HOWTO: NukeBox – An UltraViolet Sterilization Box

akiba 2020-03-25

Recently there’s been a lot of interest in using ultraviolet light as a potential way to sterilize protective equipment like N95 masks for reuse. Stanford Medicine just came out with a study on sterilization techniques and how they affect the performance of N95 masks with dry heat, steam,  and UV sterilization as potential techniques.

In this HOWTO, I’d like to go through the build and a bit of analysis of ultraviolet techniques for sterilization.


Nuke Box is a box designed for sterilizing surfaces. In putting it together, I had medical masks in mind which seem to be a shortage almost everywhere. Although re-used sterilized masks aren’t ideal, when having to decide between that and having no protection, having something is probably better than nothing. In that vein, if masks need to be sterilized, it’s important to try and make them as safe as possible. Also something like this could potentially be useful in developing countries that may need to re-use masks out of necessity and cost.

The Theory

UltraViolet Germicidal Irradiation (UVGI) is a way of attacking the DNA of pathogens so they can’t reproduce. Viruses are especially susceptible to this because they have no cell walls or cell membrane. Using high energy photons in the UV-C frequency band, it’s possible to damage the DNA in pathogens rendering them unable to reproduce and thus sterile.

With UV light, there are three frequency bands. UV-A (315–400 nm) would be the standard blacklight that makes neon colors all glow-ey. UV-B (280–315 nm) is what gives you sunburn and skin cancer. UV-C (100–280 nm) is germicidal and completely absorbed by the ozone layer in our atmosphere which also means it’s non-naturally occurring. We can produce it though using mercury fluorescent lamps and UV-C  LEDs.

With this in mind, it’s also important to wear proper safety equipment and cover up your body, eyes, and any exposed wounds since the UV light can also damage DNA in human cells.

The ability to damage pathogens isn’t dependent on the intensity of the UV bulb but rather the dosage. It’s a cumulative effect much like radiation so as long as you irradiate the surface long enough, it’s possible to render the pathogens on it sterile. This also means that you can adjust how quickly you sterilize equipment since increasing the intensity will increase the dosage rate for pathogens.

The dosage is measured in μWs/cm2 (micro Watt seconds/squared centimeter) or sometimes μJ/cm^2 (micro Joules/squared centimeter). Although the units are intimidating, the formula to obtain it is quite simple:

Dosage = (Intensity at that particular point) x (time exposed)

For example, in this particular HOWTO, I’m using a 10W germicidal fluorescent bulb with a rated output of 2.7W of UV-C light (UV light at 254 nm). This light theoretically puts out 1,364 μW/cm^2 of intensity at a point 7 cm away. I’ll go through the calculations in a bit.

The Covid-19 virus requires 5,000 μWs/cm^2 (5 mJ/cm^2) for sterilization. Based on the above formula, it would take 3.67 seconds of UV irradiation to inactivate the virus. There are other factors to consider though such as UV light warm up time, inaccuracies with the theoretical model, etc which is why you see people irradiating objects anywhere from 3 minutes to 30 minutes.

Source: Link

UVGI is a pretty straightforward approach. You take a UV-C fluorescent bulb and shine it on the surface you want to disinfect. Calculating the intensity actually depends on the shape of the light source. If we make some educated assumptions, the intensity calculation can be quite straightforward. These are the two assumptions that are important:

  1. The power radiated in the UV-C band is uniform across the whole light
  2. The distance we’re looking at is very small compared to the length of the light (r << L, where r is the distance we’re trying to calculate, L is the length of the light)

With those two assumptions, we can approximate the intensity at some point “r” from the fluorescent lamp as evenly distributed on a cylindrical surface. The formula to calculate the theoretical intensity would then be

(Total UV power) / (Surface area of a cylinder)


P / (2 * PI * r * L)

where “L” is the length of the lamp and “r” is the point that we’re trying to calculate the intensity for. I tried to make a beautiful CAD drawing to illustrate this but got frustrated and hand-drew it on a Post-It instead:

So with a 45 cm bulb that radiates 2.7W of power, assuming ideal conditions, the intensity will be 1,364 μW/cm^2 at a point 7 cm from the bulb. These happen to be the actual values for this build. We can also use this value and calculate the theoretical time it would take to sterilize a particular virus.

