Thermal and mechanical behavior

Materials and Design 111 (2016) 453–462

Contents lists available at ScienceDirect

Materials and Design journal homepage: www.elsevier.com/locate/matdes

Synthesis of SiO2/epoxy–benzoxazine ternary copolymer via sol–gel method: Thermal and mechanical behavior Cong Peng a, Jialiang Li b, Zhiwei Li b, Zhanjun Wu b,*, Dayu Zhou a a

School of Materials Science and Engineering, Dalian University of Technology, Dalian, 116024, PR China School of Aeronautics and Astronautics, Faculty of Vehicle Engineering and Mechanics, State Key Laboratory of Structural Analysis for Industrial Equipment, Dalian University of Technology, Dalian 116024, PR China

b

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Silica unit is introduced to the epoxy/ benzoxazine copolymer which has covalent bond with both epoxy and benzoxazine. • Morphology of the decomposition residues are analyzed with both naked eye and SEM. • There is no particle agglomeration observed in the matrix. • The ternary copolymer shows improved toughness and thermal stability compared with common epoxy/benzoxazine copolymer.

a r t i c l e

i n f o

Article history: Received 1 August 2016 Received in revised form 30 August 2016 Accepted 31 August 2016 Available online 07 September 2016 Keywords: Epoxy resin Benzoxazine Silica Thermal property Mechanical behavior

a b s t r a c t Trialkoxy–terminated benzoxazine monomer was synthesized using bisphenol A (BPA), 3– aminopropyltriethoxysilane (KH–550) and paraformaldehyde. Subsequently, bisphenol F epoxy resin (F51) was pretreated with 3–isocyanatopropyltriethoxysilane (IPTS) to covalently introduce trialkoxy group into the epoxy molecular. Sol–gel process was then initiated with tetraethoxysilane (TEOS) as precursor to introduce silica structure into epoxy–benzoxazine hybrid before curing reaction. The synthesized benzoxazine and epoxy resin containing trialkoxysilane group were used as silane coupling agent to connect epoxy–benzoxazine matrix as organic domain and silica units as inorganic domain. Thermal gravimetric analysis (TGA) and dynamic mechanical analysis (DMA) show that the organic–inorganic ternary copolymer possesses promoted thermal stability compared with the unmodified epoxy–benzoxazine matrix. The char residues of the ternary copolymer after decomposition test reveal a dense surface layer and unbroken original dimension which is in accordance with the TGA results. According to the results of the mechanical tests and DMA, the SiO2/epoxy–benzoxazine copolymer possesses improved toughness and is more capable of absorbing deformation energy under external force. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction

⁎ Corresponding author. E-mail address: [email protected] (Z. Wu).

http://dx.doi.org/10.1016/j.matdes.2016.08.095 0264-1275/© 2016 Elsevier Ltd. All rights reserved.

Thermosetting epoxy polymers have been widely used as engineering adhesives and matrices for composite materials due to its high modulus and strength, excellent chemical resistance and simplicity in

454

C. Peng et al. / Materials and Design 111 (2016) 453–462

processing [1–3]. However, better thermal stability and mechanical behavior are still in urgent demand for special purpose. Benzoxazine resin, as a new class of thermosetting phenolic resins with the design flexibility for various applications has been developed to overcome the shortcomings of traditional phenolic resin while keeping the advantages of cost effectiveness, heat resistance and flame retardance [4–6]. Moreover, benzoxazine has additional unique characteristics such as low water absorption, high dimensional stability and near–zero volumetric shrinkage or expansion upon curing [7–9]. The generated phenol –OH groups from the curing reaction can react with epoxy resins at elevated temperatures. In other words, they can be used as hardener of epoxy resin [10]. The higher temperature resistance and glass transition temperature, even or stronger mechanical behavior of benzoxazine compared to that of epoxy resin has proven it a reasonable way to introduce benzoxazine into the epoxy matrix to obtain copolymer of better stability, safety for the practical applications. Some researchers have investigated the copolymerization of benzoxazine resin with epoxy resin [6,11]. The crosslink density and glass–transition temperature of these copolymers were higher than those of the benzoxazine and epoxy homopolymer while the matrix became more brittle and easy to break due to the intrinsic brittleness of benzoxazine [4,12–13]. The effects of siloxane on the thermal and mechanical properties of polymer have been extensively studied in recent years [14–16]. The high dissociation energy of Si\\O bond endows it with high thermal stability and oxidation resistance. Moreover, Si element interacts with the polymer matrix, migrates to the surface of the matrix and finally forms a silicon–rich surface carbon layer during thermal decomposition due to its low surface energy. Besides, Si\\O bond can improve flexibility of polymer due to its low rotation barrier energy and high flexibility [15]. Numerous researches have been reported on enhancing the mechanical and thermal behavior of the polymer by introducing various stuff containing Si\\O structure such as nano–silica particle, polyhedral oligomeric silsesquioxane (POSS), and siloxane [17–20]. There has, however, been little research that combines these three stuffs together and investigates the relevant properties. Organic– inorganic nanocomposites obtained through sol–gel process have attracted much attention since both components act synergistically with each other to reveal unexpected hybrid properties [21]. In the present work, benzoxazine resin was introduced to the traditional epoxy matix to partially replace the common hardener and formed the epoxy–benzoxazine copolymer. Moreover, inorganic–organic hybrid was finally obtained by covalently introducing Si–O–Si network into the epoxy–benzoxazine copolymer through sol–gel method to further enhance the thermal stability and mechanical behavior.

