metal complexes with 8-Hq generally leads to a decrease in λ max PL [17]. Thus, in the row Alq 3 →Caq 2 →Liq for structurally similar polymorphic modifications, a regular shift of λ max PL is seen as 520 nm→465 nm→442 nm, respectively. Thus, it can be assumed that boron in the row B→Al→Ga→In, having the smallest molecular weight, forms complexes characterized by the smallest values of λ max PL . In this case, the coordination of one ligand leads to the generation of luminescence with the shortest λ max PL . We supposed that this effect was observed in our experiments. When using tetrahydrofuran as a solvent, taking into account the much better solubility of both boric acid and 8-Hq in TGF over the entire temperature range under study from 30 to 60 • C, the maximum PL intensity was almost instantly obtained, with a constant maximum of λ max PL = 535 ± 2 nm ( Figure 5). Taking into account that the boron oxide groups in TGF are less polymerized, it can be assumed that, in this case, complexes similar in structure to Bq 3 are formed.
It is well-known that 8-Hq complexes with Al, Ga, and In have a number of polymorphic modifications that are characterized by differences in structure, and their values differ by a maximum of 50 nm [14][15][16]. In our case, the difference in was almost 100 nm, which can hardly be explained by the formation of different polymorphic modifications. Taking into account the complex polymer formations based on boron oxide groups in the liquid, it can be assumed that at different temperatures different degrees of coordination of the boron complex with 8-Hq arise. It is known that a decrease in the mass of the coordination metal and the amount of ligands for metal complexes with 8-Hq generally leads to a decrease in [17]. Thus, in the row Alq3Caq2Liq for structurally similar polymorphic modifications, a regular shift of is seen as 520 nm465 nm442 nm, respectively. Thus, it can be assumed that boron in the row BAlGaIn, having the smallest molecular weight, forms complexes characterized by the smallest values of . In this case, the coordination of one ligand leads to the generation of luminescence with the shortest . We supposed that this effect was observed in our experiments.
When using tetrahydrofuran as a solvent, taking into account the much better solubility of both boric acid and 8-Hq in TGF over the entire temperature range under study from 30 to 60 °C, the maximum PL intensity was almost instantly obtained, with a constant maximum of = 535 ± 2 nm ( Figure 5). Taking into account that the boron oxide groups in TGF are less polymerized, it can be assumed that, in this case, complexes similar in structure to Bq3 are formed.  To exclude a solvent intercalation in a BLC structure, which could take place during solution synthesis, we analyzed the BLC preparations obtained by the heterophase reaction (1). Similar to IPA solution synthesis, we obtained two preparations that strongly differed in PL intensity and λ max PL ( Figure 6). To exclude a solvent intercalation in a BLC structure, which could take place during solution synthesis, we analyzed the BLC preparations obtained by the heterophase reaction (1). Similar to IPA solution synthesis, we obtained two preparations that strongly differed in PL intensity and ( Figure 6). Since the entire preparation was practically in the isothermal zone, we assumed that the luminescence color was influenced by the 8-Hq vapor flow and the outgoing water vapor. The preparation located closer to the 8-Hq source showed a turquoise glow (B-Hq-Dir(493)) of PQY 55%, while the preparation located far from the 8-Hq source glowed blue (B-Hq-Dir(459)) at PQY 41%.
An analysis of the PL decay kinetics showed that both preparations were characterized by the presence of two luminescence centers with different lifetimes (Figure 7, Table  2). As follows from the fundamental regularities, the complex with = 459 nm was Since the entire preparation was practically in the isothermal zone, we assumed that the luminescence color was influenced by the 8-Hq vapor flow and the outgoing water vapor. The preparation located closer to the 8-Hq source showed a turquoise glow (B-Hq- Dir(493)) of PQY 55%, while the preparation located far from the 8-Hq source glowed blue (B-Hq-Dir(459)) at PQY 41%.
