Is there a simple way to distinguish different grades of fine copper purity (on a trace metal basis) that can be carried out in a small lab?

Specifically, I would like to be able to differentiate 4N, 5N, and 6N (the commercial nomenclature for 99.99/99.999/99.9999% pure copper, respectively) grades.

The samples are in the form of pellets, and significant trace impurities include S, Zn, Ag, As, Cr, Ni, Fe, Pb, Sb, Cd, Hg, Se, Sn, Bi, Te, Ge, In, and Tl (more or less in that order).

  • 2
    $\begingroup$ You don't state if the sample is a bar, or shot / powder of Cu (lower part here). Perhaps a gravimetric analysis of a dissolved sample may discern 4N from 5N Cu if the gravimetric factor is just high enough .and. you are willing to spend enough material. The abstract of e.g., ASTM E53 looks more like electrogravimetry (electrolysis-like, e.g. from Cu foil to Pt net) than proceeding via dissolution, filtration of the precipitate, and incineration of the dedicated ash-free filter paper. $\endgroup$
    – Buttonwood
    Jul 24, 2021 at 12:07
  • 3
    $\begingroup$ Did you consider using a handheld XRF analyser? $\endgroup$
    – Loong
    Jul 24, 2021 at 17:16
  • 1
    $\begingroup$ @Loong Excellent idea! I did a quick search and both purchasing and renting of portable XRF analysers is viable, depending on one’s budget. My implicit assumption was that only simple chemical testing was of interest. $\endgroup$
    – Ed V
    Jul 24, 2021 at 17:34
  • 1
    $\begingroup$ @ Loong: If I informed myself correctly, a handheld XRF analyzer, even if accurate enough for the purpose at hand, would cost several thousand bucks, which is way too expensive for the purpose unless tons of copper would need to be handled. $\endgroup$
    – Hans
    Jul 24, 2021 at 20:05
  • 1
    $\begingroup$ Not my area ;but don't you need light spec for trace elements. ? In the stone age I worked on XRF of copper alloys and we only did the alloy elements and did not look for Sb and Bi $\endgroup$ Jul 24, 2021 at 20:37

2 Answers 2


It is unlikely that there is a simple method, not involving relatively expensive chemical analysis, for distinguishing 4N, 5N, and 6N copper specimens. Shown below is a photograph of some copper specimens I have:

Copper specimens

The evaporating dish contains copper shot, the large plate (4.76 mm thick) is OFHC (oxygen free high conductivity) copper and the third specimen is a copper bar that is 25.4 mm by 31.5 mm cross section. The standard banana is for scale. All of these copper specimens are supposedly 'pure', but how pure is 'pure'?

The problem with determining copper purity, when the specimens are already what might be considered 'pure enough,' is that the impurities are present at low levels and their identities matter a great deal. This figure, from Edward R. Tufte, The Visual Display of Quantitative Information, Graphics Press, Cheshire, CT, USA, ©1983, p. 49, shows published copper thermal conductivities on a log-log scale. The original publication reference is given as Ref.1 (page I-244).

Cu thermal conductances

This beautiful figure collects results from hundreds of studies in the published literature and shows, quite dramatically, that the large majority of those studies were incorrect, sometimes wildly so. What tripped up so many researchers is that they assumed they were measuring the thermal conductivity of copper specimens that were pure enough. But they were not pure enough. And oxygen turns out to be a major culprit in reducing the thermal conductivity of copper metal. Given that we live in an oxygen atmosphere, this is a problem. As well, trace oxygen determination is not an easy thing without appropriate instrumentation and methodology.

Bottom line: you need to know what the major impurities are (or might be) and then do trace determinations of those.


  1. C. Y. Ho, R. W. Powell, and P. E. Liley, “Thermal Conductivity of the Elements: A Comprehensive Review,” Journal of Physical and Chemical Reference Data 1974, 3 (Supp. No. 1), pp. I-244 (ISBN-13: 978-0-88318-216-1).
  • 1
    $\begingroup$ In section «Journal of Physical and Chemical Reference Data Monographs or Supplements» NIST offers direct access to the plot (as figure 60, page I-244) Tufte cites. An abridged version (including a doi and citation report) is provided by AIP. $\endgroup$
    – Buttonwood
    Jul 24, 2021 at 13:01
  • $\begingroup$ This is interesting information. If I understand well the discrepancy is predominantly attributable to oxygen content. I didn't think about this. I'm concerned with purity more on trace metal basis than oxygen content (I made an edit to the question to add this information). $\endgroup$
    – Hans
    Jul 24, 2021 at 16:54
  • 4
    $\begingroup$ For the figure, oxygen was the main problem. The example illustrates what happens if researchers get sloppy and assume what is not in evidence. In general, though, other trace species may be (and almost certainly are) problematic, e.g., S, Se, As, etc. So you need to know something about the significant trace impurities. No sane person, who needs 6N copper, would purchase and accept it without getting a serious chemical analysis certificate. The analysis would be done using, e.g., ICP-OES or ICP-MS. I see no real way around this. $\endgroup$
    – Ed V
    Jul 24, 2021 at 17:02
  • $\begingroup$ @Ed V: Say the significant trace impurities are S, Zn, Ag, As, Ni, Fe, Pb, Sb, Hg, Se, Sn, Bi, Te, Ge, In, Tl (more or less in that order). $\endgroup$
    – Hans
    Jul 24, 2021 at 20:17
  • $\begingroup$ *I forgot Cr in the list, above Ni. And Cd under Sb. $\endgroup$
    – Hans
    Jul 24, 2021 at 20:29

I don’t think there is a good chemical option to solve your problem, unless you do tests for particular contaminants, which is too much work.

A relatively simple option is magnetic sputtering, or accelerator enrichment. It may sound complex, but there are YouTube videos of people doing it with microwave parts, glass bottle, vacuum pump, silicone and some wire.

Idea is to use a magnetic field and a electric field to concentrate and deflect ions. And ions deflect differently depending on their mass. Initial evaporation of ions is usually achieved by microwave as heating when hobbyists are doing it.

It is still simpler than doing chemistry to distinguish 5N from 6N without knowing contaminants.

  • $\begingroup$ Thank you for these innovative suggestions. I'll watch some youtube videos to learn more! $\endgroup$
    – Hans
    Jul 25, 2021 at 10:02
  • 2
    $\begingroup$ A more familiar term for this technique, perhaps, is "mass spectrometry". $\endgroup$
    – zwol
    Jul 25, 2021 at 18:45
  • 1
    $\begingroup$ One thing is identification, but what about the quantitative aspect of it? $\endgroup$ Jul 25, 2021 at 20:03
  • 1
    $\begingroup$ @PeterMortensen cheap way is to measure how much residual is accumulated on a particular target spot through opacity. More precise is many electrodes and current measurments for each, on a target. But anyway, he only wants to distinguish very different examples, like 4N, 5N, 6N, with such huge difference in amount of contaminants, it is likely sensitive enough as is, just to say 'this is more likely to be X type of purity' $\endgroup$ Jul 25, 2021 at 20:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.