FACSS SciX 2025: Ochre Source Discrimination for Archaeological Provenience Studies by “Full Scan” GDMS

Abstract:

Provenience analysis (i.e., sourcing) of archaeological ochre pigments can shed light on ancient mineral resource exploitation practices, transport and trade networks, and the use and reuse of rock art sites over time.  As a prerequisite to such analyses, researchers must demonstrate that the Provenience Postulate is upheld.  The postulate states that each source must be compositionally unique compared to every other applicable source.  Put another way, inter-source compositional variability must be greater than intra-source variability.  The most common way to discriminate among sources and then assign archaeological ochre to its origin is elemental fingerprinting paired with multivariate statistical analysis.  Measuring as many elements as possible, with the greatest sensitivity possible, provides the most options for identifying a combination of variables that can uphold the postulate.  This work proposes that direct current Glow Discharge Mass Spectrometry (GDMS) with a plasma-supporting holding cathode is the optimal technique for this application. 

Building on prior work that developed Relative Sensitivity Factors (RSFs) for GDMS data reduction specific to non-conductive oxide materials like ochre, the research presented here analyzed three southern African ochre sources of archaeological interest.  Using AMETEK Nu Instruments Ltd. Astrum GDMS instruments, we performed a “full scan” of trace elements yielding conservative reporting limits ranging from 5 ppm to 0.01 ppm.  57 out of 70 measured elements have reporting limits < 0.5 ppm.  O, Si, and Fe (matrix elements), Ta (used to make the holding electrode for the non-conductive sample), and Au (interfered by 181Ta16O) were excluded.  S and the halogens F, Cl, Br, and I, which are not measurable or are poorly measurable by ICP-MS, were included in the GDMS full scan.  An additional advantage over solution ICP-MS is the ability to analyze a solid sample of just ~10 mg with no preparation beyond drying and grinding to a fine powder.  One notable finding is wide variation in Ce anomalies between the three sources, ranging from a strong negative anomaly, to virtually no anomaly, to a moderate positive anomaly.  This makes the calculated Ce anomaly a potentially informative derived variable for source discrimination, additional to the trace element concentrations themselves.