Testbeam characterization of a 3D silicon sensor read out by Timepix4

In the heart of the world's most powerful particle accelerator, the Large Hadron Collider, protons smash together at nearly the speed of light. These collisions create a chaotic storm of new particles, but to understand the fundamental laws of nature, scientists must track the paths of these fleeting fragments with extreme precision. As the accelerator prepares for an even more intense phase of operation, the detectors used to catch these particles face a daunting challenge: they must distinguish between collisions happening at the exact same moment, separated by mere billionths of a second. To do this, scientists are developing new types of sensors that act like high-speed cameras, capable of recording not just where a particle passed, but exactly when it arrived. The key to this future lies in a technology called 3D silicon sensors, where tiny pillars are etched directly into the silicon chip to collect electrical signals. By shortening the distance these signals must travel, the sensors can react much faster than traditional flat chips, offering a glimpse into a four-dimensional view of the subatomic world.

A team of researchers recently put one of these advanced sensors to the test, connecting it to a sophisticated readout chip known as Timepix4 and firing a beam of high-energy particles at it. The goal was to see how well this specific combination could measure the time and position of passing particles under different conditions. The sensor itself is a thin slice of silicon, 300 micrometers thick, filled with a grid of microscopic pillars that act as electrodes. When a charged particle zips through the silicon, it knocks loose electrons, which are then swept up by these pillars. The Timepix4 chip records the arrival time of these signals with incredible granularity, down to a fraction of a nanosecond. However, the researchers knew that the raw data would be messy. The time it takes for a signal to be recorded depends on how much charge the particle generated, and the internal clocks of the chip can drift slightly from one section to another. To get a true picture of the sensor's performance, the team had to carefully correct for these variations, essentially calibrating the clock and the signal strength for every single pixel on the chip.

The results revealed a device that is remarkably fast and precise, but with a few quirks that depend on how the particle hits it. When particles struck the sensor straight on, the team measured a time resolution of 245 picoseconds, meaning they could distinguish events separated by just a tiny fraction of a billionth of a second. The sensor also proved to be very efficient, detecting 97 percent of the particles that hit it. However, the researchers discovered that the sensor's performance is not uniform across its surface. The areas directly above the tiny pillars, where the electrical field is weaker, produced signals that were slower and less precise. In contrast, the spaces between the pillars allowed for much faster signal collection, achieving a time resolution as sharp as 153 picoseconds for the most energetic hits. This difference highlighted that the sensor's internal structure, rather than the readout chip, was the main factor limiting the overall speed.

To overcome the limitations of the pillars and improve the overall performance, the team rotated the sensor slightly relative to the beam. They found that tilting the sensor by just 8 degrees was the sweet spot. At this angle, the particles traveled through more of the active silicon between the pillars, avoiding the slower regions. This simple adjustment boosted the detection efficiency to over 99 percent and improved the time resolution by 6 percent, bringing it down to 233 picoseconds. It also sharpened the spatial resolution, allowing the sensor to pinpoint the particle's path to within about 7 micrometers. This finding suggests that a small tilt can significantly enhance the sensor's ability to track particles in the crowded environment of a future collider.

The researchers also explored what happens when particles skim the sensor at a very shallow angle, traveling almost parallel to its surface. This technique allowed them to probe the sensor at different depths, revealing that the timing performance varies depending on how deep the particle travels. Near the top and bottom surfaces, where the electrical field is weaker, the signals were slower. However, in the middle section of the sensor, where the pillars are fully active, the timing was much more consistent. By combining the timing information from many pixels that a single particle passed through, the team showed that they could further improve the precision, reaching a resolution of 170 picoseconds for large groups of hits. This demonstrated that even with the sensor's inherent imperfections, clever data processing could extract highly accurate timing information.

Ultimately, the study confirmed that 3D silicon sensors read out by the Timepix4 chip are a powerful tool for future particle physics experiments. While the sensor's internal structure creates some unevenness in performance, the researchers showed that this can be managed through careful calibration and by adjusting the angle of the detector. The work provides a clear roadmap for how these sensors can be optimized to meet the rigorous demands of the High Luminosity Large Hadron Collider, where distinguishing between billions of collisions will require detectors that are not only precise in space but also incredibly fast in time. The findings offer a solid foundation for the next generation of tracking devices, proving that with the right design and adjustments, these tiny silicon chips can keep pace with the fastest events in the universe.