Composites Part B 177 (2019) 107402 Contents lists available at ScienceDirect Composites Part B journal homepage: www.elsevier.com/locate/compositesb Expanded graphite assistant construction of gradient-structured char layer in PBS/Mg(OH)2 composites for improving flame retardancy, thermal stability and mechanical properties Hao Chen a, Ting Wang a, Yanliang Wen a, Xin Wen a, b, *, Doudou Gao b, Ronghua Yu b, Xuecheng Chen a, b, Ewa Mijowska a, Tao Tang b, ** a Nanomaterials Physicochemistry Department, Faculty of Chemical Technology and Engineering, West Pomeranian University of Technology Szczecin, Piast� ow Ave. 42, 71-065, Szczecin, Poland b State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, 130022, China A R T I C L E I N F O A B S T R A C T Keywords: Expanded graphite Magnesium hydroxide Gradient structure Flame retardancy Mechanical properties It is a great challenge to improve the flame retardancy of polymer materials by adding eco-friendly magnesium hydroxide (Mg(OH)2) but simultaneously maintain satisfied mechanical performances. In this study, expanded graphite (EG) as a synergist was introduced into poly(butylene succinate) (PBS)/Mg(OH)2 system to investigate the effect of EG on their flame retardancy, thermal stability and mechanical properties. The results showed that only addition of 5 wt% EG into PBS/20 wt%Mg(OH)2 resulted in excellent flame retardancy, including that the LOI was 29.4%, the UL-94 rating reached to V0, and the PHRR decreased by 73%. These flame retarded parameters were comparable or even better than that of PBS/40 wt%Mg(OH)2 composites. More importantly, PBS/ 20 Mg(OH)2/5 EG presented much better thermal stability and mechanical properties than PBS/40 wt%Mg(OH)2, indicating a balanced improvement on comprehensive performances of PBS. Based on the comparison with PBS/ 20 Mg(OH)2/5Graphene(Gr) system and the structure analysis for residual chars, the mechanism for improved flame retardancy was attributed to the formation of gradient-structured char layer due to the helpful thermal expansion process of EG together with the decomposition of Mg(OH)2 during combustion, resulting in better barrier action to heat, oxygen and flammable gases via “labyrinth effect”. As far as we know, it is the first report that gradient-structured char layer was discussed in flame retarded polymer systems. Hence, this work provides not only a high-efficient synergist to improve the flame retardant efficiency of Mg(OH)2, but also useful guidelines to design polymer composites with balanced comprehensive performances. 1. Introduction In recent years, with the increasing environmental concerns and stringent legislation for plastic additives, to explore eco-friendly flame retardants has drawn widespread attention [1–6]. Magnesium hydroxide (Mg(OH)2) is a good candidate owing to its advantages, such as non-toxicity, low cost and relatively high decomposition temperature [7,8]. It provides effective flame retarded effect by releasing water and producing a highly heat-resistant magnesia (MgO) char layer on the surface of polymer composites upon thermal degradation. However, it has a serious disadvantage in common that a large amount of Mg(OH)2 (generally 40-60 wt%) is required to satisfy the flame retardancy of polymers [9,10]. As a result, such high loading is detrimental to downstream processability and mechanical properties of polymer matrix. Therefore, it is really a challenge to improve the flame retardant efficiency of Mg(OH)2 for minimizing the damage to finally mechanical performances. To obtain acceptable flame retardant efficiency in polymer composites, much excellent work has been done to combine Mg(OH)2 with other synergists, mainly including organic adjuvants [11–14] and inorganic fillers [15–18]. Zhang et al. [11] synthesized aromatic boronic acid derivative 2,4,6-tris(4-boronic-2-thiophene)-1,3,5-triazine * Corresponding author. Nanomaterials Physicochemistry Department, Faculty of Chemical Technology and Engineering, West Pomeranian University of Technology Szczecin, Piast� ow Ave. 42, 71-065, Szczecin, Poland. ** Corresponding author. E-mail addresses: hgwenxin@126.com (X. Wen), ttang@ciac.ac.cn (T. Tang). https://doi.org/10.1016/j.compositesb.2019.107402 Received 10 July 2019; Received in revised form 27 August 2019; Accepted 29 August 2019 Available online 29 August 2019 1359-8368/© 2019 Elsevier Ltd. All rights reserved. H. Chen et al. Composites Part B 177 (2019) 107402 (3 TT-3BA), and demonstrated that mixing epoxy resin (EP) with both 3 TT-3BA and Mg(OH)2 resulted in better thermal stability and flame retardant properties with respect to only one of the compounds. Liang et al. [12] found that the synergistic effects between microencapsulated red phosphorus (MRP) and Mg(OH)2/Al(OH)3 on the flame retardant and tensile properties of polypropylene (PP) composites were significant. Qu group [16] reported that the addition of 5 wt% fumed silica into ethylene-vinyl acetate (EVA)/Mg(OH)2 blends, could decrease the total amount of Mg(OH)2 from 60 wt% into 50 wt% while keeping the V0 rating in UL-94 test. Recently Liu et al. [18] introduced 1 wt% nanocarbon black as synergist into EVA/brucite (a natural mineral type of Mg (OH)2) system while kept the total additives loading was 55 wt%, could improve the UL-94 rating from V1 to V0, and reduce the peak of heat release rate (PHRR) by 79% in comparison with neat EVA. Chen [17] reported that the