What conservation science found in commercial green earth pigments — and what it means for ours
When Tatiana and I first traveled to visit our Italian earth pigment supplier in 2007, we were starting a business relationship. We visited their factory in Verona, observed their operations, and came away satisfied that we were working with a serious producer. We were not looking for adulteration. It did not occur to us to look. Like most painters, like most paint manufacturers, and — as it later turned out — like museum conservators who had been using the same commercial pigments as reference materials for decades, we assumed that a pigment sold as green earth was green earth.
Our product descriptions said as much: natural mineral pigment. Independent analysis has shown that description does not tell the whole story. This article explains what that analysis found, why it happens, and what we currently believe is in these pigments.
George and Tatiana O’Hanlon in Italy, 2007 — the year Natural Pigments established its relationship with European earth pigment suppliers.
What a Conservation Scientist Found
In 2019, Gregory Dale Smith of the Indianapolis Museum of Art and Heidi Kastenholz of Butler University published a paper in the Journal of Cultural Heritage that began as a response to an unrelated study. Researchers at the University of Malta had purchased Rublev Verona Green Earth from Natural Pigments, used it in a series of crystal engineering experiments intended to improve the pigment’s opacity for conservation inpainting, and characterized it using X-ray fluorescence, infrared spectroscopy, and X-ray diffraction. Their analyses indicated a mixed mineral pigment, and they proceeded accordingly.
Before publishing, Smith contacted us directly to let us know what he and Kastenholz were finding, and asked whether we could supply samples of our other green earth pigments for further analysis. We did, including material from our Armenian sources. We cooperated because we understood this kind of color adjustment to be common practice across the pigment trade, not a problem specific to our supplier or our products.
Smith and Kastenholz looked at the same pigment using Raman microspectroscopy — a technique that can identify specific molecular structures at the particle level. What they found was not what the label said.
Rublev Colours Verona Green Earth under reflected light microscopy at 20× magnification. The dominant warm yellow-ochre matrix consists of mineral earth components. The scattered vivid teal-blue particles are copper phthalocyanine green (PG7), a synthetic organic pigment identified by Raman microspectroscopy. G.D. Smith, Indianapolis Museum of Art.
Raman spectra of Rublev Colours Verona Green Earth archived and purchased samples, dating from 2007, 2009, and 2019, compared against a phthalocyanine green (PG7) reference spectrum. Trace colors correspond to the sample year; peak positions in both Verona lot samples and the PG7 reference confirm the presence of phthalocyanine green pigment. G.D. Smith, Indianapolis Museum of Art.
The bright green crystallites dispersed throughout the pigment were neither celadonite nor glauconite. They were copper phthalocyanine green — PG7, a synthetic organic pigment introduced commercially in the 1930s. Under magnification, PG7 appears as scattered grains within a much larger field of mineral material — a minor presence by volume, not a dominant one. But synthetic organic pigments like PG7 carry far more tinting strength than mineral earths, so a small amount goes a long way toward shaping the color you see. The relationship is similar to a Cabernet Sauvignon containing a smaller portion of Merlot: the dominant material gives the body and structure, while the smaller addition measurably changes the final character. Here, the mineral component gives the pigment its body, particle structure, and much of its handling behavior in oil; the phthalo green addition shifts the color toward a more saturated, consistent green than the mineral alone would produce. Based on this analysis, we believe PG7 is present as a minor but real colorant in Verona Green Earth, as we currently source it.
They also tested other green earths from Natural Pigments’ line, from Kremer Pigments, and from Williamsburg. The full results are in their published paper, alongside the analytical data we are publishing with this article. The findings were broad. Kremer’s Verona Green Earth, labeled “genuine, pure,” contained PG7 in the sample tested. Williamsburg’s Italian Terra Verte, in the sample tested, also contained PG7. Williamsburg’s French Terre Verte, in the sample tested, contained cobalt compounds — PB36 and PG19 — which surprised the Williamsburg formulators, who believed they were using authentic natural green earth and had been told as much by their own European supplier. We mention this not to single anyone out, but because it matters for understanding the scope of the issue: this is not a problem unique to one supplier. In Natural Pigments’ own line, Antica Green Earth from Prun, Italy, was found to contain PG7 as well, mixed with yellow ochre. We believe this combination is present in our Antica Green Earth as we currently source it, and that it accounts for this pigment’s more olive tone relative to Verona. Nicosia Green Earth, from Cyprus, was found to contain cobalt compounds — PB36 and PG7 — and we believe these are present in the Nicosia Green Earth we currently source.
Two products in our line returned clean results. No synthetic colorants were detected in the Tavush Green Earth or Tavush Transparent Green Earth samples analyzed, both of which were sourced from Armenia. Based on those results, we believe these remain genuine mineral green earth pigments.