You may have noticed that I mentioned I’m using a 10W bulb but it only has 2.7W of power radiated in the UV-C spectrum. This is one of the key specs of the bulbs since the rest of the power is radiated in other frequencies (ie: the beautiful blue light that can be seen by us) as well as given up as heat.

Of course our assumptions are very ideal and we’re dealing with a real-life system. There will be fringe effects at the ends of the lamp, the power may not be evenly distributed, the bulbs take time to warm-up, etc. Hence we can take our ideal model and then add a healthy margin to it to make sure it’s safe. That’s why it may take 3.67 seconds to theoretically inactive SARS-CoV-2 under ideal conditions, but the medical studies are using 300 seconds (5 minutes) to 1,800 seconds (30  minutes). Any mistakes would be tragic so everyone is being extremely conservative on the sterilization time estimates.

Before we get to the actual build of the Nuke Box, two things:

  • Collaboration: It would be great to work with a lab that can help identify experimental results, especially for Covid-19. If you can help us test the sterilization rates, esp for Covid-19 coronavirus, please let us know. We’d love to collaborate to confirm and also improve the design.
  • Liability: This design would fall under the Good Samaritan Law. We can’t take responsibility if anything goes wrong with this design. These are strange times and we’re trying to become knowledgeable about new subjects as quickly as we can.

The Build

In this HOWTO, I’ll be building a box where the masks will be located at most 7 cm from the top and bottom light sources.

  • The first thing is SAFETY! Please cover exposed skin, any open wounds, and wear UV-blocking eye protection. Also try not to spend too much time very close to the lights while they are on.

  • In this particular HOWTO, I’ll be using off-the-shelf fluorescent fixtures. These can be easily bought and come with the required components (ballast, starter, T8 sockets) and wiring. I’ll also be using UV germicidal fluorescent tubes. In Japan, these would be the GL series (Toshiba GL-10 in this case) which stands for germicidal lamp and are made by Panasonic, NEC, and Toshiba. They come in sizes and power ratings from 10W to 40W. In the US, they would be the TUV (Tubular UltraViolet) series by Philips Lighting.

  • These are low pressure mercury fluorescent bulbs. The bulb filament ionizes the mercury atoms causing them to give off light at various frequencies, one of them being at 254 nm. This is very close to the peak absorption of DNA at 260 nm and the energy from the photons that hit the DNA damage it and render pathogens unable to reproduce.

  • They also give off some amount of visible light and generate heat. That’s what give them the cool, cyber-punkish glow…

  • This build is actually going to be quite simple. I’ll be taking two box bottoms and laying them on top of each other. I’ll also be setting up wire cord in a clothesline fashion and using spring-steel paper clips to attach them.

  • I start by drilling holes in the bottom box with a spacing that accomodates the width of an N95 mask which is about 15 cm.

  • Next is to put the fluorescent fixtures inside the box. I decided to drill a hole and cut the cable to the fluorescent lights so that I could fit the fixtures inside and have the switch on the outside. I mainly did this for my own convenience and because I’m comfortable working with electricity and soldering. Truthfully, you don’t need to do this part and can just leave the cables intact and peaking out the seam so you don’t have to do any cutting or soldering.

  • I drilled some holes in the box and added nuts and bolts to keep the fluorescent fixtures in place.

  • Then strung the cord through the holes that were drilled to make clotheslines for the mask. Once that’s done, you can just clip the masks to it. I used a double clothesline because it’s important to have the mask oriented so that the lights shine on the top and bottom directly.

  • I attached the light fixture to the other matching box bottom. I didn’t add the clotheslines though since this light will shine down on the masks.

  • Now we can put the Nuke Box together. Literally. We just put them together.

Slap on a kitchen timer and that’s the Nuke Box. I hope it’s a useful HOWTO. I have a personal stake in sterilizing masks for re-use. My partner’s mother and sister are nurses working at hospitals and clinics in Melbourne, Australia and London, England. My mother-in-law’s clinic doesn’t have enough masks and it’s likely the same in the UK. I’m hoping that if they end up having to reuse their masks, there is a way to make it safe for them and their patients.


Here are some of the improvements I’d make:

More Power: This is just a prototype. If we wanted to bulk sterilize masks, we’d be using the 40W long bulbs (think standard fluorescent lights). They’d accomodate more masks and sterilize them faster.