incorporation of an appropriate loading of lanthanum oxide (La2O3) as a catalytic synergist, could remarkably improve the flame retardancy of Mg(OH)2 filled PP composites due to the greatly improved quality of the condensed phase charred layers deposited on the surface of magnesium oxide (MgO) particles. Up to now, great progress has been made in the exploitation of novel synergists for Mg (OH)2, but in most case the actual applications of their combination are still limited by some unfavorable aspects: relatively high total loading of flame retardant, complicated multiple-step synthesis process for adjuvants, the utilization of hazardous reagents and substantial increase in cost, etc. As above mentioned, the flame retardant effect of Mg(OH)2 is based on cooling, dilution and protection of the magnesia (MgO) char layer against mass and heat transfer between the gaseous flame area and the pyrolyzing melt zone. It is acceptable that the formation of char layer belongs to the condensed phase mechanism on flame retardancy. Once the MgO char layer is effectively strengthened, its barrier action will be enhanced, which is believed to be useful for improving flame retardant efficiency. Expanded graphite (EG), a layered carbon-based material comparable to clays, made of neutral graphene nanoplatelets with potentially high aspect ratio, can greatly improve thermal and gas barrier properties in polymer matrix or foam materials [19,20]. Different from other carbon materials, it has a special accelerated thermal expansion process. Exactly, it begins to expand at about 200 � C via decomposition of interlayer compounds, and the expansion rate can reach 280 times [21]. In our opinion, due to the different density of graphene platelets and MgO particles, the expansion of EG and the decomposition of Mg(OH)2 may produce a special char layer during combustion, and further be associated to the improvement on flame retardancy. To the best of our knowledge, there are few reports regarding the investigation of char structure of EG/Mg(OH)2, especially the spatial distribution of MgO particles and graphite platelets during combustion. In this study, the aim was to design polymer/Mg(OH)2 composites with balanced comprehensive performances, mainly including flame retardancy, thermal stability and mechanical properties. To improve the flame retardant efficiency of Mg(OH)2, EG as a synergist was introduced into poly(butylene succinate) (PBS)/Mg(OH)2 system. The effect of EG content on flame retardancy of PBS/20 wt%Mg(OH)2 composites was investigated by limited oxygen index (LOI), vertical burning test (UL-94) and cone calorimeter test. Moreover, their synergistic flame retarded mechanism was discussed on the base of comparative analysis of char morphology and structure with PBS/20 Mg(OH)2/5Graphene system. In addition, thermal stability and mechanical performances of obtained PBS composites were studied by thermogravimetric analysis (TGA) and tensile/impact tests, respectively. Chemical Corp. (Toyota, Japan). Its melt flow index was 4.5 g/10 min at 190 � C under 2.16 kg of weight. Magnesium hydroxide (MAGNIFIN H-5) was provided by Albemarle Corporation (Charlotte, NC, USA) and its average diameter is about 0.5 μm. Expandable graphite (EG, with carbon content of 96 wt%) with a particle size of 80 mesh, the density of 3.25 g cm 3 and expansion rate of 220, was supplied by Zibo Deli graphite CO. (Shandong, China). It was a natural flake graphite intercalated compound, and the intercalation agents were H2SO4 and H3PO4, and the total amount was approximately 4 wt% inside the EG layers. Graphene sheets (Gr, purity > 99%) with the particle size of 325 mesh, the thickness of 6–10 nm, the surface area of 120–150 m2 g 1 and bulk density of 0.1 g cm 3, was purchased from Nanjing JCNANO Technology Co., Ltd. (Nanjing, China). Before usage, all of them were dried in oven at 110 � C for 10 h. Dicumyl peroxide (DCP) was from Shanghai Wing King Industrial Co., Ltd. (Shanghai, China). Maleic anhydride (MA) was from Tianjin Fuchen Chemical Reagent Factory (Tianjin, China). 2.2. Synthesis of maleic anhydride-grafted PBS Maleic anhydride-grafted PBS (PBS-MA) was synthesized via a reactive melt-grafting process [22]. In brief, PBS, MA and DCP with the ratio of 100/10/1.5 by weight were physically premixed, then they were reactively blended via melt compounding in an intensive mixer (Haake Rheomix 600, Germany) at 135 � C for 7 min. After that, it was refluxed in chloroform for 4 h, and the hot solution was filtered into cold methanol. To remove any unreacted reagents, the precipitated polymer was washed by methanol for 5 times. Finally the obtained sample was dried in vacuum oven at 60 � C for 24 h to get the purified PBS- MA. The grafting degree of MA monomer on PBS chains was determined by chemical titration [23], and it was approximately 4.1 wt%. 2.3. Preparation of PBS composites PBS composites consisting of neat PBS, PBS-MA, Mg(OH)2, EG or Gr were also prepared via melt blending in the intensive mixer at 135 � C with a rotor speed of 60 rpm for 8 min. In all PBS composites containing PBS-MA, the PBS-MA content was kept 15 wt% in constant. The resultant samples were designated as PBS/xMg(OH)2/yEG (or Gr), where x and y denote their weight percentage in the composites. For example, PBS/20 Mg(OH)2/5 EG means that the obtained PBS composites contained 60 wt% PBS, 15 wt% PBS-MA, 20 wt% Mg(OH)2 and 5 wt% EG. For comparison, PBS composites (without PBS-MA) contained 40 wt% Mg(OH)2 and 20 wt% Mg(OH)2/5 wt% EG were prepared under the same processing conditions, which was denoted as PBS-40Mg(OH)2 and PBS-20Mg(OH)2-5EG, respectively. Subsequently, all PBS samples were prepared by hot compression at 135 � C in 10 MPa for 6 min in a plate vulcanizing machine for various tests. 