Why It was Missed For So Long
The analytical methods most commonly used to characterize pigments — XRF, FTIR, and XRD — are not well suited to detecting small quantities of synthetic organic pigments dispersed in a mineral matrix. PG7 produces only a trace copper signal by XRF, and its infrared absorption is masked by the stronger bands from the surrounding minerals. It took Raman microspectroscopy — specifically, the ability to focus on individual bright green particles and analyze them selectively — to identify what was actually there.
The same gap explains a related problem Smith documents: the widely used Raman reference library compiled by Bell and colleagues in 1997 includes a terre verte entry whose spectrum is, in fact, a mixture of anatase, phthalocyanine green, gypsum, and calcite — not genuine green earth. The library was built partly from commercial pigment samples; if those samples were already affected by the same problem, the reference spectra were compromised at the source. Conservation researchers have been working with that reference for decades without knowing it.
Why It Happens At All
This is not a new problem, and it did not begin with synthetic organic chemistry. Pigment making has never been primarily about preserving a specific mineral’s identity; it has been about producing reliable color. Historically, colormen blended, extended, and substituted materials constantly — toning one earth with another, adding lake pigments to deepen a hue, mixing sources to standardize a batch — because the customer wanted a consistent, useful color, not a certified mineral specimen. The term "green earth" or "terre verte" has always described a color and a general category of material rather than a single, fixed mineral composition. What has changed is not the underlying practice but the tools available to detect it, and the expectation — mostly recent, mostly driven by conservation science and historical reconstruction painters — that the name on the label should correspond to a specific, verifiable mineral.
The historical green earth source at Verona was long considered among the finest terre verte in the world, mined commercially through the first half of the twentieth century. Access declined over subsequent decades; an earthquake in 1922 disrupted the most productive deposits. What replaced large-scale mining were smaller, scattered excavations — geologically inferior material from fragmented sources, difficult even for our supplier to trace back to a single origin.
The largest-volume markets for iron oxide earth pigments — architectural coatings, mulch coloring, and similar industrial uses — are indifferent to the distinction between natural and synthetic material. Those markets require consistency, which mine-sourced material cannot reliably guarantee on its own. A synthetic correction that brings variable batches to a uniform color strength serves as a quality improvement from a buyer’s perspective and is not disclosed further down the supply chain to artists’ colormen, because for an industrial customer there is nothing to disclose. The material simply meets the specification.
This is, as far as we can tell, standard practice across the trade, not a failure specific to any one supplier. Williamsburg’s Sarah Sands described a related dynamic in a Just Paint article on undisclosed substitution in ochres, siennas, and umbers — including a later, candid correction of her own where she acknowledged that even major suppliers use small synthetic additions to adjust natural earth pigments for color consistency, a practice she noted has been common in the trade for a long time. That context explains how these adjustments enter the supply chain. It does not mean painters should be left without disclosure when a pigment is sold under a natural-earth name. We are not aware of any supplier in this industry, including conservation-focused ones, that tests and discloses synthetic content in green earth pigments as a matter of course. We have chosen to do so where we know or believe it to be the case.
George O’Hanlon at a pigment excavation site in Saint-Amand-en-Puisaye, Burgundy, France, May 2023, during a trip to visit earth pigment suppliers across Europe. Sites like this one illustrate the small-scale, fragmented sourcing that characterizes much of the natural earth pigment trade today.
What We Currently Believe is in Our Green Earth Pigments
Verona Green Earth (Mt. Baldo, Italy): We believe this pigment is based on genuine mineral celadonite and glauconite, with copper phthalocyanine green (PG7) present as a minor addition that shapes the final color, as indicated by Raman microspectroscopy.
Antica Green Earth (Prun, Italy): We believe this pigment contains PG7 mixed with yellow ochre, alongside genuine mineral glauconite, based on the same analysis. This mixture is the source of Antica’s more olive tone relative to Verona.
Nicosia Green Earth (Cyprus): We believe this pigment contains cobalt blue-greenish (PB36) and PG7, alongside genuine mineral glauconite, based on the same analysis.
Tavush Green Earth and Tavush Transparent Green Earth (Armenia): No synthetic colorants were detected. We believe these remain genuine, unadulterated mineral pigments.
We are reporting what we currently believe to be true, based on the best analysis available to us. We do not test every lot we sell, and we do not claim to. Few, if any, suppliers in this trade do. What we can offer is this: when we know something, or are reasonably confident of it, we will say so.
What This Means for Painters Using These Pigments
The synthetic pigments identified are generally regarded as durable artists’ materials. Phthalocyanine green (PG7) carries an ASTM Lightfastness rating of I, the same top rating given to genuine celadonite and glauconite, and it is chemically stable and non-reactive in oil. The cobalt compounds found in Nicosia Green Earth carry the same lightfastness rating. Their presence does not indicate a known inherent fading or paint-film instability problem. The issue here is one of composition and accuracy of description, not a known conservation hazard.
What it does change is the accuracy of any claim about mineral authenticity. A painter using these pigments for historical reconstruction, conservation retouching, or because they specifically wanted the optical properties of genuine celadonite or glauconite should know that this pigment includes a synthetic color adjustment, since that may matter for that kind of work.
