Speaker
Description
A total of 24,010 AC-coupled silicon strip sensors, consisting of n-type strips in p-type silicon and referred to as ATLAS18, are currently in production for installation in the upgraded ATLAS Inner Tracker (ITk). In n-in-p strip sensors, a dense p-type region (e.g., p-stop implant) is essential for electrically isolating the n-type strips from a conductive inversion layer between the strips, caused by positive interface charges. The punch-through protection (PTP) structure is implemented at one end of each strip to protect the AC-coupling capacitor in the event of large currents flowing into the strips. During quality assurance (QA) testing, we occasionally observed a deviation in the threshold voltage ($V_{PT}$) of the PTP structure in the test chips located at a corner of the wafer perimeter, with a shift of four standard deviations or more below the mean. To evaluate the variation of $V_{PT}$ across a wafer, we selected two main sensors, each processed in a different furnace, where $V_{PT}$ measurements in the QA test chips were still near the mean value. One sensor exhibited a flat distribution of $V_{PT}$ measurements, while the other showed a non-uniform, parabolic distribution, with a shift of four standard deviations toward the wafer edge. TCAD simulations were used to analyze $V_{PT}$ as a function of p-stop density, confirming that the $V_{PT}$ correlates well with the p-stop density. By setting the nominal p-stop density ($4\times10^{12}$ cm$^{−2}$) to correspond to the mean $V_{PT}$ value, the p-stop density at the wafer edge, which is four standard deviations away from the mean, is estimated to be approximately half ($2\times10^{12}$ cm$^{−2}$). A p-stop density of $2\times10^{12}$ cm$^{−2}$ should still be sufficient to ensure isolation, even after irradiation, and in the presence of interface charges as high as $1\times10^{12}$ cm$^{−2}$.
