Micrsocope: Zeiss Standard GFL
Objective: Leitz 40/0.7 NPL Fluotar ICT
Ocular: Olympus P15
Substage: Leitz ICT
Lighting: 100 W Filament lamp + IR pass filter
Camera: Aiptek 3 Mega Cam
Micrsocope: Zeiss Standard GFL
Objective: Leitz 40/0.7 NPL Fluotar ICT
Ocular: Olympus P15
Substage: Leitz ICT
Lighting: 940 nm IR LED
Camera: Aiptek 3 Mega Cam
I have been experimenting with near IR photography using my DIC setup and have some interesting results. A little while ago I identified some glass Watson filters that allowed me to take NIR landscape photos with my Aiptek 3 Mega Cam from which I have removed the IR blocking filter. I thought I would compare the use of one of these filters with that of a "pure" NIR source - a 940 nm 24,000 mcd LED for photomicrography. The results are above. I used diatoms as test objects, not for their IR properties, but because they give useful test patterns for focusing etc.
As you can see, using a 100 W filament lamp with a NIR pass filter works OK with DIC, but using the 940 nm LED does not. Not knowing the spectral characteristics of the glass filter, I can only guess at the reasons, but I suspect that the filter is passing shorter wavelength NIR than 940 nm, and that being only just off the end of the visible spectrum, the DIC polarising filters work OK and DIC thus works. At the longer wavelength of 940 nm, I can discern no polarising effect due to the microscope's polarising filters, and I think they are transparent at this wavelength, hence no DIC effect.
There are also some other annoying problems with the 940 nm LED - it gives hotspots and very uneven field illumination with a vignette effect, even though the light is being fed into the microscope using a randomised fibre-optic cable in exactly the same manner as the filament lamp illumination. There is also often a central "hotspot". This particular example image is not too bad, but some were quite extreme. Also, the microscope needs refocussing at 940 nm, whereas using the filament lamp the image is fairly well parfocal with the visible image without the filter. Chitinous specimens show no better transparency at 940 nm, but are more transparent in NIR using either source than in visible light. It seems to me that maybe the vignetting and hotspot effects may be due to the microscope or camera optics, rather than the uniformity of the lighting.
I think, therefore that the filament lamp is a more useful source of NIR and that going as far away from red as 940 nm is probably unnecessary and risks lower resolution. The filament lamp also has the advantage that the visible component can be used to assess the image in addition to the use of a monitor for viewing the NIR image.
I'd be interested to know whether anyone else has conducted similar experiments.