A compositional analysis of silver Roman imperial coins using XRF

The author of this analysis, Rasiel Suarez, has directly given me permission to publish this article on CoinTalk. The author's full title is "A metals analysis of silver Roman imperial coins using X-ray fluorescence spectroscopy." If you have access to Academia.edu, you can read the entirety here, including the methodology and calibration--which I'll leave out below in the interest of space. Questions about these analyses should be directed to the author--PM me for his email address.

Abstract​

An overview of the compositional makeup of silver Roman coins sheds light on a number of poorly understood areas. Contrary to popular belief, the debasement of the Roman denarius was not as linearly progressive as had been initially believed. The testing performed also shows an apparently unrecorded method of counterfeiting late Roman silver coins as well as a blueprint for using this technology as a means to detect forgeries.

Introduction​

A number of studies have been performed on ancient coins to determine the composition of their alloys using a variety of means. One of the most popular has been X-ray fluorescence spectroscopy (XRF). However, as the process is both expensive and time-consuming the studies have generally been limited in scope. As a result, numismatic researchers have needed to piece together the published data to form a cohesive understanding of the various alloys in use. Gaps in the data have been filled in through extrapolation or speculative hypotheses and where published data conflicts, as is to be expected from disparate methodologies and equipment used, reconciling the differences has resulted in controversy.

Equipment and Methodology​

For the purposes of the analyses an Olympus DELTA Professional handheld gun was employed. This late-model unit comes with a 40kV X-ray source and utilizes a silicon drift detector along with multiple ionization beams for its readings. The beam area focus was approximately 10mm.

Here I will snip out a lengthy discussion of calibration. If you want to read it, go to the link above or PM me for the complete report in pdf.

​

Results​

A total of 162 Roman and Byzantine coins were analyzed. The amount of information obtained defies any easy way to portray the data holistically and, in fact, lends itself well to having different aspects published separately. The primary focus on the present paper therefore will be to show the purity of the silver and copper coinage in a chronological timeline.

Here I snip out a discussion of calibration and detectable silver content as it pertains to the surfaces of worn versus "MS" silver Roman coins, with results displayed in fig. 1. He concludes that: "We should therefore allow a ±5% margin of error when taking into account the of the readings of 84 individual coins dating from approximately 1 CE to 293 CE."​

In Fig. 2 we can finally appreciate the true extent of the degradation of the denarius and its successor, the antoninianus. The intentional debasement with copper begins in 64 CE at an estimated 1 to 20 ratio and this amount is slowly but steadily increased over the next century. It acquires a critical juncture sometime during the reign of Commodus at which time the amount of silver per coin undergoes a much more rapid decline. However, this debasement is not as linearly progressive as once thought with the fineness frequently bouncing back before a new cycle of debasement would plumb new lows.



If we take a closer look at the above data, in Fig. 3 we find that for the first two centuries of the imperial era, and despite Nero’s deliberate debasement of the denarius, only a very gradual down slope begins to develop. In fact, the silver content remains quite stable and at no point falls below what would be the equivalent today of sterling silver at 92.5%. This, however, is a snapshot of higher grade coins. A look at a worn denarius of provincial mintage dating to 68 CE reveals an interior reading of only 91.46% and several second-century worn denarii, as revealed in Fig. 1, were also found to have a silver content barely reaching the 90% mark. If we make the assumption that when they were in like new condition they should have tested 3% higher in fineness we would still be slightly short of the norm so either these are exceptional or the model needs to be less conservative to maintain the integrity of the findings. As stated before though, the 5% allowance is sufficient to cover these small deviations.



A glance at Fig. 4 below however shows us a chaotic third act for the denarius and the antoninianus clearly showing a jumble of peaks and valleys within short spans of time. The critical point is reached during the decade of the 260s at which point the level of silver has reached such a low amount that what little remains is reserved for an aesthetic coating with a nearly fully debased core composed primarily of copper, though here with a rising trend in the content of lead, tin and other metals Fig. 5.





