Highly Efficient Composite Flame Retardants for Improving the Flame Retardancy, Thermal Stability, Smoke Suppression, and Mechanical Properties of EVA

Ethylene vinyl acetate (EVA) copolymer has been used extensively in many fields. However, EVA is flammable and releases CO gas during burning. In this work, a composite flame retardant with ammonium polyphosphate (APP), a charring–foaming agent (CFA), and a layered double hydroxide (LDH) containing rare-earth elements (REEs) was obtained and used to improve the flame retardancy, thermal stability, and smoke suppression for an EVA matrix. The thermal analysis showed that the maximum thermal degradation temperature of all composites increased by more than 37 °C compared with that of pure EVA. S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA could achieve self-extinguishing behavior according to the UL-94 tests (V-0 rating). The peak heat release rate (pk-HRR) indicated that all LDHs containing REEs obviously reduced the fire strength in comparison with S-MgAl. In particular, pk-HRR of S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA and S-NdMgAl/APP/CFA/EVA were all decreased by more than 82% in comparison with pure EVA. Furthermore, the total heat release (THR), smoke production rate (SPR), and production rate of CO (COP) also decreased significantly. The average mass loss rate (AMLR) confirmed that the flame retardant exerted an effect in the condensed phase of the composites. Meanwhile, the combination of APP, CFA, and LDH containing REEs allowed the EVA matrix to maintain good mechanical properties.


Introduction
Ethylene vinyl acetate (EVA) copolymer is used extensively in many different fields, such as the wire, cable, and wrapper industries [1]. However, EVA is flammable and releases CO gas during burning [2]. In recent years, layered double hydroxides (LDHs) and intumescent flame retardants (IFRs) have been used to increase the flame retardancy and smoke suppression of the EVA matrix [3,4]. The LDH is composed of positively charged brucite-like layers, meanwhile containing charge-compensating anions and water molecules in the interlayer gallery [5]. The general formula of LDH is [M 2+ 1−x M 3+ x (OH) 2 ] x+ [A n− ] x/n ·mH 2 O [6], where M 2+ and M 3+ are divalent and trivalent cations that occupy octahedral positions in the hydroxide layers. A n− denotes the anions in the interlayer. X stands for the cationic ratio of M 3+ /(M 2+ + M 3+ ) and it can be adjusted [7]. LDH can be used in flame retardation extensively due to its advantages, such as low smoke, non-toxicity, and no harmful gas generation [8]. During burning, LDH releases water vapor and carbon dioxide that can cool the species and dilute the surrounding flammable gas to prevent the burning of composites [9]. Rare-earth elements are an important material that can be used in many fields such as global communications, medicine, agriculture, batteries, and hybrid engines due to their electronic, magnetic, and catalytic properties [14][15][16][17][18]. However, the influence of a composite flame retardant including REEs, LDHs, and IFRs on the flame retardancy of EVA has not been reported. In this paper, APP and CFA with a mass ratio of 3:1 were combined with LDH doped with trace amounts of REEs (La, Ce, Nd) to obtain a novel composite flame retardant of EVA. The new flame-retardant system not only maintained their mechanical properties well, but also achieved excellent flame retardancy and smoke suppression.

Synthesis of the Modified LDHs
A mixed aqueous solution of Mg(NO 3 ) 2 and Al(NO 3 ) 3 with Mg 2+ and Al 3+ in a molar rate of 4:1 was adjusted to a constant pH [8][9] under stirring at 70 • C. Then, a 0.25 g amount of sodium stearate was mixed with the above solution and reacted for 2 h. The prepared precursor was put in a microwave oven at 70 • C for 30 min. The precipitate was washed with deionized water to pH = 7, then filtered and dried. The sample was called S-MgAl. The LDH containing REEs (La, Ce, Nd) was obtained by the same method with a REEs 3+ :Mg 2+ :Al 3+ molar rate of 0.05:4:0.95. The samples were called S-LaMgAl, S-CeMgAl, and S-NdMgAl, respectively.

