We demonstrate that introducing a sputtered AlN buffer layer prior to epitaxial AlN-on-sapphire film growth reduces propagation losses below 0.2 dB/cm at 1550 nm and enables high- Q photonic devices, with promising prospects for short-wavelength applications.

Semiconductor membranes are widely used in research fields that target medical, biological, environmental, and optical applications. Often such membranes derive their functionality from a nanopatterning, which challenges the determination of their optical, electronic, mechanical, and thermal properties. In this work, we demonstrate the noninvasive, all-optical thermal characterization of approximately-800-nm-thick and approximately-150-& micro;m-wide membranes that consist mainly of wurtzite GaN and a stack of In0.15Ga0.85N quantum wells as a built-in light source. Because of their application in photonics, e.g., for vertical-cavity surface-emitting lasers, such photonic membranes are bright light emitters, which challenges their thermal characterization by optical means. We combine top-view two-laser Raman thermometry (2LRT) with time-resolved photoluminescence spectroscopy to extract the in-plane thermal conductivity kappa in plane of these membranes, which represents a notable difference from previous studies on epitaxial GaN films. Thus, we can disentangle the entire laser-induced power balance. Thermal imaging by Raman spectroscopy yields kappa in plane = 165+16-14 W m-1 K-1 for the best membrane. This result compares well with kappa in plane = 177 W m-1 K-1 obtained by ab initio simulations based on a solution of the linearized phonon Boltzmann transport equation, including three-and four-phonon scattering, as well as phonon-isotope and phonon-boundary scattering. Furthermore, we study how kappa in plane is affected by a roughening of the membrane's back side and additional semiconductor layers. For the membrane with the roughest back side, we observe a reduction of kappa in plane by almost 40%, which is accompanied by an anisotropy of kappa in plane due to etch channel formation. Thanks to the 2LRT approach, such variations and anisotropies of kappa in plane become accessible to the experimentalist via highly spatially resolved temperature maps.

Bright quantum emitters are key components for quantum communication systems. Radiative point defects in gallium nitride (GaN) are promising candidates for room-temperature single-photon emission, operating from the visible to the telecom O-band. Despite their potential, their integration into photonic structures has remained limited in the literature, with experimental photon extraction efficiencies far below simulated predictions. To identify the origin of this limitation, we investigate visible and near-infrared quantum emitters in GaN epilayers grown on c-plane sapphire substrates exhibiting narrow linewidths (∼4 nm), high photon count rates (>2 MHz), and strong antibunching, reaching g^(2)(0) values as low as 0.06 at room temperature. We find that these emitters are located near the GaN/substrate interface, explaining their limited coupling to optical modes. Building on this observation, we show that the insertion of a thin low-temperature GaN interlayer enables the formation of quantum emitters at arbitrary depths with sub-60 nm accuracy, independent of the substrate. The intentionally introduced emitters retain optical properties comparable to naturally occurring ones, including narrow linewidths (∼6 nm), saturation count rates exceeding 1.5 MHz, high Debye-Waller factors (0.69-0.98), and strong antibunching. The resulting epilayer fully coalesces within less than 250 nm ensuring compatibility with GaN-based cavity fabrication and enabling emitter placement within intrinsic regions of p-i-n diode architectures. These results mark a decisive step toward efficient emitter-cavity coupling, enabling the realization of cavity-enhanced quantum emission in GaN.

Off-axis electron holography enables a direct access to electrostatic potentials in semiconductor heterostructures, but its quantitative interpretation relies critically on accurate mean inner potential (MIP) values. Here, we combine off-axis electron holography with self-consistent electrostatic simulations to extract the MIP of InxGa1-xN/GaN heterostructures. By calibrating the surface Fermi-level pinning, experimental phase-shift profiles are quantitatively matched to simulations, allowing the MIP difference between GaN and InxGa1-xN to be determined. The derived MIP values deviate systematically from the linear interpolation of unstrained MIP values of GaN and InN, indicating a significant strain contribution. This strain effect is captured by an analytical strain-dependent model that reproduces the experimental MIP trend and yields a large intrinsic difference between the unstrained MIP values of GaN and InN of 3.0±1.7 V, which is consistent with density functional theory calculations. The approach provides a practical and computationally efficient route for determining strain-dependent MIP values, facilitating quantitative electron microscopy studies of strained semiconductor heterostructures.

