Influence of ceramic particle features on the thermal behavior of PPO-matrix composites

1 Introduction

In order to overcome drawbacks of the different materials to satisfy the demanding requirements of engineering applications, several types of composites have been developed, admixing different components and improving in this way several properties with respect to those of the individual components [1–6]. In particular, the dispersion of inorganic particles in an organic matrix leads to the production of composite materials characterized by increased properties in terms of thermal (e.g., thermal stability and coefficient of thermal expansion) and flame resistance, chemical and moisture resistance, barrier properties (e.g., permeability to gases, water and hydrocarbons), electrical and mechanical properties (e.g., strength, modulus and dimensional stability) and charring effect, maintaining the easy processability and the low density of the polymers [1, 7–11]. Generally, different approaches have been adopted to prepare polymer composites, as, for instance, solution blending, melt blending, and in situ polymerization [12–14]. Among them, melt blending offers an attractive route on a technological and economic point of view because it avoids the use of solvents and is compatible to industrial extrusion and blending processes. Moreover, melt blending is suitable for all thermoplastic polymers because the inorganic particles used as reinforcements are directly dispersible into the molten matrix.

In the past years, it has been demonstrated that the improvement of physical, chemical, mechanical, thermal and/or electric properties can be obtained without significantly affecting the low density of the polymeric matrix through the dispersion of nanometric reinforcements, which are required in low amounts [11, 12]. In particular, ceramic nanoparticles having three different shapes have been extensively investigated: platelets having only one dimension at the nanometer scale, elongated particles with two nanometric dimensions and equiaxial nanoparticles [11, 12]. Among them, platelets were demonstrated to be highly effective in increasing the physical, mechanical and thermal properties of the composites, thanks to their intercalation properties [11, 15, 16]. Notwithstanding this, elongated and equiaxial fillers have also been studied to improve the thermal stability and conductivity of the matrix [17–20]. Considering the extensive use of polymers in everyday life and the subsequent importance of their fire resistance, the addition of inorganic particles in organic matrices has also been investigated in view of granting an effective flame-retardant behavior [21]. In fact, ceramic reinforcements can inhibit or even stop polymer combustion, limiting the oxygen diffusion and promoting the formation of a carbonaceous surface layer called char [11, 15, 21].

In this paper, polymeric composites were prepared by adding three inorganic fillers having different shape and compositions into a thermoplastic resin, a commercial poly(phenylene oxide)/polystyrene (PPO/PS) blend. The aim of this work is to compare the effect of the different reinforcements on the thermal and combustion behavior of the matrix in order to select the best performing materials as thermal protection systems (TPSs) for aerospace applications. The employed polymeric matrix presents aromatic ring structures that favor char formation. This aspect can be exploited in an ablative TPS in order to dissipate the high heat flux, protecting aerospace vehicles during reentry.

2 Experimental

Three commercial powders were investigated as fillers, made of particles showing a platelet-like, a needle-like or an equiaxial morphology. Cloisite^® 15A (supplied by Southern Clay Products) is made of platelets of overexchanged organomontmorillonite [22]. Sepiolite (provided by Tolsa) is a hydrated magnesium silicate having a layered structure similar to that of montmorillonite [23] and an a cicular morphology [24]. The α-alumina TM-DAR Taimicron (produced by Taimei Chemicals Co.) consists of equiaxial particles with a mean diameter of 150 nm, as claimed by the supplier. The matrix is PPO/PS blend (Noryl^® 914 produced by GE Plastics) with a melting temperature of 290°C.

The ceramic powders were characterized by means of laser granulometry (Fritsch model Analysette 22 Compact, Idar-Oberstein, Germany) and field emission scanning electron microscopy (FESEM) (Hitachi S4000, Tokyo, Japan).

To prepare the composite samples, PPO/PS was first dried in a Piovan dryer for 8 h at 120°C, and then it was mixed with each inorganic filler in a Brabender internal mixer W48E at 290°C for 90 s with a rotor speed of 44 rpm. Several compositions were prepared by adding 2.5, 5.0 and 7.5 wt% of any inorganic phase. As a reference, pure PPO/PS was also processed as previously described. The related samples are labeled as PPO, whereas for the composite materials acronyms are used based on their composition, reporting the percentage and the nature of the filler. As an example, a composite containing 2.5 wt% Cloisite is hereafter referred to as PPO2.5C.

