Gold recovery for a fire assay.

[Amol Gupta]

Attached is a photograph of our sample assay sheet.
Our assay process is as follows.

1. Inquart the gold sample with silver and copper and wrap in a lead sheet and form a ball of lead containing the sample inside.
2. Cuppelling the ball to remove lead and base metals from the sample.
3. Roll the bead of the remaining gold and silver into thin sheet.
4. Parting the sheet using nitric acid.
5. Melt the parted gold sheet to recover the fine(24 kt) gold beed.

Now coming back to the assay sheet column 1 and 4 are the columns indicating the check gold we used during the assay.

Right at the bottom we have the calculation with respect to the check gold we lost during the process.
We had two samples for the check gold weighing 153.429mg and 153.286mg.
The weight of the check gold we recovered is 153.194mg and 153.039mg.

Our recovery for the check gold is

(153.194 + 153.039)/(153.429+153.286) * 100 = 99.842%.

Now my question is, is there a way to recover my lost gold(0.158% in this case) or such losses are expected and nothing can be done about it.

Where is it am I loosing my gold.

I'd be happy to clarify our process further in case of any questions.

Thanks Regards.

 

[Yggdrasil]

Attached is a photograph of our sample assay sheet.
Our assay process is as follows.

1. Inquart the gold sample with silver and copper and wrap in a lead sheet and form a ball of lead containing the sample inside.
2. Cuppelling the ball to remove lead and base metals from the sample.
3. Roll the bead of the remaining gold and silver into thin sheet.
4. Parting the sheet using nitric acid.
5. Melt the parted gold sheet to recover the fine(24 kt) gold beed.

Now coming back to the assay sheet column 1 and 4 are the columns indicating the check gold we used during the assay.

Right at the bottom we have the calculation with respect to the check gold we lost during the process.
We had two samples for the check gold weighing 153.429mg and 153.286mg.
The weight of the check gold we recovered is 153.194mg and 153.039mg.

Our recovery for the check gold is

(153.194 + 153.039)/(153.429+153.286) * 100 = 99.842%.

Now my question is, is there a way to recover my lost gold(0.158% in this case) or such losses are expected and nothing can be done about it.

Where is it am I loosing my gold.

I'd be happy to clarify our process further in case of any questions.

Thanks Regards.

I'm not sure where you have found this information.
Inquarting and cupelling are two distinctly different processes with different uses.
Inquarting is used for karat Gold to dilute the Gold so it can be parted with Nitric.
Cupelling is a process used in assaying often called fire assaying.
Depending on the starting material one first scorify and then cupel.

What is your material?
Karat Gold?

 

[Amol Gupta]

I'm not sure where you have found this information.
Inquarting and cupelling are two distinctly different processes with different uses.
Inquarting is used for karat Gold to dilute the Gold so it can be parted with Nitric.
Cupelling is a process used in assaying often called fire assaying.
Depending on the starting material one first scorify and then cupel.

What is your material?
Karat Gold?

So the assay sheet has 4 columns column 1 and 4 are check gold(this is 24kt gold) column 2 and 3 are karat gold.

The check gold is used in assaying to check for the losses along the process and the karat gold is the stuff we are assaying for which in this case is 22kt gold as is indicated by the fineness row(indicated by the 8th row in the sheet).

 

[Yggdrasil]

So the assay sheet has 4 columns column 1 and 4 are check gold(this is 24kt gold) column 2 and 3 are karat gold.

The check gold is used in assaying to check for the losses along the process and the karat gold is the stuff we are assaying for which in this case is 22kt gold as is indicated by the fineness row(indicated by the 8th row in the sheet).

Yes, but it don't explain why you are mixing processes.
Either you inquart or you cupel.

Anyway for testing metallic Gold XRF or testing acids would be better.
Fire assays are used to determine the amount of precious metal in ores or similar.

Another thing is that losses are comparatively much larger on small samples than large samples.
As if you loose x amount of milligrams with one sample that will be similar of you have a sample 10-100 times in size.
It still will be approximat the same x amount of milligrams.

 

[Amol Gupta]

but it don't explain why you are mixing processes

The silver and copper is added to the pure gold to mimic it as closely as possible to the the karat gold.
The cupelation will remove the base metals but not silver, inquartation is then carried out to remove silver and get back close to 99.9 gold back.
Cupelation will not remove silver.

