Catalyzing Benzoxazine Polymerization with Titanium-Containing POSS to Reduce the Curing Temperature and Improve Thermal Stability

Trisilanolphenyl-polyhedral oligomeric silsesquioxane titanium (Ti-Ph-POSS) was synthesized through the corner-capping reaction, and Ti-Ph-POSS was dispersed in benzoxazine (BZ) to prepare Ti-Ph-POSS/PBZ composite materials. Ti-Ph-POSS could catalyze the ring-opening polymerization (ROP) of BZ and reduce the curing temperature of benzoxazine. In addition, Ti immobilized on the Ti-Ph-POSS cage could form covalent bonds with the N or O atoms on polybenzoxazine, improving the thermal stability of PBZ. The catalytic activity of the Ti-Ph-POSS/BZ mixtures was assessed and identified through 1H nuclear magnetic resonance (1H-NMR) and Fourier-transform infrared (FTIR) analyses, while thermogravimetric analysis (TGA) and dynamic mechanical analysis (DMA) were used to determine the thermal properties of the composite. It was found that PBZ exhibited a higher glass transition temperature (Tg) and better thermal stability when Ti-Ph-POSS was added. The curing behavior of the Ti-Ph-POSS/BZ mixtures showed that the initial (Ti) and peak (Tp) curing temperatures sharply decreased as the content of Ti-Ph-POSS and the heating rate increased. The curing kinetics of these Ti-Ph-POSS/BZ systems were analyzed using the Kissinger method, and the morphology of Ti-Ph-POSS/PBZ was determined via scanning electron microscopy (SEM). It was found that the Ti-Ph-POSS particles were well distributed in the composites. When the content exceeded 2 wt%, several Ti-Ph-POSS particles could not react with benzoxazine and were only dispersed within the PBZ matrix, resulting in aggregation of the Ti-Ph-POSS molecules.


Introduction
Polyhedral oligomeric silsesquioxane (POSS) is an organic/inorganic hybrid structure with a common formula of RSiO 1.5 [1,2]. These molecular materials contain fully condensed cubic octamer clusters with partially condensed frameworks. Fully condensed POSS typically contains organic groups with variable designs for copolymerization, blending, or grafting [3]. Many recent studies have reported the preparation of polymer/POSS composites to improve the thermal stability, mechanical and dielectric properties, and flame retardancy of thermosets, resulting in superior properties compared to pure resin [4][5][6]. Notably, additional cross-linking between POSS macromers and the matrix can further improve the performance of thermosets. For example, epoxy-functionalized POSS has been shown to improve the thermal stability and dielectric properties of polybenzoxazine (PBZ) [7,8]. Yang et al. found that epoxy-functionalized POSS effectively improved the mechanical properties, thermal stability, and ablation performance of phenolic resins [9]. Compared with fully condensed POSS, partially condensed POSS has attracted interest because it can be easily functionalized with specific organic silanes and heteroelements [2,10].
have reported the synthesis of Ti-Ph-POSS to catalyze the ROP of BZ, or the preparation of Ti-Ph-POSS/PBZ composite materials.
The aim of this work was to study the reaction activity of titanium-based heterogeneous catalysts at the molecular level. In this work, Ti-Ph-POSS was synthesized and used as a titanium-based heterogeneous catalyst model to study the catalytic activity, curing behavior, and curing kinetics. In addition, Ti-Ph-POSS/PBZ nanocomposites were prepared by blending Ti-Ph-POSS with bisphenol A-based BZ (BA-a), and the structure, morphology, and thermal performance of the composites were studied. We found that Ti-Ph-POSS could catalyze the ROP of BZ, reduce the initial (T i ) and peak (T p ) curing temperatures, and offer attractive properties, especially thermal stability and flame retardancy.

