5 ms·
Oh dear. After reading the notes in the study itself about the scale this was done at, scaling up might be an issue: > A typical H₂O₂ synthesis reaction was ca
by jcranmer 5y ago
Oh dear. After reading the notes in the study itself about the scale this was done at, scaling up might be an issue:
> A typical H₂O₂ synthesis reaction was carried out using 120 mg of 1% AuPd/TiO₂ [...] The reactor system was then pressurized, typically to 10 bar. The reactor was then cooled by a water bath to 2 °C. When the reactor had reached the required pressure and the flow through the system had stabilized, the solvent (H²O, HPLC grade), typically at a flow of 0.2 ml min¯¹, was introduced into the system.
Oh hey, 0.2 mL/min of water. In Virginia, the lowest flow cutoff for waterworks is 0.5 MGD (or 5000 people), the border between Class 3 and Class 4. That flow rate is 6.5 million times that done in the tests here. The large water treatment plant I worked at is another 400 times that, and there are plants that would be another 5 or 6 times that. So the necessary scaling factor is going to be measured in the billions.
So what about those reagants? This uses about 0.6 mg of gold to process 0.2 mL/min of water (if I'm reading the notation of 1% AuPd/TiO₂ correctly). Scale that to our small plant, and that's about $225,000 of gold and $350,000 of palladium. Go to the scale I worked at, and that's $190 million of catalyst in terms of raw metal worth alone. That's comparable to the cost of a regular treatment plant of that scale--just on the catalyst. This assumes that there's a neat linear scaling, which I doubt is going to be the case (square-cube law!).
- wolverine876 5y agoThanks for the insights, though let's remember that research is just a first step. This isn't a product being announced for the market. If we looked at the original research behind a lot of mature technology, we'd find the state of the technology to be similarly far from from production. Now others can try to reproduce it, refine it, find other methods of producing the same reactive oxygen species, find substitutes that are more effective or cheaper, etc. It doesn't have to be panacea either. Maybe it will fill a niche. It could save some lives.
- 55555 5y ago> This isn't a product being announced for the market. The press release makes bold implications tbh
- aidenn0 5y agoIt wouldn't be a university press release if it didn't make a specialized laboratory experiment sound like something that is about to revolutionize life as we know it.
- dmix 5y agoMostly context on why researching such a thing would be valuable.
- Cthulhu_ 5y agoThey always do; it's a marketing statement.
- GrumpyNl 5y agoThis works better for the funding round.
- nico_h 5y agoMaybe they can use metallic foam to increase the surface area of the catalyst and increase the flow rate without scaling the noble metal usage at the same rate as the flow!
- mmazing 5y agoIf its effectiveness is as good as they say, seems like you could just use 1,000,000x less catalyst and end up with something as effective as current technology. $190 million -> 190 bucks ;) Seriously though, couldn't they just use far less catalyst and not have to deal with the prices you've calculated?
- rtpg 5y agowould it still work though? Like if the mechanism is a relation between surface area and the water, you might not have a choice to just use less material? Unless your "strategy" is to, like.... just have a tiny flow of super pure water and then mixing it into untreated water. (Your strategy might totally work of course, just might not be a given)
- honie 5y ago> Seriously though, couldn't they just use far less catalyst and not have to deal with the prices you've calculated? The flow rate is likely positively correlated with catalyst loading (in the part that we care about) and the calculation is already based on the flow rate of 0.2 ml min¯¹ they reported in the paper. The reported flow rate is also likely already optimised for the Nature Communications paper.
- neltnerb 5y agoIt is very unlikely that they could have even explored the full parameter space if they wanted to, it's such a huge undertaking. It's probably optimized under some constraints, but there's a huge number of things involved -- support material, surface area, catalyst metal, alloying, promoters, pretreatment, temperature, pressure, flow rate -- my experience has been that you can take any catalyst and make it work a lot better by just picking an understudied parameter and optimizing it. My guess is that you could drop the catalyst noble metal content by 10-100x or even substitute an alternative catalyst and find a combination of conditions and catalyst that works. It's not a sure thing, but it's not unlikely. Sometimes it's as simple as swapping one support for another and the performance goes up 100x, it's just a lot of tests to run that don't even apply to both lab and commercial scales anyway (so a lot of the tests are never practical to run anyway). I think it'd be a bigger deal if it didn't require pressurization to 10 bar personally, pumps are expensive to operate and catalysts are a one-time-cost (if you're lucky).