More Power: We can also put an array of lights on top and bottom. Having two lights installed and a single row of masks would double the intensity on each mask. This would cut the sterilization time in half.

Bigger Box: I’d also try to accomodate as many masks as possible without shading each other. Here’s a rough layout for having two rows of masks. If something like this is going to be used in a mass sterilization environment, you’d need to optimize the time and amount of masks you could sterilize in each cycle. Here’s a rough example of me testing out putting in two rows of masks into the prototype Nuke Box. You could fit in 6 masks total in this prototype box. If you did a full-sized 40W fluorescent box with multiple lamps on top and bottom, you could scale it to do a lot more masks in each run.


Here are more links to references and articles on using UVGI to sterilize masks and PPE:



I’ll add more as I find them.

Good luck to us all and much love to everyone in the world from Hackerfarm!


  1. Maybe this is a stupid question, but we have tons of tanning beds in the US- would those be use-able at all?

  2. Hi Jenny. Not a stupid question at all. The ultraviolet section of the frequency spectrum is divided into three parts. UVA is what tanning beds use. UVB is causes sunburn and in extreme cases, skin cancer. UVC is the most damaging, but is completely blocked by the earth’s ozone layer so it’s not naturally occurring on earth. They are extremely damaging to viruses, especially ones that don’t have cell walls for protection. Unfortunately tanning beds don’t have germicidal properties, which is probably for the best, since they would damage people that use them. Hope that answers your question 🙂

  3. It’s surprisingly difficult to find a florescent fixture with the right length and enough wattage by the ballast. Fixed LED fixtures have taken over that space. That’s what has been holding me up in the US.

  4. Yes, I’ve ran into the same problem. I’ve essentially started buying used fluorescent light fixtures or just fluorescent light fixture guts (ballast, starter, and sockets). Those are readily available since so many people are replacing them with LEDs. Ha ha ha. I feel like a vinyl record shopper…

  5. Hi Akiba,

    Very useful Documentation. We are working on to produce some Nuke Boxes In Brussels. (I like the name haha)
    To complete your documentation. I’ll recommend you to read those scientific publication that can help you to calcul the exposure time range.

    We’ll provide you our results as soon as we finish our tests.

  6. Totally, Awesome!

    Keep up the GREAT WORK!!

    Looking forward to the final product, once the Prototypes have been completed.

  7. Hi Akiba, This is so useful! Thank you so much!

    Everyone’s been talking about UVC but no one has explained the corelation between efficiency and wattage and distance.

    However, I have one question about calculating intensity – you say the formula of total power/surrface area of cylinder only holds true for r<l so does this formula for intensity fail in this case?

    Thanks in advance! looking forward to your reply. 🙂

    You can get in touch at

  8. @Nitish:
    For r<>L, it behaves as a point source where intensity drops as 1/r^2. If you are at r~L, exact theoretical intensity becomes harder to calculate but the formula falls somewhere in between. I’d also not rely so much on theoretical calculations and get a UV meter to measure real values.

  9. I dont have a box to house the lamp.
    What do think about using a mini fridge as a box (that has been gutted out/leaving just the shell casing)

  10. Looks like your design overcomes all of these factors
    that could hamper effectiveness.

    Diminishing power over increasing distance
    Angle of the exposed surfaces
    Surface shadowing

    You should market this

  11. Hi Akiba, thank you for the great article. We are trying to repliacate the UVC dosimeter with limited resources.

    We only have access to the following UV sensor breakout board, do you think Resistor modification to get UVC can be done on this?

  12. Yes. R3 is currently 1 Mohm. If you change it to 10 Mohm, it should be compatible with the software I wrote.

  13. A great work and very well explained.
    Do you think that aluminum foils on the walls of the box could significantly increase the power intensity or its reflectivity is not so good for uv-c wavelangth?

  14. Is there ozone produced with these bulbs and how do I test for ozone output if there is any. Thanks.

  15. Can you provide a link to the bulb fixtures? I’m having difficulty finding those. Thank you

  16. Why use a black container and not at least a reflective interior in the box, so that the light (presumably) would reflect around all angles and sides of the mask?

    Would an aluminum foil pan on the inside work? Or line the inside with aluminum foil tape? Or would that not reflect UV-C light?