aminopropyltriethoxysilane (KH–550) were purchased from Sinopharm Chemical Reagent Beijing Co. Ltd. (3–isocyanatopropyl)– triethoxysilane (IPTS) was obtained from Energy Chemical Co. Ltd. Dibutyltindilaurate (DBTDL) was purchased from Adamas Reagent Co. Ltd. Bisphenol A (BPA), paraformaldehyde, phenol and aniline were purchased from DaMao Chemical Reagent Tianjin Co. Ltd. All the reagents were analytically pure and used without further treatment. 2.2. Preparation of siloxane–modified epoxy resin (MEP) MEP was prepared following the reported method [22]. Into a 250 ml three–neck flask equipped with a stirring bar in N2 atmosphere, 120 g of F51, 12 g of IPTS and 0.4 g of DBTDL were added. The temperature remained at 50 °C until the characteristic infrared absorption peak of isocyanate group at 2271 cm−1 disappeared and colorless transparent viscous as–synthesized epoxy resin was obtained (symbolized with MEP). This process is shown in Fig. 1. It is worth mentioning that since the mole ratio of IPTS to F51 was less than 1:1, the synthesized matrix was the mixture of unmodified F51 and the modified MEP. 2.3. Preparation of trialkoxy–terminated benzoxazine (B–aptes) Trialkoxy–terminated benzoxazine (B–aptes) was synthesized as the following procedure: 11.42 g of BPA (0.05 mol) was dissolved in 100 ml dioxane and then 22.14 g of KH-550 (0.1 mol) and 6.67 g of paraformaldehyde (0.2 mol) were added. The mixture was heated to 100 °C and remained for 5 h followed by evaporation of solvent under vacuum to afford pale brown liquid. The representative structure of the reaction product (B–aptes) is shown in Fig. 2. 2.4. Preparation of silica/epoxy–benzoxazine copolymer Different weight percent of TEOS, and B–aptes were mixed with MEP and the mixture was dissolved in acetone. Equivalent water was then added with acetic acid as catalyst to initiate the sol–gel process. The mixture was remained at 50 °C for 5 h followed by evaporation of solvent under vacuum to afford viscous transparent liquid. The compositions of common epoxy–benzoxazine copolymer and epoxy– benzoxazine–silica hybrid are listed in Tables 1 and 2 respectively. The epoxy–benzoxazine–silica copolymer is symbolized with EBSX in which ‘E’, ‘B’, ‘S’ stand for ‘epoxy’, ‘benzoxazine’, ‘silica’ respectively and ‘X’ varies with the proportion of the three components. The symbols of epoxy–benzoxazine specimens are similar. 2.5. Curing process of the copolymers

2. Experimental The solid BA–a was added in the last step and the mixture was heated to 110 °C to obtain homogeneous liquid. DDS was used as the curing agent considering the high operating temperature. The mixture was further heated to 130 °C with continuous stir until DDS completely dissolved. Then the hybrids were poured into Teflon mold, vacuum degassed and gradually polymerized at 160 °C, 180 °C and 200 °C for 2 h respectively and postcured at 210 °C for 1 h to obtain the final

2.1. Materials Bisphenol F epoxy resin (F51, epoxy value 0.53–0.57) was purchased from Feicheng Deyuan Chemicals Co. Ltd. Bisphenol A type benzoxazine monomer (BA–a) was supplied by Rui Yi Chemicals Shanghai Co. Ltd. 4, 4′–diaminodiphenyl sulfone (DDS), tetraethoxysilane (TEOS) and 3–

O H C

H2C

OH

O

O

H C

CH2

+

O

C N

O Si O

DBTDL O

H2C

O

CH

F51

IPTS

O Si

O

NH C O

MEP

O CH CH2 O

Fig. 1. Preparation of siloxane–modified epoxy resin.


