Simple Spectrometer to Test Gold

[Roger Nevez]

Hi,

Recently I've watched a video [1] about how to test gems with a spectrometer and started to think whether it's possible to do the same with gold.
Does anyone know if it's possible to build a simple spectrometer [2] to identify if a gold coin/bar is fake or not?

Regards,

Roger Nevez

[1] Gem Collector Tools - Spectrometer
https://www.youtube.com/watch?v=BXk0XsylC-I

[2] Make your own CD spectrometer
https://www.youtube.com/watch?v=ZowYVDQDDZ4

[Jeffrey Warren]

Hi, Roger - can you briefly summarize how gold testing works? What do you illuminate the sample with? 

Thanks!

--
Post to this group at plots-sp...@googlegroups.com
 
Public Lab mailing lists (http://publiclab.org/lists) are great for discussion, but to get attribution, open source your work, and make it easy for others to find and cite your contributions, please publish your work at http://publiclab.org
---
You received this message because you are subscribed to the Google Groups "plots-spectrometry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plots-spectrome...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Roger Nevez]

Jeffrey,

> "can you briefly summarize how gold testing works?"

By gold testing I mean a method to test a metal coin or bar in order to be reasonably sure that the main content is the element gold.

For example, in the jewelry sector the most common gold testing is the touchstone (or acid test).

Unfortunately, this technique requires a scratch in the surface of the sample.

To avoid destructive tests, you might use high-tech spectrometers.

Such devices accurately provide the composition of the sample surface on the go.

Unfortunately such devices are very expensive.

> "What do you illuminate the sample with?"

This is one piece of puzzle I'm trying to solve:

- Is it possible to build a simple spectrometer to test gold? How this device should look like?

- what should be the source of light? Common lamps? Projector lamps? UV light?

- what should be the length of the aperture? Razor blades might work?

- analyzing the spectrum only in the visible range might work? Is it going to be necessary to analyse in UV/IR range?

Regards,

Roger Nevez

To unsubscribe from this group and stop receiving emails from it, send an email to plots-spectrometry+unsub...@googlegroups.com.

[Jeffrey Warren]

On Fri, Oct 23, 2015 at 11:56 AM, Roger Nevez <roger.nevez@gmail.com> wrote:

- analyzing the spectrum only in the visible range might work? Is it going to be necessary to analyse in UV/IR range?

I think this is the most important question for you -- the spectrometers we've designed have pretty much settled questions of length of the aperture, slit material. But do you have some literature or documentation on how gold spectrometry is done? Is there a good example to look at? 

To unsubscribe from this group and stop receiving emails from it, send an email to plots-spectrome...@googlegroups.com.

[Dave Stoft]

A quick check on the web indicates that metals (gold, silver, etc) testing (as in this example) are performed with an X-ray florescence spectrometer (some reference the acronym XRF) where an x-ray source strikes the metal and knocks some electrons from their shells which emit at another wavelength -- those wavelengths are then detected.

I rather doubt visual spectrum reflections could tell the difference between real gold and gold 'simulants' or gold mixes.

[Roger Nevez]

Dave,

> "[...] metals (gold, silver, etc) testing (as in this example) are performed with an X-ray florescence spectrometer (some reference the acronym XRF) [...]"

Yes, nowadays XRF is the most common way to test gold and other precious metals.

There are some handheld devices very easy to be used, unfortunately very expensive (>$5,000).

> "I rather doubt visual spectrum reflections could tell the difference between real gold and gold 'simulants' or gold mixes."

If there's no way to cause differences in the visible spectrum of the reflections, the most likely is that it's not possible to build a simple spectrometer for gold testing. But, just playing the role of devil's advocate:

1) if you shine a white light on a metal, small shadows might be formed in the visible spectrum of the reflections?

2) Do this shadows correlate with the size of the atoms that make the metal?

3) Is it possible to predict theoretically the size and where this shadows would be?

4) Is it feasible to build a simple spectrometer to measure that?

Regards,

Roger Nevez

[Roger Nevez]

> "[...] thought experiments are good methods."

Yes, they are.

> "[...] by "shadow" I'm assuming you mean absorption lines in the reflected spectrum."

Yes, you are right.

> "[...] I'd think such absorption lines would be very narrow so probably difficult to see with the PLab spectrometer [...]"

If the absorption lines would be too narrow to be identified in a professional spectrometer, the most likely is that it's not possible to build a simple spectrometer for gold testing.

> "[...] if you were comparing polished pure gold vs polished 50:50 mix with silver [...]"

Precious metals investors usually don't care about the beauty of the asset. They are more interested in the weight of the pure metal.

So, a spectrometer that could only work in such conditions it is, unfortunately, not that useful.

