Peculiarities of Synthesis and Properties of Lignin–Silica Nanocomposites Prepared by Sol-Gel Method

The development of advanced hybrid materials based on polymers from biorenewable sources and mineral nanoparticles is currently of high importance. In this paper, we applied softwood kraft lignins for the synthesis of lignin/SiO2 nanostructured composites. We described the peculiarities of composites formation in the sol-gel process through the incorporation of the lignin into a silica network during the hydrolysis of tetraethoxysilane (TEOS). The initial activation of lignins was achieved by means of a Mannich reaction with 3-aminopropyltriethoxysilane (APTES). In the study, we present a detailed investigation of the physicochemical characteristics of initial kraft lignins and modified lignins on each step of the synthesis. Thus, 2D-NMR, 31P-NMR, size-exclusion chromatography (SEC) and dynamic light scattering (DLS) were applied to analyze the characteristics of pristine lignins and lignins in dioxan:water solutions. X-Ray photoelectron spectroscopy (XPS) and Fourier transform infrared (FTIR) were used to confirm the formation of the lignin–silica network and characterize the surface and bulk structures of the obtained hybrids. Termogravimetric analysis (TGA) in nitrogen and air atmosphere were applied to a detailed investigation of the thermal properties of pristine lignins and lignins on each step of modification. SEM confirmed the nanostructure of the obtained composites. As was demonstrated, the activation of lignin is crucial for the sol-gel formation of a silica network in order to create novel hybrid materials from lignins and alkoxysilanes (e.g., TEOS). It was concluded that the structure of the lignin had an impact on its reactivity during the activation reaction, and consequently affected the properties of the final hybrid materials.


Introduction
The development of advanced materials from biorenewable sources is currently of high importance. Lignin, an aromatic biological macromolecule, has great potential for replacing oil-based polymers. As a cross-linked natural phenolic polymer, lignin is of great interest because of its valuable properties and abundance as a by-product in the pulp and paper industries [1]. The chemical The molecular weight of the lignin samples was characterized by size-exclusion chromatography using an SEC 1260 Infinity instrument (Polymer Standard Services, Mainz, Germany) coupled to a dual system detector (UV, RI). DMSO + 0.5% LiBr (w/w) was used as the mobile phase. The separation system consisted of a PSS GRAM Precolumn, PSS GRAM 100 Å, and PSS GRAM 10,000 Å analytical columns thermostated at 60 • C and connected in series. Pullulan standards with Mw within the range 708 kDa, 337 kDa, 194 kDa, 47.1 kDa, 21.1 kDa, 9.6 kDa, 6.1 kDa, 1.081 kDa, and 342 Da were used for the standard calibration. 31

P-NMR
The content of functional groups in lignin samples was measured by 31 P NMR [47]. Approximately 20-30 mg of lignin samples was weighed and phosphitylated using 2-chloro-4,4,5,5tetramethyl-1,3,2-dioxaphospholane. Endo-N-hydroxy-5-norbornene-2,3-dicarboximide (e-HNDI) (Sigma Aldrich, 40 mg/mL) and chromium (III) acetylacetonate (Aldrich, 5 mg/mL) were used as an internal standard and a relaxation reagent, respectively. The derivatized sample was dissolved in CDCl 3 prior to analysis. The 31 P NMR experiment was performed with a 90 • pulse angle, inverse gated proton decoupling, and a delay time of 10 s. For analysis, 256 scans with a time delay of 6 s and a total runtime of 34 minutes were collected. Measurements were performed in duplicate.

2D-Heteronuclear Single Quantum Coherence (2D-HSQC) NMR
Approximately 100 mg of sample was acetylated for better solubility [48]. The residue was dissolved in 700 µL of DMSO-d6 . The two-dimensional (2D) HSQC NMR spectrum was acquired using the Bruker, Billerica, MA, USA pulse program 'hsqcetgpsi' a relaxation delay of 1.7 s, a coupling constant of 145 Hz, an INEPT transfer delay time of 1.72 ms (d4 = 1/4 J), a spectral window of 10.5 ppm in F2, and 166 ppm in F1 with 1024 × 512 increments, 240 scans per increment, and a spectral center set at 90.0 ppm in F1 and 5.3 ppm in F2. The two-dimensional (2D) NMR dataset was processed with 2 K × 1 K data points using a π/2-shifted sine-bell window function in both dimensions. Central DMSO (δ C /δ H = 39.5/2.5 ppm) was used as an internal reference according to the adopted solvent. The particle size of the LignoBoost lignin and CleanFlowBlack lignin was measured using a dynamic light scattering meter (Zetasizer, Nano ZS, Malvern Instruments, Malvern, England). For the measurements, solutions of lignin with a concentration of 1 g·L −1 were used. A dioxane/water mixture with a volume ratio of 4:1 was used as the solvent.
FTIR Spectroscopy IR spectra were collected using a Perkin-Elmer spectrophotometer (Spotlight 400 FTIR imaging system, Waltham, MA, USA) equipped with a Spectac MKII Golden Gate system (Creekstone Ridge, GA, USA). The samples were analyzed in the range of 600-4000 cm −1 with 16 scans at a four cm −1 resolution and a one cm −1 interval at room temperature.

