Investigation of Thermophysical Properties of Zr-Based Metallic Glass-Polymer Composite

Composites based on Zr65Cu17.5Ni10Al7.5 metallic glass (MG) and polytetrafluoroethylene (PTFE) were prepared by ball milling. Different composites (30/70, 50/50 and 70/30) were produced. Samples for dynamic mechanical analysis and laser flash analysis were fabricated in the supercooled region of the metallic glass and viscous region of the polymer. Spark plasma sintering (SPS) was performed at the supercooled region for the metallic glass powder. Characteristics such as thermal, mechanical, and structural properties were studied. A formation of the Zr2Cu and Zr2Ni intermetallic was found in the metallic glass after SPS. A formation of the nanocrystalline Zr2Cu was found in composite samples. Dynamical mechanical analysis (DMA) was used to study the mechanical behavior of the material. It was concluded that the 70/30-MG/PTFE composite sample had better thermal conductivity than the other composite samples. The thermal conductivity of the metallic glass was the highest among the samples and it increased with the MG content in composites.


Introduction
Metallic glass (MG) has attracted attention worldwide due to its excellent properties. It is known for its excellent mechanical, electrical, and magnetic properties [1] and is mainly produced by rapid solidification techniques [2]. Due to its amorphous structure, it does not have crystalline defects and grain boundaries compared to crystalline materials [3]. This leads to better properties, such as high mechanical strength, wear and corrosion resistance [4]. The limitations of metallic glass are its brittle nature and the formation of shear bands during plastic deformation [5]. This limits its potential application. Polymers are ductile, with low strength and low density [6]. Composite materials have shown considerable improvement in mechanical, optical, electrical, and magnetic properties [7]. The combination of metallic glass and polymer can produce a lightweight composite with enhanced thermal and mechanical properties. Metallic glass production is cheap and can be produced by sputtering, melt-spinning techniques, powder metallurgy, liquid squat quenching, magnetron sputtering and pulsed laser quenching [8]. Zr-Cu-based metallic glass is inexpensive and has attracted attention due to its thermal stability, as well as its mechanical and corrosion properties [9]. Tang [10] et al. investigated the mechanical properties of Zr-based metallic glass and found that it has good plasticity with plastic strain. Zr 40 Ti 35 Ni 14 Nb 11 metallic glass can be used as bioimplant material [11]. Zr-Al-Fe-Cu metallic glass is known for its corrosive properties and can be used as a stent [12]. The addition of aluminum in Zr-Cu improves the mechanical properties and thermal stability of metallic glass [13][14][15]. Nickel increases the corrosion resistance, glass transition temperature and mechanical properties [16][17][18][19]. Zr-Cu-Ni-Al metallic glass is used in the average size of the MG particles was about 10 µm. The powdered metallic glass was added to the polymer for the ball milling process. Different composite samples of metallic glass/polymer ratios (mass.%) of 30/70, 50/50 and 70/30 were prepared. The mixing time for the composite samples was 30 min, each under a rotation speed of 300 rpm.
For the spark plasma sintering (SPS) measurement, the metallic glass powder was inserted inside graphite die with an inner diameter of 12.7 mm, an outer diameter of 30 mm, and a height of 30 mm. The sample was wrapped in a 0.2 mm thick graphite sheet. The sheet was used to prevent the sample from die contact. The powder was sintered using SPS [44] (Labox 650, Sinter Land inc, Nagaoka, Niigata Prefecture, Japan). The heating rate was 50 • C/min, with a maximum temperature of 400 • C, under a pressure of 50 MPa and 50 min duration.
X-ray diffraction (XRD) (Research and production enterprise "Bourevestnik", Saint Petersburg, Russian Federation) on a DRON Diffractometer, under CoKα radiation (2 angles: 10 to 120 • , step: 0.1 • , exposition time per step: 5s, beam size: 6-8 mm) was used to determine the phase and structural composition of samples. The accuracy of the phase composition was ±5%. The kinematic standard method and annealed powder use a standard to determine the percentage of amorphous phase [45].