2.4. Characterization The limited oxygen index (LOI) test was performed on an oxygen index meter (JF-3, Jiangning China), according to ASTM D2863-97 with sample dimensions 130 � 6.5 � 3.2 mm3. Vertical burning test (UL-94) was carried out on a vertical burning tester (AG5100B, China) according to ASTM D3801 with sheet dimensions 100 � 13 � 3.2 mm3. Cone calorimeter test (icone, FTT, UK) was conducted according to ISO 5660 at an incident flux of 50 kW/m2, and the size of specimens was 100 � 100 � 6 mm3; it was backed by aluminum foil and each sample was tested for three times. A digital camera was used to take photographs of the residual chars after the cone calorimeter test. The morphologies of the chars were further examined with field-emission scanning electron microscopy (FE-SEM, XL303SEM), the chemical surface characteristics were recorded by EDX. The dispersion state of inorganic fillers in polymer matrix was also examined by field-emission scanning electron microscope (FE-SEM, XL303SEM). The samples were fractured in liquid nitrogen, and the 2. Experiments 2.1. Materials Poly(butylene succinate) (PBS) was purchased from Mitsubishi 2 H. Chen et al. Composites Part B 177 (2019) 107402 fracture surfaces were coated with gold before observation. Thermogravimetric analysis (TGA) was performed on a TA STD Q600 thermal analyzer. Samples with 8.0 � 0.2 mg were heated from romm temperature to 600 � C at a rate of 10 � C/min under air atmosphere. The tensile test was carried out in a uniaxial testing machine (Instron 1121, China) according to GB/T1040.3–2006. Sample specimen was 20 � 4 � 1 mm3 with a dumbbell shape. It was performed at a tensile speed of 5 mm min 1. The impact test with V-notched specimens (55 � 6 � 4 mm3) was accomplished in an impact testing machine (CEAST, China) according to GB1843-93. All mechanical tests were performed at 50% relative humidity and room temperature. Averaged data were recorded after five runs measurement. indicating the same flame retardancy as PBS/20 Mg(OH)2/5 EG (Table 1). Therefore, the effect of PBS-MA on flame retardancy was negligible, which might be ascribed to no flame-retardant elements in PBS-MA. 3.1.2. Cone calorimeter test To further evaluate the flame retarded effect of EG in PBS/Mg(OH)2 system, the combustion behavior of pure PBS, PBS/40 Mg(OH)2, PBS/ 20 Mg(OH)2/5 EG and PBS/20 Mg(OH)2/5Gr were comparatively studied by cone calorimeter test. As is known to all, the heat release rate (HRR) is one of the most vital parameters related to the flame retardant performance of polymer materials during combustion [27]. Fig. 1 shows HRR curves of pure PBS and its composites, and specific combustion parameters are summarized in Table 2. The time to ignition (TTI) was 48 s for pure PBS, while all PBS composites displayed increments on TTI due to the decomposition of Mg(OH)2 to release water vapor that would assist in cooling and dilution flammable gases [28]. It is noted that the TTI of PBS/20 Mg(OH)2/5 EG was a little longer than that of PBS/20 Mg (OH)2/5Gr, which might be ascribed to the thermal expansion of EG to release some nonflammable gases [29], and was related to the lower heat absorption coefficient of EG than Gr [26]. Generally a concentration of heat absorption in the top layer, higher heat absorption coefficient is believed to enhance early ignition. Similar phenomena were also reported by B. Schartela group [26]. Moreover, it was observed that pure PBS was readily flammable after ignition and presented a sharp PHRR as high as 766 kW/m2. For PBS/40 Mg(OH)2, it burned a longer time than neat PBS and its PHRR was significantly decreased to 245 kW/m2, indicating a good flame retarded effect. Interestingly, PBS/20 Mg (OH)2/5 EG exhibited a nearly flat HRR curve, and its PHRR value was 173 kW/m2, which decreased to 23% of that of pure PBS. In contrast, a similar decrease trend was present in PBS/20 Mg(OH)2/5Gr, but its PHRR value (287 kW/m2) was higher than that of PBS/20 Mg (OH)2/5 EG, which suggested that EG has better synergistically flame retarded effect than Gr in our PBS system. The total heat release (THR) curves for PBS samples with combustion time were presented in Fig. 2. Neat PBS displayed a steep curve, and the resultant THR value was as high as 171 kJ/m2. In the final stage, the THR curves for PBS/40 Mg(OH)2 and PBS/20 Mg(OH)2/5 EG were almost overlapped, and correspondingly their THR values were 100 kJ/ m2 and 101 kJ/m2, respectively. However, the curve slope of PBS/20 Mg (OH)2/5Gr was larger than that of PBS/20 Mg(OH)2/5 EG, and it also exhibited bigger THR value (114 kJ/m2). This further confirmed that EG has better synergistic effect than Gr on improving the flame retardancy 3. Results and discussion 3.1. Flame retardancy 3.1.1. LOI and UL-94 It was reported that a high loading of Mg(OH)2 was needed to achieve acceptable flame retardancy in polymer systems [18]. Initially the influence of Mg(OH)2 loading on the flame retardancy of PBS was evaluated by LOI and UL-94 tests (shown in Table 1). With the increase of its loading from 20 to 40 wt%, the LOI values