Combustion measurements, such as heat release rate, critical flux, time-to-ignition, ignition temperature, thermal inertia, and kinematics—activation energy as well as preexponential factor—on epoxy polymer (Prime 20LV) with expandable graphite (EG) inorganic filler of different weight percentage composites, are conducted using the Dual Cone Calorimeter, the thermogravimetric analysis (TGA), and Linseis (Germany) THB100 Transient Hot Bridge thermal conductivity analyser. The results indicate that increasing the amount of EG in polymer composite leads to reduction in the critical flux, the time-to-ignition, the ignition temperature, the thermal inertia, the average thermal conductivity, and the activation energy (from 159.1 ± 2.3 to 145.9 ± 3.1 kJ/mol for neat epoxy to 3 wt.% EG-epoxy) of the composite samples. There is, however, an increase in the heat of gasification with increasing EG content.

1. Introduction

Polymer-based materials are competing with metallic alloys in terms of cost and functionality such as durability, strength, and other physical and chemical properties. The shortened time of fabrication, toughness, strength, lightweight, and economic considerations are increasingly making polymer-based-material products comparable and even more usage-attractive and advantageous over their metallic alloy and ceramics counterparts [1]. Epoxy resins with their curing agents have been widely used in industry as protective coatings and for structural applications in construction [2]. This thermosetting polymer can be formulated to meet specific needs for certain physical and mechanical properties. It has high chemical and corrosive resistance, good mechanical and thermal properties, and outstanding adhesion to various substrates [3]. There have been ongoing number of studies that have focused on widening the applications of epoxy resins polymer materials based on enhancing one or more of its known physical, mechanical, and chemical properties [47]. Search for different additive reinforcements in the epoxy matrix for better applications is also ongoing [3, 7].

In thermal applications, epoxy based materials are limited by their organic nature. When exposed to a heat source, the challenges of fire hazards result in undesirable changes that affect product functionality and durability. The single most important variable in fire hazard is the heat release rate (HRR) [8]. Volatiles from these polymer-based composites contribute to product ignition with possible release of toxic smoke [9] and fire spread [10].

Quite a number of studies on heat effect on polymer composites are ongoing and findings are supporting refined and better methods in fabrication and performance of these polymer-based materials [11, 12]. Furthermore, inorganic intumescent additives such as expandable graphite (EG) to organic epoxy resin have shown promising flame retardant effects [1315]. Although, there have been a number of epoxy resin EG/C composite studies that have utilized the Cone Calorimeter and TGA as measuring methods, the EG contents in the composite have been relatively high [16] and the uses of mathematical models [17, 18], to appropriately describe parametric heat values, have been few.

As part of research directed towards improved performance of resin matrix composites, the thermal behaviour of one particular epoxy polymer, Prime 20LV, with relatively low doses of EG inorganic filler of different weight percentages, was the focus of this study. The choice of Prime 20LV was simply based on its very low mixed viscosity and long working time that allows reinforcements like EG to be infused successfully. The polymer and its expandable graphite composites were investigated using the Cone Calorimeter and the thermogravimetric analysis (TGA) instruments. Thermal conductivity measurements were done using Linseis (Germany) THB100 Transient Hot Bridge thermal conductivity analyser.

2. Experimental Details

2.1. Sample Preparation

The Prime 20LV (low viscosity) infusion epoxy resin and its slow Prime hardener with mixed density of 1.084 g·cm−3, glass transition temperature of 29°C, and viscosity of 23 cP at 20°C, were obtained from the Advanced Materials Technology (Pty) Ltd., in South Africa. The inorganic filler was expandable graphite (EG) of the trade name code ES250B5 with the following specifications: minimum carbon content of 90%, expansion volume of 0.25 m3 kg−1, and average particle size of 300 μm. The EG was obtained from University of Pretoria, via Qingdao Kropfmuehl Graphite (China). The intercalating agent is suspected to be sulphuric acid.

2.2. Epoxy Resin (Curing Process)

As per specifications from the supplier, the epoxy resin and its hardener were mixed and stirred vigorously in a mass ratio of 100 : 26, respectively, and poured into a Perspex mold of dimension of 100 × 100 × 5 mm3. The epoxy resin was then left, at room temperature, to cure for 24 hours.

