Flame-retardant properties of ethylene-vinyl acetate/oil sludge/ fumed silica composites

Abstract

Ethylene-vinyl acetate (EVA)/oil sludge (OS)/fumed silica (SiO₂) composites were synthesized using OS containing CaCO₃ as raw material. Flame retardant and thermal degradation properties of the composites were characterized via limiting oxygen index (LOI), cone calorimetry test (CCT), scanning electronic microscopy (SEM), smoke density test (SDT), and thermogravimetry Fourier transform infrared spectrometry (TG-FTIR). Addition of a certain amount of SiO₂ could evidently increase LOI values. The heat release rates (HRR) of the composites were much lower than those of EVA and EVA/OS composites. The morphologies and structures of the residues, revealed by SEM, ascertained that the formed char layers on the composites were denser than those of the EVA/OS composites. SiO₂ in the material can help smoke suppression. The composites also assumed a higher thermal stability than the EVA/OS composites.

---

Introduction

Flame retardant EVA is widely used in wire and cable production.¹ As pristine EVA is very flammable because of its chemical composition, it is not suitable for many applications.² To improve the flame retardance of this material, one general way is to fill additives into EVA.³ Aluminum hydroxide (ATH) and magnesium hydroxide (MH) are used as flame retardants for EVA in most cases.^4,5 Layered double hydroxides (LDH) can be synthesized from Bayer red mud, and are also used as flame retardants for EVA. This can make use of Bayer red mud, which is a waste formed in alumina production.^6,7 Oil sludge (OS), another waste, was studied in the laboratory of Yi Qian to be turned into a flame retardant for EVA.

OS is mainly generated from the bottom of a tank, e.g., grease trap sediment, wastewater treatment plant residues, and the remainder of activated sludge.⁸ 5 000 000 tons of OS are produced in China every year. It contains hundreds of compounds and will cause serious environmental pollution if discharged without any treatments.^9,10 There are many methods to treat and utilize oil sludge, including oil sludge fueling treatment, bioremediation of oil sludge, and oil recovery from oil sludge.^8,11,12 Oil sludge containing CaCO₃ is the side-product from treatment of oilfield produced water, with great amount of production. This type of oil sludge contains so much CaCO₃ that it may be used as a substitute for CaCO₃ in some applications. Several literature reported that CaCO₃ could be used as a flame retardant.^13–16 Few investigations on OS, to the best knowledge of authors, have used it in a flame retardant system. This paper may stimulate more interest of using OS containing CaCO₃ as flame retardant.

Fumed silica (SiO₂) is usually used as an enhancing agent to increase polymers' flame-retardant properties. The synergistic effect of SiO₂ were explored with other flame retardants.^17–21 In this paper, the flame retardant and thermal properties of the EVA/OS/SiO₂ composites were studied using various techniques.

Experimental

Materials

EVA18 (18 wt% vinyl acetate) was purchased from Beijing Eastern Petrochemical Co., Ltd. (China). OS was kindly supplied by Zhongyuan Oil Field as a sponsored material. Its composition was C (11.00 wt%), O (43.92 wt%), N (1.84 wt%), H (1.64 wt%), Na (3.48 wt%), Mg (1.70 wt%), Al (0.74 wt%), Si (8.34 wt%), S (1.03 wt%), Cl (5.30 wt%), Ca (6.66 wt%), Fe (5.66 wt%), and Cu (1.02 wt%), obtained by an INCA energy X-ray energy spectrometer (EDS, Oxford) after roasted at 550 °C for 4 h. XRD data were gained at room temperature on a Philips X'Pert Pro Super apparatus (Nicolet Instrument Co., Madison, WI) using Cu K_α radiation with a nickel filter (wavelength = 1.5418 Å) at a scan rate of 0.0167° s⁻¹. The XRD spectrum of the OS material contained typical peaks of CaCO₃, ascribed to diffraction peaks of (012), (104), (110), (113), (202), (024), (116), (122), (214), and (300).^22,23 The strong and sharp peaks indicated that the OS was holding large amount of CaCO₃. Fumed SiO₂ was purchased from Shanghai Dixiang Chemical Co., Ltd. (China). Other reagents were standard laboratory reagents and used as received without further purification.

Sample preparation

Preparation of the EVA composites. All components were melt-compounded at 120 °C for 10 min using a mixer. The mixtures were compression-molded into sheets at 120 °C under a pressure of 10 MPa for 10 min, and cut into suitable size for fire testing. The constitution of the prepared composites are listed in Table 1.

Table 1 Constitution of prepared composites

Sample No.	EVA/wt%	OS/wt%	SiO2/wt%
EVA	100.0	—	—
EOS1	50.0	50.0	—
EOS2	50.0	49.0	1.0
EOS3	50.0	47.0	3.0
EOS4	50.0	45.0	5.0

---

Measurements

Limiting oxygen index (LOI). LOI was measured according to ASTM D 2863 on an HC-2 oxygen index meter (Jiangning Analysis Instrument Company, China). The specimens used for the test were of dimension as 100 × 6.5 × 3 mm.

Cone calorimeter test (CCT). Cone calorimeter (Stanton Redcroft, UK) tests were performed according to ISO 5660 standard procedures. Each specimen with a dimension of 100 × 100 × 4 mm, wrapped in an aluminum foil, was horizontally exposed to an external heat flux of 50 kW m⁻².

Scanning electron microscopy (SEM). The char residue of the samples were observed on a Hitachi X650 scanning electron microscope.

Smoke density test (SDT). The released smoke was characterized following ISO 5659-2(2006) on a smoke density test machine JQMY-2 (Jianqiao Co., Ltd. China). Each specimen with a dimension of 75 × 75 × 2.5 mm, also wrapped in aluminum foil, was horizontally exposed to an external heat flux of 25 kW m⁻² with or without applying a pilot flame.

Thermogravimetry-Fourier transform infrared spectrometry (TG-FTIR). The tests was performed from 30 to 900 °C at a linear heating rate of 20°C min⁻¹ under a nitrogen flow of 30 mL min⁻¹. The TG-FTIR instrument consists of a thermogravimeter (TG209 F1, Netzsch Instruments, Germany), a Fourier transform infrared spectrometer (Vertex70, Bruker Optics, Germany), and a transfer tube with an inner diameter of 1 mm connected to the TG and the infrared cell.

Results and discussion

Flame retardant properties of EVA/OS/SiO₂ composites

LOI values of the EVA/OS/SiO₂ composites. LOI value can be used to evaluate flame retardant properties of materials.²⁴ The LOI value of the EOS1 sample hiked from original 19.0% (pure EVA) to 23.3% in Fig. 1. And the LOI values of the EOS2–EOS4 samples were all higher than those of the EOS1 and pure EVA sample. It indicates the composites composed of EVA, OS, and SiO₂ is more reluctant to be fired. The LOI value of the EOS4 sample was the highest, the mixture of OS and SiO₂ achieving the best flame retardant performance.

Fig. 1 LOI values of all the samples.

CCT of the EVA/OS/SiO₂ composites. Besides LOI evaluation, cone calorimeter is also widely used to assess flammability of materials.²⁵ The results, in a small-scale test, were found to be well correlated with those obtained from large-scale fire tests and can predict combustion behaviors of materials in real fires.^24,26
HRR of the EVA/OS/SiO₂ composites. The HRR of the EVA/OS/SiO₂ composites with various loading of SiO₂ as a function of burning time are shown in Fig. 2, correlated data listed in Table 2.

Fig. 2 HRR curves of all the samples.

Table 2 Cone calorimeter test data

Sample code	PHRR (kW m^−2)	THR (MJ m^−2)	Time to ignition (s)	Time to PHRR (s)
EVA	2016.5	186.0	41	240
EOS1	450.0	148.2	42	395
EOS2	433.7	148.1	39	340
EOS3	438.6	152.0	47	125
EOS4	363.5	152.6	42	420

---

Pure EVA was fast burned after ignition and a sharp HRR peak appears at 240 s with a peak heat release rate (PHRR) as high as 2016.5 kW m⁻². The HRR curve of the EOS1 sample (with a PHRR of 450.6 kW m⁻²) obviously declined and prolonged with addition of OS compared to pure EVA. The possible reason for this is CaCO₃, Al₂O₃, MgO, and Fe₂O₃ in the OS can form a char layer. CaCO₃, not only as the char source, is also a gas source. It can decompose into CaO and CO₂ at high temperatures, and CO₂ dilutes the flammable gas.

With increasing amount of SiO₂, the HRR gradually decreased and the HRR of EOS4 (with a PHRR of 363.5 kW m⁻²) was the lowest and the time to the PHRR of EOS4 was the longest. Unfortunately, the first peaks of EOS2 and EOS3 were higher than that of EOS1 and the combustion time of EOS2 was shorter than EOS1. This indicated that an appropriate addition of SiO₂ could contribute to the depression in the flammability of the composites. It acted as a synergistic flame retardant of OS to retard the burning of the composites. SiO₂ can react with CaCO₃ at high temperatures, releasing CO₂ which can dilute the flammable gas. Also, SiO₂ has a large surface area and low density and tends to migrate near the regressing sample surface without sinking through the melt layer of the polymer during the gasification/burning process.²⁷ The accumulated SiO₂ will form a char layer by collaborating with Al₂O₃ and Fe₂O₃ in the OS, as a heat insulation barrier.²⁸ This layer can prevent heat transfer and transportation of degraded products between the melting polymer and open air. That is, burning is retarded.¹⁸

From Fig. 2, it can be seen the HRR curves of EOS1–EOS4 has two peaks. The times to PHRR of EOS1, EOS2, EOS4 appear in the second peak, while it appears in the first peak for EOS3. This is mainly due to the addition of SiO₂. The surface of EOS1 (without SiO₂) is well-distributed. When little SiO₂ is added, the surface will crack because CaCO₃ can react with a small amount of SiO₂. That makes EOS3 release more heat in the beginning. However, it will form a more compact structure of silicate when the content of SiO₂ is higher than 3 wt%. So, the time to PHRR of EOS3 decreases a lot.

The HRRs of all of the samples assume multiple separate peaks during the burning. It means that a gradual burning of the specimen through the thickness occurred after the initial char layers formed. This combustion feature of multiple HRR peaks was also reported by Wu²⁹ and Fu.¹⁸ It could be concluded that there are synergistic effects between OS and SiO₂.

MASS of the EVA/OS/SiO₂ composites. Fig. 3 presents the weight of the char residues. All composites containing flame retardant possessed a higher weight of char residues than that of the pure EVA. For EOS1–EOS4, the mass loss curves could be correlated to the HRR curves before 550 s. The char layer by the addition of SiO₂ would act as a protective barrier and can limit the entrance of oxygen to the substrate. Or the decomposition rate of the samples, at least, could be slowed down. When the time exceeded 550 s, the weight of the char residues decreased with more SiO₂. It can attributed to more CO₂ released from the reaction of SiO₂ and CaCO₃ at higher temperatures. This resulted more mass loss for EOS2–EOS4, especially for EOS4.

Fig. 3 Mass loss curves of all the samples.

THR of the EVA/OS/SiO₂ composites. The slope of a THR curve can be assumed as representation of the fire spread rate.³⁰ The THR energy of the samples (EOS1 to EOS4) in Fig. 4, i.e., their flame spread rates, has decreased significantly compared to that of the pure EVA (185.99 MJ m⁻²). In the EOS2 to EOS4 samples, parts of their polymer portion were protected, without complete burning.³¹ The THR of EOS4 is the lowest in the first 600 s. During burning, a char layer (shown in Fig. 6) formed on the surface of matrix by the addition of SiO₂. It resulted a thermal insulation and might provoke extinguishment of the flame. The layer could also prevent combustible gases from feeding the flame, and separate oxygen from the burning material.²⁹ The THR of EOS4 was higher than EOS1 for the longer burning time. It may be due to the crack initiated in the char layer at prolonged time. Overall, the addition of SiO₂ into the EVA/OS composites achieved a prolonged combustion time of the composites. In other words, the flame spread rate was suppressed by SiO₂.

Fig. 4 THR curves of all the samples.

SPR of the EVA/OS/SiO₂ composites. Evaluation of the fire performance often involves quantifying heat generation and smoke generation, as the rate of smoke production (SPR).³² It can be seen in Fig. 5 that all SPR curves became flat and the peak SPR (PSPR) values decreased for the samples added with flame retardant compared to the pure EVA. The trend of SPR curves changes in accordance with the HRR curves. The SPR value of the pure EVA was lower than the value of EOS1 from 50 to 150 s. After 150 s, the SPR value of the pure EVA quickly hiked and a sharp SPR peak appeared at 225 s. The SPR values of EOS1–EOS4 gradually decreased with more added SiO₂, and the reduction was most evident for EOS4. It can be attributed to a good cohesion of combustion residues and catalysis of charring generated from SiO₂.³² The smoke emission is retarded and delayed by the proper addition of SiO₂.

Fig. 5 SPR curves of all the samples.

Digital photos of residues of the EVA/OS/SiO₂ composites. The residues of all the samples after combustion in the cone calorimeter are presented in Fig. 6. The pure EVA was burnt out and the aluminum foil became visual. The other flame-retardant EVA composites remained as residues. The char residue of EOS1 did not expand and looked like a coherent residue. There were crack and collapse in the char residues in Fig. 6c and d, because of dramatic expansion in the combustion. The residue of EOS4 expanded the most, explained as the most addition of SiO₂. Its surface was not continuous and firmly compacted, suggesting permeation of the decomposed gases through the char layer. The curve profiles of all the samples were correlated well with the HRR, THR, and SPR results.

Fig. 6 Photographs after cone calorimeter test: (a) EVA, (b) EOS1, (c) EOS2, (d) EOS3, and (e) EOS4.

Scanning electron microscopy (SEM). The surfaces of char residues of EOS1 and EOS4 after cone calorimeter test were observed using SEM, as shown in Fig. 7.

Fig. 7 SEM images of the surface of the charred residues obtained from EOS1 and EOS4 samples.

The char residue of EOS1 possessed many holes. That is, heat and flammable volatiles could easily penetrate the char layer into the flame zone during burning. The layer could not effectively segregate oxygen from combustible gases. The char residue of EOS4 was compact and held less holes than EOS1. This type of char could prevent heat transfer between the flame zone and the burning substrate, protected the underlying material from further burning, and retard the pyrolysis of polymers.²⁵ The observation proved that a suitable amount of OS and SiO₂ can improve char structure and flame retardance. This evidence can explain why the HRR, THR, and SPR of EOS4 declined.

SDT of the EVA/OS/SiO₂ composites. The combustion behavior of the samples can be reflected by the cone calorimeter results. The smoke density test provides more detailed information on the smoke production. It is helpful to evaluate the smoke suppression character of the samples.¹⁷

Specific optical density (SOD) was used to quantify the amount of the smoke production. In Fig. 8A, when the pilot flame was not used in the test, EOS4 produced less smoke than EOS1 and the pure EVA. This indicates that a suitable amount of SiO₂ in the composites synergistically worked with OS to retard the smoke production. During the first 300 s, EOS1 and EOS4 produced more smokes than the EVA sample. This can also be due to the formation of the protective char layer. The production of smoke decreases thereafter, especially for EOS4. From 400 to 600 s, EOS1 suddenly produced plenty of smokes. The reason might be that the char layer was not compact and vulnerable. It became cracked at that time period.

Fig. 8 Luminous flux of all the samples: (A) without a pilot flame, (B) with a pilot flame.

In Fig. 8B, when the pilot flame was used in the test, EOS4 performed the best among all the samples. It is noted that EOS1 and EOS4 both produced more smokes in the first 250 s than the EVA. The char layer might not firmly form in such a short time. EOS4 produced less smoke than EOS1 because of the addition of SiO₂, rendering easier way to form the char layer. The amount of smoke production was reduced more than that without the pilot flame. This can be attributed to the solid particles formed from the depolymerization of the polymers. The polymers were mostly burned out, turned into gases. They, however, did not migrate directly into the gas phase to increase the amount of smoke.³³

SiO₂ can suppress the smoke production with OS when the samples are exposed to heat. The good performance of the EVA/OS/SiO₂ composites in SDT may come from decomposition of CaCO₃ and formation of the char layer. Combustion products such as tar and soot particles are limited in the contribution to the gas phase. The smoke density can be effectively reduced.³⁴

TG-FTIR characterization of the EVA/OS/SiO₂ composites.
TG behavior of the EVA/OS/SiO₂ composites. TG-FTIR analysis is always used to study the thermal degradation behavior of flame retardant materials. It can analyze the changes in evolved gases at various temperatures.^35,36 Fig. 9 shows the TG and DTG curves of samples EVA, EOS1, and EOS4 in an atmosphere of nitrogen.

Fig. 9 TG and DTG curves of the EVA, EOS1, and EOS4 composites.

Thermal stability of a polymeric material plays a significant role when it is used as a flame retardant. It mainly concerns about the release of decomposition products and formation of char. The EVA in Fig. 9 underwent two degradation steps. The first was the loss of carboxylic acid and the second involved in random chain scission of the remaining material, forming unsaturated vapor species, e.g., butene and ethylene.^37,38 There were three weight-loss steps for EOS1 and EOS4. The first step was loss of the absorbed water in the OS on the surface of the composites. The second and third steps were simultaneous dehydroxylation and decarbonation of the OS, overlapped by decomposition of acetate groups in side chains and scission of main chains of EVA.^37,38

It is noted that EOS1 assumed a lower decomposition rate in the third step and a higher one in the first and second step than that of EVA. The incorporation of OS lowered the decomposition rate in the third step. It also accelerated the loss of carboxylic acid. When the temperature was higher than 700 °C, the EVA sample left no residue and EOS1 and EOS4 both left 30%. EOS4, with both SiO₂ and OS, exhibited higher thermal stability in all the degradation process than EOS1. As a ternary composite, EOS4 carried a better flame retardance than EVA/OS composites. This result confirms the synergistic effects. It also was ascertained by the above cone calorimeter results. Char morphology plays a more vital role in flame retardant property than the quantity of a char, the same conclusion being found in Weil and Pate's study.³⁹

FTIR characterization of the EVA/OS/SiO₂ composites. 3D TG-FTIR spectra of pyrolysis products of the composites during thermal degradation are presented in Fig. 10. The evolved gas products for the three samples assumed characteristic bands in 3400–4000 cm⁻¹, 2800–3150 cm⁻¹, 2250–2400 cm⁻¹, 1700–1850 cm⁻¹, 1250–1500 cm⁻¹, and 950–1150 cm⁻¹, fitted to H₂O (3400–4000 cm⁻¹), CO₂ (2300–2400 cm⁻¹), CO (2250–2300 cm⁻¹), carboxylic acid (1700–1850 cm⁻¹), and aliphatic hydrocarbons (2800–3150 cm⁻¹, 1250–1500 cm⁻¹, and 950–1150 cm⁻¹),^39–41 respectively. Depolymerization is as known as a process associated with thermal degradation of polymers. The main decomposition products of the composites in this research were listed as above.

Fig. 10 3D TG-FTIR spectra of the pyrolysis products of (a) EVA, (b) EOS1, and (c) EOS4 during thermal degradation.

It can be noticed that the depolymerization processes of the three samples were significantly different from their pyrolysis products after thermal degradation. The pure EVA sample decomposed drastically because it is flammable without any addition of flame retardants. It released large amount of carboxylic acids and aliphatic hydrocarbons in its decarboxylation and main chain break-down. The decomposition of EOS1 containing OS obviously slowed down. When SiO₂ was added into the composites, the decomposition of EOS4 was slightly accelerated.

The characteristic spectra obtained from 30 °C to 800 °C are shown in Fig. 11. There were almost no infrared signals below 250 °C, indicating no decomposition under that temperature. Beyond 250 °C, the release of CO, CO₂, and H₂O could be detected. When the temperature was about 350 °C, a maximum signal intensity at 1700–1850 cm⁻¹, reflecting evolution of carboxylic acid, can be observed. A maximum signal at 2800–3150 cm⁻¹, related to aliphatic hydrocarbons, appeared at 460 °C. The signal intensity of the pyrolysis products declined gradually above 460 °C, implying a slower decomposition rate of the mixture. The signals of EOS1 and EOS2 involving aliphatic hydrocarbons and carboxylic acid were less than the pure EVA. More detailed information about the pyrolysis products of the composites is shown in Fig. 12.

Fig. 11 FTIR spectra of pyrolysis products of (a) EVA, (b) EOS1, and (c) EOS4 at various temperatures.

Fig. 12 Variation of (a) H₂O, (b) CO₂, (c) CO of the composites as a function of temperature cracking in nitrogen at 20 °C min⁻¹.

The releasing of H₂O for the pure EVA, EOS1 and EOS4 were in two steps, as in Fig. 12a. The first step began at about 330 °C and reached its first peak at 370 °C. The second step was above 500 °C, where the releasing of H₂O was at a high level. This result correlates well with the TG–DTG results. EOS1 and EOS4 evidently released less water than the pure EVA and EOS4 released the least water. When SiO₂ and OS were added into the composites, a char formed to protect the composites from burning, decomposition of EVA being slowed down.

In Fig. 12b, no peaks were found in the releasing of CO₂ for the pure EVA until 500 °C. A peak can be seen at about 370 °C for both EOS1 and EOS4. This peak would mainly be caused by CO₂ when the CaCO₃ containing in OS was heated. The CO₂ releasing in EOS4 was much less than that in EOS1. It can be attributed the reduced amount of OS, therefore, less CO₂.

It is clear that the release of CO was significantly reduced when OS was added into EVA, as in Fig. 12c. Less EVA may be one reason. The main reason should be the increased CO₂. The ternary composite EOS4 produced least CO. It can be explained by more aliphatic hydrocarbons and carboxylic acid released from EOS4 than in EOS1 (Fig. 11).

Conclusions

The flammability and smoke suppression characteristics of EVA/OS/SiO₂ composites had been compared with those of EVA and EVA/OS by LOI, CCT, SEM and SDT analysis. The results showed that SiO₂ had synergistic effects with OS when used in EVA. When 5 wt% SiO₂ was added into the composites, the LOI value increased obviously. The CCT data indicated that the HRR, THR, and SPR of the EOS4 (containing 5 wt% SiO₂) was almost the lowest. The SDT results showed that the composites containing both OS and SiO₂ produced less smoke than the EVA/OS composites and pure EVA for a period of time, although at last the SOD of all samples had no many differences. The TG-IR results indicated that the thermal stability of the EVA/OS/SiO₂ composite was improved and the mechanism of the synergistic effects between OS and SiO₂ may mainly depend on the condensed phase process.

Acknowledgements

The authors gratefully acknowledge the National Natural Science Foundation of China (No. 51372129 and No. 51572138).

