Novolac cured epoxy resin/fullerene modified clay composites: applied to copper clad laminates

Abstract

Functionalized inorganic layered materials based on a modification system composed of benzalkoniumchloride-N-methyl pryrrolidine-fullerene (BEN-(C₆₀–O)) were designed and fabricated in this paper, with the aim to develop high performance fire retardant epoxy composites without halogen and phosphorus. A BEN-(C₆₀–O) modifier was used to increase the inter-layer distance between the layered material. Based on the functionalization of montmorillonite type clay (CL88), a series of novolac cured epoxy resin/BEN-(C₆₀–O)-CL88 composites have been developed. The structural morphology and dispersion properties of the composites were investigated using wide-angle X-ray diffraction, and transmission electron microscopy. The composite reached V0 rating in the UL-94 vertical burning testing. Furthermore, the incorporation of BEN-(C₆₀–O)-CL88 into the novolac cured epoxy led to a significant reduction in peak heat release rate, total heat release and total smoke production, exhibiting superior fire resistance over its counterparts at the equivalent filler loading. Additionally, the composite shows improved thermo mechanical properties compared to the novolac cured epoxy resin.

---

1 Introduction

In recent years the focus of development for epoxy based polymer nano sized composites has been aimed towards their special properties such as mechanical,^1–3 thermal,⁴ and gas barrier properties⁵ as well as flammability resistance.^6,7 Among all the properties, fire-resistance is one of the most important. Epoxy has a fatal drawback of high flammability, which has severely restricted its application in fields requiring a remarkable flame-retardant grade. Thus, to improve the fire resistance one needs to pay close attention to general electronic products and printed circuit boards (PCBs). For many years various phosphorus,^8,9 silicon,^10–12 boron-containing compounds,¹³ and halogenated compounds have been incorporated into epoxy resins to improve flame retardancy. Although halogenated-epoxy compounds are used in PCBs to improve thermal and fire resistant properties,^6,14–16 these kinds of materials produce a large amount of smoke and toxic material like dioxin during combustion. The rise of nano sized composite technology has provided a revolutionary new solution to flame retardant polymer materials which need to be replaced as far as environmental pollution is concerned. Novolac cured resins have been widely used for the purpose of impregnating fiberglass to prepare high frequency printed circuit boards (HFPCBs).

The preparation of polymer/layered inorganic nano sized composites (PLNs) was extensively studied in the past two decades with respect to clay material and C₆₀, both of which have attracted considerable attention to enhance the thermal stabilities and flame retardancy of a polymer matrix with a small amount¹⁷ of layered inorganic material. Considering the structure of fullerene in a layered inorganic compound, it may act as a barrier slowing down the release of heat and hindering the transfer of combustion gases into the flame zone and energy feedback. Accordingly, we suppose that fullerene can be used as layered nanofillers in a polymer matrix to enhance thermal stability and flame retardancy. Recently, fullerene-clay composites have been synthesized by D. Gournis et al. in which the water-soluble fulleropyrrolidines were intercalated into clay through an ion exchange process.¹⁸

For the preparation of nano sized composites, the compatibility between a polymer and an inorganic layered material must be considered. In our continuous research regarding polymer/layered material composites,^5,19–23 we described novolac cured epoxy resin/modified clay composites which are applied in copper clad laminate. In order to improve properties of copper clad laminates, firstly, the CL88 was treated with bifunctional modifiers i.e. a mixture of benzalkonium chloride (BEN) and fullerene-pyrrollidine (C₆₀–O). Furthermore, the novolac cured epoxy resin/BEN-(C₆₀–O)-CL88 with different weight% compositions are successfully prepared using an in situ cross-linking polymerization method. Among all epoxy composites, the novolac cured epoxy resin/BEN-(C₆₀–O)-CL88 3 wt% was applied to make a copper clad laminate (CCL). Prepregs were prepared by impregnating five woven fiberglass sheets with a pure novolac cured epoxy resin and also with the novolac cured epoxy resin/BEN-(C₆₀–O)-CL88 3 wt% composite as described in the experimental section. Properties like water uptake, adhesion, coefficient of thermal expansion (CTE), and thermal stability were measured and also compared with the pure novolac cured epoxy resin.

