Neutral and charged excitons interplay in non-uniformly strain-engineered WS$_2$

TL;DR

This work demonstrates a scalable method to impart non-uniform mechanical strain to WS2 monolayers by suspending the film over holes and pressurizing with inert gas, producing large strain gradients that can be mapped optically and validated with FEM. Spatially resolved PL reveals not only the expected band-gap redshift under strain but also a pronounced neutral-to-trion conversion driven by carrier funneling toward the highest-strain region, with triangular geometries enhancing the effect due to steeper non-uniformity. The findings provide a practical non-uniform strain engineering knob for tuning excitonic physics in 2D semiconductors and offer insights into potential pseudomagnetic-field phenomena, with implications for nanoscale optoelectronics and fundamental studies of carrier dynamics in strained TMDCs.

Abstract

We investigate the response of excitons in two-dimensional semiconductors subjected to controlled non-uniform strain fields. In our approach to non-uniform strain-engineering, a WS$_2$ monolayer is suspended over a triangular hole. Large ($>2\;\%$), strongly non-uniform ($>0.28\;\%/μm$), and in-situ tunable strain is induced in the monolayer by pressurizing it with inert gas. We observe peak shifts and spectral shape changes in the photoluminescence spectra of strained WS$_2$. We interpret these changes as a signature of increased free electron density and resulting conversion of neutral excitons to trions in the region of high strain. Our result establishes non-uniform strain engineering as a novel and useful experimental `knob' for tuning optoelectronic properties of 2D semiconductors.

Figures (4)

• Figure 1: (a) Measurement setup schematics. The WS$_2$ flake suspended over a hole in a SiNx membrane is pressurized from one side and optically interrogated from another. (b-c) Optical images of monolayer WS$_2$ flakes transferred over a triangular (b) and a circular (c) hole. (d-e) The spatial variation of the local strain $\varepsilon_{xx}(x,y)+\varepsilon_{yy}(x,y)$ simulated by a finite element method (FEM) simulation for a triangular (d) and circular (e) membranes, respectively. Pressure difference is $\Delta P=0.8\;bar$ for both geometries. The colorbar presents the strain magnitude.
• Figure 2: (a-b) PL spectra acquired at the point of maximum strain in a triangle (a) and a circle (b) sample. The shaded areas shown for the lowest curves in (a-b) are the trion (red) and neutral exciton (blue) fitted peaks. The solid vertical lines present the predicted energy of the center of exciton peak position (see main text). The discrepancy between the predicted exciton energy and the peak position is clear for both samples. (c,e) Spatial map of the PL peak maximum at $\sim$ 1.2 bar pressure for a triangular and circular hole, respectively. (d,f) FEM simulations corresponding to (c) and (f). Points of the maximum strain are marked with a cross. Colorbars in (c-f) represent the energy ($eV$).
• Figure 3: (a,b) Detuning plots, with the neutral exciton peak position centered at $0\;eV$, for a triangle (a) and a circle (b) samples. (c,d) The spectral position of the PL peak maximum (filled black circles), the neutral exciton position (empty blue circles), and the trion position (empty red diamonds) vs. applied pressure for the same samples as in a) and b). The transition from neutral exciton to trion-dominated regimes is more evident for the triangle sample. (e,f) Examples of experimental spectra at two different pressures (black solid curves) fitted to a neutral exciton (blue dashed curves) and a trion (red dashed curves) resulting in a close fit to the data (pink curve). In (e) the exciton is more dominant while in (f) the trion is more dominant. The asymmetric red tail of the trion is due to the electron recoil effect.
• Figure 4: (a) The ratio between the neutral exciton and trion PL peak amplitudes for different samples. The guide-to-the-eye dashed lines show the approximate linear dependency of the conversion efficiency in different samples. (b) Calculation of the free carrier density from the experimental data, see main text. Dashed lines are guide-to-the-eye linear fits. (c) Schematic illustration of the spatial electron funneling under strain-modulated band-gap.