2.3. Epoxy/Expandable Graphite Composites

Measured quantity of expandable graphite was carefully added to the mixture of the epoxy resin and its slow hardener and vigorously stirred continuously with a magnetic stirrer. When the viscosity of the composite began to increase, after 2 hours, the mixture was poured into the Perspex mold and left to cure for further 24 hours. The average thicknesses of the samples were 4.8 mm. In addition to the neat epoxy, the following sample compositions were fabricated for the various thermal measurements:Epoxy with 1 wt.% expandable graphiteEpoxy with 3 wt.% expandable graphiteEpoxy with 5 wt.% expandable graphite

The samples were subjected to the following set of external heat fluxes, which, according to [19], are heat fluxes found in developing fires: 25, 30, 35, and 50 kWm−2.

2.4. Cone Calorimeter

The Cone Calorimeter is considered as the most significant bench scale instrument in fire testing due to the small sample sizes used for the test [19]. Cone Calorimetric testing is a quick way to conduct research into the fire performance (in the development stage). From each of the set external heat fluxes mentioned above, the Fire Test Technology (FTT) Dual Cone Calorimeter, conformed to ISO 5660 standard, was used to measure the following parameters: heat release rate (); mass loss rate (); time-to-ignition (); critical heat flux (); heat of gasification () = heat of vaporization () and heat of combustion (); and average smoke density that is expressed in terms of average specific extinction area (SEA).

2.5. Thermogravimetric Analysis (TGA)

Thermal decomposition of the epoxy-EG composite was investigated further by Perkin Elmer TGA 4000 thermogravimetric analyser in an inert (N2) atmosphere for all samples, from neat to those with different EG compositions. TGA measurements of mass loss fraction versus temperature and time for different heating rates were done. Arrhenius plots for mass loss fraction () from the data obtained under nonisothermal conditions, at heating rates of 5, 10, 15, and 20 Kmin−1, were used to calculate the activation energies governing the thermal decomposition of the polymer composite samples.

2.6. Scanning Electron Microscope (SEM) Analysis

Morphological studies of the epoxy resin EG composites were carried out using SEM JEOL 700 with accelerating voltage adjusted between 15 and 20 keV. The samples were carbon coated.

2.7. Theory

Heat is the driving force for ignition and development of fire. The following approximate equations describe how the measured parameters were obtained with the Cone Calorimeter and the TGA.

The time that a polymeric material can withstand heat flux radiated by a fire before it experiences sustained flaming combustion is called the time-to-ignition [9]. Ignition occurs when the mass loss rate produces sufficient volatiles whose effective heat of combustion, at the characteristic air flow in the Cone Calorimeter setup, makes a gas mixture capable of being ignited by a spark [19]. As an important fire reaction property, time-to-ignition () is closely related to the material thermal inertial () which is the reluctance for a material to ignite when subjected to a heat source. For thermally thick material, the time-to-ignition is related to heat flux according to the following equation [18]: where is the external heat flux the material is subjected to, is the thermal conductivity, is the heat capacity, is the density, and and are the temperatures at ignition and ambient, respectively.

When exposed to a given net heat flux (), an epoxy composite sample vaporizes at a certain rate which is expressed as mass loss rate per unit area () according to the following:where is the heat of gasification (or vaporization) and is defined as the energy required to produce the fuel volatiles per unit mass of the sample.

The net surface heat flux for the gasification period can be approximated as follows [20]:where is the emissivity (=1) for surface char, is the incident flame heat flux, is the Boltzmann’s constant, and is the vaporization surface temperature = .

Equations (3) and (4) show the energy balance (convection and radiation modes) at the surface with the net heat flux and the surface temperature . When the net heat flux is theoretically zero and the surface temperature is equal to the ignition temperature, a critical heat flux responsible for bringing about ignition becomes the following:Putting (3) into (2) and assuming that incident flame heat flux constant gives analytical steady-state equation [16, 18]heat of combustion can also be determined from the heat release rate per unit area () of the sample from the following equation [17]:In analysing the thermal decomposition of the polymer composite samples, the first-order Arrhenius kinetic was assumed. The model for the decomposition rate therefore assumes an Arrhenius rate ofwhere is the preexponential factor, is the activation energy, is the universal gas constant, and is the holding temperature for decomposition.