An analysis of the PL decay kinetics showed that both preparations were characterized by the presence of two luminescence centers with different lifetimes (Figure 7, Table 2). As follows from the fundamental regularities, the complex with λ max PL = 459 nm was characterized by a faster short-lived component (2.16 ns), while the complex with λ max PL = 494 nm had a slower short-lived component (6.42 ns). For the long-lived luminescence centers, the lifetime values were close (32.3 ns vs. 29.6 ns). Since the entire preparation was practically in the isothermal zone, we assumed that the luminescence color was influenced by the 8-Hq vapor flow and the outgoing water vapor. The preparation located closer to the 8-Hq source showed a turquoise glow (B-Hq-Dir(493)) of PQY 55%, while the preparation located far from the 8-Hq source glowed blue (B-Hq-Dir(459)) at PQY 41%.
An analysis of the PL decay kinetics showed that both preparations were characterized by the presence of two luminescence centers with different lifetimes (Figure 7, Table  2). As follows from the fundamental regularities, the complex with = 459 nm was characterized by a faster short-lived component (2.16 ns), while the complex with = 494 nm had a slower short-lived component (6.42 ns). For the long-lived luminescence centers, the lifetime values were close (32.3 ns vs. 29.6 ns).    In the case of boron oxide wires, their surface is smooth enough, with separate pyramids ( Figure 8a). BLC could form on a smooth surface and on the sharp edges of wires or pyramids. Thus, we obtained two possible PL centers with different lifetimes. In the case of boron oxide wires, their surface is smooth enough, with separate pyramids ( Figure 8a). BLC could form on a smooth surface and on the sharp edges of wires or pyramids. Thus, we obtained two possible PL centers with different lifetimes. The probable reaction of BLC synthesis on a smooth surface could be presented by the scheme in Figure 9, resulting in a monomeric coordination compound (B-Hq-Dir(459)). The probable reaction of BLC synthesis on a smooth surface could be presented by the scheme in Figure 9, resulting in a monomeric coordination compound (B-Hq-Dir(459)). The probable reaction of BLC synthesis on a smooth surface could be presented by the scheme in Figure 9, resulting in a monomeric coordination compound (B-Hq-Dir(459)). On the sharp edges of ledges that existed on a smooth wire surface (Figure 8b), in accordance with steric factors, several 8-Hq molecules could interact with boron, resulting in a dimeric coordination compound with a brighter PL (B-Hq-Dir(493)) ( Figure 10).   The probable reaction of BLC synthesis on a smooth surface could be presented by the scheme in Figure 9, resulting in a monomeric coordination compound (B-Hq-Dir(459)).   BLC stability was proven by cathodoluminescence (CL) measurements of a B-Hq-Dir(493) preparation. For the first time, we detected CL for BLC based on 8-hydroxyquinoline ( Figure 11). The CL was detected in blue and green channels, but the red channel was empty (Figures S11-S14). We observed CL only on the grains. The thin layers of BLC on the smooth surfaces of B 2 O 3 wires were probably quickly sputtered by the electron beam at 20 kV.
Molecules 2022, 27, x FOR PEER REVIEW 10 of 17 BLC stability was proven by cathodoluminescence (CL) measurements of a B-Hq-Dir(493) preparation. For the first time, we detected CL for BLC based on 8-hydroxyquinoline ( Figure 11). The CL was detected in blue and green channels, but the red channel was empty (Figures S11-S14). We observed CL only on the grains. The thin layers of BLC on the smooth surfaces of B2O3 wires were probably quickly sputtered by the electron beam at 20 kV. The TEM analysis showed the presence of many different objects in BLC samples obtained by heterophase synthesis, which could be geometrically divided into four types: (1) clusters with a shell ( Figure S1); (2) clusters without a shell (Figures 12, 13, S1, and S2); (3) nanotubes (Figures 14 and S1); and (4) crystalline planes, including partially twisted ones ( Figure S3). It is most likely that nanotubes are formed from twisted planes. This means that we observed nanotubes at different stages of their formation.
