Pigments on Roman Wall Painting and Stucco Fragments from the Monte d’Oro Area (Rome): A Multi-Technique Approach

Round 1

Reviewer 1 Report

The paper by Guglielmi et al is interesting and comprehensive.

Nevertheless, there are several issues that need to be treated prior to its publication. First of all, the text needs to be revised in terms of grammar and syntax (and occasionally spelling as well). Several pertinent points are highlighted below, yet I am sure that a careful reading will reveal more.

In addition, it is apparent that the authors have examined only the surfaces of the samples, which in fact raises a series of limitations: the layered structure of the samples cannot be investigated, therefore the stratigraphy of paint layers and substrates cannot be assessed.

Detailed comments, remarks and suggestions follow.

Introduction

p.2/lines 54-55, “dating approximately … and the 1^st century A.D.”: A few lines above authors mention that the majority of the excavated pottery dated to the 1st & 2nd century, and soon after they highlight the contextless character of the wall paintings & stucco fragments; in fact there is contradiction between the dating of the pottery and the dating provided in lines 54-55. Can you please clarify how this material (: contextless wall painting fragments) was finally dated?

p.2/lines 68-70, “…the chromatic palette … and carbon black”: are all these substances indeed part of the chromatic palette or they represent some of the non-pigment materials as well (e.g. gypsum & chalk)? Moreover, please elaborate on the presumable identification of gypsum (nothing relevant to gypsum appears in the text).

Materials and methods

p.2/line 73, “10 wall painting shards…”: Authors shall definitely mention whether the analyzed samples received any conservation treatment prior to their investigation. Besides, on the basis of the pictures shown in figure 1, it seems that the samples have at least undergone cleaning (: removal of superficial contamination layers).

p.2/lines 77-81, “Style and quality…and contrasting colours”: This part should be better preserved for the introduction.

p.4/Table 1, “and the. indication…”: Please delete redundant full stop.

p.4/line 105, “…a preliminary morphological screening…”: Please clarify whether the samples were observed using the secondary or backscattered electron SEM detector.

p.5/lines 132-133, “The values … measurements” and Table 2: Are the EDS analysis results normalized to 100%?

p.6/lines 141-142, “were quite homogeneous”: it would be very interesting to present a couple of indicative SEM photomicrographs.

p.6/lines 145-146. “Since the painted … the plaster too”: It probably depends on the electron beam spot size and the state of preservation of the paint layers (presence of losses); if the beam was focused only on painting (which seems to be the case) I doubt whether there was any contribution from the underlying plaster. The electrons used for excitation purposes in EDS analysis are of rather low energy therefore do not penetrate deep into the analyzed phases (usually less than 1 micron). On the other hand, the thickness of wall paintings paint layers is often well above 10 microns, therefore it seems rather improbable that these EDS analyses are affected by the substrate(s). Thus I assume that the high Ca content detected in most of the samples reflects intentional addition of a calcareous compound into the various paint layers (slaked lime? calcium carbonate?). Yet the only way to clarify this issue is though the examination of samples cross-sections that would allow for the precise determination of both the paint layers' thickness and composition.

p.6/lines 156-157, “these results … and sulphur”: please consider revising (problematic syntax).

p.6/lines 175-176, “It is evident…doubts on the actual nature of almost all the pigments”: the SEM-EDS analyses would have offer much more valuable results if the authors were able to examine polished cross sections in moderate magnifications (~5,000 to 10,000 times): thus it would be possible to investigate accurately both the elemental composition of individual pigment grains and their morphological characteristics (the latter are of great importance as well). On this view, I suggest that the authors state (probably in section 2-Materials and methods) the reason(s) why they decided not to analyze cross sections (I assume due to strict archaeological legislation/bureaucratic procedures for issuing sampling permits).

p.6/line 182, “no band ascribable … was detected”: this finding strengthens the attribution of the high Ca-contents revealed by EDS analysis to the presence of a calcareous matrix in all pigmented layers, inferring thus employment of the fresco painting technique.