Although the conclusion was that it is not possible to develop a simple spectrometer to test gold, just to know that this is a dead end it's already useful. 

And, once more, the Aristotle density test won the battle of simple gold tests.

Thanks for all ideas.

Game Over,

Roger Nevez

Yes, thought experiments are good methods. So, by "shadow" I'm assuming you mean absorption lines in the reflected spectrum. I'm not a chemist but I suspect not. If it did, I'd think such absorption lines would be very narrow so probably difficult to see with the PLab spectrometer that has a difficult time resolving better than 5nm. If you have a nice gold ingot available you could try it ;-) Perhaps, if you were comparing polished pure gold vs polished 50:50 mix with silver the color difference might be detectable but I'd guess the sensitivity would be low. But, good thoughts for experiments to try if you have the materials.

> "I rather doubt visual spectrum reflections could tell the difference between real gold and gold 'simulants' or gold mixes."

If there's no way to cause differences in the visible spectrum of the reflections, the most likely is that it's not possible to build a simple spectrometer for gold testing. But, just playing the role of devil's advocate:

1) if you shine a white light on a metal, small shadows might be formed in the visible spectrum of the reflections?

2) Do this shadows correlate with the size of the atoms that make the metal?

3) Is it possible to predict theoretically the size and where this shadows would be?

4) Is it feasible to build a simple spectrometer to measure that?

Regards,

Roger Nevez

On Friday, October 23, 2015 at 4:16:04 PM UTC-2, Dave Stoft wrote:

A quick check on the web indicates that metals (gold, silver, etc) testing (as in this example) are performed with an X-ray florescence spectrometer (some reference the acronym XRF) where an x-ray source strikes the metal and knocks some electrons from their shells which emit at another wavelength -- those wavelengths are then detected.

I rather doubt visual spectrum reflections could tell the difference between real gold and gold 'simulants' or gold mixes.

On Friday, October 23, 2015 at 8:59:52 AM UTC-7, Jeffrey Warren wrote:

[Mathew Lippincott]

I need some convincing that colorimetry wouldn't work. I've never seen fake gold that looked like gold, or low-grade gold alloys that looked like pure gold. Gold has a very unique color and is worth a ton of money; if good fake gold existed I'd assume I'd have seen it in jewelry by now.  

Why wouldn't comparing the reflected light from two evenly finished pieces of metal, one gold, one unknown, in a spectrometer where amplitude calibration was available, give a decent answer?  

Wikipedia has a graph of gold's reflectance compared to silver and aluminum:

https://en.wikipedia.org/wiki/Reflectance

I'm pretty confident that I can tell 18-24K gold from 10-14K gold, and pure or sterling silver from lower grade silver just by eyeing it. In reference to a known gold sample I'm sure I could distinguish gold with good repeatability. I don't think I could say "that's pure gold" with total confidence, but I could say "that's at least 75% gold."  

if anyone wants to test this, please send me free gold samples and I'll happily attempt and document a blind trial ;-)

--
Post to this group at plots-sp...@googlegroups.com
 
Public Lab mailing lists (http://publiclab.org/lists) are great for discussion, but to get attribution, open source your work, and make it easy for others to find and cite your contributions, please publish your work at http://publiclab.org
---
You received this message because you are subscribed to the Google Groups "plots-spectrometry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plots-spectrome...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

[Dave Stoft]

Good thought with the wiki plot. The 400-500nm region of the Ag/Au diff region of 340-500nm of ~40% might be detectable using a ~400-450nm narrow-band source to avoid the curve variations over that band where the webcam noise floor is high and is rapidly loosing sensitivity. Since both Al and Ag reflect more than Au, relative signal response (i.e. calibrating reflection levels of pure gold and pure aluminum) might detect the degree of contamination of a gold sample. (A noisy 40% dynamic range might limit the resolution.) If the test sample were always a uniform mix and surface prep it would help. Would be an interesting experiment if one had the materials.

[Dave Stoft]

On further 'reflection' (NPI) in order to eliminate surface-texture-variation as a source or reflection signal change, all references and samples would probably have to have the same surface treatment -- i.e. as in all of them polished to the same degree and the reflection angle and illumination be very uniform. There would be no way to detect the difference between  some impurity vs a surface contour which was simply more reflective.

[Roger Nevez]

Mathew,

> "I've never seen fake gold that looked like gold, or low-grade gold alloys that looked like pure gold."

This is one of the reasons why gold was used as money for thousands of years. It was easily recognizable and difficult to counterfeit. Unfortunately, nowadays people are not that familiar with gold in order to be sure whether the piece is fake or not. Unless there is a simple method to detect fraud, the average guy will always be afraid to risk his own capital in gold investments.

