Environmental aspects of solar cell etching

His Divine Shadow · Post by **His Divine Shadow** » 2005-09-11 12:37pm

Does anyone here know in detal how the etching method works and how it creates large amounts of toxic materials?

tharkûn · Post by **tharkûn** » 2005-09-11 06:06pm

Damn wrong board, typed in the wrong tab. Could a mod please delete the above mispost by me it is a complete mispost.[But of course.. - SirNitram]

Anyway the reason why solar is so polluting is that in order to make the silicon cells you have to have:
1. Make ultrapure defect free single crystal silicon.
2. Etch that silicon.

The first requires you make molten silicon which requires ubiquitious heat and in the process dumps crap previously bound in quartz into the atmosphere.

The second requires acids of varying strengths to be applied sequentially, much the same as making computer chips. Virtually all of these acids are polluting both during this process and further upstream when they were produced in petrochemical plants or whatever.

aerius · Post by **aerius** » 2005-09-11 11:52pm

Well, first they screen a photo-resist film onto the wafers to cover the parts that shouldn't be etched, then they expose the film and dunk the whole thing in chemical bath to dissolve the unexposed portions of the film. Then they rinse the wafer and dunk it in yet another tank of caustic chemicals to etch the exposed silicon away, then you rinse it off again and dunk it in yet another chemical bath to remove the exposed photo-resist film. Then it's rinsed again. The process may have to be repeated several times to get the final solar cell depending on how it's layed out. And in between there might be some vapour deposition steps or other fun stuff like that which also use fun nasty chemicals.

The stuff used to remove photo-resist is basically super industrial strength paint thinner which dissolves plastics & rubbers on contact, very nasty stuff. I don't know what they etch the silicon with, but my guess would be some kind of fluorine based acid since that's about the only thing that'll dissolve glass or silicon. And that is really bad stuff.

Dahak · Post by **Dahak** » 2005-09-12 05:06am

aerius wrote:The stuff used to remove photo-resist is basically super industrial strength paint thinner which dissolves plastics & rubbers on contact, very nasty stuff. I don't know what they etch the silicon with, but my guess would be some kind of fluorine based acid since that's about the only thing that'll dissolve glass or silicon. And that is really bad stuff.

That'll mostly be hydrofluoric acid (HF). As acids go, it's a relatively weak acid, mostly noted for its ability to etch and dissolve glass.
The bad part, for humans, is that it also is a very potent contact toxin (it won't really etch your skin all that much), but gets resorbed by the skin and then goes its merry way inside your body.

El Moose Monstero · Post by **El Moose Monstero** » 2005-09-12 07:25am

Dahak wrote:
aerius wrote:The stuff used to remove photo-resist is basically super industrial strength paint thinner which dissolves plastics & rubbers on contact, very nasty stuff. I don't know what they etch the silicon with, but my guess would be some kind of fluorine based acid since that's about the only thing that'll dissolve glass or silicon. And that is really bad stuff.
That'll mostly be hydrofluoric acid (HF). As acids go, it's a relatively weak acid, mostly noted for its ability to etch and dissolve glass.
The bad part, for humans, is that it also is a very potent contact toxin (it won't really etch your skin all that much), but gets resorbed by the skin and then goes its merry way inside your body.

Binds on to the bones as I recall and severely weakens them, when we were using it in core sample dissolutions I was told in no uncertain terms that if I got any significant quantity on me that I'd be carted off to casualty in a heart beat.

tharkûn · Post by **tharkûn** » 2005-09-12 07:29pm

Binds on to the bones as I recall and severely weakens them, when we were using it in core sample dissolutions I was told in no uncertain terms that if I got any significant quantity on me that I'd be carted off to casualty in a heart beat.

