Study on the flammability, thermal stability and diffusivity of polyethylene nanocomposites containing few layered tungsten disulfide (WS2) functionalized with metal oxides

In this work, exfoliated tungsten disulfide (WS2) functionalized with metal oxides as a filler of polyethylene (PE) was used. An efficient exfoliation procedure resulted in the synthesis of 7–9 layered flakes of WS2. Flakes of exfoliated WS2 were functionalized by iron oxide and nickel oxide nanoparticles, respectively. The nanomaterials were mixed with polyethylene by extrusion. Methods such as Transmission Electron Microscopy (TEM), Atomic Force Microscopy (AFM), X-Ray Diffraction (XRD) or Thermogravimetric Analysis (TGA) were used to characterize the materials. Flame retardant properties were investigated by microcalorimetry. Comparing the obtained values of heat released during combustion, it can be observed that the addition of fillers reduces flammability significantly compared to neat polyethylene. It is revealed that this composite can provide a certain physical barrier and inhibit the diffusion of heat and gaseous products during combustion. Thermogravimetric analysis of composites showed increased thermal stability with addition of nanofillers and reduction of carbon monoxide generation in the whole range of the nanofiller addition (from 0.5 to 2 wt% in PE). Results suggested that the composite with Ni2O3 could endow the best flame retardance for PE. The peak heat release rate of this sample with 2 wt% nanofiller was reduced to 792 W g−1 (1216 W g−1 for PE), and the total heat release was decreased to 39 kJ g−1 (47 kJ g−1 for PE). A very significant increase in thermal conductivity for all composites was observed as well.


Introduction
The discovery of graphene and its superior properties have focused research interest on other two-dimensional (2D) groups of materials like hexagonal boron nitride (h-BN) and transition metal dichalcogenides (TMDs). 1 A monolayer of TMD, the general chemical formula of which is MX 2 , consists of a metal atom (M ¼ Mo, W) sandwiched between two chalcogenide atoms (X ¼ S, Se) 3 and it has a thickness between 6 and 7Å. 2 Depending on the coordination number and oxidation state of the metal atoms, TMDs can be metallic, semimetallic or semiconducting. 4 These materials exhibit a layered structure 5 with weak interlayer van der Waals forces which allow easy exfoliation. 1,3 Unlike graphene, the band gap of these materials, which due to a reduced number of sheets, changes from an indirect band gap to a direct band gap. 1,6-9 Therefore, these materials can have many new applications e.g. in electronics, optoelectronics, 3,8 catalysis, energy storage and sensing. 3,10 Tungsten disulde (WS 2 ) is one of the most popular compound of semiconducting TMS's. The band gap of monolayer of WS 2 is 2.1 eV, while in bulk is 1.3 eV and that results in enhancement of photoluminescence. 11,12 This compound exhibit trigonal prismatic structure. Mechanically exfoliated atomically thin sheets of WS 2 were shown to exhibit high inplane carrier mobility and electrostatic modulation of conductance similar to MoS 2 . 13 Many techniques have been reported to obtain atomically thin layers WS 2 like mechanical exfoliation, 13 chemical exfoliation 4,14 and chemical vapor deposition. 15 WS 2 nanosheets have broad applications in optoelectronics. 12,[16][17][18][19][20] Recently this group of materials has attracted increased attention in the eld of nanocomposites llers, due their graphene-like properties such as high thermal and mechanical properties. 21,22 Polymeric materials are widely used in the most important industries. However, they are known for high re risk and most of them combust with emission toxic gases. It is important to modify these materials to reduce their ammability. There are three typical strategies to achieve that: use of inherently ame retardant polymers, 21 ame retardants, 22 and surface treatment/coating. 23 Usually small amount of nanoller improves thermal properties of composites. WS 2 , as a typical layered inorganic material, is expected to disperse and exfoliate in polymers and it results in the physical barrier formation which inhibits the diffusion of heat and the decomposition of polymer products. So it is reasonable that WS 2 may improve the thermal stability, mechanical properties and re resistance of polymer composites. 24 To the best of our knowledge there is a gap in the scientic reports on the potential of few layered WS2 functionalized with metal oxides as a nanoller of polymers used for ame retardancy.
Polyethylene is one of the most common polymer used in everyday object in our houses. However, its ammability and toxicity during the re are the key motivations to perform the proposed study.
In this work we present a technique of liquid exfoliation of tungsten disulde nanosheets and its functionalization with metal oxides (nickel and iron). We use these materials as llers in polyethylene nanocomposites and investigate their ame retardants properties by pyrolysis combustion ow calorimeter (PCFC) and thermal conductivity by xenon ash method.

