Broadband Near-Infrared Emission through Monolithic Vertical Cavities Fabricated by Rotational Metal-Mask Selective-Area Growth

Near-infrared surface-emitting light sources is vital for modern life and they are inside the devices and systems we use every day and carrying data through optical fibers, scanning our faces to unlock a phone, or probing tissues in medical imaging setups. The principle behind them is simple enough: light confined between two highly reflective mirrors, known as distributed Bragg reflectors, and released through a small aperture at the surface. However, this seemingly straightforward architecture—the vertical cavity—took decades of refinement to reach its current stability and efficiency. The resulting VCSELs are compact, reliable, and surprisingly versatile. Still, as optical engineers try to build sources that emit several wavelengths on one chip, the conventional fabrication methods start to show their limitations

What once worked for single-mode devices becomes awkward when precision, scalability, and wavelength diversity must coexist within the same wafer. Traditionally, producing several emission wavelengths on one chip has relied on either spatially varying the cavity thickness during epitaxy or engineering multiple quantum wells (QWs) with slightly different bandgaps. These strategies, though elegant in principle, suffer from poor scalability and material uniformity. Gradients created across wafers cannot easily yield distinct, well-defined cavities within a confined area, and QW-based emitters possess intrinsically narrow linewidths, requiring numerous meticulously tuned layers to cover even a modest spectral range. Methods based on metal-organic chemical vapor deposition with patterned masks have offered partial wavelength selectivity, but they involve repeated regrowth and lithography steps that complicate processing and reduce yield.

To this account, new research paper was published in Applied Physics Express, conducted by Yuuki Hodson, Tatsuki Yokota, and led by Professor Nobuhiko Ozaki from Wakayama University alongside Dr.  Eiichiro Watanabe and Dr. Naoki Ikeda from the National Institute for Materials Science.The researchers simulated cavity–reflectance framework predicting the relationship between GaAs layer thickness and resonant wavelength and realized fabrications of monolithic vertical cavities

via rotational metal-mask selective-area growth. These cavities incorporated stacked InAs quantum dots with different emission peak wavelengths to achieve broadband near-infrared emission resonating at multiple discrete wavelengths. The innovation lies in producing four distinct vertical cavities during a single uninterrupted molecular beam epitaxy process, eliminating post-growth patterning or regrowth steps while maintaining precise optical control across the wafer.

The team’s central idea was to examine this broadband QD emission while locally controlling the GaAs cavity thickness through selective-area growth. By introducing a rotatable metal mask with asymmetric windows during MBE growth, they could deposit additional GaAs layers only on chosen regions of the wafer, forming cavities of different optical lengths without breaking vacuum or exposing the surface to air. This approach promised a simplified, one-step fabrication process for creating multiple vertical cavities—each tuned to a distinct resonance wavelength—within the same monolithic structure. The motivation was not merely to demonstrate an optical novelty but to propose a realistic fabrication strategy for next-generation, compact, multi-wavelength VCSEL arrays that could serve telecommunication and biosensing applications alike.