2 Experimental

2.1 Materials

CL88 (trade name CO₂), was supplied by CHINA GLAZE GROUP, with a cation exchange capacity (CEC) of 200 meq/100 g of clay and its chemical formula is [Ca_0.37Na_1.28(Al_1.98Fe_2.02)Si₈O₂₀(OH)₄·1.46H₂O]. Fullerene (C₆₀) was purchased from AdvaNanoBio CO., Ltd. (75 wt%). Benzalkoniumchloride (BEN), used as a compatibilizing agent, was purchased from Showa Chemicals. The novolac cured epoxy resin and solvent propylene glycol methyl ether (PM) were purchased from ITEQ Cooperation (Taiwan). N-Methyl glycine (sarcosine) and paraformaldehyde were purchased from Showa Organics (98%) and Aldrich (95%), respectively. Solvents were used as received without further purification. A represents the bisphenol-A novolac resin and B represents the phenolic novolac cure resin. These two types of resin are mixed together which is the so-called novolac cured epoxy resin used in this study.

2.2 Preparation of water soluble N-methyl pryrrolidine-C₆₀ (C₆₀–O)¹⁶

2.416 g (27.6 mmol) of N-methyl glycine (sacrosine) and 2.07 g (69 mmol) of paraformaldehyde were added to 10 g (13.8 mmol) of fullerene (C₆₀) contained in a 500 ml toluene solution. The resulting mixture was heated for 2 h at 120–130 °C and then the solvent was removed under vacuum. N-Methyl pryrrolidine-C₆₀ (C₆₀–O) was formed.

2.3 Preparation of the modified clay BEN-(C₆₀–O)-CL88

In the preparation of BEN-(C₆₀–O)-CL88, firstly, 4 g of clay in 200 ml of distilled water was stirred overnight at room temperature. Concurrently in another beaker, 20 ml of 0.2 M HCl aqueous solution was added to 3.196 g of N-methyl pryrrolidine-C₆₀ (C₆₀–O) and 14.6 g of BEN and the resulting solution was stirred for 24 h at room temperature. The mixture solution was then stirred, centrifuged, and washed with deionized double-distilled water several times to remove the chloride ion and finally freeze dried.

2.4 Preparation of the BEN-(C₆₀–O)-CL88/novolac cured epoxy composites

Firstly, propylene glycol methyl ether (PM) in a ratio of clay/solvent = 0.1, was added separately to the various weight fractions of modified clay (1 wt%, 3 wt%, or 5 wt%) and the resulting mixture was stirred overnight. Then, in each case, the required amount of novolac resin was added to each solution and stirred for another 24 h. The solution was outgassed in a vacuum oven for a period of time. All samples were cured at 190 °C for 3 h.

2.5 Preparation of the copper-clad laminate

Among all the epoxy composites, the 3 wt% modified clay incorporated epoxy was chosen to prepare copper-clad laminates. The typical procedure is as follows: 15 g of modified clay (3 wt%) was first mixed with the solvent propylene glycol methyl ether (PM) in a ratio of clay/solvent = 0.1 and stirred for 24 h. Then 8.07 g of the novolac epoxy resin was added to the modified clay and stirred for another 24 h. Then the solution was soaked in five pieces of glass fiber sheets separately and dried in an oven at 170 °C for about 3 min followed by stacking of the sheets. Then the stacked fiber glass sheets were hot pressed at high temperature (∼190 °C) and pressure (∼300 to 400 Psi).