Results and Discussion
2.1. The Catalytic Activity of Ti-Ph-POSS 2.1.1. 1 H-NMR Analysis P-cresol-aniline-based BZ (pC-a) was used as the model molecule for studying the ROP mechanism in this work because one reaction position (sixth position) was occupied by a CH 3 unit. This could prevent the formation of a highly cross-linked PBZ, and the obtained Ti-Ph-POSS/pC-a mixtures could be dissolved in deuterated dimethylsulfoxide (DMSO-d 6 ), allowing us to proceed with the process of polymerization. Therefore, we used pC-a to investigate the ROP mechanism of BA-a catalyzed by Ti-Ph-POSS. Considering the structural complexity of the polymer and to simplify the analysis, we focused on the CH 2 units in the polymer. It is worth noting that, as shown in Figure S4, the Ti-Ph-POSS we synthesized does not exhibit characteristic absorption peaks of Si-OH stretching and bending, so only Ti-Ph-POSS participates in the catalytic curing of benzoxazine. In Scheme 1, it can be found that several different types of CH 2 units formed in pC-a: ArO-CH 2 -OAr (a), -(Ph)N-CH 2 -N(Ph)-(b), ArO-CH 2 -N(Ph)-(c), ArO-CH 2 -Ar (d), -(Ph)N-CH 2 -Ar (e), and Ar-CH 2 -Ar (f). We predicted the possible chemical shift range of multiple methylene units in different structures, as shown in Scheme 1 [26].
Molecules 2023, 28, x FOR PEER REVIEW 3 of 15 catalyze the ROP of L-lactide [45]. However, Ti-POSS has been extensively studied by scientists due to its potential, for example, in catalyst applications [46][47][48]. Ti-Ph-POSS can also act as a Lewis acid to catalyze ring-opening reactions. As far as we know, no studies have reported the synthesis of Ti-Ph-POSS to catalyze the ROP of BZ, or the preparation of Ti-Ph-POSS/PBZ composite materials. The aim of this work was to study the reaction activity of titanium-based heterogeneous catalysts at the molecular level. In this work, Ti-Ph-POSS was synthesized and used as a titanium-based heterogeneous catalyst model to study the catalytic activity, curing behavior, and curing kinetics. In addition, Ti-Ph-POSS/PBZ nanocomposites were prepared by blending Ti-Ph-POSS with bisphenol A-based BZ (BA-a), and the structure, morphology, and thermal performance of the composites were studied. We found that Ti-Ph-POSS could catalyze the ROP of BZ, reduce the initial (Ti) and peak (Tp) curing temperatures, and offer attractive properties, especially thermal stability and flame retardancy.

The Catalytic Activity of Ti-Ph-POSS
2.1.1. 1 H-NMR Analysis P-cresol-aniline-based BZ (pC-a) was used as the model molecule for studying the ROP mechanism in this work because one reaction position (sixth position) was occupied by a CH3 unit. This could prevent the formation of a highly cross-linked PBZ, and the obtained Ti-Ph-POSS/pC-a mixtures could be dissolved in deuterated dimethylsulfoxide (DMSO-d6), allowing us to proceed with the process of polymerization. Therefore, we used pC-a to investigate the ROP mechanism of BA-a catalyzed by Ti-Ph-POSS. Considering the structural complexity of the polymer and to simplify the analysis, we focused on the CH2 units in the polymer. It is worth noting that, as shown in Figure S4, the Ti-Ph-POSS we synthesized does not exhibit characteristic absorption peaks of Si-OH stretching and bending, so only Ti-Ph-POSS participates in the catalytic curing of benzoxazine. In Scheme 1, it can be found that several different types of CH2 units formed in pC-a: ArO-CH2-OAr (a), -(Ph)N-CH2-N(Ph)-(b), ArO-CH2-N(Ph)-(c), ArO-CH2-Ar (d), -(Ph)N-CH2-Ar (e), and Ar-CH2-Ar (f). We predicted the possible chemical shift range of multiple methylene units in different structures, as shown in Scheme 1 [26]. Scheme 1. Ring-opening polymerization mechanism of benzoxazines. Figure 1 shows the 1 H nuclear magnetic resonance ( 1 H-NMR) spectra of the thermally cured mixtures of pC-a and Ti-Ph-POSS at different temperatures and times. As shown in Scheme 1. Ring-opening polymerization mechanism of benzoxazines. Figure 1 shows the 1 H nuclear magnetic resonance ( 1 H-NMR) spectra of the thermally cured mixtures of pC-a and Ti-Ph-POSS at different temperatures and times. As shown in Figure 1d, due to the differences in methylene groups, the 1 H-NMR spectrum of P 300 contains five characteristic signals, δ 5.5-5.3, 4.7-4.5 ppm (i), 4.7-4.4 ppm (ii), 4.4-4.0 ppm (iii), and 4.0-3.3 ppm (iv). Other than the peaks at 5.4 and 4.6 ppm, signal i was due to the initiation of pC-a, while signal ii corresponded to CH 2 (a) or CH 2 (b) in the phenoxy structure. Signal iii indicated the presence of CH 2 (c) and CH 2 (d) in the phenoxy structure, and signal iv may be CH 2 (e) in the phenoxy and phenolic structures, as well as CH 2 (f) in phenolic structures. As shown in Figure 1d, after 5 h of curing, all BZ rings in the mixture underwent ring opening and generated main signals (ii, iii, and iv). At 200 • C, signal ii almost disappeared and transformed into signal iv through rearrangement, and signal iii attributed to CH 2 (c, d) in the phenoxy structure completely disappeared, finally changing to peak iv due to the CH 2 units (e, f). This demonstrates that the CH 2 units in the phenoxy structure are unstable, and that the CH 2 units in the phenolic structure are the ultimate stable structure in PBZ. The mechanism of this process is shown in Scheme 1. Figure 1d, due to the differences in methylene groups, the 1 H-NMR spectrum of P300 contains five characteristic signals, δ 5.5-5.3, 4.7-4.5 ppm (i), 4.7-4.4 ppm (ii), 4.4-4.0 ppm (iii), and 4.0-3.3 ppm (iv). Other than the peaks at 5.4 and 4.6 ppm, signal i was due to the initiation of pC-a, while signal ii corresponded to CH2 (a) or CH2 (b) in the phenoxy structure. Signal iii indicated the presence of CH2 (c) and CH2 (d) in the phenoxy structure, and signal iv may be CH2 (e) in the phenoxy and phenolic structures, as well as CH2 (f) in phenolic structures. As shown in Figure 1d, after 5 h of curing, all BZ rings in the mixture underwent ring opening and generated main signals (ii, iii, and iv). At 200 °C, signal ii almost disappeared and transformed into signal iv through rearrangement, and signal iii attributed to CH2 (c, d) in the phenoxy structure completely disappeared, finally changing to peak iv due to the CH2 units (e, f). This demonstrates that the CH2 units in the phenoxy structure are unstable, and that the CH2 units in the phenolic structure are the ultimate stable structure in PBZ. The mechanism of this process is shown in Scheme 1.