X-Ray Photoelectron Spectroscopy (XPS)
XPS spectra were recorded on an ESCA apparatus with a multidetection electron analyzer (Scienta R4000, produced by VG Scienta, East Sussex, UK) in fixed analyzer transmission mode.

Thermal Analysis
Thermal analysis was carried out on a TGA/DSC 1 (Mettler Toledo, Columbus, OH, USA) instrument under the following operational conditions: a heating rate of 10 • C min −1 , a dynamic atmosphere of synthetic air or nitrogen (50 mL·min −1 ), a temperature range of 30-900 • C, and a sample mass of~2.5 mg.

Scanning Electron Microscopy (SEM)
The structural characteristics of the fabricated samples were studied with a field-emission scanning electron microscope (FE-SEM, S-4800, Hitachi, Japan). The SEM images were obtained using a secondary electron detector. The samples were coated with a 1-nm thick Pt-Pd layer sputtered with a Cressington (Watford, UK), 208HR high-resolution coater.

Characterization of the Technical Lignins
The molecular weights of LBL and CFBL, measured by the SEC method, were approximately 5600 and 3000 Da, respectively ( Table 1). The polydispersity index (PDI) was higher for the LBL lignin sample than for the CFB (see Table 1). The molecular weight distribution curves are shown in Figure 1. For both lignin samples, one major peak can be observed in chromatogram; however, a shift toward a higher molecular weight can be clearly seen for the LBL lignin. The various types of interunit linkages for lignin samples were investigated and semiquantified by the 2D-HSQC NMR method ( Figure 2). The peak assignment was performed based on the works by Capanema et al. Del Río et al. Kim and Ralph, Zhang and Gellerstedt [49][50][51][52]. The signal from the phenolic carbon-2 was used as an internal standard to quantify different linkages. As illustrated in the 2D NMR spectra of LBL and CFBL (Figure 2a,b), both lignins contain the typical for lignin signals from the β-aryl ether (β-O-4), pinoresinol (β-β), phenylcoumarin (β-5), with the coniferyl alcohol structures and the methoxy groups in similar proportions (Figure 2c).

Characterization of the Technical Lignins
The molecular weights of LBL and CFBL, measured by the SEC method, were approximately 5600 and 3000 Da, respectively ( Table 1). The polydispersity index (PDI) was higher for the LBL lignin sample than for the CFB (see Table 1). The molecular weight distribution curves are shown in Figure  1. For both lignin samples, one major peak can be observed in chromatogram; however, a shift toward a higher molecular weight can be clearly seen for the LBL lignin.   The various types of interunit linkages for lignin samples were investigated and semiquantified by the 2D-HSQC NMR method ( Figure 2). The peak assignment was performed based on the works by Capanema et al, Del Río et al, Kim and Ralph, Zhang and Gellerstedt [49][50][51][52]. The signal from the phenolic carbon-2 was used as an internal standard to quantify different linkages. As illustrated in the 2D NMR spectra of LBL and CFBL (Figure 2a The 31 P NMR method was used to quantify the amount of various functional groups in both lignin samples (Table 1, Figure S1). Table 1 shows the total content of functional phenolic and hydroxyl groups in both lignin samples. CFBL had a higher content of phenolic and aliphatic hydroxyl groups than LBL. As reported in our previous work, low Mw fractions of lignin are usually enriched in phenolic compounds [10], indicating the possibility of higher reactivity for such lignin samples. Based on the quantification, the amount of total non-condensed structures for CFBL lignin with a lower molecular weight, which was obtained by the ultrafiltration of black liquor, was 21%, and that for LBL, which was obtained by the sequential precipitation of black liquor and had a higher molecular weight, was 16%. CFBL had also a higher ratio of non-condensed to condensed aromatic units than LBL (1.59 versus 1.12), which indicates the presence of a less condensed structure and The 31 P NMR method was used to quantify the amount of various functional groups in both lignin samples (Table 1, Figure S1). Table 1 shows the total content of functional phenolic and hydroxyl groups in both lignin samples. CFBL had a higher content of phenolic and aliphatic hydroxyl groups than LBL. As reported in our previous work, low Mw fractions of lignin are usually enriched in phenolic compounds [10], indicating the possibility of higher reactivity for such lignin samples. Based on the quantification, the amount of total non-condensed structures for CFBL lignin with a lower molecular weight, which was obtained by the ultrafiltration of black liquor, was 21%, and that for LBL, which was obtained by the sequential precipitation of black liquor and had a higher molecular weight, was 16%. CFBL had also a higher ratio of non-condensed to condensed aromatic units than LBL (1.59 versus 1.12), which indicates the presence of a less condensed structure and possibly the higher reactivity of the CFBL lignin.

