Producing InAs quantum dots on InP(001) requires control over more than material composition. Growth conditions influence whether the resulting structures are three-dimensional dots, elongated features or rings, as well as their density and optical behavior. A droplet epitaxy process investigated with solid-source molecular beam epitaxy (MBE) demonstrates a route to forming InAs nanostructures directly on InP, without an intermediate InGaAs or InAlAs layer.
For semiconductor process engineers and device developers, the work highlights a familiar manufacturing challenge: parameters that help control structure formation can also affect material quality. In this case, substrate temperature shapes the indium droplets and their subsequent crystallization, while thermal treatments can improve photoluminescence but also alter the nanostructures themselves.
How droplet epitaxy forms InAs nanostructures
The process begins with an InP buffer layer grown on an InP(001) substrate. Indium is then deposited to form metallic droplets, followed by exposure to arsenic, which crystallizes the droplets into InAs structures. In the reported experiments, indium was deposited at temperatures from 50°C to 300°C, with the same nominal indium coverage and deposition rate across the temperature series.
This approach differs from conventional InAs growth on InP(001), which can produce dashes or wires elongated along the [1-10] direction. Droplet-mediated growth provides another way to obtain three-dimensional structures on the (001) surface, while avoiding the need for an intermediate semiconductor layer. That direct material configuration is relevant to device development where the layer stack is an important design consideration.
The resulting nanostructure density depends on the density of the original indium droplets. Droplet density, in turn, is influenced by the surface diffusion and coalescence of indium atoms. In the experiments, the measured density changed from approximately 5 × 107 to 4 × 108 cm-2 across the investigated temperature range. The trend followed an exponential dependence on inverse substrate temperature, with a fitted activation energy of 0.126 ± 0.010 eV.
Substrate temperature changes shape and diffusion
Atomic force microscopy showed that arsenic exposure transformed the droplets into ring-like structures. Their morphology varied with the temperature used during indium deposition and crystallization. At some temperatures, rings were surrounded by disk-shaped features; at the lowest and highest temperatures studied, these surrounding disks were not clearly observed.
The inner structures also changed in size and shape. At lower temperatures, the rings were more complete and included distinct raised features. With increasing temperature, the ring profile became lower and more fragmented. The diameter of the inner structures varied across the samples, while the height and extent of the surrounding disks also depended on growth temperature.
The researchers interpreted the surrounding disks as material formed when indium atoms spread away from the droplets and combine with incoming arsenic. Their radius therefore provides an indication of indium diffusion length. Between 150°C and 260°C, the measured diffusion length increased from about 160 to 400 nm along [1-10], and from about 100 to 300 nm along [110]. The difference between directions indicates anisotropic surface diffusion under these growth conditions.
Analysis of the temperature dependence produced activation energies of 0.28 ± 0.01 eV along [1-10] and 0.25 ± 0.02 eV along [110]. The fitted exponential prefactor was about four times larger along [1-10]. These values offer process-relevant measures of how indium migrates across the InP(001) surface in the temperature range studied. The results also show why temperature cannot be treated as a simple density control: it affects droplet size, diffusion distance and the geometry formed during crystallization.
Optical performance depends on thermal history
Photoluminescence measurements were made on selected samples after InP capping. The measured emission from the nanostructures fell in the approximate 1.3–1.5 μm range, which is of interest for telecommunications-wavelength emitters. However, the as-grown samples showed weak optical signals, which the study associated with limited crystalline quality under the low-temperature growth conditions.
Post-growth annealing at 600°C for 10 minutes substantially increased the photoluminescence intensity and shifted emission toward shorter wavelengths. The response was not simply an improvement without trade-offs. In the sample grown at 50°C, localized emission features appeared after annealing, consistent with emission from quantum dots, although spectral diffusion remained significant. In the sample grown at 180°C, the structures appeared to lose their three-dimensional character and merge into an extended layer, producing more spatially uniform emission associated with quantum wells.
This contrast is important for process development. Annealing can improve optical quality, but it can also change composition, dimensions and confinement. A thermal recipe that benefits one starting morphology may not preserve another. Growth and annealing therefore need to be assessed as a combined sequence rather than as independent steps.
Process implications for semiconductor manufacturing
The study identifies several linked control variables for droplet epitaxy on InP(001): indium coverage and flux, substrate temperature, arsenic exposure, capping conditions and post-growth thermal treatment. Within the reported process, droplet density provides a route to adjusting nanostructure density, while AFM morphology offers a way to track how diffusion and crystallization respond to temperature.
- For density control: substrate temperature changes the density of droplets and the nanostructures formed from them.
- For morphology control: temperature influences ring completion, feature height and the formation of surrounding disks.
- For optical quality: annealing can strengthen emission, but may also reshape or merge small structures.
These findings establish process relationships, not a universal production recipe. The practical task is to identify conditions that balance density, three-dimensional confinement and optical quality for the intended device. Further refinement of crystallization and capping conditions would be needed to preserve the desired structure while improving its emission properties.
Conclusion
Droplet epitaxy demonstrates a way to form InAs nanostructures directly on InP(001), with substrate temperature providing a significant means of controlling droplet density, indium diffusion and final morphology. The optical results show that thermal treatment can improve emission, but may also compromise three-dimensional confinement. For engineers developing semiconductor growth processes, the central implication is to optimize the full sequence—from droplet formation through crystallization, capping and annealing—against both structural and optical requirements.