I wrote a tweetstorm on why 222nm LEDs are not around the corner, and given that there has been some discussion related to this on Lesswrong, I thought it was worth reposting here.

People interested in reducing biorisk seem to be super excited about 222nm light to kill pathogens. I’m also really excited - but it’s (unfortunately) probably a decade or more away from widespread usage. Let me explain.

Before I begin, caveat lector: I’m not an expert in this area, and this is just the outcome of my initial review and outreach to experts. And I’d be thrilled for someone to convince me I’m too pessimistic. But I see two and a half problems.

First, to deploy safe 222nm lights, we need safety trials. These will take time. This isn’t just about regulatory approval - we can’t put these in place without understanding a number of unclear safety issues, especially for about higher output / stronger 222nm lights.

We can and should accelerate the research, but trials and regulatory approval are both slow. We don’t know about impacts of daily exposure over the long term, or on small children, etc. This will take time - and while we wait, we run into a second problem; the Far-UVC lamps.

Current lamps are KrCl “excimer” lamps, which are only a few percent efficient - and so to put out much Far-UVC light, they get very hot.  https://link.springer.com/article/10.1134/1.1448635 This pretty severely limits their use, and means we need many of them for even moderately large spaces.

They also emit a somewhat broad spectrum - part of which needs to be filtered out to be safe - https://pubmed.ncbi.nlm.nih.gov/33465817/ - further reducing efficiency. Low efficiency, very hot lamps all over the place doesn’t sound so feasible.

So people seem skeptical that we can cover large areas with these lamps. The obvious next step, then, is to get a better light source. Instead of excimer lamps, we could use LEDs! Except, of course, that we don’t currently have LEDs that output 222nm light.

(That’s not quite true - there are some research labs that have made prototypes, but they are even less efficient than Excimer lamps, so they aren’t commercially available or anywhere near commercially viable yet, as I’ll explain.)

But first, some physics! 
The wavelength of light emitted by an LED is a material property of the semiconductor used.  Each semiconductor has a band-gap which corresponds to the wavelength of light LEDs emit. 

It seems likely that anything in the range of between, say, 205-225nm would be fine for skin-safe Far-UVC LEDs. So we need a band-gap of somewhere around 5.5 to 6 electron-volts. And we have options. Here’s a list of some semiconductors and band-gaps;  https://en.wikipedia.org/wiki/List_of_semiconductor_materials

Blue LEDs use Gallium nitride, with a band-gap of 3.4 eV. Figuring out how to grow and then use Gallium nitride for LEDs won the discoverers a Nobel Prize - so finding how to make new LEDs will probably also be hard. https://www.nature.com/articles/nphoton.2014.291 

Aluminum nitride alone has a band gap of 6.015 eV, with light emitted at 210nm. So Aluminum nitride would be perfect… but LEDs from AlN are mediocre. https://physicsworld.com/a/leds-move-into-the-ultraviolet/ 

Current tech that does pretty well for Far-UVC LEDs uses AlGaN; Aluminium gallium nitride. And when alloyed, AlGaN gives an adjustable band-gap, depending on how much aluminum there is. 

Unfortunately, aluminum gallium nitride alloys only seem to work well down to about 250nm, a bunch higher than 222nm. This needs to get much better. Some experts said a 5-10x improvement is likely, but it will take years.

That’s also not really enough for the best case, universal usage of really cheap disinfecting LEDs all around the world. It also might not get much better, and we’ll be stuck with very low efficiency Far-UVC LEDs, at which point it’s probably better to keep using Excimer lamps. 

But fundamental research into other semiconductor materials could allow much better Far-UVC LEDs. One candidate is hexagonal Boron nitride crystals. Another is diamond - which I don’t think will be practical to work with or build LEDs from, but “Diamond LEDs” sound awesome.

If we do find a new promising material, getting a good manufacturing process to make it and create the PN junctions will be critical. And unlike AlGaN, advances in other areas won’t provide benefits for a new material. 

Plus we won’t have the existing knowledge of how to make it work. Remember the Nobel-prize for Blue LEDs? It’s hard to figure this stuff out. But people haven’t had a strong reason to do so - disinfecting air changes that.

There’s a bunch of cool physics and simulation tech that lets research explore which possible semiconductors could be viable. That seems very worth doing, in case AlGaN doesn’t work, or something better can be found.

Unfortunately, there’s another (half) problem, which is really the first problem again. Remember, whichever LED semiconductor material we find that works, if we find any, probably won’t emit light at the same wavelengths as KrCl excimer lamps.

How do different light sources affect safety? We don’t know exactly. A better LED is likely to be higher output than 222nm lights, and will be at a slightly different wavelength. We might even need entirely new safety studies done at whichever new wavelength we find.

And even if we get those safety studies, getting from there to commercial viability will take time - and it’s unclear how expensive or difficult it will be to make these new LEDs.

This is not to say I’m pessimistic about the idea! I think there’s a >50% chance we find LED materials that work at significantly better than 10% efficiency, are cheap, and are safe for humans. (Conditional on 222nm being found safe.) But it’s a decade or more away.