Heating of semiconductor devices limits their performance and lifetime, which must be addressed by thermal management starting at the heat source. It is a common assumption that the heat source and the resulting heat spot locally coincide, if their size exceeds the mean free paths of the main heat carriers, the phonons. We show that this paradigm of heat locality breaks down on length scales spanning several micrometers. As a consequence, non-local heating occurs in contradiction to Fourier's law. Therefore, we heat laterally structured semiconductor membranes that feature a rising number of interfaces with a well-focussed laser and map-out lattice temperatures by Raman thermometry. Remarkably, the non-local heating can exceed the laser-induced local heating, which we attribute to ballistic phonon transport far above cryogenic temperatures.

InGaN/GaN-based light-emitting diodes (LEDs) exhibit remarkable tolerance to high threading dislocation densities (TDDs), maintaining internal quantum efficiencies above 90% at room temperature even with TDD exceeding 10 8 cm - 2. This robustness is attributed to V-shaped pits (V-pits), which form around threading dislocations (TDs) in InGaN/GaN quantum wells (QWs) and create potential barriers that inhibit carrier diffusion toward non-radiative dislocation cores. However, assessing the intrinsic non-radiative activity of dislocations in InGaN QWs is challenging, as standard metalorganic vapor-phase epitaxy-grown LED structures tend to exhibit V-pits. Those V-pits originate from the InGaN underlayer (UL) used to trap surface defects from high-temperature GaN growth, ensuring high QW efficiency. To investigate the genuine properties of TDs, we developed an InGaN UL that suppresses V-pits while maintaining surface defect-trapping capability. By adjusting growth conditions, we fabricated QWs with and without V-pits and compared their optical properties. Photoluminescence and cathodoluminescence measurements confirmed that TDs in InGaN QWs act as non-radiative centers. Time-resolved photoluminescence further revealed comparable capture cross sections for dislocations and point defects.

Abstract Off-axis electron holography enables a direct access to electrostatic potentials in semiconductor heterostructures, but its quantitative interpretation relies critically on accurate mean inner potential (MIP) values. Here, we combine off-axis electron holography with self-consistent electrostatic simulations to extract the MIP of InxGa1−xN/GaN heterostructures. By calibrating the surface Fermi-level pinning, experimental phase-shift profiles are quantitatively matched to simulations, allowing the MIP difference between GaN and InxGa1−xN to be determined. The derived MIP values deviate systematically from the linear interpolation of unstrained MIP values of GaN and InN, indicating a significant strain contribution. This strain effect is captured by an analytical strain-dependent model that reproduces the experimental MIP trend and yields a large intrinsic difference between the unstrained MIP values of GaN and InN of 3.0±1.7 V, which is consistent with density functional theory calculations. The approach provides a practical and computationally efficient route for determining strain-dependent MIP values, facilitating quantitative electron microscopy studies of strained semiconductor heterostructures.

Polarization engineering has played an important role in advancing III-nitride semiconductor devices over recent decades. However, the exact magnitude and orientation of polarization remain highly debated, as significant discrepancies persist among different theoretical approaches and experiments. Therefore, we combine off-axis electron holography with surface potential calibration and self-consistent electrostatic modeling to directly measure quantitatively polarization changes at InxGa1-xN/GaN interfaces. The results reveal a pronounced nonlinear, cubic dependence of spontaneous polarization on indium composition responsible for theoretical overestimation of polarization changes. Moreover, we show that only a layered hexagonal reference phase, rather than the zinc-blende phase, provides an accurate description of the experimental bowing of spontaneous polarization and hence a consistent theoretical framework of polarization in group-III nitrides. These results reconcile experiment and theory and provide a reliable basis for polarization engineering in InGaN alloys.