455

Fig. 2. Preparation of trialkoxy–terminated benzoxazine.

specimens. The preparation process of epoxy–benzoxazine–silica hybrid is shown in Fig. 3. (Note: the values in the brackets indicate the mole ratio of each composition).

after impact test were obtained using a QUANTA 450 tungsten filament scanning electron microscope (SEM) with a voltage of 20 kV. The EDX spectra were obtained using an OXFORD X–Max equipment installed in the SEM instrument.

2.6. Characterizations 3. Results and discussion Fourier transform infrared (FTIR) spectroscopy was carried out to monitor the synthesis process of MEP with a Perkin Elmer Spectrum One FTIR. The infrared spectrum was recorded in the wave number ranged from 650 to 4000 cm−1. 1H NMR was carried out to characterize B–aptes on a NMR spectrometer of Varian INOVA at frequency of 500 MHz in CDCl3. Thermal gravimetric analysis (TGA) was carried out with a TGA Q500 thermogravimetric analyzer (TA Instruments) at heating rate of 20 °C/min from 35 °C to 600 °C under N2 atmosphere. The dynamic mechanical analysis (DMA) of cured samples was carried out with a DMA 2980 dynamic mechanical thermal analyzer (TA Instruments) in single cantilever mode at a frequency of 10 Hz. The samples were heated from room temperature to 250 °C at a heating rate of 3 °C/min. Notched impact test was performed according to ASTM D5942 on an impact tester with pendulum energy of 0.5 J and span of 60.0 mm. A 45° V–shaped notch was made at the central part of the impact bar with a notch–tip radius of 0.25 mm. For each group of specimens, five specimens were tested to obtain the average value as the final impact strength. The fracture toughness test was carried out with a servo–hydraulic testing machine with a displacement rate of 1.0 mm/min following ASTM D5045–14. The micromorphology images of the residues after decomposition test and the fracture surfaces Table 1 Compositions of epoxy–benzoxazine hybrid (wt.%). System

F51

BA–a

DDS

Pure EP EB1 EB2 EB3

75.9 (1) 71.4 (1) 67.6 (1) 64.1 (1)

0 8.1 (0.1) 15.2 (0.2) 21.7 (0.3)

24.1 (0.25) 20.5 (0.45) 17.2 (0.4) 14.2 (0.35)

The traditional bisphenol F glycerol ether (business code: F51) was pretreated with IPTS to introduce hydrolysable alkyl group to the epoxy molecule. FTIR spectrum was used to monitor the reaction between IPTS and F51. Fig. 4a and b are the FTIR spectra of the mixture of F51 and IPTS before reaction and after complete reaction. The strong absorption peak around 2271 cm−1 in the spectrum of the mixture of F51 and IPTS before reaction, which corresponds to the C = N = O stretching band, vanished after about 5 h reaction. A newly generated absorption peak around 1720 cm− 1 is found, which corresponds to the O–C = O linkage (see Scheme 1). The broad peak around 3500 cm−1 which corresponds to –OH stretching vibration in F51 shifted to 3400 cm−1 due to the generation of –NH group in the MEP. The results above indicate a complete reaction between F51 and IPTS. It is worth mentioning that the mole ratio of F51 to IPTS is about 10 in purpose of a complete reaction between them. The structure of synthesized trialkoxy–terminated benzoxazine monomer has been confirmed by 1H NMR in Fig. 5. 1H NMR shows the characteristic peaks which belong to methylene protons of –O–CH2– N– (a: 4.82 ppm) and Ar–CH2–N– (b: 3.92 ppm) of oxazine structure. The peaks which are assigned to the protons of –Si–O–CH2– appear as quartet at 3.81 ppm (f) and protons of –CH3 as triplet at 1.22 ppm (g). The peaks assigned to the protons of propyl chain appear at 2.73– 2.76 ppm (c), 1.64–1.71 ppm (d) and 0.63–0.68 ppm (e) for–N–CH2– C–, –C–CH2–C and –C–CH2–Si– respectively. The single peak at 1.59 ppm (h) corresponds to protons of –CH3–C–CH3– of the bisphenol A structure [5]. The detailed 1H NMR data is as follows: 1 H NMR (500 MHz, CDCl3) δ4.82 (s, 5H), 3.92 (d, J = 6.8 Hz, 5H), 3.81 (q, J = 7.0 Hz, 16H), 2.75 (dd, J = 12.2, 4.9 Hz, 5H), 1.71–1.64 (m, 5H), 1.59 (s, 8H), 1.22 (t, J = 7.0 Hz, 24H), 0.68–0.63 (m, 5H).