Dynamic Light Scattering
Dynamic light scattering was applied to investigate the size distribution of lignin macromolecules in a dioxane/water mixture with a ratio of 4:1, as this solvent was used for the synthesis. The plots of the Z-average size distributions by number of the LBL and CFBL macromolecules are presented in Figure 3. According to the Z-average data, the hydrodynamic diameter of the lignin's particles (or micelles) range in the chosen solvent mixture was 1.3-12 nm for LBL and 0.7-2.4 nm for CFBL, which is in good agreement with the existing literature data [53,54].

Dynamic Light Scattering
Dynamic light scattering was applied to investigate the size distribution of lignin macromolecules in a dioxane/water mixture with a ratio of 4:1, as this solvent was used for the synthesis. The plots of the Z-average size distributions by number of the LBL and CFBL macromolecules are presented in Figure 3. According to the Z-average data, the hydrodynamic diameter of the lignin's particles (or micelles) range in the chosen solvent mixture was 1.3-12 nm for LBL and 0.7-2.4 nm for CFBL, which is in good agreement with the existing literature data [53,54].
In organic solvents, lignin has a swollen, extended structure. Due to their hydrophobic nature, lignin molecules fold when in contact with water. In our opinion, the immobilization (distribution) of lignin molecules in a silica network would prevent the folding of the molecules upon contact with water; thus, better accessibility of the functional groups can be achieved.

Design and Synthesis of Sorbent
The sol-gel method has proven to be useful for generating polymer-silica hybrid materials. The methods of modification of a silica surface with phenol-containing compounds by the Mannich reaction was developed earlier in the works of Tertykh et al. and Yanovska et al. [55,56]. Conducting the in situ formation of silica in the presence of a solution containing macromolecules facilitates the formation of hybrid materials with covalent bonds between the organic and inorganic phases [42,57,58]. However, the sol-gel reaction requires the addition of water (Scheme 1).

Scheme 1.
General scheme for the synthesis of silica by the sol-gel reaction.
Incorporating pristine (nonactivated) lignin into the sol-gel process resulted in a heterogeneous mixture of silica and lignin, where lignin was easily removed by washing with dioxane. To improve the reactivity of the lignin samples, which are not capable of undergoing a polycondensation reaction with TEOS, the lignin was activated by an aminomethylation (Mannich) reaction with APTES. Due to the presence of active protons in the lignin structure, formaldehyde forms a methylene bridge between APTES and lignin (Scheme 2).  In organic solvents, lignin has a swollen, extended structure. Due to their hydrophobic nature, lignin molecules fold when in contact with water. In our opinion, the immobilization (distribution) of lignin molecules in a silica network would prevent the folding of the molecules upon contact with water; thus, better accessibility of the functional groups can be achieved.

Design and Synthesis of Sorbent
The sol-gel method has proven to be useful for generating polymer-silica hybrid materials. The methods of modification of a silica surface with phenol-containing compounds by the Mannich reaction was developed earlier in the works of Tertykh et al. and Yanovska et al. [55,56]. Conducting the in situ formation of silica in the presence of a solution containing macromolecules facilitates the formation of hybrid materials with covalent bonds between the organic and inorganic phases [42,57,58]. However, the sol-gel reaction requires the addition of water (Scheme 1).