p.6/line 184, “…and evident different…”: I suggest using “evidently”.

p.7/line194, “…that it be basically…”: please revise.

p.7/line 2020, “871 e 712 cm^-1”: please revise.

p.7/line 207, “suggest”: revise to “suggests” (single band).

p.8/lines 212-213, “the results of Raman…samples 1 (figure 3d) and 3 (figure 3e)”: Figures 3e and 3d correspond to FTIR spectra, not Raman. Please revise accordingly.

p.8/line 218, “In sample 2…”: please add a reference to the relevant figure (3b).

p.8/lines 223-225, “The weak peak…appear in the EDS spectrum”: these data imply mixing of cinnabar with red ochre, which to my view is a notable finding that shall be highlighted and commented.

p.10/Figure 5: I think that it would be better to combine figures 5 and 6 into a single one (thus one may readily assess the pertinent FTIR and Raman spectra).

p.10/lines 257-259, “it must be stressed…SEM-EDS techniques”: this passage highlights one of the reasons why it is highly advisable to examine painting samples using the SEM's BSE detector (compounds containing heavy elements such as lead are readily discernible).

p.11/line 271, “overlapping of yellow and red painting”: overlapping of various paintings or simply a mixture of pigments?

p.12/line 302, “are due another time to calcite”: please revise.

p.13/line 314, “the typical chemical composition of Egyptian blue”: these spectra and analyses are not really typical of Egyptian blue: the grains of the latter contain notably more copper and silicon, and significantly less calcium. It is obvious that this inconsistency arises from the fact that the EDS analyses were conducted on the surface of the samples, rather than on individual pigment grains. Therefore I suggest rephrasing to "The SEM-EDS analyses clearly hind towards Egyptian blue" or similar. Also, to my view the EDS analyses results present in Table 2 are sufficient, and the spectra shown in Figure 9 do not add more info.

p.13/lines 317-318, “…were generally prepared with similar plasters”: in fact the paper contains no plaster analysis results. The incorporation of pertinent data (if available) would be a valuable addition to the paper.

p.15/line 346, “the presence of a patina”: What do the authors infer by using the term patina? I think that this word is mostly used to denote alteration products, while here authors are dealing with a superficial gold leaf (that was intentionally adhered onto the substrate by a mordant/glue).

p.16/line 352, “by an illustrious and wealthy person in since”: please revise.

p.16/line 368, “the absence of organic binders”: The absence of organic binders is indeed an important feature of the samples that strongly hints towards the employment of the fresco technique. However, authors shall take into account degradation phenomena that can eventually lead even to the complete degradation (and therefore loss) of organic binders, especially in case of archaeological samples. In this view the fact that calcite is mixed with all the identified pigments must be highlighted and utilized as a further evidence of fresco technique employment (see also a relevant comment above).

p.16/line 368, “to differentiate specific spectral features”: Could you please elaborate on these features and on their contribution towards revealing the painting technique?

p.16/line 373, “totally conservative way”: maybe it would be better to use “non-destructive way”.

p.17/lines 382-382, “the colourimetric investigation”: In order for the colourimetric investigation to give back valuable data it is essential that the analyzed samples are free of accumulation layers. This fact eliminates considerably field applications (e.g. in situ examination of un-cleaned painting fragments).

p.17/line 395, “also different painting layers”: To my view, authors offer no evidence as regards the analytical identification of successive paint layers. In order to explore the presence of overlapping paint layer one may either examine samples' cross-section, or use confocal-based techniques. An alternative approach would be the surface examination of samples that bear partially lost surface layers that reveal their substrates (yet this seems not to be the case here as it is evident by fig. 1).

p.17/lines 398-399, “a few collected painting fragments”: To be accurate, the presence of Au-leaf has been confirmed in the case of a single fragment.

p.17/line 401, “a red ochre layer beyond…”: in the end, the red ochre lays above, below or within the cinnabar layer? To my view the presented data may support only the latter hypothesis.

p.17/lines 402-406, “As we well know…in this study”: This part better fits into the discussion.

p.17, lines 408-409, “the result could…to different roman building”: The mixing of these two pigments in purpose of adjusting the hue/tonality of this specific paint layer shall be considered as an alternative explanation.