Semiconductor membranes are widely used in research fields that target medical, biological, environmental, and optical applications. Often such membranes derive their functionality from a nanopatterning, which challenges the determination of their optical, electronic, mechanical, and thermal properties. In this work, we demonstrate the noninvasive, all-optical thermal characterization of approximately-800-nm-thick and approximately-150- μ m -wide membranes that consist mainly of wurtzite Ga N and a stack of In 0.15 Ga 0.85 N quantum wells as a built-in light source. Because of their application in photonics, e.g., for vertical-cavity surface-emitting lasers, such photonic membranes are bright light emitters, which challenges their thermal characterization by optical means. We combine top-view two-laser Raman thermometry (2LRT) with time-resolved photoluminescence spectroscopy to extract the in-plane thermal conductivity κ in~plane of these membranes, which represents a notable difference from previous studies on epitaxial Ga N films. Thus, we can disentangle the entire laser-induced power balance. Thermal imaging by Raman spectroscopy yields κ in plane = 165 − 14 + 16 W m − 1 K − 1 for the best membrane. This result compares well with κ in plane = 177 W m − 1 K − 1 obtained by simulations based on a solution of the linearized phonon Boltzmann transport equation, including three- and four-phonon scattering, as well as phonon-isotope and phonon-boundary scattering. Furthermore, we study how κ in plane is affected by a roughening of the membrane’s back side and additional semiconductor layers. For the membrane with the roughest back side, we observe a reduction of κ in plane by almost 40%, which is accompanied by an anisotropy of κ in plane due to etch channel formation. Thanks to the 2LRT approach, such variations and anisotropies of κ in plane become accessible to the experimentalist via highly spatially resolved temperature maps.

In recent years, aluminum nitride (AlN) has emerged as an attractive material for integrated photonics due to its low propagation losses, wide transparency window, and presence of both second- and third-order optical nonlinearities. However, most of the research led on this platform has primarily focused on applications rather than material optimization, although the latter is equally important to ensure its technological maturity. In this work, we show that voids, which are commonly found in crystalline AlN-on-sapphire epilayers, have a detrimental role in related photonic structures, as they can lead to propagation losses exceeding 30 dB cm-1 at 1550 nm. Their impact on light propagation is further quantified through finite-difference time-domain simulations that reveal void-related scattering losses are strongly dependent on their size and density in the layer. As a possible solution, we demonstrate that when introducing a thin sputtered AlN buffer layer prior to initiating AlN epitaxial growth, void-free layers are obtained. They exhibit intrinsic quality factors in microring resonators as high as 2.0 × 106, corresponding to propagation losses lower than 0.2 dB cm-1 at 1550 nm. These void-free layers are further benchmarked for high-power applications through second-harmonic and supercontinuum generation in dispersion-engineered waveguides. Such layers are highly promising candidates for short-wavelength photonic integrated circuit applications, particularly given the strong potential of AlN for visible photonics. Given that volumetric scattering losses scale as λ-4, the platform quality becomes increasingly critical in the visible and ultraviolet range, where our improved layers are expected to deliver enhanced performance.

Applications such as optical clocks, gas spectroscopy, and optical coherent tomography stand to benefit significantly from the integration of on-chip supercontinuum sources. Materials exhibiting both second- and third-order nonlinearities are particularly attractive for this purpose, enabling the implementation of f-to-2f interferometry within a single device. Aluminum nitride is an especially promising material, featuring a wide band gap of 6.2 eV and transparency down to 200 nm. Through advancements in AlN epilayer quality achieved via metalorganic vapor-phase epitaxy on sapphire and the optimization of waveguide fabrication process, we drastically reduced absorption and scattering losses compared to their polycrystalline AlN counterparts. This achievement translates into propagation losses below 1 dB/cm at 1550 nm. Moreover, by tailoring the dispersion characteristics of the waveguide, we can precisely control the positions of short- and long-wavelength dispersive waves. This enabled the generation of gap-free supercontinuum spanning from visible to mid-infrared wavelengths, by pumping with telecom femtosecond laser. When pumped with TM polarization, efficient second-harmonic generation was achieved with the phase-matched higher-order mode TM20. For specific waveguide cross sections, the SH generated component overlapped with the dispersive wave, creating ideal conditions for f-to-2f interference.