If you are lucky it goes to the bones. Hydrofluoric acid will also attack the kidneys, nerves, heart, liver, and a few other organs. Anything that either has high calcium concentrations or requires fine calcium ion concentration control is up for failure. Depending on the concentration you are dealing with you may survive long enough for treatment (and by many accounts wish you hadn't as administration of pain releivers is contraindicated), or you may go immediately to acute hypocalcemia and play dice to see if you die from a heart attack or CNS failure.

That'll mostly be hydrofluoric acid (HF). As acids go, it's a relatively weak acid, mostly noted for its ability to etch and dissolve glass.

Normally you use HF, HNO3, acetic acid, NH4F, KOH, EDA, and iodine for pure silicon. Depending on what process you are working with you may have anahydrous HF, dilute HF, or popular (and quite lethal) mixes of HF/NH4F and some other acid, base, or metal buffer/chelator depending.

Every single one of those chemicals can kill. Every single one requires pollution when making it and most pollute when you dispose of them.

When you move beyond the silicon and work with etched metal, as in some of the highly efficient solar designs you get to play with H3PO4 and various cyanides as well.

Oh and HF is only weak in terms of Arrhenius acidity (tendency to disassociate) which is really a fluke of screwy hydration (once you put HF in water it no longer is really HF, but the jury is out on exactly what it is). In a Bronsted-Lowry or Lewis measure (tendency to transfer protons or accept electrons in acid-base reactions respectively) anhydrous HF ranks as quite strong. These scales measures normally differ only trivially from Arrhenius measures, but for cases like HF, superacids (like HF/SbF5), and a few other oddballs the difference is ridiciously pronounced.

I don't know what they etch the silicon with, but my guess would be some kind of fluorine based acid since that's about the only thing that'll dissolve glass or silicon. And that is really bad stuff.

I think HF/NH4F with buffering is the most common etchant, but I am out of date. There are very few alternatives to HF analogues (all which are class 3 toxins or above), none that I know of are remotely close to cheap enough to use for mass production.

drachefly · Post by **drachefly** » 2005-09-13 10:53am

I wonder how dirty the new 'plastic' solar cells are.

As for these etching chemicals, I wonder how possible it would be to recycle them so you don't just dump them.
In an industrial setting, it would certainly be tempting to use it for a long time, even if you had to 'send it to the cleaners' for a while.

In any case, if your finest lateral features are no littler than tens of microns or so, you can use a solid, reusable shadow mask for the evaporations; you don't need to use chemical resists.

I don't know how much lateral fine feature work solar cells really need. Anyone care to fill us in?

tharkûn · Post by **tharkûn** » 2005-09-13 01:33pm

As for these etching chemicals, I wonder how possible it would be to recycle them so you don't just dump them.

How much are you willing to pay? In most cases the chemicals are reacted away in the etching process - for instance some of the hydrofluoric acid will go to silicon tetrafluoride . You can go back to HF, but you will need input energy (either direct or upstream with another chemical reagent). Such recycling could easily double the energy payback time for solar cells and likely would push the economic payback time out beyond their useful lifetime in most climates.

There is active research into reusing chemicals, maximizing their useable lifetime, etc. But unless I missed something major you are still using 100g or so of chemicals per g of etched Si.

In any case, if your finest lateral features are no littler than tens of microns or so, you can use a solid, reusable shadow mask for the evaporations; you don't need to use chemical resists.

As a general rule finer etching means better efficiency. You could build a solar plant that works off boiling water in a glass turbine, but the efficiency will be crap compared to photovoltiacs.

I don't know how much lateral fine feature work solar cells really need. Anyone care to fill us in?

Not off the top of my head, I might be able to look it up later.

drachefly · Post by **drachefly** » 2005-09-14 11:53am

The thing is, I think the important part is to make it very THIN, and not so important to make it laterally fine.

That's just my totally wacky guess, but it makes sense to me.

As for the chemicals: if you're going to use them fresh, then you're going to have to spend energy to make them in the first place. What is the comparative energy of recycling and producing? It might not be such a big difference. That said, if energy is more expensive, then it might be more expensive either way.

