How far beyond visible spectrum can we go?

[MacM]

Could a new sensor expand the light spectrum for cameras? Also, have you ever thought about narrowing down a specific light spectrum, so that for example, the camera only sees from 300 to 310 NM? This is probably tricky, because the increased exposure means that there will be light leakage. The reason I ask is because narrowing down a very narrow range of light could make for higher quality, since the lens is only focusing a specific range of light. Instead of expanding spectrum, we narrow it down.
A 720nm IR filter on top of a 300nm UV filter, what light gets through? I know that it doesn't seem as though any light should be visible, except that something tells me that there is light getting through. I mean, it's not as if the glass becomes totally opaque, right?

 

[Tom Axford]

Could a new sensor expand the light spectrum for cameras?

The lens is much more of a problem than the sensor.

Ordinary camera lenses normally block nearly all of the UV light. Special lenses are used in industrial applications that use UV light (e.g. here), but they are expensive.

 

[ProfHankD]

Could a new sensor expand the light spectrum for cameras? Also, have you ever thought about narrowing down a specific light spectrum, so that for example, the camera only sees from 300 to 310 NM? This is probably tricky, because the increased exposure means that there will be light leakage. The reason I ask is because narrowing down a very narrow range of light could make for higher quality, since the lens is only focusing a specific range of light. Instead of expanding spectrum, we narrow it down.
A 720nm IR filter on top of a 300nm UV filter, what light gets through? I know that it doesn't seem as though any light should be visible, except that something tells me that there is light getting through. I mean, it's not as if the glass becomes totally opaque, right?

You can buy notch (band-stop) filters with very narrow wavelength ranges rejected. They tend to be a little pricey, but <50nm notch widths aren't too hard to find. You also can buy band-pass interference filters with transmission ranges as small as 10nm.

As for how wide a spectrum you can sample, typical sensors are good from something like 1200-300nm, with NIR blocking filters taking out most of the long end. Conventional lenses clip NUV pretty close to where the sensor response falls off, but I think they can go as long as about 2500nm NIR. Without changing the sensor, it also would be possible to use an anti-Stokes phosphor to convert 1500nm into 800-1200nm, thus enabling a conventional sensor to record it.

Longer wavelengths (true IR) are normally sensed using (cooled) microbolometers, which essentially sum energy from photons rather than getting a unit charge per photon. Shorter wavelength UV tends to require Quartz lenses, and even then it's pretty hard to get good sensitivity at 250nm.

Of course, there are all sorts of other mechanisms for sensing other EM bands... but shorter wavelengths tends to be harmful to living things and longer ones quickly limit resolution. One doesn't really think about it much, but diffraction spreads NIR about 2.5X as much as NUV; the CoC you get at f/11 for NUV is hit at f/8 for green and nearly f/4 for NIR. Thus, you need a very fast lens and/or large pixels for EM much longer than NIR.

 

[MacM]

How did they obtain the image at far right? https://en.wikipedia.org/wiki/File:Infrared_portrait_comparison.jpg
Can I get a camera modified specially for exploring the far extremes of the spectrum of light. I think it would be compelling, maybe I could discover anomalous phenomena. I've written before about wanting a medium format camera with super low resolution so that I can discover anything invisible to us. It's a truly astounding topic.

 

[ProfHankD]

How did they obtain the image at far right? https://en.wikipedia.org/wiki/File:Infrared_portrait_comparison.jpg
Can I get a camera modified specially for exploring the far extremes of the spectrum of light. I think it would be compelling, maybe I could discover anomalous phenomena. I've written before about wanting a medium format camera with super low resolution so that I can discover anything invisible to us. It's a truly astounding topic.

1500-1700nm is beyond a normal CMOS/CCD sensor, but viable with anti-stokes phosphor and a conventional lens. I think the easiest would be to get a sheet coated with anti-stokes phosphor, image on the sheet, and then photograph the sheet. Of course, a real IR camera (night vision scope, FLIR) can do that and longer wavelengths directly using microbolometers....

To 1100nm is a lot easier with ordinary cameras; take a look at http://www.maxmax.com/

 

[petrochemist]

Silicon becomes transparent above 1150nm so can't be used for detectors going further into the infra red. Indium/Galium/Arsenide is used for detectors going further into the IR, typically covering 800-1700nm. Other technologies are available taking detection right through the IR band but there isn't a one suits all detector. IIRC our spectrometer at work uses a lead selinide photomultiplier which isn't suitable for imaging at all.

