[Paper Review] Magneto-excitons in large area CVD grown monolayer MoS$_{2}$ and MoSe$_{2}$ on sapphire

Anatolie Mitioglu, Krzysztof Gałkowski+6 more|HAL (Le Centre pour la Communication Scientifique Directe)|Feb 3, 2016

Quantum and electron transport phenomena5 citations

TL;DR

This study investigates valley Zeeman splitting in large-area CVD-grown monolayer MoS₂ and MoSe₂ on sapphire using magneto-optical spectroscopy. It finds valley g-factors of approximately -4.5, indicating minimal influence from intrinsic strain (~0.2%), and confirms electronic properties are indistinguishable from exfoliated samples, supporting CVD-grown TMDs for scalable valleytronic devices.

ABSTRACT

Magneto transmission spectroscopy was employed to study the valley Zeeman effect in large-area monolayer MoS$_{2}$ and MoSe$_{2}$. The extracted values of the valley g-factors for both A- and B-exciton were found be similar with $g_v \simeq -4.5$. The samples are expected to be strained due to the CVD growth on sapphire at high temperature ($700^\circ$C). However, the estimated strain, which is maximum at low temperature, is only $\simeq 0.2\%$. Theoretical considerations suggest that the strain is too small to significantly influence the electronic properties. This is confirmed by the measured value of valley g-factor, and the measured temperature dependence of the band gap, which are almost identical for CVD and mechanically exfoliated MoS$_2$.

Motivation & Objective

• To investigate the valley Zeeman effect in large-area, CVD-grown monolayer MoS₂ and MoSe₂ on sapphire substrates.
• To determine whether strain from lattice mismatch during high-temperature CVD growth significantly alters electronic properties.
• To compare the valley g-factors and temperature-dependent band gap evolution in CVD-grown TMDs with those of mechanically exfoliated samples.
• To assess the role of strain in modifying the valley magnetic moment and excitonic properties in transition metal dichalcogenides.

Proposed method

• Performed polarization-resolved magneto-optical absorption spectroscopy using circularly polarized light (σ⁺, σ⁻) to selectively probe ±K valley transitions.
• Applied magnetic fields up to 65 T to measure linear scaling of excitonic transition splitting with field, enabling extraction of valley g-factors.
• Used the single-oscillator model of O’Donnell and Chen to fit the temperature-dependent band gap evolution in MoS₂, incorporating phonon coupling and exciton binding energy.
• Estimated strain from thermal contraction differences between MoS₂ and sapphire, assuming a 70 meV/% strain shift in band gap.
• Compared experimental g-factors and band gap temperature dependence with theoretical predictions and data from exfoliated MoS₂ on SiO₂/Si substrates.
• Conducted DFT calculations to assess intercellular contributions to valley magnetic moment, though these were found inconsistent with experiment.

Experimental results

Research questions

• RQ1 What is the value of the valley g-factor in CVD-grown monolayer MoS₂ and MoSe₂ under high magnetic fields?
• RQ2 To what extent does strain from lattice mismatch with sapphire substrates affect the valley Zeeman splitting in CVD-grown TMDs?
• RQ3 How does the temperature dependence of the A-exciton band gap in CVD-grown MoS₂ compare to that of exfoliated MoS₂ on SiO₂/Si?
• RQ4 Can the observed valley g-factor values be explained by theoretical models that include strain and intercellular contributions?
• RQ5 Is the electronic behavior of CVD-grown TMDs sufficiently similar to exfoliated TMDs to support their use in scalable valleytronic devices?

Key findings

• The low-temperature valley g-factor for the A-exciton in MoS₂ is gᵥ ≈ -4.5 ± 0.1, and for the B-exciton gᵥ ≈ -4.3 ± 0.1.
• For MoSe₂, the A-exciton valley g-factor is gᵥ = -4.4 ± 0.1 at low temperatures, indicating consistency across materials.
• The valley g-factor remains nearly constant from 2 K to 120 K, indicating weak temperature dependence of the valley splitting.
• Estimated strain in CVD-grown MoS₂ is ~0.2% at 2 K, but this has negligible impact on electronic properties, as confirmed by band gap measurements.
• The temperature dependence of the A-exciton band gap in CVD MoS₂ matches that of exfoliated MoS₂ on SiO₂/Si, with a fit using the O’Donnell and Chen model yielding E(0) = 1.948 eV and ⟨ħω⟩ = 24.25 meV.
• The expected strain-induced band gap shift (~8 meV) is much smaller than the observed 70 meV change with temperature, confirming strain is not the dominant factor in gap evolution.