Preparation of CFA
The synthetic scheme of CFA is shown in Figure 2 [19]. The experimental device consisted of a 1000 mL four-neck flask, a stirrer, a thermometer, a reflux condenser, and a dropping funnel. First, 0.4 mol cyanuric chloride and 200 mL acetone were added into the four-neck flask. A solution of ethanolamine in acetone and a solution of sodium hydroxide were simultaneously added dropwise into the four-neck flask and the experimental temperature was kept at 0-5 • C. The obtained reactive system was called intermediate I.
into the four-neck flask and the experimental temperature was kept at 0-5 °C. The obtained reactive 86 system was called intermediate I. and 1.0 mol sodium hydroxide in 500 mL water were added into a 2000 mL four-neck flask provided 92 with a stirrer, a thermometer, and a reflux condenser, and the reacting solution was kept under reflux 93 for 10 h. Then, the obtained solution was filtered, washed, and dried to obtain the charring-foaming 94 agent [20]. The structure of CFA has already been demonstrated [21].      A mixed aqueous solution containing 0.5 mol ethylenediamine and 1.0 mol sodium hydroxide was added into intermediate I. The reaction temperature was kept at 40-50 • C for over 10 h. After cooling, the product was filtered, washed, and dried to obtain intermediate II (yield: 99.4%).
A 0.5 mol amount of intermediate II and a mixed solution containing 0.5 mol ethylenediamine and 1.0 mol sodium hydroxide in 500 mL water were added into a 2000 mL four-neck flask provided with a stirrer, a thermometer, and a reflux condenser, and the reacting solution was kept under reflux for 10 h. Then, the obtained solution was filtered, washed, and dried to obtain the charring-foaming agent [20]. The structure of CFA has already been demonstrated [21].

Preparation of Composites
The mass of the flame retardant accounted for 20%. The composites' ingredients are listed in Table 1. The composites were prepared via the melt-blending method at 140 • C in an RM-200A torque rheometer (Harbin University of Science and Technology, Harbin, China) for 6 min with a rotor speed of 60 rpm. The composites were called S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, S-NdMgAl/EVA, APP/CFA/EVA, S-MgAl/APP/CFA/EVA, S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA, respectively. All composites were treated in an oven (Beijing Xianghu Science and Development Company, Beijing, China) at 60 • C for 4 h before testing to remove adsorbed water from the surface.