Author Response

We have revised the work taking into account the valuable suggestions of the reviewer and we believe that the work has now improved.

Author Response File: Author Response.pdf

Reviewer 2 Report

Many incertainties could be improved. I suggest

Line 38 delete “made”

Line 42 after widespread a word is lacking, perhaps “presence”?

Line 43 fresco (and wherever the term is used): better “wallpainting” that is more general, since in roman painting we have also “mezzo fresco”, mixing of colours, overlayers, metal foil on bole, …

Mezzo fresco is frequently applied on fresco when relief decorations are present (see line 76)

Line 65 “in-field” many authors prefer “in situ”

Line 81 “stamps” are properly moulds as testified by “mouldings” of other lines

Line 93 Table 1. could be compacted in the caption , the samples being all from the same site

Line 122 measurements instead of “measures”

Line 145 “lied” past tense of to lie is “lay”

Line 152 “amount” is not a chemical term, I suggest percentage or concentration

Line 153 “Ba” it should be better the complete word of the element Barium, like all the other ones

Line 165 “antiques” is a term for antiquary, it is better “ancient artifact”

Line 167 “on the base” change to “on the basis” ; the same in line 211

Line 169 “blue and yellow matters [8]”: shades of green in roman paintings were obtained also by mixing Egyptian blue with a pale green, such as green earts.

Line 211 “The spectra of sample 1 (figure 3d) and 3 (figure 3e)” must be referred to Fig. 4d and 4e

Line 219 The attribution of bands to kaolinite could be properly done on the basis of the OH bands at around 3500 cm-1 and of some bands at wavenumber lower than 400 cm-1

Line 224 “a very low quantity of cinnabar” should be described as “low % or concentration of cinnabar, not uniformly distributed in the painting layer ”

Line 227: “the relatve highness” change to” the relative intensity”

Line 236 “positioning” change to wavenumber

Line 237 “silica minerals” change to silicate minerals

Line 250 “peak at 391 cm-1 attributed to yellow ochre”: in the yellow colours the authors never attribute it to a chrystallographic species of FeOOH. In this case Goethite was identified.

Line 259 and 262 measures > measurements

Line 273 and followings: the technique is suitable for pure pigments and is not ideal for mixtures or for low concentration items, as are some of the samples analysed. Red and orange ochres are frequently mixtures and can be better characterized by Raman and FTIR spectroscopies. In Fig. 7 the abscissa is wrongly  reported in wavenumbers

Line 302 “are due another time” > are “ascribable once more”

In Fig.8 the peaks attributed do not correspond to the numbers reported: 1160 should be the weak peak near 1200; 1045 should correspond to the peak with the number 1160; the third peak having 1045 should be 1008 cm-1

Line 312 “Otherwise” could be deleted

Figure 10. Raman spectra of a) …

Line 343 The study of sample 1 was particularly interesting because it showed…

Line 360 with each analytical technique employed

Line 367 commitment > commission or client

Line 368 Gold leaf was laid on a wall or wood or ivory through an armenian bole (yellow or red ochre) by means of a binder. In this case the binder could have been lost or still present below the gold foil. It is to emphasize that the surface here is all red by ochres. Microanalyses could have shown some more details of this gilding.

Line 389 “indicate in which…”  > identify where also an in-depth

Line 408 “could eventually be related…” > might be related to diachronic painting fragments and /or their pertinence to different roman buildings.