gradually increased from 26.2 to 31.9% (the LOI of neat PBS was 21.0%). However, only when 40 wt% Mg(OH)2 was added, the UL-94 rating could reach to V0. In this regard, the least required loading of Mg(OH)2 was 40 wt% in our PBS system (abbreviated to PBS/40 Mg(OH)2), implying its low flame retardant efficiency. Subsequently, EG as a synergist was added into PBS/20 wt% Mg (OH)2 composites. With the increase of EG content from 2.5 to 10 wt%, the LOI value increased from 28.5 to 32.6% (Table 1). When the EG content was more than 5 wt%, the UL-94 could pass V0 ratting. In consequence, the combination of 5 wt% EG and 20 wt% Mg(OH)2 was identified as the optimal ratio in our PBS system, which was denoted as PBS/20 Mg(OH)2/5 EG for convenience of discussion later. In addition, as a comparison with EG, 5 wt% graphene sheets were also added into PBS/20 wt% Mg(OH)2 (denoted as PBS/20 Mg(OH)2/ 5Gr). It is well-known that graphene was considered to be one efficient synergist in polymer composites [24,25], although it was much more expensive than EG. As shown in Table 1, it was disappointing that the LOI of PBS/20 Mg(OH)2/5Gr was only 28.2%, and the resultant UL-94 rating reached to V1. These results suggested that EG was a more efficient synergist than Gr in improving the flame retardancy in PBS/Mg (OH)2 system. Similar results were also reported by B. Schartela group in PP system [26]. It was related to the dispersion of carbon fillers in polymer matrix, which will be discussed later by SEM. Moreover, the thermal decomposition of EG might be favorable to improve the flame retardancy in LOI and UL-94 due to the “dilutive effect” from intercalation agents. Additionally, the effect of PBS-MA on LOI and UL-94 was also discussed. For PBS-20Mg(OH)2-5EG (corresponding to the sample without PBS-MA), its LOI value was 29.2% and UL-94 rating was V0, Table 1 Combustion properties of PBS samples from LOI and UL-94 tests. Samples LOI (%) PBS PBS/20 Mg(OH)2 PBS/30 Mg(OH)2 PBS/40 Mg(OH)2 PBS/20 Mg(OH)2/2.5 EG PBS/20 Mg(OH)2/5 EG PBS/20 Mg(OH)2/7.5 EG PBS/20 Mg(OH)2/10EG PBS/20 Mg(OH)2/5Gr PBS-20Mg(OH)2-5EG 21.0 � 0.1 26.2 � 0.1 28.3 � 0.2 31.9 � 0.3 28.5 � 0.2 29.4 � 0.2 30.7 � 0.3 32.6 � 0.3 28.2 � 0.3 29.2 � 0.3 UL-94 (3.2 mm) Rating Fail Fail V1 V0 V1 V0 V0 V0 V1 V0 Dripping Yes No No No No No No No No No Fig. 1. Heat release rate curves of PBS and its composites by cone calorimeter test at an external radiant flux of 50 kW/m2. 3 H. Chen et al. Composites Part B 177 (2019) 107402 Table 2 Combustion parameters of PBS samples from cone calorimeter test. Samples TTIa (sec) PHRRb (kW/ m2) THRc (MJ/ m2) Residual chard (%)a PBS PBS/40 Mg(OH)2 PBS/20 Mg (OH)2/5 EG PBS/20 Mg(OH) 2/5Gr 48 � 1 56 � 2 55 � 2 766 � 11 245 � 13 173 � 10 171 � 4 100 � 6 101 � 5 0.1 � 0.1 28.8 � 0.5 18.5 � 0.4 52 � 2 287 � 12 114 � 6 18.9 � 0.5 a TTI: time to ignition. PHRR: peak heat release rate. c THR: total heat release. d The residual char was calculated according to the value at 460 s for neat PBS and 750s for its composites in mass loss rate curves. b Fig. 3. Mass loss rate curves of PBS and its composites by cone calorimeter test at an external radiant flux of 50 kW/m2. mechanism of polymer materials [30]. The digital photographs of char residues from PBS samples are shown in Fig. 4. Neat PBS was fully burned without any carbon residue (not shown here). In Fig. 4a, the residues from PBS/40 Mg(OH)2 formed a strong and continuous char layer due to the space stacking of MgO particles [31]. Surprisingly, as shown in Fig. 4b, PBS/20 Mg(OH)2/5 EG presented a cancellated and compact char layers with no surface cracks. Similarly, PBS/20 Mg (OH)2/5Gr exhibited a complete char layer, but its color looked lighter with a smooth surface (Fig. 4c). However, from the side view, their char structure was quite different. For PBS/20 Mg(OH)2/5 EG, the top half was black color, while the bottom half was grey (Fig. 4d). This means most worm-like graphite sheets are enriched in the top zone, while MgO particles are mainly distributed in the bottom zone. In Fig. 4e, no distinct stratification was observed from PBS/20 Mg(OH)2/5Gr, suggesting the char was composed of both graphite sheets and MgO particles with uniform mixing. Furthermore, the char of PBS/20 Mg(OH)2/5 EG was three times thicker than PBS/20 Mg(OH)2/5Gr, which could provide better barrier action as protective char layer for heat and mass transfer. Whether does the different distribution of graphite sheets and MgO particles determine the resultant flame retardancy? To make it clear, the char microstructure of PBS/20 Mg(OH)2/5 EG was further observed by SEM. Within the top area of the char layer (showed in Fig. 5a and b), it was mainly composed of graphite sheets, and a small number of MgO particles were filled in their cracks, which could inhibit the aggregate of graphite sheets during combustion. Within the middle area (Fig. 5c and d), the amount of graphite sheets decreased while that of MgO increased. Moreover, within the bottom area of the char (Fig. 5e and f), it only contained MgO particles but no graphite sheets. EDX spectra were also employed to analyze the residual chars from PBS/20 Mg(OH)2/5 EG. As shown in Fig. 6, the chars from three areas were composed of C, O and Mg elements, but their content was quite different. In the upper area, the percent for C and Mg was 73.43% and 15.18%, respectively. In the middle area, the value changed into 22.32% and 46.08%. In the