3.1 Morphology and structure of filler networks and their composites
The structure of the filler network was of vital importance for the thermal transport of the composite. Figure 2a-c display the microstructure of the cross-section of the compound filler (3D-BTA) networks. As can be seen, the filler network was a kind of 3D connected porous structure. With the increasing of the filler content, the size and number of pores in the network reduced. It has to be noted that this continuous filler network could serve as an effective way for the heat dissipation of composites [24–26].
When this sample of porous filler structure was infiltrated by epoxy resin, the epoxy composites with 3D filler network were obtained as shown in Fig. 2d-f. The filler network was maintained as an integrated 3D structure, and this could contribute to the enhancement of the thermal conductivity of epoxy composites.
3.2 Thermal conductivity of the BTA/EP composites
Figure 3a exhibited the thermal conductivity of the epoxy composites with various filler loading and distribution. The thermal conductivity of the composites increased with the increase of filler loading as expected. With the introduction of the filler network 3D-BTA, the thermal conductivity of the resin was remarkably enhanced, which was significantly higher than that of the composites (BTA/EP) with randomly distributed fillers. The thermal conductivity of 3D-BTA3/EP reached 3.08 Wm− 1K− 1, approximately 15 times higher than that of pure epoxy resin, and 3 times higher than that of BTA/EP composites with randomly distributed fillers.
The thermal transfer of the different composites during heating and cooling processes were also monitored by infrared thermography. Figure 3b shows that the thermal response speed of the composites with a 3D filler network structure during heating was also faster than that of the composites with randomly distributed filler. Figure 3c-d show the temperature change of the composites during heating and cooling. It can be found that the temperature of composite with integrated filler networks (3D-BTA3/EP) can rapidly rise from 29.8°C to 78.3°C within 150 s, which was much more efficient than that of samples with randomly distributed fillers (R-BTA3/EP), further reflecting the superiority of structural design. The thermal response of R-BTA3/EP sample was similar with that of 3D-BTA1/EP sample.
3.3 Flame retardancy of the BTA/EP composites
EP, R-BTA/EP and 3D-BTA/EP samples were ignited by alcohol lamp to observe the ignition time and combustion state of samples. It can be found from Fig. 4a that pure EP burned within 20 s and continued to burn intensively even without ignition. In contrast, the ignition time of R-BTA3/EP was extended to 55 s (Fig. 4d), which was attributed to the fact that the filler combination of BN, talc and APP played a positive role in the flame retardancy of the composites, but self-extinguishing property could not achieve when the flame left. Interestingly, the flame retardancy of R-BTA3/EP sample was also similar with that of 3D-BTA1/EP sample, just like thermal response shown in Fig. 3. Good heat dissipation has been found to be essential in retarding the combustion of composites [27]. When certain part of the composites was overheated, that part of heat can be dissipated in time to avoid excessive accumulation of local heat, leading to the retardation of combustion. Meanwhile, the ignition time of 3D-BTA2/EP and 3D-BTA3/EP was further extended to 64s and 75 s, respectively, and self-extinguishing occurred after leaving the ignition source as shown in Fig. 4f-g, which could be related to the thermal inertia (Qi) of the material.
When a material has a higher thermal inertia (Qi), it is harder to ignite and took more time to reach the ignition temperature as less heat accumulation on the material surface during the ignition process [30]. Qi of a material is positively related to the thermal conductivity (K), density (ρc) and specific heat capacity (Cp). As can be seen from Table 3, Qi increased linearly with the filler content of the composite. Moreover, the construction of filler network has played a key role in improving the thermal conductivity of the composites, so the Qi increased significantly. According to Tian's study [29], the 3D filler network can stabilize the morphology of the char layer, inhibiting the spread of fire sources [31], thus enabling the self-extinguish property of 3D-BTA/EP composites with certain filler loadings.
Table 3
Heat-conductive parameters of EP, R-BTA/EP and 3D-BTA/EP
Samples | ρc (g·cm− 3) | Cp (J·g− 1·K− 1) | K (W·m− 1·K− 1) | Qi (J2·cm− 4·K− 2·s− 1) |
Epoxy | 0.98 | 0.55 | 0.20 | 0.001 |
R-BTA1/EP | 1.31 | 1.10 | 0.46 | 0.006 |
3D-BTA1/EP | 1.35 | 0.76 | 0.53 | 0.006 |
R-BTA2/EP | 1.45 | 0.84 | 0.58 | 0.007 |
3D-BTA2/EP | 1.45 | 0.78 | 3.04 | 0.034 |
R-BTA3/EP | 1.53 | 0.71 | 0.90 | 0.010 |
3D-BTA3/EP | 1.57 | 0.79 | 3.08 | 0.038 |
The limited oxygen index (LOI) of samples was also measured and the results are presented in Table 4. Significant increase in LOI was obtained after adding BN, talc and APP. Generally, under the same filler content, the LOI of 3D-BTA/EP was higher than that of R-BTA/EP, and the best one (3D-BTA/EP) can reached up to 37.8%.
Table 4
LOI results of different samples
Samples | LOI (%) |
Epoxy | 20.0 ± 0.1 |
R-BTA1/EP | 25.5 ± 0.1 |
R-BTA2/EP | 25.8 ± 0.3 |
R-BTA3/EP | 26.6 ± 0.2 |
3D-BTA1/EP | 28.0 ± 0.0 |
3D-BTA2/EP | 37.2 ± 0.2 |
3D-BTA3/EP | 37.8 ± 0.1 |
Figure 5 presents a summary regarding the thermal conductivity and LOI of the composites which were both reported in different systems [28, 32–40]. It can be seen that BTA/EP composites possessed an advantage of thermal conductivity and flame resistance over other composites. Constructing BTA-filler network in composites was a more effective way to syngerstically enhance the thermal conductivity and flame resistance in composites.