Characterization
The structures of materials before and aer exfoliation/ functionalization were studied by Transmission Electron Microscopy (Tecnai F20-based at 200 kV accelerating voltage). Atomic Force Microscopy (AFM NTEGRA Aura (NT-MTD) microscope) provide information about thickness and number of layers of WS 2 . X-ray diffraction (XRD) Philips X'Pert PRO X-ray diffractometer witch Cu Ka radiation was employed to identify phase identication of the samples. Thermogravimetric analysis (TGA) was carried out using a SDT Q 6000 thermoanalyzer instrument (TA Instruments Inc.) under air ow of 100 mL min À1 . The samples with mass of about 5.0 mg were heated from room temperature to 900 C at a linear heating rate of 10 C min À1 . During heating the sample in thermobalance, the generated gases were analyzed in situ by Quadrupole Mass Spectrometer QMS 422. All PE and its composites were fabricated using a twin screw extruder (Zamak Mercator EHP 2x12).  Micro Calorimeter (FAA MICRO Calorimeter) was used to investigate the ammability properties of PE nanocomposites. Samples of about 2.0 mg were heated in air atmosphere (80% of nitrogen and 20% of oxygen) at a constant heating rate of 1 C s À1 from room temperature to 700 C. This method allowed to determinate parameters such as heat release rate (W g À1 ), heat release capacity (J g À1 K À1 ), total heat release (kJ g À1 ). The thermal diffusivity of the composites was measured via xenon ash technique using Linseis XFA 300 laser ash apparatus.

Exfoliation of WS 2
Few-layered WS 2 akes were synthesized by liquid exfoliation of bulk WS 2 . First, 1 g of bulk WS 2 and cetyltrimethylammonium bromide (0.5 g) were mixed, dissolved in distilled water and sonicated for 3 h. Then, the mixture was centrifuged at 10 000 rpm for 0.5 h. Aer that, the solution was washed few times with distilled water and dried 24 h in 60 C.

Functionalization of WS 2
Two samples of WS 2 modied by metal oxide nanoparticles (WS 2 _Ni 2 O 3 and WS 2 _Fe 2 O 3 , respectively) were prepared according to the following procedure: 0.5 g of WS 2 and 0.5 g nickel(II) acetate tetrahydrate (product referred to as WS 2 _Ni 2 O 3 ) and iron(II) acetate (product referred to as WS 2 _Fe 2 O 3 ), respectively, were dispersed in 250 mL of ethanol and sonicated for 3 h. Aerwards, the mixtures were stirred for next 24 h. Finally, the samples were dried in high vacuum at 440 C for 10 min.

Preparation of nanocomposites
Polyethylene (PE) was used as the polymer matrix. The content of MoS 2 , MoS 2 _Ni 2 O 3 and MoS 2 _Fe 2 O 3, were 0.5%, 2%, 3%, respectively. The composites were prepared by extrusion molding at a temperature of 120 C.