That’s OK. We can plan for a decade or more in the future. As attention to the areas grows, people are doing exactly that. So as usual, I’m excited that the future will be awesome, and can be made much safer than the present, at least from biorisks.

But we definitely don’t want to fool ourselves into assuming there is a silver bullet around the corner. And even once it’s around, it won’t eliminate the need for multi-layered protection against future pandemics - and we should be investing in those other parts now as well.


New Comment
27 comments, sorted by Click to highlight new comments since: Today at 12:53 PM

I have a couple disagreements with this:

  1. Regarding regulatory approval, 222nm far UV-C irradiation is already legal (in the US) to levels that probably significantly reduce transmission (8-hour limit of 479 mJ/cm2 for skin). Various people I know think that the limits should be much higher, but even irradiation at current US limits seems very valuable -- & very safe -- to me.
  2. While KrCl lamps are expensive, I think this post overstates how unviable they are. I think an interested organisation could afford to install & run a bunch of these in an office (within the legal limits) basically right now, and see benefits that are worth the cost. (Someone throwing cost numbers at me could ofc change my mind here.)

I agree that the LEDs seem pretty hard.

479 mJ/cm2 for skin

Source for this? I see just 23 mJ/cm^2 given in this paper

That's the old limit; it was changed last year. See e.g. this figure from Blatchley et al

Seems like we still need to be be pushing for further regulatory changes. I saw a report from Kevin's lab that the limit for is 478 mJ/cm^2 per 8-hour day, and that ~5 minutes of light at around 100mJ/cm^2 kills ~99% of (bacterial) pathogens, whereas ~5 minutes at closer to 10mJ/cm^2 kills ~90% of (bacterial) pathogens, which means we would need a higher limit. Alternatively, the current limits seem to show that AlN LEDs at 210nm would be allowed to be much stronger than 222nm, (around double?) which seems like a good reason to try to work on improving them, rather than pushing for 222nm LEDs - or pushing for KrBr Excimer lamps at 208nm, though I understand they have other problems and aren't commercially viable at present.

(Also, why is UVA is considered so much safer for eyes than far-UVC? It seems from basic skimming that it causes long-term retinal damage...)

It seems to me like hospitals would be the kind of organizations that care a lot about reducing infections. Is there any analysis about what a hospital could achieve if they deployed it everywhere in reducing hospital-acquired infections?

Given the amount of money hospitals currently invest into keeping everything clean, they should be the kind of organizations that are first to adopt the technology even when it's relatively expensive. 

When Googling I found that Amazon.com doesn't sell any of the KrCl lamps. https://www.waronflu.com/product/222-nm-krypton-chloride-krcl-40-watt-far-uvc-excimer-bulb-222nm-first-uvc-f-series-40w-far-uv-light-24v-dc/ seems to sell them directly. On aspect of the page is that they expect the lamps to work for 4000 hours. That means that you would likely have to reinstall new lamps every year which adds to costs. 

It seems to me like that price should still be okay for health care applications. Health care applications just need more clinical evidence. Without patent protection the manufactors of the lamps likely don't have a good business model to pay for the lamps. Advances in lamp technology would have the nice benefit of producing a technology that can be defended with patents and thus making the business model work. 

While KrCl lamps are expensive, I think this post overstates how unviable they are. I think an interested organisation could afford to install & run a bunch of these in an office (within the legal limits) basically right now, and see benefits that are worth the cost.

Well it depends how much UVC power you need. If a KrCl lamp with a bandpass filter is 2% efficient you can still run it at 500 Watts electrical and you have a 10-Watt source. 500 Watts electrical is totally feasible from standard electrical outlets (240 Volts @ 2 Amps), and electricity is cheap compared to the cost of lost productivity and the economic cost of death.

I would be interested to hear from some more qualified people on what power density you really need for it to work.

It seems like a large amount of work of this post is being done by:

So people seem skeptical that we can cover large areas with these lamps.

Maybe the experts are thinking of large-scale deployments in schools, hospitals, airports, conference centers? I feel like numbers seem important.

Deploying UVC to disinfect large spaces might be infeasible, but would it be easier to have smaller UVC lamps inside ventilation ducts and let the air pass under them at a higher rate? You get a much closer lamp, much more airflow, and don't have to expose anyone to UVC directly.

They already have UVC disinfection in ventilation systems - but you need to circulate a ton of air for it to be really effective, so one of the key benefits of Far-UVC - not needing to change existing buildings or install expensive new systems - is lost.

Also, if I were going to put a UV lamp in an air duct, I wouldn't make it 222nm. IIRC other wavelengths (e.g. 254nm) are more effectively germicidal and are mainly bottlenecked by safety issues, which don't apply in this context.

Forecasting is hard. Maybe conventional LEDs wouldn’t be that easy, but there may be other approaches superior to excimer lamps we could use for pathogen control. Only one of them has to work, making this a disjunctive claim. For example, frequency-doubling solid-state lasers can kick blue light up to the UVC range. Also, quantum dots can be tuned very precisely, even without changing the component materials.

