Assaying silver content of a mixed metal/semiconductor powder

[SunSilver1993]

Hi All,

I have a silicon powder from crushed up solar panels that I am trying to determine the silver content of. The powder contains Silicon, Copper, Aluminium Iron and Tin. We already have an XRF analysis showing that it contains 0.4wt% silver.

I would like to confirm this result via chemical assay. This would also have the added benefit of having a way to test silver in house without XRF.

On that matter I've already tried a couple of things and am looking for some advice. In short, I've tried the silver chloride method a couple of times and tried once to plate silver out onto copper but need advice on both. I am also wondering if there is a relatively simple titration I could perform to determine silver content of the solution without having to isolate the powder (something like reverse argentrometry?)

Silver Chloride method

I've added silicon powder to excess 30% Nitric acid and heated under stirring until evolution of NO2 stopped. I then filtered the solution and rinsed the powder to collect a clear blue solution to which I added excess NaCl. This caused a colour change to lime green and the formation of a very fine precipitate.

Attempts to filter this solution and collect the insolubles resulted in no collected powder and a milky solution in the buchner flask below. It appears that the precipitate is so fine that it can't be filtered?

Is there anything I'm missing here? Should I be adjusting the pH prior to adding NaCl?

Silver Plating

I followed the same procedure to collect the clear blue solution above. I then attempted to neutralise this with a concentrated NaOH solution. The reason I did this is that I wanted to take the pH to a point where the nitric didn't react with the copper so the plating would be obvious to observe and easier to collect for weighing. Addition of NaOH caused the temporary formation of a black precipitate which would go back into solution. I kept adding NaOH at this point hoping to stop the formation of the precipitate but pushed it too far and a permanent precipitate formed . At this point I was discouraged so I discarded the solution.

So for this experiment I have the questions;
- What pH should I neutralise to ? I don't have an easy way of testing pH so would rather just add a fixed amount of NaOH or bicarb
- Should I use sodium bicarbonate instead?
- Should I just keep adding copper that will dissolve until silver starts to plate out on it? Seems quite wasteful and generates a lot of soluble copper which obviously isn't ideal from a disposal point of view.
- What is that black precipitate?
- Will this method be an accurate way to determine silver content once I've collected the silver?


Titration

Looking for any suggestions here, anyone have any experience with this? The more I think about it the better of an idea this seems as there will be no losses from filtering and likely I can use a much smaller sample.


Thank you so much in advance for any help

 

[Martijn]

With titration, you could add and weigh HCL or NaCl in known concentration to make AgCl and calculate the silver percentage.

for example: how i tritate H2SO4 with NaOH:
it takes 2 moles of NaOH to neutralize 1 mole of H2SO4, i take a beaker with 1Molar solution of Lye and weigh it with the pipette, before and after al is neutralized (measured with tiny pH paper with a pincet to avoid too much loss of liquid during tritation) so i calculate how much H2SO4 is neutralized by stoichiometry.
for silver chloride it can be another ratio, but that can be figured out by stoichiometry.
If you have the amount of grams of silver you have converted to AgCl, you can calculate the percentage pretty easy if you weighed the sample you dissolved.
How much powder do you have to process?

Attempts to filter this solution and collect the insolubles resulted in no collected powder and a milky solution in the buchner flask below. It appears that the precipitate is so fine that it can't be filtered?

If the AgCl is that fine, there wil not be enough silver in the powder to be cost effective, i'm afraid. Let it settle, and siphon off the clear liquid, rinse the rest with boiling water a couple times, it tends to clump and settle better after that.
But if there are no clumps of silver chloride, i would not bother separating it. silver Chloride has a lot of volume compared to it's silver amount.

If you keep using the same filter over and over, in time it will catch the silver chloride. it gets clogged by the AgCl which helps it filter slower, but better.

 

[Martijn]

As far as pH goes: that's irrelevant. Add HCl until no more AgCl is formed. This can take time and you can do a rough determination first to see what is enough to get it all out fast, and the repeat it with a new batch but slower addittions at the end.

in fact, you will need to tritate your HCl to know how much moles are in that exactly first or use NaCl instead.

 

[SunSilver1993]

Thanks Martijn,

I'm aware it will be uneconomical to recover using Nitric but want a way to assay samples so that I can test other cheaper methods to get the silver out. I have lots of powder to play with but could have much more in the future so I need the assay methodology.

