One can see that the photodiode extends well underneath the transistor region. Apparently, Omnivision did a great job to achieve lag-free operation with such a complex shape of the fully depleted photodiode.
The earlier Omnivision BSI applications, such as WO/2009/099483, WO/2009/09949, WO/2009/099494, WO/2009/099491 do not mention the extended photodiode profile. By the way, the last of these has an interesting part describing optimizations of the backside implant and the sensor's substrate thickness.
A patent application does not necessarily mean that the idea is present in the products ....
ReplyDeleteA.T.
True. However, if the same idea appears in 3 applications on different aspects of BSI pixel, there should be something in it. For example, the black pixel application has only extending PD pictures. It describes p- and n-type versions of PD, but they both have the same extending deep portion.
ReplyDeleteI hope Chipworks can say something more definite about it, they have a complete reverse engineering report on Omni BSI sensor.
interesting typo I guess in 099494, where they talk about a suiciding layer instead of a siliciding layer. A.T.
ReplyDeleteThose are old US patents.
ReplyDeleteYes, the patent applications are from 2008. However, they are published only in August 2009, two months ago. In that sense they are quite new.
ReplyDeleteA.T. - funny typo!
Fig.4 : Very difficult structure to make it fully depleted. I assume PD is implanted from the back to make the PD profile a shallow sandwich. The question is how much is the bridge doped to extract all electrons? Or they can leave with N+ and a few e- kTC? The lag could be flushed out...
ReplyDeletehow can you aneal the deep-diode if it is implanted from back-side?
ReplyDeleteOh, that's easy. First you need to switch Cu interconnects for carbon nanotubes. Then switch MOSFETs for silicon-carbide JFETs and bipolars. Then you can do whatever annealing you want. Most important, don't forget to move CF and ulens till after annealing step.
ReplyDelete