Speaker
Description
Low-Gain Avalanche Detectors (LGADs) have seen prominent usage and development over the last 10 years, particularly within High Energy Physics. These devices are planned to be utilised for track timing as part of the High-Luminosity upgrade proposed for the Large Hadron Collider (HL-LHC). Within the Photon Science Community however, progress has been made in developing pixellated LGAD devices for the detection of and imaging with low energy X-Rays, particularly in the range $250 \, eV$ to $2 \, keV$.
The difficulty with pixellated LGADs comes from segmenting the gain layer to isolate the pixels, as early device breakdown becomes a problem due to the abrupt change in potential difference. It has become common to use a moderately doped silicon implant extending from the edge of the pixels, known as a Junction Termination Extension (JTE), to reduce the electric field at these edges by spreading out the space over which voltage drop occurs. The consequence of this however is a reduced fill factor (the ratio of detector area with gain to total detector area) due to area taken up by the JTE. Additionally, due to the spreading of this electric field, a limitation in pixel pitch is imposed as charge carriers (electrons in this case) are diverted from the gain region and towards the JTE. These carriers are then collected at the pixel edge where there is no gain amplification, resulting in an effective $0$ fill factor for the detector when sensing. This has currently been observed in LGAD arrays on a pixel pitch of around $55 \, \mu m$.
Trench-isolated LGADs are currently being studied as an alternative method of pixellation by cutting through the gain layer and filling the trench with an insulating oxide. This would isolate the pixels without need for a JTE to dissipate the voltage drop and henceforth remove the limitation in pixel pitch. To investigate this, on a single silicon wafer we had a number of conventional LGAD detectors and trench-isolated LGAD detectors produced, as well as some devices identical in layout but with no gain layer for direct comparison such that gain layer effects could also be isolated. These devices were produced by Micron Semiconductor Ltd on $500 \, \mu m$ float zone silicon, from our own custom mask design and a known doping recipe. For each detector type we had a variety of different pixel arrangements, pitches, and total detector areas.
We primarily tested a selection of $2 \, cm \, x \, 2 \, cm$ area 80x80 pixel hybrid detectors on a HEXITEC ASIC of $250 \, \mu m$ pitch, with a fill factor of $0.81$ for the conventional LGADs. The performance of these LGAD HEXITEC hybrid devices will be shown, including measurements of leakage current for increasing reverse bias and energy spectra for the irradiation of such devices with an Fe-55 source. From these results we have quantified the uniformity across the detector area, decoupled gain specific effects in the resolved energy spectra and assessed the feasibility of trench-isolated LGADs for pixellated soft-x-ray detector applications.
To support the LGAD HEXITEC results and to inform on fundamental properties of these LGAD designs, a selection of small test devices which were manufactured on the same silicon wafer were tested, consisting of 2x2 pixels on a $1.3 \, mm$ pitch, and a fill factor of $0.97$ for the Conventional LGADs. The I-V and C-V measurements conducted on these devices, indicating the gain layer depletion voltage and the overall doping profiles of this batch, will be presented. Additionally, by irradiating with alpha particles from an Am-241 source under vacuum, the response of the gain layer at increasing reverse bias was measured. These results are compared with the non-gain detectors of the same geometry and specifications to again isolate gain layer only effects and better understand the influence of gain on charge transport.