Dynamic Light Scattering
Dynamic light scattering was applied to investigate the size distribution of lignin macromolecules in a dioxane/water mixture with a ratio of 4:1, as this solvent was used for the synthesis. The plots of the Z-average size distributions by number of the LBL and CFBL macromolecules are presented in Figure 3. According to the Z-average data, the hydrodynamic diameter of the lignin's particles (or micelles) range in the chosen solvent mixture was 1.3-12 nm for LBL and 0.7-2.4 nm for CFBL, which is in good agreement with the existing literature data [53,54].
In organic solvents, lignin has a swollen, extended structure. Due to their hydrophobic nature, lignin molecules fold when in contact with water. In our opinion, the immobilization (distribution) of lignin molecules in a silica network would prevent the folding of the molecules upon contact with water; thus, better accessibility of the functional groups can be achieved.

Design and Synthesis of Sorbent
The sol-gel method has proven to be useful for generating polymer-silica hybrid materials. The methods of modification of a silica surface with phenol-containing compounds by the Mannich reaction was developed earlier in the works of Tertykh et al. and Yanovska et al. [55,56]. Conducting the in situ formation of silica in the presence of a solution containing macromolecules facilitates the formation of hybrid materials with covalent bonds between the organic and inorganic phases [42,57,58]. However, the sol-gel reaction requires the addition of water (Scheme 1).

Scheme 1.
General scheme for the synthesis of silica by the sol-gel reaction.
Incorporating pristine (nonactivated) lignin into the sol-gel process resulted in a heterogeneous mixture of silica and lignin, where lignin was easily removed by washing with dioxane. To improve the reactivity of the lignin samples, which are not capable of undergoing a polycondensation reaction with TEOS, the lignin was activated by an aminomethylation (Mannich) reaction with APTES. Due to the presence of active protons in the lignin structure, formaldehyde forms a methylene bridge between APTES and lignin (Scheme 2). Incorporating pristine (nonactivated) lignin into the sol-gel process resulted in a heterogeneous mixture of silica and lignin, where lignin was easily removed by washing with dioxane. To improve the reactivity of the lignin samples, which are not capable of undergoing a polycondensation reaction with TEOS, the lignin was activated by an aminomethylation (Mannich) reaction with APTES. Due to the presence of active protons in the lignin structure, formaldehyde forms a methylene bridge between APTES and lignin (Scheme 2). The modification of lignin by the Mannich reaction is a particular type of amination process where typically one dialkylaminomethyl group is introduced at the ortho position of a free phenolic hydroxyl group through a reaction with formaldehyde and amine [59]. According to the study of Du et al. [60], there is the primary product, the so-called Mannich condensation product, where the nitrogen-containing group is introduced at the C5 position of the benzene ring; however, some amount of side reactions could take place [61]. To eliminate the risk of involving side products in the second stage of the synthesis of hybrid material, the product of the Mannich reaction was precipitated and washed by diethyl ether. In the Supplementary Materials ( Figure S2), the FTIR spectra of lignin before and after washing by diethyl ether are presented. It can be seen that a characteristic band at 1738 cm −1 , which corresponds to the presence of side products (preferably imines), completely disappeared after washing with ether. Thus, activated lignin washed with diethyl ether was cohydrolyzed with TEOS in an acidic medium (Scheme 3). According to the well-established sol-gel reaction, a silica network was formed.