Unfortunately, we failed to detect structures containing 8-hydroxyquiniline, whose crystal lattice parameter is about 6-8 Å in coordination compounds [15]. This could be explained by very the small thickness of BLC, about several monolayers. These thin structures could be easily vaporized by the e-beam of TEM, similar to the process we observed when we analyzed tris(8-hydroxyquiniline) aluminum crystals by SEM. We observed the The TEM analysis showed the presence of many different objects in BLC samples obtained by heterophase synthesis, which could be geometrically divided into four types: (1) clusters with a shell ( Figure S1); (2) clusters without a shell (Figures 12, 13, S1 and S2); (3) nanotubes (Figures 14 and S1); and (4) crystalline planes, including partially twisted ones ( Figure S3). It is most likely that nanotubes are formed from twisted planes. This means that we observed nanotubes at different stages of their formation. lattice form a rhombus with an angle of 60 degrees ( Figure 13) and clusters with a shell (dot sizes of 8-12 nm and core sizes of 4-6 nm, Figure S2). 4. Nanotubes are formed only by a crystal lattice with a parameter of 3.2 Å (Figure 14).
The dimensions of the tubes are small; the length range is from 20 to 50 nm, with diameters of 4 to 15 nm. We also observed multilayer tubes (up to 16 layers, with 2.4 A between layers).   lattice form a rhombus with an angle of 60 degrees ( Figure 13) and clusters with a shell (dot sizes of 8-12 nm and core sizes of 4-6 nm, Figure S2). 4. Nanotubes are formed only by a crystal lattice with a parameter of 3.2 Å (Figure 14).
The dimensions of the tubes are small; the length range is from 20 to 50 nm, with diameters of 4 to 15 nm. We also observed multilayer tubes (up to 16 layers, with 2.4 A between layers).

BoronLightSil
An analysis of the physicochemical properties and structural parameters of silica aerogels and BoronLightSil samples (Table 3) showed that the bulk density (ρbulk) of silica aerogels was half that of the hydrophilic BoronLightSil samples, while the skeletal densities Unfortunately, we failed to detect structures containing 8-hydroxyquiniline, whose crystal lattice parameter is about 6-8 Å in coordination compounds [15]. This could be explained by very the small thickness of BLC, about several monolayers. These thin structures could be easily vaporized by the e-beam of TEM, similar to the process we observed when we analyzed tris(8-hydroxyquiniline) aluminum crystals by SEM. We observed the crystal vaporization under e-beam action in real time for several seconds.
The observed crystal lattices can be classified into two types: 1. Lattice parameter 3.2 ± 0.5 Å. Taking into account the limited number of elements in the reaction components (H, C, N, O, and B), we searched for appropriate crystalline structures and came to the conclusion that the only suitable phase with a 3.2 ± 0.5 Å lattice parameter could be C 3 N 4 [18]. There are several polymorphs of C 3 N 4 that meet the search lattice parameter. The small clusters of C 3 N 4 could be formed as a result of 8-Hq decay at high temperatures and crystal lattice compression. This assumption needs further investigation, but it could be very interesting to synthesize C 3 N 4 in such soft conditions. 2.
The main conclusions after image processing: 1.
Single-crystal planes are mainly formed by a lattice with a lattice parameter of 3.2 Å. The fraction of planes with a lattice of 4.2 Å is very small and their size is an order of magnitude smaller than that of the former.

2.
The same 3.2 Å lattice forms nanosized clusters with diameters of up to 20 nm of a rather spherical shape ( Figure 12).

3.
A crystal lattice with a parameter of 4.2 Å also forms clusters, including two types of small irregularly shaped clusters up to 10 nm in size: four neighboring atoms in this lattice form a rhombus with an angle of 60 degrees ( Figure 13) and clusters with a shell (dot sizes of 8-12 nm and core sizes of 4-6 nm, Figure S2).

4.
Nanotubes are formed only by a crystal lattice with a parameter of 3.2 Å (Figure 14). The dimensions of the tubes are small; the length range is from 20 to 50 nm, with diameters of 4 to 15 nm. We also observed multilayer tubes (up to 16 layers, with 2.4 A between layers).