The engineering of aluminum nitride (AlN) waveguides allows for broad supercontinuum generation (SCG) within its wide transparency window (200 nm - $5.5 \mu \mathrm{m}$) enabling applications in precision spectroscopy and metrology fields. An efficient coverage from 500 to 3500 nm was shown by pumping crystalline AlN waveguides with a femtosecond laser centered at 1560 nm [1]. UV and visible coverage was reached by pumping around 800 nm and by chirping the waveguide [2] or by exploiting the modal dispersion of higher-order modes [3]. An effective approach to further expand the SC range to the MIR lies in engineering the waveguide dispersion adjusting both its height and width, to extend the anomalous dispersion around the pump and to increase the spectral separation of long and short-dispersion waves (LWDW, SWDW). While the waveguide width is determined by the lithographic process used to transfer the chip design to the material substrate and is easily tunable with high precision, the height is fixed by the epitaxial growth process of the epilayer, limiting the choice of this dimension to the available commercial standards. Thanks to in-house crystalline AlN epitaxial layer growth facilities, we obtain high quality AlN layers up to $1.3\mu \mathrm{m}$ thick on $c$ -plane sapphire. This gives us an additional degree of freedom in the dispersion engineering of waveguides for SCG. In addition, we developed the fabrication process to obtain fully etched waveguides to improve mode confinement. This required adjusting the thickness of the lithographic mask without compromising the quality of the sidewalls. We achieve propagation losses of 0.24 dB/cm at 1550 nm as shown in the ring $Q$ -factor measurement in Fig. 1(a).

In the last decade, aluminum nitride (AlN) has proven to be an attractive material for both linear and nonlinear integrated photonics, due to its large transparency window, its low propagation losses and the presence of both second- and third-order nonlinear optical susceptibilities. However, the investigation and optimization of AlN layers for photonic applications have been scarcely explored in the literature so far, although it represents an important stepping stone upon which will impact the final performance of AlN-based devices. Here we present a systematic comparison of various types of AlN epilayers grown on sapphire by metalorganic vapor-phase epitaxy. Optical losses are measured from waveguides and microring resonators fabricated from different AlN epilayers. The results are analyzed through a comprehensive material characterization study, showing the need to optimize growth techniques to further push the performance of AlN-on-sapphire photonic devices.

We report an efficient extension of supercontinuum generation through dispersion engineering of crystalline aluminum nitride (AlN)-on-sapphire waveguides. Using a tailored epitaxial regrowth of AlN epilayers and an optimized fabrication protocol, the dispersion sensitivity to the waveguide cross-section was enhanced allowing for a significant reach extension of both short and long dispersive wave with optimized pumping conditions, reaching down to 550 nm in the visible and up to 4.5 µm in the mid-infrared.

In this paper, we investigate the nature of surface defects originating from the high-temperature (HT) GaN buffer and their incorporation into InGaN quantum wells (QWs) grown using the metalorganic vapor phase epitaxy technique. In particular, we conduct a detailed examination of the migration process of these defects from the HT-GaN buffer to the QWs, focusing on two potential pathways: diffusion versus surface segregation. A careful study confirms surface segregation as the dominant migration mechanism. To further understand the defect nature, we evaluate the stability of the HT-GaN surface under different annealing conditions, including different combinations of temperature and ammonia flow. We find that higher annealing temperatures or reduced ammonia flows significantly enhance the formation of defects, which speaks in favor of nitrogen vacancies (VN). Finally, we propose that these VN vacancies segregate toward the surface and interact with indium vacancies (VIn) in InGaN layers, forming VN–VIn divacancies. These VN–VIn divacancies could be the primary defects incorporated into InGaN layers acting as the main non-radiative recombination centers in InGaN QWs.