After the collapse of the silver coinage a new high-purity coin is introduced in 294 CE but it too rapidly succumbs to the ravages of inflation until, as before, it remains only at trace levels. While high quality silver coins will be sporadically minted from this point forward most will be issued in extremely limited circumstances as ceremonial distributions. Silver coinage will not circulate to any great extent until a new denomination, the siliqua, is mass-produced towards the middle of the fourth century. Unfortunately, an insufficient number of these coins were available for this study. All the same, a look at the few tested hints at significant debasement as Rome’s economy continued to falter (Fig. 6) but the most surprising revelation is the previously unknown use of a tin-copper-antimony alloy as a faux stand-in for silver (Fig. 7).





Addressing the issue of counterfeiting, XRF is quickly gaining ground in archaeometry as an important tool in revealing ancient and modern deception schemes alike. The opportunity to study samples of these was limited in this case to just three coins. A Caligula denarius of good imperial style and showing considerable circulation wear, which was already suspected of being a silver-plated contemporary copy due to a small but telltale surface break, was confirmed as such when the XRF reading showed an anomalously low silver content of 89.55%. Two denarii, a Marciana and a Pertinax, both particularly desirable by collectors for their rarity, of reasonably good design and carefully weathered to conceal their inauthentic nature, were quickly unmasked by the analyzer which reported unacceptably low readings of 82.00% and 80.09%, respectively.

Conclusion​

The steady decline in the fineness of the silver used in the denarius had been known well before the advent of analytical equipment able to confirm this. However, where the decline had been suspected of occurring at a measured pace this study shows only modest drops in the fineness during the first and second centuries. This much more closely correlates to the historical period when the Roman economy was at its strongest. The integrity of the coin then begins to quickly fall apart over the course of half a century until the nominally silver coinage had but traces of this metal left. While the trend is inexorably towards decreasing silver content many individual issues prove to have anachronistically high levels of silver; perhaps due to abortive policy attempts at stalling inflation but possibly simply explained as inconsistencies in the available supply and shoddy quality control.
While admittedly not a conclusive standalone tool, the use of XRF analysis proves its worth in telling apart counterfeit coins from genuine ones whether ancient or modern in origin.

Acknowledgements​

The author would like to show his deep appreciation to Olympus for providing the generous loan of the analytical equipment used in this test.

References​

A. Italiano, L. T. (n.d.). A comparative analysis of old and recent Ag coins by XRF methodology. ESE - Salento University Publishing.
Carter G. F., K. M.-B. (1978). Chemical Compositions of Copper-Based Roman Coins. V. Imitations of Caligula, Claudius and Nero. Revue numismatique, 6e série - Tome 20, 69-88.
L. Beck, S. B. (2004). Silver surface enrichment of silver–copper alloys: a limitation for the analysis of ancient silver coins by surface techniques. Elsevier B.V.
Wanhill, R. J. (2001). Microstructurally-induced embrittlement of archaeological silver. Amsterdam, Netherlands: National Aerospace Laboratory NLR.

The coins used in testing are available for further study online (contact me for the link)

 

Log in or Sign up to hide this ad.

Thank you very much!

 

Thank you KurtS for a very interesting article.

 

That is very scary and unnerving.
I hope the Imperial government strung those cheats up high!

 

I fact-checked the reference to "Babbitt Metal"--the composition of the counterfeit is very close to a common bearing alloy used today.

 

So it's dammed near impossible to determine if a Siliqua is made of Babbitt or Silver without XRF equipment?

 

I didn't do the analysis myself, but I suspect that XRF is the least destructive method--chemical tests for silver would leave some damage.

 

Thank you Kurt for bringing this to my attention. I will have to be more careful in the future, and only look for Siliqua with confirmed die matches and an obvious silver tone.

 

Kudos Kurt! - Academia.edu is a site I use to post some of my research papers -- on other subjects. Usually great stuff!

Gary in Washington

 

And what if it is happening in your present government?

 

Various Methods for the Metallurgical Analyses of Coins
(In Nuce)


(1)

SEM/EDS (Micro-Analysis):
Surface analyses, chemical analysis and imaging on a variety of materials are performed using a Scanning Electron Microscope (SEM). The Scanning Electron Microscope is equipped with an Energy Dispersive Spectrometer (EDS). SEM/EDS provides chemical analysis of the field of view or spot analyses of minute particles. More than 90 elements can be detected with most low-atomic number detectors using the SEM/EDS method. This micro-chemical analysis is also a NON-DESTRUCTIVE test. The equipment, takes up some space, and is quite expensive. But delivers great accuracy, and it can detect numerous metals at once. I personally, am not sure how well it works for a coin! Input invited!


