Fujifilm patents hybrid organic/CMOS sensor
Fujifilm has been granted a patent for an innovative organic-hybrid sensor technology. However, while interesting, it may not offer a compelling advantage over existing designs, according to sensor technologist Professor Eric Fossum. The company has recently been granted a patent for its work on a sensor that uses an organic (carbon-chemistry-based) material on top of silicon circuitry. Speculation about Fujifilm's forthcoming mirrorless camera has latched onto a technical paper the company published in late 2009, but both Fossum and the company say the work shows more promise for small-scale sensors.
How the technology differs
The Fujifilm sensor differs from exisiting CMOS and CCD designs in that it uses a photoelectric organic coating to convert light into electrons, rather than silicon photodiodes. The electrons produced by the organic layer are then 'read' by CMOS circuitry built underneath. Fujifilm's internal paper on the technology claims several advantages to the design. The first is that the organic layer can be coated across the entire sensor, making much more of the surface light-sensitive. This means there's no need for microlenses to redirect light onto specific light-sensitive regions. There's also no need to divide the photosensitive layer into individual photosites. Both these features could make the sensor less expensive to manufacture, Fujifilm's team says. The organic material's sensitivity to only visible light also avoids the need for a IR-filter in front of the sensor (silicon sensors are sensitive to IR light).
The main stated aim, though, is to be able to make sensors that perform well, despite the demand for ever-higher pixel counts and the consequent smaller pixels. As the authors of the Fujifilm paper explain: 'Despite the tireless efforts to improve technologies to achieve smaller pixel size, the light capture efficiency and sensitivity of CCD and CMOS image sensors have been decreasing.'
|Rather than using the light-sensitivity of silicon p-n junctions, the Fujifilm sensor uses a coating of a light-sensitive carbon-based chemical (shown as the darker purple) to convert light into electrons. (image from Fujifilm)|
Sensor technologist (and inventor of the active-pixel CMOS sensor that underpins most modern cameras), Fossum is impressed by the work, but not fully convinced by the company's list of benefits. 'As a fellow image sensor technologist, I would like to congratulate Fujifilm on the fine R&D progress they have made and I find their work interesting. If the report is a good assessment of the state-of-the-art at Fujifilm in 2009, then I would say it is possible the technology could be ready for prime time in just a few years from now, or sooner if they have had good luck.
Aside from issues like reducing noise (from the reported 38 electrons rms to below 5 e- rms), they also need to deal with the issues of large scale manufacturability of the material layer. Of course they are well-poised to do this since most RGB filters on image sensors probably employ Fuijifilm product. The RGB filter layer material is passive however, and not an electronically-active material, so there additional manufacturability concerns.'
A great leap forward?
'As far as a prognosis for this R&D to find its way to commercially-competitive products, one has to look at what the compelling advantage of the product is, once the issues above are resolved, compared to where the rest of the industry will be at the time of product roll-out,' he says. 'Their technology might allow smaller pixels with functionality like global shutter but this is already being demonstrated in monolithic silicon devices in small pixel formats. I do not believe manufacturing cost will be a compelling advantage for this technology but it could be important. I don’t believe that there will be a compelling advantage in the usual metrics for image sensors, such as sensitivity, read noise, full well, dynamic range etc. but it is difficult to predict these things'
|The sensor is only sensitive to light in the visible range (around 380 -740nm), removing the need for Infrared filters. (image from Fujifilm)|
'The one compelling advantage for Fujifilm itself is that is it is a homegrown technology and fully captive. On the other hand, that means unlike the quasi-cooperative nature of the CMOS image sensor industry at large (not intentional cooperation of course) where advancements by one company spur on other companies, Fuijifilm would be an island unto itself and this will lead to slow technology advancement compared to the rest of the industry.'
'I really like the idea of highly absorptive photodetector layers. I think such a layer, be it the Invisage Quantum Film, the Fujifilm organic photodetector layer, or some other to-be-invented material, will be helpful in realizing my own vision for a Quanta Image Sensor where every photoelectron is counted. But, I am not so optimistic about the compelling commercial advantages of the Fujifilm approach. Nevertheless, it is always exciting to me to see new technology being invented and developed.'