A differential scanning calorimeter (DSC) (NETZSCH DSC 204 F1) (Netzsch Erich Netzsch GmbH & Co. Selb, Upper Franconia, Bavaria, Germany) was used for thermal analysis. The heating rate was 10 • C/min under an argon atmosphere. The sample mass was between 10 and 15 mg. The maximum temperature of the DSC was 600 • C to study the supercooled region and crystallization and 320 • C for the measurement of heat capacity of the samples. It was used to determine the glass transition temperature (T g ) and the onset of the crystallization temperature (T x ) using a computer application.
A dynamic mechanical analyzer model Q800 (TA Instruments, New Castle, DE, USA) was used to measure the viscoelastic properties as a function of temperature. For dynamic mechanical analysis (DMA) [46], a dual cantilever (bending) was used for composite samples and PTFE and a tensile clamp for the metallic glass ribbon. The dimensions for the specimens were a length of 30 mm, a width of 5.05 mm, and a thickness of 1.30 mm (30 × 5.05 × 1.30). The temperature for the samples was 400 • C at a heating rate of 2 K/min. The samples were examined at 0.1, 1 and 10 Hz with a dynamic strain of 0.1%.
Loss factor (Tan δ) was used to calculate the internal friction and damping of the samples, which was calculated using the following equation: where E" is loss modulus, and E is storage modulus. DMA was used to study the storage modulus (E ) and loss factor (Tan δ) under heating. Thermal diffusivity of the samples was studied by Netzsch LFA 447 NanoFlash (Netzsch Erich Netzsch GmbH & Co., Selb, Upper Franconia, Bavaria, Germany). The temperature range was between 25 and 300 • C.
The calculation for thermal conductivity was performed using the following equation: where α is the thermal diffusivity (mm 2 /s), C p is the specific heat capacity J/g·K, and ρ is the sample density (g/cm 3 ). SPS was used to produce a metallic glass sample in a bulk state. The samples for thermal diffusivity were obtained in the form of a circular disk 12.7 mm in diameter. A shop press was used to prepare the similar circular disk from composite powders, and subsequent sintering was performed in a vacuum furnace (PT 200) at 400 • C for 4 h and air-cooled for 1 h. The densities of the samples were measured by the hydrostatic weighing method using the AND GR-202 analytical balance and by using the AND AD-1653 set for density determination in ethanol.
A scanning electron microscope SEM, (Hitachi TM-1000, Tokyo, Japan) was used to study the microstructure of the samples at 15 kV in backscattered mode.

X-ray Diffraction Analysis (XRD)
The broad diffraction pattern (Figure 1a) of the Zr 65 Cu 17.5 Ni 10 Al 7.5 metallic glass shows its amorphous nature. Amorphous broad halo diffraction can be seen at 2 For the spark plas inserted inside graphite mm, and a height of 30 The sheet was used to using SPS [44] (Labox heating rate was 50 °C/ 50 MPa and 50 min dur X-ray diffraction ( Petersburg, Russian Fed Ɵ angles: 10 to 120°, used to determine the p phase composition was a standard to determine A differential scan Netzsch GmbH & Co. analysis. The heating ra was between 10 and 15 the supercooled region pacity of the samples. I the onset of the crystall A dynamic mecha USA) was used to mea dynamic mechanical a composite samples and mensions for the specim of 1.30 mm (30 × 5.05 × rate of 2 K/min. The sam 0.1%.