Table 2 Compositions of epoxy–benzoxazine–silica hybrid (wt.%). System EBS1 EBS2 EBS3

MEP 415 (1) 415 (1) 415 (1)

B–aptes 17.3 (0.025) 34.5 (0.05) 51.8 (0.075)

b

BA–a

TEOS

DDS

SiO2 a(%)

H2O

32.7 (0.075) 65.3 (0.15) 97.9 (0.225)

31.2 (0.15) 41.6 (0.25) 62.4 (0.3)

111.6 (0.45) 99.2 (0.4) 81.0 (0.33)

3.6 5.0 5.4

18.9 31.5 37.8

(Note: a. the theoretical SiO2 ratio assuming a complete hydrolysis and condensation. b. the values in the brackets indicate the mole ratio of each composition, the sum amount of B–aptes and BA–a of EBSX equals to the corresponding amount of BA–a in EBX).

456


Fig. 3. Preparation of SiO2/ epoxy–benzoxazine ternary copolymer.

Following the preparation of epoxy–benzoxazine–silica copolymer (Section 2.4), the mixture after sol–gel reaction was investigated again using FTIR as shown in Fig. 4c. The intensity of the absorption at 1078 cm− 1 assigned to Si–O–C stretching decreases while those at 788 cm−1 and 1112 cm−1 due to –Si–O–Si– symmetric and asymmetric vibration increases, indicating the formation of silica particles within the matrix [23,24]. The DMA results of the cured ternary copolymer (EBS1–EBS3), pure BA–a, F51/DDS and the copolymer of epoxy and benzoxazine (EB1–EB3) are shown in Fig. 6. For pure resin systems, i.e. the BA–a and F51/DDS, the cured BA–a shows high storage modulus (E′) as shown in Fig. 6a and a much higher glass transition temperature (Tg) compared to that of F51/DDS as calculated from the mechanical loss factor (tanδ) in Fig. 6b which is due to the intrinsic properties of these two resin. The incorporation of the benzoxazine into epoxy resin has obvious effect on

enhancing E' and Tg of the matrix. The E' of EB1 to EB3 is in order of EB3 N EB2 N EB1 and just between that of pure epoxy and benzoxazine before 130 °C, whereas, the order of Tg is just the opposite, i.e. EB1 possesses the highest Tg and EB3 the lowest. This phenomenon can be attributed to the reaction between the oxazine and the primary amine group which is to say that the EP/diamine ratio was no more stoichiometric and thus crosslink density was affected [6]. The introduction of silica structure into the network of epoxy/benzoxazine has obvious influence on the storage modulus. The results of EBS1 have confirmed the good enhancing effect of the mechanical and thermal properties of the copolymer. With the minimum content of silica units, the storage modulus of EBS1 shows a sharp decline, which is even smaller than that of the pure epoxy system. As expected, the introducing of flexible Si\\O can improve the flexiblility of the matrix as well as the thermal stability. EBS1 not only possesses a relatively low storage modulus but


457

Fig. 4. FTIR spectra of (a.) blend of F51 and IPTS before reaction (b.) after complete reaction and (c.) hybrid after sol–gel process.

also has a promoted glass transition temperature (Tg), which, is even higher than that of pure benzoxazine system. As for EBS2, the E' is very close to that of EB2 before 150 °C as shown in Fig. 6a and the Tg is higher than EB2 due to the increase of crosslink density as shown in Fig. 6b. In the case of EBS3, the E′ is much higher than that of EB2 and the increasing in Tg is not so obvious. In the discussion on the results of EBS1, the introduction of silica units have the trend of reducing E′ and prompting the Tg. However, this trend reverses with increasing amount of silica content. The order of E′ is EBS3 N EBS2 N EBS1 and that of Tg is EBS1 N EBS2 N EBS3. This can be explained that the introduction of silica into the copolymer network is not the sole influence factor on the E′ and Tg. As mentioned above, the EP/diamine ratio was no more stoichiometric due to the reaction between the oxazine and the primary amine group which will reduce the crosslink density. The effect of introducing silica on crosslink density will became secondary when the benzoxazine ratio exceeds a critical value. As a result, the Tg decreases with the increase of benzoxazine ratio due to the decline of the crosslink density. Concerning the high value of E′ of EBS3, excessive precursor of silica will cause a heavy degree of self polymerization which prevents the formed silica monomer from copolymerizing with the epoxy/ benzoxazine matrix. As a result, part of the silica dispersed in the matrix in the form of nanoparticle and the stiffness is developed as many researches mentioned [18,25,26].

Fig. 5. 1H NMR spectra of B–aptes.