The fraction of mass loss rate is assumed to be related to the decomposition rate of the polymer matrix as follows [17]:where , with and being final and initial sample mass and the conversion factor (), in terms of the residual, isPutting (8) and (10) into (9) and solving giveswhere is the heating rate.

A similar analytical equation follows the Kissinger method and supposes a first-order kinetic was used by Régnier and Fontaine [21]:where is the temperature at the maximum of the heating derived mass loss curve.

Equation (11) can further be simplified if is plotted against (from ) for different heating rates (). A set of the decomposition rate constants and its associated temperature could be tabulated from a particular mass fraction (say, ) to enable the determination of and [22].

3. Results and Discussion

Figure 1 clearly shows that the intumescent nature of the EG is more pronounced with increasing weight percentage; the residual swelling and char formation of the 5 wt.% EG is seen as the highest.

3.1. Heat Release Rate (HRR) Measurements with Time from Different External Heat Fluxes on the Epoxy-EG Composition Samples

For a particular HRR profile (see Figure 2) the following could be distinguished: time-to-ignition (OA), decomposition of the organic components (small hump, B, after ignition) and thereafter the rapid release of flammable volatiles that leads to main peak HRR (C) and HRR contribution from char (small peak to the right side of main peak, D), and HRR (E, about 32 kWm−2) from background residual and negligible char. The description fits the four stages of fire development which includes growth and preflashover stage, flashover, fully developed fire, and decay stages [23].

As expected, the profile from the highest external heat flux (50 kWm−2) leads to the shortest ignition time (57 s), the highest HRR value (1359 kWm−2), and a more pronounced peak. The peak HRR values decrease with decreasing external heat fluxes [10]. Similar profiles on sample V2FL30 irradiated with different heat fluxes were reported by Schartel and Hull [19].

External heat flux (50 kWm−2) leads to the shortest ignition time (37.5 s) and the highest HRR value (817.4 kWm−2). Similar to Figure 1, the peak HRR values decrease with decreasing external heat fluxes and the time-to-ignition decrease with increasing heat flux. There is a slight increase in the HRR value from residue char (about 51 kWm−2) for the 25 kWm−2 external heat flux due to the 1-wt.% EG in the composite as compared to the same heat flux in Figure 1. However, the HRR values from residue char in the higher external fluxes seemed lower than the 25 kWm−2 external heat flux. The higher external heat fluxes seemed to have burnt the organic resin completely.

In Figure 4, the external heat flux (50 kWm−2) indicates the shortest ignition time (25.9 s) and the highest HRR value (342.4 kWm−2). The effect of increase in EG composition is seen in the lowering of the peak HRR values which again shows the effects of the fire retardant inorganic EG. Char of EG gives off HRR values of about 17.6 kWm−2.

External heat flux (50 kWm−2) again leads to the shortest ignition time (23.2 s) and the highest HRR value (360.8 kWm−2) in Figure 5. Char of EG gives off average HRR values of about 27.2 kWm−2 (higher than in Figure 3). As reported by Schartel and Hull [19], char influences the amount of combustible material and works as a barrier between pyrolysis and flame zone.

The summary of sample composition, time-to-ignition, and external heat flux are displayed in Figure 5 and Table 1. Though not a surface plot, four ranges of time-to-ignition in relation to different external heat fluxes and EG composition are clearly displayed in Figure 6. As the external heat fluxes increase from 25 to 50 kWm−2, the time-to-ignition decreases with respect to increased EG content. The heat induced swelling of EG (intumescent) coupled with its good thermal conductance enhances the decomposition of the resin but at the same time reduces the heat release rate and curtail possible fire spread. There is no linear relationship between time-to-ignition at 50 kWm−2 heat exposure and the increasing EG composition in the polymer. See Table 1.

The decreasing time-to-ignition in increasing content of the EG findings, as in Table 1, is countered by the Fire Growth Rate Index (FIGRA). FIGRA is calculated by dividing the peak HRR by the time to peak HRR. The lower the FIGRA value is, the slower the flame spread and flame growth are assumed to be [24]. The term Fire Performance Index (FPI) is also used in the literature to describe fire spread [25]. Thus, the addition of EG decreases the HRR and also lowers flame spread and growth prospers of the epoxy-EG composites.

