Study on thermal degradation mechanism of a cured aldehyde-functional benzoxazine

Abstract

A cured product of aldehyde-functional benzoxazine has good heat-resistant performance. The thermal degradation process of this polybenzoxazine was actively studied by TGA-FTIR and Py-GC/MS. The temperature range of pyrolysis and the major products were determined by TGA-FTIR. A stepwise-temperature testing method based on Py-GC/MS was employed to identify the structures and contents of the pyrolysis products at different temperature stages. Transformation and chemical structures of the polymer bulk during the pyrolysis process were speculated. The results showed that the reactions of the aldehyde groups can form special crosslinking structures which effectively prevent the release of phenols during the pyrolysis process. Additionally, benzophenone compounds and carbon monoxide were detected.

---

1. Introduction

Benzoxazine is a novel class of high-performance thermosetting resin. Polybenzoxazines formed by ring-opening polymerization have various unique properties such as high mechanical strength, low water absorption, good flame-resistance and excellent thermal stability.^1–3 They can be used in several fields. With the development of aerospace technology, ablative-resistant materials are widely used leading to increasing requirements of the thermal stability of the materials.

Studying the pyrolysis of polybenzoxazines can help us to understand the essence of the thermal stability and guide us to design new structures to further improve the thermal resistance of materials. Ishida investigated the thermal decomposition processes of aromatic amine-based polybenzoxazines by TGA and GC/MS,⁴ and proposed that the thermal degradation products of this polybenzoxazine can be grouped into two categories including primary decomposition products originated from polybenzoxazines itself and secondary decomposition products resulting from the combination or further decomposition of those primary decomposition products. Although GC/MS tests can give structure information of decomposition products during the entire process, this test fails to tell the exact time and temperature of the appearance of the decomposition products. TGA-FTIR can give structure information at some time and temperature, but it is hard to distinguish the decomposition products because of the complexity of IR spectra. To solve this problem, a stepwise-temperature testing method based on Py-GC/MS has been applied by our research group to study the thermal degradation processes of two polybenzoxazines containing sulfone groups.⁵ This method can give lots of information about degradation products and the variation of chemical structures in bulk.

Reactive functional groups are usually employed to increase crosslinking densities and the thermal stability of polybenzoxazines. It has been shown that the thermal stability of the cured benzoxazine monomers containing acetylene, nitrile, furan, and maleimide groups can be greatly improved.^6–15 A novel aldehyde-functional benzoxazine monomer (Ald-B) has been synthesized in our group and its polymer has superior thermal stability.¹⁶ Also, its curing reaction has been investigated in detail.¹⁷ It was inferred that the aldehyde groups reacted with ortho positions of phenol and participated in oxidation and decarboxylation reactions to create new crosslinking structures. But the effect of these new structures on the degradation process is unclear.

In this study, the thermal degradation process of cured Ald-B was studied as a part of our research series on this resin. The temperature range of pyrolysis and main products were determined by TGA-FTIR. A stepwise-temperature testing method based on Py-GC/MS was employed. Several temperatures were selected to divide the main weight-losing temperature range into small ones. Structures and contents of the degradation products at different temperatures were obtained. The entire process of thermal degradation is discussed in detail involving the structure evolution of gas phase and bulk phase. The chemical structure of polybenzoxazine (poly-A) was further speculated from the perspective of the degradation.

2. Experimental

2.1 Materials

3-Phenyl-3,4-dihydro-2H-benzo[e][1,3]oxazine-6-carbaldehyde (Ald-B, Fig. 1) was synthesized according to ref. 16.

Fig. 1 Chemical structure of Ald-B.

2.2 Preparation of polymers

Ald-B was melted in a mold at 130 °C. The resin-filled mold was initially degassed at 130 °C for 3 h to remove trace solvent from the resin. Then a stepwise curing process was adopted as follows: 140 °C/2 h, 160 °C/4 h, 180 °C/2 h and 200 °C/4 h. After that, a dark red-brown porous polybenzoxazine was obtained and marked as poly-A.

