Summary of quality control (QC) of ATLAS18 production ITk strip sensors

Imagine the universe as a giant, chaotic dance floor where subatomic particles are the dancers, zooming around at nearly the speed of light. To understand the secrets of the cosmos, scientists built a massive, circular racetrack called the Large Hadron Collider (LHC) to smash these particles together. But the LHC is getting a super-charged upgrade, the "High-Luminosity" version, which means the dance floor is going to get incredibly crowded and intense. The particles will collide so frequently that the detectors watching them will be bombarded with a level of radiation that would fry a normal camera sensor in seconds. To survive this, scientists are building a new, ultra-rugged "eye" for the experiment called the ATLAS Inner Tracker. This new eye is made entirely of silicon, the same material in your computer chips, but engineered to be tough enough to withstand the cosmic radiation storm. The challenge isn't just making these silicon sensors; it's making thousands of them perfectly, because even a tiny scratch or a microscopic flaw in the electrical wiring could ruin the picture of the universe.

This paper is the story of how a global team of scientists acted as the ultimate quality control inspectors for this massive silicon sensor project. They didn't just build the sensors; they put every single one through a grueling "boot camp" to ensure it was ready for the job. The team tested over 23,000 sensors, checking them for physical damage, measuring how much electricity leaked through them, and ensuring they could handle high voltages without breaking. They discovered that while the sensors were mostly excellent, a few had issues like static electricity buildup from being rubbed against their packaging during shipping, or uneven chemical doping that made them weak in certain spots. The paper details how they fixed the static problems with special light treatments and heat, and how they had to reject a small number of batches that were just too flawed to save. By the end of their long inspection process, they confirmed that over 91% of the sensors were perfect and ready to be installed, ensuring the ATLAS experiment will be able to see the universe clearly even in the most extreme conditions.

The Silicon Sensor Boot Camp

Think of the ATLAS experiment as a giant, high-speed camera trying to take a picture of a fireworks display, but the fireworks are exploding with the force of a nuclear bomb. The "lens" of this camera is the Inner Tracker, a new system made of 165 square meters of silicon sensors. That's a lot of silicon—enough to cover a small house floor-to-ceiling. These sensors are being built by a company in Japan and shipped to seven different labs around the world for testing. The goal? To make sure every single one of the roughly 18,000 sensors destined for the outer layers of the tracker can survive a radiation dose that would turn a normal electronic device into a brick.

The scientists in this paper describe the "Quality Control" (QC) process, which is essentially a multi-stage obstacle course for the sensors. Before a sensor is allowed to join the team, it has to pass a series of tests that check its body, its brain, and its stamina.

The Physical Exam
First, the sensors get a visual inspection. Scientists look at them with the naked eye and powerful microscopes, hunting for scratches, chips, or any sign of physical trauma. It's like checking a new car for dents before you buy it. They also take high-resolution photos to create a "before" picture, just in case something goes wrong later. Then, they measure the sensor's "bow" or curvature. Imagine a ruler that is supposed to be perfectly flat; if it bends too much, you can't build a precise module on top of it. The sensors must be flatter than 200 micrometers (which is about the width of two human hairs). If they are too curved, they get sent back. They also check the thickness, ensuring every sensor is exactly 320 micrometers thick, give or take a tiny bit.

The Electrical Stress Test
Once the physical check is done, the sensors go into the electrical gym. Here, they are subjected to high voltages to see how they hold up.

• The Leakage Test: Scientists measure how much electricity "leaks" through the sensor when it's turned on. A good sensor should be a tight seal, letting almost no current escape. The limit is set at less than 100 nanoamperes per square centimeter. If a sensor leaks too much, it's like a bucket with a hole in it—it won't work.
• The Depletion Test: This checks how much voltage is needed to turn the sensor "on" so it can detect particles. The sensors need to be fully active at less than 350 volts. If they need more, they might not be able to keep up with the fast-moving particles in the collider.
• The Breakdown Test: This is the ultimate stress test. Scientists crank up the voltage to see when the sensor finally gives up and breaks (breakdown). The sensors must survive at least 500 volts without exploding electrically.

The Strip Check
The sensors aren't just solid blocks of silicon; they are covered in thousands of tiny metal strips, like the strings on a guitar, that read the signals. The team performs "Full Strip Tests" on a sample of sensors to make sure every single string is working. They check for shorts (where two strings touch when they shouldn't), broken connections, or faulty resistors. If too many strips in a row are broken, the whole sensor is rejected.

The Static Electricity Surprise

One of the most interesting discoveries in the paper was a case of "static shock." When the sensors arrived at the labs, some of them were covered in a high level of static electricity. The scientists suspected this happened because the sensors were rubbing against their packaging materials during transport, like a balloon rubbing against your hair. This static charge was causing the sensors to fail their electrical tests, making them look broken when they were actually fine.

The team realized this was a reversible problem. They treated the sensors with various "cures":

1. UV Light: Shining specific types of ultraviolet light on them for hours.
2. Ionized Air: Blowing ionized air over them to neutralize the charge.
3. Baking: Putting them in an oven at 160°C for more than 16 hours.

These treatments worked like magic, clearing the static charge and allowing the sensors to pass their tests. The team confirmed that this fix wasn't just temporary; the sensors stayed healthy for at least a year after the treatment. This was a huge win, saving many sensors that would have otherwise been thrown away.

The Rejection Rate and the "Bad Batches"

Despite the rigorous testing, not every sensor made the cut. The paper reports that out of the thousands of sensors tested, only about 2.8% were rejected. That means over 91% of the production was accepted and is ready for use. This is a very high success rate for such a complex manufacturing process.

However, the team did have to say "no" to a few specific batches.

• Two batches were rejected because they had electrical instability that couldn't be fixed. These sensors had a "non-recoverable breakdown," meaning they were fundamentally flawed and would fail under pressure.
• Four batches were rejected due to a chemical issue called "p-stop doping." Imagine the silicon sensor as a cake, and the "p-stop" as a special ingredient mixed in to keep the electricity flowing in the right direction. In these four batches, the ingredient wasn't mixed evenly. On one side of the sensor, the concentration was too low, causing the electrical protection to fail. The scientists found this by measuring the "punch-through protection voltage," which acts like a safety valve. If the valve opens too early (below 12 volts), the sensor is unsafe. Because this was a systematic issue affecting whole batches, those batches were discarded to ensure the final detector is reliable.

The Final Verdict

By September 2025, the team had delivered over 98% of the required sensors. The quality control process, which involved testing over 22,000 sensors across seven international labs, proved to be robust and effective. The team successfully identified and fixed issues like static charge and caught the few batches with chemical inconsistencies before they could cause problems in the final detector.

The paper concludes that the production of these silicon sensors is a massive success. With nearly 21,800 sensors accepted and ready for installation, the ATLAS experiment is well-prepared for the High-Luminosity upgrade. The sensors are tough, the testing is thorough, and the team has ensured that the new "eye" of the ATLAS detector will be sharp enough to see the universe's deepest secrets, even in the most intense radiation environment ever created by humans. The journey from the factory floor to the final detector is complete, and the silicon sensors are ready for their big moment.