Tuesday, February 01, 2011

UV-Transparent Coating for Image Sensors

Nanowerk: CMOS devices are normally covered with a silicon nitride coating to protect from moisture and other impurities. The silicon nitride coating limits the range of optical applications because it is impermeable to UV light.

Scientists at the Fraunhofer Institute for Microelectronic Circuits and Systems IMS in Duisburg have found a solution to this problem: "We've developed a new process step," says Werner Brockherde, head of department at Fraunhofer IMS, "that allows us to produce a protective coating with the same properties but which is permeable to blue and UV light."

The trick is to increase the proportion of nitrogen in the coating by fine-tuning the deposition parameters such as pressure and temperature. "This reduces the absorption of shortwave light," explains Brockherde. "This makes CMOS image sensors suitable for use in wavelength ranges down to 200 nanometers," states Brockherde. "With standard passivation the limit was about 450 nanometers."


  1. Any data on stability of the coating under UV light? Buildup of positive charges on the layer with a certain dose of light below 250 nm?

  2. -what applications would like to see UV?
    -was under the impression that silicon could not absorb wavelengths under 400um anyway? not correct?

  3. Last I checked, the back thinned scientific CCDs (whish as far as I know are still made of silicon) managed to absorb below 200nm.

    Silicon w/o glass does fine under 400nm.

  4. Spectroscopic devices have always worked down to 190 nm. The problem is not that the UV is not absorbed, it is that it is absorbed very close to the surface so in most devices the generated charge recombines without being captured for readout.

  5. Take into consideration that absorption length is a few mm in the uv...

  6. The absorption length is long in IR, in UV it's very short.

  7. Silicon imagers can deliver reflection-limited performance 100nm with Delta doping (e.g. QE>30%). Delta doping is an MBE process.

    Enhanced UV can also be had with boron implant and laser anneal or in-situ doped poly, but it's not as good as delta doping.

    These are all extra steps not typically incorporated with silicon imagers. But they are well established.

  8. I meant they perform well down to at least 100nm.


All comments are moderated to avoid spam and personal attacks.