The structure of hexagonal boron nitride (h-BN) is similar to that of graphite before functionalization and exfoliation. For applications in polymer nanocomposites, chemical exfoliation is a more economically attractive route to few-layer h-BN nanosheets. A thermal oxidation process of h-BN powder could achieve large scale exfoliation and hydroxylated functionalization, as described in prior literature. In this report, hydroxylated h-BN (BNO) was prepared by heating h-BN under air, and then covalently incorporated into epoxy resin modified with (3-isocyanatopropyl)triethoxysilane to prepare epoxy resin (EP) nanocomposites by sol–gel process. The structure and morphology of BNO were well characterized. BNO was dispersed in the EP matrix with the form of mainly exfoliated and intercalated structures, and formed strong interfacial interaction with the matrix. Thermogravimetric analysis results revealed that BNO significantly improved thermal stability and thermal oxidative resistance of EP nanocomposites at high temperature. The char yield and the temperature at 50 wt% mass loss were increased and the maximum mass loss rate was remarkably reduced. Moreover, the addition of 3 wt% BNO led to extremely high T_g of EP nanocomposite, 42.7 °C higher than that of pure EP, due to improved crosslinking density and confinement effect of BNO sheets on the mobility of polymer networks. Cone calorimeter test results indicated that fire safety properties of EP nanocomposites were also enhanced by the addition of BNO, such as 53.1% reduction in peak heat release rate and 32.6% decrease in total heat release, and decreased release of smoke and toxic gases. The mechanism for enhanced fire retardancy is that thermally stable condensed barrier consisting of h-BN sheets and silicon dioxide for heat and mass transfer protects the matrix from further combustion.

Functionalized reduced graphene oxide (FRGO) wrapped with a phosphorus and nitrogen-containing flame retardant (FR) was successfully prepared via a simple one-pot method and well characterized. Subsequently, FRGO was covalently incorporated into epoxy resin (EP) to prepare flame retardant nanocomposites. The FRGO was well dispersed in the matrix and formed strong interfacial adhesion. Thermogravimetric analysis results revealed that the presence of RGO, FR or FRGO in an EP matrix led to a slight thermal destabilization effect under air and nitrogen, which increased the char yield at 700 °C and reduced the maximum mass loss rate. Furthermore, the glass transition temperature of the FRGO/EP nanocomposite with an FRGO loading of 4 wt% (FRGO/EP4) was remarkably increased by 29.6 °C, probably due to the improved crosslinking density and confinement effect of graphene sheets on the mobility of polymer networks. The evaluation of combustion behavior demonstrated that a 43.0% reduction in the peak heat release rate (PHRR) for the FRGO/EP nanocomposite containing 2 wt% FRGO and a 30.2% reduction in the total heat release (THR) for FRGO/EP4 over pure EP were achieved by the addition of FRGO. These notable reductions in fire hazards were mainly due to the synergistic effect of FRGO and the flame retardant: the wrapped flame retardant accelerated the degradation of the EP matrix, promoting the formation of additional char residues; the flame retardant improved the thermal oxidative resistance of the graphene; a high-thermal-stability char layer, consisting of graphene sheets, retarded the permeation of heat and the escape of volatile degradation products.

A novel strategy based on functionalized graphene oxide (FGO)/phosphoramide oligomer flame retardant was developed to overcome the challenges of the dispersion of graphene sheets in polymer matrix and the ease of the burn-out of graphene under air atmosphere. Graphene oxide (GO) was modified by 4,4-diaminodiphenyl methane (DDM) and then in situ incorporated into phosphoramide oligomer, resulting in a nanocomposite flame retardant (FRs-FGO) containing exfoliated graphene. Subsequently, the flame retardant (FRs-FGO) was incorporated into polypropylene (PP) and simultaneously compatilized with PP-grafted maleic anhydride. TEM results showed that the FGO was dispersed more uniformly in PP than the bare GO because of the strong interfacial interaction and previous exfoliation of FGO in FRs before blending. The thermal properties investigated by thermogravimetric analysis (TGA) indicated that the addition of FRs-FGO into PP resulted in a significant improvement of thermal stability at elevated temperature with higher char yields. Moreover, the crystallization and fire safety properties of PP composites were also improved by the incorporation of FRs-FGO, including increased crystallization temperature (11.4 °C increase), reduced peak heat release rate (66.9% reduction) and decreased total heat release (24.4% decrease), and decreased fire growth rate index (73.0% decrease). The cone results indicated the simple blending of GO with FRs and exhibited less improvement in fire safety properties than FRs-FGO, which resulted from the improved dispersion and thermal stability of FGO sheets. The flame retardant mechanism was because of the shielding effect of FGO and char layers, which could reduce the release of combustible gases and inhibit the mass and heat transfer between the gas phase and condensed phase.

A variety of environmental analysis applications have been advanced by the use of satellite remote sensing. Smoke detection based on satellite imagery is imperative for wildfire detection and monitoring. However, the commonly used smoke detection methods mainly focus on smoke discrimination from a few specific classes, which reduces their applicability in different regions of various classes. To this end, in this paper, we present a new large-scale satellite imagery smoke detection benchmark based on Moderate Resolution Imaging Spectroradiometer (MODIS) data, namely USTC_SmokeRS, consisting of 6225 satellite images from six classes (i.e., cloud, dust, haze, land, seaside, and smoke) and covering various areas/regions over the world. To build a baseline for smoke detection in satellite imagery, we evaluate several state-of-the-art deep learning-based image classification models. Moreover, we propose a new convolution neural network (CNN) model, SmokeNet, which incorporates spatial and channel-wise attention in CNN to enhance feature representation for scene classification. The experimental results of our method using different proportions (16%, 32%, 48%, and 64%) of training images reveal that our model outperforms other approaches with higher accuracy and Kappa coefficient. Specifically, the proposed SmokeNet model trained with 64% training images achieves the best accuracy of 92.75% and Kappa coefficient of 0.9130. The model trained with 16% training images can also improve the classification accuracy and Kappa coefficient by at least 4.99% and 0.06, respectively, over the state-of-the-art models.