DIY Spectroscopy: Ideas RE test rig for UV-C reflective materials (E=R+A+T)

[pipe2null]

I'm trying to figure out a test rig to measure UV-C absorbance/reflectance/transmittivity of readily available materials.  This is not an easy subject to tackle in a home lab/workshop, but I'm attempting to build a test rig that gets at least some usable ballpark numbers.  I have my initial idea illustrated below.  If you know of any good reference books on the measurement side of things, please recommend.

Background: I have a different project that intends to use UVC to kill viruses and whatnot, but I've realized that information on UV-C absorbance/reflectance/transmissivity of readily available materials is virtually non-existent.  I do not have the tools to effectively use sheet aluminum for UVC light housing and reflectors (aka bioreactor virus killer).  Normal UVGI R&D would only use materials with known characteristics, and aside from aluminum, little if any shopping could be done at a local hardware store AFAIK.  But my project isn't "normal", and the whole point of the project is to only use materials commonly available at the hardware store (and possibly common 3DP filaments) and UVC sources easily obtained on eBay or Amazon with somewhat low lead times.

I've been using Wladyslaw and Kowalski's "Ultraviolet Germicidal Irradiation Handbook" for reference and it covers quite a bit, but it does not list material characteristics for every substance in the world, nor do consumer product labels typically include statements like "Great for use as a UVC reflector/absorber!".   



ANYway, on to my tentative first draft of a first draft of the test rig design.  Since transmissivity, absorbance, and reflectance are relative to the total irradiance, it is not necessary to use a UVC power measurement device that takes accurate absolute measurements, and you also do not need a UVC source with known UVC wattage.  As long as the UVC wattage is sufficient to get readings on a photodiode that has linear response, and your test rig has a fixed known distance between source and detector, you can calculate a material's properties based on measurements relative to the open-air photodiode current since open air has roughly 100% transmittance (I think?).  That gives you an arbitrary number of arbitrary units that represents the total irradiance incident on detector, but since the result is a ratio, the units and arbitrary numbers cancel out resulting in a percentage.  Yes, still need to subtract dark and/or ambient current first.  NOTE:  I'm still working on a cheap but effective UVC measurement probe, so references to "photodiode current" are interchangeable with whatever arbitrary "number" of arbitrary "unit" is obtained from a UVC-sensitive device.

At least that's the idea...

Please correct me if I'm wrong:
To get transmissivity: Cover the detector active area with a material sample of known thickness and calculate ratio of photodiode currents compared to open air diode current at same distance from UVC source.
To get absorption: Calculate by 100% - (transmissivity% + reflectivity%)
To get reflectivity:  This one is more of a headache for many reasons, including diffusion...  Here's my initial thought on a test rig:


The goal is not lab-grade materials analysis, it's to get a good enough approximation to make sure wall thicknesses are enough to prevent UVC from leaking from chamber and get a rough idea of how much reflection is going on inside a chamber of arbitrary dimension to help size UVC source in a "conservative but not overkill" way.  The same issue applies for coating the inside or outside of the chamber with different types of easily available material (products that claim to be "aluminum" foil tape, flat black or "aluminum" paint, etc.).  Case in point: types of "PTFE" are known to have excellent UVC reflectivity, but types of "Teflon" are sometimes used to coat bulbs because it is essentially transparent to UVC.  To a lay-person like myself, I used to believe that PTFE and Teflon were essentially the same thing, but I stand corrected...

Thoughts?

[Kleinstein]

The transmissive part is relatively easy. Most photo-diode and similar detectors are reasonable linear, so not problem there. The slight difficulty is more with scattering. So the detector should be relatively close behind the sample, so that most of the scattered light from the backside is captured too. One could in addition move the detector sideways to get an idea of the scattering as additional information.

To get rid of background (e.g. from sun-light) one could modulate the UV source and only look at the AC part. This is standard practice for optical measurements and can use a lockin amplifier for detection of even weak signals. A LED as source makes the modulation very easy.