Results and discussion
The morphologies of WS 2 aer exfoliation investigated by TEM (Fig. 1A) and AFM (Fig. 1B) microscopy. Fig. 1 shows that WS 2 was successfully exfoliated from bulk to few nanosheets. Tapping-mode of AFM was used to determinate the size and thickness of exfoliated WS 2 . The AFM samples were prepared by dropping a few drops of exfoliated WS 2 dissolved in ethanol on silica wafer and le to evaporate the solvent. AFM analysis shows that exfoliated WS 2 had diameter $0.74-1.4 mm and thickness about 4.6-5.9 nm which corresponds to 7-9 layers of WS 2 . Aer exfoliation the product was functionalized with nickel and iron compounds. The morphology and structure was characterized by TEM. Fig. 2A and C shows, that metal particles have covered the whole surface of WS 2 , what proves very good dispersion of metal oxides on WS 2 surface. Particle size distribution of iron oxide (Fig. 2D) and nickel oxide (Fig. 2B) was estimated basing on $100 of particles pictured in TEM. Detailed analysis reveals very uniform diameters of samples. Particle size distribution of the iron compound and nickel compound is $35 and $25 nm, respectively.
The composition of WS 2 _Fe 2 O 3 and WS 2 _Ni 2 O 3 were examined by powder X-ray diffraction, as it is demonstrated in Fig. 3. In both patterns one intense pick at value 2q 14 is present. This peak is assigned to WS 2 . Few peaks with lower intensity at 2q value 29 , 39.5 , 44 and 60 are also associated with pristine WS 2 . Fig. 3A shows XRD pattern of WS 2 _Fe 2 O 3 . Relatively strong peak is present at 50 C. This signal belongs to iron(III) oxide. Peaks at 33.5 , 36 , 56 , 58 , 70 , 72 and 75 are consistent with the data for the a-Fe 2 O 3 . Fig. 3B presents XRD pattern of WS 2 _Ni 2 O 3 . Besides peaks which correspond to WS 2 , there are peak, as well a 2q value 34 and 60 assigned to NiO 2 . Peak at 49.5 is associated with presence of Ni 2 O 3 phase. There are also peaks of metallic nickel (peaks at 2q values 58.5 , 72 and 75 ). The above analysis proves that WS 2 was successfully functionalized with nickel and iron oxides, respectively.
TGA was performed to investigate the general thermal stability of PE and WS 2 nanocomposites. Fig. 4 shows thermograms of PE and PE_WS 2 nanocomposites. The corresponding TGA data are presented in Table 1. The temperature where the weight loss is 10 wt% is denoted as T 0.1 and the temperature there half of the sample was loss is T 0.5 . Char yield is the weight Paper percent obtained at the end of pyrolysis. Thermal stability of composites has decreased for most samples. The most signicant drop was observed for composite PE_WS 2 _Ni 2 O 3 _2% for which the temperatures of T 0.1 and T 0.5 has increased by 27 C and 11 C, respectively, compared to neat PE. The position of PE and WS 2 _Fe 2 O 3 has not satisfactory inuenced thermal stability of the nanocomposites. The lowest decrease of T 0.1 and T 0.5 was obtained for the composites with 2 wt% (11 C lower than in pristine PE). T 0.1 and T 0.5 has increase by 8 C and 16 C, respectively, in the PE_WS 2 with 2 wt% of WS 2 . Additionally, char yield increased of $7% in composite PE_WS 2 _0.5% composite with PE_WS 2 _Ni 2 O 3 consisting 2 wt% exhibit the best value and it is increased $51% compared with PE. For materials of PE_WS 2 _Fe 2 O 3 the best value was obtained for 1 wt% and the value of char yield has not changed. Char yield and carbon monoxide generation of nanocomposites with metal oxides functionalized few layered WS 2 are lower than the neat PE. We suppose that inorganic oxides assist in conversion of CO to CO 2 and they limit generation of soot. 25 During the thermogravimetric analysis, mass spectrometer was coupled and in situ gas analysis was carried out. The emission of toxic gases is considered as important parameter for ame-retardant materials. Mass spectra of all samples ( Fig. 5) exhibit very signicant peak at position of 28 amu, which corresponds to CO emission. The lowest value of carbon oxide emission was obtained for composite consisting PE_WS 2 _Ni 2 O 3 with 2 wt% compared to the polyethylene the value of emission decreases by 47%. Good efficiency of reduced CO emission was also obtained for PE_WS 2 _Fe 2 O 3 _2% ($44% compared to pristine polyethylene). For composites with exfoliated WS 2 the amount of ller has not signicantly inuenced the reduction CO emission. This value has decreased from 30% (for 2 wt%) to 35% (for 1 wt%) compared to polyethylene.
For assessment ammability properties of polyethylenebased nanocomposites microcalorimeter was used. The microscale combustion calorimeter (MCC) is a small-scale instrument that measures the heat release of a material by oxygen consumption calorimetry. Using this technique, the samples are exposed to a fast heating rate to mimic re-type conditions. During MCC measurement several parameters are obtained, such as total heat release (THR), heat release rate (HRR) and heat release capacity (HRC). These parameters are crucial for assessing the re risk of materials. The results of the MCC testing are summarized in Fig. 6 and Table 2.