FTIR Analysis
The catalytic activity of Ti-Ph-POSS was studied using Fourier-transform infrared (FTIR) spectroscopy. FTIR spectra of Ti-Ph-POSS/BZ with a Ti-Ph-POSS content of 2 wt% at different curing temperatures are shown in Figure 2. We observed that the characteristic peak of the benzoxazine ring (945 cm −1 , 1230 cm −1 ) gradually decreased during the curing process and eventually disappeared. The peak located at 1497 cm −1 indicates that the benzene ring was trisubstituted, which gradually moved to 1485 cm −1 after curing at 180 °C and 200 °C. There was a significant formation of tetrasubstituted benzene after the ROP of the BZ ring. In addition, a new broad absorption peak appeared at 3400 cm −1 at curing temperatures of 160 °C, 180 °C, and 200 °C, indicating the formation of hydroxyl groups. Furthermore, the characteristic absorption of Si-O-Si at 1107 cm −1 and 1170 cm −1 appeared during the curing process. The new peak at 1170 cm −1 belongs to the silanol group, which catalyzed the ROP of BZ and the formation of C-O-Si.

FTIR Analysis
The catalytic activity of Ti-Ph-POSS was studied using Fourier-transform infrared (FTIR) spectroscopy. FTIR spectra of Ti-Ph-POSS/BZ with a Ti-Ph-POSS content of 2 wt% at different curing temperatures are shown in Figure 2. We observed that the characteristic peak of the benzoxazine ring (945 cm −1 , 1230 cm −1 ) gradually decreased during the curing process and eventually disappeared. The peak located at 1497 cm −1 indicates that the benzene ring was trisubstituted, which gradually moved to 1485 cm −1 after curing at 180 • C and 200 • C. There was a significant formation of tetrasubstituted benzene after the ROP of the BZ ring. In addition, a new broad absorption peak appeared at 3400 cm −1 at curing temperatures of 160 • C, 180 • C, and 200 • C, indicating the formation of hydroxyl groups. Furthermore, the characteristic absorption of Si-O-Si at 1107 cm −1 and 1170 cm −1 appeared during the curing process. The new peak at 1170 cm −1 belongs to the silanol group, which catalyzed the ROP of BZ and the formation of C-O-Si.

