A lower array-average dark current does not by itself prove that rare high-dark-current pixels improved. The average summarizes the population, while the tail describes a small subset. Both can matter to image quality, and they can move differently.
Separate population and outlier questions
Local defects and process variation can cause substantial pixel-to-pixel differences. A few pixels can create visible white-spot blemishes even when most pixels have a small dark signal. The pinned-photodiode review discusses both low average dark current and problematic local behavior.
Consider a conceptual array in which most pixels improve but a rare defective subset stays unchanged. Its average can fall while its outlier count does not. Conversely, reducing a few extreme pixels can improve the tail without an equally dramatic shift in the central population.
Process checkpoint
Understand P-Pinning Ion Implantation in context
Explain surface pinning as one boundary-control mechanism and identify why transfer-gate defects and local variation can still produce outliers.
Process context for “Average Dark Current and the Hot-Pixel Tail Need Separate Evidence”: 40nm BSI CMOS Image Sensor · PD · Step 99
A mechanism can help without covering every site
Surface pinning changes the carrier population near the photodiode interface. It addresses one important boundary. The neighboring transfer gate and defects beneath or around it can still generate carriers collected by the storage well. Local implantation-related defects are another possible contribution, rather than a universal explanation for every outlier.
This is why the success of one intended passivation mechanism cannot certify every pixel. The location of generation and the path to collection both matter.
Keep conditions comparable
Comparing averages or tails requires a consistent observation basis, including accumulation time, temperature, bias and the definition of an outlier. A changed criterion can alter the reported count without changing the device population.
A single dark image cannot establish the cause of each extreme pixel. Repeated observations and mechanism-sensitive evidence are needed to separate persistent high generation, offsets and time-varying behavior. This is a reasoning boundary, not a screening recipe.
Read the pinning station as one contribution
Identify the surface boundary that pinning manages. Then identify what remains outside that statement: neighboring transfer-gate behavior, local defects and the array distribution. Explain why a smaller average is not enough to sign a claim that all white spots disappeared.
The existing dark-current–read-noise article separates temporal sources. This article adds the population-versus-tail acceptance question; it does not relabel every hot pixel as a pinning failure.
References
A Review of the Pinned Photodiode for CCD and CMOS Image Sensors
E. Fossum, Donald B. Hondongwa · IEEE Journal of the Electron Devices Society
A Review of Ion Implantation Technology for Image Sensors †
N. Teranishi, G. Fuse, M. Sugitani · Italian National Conference on Sensors