Characterizations
X-ray diffraction (XRD) was carried out on a D/MAX 2200VPC diffractometer (Rigaku, Tokyo, Japan) with Cu Kα. Data were collected at the rate of 4 • /min and step 0.02 • at 40 kV and 30 mA. Fourier transform infrared (FTIR) spectra of LDH were collected using a Nicolet FTIR spectrometer (Thermo Nicolet, Madison, WI, USA) (KBr pellet method, 4 cm −1 resolution, number of scans was 32). The samples were scanned in the 2θ range of 10 • to 70 • . Scanning electron microscope (SEM) observations were carried out on an FEI-Sirion (FEI, Eindhoven, Holland) with a field emission of 20 kV. The transmission electron microscopy (TEM) was carried out with a HITACHI 1H-7650 (Hitachi, Tokyo, Japan) operating at the accelerating voltage of 100 kV. Thermal analysis was carried out with a thermogravimetric analyzer (TGA pyris 1, PerkinElmer Co, Ltd., Waltham, MA, USA) using a constant heating rate of 10 • C/min under a pure nitrogen atmosphere from 50-700 • C. The weight of the samples was kept within 3 and 4 mg. Vertical burning ratings (UL-94) were measured on an instrument (CZF-2, Jiangning Analysis Instrument Factory, Jiangning, China) with sample dimensions of 125 × 12.5 × 1.6 mm 3 according to ISO 1210-1992. All cone calorimeter tests were carried out in a cone calorimeter (FTT0007, Hongkong Universal Technology Co, Hongkong, China) at an incident heat flux of 50 kW/m 2 according to the ISO 5660-1 standard. The polymer sample (100 × 100 × 5 mm 3 ) was placed horizontally on a balance holder. Before testing, aluminum foil was used to prevent ignition of the sides for the samples. The tensile strength and elongation at break of composites were evaluated according to the national standard GB/T 16421-1996 by using a material test machine (RGD-20A, Shenzhen Regear Instrument Cooperation, Shenzhen, China). A 1 mm thick sample was placed in a special mold to make a dumbbell-shaped standard sample before testing the tensile strength and elongation at break.  (113), which were indexed in a hexagonal lattice with an R3m rhombohedral space group symmetry. However, compared with S-MgAl, the diffraction peak resolutions of (110) and (113) of the other three LDHs were not high; this was attributed to the addition of REEs that reduced the orderly arrangement of anions within the brucite-type layers [22]. The basal spacings (d basal ) of S-MgAl, S-LaMgAl, S-CeMgAl, and S-NdMgAl were 7.65 Å, 7.84 Å, 7.66 Å, and 7.84 Å calculated from the (003) diffraction, respectively. These values indicated that the interlayer anions were CO 3 2− [23] besides OH − , without the coexistence of NO 3 − or C 18 H 35 O 2− [24]. Figure 3b presents the FTIR spectra of the stearate surface-modified LDHs. All samples showed the band at 3455 cm −1 , which was attributed to a large amount of OH − between LDH layers. The band at 1631 cm −1 illustrated the bending vibration of -OH from crystal water. The appearance of this band was due to the electrostatic action and the hydrogen-bonding action between -OH on the layer board of the LDH and the anion or water molecule. The absorption peak at 1377 cm −1 was assigned to the stretching vibrations of C-O in CO 3 2− . Two bands at 661 cm −1 and 861 cm −1 were the in-plane bending vibration of C-O in CO 3 2− . The appearance of the band at 444 cm −1 was attributed to the lattice vibration of the metal -O. The band at 1489 cm −1 was due to the bridging-two-tooth complexation of CO 3 2− -CO 3 2− . The bands at 2923 cm −1 and 2854 cm −1 belonged to -CH 3 and -CHgroups in sodium stearate [25]. According to the results of XRD and FTIR, the interlayer anions of synthetic samples were CO 3 2− and sodium stearate absorbed on the surface of the LDH.   162 Figure 5 shows the TEM images of the four modified LDHs. All LDH samples showed the flake   Figure 4 presents the SEM images of the four modified LDHs. The surfaces of all LDHs were uniformly coated by the stearate. The morphology results also confirmed that the stearate was uptaken on the surface of the LDHs, which was consistent with XRD and FTIR analysis.      Figure 5 shows the TEM images of the four modified LDHs. All LDH samples showed the flake particle morphology. Furthermore, the four modified LDHs possessed the uniform and nanoscale crystal sizes. Moreover, their TEM images appeared aggregated, which was attributed to the effect of stearate on the surface of the LDH.

Thermal Analysis of Pure EVA and Its Composites
The thermogravimetric analysis (TGA) and differential thermogravimetry (DTG) curves of pure EVA and its composites are shown in Figure 6. From Figure 6a, it can be seen that pure EVA underwent two stages of decomposition, whereas all composites underwent three stages of decomposition. This could be attributed to the removal of the interlayer water molecules, dehydroxylation, and decomposition of the interlayer anions at 120-309 • C [26,27].  Table 2. The incorporation of a flame retardant changed the thermal degradation behavior to allow composites to have a lower T 5% in comparison with pure EVA; meanwhile, this promoted the formation of more char residue, and the F m for composites was thus higher than that of pure EVA. However, the T 5% of S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA was significantly higher than that of S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, S-NdMgAl/EVA, APP/CFA/EVA, and S-MgAl/APP/CFA/EVA. This fact indicates that the combination of APP, CFA, and LDH containing REEs can allow composites to achieve better thermal stability. From the DTG curves (Figure 6c,d), the maximum thermal degradation temperature of pure EVA was 437 • C, and the temperature at maximum weight loss rate (T peak ) of all composites increased by more than 37 • C compared with that of pure EVA. This result confirmed that composites can achieve better thermal stability with the addition of flame retardant.

Flame Retardancy of EVA and its Composites
From Table 1

203
Before testing, aluminum foil was used to prevent ignition of the sides of the sample. The key data 204 are listed in Table 2.

205
It is evident that the time to ignition (TTI) of composites was reduced compared with that of 206 pure EVA. This was attributed to the decomposition of EVA, which led to an earlier evolution of 207 volatiles, mainly acetic acid. The peak value of HRR (pk-HRR) of pure EVA was 1119.7 KW/m 2 .