The HRR curves of nanocomposites derived from MCC are plotted in Fig. 6. The pHRR value has been oen regarded as the most accurate indicator of ame. As it is shown, the addition of small amount of llers leads to signicant decrease of pHRR value in all composites. Compared to the PE, values of pHRR of composites with PE_WS 2 and PE_WS 2 _Fe 2 O 3 have decreased with decreasing loading of llers. The best improvement was obtained for composites consisting of 0.5 wt% of WS 2 _Fe 2 O 3 and it was about 33%. For composites consists PE_WS 2 pHRR has decreased signicantly compared with pristine PE. The largest decrease was observed for the sample containing 2 wt% of the ller. The value pHRR has decreased by 28%. Diez-Pascual et al. explains these improvements that WS 2 could behave like a mass transport barrier which prevent the escape of volatile products generated during the burning and also hinders access to the matrix. 26 The addition of llers increase the temperature of maximum value peak HRR, which indicates the improvement in ame retardancy of nanocomposites. The most interesting value was obtained for composite the PE_WS 2 _Ni 2 O 3 containing 2 wt% of metal oxide. The peak heat release rate of PE_WS 2 _Ni 2 O 3 was reduced to 792 W g À1 , and the total heat release was decreased to 39 kJ g À1 . For almost each composite the temperature of pHRR increased by at least 10 C. Interesting results was obtained also for PE_WS 2 _Ni 2 O 3 containing 1 wt% and 2 wt% the temperature has increased by 12 C.
As is known to all, incorporation of 2D layered nanollers usually increase the thermal stability of a polymer matrix due to the physical barrier effect which retards the diffusion of degradation products, gases and heat. MoS 2 nanosheets must present better physical barrier effects compared to pristine PE. A lot of attention is devoted to MoS 2 and its composite from PE, 27 PS, 28 PVA. 29 However, composites with WS 2 have much better re retardancy and thermal stability properties. Compared data with MoS 2 dispersed in the PE matrix, WS 2 shows a decrease of ame retardancy. For the composites contained 2% nanollers Ni 2 O 3 . The pHRR for composites with WS 2 and MoS 2 decreased $35% and 30%, respectively. The ame retardance analysis, it is reasonably speculated that independent WS 2 nanosheets in the PE matrix act as nanobarriers to restrain the permeation of heat and oxygen and inhibit the effusion of volatile toxic materials. In comparison to the composites containing carbon nanotubes as llers they did not exhibit ame retardant properties as good as WS 2 . Polyethylene with 2 wt% of CNT's decrease pHRR value about 24.5% compared to neat PE, but that nanocomposite reaches mu higher char yield. 30,31 To measure the thermal conductivity of composites xenon ash method was used. The measurement was carried out along the thickness direction of each sample. The samples were coated with a thin layer of graphite to facilitate the absorption of the laser light at the surface of the sample. Measurements were conducted in the vacuum (1.0 Â 10 À2 bar). Three laser shots were applied to each sample at a room temperature. The nal result for each parameter was presented as the average value of three partial measurements. The results are shown in the Table  3. Compared to the value of pristine polyethylene, thermal conductivity increased for all composites. The addition of small amount of ller brings the increase value of thermal conductivity about 100% for all the composites. Furthermore, the addition of 0.5 wt% of WS 2 _Fe 2 O 3 into PE exhibits the most signicant increase in thermal conductivity. Compared to neat PE the value has increased about $240%. The most signicant difference of value for composites containing WS 2 _Ni 2 O 3 was obtained for 2 wt% and it was 290% higher than for pure PE. The interesting value was obtained for composite PE_WS 2 _1%. In this case, the improvement of thermal conductivity was 230% compared to pristine PE.

Conclusion
In this study, the aqueous phase exfoliated WS 2 nanosheets were successfully functionalized with metal oxides (nickel and iron) and incorporated into polyethylene matrix by extruder melting. Adding small amount of layered nanollers improved the thermal stability and re resistance of composites signicantly. MCC, TGA and thermal conductivity measurements indicate that the reduction of ammability is dependent on the content of WS 2 llers. The best improvement in thermal degradation was obtained for PE_WS 2 _Ni 2 O 3 _2% for which the temperature T 0.1 and T 0.5 increased by 27 C and 11 C, respectively, compared to neat PE. Signicant decrease in value of peak heat release rate and CO emission was obtained for PE_WS 2 _Ni 2 O 3 _0.5% which was 33% and 44%, respectively. Thermal conductivity of this composite has increased by 144% compared to pristine polyethylene. The reduction of re hazard was attributed to the physical barrier effect of WS 2 . Therefore, it

Conflicts of interest
There are no conicts to declare.