2.6 Characterization of the layered materials/novolac cured epoxy composites

After curing all the slurries at 190 °C for 3 h, samples were made in a thickness, length, and width within 2.5, 40, and 25 mm, respectively. Wide angle X-ray diffraction (WAXRD) patterns of the samples were recorded using a PANalytical, PW3040/60 X’Pert Pro with Cu K_α radiation (45 kV, 40 mA) and wavelength λ = 1.54 Å. The scan angle covered 2° < 2θ < 80° for clays and composites at a scan speed of 3° min⁻¹. A thermo-gravimetric analyzer (TGA, Thermal Analysis, and TA Q50) was used to measure the amount of modifying agent intercalated within the clay layers and the decomposed temperature of the composites. Thermogravimetric analysis (TGA) was carried out using a SII TG/DTA6200 thermo-analyzer instrument at a linear heating rate of 10 °C min⁻¹ under nitrogen flow. The dynamic mechanical analysis (DMA) measurements were performed using a TA-Q800 instrument in air at a scanning range of 30 °C to 250 °C with a heating rate of 3 °C min⁻¹. Details of the morphology and microstructures of the composite dispersions of clay were characterized using transmission electron microscopy (TEM JEM2010, 200 kV, and JEOL). Finally, a limiting oxygen index (LOI) test was carried out according to ASTM D2863 and the peak heat release rate (PHRR) was measured using cone calorimeters (Atlas Technology Corp. CONE 2 instrument). The UL94 test was carried using the ASTM D3801 method.

2.7 Characterization of the novolac cured epoxy resin/clay composites in copper clad laminates

Physical and mechanical properties of the different weight % of modified clay containing epoxy composites have been characterized. It has been observed that material doping with 3 wt% of modified clay into the epoxy polymer shows excellent physical and mechanical properties. Therefore, we expected that these composites could be applied in copper clad laminates. We measured the important properties of copper clad laminates (CCLs).

2.8 Water uptake

The CCL was cut to 2 inches in width and length. The sheet was then dried in an oven at 70 °C to remove any water in the sheet followed by measuring the weight. Then it was kept in a pressure cooker containing water and was heated at 120 °C for 3 h. Then the weight of the CCL was again measured and the difference of the weights is the amount of water uptake. The same procedure was followed for all other composites and the pure novolac cured epoxy resin to obtain the water uptake values and then these values were compared with the water uptake value of the pure novolac cured epoxy resin.

2.9 Dipping

The CCLs, with the same width and length as used in the determination of the water uptake, were suspended in a solder bath at 288 °C for 10 min. The surface of the CCL was checked about once per minute. If there was no blister on the surface for over 10 min, the CCL passed the test.

2.10 Adhesion

After the final stage, both sides of the CCL were covered with copper-foil. The peel strength between the surface of the copper clad laminate and the copper-foil was measured using a Shimadzu AGS 500 G tester with a load cell of 500 g. The CCL was then etched to 15 mm in width and 100 mm in length and the copper foil was pulled out to separate it from the CCL and the tensile strength was measured. The operation condition was 50.8 mm min⁻¹. The same process was applied in the case of the pure novolac cured epoxy resin and all other composites and compared with the storage modulus of the pure novolac cured epoxy resin.

2.11 Coefficient of thermal expansion (CTE)

Thermomechanical analysis (TMA) is one of the important characterization techniques in the field of thermal analysis. In the TMA technique, a number of different probe configurations are offered in order to optimize the test conditions for a specific sample and/or application. The most commonly used TMA probe is the expansion probe. This probe rests on the surface of the test specimen under low loading conditions. As the sample expands, during heating, the probe is pushed up and the resulting expansion of the sample is measured. The apparatus used in this study is a Perkin-Elmer TMA7 equipped with a classical expansion quartz probe. Its weight is fixed at 50 mN and the displacement sensibility is 50 nm. The thermal calibration of the instrument was performed with zinc and indium. The heating rate was 10 °C min⁻¹. The CCL samples of both the pure novolac cured epoxy resin and the novolac cured epoxy resin/modified clay (3 wt%) composites were cut into ca. 5 × 5 × 5 mm in length, width and thickness and dried before measurement.