The reflection part is really difficult, as there is no good 100% reflective reference material. So it may be a 2 step process: with a setup similar to shown measure relative to a reference material (like PFTE, surface Al mirror) and in a separate setup measure that reference material. For mirror one could compare the direct path with a mirror image. This still has the loss due to scattering though. For a more scattering surface (e.g. PTFE) one could assume full scattering and calculate from the geometry. Instead of the math one could do an analog measurement with visible light (e.g. red LED) and a known white color (e.g. PTFE, paint).

Teflon is the brand name (I think DuPont) for PTFE.  As a test material a few layers of the plumbers tape (sealant) is probably a good source.   

[pipe2null]

Thanks for the input!

The UVGI Handbook I referenced has a table that includes a small assortment of materials and their UVC reflectivity.  "ePTFE (tm WLGore)" is 99+% reflective and "Spectralon (PTFE)" is 95+% reflective.  But another section of the book that discusses UVC source bulb characteristics mentions Teflon being used to coat bulbs for protection since some types of Teflon are nearly transparent to UVC, ie ~0 % reflective.  So, that's where my confusion came from, and it prompted a learning experience that Teflon is just a brand name as you mentioned.   Heh. 

For "reference" measurements, I intended on only using open-air measurement to get total irradiance on the photo-diode active area at that specific distance from source instead of trying to obtain materials with known characteristics for comparison.  Plus the obligatory dark current measurement plus UVC source off for ambient measurement...  If I've missed something, please correct me.

Thanks for pointing out the transmissive scattering, I totally forgot that part.  My intention was to place the sample material directly against the photo-diode, so hopefully that will cover it (pun-not-intended).  I'm considering the light modulation you mentioned, thanks for the tip.

What do you think about this for reflective measurement, using open-air as reference.  Is this along the lines of what you were saying, using mirror image?
Completely different test rig than illustrated before, so ignore previous idea: Use an aperture that is ~10x the active area of the sensor.  Mount sensor 2cm behind the center of the aperture, and take open-air measurement.  Rotate sensor 90 degrees with a pivot point that is 1cm behind the center of the aperture so the detector is now 1cm back and 1cm to the side (from aperture center) and blocked from direct exposure from UVC source, and then place a planar sample of material-under-test (MUT) at a 45 degree angle behind the aperture oriented so the "virtual image" of the sensor is located identically as the physical image was during the open-air measurement.  If the MUT is 100% reflective with no scattering/diffusion, then the sensor readings would be identical for both measurements.  As long as the aperture is not too small, there should be enough active area of the sensor to pick up some scattering coming from indirect paths as well...

[aussie_laser_dude]

I've got a suggestion for transmission + absorption measurements, reflection may require some additional optics like a lens / mirror.

PARTS:
- cheap silicon photoconductive detector (~$2 from ebay or digikey) just make sure it detects your wavelength (about 200 nm)
- A ~50 kohm resistor
- Voltmeter / multimeter
- Battery power source (whatever voltage, 3V, 5V, 9V etc)
- Breadboard / pcb prototyping board / wires + hot glue or whatever to connect the stuff together
- uv-c diode with it's own power + switch
- calculator



Steps.
1. Wire the battery, photoconductive sensor & resistor in series.
2. Point the uv-c diode at the sensor, have them close.
3. Turn uv-c diode on. Make sure battery to sensor is also connected.
4. Connect voltmeter across the resistor, set to volt range if multimeter.
5. You should see a voltage display on the voltmeter, put some metal in between diode and sensor, the voltage should change with the increasing/decreasing uv-light. The change will be proportional to the light falling on the detector.

That's your basic uv-c transmission + absorption detector.
  Lights / sunlight may affect your measurements. A lock-in amplifier or an enclosed uv-vis spectrometer would solve this problem but sounds like you want a cheaper dollar project, so doing measurements in a dark room or covering with a cardboard box or something should work ok.

(I made this circuit just yesterday btw)

For a slightly more advanced design that isn't as susceptible to background light, and still cheap. You could make the uv-c diode blink on and off at ~10Hz, just switch the multimeter to AC voltage.


