Dr. Murzy Jhabvala

Visible light is only one narrow band of the electromagnetic spectrum, and doesn't always tell scientists what they need to know. Infrared, which is outside the range of human eyesight, has for years been used to delve out mysteries of distant stars or to allow users to see in the dark. NASA scientists have now improved the Quantum Well Infrared Photodetector (QWIP) array infrared technology to gain more detail than ever before. NASA engineer Dr. Murzy Jhabvala led the project.

NASA Tech Briefs:

What is the Quantum Well Infrared Photodetector Array, and how does it work?

Murzy Jhabvala: It's a gallium arsenide based structure. And gallium arsenide, in terms of technology and using it in fabricating devices, is very similar to silicon technology, which is used to make all the computer chips. So because we are using a technology that is comparable to that makes it very versatile, very easy to work with, because we don't have to invent new equipment to deal with it.

But it is a gallium-arsenide based technology. With the Quantum Well Infrared Photodetector, or QWIP, the quantum well part of that, what we do is we start with a gallium-arsenide wafer, which looks very similar to a blank silicon wafer, and on it we grow alternating layers of different materials. In our case, it's alternating layers of aluminum gallium-arsenide, gallium-arsenide, and indium gallium-arsenide. Each layer is on the order of forty or fifty atoms thick. And we alternate them and we grow them about 106 periods of this. So in our case, we had a bottom layer of aluminum gallium-arsenide, then another layer of gallium-arsenide, then a layer of indium gallium-arsenide, then a top layer of gallium arsenide. Those four layers were repeated over 100 times, and each of them are about 40 to 50 atoms thick.

So we create this structure on the surface of the gallium-arsenide, and this is the quantum well part of the device. Once the wafer is constructed, then we pattern it into discrete detector elements. And in our case, we patterned it into arrays that were 1,000 by 1,000 square array, with each element being 25 microns in size. And when you have 1,000 by 1,000, the total number is going to be a million. So we make this one million pixel array.

And what is unique about this array is that QWIPS, by virtue of their structure, tend to want to be responsive to a specific wavelength, a specific narrow band wavelength, because of their quantum nature. And what we did here that is somewhat unique, is we were able to make the QWIP responsive to a broad band of wavelengths; in our case, 8 to 12 microns. And that's important in terms of spectroscopy, were you want to work with an object and see its characteristics at a variety of wavelengths, not just one wavelength. That's where the information is, in the spectral content of the image that you are looking at.

And QWIPS, like many infrared detectors, have to be cooled, and in our case, these were cooled to about 60 kelvin, which is quite cold. That would about -231 C. But infrared detectors, particularly sensitive infrared detectors, have to be cooled. There are a number of reasons; some just won't work until they get to a certain temperature. And they tend to make their own signals when they are warm, or warmer, and these signals will overwhelm any incoming signal, and you can't discriminate the signal your looking at and the signal the detector is making. It's "noisy." So you have to keep cooling it until you make the detector signal go away. And so we ran ours around 60 kelvin, and what is new for us is that we were able to capture imagery in video format. We have videos! And further, many detectors which are not silicon-based, of which there are quite a few in the infrared spectrum-there's mercury-cadmium-telluride and indium antimonide and a whole range-all those detectors convert incoming signal in the infrared into electrons. And that is what the detector does. But you have to take those electrons and get them out of the detector. And in order to do that, the detector has to be mated, physically mated, to a read-out chip. And the read-out chip has the same exact footprint as the QWIP detector array. So if you have a million pixels, each 25 microns square, you need a read-out chip with a million cells, with about three or four transistors in them, each 25 microns square as well, and then you physically mate the detector to this read-out chip and the electrons are extracted once the incoming photons hit the detector array, they are converted to electrons, those electrons are extracted.