FTIR Analysis
To confirm the formation of silica as part of the hybrid lignin-silica composites, the FTIR spectra of the initial and modified lignins and synthesized composites were compared (Figure 4a The modification of lignin by the Mannich reaction is a particular type of amination process where typically one dialkylaminomethyl group is introduced at the ortho position of a free phenolic hydroxyl group through a reaction with formaldehyde and amine [59]. According to the study of Du et al. [60], there is the primary product, the so-called Mannich condensation product, where the nitrogen-containing group is introduced at the C 5 position of the benzene ring; however, some amount of side reactions could take place [61]. To eliminate the risk of involving side products in the second stage of the synthesis of hybrid material, the product of the Mannich reaction was precipitated and washed by diethyl ether. In the Supplementary Materials ( Figure S2), the FTIR spectra of lignin before and after washing by diethyl ether are presented. It can be seen that a characteristic band at 1738 cm −1 , which corresponds to the presence of side products (preferably imines), completely disappeared after washing with ether. Thus, activated lignin washed with diethyl ether was cohydrolyzed with TEOS in an acidic medium (Scheme 3). According to the well-established sol-gel reaction, a silica network was formed. The modification of lignin by the Mannich reaction is a particular type of amination process where typically one dialkylaminomethyl group is introduced at the ortho position of a free phenolic hydroxyl group through a reaction with formaldehyde and amine [59]. According to the study of Du et al. [60], there is the primary product, the so-called Mannich condensation product, where the nitrogen-containing group is introduced at the C5 position of the benzene ring; however, some amount of side reactions could take place [61]. To eliminate the risk of involving side products in the second stage of the synthesis of hybrid material, the product of the Mannich reaction was precipitated and washed by diethyl ether. In the Supplementary Materials ( Figure S2), the FTIR spectra of lignin before and after washing by diethyl ether are presented. It can be seen that a characteristic band at 1738 cm −1 , which corresponds to the presence of side products (preferably imines), completely disappeared after washing with ether. Thus, activated lignin washed with diethyl ether was cohydrolyzed with TEOS in an acidic medium (Scheme 3). According to the well-established sol-gel reaction, a silica network was formed.

FTIR Analysis
To confirm the formation of silica as part of the hybrid lignin-silica composites, the FTIR spectra of the initial and modified lignins and synthesized composites were compared (Figure 4a

FTIR Analysis
To confirm the formation of silica as part of the hybrid lignin-silica composites, the FTIR spectra of the initial and modified lignins and synthesized composites were compared (Figure 4a

XPS Analysis
XPS was used to study the hybrid lignin-silica composite surface structure. This method is based on changes in the electronic states of the elements. Moreover, XPS analysis could be used to confirm the successful aminomethylation of lignin and the formation of a silica network during the sol-gel reaction.
The original and deconvoluted XPS spectra of CFBL modified by the Mannich reaction (CFBL-Mannich), and the corresponding lignin-silica (CFBL-Silica) hybrid are presented in Figure 5a

XPS Analysis
XPS was used to study the hybrid lignin-silica composite surface structure. This method is based on changes in the electronic states of the elements. Moreover, XPS analysis could be used to confirm the successful aminomethylation of lignin and the formation of a silica network during the sol-gel reaction.
The original and deconvoluted XPS spectra of CFBL modified by the Mannich reaction (CFBL-Mannich), and the corresponding lignin-silica (CFBL-Silica) hybrid are presented in Figure 5a,b, respectively. The N1s spectra of the CFBL-Mannich and CFBL-silica samples are represented by three nonequivalent nitrogen atoms. N1s with binding energy E bind = 399.5 eV can be assigned to the secondary amino groups of APTES radical (−CH 2 NHCH 2 ) as well as to the −C−N− bond between APTES and lignin residues after the Mannich reaction. Nitrogen with a binding energy E bind = 400.7 eV could be related to the nitrogen in the −C=N bond after an APTES amino group reaction with formaldehyde (possible azomethine bond formation). The N with E bind = 402.1 eV corresponds to the quaternary N, which was possibly obtained after amino group protonation (R 2 NH + ). Compared with a CFBL-Mannich composite, the ratio between the −C−N− and −C=N bonds in CFBL-silica changes in favor of the −C−N− of the CFBL-silica composite. The N/Si ratio for the CFBL-Mannich composite is 0.614 and decreased up to 0.307 for the CFBL-silica composite.

XPS Analysis
XPS was used to study the hybrid lignin-silica composite surface structure. This method is based on changes in the electronic states of the elements. Moreover, XPS analysis could be used to confirm the successful aminomethylation of lignin and the formation of a silica network during the sol-gel reaction.
The original and deconvoluted XPS spectra of CFBL modified by the Mannich reaction (CFBL-Mannich), and the corresponding lignin-silica (CFBL-Silica) hybrid are presented in Figure 5a,b, respectively. The N1s spectra of the CFBL-Mannich and CFBL-silica samples are represented by three nonequivalent nitrogen atoms. N1s with binding energy Ebind = 399.5 eV can be assigned to the secondary amino groups of APTES radical (−CH2NHCH2) as well as to the −C−N− bond between APTES and lignin residues after the Mannich reaction. Nitrogen with a binding energy Ebind = 400.7 eV could be related to the nitrogen in the −C=N bond after an APTES amino group reaction with formaldehyde (possible azomethine bond formation). The N with Ebind = 402.1 eV corresponds to the quaternary N, which was possibly obtained after amino group protonation (R2NH + ). Compared with a CFBL-Mannich composite, the ratio between the −C−N− and −C=N bonds in CFBL-silica changes in favor of the −C−N− of the CFBL-silica composite. The N/Si ratio for the CFBL-Mannich composite is 0.614 and decreased up to 0.307 for the CFBL-silica composite.  (Table 2).
According to the atomic concentrations (%) of Si and C atoms in the activated lignin (CFBL-Mannich sample), one atom of Si is incorporated per 12 atoms of C (C:Si = 9:0.75). This result led us to conclude that almost every C9 phenolic unit of lignin reacted with an aminosilane (APTES) molecule. In the case of the CFBL-silica hybrid, the ratio of carbon to silicon changes to 6:1.

