How do you determine the purity of a silver bar?

[orvi]

I just use some analytical grade nitric acid, and look to see that I have a clear solution.
Even the slightest impurity will affect the colour.
I do have a spectrum analyser, but have never really needed to use it.

When we are at the level of purity 99.95+%, when alloy contain say 0,01-0,02% of contaminant, aside of copper, i don´t think you could reliably tell only by looking at the solution. For 99,9 % could be clearly possible (to see mainly copper-ammonia). I encountered tin once as the contaminant in the silver bar, if I remember correctly it was somewhere around 0,05%. We refined this anyway in the cell. Small piece of this (not completely) pure silver was dissolved in nitric.
And i can say that solution was crystal clear, without milkiness.

 

[justinhcase]

Unlike other nitrates, Sn reacts with water to produce an oxide which should be very visible with a large enough sample.
Sn + 2 H2O -> SnO2 + 2 H2

 

[Lou]

1000 mg/kg vs 100 mg/kg is not a job for visual estimation.

Use ICP.

 

[justinhcase]

1000 mg/kg vs 100 mg/kg is not a job for visual estimation.

Use ICP.

Emission spectroscopy (ICP-AES) and mass spectrometry (ICP-MS) are both highly effective chemical analysis techniques,
But that costs between £80 and £120 a run at a good lab.
What are you doing with the silver?
I sell silver nitrate back to my chemical provider, who runs quite extensive quality control.
He informed me that my product is all laboratory grade used in chloride test kits.
I must admit that I had to wait until I had a window in my processing that would allow me to eliminate all chloride based reactions in the work, aria and twice distil all my H2O before I managed a crystallisation I felt would hold up to testing.
That is an exceptional case, mostly if a buyer is testing my silver they are using a fire assay which will not really register trace contamination at that level.
How much money are you making buying the silver?
These are all relevant factors to consider when choosing a test method.

 

[sar]

ISO 15096-2020
This document specifies a method intended to be used as the recommended method for the determination of silver of fineness of and above 999 ?. For the determination of fineness of and above 999,9 ?, modifications described in Annex B apply.
Also for assurance take an ultrasonic test for homogeneity of inside the bar.

 

[Elemental]

I was going to publish ISO 15096-2020, but I see that it is has a copywrite. ISO 15096:2020
Cost: 58CHF (Swiss Francs) or $63.31 (US Dollars)

I am not advocating anyone purchase this standard, just providing more information about it.

Elemental

 

[orvi]

I trust you produce a high quality product. But in that trace quantities, visual isn´t the most reliable technique. Maybe some spectroscopic tester for Cu+indicator... Or nephelometry for trace tin content.
I´ve done a lot of "eye" analyses in my work. My opinion is that trace impurities could be detected reliably with trained eye to 0,05% level. Very specific test, if applyied (like Ni-DMG for example) could lower this number.

With unknown bar, refined with unknown process, it would be difficult to asses purity with just basic methods, without sensitive instrumental analyses.

 

[sar]

Here is 2014 edition
Unfortunately I don't have 2020 edition

Sar, I have removed the document you attached because it is protected by copyright. We do not allow copyrighted material to be posted here. FrugalRefiner

 

[campbellj46]

A mass spectrometer analysis is the only way I would buy such a large bar. The lab cost is negligible if you're dropping that much coin for a large bar of silver.

 

[Lou]

I'm bolder I guess. If I can melt it and it dissolves clear in nitric and doesn't "weep" during the melt and moreover, dissolves in ammonia, it's pretty damn pure silver.

Lot of people don't want their silver melted or drilled. In small format, the density/weight you can determine well enough, but larger bars, there's little to no premium so it gets grained up and sampled as grain and run on ICP-OES.

Btw, Nickvc, a lot of bar producers are jerks that won't even look up their serial numbers.