Dialing back time on Timepix3: A study on the timing performance of Timepix3

Abstract

For the advancement of the understanding of timing systematics in pixelized readout chips and for the benefit of future fast timing detectors to aid 4D tracking technology in the HL-LHC, we have performed detailed studies of the timing properties of the Timepix3, a hybrid pixel detector developed by the Medipix collaboration. These studies use three different measurement techniques to investigate the timing systematics of this detector: test pulses, testbeam, and a laser setup that we build for this work. The average delay over the pixel matrix is determined with the testbeam and the laser and shows the same structure for different Timepix3 chips. The difference in delay results in a maximum difference over the pixel matrix of around 4 ns, which is large compared to the time bins of 1.56 ns. By correcting for this difference on a per-pixel level, the time resolution is on average improved with 159 ps depending on the sensor. The best time resolution that is achieved after a timewalk correction and a correction for the difference in the average delay per pixel is 686.4$\pm$0.2 ps for a single Timepix3 chip with a 200 $\mu\text{m}$ planar silicon sensor, compared to a time resolution of 892.9$\pm$0.3 ps without any correction. However, this improved time resolution is not yet the naively expected time resolution of 451 ps. The origin of this delay is determined with this laser setup, and is due to a combination of signal propagation and a difference of the start-up time of the fast oscillator within the pixels. This work indicates that it is vital for next generation pixel detectors to correct for these systematic timing effects in order to reach a better time resolution to aid 4D tracking technology in the HL-LHC.