Improving flame-retardant, thermal, and mechanical properties of an epoxy using halogen-free fillers

2.1 Materials

AG with average pore size of 20 nm, porosity of >95%, and density of 0.06 g/cm³ was supplied by Vaspart Co. Ltd. (Isfahan, Iran). Fine powders of APP (average size of 5 μm, density of 1.98 g/cm³) were purchased from Pars Idea Chem Co. Ltd., Tehran, Iran. Diglycidyl-ether bisphenol A (Epikote 828) resin and tetra-amine based hardener (HA-11) with a mass ratio of 10:1 were used (Shimi Afsoon Co. Ltd., Tehran, Iran). A 15 wt% reactive diluent, Cardura E10P, existed in the resin.

2.2 Preparation

Composites were prepared using a solution processing method [13]. The samples contained (A) only APP, (B) only AG, or (C) both APP and AG. The volume fractions of fillers in composite types A and B were 0%, 5%, 10% and 20%. For composite type C, the filler contents were 15 vol% (APP:AG volume ratio of 1:2) and 20 vol% (APP:AG volume ratios of 3:1, 1:1 and 1:3). Hereafter, samples are referred to by the volume percentage of each phase followed by the phase name, e.g. sample containing 10 vol% AG and 5 vol% APP is 10AG5APP.

In order to obtain a better dispersion of the constituent materials, the epoxy resin was heated to 50°C to lower its viscosity. Subsequently, adequate volume fractions of AG and APP were added and the mixture was mechanically stirred for 1 h at 2000 rpm. Subsequently, the obtained mixture was degaussed in a vacuum oven at room temperature for 30 min. The hardener (with a weight ratio of 10:1 resin to hardener) was then added and the mixture was mechanically stirred for another 10 min. Finally, the prepared mixture was poured into molds and the whole system was cured in ambient temperature in a vacuum oven for 12 h.

2.3 Characterization

A flame chamber was used to perform the UL-94 (plastics flammability standard) horizontal burning test following the ASTM D 635-10. The anti-dripping and combustion properties of the composites were extracted primarily from this test. The standard samples for UL-94 were fabricated to dimensions of 100×12×2 mm³. The LOI values were measured using an HC-2 oxygen index meter (Jiangning Co, China), according to the ASTM D 2863 with a sample size of 130×6.5×3.2 mm³. The decomposition behavior and thermal stability of the composites were quantified by thermo-gravimetric analysis (TG/DTA 6300, SII Nanotechnology, Japan). TGA was performed in air at a heating rate of 10°C/min from room temperature to 700°C following the ASTM E 1131. Differential scanning calorimetry (DSC, 2000 FC by Netzsch Co., Germany) and dynamic mechanical analysis (DMA, Q800 by TA Instruments, USA) were used to evaluate the influence of FR additives on the glass transition temperature (T_g) and mechanical performance of the composites. In accordance with ASTM D7028-07, the DMA test specimens had a span to thickness ratio of >10. Tests were performed at a frequency of 1 Hz, strain amplitude of 0.1%, heating rate of 5°C/min, and temperature range of approximately 30°C–120°C. For DSC, 10 mg of each sample was heated from ambient temperature to 300°C at a heating rate of 10°C/min under a nitrogen flow (ASTM D 3418). The average values of the measured properties for a minimum of five different samples were reported for each composite to ensure reproducibility and accuracy.

Thermal conductivity of the samples was measured using a discovery DXF-200 flash diffusivity system (TA instruments, USA). Tablets with a diameter of 1.27 and a thickness of 1–2 mm were cut from the samples. Xenon flash induces a uniform heat pulse through the sample if it is opaque. Samples in this study were translucent, thus a graphite spray coat (Rust-Oleum graphite spray) was applied to one side of the sample. The comparative mode of the DXF-200 was used to simultaneously measure the specific heat and thermal diffusivity of the samples. In this measurement the flash method is simultaneously applied to a reference (Austenite stainless steel in this study) and an unknown sample where identical flash power and thermal conditions are applied to both samples. By definition, thermal conductivity was then calculated from the following equation:

where λ is the thermal conductivity, α is the thermal diffusivity, ρ is the density, and C_p is the specific heat. Density was calculated manually by measuring the mass (Nevada Radwag scale, USA: 0.0001 g) and volume (dimensions were measured using a Mitutoyo Caliper, USA: 0.1 microns) of a rectangular sample.

An FEI Q3D scanning electron microscope (SEM, Zeiss, Germany) was used to acquire micrographs of samples. For SEM imaging, samples were immersed in LN2, fractured immediately after, and their fractured surfaces were coated with 4 nm of carbon.