Morphology and Textural Characteristics
The morphologies of the initial lignins and obtained hybrid materials were characterized via SEM analysis. Figures 6-8   As can be observed from the Si2p XPS spectra presented in Figure 5c, there are two types of silicon atoms in the CFBL-Mannich composite. The Si2p 3/2 band with a binding energy of 102.3 eV can be assigned to the silicon in siloxane bonds as well as the unreacted Si(OEt) 3 residue of APTES or TEOS. After the addition of TEOS to the CFBL-Mannich composite (Figure 5d), a third, nonequivalent Si electronic state at E bind = 103.3 eV can be observed. This state can be related to the SiO 2 being formed after the introduction of TEOS and polycondensation ( Table 2). According to the atomic concentrations (%) of Si and C atoms in the activated lignin (CFBL-Mannich sample), one atom of Si is incorporated per 12 atoms of C (C:Si = 9:0.75). This result led us to conclude that almost every C9 phenolic unit of lignin reacted with an aminosilane (APTES) molecule. In the case of the CFBL-silica hybrid, the ratio of carbon to silicon changes to 6:1.

Morphology and Textural Characteristics
The morphologies of the initial lignins and obtained hybrid materials were characterized via SEM analysis. Figures 6-8 present the structures of the original and silica-containing lignin (Figure 6a-f, respectively), CFBL lignin (Figure 7a-f, respectively), and reference silica particles (Figure 8).       As seen from Figure 6a-c and 7a-c, the structures of the original lignins are rather dense, and most of the nanoparticles are in close contact with each other. After modification Figure 6d-f and 7d-f), the morphologies of the lignin-silica hybrid composites become rougher and more friable, which is somewhat typical for hybrid materials obtained by the sol-gel method. This observation presumably confirms the creation of a silica network between the particles of lignin. Additionally, the formation of a rough and irregular surface with plenty of holes, canals, and cavities most likely causes a desirable increase in the specific surface area of the fabricated hybrid materials. The nanosized structure of the hybrids' surface could be observed in Figures 6 and 7.  Figure 7. SEM images of the original CleanFlowBlack lignin (a-c) and the CleanFlowBlack ligninsilica hybrid (d-f) (high-resolution topographic contrast secondary electron imaging, parameters: accelerating voltage = 1 kV; working distance: (a-c) 1.4 mm, (d-f) 1.7 mm; magnification: (a,d) ×10,000, (b,e) ×50,000, (c,f) ×150,000).  According to the data from the nitrogen adsorption/desorption isotherms presented in Budnyak et al. [46], the LBL-silica and CFBL-silica composites have a Brunauer-Emmett-Teller (BET) specific surface areas of 74 m 2 ·g −1 and 92 m 2 ·g −1 , respectively, and belong to mesoporous materials with some amount of micropores. A detailed investigation of the textural characteristics of lignin-silica hybrids and initial polymers is presented in Budnyak et al. [46]. By comparing the data for the original lignins with those of the hybrid composites, it is evident that the creation of a silica network in the presence of lignin results in an increase in porosity, which causes an increment of the specific surface area of the obtained materials. However, the specific surface area and porosity of the obtained hybrids are relatively low in comparison with those of the inorganic materials or hybrids with a high content of the inorganic component, [42,58,62] which is in agreement with the data obtained by other research groups [20,63,64].