(You can get thermal cameras which are made to record longer wavelength IR, they are however both very expensive & low resolution (generally under 1MP)

At the other end of the spectrum, as has been mentioned already glass doesn't transmit UV well, but even using exotic lenses made from Quartz you rapidly run into problems. Top surface mirrors work slightly better, having no significant absorbance themselves. but Air itself blocks UV below 190nm in fairly short distances (no more than a few meters). This is really just as well otherwise sunlight would kill all life on Earth 8)

 

[drh681]

is imaged with radio waves.

 

[MacM]

I read that sensors keep getting more sensitive to the infrared spectrum. Is that also true for Ultraviolet? Could I build a camera specially for this, or is there no practical way of doing so? I guess it would probably be extremely expensive.

 

[ProfHankD]

I read that sensors keep getting more sensitive to the infrared spectrum. Is that also true for Ultraviolet? Could I build a camera specially for this, or is there no practical way of doing so? I guess it would probably be extremely expensive.

Like I said above, Anti-Stokes Phosphors are the only way I know to get longer than about 1200nm with a conventional sensor, and for that you could still use a conventional lens. Heck, low-resolution IR microbolometers are also pretty cheap, so deeper IR is usable. Getting UV much shorter than around 300nm is much harder. Your lens can't be made of glass, and exposure to that range of UV kills live cells (they are often used for sterilizing surfaces)... you don't want to spend time around bright UV lights because they can seriously damage human skin.

 

[OpticsEngineer]

see page 4, covering the wide range of technologies available

http://www.hamamatsu.com/resources/pdf/ssd/image_sensor_kmpd0002e.pdf

lest one thinks imaging really short wavelengths is exotic, remember x-rays have been taken on film for a very long time.

 

[MacM]

http://www.randombio.com/uv.html
What does the author mean when he writes, "Unlike in infrared, there is no partial conversion for ultraviolet. Only a full conversion extends the camera's response into the UV region"

 

[BobORama]

I read that sensors keep getting more sensitive to the infrared spectrum. Is that also true for Ultraviolet? Could I build a camera specially for this, or is there no practical way of doing so? I guess it would probably be extremely expensive.

Like I said above, Anti-Stokes Phosphors are the only way I know to get longer than about 1200nm with a conventional sensor, and for that you could still use a conventional lens. Heck, low-resolution IR microbolometers are also pretty cheap, so deeper IR is usable. Getting UV much shorter than around 300nm is much harder. Your lens can't be made of glass, and exposure to that range of UV kills live cells (they are often used for sterilizing surfaces)... you don't want to spend time around bright UV lights because they can seriously damage human skin.

Anologous to your frequency doubling phosphors and similar systems on the IR end, many moons ago we wanted to image items using the low pressure Hg line around 270nm. We had a fluorite-quartz lens element focused onto a phosphor coated ground glass screen ( think a focusing screen ) - which itself served as a UV cut filter. This provided a 6" diameter image circle which was then photographed to slide film.

We eventually improve this but using a 3 phosphor system each of which responded selectively to different spectral lines over the UV spectrum. So a false color UV imaging system.

The resolution was, not great. But it was sort of magical as you could "see" in UV by looking at the screen or through the view finder on the camera.

-- Bob
http://bob-o-rama.smugmug.com -- Photos
http://www.vimeo.com/boborama/videos -- Videos
http://blog.trafficshaper.com -- Blog

 

[petrochemist]

http://www.randombio.com/uv.html
What does the author mean when he writes, "Unlike in infrared, there is no partial conversion for ultraviolet. Only a full conversion extends the camera's response into the UV region"

An interesting site that I'd not seen before in my frequent googling the subject. TFS

I think what was meant is that the standard conversions include multiple options from NIR 720nm, 860nm etc. But the only standard conversion that is commonly available that helps much UV is the replacement of the hot mirror with quartz (Known as Full spectrum conversion). Despite the claims it is certainly possible to find filters that will only allow smaller sections of the UV to be recorded - finding affordable ones that block visible & IR but not UV is quite a challenge. combinations of technical glasses can get quite close but reduce the UV significantly too.

One thing I spotted that may be misleading is the transmission of acrylic lenses. Plastics including acrylic are usually sold with a UV absorbing additive, but you can get acrylic without this and it's UV transmission is considerably better than glass. It might need some considerable searching to find lenses made of the UV transmitting grades however.