bottom area, it was 7.62% and 54.90%. These results indicated the formation of a special gradient-structured char layer consisting of graphite sheets and MgO particles. From top to bottom, the amount of EG sheets gradually decreased while that of MgO particles increased correspondingly. Generally compared to zero-dimensional MgO particles, twodimensional EG sheets had better “labyrinth effect” [32,33], so this gradient-structured char layer was favorable to improve barrier effect for the entry and exit of gaseous substances. Based on above analysis, a possible flame retarded mechanism for PBS/20 Mg(OH)2/5 EG was presented in Fig. 7. When PBS/20 Mg(OH)2/ Fig. 2. Total heat release curves of PBS and its composites by cone calorimeter test at an external radiant flux of 50 kW/m2. of PBS/Mg(OH)2 composites. The mass loss rate (MLR) curves of above PBS samples were shown in Fig. 3. The trend of MLR in the later stage of combustion was consistent with the reduction on PHRR, because lower slope of MLR curve means slower burning rate and smaller value of PHRR. It was apparent that a quickly dropped MLR curve for neat PBS represented rapid combustion of fuel. Differently, PBS/40 Mg(OH)2 and PBS/20 Mg(OH)2/5 EG presented gentler MLR curves, implying their slower burning rate. Furthermore, PBS/20 Mg(OH)2/5 EG exhibited a smaller slope than PBS/20 Mg(OH)2/5Gr, implying better synergistically flame retarded effect from EG. As listed in Table 2, the yield of residual chars from PBS/ 40 Mg(OH)2 was 28.8%, which is well consistent with the MgO yield from Mg(OH)2. Meanwhile, PBS/20 Mg(OH)2/5 EG and PBS/20 Mg (OH)2/5Gr had very close value on the yield of residual chars. More exactly, the yield from PBS/20 Mg(OH)2/5 EG was a little lower, which might be connected to the volatilization of intercalation agent in EG during combustion. 3.1.3. Analysis for flame retarded mechanism According to the above results, PBS/20 Mg(OH)2/5 EG showed better flame retardancy than PBS/20 Mg(OH)2/5Gr, and it was comparable to (or even better than) PBS/40 wt%Mg(OH)2. Why is EG more efficient as synergist than Gr? And what is the flame retarded mechanism of EG in PBS/20 Mg(OH)2 system? To answer these questions, the morphology and structure of the residual chars after cone calorimeter test was investigated, which is beneficial to illuminate the combustion behavior and flame retarded 4 H. Chen et al. Composites Part B 177 (2019) 107402 Fig. 4. Photographs of the residual chars after cone calorimeter test from top view for (a) PBS/40 Mg(OH)2, (b) PBS/20 Mg(OH)2/5 EG, (c) PBS/20 Mg(OH)2/5Gr and from side view for (d) PBS/20 Mg(OH)2/5 EG, (e) PBS/20 Mg(OH)2/5Gr. Fig. 5. SEM micrographs of the residual chars after cone calorimeter test from PBS/20 Mg(OH)2/5 EG: (a, b) upper area; (c, d) middle area; (e, f) bottom area. 5 H. Chen et al. Composites Part B 177 (2019) 107402 matrix, leading to a substantial reduction in molecular weight. Therefore, it is necessary to decrease the loading of Mg(OH)2 so as to minimise its damage to thermal stability and mechanical performances. The thermal stability of PBS/20 Mg(OH)2/5 EG was investigated by TGA in air atmosphere, meanwhile neat PBS and PBS/40 Mg(OH)2 were selected as references. Fig. 8 presents their TGA and DTG curves, and the detailed parameters are listed in Table 3. The T5wt% and T10 wt% corresponded to the temperature with 5 wt% and 10 wt% weight loss respectively, which is essential to evaluate thermal decomposition behavior of polymer on the onset stage. Compared to neat PBS, it seemed that PBS/40 Mg(OH)2 and PBS/20 Mg(OH)2/5 EG exhibited approximately 60 � C decrease on both T5wt% and T10 wt%. It is noted that the decomposition region of Mg(OH)2 was approximately from 290 to 350 � C, so the recorded weight loss was from Mg(OH)2, but not from the real initial degradation temperature of PBS chains. Considering the processing temperature of PBS at 135 � C, here the starting decomposition temperatures for Mg(OH)2 (or degradation temperatures for PBS) was higher than 250 � C, indicating no apparent damage to thermal stability of PBS. As shown in Table 3, another important thermal parameter, Tmax corresponds to the temperature with maximum weight loss rate was present. The Tmax for neat PBS presented at 396.9 � C. For PBS/40 Mg (OH)2 and PBS/20 Mg(OH)2/5 EG, there were two peaks in DTG curve (denoted as Tmax1 and Tmax2), assigned to the maximal decomposition temperature of Mg(OH)2 and PBS, respectively. It was found that the Tmax2 in PBS/40 Mg(OH)2 was 423.1 � C, which was much higher than that of neat PBS, indicating that the presence of Mg(OH)2 could inhibit the thermal decomposition of PBS with better thermal stability. This also confirmed that the decrease on both T5wt% and T10 wt% resulted from the weight loss of Mg(OH)2 (If it is from the initial degradation of PBS chains, it means the presence of Mg(OH)2 will promote the thermal decomposition of PBS, so subsequently the Tmax will be greatly decreased). Moreover, the Tmax1 of PBS/20 Mg(OH)2/5 EG was a little higher than that of PBS/40 Mg(OH)2, which might be ascribed to the positive effect of graphite sheets on delaying the decomposition of Mg (OH)2 [39]. Meanwhile, the Tmax2 for PBS/20 Mg(OH)2/5 EG shifted to 443.7 � C, which was 20 � C higher than that of PBS/40 Mg(OH)2, further indicating the enhanced barrier effect of graphite sheets on inhibiting the decomposition of PBS. In brief, the incorporation of