 

[Yggdrasil]

The silver and copper is added to the pure gold to mimic it as closely as possible to the the karat gold.
The cupelation will remove the base metals but not silver, inquartation is then carried out to remove silver and get back close to 99.9 gold back.
Cupelation will not remove silver.

I know.
But it puzzled me why you were mixing processes.
So you alloy the Gold then you cupel it.
Next you inquart and part it.
So how do you know how much Silver to add?

Each of these steps will cause some losses.
So the more steps you have the more loss you will have.

 

[Amol Gupta]

Next you inquart and part it.

No the inquartation process is taking place before cupelation.

how do you know how much Silver to add

This is based on the sample we want to mimic.

 

[Yggdrasil]

No the inquartation process is taking place before cupelation.




This is based on the sample we want to mimic.

I really do not understand why you do this.
Karat scrap are never cupelled.
It is inquarted and parted.

If it is to test a sample, it has to be done on the actual unknown sample as any alloy you make are known.

 

[Amol Gupta]

I really do not understand why you do this.

With all due respect I am in no position to question the procedure, in the country I live this is the procedure laid down by law and we are in no position to question or abstain from the procedures.

If we could just ignore the procedures involved and assess where I am losing gold and if any of it is recoverable.

I'd like to start off by having a reference.
What is the gold recovery I can expect if I refine 100gm(approx 3 ounce) of karat scrap if it is inquarted and parted.
My current recovery is 99.8%~, can I do any better....?

 

[4metals]

A proof of a fire assay is to measure and correct the quantity of silver retained in a fire assay. Try as you will, you can never remove 100% of the silver by parting. This is why your results puzzle me. First off the coronet weight of the proof is always greater than the weight of gold you started with because you can never remove all of the Silver with nitric parting. This is not the case on your data sheet.

Second you did not inquart the gold with Silver at a 3:1 ratio but the rolling out of the bead before parting allows slightly less Silver.

When adding Silver to proof gold the goal is to have the Silver content approximately equal to the Silver content in the sample before assay. So in your case you should have added more Silver in the proofs to have it as close as possible to the same Silver content in the samples and in the proofs.

The slight additions of copper is an often duplicated technique to mimic the effect the Copper has on the cupellation and should all be removed by cupellation.

When the proof sample is weighed, the start weight of the proof gold is divided by the finished coronet weight. You will notice the weight of the gold has increased, this is normal due to silver retention. By dividing the finished weight into the start weight you will come up with a number usually similar to .9985. (Once you are proficient at this.) That number is your correction factor, which is multiplied by your normal assay result to obtain a corrected fire assay result.

Your results are different than what is widely practiced in the world of fire assay. We can walk you through a few test proofs to be sure you get the process correct, but as you described it something is off.

With all due respect I am in no position to question the procedure, in the country I live this is the procedure laid down by law and we are in no position to question or abstain from the procedures.

This is something we can do little about but we can be sure you are interpreting it properly to try to identify the problem.

 

[Amol Gupta]

Try as you will, you can never remove 100% of the silver by parting.

You are right about this we start of with 99.99% gold but what we get back is 99.80-99.85% gold as the check gold.

Second you did not inquart the gold with Silver at a 3:1 ratio

We have used a ratio of 2.5:1.

So in your case you should have added more Silver in the proofs to have it as close as possible to the same Silver content in the samples and in the proofs.

The sample in this case was 22 kt gold with negligible amount of silver, the main contaminant being copper.

By dividing the finished weight into the start weight you will come up with a number usually similar to .9985. (Once you are proficient at this.)

Now this is where we need to talk a bit more, so you are telling me the weight of the check gold should increase because not all the silver could be removed.

My question is will I be loosing gold from the check gold along the process and if so by how much....?

 

[4metals]

Now this is where we need to talk a bit more, so you are telling me the weight of the check gold should increase because not all the silver could be removed.

The reason behind a proof assay (for gold) is to determine just how much silver is retained by the gold so you can lower your assay of the sample appropriately. This is so you do not pay gold prices for Silver retained. So the result of a proof assay is always less than 1 and it lowers the assay result. This is why I do not understand why your proof result came out weighing less than it started. There are some unexplained losses that I have never personally experienced nor have any of the assayers I have trained mentioned happening to them.