2.3 Characterization

TGA-FTIR was performed on a TA Instruments TA-Q500 thermogravimetric analyzer interfaced to a Nicolet 6700 Fourier transform infrared spectrometer. Thermal degradation of the sample was under nitrogen atmosphere from 30 to 800 °C at a heating rate of 20 °C min⁻¹. The spectral resolution of the spectrometer was 8 cm⁻¹ and about 20 mg powder of poly-A was used.

Py-GC/MS analysis was carried out using a combination of single-point pyrolyzer PY-2020i (Frontier, Japan) and chromatograph-mass spectrometer GC/MS QP2010 (Shimadzu, Japan) equipped with a pyrolysis injection system. About 1 mg of sample was pyrolyzed at setting temperature immediately. Helium was used as a carrier gas. The energy of MS used for electron ionization (EI) was 70 eV. Masses were scanned from m/z 29 to 500. Identification of compounds was performed with a Nist11 MS library.

The process of the stepwise-temperature method by Py-GC/MS is detailed in Fig. 2. As shown below, the sample poly-A first degraded in pyrolyzer at T₁ to obtain Volatiles #1 and Char #1. Volatiles #1 were separated and characterized by GC/MS, while Char #1 was weighted and recorded as m₁. After that, Char #1 degraded further at next temperature T₂ similarly and degradation at T₃ and T₄ would be conducted by analogy.

Fig. 2 Stepwise-temperature method conducted by Py-GC/MS.

3. Results and discussion

3.1 TGA-FTIR

3.1.1 Thermal stability of poly-A. TGA thermogram and its derivative DTA curve from poly-A are shown in Fig. 3. Three obvious peaks centered at 310 °C, 449 °C and 563 °C were observed on the derivative curve. The main weigh-losing temperature interval ranged from 300 to 650 °C. It is generally considered that the thermal degradation process of polybenzoxazine has two stages including a low temperature stage for the release of amines and a high temperature stage for phenols, due to C–N bonds in the network of polybenzoxazine breaking easily when exposed to high temperatures. For poly-A, 449 °C and 563 °C represented the release of amines and phenols respectively, while 310 °C was probably due to reactions of aldehyde groups. The char yield at 800 °C of poly-A was 55.4% which is higher than that of cured Ph-a (35%) whose monomer has no reactive functional groups.⁴ It is proved that the aldehyde groups can effectively improve the thermal stability of polybenzoxazine.

Fig. 3 Thermal degradation pattern of poly-A.

3.1.2 FTIR spectra of volatiles. Fig. 4 gives the three-dimensional FTIR spectra of the volatiles from the thermal degradation of poly-A. The height of the peaks represents the strength of absorbance. The dotted line in Fig. 4 represents 350 °C. It can be seen clearly that the degradation of poly-A began to occur around 350 °C. Eight IR spectra between 100 and 800 °C with an interval of 100 °C (5 min) in Fig. 4 were picked up and collected in Fig. 5. Main absorption bands were assigned and classified. As a result, some kinds of volatiles involving amines (1265 cm⁻¹), phenols (3648 cm⁻¹), and carbon monoxide (2179, 2117 cm⁻¹) were detected during the thermal degradation of poly-A. To clearly analyse the evolution of these volatiles, the absorbance of the characteristic bands at 1265 cm⁻¹, 3648 cm⁻¹ and 2117 cm⁻¹ vs. temperature was plotted in Fig. 6.

Fig. 4 Three-dimensional FTIR spectra of volatiles from the thermal degradation for poly-A.

Fig. 5 FTIR spectra of volatiles from 100 to 800 °C.

Fig. 6 Release of amines, phenolic compounds and carbon monoxide.

As seen from Fig. 5 and 6, at the beginning of thermal degradation (350–400 °C), the absorption band at 1265 cm⁻¹ assigned to amines was obviously strong, while 3648 cm⁻¹ belonging to phenols was still weak. Phenols began to evolve at about 500 °C and reached a maximum at 600 °C. This result suggested that the release of phenols was greatly delayed. It is inferred that the reactive aldehyde groups located in the para position of phenolic hydroxyl may participate in some kind of special crosslinking reactions during the polymerization process, and these crosslinking structures can “lock” phenols firmly in polybenzoxazine.