With your titration suggestion there's no clear cut off point unfortunately as it will just get cloudier and cloudier but there isn't a color change or something like that to know when to stop. I'm now looking into potentiometric titration as an accurate method

 

[kurtak]

Leaching this type of material (with acid) is most certainly NOT the way to go - NOT even for testing

you would be MUCH better served doing a fire assay on this

IF (the VERY BIG IF) fire assay shows this is worth going after for the silver then smelting would be the way to go

Kurt

 

[FrugalRefiner]

Volhard Titration

How it Works - Potassium thiocyanate reacts with silver ions to create an insoluble precipitate of silver thiocyanate. It also reacts with the ferric (Fe3+) ions of the indicator solution to create a blood/wine red solution. Silver ions have a greater affinity for the thiocyanate than ferric ions, so the thiocyanate will react preferentially with the silver ions until there are no more of them in solution. When there are no more silver ions in solution, the red color will remain, indicating the endpoint of the titration.

As mentioned above, two reagents are needed for the titration, the Indicator, and the Titrant. They are prepared as follows.

Ferric Ammonium Sulfate Indicator - The indicator solution is easy to prepare. Only a small amount (approximately 1 ml.) of the indicator solution is needed for each test, so any convenient amount can be made depending on your needs. Dissolve about 1 gram of ferric ammonium sulfate into about 10 ml. of distilled water. The amounts are not critical. It is only important that there be sufficient ferric (Fe3+) ions in the solution to combine with excess thiocyanate ions at the endpoint of the titration. The ferric ions form a dark red complex with the thiocyanate ions to indicate the endpoint has been reached. For best results, the indicator solution should be prepared the same way each time, and the amount used should be consistent.

Potassium Thiocyanate Titrant - More care is required when preparing the titrant, as the precise concentration is important to achieve accurate results. The solution should be prepared in a volumetric flask to ensure accuracy. The molecular weight of potassium thiocyanate is 97.181 grams per mole. We want a 0.1 molar solution, so add 9.718 grams of potassium thiocyanate to a 1,000 ml. volumetric flask. Add about 500 ml. of distilled water to the flask and swirl the solution in the flask until all the thiocyanate is dissolved. When it is all dissolved and at room temperature, add distilled water to bring the solution up to the mark on the neck of the flask. This creates a 0.1 molar solution of potassium thiocyanate.

Note: Potassium thiocyanate is hygroscopic to the point of being deliquescent. Be sure to keep the bottle tightly capped and work quickly when weighing a sample as the sample will gain weight the longer it is exposed to any moisture in the air.

Be meticulous when preparing this solution. Any error in this step will be magnified when the solution is used. My inexpensive scales can drift during use, so I weigh the thiocyanate multiple times to achieve an accurate weight. Always allow all solution to return to the temperature specified on your volumetric flask when topping up the volume to the line on the flask. Be sure to thoroughly mix the final solution to ensure the concentration is consistent throughout.

Performing The Titration on a Solution - Precision is important in titrations. Pipettes and burettes must be clean and dry. Using a quality measuring pipette, add exactly 1 ml. of the solution to be tested to a suitable, clean container. Erlenmeyer flasks are ideal for this work, as they allow you to swirl the solution in the flask to ensure thorough mixing while minimizing the risk of splashing any of the solution out of the flask. A 250 ml. flask is a good size. A beaker can be used, but extra care must be exercised to prevent any loss of solution which would render the results inaccurate. Add about 1 ml. of the indicator solution, and about 100 ml. of distilled water. These quantities are not critical as they do not affect the accuracy of the results, but be consistent from one test to the next.

Fill a burette with the titrant and adjust the level to the zero mark. Begin adding the titrant from the burette to the flask containing the sample, swirling the flask continuously to mix the solutions. At first, you’ll see a red reaction where the titrant contacts the sample, but the color disappears quickly and a precipitate forms as the solutions mix together. As you approach the endpoint, the color remains longer and the additions of titrant should be slowed to avoid overshooting the endpoint. At the end, you should only add the titrant one drop (or less) at a time, thoroughly mixing between each addition. The endpoint is reached when the last addition of titrant causes the solution to remain a slight pink/orange/purple/red color (the color will vary depending on the colors of any other ions in the solution). Read the amount of titrant that has been used from the burette and use it in the following calculations.

Calculations - The titrant contains 0.1 moles of potassium thiocyanate per liter, or .0001 mole per ml. Multiply the number of ml. of titrant used by .0001 to determine the moles of potassium thiocyanate used. The thiocyanate ions combine with the silver ions in a 1:1 relationship as shown in the formula below.

Ag+ (aq) + SCN- (aq) => AgSCN (s)

So the moles of thiocyanate used equals the moles of silver in the solution.

If you used 10 ml. of titrant, you would have used 10 * .0001 = .001 moles thiocyanate = .001 moles Ag. The atomic weight of silver is 107.8682 grams per mole. Multiply the number of moles of silver by the atomic weight to determine the number of grams of silver in the sample, .001 * 107.8682 = 0.1079 grams of silver. Finally, since our sample was 1 ml., multiply this result by 1,000 to find the number of grams of silver per liter of solution, 0.1079 * 1,000 = 107.9 grams of silver per liter.