Curing Behavior of Ti-Ph-POSS/BZ
The isothermal differential scanning calorimetry (DSC) curves of pure BZ and Ti-Ph-POSS/BZ with different mass ratios at a heating rate of 10 • C/min are shown in Figure 3. The Ti-Ph-POSS/BZ systems showed significantly lower initial and peak curing temperatures than the pure BZ resin, which indicated that Ti-Ph-POSS as the Lewis acid could catalyze BZ ring opening and promote the reaction. Our previous work showed that the initial and peak curing temperatures of pure benzoxazine were 252 • C and 258 • C, Molecules 2023, 28, 5450 5 of 15 respectively [49]. As shown in Figure 4, T i decreased with the increase in the Ti-Ph-POSS mass ratio from 0 wt% to 3 wt%. According to Table 1, when the Ti-Ph-POSS mass ratio was 3 wt%, the initial and peak curing temperatures decreased by almost 44.6 • C and 27.8 • C, respectively.

Curing Behavior of Ti-Ph-POSS/BZ
The isothermal differential scanning calorimetry (DSC) curves of pure BZ and Ti-Ph-POSS/BZ with different mass ratios at a heating rate of 10 °C/min are shown in Figure 3. The Ti-Ph-POSS/BZ systems showed significantly lower initial and peak curing temperatures than the pure BZ resin, which indicated that Ti-Ph-POSS as the Lewis acid could catalyze BZ ring opening and promote the reaction. Our previous work showed that the initial and peak curing temperatures of pure benzoxazine were 252 °C and 258 °C, respectively [49]. As shown in Figure 4, Ti decreased with the increase in the Ti-Ph-POSS mass ratio from 0 wt% to 3 wt%. According to Table 1, when the Ti-Ph-POSS mass ratio was 3 wt%, the initial and peak curing temperatures decreased by almost 44.6 °C and 27.8 °C, respectively.

Curing Kinetics of Ti-Ph-POSS/BZ Composite Systems
The curing kinetics model based on DSC data has been widely used in the study of thermosetting resins. We used Ti-Ph-POSS/BZ with different contents and discussed the thermal curing kinetics of Ti-Ph-POSS/BZ mixtures at heating rates of 5, 10, 15, and 20 • C/min using the DSC technique. The Kissinger method was used under the anisothermal condition to calculate the activation energy (E a ) and frequency factor (A), due to its simplicity in handling the dynamic curing process of the BZ system. The kinetic parameters obtained using the Kissinger method did not require assumptions regarding the conversion-dependent equation. Kissinger's equation can be expressed by where β is the heating rate, T p is the curing peak temperature of Ti-Ph-POSS/BZ, and R is the constant of an ideal gas. Thus, E a could be obtained from the slope of the ln(β/T p 2 ) vs. 1/T p plot, and A could be obtained from Figure 5. In addition, the curing reaction order (n) can be calculated using n = 1.26S 1/2 , where S is the peak shape index of the DSC curves and is equal to the absolute value of the slope of the left inflection point vs. the slope of the right inflection point in the curves. We took the average value of S among the four different heating rates, and the curing reaction order (n) was obtained.

Curing Kinetics of Ti-Ph-POSS/BZ Composite Systems
The curing kinetics model based on DSC data has been widely used in the study of thermosetting resins. We used Ti-Ph-POSS/BZ with different contents and discussed the thermal curing kinetics of Ti-Ph-POSS/BZ mixtures at heating rates of 5, 10, 15, and 20 °C/min using the DSC technique. The Kissinger method was used under the anisothermal  the slope of the left inflection point vs. the slope of the right inflection point in the curves. We took the average value of S among the four different heating rates, and the curing reaction order (n) was obtained. As shown in Figure 4, the Tp of the curing curve increased with the increase in β. The relevant DSC data are summarized in Table 1. The Ea and A values of the Ti-Ph-POSS/BZ systems were obtained using Equation (1). The results of the dynamic DSC experiments are summarized in Table 2.  As shown in Table 2, the value of Ea decreased as the Ti-Ph-POSS content increased. When the content of Ti-Ph-POSS was below 2 wt%, Ti-Ph-POSS was uniformly dispersed in Ti-Ph-POSS /BZ, effectively catalyzing the solidification of benzoxazine. It is worth noting that the Ea value of 2 wt% is similar to that of 3 wt%. When the mass ratio of Ti-Ph-POSS exceeded 2 wt%, Ti-Ph-POSS aggregated in Ti-Ph-POSS/BZ and had no significant effect on the catalytic effect of BZ ring opening. As shown in Figure 4, the T p of the curing curve increased with the increase in β. The relevant DSC data are summarized in Table 1. The E a and A values of the Ti-Ph-POSS/BZ systems were obtained using Equation (1). The results of the dynamic DSC experiments are summarized in Table 2. As shown in Table 2, the value of E a decreased as the Ti-Ph-POSS content increased. When the content of Ti-Ph-POSS was below 2 wt%, Ti-Ph-POSS was uniformly dispersed in Ti-Ph-POSS /BZ, effectively catalyzing the solidification of benzoxazine. It is worth noting that the E a value of 2 wt% is similar to that of 3 wt%. When the mass ratio of Ti-Ph-POSS exceeded 2 wt%, Ti-Ph-POSS aggregated in Ti-Ph-POSS/BZ and had no significant effect on the catalytic effect of BZ ring opening.

