Reducing parasitic capacitance in single‐photon avalanche diodes

Abstract

Page ***, ‘Triple epitaxial single-photon avalanche diode for multichannel timing applications’, Ivan Labanca, Francesco Ceccarelli, Angelo Gulinatti, Massimo Ghioni and Ivan Rech. Cross section of the new SPAD structure. A lightly doped n- Intermediate Layer has been introduced between the Buried Layer and the Substrate to reduce the parasitic capacitance. An array of SPADs that have been fabricated using the new custom technology. A single-photon avalanche diode with picosecond temporal resolution and low parasitic capacitance is presented by researchers at the Politecnico di Milano, Italy. Single-photon avalanche diodes (SPADs) are solid-state devices that provide a digital current pulse every time they are hit by a single photon. They record not only the number of photons impinging on the detector but they also determine the arrival time of each photon within an accuracy of tens of picoseconds. SPAD development is driven by the widespread and growing interest in low-level light detection in several scientific and industrial applications, such as fluorescence spectroscopy in life and material sciences, quantum computing and cryptography, profiling of remote objects with optical radar techniques, particle sizing, and more. In particular, the use of fluorescence lifetime spectroscopy as an analytical research tool has increased substantially in recent years, finding remarkable applications in chemistry, biochemistry and biology. In this Issue of Electronics Letters, Francesco Ceccarelli and colleagues have modified the typical structure of SPADs to introduce an additional layer that reduces the parasitic capacitance between the detector anode and the chip substrate. This was made possible due to the use of a fully custom technology, namely a fabrication process in which each step can be finely optimised in order to achieve the best overall device performance. The new structure reported by the authors increases the fraction of avalanche current that can be fed into the front-end electronics. This means that the same timing jitter can be attained with a higher detection threshold and is especially important when more than one SPAD is integrated on the same chip and the disturbances generated by a detector during operation can degrade the timing jitter on adjacent pixels. “By simplifying the problem of reading the avalanche current with a low timing jitter, the device structure reported in this Letter opens the way to fabricating photon detection systems that are based on custom-technology SPADs with a large number of pixels and picosecond temporal resolution” explains Ceccarelli. The fabrication of large arrays of SPADs, which are truly parallel and attain high detection efficiency and picosecond timing resolution, is certainly a long-term objective in this field. This work represents one of the milestones towards achieving this goal. Ceccarelli says “we aim at arrays with many thousands of pixels which can have a dramatic impact on very critical fields like new drugs discovery.” Ceccarelli and his colleagues at the Politecnico di Milano will now work on combining this new technology with a recently developed integrated front-end circuit to build a multichannel time-resolved single-photon detection module. They will also continue to develop the detector to improve its performance. Ceccarelli explains “we are looking in particular at very high detection efficiency in the visible / near infrared region and at removing the slow components in the temporal response.” The interest in single-photon detection has grown considerably in the last few years and is expected to increase even further in the coming decade, due to emerging consumer and scientific applications, like autonomous cars and new drug discoveries. Most of the attention will likely be devoted to the development of a high-throughput, time-resolved, single-photon imager with a large number of pixels. On the one hand, this requires the design of detectors with always increasing performance, especially in terms of detection efficiency, temporal response and scalability. On the other hand, a lot of work must be done to devise new architectures for handling the huge amount of data produced by large, time-resolved arrays. “Beside this mainstream development, we expect also a lot of activity in emerging fields, especially in quantum information processing” concludes Ceccarelli.