EG into the PBS/Mg(OH)2 composites was beneficial to further enhance the thermal stability of PBS in comparison with the addition of Mg(OH)2 alone. Fig. 6. EDX spectra of the residual chars after cone calorimeter test from PBS/ 20 Mg(OH)2/5 EG: (a) upper area; (b) middle area; (c) bottom area. 5 EG was put into the flame with high-temperature, the thermal expansion of EG was almost simultaneous with the decomposition process of Mg(OH)2. The movement of in situ generated lightweight graphite sheets was promoted by the rapid evaporation of water vapor from bottom to up [34,35], so the aggregate trend was limited and graphite sheets were isolated, resulting in the formation of gradient-structured char layer with large volume. Furthermore, the gradient-structured char layer could display some positive effects as follows: i) the increase on the total thickness of char layer could effectively delay the penetration time of the outer heat/oxygen into the inner polymer degradation domain; ii) the specially gradient distribution of graphite sheets could enhance “labyrinth effect” to prolong the spatial diffusion distance of flammable gas products into combustion zone (shown in Fig. 7). Therefore, the gradient-structured char layer could play an important role to maximize flame retardant effect in the condensed phase. 3.2. Thermal stability Thermal stability of polymer composites is very significant because it is greatly related to their processing temperature range and resultant mechanical properties [36–38]. Generally when a high loading of inorganic fillers (such as Mg(OH)2) was added, the rapidly rising viscosity and temperature will cause thermal degradation of polymer 3.3. Mechanical properties The mechanical properties of neat PBS, PBS/40 Mg(OH)2 and PBS/ 20 Mg(OH)2/5 EG were evaluated by tensile and impact tests. As a Fig. 7. Possible mechanism of the proceeding combustion of PBS composites. 6 H. Chen et al. Composites Part B 177 (2019) 107402 Fig. 8. TGA(a) and DTG (b) curves of pure PBS and its composites at a heating rate of 10 � C min Table 3 Thermal decomposition properties of PBS samples in air. 1 in air atmosphere. Table 4 Mechanical properties for PBS and its composites from tensile and impact tests. Samples T5wt% (oC) T10 wt%(oC) Tmax (oC) Tpeak1 Tpeak2 PBS PBS/40 Mg(OH)2 PBS/20 Mg(OH)2/5 EG 352.1 294.2 294.3 363.7 301.6 303.2 – 309.9 311.9 396.9 423.1 443.7 comparison, the mechanical properties of PBS-40Mg(OH)2 (corresponded to PBS composites contained 40 wt% Mg(OH)2 without PBSMA) was also evaluated. Their stress-strain curves are shown in Fig. 9, the detailed data for Young’s modulus, tensile strength, yield strength, elongation at break and notched-impact strength are summarized in Table 4. For PBS-40Mg(OH)2 and PBS/40 Mg(OH)2, Young’s modulus increased due to the reinforced effect of rigid inorganic filler, but tensile strength, elongation at break and impact strength were greatly decreased. Especially, the elongation at break of them was only 14% and 17%, which was much lower than that of neat PBS with 402%. It was reported that higher content fillers resulted in the decreased ductility of polymer matrix due to the disability of fillers to sustain the forces transferred by the matrix, which disrupted the propagation of stress and agitate brittleness in polymer composites [40–42]. In this case, 40 wt% Mg(OH)2 was a too high amount to seriously deteriorate the strength and toughness of PBS. Relatively, the mechanical properties of Samples Young’s modulus (MPa) Tensile strength (MPa) Yield strength (MPa) Elongation at break (%) Impact strength (kJ/m2) PBS PBS40Mg (OH)2 PBS/ 40 Mg (OH)2 PBS/ 20 Mg (OH)2/ 5 EG 320 � 25 542 � 28 32.8 � 1.2 24.3 � 1.0 28.5 � 1.1 24.3 � 1.0 402 � 12 14 � 2 8.7 � 0.7 2.1 � 0.3 561 � 36 27.6 � 1.0 27.6 � 1.0 17 � 1 2.4 � 0.2 475 � 32 28.9 � 1.1 28.4 � 1.2 289 � 9 5.0 � 0.4 PBS/40 Mg(OH)2 were better than that of PBS-40Mg(OH)2 due to the compatible effect of PBS-MA. However, for PBS/20 Mg(OH)2/5 EG, except for a little lower value on Young’s modulus, other performance indexes were much better than PBS/40 Mg(OH)2, so that a balance on mechanical performances with good stiffness and toughness was obtained. Especially, although PBS/20 Mg(OH)2/5 EG exhibited lower tensile strength, its yield strength was very close to that of neat PBS. These results suggested that only when an appropriate content of Mg (OH)2 (less than 20 wt%) and PBS-MA as compatibilizer were added, satisfied mechanical performances of PBS composites could be achieved. It is acceptable that the dispersion of inorganic fillers and interface adhesion with polymer are two of the most important factors to determine mechanical properties of polymer composites [43–45]. In our PBS system, 15 wt% PBS-g-MA was added as compatibilizer. To evaluate its positive effect, the fracture surfaces of the PBS composites were comparatively investigated by SEM (shown in Fig. 10). For PBS-40Mg (OH)2, it exhibited a rough surface (Fig. 10a1), and some protruded particles were apparent (blue circles in Fig. 10a2). Differently, on the help of PBS-MA, PBS/40 Mg(OH)2 showed a relatively flat surface and presented more uniformly distributed particles (Fig. 10b1). Meanwhile, these particles adhered firmly to the polymer matrix, indicated a stronger interfacial interaction (red circles in Fig. 10b2). Moreover, for PBS/20 Mg(OH)2/5 EG, the surface was also flat (Fig. 10c1), and micron-sized EG flakes were visible with good dispersion (green circles in Fig. 10c2). These results indicated that PBS-MA played an important role to increase dispersion of fillers and enhance interface adhesion with polymer matrix, which is well related to above mechanical properties. 