A New Approach to Achieving High Granularity in Low-Gain Avalanche Detectors

The Low Gain Avalanche Detector LGAD), a new type of solid-state detector, has achieved a timing resolution of better than 20 psec, which has enabled the development of fast timing layers for the ATLAS and CMS detectors. The high speed of LGAD signals have also drawn the attention of the nuclear and photon science communities. In their current realization, however, the high fields needed to induce the LGAD gain are produced by a highly doped layer at or near the segmented implant layer. This leads to breakdown and requires the addition of an isolation structure the Junction Terminating Extension, or JTE). However, the JTE generates a dead region in the device, limiting the granularity of conventional LGADs to the 1x1 mm2 scale. On the other hand, the requirements of X-ray imaging, and of true four-dimensional tracking of beam collision products rather than the currently-planned time-stamp provided by a timing layer at the outside of the tracker) will require granularity at the 50x50 mm2 scale – over an order of magnitude reduction in the linear dimension of the LGAD pixel size relative to the sensors currently under development for the HL LHC. While there are several approaches to increasing LGAD granularity that are under exploration AC LGADs, Trenched LGADs, ILGADs), none of these have yet been shown to provide an adequate solution to the X-ray imaging and 4D tracking granularity challenge. Here, we propose to fabricate the first-ever prototype of a proprietary new approach to the production of high- granularity LGADs that permits conventional approaches to silicon diode pixilation to be employed, effectively removing the granularity limit suffered by the LGAD sensors under development for the ATLAS and CMS detectors at the LHC. Simulations performed with the Sentaurus Device simulation package suggest that the resulting sensor will satisfy essential goals of Topic 34b, achieving deep sub-millimeter granularity with GHz counting rate capability. In addition, Figure 1 shows the simulated gain variation for the current device baseline; a gain uniformity across the profile of the device of better than ±5% is expected. As for all solid-state sensors, the dynamic range will be limited only by space-charge saturation of the bias field. Furthermore, a capability unique to LGAD sensors is a dynamic range tunable by up to two orders of magnitude via the externally-applied bias voltage.