BoronLightSil
An analysis of the physicochemical properties and structural parameters of silica aerogels and BoronLightSil samples (Table 3) showed that the bulk density (ρ bulk ) of silica aerogels was half that of the hydrophilic BoronLightSil samples, while the skeletal densities (ρ sk ) differed only by 15%. Eventually, the total pore volume of the silica aerogel was twice as large as in the BoronLightSil samples due to the pores filling with boron oxide and BLCs. Another situation was observed for hydrophobic BoronLightSil samples. The bulk densities for BoronLightSil were two times higher than in hydrophobic silica aerogel. The skeletal densities were nearly the same as for the hydrophilic BoronLightSil samples and just a bit lower the hydrophobic silica aerogel. The total pore volume was bigger than for the hydrophilic samples for both the silica aerogel and BoronLightSil samples.
The PL spectra of hydrophilic BoronLightSil samples ( Figure 15) could be characterized as simple and attributed to a single chemical compound, contrary to the solution synthesis of BLC (see Section 3.1). The λ max PL of the hydrophilic BoronLightSil samples was 513 ± 5 nm (Table 4) and was slightly shifted to the red wavelength region on boron concentration increase.
having a size of 5-7 Å , are formed at hydrophobic ends and also pene of aerogels and fill them. When interacting with 8-hydroxyquinoline, tion (1) results in the formation of organo-element monomeric [6], d more complex clusters, which have sizes of about 2-4 nm. These comp cated on hydrophobic ends and inside pores. However, if the pores are boron oxy groups, then the reaction of BLC formation proceeds efficien the pores and does not propagate deep inside due to steric limitation groups inside pores are not involved in the formation of luminescent reduce the luminescence intensity, which we observed experimentally In the case of hydrophobic BoronLightSil samples, we observed th one order higher ( Figure 16, Table 4). The values were shorter compared to the hydrophilic samples, but were very close to those of sized BLC (see Table 2).   An analysis of the PL spectra of the BoronLightSil samples showed that an increase in the total boron concentration in the hydrophilic BoronLightSil samples first led to an increase in the PL intensity ( Figure 15) and then to a noticeable decrease.
This character of dependence can be explained by the fact that boron oxo groups, having a size of 5-7 Å, are formed at hydrophobic ends and also penetrate into the pores of aerogels and fill them. When interacting with 8-hydroxyquinoline, the exchange reaction (1) results in the formation of organo-element monomeric [6], dimeric [7], or even more complex clusters, which have sizes of about 2-4 nm. These complexes are easily located on hydrophobic ends and inside pores. However, if the pores are heavily filled with boron oxy groups, then the reaction of BLC formation proceeds efficiently at the exit from the pores and does not propagate deep inside due to steric limitations. Thus, boron oxy groups inside pores are not involved in the formation of luminescent complexes and can reduce the luminescence intensity, which we observed experimentally ( Figure 15).
In the case of hydrophobic BoronLightSil samples, we observed the PL intensity to be one order higher ( Figure 16, Table 4). The λ max PL values were shorter (494 and 498 nm) compared to the hydrophilic samples, but were very close to those of the direct-synthesized BLC (see Table 2). Figure 15. PL spectra of hydrophilic BoronLightSil samples with different concentra during gel synthesis: (1) 0.01 wt% for BLS-0.01-HL; (2) 0.05 wt% for BLS-0.05-HL; ( BLS-0.10-HL (λ ext = 360 nm).
In the case of hydrophobic BoronLightSil samples, we observed the PL in one order higher ( Figure 16, Table 4). The values were shorter (494 compared to the hydrophilic samples, but were very close to those of the d sized BLC (see Table 2). This could indicate that the synthesis reactions of BLC in hydrophobic very similar to those that occurred in the direct synthesis at high temperat 2.2). This could indicate that the synthesis reactions of BLC in hydrophobic aerogels are very similar to those that occurred in the direct synthesis at high temperatures (Section 2.2).