And of course if we're working on plastic, etching is done completely differently.

tharkûn · Post by **tharkûn** » 2005-09-14 07:07pm

I checked a few of the higher end schemes floating around and they are calling for fine features in the nm range. Nanometer scale defects (mostly metal impurities, but also silicon packing defects) are one of the major sources of inefficiency on high end cells.

The thing is, I think the important part is to make it very THIN, and not so important to make it laterally fine.

Thin is important, however the most important thing appears to be sequestering your exciton -> hole + electron transition. Metal impurities make for "lost" electrons and weird hole migration.

As for the chemicals: if you're going to use them fresh, then you're going to have to spend energy to make them in the first place. What is the comparative energy of recycling and producing? It might not be such a big difference. That said, if energy is more expensive, then it might be more expensive either way.

True, but you spend much less energy converting calcium fluoride (a mineral that can be mined) to hyrdoflouric acid than recovering potassium flouride from very dilute solution.

Last I checked every major semi-conductor firm either had in house or was sponsoring research on economicly recycling its chemicals. The big problems seem to be that you not only to drive many of the chemical reactions completely in reverse, but that you also have to purify out some ugly azeotropes and recover from dilute concentrations. As an added bonus wherever you do this work you incur all the normal costs of dealing with highly toxic chemicals.

If you can find a way to recycle these chemicals on the cheap you can sell it for millions or perhaps billions.

His Divine Shadow · Post by **His Divine Shadow** » 2005-12-14 02:53am

Sorry for the necromancy but i was discussing this on another forum and someone there claimed that all those acids can be easily neutralized with limestone. Any truth to that?

This is what he said:

All of them can be reacted with limestone to neutralize them. The fluoride compounds are by far the most toxic, but when reacted with limestone they form essentially insoluble/inert calcium fluoride, which is about as dangerous as quartz. Neutral salts of the others pose very little threat at all. Calcium nitrate and phosphate are used as fertilizers. Calcium acetate doesn't currently have any major uses to my knowledge, but neither is it dangerous; acetic acid and its salts are common food ingredients.

I agree that solar power cannot substitute for a large fraction of existing power sources, but that belief is pragmatic/economic in nature. The environmental impact of solar cells is not very high. They're just too expensive and sunlight is too diffuse for them to become major players without major technological advancements.

If he's right I am not going to argue, but I don't know any chemistry myself.

Stormin · Post by **Stormin** » 2005-12-14 03:06am

I don't know the exact science but I have to look at it this way, limestone is cheap and containment of toxic chemicals is expensive. If the stuff could just be neutralised with limestone or other non-exotics then it would have been done. At the least, if that worked it would create less of a storage/treatment/transport problem.

His Divine Shadow · Post by **His Divine Shadow** » 2005-12-14 03:21am

Well.... are the manufacturers using limestone or not? Is there a storage/treament problem?

arigo · Post by **arigo** » 2005-12-14 05:52am

Does anyone know where to find the figures with an estimate on the amount of pollution creating solar cells creates in relation to the total ouput of energy it is able to convert over its entire lifetime? Because I was curious about this a while ago and never found what I wanted in order to compare to other sources. It seems the fact that it's a one time thing for 30+ years of use allows for a relatively large amount of pollution to me made in creating it, and still come out ahead evironmentally over burning fossil fuels.

Zero · Post by **Zero** » 2005-12-14 08:14am

Could diamond replace silicon crystals in solar cells? I'm just asking primarily out of curiousity, I don't believe it would probably be economically viable. How would that effect the efficiency?

Shroom Man 777 · Post by **Shroom Man 777** » 2005-12-14 08:45am

I know shit, but I've heard that companies only artificially inflate diamond prices. They're actually as common as flies on shit...

GrandMasterTerwynn · Post by **GrandMasterTerwynn** » 2005-12-14 10:22am

His Divine Shadow wrote:Sorry for the necromancy but i was discussing this on another forum and someone there claimed that all those acids can be easily neutralized with limestone. Any truth to that?

