## Introduction

Greater use of conventional petroleum based commodity resins in various applications have led to inevitable rise of its ecological footprint. Various new eco-friendly biodegradable polymers were developed in order to counteract this effect. With its biodegradability, low emission of greenhouse gas and low production energy PLA currently stands as one of popular alternatives for conventional polymer materials1. Due to its high degree of transparency, good mechanical properties, low toxicity and ability to process using equipment as well as relatively low cost and large production volume it shows high potential for packaging, household and biomedical applications. Nevertheless, other potential applications such as in electronics or automotive industry require PLA grades with high impact strength, better processing ability and improved flame retardant behaviour2. In order to meet that demand novel methods of modification need to be developed and applied.

A common way of improving flame retardancy of polymer is through combining its matrix with flame retardant filler. In the past improved flame resistance was achieved by introduction of halogenated flame retardants (HFR). Nowadays, it faces severe restrictions in Europe and North America due to the release of toxic substances and large amounts of smoke during combustion3. Instead, halogen-free flame retardant formulations function more commonly now as an environmentally friendly alternatives for HFR. These include intumescent flame retardants4,5, phosphorus- and nitrogen-containing micro- and nanoparticles6,7,8, inorganic substances and silica derivatives9 which can be used either individually or in combination10 to achieve optimal flame retardant performance.

Polymer/layered inorganic materials composites have high potential for improving thermal properties of polymers. Molybdenum disulfide is a member of a family of transition metal dichalcogenides (TDM). Structurally its crystals are characterized as hexagonal layered configurations. Atoms in the layer are bonded with strong covalent bonding, while layers are packed together to form a sandwich structure with weak Van der Waals forces similarly to graphite or boron nitride11. With their unique electrical, optical, thermal and mechanical properties MoS2 nanosheets can be potentially used for application as fillers for improving properties of polymers. Being a representative of layered inorganic materials MoS2 is expected to disperse and exfoliate in polymers, which results in the physical barrier effects that inhibits the diffusion of heat and gaseous decomposition products12,13. Molybdenum, a transition metal element, promotes the formation of charred layer during the combustion which acts as a physical barrier that slows down the heat and mass transfer11. In order to achieve high-performance homogenous dispersion of MoS2 nanosheets in the polymer hosts and proper interfacial interactions need to be established14. Similarly to metal oxide/graphene hybrid materials addition of metal oxide/MoS2 nanoparticles might prevent the aggregation of MoS2 flakes during preparation of polymer nanocomposites and result in improved flame retardant performance15,16. In addition they can also lead to improved char generation due to catalytic activity of metal oxide as well as suppressed smoke production and reduced toxicity due to catalytic conversion of carbon oxide (CO)17.

Another possible alternative to the use of conventional flame retardants are carbon nanotubes (CNT). These can be introduced into the polymer matrix in pristine form of a small diameter (1–2 nm) single-walled carbon nanotubes or a large diameter (10–100 nm) multi-walled carbon nanotubes (SWCNT and MWCNT, respectively)18,19,20,21,22,23. In addition to this functionalization of CNT can also be carried out in order to significantly improve the flame retardant performance17. This can be performed in three different ways. Surface modification by coupling agents allows for enhancement of dispersion state of CNT24,25,26. For example, addition of 9 wt% content of CNT functionalized with vinyltriethoxysilane into epoxy composite allowed for increase of limiting oxygen index (LOI) from 22 to 27% and improvement of UL-94 rating from V-1 to V-025. This also resulted in increase of char yield at 750 °C by 46.94%. Char yield can be also enhanced through covalent linkage of organic flame retardants to CNT27,28,29,30 following surface treatment. Functionalization of CNT can facilitate their dispersion within the polymer matrix and enhance interfacial adhesion between the CNT and the polymer30. Formation of genuine composites can be confirmed by an increase in the Young’s modulus and flame resistance of compositions containing pristine CNT28. Finally, hybridization of CNT by inorganic particles can allow for superior flame retardant properties31,32,33. In addition to that other key parameters, such as thermal stability and dielectric properties, can be also enhanced31. Generally speaking, flame retardant actions of CNT/polymer composites involve the condensed phase action. Char layer formed on the entire sample surface acts as insulation layer that reduces the amount of volatiles escaping to the flame. The formation of continuous layer is obtained by formation of three-dimensional network structure when the content of CNT reaches a threshold value17. Good quality char plays major role in reduction of peak heat release rate (pHRR)34,35.

During the scope of presented research few-layered MoS2 nanosheets functionalized with MxOy nanoparticles served as catalysts for growth of CNT in the CVD process were prepared. The obtained nanomaterials were used as flame retarding agents in PLA composites. Detailed description of synthesis, as well as full characteristics and analysis of properties of obtained PLA-based composites were provided. Additionally, a thorough analysis of thermal stability, fire performance and thermal conductivity of these materials was also performed and discussed in details.

## Methods

### Materials

Bulk MoS2 (powder), N-Methyl-2-pyrrolidone (NMP) (anhydrous, 99.5%), cobalt(II) acetate tetrahydrate (99%), iron(II) acetate (95%) and nickel(II) acetate tetrahydrate (98%) were purchased from Merck. PLA was obtained from Goodfellow. Hydrogen peroxide (30%) and ethanol (96%) were purchased from Chempur. Gaseous nitrogen and ethylene were purchased from Messer and Air Liquide, respectively.

### Preparation of few-layered MoS2

1 g of bulk MoS2 powder was transferred into a 250 mL flat-bottomed beaker filled with solution of 95 mL of NMP and 5 mL of hydrogen peroxide. This was followed by 30 minutes of continuous sonication in an ultrasonic washer, after which solution was transferred to a 250 mL round bottomed flask, plugged to reflux and consciously stirred at 360 rpm and 35 °C for 24 h. Final dispersion was centrifuged four times at 10000 rpm for 20 minutes and washed with ethanol.

