Leakage current evolution in LHCb VELO sensors during Run 1-2 LHC data taking period

In the heart of the Large Hadron Collider, where protons collide at speeds approaching that of light, a specialized experiment called LHCb watches closely to understand the behavior of heavy-flavor particles. To see these fleeting particles, the experiment relies on a delicate layer of silicon sensors positioned just millimeters away from the collision point. These sensors act like high-speed cameras, capturing the paths of particles as they fly out from the crash. However, living this close to the beam is a harsh existence. The sensors are constantly bombarded by a storm of subatomic particles, a bombardment that slowly damages the internal structure of the silicon crystal. This damage does not destroy the sensors immediately, but it changes how they work, causing them to leak a small, steady electrical current even when no particles are passing through. This leakage current grows over time, and if it becomes too large, it can overwhelm the sensor's ability to see the real signals, effectively blinding the camera. Understanding exactly how this current grows, and why, is essential for keeping these sensors alive and for predicting how future detectors will survive in such extreme environments.

A team of researchers recently spent years analyzing the history of these sensors during the first two major runs of the collider, a period spanning from 2011 to 2018. They gathered a massive amount of data, recording the temperature and the electrical current of every single sensor module over eight years. Their goal was to test a widely used prediction model, known as the Hamburg model, which attempts to calculate how much damage radiation causes and how that damage changes the electrical current. By comparing the model's predictions against the actual, continuous records from the sensors, the team could see if their understanding of radiation damage was accurate or if something was missing from the picture.

The researchers found that the model generally worked very well. It successfully predicted the overall shape of how the leakage current grew over time, including the periods when the collider was shut down for maintenance. During these breaks, the sensors cooled down, and the damage within the silicon partially repaired itself, causing the current to drop. The model captured this healing behavior, known as annealing, with impressive precision. However, when the team looked closer, they noticed a consistent gap between what the model predicted and what the sensors actually reported. The sensors were leaking more current than the model expected, even after accounting for the known variables.

Upon closer inspection, the team realized the culprit was likely the temperature readings themselves. The sensors are cooled by a system that keeps them at roughly minus eight degrees Celsius, but the tiny thermometers used to measure this temperature are not placed directly on the silicon crystal. Instead, they sit on the surrounding electronics. Because the silicon heats up slightly more than the nearby components during operation, the thermometers were reporting a temperature that was a few degrees lower than the actual temperature of the crystal. Since electrical leakage is extremely sensitive to temperature, this small difference in measurement led to a significant error in the calculations. When the researchers adjusted their data to reflect a temperature that was 3.2 degrees higher than what the thermometers showed, the gap between the prediction and the reality almost vanished. The model and the measurements aligned perfectly, confirming that the physics of the damage was understood, but the environmental data needed a small correction.

There was one remaining puzzle. The researchers expected that sensors closer to the center of the beam would suffer much more damage than those further away, simply because they are hit by more particles. While this trend was visible, the difference between the sensors was not as dramatic as the computer simulations had predicted. The model suggested that the damage should drop off sharply with distance, but the real sensors showed a more gradual change. The team traced this discrepancy not to the damage model itself, but to the simulation used to estimate how many particles hit each sensor. The computer program that calculated the particle storm, which is based on complex physics of how particles interact with matter, seemed to underestimate the number of particles reaching the outer sensors. This suggests that while the theory of how radiation damages silicon is solid, the tools used to predict exactly where and how much damage occurs in such a complex environment still need refinement.

This study highlights the importance of long-term, real-world monitoring in high-energy physics. By collecting data continuously over eight years, the researchers were able to spot subtle patterns that short-term tests would have missed. They proved that the standard model for radiation damage is robust, provided that the temperature is measured accurately. They also identified a specific weakness in the simulation tools used to map the radiation environment, a finding that will help improve the design of future detectors. As scientists plan the next generation of experiments, which will face even more intense radiation, these lessons ensure that their silicon sensors will be built with a clearer understanding of the harsh conditions they must endure.