This is what he said:
All of them can be reacted with limestone to neutralize them. The fluoride compounds are by far the most toxic, but when reacted with limestone they form essentially insoluble/inert calcium fluoride, which is about as dangerous as quartz. Neutral salts of the others pose very little threat at all. Calcium nitrate and phosphate are used as fertilizers. Calcium acetate doesn't currently have any major uses to my knowledge, but neither is it dangerous; acetic acid and its salts are common food ingredients.

I agree that solar power cannot substitute for a large fraction of existing power sources, but that belief is pragmatic/economic in nature. The environmental impact of solar cells is not very high. They're just too expensive and sunlight is too diffuse for them to become major players without major technological advancements.
If he's right I am not going to argue, but I don't know any chemistry myself.

Limestone is simply calcium carbonate (CaCO3), which is the same active ingredient that's found in many antacids. The chemical reaction is found below:

(2)HF + CaCO3 = CaF2 + H2O + CO2

Immediately, you can see at least one environmental problem with using limestone to neutralize the surplus acid. While two of the byproducts are calcium fluoride and water, the last byproduct is carbon dioxide. Which will be added to the carbon dioxide produced burning the coal or natural gas required to generate the energy needed to make the solar cells. Carbon dioxide, incidentally, is among the chief culprits in global warming, as non-toxic as it is.

Really, this just proves that there's no such thing as a free lunch. You get to choose between polluting the environment with toxic chemicals, or polluting the environment with greenhouse gasses.

drachefly · Post by **drachefly** » 2005-12-14 12:03pm

Well, we're looking at something like 3 kilograms of CO2 for every 2 kilos of HF... Considering that we're going to use less mass of HF than the mass of the solar cell, and the solar cell will last for N years, where N is much larger than 1... and so it will produce over its lifetime much much more energy than the same mass of fossil fuels, which would thus release much much more CO2

A five cent lunch is, I suppose, technically not free.

And I'm still wondering what the ecological impact of these photovoltaic plastics will end up like...

tharkûn · Post by **tharkûn** » 2005-12-14 08:39pm

All of them can be reacted with limestone to neutralize them

Rather stupid thing to say given that only acids need be neutralized and that can be done far more cheaply with any simple base.

The fluoride compounds are by far the most toxic, but when reacted with limestone they form essentially insoluble/inert calcium fluoride, which is about as dangerous as quartz.

Considering that quartz of similar dimensions is a carcinogenic according to IARC that is not particularly reassuring.

In any event calcium flouride has an acute LD-50 of around 1-4 g/kg (depending on uptake mechanism), which while not earth shatterring, is far from harmless. Chronic exposure leads to osteosclerosis and fluorosis.

Yeah you can react away HF with limestone, of course given that limestone has piss for fine consistency you risk having trace contaminants generate all manner of fun crap and when all is said and done you are still prohibited from discharging (at least in any country that gives a damn about public health). Then of course there is the problem with the waste which isn't simple inorganic compounds that react readily with calcium carbonate. Of course I'm being generous and ingoring all kinetic/process concerns, but anyone who thinks dumping limestone into silicon etching waste is sufficient knows nothing.

Sorry but if you could just dump limestone into the waste semiconductor fabrication wouldn't be the most polluting (on a per mass basis) industry on the planet.

The environmental impact of solar cells is not very high.

BS. According to a full lifecycle analysis the bad PV cells emit more carbon dioxide equivalents than some natural gas plants (and the worst ones are almost into friggen coal territory).

