Tunable electronic structure in gallium chalcogenide van der Waals compounds

Transition-metal monochalcogenides comprise a class of two-dimensional materials with electronic band gaps that are highly sensitive to material thickness and chemical composition. Here, we explore the tunability of the electronic excitation spectrum in GaSe by using angle-resolved photoemission spectroscopy. The electronic structure of the material is modified by in situ potassium deposition as well as by forming ${\mathrm{GaS}}_x{\mathrm{Se}}_{1-x}$ alloy compounds. We find that potassium-dosed samples exhibit a substantial change of the dispersion around the valence-band maximum (VBM). The observed band dispersion resembles that of a single tetralayer and is consistent with a transition from the direct-gap character of the bulk to the indirect-gap character expected for monolayer GaSe. Upon alloying with sulfur, we observe a phase transition from AB to ${\text{AA}}^{\prime }$ stacking. Alloying also results in a rigid energy shift of the VBM towards higher binding energies, which correlates with a blueshift in the luminescence. The increase of the band gap upon sulfur alloying does not appear to change the dispersion or character of the VBM appreciably, implying that it is possible to engineer the gap of these materials while maintaining their salient electronic properties.

Figure 1

Real-space structure and HRSTEM images of the two most common ${\mathrm{GaS}}_x{\mathrm{Se}}_{1-x}$ polytypes. Panels (a) and (b) show the crystal structure of the $\varepsilon$ phase in the planes defined by the axes in the labels. The corresponding HRSTEM image in panel (c) shows a trigonal lattice, resulting from the projection of Bernal (AB) stacked honeycomb layers. The intensity cut in panel (d) was obtained from the red rectangle in panel (c) and shows bright atomic columns with approximately twice the intensity of dim atomic columns. Analogous images in panels (e) and (f) show the geometry of the $\beta$ phase. The HRSTEM image in panel (g) shows a honeycomb mesh, with no intensity in the interstitial columns between hexagonal cells.

Figure 2

ARPES measurements at a photon energy of 94 eV of the electronic structure of pristine GaSe around the $\mathrm{\Gamma }$ point. (a) Sketch of the BZs and their orientation in the experiment with $\mathrm{\Gamma },K$, and $M$ symmetry points labeled. (b) Measured dispersion in the $\mathrm{\Gamma }\text{−}K\text{−}M$ direction obtained along the blue dashed line shown in panel (a). (c) Measured dispersion in the $\mathrm{\Gamma }\text{−}M\text{−}\mathrm{\Gamma }$ direction obtained along the red dashed line shown in panel (a). The double headed arrows in panels (b) and (c) indicate the planes of constant energy from which the cuts in panels (d)–(h) are taken. The energy of each cut is also listed in bottom of each panel in units of electron volts (eV). High-symmetry points $\mathrm{\Gamma },K$, and $M$ have been labeled in panels (e)–(g), respectively.

Figure 3

ARPES measurements of the $k_z$ dispersion of GaSe. (a) Sketch of bulk BZ highlighting the $\mathrm{\Gamma }\text{−}A\text{−}H\text{−}K$ plane. (b) Dispersion along the $\mathrm{\Gamma }\text{−}A$ direction, as marked by the blue dashed line in panel (a). Fits of the position of the VBM intensity are shown via open circles. The solid curve following the fitted dispersion is a guide for the eye. The BZ size of $2\pi /c$ in the $z$ direction is marked by a double-headed arrow and two vertical dashed lines. (c), (d) Dispersion extracted at the given $k_z$ values, corresponding to the $\mathrm{\Gamma }\text{−}K$ direction (red dashed line). (e) Sketch of first and second BZs with an outline of the $\mathrm{\Gamma }\text{−}A\text{−}L\text{−}M$ plane. (f) Constant-energy contour of the plane sketched in panel (e) extracted at a binding energy of 0.2 eV with high-symmetry points annotated by tick marks. (g) Dispersion along the $\mathrm{\Gamma }\text{−}M\text{−}\mathrm{\Gamma }$ direction obtained at a $k_z$ value of 4.75 Å${}^{-1}$.

Figure 4

Effect of K dosing on the GaSe electronic structure. (a), (b) ARPES measurements of the dispersion in the $\mathrm{\Gamma }\text{−}K\text{−}M$ direction (a) before and (b) after K dosing. The dashed curves present the fitted location of the VBM peak intensity. (c) Location of the VBM peak intensity (open circles) fit with a second-order and a fourth-order polynomial for the clean and K dosed samples, respectively. (d) EDCs (open circles) and results of Voigt function fits (curves) in the clean and K-dosed cases. The EDCs were extracted at $\mathrm{\Gamma }$ along the lines shown in panels (a) and (b). The position of the top-most peak is plotted via dashed lines in panels (a) and (b) and open circles in panels (c). (e) Evolution of the photoemission intensity at $\mathrm{\Gamma }$ as a function of K-dosing steps. The open circles and guidelines mark the fitted peak positions obtained by performing the EDC analysis shown in panel (d) in each K-dosing step. Ga $3d$ core-level measurements (f) before and (g) after K dosing. The data (thick red curves) have been fit (dashed curves) with Doniach-Sunjic lineshapes (filled peaks) for each of the observed components.

Figure 5

Composition of ${\mathrm{GaS}}_x{\mathrm{Se}}_{1-x}$ alloys from core-level data. Core-level measurements (thick red curves) in the Se $3p$ and S $2p$ binding energy region for (a) GaSe and (b), (c) ${\mathrm{GaS}}_x{\mathrm{Se}}_{1-x}$ alloys. Fit results (black curves) to Doniach-Sunjic lineshapes (filled peaks) are provided in each panel. The composition $x$ of the alloys determined via the fitted peak components is stated in panels (b) and (c). Note the intensity scale varies between the panels.

Figure 6

VBM and PL peaks positions for ${\mathrm{GaS}}_x{\mathrm{Se}}_{1-x}$ alloys. ARPES measurements of the upper VB region for (a) GaSe, (b) ${\mathrm{GaS}}_{0.27}{\mathrm{Se}}_{0.73}$, and (c) ${\mathrm{GaS}}_{0.61}{\mathrm{Se}}_{0.39}$. The dotted line presents the fitted position of the VBM and the dashed curves exhibit the fitted dispersion around the VBM in each case. The corresponding VBM binding-energy positions are given in units of eV with an error bar of $\pm 0.03$ eV. (d) PL intensity for the same samples as in panels (a)–(c) and Fig. 5. Positions of the (e) PL peaks and (f) VBM as a function of composition.