The full calculation is ml. of titrant used * .0001 * 107.8682 * 1,000. Let’s simplify that. The multiplication by .0001, the atomic weight of silver at 107.8682, and the multiplication by 1,000 are all constants. .0001
* 107.8682 * 1,000 = 10.7868. So simply multiply the number of ml. of titrant used by 10.7868 and you have the number of grams of silver per liter of solution, 10 * 10.7868 = 107.868 grams per liter.

Accuracy - While we can calculate the result in the example above to a thousandth of a gram, we can’t actually achieve that accuracy without some very expensive equipment. The results of any calculation are only as accurate as the least accurate part of the calculation. If our sample is 1 ml ± .01 ml, then the accuracy is to within 1%. If our sample is within 1%, we can say we have somewhere around 107 grams give or take 1.07 grams (1%). If we could measure our sample to within 0.1%, then the accuracy of the calculation would be 107.8 grams plus or minus .107 grams.

Assuming every part of our test is within 1% accuracy, can we assume our result is within 1% as well? Unfortunately, the answer is No. Error is cumulative. Let’s say that when we mixed our titrant, we were within 1%, our sample was within 1%, and the measurement of the titrant used was within 1%. Our total accuracy could be off by as much as 3%. Fortunately, with even a .01 gram scale and a 1,000 ml. volumetric flask, we should be able to get our titrant within a couple of tenths of a percent, so the least accurate factors will probably be the sample size and the measurement of the volume of the titrant, and if we can get both of those to within 1%, the results are accurate enough to balance the electrolyte of a silver cell.

Performing The Titration on a Solid - A solid silver alloy, such as a bar, can be tested for purity with this method. The size of your sample depends on the accuracy of your scales as described above vs. the amount of titrant that will be needed vs. the accuracy you require. If your scale is only accurate to .01 grams, it would require a 1 gram sample to be within 1%. Add the error of the burette described above and the error grows to 2%. A bit more error when mixing the titrant and the error approaches 2.5%. This may be accurate enough to balance the electrolyte in a silver cell, but could add up to significant money when buying silver.

For truly accurate results, an analytical balance that can weigh to .0001 gram or better is needed, and a smaller sample can be used. A standard sample is 0.2 grams weighed on a .0001 scale. A micro burette is also used to measure the volume of titrant used, and the titrant should be prepared as accurately as possible.

Start with a sample of the metal you want to test by taking drillings of a bar or otherwise gathering a representative sample. Add the weighed sample to a flask as we did with the solution above. Add a couple of ml. of distilled water and just enough HNO3 to the flask to dissolve the sample. At this point, you are simply working with a solution as described above. Add the indicator solution and distilled water as described above and continue with the titration.

Calculations - The calculation is similar to the one shown earlier. We’ll assume we started with a 0.2000 gram sample and we used 18.00 ml. of titrant in this example. 18.00 ml. * .0001 = .0018 moles of thiocyanate = .0018 moles silver. .0018 moles of silver * 107.8682 = 0.19416 grams of silver in the sample. If our original sample weighed 0.2000 grams, the purity would be 0.19416 grams of silver / 0.2000 grams original weight = .9708 * 100 = ~97.08% or ~970.8 fine.

Dave

 

[Martijn]

Thanks Martijn,

I'm aware it will be uneconomical to recover using Nitric but want a way to assay samples so that I can test other cheaper methods to get the silver out. I have lots of powder to play with but could have much more in the future so I need the assay methodology.

With your titration suggestion there's no clear cut off point unfortunately as it will just get cloudier and cloudier but there isn't a color change or something like that to know when to stop. I'm now looking into potentiometric titration as an accurate method

Yes there is. but it takes time. Add HCL and watch for a white cloud to form. if there is too much white, wait to let it settle to see clearly, hence the "it will take time" comment.
keep track of how much you added, and if you think you overshot it repeat with fresh batch. go slower at the end.
how slow and precise you will get a result, depends on you, but how precise do you want results? Within 5 to 10% tolerance of what is there will be a pretty good indicator imo. especially if the expected percentage is 0,4%.
It's the difference between "not worth your time" and "almost getting close to being profitable" a 4 gram per kilo ore is worth it, if processing costs are low.

As Kurt stated, a fire assay will get a higher percentage out of the powder and will be more precise. but for estimation purposes, leeching and tritation will give you an idea.

if you can go the smelting route, do so with a measured amount of powder and there you will have the result you can expect from your process. Your process may not be as efficient as an assay.