Fig. 6. DMA thermograms of different systems.

Impact strength of the unmodified resin blend and the epoxy/benzoxazine/silica copolymers is shown in Fig. 7. For the unmodified specimens, the impact strength shows perfect downward trend with the increase of the benzoxazine ratio which reveals the intrinsic brittleness of benzoxazine resin. The silica containing copolymers (EBS1 and EBS2) show obvious enhancement in the impact strength compared with EB1 and EB2 respectively. With the same benzoxazine mole ratio, the impact strength of EBS1 is improved by 26% compared to that of EB1 and even higher than that of pure epoxy resin. That of EBS2 is improved by 10.8% compared to EB2. This result is in accordance with discussion of the DMA test. Moreover, the molecular weight increases during the sol– gel process as illustrated in Scheme 3 which has positive effect on the mechanical behavior of polymer according to relevant research [27]. However, the impact strength of EBS3 declines obviously compared with the former two groups, which is even lower than that of EB3. According to the formula shown in Table 2, with the increase of benzoxazine mole ratio, the content of silica precursor increases as well. The TEOS of small molecular is more likely to self–polymerize rather than grafting onto the epoxy–benzoxazine network. It is reasonable for the deterioration of mechanical behavior when the silica quantity reaches a critical point. The results of the mechanical tests are listed in Table 3.

Fig. 7. Impact strength and fracture toughness of each group of specimens.

458


Table 3 Test results of each resin system. E′ (GPa)

Impact

Fracture

System

Tg (°C)

50 °C

Tg + 30 °C

T2% (°C)

Char yield at 600 °C (wt.%)

Strength (kJ/m2)

toughness (MPa·m1/2)

BA–a F51/DDS EB1 EB2 EB3 EBS1 EBS2 EBS3

178.2 143.9 157.9 156.3 151.7 185.6 171.6 157.2

1.72 1.83 1.57 1.73 1.76 1.43 1.83 2.07

1.87E–2 5.84E–2 1.73E–2 2.26E–2 4.22E–2 2.43E––2 3.48E–2 6.79E–2

281.3 372.7 361.1 298.1 292.5 295.8 356.3 335.7

31.7 21.7 23.4 28.2 28.4 29.5 31.9 34.7

0.78 ± 0.06 0.74 ± 0.05 0.91 ± 0.05 0.72 ± 0.02 0.82 ± 0.04 0.68 ± 0.03 0.67 ± 0.09 0.64 ± 0.02

1.06 ± 0.09 1.07 ± 0.06 1.31 ± 0.09 1.09 ± 0.04 1.19 ± 0.07 1.12 ± 0.02 1.15 ± 0.11 1.22 ± 0.08

The fracture toughness of benzoxazine shows a different trend which is higher than that of epoxy resin. As a result, with increasing of the benzoxazine content, the order of fracture toughness of EB1 to EB3 is EB3 N EB2 N EB1. The different results of impact strength and fracture

toughness tests could be attributed to the different loading mode of the two tests. Despite the opposite trend of EB1–EB3, the fracture toughness of EBS1–EBS3 shows promoted values compared to that of EB1–EB3 respectively. Moreover, fracture toughness of EBS1 shows

Fig. 8. SEM images of fracture surface of pure BA–a (a), F51 (b), EB2 (c) and EBS2 (d) after impact test (e) high resolution image of the relevant area in (d).


Fig. 9. TGA thermogram of different systems.

the highest value as it does in the impact strength. This result has further confirmed the strengthening effect of the covalent introduction of silica unit into the structure of epoxy/benzoxazine copolymer.

459

Representative SEM micrographs of crack formation zones of pure BA–a, F51, EB2 and EBS2 after impact test are shown in Fig. 8. F51/ DDS and BA–a exhibit quite similar fracture pattern which are smooth with cracks almost parallel to the crack–propagation direction. This indicates a typical brittle fracture behavior, which accounts for the relatively low impact strength of these two pure resin systems. It should be noted that well developed white river–like zone can be observed for pure F51. The white color is a result of stress–whitening effect associated with a large number of micro–cracks formed inside the damage zones [28]. Such white river–like zone is less perceptible in BA–a fracture surface which should be attributed to the better toughness of epoxy resin compared to benzoxazine resin. It is well accepted that the larger and the rougher the plastic zone, the higher the fracture toughness and energy values are. The morphology of the fracture surface of EBS2 is obviously different. The surface appeared coarser and ditches with characteristic parabolic feature from crazes intersecting the main feature plane, which indicates that crack deflection process occurred. This generated an increase in the total fracture surface area resulting in greater energy absorption as compared to the pure resin system. Another important aspect should be noted that no separate phase could be detected in Fig. 8d which means that either epoxy, benzoxazine and TEOS are copolymerized, or the homopolymerized phase is nanoscale dispersed in the EP matrix and therefore cannot be

Fig. 10. Digital images of the thermal decomposition test for specimens (a) before decomposition (b) the outer appearance after decomposition (c) the inner appearance after decomposition.

