Unraveling Thermal Transport Properties of MoTe2 Thin Films Using the Optothermal Raman Technique

Understanding phonon transport and thermal conductivity of layered materials is not only critical for thermal management and thermoelectric energy conversion but also essential for developing future optoelectronic devices. Optothermal Raman characterization has been a key method to identify the properties of layered materials, especially transition-metal dichalcogenides. This work investigates the thermal properties of suspended and supported MoTe2 thin films using the optothermal Raman technique. We also report the investigation of the interfacial thermal conductance between the MoTe2 crystal and the silicon substrate. To extract the thermal conductivity of the samples, temperature- and power-dependent measurements of the in-plane E2g1 and out-of-plane A1g optical phonon modes were performed. The results show remarkably low in-plane thermal conductivities at room temperature, at around 5.16 ± 0.24 W/m·K and 3.72 ± 0.26 W/m·K for the E2g1 and the A1g modes, respectively, for the 17 nm thick sample. These results provide valuable input for the design of electronic and thermal MoTe2-based devices where thermal management is vital.


Nonlinearity of the power absorbed and phase change of the supported MoTe 2
Power-dependent measurements of the main text have been performed considering the linear region of the laser power absorbed (see Figure S1a). This linear evolution of the sample differs whether the sample is suspended or supported, as well as the thickness of the flake.
For example, by using a 100× objective in the supported part of sample 1 (17 nm), the nonlinearity of the Raman modes occurs above 1 mW of the laser power. Then the observed redshift of E 1 2g (blue circles) and A 1g (red circles) saturates, and the trigonal 2H phase is distorted into the octahedral 1T' structure ( Figure S1b). As a result of the absorbed laser power, the E 1 2g and A 1g disappear, and two new features emerge in the Raman spectra at 126 and 140 cm −1 that corresponds to the modes of the induced 1T'-MoTe 2 . For the purpose of comparison, the phase change by laser irradiation has been performed in a different sample with similar thickness. Figure S1: a) Power dependent measurements of the supported sample 1 of the E 1 2g and A 1g for laser power up to 6.6 mW b)Beyond 6 mW the lattice is deformed, and the 1T' is induced by laser irradiation.
Phase change by laser irradiation below 10-15 nm of thickness The selected samples' thicknesses for the optothermal analysis have been selected because those are more robust to ambient perturbations and can be considered in the polymorphic engineering of MoTe 2 . In our experiments, samples with thicknesses below 10-15 nm thick cannot be laser-driven from the 2H to 1T' phase since most of the material is evaporated and they are environmentally unstable using the optothermal Raman technique. Figure S2: Optical microscope image of few-layer MoTe 2 before a) laser irradiation and b) after laser irradiation.
Temperature dependence measurements of sample 2 Figure S3: a) Temperature-dependent Raman spectra of a) A 1g and b) E 1 2g for supported MoTe 2 (sample 2, 37nm) with temperature range from room temperature to 500 K. c) Temperature-dependent Raman peak positions of the characterized E 1 2g (red square are related to the supported and red circles to the suspended) and A 1g (blue square are related to the supported and blue circles to the suspended) phonon modes for the sample 2 MoTe 2 . All the points have been individually extracted from the Lorentzian fit of the Raman peaks. Figure S4: A 1g and E 1 2g phonon mode frequency of sample 1 (17 nm) for a) suspended and b) supported MoTe 2 as a function of absorbed laser power using 50× objective with the linear fit. Figure S5: A 1g and E 1 2g phonon mode frequency of sample 2 (37 nm) for suspended and supported MoTe 2 as a function of absorbed laser power using a-b) 50× and c-d) 100× objective with the linear fit.   Figure S6 (a) presents the thermal resistance for the 17 nm sample. The solution given by the experimental and theoretical spot sizes differ by one order of magnitude; thus, determining the actual spot sizes of the heating laser beam is essential to get reliable results with this method. The upper limit of the experimental uncertainty of the R m (100×)/R m (50×) is on the region where no solution for the k 2 /G is obtained due to the Si substrate having too big interfacial thermal conductance with the MoTe 2 . One option to decrease the uncertainties of the thermal properties would be to transferMoTe 2 onto different substrates with lower thermal conductivities, such as silica. In general, the interfacial thermal conductance analysis with distinctive substrates would contribute to a better understanding of thermal management. Figure S6 (b) gives the corresponding curves for the 37 nm sample. The difference of k 2 /G between the theoretical and experimental spot sizes is substantially smaller. The uncertainties in the experimentally obtained ratio cause much smaller uncertainty in the obtained k 2 /G, as the position on the curve is at a much steeper part in contrast to the case for the 17 nm sample, where the solution is on the saturated part of the curve. However, the uncertainties are still quite large due to the logarithmic scale of the x-axis, which means that the theoretically obtained ratio is really sensitive to the experimental value of the thermal resistances R m .

Temperature profile of sample 1
In the case of suspended MoTe 2 , the sample is supported in the region r ≥ R, in which Eq. (3) from the main text applies, with proper boundary conditions to conform with the temperature profile of the suspended part. The temperature profile for the supported part, in this case, is This temperature profile is a decaying function with the limit T (r → ∞) = T a , and G and k 2 determine the decay length. In the case of MoTe 2 , the low thermal conductivity k 2 and large interfacial thermal conductance G mean that the decay length is really short, and the temperature has decayed to the room temperature at r = 4.8 µm, namely 0.3 µm from the edge, for 17 nm MoTe 2 . In the case of 37 nm MoTe 2 , the temperature has decayed to room temperature at r = 6 µm, 1 µm from the edge, due to lower G and higher k 2 compared to 17 nm MoTe 2 . For both samples, the temperature at r = R is quite close to the room temperature due to the high interfacial thermal conductance between Si and MoTe 2 . However, if the substrate is a low thermal conducting material, such as silica, the temperature at the edges is expected to be higher.