3 Results and discussion

3.1 Characterization of the fullerene and modified clay

To prove that the interlayer spacing of the clay was indeed expanded by the introduction of C₆₀–O, we employed X-ray diffraction to determine the d₀₀₁ spacing. The characterization of C₆₀–O and the modified clay was determined by Wide angle X-ray diffraction (WAXRD). Fig. 1(A) shows the WAXRD graphs of powder C₆₀ and C₆₀–O. In Fig. 1(A), the new peak observed at 2θ = 19.57° is the (211) phase of N-methyl pryrrolidine. The d-spacing of pristine CL88 is about 12.63 Å, but when CL88 was intercalated with the mixture of N-methyl pryrrolidine-C₆₀ (C₆₀–O) and BEN, the d-spacing is increased to about 19.81 Å as shown in Fig. 1(B). The variation of d-spacing is caused by the orientation of C₆₀–O and BEN, of which the theoretical molecular sizes are 7–7.5 Å and 13–24 Å, respectively. The thickness of the brucite-like clay sheet is 2.54 nm and the average diameter of CL88 is 500–600 nm. C₆₀–O and BEN structures are schematically represented in Fig. 2.

Fig. 1 WAXRD graphs for (a) C₆₀ and C₆₀–O (b) pure CL88 and modified CL88.

Fig. 2 Schematic representation of the modified clay material, epoxy composite morphology and burning behavior.

In the FT-IR spectra of pure C₆₀ and modified C₆₀ shown in Fig. 3(A), bands that appeared at 3445 and 2918 cm⁻¹ are due to free water and the –CH₂– asymmetric stretching frequency of C₆₀. The vibrational bands above ∼1000 cm⁻¹ are predominantly due to displacement tangentials to the C₆₀ surface. A medium band appearing at 2931 cm⁻¹ and a weak band at 2768 cm⁻¹ in the modified C₆₀ (C₆₀–O) are assigned to υ_(–CH2) and υ_(N–CH3), respectively. Raman spectroscopy has been used to evaluate the effect of the modifier on the properties of the used C₆₀. In Fig. 3(B), the Raman spectrum of C₆₀ and C₆₀–O shows that, except for a partial contribution of the background signal, both the C₆₀ samples exhibit similar D-band (∼1300 cm⁻¹) and G-band (∼1600 cm⁻¹) peaks, which are associated with disordered and fullerene carbon, respectively. The D-band/G-band peak intensity ratio of C₆₀ and C₆₀–O are 2.529 and 4.420 respectively, indicating that additional disorder is introduced during the modification process.

Fig. 3 (A) FT-IR and (B) Raman spectra of C₆₀ and C₆₀–O.

The thermal stability of pristine CL88 and modified CL88 has been investigated and the important results are presented in Fig. 4. The decomposition temperature of pure CL88 is 450 °C. The weight loss observed at 230 °C in the TGA curve of modified-CL88 is due to the loss of crystalline water and the decomposition of BEN. The decomposition temperature of C₆₀–O is ∼340 °C. Fig. 4 shows that the 5% decomposition temperature (T_5d) of C₆₀–O is increased to about 390 °C after intercalation. This is evidence of the intercalation of C₆₀–O into the clay layer. The above results reveal that the modified clay is successfully synthesized. TGA not only measures the thermal stability of modified clays but also determines the amount of modifying agent intercalated into the clay layer and the results show that the intercalation amount of BEN is 27.09% and that of C₆₀–O is 47.41%.

Table 1 TGA and solution gel time results of novolac pure-epoxy and epoxy composites

Sample	T5d (°C)	Char yield	SG (sec.)
Novolac pure-epoxy	395.6	0	>1500↑
CL88-C60–O/BEN-1 wt%	389.7	0.5	461
CL88-C60–O/BEN-3 wt%	395.6	1.5	262
CL88-C60–O/BEN-5 wt%	398.0	2.1	170

---

Fig. 4 TGA curves of C₆₀–O, pure CL88 and BEN-(C₆₀–O)-CL88.