(2)

Typical SPECTROGRAPHIC ANALYSIS:

(Bulk material and alloy certification):
An alternative to SEM/EDS, the Optical Emission Spectrometry (OES) technique utilizes a high-energy spark created across an argon-filled gap between an electrode and a sample of the material to be analyzed. The spark creates an emission of radiation from the excited sample surface with wavelengths characteristic of the elemental composition. The spectrum of radiation is separated into the distinct element lines and the intensity of each line is measured. Finally, these are precisely converted into concentration values for each element present. Typical applications involve determination of the alloying content of iron and steel, aluminum, copper, nickel, zinc, lead and many other metals and alloys. Optical Emission Spectrometry continues to be the reference technique for direct chemical analysis of solid metallic samples. The unmatched combination of accuracy, high speed, precision, stability and reliability have made it an indispensable tool for production and verification of quality metallurgical products. Since the metal to be tested must first be ground and inserted into an electrode it is obviously, DESTRUCTIVE.


(3)

X-RAY FLUORESCENCE (XRF)

Elements can absorb very specific frequencies of X-rays and, in turn, through a reshuffling of their electrons, kick out other, characteristic bands of X-rays. Xray fluorescence spectrometry utilizes this process to identify and quantify the elements found in a sample.The test is NON-DESTRUCTIVE. First, the X-rays are produced and sent to the sample. Fluorescent X-rays are then sent back to a detector, and the detector counts them. The algorithms then turn these counts into a concentration.

While laboratory XRF instrumentation can be large, powerful, and imbued with complex and sophisticated competencies, the breed of portable instruments is designed to be taken to a work or research site. For the most popular of these—the handheld radar gun-like instruments—taking a measurement is essentially a matter of pointing at a sample and pulling the trigger.


(4)

Optical Emission Spectroscopy (OES and AES)

Optical emission spectroscopy (OES) or atomic emission spectroscopy (AES) are an important tool for fast and accurate elemental analysis of metals. Optical emission spectroscopy (OES) provides a non-evasive probe to investigate atoms, ions and molecules within a plasma. (AES) is a method of chemical analysis that uses the intensity of light emitted from a flame, plasma, arc, or spark at a particular wavelength to determine the quantity of an element in a sample. The wavelength of the atomic spectral line gives the identity of the element while the intensity of the emitted light is proportional to the number of atoms of the element. As usual in most spectrographic analysis techniques, the material tested is ground to a power, put into an electrode, or diluted with various acids and made into a plasma or spray. The light given off when "sparked" is measured on a or in a calibration scale. These two tests are DESTRUCTIVE.



(5)

COIN DRY WEIGHT AND DENSITY:

Density = mass/volume

These tests are usually the least expensive to perform. And are NON-DESTRUCTIVE. Pure metal compositions are the easiest to evaluate. But with accurate charts and prior measurement comparisons, various alloys can be perceived, or at least recognized. Density can indicate a change in the composition of a material, or a defect in a product, such as a crack or a bubble in cast parts (known as voids), for instance in coins or in bars et cetera.

The production of accurate charts of various coins and their alloys -- showing each coins specific gravity, is a first step. Very little work has been done in this area of numismatics.

A general rule is: The higher the temperature, the lower the density. Coins should be tested at normal room temperatures 62-75 degrees F. as well as the distilled water. Scales should be in a very calm environment, no drafts, best if inside a glass chamber! Both the buoyancy and the displacement methods are suitable for density determination on porous solids. To determine the bulk density of porous material, the sample can also be covered in a wax or latex coating or layer to prevent liquid from entering open pores.


Accuracy requires that the temperatures be stable, and that the porosity of the coin is evaluated, if needed. There are formulas which can be incorporated to consider/calculate the effect of the coin's pores in the specific gravity test. If a high degree of precision is required, the existing conditions must be allowed for. With care a coin could be coated with a wax to seal all pores, but this is undesirable, especially with rare coins. One can simply agitate the water to loosen the air bubbles, et cetera. Technically speaking, the weighing instrument does not show the mass of the samples–the value given in the equations–but rather the weight value for the samples in air. For more precise results, use the weight values obtained after air buoyancy correction. When using the buoyancy method, the immersion level of the pan hanger assembly is affected when the sample is immersed, which produces additional buoyancy. This must also be considered in more precise calculations. In general, density determination procedures require very careful work; it is especially important to make sure the temperature remains constant during testing. With porous material: bulk density relates to the total volume including open and closed pores – true density relates to the solid volume. Most coins have very very tiny pores, which should have a minimal effect upon the outcome. Ancient and very corroded coins may require numerous measurements and averaging to reflect the actual specific gravity.

