The III-Se layered semiconductors, including InSe and GaSe, are promising optoelectronic materials due to their relatively high electron mobilities at room temperature, nonlinear optical responses, ferroelectricity, self-passivated van der Waals surfaces, and ability to alloy and synthesize heterostructures for bandgap engineering. Adsorption control is a widely utilized strategy for controlling the stoichiometry and phase formation of these materials; however, the bounds of the adsorption-controlled growth window for GaSe have not been systematically established. Additionally, challenges with control over polytype and twinning remain. Here, we use molecular beam epitaxy to experimentally map the adsorption-controlled growth window of GaSe films on vicinal GaAs (111)B substrates. The observed phase boundaries show qualitative agreement with Ellingham diagram predictions. All films crystallize in the $γ$ ($R3m$) polytype. Increasing growth and annealing temperature leads to decreased mosaicity measured by x-ray rocking curve and smoother surfaces measured by atomic force microscopy, at the expense of a transition from singly oriented $γ$ to twinned $γ$ with $60\degree$ rotated domains.
Alex: Welcome to another episode of ResearchPod. Sam, what paper are we diving into today?
Sam: This is a study from Joshua Eickhoff and colleagues on growing thin films of a material called gallium selenide, or GaSe, using a technique known as molecular beam epitaxy. The central puzzle they tackle is how to make these films with the exact right balance of atoms and without harmful defects that ruin their performance in electronic devices.
Alex: So this paper is basically mapping out the conditions where GaSe grows perfectly, right? And the core problem before was getting mixed-up phases or these twin defects?
Sam: Yes, exactly. Prior attempts at growing GaSe this way often ended up with a mix of GaSe and another phase called Ga2Se3, or uncontrolled twinning where parts of the crystal rotate and misalign. These issues scatter electrons and trap light, hurting the material's use in things like fast transistors or light detectors.
Alex: Twinning sounds like a big deal—what makes it so problematic for devices?
Sam: Twin boundaries are like mismatched seams in a fabric; they block electrons from moving freely and cause light energy to get lost as heat instead of being useful. The study promises materials with high electron mobility—meaning electrons zip through easily—and good light-handling, but only if you avoid these twins and keep the atom ratio locked at one-to-one for GaSe.
Alex: Okay, so the practical challenge is nailing that perfect recipe during growth. Like trying to bake a cake where one ingredient evaporates easily, and too much or too little ruins it?
Sam: That's a solid way to picture it. They use beams of gallium and selenium atoms shot onto a heated surface—think of it as gently layering ingredients in a vacuum oven. The key is finding the "window" of temperatures and selenium amounts where extra selenium just floats away, self-locking the GaSe composition perfectly. Researchers call this approach adsorption-controlled epitaxy, similar to how GaAs grows reliably.
Alex: Right, and without that map, you're guessing in the dark. What did past work miss?
Sam: Earlier efforts didn't systematically test the boundaries, so films had drifting ratios—too much gallium makes Ga-rich junk, too much selenium shifts to Ga2Se3. This paper maps it experimentally, matching predictions from a thermodynamic chart called an Ellingham diagram, which plots when phases are stable based on temperature and vapor pressure.
Alex: Huh, so it's like a recipe chart showing safe cooking temps to avoid burning or sogginess.
Sam: Precisely. And there's a temperature tradeoff: cooler growth around 400 degrees Celsius gives single-orientation crystals without twins, but rougher surfaces; hotter temps smooth it out and align better, but introduce those 60-degree twins. The study confirms all films are the gamma polytype—a specific stacking of layers like Se-Ga-Ga-Se blocks with weak gaps between.
Alex: That tradeoff makes sense for why control matters. So they grow on GaAs substrates to leverage similar crystal habits?
Sam: Yes, they use vicinal GaAs substrates—slightly tilted GaAs crystals cut at a small angle to create steps that guide the GaSe layers to align properly. This helps the GaSe grow in large, uniform sheets matching the GaAs underneath.
Alex: So the substrate sets the stage for good alignment from the start. How do they keep an eye on what's happening during growth?