Loss factor (Tan δ samples, which was cal where E″ is loss modul age modulus (E′) and lo ≈ 38-50 • in the XRD pattern. After ball milling (Figure 1b), the metallic glass stayed amorphous. The XRD pattern of the metallic glass after SPS (Figure 1c) shows that the metallic glass was semicrystalline. After SPS, there was a formation of crystalline Zr 2 Ni and Zr 2 Cu phases in the metallic glass [47]. As compared to our previous results, it should be concluded that the intermetallic phases were formed during the SPS procedure. PTFE ( Figure 1d) shows a crystalline nature with the most intensive peak at 2 isette 5, Berlin. Germany) with a rotation speed o an argon atmosphere. The total duration of ball m The average size of the MG particles was about 10 added to the polymer for the ball milling process. lic glass/polymer ratios (mass.%) of 30/70, 50/50 time for the composite samples was 30 min, each u For the spark plasma sintering (SPS) measu inserted inside graphite die with an inner diamet mm, and a height of 30 mm. The sample was wra The sheet was used to prevent the sample from using SPS [44] (Labox 650, Sinter Land inc, Nag heating rate was 50 °C/min, with a maximum tem 50 MPa and 50 min duration. X-ray diffraction (XRD) (Research and produ Petersburg, Russian Federation) on a DRON Diffr Ɵ angles: 10 to 120°, step: 0.1°, exposition time used to determine the phase and structural comp phase composition was ±5%. The kinematic stand a standard to determine the percentage of amorph A differential scanning calorimeter (DSC) (N Netzsch GmbH & Co. Selb, Upper Franconia, Ba analysis. The heating rate was 10 °C/min under a was between 10 and 15 mg. The maximum tempe the supercooled region and crystallization and 3 pacity of the samples. It was used to determine th the onset of the crystallization temperature (Tx) us A dynamic mechanical analyzer model Q8 USA) was used to measure the viscoelastic prope dynamic mechanical analysis (DMA) [46], a du composite samples and PTFE and a tensile clamp mensions for the specimens were a length of 30 m of 1.30 mm (30 × 5.05 × 1.30). The temperature fo rate of 2 K/min. The samples were examined at 0. 0.1%.
Loss factor (Tan δ) was used to calculate th samples, which was calculated using the followin Tan δ = E where E″ is loss modulus, and E′ is storage modu age modulus (E′) and loss factor (Tan δ) under hea Ball milling of the samples was performed using a p isette 5, Berlin. Germany) with a rotation speed of 300 rota an argon atmosphere. The total duration of ball milling for The average size of the MG particles was about 10 µm. The added to the polymer for the ball milling process. Different lic glass/polymer ratios (mass.%) of 30/70, 50/50 and 70/30 time for the composite samples was 30 min, each under a ro For the spark plasma sintering (SPS) measurement, th inserted inside graphite die with an inner diameter of 12.7 mm, and a height of 30 mm. The sample was wrapped in a The sheet was used to prevent the sample from die conta using SPS [44] (Labox 650, Sinter Land inc, Nagaoka, Ni heating rate was 50 °C/min, with a maximum temperature 50 MPa and 50 min duration. X-ray diffraction (XRD) (Research and production ent Petersburg, Russian Federation) on a DRON Diffractometer Ɵ angles: 10 to 120°, step: 0.1°, exposition time per step used to determine the phase and structural composition of phase composition was ±5%. The kinematic standard meth a standard to determine the percentage of amorphous phas A differential scanning calorimeter (DSC) (NETZSCH Netzsch GmbH & Co. Selb, Upper Franconia, Bavaria, Ge analysis. The heating rate was 10 °C/min under an argon a was between 10 and 15 mg. The maximum temperature of the supercooled region and crystallization and 320 °C for pacity of the samples. It was used to determine the glass tr the onset of the crystallization temperature (Tx) using a com A dynamic mechanical analyzer model Q800 (TA In USA) was used to measure the viscoelastic properties as a dynamic mechanical analysis (DMA) [46], a dual cantile composite samples and PTFE and a tensile clamp for the mensions for the specimens were a length of 30 mm, a widt of 1.30 mm (30 × 5.05 × 1.30). The temperature for the sam rate of 2 K/min. The samples were examined at 0.1, 1 and 1 0.1%.
