Nikkei: Panasonic developed an image sensor 100 times more sensitive to light than an existing technology. The new sensing layer is made from an organic material film rather than silicon. Panasonic developed the basic technology with Fujifilm. A proprietary readout circuit reduces noise, while a modified electrode structure helps detect even faint light efficiently.
The company says the new sensor does not suffer from blooming, and is aimed to automotive applications.
 |
| Blooming demo, Panasonic sensor is on the right. |
Possibly, the article material is based Panasonic presentation at ISSCC 2016. The company aims to have a practical version ready as early as 2020 or so.
How does one get 100x sensitivity when the integrated QE of current CMOS sensors, across the visual band, is probably close to 50%?
ReplyDeleteAre they able to detect 50x more photons than those that actually reach the sensor?
They might be referring to the photoconductive gain in their pixel. It increases the signal by the gain factor, but also amplifies the photon noise by the same or (most probably) larger factor.
DeleteIn the conference, they were talking about extending the range, not the sensitivity. They did extend the range by having 100x less sensitive pixels next to regular pixels ( regular with OPF is in the eye of the beholder....)
DeleteIn this case, there should be a large image lag ...
DeleteOr similar to the statement of Invisage : the absorption coefficient of the organic layer is 100 x higher than the one of silicon. So finally you do need a much thinner layer, but of course, you do not get 100 times the QE of silicon. Actually a very misleading statement of Panasonic.
Delete@ In this case, there should be a large image lag ...
DeleteNot necessarily a visible lag. The photoconductive gain is roughly equal to the carrier life time over the carrier flight time (not going too much in detail here). If the flight time is, say, 1us, considering low carrier mobility in organic material, and the life time is 100us, the gain is 100x, while the image lag is 100us, imperceptible to the naked eye.
Could someone share a pointer to a good reference on the photoconductive gain phenomenon and its limitations? I recall that some time ago SiOnyx used to make those claims about incredible sensitivity, quote similar to what we are seeing here. IIRC this may have been caused by photoconductive gain but apparently did not find it useful. There was some paper along the way, but not very illuminating. It would be good to understand this better.
DeleteThey better called the demo "wide dynamic range demo" or so. Blooming is usually not really an issue anymore...
ReplyDeleteYes, you're right.
DeleteThe Panasonic PR makes more sense than the Nikkei.
ReplyDeleteThey use a old Fujifilm like EXR-like feature (with two pixels of different sizes and sensitivities to image very bright (highlights) and normal light (midtone and shadows) rather than a high QE.
See this link for more details
http://news.panasonic.com/press/news/data/2016/02/en160203-5/en160203-5.html
Panasonic also announced a global shutter version of their organic + CMOS sensor (in addition to the HDR version).
ReplyDeleteSee this PR
http://news.panasonic.com/press/news/data/2016/02/en160203-6/en160203-6.html