To understand the nature of BLC formation, we analyzed the PL decay kinetics of BoronLightSil samples (Figure 17, for details see Figures S6-S10). According to an analysis, for all BoronLightSil samples the PL decay kinetics were successfully described by three exponents (Table 4, Figures S6-S10), contrary to the solid direct-synthesized BLC samples (Table 2), where we found two luminescent centers. For most of the BoronLightSill samples, the short-lived centers had lifetimes of about 2.4 ± 0.1 ns. The exception was the BLS-0.05-HB-5 sample, in which hydrophobization was more complete due to the high concentration of TMCS. The middle-lived centers had lifetimes of about 11 ± 2 ns for all samples, and the long-lived centers for hydrophilic BoronLightSil had lifetimes of 168 ± 10 ns, while for hydrophobic BoronLightSil they had shorter lifetimes of 74-148 ns.  According to an analysis, for all BoronLightSil samples the PL decay kinetics were successfully described by three exponents (Table 4, Figures S6-S10), contrary to the solid direct-synthesized BLC samples (Table 2), where we found two luminescent centers. For most of the BoronLightSill samples, the short-lived centers had lifetimes of about 2.4 ± 0.1 ns. The exception was the BLS-0.05-HB-5 sample, in which hydrophobization was more complete due to the high concentration of TMCS. The middle-lived centers had lifetimes of about 11 ± 2 ns for all samples, and the long-lived centers for hydrophilic BoronLightSil had lifetimes of 168 ± 10 ns, while for hydrophobic BoronLightSil they had shorter lifetimes of 74-148 ns.
The QY of hydrophobic BoronLightSil was compared with BLC samples, while for hydrophilic BoronLightSil the QY was lower at times. We explained the obtained results due to the known poor resistance of 8-hydroxyquinolne-based coordination compounds under humidity exposure [20]. The presence of OH groups in hydrophilic BoronLightSil decreased the PL intensity by more than one order of magnitude, and the QY reduced correspondingly.
We assume that the observed results could be explained by the differences in structures of boron oxide wires and silica aerogels. In the case of BoronLightSil, its aerogel matrix has a much more complex structure with bridges and pores of different sizes (Figure 8b). We could identify at least three different locations for BLC formation, namely, smooth bridge wires, pore exits, and pore walls. Therefore, the BLC configuration at these sites is different, resulting in different lifetimes. We assumed that the shortest lifetime would be for BLC located on a smooth surface and that the BLC located on the walls of the pores would have the longest lifetime.
The hydrophobization of BoronLightSil aerogels made it possible to significantly increase the stability of luminescent characteristics. In the case of a hydrophilic BoronLightSil sample, the loss of PL intensity was almost 100% within 6 months of exposure to a natural environment (Figure 18a), while for the hydrophobic BoronLightSil, the decrease in the PL intensity after the same 6 months of exposure was only 15% (Figure 18b). It should be notice that a LightSil hybrid based on Alq 3 [3] had a PQY of 45% (measured in this research), but it decayed in several days under natural environmental conditions.

