Extremely high excitonic $g$-factors in 2D crystals by alloy-induced admixing of band states

TL;DR

This study demonstrates that alloying Mo and W in monolayer MoWSe2 enables extremely large excitonic g-factors, with neutral excitons reaching g-factors around $g \approx -10$ at Mo content $x \approx 0.23$, in contrast to the conventional $g \approx -4$ seen in unary S-TMDs. Using magneto-optical measurements on hBN-encapsulated MoWSe2 alloys and first-principles band-structure calculations, the authors show that alloy-induced admixture of K- and Q-valley conduction-band states drives a non-monotonic g-factor evolution for neutral excitons, while trion g-factors remain near the conventional values. Theoretical analysis reveals that conduction-band mixing, modulated by lattice parameters and strain, underpins the giant g-factors, and that local strain contributes only modestly to the observed effects. Collectively, the work establishes alloying as a practical route to tailor valley Zeeman physics in 2D crystals, with significant implications for valleytronics and strain-engineered optoelectronic devices.

Abstract

Monolayers (MLs) of semiconducting transition metal dichalcogenides (\mbox{S-TMDs}) emit light very efficiently and display rich spin-valley physics, with gyromagnetic ($g$-) factors of about -4. Here, we investigate how these properties can be tailored by alloying. Magneto-optical spectroscopy is used to reveal the peculiar properties of excitonic complexes in Mo$_{x}$W$_{1-x}$Se$_2$ MLs with different Mo and W concentrations. We show that the alloys feature extremely high $g$-factors for neutral excitons, that change gradually with the composition up to reaching values of the order of -10 for $x \approx 0.2$. First-principles calculations corroborate the experimental findings and provide evidence that alloying in S-TMDs results in a non-trivial band structure engineering, being at the origin of the high $g$-factors. The theoretical framework also suggests a higher strain sensitivity of the alloys, making them promising candidates for tailor-made optoelectronic devices.

Figures (24)

• Figure 1: (a) Low-temperature ($T$=10 K) example PL spectra of MoSe$_2$, WSe$_2$, and Mo$_x$W$_{1-x}$Se$_2$ MLs encapsulated in hBN flakes. (b) Extracted energy of neutral exciton (X) as a function of molybdenum (Mo) concentration. The dashed lines are guided to the eye.
• Figure 2: (a) Helicity-resolved PL spectra of an hBN-encapsulated Mo$_{0.49}$W$_{0.51}$Se$_2$ ML at $T$=4.2 K measured at selected values of the applied out-of-plane magnetic field. The red (blue) color corresponds to the $\sigma^+$ ($\sigma^-$) polarized spectra. The measurements were performed under excitation energy of 2.41 eV and power of 2.5 $\mu$W. The $\sigma^+$-polarized spectra were normalized to the X intensity, while the $\sigma^-$-polarized spectra were multiplied by scaling factors to make them better visible. The spectra are vertically shifted for clarity. (b) Magnetic-field evolutions of the energy differences ($\Delta E$) between the two circularly polarized split components of the X and T transitions. The solid lines represent fits according to the equation described in the text.
• Figure 3: Experimental values of $g$-factors extracted for the neutral and charged excitons measured on the MoSe$_2$, WSe$_2$, and MoWSe$_2$ MLs with different Mo concentration. The dashed curves are guide to the eye.
• Figure 4: Electronic band structures calculated for the studied Mo$_x$W$_{1-x}$Se$_2$ MLs with $a_0(x)$. The color code reveals the wavefunction localization at the W and Mo atomic sites. Band structures using the 2$\times$2 supercells for pure MoSe$_2$ and WSe$_2$ crystals are shown with thin grey lines for comparison. The lowest energy interband transitions with circular polarization are indicated with vertical arrows. The unfolded band structures are shown in Fig. S7.2 of the SM SM.
• Figure 5: (a) Exciton $g$-factors calculated from first principles. Colored lines correspond to different lattice constants of MoSe$_2$. The non-monotonic dependence reproduces the experimental trend shown in Fig. \ref{['fig:gfactor_exp']}. The dashed lines indicate the results without optimization of the atomic positions (n.o.), which has a minor impact on the final values. Calculated $g$-factors for the (b) conduction, $g_c$, and (c) valence, $g_v$, bands at the K point, allowing us to pinpoint the electronic origin of the non-monotonic dependence observed in the exciton $g$-factor.