  17. @Vince: The low pressure mercury bulbs use a quartz coating that filters out the wavelength that produces ozone. This may or may not what you desire. There are also specialty bulbs that will produce ozone.

  18. @Jeff: I think you may be having problems finding the fluorescent bulb fixtures which would pertain to whatever bulb you are using. Unfortunately recently manufacturers have moved to LED lights instead of fluorescent so those are getting harder to find. You may need to wire up your own light using an LED fixture and a ballast. You can find many tutorials on Youtube on how to do this and ballasts can be purchased via ebay or Amazon.

  19. @Brian: Aluminum foil would likely work. This post was written as a sample to demonstrate how to put together a UV light box. But please feel free to add foil, change colors, etc.

  20. We have found the light bulbs but not the basic fluorescent off the shelf mounting: “l be using off-the-shelf fluorescent fixtures. These can be easily bought and come with the required components (ballast, starter, T8 sockets) and wiring.”. Can you provide any pointers to locations to purchase? We have tried Lowes and Home Depot. When we say we are building a NukeBox, they look at us very strangely:) Thanks!

  21. First of all, I would like to thank Akiba for the excellent guidance and sharing those very useful information and resources. I managed to build a box with two 8W T5 UVC tubes from ebay. I used petri dishes to confirm the lights are working (at least on bacteria); I have tested with a 18 minutes exposure, you can see the result here ( I am waiting for the result for a 6 minutes exposure.

    I am concern if the tubes (no manufacturer marks) are producing ozone; there is a strange smell every time they are on, it is not a smell that I can associate with new electronics or glues of new builds. I have added an exhaust fan just in case.
    Here ( explains UV-C do produce ozone, but with “unbranded” tubes may not have the same “quality” built with doped quartz glass blocking 185nm wavelength from escaping.
    The question is that should I be concerned with the smell (associate withe ozone)?
    I am starting a new build with a larger box and 2 x 10W T5 tubes (from ebay/ China), should I add an exhaust vent (to outside)?

    Here are some photos of my build if you are interested.

  22. Thanks for another useful piece of information; now the smell makes sense (no pun intended)! I will still add an exhaust fan to my new build to minimise the smell absorbed by the masks.

  23. Hi great article. May I ask a question? If I use an e10 t8 gl 10w lamp in a box. How should I turn it on to sterilize? And do we need to line the box up with aluminum foil? Thanks!

  24. @Janus: You would need to set up a ballast and starter lamp/coil to get it going. It would be the same as in a standard fluorescent light fixture. I recommend just buying an old fluorescent light fixture to start and get the lamp going first.

  25. Also attempted to do this built. Due to the non availability of small UVC tubes, I used a 75 watt (25w UVC) water Sterilisation tube.
    Using a chest freezer clad with reflective polished aluminum all around.

    I don’t have a dosimetry testing, so doing 30 minutes per set of masks – in the enclosed environment I do get a build up of Ozone (at least that’s what I think it is, smells like a Old photocopier machine)

    Interestingly, that smell remains on the masks even days after Sterilisation, anyone else make that experience?

  26. “This light theoretically puts out 1,364 μW/cm^2” need to be changed to “This light theoretically puts out 1,364 mW/cm^2”

  27. @Akiba: After working out of nukemeter on UVC LED, I decided to try UVC tube. While testing with the nukemeter output, I would like to know how much my tested UVC tube discrepancy between nukemeter vs Inverse Square Law. I noticed you used 45cm as bar light length for GL10 to the calculation P / (2 * PI * r * L), but GL10 length is ~33cm according to spec, how to get the length value? Thanks.

  28. Thank you Akiba. Your input has been very valuable.
    One year on, I am building a UVC-box out of an old pc tower with four 9W lamps. The fixture is actually from a nail lamp featuring four G23 sockets. FFP2/KN95 masks are available again but getting some with reliable sources and quality is actually hard. China is no source for real FFP2/KN95 masks these days. Even pharmacies and agency distribute fake masks. 🙁 So reusing known-good masks is an option right now, with all those new covid-19 variants coming up.
    This box comes with the sensor you’re using and an esp-8266 to make up a frankenstein-dosimeter.
    Thanks and may you all stay sound and safe.

Post a new comment