### Preparation of MoS2/MxOy/CNT

In order to obtain MoS2/MxOy/CNT modified nanomaterials following procedure was applied. First, 150 mg of few-layered MoS2 and 150 mg of respective source of metal oxide (cobalt(II) acetate tetrahydrate, iron(II) acetate or nickel(II) acetate tetrahydrate) were dispersed in ethanol and sonicated in an ultrasonic washer for 2 h. Next, dispersions were stirred for 48 h, after which they were dried under high vacuum at 440 °C for 3 h. This resulted respectively in MoS2/Co2O3, MoS2/Fe2O3 and MoS2/Ni2O3 modified nanomaterials. Next CVD was performed in tube furnace under constant 100 mLmin−1 flow of nitrogen at 850 °C with a 15 minute long, 60 mLmin−1 flow of ethylene as a source of carbon.

### Extrusion of MoS2/MxOy/CNT modified PLA composites

PLA was utilized as a polymer matrix. Four composites were prepared, containing addition of few-layered MoS2, MoS2/Co2O3/CNT, MoS2/Fe2O3/CNT or MoS2/Ni2O3/CNT, respectively. For each composite three samples containing different amounts of nanomaterials additives were prepared − 0.5 wt%, 1 wt% and 2 wt%, respectively. Following 12 h of drying the nanomaterial ware blended with PLA using twin-screw extruder (Zamak Mercator EHP 2 × 12). For reference a sample of pristine PLA was also extruded.

### Characterization

Morphology of nanomaterials obtained during each stage was analyzed using transmission electron microscopy (TEM) (Tecnai F20, FEI) with 200 kV accelerating voltage. Raman analysis was performed in microscope mode (InVia, Renishaw) with a 785 nm laser in ambient air. Number of layers in few-layered MoS2 samples was determined using atomic force microscopy (AFM) (MultiMode 8, Bruker).

For the obtained composites and pristine PLA following analyses were performed. Thermogravimetric analysis (TGA) was performed using thermal analyzer (SDT Q600, TA Instruments) under airflow of 100 mLmin−1. In each case individual sample (ca. 5 mg in an alumina crucible) was heated from room temperature to 800 °C at a linear heating rate of 10 °Cmin−1. In addition to this, gaseous products from the heating process were analyzed in situ with a mass spectrometer (ThermoStar, Pfeiffer Vacuum) coupled with the TGA, under 100 mLmin−1 argon flow. Microscale combustion calorimetry (MCC) was employed for measurement of flame retardancy using FAA Micro Calorimeter, from FTT. This allowed for determination of pHHR, heat release capacity (HRC) and total heat release (THR) from 2 mg specimens. Thermal conductivity of the obtained samples was measured using laser flash apparatus (XFA 300, Linseis). Prior to this measurement the samples were spray coated with thin layer of graphite in order to facilitate the absorption of laser at the surface.

## Results and Discussion

Successful preparation of few-layered MoS2 samples was confirmed using TEM and AFM (Fig. 1). Number of MoS2 layers was determined through analysis of high profile with AFM (Fig. 1C). The flakes were typically 5 nm in height, which corresponded to approximately 7 layers of MoS2 (assuming average thickness of single layer ca. 0.7 nm36 (Dependence between lateral size and aspect ratio in Supplementary information)). This was verified later with Raman spectroscopy as the intensity of $${E}_{2g}^{1}$$ and $${A}_{1g}$$ peaks changes between bulk MoS2 and few-layered MoS2. With the increased number of layers the frequency of $${E}_{2g}^{1}$$ (~382 cm−1 for bulk MoS2) decreases while that of the $${A}_{1g}$$ (~407 cm−1 for bulk MoS2) increases. This effect is caused by the interlayer van der Waals force in MoS2 suppressing the atom vibration, which results in higher force constants36,37.

Dispersion of metal oxide nanoparticles on the surface of MoS2 was determined from collected high-resolution TEM images presented in Fig. 2. Metal oxide nanoparticles appeared to be well dispersed on the surface and deposited homogenously. The nanoparticles size was measured and ranged between 5 and 25 nm for all metal oxide nanoparticles. Presence of CNT on samples from the CVD process was confirmed with TEM (Fig. 3). As observed, internal diameter of nanotubes matched the diameter of specific metal oxide nanoparticles. Hallow core and an open end were visible and the surface appeared to be smooth.

### Thermal stability

CO content analysis performed with use of mass spectrometry during the TGA pyrolysis cycle in argon atmosphere revealed promising results. In each case the amount of CO released was lower in comparison to that observed for pristine PLA (Fig. 5). For PLA modified only with addition of few-layered MoS2 (Fig. 5A) the reduction in CO release ranged from 38 up to 70%, with the highest value recorded for sample containing 1 wt% of nanomaterials which also corresponded to the largest recorded charred residue value. Previously we observed that introduction of few-layered MoS2 into the poly(vinyl alcohol) (PVA) composite resulted in significant decrease (up to 94%) in permeability of hydrogen gas in comparison to pristine PVA39. This effect was attributed to the incorporation of impermeable MoS2 nanosheets with high aspect ratio. As a result a high number of tortuous pathways was created which restricted the diffusion of penetrants, such as H2 or CO39,40. Reduction of CO emission during pyrolysis was further enhanced in case of MoS2/MxOy/CNT composites. Highest reduction of CO emission in respect to pristine PLA was observed for all samples containing MoS2/Fe2O3/CNT (Fig. 5C), and it was always above 90%. This was attributed to a well developed and thermally stable char barrier, which protected the molten zone from the burning zone, as reported by Attia et al.41 in case of polystyrene composites containing addition of MWCNT.