PG7 has considerably higher tinting strength than celadonite or glauconite, so even a minor addition can noticeably shift the color toward a more saturated, consistent green than the mineral alone would produce. Because PG7 is present only as scattered grains within the mineral matrix, we would not expect it to meaningfully change this pigment’s oil absorption, rheology, or behavior in mixtures — those properties are governed by the mineral component, which makes up the bulk of the material. The difference to expect is in color, not handling.
If the specific colorant content of a green earth pigment is critical to your work — for conservation retouching, historical reconstruction, or any application where you need to know precisely what you are using — contact us. We will do what we can to help you find the right material for that purpose, whether that means Tavush Green Earth, Tavush Transparent Green Earth, or further information about a specific product.
The Broader Picture
Smith’s paper examined fifteen commercial green earth products from four suppliers. Eight of those fifteen contained synthetic colorants not disclosed on their labels. This tells us that the issue is not confined to a single supplier or country of origin. It is, as far as the evidence shows, how this part of the trade generally operates: pigment names like “green earth” function as a color and category description more than a guarantee of mineral content, and most suppliers do not test for or disclose the difference.
For painters, the practical implication is straightforward: a label that says “natural green earth” or names a geographic source is not, by itself, proof of what is in the tube or jar. We are telling you what we currently believe is in our pigments, based on the best analysis available to us. We believe this level of disclosure remains uncommon in the artist pigment trade, including among suppliers serving conservation-minded customers. We think that’s worth knowing, even if it isn’t the norm.
The two products in our line confirmed clean in testing, Tavush Green Earth and Tavush Transparent Green Earth, remain in the line. Their lower tinting strength compared to the other green earths in our catalog is not a shortcoming; it is one of the visual signatures of genuine, mineral celadonite-type material, and a reminder of how historical green earth actually behaves in use.
Questions We Expect Painters to Ask
Is this pigment unsafe to use? The synthetic pigments identified do not indicate a known paint-film stability problem or an unusual hazard beyond normal dry pigment precautions. As with all dry pigments, avoid inhaling dust. For Nicosia Green Earth, the cobalt-containing pigments identified make dust control especially important.
Will paintings I have already made with these pigments fade or change? PG7 and the cobalt compounds identified all carry top ASTM lightfastness ratings. We have no evidence suggesting a paint-film stability problem from these synthetic additions.
Why didn't Natural Pigments know sooner? The standard analytical methods used to characterize commercial pigments are not well-suited to detecting small quantities of synthetic organic colorants in a mineral matrix. That gap is common across the trade, not specific to us.
Is this just a Natural Pigments problem? No. The available evidence, including testing of products from other suppliers, indicates this is common practice across the green earth trade. We are disclosing what we currently believe about our own products because we think painters are better served knowing it, not because we believe we are unusual in this respect.
Is Tavush Green Earth genuine? No synthetic colorants were detected in the samples analyzed, and we believe it to be a genuine mineral pigment.
What if the colorant content matters for my work? Contact us. If you need to know precisely what’s in the pigment you’re using, we’ll help you find the right material, which may mean Tavush Green Earth or Tavush Transparent Green Earth.
References
Smith, G.D. and Kastenholz, H. (2019). "A note on the use of crystal engineering to increase the opacity of terra verde inpainting pigment: Adulteration of green earths." Journal of Cultural Heritage. Available online October 17, 2019. DOI: 10.1016/j.culher.2019.10.004
Micallef, D., Spiteri, R., Baisch, U., and Vella-Zarb, L. (2019). “Terra verde: overcoming the problem of transparency by crystal engineering.” Journal of Cultural Heritage 38, 1–7.
Fanost, A., Gimat, A., de Viguerie, L., et al. (2020). “Revisiting the identification of commercial and historical green earth pigments.” Colloids and Surfaces A 584, 124035.
Uebele, C.L. (1913). Paint Making and Color Grinding: A Practical Treatise for Paint Manufacturers and Factory Managers. Chicago.
Auf der Mauer, M. (2024). “Availability Issues with Natural Earth Pigments.” Just Paint, Golden Artist Colors. justpaint.org.
Sands, S. (2017). “Sorry, Wrong Umber — Part I.” Just Paint, Golden Artist Colors. justpaint.org.
Grissom, C.A. (2012). “Green Earth.” In R.L. Feller (ed.), Artists’ Pigments: A Handbook of Their History and Characteristics, Vol. 1. Archetype Publications, London, pp. 141–168.
Cáceres-Rivero, C., Tupa-Quispe, A.L., Torres-Casas, R., and Bedregal, P. (2022). “Identification of adulterants in artistic earth pigments using a multi-technique approach.” Results in Chemistry 4, 100561. DOI: 10.1016/j.rechem.2022.10056
George O’Hanlon is the technical director of Natural Pigments. The full Raman analysis data for each affected product, including spectral comparisons, is published on the respective pigment product pages linked above.
