Strain Engineering of MoS₂ by Tuning the Transfer Process for Improving Its Electrical Performance

Two-dimensional semiconductors, such as MoS₂, have garnered significant attention due to their potential as alternatives to silicon in electronic devices. In the case of MoS₂, reported experimental carrier mobility is an order of magnitude lower than expected, which hinders its future applications. Strain engineering is a common method in the semiconductor industry to enhance the carrier mobility of silicon and similar improvements have been expected for MoS₂: applying tensile strain into MoS₂ enhances its electron mobility. Therefore, strain engineering is a possible method to achieve future high-performance MoS₂-based electronic devices. One of the challenges in strain engineering of MoS₂ is achieving strained MoS₂ on rigid substrates, such as surface-oxidized silicon (SiO₂/Si): the surface inertness of MoS₂ makes it challenging to induce and maintain strain through lattice mismatch with the substrate. In this work, we report another approach of introducing strain into MoS₂ by tuning the wet transfer process. By optimizing vacuum annealing conditions during transfer process of CVD MoS₂, we successfully introduced approximately 0.5% tensile strain into CVD MoS₂, resulting in a 19-fold increase in carrier mobility. Our results show that controlling of the transfer process is an important way to achieve strain-engineered MoS₂ for future high-performance MoS₂-based devices.
MoS₂ flakes were synthesized on a SiO₂/Si substrate through CVD using MoO₂ and elemental S. For the growth, we added KBr as a growth promoter. Next, MoS₂ was transferred using PMMA-assisted wet-transfer method, with KOH solution as the etchant. After transferring the PMMA/MoS₂ stack onto another SiO₂/Si substrates, we annealed the samples at 50°C or 160°C under vacuum for 30 min to improve adhesion between the MoS₂ and SiO₂ (referred to as LTT and HTT MoS₂, respectively). Then, we measured the Raman mapping of these two samples to evaluate the strain and doping introduced by the transfer process. By plotting the relationship between the E′ and A′₁ peak positions, the strain and carrier density of MoS₂ can be evaluated. The tensile strain in HTT MoS₂ increased approximately 0.5% from the LTT MoS₂. This strain remained after the PMMA removal process, indicating the successful transfer of strained MoS₂ onto a rigid substrate. By estimating the thermal expansion of MoS₂, SiO₂, and PMMA, we found that the strain in HTT MoS₂ could not be achieved by the thermal expansion of MoS₂ itself (~0.07%), while this value is almost same as the thermal expansion of PMMA (~0.8%). Thus, we speculated that HTT MoS₂ was expanded by the thermal expansion of PMMA, and the adhesion between MoS₂ and SiO₂ was improved. As a result, strained MoS₂ was successfully transferred onto SiO₂/Si. Note that the carrier density of HTT MoS₂ was 1.6 times higher than that of LTT MoS₂. This increase could be caused by carrier doping from the substrate due to improved adhesion between MoS₂ and SiO₂ during annealing at 160°C. Field-effect transistors were fabricated using photolithography, followed by thermal deposition of Ni and Au as contact metals. We measured the performance of over 30 devices fabricated by each process and found the average carrier mobilities of 0.48 and 9.1 cm²V⁻¹s⁻¹ for LTT and HTT MoS₂, respectively. The increase in carrier mobility is attributed to both the tensile strain and the increased carrier density.
In summary, we successfully demonstrated that strain-engineered MoS₂ can be achieved onto SiO₂/Si by optimizing the transfer process. As a result, the carrier mobility of MoS₂ improved by 19 times. Our results suggest that the transfer process is one of the important parameters for tuning the electronic and optoelectronic performance of MoS₂-based devices.