Assaying gold by torch, chemicals and .001g scale.

[autumnwillow]

This is due to the lack of equipment.
I believe a fire assay lab would require at least 0.0001g scale accuracy, a furnace, cupels, a cone mold and different fluxes.

I had a case where I refined some karat scrap and relied on a touchstone assay. If the molten mixed karat scrap contains platinum, this would definitely ruin the touchstone assay accuracy. I have only found out that platinum was present when I checked the AR solution with stannous after the gold precipitated.

The purpose of the assay is to just get a very near estimate of gold content of the material that is to be refined.

Therefore if I were to perform such an assay with limited equipment, I would do the following:
1) Take 1g sample of the of the material to be refined.
2) Mix with 2 parts silver.
3) Melt and roll.
4) Digest in 50/50 nitric (in excess amounts).
5) Wash the powder several times.
6) Dry and melt.

I am actually doing the process already but I had to stop because I had to do something else, I will continue on Monday.
The sample was 1.088g of fine gold, 1g of silver and 1g of rose gold alloy containing one percent iridium.

What do you think would be the accuracy of my assay?
Any suggestions would be great.

Keita

 

[FrugalRefiner]

Depending on your skills and lab hygiene, I would think your results should be within a percent or two. While your 3 digit scale (.001) would theoretically be within 0.1 percent, we have to allow for a bit of error in the scale (maybe it's off just a bit), and that error compounds when you weigh the original sample, then weigh the finished button. Then allow for errors in processing; maybe there's a little silver that doesn't dissolve, or maybe you lose a bit of gold powder while decanting, or maybe a bit of platinum that doesn't dissolve, or... So the theoretical 0.1 percent quickly becomes 1.0 percent. Maybe a little more. Just my quick thoughts.

The sample was 1.088g of fine gold, 1g of silver and 1g of rose gold alloy containing one percent iridium.

This part confuses me. At first you said you were using a 1 gram sample with twice that amount of silver. But here you say the sample is a combination of fine gold, silver, and rose gold. I probably just don't understand what you meant. Maybe you can explain this part?

Dave

 

[4metals]

You are not adding enough silver to properly inquart the sample. If you do not add enough silver your parting will be incomplete and your errors will be on the high side as you will retain silver. Strive for 3 times the gold weight in silver.

Part first in 2 parts distilled water to 1 part nitric acid, followed by 1 part nitric acid to 1 part distilled water.

Then anneal the sponge by heating it until it glows but do not melt it. Annealing is important because it will allow you to see the impurity in the sponge from the iridium. Obviously the iridium will be weighed with the gold producing an error.

And you are correct in assuming that you need a balance accurate to 0.0001 g.

 

[autumnwillow]

This part confuses me. At first you said you were using a 1 gram sample with twice that amount of silver. But here you say the sample is a combination of fine gold, silver, and rose gold. I probably just don't understand what you meant. Maybe you can explain this part?

Right, my bad here. My real sample that I am processing is actually 1.088g fine gold, 1g rose gold alloy containing 1 percent iridium, 1g yellow gold alloy and 1g silver.

You are not adding enough silver to properly inquart the sample. If you do not add enough silver your parting will be incomplete and your errors will be on the high side as you will retain silver. Strive for 3 times the gold weight in silver.

The reason that I want to change the ratio to 2:1 is to avoid losses during decanting and melting. I have read somewhere here that the mint uses 2:1 ratio and rolls the sample very thin in order for the gold to maintain as large chunks.

My balance has 0.002g margin for error. So thats 0.02 percent.

I'll redo the experiment twice.

Thank you!

Keita

 

[g_axelsson]

My balance has 0.002g margin for error. So thats 0.02 percent.

To get 0.02 percent error you need to weigh a 100/0.02*0.002g = 10g sample.

If your final gold sample is 0.5g then the error will be 100*0.002/0.5 = 0.4 percent.

I still can't tell with certainty if your have four individual samples or one large.

Göran

 

[autumnwillow]

I tried the 2:1 ratio but it ended up not digesting well, so I made another alloy.

Here are the data:
0.650g 999.9 gold
0.380g rose gold alloy containing 1 percent iridium, 0.0038g iridium
0.774g white gold alloy
0.960g silver
total weight: 2.734g
gold content by percentage: 23.77 percent
iridium content: 0.0038g = 0.13 percent
melted and rolled weight: 2.654g
chemically parted gold weight(melted as button): 0.646g

I am not sure whether I should get the percentage of gold and iridium from the melted and rolled weight then compute from there or should I just consider the loss from melted and rolled weight as part of the cupellation? I just used borax though. Torched the bottom of the crucible until glowing red before I applied the torch to the metals.

Should I base it on the melted weight or non-melted weight?

If based on the non-melted and rolled weight.
Fine gold wt + iridium wt = 0.6538g
Final metal wt: 0.646g
This is a 1.19 percent error.

If based on melted and rolled weight.
Total metal weight: 2.654g
Gold content based on 23.77 percent: 0.630g
Iridium content based on 0.13 percent: 0.0034g
Fine gold + iridium weight based on percentage: 0.6334g
Final metal wt.: 0.646g
A 1.95 percent error here. I think. Not sure.

Which one is accurate?

 

[autumnwillow]

I'm sorry for asking that stupid question above. Of course the answer would be whichever is closer from the 999.9 gold sample. Now just replicate these a few more times to check its consistency and margin for errors.