Monolayer transition metal dichalcogenides (TMDs) host excitons, bound states of electrons and holes, existing as superpositions of conduction and valence band states in the K and K' valleys. Circularly polarized light selectively generates K and K' valley excitons of opposite helicities, while linearly polarized excitation produces a valley-coherent K-K' exciton hybrid. However, a major challenge lies in the rapid degradation of valley coherence, typically within a sub-picosecond timescale (98–520 fs), primarily due to scattering and inter-valley exchange interactions. This short coherence time, significantly shorter than the exciton radiative lifetime (~1 ps), hinders optical readout of strong exciton valley coherence. To utilize these coherent excitons as qubits in quantum information processing, substantial enhancement of valley coherence time is crucial. This study aims to demonstrate a method for significantly enhancing this coherence time, a goal of high scientific importance for quantum technologies. The researchers demonstrate a 100% degree of linear polarization (DOLP) in the photoluminescence (PL) peak of the A_1s exciton in monolayer MoS₂ encapsulated with few-layer graphene (FLG) on top and bottom. This complete retention of valley coherence in steady-state PL signifies a significantly enhanced valley coherence time, essentially limited only by the exciton lifetime.

Previous studies on valley coherence in monolayer TMDs have reported coherence times in the range of 98–520 fs, significantly shorter than the exciton radiative lifetime. This short coherence time is attributed to a combination of fast scattering and inter-valley exchange interactions. These limitations make the optical readout of strong exciton valley coherence challenging. Several studies have investigated techniques to extend the coherence time, but significant breakthroughs were lacking before this current research. The existing literature highlighted the need for strategies to minimize scattering and reduce the exchange interaction to achieve longer valley coherence times, essential for potential applications in quantum computing.

The researchers employed four different stacks of monolayer MoS₂ with varying degrees of dielectric screening and exciton lifetimes to investigate the interplay between these factors and valley coherence. The stacks were: (1) hBN-MoS₂-hBN (HMH), (2) FLG-hBN-MoS₂-hBN-FLG (GHMHG), (3) MoS₂-FLG-hBN (MGH), and (4) FLG-MoS₂-FLG (GMG). They measured the degree of linear polarization (DOLP) of the A_1s exciton in the photoluminescence (PL) spectra under linearly polarized excitation. They also conducted polarization-dependent time-resolved photoluminescence (TRPL) measurements to determine the exciton lifetime. To understand the underlying mechanisms, they utilized theoretical calculations based on solutions of Bethe-Salpeter and Maialle-Silva-Sham (MSS) equations. These calculations modeled the exciton dynamics, considering the effects of scattering, exchange interaction, and dielectric screening. The experimental results were compared with these theoretical simulations to understand the effects of screening and filtering on valley coherence. The Bethe-Salpeter equation was solved numerically to determine the exciton binding energy in different stacks. The long-range exchange potential was calculated to evaluate its dependence on screening. The steady-state form of the MSS equation was solved to model the DOLP as a function of the scattering rate, providing a quantitative understanding of the motional narrowing regime.

The study observed a significant enhancement in the DOLP of the A_1s exciton in the GMG stack (~100%), compared to the HMH stack (44.5%). The GMG stack, with monolayer MoS₂ encapsulated by FLG on both sides, exhibited nearly complete valley coherence. The enhanced DOLP in the GMG stack is attributed to three factors: (a) suppressed exchange interaction due to increased dielectric screening from the graphene layers, (b) reduced exciton lifetime due to fast inter-layer transfer to graphene (acting as a filter), and (c) operation in the motional narrowing regime. The experimental DOLP values for the four stacks were: HMH (44.5 ± 10%), GHMHG (37 ± 9%), MGH (77 ± 5%), and GMG (96 ± 6%). In several instances, the GMG stack exhibited ~100% DOLP. The A_2s-A_1s energy separation, indicative of screening, decreased from 144.5 meV in HMH to 43.3 meV in GMG. The calculated exciton binding energy also decreased significantly from 379 meV in HMH to 122 meV in GMG. TRPL measurements showed exciton lifetimes of <5 ps in HMH and GMG, and 6-8 ps in GHMHG. The theoretical modeling based on MSS equations confirmed that the system operates in the motional narrowing regime in both HMH and GMG stacks, where increased scattering rate leads to longer pseudospin coherence time.

The findings demonstrate a significant advancement in controlling valley coherence in monolayer TMDs. The ~100% DOLP observed in the GMG stack represents a substantial improvement over previous reports, showing for the first time that excitons can maintain valley coherence for their entire lifetime. The combined effects of enhanced dielectric screening by graphene, fast exciton transfer to graphene, and motional narrowing enable the achievement of this high degree of valley coherence. The results highlight the importance of dielectric screening in suppressing the exchange interaction and the role of fast filtering in reducing the exciton lifetime. The excellent agreement between experimental results and theoretical simulations based on the MSS equation strengthens the understanding of the underlying physical mechanisms. These findings have important implications for the development of valleytronic and quantum information processing devices based on TMDs.

This study achieved near-perfect valley coherence in monolayer MoS₂ by encapsulating it with graphene. This unprecedented ~100% DOLP is a result of the synergistic effects of enhanced screening, reduced exciton lifetime, and motional narrowing. The findings provide a pathway towards creating robust valleytronic devices and advancing quantum information technologies. Future research could explore other TMD materials and investigate different encapsulation strategies to further optimize valley coherence. Investigating the detailed dynamics of exciton transfer to graphene and the impact of different types of graphene would also be valuable.

The study focuses on a specific set of TMD materials and graphene encapsulation methods. The generalizability of these findings to other TMD materials and different environmental conditions needs further investigation. While the theoretical model accurately describes the observed trends, it involves certain approximations (e.g., neglecting exciton-phonon scattering at 5K). The TRPL measurements have limitations in determining precise lifetimes shorter than the instrument response function, introducing uncertainties in lifetime estimations for some of the stacks.