R REVIEW 15 of 17
for BLC located on a smooth surface and that the BLC located on the walls of the pores would have the longest lifetime. The hydrophobization of BoronLightSil aerogels made it possible to significantly increase the stability of luminescent characteristics. In the case of a hydrophilic BoronLightSil sample, the loss of PL intensity was almost 100% within 6 months of exposure to a natural environment (Figure 18a), while for the hydrophobic BoronLightSil, the decrease in the PL intensity after the same 6 months of exposure was only 15% (Figure 18b). It should be notice that a LightSil hybrid based on Alq3 [3] had a PQY of 45% (measured in this research), but it decayed in several days under natural environmental conditions. An analysis of the thermostability of BoronLightSil was conducted by the step-by-step heating of a sample placed into a thermostabilized quartz-glass cuvette and in situ PL measurements. Ten minutes of exposure was enough to reach a constant PL intensity value at a given temperature. We observed a continuous decrease in PL intensity during a temperature increase ( Figure 19). However, up to 90 °C the aerogel structure of the Bo-ronLightSil sample was preserved, while the ordinary SiO2 aerogel decayed at 80 °C. Taking into consideration the thermal stability of Mq3 (M = Al, Ga, and In) compounds [14], we assume that a thin layer of BLC on a highly developed SiO2 aerogel surface strongly stabilized the aerogel structure and noticeably increased its thermal stability. An analysis of the thermostability of BoronLightSil was conducted by the step-by-step heating of a sample placed into a thermostabilized quartz-glass cuvette and in situ PL measurements. Ten minutes of exposure was enough to reach a constant PL intensity value at a given temperature. We observed a continuous decrease in PL intensity during a temperature increase ( Figure 19). However, up to 90 • C the aerogel structure of the BoronLightSil sample was preserved, while the ordinary SiO 2 aerogel decayed at 80 • C. Taking into consideration the thermal stability of Mq 3 (M = Al, Ga, and In) compounds [14], we assume that a thin layer of BLC on a highly developed SiO 2 aerogel surface strongly stabilized the aerogel structure and noticeably increased its thermal stability.
value at a given temperature. We observed a continuous decrease in PL intensity during a temperature increase ( Figure 19). However, up to 90 °C the aerogel structure of the Bo-ronLightSil sample was preserved, while the ordinary SiO2 aerogel decayed at 80 °C. Taking into consideration the thermal stability of Mq3 (M = Al, Ga, and In) compounds [14], we assume that a thin layer of BLC on a highly developed SiO2 aerogel surface strongly stabilized the aerogel structure and noticeably increased its thermal stability.

Conclusions
A new luminescent hybrid material based on silica aerogel and a boron-containing coordination compound with 8-hydroxyquinoline, named BoronLightSil, was created. We developed a simple scheme for the synthesis of a hybrid boron-based luminescent material that combined the processing of hydrophobic silica aerogel fabrication by supercritical drying in CO 2 flow at 120 atm and the simultaneous synthesis of a boron-based phosphor in situ by the chemical reaction of boric acid with 8-hydroxyquinilinol in isopropanol. The study of the chemical reaction in isopropanol and tetrahydrofuran media showed that the nature of the luminescent complexes based on boron with 8-hydroxyquinoline is very complex. At least two basic products could be formed, depending on the reaction conditions.
It was found that the hydrophobization of silica aerogels by a treatment in n-hexane resulted in a BoronLightSil material with stable luminescent properties and higher thermostability.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/molecules27165226/s1, Table S1: Concentration of impurities in aerogel determined by ICP-MS analysis; Figure S1: TEM image of Object 3. Shell clusters. The lattice parameter for Object 3 is 4.2 A. For Object 4, the parameter is not defined. Object diameter 3-12 nm (core-6 nm), Object diameter 4-8 nm (core-4 nm); Figure S2: TEM image of Object 5 is a nanotube. Length-25 nm, diameter-5 nm; Figure S3: TEM image of Object 6-several crystalline planes, partially twisted at the ends. The lattice parameter for all objects in the image is 3.2 A; Figure S4: PL decay kinetics of B-Hq-Dir(459) sample obtained by direct synthesis; Figure S5: PL decay kinetics of B-Hq-Dir(493) sample obtained by direct synthesis; Figure S6: PL decay kinetics of BLS-0.01-HL sample; Figure S7: PL decay kinetics of BLS-0.05-HL sample; Figure S8: PL decay kinetics of BLS-0.10-HL sample; Figure S9: PL decay kinetics of BLS-0.05-HB-3 sample; Figure S10: PL decay kinetics of BLS-0.05-HB-3 sample; Figure S11. SEM image of BLC preparation fabricated by direct synthesis at 20 kV accelerating voltage; Figure S12: Cathodoluminescence (blue channel) of BLC preparation fabricated by direct synthesis at 20 kV accelerating voltage; Figure S13: Cathodoluminescence (green channel) of BLC preparation fabricated by direct synthesis at 20 kV accelerating voltage; Figure S14