How far beyond visible spectrum can we go?

D Cox said:

ProfHankD said:

MacM said:

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.

Click to expand...

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.

Click to expand...

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.

Click to expand...

Exposures through pinholes & zone plates are long even without UV sensitivity issues. So they're not very practical having just the same issues of wanting more light & in the case of UV it's germicidal effects. The ideal diameter of a pinhole is also a function of the lights wavelength. Most are designed for green light since that's half way through the spectrum.

No variation in the lens mechanism will get you round sensor limits so getting round those will require multiple sensors.

petrochemist said:

petrochemist said:

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.

Click to expand...

Exposures through pinholes & zone plates are long even without UV sensitivity issues. So they're not very practical having just the same issues of wanting more light & in the case of UV it's germicidal effects. The ideal diameter of a pinhole is also a function of the lights wavelength. Most are designed for green light since that's half way through the spectrum.

No variation in the lens mechanism will get you round sensor limits so getting round those will require multiple sensors.

Click to expand...

I have seen (on the Getdpi web site) hand held pinhole photos, taken with a Sony A7s.

So I think it's quite possible that with current high sensitivity sensors you could get exposures of less than a minute with a zone plate designed for UV or for IR.

It's possible yes. But the usual softness of a pinhole image will be further degraded by the pinhole being used away from it's design spot. You cant design it for UV & IR, pick one, or accept the average which will be the visual design...

FWIW I found the formula I'd half remembered. the optimum focal length of a pinhole is equal to the square of the pinhole diameter times a factor base on wavelength.

Some of the factors I have noted are:

UV @370nm 1134
IR @ 750nm 561
IR @ 950nm 443

Visual range ~950-600

Using lenses few have any problems with IR throughout the NIR range sensors respond to.

The UV side is more of a problem but there's no shortage of lenses that transmit well down to at least 350nm. I've tested 15 of my cheap lenses & six of them transmit more than 1/4 of the light at 350nm. Nothing significant gets through down at 300nm but the El Nikkor 80mm does still manage 1% transmission at 306nm. According to Dr Klaus Schmitt this lens has minimal focus shift when going from visible to UV too.

I think you'll find 1% of f/4 works out about ten times brighter than f/128 (typical pinhole).

If you need to image below 300nm the pinhole/plate option may be the way to go, but I suspect a basic lens built with silica elements would be well worth the effort

petrochemist said:

It's possible yes. But the usual softness of a pinhole image will be further degraded by the pinhole being used away from it's design spot. You cant design it for UV & IR, pick one, or accept the average which will be the visual design...

Click to expand...

Yes, a pinhole has to be designed for the waveband you are using.

petrochemist said:

FWIW I found the formula I'd half remembered. the optimum focal length of a pinhole is equal to the square of the pinhole diameter times a factor base on wavelength.

Some of the factors I have noted are:

UV @370nm 1134
IR @ 750nm 561
IR @ 950nm 443

Visual range ~950-600

Using lenses few have any problems with IR throughout the NIR range sensors respond to.

The UV side is more of a problem but there's no shortage of lenses that transmit well down to at least 350nm. I've tested 15 of my cheap lenses & six of them transmit more than 1/4 of the light at 350nm. Nothing significant gets through down at 300nm but the El Nikkor 80mm does still manage 1% transmission at 306nm. According to Dr Klaus Schmitt this lens has minimal focus shift when going from visible to UV too.

Click to expand...

By "cheap lenses" do you mean ones with a simple design ? I would expect the total thickness of the glass to be a major factor.

I have a 63mm EL-Nikkor, but not the 80mm. Time to look out for one.

petrochemist said:

I think you'll find 1% of f/4 works out about ten times brighter than f/128 (typical pinhole).

Click to expand...

OK. In that case the only advantage of a pinhole would in focussing.

petrochemist said:

If you need to image below 300nm the pinhole/plate option may be the way to go, but I suspect a basic lens built with silica elements would be well worth the effort

Click to expand...