The electron optical phase contrast probed by electron holography at n-n+ GaN doping steps is found to exhibit a giant enhancement, in sharp contrast to the always smaller than expected phase contrast reported for p-n junctions. We unravel the physical origin of the giant enhancement by combining off-axis electron holography data with self-consistent electrostatic potential calculations. The predominant contribution to the phase contrast is shown to arise from the doping dependent screening length of the surface Fermi-level pinning, which is induced by FIB-implanted carbon point defects below the outer amorphous shell. The contribution of the built-in potential is negligible for modulation doping and only relevant for large built-in potentials at e.g. p-n junctions. This work provides a quantitative approach to so-called dead layers at TEM lamellas.

Thermal healing of focused ion beam-implanted defects in GaN is investigated by off-axis electron holography in TEM. The data reveal that healing starts at temperatures as low as about 250 °C. The healing processes result in an irreversible transition from defect-induced Fermi level pinning near the VB toward a midgap pinning induced by the crystalline-amorphous transition interface. Based on the measured pinning levels and the defect charge states, we identify the dominant defect type to be substitutional carbon on nitrogen sites.

We present a butt-coupled InGaN fiber Bragg grating (FBG) semiconductor laser diode operating below 400 nm in the single-mode emission regime. This compact coherent laser source exhibits an intrinsic linewidth of 14 kHz in the near-UV range and a side-mode suppression ratio reaching up to 40 dB accompanied by almost 2 mW output power. Furthermore, the properties of the FBG, including its central wavelength, bandwidth, and reflectivity, can be readily customized to fulfill specific requirements. As a result, the small footprint design of this laser is compatible with integration into a standard butterfly package to ease the lab-to-market technology transfer. The combination of low-frequency noise and fibered output signal positions these FBG laser systems as strong candidates for hybridization with integrated photonic platforms tailored for quantum information processing and metrology.

In recent years, point defects (PDs) have been unveiled as critical nonradiative recombination centres in InGaN/GaN quantum wells (QWs). When left unchecked, these nonradiative PDs can lead to at least an order-of-magnitude reduction in the internal quantum efficiency of blue light-emitting diodes. While macroscale studies have provided some information on such critical PDs, much deeper insight could be obtained by directly accessing the nanoscale impact of PDs on QW optical properties. Here, we present a detailed investigation of nonradiative PDs in a series of single InGaN/GaN QWs. Applying time-resolved cathodoluminescence (TRCL), we map the evolution of QW CL intensity spatially and temporally with nanometre and sub-nanosecond resolution, pinpointing individual PD locations. We fit the CL decays around single PDs with a carrier diffusion-recombination model to fully quantify their intrinsic properties, including novel phonon-limited relaxation times.

In this work, we report on the efficiency of single InGaN/GaN quantum wells (QWs) grown on thin (<1 µm) GaN buffer layers on silicon (111) substrates exhibiting very high threading dislocation (TD) densities. Despite this high defect density, we show that QW emission efficiency significantly increases upon the insertion of an In-containing underlayer, whose role is to prevent the introduction of point defects during the growth of InGaN QWs. Hence, we demonstrate that point defects play a key role in limiting InGaN QW efficiency, even in samples where their density (2–3 × 109 cm−2) is much lower than that of TD (2–3 × 1010 cm−2). Time-resolved photoluminescence and cathodoluminescence studies confirm the prevalence of point defects over TDs in QW efficiency. Interestingly, TD terminations lead to the formation of independent domains for carriers, thanks to V-pits and step bunching phenomena.