3.1 Morphology

Representative SEM images of the fractured surface of the different samples are shown in Figure 1. Distribution of APP in 20APP/epoxy composite is shown in Figure 1B. Although larger sized APP particles are observed in the micrograph, APP is still homogenously dispersed in the epoxy matrix. The SEM micrograph of the 20AG/epoxy composite (Figure 1C) shows that AG particles are engulfed and encapsulated in the matrix. This is because AGs are composed of an interconnected network of channels. Furthermore, compared to the 20APP sample, some large pores are visible on the surface of the 20AG composite. These pores are the origin of dropping mechanical and physical properties of the composites. Morphology of the sample containing APP and AG is characterized in Figure 1D (5AG15APP). A close observation reveals that there are no considerable differences in the dispersion degree of the constituent materials, compared to the sample containing APP or AG alone.

Figure 1:

Representative SEM micrographs of the: (A) neat epoxy, (B) 20APP, (C) 20AG and (D) 5AG15APP samples showing effective dispersion of AG and APP phases.

3.2 Thermal properties

Table 1 summarizes the results of UL-94 test, i.e. burning length, burning time, and burning rate. Only the neat epoxy samples exhibited dripping without significant formation of char, while the addition of AG and/or APP eliminates dripping for the duration of the experiment. In general, dripping of flaming polymer during combustion promotes the progression of fire [14]. On the contrary, formation of stable surface layer of char blocks further ignition by limiting the heat and mass transfer from the flame to the polymer and from the polymer to the flame, respectively [14].

The addition of AG to the epoxy matrix decreases the burning rate of the epoxy, however, AG loading in excess of 10 vol% has no considerable effect on the combustion rate of the composites. The flammability resistance of the composite shows dramatic improvement when infused with APP. For the 5APP sample, flame extinguished within the first 25 mm, while in 10APP and 20APP samples no flame was produced in atmospheric conditions. The APP/AG/epoxy composites with 10 vol% APP or higher also did not burn in the horizontal burning test. Lastly, the 5APP system with the addition of 10 vol% of AG shows a reduction on the burning time from 30 to 17 s.

In order to investigate the effect of APP and AG addition on the epoxy composite, the 20APP and 15APP5AG samples were examined under a modified burning procedure. In this experiment, the sample was positioned horizontally above a 7 cm flame for 60 s and subsequently examined. As shown in Figure 2, the addition of 5 vol% AG results in a noticeable reduction of charring on the surface of the sample. These carbonaceous chars seem to retard the heat transfer and substantially improve the flammability properties of the fabricated composites. Also it seems that the existence of AG in the composite leads the flame toward the edges of the sample, thus lowering the internal sample degradation. This phenomenon is most probably ascribed to the excellent thermal insulation properties of AG.

Figure 2:

The char residue in: (A) 20APP and (B) 15APP5AG.

To better understand the combined effect of the fillers, the LOI of the samples was measured. The LOI with respect to various filler concentrations is presented in Table 1. Neat epoxy, with the LOI value of 20.2%, is easily combustible in atmospheric condition. With the addition of 5 vol% and 10 vol% AG, the LOI value slightly increases to 23% and 24.1%, respectively. Further increase in the AG above 10 vol% did not improve the LOI. A material must have a minimum LOI of 26% in order to be categorized as self-extinguishing [10], [15]; the addition of the AG did not make the composite self-extinguishing. Compared with epoxy composites containing only AG, the LOI values of the APP/epoxy composites show substantial improvement. The LOI values increase dramatically with increasing concentration of APP up to 20 vol%, where the LOI reaches a maximum of 41%. It is therefore clear that APP is more capable of improving self-extinguishing behavior in epoxy, especially at higher loadings.

Investigation of the LOI data of the APP/AG/epoxy composites provides some interesting insight into the trends associated with various filler ratios. At a constant concentration of AG (10 vol%), increasing APP loading increases the LOI. Conversely, at 10 vol% APP, the addition of AG has no effect on the LOI. A slight decrease of LOI is observed for the 5APP15AG sample compared with the 5APP10AG. This is likely due to difficulty in dispersion of AG in epoxy matrix at higher loadings. The 5AG15APP exhibited a large LOI of 38.6%.

Figure 3 shows TGA and differential thermal analysis (DTA) profiles for the different samples. The primary results of Figure 3 are summarized in Table 2. Three temperatures are extracted from TG curves: (1) T₅, which demonstrates the temperature of a small weight loss of 5%, (2) T_max, exhibits the temperature at which the maximum weight loss rate occurs, and (3) T_f, temperature at which constant residual weight of sample is achieved. T₅ also represents apparent thermal stability of composites [16]. The initial mass loss below 150°C is due to the possible evaporation of the absorbed water and the remaining reactive diluent (Cardura E10P) [16]. The possible monomeric compounds remaining from the incomplete polymerization of the epoxy is also evaporated below 150°C. The T₅ of the 20APP sample is 186°C, which is about 21°C lower than that of the neat epoxy. Addition of 5 vol% AG in the 15APP5AG system increases T₅ to 202°C. In line with this trend, 20AG sample shows a 6°C increase in T₅ over the neat epoxy. It can be concluded that the addition of APP decreases the initial thermal stability of epoxy, while AG can slightly improve it. In fact, the presence of AG particles in the epoxy can be associated to hinder the evolution of volatile decomposition substances from the epoxy matrix [14], [17].

