The core parts of the power transmission system, the source and the receiver, are fully GaN-based. The receiver is a 1 mm × 1 mm high periodicity MQW-based detector grown on c-plane single-side polished sapphire substrate by metal-organic chemical vapor deposition (MOCVD). Over an unintentionally doped GaN buffer layer, a 2 µm thick n-GaN:Si (5 × 1018
) is grown before the active region. The high periodicity MQW structure consists in 25 pairs of nominally undoped In0.15
N/GaN quantum wells, 2.2 and 4.8 nm thick, respectively. The p-type layer is 100 nm thick and doped with Mg (5 × 1017
). A schematic sketch of the structure is reported in Figure 1
a. The contact region is composed of a semi-transparent Ni/Au current spreading layer with Ni/Au grids. Additional details on device structure, growth and processing conditions, quality and performance can be found in [7
]. The source is a high power 405 nm laser diode (nominal optical output power > 2 W at 1.5 A, manufacturer, city, state, country), chosen in order to excite the useful spectral range of the detector, shown in Figure 1
The external quantum efficiency of the receiver is about 17% at the 405 nm working wavelength [7
], significantly limiting the performance of the whole system. In order to understand the origin of the losses, we carried out a characterization of the epitaxial material before processing. Photoluminescence (PL) measurements, both integrated and time-resolved, were performed exciting the samples with the third-harmonic (λ = 355 nm) of a 5 ns pulsed Nd:YAG laser (Brilliant, Quantel, Les Ulis, France) at a repetition-rate of 10 Hz. The luminescent emission was spectrally selected by a single grating monochromator (Oriel MS257, Newport Corporation, Irvine, CA, USA) and detected by a photomultiplier tube (R928, Hamamatsu, Hamamatsu City, Japan). A transient digitizer (TDS 7104, Tektronix, Beaverton, OR, USA) was used to record the PL signal evolution as a function of time to obtain the temporal decay curves. The photoluminescence (PL) spectrum, reported in Figure 2
a, shows a significant emission from the deep levels responsible for the yellow luminescence band in GaN, supposed to be gallium vacancies (VGa
], enhanced by the presence of impurities such as carbon [9
] or oxygen [12
]. The luminescence may originate from the quantum wells, suggesting a high defect density that may lower the efficiency due to defect-assisted recombination, or from the upper p-GaN layer, therefore leading to a reduction in efficiency caused by defect-mediated absorption of the transmitted power before the quantum wells. By analyzing the time-resolved decay of the PL at 460 nm, i.e., in a spectral region compatible with band-to-band recombination inside the quantum wells (see Figure 2
b), it is possible to notice a long time constant that lowers at higher excitation levels, possibly caused by the low residual carrier density inside the quantum wells after the discussed defect-assisted absorption in the p-GaN layer.
The assembled system is shown in Figure 3
. The laser diode is housed in a Peltier-cooled TCLDM9 high power mount connected to an ITC4005 combined laser diode and TEC controller. The mount also holds the C440TMD-A collimating lens. The laser beam is split by a BSF10-A UV fused silica beam sampler. A small part of the light is diverted to a PDA36A-EC photodetector, and a DG10-1500-A N-BK7 ground glass diffuser is used in order to reduce the optical power density on the photodiode, leading to improved performance and stability. The photodetector optical branch is used as a feedback system connected to the ITC4005 controller, ensuring that the optical power reaching the receiver is constant over time. The laser beam passing through the beam sampler is then directed to the receiver by a CM1-E02 dielectric coated turning prism mirror. The receiver is housed in a custom sample holder, mainly composed of a HT24S metal ceramic heater and a TH100PT 100 Ohm platinum resistance temperature detector connected to a TC200-EC temperature controller, used to stabilize the temperature of the device. The various optical power densities were obtained by changing the optical power setpoint in the feedback branch, i.e., by varying the bias current of the laser. For this reason, the size of the laser spot on the device under test is not constant and was measured for every setpoint. The power of the optical beam on the target plane was obtained for every setpoint by means of a calibrated S130VC photodiode power sensor and of a calibrated S142C integrated sphere photodiode power sensor, both connected to a PM100USB USB power meter interface. All the part numbers refer to products of the manufacturer Thorlabs (Newton, NJ, USA). The total length of the optical path, in this case, is 40 cm, but taking into account the good collimation and spatial coherence properties of a laser beam it can be easily increased.