3.2. Heat Release Rate (HRR) Measurements with Time from Different Epoxy-EG Composition Samples at Same External Heat Flux
3.2.1. Irradiation of 25 kWm−2 on Different Epoxy-EG Composites

Figure 7 clearly shows that increasing EG composition decreases the peak HRR value but widens the peak width when samples were irradiated with 25 kWm−2.

For example, the epoxy-EG 5-wt.% has the least peak HRR value of around 200 kWm−2, peak width with a plateau over 375 s, before average background charring of 38 kWm−2. The residue charring increases somewhat with increasing EG content (see Figure 1), while flame inhibition (FIGRA values in Table 2), which plays an important role for the fire performance of the composites, decreases (from 4.0, neat, to 1.9 kWm−2s−1, 5 wt.% EG). The profiles of higher EG content shift to the left (not much, though) and indicate shorter time for their thermal decomposition. When the composite is irradiated, the intumescent behaviour of the EG somewhat eases the surface decomposition of the epoxy resin leading to shorter ignition time. Thereafter, the formation of char from the EG influences the amount of combustible material and works as a thermal barrier. Similar trends were observed in other external heat fluxes. See Figure 8, where the external heat flux was 50 kWm−2.

Furthermore, the total heat release per unit area (THRR) for the composites subjected to assigned external flux, which is equivalent to the area under a HRR versus time profile, shows some variations, due possibly to the fact that the samples thicknesses were not the same. The general trend, however, shows that THRR decreases with increasing EG addition. See Table 3.

3.2.2. Irradiation of 50 kWm−2 on Different Epoxy-EG Composites

The THRR value of the neat is the highest (134.2 MJm−2) as compared to the 5 wt.% EG composite of 118.6 MJm−2.

3.3. Determination of Critical Heat Flux

From (1), the time-to-ignition () is related to the square of the reciprocal of the external heat flux; the sample (assumed to be thermally thick) is subjected to.

Figure 9 shows clearly that the time-to-ignition decreases as external heat flux increases. For lack of experimental points at lower external heat fluxes, the curve in Figure 9 would have been closed to a vertical asymptote () that defines the critical heat flux for ignition to occur. The critical heat fluxes for the other composite samples were determined in a similar way. See Table 4.

Furthermore, from the slopes of plots of reciprocal of square root of time-to-ignition and external heat flux, according to (1), the thermal inertia for each of the samples was determined [16].

Other similar plots, as in Figure 10, were plotted for 1, 3, and 5-wt.% EG in the polymer. The results are found in Table 4.

All the values of the three parameters, namely, the critical heat flux, the ignition temperature, and the thermal inertia, seem to be decreasing with increasing content of EG in the composition. In the case of the critical heat flux, with the exception of the “odd” epoxy 1-wt.% EG sample (8.0 kW·m−2), the value of the neat epoxy was found to be 6.9 (±0.5) kW·m−2 decreasing to 6.7 (±0.2) kW·m−2 and further to 5.8 ± 0.4 kW·m−2 in the epoxy 3-wt.% EG and epoxy 5-wt.% EG, respectively. The overall decrease could be attributed to the decrease in the thermal inertia (and corresponding ignition temperature), from neat epoxy value of 7.0 kW2sm−4 K−2 to 4.7 kW2sm−4 K−2 for the epoxy 5-wt.% EG sample. Similar conclusions that critical heat flux decreases (from 435 to 370 kW·m−2) with increasing carbon fibre (from 56 to 59 vol%) in epoxy were reported by Dao et al. [16] in their study on thermal decomposition of epoxy resin/carbon fibre composites. In addition, the decrease in the ignition temperature from the neat to increasing EG composition stems from the increasing conductivity (see Table 4) and the intumescent property of the dispersed EG.

3.4. Determination of Heat of Gasification and Combustion

When each of the composite samples was subjected to a given irradiance, it vaporized at a certain rate. The steady-state mass loss per unit area () of the sample at a given net heat flux () defines the rate of vaporization, as in (6) [16]. The heat of gasification () is the energy required to produce the fuel volatiles per unit mass of the material.