The following are lifecycle figures from the International Energy Agency (coal will only refer to modern western coal power generation and natural gas will be similarly restricted):

CO2 equivalents g/kWh:
nuclear - 2-59
wind - 7-124
natural gas - 389-511
coal - 790-1182
PV - 13-731

SO2 mg/kWh:
nuclear - 3-50
wind - 21-87
natural gas - 4-15000+ *
coal - 700-32321+
PV - 24-490

NOx mg/kWh:
nuclear - 2-100
wind - 14-50
natural gas - 13-1500
coal - 700-5273
PV - 16-340

Non-methane volatile organic vapor mg/kWh:
nuclear - ~0
wind - ~0
natural gas - 72-164
coal - 18-29
PV - 70

Particulent matter mg/kWh:
nuclear - 2
wind - 5 - 35
natural gas - 1-10
coal - 30-663+
PV - 12-190

*Natural gas reserves are contaminated with as much as 1% hydrogen sulphide, under normal circumstances virtually all sulphur is sequestered prior to combustion, however on occassion some gas is burnt directly and that is the only time the high end is seen.

Oh look some PVs contribute more greenhouse gas than some friggen natural gas plants. Most PVs are decidedly inferior to most nuclear and wind plants across the board. Of and in the emissions categories with the highest health risks to humans? PVs might beat out coal. Color me unimpressed.

Now if you exclude old technology and fabrication methods PVs get better, so too does everything else. If by "not very high" the dumbass means in comparison to fossil fuel he may be correct, if he means compared to nuclear, wind, and a few other clean power sources, he's talking out his ass.

hey're just too expensive and sunlight is too diffuse for them to become major players without major technological advancements.

A major part of the expense is reducing their upstream emissions. Working with silicon is a major bitch from purification to etching to cleanup.

Does anyone know where to find the figures with an estimate on the amount of pollution creating solar cells creates in relation to the total ouput of energy it is able to convert over its entire lifetime? Because I was curious about this a while ago and never found what I wanted in order to compare to other sources. It seems the fact that it's a one time thing for 30+ years of use allows for a relatively large amount of pollution to me made in creating it, and still come out ahead evironmentally over burning fossil fuels.

Phylipsen and Alsema out of Utrect have done several lifecycle analysis, but most of their work is with new technology in good operating conditions (i.e. no snow, cleaning, etc.).

Could diamond replace silicon crystals in solar cells?

No. The band gap in diamond is far too wide for anything except the most energetic photons to span. Even if the EM spectrum suddenly shifted down to those wavelengths, diamond doesn't conduct and you'd have to do all manner of crap to get electricity out.

How would that effect the efficiency?

Orders of orders of magnitude decrease.

Well, we're looking at something like 3 kilograms of CO2 for every 2 kilos of HF... Considering that we're going to use less mass of HF than the mass of the solar cell

You may well use more HF than the mass of the cell, though such methods should be falling by the wayside. In any event in order to get sufficiently pure calcium carbonate, so you don't have trace reactions liberating fluoride and other fun, you need to expend much energy. I have already posted comparable carbon dioxide emissions. PVs suck ass compared to nuclear.

A five cent lunch is, I suppose, technically not free.

Compared to a less than a cent nuclear lunch it really means piss. Yes PV's under optimistic conditions will emit less carbon dioxide over their lifetime than fossil fuels. However emissions now are not the same as emissions later, there must be discounting to account for compounding effects and progress. Likewise one must remember that much of Europe, North America, China, and Japan are not optimal areas for solar cells. Snow removal adds pollution. So too does cleaning off acumulated dust.

Sure PVs beat coal on pollution, what doesn't?

drachefly · Post by **drachefly** » 2005-12-15 01:31pm

You may well use more HF than the mass of the cell, though such methods should be falling by the wayside.

Of course I wasn't advocating using inefficient methods... :roll:

PVs suck ass compared to nuclear.

I don't disagree that Nuclear is less carbon-emitting.
Let's look at those ranges...