Thermal Properties of Ti-Ph-POSS/PBZ
According to Rosa María Sebastián et al. [50], polymers with a predominantly phenolic structure will have stable and rigid structures with a low surface energy due to the predominance of the intramolecular hydrogen bonds. Dynamic mechanical thermal and thermogravimetric analyses in subsequent work demonstrated that BZ catalyzed by Ti-Ph-POSS had a higher T g and thermal stability than pristine BZ.

Dynamic Mechanical Thermal Analysis
The loss (tanδ) and storage modulus (E ) curves of the Ti-Ph-POSS/PBZ composites with various contents of Ti-Ph-POSS are displayed in Figure 6a,b. Figure 6a presents the loss tanδ peak temperatures obtained using dynamic mechanical analysis (DMA), indicating the T g of these materials. The T g value of pure BZ was 168.8 • C, while the T g values of the composites containing 0.5, 1, 2, and 3 wt% Ti-Ph-POSS were 194.8 • C, 196.3 • C, 199.3 • C, and 201.8 • C, respectively, indicating that T g increased with increasing Ti-Ph-POSS content. T g was affected by the backbone rigidity of the BZ monomer and cross-linking density; thus, the T g of the cured polymers could be drastically elevated by introducing rigid groups into the backbone or by increasing the cross-linking density of the cured polymer [7]. As shown in Figure 6 and Table 3, the Ti-Ph-POSS catalysts within the composites were separated into two parts. Ti immobilized on the POSS cage participated in forming covalent bonds with the N or O atoms of BZ. Thus, the Ti-Ph-POSS particles were well distributed within the polymer matrix. When the content of Ti-Ph-POSS was 0.5 wt%, Ti-Ph-POSS was fully involved in cross-linking, and T g rapidly increased. However, as the content of the Ti-Ph-POSS catalyst increased, some of the Ti-Ph-POSS particles could not react with the benzoxazine rings and were only dispersed within PBZ, leading to the aggregation of the Ti-Ph-POSS molecules (Scheme 2). As expected, the strong covalent bonding and presence of the rigid POSS particles hindered the mobility of the polymer chains and resulted in higher T g values. When the Ti-Ph-POSS mass ratio exceeded 0.5 wt%, the T g values slowly changed with the Ti-Ph-POSS content. When the content increased from 0 wt% to 0.5 wt%, the storage modulus increased, mainly due to the nano-strengthening effect of the Ti-Ph-POSS cage structure, which makes it difficult for PBZ macromolecular chains to move and increases the modulus of the composite material. When the content exceeded 0.5 wt%, with the increase in the Ti-Ph-POSS content, the presence of hollow structures dominated, leading to a decrease in the material density and storage modulus.
with various contents of Ti-Ph-POSS are displayed in Figure 6a,b. Figure 6a presents the loss tanδ peak temperatures obtained using dynamic mechanical analysis (DMA), indicating the Tg of these materials. The Tg value of pure BZ was 168.8 °C, while the Tg values of the composites containing 0.5, 1, 2, and 3 wt% Ti-Ph-POSS were 194. 8 °C, 196.3 °C, 199.3 °C, and 201.8 °C, respectively, indicating that Tg increased with increasing Ti-Ph-POSS content. Tg was affected by the backbone rigidity of the BZ monomer and cross-linking density; thus, the Tg of the cured polymers could be drastically elevated by introducing rigid groups into the backbone or by increasing the cross-linking density of the cured polymer [7]. As shown in Figure 6 and Table 3, the Ti-Ph-POSS catalysts within the composites were separated into two parts. Ti immobilized on the POSS cage participated in forming covalent bonds with the N or O atoms of BZ. Thus, the Ti-Ph-POSS particles were well distributed within the polymer matrix. When the content of Ti-Ph-POSS was 0.5 wt%, Ti-Ph-POSS was fully involved in cross-linking, and Tg rapidly increased. However, as the content of the Ti-Ph-POSS catalyst increased, some of the Ti-Ph-POSS particles could not react with the benzoxazine rings and were only dispersed within PBZ, leading to the aggregation of the Ti-Ph-POSS molecules (Scheme 2). As expected, the strong covalent bonding and presence of the rigid POSS particles hindered the mobility of the polymer chains and resulted in higher Tg values. When the Ti-Ph-POSS mass ratio exceeded 0.5 wt%, the Tg values slowly changed with the Ti-Ph-POSS content. When the content increased from 0 wt% to 0.5 wt%, the storage modulus increased, mainly due to the nano-strengthening effect of the Ti-Ph-POSS cage structure, which makes it difficult for PBZ macromolecular chains to move and increases the modulus of the composite material. When the content exceeded 0.5 wt%, with the increase in the Ti-Ph-POSS content, the presence of hollow structures dominated, leading to a decrease in the material density and storage modulus.