My question is will I be loosing gold from the check gold along the process and if so by how much....?

The old classic fire assay books suggest adding a little copper to the assay because as copper leaves surface of the cupel it may take an ever so slight amount of gold into the cupel with it. Since you want to duplicate exactly what happened in your assay cupel in your proof cupel, the copper addition makes them all behave the same. From what I had been taught only the last trace of copper has the ability to cause some gold to enter the cupel and not remain with the bead.

I have, in my younger days, actually ball milled and assayed spent cupels. They are a pain in the derriere to do because you have to meticulously clean the ball mill to avoid contamination before milling and sifting the spent cupels and the bone ash cupels that I always used are a naturally refractory material so they are difficult to assay to begin with. In the multiple times I have done this I had never detected as much as a milligram of gold and I have assayed it in assay ton sample proportions where 1 milligram equals one ounce per ton of material. So, I never paid those losses much mind.

Surely some other members here have some practical fire assay experience here, can we get some other opinions? Possibly @Lou ?

 

[4metals]

5. Melt the parted gold sheet to recover the fine(24 kt) gold beed.

Usually the folded coronet that is parted comes out intact and it is annealed until it glows red before it is weighed. What technique do you use to melt the coronets after parting, perhaps that causes some losses?

 

[4metals]

This is a copy of the write up I made to distribute to fire assayers I have trained.

Proofs for fire assays​

A classic fire assay purifies the gold in the sample by two mechanisms, the first is cupellation; which absorbs the base metals into the cupel, and the second is inquartation and subsequent parting in nitric acid. The limit to the accuracy of this method is at the very high purity end and as a result this method is used for alloys up to about 99.5% pure. One reason for this is the gold parted in nitric acid always retains some of the inquarted silver. If this is not corrected the refiner will over-pay the customer as they will be paying for silver retained weight and calculating it as gold. For this reason, a corrected fire assay is what is used to settle where the highest accuracy is required.

The goal in an assay correction is to make the sample and the proof as close as possible in the gold to silver ratio as possible. In a modern fire assay lab equipped with an XRF this is much easier than it was years ago.

First you need to get an XRF reading for both gold and silver in the alloy for assay. I always assay bullion with a 0.5 gram sample but any size will work. For this example I will use a .5 gram sample. If the XRF result is 54.31% Au for example weigh out (.5431 X .5 = .2715) grams of .9999 proof gold. This is very close to the quantity of gold that will remain in the cupel after your sample is done. Add the proof gold to the lead boat and then do the same for silver. If the XRF for silver was 7.54% weigh out (.0754 x .5 = .0377) grams of proof silver and add it to the lead boat. Now weigh out the same quantity of silver you will be using to inquart every sample and add that to the lead boat. Finally, add a small quantity of pure copper wire .005 or so to the lead boat. The copper will alloy with the gold and act the same as the copper in the actual sample acts in absorbing an ever so slight quantity of gold into the cupel. Now we have produced a proof sample, which will have the same gold to silver ratio as out sample at the end of the cupellation and behave the same way in parting.

The proof sample should be run in the same area of the furnace as the assay sample and be exposed to the same conditions. The samples from here on out are parted in the same strength acids for the same times and otherwise handled identically.

When the proof sample is weighed, the start weight of the proof gold is divided by the finished coronet weight. You will notice the weight of the gold has increased, this is normal due to silver retention. By dividing the finished weight into the start weight you will come up with a number usually similar to .9985. (Once you are proficient at this.) That number is your correction factor, which is multiplied by your normal assay result to obtain a corrected fire assay result. These results will vary depending on the karat of your starting assay. Typically an assayer running a large oven of samples will run a 10K, 14K and occasionally 18K proof samples to cover the entire furnace load of work.

The result is reported as a corrected fire assay.

In today’s world of cost cutting and speed, running proofs on every furnace load will be costly and time consuming. But after enough runs you will start to see a consistency and it is not unusual to use that average on all assays of the same karat range. A good assayer will check on this number occasionally. In today’s world a lot of assayers just use .995 which is actually ½ % in favor of the refiner. But I am here to teach refining and assaying, not morals!

 

[Amol Gupta]

I have a few more assay sheets I'll get back to you after checking the same.