3.2 Morphology of the novolac cured epoxy resin modified clay composite

The morphology of the pure novolac cured epoxy resin and novolac cured epoxy nano sized composites containing CL88-(C₆₀–O)-BEN 1, 3 and 5 wt% was confirmed by XRD and TEM and the results are shown in Fig. 5 and 6. XRD patterns of the pure novolac cured epoxy resin show two broad peaks at 2θ = ∼4.4° and 19.5°. After incorporation of 3 wt% and 5% of CL88-C₆₀–O-BEN in the novolac cured epoxy, the characteristic peak of clay completely disappeared, which is evidence that the clay is well dispersed in the novolac cured epoxy polymer matrices. However, XRD patterns cannot be considered as conclusive evidence of monolayer exfoliation because they detect average diffractions from the overall X-ray irradiated sample area. Therefore, TEM is the best way to describe the morphologies of the composites. Among all the samples we chose the 3 wt% composite for TEM and the dispersion of the clay in the novolac cured epoxy composites are shown in Fig. 6 which indicates primarily an intercalated and partly exfoliated morphology. The dark black spot represents fullerene.

Fig. 5 WAXRD patterns of pure novolac epoxy resin and BEN-(C₆₀–O)-CL88/novolac cured epoxy composites.

Fig. 6 TEM image of the BEN-(C₆₀–O)-CL88/novolac cured epoxy 3 wt% composites.

3.3 Solution gel time measurement

Table 1 shows the solution gel (SG) time of the slurries as the precursor of the novolac cured epoxy resin/CL88-(C₆₀–O)-BEN composites with 3 wt% of modified clay. The solution gel time for the pure epoxy is longer than 1500 seconds. The SG time is decreased to 262.2 seconds with 3 wt% of modified clay containing epoxy resin. The SG time longer than 1500 seconds shows no reactivity while the shorter SG time indicates good reactivity. This is because the incorporation of the layered material in the epoxy polymer affects the cross linking reaction i.e., the reaction between the layered material and epoxy resin becomes fast which decreases the SG time.

3.4 Thermal stability of the epoxy composite

The thermal stability of the functionalized BEN-(C₆₀–O)-CL88 and BEN-(C₆₀–O)-CL88 based epoxy composites has been investigated in Fig. 7(A). Table 1 summarizes the 5% weight-reduction temperature (T_5d) of the pristine pure novolac cured epoxy and novolac cured epoxy composites. It was found that the thermal stability of the pristine novolac cure epoxy is 395 °C, whereas the T_5d for novolac cured epoxy/BEN-(C₆₀–O)-CL88 3 wt% composites was slightly increased to 398 °C. Thermal stability increased due to the introduction of organic structures in the modifiers. The incorporation of the functional BEN-(C₆₀–O)-CL88 led to an increase in the char yield at 800 °C. The pure EP has 0% residue at 800 °C, whereas epoxy/BEN-(C₆₀–O)-CL88 1 wt%, epoxy/BEN-(C₆₀–O)-CL88 3 wt%, and epoxy/BEN-(C₆₀–O)-CL88 5 wt% composites have 0.5, 1.5 and 2.1% residues, respectively.

Fig. 7 (A) TGA, (B) DSC, (C) DMA, and (D) TMA curves of the pure novolac cured epoxy and BEN-(C₆₀–O)-CL88/novolac cured epoxy composites.

3.5 Mechanical properties

Differential Scanning Calorimetry (DSC) of the pure novolac cured epoxy and novolac cured epoxy/BEN-(C₆₀–O)-CL88 3 wt% composite results are summarized in Table 1. DSC thermograms are shown in Fig. 7(B). The glass transition temperature (T_g) of pure epoxy is 188.1 °C whereas the T_g for novolac cured epoxy/BEN-(C₆₀–O)-CL88 3 wt% composites is decreased to 184.2 °C. The lower T_g was attributed to the enlarged free volume arising from the interface between the fillers and epoxy resin, which provides more space for polymer chain segments to move even at a lower temperature.

