Portable X-Ray Fluorescence as a Proxy for Aerinite in Pigments of Medieval Alto Aragón Cultural Heritage

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This article presents an application of portable X-ray fluorescence spectroscopy (pXRF) for the noninvasive screening of aerinite in medieval artworks from Alto Aragón, Spain. The authors compare elemental compositions of aerinite reference samples from the La Soriana quarry with those of lapis lazuli and numerous heritage objects, applying multivariate analysis to establish compositional fingerprints. The application of pXRF to aerinite is novel for the Alto Aragón region and contributes to heritage science practice; however, the fundamental analytical principle (elemental screening via pXRF combined with PCA) is well established. The contribution is specialized and regionally significant rather than methodologically groundbreaking. The study acknowledges the limitations of pXRF for definitive pigment identification and frames the technique as a screening tool requiring subsequent confirmation by micro-Raman or micro-XRD spectroscopy. The manuscript is well motivated and methodologically transparent; however, before it is accepted, the authors should clarify several points:

• Can you confirm the X-ray tube voltage/current range and detector type (Si-drift or PIN) for the Niton XL3t GOLDD in mining mode? This information would improve metrological traceability.
• For the reference materials (Tables S1–S2), how many independent measurement spots were acquired for each sample? Do the reported values represent mean ± SD, or single measurements? For key reference materials, please add columns reporting the number of measurement replicates (n), and the standard deviation (SD) or range for major elements. For example, modify Table S1 to show “Fe 4.8 ± 0.3 wt%, n = 3” or similar.
• Please provide representative raw XRF spectra (counts vs. energy) in the Supplementary Materials; even 2–3 examples would substantially enhance transparency.
• Line 123: Clarify “live acquisition times exceeding 120 s” — does this include or exclude dead-time correction?
• Table S1–S7 captions: Include column descriptions (e.g., “elemental concentrations in wt%, as reported by Niton XL3t GOLDD software; values below LOD reported as <LOD”).
• Statistical tests of PCA group separation (e.g., permutation tests, MANOVA) are not explicitly performed. For heritage-science screening, visual interpretation of PCA plots is widely accepted; however, formal statistical testing would strengthen claims of significant separation.
• The intermediate PCA positions of some heritage paint blues (e.g., Chrismon, Portaspana panels) between aerinite and lapis lazuli reference clusters are attributed to “mixture complexity” (paint containing aerinite + lead white + possible contributions from underlying layers). While this explanation is reasonable, the text could more explicitly rule out the alternative hypothesis that these intermediate positions reflect genuine lapis lazuli admixture. The authors address this indirectly (by noting the absence of diagnostic Na–S signatures), but an explicit discussion of how one would distinguish mixture complexity from minor lapis-lazuli additions would strengthen the interpretation.
• Providing measurement spot sizes, approximate beam penetration depths, or a discussion of how multiple layers contribute to the integrated signal would improve understanding of the spatial and stratigraphic information content.
• The claim that measurements were “carefully located on relatively homogeneous color areas” is qualitative; the authors should clarify whether homogeneity was assessed visually or by other means.

I understand that some of these clarifications may be somewhat excessive and not strictly required in heritage science, but the manuscript was submitted to the Spectroscopy Journal, and therefore it should pay greater attention to the analytical aspects.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Although the study presents elements of interest and some potential for a broader contextualisation of the use of aerinite in cultural heritage, it nevertheless exhibits significant methodological shortcomings.

The use of raw data produced by the Niton XL3t GOLDD+ spectrometer, without any subsequent processing, does not allow quantitatively reliable results to be obtained, thereby rendering the entire development of the proposed correlations largely conjectural.

The presence of instrumental nickel peaks, characteristic of this model and, more generally, of most handheld devices, has not been adequately considered. The intensity of this signal varies according to the mass of the matrix and, without appropriate data processing (for example, through fitting software that accounts for matrix effects), a reliable quantification of nickel cannot be achieved. It should also be noted that the low values reported (0.5–1 wt%) fall below the experimental detection limit for this element with this instrument.

Additional contributions arising from instrumental interferences occur in proximity to the characteristic energies of chromium and vanadium. These interferences, often mistaken for those elements, are not discriminated by the device’s internal software and require the use of external analytical tools based on peak-deconvolution algorithms, such as PyMCA.

The quantitative discrimination between lead and arsenic is likewise not handled accurately by the Thermo Scientific software and should instead be addressed through appropriate signal processing.

The subsequent normalisation of the data employed in the multivariate analyses further indicates an incomplete understanding of the analytical process involved. XRF spectroscopy provides an absolute quantification of the matrix: the BAL value represents the sum of the fractions corresponding to elements too light to be detected by the instrument. In this context, post-hoc normalisation is devoid of meaning. To minimise the contribution of BAL, it would have been sufficient to express the detected elements in the form of their oxides, as is standard practice when dealing with mineralogical materials or light glass matrices.

In light of these considerations, the study, in its current form, does not meet the minimum requirements for scientific publication on this topic.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf