Investigating into the liquid oxygen compatibility of a modified epoxy resin containing silicon/phosphorus and its mechanical behavior at cryogenic temperature

Abstract

A 9,10-dihydro–9–oxa–10–phosphaphenanthrene–10–oxide (DOPO) derivative (DOPO–TVS) was synthesized through a reaction between DOPO and triethoxyvinylsilane (TVS). To modify the common epoxy resin molecular without consuming the epoxy group, the general bisphenol F epoxy resin was first treated with isocyanate propyl triethoxysilane (IPTS). In the next step, the pretreated epoxy resin and DOPO–TVS were mixed to initiate the sol–gel process to generate the organic–inorganic hybrid Si–O–Si network within the epoxy matrix. The characterization of each reaction product was confirmed by Fourier transform infrared (FTIR) spectroscopy and nuclear magnetic resonance (¹H NMR, ³¹P NMR) spectroscopy. The liquid oxygen compatibility of the cured epoxy resin was evaluated through mechanical impact in accordance with ASTM D2512-95. The surface elemental composition of the specimen before and after mechanical impact was investigated by X-ray photoelectron spectroscopy (XPS). The results of liquid oxygen mechanical impact showed that the liquid oxygen compatibility of the silicon/phosphorus containing epoxy resin was obviously enhanced. Moreover, the surface element composition also confirms the migration of the silicon to the surface of the cured epoxy resin and the generation of phosphoric oxyacid, which belongs to the condensed phase flame retardant mechanism. The tensile test and fracture toughness test under cryogenic conditions (77 K) were also carried out. The results showed that the modified epoxy resin possesses an elevated toughness property, which was attributed to the flexible Si–O–Si network in the cured modified epoxy resin.

---

1. Introduction

Epoxy resins continue to be primary candidates of the matrix for composite materials used in many applications,¹ especially in space applications wherein the payoff from weight saving is typically the highest.² It would make a great contribution to the weight saving to use polymer matrix composites for the cryogenic propellant tanks on launch vehicles instead of the traditional aluminum and stainless steel. Study shows that the total tank weight could be decreased by as much as 27–35% if the polymer matrix composites are used as the tank material instead of metals.³ In most respects, epoxy resins are chosen as the matrix of composites for use in cryogenic tanks because of their easy handling and processability, low shrinkage after curing, and mechanical properties.^4–7

The two most important issues for epoxy resin as a lighter weight option for the cryogenic liquid oxygen tank are its poor compatibility with liquid oxygen and brittle behavior under cryogenic conditions. This may ignite or initiate reactions when being in contact with liquid oxygen and subjected to external stimuli such as mechanical impact, frictional heating, and static electricity.^2,8 Unfortunately, little study concerning these two issues at the same time has been reported. To make the epoxy resin a safe option for the matrix material of liquid oxygen tank, it should be modified to maintain its original properties in the presence of liquid or high concentration gaseous oxygen. Studies have suggested that the oxidation reactions may be the key factors for the incompatibility for the epoxy resin with liquid oxygen and enhancing the flame retardancy of epoxy resins could be an effective method to enhance the compatibility.^9,10 This has been proven to be an effective way of introducing a flame retardant element to improve the liquid oxygen compatibility of epoxy resins.

9,10-Dihydro-9-oxa-10-phosphaphenanthrene-10-oxide (DOPO) is a type of widely used phosphorus-containing flame retardant with high reactivity and good flame retardant performance. Tremendous achievements on enhancing the flame retardancy have been reached by incorporating DOPO and its derivatives into epoxy resin. However, unmodified DOPO consumes the epoxy group of the epoxy resin in high temperature during the curing process, which will affect the mechanical properties of the resin.^11–13

The dissociation energy of the Si–O bond is 460 kJ mol⁻¹, which is higher than that of the C–C bond (332 kJ mol⁻¹) and C–O bond (326 kJ mol⁻¹). The high thermal stability of the Si–O bond can play a role of improving the oxidation resistance of the polymer during thermal decomposition.¹⁴ During thermal degradation of the polymer materials, the silicon element interacts with the polymer matrix and migrates to the surface of the matrix due to its low surface energy and finally forms a silicon-rich surface carbon layer.¹⁵ In addition, the Si–O bond could improve the flexibility and impact strength of the polymer due to its low rotation barrier energy and high flexibility.¹⁶ Numerous achievements have been made to enhance the mechanical behavior of epoxy material through various ways of introducing silicon-containing substance such as nanosilica, polyhedral oligomeric silsesquioxane (POSS), and siloxane.^17–19 Therefore, it is reasonable to introduce Si–O to enhance the mechanical behavior of epoxy resins at cryogenic temperatures, which will contribute a lot to improving the structural safety of the composite liquid oxygen tank.^20–22

To further enhance the oxidation resistance, introducing phosphorus within silicon-containing epoxy resins to bring a P/Si synergistic effect is a promising method. During oxidative decomposition, phosphorus provides a tendency for char formation, and silicon enhances the thermal stability of the char. Thus, introducing both elements may combine these two advantages, which will lead to strong oxidation resistance.

In this study, a trisiloxane (DOPO–TVS) containing silicon and phosphorus was synthesized through a reaction between DOPO and TVS to enhance the liquid oxygen compatibility of epoxy resin (F51) and its mechanical behavior in cryogenic temperature. To form the Si–O–Si network within the original curing resin without consuming the epoxy group, the traditional epoxy resin F51 was pretreated with (3-isocyanatopropyl)-triethoxysilane (IPTS). In the last step, the Si–O–Si network was obtained via a sol–gel reaction of DOPO–TVS and the pretreated resin (F51–IPTS).

2. Experimental

2.1 Materials

Bisphenol F epoxy resin (F51, epoxy value 0.53–0.57) was purchased from Feicheng Deyuan Chemicals Co. Ltd. 2,2′-Azobisisobutyronitrile (AIBN) and 4,4′-diaminodiphenylmethane (DDM) were purchased from TianJin Guang Fu Fine Chemical Research Institute. 9,10-Dihydro-9-oxa-10-phosphaphenanthrene-10-oxide (DOPO, industrial grade, 97%) was purchased from Huizhou Sun Star Technology Co. Ltd. China. Triethoxyvinylsilane (TVS) was obtained from AB Specialty Silicones. (3-Isocyanatopropyl)-triethoxysilane (IPTS) was supplied by Energy Chemical Co. Ltd. Dibutyltindilaurate (DBTDL) was purchased from Adamas Reagent Co. Ltd. All other materials were analytically pure and used without any treatment.

2.2 Synthesis of the trisiloxane DOPO–TVS containing phosphorus element

21.6 g of DOPO (0.1 mol), 21 g of TVS (0.11 mol) and 30 mL of toluene were mixed in a three-neck flask equipped with a stirrer, reflux condenser. The mixture was then heated to 80 °C to allow DOPO to dissolve. 0.2 g of AIBN (5000 ppm) dissolved in 20 mL methylbenzene was added dropwise to the flask with a drop funnel in about 2 hours. The temperature was maintained to 80 °C for another 2 hours. Finally, light yellow transparent viscous liquid was obtained (which is symbolized with DOPO–TVS) after methylbenzene and excess TVS were distilled off under reduced pressure. The product of this section is symbolized with DOPO–TVS. A schematic of the synthesis of DOPO–TVS is demonstrated in Scheme 1.

Scheme 1 Schematic of synthesis process.

2.3 Pretreatment process of the epoxy resin

120 g of F51, 12 g of IPTS and 0.4 g of DBTDL were blended in a 250 mL three-neck flask equipped with a stirring bar in a N₂ atmosphere. The mixture was then heated to 50 °C. The temperature remained at 50 °C until the characteristic infrared absorption peak of isocyanato at 2271 cm⁻¹ disappeared and a colorless transparent viscous pretreated epoxy resin was obtained (symbolized with IPTS–F51).²³ This process is shown in Scheme 1.

2.4 Synthesis of the final silicon–phosphorus containing EP

IPTS–F51 and DOPO–TVS were blended at different ratios. Acetone was used as a solvent. Distilled water was added to initiate the sol–gel process. The molar ratio of water to ethoxy was 1.5 : 1.²⁴ Acetic acid was added as a catalyst for the hydrolysis reaction to adjust the pH to 3–4.²⁵ The mixture was heated to 50 °C and remained for 5 hours. Finally, the mixture was distilled under reduced pressure to accelerate the condensation reaction and remove water and the solvent. Finally, light yellow transparent viscous silicon–phosphorus containing epoxy resin was obtained. This process is shown in Scheme 1. The final product is symbolized with EPX in which X varies with the proportion of IPTS–F51 and DOPO–TVS. Detailed contents of each ingredient is listed in Table 1.

Table 1 Content of each component

Specimen ID	IPTS–F51 (%)	DOPO–TVS (%)	Si (%)	P (%)	Water (%)	DDM (%)
Pure F51	—	—	—	—	—	27.7
EP5	95	5	4.94	0.38	1.78	23.7
EP7.5	92.5	7.5	4.99	0.56	1.96	23.0
EP10	90	10	5.14	0.75	2.49	22.4
EP15	85	15	5.21	1.13	2.51	21.2

---

2.5 Curing process of the silicon–phosphorus containing the epoxy resin

DDM and the silicon–phosphorus containing epoxy resin were blended at the molar ratio that one amino hydrogen corresponds to one epoxy group. The mixture was stirred vigorously at 80 °C until DDM had completely dissolved. The liquid mixture was poured into the mold after vacuum degassing and cured in an oven under the process of 100 °C (2 h)–130 °C (2 h)–160 °C (4 h). Finally, the mold was cooled gradually to room temperature and silicon–phosphorus containing epoxy resin was obtained. The detailed content of each ingredient is listed in Table 1.

2.6 Measurements

Fourier transform infrared spectroscopy was carried out to investigate the synthesis outcome of IPTS–EP with a Perkin Elmer Spectrum One FTIR. The infrared spectrum was obtained in the wavenumber ranged from 650 to 4000 cm⁻¹. ¹H NMR and ³¹P NMR were used to confirm the synthesis of DOPO–TVS on Varian INOVA at a frequency of 400 MHz and 202 MHz, respectively, in CDCl₃. The liquid oxygen compatibility of the cured epoxy resin was determined under an impact energy using the Army Ballistic Missile Agency-type impact tester, as described in ASTM D2512-95 (2008). The surface elemental composition of the epoxy resin before and after liquid oxygen impact was determined by X-ray photoelectron spectroscopy (XPS) using an X-ray photoelectron spectrometer (Perkin-Elmer, PHI 5300) equipped with a magnesium X-ray source. The tensile test and fracture toughness test under cryogenic conditions at liquid nitrogen temperature (77 K) were carried out with a 50 kN capacity servo-hydraulic testing machine with a displacement rate of 1.0 mm min⁻¹ following ASTM D638-99 and ASTM D5045-14, respectively. The ultra-low temperature was obtained through spraying liquid nitrogen into a sealed cryostat. The mechanical test was carried out immediately when the temperature dropped to 77 K.

3. Results and discussion

3.1 Characterization of DOPO–TVS, IPTS–EP, and silicon–phosphorus containing epoxy resin

The reaction product of DOPO–TVS was characterized by ¹H NMR and ³¹P NMR. The ¹H NMR and ³¹P NMR spectra of DOPO and DOPO–TVS are shown in Fig. 1 and 2, respectively. Peak ‘A’ (δ 8.65) in Fig. 1a is the characteristic peak of hydrogen in the ‘P–H’ bond, which disappears in Fig. 1c, indicating the complete consumption of DOPO.²⁶ Multiplet ‘C’ (δ 6.18–5.80) in Fig. 1b is the characteristic peak of atom hydrogen of the ‘C=C’ bond in TVS, which disappears in Fig. 1c, indicating the complete reaction of TVS.²⁷ Moreover, two new multiplets (F and G) are detected in Fig. 1c, which should correspond to the newly generated hydrogen in DOPO–TVS. To confirm the degree of the reaction of DOPO–TVS, the ³¹P NMR spectrum was obtained, as shown in Fig. 2. There was only 1 peak in either Fig. 2a or b along with an obvious chemical shift from δ 14.65 to δ 39.53, which is further proof of the reaction.

Fig. 1 ¹H NMR spectrum of DOPO (a), TVS (b) and DOPO–TVS (c).

Fig. 2 ³¹P NMR spectrums of DOPO (a) and DOPO–TVS (b).

Fig. 1a: ¹H NMR (400 MHz, CDCl₃) (A) δ 8.65 (s, 1H), (B) δ 8.11–7.13 (m, 17H).

Fig. 1b: ¹H NMR (400 MHz, CDCl₃) (C) δ 6.18–5.80 (m, 1H), (D) δ 3.82 (dd, J = 6.9, 0.8 Hz, 2H), (E) δ 1.20 (td, J = 6.9, 0.8 Hz, 3H).

Fig. 1c: ¹H NMR (400 MHz, CDCl₃) (B) δ 7.97–7.17 (m, 7H), (C) δ 6.05–5.93 (m, 0H), (D) δ 3.75 (q, J = 7.0 Hz, 2H), (E) δ 1.18–1.12 (m, 4H), (F) δ 2.09 (ddd, J = 12.3, 8.7, 5.9 Hz, 1H), (G) δ 0.93–0.76 (m, 1H).

FTIR spectrum was used to investigate the reaction between IPTS and F51. The obvious strong absorption peak at 2271 cm⁻¹ in Fig. 3b, which corresponds to the C=N=O stretching band²⁸ disappears in Fig. 3c and a newly generated absorption peak at 1722 cm⁻¹ is found, which corresponds to the O–C=O linkage.²⁹ Moreover, an obvious peak at 3390 cm⁻¹, which does not exist in Fig. 3a and b is found, i.e., the N–H stretching vibration peak of the secondary amine, which is another proof of the expected reaction. It is important to note that the intrinsic absorption for the epoxy group is observed at 916 cm⁻¹ in both Fig. 3a and c, which indicates that the epoxy group was not consumed in this step³⁰ and therefore confirm that IPTS was grafted successfully onto the molecule of F51.

Fig. 3 FTIR spectra of F51 (a), IPTS (b) and F51–IPTS (c).

Fig. 4 is the FTIR of the final product F51–IPTS + 10% DOPO–TVS (EP10) after sol–gel. The partial enlarged detail represents the obvious absorption peak of the Si–O–Si asymmetric stretching within the range of 1050–1130 cm⁻¹ according to relevant research, which confirms the existence of the silica network by the sol–gel reaction.²¹

Fig. 4 FTIR spectrum of the final product F51–IPTS + 10% DOPO–TVS (EP10) after sol–gel.

3.2 Phenomenon of liquid oxygen compatibility test

The unmodified F51/DDM curing product is not compatible with liquid oxygen according to relevant research. It may flash, explode, burn during the 20 times mechanical impact when immersed in liquid oxygen.^31,32 The impact sensitivity is defined by eqn (1), where ‘S’ indicates the impact sensitivity, ‘i’ is the times for each type of phenomenon, ‘C_i’ is the coefficient of sensitivity, and ‘N’ is the total test times. The prescript of ‘C_i’ is as follows: one explode and combustion prorates 1, one single explode prorates 0.9, one light prorates 0.7, one spark prorates 0.6, one char prorates 0.4.⁸ According to the research of Wu et al., four times of flash, and one time of charring were observed during the 20 times mechanical impact for the unmodified F51/DDM system, which means the impact sensitivity is as high as 0.16. In addition, one time flash was also observed for the F51/DDM system modified with DOPO, which indicates that the impact sensitivity for phosphorus-modified F51/DDM system is 0.035.³² It is worth mentioning that all the specimens containing P/Si did not show any visible phenomena during the 20 times impact, which indicates that the impact sensitivity is zero for the P/Si synergistically modified F51/DDM system. The result confirms that the liquid oxygen compatibility is obviously enhanced through the synergetic effect of silicon and phosphorus.

3.3 Surface element analysis

The surface elements of the specimens before and after mechanical impact in liquid oxygen were investigated by XPS. The overall XPS spectra of the specimen EP10 is shown in Fig. 5. Two obvious peaks around 100 eV and 150 eV appear after the mechanical impact, as shown in Fig. 5, which correspond to the Si_2p and Si_2s signal, respectively. During the mechanical impact, a large amount of instantaneous energy was released, which led to the rapid increase in the local temperature and finally initiate the oxidation reaction. It was reported that during thermal degradation of the polymer materials, silicon interacts with the polymer matrix and migrates to the surface of the matrix due to its low surface energy.³³ The obvious change in the content of Si after mechanical impact perfectly confirms the migration of Si and thus explains the role of Si element during the oxidation reaction, i.e., it formed a silicon-rich surface carbon layer and protected the inner material from further oxidation.

Fig. 5 Overall XPS spectra of EP10 before impact and after impact.

To analyze the changes in the phosphorus element on the surface of specimens after the mechanical impact, the P_2p spectra were further investigated. Fig. 6 shows the P_2p peak fitting curves of specimen EP10 before (a) and after (b) the mechanical impact. There are two peaks in the P_2p spectra before mechanical impact: the peak around 133.5 eV was attributed to C–P and the peak around 134.0 eV was assigned to O=P–O in the DOPO group. After mechanical impact, the peaks of C–P and O=P–O are still detected.³⁴ Moreover, a new peak at 135.1 eV was observed. This was attributed to the –PO₃ group in the pyrophosphate and polyphosphate generated during the mechanical impact,³⁵ which confirms that the phosphorus-containing groups thermally decomposed under the rapid increase in temperature caused by the mechanical impact and formed phosphoric oxyacid. This covered the surface of the specimen and prevented the reaction between the specimen and liquid oxygen.

Fig. 6 P_2p peak fitting curves of the specimen EP10 before and after mechanical impact.

The surface element content of all groups before and after the mechanical impact are listed in Table 2. The element content changed significantly after the mechanical impact in liquid oxygen. In addition to the obvious increased Si concentration, a common phenomenon that the O/C ratio of all the specimens increases at different degrees was detected, which reveals the oxidation reaction during the mechanical impact.

Table 2 Surface element content of the specimens before and after the mechanical impact

	Specimen ID	C (%)	O (%)	Si (%)	P (%)	N (%)	O/C
Before impact	EP5	79.15	16.34	1.04	0.34	3.13	0.21
	EP7.5	78.95	16.3	1.5	0.38	2.87	0.21
	EP10	80.18	15.23	1.52	0.41	2.66	0.19
	EP15	79.34	15.2	1.84	0.68	2.94	0.19
After impact	EP5	70.03	17.97	9.41	0.33	2.26	0.26
	EP7.5	66.74	19.31	10.01	0.39	3.55	0.29
	EP10	68.13	18.24	11.48	0.45	1.71	0.27
	EP15	65.88	19.58	11.51	0.70	2.33	0.25

---

3.4 Mechanical behavior under cryogenic condition

The diagrammatic sketch of the existence form of Si–O–Si network is demonstrated in Scheme 2. The crosslink density of the cured epoxy network will decrease due to the generated SiO_x bulk within it. Thus, the free volume increases, which will decrease the brittleness of the matrix under cryogenic conditions. Moreover, the effective load transfer between the three-dimensional Si–O–Si structure and epoxy matrix will contribute to the toughness of the hybrid material according to relevant research.¹⁷

Scheme 2 Schematic of the Si–O–Si network in the curing system before (a) and after modification (b).

Fig. 7 shows the stress/strain curves of the pure EP and modified epoxy resins with different ratio of DOPO–TVS. The stress/strain curves of all specimens exhibit an approximate proportional relationship, which is the typical feature of brittle material. It is noteworthy that the slope of the curves, which corresponds to the Young's modulus ‘E’, shows obvious regularity. The Young's modulus shows a decreasing trend along with the increase in DOPO–TVS content. The tensile strength of the modified epoxy resin declines slightly, while the elongation at break exhibits an opposite tend. The optimal mechanical behavior appears when the DOPO–TVS content ranges from 7.5% to 10%. The elongation of EP7.5 increases by 15% compared to that of the unmodified epoxy resin. The tensile strength of EP10 merely decreases by 6% while obtaining an 11.5% increase in elongation. The long bond length of Si–O and the large bond angle of Si–O–Si endows it with low rotation barrier energy and high flexibility,³⁶ which contributes to the improvement of the toughness. Small molecules of DOPO–TVS are more likely to self-polymerize rather than polymerize between EP–IPTS and DOPO–TVS during the sol–gel process. When the ratio of DOPO–TVS reaches a certain degree, the negative effect of the agglomeration of the generated silica particle becomes dominant, thus influencing the mechanical behavior of EP15. Fig. 8 summarizes the fracture toughness and elongation results. These two indicators show the same trend, which reaches the optimum when the trisiloxane content is at the range of 7.5–10%. The results of the mechanical test in a liquid nitrogen environment are listed in Table 3. The experimental data of mechanical tests were chosen from the average value of 5 repeated tests.

Fig. 7 Stress–strain curves of stretching test in cryogenic temperature.

Fig. 8 Mechanical behaviors at cryogenic temperatures.

Table 3 Mechanical properties of the samples

Sample	Stress (MPa)	Strain (%)	Young's modulus ‘E’ (GPa)	Fracture toughness (MPa m^1/2)
Pure EP	177.1 ± 4.2	2.25 ± 0.15	7.8 ± 0.4	1.32 ± 0.12
EP5	161.6 ± 3.1	2.22 ± 0.16	7.3 ± 0.3	1.47 ± 0.11
EP7.5	157.1 ± 6.2	2.60 ± 0.20	6.0 ± 0.3	1.61 ± 0.15
EP10	165.8 ± 4.8	2.51 ± 0.22	6.6 ± 0.4	1.63 ± 0.08
EP15	129.8 ± 5.2	2.25 ± 0.19	5.8 ± 0.3	1.21 ± 0.12

---

3.5 Morphology of the fracture surface

To further investigate the mechanism of the modification of the mechanical behavior, SEM was used to observe the fracture surface of pure EP and EP10 specimens, as shown in Fig. 9. The fractograph for the pure epoxy resin (Fig. 9a) reveals a smooth, glassy, and homogenous microstructure without any plastic deformation, which is the feature of typical brittle fracture. Compared to the case of the pure epoxy resin, the fracture surface of the modified resin presented considerably different fractographic features. Fig. 9b shows the fracture surface of EP10 under low magnification (×200), which shows a much rougher fracture surface. The roughness of the fracture surface has two explanations according to relevant research.^37,38 First, it is an indication of crack path deflection, i.e., the crack deviated from its original plane increasing the area of the crack. Therefore, the energy required for the propagation of the crack increased. Second, the roughness indicates the ductile nature of the crack. Interleaved ridges are observed in the fracture surface of the modified epoxy resin, which is shown in Fig. 9c, and Fig. 9d shows the magnification of the ridge area at 1600 times. The multi-planar crack propagation and interleaved ridges increase the fracture surface drastically thus absorbing more fracture energy and then improving the mechanical strength of the epoxy thermoset. It is notable that there are no visible signs of particle formation even under high magnification. Fig. 9e shows the morphology of the ridge area under a magnification of 25 thousand times, which shows a homogeneous single-phase feature and reveals a low self-polymerization ratio of DOPO–TVS.

Fig. 9 Typical SEM fractograph of neat epoxy resin (a), EP10 in 200 times (b), EP10 in 400 times (c), EP10 in 1600 times (d), EP10 in 25 000 times (e).

4. Conclusions

A DOPO based trisiloxane containing phosphorus and silicon was prepared with DOPO and TVS. The hybrid epoxy resin containing phosphorus and silicon was synthesized. Phosphorus and silicon were introduced to the matrix in the form of silica network via a conventional sol–gel reaction. The products were characterized by FTIR, ¹H NMR and ³¹P NMR. The liquid oxygen compatibility of the synthesized epoxy resin was confirmed by the mechanical impact. There was no visible phenomenon during the 20 times impact, which suggests that the modified epoxy resin system is compatible with liquid oxygen. Moreover, the overall XPS spectra revealed the migration of Si to the surface of the specimen during the mechanical impact in liquid oxygen. This phenomenon confirms the enhancement mechanism of Si on the liquid oxygen compatibility of the matrix. A new peak at 135.1 eV was observed in the P_2p spectra after the mechanical impact, which was assigned to the –PO₃ group in the pyrophosphate and polyphosphate and reveals the role of phosphorus in the oxidation enhancement. The tensile test and fracture toughness tests were carried out to explore the influence of the modification on the mechanical properties of epoxy resin under cryogenic conditions. The elongation of EP5 increases to 2.6%, which is 15.5% higher than that of the pure resin and the fracture toughness of EP10 increased to 1.61, which is 22% higher than that of the unmodified resin. The elongation curves of the modified groups show a slighter slope, which indicates a decreased elastic modulus and reduced brittleness. The results showed that the SiO_x network has an obvious effect on enhancing the toughness of the epoxy resin at low temperatures. Overall, the introduction of a three dimensional SiO_x network containing phosphorus endows the epoxy resin with good liquid oxygen compatibility, as well as enhanced toughness under cryogenic conditions. The method mentioned in this study is a promising way to obtain a liquid oxygen compatible epoxy matrix for the cryogenic tank on launch vehicles.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (91016024).

References

1. J. Guo, X. Zhang, H. Gu, Y. Wang, X. Yan, D. Ding, J. Long, S. Tadakamalla, Q. Wang, M. A. Khan, J. Liu, X. Zhang, B. L. Weeks, L. Sun, D. P. Young, S. Wei and Z. Guo, RSC Adv., 2014, 4, 36560–36572.
2. L. Guadagno, M. Raimondo, V. Vittoria, L. Vertuccio, C. Naddeo, S. Russo, B. De Vivo, P. Lamberti, G. Spinelli and V. Tucci, RSC Adv., 2014, 4, 15474–15488.
3. R. K. Mark Tudela, in 46th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics & Materials Conference, 2005.
4. A. Toldy, B. Szolnoki and G. Marosi, Polym. Degrad. Stab., 2011, 96, 371–376.
5. D. Duraibabu, M. Alagar and S. A. Kumar, RSC Adv., 2014, 4, 40132.
6. H. Jin, C. L. Mangun, A. S. Griffin, J. S. Moore, N. R. Sottos and S. R. White, Adv. Mater., 2014, 26, 282–287.
7. J. K. W. Sandler, J. E. Kirk, I. A. Kinloch, M. S. P. Shaffer and A. H. Windle, Polymer, 2003, 44, 5893–5899.
8. G. Wang, X. Li, R. Yan and S. Xing, Mater. Sci. Eng., B, 2006, 132, 70–73.
9. J. Musil, Surf. Coat. Technol., 2012, 207, 50–65.
10. C. Binder, K. Arlt, T. Brock, P. Hartwig, O. Hesse and T. Tillack, J. ASTM Int., 2009, 6, I102308–I102309.
11. J. Hu, J. Shan, D. Wen, X. Liu, J. Zhao and Z. Tong, Polym. Degrad. Stab., 2014, 109, 218–225.
12. W. Zhang, X. Li and R. Yang, Polym. Degrad. Stab., 2014, 99, 118–126.
13. L. Qian, Y. Qiu, J. Liu, F. Xin and Y. Chen, J. Appl. Polym. Sci., 2014, 131, 39709.
14. Y. Meng, Z. Wei, L. Liu, L. Liu, L. Zhang, T. Nishi and K. Ito, Polymer, 2013, 54, 3055–3064.
15. L. Deng, M. Shen, J. Yu, K. Wu and C. Ha, Ind. Eng. Chem. Rem., 2012, 51, 8178–8184.
16. S. Park, F. Jin, J. Park and K. Kim, Mater. Sci. Eng., A, 2005, 399, 377–381.
17. H. Gu, J. Guo, H. Wei, X. Yan, D. Ding, X. Zhang, Q. He, S. Tadakamalla, X. Wang, T. C. Ho, S. Wei and Z. Guo, J. Mater. Chem. C, 2015, 3, 8152–8165.
18. H. Dodiuk, S. Kenig, I. Blinsky, A. Dotan and A. Buchman, Int. J. Adhes. Adhes., 2005, 25, 211–218.
19. S.-T. Lin and S. K. Huang, J. Polym. Sci., Part A: Polym. Chem., 1996, 34, 1907–1922.
20. C. Chen, M. Li, Y. Gao, J. Nie and F. Sun, RSC Adv., 2015, 5, 33729–33736.
21. Q. Zhang, H. Liu, X. Zhan, F. Chen, J. Yan and H. Tang, RSC Adv., 2015, 5, 77508–77517.
22. J. Chen, A. J. Kinloch, S. Sprenger and A. C. Taylor, Polymer, 2013, 54, 4276–4289.
23. E. Burunkay, Ö. Kesmez, N. Kiraz, H. E. Çamurlu, M. Asiltürk and E. Arpaç, Thin Solid Films, 2012, 522, 238–246.
24. K. M. Seeni Meera, R. Murali Sankar, S. N. Jaisankar and A. B. Mandal, J. Phys. Chem. B, 2013, 117, 2682–2694.
25. C. Fu, X. Hu, Z. Yang, L. Shen and Z. Zheng, Prog. Org. Coat., 2015, 84, 18–27.
26. W. Lee, L. Liu, C. Chen and J. Lin, Polym. Adv. Technol., 2014, 25, 36–40.
27. C. Tasdelen-Yucedag and A. T. Erciyes, Prog. Org. Coat., 2014, 77, 1750–1760.
28. R. Gallego, J. F. Arteaga, C. Valencia and J. M. Franco, Carbohydr. Polym., 2013, 98, 152–160.
29. G. Sun, L. Liu, J. Wang, H. Wang, W. Wang and S. Han, Polym. Degrad. Stab., 2015, 115, 1–15.
30. J. Liu, J. Tang, X. Wang and D. Wu, RSC Adv., 2012, 2, 5789.
31. Z. Wu, S. Li, M. Liu, Z. Wang and X. Liu, RSC Adv., 2015, 5, 11325–11333.
32. Z. Wu, J. Li and Z. Wang, Polym. Adv. Technol., 2015, 26, 153–159.
33. A. Fina, D. Tabuani, F. Carniato, A. Frache, E. Boccaleri and G. Camino, Thermochim. Acta, 2006, 440, 36–42.
34. X. Wang, Y. Hu, L. Song, W. Xing, H. Lu, P. Lv and G. Jie, Polymer, 2010, 51, 2435–2445.
35. J. Wang, D. Wang, Y. Liu, X. Ge and Y. Wang, J. Appl. Polym. Sci., 2008, 108, 2644–2653.
36. Y. Kawakami and I. Imae, Macromol. Chem. Phys., 1999, 200, 1245–1256.
37. B. Francis, S. Thomas, J. Jose, R. Ramaswamy and V. Lakshmana Rao, Polymer, 2005, 46, 12372–12385.
38. Q. He, T. Yuan, X. Yan, D. Ding, Q. Wang, Z. Luo, T. D. Shen, S. Wei, D. Cao and Z. Guo, Macromol. Chem. Phys., 2014, 215, 327–340.

---