Atomistic Origin of Photoluminescence Quenching in Colloidal MoS2 and WS2 Nanoplatelets

Abstract

Large chemical tunability and strong light-matter interactions make colloidal transition metal dichalcogenide (TMD) nanostructures particularly suitable for light-emitting applications. However, ultrafast exciton decay and quenched photoluminescence (PL) limit their potential. Combining femtosecond transient absorption spectroscopy with first-principles calculations on MoS2 and WS2 nanoplatelets, we reveal that the observed sub-picosecond exciton decay originates from edge-located optically bright hole traps. These intrinsic trap states stem from the metal d-orbitals and persist even when the sulfur-terminated edges are hydrogen-passivated. Notably, WS2 nanostructures show more localized and optically active edge states than their MoS2 counterparts, and zigzag edges exhibit a higher trap density than armchair edges. The nanoplatelet size dictates the competition between ultrafast edge-trapping and slower core-exciton recombination, and the states responsible for exciton quenching enhance catalytic activity. Our work represents an important step forward in understanding exciton quenching in TMD nanoplatelets and stimulates additional research to refine physicochemical protocols for enhanced PL.

Figures (5)

• Figure 1: High resolution TEM-image of (a) WS$_2$ NPLs and (b) MoS$_2$ NSs, different in shape, size and orientation. A single exemplary NPL is highlighted with a white border. The hexagonal crystal structure of the semiconducting 2H-phase can be seen in the FFT-inset (yellow) with the underlying graphene structure indicated by the black arrow. Size distributions of (c) WS$_2$, and (d) MoS$_2$ NPLs and NSs respectively.
• Figure 2: (a, b) Spectral line cuts of MoS$_2$ (a) and WS$_2$ (b) at $t_0$ showcasing shift, broadening and weakening of the A-excitonic transition under lateral confinement. The spectral position of each A-exciton is marked by a vertical line, while the areas obscured by the pump pulse are marked in the resp. color of the corresponding graph. (c) Decay dynamics of the A-exciton of all samples. The time scale is linear between -0.5 and 10 ps and logarithmic from 10-7500 ps to allow both high resolution of the early response and maximum coverage.
• Figure 3: (a) Optimized structures of hydrogen-passivated MoS$_2$ nanoplatelets (NPLs) of different sizes with zigzag (ZZ)and armchair (AC) S-edges. Mo atoms are shown in purple, S in yellow, and H in gray. Black arrows indicate the size. (b) Formation energies of MoS$_2$ and WS$_2$ NPLs as a function of size. (c) Inverse participation ratio for states around Fermi energy (E$_F$ ) for MoS$_2$ and WS$_2$ NPLs. (d) Net Bader charges for calculated MoS$_2$ (top) and WS$_2$ (bottom) NPL sizes. (e) Projected density of states (PDOS) near the Fermi level for 2.1 nm MoS$_2$ (left) and WS$_2$ (right) NPLs.
• Figure 4: Calculated core-edge contributions to energy levels near the Fermi energy in (a) MoS$_2$ and (b) WS$_2$ NPLs. Blue stars (red triangles) mark states with $>50\%$ edge (core) contribution for the core-edge partitioning. (c) Relative contributions of zigzag (ZZ) and armchair (AC) edges to electronic states near the Fermi energy (E$_F$) in 2.1 nm MoS$_2$ (left) and (b) WS$_2$ (right) NPLs.
• Figure 5: Calculated oscillator strengths for transitions below 4 eV in (a) MoS$_2$ and (b) WS$_2$ NPLs, colored based on state character: red (core-originated, both states $<50\%$ edge character), blue (edge-originated, at least one state has $>50\%$ edge character).