Loss factor (Tan δ) was used to calculate the internal samples, which was calculated using the following equation

Tan δ = E″/E′
where E″ is loss modulus, and E′ is storage modulus. DMA age modulus (E′) and loss factor (Tan δ) under heating.

Differential Scanning Calorimetry (DSC)
The DSC thermogram of the Zr 65 Cu 17.5 Ni 10 Al 7.5 metallic glass (Figure 2a) shows that T g is 374 • C and T x is 461 • C. The Zr 65 Cu 17.5 Ni 10 Al 7.5 metallic glass has a large supercooled region of 87 • C. This is in good agreement with the work of Abrosimova et al. [48]. The heating rate for the metallic glass sample is 10 • C/min. The melting temperature of the PTFE is 341 • C (Figure 2b). The composite samples have a melting point near the PTFE. The criteria for choosing the polymer were its physical properties and its melting temperature near the supercooled region of the metallic glass. This criterion was also used in our previous works [41][42][43]. The composites have a melting temperature of 346 • C for 30/70, 347 • C for 50/50 and 348 • C for 70/30 (Figure 2b) composites.

Dynamic Mechanical Analysis
DMA was used to study the atomic ability [49] of the metallic glass, PTFE and their composites. Atomic mobility is decreased by the structural relaxation due to a decrease in the defect concentration. The loss factor (Tan δ) increases with an increase in defect mobility. The change in the atomic mobility influences the structural relaxation and internal friction that caused α and ß relaxation in the composite and PTFE samples. For the metallic glass (Figure 3a), the storage modulus is maximum at 250 • C at 0.1 Hz, 270 • C at 1Hz and 301 • C at 10 Hz. The loss factor (Tan δ) increases with an increase in temperature. The maximum value for the metallic glass is 31,691 ± 317 MPa at 0.1 Hz, 31,194 ± 312 MPa at 1 Hz and 31,428 ± 314 MPa at 10 Hz. The 70/30 composite (Figure 3e) has the highest storage modulus compared to the 30/70 and 50/50 composites (Figure 3c,d). PTFE has the lowest storage modulus among the samples (Figure 3b).

Laser Flash Analysis (LFA)
The thermal conductivity of the metallic glass ( Figure 4) increases with an increase in temperature. It has the highest thermal conductivity (Table 1)

Scanning Electron Microscope
SEM analysis was performed to study the microstructure of the composite. Figure 5 shows that the bonding between the particles was not adequate. The presence of coarse metallic glass particles and PTFE can also be seen in the microstructure images. These influenced the resulting properties of the composites.

Discussions
The XRD pattern of the Zr 65 Cu 17.5 Ni 10 Al 7.5 metallic glass shows a broad halo between the diffraction angles of 2 For the spark plasma sintering (SPS) measurement, the metallic glas inserted inside graphite die with an inner diameter of 12.7 mm, an outer mm, and a height of 30 mm. The sample was wrapped in a 0.2 mm thick The sheet was used to prevent the sample from die contact. The powde using SPS [44] (Labox 650, Sinter Land inc, Nagaoka, Niigata Prefectur heating rate was 50 °C/min, with a maximum temperature of 400 °C, unde 50 MPa and 50 min duration. X-ray diffraction (XRD) (Research and production enterprise "Boure Petersburg, Russian Federation) on a DRON Diffractometer, under CoKα Ɵ angles: 10 to 120°, step: 0.1°, exposition time per step: 5s, beam size used to determine the phase and structural composition of samples. The a phase composition was ±5%. The kinematic standard method and anneal a standard to determine the percentage of amorphous phase [45].