Author Response

We have revised the work taking into account the valuable suggestions of the reviewer and we believe that the work has now improved.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper is enough interesting to be published. However, before of accepting it, for publication, the authors should improve their work showing further details about their Raman characterization of the investigated samples.

Indeed, one the most important point stressed in the conclusions by the authors is:
"Furthermore, from the point of view of the analytical approach applied, it has been demonstrated that the portable instrumentation (Raman spectroscopy and reflectance spectroscopy investigations) can successfully be applied for highlighting the nature of the pigment without to invoke more expensive tools. The employment of portable devices might provide a rapid and effective overview directly on vast wall painting giving the opportunity of performing either totally non-destructive analysis (without any sampling from the artwork itself) or at least targeting the sampling in order to do it as fewer as possible".

Also at the end of the discussion's section, it is claimed that:

"It is worth to notice that in case of pigmented surfaces, in particular in the case of great dimension ones, both portable Raman spectroscopy and colourimetric analysis can be employed to carry out a preliminary survey useful to indicate in which areas an in-depth study with destructive techniques could be useful";

"Thanks to the synergic use of both a portable Raman spectrophotometer and a conventional instrument such as a bench tool, the potentiality of these two systems have been exploited. In particular, the use of the portable facility might allow us, in the future, to perform chemical analysis in a totally conservative way, i.e. without the necessity to withdraw some samples. It is
absolutely true that the low spatial resolution achievable with the use of the portable tool equipped with a fibre optic probe might not match that of a Raman spectrometer equipped with a microscope, as proved in several of the performed analyses (for example results obtained on sample 2).
Nevertheless, in this case completely not-destructive analyses can be performed. Furthermore, it must be stressed that in the case of fresco samples the aforementioned portable Raman instrument could be the actual goal, most of all in the hypothesis of its application in the field, i.e. directly on frescoes".

In fact, as underlined by the authors, this aspect is very important for archeological research. However, the comparison between portable Raman instrument's results and laboratory's ones has been not sufficiently demonstrated in the paper.

Indeed, only for the sample 2 the two Raman spectra (portable and laboratory instrument) have been reported together (Figure 4), to allow a direct comparison. If, as claimed, "in the case of fresco samples the aforementioned portable Raman instrument could be the actual goal", why the comparison between the two Raman spectra (785 and 1064 nm) has not been shown always for all the samples?

If these measurements are available, for the authors it should not be a great work to introduce and briefly discuss them in their paper. In this way the reader interested to archeological research could have a clear survey of the potentialities and limits of the portable Raman instrument, compared with the laboratory equipment, when frescoes have to be studied not in laboratory.

If these measurements are not available, some conclusions and statements should be reconsidered, as, for example, that at the end of the introduction:

"Among the analytical techniques implemented during this measurement campaign, it is worth to notice that Raman spectroscopy has been successfully applied by means of both a benchtop device and a more versatile portable Raman spectrophotometer. Furthermore, the comparison of the considerable analytical results achieved from either the two pieces of equipment have been really encouraging in the perspective of an “in-field” application (i.e. directly on wall paintings) of this powerful and totally non-destructive technique".

Indeed, "the comparison of the considerable analytical results achieved from either the two pieces of equipment" implies that both Raman spectra are available for all the samples, or for many of them. Therefore why not show  both Raman spectra in the paper for all the investigated samples? 

In order to facilitate the revision, the authors could  insert in the work a specific section dedicated to the comparison of Raman spectra obtained by the two instruments, to highlight potentialities and limits of the portable equipment.

Author Response

We have revised the work taking into account the valuable suggestions of the reviewer and we believe that the work has now improved.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

personally I am satisfied about the corrections and notes added. 

Reviewer 3 Report

I would have preferred to see the suggested comparison between
the Raman measures obtained by the two instruments. Nevertheless, the changes made to the revised version of the paper
are suitable to make the discussion and conclusions appropriate.
Therefore the paper can be published as it is.