The thermomechanical properties of epoxy composites are measured by dynamic mechanical analysis (DMA) (Table 2). The storage modulus (E′) generally increases with increasing crosslinking density. When the modified clays contain a curing agent the storage modulus E′ is increased from 1950.60 to 2186.1 Mpa for the pure novolac cured epoxy and novolac cured epoxy/BEN-(C₆₀–O)-CL88 3 wt% respectively. Also the glass transition temperature (T_g) of composites can be estimated by plotting tan δ versus temperature curves (Fig. 7(C)) from DMA. From this result, T_g of pure novolac cured epoxy is higher (194.4 °C) than that of the corresponding epoxy composites (197.2 °C). This might be due to the fact that the clay inhibits the growth of the polymer chain and consequently it becomes short.

Table 2 DMA, LOI, UL-94 and combustion parameters obtained from the cone calorimeter test results of EP and EP composites

Sample	Novolac pure-epoxy	Epoxy/CL88-C60–O/BEN-3 wt%
Storage modulus (Mpa)	1950.86	2186.1
Tg	194.4	197.2
LOI	22.5	30
UL-94	V-1 test is not achieved	V-1 test is passed
pHRR (kW m^−2)	599.03	567.77
THR (MJ m^−2)	92.01	86.45

---

3.6 Burning behaviour

The LOI is the minimum percentage of oxygen in an oxygen–nitrogen mixture that will initiate and support for 3 min a candle like burning of a polymer sample. The fire-resistance properties of novolac cured epoxy composites are listed in Table 2. From the LOI results it is observed that, after the incorporation of modified clay in the epoxy resin, the LOI value increased from 22.5 in the pure epoxy resin to 30 in the epoxy composite. In the pure epoxy polymer and novolac cured epoxy resin/BEN-(C₆₀–O)-CL88 3 wt% composite, the total heat released is decreased from 92.01 to 86.45 MJ m⁻¹. The UL94 test was carried out for both the pure novolac cured epoxy and novolac cured epoxy resin/BEN-(C₆₀–O)-CL88 3 wt% composite. These results are also summarized in Table 2. The nano sized composites passed the UL94-V1 test. In the case of epoxy composites, the increasing LOI value is due to the incorporation and dispersion of CL88, which provides the silicate layer and C₆₀–O which contains a number of C=C groups. During combustion, the pathway from the air into the composites is increased by the silicate layer and the hydroxyl group producing the mist, and we expect that C=C groups act as a free radical trap. However, it was noted that adding BEN-(C₆₀–O)-CL88 into the novolac cured epoxy resulted in a great improvement in fire resistance and the UL-94 V0 rating was achieved in the vertical burning test. These composites with excellent improvement of the fire-resistance property can be applied to making a printed circuit board.

Cone calorimeter tests were carried out for measuring the burning behavior of polymeric materials in bench-scale tests. It was found that the pure epoxy burnt very rapidly after ignition and the peak heat release rate value is 599.03 kW m⁻². Compared with pure epoxy, epoxy composites burnt relatively slowly and the peak HRR decreased to 567.77 kW m⁻². The improved fire retardancy of epoxy composites led to a good dispersion of the modified clay in the epoxy matrix; additionally, C₆₀ species improved the char yield of the epoxy composites during combustion. Pure epoxy released a total heat of 92.01 MJ m⁻², while the epoxy composite released heat at 86.45 MJ m⁻². The significant reduction in THR meant that more organic structures in the epoxy resin participated in the carbonization process and were kept in the condensed phase, rather than being converted to “fuel” in the gas phase. This was also evident from the increased char residues.