The team began with numerical simulations using the Cavity Modeling Framework to predict how the reflectance spectrum of a GaAs-based cavity changes with thickness. A model consisting of 10.5 alternating AlAs/GaAs DBR pairs (96 nm / 82 nm) topped by a GaAs cavity containing stacked InAs QDs was analyzed. The simulation revealed a broad photonic stopband between 1.0 µm and 1.2 µm in wavelength, within which specific dips in reflectance corresponded to resonant cavity modes. Four target cavity thicknesses—347, 362, 377, and 392 nm—were selected to align these resonance wavelengths with the emission spectrum of InAs QDs, ensuring that each cavity would favor a slightly different mode around 1.13–1.18 µm. The researchers grew the structures on n-type GaAs(100) substrates using MBE. The bottom DBR was first deposited, followed by a 347 nm GaAs layer incorporating three stacked QD layers. Each QD layer was formed by supplying 2 monolayers of InAs at 480 °C, but capped differently to tune emission: one with GaAs at 0.5 ML/s, another at 1.0 ML/s, and the third with In₀.₂Ga₀.₈As at 1.25 ML/s. This combination produced broadband emission covering the simulated stopband. The authors then introduced a rotational metal mask containing asymmetric windows mounted on a rotatable holder inside the MBE chamber to generate cavities of variable thickness and by rotating the mask 90° between depositions, additional GaAs layers were selectively grown in four distinct regions, creating the four designed cavity thicknesses in one uninterrupted process. Finally, a three-pair AlAs/GaAs top DBR was deposited across the surface. Afterward, they performed optical characterization through reflectance and micro-photoluminescence (PL) mapping which confirmed the success of this selective-area growth. Distinct dips appeared in the reflectance spectra at 1118, 1129, 1137, and 1157 nm for VC1–VC4, and PL peaks matched these values within experimental error. Each cavity’s emission red-shifted progressively with increasing thickness, exactly as predicted. PL intensity and wavelength mapping further revealed homogeneous emission within each selective-growth area, with minor gradients attributed to natural MBE thickness variation. PL intensity from the MM-SAG sample rose almost linearly and reached nearly 100-fold enhancement relative to a QD reference sample without a cavity. This dramatic gain confirmed that the light was effectively confined and amplified by the vertical cavity modes. The emission remained stable and continuous across all regions, suggesting consistent optical thresholds suitable for future VCSEL operation.

In conclusion, the new study by Professor Nobuhiko Ozaki and colleagues is an important advancement toward practical broadband, multi-wavelength NIR light sources fabricated by a single epitaxial process. The rotational MM-SAG approach eliminates the multi-step lithography and regrowth traditionally required for wavelength multiplexing. By simply varying the GaAs cavity thickness within one continuous MBE run, the team realized four distinct optical resonators—each resonating at a different wavelength yet seamlessly integrated on the same wafer. This efficiency could significantly reduce manufacturing complexity and cost for VCSEL arrays used in compact optical communication modules and biometric sensors. The innovative work also demonstrates the remarkable synergy between InAs quantum dots and the selective-area growth concept. The broadband emission of the stacked QDs allowed simultaneous resonance with multiple cavity modes, achieving roughly 40 nm of wavelength separation—double that of conventional quantum-well-based VCSELs. This capability is critical for expanding data-transmission bandwidths in wavelength-division-multiplexed systems, where precise yet broad coverage of the NIR region is essential. Moreover, the use of excited-state (ES) emissions, which are less susceptible to ground-state reabsorption, provides an intrinsic advantage for achieving efficient lasing once electrical injection is introduced.

Additionally, the MM-SAG method paves the way for dense, chip-scale VCSEL integration. The size of the metal-mask windows can be miniaturized to several hundred micrometers, enabling compact arrays emitting at programmable wavelengths without device-to-device processing variability. Such architecture could underpin future “smart-sensing” platforms—devices capable of multi-spectral illumination for tissue diagnostics, chemical detection, or environmental monitoring. Additionally, the approach is inherently compatible with standard GaAs-based optoelectronic fabrication lines, making industrial translation plausible. We believe the innovative method also offers a new way into cavity–emitter coupling control via geometric confinement rather than compositional tuning. The pronounced photoluminescence enhancement and consistent mode behavior observed across different regions validate the precision of MM-SAG for tailoring optical microcavities. Importantly, the results illustrate that broadband QD ensembles can replace complex multi-QW stacks without sacrificing spectral flexibility. Indeed, such simplification could re-shape the design philosophy of multi-wavelength photonic devices and shift emphasis from material diversification to cavity-geometry engineering. Ultimately, the study establishes a clear pathway for developing compact, broadband, and tunable NIR emitters using rotational selective-area growth, and offers a bridge between fundamental semiconductor optics and applied photonic device engineering.