- yxhuvud 5y agoI wonder if it could replace clorinated water in swimming pools though.
- KSteffensen 5y agoThis can already be done using UV light possibly combined with ozone treatment. I'm not sure what the hydrogen peroxide method brings to the table compared with UV/ozone (besides a university press release). [0]: https://www.espwaterproducts.com/ultraviolet-uv-systems/ https://www.espwaterproducts.com/ultraviolet-uv-systems/ [1]: https://www.socotec.co.uk/media/faqs/faqs-ozone-water-treatment https://www.socotec.co.uk/media/faqs/faqs-ozone-water-treatm...
- a_conservative 5y agoChlorine in the water provides disinfection all throughout the pool, this is very desirable when you have a pool. Imagine you see someone's bandage floating next to you in the water! Do you want to wait for the water to be pulled through the filtration system to get cleaned up? It typically takes a pool pump many hours to bring the full volume of the pool through the filtration system. By having the disinfection agent spread through the pool, any bacteria or other contaminant will be acted on immediately. Algae is another big problem in many pools. Oftentimes there is some corner of the pool (or ladder, etc) that will harbor a little bit of algae that is waiting to take over the rest of the pool. Having chlorine dispersed through the water is very effective at preventing algae. Also, shocking the pool water with high levels of chlorine is oftentimes needed to keep the algae under control. The common approach for making a pool need less daily maintenance is to get a salt water chlorine generator. A small amount of chlorine is generated by running an electrical current through electrodes. The electrodes need to be replaced every so often. Liquid chlorine is generally considered to be cheaper, but requires a lot of diligence because the chlorine dissipates pretty quickly. I've seen hot tubs that use ozone for disinfection but not pools. I'm not sure what makes it a reasonable choice for hot tubs but not pools. Ozone would disperse through the water like chlorine, but it dissipates even more quickly. One more interesting tidbit I picked up alone the way: the chlorine smell in a pool is not from the free chlorine. Free chlorine is what's available for disinfection. After a free chlorine molecule bonds to it's target, it's called combined chlorine. Combined chlorine is what is responsible for the nasty smell in pools. The solution for combined chlorine is to add dd more free chlorine!
- azalemeth 5y agoWhat does MGD mean, and what is it in SI units? Million gallons per day? US gal or Imperial ones? (I'm not being facetious -- I genuinely don't know and it's my understanding that the US mix parts of the old British Imperial system, e.g. BTU, together with their own)
- yesenadam 5y agoA minute's googling revealed that, yes, it's million gallons/day, and it seems both Imperial and US gallons are commonly used, so depends on author's nationality which is meant, I suppose.
- jcranmer 5y agoYes, MGD means million gallons / day, and this is US gallons not Imperial (you should have been able to guess that from my reference to Virginia's standards--Virginia being a US state).
- azalemeth 5y agoThank you. (So about 0.0526 m^3/s.)
- aardvarkr 5y ago“When the reactor had reached the required pressure and the flow through the system had stabilized, the solvent (H²O, HPLC grade), typically at a flow of 0.2 ml min¯¹, was introduced into the system.“ You’re making such a big deal of the 0.2 ml/min that you didn’t stop to read that it’s talking about the amount of solvent… you’re just stirring up drama for the sake of drama
- kragen 5y ago(I should preface this by saying that, unlike you, I have no real water disinfection expertise, so I may be overlooking significant fundamentals in what follows.) Those costs sound reasonable for niche uses (though I do think they would prevent it from "revolutionizing water disinfection technologies" as they claim in the Conclusions), and being able to operate in practical terms at a very small scale seems to me like an advantage rather than a disadvantage. It clearly is more expensive than conventional water treatment and would have difficulty scaling to replace it for reasons of material availability. Other alternative very-small-scale disinfection approaches are probably cheaper than their process in its current form; they mention ozonation, UV irradiation with germicidal lamps, photocatalytic disinfection, and the Fenton process in the paper and supplement. I don't see how the square-cube law plays into it; the reason they're supporting the catalyst on rutile (anatase?) instead of just using solid bars of metal alloy is precisely to ensure neat linear scaling, and the reason it's 1% metal instead of 0.1% or 0.001% is that the rutile is in the form of 1–100 nm particles. (See