4. Conclusions EG assistant construction of gradient-structured char layer in PBS/ Fig. 9. Stress-strain curves of PBS and its composites from tensile test. 7 H. Chen et al. Composites Part B 177 (2019) 107402 Fig. 10. FE-SEM micrographs of the brittle-fractured surfaces of PBS composites: (a1,a2) PBS-40Mg(OH)2, (b1,b2) PBS/40 Mg(OH)2 and (c1,c2) PBS/20 Mg (OH)2/5 EG. References Mg(OH)2 system was developed. On the premise of comparable flame retardancy to PBS/40 Mg(OH)2, PBS/20 Mg(OH)2/5 EG displayed much better thermal stability and mechanical properties. The enhanced mechanism on flame retardancy was mainly ascribed to the formation of gradient-structured char layer during combustion, which could enhance barrier action in condensed phase to inhibit the permeation of heat, oxygen and flammable gas products via“labyrinth effect”. Meanwhile, as an effective compatibilizer, PBS-MA played an important role to increase the dispersion of fillers and enhance interface adhesion with polymer matrix, resulting in the improvements on mechanical properties. To the best of our knowledge, it is the first report to discuss the effect of gradient-structured char layer on the flame retardancy in polymer systems. In summary, this work provides a facile and efficient method to improve the flame retardant efficiency of Mg(OH)2, which is helpful to fabricate high-performance polymer composites. [1] Wang X, Kalali EN, Wan J-T, Wang D-Y. Carbon-family materials for flame retardant polymeric materials. Prog Polym Sci 2017;69:22–46. [2] Ran S, Fang F, Guo Z, Song P, Cai Y, Fang Z, Wang H. Synthesis of decorated graphene with P, N-containing compounds and its flame retardancy and smoke suppression effects on polylactic acid. Compos B Eng 2019;170:41–50. [3] Liu L, Xu Y, Xu M, Li Z, Hu Y, Li B. Economical and facile synthesis of a highly efficient flame retardant for simultaneous improvement of fire retardancy, smoke suppression and moisture resistance of epoxy resins. Compos B Eng 2019;167: 422–33. [4] Fang F, Ran SY, Fang ZP, Song PG, Wang H. Improved flame resistance and thermomechanical properties of epoxy resin nanocomposites from functionalized graphene oxide via self-assembly in water. Compos B Eng 2019;165:406–16. [5] Sun JH, Li L, Li J. Effects of furan-phosphamide derivative on flame retardancy and crystallization behaviors of poly(lactic acid). Chem Eng J 2019;369:150–60. [6] Yang H, Yu B, Song P, Maluk C, Wang H. Surface-coating engineering for flame retardant flexible polyurethane foams: a critical review. Compos B Eng 2019;176: 107185. [7] Ma YH, Qi F, Chen M, Chen XL, Qin J, Tang MQ. Study on viscoelastic, crystallization, and mechanical properties of HDPE/EVA/Mg(OH)(2) composites. Polym Compos 2017;38(6):1221–6. [8] Yang YR, Niu M, Li JJ, Xue BX, Dai JM. Preparation of carbon microspheres coated magnesium hydroxide and its application in polyethylene terephthalate as flame retardant. Polym Degrad Stab 2016;134:1–9. [9] Wang CF, Wang YL, Han ZD. Enhanced flame retardancy of polyethylene/ magnesium hydroxide with polycarbosilane. Sci Rep 2018;8:10. [10] Chen H, Wen X, Guan YY, Min JK, Wen YL, Yang HF, Chen XC, Li YH, Yang XY, Tang T. Effect of particle size on the flame retardancy of poly(butylene succinate)/ Mg(OH)(2) composites. Fire Mater 2016;40(8):1090–6. Acknowledgements This work is supported by the National Science Centre Poland OPUS UMO-2018/29/B/ST8/01265 and the National Natural Science Foundation of China (51773202, 21204079, 51073149 and 50873099). 8 H. Chen et al. Composites Part B 177 (2019) 107402 [28] Huang H, Tian M, Liu L, Liang W, Zhang L. Effect of particle size on flame retardancy of Mg(OH)2-filled ethylene vinyl acetate copolymer composites. J Appl Polym Sci 2006;100(6):4461–9. [29] Focke WW, Badenhorst H, Mhike W, Kruger HJ, Lombaard D. Characterization of commercial expandable graphite fire retardants. Thermochim Acta 2014;584:8–16. [30] Schartel B, Bartholmai M, Knoll U. Some comments on the main fire retardancy mechanisms in polymer nanocomposites. Polym Adv Technol 2006;17(9–10): 772–7. [31] Liu L, Hu J, Zhuo J, Jiao C, Chen X, Li S. Synergistic flame retardant effects between hollow glass microspheres and magnesium hydroxide in ethylene-vinyl acetate composites. Polym Degrad Stab 2014;104:87–94. [32] Lecouvet B, Sclavons M, Bourbigot S, Bailly C. Thermal and flammability properties of polyethersulfone/halloysite nanocomposites prepared by melt compounding. Polym Degrad Stab 2013;98(10):1993–2004. [33] Wang W, Pan HF, Shi YQ, Yu B, Pan Y, Liew KM, Song L, Hu Y. Sandwichlike coating consisting of alternating montmorillonite and beta-FeOOH for reducing the fire hazard of flexible polyurethane foam. ACS Sustain Chem Eng 2015;3(12): 3214–23. [34] Duquesne S, Delobel R, Le Bras M, Camino G. A comparative study of the mechanism of action of ammonium polyphosphate and expandable graphite in polyurethane. Polym Degrad Stab 2002;77(2):333–44. [35] Zheng Z, Liu Y, Zhang L, Wang H. Synergistic effect of expandable graphite and intumescent flame retardants on the flame retardancy and thermal stability of polypropylene. J Mater Sci 2016;51(12):5857–71. [36] Rajaei M, Wang DY, Bhattacharyya D. Combined