Furthermore, the absorption bands at 2179 cm⁻¹ and 2117 cm⁻¹ assigned to carbon monoxide started to appear at approximately 500 °C and reached a maximum at 600 °C as phenolic compounds. As we know, oxygen elements in polybenzoxazine originate from phenolic hydroxyl and no carbon monoxide could be detected during the thermal degradation of those polybenzoxazines without reactive functional groups.^4,18,19 In addition, for poly-A, few volatiles containing aldehyde groups were detected during the thermal degradation. So, it is believed that carbon monoxide probably comes from the degradation of the crosslinking structures formed by aldehyde groups. Moreover, the release of the phenols and carbon monoxide showed a similar tendency meaning there is a close relationship between them.

3.2 Py-GC/MS

A stepwise-temperature testing method based on Py-GC/MS was employed to identify the chemical structures and contents of the volatiles in different thermal degradation stages. According to the TGA results of poly-A, the main degradation stage was from 300 °C to 650 °C. Four temperatures, 350 °C (T₁), 400 °C (T₂), 500 °C (T₃) and 600 °C (T₄), were selected. The original sample and the chars at different temperatures were weighted, and losing-weight percentage Δm_n was calculated using Eqn (1). All the results are shown in Table 1.

Table 1 Char weights and losing-weight percentages at different temperatures

	25 °C	350 °C	400 °C	500 °C	600 °C
mn	1.410 mg	1.357 mg	1.265 mg	1.076 mg	0.984 mg
Δmn	—	3.8%	6.5%	13.4%	6.5%

---

In this equation, m₀ is the original sample weight while m_n represents the weight of the char at different temperature stages, such as 350 °C, 400 °C, 500 °C and 600 °C. For example, “m₁ − m₂” means the difference value of the weight of the char at 350 °C and 400 °C, it also means the mass of volatiles at 400 °C.

Based on the results of MS and the classification method used before,⁴ volatiles were grouped into eleven types (Table 2) including small molecule gases (abbr. as gas), phenolic compounds (–OH), amines (–NH), Mannich bases (Mannich), Schiff bases (Schiff), benzene derivatives (Ph), biphenyl compounds (Biph), 2,3-benzofuran derivatives (Bf), isoquinoline derivatives (Iq), phenanthridine derivatives (Pd) and benzophenones derivatives containing carbonyl groups (Bp).

Table 2 Chemical structures of volatiles detected by the stepwise temperature testing method using Py-GC/MS

Classification of volatiles	Chemical structures
Small molecule gas (Gas)	CO, CH4, NH3, H2O
Phenolic compounds (–OH)
Amines (–NH)
Mannich bases (Mannich)
Schiff bases (Schiff)
Benzene derivatives (Ph)
Biphenyl compounds (BiPh)
2,3-Benzofuran derivatives (Bf)
Isoquinoline derivatives (Iq)
Phenanthridine derivatives (Pd)
Benzophenone derivatives containing carbonyl groups (Bp)

---

These eleven types of volatiles were further grouped into three categories in Table 3 named primary degradation products, secondary degradation products (aromatic and heterocyclic compounds) and benzophenones derivatives containing carbonyl groups. The primary degradation products are obtained directly from the cleavage of the original network of polybenzoxazines, while the secondary degradation products come from some structures which result from further reactions during the degradation. Restricted to the Py-GC/MS method, only content percentages (abbr. as CP_n) of the volatiles at a certain temperature can be obtained by gas chromatography. Considering the weight-losing amounts are different at each temperature stage, we introduced the percentage Δm_n to get absolute release percentages by Eqn (2) and (3). All the results can be found in Table 3.