The composites and the pristine polymer microstructures were characterized by scanning electron microscopy (SEM) (Hitachi S2300, Tokyo, Japan). Thermogravimetric analyses (TGA) (Mettler-Toledo TGA analyzer, Novate Milanese, Italy) were carried out in an inert atmosphere (argon, flux of 50 ml/min), heating up to 1500°C with a heating rate of 20°C/min to evaluate the overall thermal degradation of the materials. Because all the investigated materials underwent a relevant mass loss below 550°C, the enthalpy of the degradation reactions of the pristine polymer and the composite materials containing 5.0 wt% of each filler was evaluated by differential thermal analyses (DTA) (Netzsch STA 409, Selb, Germany) performed up to this temperature with a heating rate of 10°C/min.

Finally, cone calorimeter tests (Dark Star Research Ltd., Park Lane, UK) were carried out on samples (5.0 cm×5.0 cm×0.3 cm), with a heat flux of 35.00 kW/m² and an air flow rate of 24.00 l/s.

3 Results and discussion

The morphologies of the three fillers are compared in Figure 1.

Figure 1

FESEM micrographs of the three fillers: (A) Cloisite^® 15A; (B) Sepiolite, with a high-magnification image of the needles; (C) α-alumina.

Cloisite^® 15A, labeled hereafter as C, consisted in agglomerates of platelet-like nanoparticles (shown in Figure 1A). However, in the case of Sepiolite, here abbreviated as S, nanometric needles were observed within the platelets (Figure 1B). Finally, α-alumina powder, in the following named A, presented large agglomerates of equiaxial particles with a mean diameter of 150 nm (Figure 1C).

The agglomeration of the ceramic powders was confirmed by granulometric analyses and summarized in Figure 2. A mean diameter of about 9.3 and 11.1 μm was detected for the agglomerates of C and S, respectively, whereas the aggregates of A particles were of about 29.5 μm in average.

Figure 2

Particle size distributions of the three fillers.

The mixing step yielded to a homogeneous dispersion of the ceramic particles in the organic matrix, as shown in Figure 3, in which the microstructures of the composite materials are compared to the pure polymer. SEM observations showed that the larger agglomerates of the fillers, revealed by granulometric analyses of the as-received powders, were not visible in composite microstructures. In particular, for all the fillers equiaxial aggregates with diameters of a few micrometers were observed. This demonstrated that the process employed for the production of the composite materials is not able to fully disaggregate the filler powders, probably because of the matrix properties or the working conditions [25].

Figure 3

SEM micrographs of (A) PPO, (B) PPO2.5C, (C) PPO5.0C, (D) PPO7.5C, (E) PPO5.0S, (F) PPO5.0A.

The increase in the filler amount did not affect the homogeneity of the composite microstructure, as highlighted in Figure 3B–D using PPOC materials as an example. The same microstructural features were observed in the S and A composites, even if in the latter case the inorganic particles were less easily observable due to the lower volume percentages of the inorganic phase in the PPOA composites because the density of the alumina (3.96 g/cm³) is higher than those of the other powders (1.66 and 2.15 g/cm³ for C and S, respectively).

Figure 4 A reports TGA curves and their derivatives (DTGA) of PPO. The polymer blend presented a single-step decomposition, with an important mass loss between 360°C and 500°C, reaching the maximum rate at 460°C. Even if the same trend in the TGA curve was recorded for all the investigated composites, only for a few compositions the effect of the inorganic particles was clearly highlighted in the related DTGA curves (Figure 4B).

Figure 4

(A) TGA and DTGA curves of PPO. (B) DTGA trends of the investigated materials.

In particular, the DTGA curves of the PPO7.5S and PPO5.0C (Figure 4B) as well as of PPO7.5C (not shown in this figure) materials presented a shoulder, which is reasonably imputable to a different degradation mechanism [26]. From the DTGA curves, the values of T_onset (the onset decomposition temperature), T_max (the temperature at which the maximum rate of the decomposition is achieved) and T_end (the end decomposition temperature) for each composite material were determined and compared with those of the pristine matrix in Figure 5. The presence of the inorganic particles influenced the temperature range in which the degradation processes of the polymer occurred. In fact, even if for each filler and amount the T_onset was not modified (and, for this reason, not shown in Figure 5), an increase in T_end was observed. In particular, a more gradual rise was recorded for the PPOC materials as the filler amount increased. However, in the case of S and A composites, T_end changed with the addition of the lower filler amount, and it remained constant as the inorganic phase increased.