3.4 Thermal stability of the BTA/EP composites
The thermal stability and thermal degradation properties of the various composites were investigated by TG analysis. Figure 6 has shown the TG curves of different composites, and the corresponding data can be found in Table 5. Among them, T5% has the decomposition temperature of 5% mass loss, and Tmax has the temperature corresponding to the maximum weight loss. The mass reduction of talc was around 17.4% by 750°C, which could be attributed to the generation of water, vulgar stone, and amorphous silica [41]. APP experienced a two-step thermodegradation, which corresponded to the emission of ammonia and hydrovapor and the decomposition of the P-O compound [42]. It was worth noting that both 3D-BTA3/EP and R-BTA3/EP composites underwent decomposition of APP and EP [43], at which the temperatures corresponding to 3D-BTA3/EP and R-BTA3/EP (Tmax) were 378°C and 351°C, respectively. Further, the three-dimensional filler structure provided structural support for char formation at high temperature, leading to a more stable char layer, and inhibited the decomposition of composites [33], which was consistent with the results of the combustion experiment.
Table 5
TG data of epoxy and its composites in nitrogen
Samples | T5% (°C) | Tmax (°C) | Residue at 750°C |
EP | 385 | 419 | 4.01 |
BN | - | - | 99.92 |
Talc | 530 | - | 82.64 |
APP | 334 | - | 20.46 |
3D-BTA3/EP | 355 | 378 | 52.81 |
R-BTA3/EP | 336 | 351 | 47.64 |
3.5 Dimensional stability and strength of ablated residue from 3D-BTA/EP composite
Aiming to assess the stability of residues, the 3D-BTA3/EP sample was ablated in a muffle furnace at 600, 700, 800, 900 and 1000°C for 30 min. It can be found from Fig. 7a that the sample can maintain its original form without cracks under different ablation temperatures. The colour of the residue gradually turned white as the temperature increased, due to the oxidation of the carbon residue at high temperatures [44].
Figure 7b-c showed the volume shrinkage and compress strength of 3D-BTA3/EP composite at different temperatures. The degree of volume shrinkage decreased with increasing of ablation temperature, while the compress strength of the residues increased, which was attributed to the ceramifiable reaction [45]. The ceramic skeleton was formed after ablation, which provided mechanical properties for the char layer, and further enhancing the flame resistance of the composite.
3.6 XRD of ablated residue from 3D-BTA/EP composites
As the XRD patterns of BN, talc and 3D-BTA3/EP burned under various temperatures in Fig. 8 were depicted. The typical diffraction peaks of BN and talc can be noticed in the XRD pattern. It can be found that when the ablation temperature increased to 900°C, a new diffraction peak occurred at 29.7°, corresponding to magnesium pyrophosphate (Mg2P2O7) [46]. Diffraction peaks corresponding to magnesium phosphate (Mg3(PO4)2) appeared at 20.18, 21.40 and 24.08° (PDF No: 33–876). A new diffraction peak also appeared at 31.93°, indicating the formation of cristobalite phase [47]. When the ablation temperature increased to 1000°C, the characteristic peaks of each crystal phase gradually intensified. The characteristic peak of BN was found to remain unchanged at this temperature, which indicated that talc and APP in the system could have ceramifiable reaction when the ablation temperature was at 900°C.
3.7 FTIR of ablated residue from 3D-BTA/EP composites
To further verify the ceramifiable reaction of the composite, FTIR tests were carried out on the ablation residues at different temperatures. The results were displayed in Fig. 9. with regard to talc, the peak at 1020 cm− 1 and 536 cm− 1 can be designed as Si-O and Si-O-Mg stretching vibrations, respectively [41]. In the spectrum of BN, the peak at 803 cm− 1 corresponded to the external extension vibration of B-N-B [48]. When the ablation temperature reached 900°C, three new characteristic peaks appeared at 1045 cm− 1, 986 cm− 1, and 620 cm− 1. Based on Gong’s [46, 49] work, those peaks at 1045 cm− 1, 986 cm− 1, and 896 cm− 1 can be ascribed to phosphate tetrahedra and asymmetrical stretching patterns of P-O-P bonds, indicating the presence of phosphates and pyrophosphates, respectively. In addition, the peaks measured at 1087 cm− 1 and 620 cm− 1 can be assigned to quartz and rhodochrosite [50]. Thus, these results pointed to the fact that the structure of talc was actually broken down and reacted with the products of APP decomposition at high temperature, indicating that ceramization have occurred under such circumstances. The FTIR results were in agreement with the XRD analysis.
3.8 Potential mechanisms of thermal conductivity and flame retardancy
A possible mechanism for syngerstically enhancing the thermal conductivity and flame resistance in composites with integrated filler networks was suggested in Fig. 10. When a certain amount of heat was given to the surface of the material, it was possible to diffuse the heat on the plane and to the interior in time, avoiding excessive accumulation of local heat in the sample, leading to the retardation of combustion. During combustion, the composite started to disintegrate and release non-combustible gases (NH3, H2O and CO2) at a certain temperature. The decomposition products of APP, mostly polyphosphates and their derivatives, can be covered on the char of epoxy decomposition, forming a phosphoric acid-char layer structure. A ceramifiable reaction took place as the char layer formed and oxidized with further rise in temperature. In the end, a ceramics-like foam residue was formed containing magnesium phosphate, magnesium pyrophosphate, cristobalite and quartz. This strategy has led to epoxy composites with excellent thermal conductivity at room temperature as well as significantly enhanced flame resistance and fire proofing properties.