EMP_n = Δm_n × CP_n (n = 1, 2, 3, 4) (2)

Table 3 Contents of thermal degradation volatiles at different temperatures

Degradation volatiles		350 °C		400 °C		500 °C		600 °C		TMP (%)
		CP1 (%)	EMP1 (%)	CP2 (%)	EMP2 (%)	CP3 (%)	EMP3 (%)	CP4 (%)	EMP4 (%)
Primary degradation products	Gas	2.4	0.092	4.4	0.283	2.1	0.279	22.0	1.432	6.9
	–OH	3.9	0.148	9.7	0.631	32.4	4.347	27.6	1.791	22.9
	–NH	75.1	2.854	77.5	5.037	48.9	6.554	4.8	0.314	48.9
	Mannich	0.0	0.000	0.8	0.051	2.3	0.302	0.0	0.000	1.2
	Ph	0.2	0.008	0.8	0.051	2.5	0.336	15.0	0.974	4.5
Secondary degradation products	Biph	0.0	0.000	2.5	0.159	1.1	0.151	2.3	0.151	1.5
	Schiff	5.4	0.206	0.0	0.000	0.0	0.000	0.0	0.000	0.7
	Iq	0.0	0.000	0.0	0.000	0.0	0.000	0.6	0.036	0.1
	Bf	0.0	0.000	0.0	0.000	1.4	0.193	10.9	0.710	3.0
	Pd	12.5	0.474	1.2	0.081	5.6	0.745	16.8	1.093	7.9
Bh		0.5	0.019	3.2	0.206	3.7	0.493	0.0	0.000	2.4

---

EMP_n above represents mass percentage of each volatile type at a certain temperature when we regard the original sample mass as 100%. While TMP represents the total mass percentage of a type during the whole temperature stage (350 °C to 600 °C) when we regard the total losing weight as 100%. Taking the gas for example, “EMP₁ = 0.092%”, means that at 350 °C the mass of the released gas accounts for 0.092% of the original sample mass, while “TMP = 6.9%”, means that during the whole temperature stage the content of the released Gas accounts for 6.9% of the total losing weight or total mass of the volatiles. “n = 1, 2, 3, 4” here corresponds to 350 °C, 400 °C, 500 °C and 600 °C, respectively.

As we can see from Table 3, the primary degradation products including gas, –OH, –NH, Mannich and Ph accounted for 84% of the total amount meaning the primary degradation products are the main volatiles during the whole degradation process. These products were obtained directly from the cleavage of C–C and C–N bonds in the Mannich bridge of polybenzoxazine which usually began around 350 °C. Among these primary degradation products, –OH and –NH had the biggest release amount with 72%, and showed a similar release tendency. Both of them formed at 350 °C and their EMP values increased with temperature and reached the peak at 500 °C. After that the values decreased sharply. However, it should be noted that the EMP values of –OH compounds at 350 and 400 °C were 0.148% and 0.631%, respectively, which were no more than one ninth of those of –NH compounds (2.854% and 5.037%). These results suggest that the release of phenolic compounds has been greatly delayed, which is also observed in the results of TGA-FTIR. Combined with the MS results that the main released phenolic compounds are disubstituted and trisubstituted phenols, we can deduce that lots of new crosslinking sites in poly-A are formed on the phenol rings because of the reactions of aldehyde groups.

Additionally, the TMP values of the Mannich and Ph compounds were quite small and only accounted for 1.2% and 4.5%, respectively. The Mannich compounds were only detected at 400 and 500 °C. The release of Ph needs higher temperatures, like 500 °C or 600 °C, because of the requirements for the breaking of Ar–O and Ar–N bonds. Furthermore, gas showed an increasing trend and reached a maximum at 600 °C. They accounted for 6.9% of the total volatiles during the whole degradation process. Overall, it should be noted that not all the primary degradation products were released directly in the form of gas, some were kept at the end of the molecular chain and may be released gradually as temperature rises. The formation and volatilization of the secondary degradation products need higher temperatures. So, as shown in Table 3, most of them were detected at 500 °C and 600 °C except for small amounts of Schiff and Pd derivatives. The Schiff bases appeared only at 350 °C which are probably formed by the recombination of amines and phenols containing aldehyde. While the phenanthridine (Pd) derivatives detected at each temperature were the most abundant of the secondary degradation products with a TMP of 7.9%. They are probably generated from further degradation of the Mannich base compounds. Moreover, the Biph compounds were detected at 400, 500 and 600 °C with a similar EMP value. 2,3-Benzofuran (Bf) and isoquinoline (Iq) derivatives were found at 500 °C or a higher temperature. Overall, the secondary degradation products were more abundant at high temperatures, but their absolute quantities were very low because of less weight loss of polymer in this temperature stage. Compared with the primary products throughout the pyrolysis process, the total mass percentage of the secondary degradation products was less and only accounted for 13%. This is because most of the secondary products contain a thermally stable aromatic heterocyclic or polymeric structures with a large molecular weight and a high boiling point. These species in bulk are vital to form chars.