Can quantum dots emit light at intensities sufficient to disinfect a room? I always see them made and used on the nanoscale. And can a laser be wide enough so that you could make a “laser sheet” to disinfect any air that passes through it? Otherwise it seems like quantum dots are too small and lasers are too focused to work. But I am not an optics person.

Lasers can be widened with optics, like curved reflectors. Fiber could potentially distribute an intense source to multiple endpoints, although UVC would require the use of special materials. I’m not an optics specialist either. I don’t know of quantum dots in the UVC range, maybe it hasn’t been done yet. For visible wavelengths they can be pretty bright, so maybe? I don’t think these alternatives exist yet, but so many approaches seem potentially viable that I’m not sure it will take ten years.

Whilst the LEDs are not around the corner, I think the Kr-Cl excimer lamps might already be good enough.

When we wrote the original post on this, it was not clear how quickly covid was spreading through the air, but I think it is now clear that covid can hang around for a long time (on the order of minutes or hours rather than seconds) and still infect people.

It seems that a power density of 0.25W/m^2 would probably be enough to sterilize air in 1-2 minutes, meaning that a 5m x 8m room would need a 10W source. Assuming 2% efficiency that 10W source needs 500W electrical, which is certainly possible and in the days of incandescent lights you would have had a few 100W bulbs anyway.

EDIT: Having looked into this a bit more, it seems that right now the low efficiency of excimer lamps is not a binding constraint because the legally allowed far-UVC exposure is so low.

"TLV exposure limit for 222 nm (23 mJ cm^−2)"

23 mJ per cm^2 per day is just 0.002 W/m^2 , so you really don't need much power until you hit legal limitations.


I think your math on the energy is wrong / incomplete. The energy hitting a person isn't the same as the energy emitted from the source, and it falls off drastically with distance. So you can't just divide wattage at the source by room area.

Also, the area being reached depends significantly on room geometry - meaning that it's not easy to have a single lamp cover a room, especially the type of large, high occupancy room which is most likely to lead to large exposure events.

Is there a cost analysis of how effective UVC sterilization is compared to other methods such as increasing ventilation?

I don't know of a formal analysis, but informally it seems that it will be much easier to enforce far-UVC than ventilation. People don't like ventilation because it makes them cold or costs money, so they will tend to shut it off when they think they can get away with it.

Far-UVC is invisible and won't cost much compared to heating costs.

I also suspect that it's inherently more effective at the right doses because ventilation can't really stop transmission, only reduce the rate somewhat, so there will still be transmission and we may eventually find a pathogen that is contagious enough to still cause problems.

Ventilation costs can be reduced almost by an order of magnitude using enthalpy recovery systems, which have an upfront capital cost though. 

There are additional benefits that helps mitigate that though, such as more alert staff due to lower CO2 concentrations, lower VOC levels from off gassing materials, better humidity control to keep indoor humidity in the optimal zone, etc.

I'm not entirely convinced that UVC systems can be procured and installed cheap enough to be obviously better in 100% of hospitals. 

Enthalpy recovery systems really seem like they should be implemented everywhere. Given the benefits for both climate change prevention, pandemic preparedness, mold prevention, VOC health benefits, nightly sound pollution, and cognitive performance via lower CO2 in the room the case seems to me to be very strong. In the European case, reducing dependency on gas is also a nice side effect. 

https://www.energyvanguard.com/blog/will-balanced-ventilation-be-required-code/ suggests that Aspen already put it into their building code. 

I'm not entirely convinced that UVC systems can be procured and installed cheap enough to be obviously better in 100% of hospitals.

In a hospital setting, you likely care about any additional reduction in hospital-acquired infections. If there's clear evidence that you can reduce hospital-acquired infections this way, hospitals that don't adopt the technology can have a problem. You can sue them for malpractice if you get a hospital-acquired infection and they didn't implement all interventions that can reasonably be used to prevent the infection. 

Agree that indoor air systems are an obvious health benefit, but re: hospitals, it's only malpractice if using the systems is considered standard of care, which requires more than just evidence of efficacy. That said, I suspect that once the evidence is in, unless there is a new safety concern, it will start to get used in most hospitals very quickly, and become the standard of care.

Maybe in the U.S., but even then aren't there lots of hospitals infamous for their low quality of care, high infection rate, etc.? 

Why haven't they all disappeared yet if it were so easy to sue hospitals into adopting superior practices?

It's not. You need to show that they didn't meet standard of care, and if no-one does it, it's not actionable.

Also worth pointing out that multiple companies seem to already be selling far-UVC lamps in the 50W electrical power range (though some go all the way up to 500W)

I do suspect that this is a technology we can roll out right now (with some care about safety!)



Another alternative is to use a 440nm light source and a frequency doubling crystal https://phoseon.com/wp-content/uploads/2019/04/Stable-high-efficiency-low-cost-UV-C-laser-light-source-for-HPLC.pdf although the efficiency is questionable, there are also other variations based on frequency quadrupling https://opg.optica.org/oe/fulltext.cfm?uri=oe-29-26-42485&id=465709.

Is there existing work on how those Far-UVC Lights affect the human skin microbiome?

New to LessWrong?