Sam: They shine a beam of high-energy electrons at the growing surface and watch the diffraction pattern it makes—like a real-time X-ray of the top layer to spot when the structure is forming right. This reflection high-energy electron diffraction, or RHEED, shows clean GaSe patterns in the right conditions, but mixed or wrong phases if the temperature drops too low or selenium flux rises. It confirmed the phase changes match predictions: pure GaSe in the sweet spot, shifting to mixtures or pure Ga2Se3 otherwise.
Alex: That real-time check sounds crucial for nailing the window. What about making the crystals even better ordered?
Sam: Higher growth temperatures smooth the surface and align the crystal blocks better—the bumps get smaller, and the spread in orientations narrows, as seen in atomic force microscope images and a measure called rocking curve width from X-rays, where a skinnier curve means the whole film points more uniformly. But above 450 degrees Celsius, sticking drops so atoms don't land well, leading to no growth. To fix that, they grow at 400 degrees where atoms stick reliably, then anneal—heat the film to 520 degrees under balanced gallium and selenium fluxes to reorder without much loss.
Alex: So the anneal polishes it up without starting over. Does that keep the single orientation, or bring back twinning?
Sam: Not quite—the 400-degree growth alone gives a single orientation of the gamma polytype, with layers stacked ABC-ABC like offset blocks of Se-Ga-Ga-Se. But higher growth temps or the anneal introduce twins, confirmed by direct atom images from HAADF-STEM microscopy—showing bright gallium and dimmer selenium atoms—and electron diffraction patterns over larger areas revealing rotated domains.
Alex: Huh, so there's still that temperature compromise: better quality but twins creeping in. The paper notes further work needed on surfaces to control it?
Sam: Exactly—a meaningful step in mapping the tradeoffs, but smoothing the substrate more or defining steps better may help suppress twinning. The evidence points to kinetic barriers overcome at high temperatures driving those rotations, with the anneal suggesting a solid-state reorientation.
Alex: How did they actually spot the difference between single orientation and twinned crystals?
Sam: They used a tool that shoots electrons through a tiny selected spot on the sample and looks at the pattern of scattered electrons, like taking a fingerprint of the local crystal arrangement to see if all parts face the same way. For films grown at 400 degrees Celsius, the pattern shows sharp spots in just one set of positions, meaning a single orientation. At 450 degrees or after annealing to 520 degrees, extra spots appear at 60-degree rotations, confirming twinned domains. Researchers call this selected area electron diffraction, or SAED.
Alex: Okay, so the pattern splits when twins form—like extra shadows showing misaligned sections.
Sam: Exactly. They also rotated the sample under X-rays and measured the intensity around a specific reflection, referenced to the substrate. A single sharp peak means uniform orientation; broadening or splits indicate twins. These azimuthal phi scans matched the SAED, showing clean alignment at low temperature but twinning higher up.
Alex: That lines up with the temperature tradeoff you mentioned. What about the exact layer stacking in these gamma films?
Sam: All films showed the gamma polytype, where layers stack in an ABC-ABC repeat—like shifting books on a shelf so each tetralayer of Se-Ga-Ga-Se sits offset from the one below, held by weak van der Waals forces between them. Simulations of electron diffraction patterns confirmed this rhombohedral structure over other stackings. Direct images from a microscope that scans electrons and highlights heavy atoms like gallium brighter also visualized the ABC repeat without twins in the low-temperature sample.
Alex: Huh, so HAADF-STEM gives that atom-by-atom view to nail the structure.
Sam: Yes, and it proved the film is purely gamma GaSe in that case. The annealing at 520 degrees improved ordering without thinning the film—X-ray reflectivity showed thicknesses around 43 to 48 nanometers before and after, meaning minimal atom loss.
Alex: Stable thickness during anneal keeps the benefits without sacrificing material. Does this mapping open doors for better devices?
Sam: It does—the precise window ensures self-stoichiometric GaSe, avoiding phase mixes that scatter carriers. This supports high-mobility optoelectronics, as uniform gamma films reduce recombination losses. The paper suggests it's a notable advance for van der Waals semiconductors, though suppressing twins fully remains key.
Alex: Well put, Sam. This careful mapping shows the path forward without overpromising. Thanks for breaking it down so clearly—that's it for our look at growing GaSe thin films with molecular beam epitaxy. Thanks for listening to ResearchPod.