References

1. S. T. Bee, A. H. San, C. T. Ratnam, T. T. Tee and L. T. Sin, Polym. Compos., 2012, 33, 1883–1892.
2. C. Nyambo and C. A. Wilkie, Polym. Degrad. Stab., 2009, 94, 506–512.
3. S. Fang, Y. Hu, L. Song and J. Wu, J. Appl. Polym. Sci., 2009, 113, 1664–1670.
4. C. M. Jiao and X. L. Chen, J. Appl. Polym. Sci., 2010, 116, 1889–1893.
5. N. H. Huang, Z. J. Chen, C. H. Yi and J. Q. Wang, eXPRESS Polym. Lett., 2010, 4, 227–233.
6. L. Li, Y. Qian and C. M. Jiao, Iran. Polym. J., 2012, 21, 557–568.
7. L. Li, Y. Qian and C. M. Jiao, J. Therm. Anal. Calorim., 2013, 114, 45–55.
8. G. C. Jiang, S. X. Xie, M. Chen, R. S. Wang, Z. Y. Li, H. Mao and L. F. Zeng, Pet. Sci. Technol., 2013, 31, 174–184.
9. G. L. Jing, M. M. Luan and T. T. Chen, Environ. Prog. Sustainable Energy, 2013, 32, 99–102.
10. B. Hou, S. X. Xie, M. Chen, Y. Jin, D. Hao and R. S. Wang, Pet. Sci. Technol., 2013, 31, 458–464.
11. M. Dhote, A. Juwarkar and A. Kumar, ABS J. Sustain. Biotech., 2013, 1, 1–18.
12. Y. Q. Jin, X. Y. Zheng, X. L. Chu, Y. Chi, J. H. Yan and K. F. Cen, Ind. Eng. Chem. Res., 2012, 51, 9213–9217.
13. A. Hermansson, T. Hjertberg and B. Sultan, J. Appl. Polym. Sci., 2006, 100, 2085–2095.
14. Y. H. Jiao and J. Z. Xu, J. Appl. Polym. Sci., 2009, 112, 36–43.
15. L. Karlsson, A. Lundgren, J. Jungqvist and T. Hjertberg, Polym. Degrad. Stab., 2009, 94, 527–532.
16. J. Z. Xu, Y. H. Jiao, B. Zhang, H. Q. Qu and G. Z. Yang, J. Appl. Polym. Sci., 2006, 101, 731–738.
17. T. Kashiwagi, J. W. Gilman, K. M. Butler, R. H. Harris, J. R. Shields and A. Asano, Fire Mater., 2000, 24, 277–289.
18. M. Z. Fu and B. J. Qu, Polym. Degrad. Stab., 2004, 85, 633–639.
19. F. Yang, I. Bogdanova and G. L. Nelson, Polym. Adv. Technol., 2008, 19, 602–608.
20. F. Yang and G. L. Nelson, Polym. Adv. Technol., 2006, 17, 320–326.
21. L. C. Du, Y. C. Zhang, X. Y. Yuan and J. Y. Chen, Polym.–Plast. Technol. Eng., 2009, 48, 1002–1007.
22. X. Q. Li, Z. W. Feng, Y. Y. Xia and H. C. Zeng, Chem.–Eur. J., 2012, 18, 1945–1952.
23. M. Ni and B. D. Ratner, Surf. Interface Anal., 2008, 40, 1356–1361.
24. C. M. Jiao and X. L. Chen, J. Therm. Anal. Calorim., 2009, 98, 813–817.
25. X. L. Chen and C. M. Jiao, Polym. Eng. Sci., 2010, 50, 767–772.
26. X. L. Chen, C. M. Jiao and J. Zhang, Polym. Eng. Sci., 2012, 52, 414–419.
27. Y. Tang and M. Lewin, Polym. Degrad. Stab., 2008, 93, 1986–1995.
28. C. M. Jiao and X. L. Chen, J. Appl. Polym. Sci., 2011, 120, 1285–1289.
29. X. F. Wu, L. C. Wang, C. Wu, G. L. Wang and P. K. Jiang, J. Appl. Polym. Sci., 2012, 126, 1917–1928.
30. X. L. Chen and C. M. Jiao, Polym. Adv. Technol., 2010, 21, 490–495.
31. X. Y. Li, T. R. Tan, X. P. Hu, K. Wang and Q. Fu, Polym. Eng. Sci., 2013, 53, 410–416.
32. L. C. Wang and P. K. Jiang, J. Appl. Polym. Sci., 2011, 119, 2974–2983.
33. L. Shi, D. Q. Li, S. F. Li, J. R. Wang, D. G. Evans and X. Duan, Chin. Sci. Bull., 2005, 50, 1101–1104.
34. T. M. Shukri, J. Mosnáček, A. A. Basfar, M. A. Bahattab, P. Noireaux and A. Courdreuse, J. Appl. Polym. Sci., 2008, 109, 167–173.
35. S. Hirose, K. Kobashigawa, Y. Izuta and H. Hatakeyama, Polym. Int., 1998, 47, 247–256.
36. J. Salgado and M. I. Paz-Andrade, J. Therm. Anal. Calorim., 2009, 95, 837–842.
37. C. X. Zhao, Y. Liu, D. Y. Wang, D. L. Wang and Y. Z. Wang, Polym. Degrad. Stab., 2008, 93, 1323–1331.
38. M. Zanetti, T. Kashiwagi, L. Falqui and G. Camino, Chem. Mater., 2002, 14, 881–887.
39. E. D. Weil and N. G. Patel, Polym. Degrad. Stab., 2003, 82, 291–296.
40. V. G. Gregoriou, G. Kandilioti and S. T. Bollas, Polymer, 2005, 46, 11340–11350.
41. Y. J. Chen, J. Zhan, P. Zhang, S. B. Nie, H. D. Lu, L. Song and Y. Hu, Ind. Eng. Chem. Res., 2010, 49, 8200–8208.

---