Abstract
The mid-wave infrared (MWIR) spectrum, typically referring to the wavelength range of 2 - 5 µm, has the potential to enable next-generation breakthroughs in photonics applications. Therefore, high-performance photodetectors that operate in this range are of high interest. However, high signal-to-noise ratio (SNR), a critical figure of merit for photodetectors, is extremely difficult to achieve in the MWIR range. The challenge is twofold, one is the low signal power level, determined by the application scenario in the MWIR range, and the other is the high noise level, which originates from the high dark current of the narrow bandgap material. To address these challenges, my PhD research focuses on developing high performance MWIR photodetectors based on the AlInAsSb digital alloy materials system. The primary goal is to achieve high SNR. The solution is twofold.
The first approach is to achieve a high signal level by using the internal gain of avalanche photodiodes (APDs). To maximize the SNR performance, I designed a new separate absorption, charge, and multiplication (SACM) APD with an ultra-thin absorber to suppress dark current, while the quantum efficiency of the device is maintained at a high level by using photo-trapping structures. The device exhibits dark current level more than 2 order lower than the previous record result at 2 µm, and the quantum efficiency reaches ~ 22 % by using surface metal gratings and ~ 24 % through edge coupling. Moreover, the maximum gain of the device is as high as ~ 1000 at 240 K. Consequently, record high SNR performance is demonstrated by the thin SACM APD. Additionally, the thin absorber design also improved the speed performance. The frequency response measurement shows that the maximum bandwidth reaches ~ 7 GHz, and the gain-bandwidth product is over 200 GHz. Both of these two values are more than 4 time higher than the previously record bandwidth for 2 µm APDs. In the future, this idea of using the combination of an ultra-thin absorber and photon trapping structures can be applied to longer wavelength cut-off APDs in the MWIR range.
Instead of boosting up the signal level by using an APD, another solution to improve the SNR is to suppress the dark current by using an n-barrier-n (nBn) photodetector design. Although nBn photodetectors have been substantially improved using superlattice materials systems, their performance is limited by their valence band discontinuity, especially in the long wave infrared range. However, the AlInAsSb digital alloy materials system has been found to have minimal valence band discontinuity, which provides the potential to improve the performance nBn photodetectors fabricated from conventional materials. Therefore, I designed and demonstrated the first AlInAsSb nBn, and the detectivity of the device at room temperature achieved record high value at 2 µm. Additionally, since the target of the previous research on nBn photodetectors is high SNR, the frequency behavior and bandwidth have not been studied. To enable high-speed applications, I proposed the first equivalent circuit model and the bandwidth theory for the nBn photodetector. The new model provided good fits to measurements, and moreover, it reveals the limiting factor of the bandwidth of nBn photodetector and the methods to improve the frequency response. This work lays the foundation for extending the application sphere of nBn photodetectors to high-speed scenarios in the MWIR range.
Additionally, I developed a special double mesa approach that is easy and straightforward to determine the background doping density and polarity of an unintentionally doped layer, which is a critical parameter for the design of both the SACM APD and the nBn photodetectors in my PhD work. Moreover, my research on the double mesa also reveals its potential to suppressing surface leakage dark current and extend surface breakdown voltage, which will be a promising area of research in the future.