460


detected at low magnifications of SEM. Fig. 8e, the higher resolution SEM from Fig. 8d, which shows homogeneous morphology without granule and agglomeration confirms the first assumption. The TGA curves of each group of specimens are shown in Fig. 9. The temperature values linked to 2 wt.% loss and the char yield at 600 °C are listed in Table 3. As expected, the BA–a system exhibits good char formation behavior and the pure epoxy system F51/DDS shows the lowest char yield which coincide with relevant research [6,29]. There is one interesting phenomenon that the onset degradation temperature (2 wt.% loss) of pure epoxy system F51/DDS is the highest above all the specimens while that of the pure benzoxazine system BA–a is almost the lowest which indicates that the epoxy resin has better thermal stability than benzoxazine in low temperature [30]. Thermal degradation of polybenzoxazine below 300 °C is due to pyrolytic evolution of aniline

derivative of Mannich bridge [31]. By the respective comparison between TGA curves of EB2 and EBS2, EB3 and EBS3, it can be concluded that the onset degradation shifts to higher temperature due to the introducing of silica units. Moreover, from the rank of char yield at 600 °C shown in Fig. 9, the ternary copolymers show excellent char forming behavior some of which are even better than pure benzoxazine even with a low content of benzoxazine monomer. A polymeric material's resistance to thermal decomposition is important to measure its thermal stability. Fig. 10 shows the digital images of the specimens before and after thermal decomposition. The cured specimens were heated to 600 °C under air atmosphere at a rate of 5 °C/min in a tube furnace to acquire the residues. The heating process stopped once the temperature reached 600 °C to avoid heavy decomposition and to get char residues with relatively complete

Fig. 11. SEM images of char residues of EBS2 (a–c) and of EB2 (d–e) and EDX spectra (f) of inner and external elements of EBS2 residue.


structure. The char residue of pure EP reveals obviously hollow and intumescent appearance and only a thin and glossy surface layer left which is very loose and fragile. It implies that the matrix melted and volatilized during heating and condensed after cooling down. It can be noted that the char layer of BA–a is more compact than that of pure EP and the residue retains the original contour which indicates better thermal stability during thermal decomposition. In other words, benzoxazine has slower decomposition rate than epoxy resin due to the higher thermal stability of aromatic benzene ring in the molecular structure. The degree of volume expansion of EB2 is smaller than that of pure EP due to the incorporation of BA–a. As a strong contrast, the residual of EBS2 shows dense internal structure and close volume compared to the original status before heating. Obviously, the residue of EBS2 shows elevated thermal stability compared with that of EB2 even though the benzoxazine content is identical. The contained silicon element and engrafted silica has played an important role in enhancing the thermal stability. During thermal degradation of the polymer materials, silicon element interacts with the polymer matrix, migrates to the surface of the matrix due to its low surface energy, and finally forms a silicon– rich surface layer which protects the matrix from the external heat and oxygen. The results of the decomposition test are in good agreement with the TGA curves. Combining the digital images, SEM was carried out to investigate the detailed information of the char residues. As can be seen from Fig. 11a, the external surface of EBS3 residue is continuous and compact which explains the good shape–holding behavior shown in Fig. 10. The inner structure of EBS2 residue is shown in Fig. 11c which shows a continuous and multihole feature. The multihole structure can be attributed to the decomposition and vaporization of matrix where weight loss is quicker. As an obvious contrast, the morphology of EB2 residue exhibits different character. The external surface as shown in Fig. 11d is glossy with obvious ripples and folds which reveal that the matrix melted during heating and condensed after cooling down. There is almost a thin layer left for EB2 after heating. Fig. 11e is the morphology of inner structure of EB2 residue which shows the similar glossy feature as for the external surface. It should be noticed that, compared with Fig. 11b, the gap on the surface layer of EB2 is very shallow which does not run through the whole depth of the skin. Combining the inflated feature of the char residue for all samples except EBS2, it can be assumed that the majority of the inner matrix had already decomposed during heating process and migrated to the surface. Since the heating process stopped once the temperature got to 600 °C, the decomposed product condensed and gathered on the surface and formed the final morphology. EDX spectra of the external surface and the inner structure can further illustrate the good thermal stability of EBS2 (other elements except ‘C'’ ‘O'’ ‘Si'’ were ignored and the content is normalized). As seen in Fig. 11f, ‘Si'’ content of the external surface is obviously higher than that of the inner structure and the situation of ‘O'’ content is just the opposite. This result is in accordance with relevant researches that silicone moved to the surface, accumulated on the surface of the polymer to form a barrier to mass transport and to protect the underlying polymer from the flame [32]. 4. Conclusions Inorganic–organic copolymer was obtained by covalent introduction of silica unit into the epoxy–benzoxazine copolymer through sol–gel method to further enhance the thermal stability and mechanical behavior. The results of impact strength and fracture toughness test exhibit a common conclusion that the mechanical behavior is improved by introducing silica unit into the epoxy–benzoxazine network. Among all silica–containing specimens, EBS1, whose theoretical content of silica is 3.6 wt.% shows the best mechanical properties as well as the best thermal stability as indicated in DMA results. The impact strength and fracture toughness of EBS2, with silica content of 5.0 wt.%, decline compared with that of EBS1 which implies that the optimal silica content is

