Question: Is there a way to make a cheaper "UV Hound"?

liz is asking a question about air-quality
Follow this topic

by liz | October 24, 2019 19:03 | #21284

I'm posting on behalf of the amazing Houston organizer Bryan Parras, who shared some insights with me on the phone the other day.

In a settlement from a lawsuit against local industry, decorated Port Arthur environmental activist Hilton Kelly (a Goldman award winner) received 2 UV mass spectrometers, known as "UV Hound": . They each cost between 20K and 30K, and each can monitor abt 200'-500' of airspace.

image description

Bryan Parras mentioned that setting devices like this along the fenceline to catch any emissions leaving the industrial site and entering fenceline communities could be quite effective.

  • are any other devices that are cheaper and similar?
  • are any other groups are using these and, if so, what are their experiences?


Gwen Ottinger (who is a member of Public Lab's board of Directors) mentions the UV Hound in her book Refining Expertise: How Responsible Engineers Subvert Environmental Justice Challenges.. I've linked to the pages in Google Books if you want to see the relevant pages). She talks about it in relationship to the buckets used by the Bucket Brigade (her opinion on the buckets is quite favorable).

Is this a question? Click here to post it to the Questions page.

oh ok! yes, in this section of the book, looks like Gwen is using a UV Hound that was in fact rented from Hilton Kelly's organization Community In-Power and Development Association. Can the world get any smaller...

Id love to hear of how the "hound" data has been used effectively in communities. After the data was collected, what did they do with it...Are there examples in Gwen's book?

Is this a question? Click here to post it to the Questions page.

I also want to know about this type.

Reply to this comment...

Hi, just found it interesting when i read about the UV Hound - it's actually not a mass spectrometer but an optical spectrometer not too dissimilar to our kits, but in the UV range. The technique is called Ultra-Violet Differential Optical Adsorption Spectrometer (UV-DOAS) and they said they use Beer's Law just like our activity here: #beers-law

The readout looks really similar to although it's in a range we can't detect - UV, so down to 200 nanometers (our kits only go down to 400, pretty much same as human vision)


Also it needs to pass through 1000+ feet of open air, and get reflected back in. probably because air is so transparent:


I bet the UV image sensor is super expensive, driving the cost up.

They mention it draws on EPA Method TO-16R - which is actually for infrared -- - "Long-Path Open-Path Fourier Transform Infrared Monitoring Of Atmospheric Gases"

Here's a gas industry video with more info: -- it's about the OPSIS UV-DOAS, which is another product that uses the same technique. But I wasn't able to figure out the price.

The OPSIS site has a nice diagram of how their spectrometer works; it's basically the same as our kits but in UV, with a really long beam path:


They even use the same spectral matching technique as us. But their spectrometers are probably really well intensity calibrated or have a great baseline comparison spectrum and very little noise or drift.

Is this a question? Click here to post it to the Questions page.

Deep UV or DUV is pretty hard to detect. This article talks about it:

(These are some cameras that can do it, but I'm not sure if it's necessary for them to be cryogenically cooled or if these are that for a different reason:

One idea that the article explores is using a fluorescent coating that will glow, so the camera doesn't have to detect DUV, but can photograph the coating:

Another method involves coating the substrate with a wave-shifting coating such as Metachrome II or various phosphors. These coatings fluoresce in the visible portion of the spectrum. The fluorescence signal can traverse the substrate more easily than short-wavelength UV and, therefore, is detected. The disadvantage is that half the light generated by the fluorescence process radiates away from the sensor and is never detected. The coatings also degrade the MTF of the camera system and can introduce image-retention issues if the camera is running at a high frame rate.

And the lenses have to be made of quartz:

Standard glass lenses will not work well in the deep-UV band, since the typical lens glass absorbs strongly below 300 nm. Special lenses made with quartz (fused silica) or calcium fluorite are required. There are few off-the-shelf sources for these lenses, and the cost is quite high compared to visible-light optics. Microscope objectives are available for confocal microscopes operating in the DUV band.

Big thanks for unpacking all this @warren


I think the takeaway is that searching for a price for the OPSIS, or pricing for a different UV-DOAS device that could substitute for the UV-Hound, could get you a moderately lower price, BUT the need for an expensive Deep UV (DUV) camera sensor and lenses makes this device reasonably expensive to start with.

BUT, for those interested in DIY spectrometry for UV, there are some experiments at #uv-imaging, but this hobby imagery website also has a good overview, saying some off the shelf cameras can actually photograph down to 200 nanometers:

LifePixel offers some UV conversion cameras for ~$475, but it's not clear that they'd go much below 300 nanometers, according to this chart from


This USB microscope camera goes down to 200nm, but i can't find a price:

OK, this is weird and interesting. Maybe it relates to the above suggestion of a fluorescent coating, but this product seems to 'convert' UV from the 193 - 360nm range into visible light for a CCD camera:

Yes: "The UV image converters are fluorescent plates that convert UV radiation that is poorly imaged by silicon cameras into visible light that is then imaged onto the CCD of the camera. " -- more at


Here's one which converts 193-360nm light to visible, and it costs $3,272

So i guess the modified camera approach might be cheaper; especially since we don't need a lens.

I wonder if any CMOS sensor could work, like a cheap webcam, as long as our diffraction grating was reflective in the UV range. Those seem to be available for about $100:

Or maybe a scanning UV sensor like this:

For all this to work together, it'd be something like:

  • a really really powerful Xenon spotlight, aimed across a thousand feet of open air
  • reflected back with a first-surface mirror
  • bounced off of a reflective diffraction grating in a dark box like one of our spectrometers
  • aimed at a CMOS camera OR a spinning mirror plus a UV sensor
  • wait a long time to get a long exposure
  • then do it again when there is suspected benzine and other PAHs, kind of like in this post (image below)

I think it's a pretty tough challenge, and you'd have to try to separate out just the benzene detection from any other things you detected, just like we struggled to ID fluorescence spectra in a mixed sample of liquid.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

I wonder how the Hound is used if it needs a reflector to bounce the beam back. maybe you have to be on 2 sides of a site? You can be 1000' or more away, but it has to get back to the sensor again somehow.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Very Helpful information,thanks

Reply to this comment...

Sharing back a response from Bryan Parras with permission:

I'd love to be a part of developing a more open sourced remedy for folks. I'm working with folks right now, mostly looking for cost effective ways to do our own community based monitoring.

Reply to this comment...

Log in to comment