208
Compared with pure EVA, when incorporating LDH into the EVA matrix, the pk-HRR of S- The THR results demonstrated that the composite flame retardants reduced the spread of heat [29].

218
In a fire, the deadliest factor is not the heat but the smoke and toxic gases produced, such as CO 219 [30]. The peak values of SPR (pk-SPR) and peak values of COP (pk-COP) for pure EVA, S-MgAl/EVA,   Table 1, pure EVA, S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, and S-NdMgAl/EVA produced many drippings with flame during combustion. Therefore, these composites having a thickness of 1.6 mm failed to pass the UL-94 tests. In contrast, S-MgAl/APP/CFA/EVA, S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA successfully passed UL-94 V-0 rating and produced no dripping. This could be attributed to the common flame-retardant effect between APP, CFA, and LDH containing REEs.
Cone calorimetry tests can simulate a real fire scenario [28]. Therefore, they were used to evaluate the materials' flame retardancy. The curves of the heat release rate (HRR), total heat release (THR), smoke production rate (SPR), and production rate of CO (COP) are shown in Figures 7 and 8. Before testing, aluminum foil was used to prevent ignition of the sides of the sample. The key data are listed in Table 2.
It is evident that the time to ignition (TTI) of composites was reduced compared with that of pure EVA. This was attributed to the decomposition of EVA, which led to an earlier evolution of volatiles, mainly acetic acid. The peak value of HRR (pk-HRR) of pure EVA was 1119.7 KW/m 2 . Compared with pure EVA, when incorporating LDH into the EVA matrix, the pk-HRR of S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, and S-NdMgAl/EVA was reduced to 617.9, 523.0, 521.4, and 404.8 KW/m 2 , and decreased by 44.8%, 53.3%, 53.4%, and 63.8%, respectively. All LDH containing REEs obviously reduced the fire strength in comparison with S-MgAl. The pk-HRR of APP/CFA/EVA and S-MgAl/APP/CFA/EVA was 249.0 KW/m 2 and 225.0 KW/m 2 , compared with that of pure EVA, and the pk-HRR decreased by 77.8% and 79.9%, respectively. However, it is worth noting that the pk-HRR of S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA decreased to 193.7, 198.2, and 202.8 KW/m 2 , respectively, and were all reduced by more than 82%. Moreover, the THR curves of all composites showed a decrease compared with that of pure EVA. The THR results demonstrated that the composite flame retardants reduced the spread of heat [29].

226
The results showed that the addition of REEs allowed composites to achieve a better flame retardancy, 227 smoke release, and CO suppression. The char residues of composites were higher than those of pure 228 EVA, which all increased more than three times. The increases in char residues were one of the main 229 reasons of the improvement noted for HRR, THR, SPR, and COP.

226
The results showed that the addition of REEs allowed composites to achieve a better flame retardancy, 227 smoke release, and CO suppression. The char residues of composites were higher than those of pure 228 EVA, which all increased more than three times. The increases in char residues were one of the main 229 reasons of the improvement noted for HRR, THR, SPR, and COP.

236
The average mass loss rate (AMLR) values of pure EVA and its composites are shown in Table In a fire, the deadliest factor is not the heat but the smoke and toxic gases produced, such as CO [30]. The peak values of SPR (pk-SPR) and peak values of COP (pk-COP) for pure EVA, S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, S-NdMgAl/EVA, APP/CFA/EVA, S-MgAl/APP/CFA/EVA, S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA are listed in Table 2. It can be seen that S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA showed lower values of pk-HRR, THR, pk-SPR, and pk-COP. Furthermore, the pk-HRR, pk-SPR and pk-COP of S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA were lower than those reported for APP/CFA/LDH/EVA without REEs [13]. The results showed that the addition of REEs allowed composites to achieve a better flame retardancy, smoke release, and CO suppression. The char residues of composites were higher than those of pure EVA, which all increased more than three times. The increases in char residues were one of the main reasons of the improvement noted for HRR, THR, SPR, and COP.
The average mass loss rate (AMLR) values of pure EVA and its composites are shown in Table 2. The AMLR values decreased by 43%, 33%, 53%, 50%, 71%, 73%, 80%, 82%, and 78% compared with that of pure EVA, respectively. The results of AMLR showed the condensed-phase fire retardant action of APP, CFA, and LDH containing REEs in the composites.