Thermogravimetric Analysis
Thermogravimetric analysis (TGA) was used to assess the effect of Ti-Ph-POSS on the thermal properties of the Ti-Ph-POSS/PBZ composites, as shown in Figure 7. To compare the thermal stability, we used the 10% weight loss temperature (T d ) as the standard. The TGA results showed the char yield at 700 • C (Y c ) and T d , as shown in Table 4. Compared with pure BZ, the T d of the Ti-Ph-POSS/PBZ composite material increased. In addition, when Ti-Ph-POSS was added to the matrix, the Y c improved. The difference in decomposition temperature was possibly due to the effect of Ti-Ph-POSS forming covalent bonds with the BZ molecule. When the content of Ti-Ph-POSS was less than 2 wt%, Ti-Ph-POSS was uniformly dispersed in the benzoxazine matrix, and all of Ti-Ph-POSS participated in cross-linking. The thermal motion of the chain segment was severely blocked, and the T d of the composite rapidly increased. When the content of Ti-Ph-POSS exceeded 2 wt%, part of Ti-Ph-POSS aggregated in the PBZ matrix, the restriction of the thermal motion of the chain segment decreased, and the increase in T d slowed down. As a result, the thermal motion of the chain segment was limited, thus reducing the decomposition path of organic matter. In addition, inorganic components (Ti-Ph-POSS) can provide additional heat capacity to increase the thermal stability of the material [51]. In general, the T d and Y c of the composites were higher than those of pristine PBZ. According to the test results, the T d and Y c of the Ti-Ph-POSS molecule were 442.0 • C and 55.9%, respectively. Considering its high char yield, we determined that Ti-Ph-POSS could act as a flame retardant in the composite.

Thermogravimetric Analysis
Thermogravimetric analysis (TGA) was used to assess the effect of Ti-Ph-POSS on the thermal properties of the Ti-Ph-POSS/PBZ composites, as shown in Figure 7. To compare the thermal stability, we used the 10% weight loss temperature (Td) as the standard. The TGA results showed the char yield at 700 °C (Yc) and Td, as shown in Table 4. Compared with pure BZ, the Td of the Ti-Ph-POSS/PBZ composite material increased. In addition, when Ti-Ph-POSS was added to the matrix, the Yc improved. The difference in decomposition temperature was possibly due to the effect of Ti-Ph-POSS forming covalent bonds with the BZ molecule. When the content of Ti-Ph-POSS was less than 2 wt%, Ti-Ph-POSS was uniformly dispersed in the benzoxazine matrix, and all of Ti-Ph-POSS participated in cross-linking. The thermal motion of the chain segment was severely blocked, and the Td of the composite rapidly increased. When the content of Ti-Ph-POSS exceeded 2 wt%, part of Ti-Ph-POSS aggregated in the PBZ matrix, the restriction of the thermal motion of the chain segment decreased, and the increase in Td slowed down. As a result, the thermal motion of the chain segment was limited, thus reducing the decomposition path of organic matter. In addition, inorganic components (Ti-Ph-POSS) can provide additional heat ca-

Morphology of Ti-Ph-POSS/PBZ
The morphology of Ti-Ph-POSS/PBZ was assessed using scanning electron microscopy (SEM).