Thermal Analysis
TGA was performed to characterize the thermal properties of the original lignins and synthesized hybrids. The destruction processes of the initial LBL and CFBL lignins and the hybrid composites under a N 2 atmosphere are presented in Figure 9. The obtained TG and DTG curves have similar thermal behaviors; however, there are some significant differences in the thermal characteristics of the two types of original technical lignins compared to the immobilized ones. At the first stage of thermal decomposition of the initial LBL and CFBL, gradual moisture evaporation and water evaporation due to self-condensation reactions (peaks at 51 • C and 184 • C, 62 • C and 150 • C, respectively) occur. The maximum temperatures for the conversion of phenols into pyrocatechols [65] and the conversion of short substituents of the benzene rings [66] were found at 390 • C and 345 • C for the original LBL and CFBL, respectively. The rearrangement of the backbone and carbonization of the studied lignins takes place at temperatures above 400 • C. As seen from figures 6a-c and 7a-c, the structures of the original lignins are rather dense, and most of the nanoparticles are in close contact with each other. After modification (figures 6d-f and 7d-f), the morphologies of the lignin-silica hybrid composites become rougher and more friable, which is somewhat typical for hybrid materials obtained by the sol-gel method. This observation presumably confirms the creation of a silica network between the particles of lignin. Additionally, the formation of a rough and irregular surface with plenty of holes, canals, and cavities most likely causes a desirable increase in the specific surface area of the fabricated hybrid materials. The nanosized structure of the hybrids' surface could be observed in figures 6f and 7f.
According to the data from the nitrogen adsorption/desorption isotherms presented in Budnyak et al [46], the LBL-silica and CFBL-silica composites have a Brunauer-Emmett-Teller (BET) specific surface areas of 74 m 2 ·g −1 and 92 m 2 ·g −1 , respectively, and belong to mesoporous materials with some amount of micropores. A detailed investigation of the textural characteristics of lignin-silica hybrids and initial polymers is presented in Budnyak et al [46]. By comparing the data for the original lignins with those of the hybrid composites, it is evident that the creation of a silica network in the presence of lignin results in an increase in porosity, which causes an increment of the specific surface area of the obtained materials. However, the specific surface area and porosity of the obtained hybrids are relatively low in comparison with those of the inorganic materials or hybrids with a high content of the inorganic component, [42,58,62] which is in agreement with the data obtained by other research groups [20,63,64].

Thermal Analysis
TGA was performed to characterize the thermal properties of the original lignins and synthesized hybrids. The destruction processes of the initial LBL and CFBL lignins and the hybrid composites under a N2 atmosphere are presented in Figure 9. The obtained TG and DTG curves have similar thermal behaviors; however, there are some significant differences in the thermal characteristics of the two types of original technical lignins compared to the immobilized ones. At the first stage of thermal decomposition of the initial LBL and CFBL, gradual moisture evaporation and water evaporation due to self-condensation reactions (peaks at 51 °C and 184 °C, 62 °C and 150 °C, respectively) occur. The maximum temperatures for the conversion of phenols into pyrocatechols [65] and the conversion of short substituents of the benzene rings [66] were found at 390 °C and 345 °C for the original LBL and CFBL, respectively. The rearrangement of the backbone and carbonization of the studied lignins takes place at temperatures above 400 °C. It was found that the maximum temperature of the polymer destruction decreased after each stage of lignin modification. Thus, for activated lignins by Mannich reaction (samples LBL-Mannich and CFBL-Mannich), the maximum for condensation and elimination of the -OH groups and conversion of phenols into pyrocatechols occurs at 338 °C and 333 °C, respectively. In the case of lignin-silica composites, the maximum for this processes was found at 290 °C for both samples. The maximum temperature of the conversion of short substituents of the benzene rings, the rearrangement of the backbone, and carbonization of LBL-Mannich and CFBL-Mannich was detected at 464 °C and 422 °C, respectively, and for LBL-based and CFBL-based composites, the maximum shifted to 416 °C and 414 °C. The concentrations of moisture in both lignin-silica composites are below 1% (Figure 9b). The total weight loss for the LBL-silica and CFBL-silica composites was 45% and 46%, respectively, and for the original LBL and CFBL samples, it was 61% and 59%, respectively. The main destruction regions are presented in Table 3.  It was found that the maximum temperature of the polymer destruction decreased after each stage of lignin modification. Thus, for activated lignins by Mannich reaction (samples LBL-Mannich and CFBL-Mannich), the maximum for condensation and elimination of the -OH groups and conversion of phenols into pyrocatechols occurs at 338 • C and 333 • C, respectively. In the case of lignin-silica composites, the maximum for this processes was found at 290 • C for both samples. The maximum temperature of the conversion of short substituents of the benzene rings, the rearrangement of the backbone, and carbonization of LBL-Mannich and CFBL-Mannich was detected at 464 • C and 422 • C, respectively, and for LBL-based and CFBL-based composites, the maximum shifted to 416 • C and 414 • C. The concentrations of moisture in both lignin-silica composites are below 1% (Figure 9b). The total weight loss for the LBL-silica and CFBL-silica composites was 45% and 46%, respectively, and for the original LBL and CFBL samples, it was 61% and 59%, respectively. The main destruction regions are presented in Table 3.   The characteristics of the thermal decomposition of studied materials in O 2 atmosphere are presented in Table S1 (Supporting Information). TG and DTG curves of thermal decomposition in N 2 and O 2 atmosphere for initial and modified lignins, hybrids, and silica presented in Figures S4 and S5 of the Supporting Information, respectively.