The coefficient of thermal expansion (CTE) is a very important parameter for evaluating the usage stability and reliability of the electrical/electronic substrate. It can be easily measured by TMA using an expansion probe. The CTE is a quantitative assessment of expansion of a material over a temperature interval. The CTE values are measured below the T_g value on copper clad laminates containing novolac cured epoxy resin with 3 wt% modified clay. The copper clad laminate of the pure novolac cured epoxy has a CTE value of 62.10 ppm °C⁻¹ and the CTE value is decreased to 60.60 ppm °C⁻¹ for novolac cured epoxy/BEN-(C₆₀–O)-CL88 3 wt% composites below T_g (α₁) in the thickness (Z) direction. In addition, the water uptake was decreased from 0.34% for pure novolac cured epoxy to 0.26% for composites using BEN-(C₆₀–O)-CL88 as modified clays. This result is very desirable especially in electronic applications where water absorption is detrimental to dielectric performance. As for the thermal stability, the durable time in a 288 °C solder bath was significantly increased over 10 min.

4 Conclusion

A series of nano sized composites that are composed of novolac cured epoxy/modified-fullerenes-clay (BEN-(C₆₀–O)-CL88) has been synthesized and characterized using WAXD, TGA, DMA and TEM. Due to the introduction of the BEN-(C₆₀–O)-CL88 moiety, the obtained thermosets had a high T_g and exhibited excellent thermal stability and improved flame retardant properties as shown by the evaluation of the LOI measurement and the UL-94 vertical test. The material is highly flame retardant due to the presence of clay and fullerene which acts as a free radical trap. The copper clad laminates containing modified clays pass strict tests, exhibiting low water uptake, low coefficient of thermal expansion, high thermal stability, and strong copper adhesion and all of these improved properties are due to the presence of the modified clay using modified C₆₀–O and BEN. BEN acts as a compatibilizing agent as well as a curing agent. This result suggests that the novolac cured epoxy resin/clay composites with balanced properties can be potentially applied in electrical/electronic industries to make copper clad laminates.

Acknowledgements

The authors are grateful to National Science Council Taiwan, and the Center-of-Excellence Program on Membrane Technology, the Ministry of Education, Taiwan, R. O. C. under Grant NSC-95-2113-M-033-008-MY3.