> "Gold has a very unique color [...]"

Agree, gold is the only metal element that is yellow. However, there are some alloys with similar color (at least for a naked eye) which are not that easy to be spotted. Fortunately, most of this cases can be identified with simple tests such as:

- signs of oxidation

- signs of discoloration

- magnet (gold is diamagnetic)

- bite (gold is relatively soft)

- ping (mainly for coins)

- density

Even so, there is no way of a serious goldsmith pay full price in a gold piece without doing further analysis such as acid test, XRF or even melting the piece.

> "Why wouldn't comparing the reflected light from two evenly finished pieces of metal [...] give a decent answer?"

There are some practical problems with that:

1) Gold coins

There are several parties minting gold coins. Each one with different textures, images and sizes. So, this would require a considerable effort to build a database for each gold coin reflected light.

Nowadays, the only characteristic that is roughly the same is the 99.9% purity.

2) Gold bars

Gold bars for investment usually don't have special surface treatment. The molten metal goes into a mold, cools down, receives a foundry mark and it's basically ready for delivery. So, due to this manufacturing process, there will always be some irregularities over the surface, like small waves in the middle of a lake.

> "In reference to a known gold sample I'm sure I could distinguish gold with good repeatability."

Ok, but no goldsmith rely solely on visual inspection before buying a gold piece and, probably, there's a reason for that.

Therefore, it's wise to do the same when investing your own money in gold.

> "I don't think I could say "that's pure gold" with total confidence, but I could say "that's at least 75% gold.""

Most investors would not pay full price in a gold bar with such margin of error for the purity.

Dave,

> "[...] the webcam noise floor is high and is rapidly loosing sensitivity."

If the test could be done with a webcam or with a smartphone, it would be perfect!

> "If the test sample were always a uniform mix [...]"

Nowadays the only purity that matters is 99.9% gold. Bellow that it's more likely to be jewelry or numismatic, and not investment.

> "[...] and surface prep it would help."

If this surface prep is as simple as cotton + isopropanol, it would be nice!

> "i.e. as in all of them polished to the same degree and the reflection angle and illumination be very uniform."

Unfortunately, to do such surface treatment means that the test would be destructive, instead of non-destructive.

> "There would be no way to detect the difference between  some impurity vs a surface contour which was simply more reflective."

Unfortunately, there is no simple way to avoid random surface contour over gold bars.

[Dan Beavers]

I am confused. Why would the surface finish have such a detrimental
effect? Wouldn't you just look at what frequency the large amplitude
change happens? That would be independent of the amplitude.

> --
> Post to this group at plots-sp...@googlegroups.com
>
> Public Lab mailing lists (http://publiclab.org/lists) are great for
> discussion, but to get attribution, open source your work, and make it
> easy for others to find and cite your contributions, please publish your
> work at http://publiclab.org
> ---
> You received this message because you are subscribed to the Google
> Groups "plots-spectrometry" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to plots-spectrome...@googlegroups.com

> <mailto:plots-spectrome...@googlegroups.com>.

[Dave Stoft]

Maybe I'm missing your point, but I'm making the assumption that, as in the referenced wiki curves, gold reflects ~40% less than silver in the 350-450nm band and that difference is the source of the visible spectrum measurement. To make a measurement of an unknown mix of gold/silver (again using the simple example used for this discussion) one would need to take 2 readings in a fixed source-sample-spectrometer configuration -- pure silver and pure gold. (Aside: There's also the problem that pure silver oxidizes rapidly so its reflection efficiency may not remain stable). If a 50:50 test sample of the same surface characteristics replaces the reference, we'd expect to see a spectral amplitude change in the 350-450nm band midway between the references. But what if the reference and test samples did not have the same reflectivity simply due to the differences of each of their surface diffuse scattering of light. The measurement is a _relative_ measurement (not absolute and not just visualizing the 'shape' of the broadband curve) and so, to eliminate extraneous errors, the nature of how the light get's reflected/scattered should be the same for all references and test samples. (Aside: Because of these issues, and others, my hypothesis is that detecting 99.9% purity is simply not possible using a PLab device.)

[Dan Beavers]

I saw the light. I was just thinking of detecting gold vs silver, not
the alloying of gold. My bad, too many things on my mind. But thanks
for the response.

[Roger Nevez]

> Mathew: "if anyone wants to test this, please send me free gold samples and I'll happily attempt and document a blind trial ;-)"

> Dave: "Would be an interesting experiment if one had the materials."

Unfortunately, it doesn't seem that spectometry is the right experiment for gold testing.

Does Public Lab also provides other non-destructive experiments that might be used for gold testing? (example: eddy current, ultrasonic, ...)

ps: let me know if this is out of subject of this mailing list