Stripped-back representative VCSEL devices with a simple fabrication process that very closely approaches the performance of standard BCB-planarised devices have been produced. These VCSEL Quick Fabrication (VQF) devices achieve threshold currents only 0.3 mA higher than that of a standard device produced from the same material. The predictability of standard performance from VQF performance is also robustly assessed in terms of temperature effects to account for the observed disparities. These VQF devices are then processed across a 6-inch (152 mm) wafer and the resulting device-level characteristics are mapped. From this, it is apparent that there is an approximately radial decrease in oxide aperture diameter from centre to edge, found to be driven by the strain-induced wafer bow. After corrections, a residual spatial variation across the wafer remains, which, in conjunction with temperature dependent measurements, is shown to be a result of epi-material variation. By observation at 50 °C, that is, at a temperature closely resembling that of intended application, the residual centre-to-edge variation in threshold current density is found to be only 0.2 kA/cm 2 , compared to 1.3 kA/cm 2 when observing the room temperature variation of devices of nominally equivalent active volumes.
13 March 2023 in Journal of Physics D: Applied Physics
Vertical cavity surface emitting laser (VCSEL) devices and arrays are increasingly important in meeting the demands of today’s wireless communication and sensing systems. Understanding the origin of non-uniform wet thermal oxidation across large-area VCSEL wafers is a crucial issue to ensure highly reliable, volume-manufactured oxide-confined VCSEL devices. As VCSEL wafer diameters approach 200 mm, germanium (Ge) is emerging as an alternative substrate solution. To this end, we investigate the uniformity of 940 nm-emitting VCSEL performance across 150 mm diameter GaAs- and Ge-substrates, comparing the oxidation method in each case. Nominally identical epitaxial structures are used to evaluate the strain induced wafer bow for each substrate type with Ge exhibiting a reduction of over 100 μm in the peak-to-valley distortion when compared with GaAs. This wafer bow is found to be the principal cause of centre-to-edge oxidation non-uniformity when utilising a conduction-heated chuck furnace, in comparison to a convection-heated tube furnace. Using on-wafer testing of threshold current, differential resistance, and emission wavelength, device performance is demonstrated for the first time across a 150 mm Ge wafer, and is shown to be comparable to performance on GaAs substrates, when the effects of oxidation uniformity are removed. These results provide evidence that there is a realistic path to manufacturing high yield VCSELs, over wafer diameters approaching those used in Si-photonics, via Ge substrates.
This letter reports on InP/GaInP quantum dot mode-locked lasers emitting in the 730 nm wavelength region, extending the spectral range of previously reported monolithic mode-locked edge-emitting lasers. Modal gain and absorption measurements were used to identify a relatively broad spectrum which is utilised to support passive mode-locking in a monolithically integrated two-section ridge laser. The conditions for mode-locking were explored by varying the current to the gain section and reverse bias to the absorber section. For a total cavity length of 3 mm, the shortest pulse train observed was 6 ps in duration with a repetition rate of 12.55 GHz.
We demonstrate high temperature operation of InAs quantum dot lasers through active region engineering. We grew an n-doped region within the InAs quantum dot (QD) layer and incorporated a 10 nm p-GaAs modulation-doped layer within a 43 nm GaAs spacer layer, both of which are essential components of the active region of the InAs QD laser device. The co-doping scheme enables the demonstration of InAs QD laser device with only seven layers of InAs QDs in the active region for high temperature operation, which is compared with a conventional undoped InAs QD laser device. A Fabry-Pérot laser device with as-cleaved facets and a co-doped InAs QD active region enables ultra-high temperature pulsed-biased O-band laser operation up to 202 °C, compared to 180 °C for conventional undoped InAs QD laser, and without requiring high-reflective facet coatings.
This Letter reports on the temperature-dependent optical gain and absorption features, including the quantum confined Stark effect, of an InAs/InGaAs quantum-dash laser directly grown on a (001) Si substrate, with the lasing wavelength within the range of 1.5–1.6 μm. The maximum optical net gain was 22 cm−1, and the internal optical loss was ∼–17 cm−1 at 20 °C. Measurements as a function of injection level indicate that while the required current densities are still high, the intrinsic performance is significantly better than that of similarly operated InAs quantum dots operating at 1.3 μm, and further effort on growth could be made to reduce the internal optical losses and non-radiative current density. Optical modal absorption spectra were measured as a function of reverse bias from 0 V to 6 V, and a 40 nm redshift was observed in the absorption edge due to the quantum confined Stark effect, suggesting potential applications of these materials in electro-absorption modulators grown on silicon.
31 January 2023 in Journal of Physics D: Applied Physics
We report direct measurements of the optical gain on vertical-cavity surface-emitting laser (VCSEL) material using a stripe-length method featuring segmented contacts. We utilise the similarity of the in-plane transverse electric (TE) polarised matrix element and that of the VCSEL lasing mode and a simple method to reduce round trip effects. The confinement factor is determined from cold-cavity simulations of the in-plane TE polarised slab waveguide mode and used to convert the measured in-plane modal gain into the vertical-cavity modal gain, as required for the VCSEL structure. This gives a threshold material gain of 1440 ± 140 cm−1 at 30 °C for this structure. A comparison with the threshold material gain values determined from the lasing condition, where internal optical losses due to doping induced absorption is included using parameters taken from the literature, indicates the presence of an additional source of optical loss in the experiment which increases the threshold material gain by ∼450 cm−1. A best fit is obtained by increasing the optical loss in the n-DBR (distributed Bragg reflectors) layers to 40 cm−1, which is consistent with previous work on additional scattering losses due to interface roughening in the n-DBR layers. To further demonstrate the utility of this method for rapid optimisation, the gain-peak wavelength is measured directly, and its temperature dependence is compared to the lasing wavelength.