In addition, several benzophenone (Bh) derivatives containing carbonyl groups were generated mainly at 400 and 500 °C with a total mass percentage of 2.4%. They do not come from primary or secondary degradation processes, but are rather likely to come from direct fracture of the original structures of polybenzoxazine. So we inferred that the benzophenone structures existed originally in polybenzoxazine.

3.3 Thermal degradation process of poly-A

Time and temperature information was introduced in the form of a stepwise-temperature testing method using Py-GC/MS. Based on this information, we know the chemical structures and their respective contents of degradation products at different temperatures. Furthermore, combined with the information of small molecule gas obtained from the TGA-FTIR results, polymer structures and the degradation process of poly-A could be speculated.

The chemical structure of poly-A has been inferred by investigating the polymerization process in a previous study by our group.¹⁷ As supplement, the results of studying the thermal degradation process were used to further speculate the structure from the other side. As shown in Fig. 7, the Mannich bridge is a basic structure of polybenzoxazine, which is closely related to the formation of primary and secondary degradation products of polybenzoxazine. Beyond this, phenolic compounds containing an aldehyde group could be detected in Py-GC/MS, indicating that poly-A still contains some aldehyde groups without participating in any of the reactions.

Fig. 7 Chemical structure of poly-A and formation of benzophenone structure.

Furthermore, hydroxymethyl groups and carboxyl groups also exist in poly-A. These two groups are obtained through reductive or oxidation reactions of aldehyde groups. Additionally, there exists a special crosslinking structure containing benzophenone species in poly-A, which is closely related to aldehyde groups. It is reasonable to assume that some carboxyl groups can be generated from the oxidation of aldehyde groups. It is believed that some of the carboxyl groups take part in the decarboxylation reactions to create new reactive crosslinking points on the para positions of phenolic hydroxyls. Other carboxyl groups, which are not involved in any of the reactions, may react with the ortho or para position of phenolic hydroxyls to generate benzophenone structures with removal of water. Besides that, the para positions of aniline may be also be involved in this reaction. Formation of these structures may also occur in the early stages of degradation. The small weight loss at 310 °C in TGA is probably attributed to the decarboxylation and dehydration reactions.

The thermal degradation process of poly-A is outlined in Fig. 8 and is divided into four parts. The evolution of bulk structures is presented on the left and the variation of volatiles is shown on the right.

Fig. 8 Evolution of gas and bulk during thermal degradation of poly-A.

During the temperature range 350–400 °C of the thermal degradation process, the weakest C–N and Ar–C bonds in polybenzoxazine break first. The molecular weight rapidly decreases and lots of small molecule radicals are generated. These radicals react with hydrogen to form homologous primary degradation products such as amines, phenolic compounds and Mannich base. This is the main reason for the weight loss. Due to the generation of more thermally stable structures, like carbonyl groups on phenolic rings from the crosslinking reactions of the aldehyde groups as mentioned above, the amount of phenolic compounds is really small. The weak bonds adjacent to the benzophenone structure may break to release a small amount of benzophenone compounds, and most of them still remain in bulk. Moreover, some polymer chain fragments terminated with primary degradation products have enough good thermal stability and also are kept in bulk.

In the temperature range 400–500 °C, polymer chain fragments further degrade to release amines, phenolic compounds and other primary products. The rising temperature increases the breaking possibility of adjacent chemical bonds, so Ar–O and Ar–N bonds may be broken and create the ammonia and benzene derivatives. At the same time, some primary degradation products can either combine with each other to form the secondary products like biphenyl compounds and phenanthridine derivatives or degrade further. Those polymer chain fragments terminated with primary products can also react like this and make the species of biphenyl and phenanthridine to remain in the bulk. These species are vital to the carbonization reaction at higher temperatures.