D Cox said:

petrochemist said:

It's possible yes. But the usual softness of a pinhole image will be further degraded by the pinhole being used away from it's design spot. You cant design it for UV & IR, pick one, or accept the average which will be the visual design...

Click to expand...

Yes, a pinhole has to be designed for the waveband you are using.

D Cox said:

FWIW I found the formula I'd half remembered. the optimum focal length of a pinhole is equal to the square of the pinhole diameter times a factor base on wavelength.

Some of the factors I have noted are:

UV @370nm 1134
IR @ 750nm 561
IR @ 950nm 443

Visual range ~950-600

Using lenses few have any problems with IR throughout the NIR range sensors respond to.

The UV side is more of a problem but there's no shortage of lenses that transmit well down to at least 350nm. I've tested 15 of my cheap lenses & six of them transmit more than 1/4 of the light at 350nm. Nothing significant gets through down at 300nm but the El Nikkor 80mm does still manage 1% transmission at 306nm. According to Dr Klaus Schmitt this lens has minimal focus shift when going from visible to UV too.

Click to expand...

By "cheap lenses" do you mean ones with a simple design ? I would expect the total thickness of the glass to be a major factor.

Click to expand...

No I mean as in inexpensive, I think only one was over £100

They ranged from modern telephoto zooms to a fisheye (where due to alignment issues I couldn't get more than 10% transmission at any wavelength) There was some weighting towards lenses that might be useful for UV, but all the ones in my MFT bag got tested too.

Modern coated lenses tended to be the worst, then particularly fast glass (with thicker lenses).

I really must get round to testing some more before the spectrometer at work is replaced with a model that doesn't have room in the sample compartment for lenses!

D Cox said:

I have a 63mm EL-Nikkor, but not the 80mm. Time to look out for one.

Click to expand...

Klaus did specify that it should be the early metal bodied version. It might have taken a few weeks for me to see one on e-bay, but it wasn't a difficult search.

D Cox said:

D Cox said:

I think you'll find 1% of f/4 works out about ten times brighter than f/128 (typical pinhole).

Click to expand...

OK. In that case the only advantage of a pinhole would in focussing.

Click to expand...

With modern cameras having Live-view or better still an EVF focusing isn't much of a problem if the sensor can see the light. If you need extreme DOF pinholes come into there own, but having played with them for visual shots I find them worthwhile only to be using an alternative approach.

I considered switching to film (which is quite sensitive to UV) but realized it would be a major pain wrt focus & metering...

The problem of finding materials suitable for wideband transmission is difficult. However for wideband reflection is a solved problem. So make a mirror lens.

• Highly reflective ( > 70% ) from 100nm - 8000nm
• Already proven for multi spectral imaging in astronomy
• No CA
• Commodity off the shelf components already exist

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

BobORama said:

The problem of finding materials suitable for wideband transmission is difficult. However for wideband reflection is a solved problem. So make a mirror lens.

• Highly reflective ( > 70% ) from 100nm - 8000nm
• Already proven for multi spectral imaging in astronomy
• No CA
• Commodity off the shelf components already exist

Click to expand...

That might be a great idea. I would prefer a lens that's not telephoto, could you give some advice? I think a micro 4/3 camera is what I should get I'm not sure I know if I want full spectrum or Ultraviolet, or if I should get a debayered camera.

BobORama said:

The problem of finding materials suitable for wideband transmission is difficult. However for wideband reflection is a solved problem. So make a mirror lens.

• Highly reflective ( > 70% ) from 100nm - 8000nm
• Already proven for multi spectral imaging in astronomy
• No CA
• Commodity off the shelf components already exist

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

Click to expand...

With the right design that would work really well!

Do be aware though that mirror lenses you can buy for photography are catadiotric designs using both mirrors & lenses. They might work better than standard lenses but will still have absorbance issues. A design more like a Newtonian telescope used at prime focus (no eye piece or camera lens) should fit the bill.

I've used my 600mm cat. for visual & IR shots (even both together) but haven't tried it with UV. If I can arrange a way to fit suitable filters into the coupling between my camera & my Newtonian scope I'll give that ago too sometime