Morphology of Ti-Ph-POSS/PBZ
The morphology of Ti-Ph-POSS/PBZ was assessed using scanning electron microscopy (SEM).

Materials
Trisilanolphenyl   33 (m, Ar)). BA-a was synthesized in our laboratory using the solvent method described in our previous work [7]. As shown in Figures S1 and S2, the structure of BA-a was characterized using FTIR and 1 H-NMR, and the cyclization rate of the benzoxazine monomer was 97%. pC-a was prepared according to the procedure reported in the literature [50], and 4,4 -(1-methylethylidene) bisphenol, P-cresol, aniline, paraformaldehyde, acetone, ethanol, dichloromethane (DCM), and tetrahydrofuran (THF) were chemically pure. THF was dried and distilled over CaH 2 . All other reagents were used as received without further purification.

Syntheses of Trisilanolphenyl-POSS Titanium (Ti-Ph-POSS)
The synthesis of trisilanolphenyl-POSS titanium (Ti-Ph-POSS) was based on the cornercapping reaction of trisilanolphenyl-POSS with titanium(IV) isopropoxide [Ti(OiPr) 4 ] (Scheme 3) [52]. A certain proportion Ti(OiPr) 4 was added to a solution of trisilanolphenyl-POSS, which was dissolved in acetone under stirring in a nitrogen atmosphere. Then, it was purged in a nitrogen atmosphere for 30 min to remove air and moisture. The solution was reacted on a preheated 80 • C air bath heating jacket for 6 h. Upon cooling, the crude product was collected through suction filtration and allowed to evaporate in a vacuum oven for 1 day. Then, the crude product was dissolved in DCM, reprecipitated from methanol, and vacuum-dried at 50 • C for at least 24 h, yielding approximately 80% white powder. Ti-Ph-POSS was characterized using FTIR and NMR, as shown in Figures S3 and S4. In addition, Figure S5 can confirm our inquiry and indirectly show that Ti-Ph-POSS is relatively pure. FTIR characterization showed peaks at 1100 cm −1 (Si-O-Si) and 935 cm −1 (Si-O-Ti) [12], while 29 Si-NMR (in CDCl 3 at 25 • C standard tetramethyl silane) indicated −78.55 ppm (relative to the three silicon nuclei next to the Ti-containing silicon), −78.30 ppm (one silicon in the clinodiagonal of the Ti atom), and −71.76 ppm (Ti-containing silicon), with the intensity ratio approaching 3:1:3.

Ti-Ph-POSS/PBZ Composites
Ti-Ph-POSS and the BA-a molecule were dissolved in THF at mass ratios of 0.5:99.5, 1:99, 2:98, and 3:96 to obtain Ti-Ph-POSS/BZ mixtures with contents of 0.5 wt%, 1 wt%, 2 wt%, and 3 wt%, respectively. Most of the THF was evaporated at room temperature, and the rest was dried under vacuum to obtain a homogeneous Ti-Ph-POSS/BZ mixture. These Ti-Ph-POSS/BZ blends were molded and solidified. The curing process of Ti-Ph-POSS/BZ was performed at 100 • C for 1 h, 120 • C for 2 h, 140 • C for 2 h, 160 • C for 2 h, and 180 • C for 2 h, followed by post-curing at 200 • C for 2 h.

Ti-Ph-POSS/pC-a Composites
Similarly, P-cresol-aniline-based BZ and Ti-Ph-POSS (1 wt%) were dissolved in THF. The mixture was vacuum-dried at room temperature for 1 day and then heated at 150 • C for 15 min, 150 • C for 0.5 h, 150 • C for 1 h, 150 • C for 5 h, and 200 • C for 2 h. The prepared mixtures were dissolved in DMSO-d 6 , which was only used for catalytic activity analysis. ppm (one silicon in the clinodiagonal of the Ti atom), and −71.76 ppm (Ti-containing silicon), with the intensity ratio approaching 3:1:3.