In the temperature range 500–600 °C, the release of primary products becomes slow. This means that the basic skeleton of Mannich bridges in polybenzoxazine nearly disappears. The secondary products recombine with each other to generate aromatic heterocyclic or fused rings structures. The dehydrogenation and aromatization of these structures carries on further. Additionally, high temperatures cause Ar–O bonds in the benzophenone structure to break and release “locked” phenolic species.

Above 600 °C, the weight loss of the bulk almost reaches equilibrium. Aromatization and dehydrogenation in the bulk are the main reactions at this temperature range, and the char is formed finally.

4. Conclusion

The thermal degradation of polybenzoxazine-containing aldehyde groups was studied by TGA-FTIR using a stepwise-temperature testing method based on Py-GC/MS. The entire process of thermal degradation was discussed in detail and the crosslinking structures of poly-A were further proved from the perspective of the degradation. The results indicate that special crosslinking benzophenone structures caused by aldehyde groups exist in poly-A. These extra crosslinking structures improve the thermal stability of poly-A. During the thermal degradation process, the primary degradation products include amines, Mannich base and phenolic compounds were detected and the release of phenolic compounds was significantly delayed. These primary products and polymer chain fragments terminated with these species can recombine with each other to generate secondary degradation products or degrade further. The secondary degradation products and their species are vital to form chars at higher temperature. In addition, the segments containing benzophenone structures may degrade at a high temperature to release carbon monoxide and unlock the phenolic species at the same time.

Acknowledgements

This work was supported by the Research Foundation of Sichuan University, Fundamental Research Funds for the Central Universities (no. 2013SCU04A27) and financially supported by State Key Laboratory of Polymer Materials Engineering (Grant no. sklpme2014-3-13).

References

1. X. Ning and H. Ishida, J. Polym. Sci., Part A: Polym. Chem., 1994, 32, 1121–1129.
2. H. Ishida and D. J. Allen, J. Polym. Sci., Part B: Polym. Phys., 1996, 34, 1019–1030.
3. Y.-X. Wang and H. Ishida, Polymer, 1999, 40, 4563–4570.
4. K. Hemvichian and H. Ishida, Polymer, 2002, 43, 4391–4402.
5. H. Zhang, W. Gu, R. Zhu, Q. Ran and Y. Gu, Polym. Degrad. Stab., 2015, 111, 38–45.
6. T. Agag and T. Takeichi, Macromolecules, 2001, 34, 7257–7263.
7. Z. Brunovska, R. Lyon and H. Ishida, Thermochim. Acta, 2000, 357, 195–203.
8. G. Cao, W. Chen and X. Liu, Polym. Degrad. Stab., 2008, 93, 739–744.
9. H. Ishida and S. Ohba, Polymer, 2005, 46, 5588–5595.
10. H. Kim, Z. Brunovska and H. Ishida, Polymer, 1999, 40, 6565–6573.
11. Y. L. Liu and C. I. Chou, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 5267–5282.
12. Y. L. Liu and J. M. Yu, J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 1890–1899.
13. Y. L. Liu, J. M. Yu and C. I. Chou, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 5954–5963.
14. J. Wang, M.-q. Wu, W.-b. Liu, S.-w. Yang, J.-w. Bai, Q.-q. Ding and Y. Li, Eur. Polym. J., 2010, 46, 1024–1031.
15. F. Zuo and X. Liu, J. Appl. Polym. Sci., 2010, 117, 1469–1475.
16. Q. Ran, Q. Tian and Y. Gu, Chin. Chem. Lett., 2006, 17, 1305–1308.
17. Q. Ran and Y. Gu, J. Polym. Sci., Part A: Polym. Chem., 2011, 49, 1671–1677.
18. H. Y. Low and H. Ishida, J. Polym. Sci., Part B: Polym. Phys., 1998, 36, 1935–1946.
19. H. Y. Low and H. Ishida, J. Polym. Sci., Part B: Polym. Phys., 1999, 37, 647–659.

---