A differential scanning calorimeter (DSC) (NETZSCH DSC 204 F1) Netzsch GmbH & Co. Selb, Upper Franconia, Bavaria, Germany) was us analysis. The heating rate was 10 °C/min under an argon atmosphere. Th was between 10 and 15 mg. The maximum temperature of the DSC was 6 the supercooled region and crystallization and 320 °C for the measurem pacity of the samples. It was used to determine the glass transition tempe the onset of the crystallization temperature (Tx) using a computer applicat A dynamic mechanical analyzer model Q800 (TA Instruments, Ne USA) was used to measure the viscoelastic properties as a function of tem dynamic mechanical analysis (DMA) [46], a dual cantilever (bending) composite samples and PTFE and a tensile clamp for the metallic glass r mensions for the specimens were a length of 30 mm, a width of 5.05 mm, a of 1.30 mm (30 × 5.05 × 1.30). The temperature for the samples was 400 ° rate of 2 K/min. The samples were examined at 0.1, 1 and 10 Hz with a dy 0.1%.
Loss factor (Tan δ) was used to calculate the internal friction and d ≈ 38 • to 50 • , without any sharp diffraction peaks. This indicates the amorphous behavior of the metallic glass. After ball milling of the metallic glass, a broad halo spectrum indicates the amorphous nature of the sample. During SPS, some crystalline peaks occurred, which corresponded to the formation of crystalline precipitates. Zr 2 Ni and Zr 2 Cu were indexed as the crystalline phase. In [50], the authors showed that metallic glass is semicrystalline. It should also be noted that after SPS at 400 • C, the crystalline phase (Table 2) was increased to 30 ± 5% in the MG sample. The most intense crystalline peaks in the polymer samples correspond to the PTFE. The metallic glass was mixed with a polymer, and different proportions of the composite samples were prepared. The crystalline peaks corresponded to the PTFE for all composite samples. A broad halo diffraction spectrum similar to the metallic glass at 2 Metals 2021, 11, x FOR PEER REVIEW ness of the ribbons was 2 age size of the polytetraflu Ball milling of the sa isette 5, Berlin. Germany) an argon atmosphere. The The average size of the M added to the polymer for lic glass/polymer ratios (m time for the composite sam For the spark plasma inserted inside graphite d mm, and a height of 30 m The sheet was used to pr using SPS [44] (Labox 65 heating rate was 50 °C/mi 50 MPa and 50 min durati X-ray diffraction (XR Petersburg, Russian Feder Ɵ angles: 10 to 120°, st used to determine the pha phase composition was ±5 = 38 • to 50 • was visible in the composite samples, which showed their amorphous nature. There was the formation of Zr 2 Cu in the composite samples, but this was not seen in different XRD intensity peaks. The sizes of this Zr 2 Cu should be nanocrystalline, but the amount of this phase should not be more than 5%. The DSC scan of the Zr 65 Cu 17.5 Ni 10 Al 7.5 metallic glass shows it was heated at 10 K/min. The onset glass transition temperature (T g ) is 374 • C, and the onset crystallization transition temperature (T x ) is 461 • C. The supercooled region T sc = T x − T g is 87 • C. Zr-based metallic glass has high thermal stability, which leads to a stable supercooled region [51]. PTFE and 30/70, 50/50 and 70/30 composites were heated at 10K/min. The composites have a melting temperature near PTFE. This result is similar to our previous research [43]. The thermal conductivity of the metallic glass is higher than the other samples and results in a decrease in the melting point of the composites. During DSC heating, the composites start to melt faster (with increasing MG content) when heat is transferred to the metallic glass. Hence, the melting peaks of the composite start to shift toward PTFE. Senatov et al. [52] showed that the dispersed filler in the polymer is influenced by the melting point of the polymer. The above factors could affect the melting point of the composites. Structural relaxation [53] is an important aspect of metallic glass. This is caused due to the change in the microstructure. It normally occurs near the glass transition (T g ) of the metallic glass. When heat is given to metallic glass, a local rearrangement of atoms occurs for a short-range order leading to a nonhomogenous region. This causes fluctuation in the electron density. These fluctuations are a result of structural relaxation during heating and can be seen in composites.
