It will depend entirely on the glass used in the optics I imagine.

My Sony digital 35mm-alike seems to take photos in or augmented with (presumably) near-IR in its "nightshot" mode.

Rgds

Damon

Apparently, according to the font of all knowledge Wikipedia, Sony had a sensor that can cope with IR. Why I thought of pulling my broken Alpha apart.

I read somewhere that IR film is between 700 and 900 nm, not quite sure how that related to degrees C.

I've just come off the 'phone to an ex-colleague (who rang to tell me he's just retired, too, and isn't it great?!). anyway, it turns out he worked on MEMS (Micro Electronic Mechanical Systems) devices at Malvern years ago and one of the sensors they developed was for imaging long wavelength IR. They made flat plane arrays of microscopic heat-sensitive resistors (thermistors), by etching them from doped silicon wafers. These were then used just like the sensor in a digital camera. The trick to make them sensitive enough to work at room temperature was to etch away the silicon underneath these microscopic sensors, so they looked like tiny flat topped umbrellas. This then thermally decouples them from their substrate and makes them respond almost entirely from the IR that impinges on them.

He reckoned the technology was now mature (it was around 20 years ago that he worked on it last) and is probably the same as that used in commercial IR cameras now. He did say there was no good reason for the sensors being expensive, as they are, apparently, as easy to fabricate as any other medium scale semiconductor device.

I'll do some digging around and see if the sensors are available on their own, as if they are then the chances are that a cheap DIY IR imager could probably be put together.

<blockquote><cite>Posted By: bot de paille</cite>What kind of technology is required to do that kind of etching of wafers??</blockquote>

The kit is pretty expensive and needs a hefty investment in things like clean rooms, but the process seems essentially quite simple. The process is a bit like that needed to make semiconductor chips, where bits of silicon are etched away, other materials are impregnated or sputtered on to the silicon wafer and more selective etching takes place, so that a 3D construct can be built on the surface of the silicon. This paper gives and overview of how these things are made: http://www.ee.kth.se/php/modules/publications/reports/2007/IR-EE-MST_2007_132.pdf

So far I've not found any really cheap sensors, but there are a few DIY thermal imager projects around on the web. In principle it seems pretty easy to make a scanned imager, using a relatively cheap sensor (and I happen to have some Melexis MLX90247 sensors left from another project that might do the job). Looking around the web making an imager to scan a sensor mechanically using a small microcontroller and a pair of model aircraft servos seems very straightforward. The same microcontroller could read the sensor output and store a grid of values for the whole image. I'm reasonably sure I could get this bit to work OK, maybe storing images as a file on an SD card.

Where I'd come unstuck is in writing the code for a PC to display the stored images.

I've no doubt that a project like this might well prove useful when investigating house thermal losses, so might be worth playing with.

Isnt it the crystal lense which makes these cameras expensive?

<blockquote><cite>Posted By: bot de paille</cite>Isnt it the crystal lense which makes these cameras expensive?</blockquote>

It can, yes, but there are alternatives, like ceramic lenses, From what I've been reading very recently it seems that the easy and cheap way to do it is to use a scanning mirror instead of a lens, as pretty much any sort of reflective surface will work OK at these long wavelengths. My guess is that an IR optical system like a reflecting telescope might also work well and avoid the need for expensive lenses.