p. 8/12 in the article and Fig. 4 on p. 7/12). The only square-cube thing that occurs to me is that a large ice bath requires much less ice input than a small ice bath, but that favors scaling the process up, not down. A couple of other points: ⓪ Your calculations seem to be correct. ① People's drinking water needs are lower than you suggest by a factor of about 65, although conventional surface-water treatment plants cannot take advantage of this. ② You're not taking into account the supply-chain issues that plague smaller-scale treatment facilities. ③ The cost of consumables that this process would eliminate would still be lower than the cost of the catalyst. ④ There are plausible ways the process might be improved that could make it economic. Thanks to mkr-hn I found the paper at https://www.researchgate.net/publication/352882305_A_residue-free_approach_to_water_disinfection_using_catalytic_in_situ_generation_of_reactive_oxygen_species; https://www.researchgate.net/publication/352882305_A_residue... the canonical paper link (unforgivably missing from the original press release) is https://www.nature.com/articles/s41929-021-00642-w https://www.nature.com/articles/s41929-021-00642-w, and the supplementary material is at https://static-content.springer.com/esm/art%3A10.1038%2Fs41929-021-00642-w/MediaObjects/41929_2021_642_MOESM1_ESM.pdf https://static-content.springer.com/esm/art%3A10.1038%2Fs419.... — ⁂ — ⓪ Scaling calculations. Here are the calculations in more detail, since I misread yours badly at least twice. You say 0.5 MGD is 6.5 million times higher than the 0.2 mℓ/min (3.3 μℓ/s) in the experiment. 0.5 MGD is 21.9 ℓ/s, which is 6.6 million times higher than 3.3 μℓ/s. I think you are reading it correctly; they do say their catalyst is 0.5% Au and 0.5% Pd on a TiO₂ support: 0.6 mg of gold and 0.6 mg of palladium to process 0.2 mℓ/min of water. The supplementary material has a plot of different catalyst mixes they tried (Sup. Fig. 14, p. 12, 13/32). If we normalize that amount of catalyst metal to SI units, that's 180 gram seconds per liter (g·s/ℓ) each of gold and palladium. Your 0.5 MGD (22 ℓ/s) for 5000 people is 4.4 mℓ/s per person, which works out to 790 mg per person each of gold and palladium. Palladium at US$2800 per troy ounce (https://www.kitco.com/charts/livepalladium.html https://www.kitco.com/charts/livepalladium.html — that is per troy ounce, isn't it?) is US$90/g. Gold at US$1800 per troy ounce is US$58/g. Multiplying it out, that's US$71 of palladium per person and US$46 of gold per person, for a total of US$117 of catalyst metals per person. This seems likely to be the dominant cost of the whole shebang at anything larger than laboratory scale; the actual preparation of the catalyst with these metals by the process they reported in the paper might be even more expensive, but presumably cheaper methods are possible. The 5000-person 0.5 mgd plant you cite as an example would need 3.9 kg of gold (US$230k) and of palladium (US$350k) for its catalysts, a total of US$580k, which is the number you gave for "our small plant". 790 mg of gold per person, times 7.7 billion people, would be about 6100 tonnes, about 3% of all the 200 000 tonnes of gold that has been mined so far. However, the corresponding 6100 tonnes of palladium would vastly exceed the above-ground stocks of palladium, estimated at 4.5 "moz", which I think means "million troy ounces", or 140 tonnes. However, they also tried 1% gold without any palladium, and although this produced lower H₂O₂ concentrations, they were still high enough to be somewhat effective (≈80 ppm rather than 220 ppm, resulting in a 1.6 log₁₀ reduction rather than the 8.1 they were so satisfied with). So in the face of resource limits you could trade off a larger amount of gold and a longer residence time against scarce palladium. ① Potable water needs are 5.7 ℓ/day/person, not 380. Your figure of 4.4 mℓ/s/person is 380 liters per day per person (100 gallons per day per person), but Burning Man recommends 1.5 gallons per day per person (5.7 liters/day/person, 0.066 mℓ/s/person), which includes water for showers, for cooking, and for drinking in a very dry environment with extensive physical exertion, though not for bidets. My experience at Burning Man is that you can get by on less. That's 67 times less water than your figure. Maybe your waterworks is supplying not only drinking water but also toilet-flushing water and lawn-irrigating water? Those don't normally need antibacterial treatment. Even the 0.2 mℓ/min benchtop catalyst they used in the paper (containing 5.4¢ of palladium and 3.5¢ of gold) would supply 0.29 liters per day; it would only need to be scaled up by a factor of 20 (to 12 mg gold (US$0.69) 12 mg palladium (US$1.10), 2.4 g rutile) to supply the requisite 5.7 liters per day per person. (continued below)