CO2 equivalents g/kWh:
nuclear - 2-59
wind - 7-124
natural gas - 389-511
coal - 790-1182
PV - 13-731

yes, that's a range of a factor of around 60 for PV. Obviously, we should take the cleaner methods rather than the dirty ones? In which case we get

CO2 equivalents g/kWh:
nuclear - 2
wind - 7
natural gas - 389
coal - 790
PV - 13

and then everything BUT fossil fuels is basically negligible.

tharkûn · Post by **tharkûn** » 2005-12-15 03:44pm

Of course I wasn't advocating using inefficient methods...

Do you want it cheap or do you want it efficient? TANSTAAFL still applies, if you use certain techniques to lower HF consumption you have to pay more for it. In the simplest case you can get away with coarser features, which impairs efficiency and increases energy payback time.

yes, that's a range of a factor of around 60 for PV. Obviously, we should take the cleaner methods rather than the dirty ones?

A good portion of the range derives from where the panels are operating. Not all panels operate in the Mojave desert and those which operate in less friendly climes pollute far more per kWh. Argueably the high end of the range is more applicable for PVs as all their pollution is front loaded and thus gets no discounting like fossil fuels (note to some degree the same applies to nuclear and win).

and then everything BUT fossil fuels is basically negligible.

Negligable? No. That amount of CO2 production is still huge, frontloading also means you are going to have an initial spike which will have disproportionate ecological impact. It will handily beat coal, but natural gas is still a contender.

drachefly · Post by **drachefly** » 2005-12-16 11:02am

yes, that's a range of a factor of around 60 for PV. Obviously, we should take the cleaner methods rather than the dirty ones?
A good portion of the range derives from where the panels are operating. ...

Fair enough. I agree that PV should not be used as the main grid source except in deserts and such places.

and then everything BUT fossil fuels is basically negligible.
Negligable? No. That amount of CO2 production is still huge, frontloading also means you are going to have an initial spike which will have disproportionate ecological impact. It will handily beat coal, but natural gas is still a contender.

The frontloading argument does make sense, but if you divide human CO2 production by a factor of 6 or so I think the ecological impact WOULD be pretty minor.

tharkûn · Post by **tharkûn** » 2005-12-16 02:22pm

Fair enough. I agree that PV should not be used as the main grid source except in deserts and such places.

Even in deserts you have problems. Windborn grit plays merry hell with efficiency, lifespan, and operating costs. There are places where solar isn't that bad, but very few are remotely comparable to nuclear.

The frontloading argument does make sense, but if you divide human CO2 production by a factor of 6 or so I think the ecological impact WOULD be pretty minor.

The health effects of a spike in NMVOV, Sulfur dioxide, NOx, and particulent emissions are not to be overlooked.

Yeah if we killed coal to bring in solar we wouldn't be doing too terrible on the tradeoff. But it is doubtful that it would go down that way, coal is so cheap in comparison to virtually everything that it may well be kept while natural gas is 'cut' for PV.

Nephtys · Post by **Nephtys** » 2005-12-16 02:32pm

drachefly wrote:
yes, that's a range of a factor of around 60 for PV. Obviously, we should take the cleaner methods rather than the dirty ones?
A good portion of the range derives from where the panels are operating. ...
Fair enough. I agree that PV should not be used as the main grid source except in deserts and such places.

and then everything BUT fossil fuels is basically negligible.
Negligable? No. That amount of CO2 production is still huge, frontloading also means you are going to have an initial spike which will have disproportionate ecological impact. It will handily beat coal, but natural gas is still a contender.
The frontloading argument does make sense, but if you divide human CO2 production by a factor of 6 or so I think the ecological impact WOULD be pretty minor.

Why would the Desert be the only place? We shouldn't use Solar at all and go full Nuclear, really. Wouldn't erosion and harsher conditions in places like Nevada and Arizona require more frequent replacement of said solar cells anyway?

Zero · Post by **Zero** » 2005-12-18 02:57am

http://exploration.vanderbilt.edu/news/news_diamond.htm

Found my own answer to my question about the usage of diamond in solar cells. Does anyone think such a thing would be viable in terms of economics and whatnot?