Ti-Ph-POSS/PBZ Composites
Ti-Ph-POSS and the BA-a molecule were dissolved in THF at mass ratios of 0.5:99.5, 1:99, 2:98, and 3:96 to obtain Ti-Ph-POSS/BZ mixtures with contents of 0.5 wt%, 1 wt%, 2 wt%, and 3 wt%, respectively. Most of the THF was evaporated at room temperature, and the rest was dried under vacuum to obtain a homogeneous Ti-Ph-POSS/BZ mixture. These Ti-Ph-POSS/BZ blends were molded and solidified. The curing process of Ti-Ph-POSS/BZ was performed at 100 °C for 1 h, 120 °C for 2 h, 140 °C for 2 h, 160 °C for 2 h, and 180 °C for 2 h, followed by post-curing at 200 °C for 2 h.

Ti-Ph-POSS/pC-a composites
Similarly, P-cresol-aniline-based BZ and Ti-Ph-POSS (1 wt%) were dissolved in THF. The mixture was vacuum-dried at room temperature for 1 day and then heated at 150 °C for 15 min, 150 °C for 0.5 h, 150 °C for 1 h, 150 °C for 5 h, and 200 °C for 2 h. The prepared mixtures were dissolved in DMSO-d6, which was only used for catalytic activity analysis.

Characterization
FTIR measurements were conducted using a NEXUS 6700 spectrophotometer at room temperature (25 °C) with a range of 4000-400 cm −1 and a resolution of 1.0 cm −1 . All samples were prepared using spectral-grade KBr.
The NMR spectra were obtained using a Bruker AV600 spectrophotometer at room temperature, and tetramethylsilane (TMS) was used as the internal standard. The 1 H-NMR samples were dissolved in DMSO-d6, and the 29 Si-NMR samples were dissolved in CDCl3.
DSC was tested on the PerkinElmer Pyris-1 instrument and calibrated with standard indium. All samples were heated from 25 °C to 300 °C, and the entire experiment was conducted in a dry nitrogen atmosphere.

Characterization
FTIR measurements were conducted using a NEXUS 6700 spectrophotometer at room temperature (25 • C) with a range of 4000-400 cm −1 and a resolution of 1.0 cm −1 . All samples were prepared using spectral-grade KBr.
The NMR spectra were obtained using a Bruker AV600 spectrophotometer at room temperature, and tetramethylsilane (TMS) was used as the internal standard. The 1 H-NMR samples were dissolved in DMSO-d 6 , and the 29 Si-NMR samples were dissolved in CDCl 3 .
DSC was tested on the PerkinElmer Pyris-1 instrument and calibrated with standard indium. All samples were heated from 25 • C to 300 • C, and the entire experiment was conducted in a dry nitrogen atmosphere.
DMA was conducted on a Rheometrics ScientificTM DMTA V at 1 Hz, from ambient temperature to 300 • C at a heating rate of 5 • C/min. TGA was carried out using a Toshiba Netzsch 209 C thermogravimetric analyzer, from room temperature to 700 • C, and the experiment was carried out under a nitrogen atmosphere, with a heating rate of 10 • C/min. SEM images were taken using a ZEISS Gemini 300 operated at 3 kV. All samples were cut at room temperature, and the distributions of Si atoms in the hybrids were obtained using SEM-EDX Si mapping, which is operated using Xplore 30 instruments at 15 KV, Oxford Instrument Technology (Shanghai) Co., Ltd. (Shanghai, China), energy spectrum model is Smartedx.

Conclusions
In this work, Ti-Ph-POSS catalyzed the BZ ring-opening reaction, allowing the Ti atom, which was immobilized on the corner of a POSS structure, to form a covalent bond with PBZ. Several systems with different Ti-Ph-POSS/BZ contents were prepared, and the ring-opening curing process and catalytic activity were assessed and identified using FTIR and 1 H-NMR. The curing behaviors of Ti-Ph-POSS/BZ indicated that the initial (T i ) and peak (T p ) curing temperatures sharply decreased with increasing Ti-Ph-POSS mass ratio and heating rate. The curing kinetics of these Ti-Ph-POSS/BZ systems were investigated using the Kissinger method. The results indicated that Ti-Ph-POSS could be used as a type of catalyst for BZ. In addition, with the introduction of Ti-Ph-POSS, the composites had a higher T g and better thermal stability.