Impure Elemetal, OPM, NTR Metals & Provident Bars

This is a placemarker for a page that I expect to put up regarding potentially impure Elemetal, OPM (Ohio Precious Metals), NTR Metals, and Provident Precious Metals bars.

NTR, OPM, Elemetal Bar History

July 17, 2017

	2010	2011	2012	2013	2014	2015	2016	2017
NTR Metals (YY-#####)			2012 to ~Aug 2015
Elemetal (YY-#####)							~Aug 2015-2017
OPM Old (YY-#####)	2010 to ~Jan 2013
OPM New (A########A)				~Jan 2013 to ~Aug 2015

From this chart, we see that OPM made their old-style bars (with a serial number format including the year, "YY-#####") from some time in 2010 through early 2013. They then changed the design to the newer design, with a serial in the format A########A (where 'A' is a letter from A-Z), that they made through about August, 2015.

NTR Metals had bars in the YY-##### format from sometime in 2012 through around August, 2015.

Elemetal bars replaced the OPM and NTR bars around August, 2015, using the old YY-##### format.

Data Matrix ("QR") Codes

July 17, 2017
Starting around January, 2013, OPM 1oz gold bars started including Data Matrix codes (often confused by "QR" codes). These 2D barcodes would appear on both the bar itself and the packaging. These data matrix barcodes were discontinued around August, 2015, when OPM bars were replaced with Elemetal bars.

While some Provident bars have a standard QR code on the packaging ("http://www.providentmetals.com"), the OPM bars have a more thorough Data Matrix code laser-etched on the bar and printed on the packaging. It seems like nobody on the Internet has ever bothered looking into them. The one from the bar I am sending for lab testing works out to:


OPM Metals
K38124632S
1 oz
8/21/2015
Minted in the USA

That's pretty cool. It doesn't have the fineness of the bar, but it does have the serial number and the date that it was made. And of course, from the serial number, OPM could in theory check their records to determine further information about the bar (perhaps the batch of gold used to create it, the fineness, etc.).

Testing Purity

July 17, 2017
Testing the purity of a gold bar is not easy!

Most gold tests simply cannot distinguish between .997 gold and .9999 gold. Measuring the weight/dimensions, the specific gravity test, touchstones, electrical conductivity, acid tests and ultrasonic thickness gauges typically aren't accurate enough.

The cheapest test that can often do an acceptable job with purity is standard XRF (starting at a whopping $15K or so), but it requires calibration, proper usage (e.g. the amount of time it takes to get a reading), and can be fooled by gold or silver plating. So it is a start, but not a guarantee of a pure/impure bar.

To properly determine the purity of gold, a high-tech spectrometry process (such as Inductively Coupled Plasma Spectrometry) is really needed. While a fire assay can do a very good job, it may not even be able to determine if a bar is truly .9999 fine (see e.g. "Analytical Methods" in Gold Bulletin).

.997 versus .9999+

July 17, 2017
Some people have commented that they don't really care whether their gold marked as .9999+ is really .997. For the purposes of this post, I'm going with .997 because that is the one firm number I have seen (from XRF). Remember, of course, that XRF isn't always very accurate.

A study of high purity gold in 1998 tested 7 bars labelled ".9999+". The least pure was .99993, with the average about .999946. So let's define ".9999+" as .999925. That means that out of every 1,000,000 parts, 999,925 are gold and 75 are something else.

Compare that with .997 gold, where out of every 1,000,000 parts, 997,000 are gold and 3,000 are something else. That .997 bar would have 40 times the impurities of the .9999+ it claims to be.

Or, in images:

.9999+	.997

Each image has 40,000 pixels. The .9999+ image has 3 dark pixels representing impurities (.999925). The .997 image has about 120 dark pixels representing impurities (.997). If your monitor is like mine, you'll have troubles telling if the impurities in the .9999+ fine example are dust on your screen or impurities in the image. But the .997 is obviously impure. Note that unlike this picture, you will normally not see impurities in a gold bar (although the color may be slightly different at lowering purities).

.997 fine gold works out to just 0.3% less gold than .9999+ fine gold, or roughly $4 worth of gold. It's not a huge amount. But, just like the Salami Embezzlement Technique, it could potentially add up to a lot of money.

Page Started

July 17, 2017
I started this page due to allegations of impure Elemetal, NTR, OPM, and Provident gold and/or silver bars.

The allegations started with an anonymous tip that Dillon Gage's refining department tested impure bars, and that Dillon Gage stopped buying back NTR, OPM, and Elemetal bars as a result. Dillon Gage declined to confirm this or comment.

After reporting on this and asking for tips, I received a handful of reports, mostly anonymously. There was a report of a jeweler buying OPM gold bars that tested .999 with XRF that turned out to be less pure when melted. Then a report that Elemetal bars are usually pure, with the explanation that the issue was due to irregular minting processes causing bullion verification devices to fail. Then a report of XRF showing bars as .992-.997 fine. Unfortunately, those were all anonymous reports without evidence backing them.

But I also got a report from someone I know of an Elemetal gold bar marked .9999+ showing as .9970 with XRF (with a photo of the results). And an OPM gold bar marked .9999 that failed a Simga Metalytics precious metals verifier for .9999 fine gold (and outside the expected range for .999 fine).

Then, Kitco declared that they would offer a "large discount to spot" for Elemetal, OPM, NTR, and Provident gold and silver bars. They explained that this was due to the fact that Elemetal's COMEX status being revoked, which does make some sense: but not quite enough. Why drag OPM, NTR, and Provident into this, and why not just specify that bars from non-COMEX-approved refiners will be melted and paid lower prices? Kitco doesn't have a history of being entirely forthcoming (like when they explained delays due to a police raid and gold seizure as due to a "major power outage" and then later as "technical difficulties" -- off topic here, but true).

The actual evidence (XRF and Sigma Metalytics tester) isn't quite good enough to prove a problem. But with all the anecdotal reports and the evidence that I did have, it justified having the bar tested at a lab designed for such purposes. And it justified starting this page, regardless of the results of the lab testing.