References

1. A. J. Kinloch and A. C. Taylor, Mechanical and Fracture Properties of Epoxy/Inorganic Micro and Nano composite, J. Mater. Sci. Lett., 2003, 22, 1439–1441.
2. A. Haque, M. Shamsuzzoha, F. Hussain and D. S. Dean, S2-glass/epoxy polymer nanocomposites: manufacturing, structures, thermal and mechanical properties, J. Compos. Mater., 2003, 37, 1821–1837.
3. Y.-Q. Zhang, J. H. Lee, J. Jang and C. W. Nah, Preparing PP/clay nanocomposites using a swelling agent, Composites, Part B, 2004, 35, 133–138.
4. O. Becker, R. J. Varley and G. P. Simon, Thermal stability and water uptake of high performance epoxy layered silicate nanocomposites, Eur. Polym. J., 2004, 40, 187–195.
5. T.-Y. Tsai, C.-H. Li, C.-H. Chang, W.-H. Cheng, C.-L. Hwang and R.-J. Wu, Preparation of Exfoliated Polyester/Clay Nanocomposites, Adv. Mater., 2005, 17, 1769–1773.
6. M. Zanetti, G. Camino, D. Canavese, A. B. Morgan, F. J. Lamelas and C. A. Wilkie, Fire Retardant Halogen–Antimony–Clay Synergism in Polypropylene Layered Silicate Nanocomposites, Chem. Mater., 2002, 14, 189–193.
7. M. Hussain and G. P. Simon, Fabrication of phosphorus–clay polymer nanocomposites for fire performance, J. Mater. Sci. Lett., 2003, 23, 1471–1475.
8. C. H. Lin, T. Y. Hwang, Y. R. Taso and T. L. Lin, Phosphorus-Containing Epoxy Curing Agents via Imine Linkage, Macromol. Chem. Phys., 2007, 208, 2628–2641.
9. X. Wang, Y. Hu, L. Song, W. Y. Xing, H. D. Lu, P. Lv and G. X. Jie, Effect of a triazine ring-containing charring agent on fire retardancy and thermal degradation of intumescent flame retardant epoxy resins, Polym. Adv. Technol., 2011, 22, 2480–2487.
10. Y. Zheng, K. Chonung, X. L. Jin, P. Wei and P. K. Jiang, Study on the curing reaction, dielectric and thermal performances of epoxy impregnating resin with reactive silicon compounds as new diluents, J. Appl. Polym. Sci., 2008, 107, 3127–3136.
11. L. A. Mercado, M. Galia and J. A. Reina, Silicon-containing flame retardant epoxy resins: synthesis, characterization and properties, Polym. Degrad. Stab., 2006, 91, 2588–2594.
12. X. Wang, Y. Hu, L. Song, W. Y. Xing and H. D. Lu, Thermal degradation behaviors of epoxy resin/POSS hybrids and phosphorus–silicon synergism of flame retardancy, J. Polym. Sci., Part B: Polym. Phys., 2010, 48, 693–705.
13. H. Y. Yang, X. Wang, B. Yu, L. Song, Y. Hu and R. K. K. Yuen, Effect of borates on thermal degradation and flame retardancy of epoxy resins using polyhedral oligomeric silsesquioxane as a curing agent, Thermochim. Acta, 2012, 535, 71–78.
14. M. L. Christopher Preston, G. Amarasinghe, L. J. Hopewell, R. A. Shanks and Z. Mathys, Evaluation of polar ethylene copolymers as fire retardant nanocomposite matrices, Polym. Degrad. Stab., 2004, 84, 533–544.
15. S. Wang, Y. Hua, R. Zong, Y. Tang, Z. Chen and W. Fan, Preparation and characterization of flame retardant ABS/montmorillonite nanocomposite, Appl. Clay Sci., 2004, 25, 49–55.
16. J. Wang, J. Du, J. Zhu and C. A. Wilkie, Chemiluminescence of polyamides III. Luminescence accompanying thermooxidation of lactam-based polyamides stabilized by antioxidants, Polym. Degrad. Stab., 2002, 77, 249–258.
17. M. Maggini and G. Scorrano, Addition of azomethine ylides to C₆₀: synthesis, characterization, and functionalization of fullerene pyrrolidines, J. Am. Chem. Soc., 1993, 115, 9798–9799.
18. D. Gournis, V. Georgakilas, M. Karakassides, T. Bakas, K. Kordatos, M. Prato, M. Fanti and F. Zerbetto, Incorporation of Fullerene Derivatives into Smectite Clays:  A New Family of Organic–Inorganic Nanocomposites, J. Am. Chem. Soc., 2004, 126, 8561–8568.
19. T.-Y. Tsai and S.-T. Lu, Mechanochemical treatment of vermiculite in vibration milling and its effect on lead(II) adsorption ability, Clay Sci., 2005, 12, 74–78.
20. T.-Y. Tsai, S.-J. Jong, A.-C. Yeh, C. J.-B. Chiang and R. Fang, US Pat., US 6548159B2, 2003.
21. T.-Y. Tsai, W.-H. Lin, Y.-Y. Lin, Y.-C. Hsu, U. Ray, Y.-T. Lin and M.-J. Ou, Permeability property of Nylon 6 nanocomposite membranes with various clay minerals, Desalination, 2008, 233, 183–190.
22. T.-Y. Tsai, S.-T. Lu, C.-H. Li, C.-J. Huang, J.-X. Liu and L.-C. Chen, Effect of bifunctional modifiers of the clay on the morphology of novolac cured epoxy resin/clay nanocomposites, Polym. Compos., 2008, 29, 1098–1105.
23. T.-Y. Tsai, S.-T. Lu, C.-J. Huang, J.-X. Liu, C.-H. Li, L.-C. Chen and U. Ray, The structure–property relationship of novolac cured epoxy resin/clay nanocomposites, Polym. Eng. Sci., 2008, 48, 467–476.

---