The influence of expandable graphite in the epoxy-EG composite was further demonstrated from plots of external heat flux versus average specific mass loss rate (gs−1 m−2) from the Cone Calorimeter data, according to (6), and the results are shown in Figure 11.

Using the slopes in Figures 11 and 12, the calculated heat of gasification () and heat of combustion () for the samples were determined. The net flame heat flux, on the other hand, was calculated from -intercepts of the plots in Figures 11 and 12. The results are shown in Table 5.

The values of the heat of gasification obtained from Figure 11 were used to determine the heat of combustion from the slope of Figure 12.

The heat of gasification of neat epoxy (3.4 kJg−1) compares well with 2.4 kJg−1 obtained by Tewarson [26]. However, the values of the heat of combustion of epoxy are about one-half that found in literature [27].

The increasing content of EG (as explained earlier) in the resin assists the initial ignition of the epoxy volatiles and, thereafter, forms thermal barrier that seems to prevent further gasification of resin molecules and, thereby, increases the heat of gasification () required to vaporize the composites from their initial state (shorter time-to-ignition). There is the corresponding increase in the swelling and char formation (see Figure 1) associated with the EG which tend to decrease the heat of release rate (HRR). The char formed a protective layer on the surface of the epoxy that slowed down the heat and mass transfer between the burning polymer and the flame [28]. Thus, the higher EG content, the better flame retardancy and the higher heat of combustion () (with the exception of the 1-wt.% EG) as shown in Table 5.

The net flame heat flux () results calculated from the slopes of plots in Figures 11 and 12 are shown in Table 5. The epoxy 5-wt.% EG composite gives a value of 18.9 ± 3.2 kWm−2.

The choice of mean peak heat release rate data from the Cone Calorimeter readings for each sample seems to give an energy release rate that is more consistent with steady burning as opposed to peak heat release. The intercept on the mean peak heat release axis gives the following:For the epoxy 5 wt.% EG composite linear graph, in Figure 12, for example,If the corresponding values of and in Table 5 are substituted in the above equation, the net flame heat flux is obtained. Similar calculations, from Figure 11, give 16.7 kWm−2 as the net flame heat flux value for the same sample. The calculated values for the other samples are given in Table 5 and show a partial decrease with increaseing EG content in the epoxy.

These results show that EG fraction increase favours the transfer and the distribution of heat within the composite material. That is, there is decrease in inflammability and rather increase in the combustibility of the epoxy-EG composite sample.

3.5. Smoke Formation from External Heat Flux

Apart from heat release rate being the single most important parameter in characterizing material hazard in a fire, smoke density as a by-product of thermal decomposition of the polymer matrix is the other hazard that affects visibility and produces toxic gas that needs to be analysed [28].

The variation of average smoke densities, measured by the decrease in transmitted light intensity of a He-Ne laser beam located within the Cone Calorimeter fume extraction duct and expressed in terms of average specific extinction area (SEA) [9], is determined when the epoxy-EG composite samples were irradiated with different external heat fluxes. The results are shown in Figure 13.

The profiles in Figure 13 suggest two general results. Firstly, for each of the epoxy-EG composite, there is a linear relationship between smoke density and increasing external fluxes. Secondly, the increasing content of EG in the epoxy produces lesser smoke yield. See Table 6. This is most likely due to the higher char yield from higher content intumescent EG in the epoxy. Similar results were reported by Mouritz et al. [9].

3.6. Thermogravimetric Analyser (TGA) Results

Further analyses of the thermal decomposition of the epoxy-EG composites were done with the TGA to evaluate their thermal stability [21]. Each epoxy-EG composite (including the neat epoxy) sample was subjected to the same nonisothermal conditions (from 25–800°C), at heating rates of 5, 10, 15, and 20°C min−1.

3.7. Mass Loss versus Temperature over Different Heating Rates

A typical mass loss versus temperature profiles for neat epoxy and 1-wt.% EG-epoxy composite sample is seen in Figure 14.

Both thermographs in Figures 14(a) and 14(b) show that the decomposition temperatures increase with increasing heating rate. In Figure 14(b), for example, the sample under the slowest heating rate (5 Kmin−1) started decomposing at 326.8°C while that of the fastest heating rate (20 Kmin−1) occurs at 371.5°C. (These values complement the ignition temperature values shown in Table 1.) The slowest rate has more time, at a particular temperature, for heat transfer and absorption which lead to early decomposition.