Bubbles entering the liquid when the sample is immersed can also affect results; bubbles adhering to the test piece will distort the measurement results. This can be reduced with practice, and with certain additives which can assist with breaking the H20's surface tension, and as mentioned, via agitation.



The following link gives the formulas required, and the necessary steps to achieve very accurate specific gravity tests of various materials. These are only useful if you have a laboratory set-up. Home-made kits can work just fine, with care, and can give useful results.

www.dcu.ie/sites/default/files/mechanicalengineering/PDFS/manuals/De-nsityDeterminationManual.pdf

or search Google for Densitydeterminationmanual.pdf

Wikipedia also has some good basic information:

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


Here is a periodic table listed as per descending densities!!

http://www.science.co.il/PTelements.asp?s=Density


courtesy of Mr. Gary Dykes

 

Why though? Certainly one never wants to unknowingly acquire modern forgeries, but it's entirely possible that faux silver siliquae were officially sanctioned by Roman authorities. The Romans had a remarkably sophisticated grasp of metallurgy - the OP article is only one example of the evidence. The biggest cheats are always government officials, because they have the power to cheat with impunity.

I've never met a collector of ancient coins that was interested in the bullion value of his coins, so if a particular specimen is authentic, why would it matter if it was good silver or some sort of silvery alloy? This all begs for an analysis of the siliquae...

 

+1

 

I think it's interesting that it's easier to believe the ancient Romans were debasing silver than by modern governments--as if people today are that much different. After taking that margin of error into account, the XRF data is still compelling--and probably the best data we can hope for. I'll will also be posting data for modern coins in the world forum--check back there for updates.

 

Downloaded. Thanks a lot

 

Lol, I knew one guy who was like that, greedy git.
I assumed that the "Fouree" Siliqua were manufactured by a counterfeiter, but after I read RS's description for the Valens Siliqua last night I kinda agreed with him; the Imperial Gov. were the only ones with that kind of skill and intelligence to pull that off.
This brings many new question to light though: how common are these?, Under which Emperor did the production of them begin under? Were the Legions paid in them(a very dangerous move)? How could the Government tell which coins were real and which were fake when collecting taxes in Siliqua?
This period of time (4th Cent) needs so much more study and analysis.

All this talk of Siliquae makes me want to get some!

 

Btw, I carefully calculated the specific gravity using the paper's data for the fake Valens siliquae--it's 7.32. Compare that to the genuine Theodosius III (fairly typical of genuine siliquae in that chart), where the SG is 10.42. That's a large difference that should be easily detected even with careful SG testing.

 

This link takes you to a page which demonstrates the simple math used to calculate the SG for alloys. There are various methods (as Kurt shared with me his preferred method), this method below uses the "reciprocal" method:

www.tmtpages.com/specific-gravity.htm

Believe me, going out to 5 or even 6 decimal places can increase accuracy!
Also below (again) the link to my 12 page paper on World Coin Alloys and Their Specific Gravities - (which is always available and up to date)

www.Biblical-data.org/GSDykes_specific_gravity_tests.pdf

Gary in Washington

 

Nice work--I see you've made a lot of updates!

 

Yes, I just added 4 more silver coins, .999 and .900 coins of the US. Three of them were not .900!! The 2015 ASE was .999. The other supposedly .900 alloys were 7 to 8 percent below that mint specification! And with your prompting, I did my own math, and corrected a few of the "standard" settings, on the introductory page. The updated version is always available here:

one more time:
www.Biblical-data.org/GSD_specific_gravity_tests.pdf

So Kurt, we have more evidence, of possible debasing. I have more material for future XRF tests. I tested each new Ag coin, 3 or 4 times, everything was at 72 degrees. Time and time again, we are finding low percentages of silver in our US minted coinage!! Older junk silver, and current mintages. Even the 50th Anniversary Kennedy 4 coin set is suspicious, per my testing!