 

[4metals]

you would be MUCH better served doing a fire assay on this

I agree with Kurt, a sweeps fusion followed by cupellation will yield a bead of Silver to weigh. You may have to experiment to get the right flux to get a good fusion but I think it is the way to go.

Leaching to do a volhard titration is likely to hold up some values and not give a true representation.

 

[SunSilver1993]

Volhard Titration

How it Works - Potassium thiocyanate reacts with silver ions to create an insoluble precipitate of silver thiocyanate. It also reacts with the ferric (Fe3+) ions of the indicator solution to create a blood/wine red solution. Silver ions have a greater affinity for the thiocyanate than ferric ions, so the thiocyanate will react preferentially with the silver ions until there are no more of them in solution. When there are no more silver ions in solution, the red color will remain, indicating the endpoint of the titration.

Thanks for this Dave, I was actually coming to a similar conclusion myself by doing more research however I am concerned about the presence of other metal ions in the solution causing problems. For example there is a significant amount of copper and aluminium which could also form insoluble thiocyanates which would invalidate the result of the titration. As far as I can see copper thiocyanate is also insoluble in water.

I think that a potentiometric titration of sodium chloride into the leaching solution may be best for this as we know the chloride is selective for the silver which makes analysis much simpler

eaching this type of material (with acid) is most certainly NOT the way to go - NOT even for testing

you would be MUCH better served doing a fire assay on this

IF (the VERY BIG IF) fire assay shows this is worth going after for the silver then smelting would be the way to go

Kurt

Thanks for the input Kurt, can you let me know some reasons why acid leaching isn't the way to go for this? I don't have any equipment for a fire assay so would prefer to go the acid route if possible.

I agree with Kurt, a sweeps fusion followed by cupellation will yield a bead of Silver to weigh. You may have to experiment to get the right flux to get a good fusion but I think it is the way to go.

Leaching to do a volhard titration is likely to hold up some values and not give a true representation.

Thanks also for your input 4metals, do you have a link to a good resource for how to do such things? As said to Kurt would rather do leaching if possible but am open to all options.

 

[4metals]

First off, the XRF results can vary wildly on powder. You are wise to be seeking an analytical result that is reproducible and accurate. As it is at 0.4% Silver the value per ton is an the neighborhood of $3700. And that's based on content not cost to extract it. I would guess the number from a reliable assay will increase and it is worth the effort to accurately quantify it.
I would seek out a reliable fire assay lab and have them fire assay the material, with the understanding that when they come up with a reproducible result they will give you the flux formulation.

To do this in house will require an assay oven. A 6X6 Vcella kiln is about $1600. You will need an exhaust hood and an analytical balance as well. The process will involve a fusion and a cupellation. It will not require parting because the bead is pure Silver. the actual time involved is about 45 minutes for the fusion and about 45 minutes for the cupellation. With scale time and prep, you will be in for about 2 hours total time. Much of which is waiting.

 

[kurtak]

Thanks for the input Kurt, can you let me know some reasons why acid leaching isn't the way to go for this?

Because with this type of material (powdered silicon) it is next to impossible to get all of the values (silver) in the leach washed out of silicon powder - so - some values are going to stay tied up in the silicon powder

Therefore - all titration of the leach solution is going to tell is how much silver is in the leach solution - not how much silver is actually in the silicon powder

That is why - for this type material fire assays are done (as an industry standard) for best most/more accurate test results

Here are some places you can contact to ask about having fire assays done -----------

https://www.hazenresearch.com/capabilities/analytical-services/precious-metals-fire-assaying

https://fireassayer.com/index.html

https://www.actionmining.com/

Kurt

 

[kurtak]

I'm aware it will be uneconomical to recover using Nitric but want a way to assay samples so that I can test other cheaper methods to get the silver out.

Per the bold print - there is no need to "test" other methods "for getting the silver out"

You really only have 3 options ----------

1) acid leaching (nitric)

2) cyanide leaching

3) smelting

None of which will be "cheap" - especially if you intend to try to process in large amounts

Silicon, Copper, Aluminium Iron and Tin. We already have an XRF analysis showing that it contains 0.4wt% silver.

How much copper did your XRF test say ?

Kurt

 

[kurtak]

Two more questions -----

Is someone trying to sell you this silicon powder or are you creating this powder yourself ?

And ------------

How much of this do you have &/or have available to you ?

Kurt

 

[SunSilver1993]

Hi Kurt,

Thanks very much for your reply I am not US based but have found someone to do the fire assay for us. Appreciate the expert input! Also thanks for clearing up the meaning of 'values' in this context, I was getting confused there.

Copper is 1.5% Aluminium is 3.5% Tin is 1%

We are producing the powder and will have multiple tons