As the samples begin to decompose and lose mass, they also vaporize [22]. A first-order Arrhenius kinetic was assumed and, as approximation, the temperatures for 60% decomposition of each heating rate profile (here, assumed to be linear with the decomposition rate ()) were determined from Figure 14. Data of decomposition rate with their corresponding temperature at 60% decomposition (mass fraction of 0.6) were compiled. The Arrhenius equation (8) was then used to determine the activation energy and the preexponential term of the melt-vapour thermal diffusion transition.

Similar graphs, as in Figures 14 and 15, were plotted for 3 and 5-wt.% EG epoxy composite samples and the summary of their activation energies as well as the preexponential factors is found in Table 7.

The activation energy, which is a measure of the reduction in molecular mobility within the resin structure [29], seems to decrease with increasing EG content. Thus, the mobility of the molecules (thermal diffusion) would be decreased with the addition of EG.

3.8. Mass Loss versus Temperature for Different Composition at the Same Heating Rate

Further analysis involving different EG compositions in the epoxy samples at the same heating rates was done. Typical example of one such thermogram, where the composites had the same heating rate of 5°C/min, is shown in Figure 16.

The neat expandable graphite (EG) thermograph shows a multiorder kinetic with the first decomposition occurring from 210°C (dilution of the EG yielding water and sulphur dioxide). The effect of EG in the epoxy-EG composite is further displayed from temperatures above 400°C. There is increase in the residue mass fraction (char) for increasing EG content which is in agreement with the Cone Calorimeter results, as shown in Figures 8 and 9. The decomposition rate of the composites also decreases with increasing EG content between 300°C and 425°C which further explains the activation energy values in Table 7.

3.9. Thermal Conductivity Analysis

The results of direct thermal conductivity measurement, at 24°C, using the Linseis THB100 Transient Hot Bridge thermal conductivity analyser show that the thermal conductivity of the epoxy-EG composite increases with increasing EG content. The results are shown in Figure 17.

Perhaps a well-homogenous dispersion of the EG within the epoxy resin matrix would have given a better linear graph.

The thermal degradation status of the samples is further demonstrated with Scanning Electron Microscope (SEM) micrographs.

3.10. SEM Analysis: Thermal Degradation

SEM micrographs of the EG/epoxy composite before and after the Cone Calorimeter fire testing measurements are shown in Figures 1821.

Figures 18 and 19 show a typical rough surface patches of an epoxy resin micrograph before and after Cone Calorimeter measurements, respectively.

From the fire testing observations, the decomposition of the epoxy resin (solid) results in the release of volatile fuels which cause the resin to change phase from solid to liquid to gaseous under the constant heat flux leaving thin char remains. The SEM image of the char remain is shown in Figure 19.

Figure 20 shows the neat expandable graphite, ES250 B5, in flake form of average size of about 150 μm, prior to mixing with the epoxy resin.

The residue of one of the composite samples after the Cone Calorimeter measurement is shown in Figure 21. The normal worm-like image indicating the intumescent behaviour of the EG is clearly seen.

4. Conclusions

Thermal measurements on decomposition processes of inorganic intumescent additives of EG in organic epoxy resin composite samples were done. Upon subjecting the composite samples to different heat fluxes (25, 30, 35, and 50-kWm−2), it was found that both the time-to-ignition and the peak heat release rate decrease from the lowest to the highest external heat flux, for a particular EG content composite.

Increasing the EG content in the epoxy composites, at a particular external heat flux, was found to lower the time-to-ignition, the critical heat flux, the ignition temperature, the thermal inertia, the smoke yield, and the peak heat release rate.

The activation energy of the decomposition of the epoxy-EG samples is found to decrease with increasing EG content. The measured activation energy of the neat epoxy was 159.1 kJmol−1 as compared to 145.9 kJmol−1 of 3-wt.% EG-epoxy composite.

Competing Interests

The authors declare that they have no competing interests.


The authors wish to thank the Research and Innovation Committee, TUT, for the generous support rendered in the research into thermal measurements on polymer composites.