Hope this helps

[Kleinstein]

What do you think about this for reflective measurement, using open-air as reference.  Is this along the lines of what you were saying, using mirror image?
Completely different test rig than illustrated before, so ignore previous idea: Use an aperture that is ~10x the active area of the sensor.  Mount sensor 2cm behind the center of the aperture, and take open-air measurement.  Rotate sensor 90 degrees with a pivot point that is 1cm behind the center of the aperture so the detector is now 1cm back and 1cm to the side (from aperture center) and blocked from direct exposure from UVC source, and then place a planar sample of material-under-test (MUT) at a 45 degree angle behind the aperture oriented so the "virtual image" of the sensor is located identically as the physical image was during the open-air measurement.  If the MUT is 100% reflective with no scattering/diffusion, then the sensor readings would be identical for both measurements.  As long as the aperture is not too small, there should be enough active area of the sensor to pick up some scattering coming from indirect paths as well...

This sound perfectly reasonable. There is still a little uncertainty as the scattering paths use slightly different angles and longer paths, so there may be some error in not getting enough scattered light. It would not need a large detector, more like a large enough sample and a detector small in comparison to the distance to the sample.  to avoid the direct light it may take a little more than 45 degree angle, but the principle sound good.

For the transmission is could be enough to similar have a large sample and a detector that also detects light from an angle.

There is one more point to watch for: the LEDs are likely not pure UV-C but will also have some intensity at longer wavelength, though not much. So with relatively strong absorbance the little light that still comes through could be longer wavelength. Similar there may be fluorescence light that may cause errors, if the detector is also sensitive to longer wavelength. A SiC or similar sensor could thus be a good idea.

As for the reflectance of PTFE: the white material is not 100% dense, but has tiny pores and these interfaces scatter the light.  If 100 % dense PTFE would be essentially transparent like glass.

[pipe2null]

As for the reflectance of PTFE: the white material is not 100% dense, but has tiny pores and these interfaces scatter the light.  If 100 % dense PTFE would be essentially transparent like glass.

So...  Let's see if I finally understand this.  Is the following statement correct?:  100% dense PTFE and Teflon (yes, its a brand, but the two are often thought of as synonymous) are both ~100% clear to UV like a pane of glass is to the visible spectrum, which is why it is sometimes used to coat and protect UV lamp bulbs.  But when the "density" of microscopic air bubbles inside the PTFE material is sufficiently high, the PTFE/Teflon becomes ~100% reflective due to the interaction of light between the many "air<->PTFE" interfaces, thus why ePTFE (AKA "expanded" PTFE) has 99+% reflectance.  So, most/all the reflection comes from something like Snell's Law, and is not "directly" due to the material itself?

I'm putting together a Digikey order today for a minor grab bag of part candidates to try out.   I already ordered my 6.5 digit bench DMM (0.1 nA max resolution) and should have it in a few days, so I should be able to try a proof of concept without needing circuit design ahead of time.

'Course, I should PROBably figure out the most convenient way to protect myself from UVC without requiring an enclosure to start with.  Make sure I don't bake myself and/or give myself skin cancer while I tinker with various unshielded UVC sources over an extended period of time, and add the shielded enclosure later to minimize ambient (and for safety) after I know it's working...    My workbench surface is made of 3/4" wood, which I think is pretty absorbent, but the walls are white paint of unknown type, so not sure if leather gloves/skin covering + some type of face shield or plain old sheet of glass (not sure minimum thickness) as an open air movable barrier to sit behind would be sufficient.  Hmm...

[Kleinstein]

For the reflection some plastics are also locally inhomogeneous (e.g crystals) but AFAIK PTFE is amorphous.
White material is usually inhomogeneous like a powder of material that is otherwise transparent. Usually a high index of diffraction also helps.

For the detection I would really consider an extra amplifier (trans-impedance amplifier TIA, OP circuit). These are not that complicated and can easily be lower noise than the usual DMM.  The TIA should be close to the detector to minimize parasitic capacitance.
For comparison I would also (first) try a visible LED - here one can find better compassion values and can see the adjustment.
For the UV part, just cover the setup with a box (even if just cardboard). One would not see much from the UV anyway. One may need this anyway to reduce background light.
The usual LEDs are usually not as bright to worry about the skin, if at all the eyes have a problem.