Dynamic mechanical properties were used to study the structural evolution of the sample. Storage modulus (E') and loss factor (Tan δ) were examined under constant heating and different frequencies (0.1 Hz, 1 Hz and 10 Hz). Tan δ or internal friction was used to measure the relaxation process in the samples. For the metallic glass, the storage modulus is constant up to 115 • C. It increases from 115 • C to 311 • C at 10 Hz, to 270 • C at 1 Hz and 246 • C to 0.1 Hz. There is a decrease in the storage modulus near the super liquid region. This is associated with a decrease in the viscosity and solid to liquid behavior translation takes place. During this process, mechanical relaxation occurs and is defined by viscoelastic loss [54]. The internal friction (Tan δ) increases with an increase in temperature. The storage modulus of the PTFE decreases with temperature due to the mechanical losses that occur during the secondary relaxation [55]. The local movement of the polymeric chains affects the mechanical rigidity of PTFE [56]. For PTFE, at 0.1Hz and 1 Hz, there is α transaction from 130 to 145 • C. The other peaks correspond to the glassy phase of PTFE. At 10 Hz, α translation occurring from 130 to 145 • C could be seen. The storage modulus of the 30/70, 50/50 and 70/30 samples was decreased with an increase in temperature. The mobility of the PTFE increases with an increase in temperature as it goes from the glass state to the viscous flow. This increases the deformation state of the composite [57]. The storage modulus of 70/30 is highest among the 30/70 and 50/50 composite samples. This can be attributed to the increase in the metallic glass content. For composite samples, the loss factor with frequencies 0.1 Hz and 1 Hz increases with an increase in temperature. For 10 Hz, the loss factor decreases with an increase in temperature. The α transaction peak at 130-145 • C could be seen in all the composites similar to the PTFE. Another peak in the composites around 350 • C shows the melting of the polymer. The PTFE peaks represent ß-relaxation before α relaxation. The composites have similar peaks to PTFE and can be explained by the PTFE peaks [58].
Metallic glass has the highest thermal conductivity, and PTFE has the least thermal conductivity among samples. The thermal conductivity of the metallic glass is temperature dependent and increases with an increase in temperature. Heat transfer is influenced by the lattice structure. Thermal diffusivity is dependent on the temperature and related to the mean free path changes. The translation enthalpy causes a change in the structure of the PTFE. This reduces the thermal conductivity of the PTFE at higher temperatures [59]. The 70/30 composite has the most thermal conductivity between the 50/50 and 30/70 composites. This is because the composite has more metallic glass content. The 70/30 composite shows an increase in conductivity with temperature. The 50/50 composite has slightly more thermal conductivity compared to the 30/70 composite. The change in thermal conductivity in the 30/70 composite and 50/50 composite is due to the change in the structure of the PTFE during the enthalpy process. This concludes that the addition of metallic glass into PTFE can improve the thermal conductivity of the composite.
The morphology of the composites was studied. Coarse particles, the presence of voids and lack of bonding result in the degradation of the properties of the composites. In this regard, it will be better to use organofunctional lubricants (for example, silanes) before the mechanical mixing of polymers and metallic glass. This will help to increase the interaction between the particles and improve interface bonding [40]. Additionally, for composite preparation, it will be better to use small glass particles with similar sizes to avoid varied particle sizes (especially coarse particles).

Conclusions
In the present study, metallic glass (Zr 65 Cu 17.5 Ni 10 Al 7.5 ) and PTFE composites were produced by ball milling. Different composites (30/70, 50/50 and 70/30) were prepared. Spark plasma sintering was used to produce a sample from metallic glass powder. It showed the formation of intermetallic crystalline Zr 2 Ni and Zr 2 Cu in the amount of about 30% for the MG sample. DMA was used to study the storage modulus and internal friction during heating of the samples. The storage modulus of the composites was decreased with an increase in temperature due to the increase in the mobility of the PTFE. α-structural relaxation occurred in the composites. The 70/30 composite had the highest thermal conductivity among the composite samples. The obtained composites will be promising in the field of dielectric materials for thermal applications.  Data Availability Statement: Data presented in this article are available at request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.