Figure 5

T_max and T_end values of the investigated materials as a function of the filler amounts.

Moreover, the addition of the fillers also affected the temperature at which the maximum rate of the decomposition was achieved. In fact, a shift of T_max at higher values was recorded. This is very limited in the case of A samples and more evident in the other composites, especially in the PPOC materials.

As shown in Figure 6, the DTA analyses revealed that the three fillers modified the thermal behavior of PPO in a different way without affecting its melting (T_m) and decomposition (T_d) temperatures recorded at about 255°C and 440°C in all the prepared materials. The DTA curves of PPO and the PPOA composites revealed a very similar trend, which suggested a negligible effect of the ceramic powder on the thermal behavior of the polymeric matrix.

Figure 6

DTA curves of the investigated materials.

On the contrary, a very different response was observed in the materials containing S or C particles, leading to a displacement of the DTA curves. The addition of the S filler implied a more endothermic signal during all the analyses, whereas in the case of the PPOC composites this effect was observed only after the peak related to the PPO decomposition.

Finally, the flame behavior of the composite materials was evaluated in order to investigate a possible effect of the ceramic particles on the char formation. In fact, PPO presents an aromatic structure that implies a high tendency to form a char layer [27], as confirmed by the shape of its heat release rate (HRR) curve, compared in Figure 7 to those of the studied composites.

Figure 7

Comparison between the HRR curves of PPO and of the composites containing (A) 2.5 wt% of filler, (B) 5.0 wt% of filler, (C) 7.5 wt% of filler. (D) Digital image of the surface of PPO7.5S after the cone calorimeter test.

During the exposure to a constant heat flux, PPO yielded a stable char layer, with a decrease of the HRR, as usually observed for the thermally thick-charring samples [28]. No relevant changes in HRR curves were observed at the lower amount of the ceramic fillers. In Figure 7A it is evident that the A particles did not modify the flame behavior of the polymeric matrix, whereas the other reinforcements implied a slight decrease of HRR peaks, imputable to the formation of a thin char layer [28]. Moreover, the presence of S and C also induced a higher combustion time, which represents an increase in the duration of the degradation processes. When a filler amount of 5.0 wt% was employed in the composite preparation (Figure 7B), the HRR curves of the PPO5.0A and PPO5.0C materials did not change with respect to the composites containing the lower content, even if these materials presented a slightly higher time to ignition (TTI), that is, the response time of the material to the heat irradiation. On the contrary, 5.0 wt% S implied an increase in the charring ability of the matrix. In fact, a decrease of the peak of the HRR curve was recorded, suggesting the development of a thicker char layer, and the combustion phenomena were prolonged. Finally, at the higher amounts (Figure 7C), the A particles were able to influence the flame behavior of PPO, with an increase in the TTI value and in the combustion time, promoting the development of a thin char layer as deducible from the decrease in HRR peak. The effect of A observed only at the highest content was probably imputable to the lower volume percentages of this inorganic phase with respect to C and S. Moreover, the higher amount of the C and S fillers showed a more evident effect. Higher combustion times were recorded, particularly in the case of the PPO7.5S material. After the test a visible carbonaceous coating was present on the surface of the samples, as shown in Figure 7D.

4 Conclusions

Clay (C), sepiolite (S) and α-alumina (A) powders were used as fillers in the preparation of the PPO-matrix composites through a melt blending process. Even if nanometric fillers were selected, the composite materials contained homogeneously distributed, micrometer-sized agglomerates of inorganic particles in different amounts in the range 2.5–7.5 wt%.

The C materials presented a higher thermal stability and a more endothermic decomposition at high temperatures. This ceramic powder, particularly at higher amounts, was able to promote the development of a protective char layer during combustion.

S improved the temperature range and the endothermic nature of the decomposition reactions. In addition, an increased char ability was demonstrated and a visible char layer formed after cone calorimeter tests.

Finally, the A filler positively affected the PPO decomposition and combustion phenomena, but with more limited changes in its behavior with respect to the other reinforcements.

Notwithstanding the partial dispersion of the as-received powders reached through the melt blending, the selected fillers were able to modify the thermal and combustion behavior of PPO. Further investigations will be carried out in order to evaluate the behavior of the corresponding nanocomposites, optimizing the preparation method in view of a good dispersion of the ceramic fillers.