A customizable imaging spectrometer – otherwise known as a “hyperspatial imager”

By Michael P C Watts

A camera that can detect any color even IR and UV would be very useful for any number of inspection or detection tasks….. a “hyperspatial imager”. The classic RGB color camera is limited by the RGB color filters which detect a limited range of visible wavelengths, known as the “color gamut” of the camera. At Photonics West 2012 in San Francisco, a team from Imec in Belgium reported on a customizable imaging chip, where they can control the spectral response on a pixel by pixel basis. This is both an interesting fabrication trick and opens up a number of interesting applications.

The fabrication uses the recent progress in very well controlled timed or blind etches used in making phase masks. The filter in the Imec chip is a Fabry – Perot filter, basically 2 partial mirrors held apart by a dielectric where the distance between the mirror defines the resonance frequency of the filter. The Imec team deposit the first mirror and a layer of dielectric on a completed wafer of CMOS imagers. They then open a photoresist mask over the pixels that they want to define with the same filter, and then do a timed etch to get the target gap between the mirrors. They repeat the mask and etch for as many different filters as they want, and finally they deposit the second mirror.

There are several different possible imaging architectures, for instance;

1) a varying color gamut, 3 color camera detecting .
2) a N color camera, where N different color filters are grouped together.
3) a spectral line camera, where small changes in filter occur in one axis and the other axis provides imaging from a line image. You can then scan the object under the line camera to get a complete spectral image.

The idea of customizing standard CMOS devices using further processing in a full fab environment is the basis for Imec’s CMORE strategy. In this case the alignment of the filter to the pixels is greatly enhanced by post processing on the completed wafer.

There was a second paper looking at a specific bio application where a line scan camera that measured 425 different wavelength was used to distinguish growths in a petri dish as part of a food safety analysis.

The picture below compares the processed images of growths in a petri dish. The images from the hyperspatial camera were processed through 2 different piplines., in both cases the growths were clearly identified as “E-coli”.

My take is that further customization in MEMS, optics, or rerouting layers for different pad and package configurations is a growing opportunity in applications specific devices. Imec seem to be using this a way to get them in the small volume foundry business.

About the author

Mike Watts has been patterning since 1 um was the critical barrier, in other words for a longtime. I am a tall limey who is failing to develop a Texas accent here in Austin. I have a consulting shingle at www.impattern.com.
My blog “ImPattering” will focus on the latest developments in the business and technology of patterning. I am particularly interested in trying to identify how the latest commercial applications evolve.