PBS (trade name GS PLA, Japan) was bought from Mitsubishi Chemical Corp. (Toyota, Japan). The melt flow index was 4.5 g/10 min at 190°C under 2.16 kg of weight. Aluminum diethylphosphinate (AlPi) was provided by Qingdao Fuslin Chemical Technology Co., Ltd. Maleic anhydride-grafted poly(butylene succinate) (PBS-g-MA) was synthesized via reactive melt-grafting process according to previous literature (Phua et al., 2013b). The ratio of PBS, maleic anhydride, and dicumyl peroxide (DCP) was 100/10/1.5 by weight, and the grafting degree of MA onto PBS (G _d) was 4.2 wt%.

PBS/AlPi composites were prepared by melt compounding in a HAAKE batch intensive mixer (HAAKE Rheomix 600, Karlsruhe, Germany) at 135°C with a rotor speed of 80 rpm; the mixing time was 6 min for each sample. The content of PBS-g-MA (as a compatibilizer) was kept as 15 wt% in all PBS composites, and the content of AlPi was changed from 5 to 25 wt%. For convenience, the obtained samples were designated as PBSxAlPi. Here, x denotes the weight percentage of AlPi in the PBS composites.

Thermogravimetric analysis (TGA) was performed on a TA STD Q600 thermal analyzer. The PBS samples with mass 8.0 ± 0.2 mg were heated from room temperature to 600°C at 10°C min⁻¹ under nitrogen and air atmosphere, respectively. The limited oxygen index (LOI) was tested on a JF-3 oxygen index meter (Jiangning, China) with sheet dimensions of 130 × 6.5 × 3.2 mm³, according to ISO4589-1984. The vertical burning testing was carried out according to the UL-94 (ANSI/ASTMD635-77) with sheet dimensions of 125 × 12.7 × 3.2 mm³. Cone calorimeter testing (icone, FTT, United Kingdom) was conducted according to ISO 5660-1. The sample dimension was 100 × 100 × 6 mm³; it was backed by aluminum foil and irradiated at a heat flux of 50 kW m⁻². The photographs of residual chars after cone calorimeter testing were collected by a digital camera. The dispersion of AlPi in PBS matrix was examined with the scanning electron microscope (XL30 FESEM FEG, FEI Co.). The samples were fractured in liquid nitrogen, and the fracture surfaces were coated with gold before SEM observation. Uniaxial tensile tests were performed at room temperature with an Instron 1,121 testing machine (Canton, MA). Specimens were compression-molded into sheets with 1 mm thickness, then cut into a dumbbell shape with gauge dimensions of 20 mm × 4 mm × 1 mm. The measurements were conducted at a crosshead speed of 20 mm min⁻¹. At least five runs for each sample were measured and averaged.

The influence of AlPi on the thermal stability of PBS under nitrogen was investigated by TGA, which reflects the thermal degradation behavior of materials with the increase of temperature (Wang et al., 2020b; Xu D. et al., 2021). As shown in Figure 1A, with the increase of AlPi content, the TGA curves slowly shifted to a low-temperature range. The temperatures corresponding to 5 and 10 wt% weight loss (T _5wt% and T _10wt%) are essential to evaluate the thermal decomposition of polymers on the onset stage. For PBS5AlPi, T _5wt% and T _10wt% decreased less than 1°C in comparison with that of neat PBS, indicating that the addition of AlPi with low content has a small impact on the thermal stability of PBS. However, the reduction in thermal stability gradually became larger with high content fillers (Table 1). For example, T _5wt% and T _10wt% of PBS25AlPi decreased to 15.8 and 14.9°C, respectively. Meanwhile, with the increase of the AlPi loading, Figure 1B exhibited a gradual decrease for T _max1, corresponding to the maximum weight loss rate of the polymer. Herein, the decrease on T _5wt%, T _10wt%, and T _max1 was ascribed to the thermal degradation of PBS during melt compounding because the increase in viscosity with high content of inorganic fillers could result in the sharp rise of temperature under high shear force, promoting the thermal degradation polymer (Fong et al., 2021). In addition, another characteristic peak was present at approximately 472°C (denoted as T _max2 in Table 1), which should be assigned to the thermal decomposition of AlPi.

The thermal stability of PBS/AlPi composites under air was also evaluated, which is more important because the melt processing of polymer materials is usually performed in an air atmosphere. As shown in Figure 2A and Table 2, T _5wt%, T _10wt%, and T _max1 exhibited a similar trend to the situation of nitrogen, but the values were relatively smaller owing to the oxidation degradation (Wen et al., 2011). Furthermore, as shown in Figure 2B, T _max2 appeared at approximately 427°C (45°C lower than that in nitrogen), and the intensity of peaks became much weaker. It is suggested that part of AlPi gradually degraded before T _max2, which is accompanied by the degradation of PBS chains. Based on these results, the content of AlPi could influence the thermal stability of PBS/AlPi composites, and the decrease in thermal stability became stronger under air atmosphere at higher AlPi loadings.

The effect of AlPi on the flame retardancy of PBS matrix was investigated by LOI and UL-94 tests. As listed in Table 3, the LOI values gradually increased with the addition of AlPi content, of which the highest value corresponding to PBS25AlPi could achieve 29.5. Moreover, only when the content of AlPi was 20wt%, the UL-94 rating reached V1. With further increasing the AlPi to 25wt%, the UL-94 rating could pass V0. Besides the flame retardant effect, AlPi displayed a positive effect on the inhibition of dripping. For all PBS/AlPi composites, there was no melt dripping, indicating that AlPi was an effective dropping inhibitor for PBS matrix.