Application as Sorbzents in Water Treatment
Synthesized hybrid materials were found to be effective as adsorbents for organic pollutants from aqueous solutions. Figure S6 presents a comparison of the adsorption capacity of the hybrids, initial lignins, and silica toward methylene blue dye. The interesting observations are that the adsorption activity of the lignins is improved, and that the difference in the lignin structure affects the performance of the lignin to a more substantial degree as a result of immobilization on a silica surface. Our detailed study of the adsorption activity of synthesized materials for the removal of methylene blue dye from aqueous solutions was presented in previous work [46]. In the study, the effects of parameters such as the contact time, initial concentration of dye, and initial pH on the adsorption capacity was evaluated. Thus, the lignin-silica composites were proved to be potential materials for application in wastewater treatment.

Conclusions
The proposed method for lignin modification using APTES according to the Mannich reaction was found to be an effective way to activate kraft lignin. As demonstrated in the current study, the activation of lignin is necessary for the sol-gel formation of a silica network to create novel hybrid materials from lignins and alkoxysilanes (e.g., TEOS). According to the XPS data, the carbon:silicon ratio in the activated lignin was found to be 12:1 (9:0.75), which led us to conclude that almost every C9 phenolic unit of lignin reacted with APTES. In the case of the CFBL-silica hybrid, the ratio of carbon to silicon changed to 6:1. A comparison of the FTIR spectra of the initial and modified lignins and the synthesized composites confirmed the creation of a silica network as a part of the hybrid lignin-silica composites.
The proposed method allowed us to obtain porous hybrid materials with an increased surface area and a somewhat rough and friable structure with plenty of holes, canals, and cavities. It was found that a silica network was created between the lignin particles. This structure was confirmed by SEM analysis, where it was observed that in the case of the hybrid materials, the nanoparticles of lignin were better separated from each other. The fact that the modification of lignin with alkoxysilanes (APTES and TEOS) promotes the formation of cross-linked lignin particles interconnected by an inorganic matrix was also confirmed by the textural characteristics.
It was concluded that the structure of the lignin had an impact on its reactivity during the activation reaction, and consequently affected the properties of the final hybrid materials. Thus, the thermal stability of the immobilized lignin increased in comparison to the original lignins: the maximum destruction temperature of the original CFBL and LBL samples shifted from 345 • C to 414 • C and from 390 • C to 416 • C, respectively, after their immobilization in the hybrid composites. It was demonstrated that the synthesized hybrid materials could be used as effective sorbents for organic molecules from aqueous solutions.
Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4991/8/11/950/s1, Figure S1: 31 P-NMR spectra of the CFBL and LBL; Figure S2: FTIR spectra of the LBL-Mannich and CFBL-Mannich samples before and after precipitation and washing by diethyl ether; Figure S3: FTIR spectra (in range 4000-600 cm −1 ) of the materials based on (a) LBL and (b) CFBL: initial lignin; lignin modified by the Mannich reaction; lignin-silica hybrids and synthesized pure silica; Figure S4: TG-and DTG-curves of thermal decomposition in N 2 atmosphere for initial and modified lignins, hybrid composites and silica; Figure S5: TGand DTG-curves of thermal decomposition in O 2 atmosphere for initial and modified lignins, hybrids and silica; Figure S6: Adsorption of methylene blue dye; Table S1: Characteristics of thermal decomposition in O 2 atmosphere of studied materials.