Figure 3:

TG and DT curves of the neat epoxy, 20APP, 20AG and 15APP5AG.

In neat epoxy, T_max occurs at 368°C. Addition of 20 vol% AG increases the T_max to 395°C. Moreover, total weight loss percentages for the neat epoxy and 20AG composite between 250°C and 450°C, are 78% and 70%, respectively. Accordingly, it is unambiguous that the presence of AG leads to major delay in thermo-degradation of epoxy due to physical barrier preventing of AG particles. With the addition of 20 vol% APP, T_max is sharply reduced to 313°C. In this case, the decomposition rate is also decreased, which facilitates the formation of heat-resistant char [18], [19]. The underlying mechanism of thermal degradation of APP/polymer composites has been comprehensively investigated in the literature [17], [20], [21]. As a brief explanation, chemical reactions of APP and epoxy resin lead to formation of phosphorous-rich layer, resulting in the lower degradation temperature and decomposition rate of APP/epoxy composites. At higher temperatures, the protective layer decomposes to form a compact char coating surface, which enhances the thermal stability of APP/epoxy composite over 450°C [20]. The remaining char of 20AG, 20APP and 15APP5AG samples is respectively ~6%, ~17% and ~17% higher than the neat epoxy. This residual char is a useful indicator for forecasting fire retardancy of the materials. The APP remarkably improves the char-forming behavior.

The T_f is evidently increased in the composites containing APP as the FR. Based on the results, one can conclude that addition of APP to the epoxy matrix results in a much better thermal performance at higher temperatures compared to the AG and neat epoxy samples. The AG integrated with APP, however, shows effective FR efficiency in epoxy matrix.

Incorporation of APP and AG in the epoxy matrix also affects its glass transition temperature (T_g). Using DSC (Figure 4), the extracted T_g values for the different samples are summarized in Table 3. The T_g of the reference neat epoxy is 66°C. When 10 vol% AG is incorporated into the epoxy matrix, T_g increases to 74°C. The observed increase is ascribed to well-dispersed AG particles in the matrix. In fact, improving dispersion uniformity of additives can hinder the segmental movements of cross-linked epoxy matrix, hence the increase in T_g [22]. Conversely, at higher volume fractions of fillers, increase in interactions between the monomers and fillers during the polymerization process leads to the reduction in the degree of crosslinking. Therefore, a drop in T_g is expected [22], [23]. For the 20AG, the T_g decreases to the T_g of neat epoxy. A slight decrease in T_g is also observed in the 20APP composite compared to the 10APP sample. Szolnoki et al. [22] reported significant improvement in T_g of the APP/epoxy composites due to the addition of small amounts of APP. The addition of AG to the APP/epoxy system also resulted in increase in the T_g. These results authenticate the synergistic effect of the APP and AG on the T_g of the epoxy composites.

Figure 4:

DSC curves of different samples.

As shown in Figure 5, storage modulus at different temperatures for the 20AG sample decreases compared to the neat epoxy sample and increases for the 20APP sample. As expected, the highly porous AG particles deteriorate the mechanical performance of the epoxy. The net effect of APP and AG additions to the next epoxy in the 5AG15APP sample results in an overall improvement in the performance over the neat epoxy sample. It is evident from the DMA curves that the transition to a rubbery state for both the 20APP and 5AG15APP samples occurs at a relatively higher temperature compared with the neat and 20AG samples.

Figure 5:

Temperature-dependent storage modulus for different samples.

Density, thermal diffusivity, specific heat, and the calculated thermal conductivity values for the different samples are summarized in Table 4. In general, the addition of APP and AG increases the thermal conductivity. Addition of 20 vol% AG to epoxy increases the thermal conductivity by only 20%, while the addition of 20 vol% APP increases it by 115%. This could be explained by the structural differences between the APP- and AG-based composites where AG contains pores and therefore has a much smaller thermal conductivity compared with APP. Interestingly, combining AG and APP in the 5AG15APP sample only increased the thermal conductivity by 40%. This can be responsible for the burning behavior of the 5AG15APP sample where the insulating properties of the AG hinder heat transfer to the bulk epoxy.

Using the rule of mixture a density of 590 kg/m³ is calculated for the AG phase in the 20AG sample. This value is much larger than the density of AG itself (~60 kg/m³). It is therefore concluded that AG particles are partially infused by the epoxy. As a result, although it was expected that addition of AG lowers the thermal conductivity of the base epoxy, it slightly increased it.