Further, the flame retardancy of PBS/AlPi composites was investigated by cone calorimeter testing, which is useful to provide various important information about fire risk during combustion (Zanetti et al., 2002; Wang et al., 2020a; Xue et al., 2020b; Xu et al., 2020), such as time to ignition (TTI), heat and smoke release, and mass loss. First, heat release rates (HRR) and combustion time curves for PBS samples are shown in Figure 3A, and detailed parameters are listed in Table 3. With the increase of AlPi loading in the PBS system, the ignition time (t _ign) became shorter, implying that the PBS/AlPi composites were easier to be ignited than neat PBS. However, a gradual decrease for the peaks of HRR (PHRRs) was present. For instance, the PHRR for PBS25AlPi was reduced to 388 kW/m², which was reduced by 49.3% compared to that for neat PBS (765 kW m⁻²). Further, Figure 3B shows the curves of total heat release (THR). It is significant that the slope flattened out at the latter half of combustion time, and the final THR value became smaller to a certain extent. These results suggest that AlPi is an effective flame retardant for reducing PHRR and THR.

Figure 3C shows the normalized mass loss (ML) curves of PBS samples with combustion time. All PBS/AlPi composites exhibited similar curves, but the slope became smaller with more AlPi fillers, indicating that AlPi as flame retardant could delay the combustion of PBS. Meanwhile, the residual char gradually increased with the increase of AlPi content. Furthermore, the total smoke production (TSP) was evaluated. As shown in Figure 3D, the TSP increased with the addition of AlPi (detailed data are listed in Table 3). The increase in TSP should be related to the radical trapping effect in the gas phase. The decomposition of AlPi could produce P• and PO• radicals (Tang et al., 2020), which could quench H• and HO• radicals so that some derived products were generated as smoke particles.

Besides the trapping radicals in the gas phase, AlPi also contributed to the condensed phase through char formation. As shown in Figure 4A, neat PBS was typical no-char polymers (burned completely without residual char). With the increase of AlPi concentration, the char amount increased and the carbon layer became thicker. It is clear that the cracks gradually became smaller until they disappeared (Figures 4B–F). To further evaluate the microstructure, SEM observation was carried out for the char from PBS25AlPi. Interestingly, the char morphologies from the outer surface and inner surface were different. As shown in Figures 5A,B, the outer surface was covered by flocculent porous carbon, which may come from the deposition of P•/PO•-derived solid products. However, the inner surface was compact and cohesive with high supporting strength (Figures 5C,D). It was a whole bulk with some cavities, which may result from the release of flammable gases into the gas phase zone.

Based on the above analysis, a possible enhancement mechanism of AlPi in the PBS system was proposed in Figure 6. It is reported that AlPi could be thermally decomposed to oligomers of phosphinates, diethylphosphinic acid, and aluminum phosphate (Kaya and Hacaloglu, 2014; Vothi et al., 2020). On the one hand, oligomers of phosphinates and diethylphosphinic acid can further be decomposed to P• and PO• radicals in the gas phase, which could quench H• and HO• radicals in the combustion zone, and some derived solid products were generated as the char of outer layer. On the other hand, aluminum phosphate could construct continuous and compact char as the inner layer. The combined char layer can effectively reduce the heat and mass transfer rate and protect the underlying material from burning (Xu Y.-J. et al., 2021). As a result, the improvements on flame retardancy with LOI of 29.5%, V0 in UL-94, and 49.3% reduction on PHRR were presented. In brief, the enhanced flame retardancy of PBS was attributed to the gas–solid flame retardancy mechanism of AlPi.

The effect of AlPi on the mechanical properties of PBS was investigated by tensile testing. Figure 7 shows the stress-strain curves of neat PBS and PBS/AlPi composites. The detailed data for Young’s modulus, tensile strength, and elongation at break are listed in Table 4. With the increase of AlPi loading, Young’s modulus gradually increased due to the reinforcing effect of AlPi as rigid inorganic particles. The tensile strength firstly increased and then decreased with the increase of AlPi content. For PBS15AlPi, it exhibited the highest value of tensile strength of 35.8 MPa. However, the tensile strengths for all PBS composites were higher than that of neat PBS, which should be ascribed to good compatibility between PBS matrix and AlPi particles. In addition, the elongation at break showed a decreased trend with the addition of AlPi loading, indicating a decreased ductility of the PBS matrix. Similarly, the impact strengths of PBS composites also decreased gradually with the increase of AlPi content (Table 4).

In polymer/filler composites, the dispersion of fillers and interfacial interaction between two components are the most important factors to determine the final mechanical properties (Sun et al., 2017; Wen et al., 2020b; Xue et al., 2020a; Yang et al., 2019). As shown in Figures 8A–F, the fracture surface morphologies of PBS samples were investigated by SEM. It was apparent that AlPi particles were uniformly distributed in the PBS matrix, and no big aggregates were detected. Further, AlPi particles were firmly adhered to the PBS matrix, indicating their strong interfacial interaction (Wen et al., 2020a; Wen et al., 2012). Even for PBS25AlPi with the highest AlPi content, most AlPi particles still exhibited good dispersion without big aggregates, and no debonding cavities were present (blue circles in Figure 8F). In our PBS/AlPi system, 15 wt% PBS-g-MA was added as a compatibilizer, which was helpful to improve the dispersion of fillers and the matrix-particles interaction (Phua et al., 2013a; Chen et al., 2015). As a result, our PBS/AlPi composites exhibited good mechanical properties with high tensile strength and Young’s modulus.