Digital Photographs of Residues after Cone Calorimeter Tests
The char layer was considered an important influence on flame retardancy. In order to know the level of flame retardancy, the digital photographs of residues after the cone calorimeter tests are shown in Figure 9. Pure EVA had almost no residue remaining. The char layers of S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, and S-NdMgAl/EVA were fragmentary and not continuous. The char layer of APP/CFA/EVA was continuous, but the char layer was loose and had some small holes on the surface. Finally, the char layer of S-MgAl/APP/CFA/EVA was not continuous. Relatively speaking, the char layers of S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFAEVA were continuous and compact. The more compact and uniform char layers could reduce the volatile release and better limit the heat and gas transfer [31]. Hence, S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA achieved a better flame retardancy.

Morphological Analysis of Cone Calorimetric Residues
The SEM images of the cone calorimetric residues are shown in Figure 10.

Mechanical Property Analysis of Pure EVA and Its Composites
Before testing the tensile strength and elongation at break for all samples, a 1 mm thick sample was placed in a special mold to make a dumbbell-shaped standard sample. The tests were conducted at the speed of 50 mm/min and gauge was 25 mm. All data points regarding both the tensile strength and the elongation at break were the mean values of 10 test data. Further, the 10 test data were within the error of the measurements. The tensile strength and elongation at break of pure EVA and its composites are presented in Figure 11.
As is well known, the incorporation of a flame retardant can result in changes to the mechanical properties for a polymer matrix (as shown in Figure 11). Due to the addition of 20% flame retardants, compared with pure EVA, the tensile strength values of S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, S-NdMgAl/EVA, APP/CFA/EVA, S-MgAl/APP/CFA/EVA, S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA decreased significantly. However, it should be pointed out that, compared with S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, S-NdMgAl/EVA, APP/CFA/EVA, and S-MgAl/APP/CFA/EVA, composites containing REEs (S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA) still retained relatively good strength. As Figure 11b shows, S-MgAl/APP/CFA/EVA obtained the highest ductility. With the incorporation of REEs, the elongation at break of S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA decreased. Thus, REEs reduced the compatibility between APP, CFA, and LDH. However, the elongation at break of S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA was higher than that of S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, and S-NdMgAl/EVA. This confirmed that the combination between APP, CFA, and LDH containing REEs not only allowed composites to maintain favorable mechanical properties but also to retain the required flame retardancy and smoke suppression.
than that of S-MgAl/EVA, S-LaMgAl/EVA, S-CeMgAl/EVA, and S-NdMgAl/EVA. This confirmed 294 that the combination between APP, CFA, and LDH containing REEs not only allowed composites to 295 maintain favorable mechanical properties but also to retain the required flame retardancy and smoke

301
The combination of XRD, FTIR, and SEM results indicated that stearate was grafted onto the 302 surfaces of the LDHs. Compared with EVA, the pk-HRR, THR, pk-SPR, and pk-COP values for S-

303
LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA all decreased, 304 especially the pk-HRR that decreased by more than 82%. This was attributed to the combination of

Conclusions
The combination of XRD, FTIR, and SEM results indicated that stearate was grafted onto the surfaces of the LDHs. Compared with EVA, the pk-HRR, THR, pk-SPR, and pk-COP values for S-LaMgAl/APP/CFA/EVA, S-CeMgAl/APP/CFA/EVA, and S-NdMgAl/APP/CFA/EVA all decreased, especially the pk-HRR that decreased by more than 82%. This was attributed to the combination of APP, CFA, and LDH containing REEs which formed compact char layers and released more flame-retardant gases to improve the flame retardancy and smoke suppression of the EVA matrix. From the SEM images of residues, the values of AMLR indicated the condensed-phase fire retardant action in the composites. Meanwhile, after incorporating the composite flame retardant, the mechanical properties were still well maintained. Hence, the formed novel flame retardants are promising additive agents in terms of flame retardancy.