effects of ammonium polyphosphate and talc on the fire and mechanical properties of epoxy/glass fabric composites. Compos B Eng 2017;113:381–90. [37] He WT, Gao J, Liao ST, Wang XX, Qin SH, Song PA. A facile method to improve thermal stability and flame retardancy of polyamide 6. Compos Commun 2019;13: 143–50. [38] Song PA, Liu LN, Huang GB, Yu YM, Guo QP. Largely enhanced thermal and mechanical properties of polymer nanocomposites via incorporating C-60@ graphene nanocarbon hybrid. Nanotechnology 2013;24(50):8. [39] Khalili P, Tshai KY, Kong I. Natural fiber reinforced expandable graphite filled composites: evaluation of the flame retardancy, thermal and mechanical performances. Compos Appl Sci Manuf 2017;100:194–205. [40] Zaghloul MMY, Zaghloul MMY. Influence of flame retardant magnesium hydroxide on the mechanical properties of high density polyethylene composites. J Reinf Plast Compos 2017;36(24):1802–16. [41] Song P, Wang H. High-performance polymeric materials through hydrogen-bond cross-linking. Adv Mater 2019;0(0):1901244. [42] Song P, Xu Z, Wu Y, Cheng Q, Guo Q, Wang H. Super-tough artificial nacre based on graphene oxide via synergistic interface interactions of π-π stacking and hydrogen bonding. Carbon 2017;111:807–12. [43] Guo SM, Ma LC, Song GJ, Li XR, Li PY, Wang MY, Shi LL, Gu Z, Huang YD. Covalent grafting of triazine derivatives onto graphene oxide for preparation of epoxy composites with improved interfacial and mechanical properties. J Mater Sci 2018; 53(24):16318–30. [44] Qian M, Sun Y, Xu X, Liu L, Song P, Yu Y, Wang H, Qian J. 2D-alumina platelets enhance mechanical and abrasion properties of PA612 via interfacial hydrogenbond interactions. Chem Eng J 2017;308:760–71. [45] Song PG, Dai JF, Chen GR, Yu YM, Fang ZP, Lei WW, Fu SY, Wang H, Chen ZG. Bioinspired design of strong, tough, and thermally stable polymeric materials via nanoconfinement. ACS Nano 2018;12(9):9266–78. [11] Zhang T, Liu W, Wang M, Liu P, Pan Y, Liu D. Synergistic effect of an aromatic boronic acid derivative and magnesium hydroxide on the flame retardancy of epoxy resin. Polym Degrad Stab 2016;130:257–63. [12] Liang JZ, Feng JQ, Tsui CP, Tang CY, Liu DF, Zhang SD, Huang WF. Mechanical properties and flame-retardant of PP/MRP/Mg(OH)2/Al(OH)3 composites. Compos B Eng 2015;71:74–81. [13] Liu JC, Peng S, Zhang YB, Chang HB, Yu ZL, Pan BL, Lu C, Ma JY, Niu QS. Influence of microencapsulated red phosphorus on the flame retardancy of high impact polystyrene/magnesium hydroxide composite and its mode of action. Polym Degrad Stab 2015;121:208–21. [14] Ye L, Miao YY, Yan H, Li Z, Zhou YL, Liu JX, Liu H. The synergistic effects of boroxo siloxanes with magnesium hydroxide in halogen-free flame retardant EVA/MH blends. Polym Degrad Stab 2013;98(4):868–74. [15] Ye L, Wu Q, Qu B. Synergistic effects and mechanism of multiwalled carbon nanotubes with magnesium hydroxide in halogen-free flame retardant EVA/MH/ MWNT nanocomposites. Polym Degrad Stab 2009;94(5):751–6. [16] Fu M, Qu B. Synergistic flame retardant mechanism of fumed silica in ethylenevinyl acetate/magnesium hydroxide blends. Polym Degrad Stab 2004;85(1):633–9. [17] Shen L, Chen Y, Li P. Synergistic catalysis effects of lanthanum oxide in polypropylene/magnesium hydroxide flame retarded system. Compos Appl Sci Manuf 2012;43(8):1177–86. [18] Liu ZQ, Zhang YL, Li N, Wen X, Nogales LA, Li LJ, Guo F. Study of nanocarbon black as synergist on improving flame retardancy of ethylene-vinyl acetate/brucite composites. J Therm Anal Calorim 2019;136(2):601–8. [19] He D, Zhao CX, Gou HL, Li YT, Xiang D. Flame retardancy and toughening properties of epoxy composites containing ammonium polyphosphate microcapsules and expanded graphite. High Perform Polym 2018;30(10):1247–59. [20] Wang SJ, Qian LJ, Xin F. The synergistic flame-retardant behaviors of pentaerythritol phosphate and expandable graphite in rigid polyurethane foams. Polym Compos 2018;39(2):329–36. [21] Ming G, Chen S, Sun YJ, Wang YX. Flame retardancy and thermal properties of flexible polyurethane foam containing expanded graphite. Combust Sci Technol 2017;189(5):793–805. [22] Phua YJ, Chow WS, Ishak ZAM. Reactive processing of maleic anhydride-grafted poly(butylene succinate) and the compatibilizing effect on poly(butylene succinate) nanocomposites. Express Polym Lett 2013;7(4):340–54. [23] Detyothin S, Selke SEM, Narayan R, Rubino M, Auras R. Reactive functionalization of poly(lactic acid), PLA: effects of the reactive modifier, initiator and processing conditions on the final grafted maleic anhydride content and molecular weight of PLA. Polym Degrad Stab 2013;98(12):2697–708. [24] Ming G, Li JF, Zhang XQ, Yue LN, Chai ZH. The flame retardancy of epoxy resin including the modified graphene oxide and ammonium polyphosphate. Combust Sci Technol 2018;190(6):1126–40. [25] Dittrich B, Wartig KA, Mulhaupt R, Schartel B. Flame-retardancy properties of intumescent ammonium poly(phosphate) and mineral filler magnesium hydroxide in combination with graphene. Polymers 2014;6(11):2875–95. [26] Dittrich B, Wartig KA, Hofmann D, Mulhaupt R, Schartel B. Carbon black, multiwall carbon nanotubes, expanded graphite and functionalized graphene flame retarded polypropylene nanocomposites. Polym Adv Technol 2013;24(10):916–26. [27] Zanetti M, Kashiwagi T, Falqui L, Camino G. Cone calorimeter combustion and gasification studies of polymer layered silicate nanocomposites. Chem Mater 2002; 14(2):881–7. 9