A few ordinary lenses are usable for UV work - generally these are slower primes (less glass means more UV) Early examples of the El Nikkor 80mm enlarger lenses are apparently good for this - according to extensive testing by Dr Klaus Schmit .

 

[BobORama]

One thing I spotted that may be misleading is the transmission of acrylic lenses. Plastics including acrylic are usually sold with a UV absorbing additive, but you can get acrylic without this and it's UV transmission is considerably better than glass. It might need some considerable searching to find lenses made of the UV transmitting grades however.

This is true, but polyacrylates are generally terrible ( opaque ) below 300nm. polycarbonates and others even worse. There are some that get you down to 250nm.

Sapphire has the widest usable range from something like 100nm to 5000nm - but its not a practical material compared with quartz which is easier to work. It is a great window material, however.

An eyeglass lab can make you these lenses if you can get a slab. Make friends with an optician at a local eyeglass mill. One guy had a lens made out of an exotic plastic for $75, the big problem was getting the slab of plastic down to the "standard" blank size the mill wanted. The plant operations guys did that on their lathe. The whole thing was fabricated in about 2 hours. You just need to befriend an optician at one of these eyeglass mills.

-- Bob
http://bob-o-rama.smugmug.com -- Photos
http://www.vimeo.com/boborama/videos -- Videos
http://blog.trafficshaper.com -- Blog

 

[MacM]

I have a film camera called a Zenit. I'm sure the lens transmits quite well, except that it's not practical for modern cameras.

 

[MacM]

http://randombio.com/uv2.html

 

[petrochemist]

I have a film camera called a Zenit. I'm sure the lens transmits quite well, except that it's not practical for modern cameras.

In what way ??

My Zenit has M42 mount (I believe some where M39).

M42 lenses are fully manual but are used on Canon & Pentax DSLRs as well as most types of mirrorless cameras.

Being relatively old lenses most M42 lenses are uncoated (some coatings can reduce UV) but still have quite a bit of glass. I see no reason to suspect the lens will transmit UV particularly well and have measured the common Helios 44m on the spectrometer at work. It's transmission dropped to 1% at 342nm & was under 5% at 350nm. Both the Tamron adaptall2 lenses tested performed better.

Results are online at http://global-infrared.freeforums.net/thread/80/lens-transmission-wavelength-repost-forum

 

[Joe Pineapples II]

is imaged with radio waves.

Thats right - mostly at 21 cm with various degrees of red shift.

Joe

 

[D Cox]

I read that sensors keep getting more sensitive to the infrared spectrum. Is that also true for Ultraviolet? Could I build a camera specially for this, or is there no practical way of doing so? I guess it would probably be extremely expensive.

Like I said above, Anti-Stokes Phosphors are the only way I know to get longer than about 1200nm with a conventional sensor, and for that you could still use a conventional lens. Heck, low-resolution IR microbolometers are also pretty cheap, so deeper IR is usable. Getting UV much shorter than around 300nm is much harder. Your lens can't be made of glass, and exposure to that range of UV kills live cells (they are often used for sterilizing surfaces)... you don't want to spend time around bright UV lights because they can seriously damage human skin.

Including the cornea.

One way around the lens problem could be to use a pinhole or a zone plate. There are also those flat "lenses" that use nanostructures.

 

[D Cox]

I have a film camera called a Zenit. I'm sure the lens transmits quite well, except that it's not practical for modern cameras.

In what way ??

My Zenit has M42 mount (I believe some where M39).

M42 lenses are fully manual but are used on Canon & Pentax DSLRs as well as most types of mirrorless cameras.

Being relatively old lenses most M42 lenses are uncoated (some coatings can reduce UV) but still have quite a bit of glass. I see no reason to suspect the lens will transmit UV particularly well and have measured the common Helios 44m on the spectrometer at work. It's transmission dropped to 1% at 342nm & was under 5% at 350nm. Both the Tamron adaptall2 lenses tested performed better.

Results are online at http://global-infrared.freeforums.net/thread/80/lens-transmission-wavelength-repost-forum

That lens would also work on a Sigma SDQ with a simple adapter.

The Nikon enlarger lenses have a reputation for being good in the near UV, at least down to 350nm. See here for examples:

http://www.savazzi.net/photography/el-nikkor_uv.htm