461

supposed to be in a range below 5.0 wt.%. Results of TGA and decomposition test show that the silicon–containing matrix possesses better thermal resistance due to the formation of the silicon–rich protecting layer. Acknowledgments This work was supported by the National Natural Science Foundation of China (11172053, 91016024, 51102031 and 11302036), the New Century Excellent Talents in University (NCET110055) and the Fundamental Research Funds for the Central Universities (DUT12LK33, DUT14LK26). References [1] P.S. Shivakumar Gouda, V. Chatterjee, P.K. Barhai, D. Jawali, S. Rahatekar, M.R. Wisnom, Improved fracture toughness in carbon fibre epoxy composite through novel pre–preg coating method using epoxy terminated butadiene nitrile rubber, Mater. Des. (1980–2015) 62 (2014) 320–326. [2] D. Carolan, A. Ivankovic, A.J. Kinloch, S. Sprenger, A.C. Taylor, Toughening of epoxy– based hybrid nanocomposites, Polymer 97 (2016) 179–190. [3] P. Luangtriratana, B.K. Kandola, P. Myler, Ceramic particulate thermal barrier surface coatings for glass fibre–reinforced epoxy composites, Mater. Des. 68 (2015) 232–244. [4] M. Xu, X. Yang, R. Zhao, X. Liu, Copolymerizing behavior and processability of benzoxazine/epoxy systems and their applications for glass fiber composite laminates, J. Appl. Polym. Sci. 128 (2013) 1176–1184. [5] T. Agag, T. Takeichi, Synthesis and characterization of benzoxazine resin–SiO2 hybrids by sol–gel process: the role of benzoxazine–functional silane coupling agent, Polymer 52 (2011) 2757–2763. [6] S. Grishchuk, Structure, thermal and fracture mechanical properties of benzoxazine– modified amine–cured DGEBA epoxy resins, Express Polym Lett 5 (2011) 273–282. [7] M. Arslan, B. Kiskan, Y. Yagci, Post–modification of polybutadienes by photoinduced hydrogen abstraction from benzoxazines and their thermally activated curing, Macromolecues 49 (2016) 5026–5032. [8] Y. Wang, G. Wu, K. Kou, C. Pan, A. Feng, Mechanical, thermal conductive and dielectrical properties of organic montmorillonite reinforced benzoxazine/cyanate ester copolymer for electronic packaging, J. Mater. Sci. Mater. Electron. 27 (2016) 8279–8287. [9] B. Del Saz–Orozco, D. Ray, A. Kervennic, M.G. PT, S. WF, Toughening of carbon fibre/ polybenzoxazine composites by incorporating polyethersulfone into the interlaminar region, Mater. Des. 93 (2016) 297–303. [10] H. Kimura, A. Matsumoto, K. Hasegawa, K. Ohtsuka, A. Fukuda, Epoxy resin cured by bisphenol A based benzoxazine, J. Appl. Polym. Sci. 68 (1998) 1903–1910. [11] S. Rimdusit, P. Kunopast, I. Dueramae, Thermomechanical properties of arylamine– based benzoxazine resins alloyed with epoxy resin, Polym. Eng. Sci. 51 (2011) 1797–1807. [12] Z. Wang, L. Li, Y. Fu, Y. Miao, Y. Gu, Reaction–induced phase separation in benzoxazine/bismaleimide/imidazole blend: effects of different chemical structures on phase morphology, Mater. Des. 107 (2016) 230–237. [13] S. Grishchuk, S. Schmitt, O.C. Vorster, J. Karger–Kocsis, Structure and properties of amine–hardened epoxy/benzoxazine hybrids: effect of epoxy resin functionality, J. Appl. Polym. Sci. 124 (2012) 2824–2837. [14] S. Najafi–Shoa, H. Roghani–Mamaqani, M. Salami–Kalajahi, R. Azimi, M. Gholipour– Mahmoudalilou, Incorporation of epoxy resin and carbon nanotube into silica/siloxane network for improving thermal properties, J. Mater. Sci. 51 (2016) 9057–9073. [15] X. Lan, W. Huang, Y. Yu, Synthesis, characterization and properties of the polysiloxane–based episulfide resin, Eur. Polym. J. 46 (2010) 1545–1556. [16] E. Franchini, J. Galy, J. Gérard, D. Tabuani, A. Medici, Influence of POSS structure on the fire retardant properties of epoxy hybrid networks, Polym. Degrad. Stab. 94 (2009) 1728–1736. [17] L. Wang, J. Zhang, X. Yang, C. Zhang, W. Gong, J. Yu, Flexural properties of epoxy syntactic foams reinforced by fiberglass mesh and/or short glass fiber, Mater. Des. 55 (2014) 929–936. [18] C. Chen, R.S. Justice, D.W. Schaefer, J.W. Baur, Highly dispersed nanosilica–epoxy resins with enhanced mechanical properties, Polymer 49 (2008) 3805–3815. [19] Z. Yu, S. You, H. Baier, Effect of organosilane coupling agents on microstructure and properties of nanosilica/epoxy composites, Polym. Compos. 33 (2012) 1516–1524. [20] E.S.A. Rashid, K. Ariffin, C.C. Kooi, H.M. Akil, Preparation and properties of POSS/ epoxy composites for electronic packaging applications, Mater. Des. 30 (2009) 1–8. [21] M. May, H.M. Wang, R. Akid, Effects of the addition of inorganic nanoparticles on the adhesive strength of a hybrid sol–gel epoxy system, Int. J. Adhes. Adhes. 30 (2010) 505–512. [22] X. Qian, L. Song, Y. Hu, R.K.K. Yuen, Thermal degradation and flammability of novel organic/inorganic epoxy hybrids containing organophosphorus–modified oligosiloxane, Thermochim. Acta 552 (2013) 87–97. [23] S. Hou, Y. Chung, C. Chan, P. Kuo, Function and performance of silicone copolymer. Part IV. Curing behavior and characterization of epoxy–siloxane copolymers blended with diglycidyl ether of bisphenol–A, Polymer 41 (2000) 3263–3272. [24] T. Nazir, A. Afzal, H.M. Siddiqi, Z. Ahmad, M. Dumon, Thermally and mechanically superior hybrid epoxy–silica polymer films via sol–gel method, Prog. Org. Coat. 69 (2010) 100–106.

462


[25] A. Allahverdi, M. Ehsani, H. Janpour, S. Ahmadi, The effect of nanosilica on mechanical, thermal and morphological properties of epoxy coating, Prog. Org. Coat. 75 (2012) 543–548. [26] Y. Rostamiyan, A. Hamed Mashhadzadeh, A. SalmanKhani, Optimization of mechanical properties of epoxy–based hybrid nanocomposite: effect of using nano silica and high–impact polystyrene by mixture design approach, Mater. Des. 56 (2014) 1068–1077. [27] P. Czub, Synthesis of high–molecular–weight epoxy resins from modified natural oils and Bisphenol A or Bisphenol A–based epoxy resins, Polym. Adv. Technol. 20 (2009) 194–208. [28] Y. Ye, H. Chen, J. Wu, L. Ye, High impact strength epoxy nanocomposites with natural nanotubes, Polymer 48 (2007) 6426–6433.

[29] X. Li, Y. Xia, W. Xu, Q. Ran, Y. Gu, The curing procedure for a benzoxazine–cyanate– epoxy system and the properties of the terpolymer, Polym. Chem. UK 3 (2012) 1629–1633. [30] M. Spontón, J.C. Ronda, M. Galià, V. Cádiz, Cone calorimetry studies of benzoxazine– epoxy systems flame retarded by chemically bonded phosphorus or silicon, Polym. Degrad. Stab. 94 (2009) 102–106. [31] S.R. Kumar, J. Dhanasekaran, S.K. Mohan, Epoxy benzoxazine based ternary systems of improved thermo–mechanical behavior for structural composite applications, RSC Adv. 5 (2015) 3709–3719. [32] J. Du, Y. Bu, Z. Shen, X. Hou, C. Huang, Effects of epoxy resin on the mechanical performance and thickening properties of geopolymer cured at low temperature, Mater. Des. 109 (2016) 133–145.