Assessing Laser Cleaning of a Limestone Monument by Fiber Optics Reflectance Spectroscopy (FORS) and Visible and Near-Infrared (VNIR) Hyperspectral Imaging (HSI)

Round 1

Reviewer 1 Report

Manuscript Minerals-966097 aims to investigate FORS and VNIR-HIS as “monitoring” techniques of laser cleaning.

The paper can be enhanced as regards its research and English content and must be significantly revised prior to its publication.

Important issues to reconsider are:

1. 1)The term “monitoring” which is misused; all experiments were performed off-line to the cleaning procedure and thus the term “monitoring” is rather misleading. Alternative terms recommended are “evaluation”, “assessment”, “validation”, terms that refer to systematic studies to characterize the cleaning level and result, but they do not include the concept of on-line control which is clearly implied by the term “monitoring”. Thus the term “monitoring” must be revised through the manuscript including the title and the keywords.
2. Another key issue refers to “the details of the cleaning procedure”. The authors must provide the details of their irradiation tests as they are of importance e.g. F-values, number of pulses, wetting –if it was applied-, and so on. As an example one could say that “treatment b” is effective because the total F applied is higher than the F applied for “treatment a” and in the condition that areas irradiated using the “treatment a” -only with 1064nm- at higher F values (equal to F₍₁₀₆₄₎ +F₍₅₃₂₎) the result would be equivalent. Actually in several areas treated with “treatment b” one could say that excessive cleaning was performed and thus the “cleaning degree” discussed within the paper is debateable. So it is very important to give analytical information, by adding to the manuscript a relevant table, and also specify the areas that were under-cleaned, over-cleaned and ideally treated.
3. Figure 1b must be revised as it is rather dark, not focused (the actual details of laser cleaning, which is a very important aspect of this paper, cannot be read easily), while there is a lot of dark background image that does not contribute to the figure. As already mentioned the detailed laser cleaning parameters must be given and a photo with higher resolution must be provided. Similar applies to Figures 3a and 6b where it seems that the black background was added digitally (as one can see from the top of the image, above the stone surface …)
4. The authors must further discuss and argue on “which is the information revealed using the two analytical techniques”. Figures 5 and 6 do not clearly indicate which is the additional knowledge that was revealed through this study. The question to be answered is "Why a conservator would follow this assessment methodology, given that the studied areas are clearly discernible also through visible imaging?" The authors must indicate which is the different information shown in Figure 6a when compared to the info offered by Figure 6b. This is a very important point and must be well discussed. Possibly this methodology works better in another application of cleaning but in this specific case/sample it is not clear which is the additional info revealed while the concept of monitoring cannot be discussed.
   Specifically in the text: “ As the HSI spectra were qualitatively comparable with those measured with FORS, both Vis and NIR data cubes were processed by PCA which enabled to extract new images and maps that allowed to visualize the materials distributions and details on the surface that could not be detected by naked eye.” [lines 243-6] the authors are encouraged to explain further and in details which are the materials that cannot be detected by naked eye.
5. [lines 202-204]. The approach given to explain the yellowing is rather superficial and must be revised so to properly address (briefly, in one sentence) the yellowing reason. Yellowing is not caused due to rapid temperature increase as this would also happened upon 532nm laser ablation. Furthermore, the 532nm does not remove the “dust” particles but instead the remaining gypsum matrix. This paragraph must be significantly revised to properly discuss the actual situation.
6. [lines 208-209] Could the authors explain how they “assess the cleaning accuracy”? How do they define the “cleaning accuracy” and how this is assessed (qualitatively or quantitatively) through this paper?
7. [lines 214-225] why 13 sets of FORS spectra were chosen? Which are the four groups? Please describe them. What does the term “bulk” mentioned in the figure 2 caption refers to; crust or un-weathered stone?
8. A reference spectrum from gypsum or clear indications to the gypsum peaks will enhance Figure 2
9. Discussion as regards the contribution of crust remains and the micro-relief that can be seen on this stone fragment, which appear not cleaned upon certain laser cleaned areas (i.e. 6 or 11) will benefit this manuscript. How do they affect your measurements?
10. [lines 296-298] “… clearly demonstrated the efficiency of FORS-HSI to monitor the effects of laser cleaning procedures, allowing to choose the best parameters for cleaning by performing additional tests to compare the effect of ultraviolet and infrared Q-switched pulses…”
    How the images enabled the choice of the best cleaning parameters? "Ultraviolet"??
11. [lines 299-303] “The results obtained in this work confirmed those obtained in previous studies [36,42] in that the black crust deposit consisted mainly of cryptocrystalline gypsum mixed with calcite combined with quartz silt and soot and dust deposits. The formation of a gypsum crust and related stone surface modification could be attributed to particulate (dust) deposition especially in areas sheltered from rain and rain-wash, as it is the case of the monument stone fragment studied in this work”.
    Could you please revise this sentence and specify how the images resulted from your research confirmed the black crust composition (how did you identify the quartz, the soot and the dust deposits?] and how you reached the result that dust deposition in sheltered areas caused the gypsum crust?
12. The introductive section is long and rather rambling. All the info on limestone, its weathering and patinas, is rather not lengthy for the concept of the paper. A considerable revision is recommended to keep only the information that is important for the discussion of this paper. Also line 69 [portable..?] necessitates revision, while there are more than the 2 studies [Ref 34-35 in lines 93-96,] that have been published as regards assessment of laser cleaning that could help the authors to further argument on their work.
13. [lines 124-126] The authors could further give information on the black crust. Is there any biological agent within the crust? Judging by the figures one could say that biological species may also be part of this crust.
14. Author contribution section must be filled properly.

Author Response

Responses to Reviewer 1 Comments

Manuscript Minerals-966097 aims to investigate FORS and VNIR-HIS as “monitoring” techniques of laser cleaning.

The paper can be enhanced as regards its research and English content and must be significantly revised prior to its publication.

Important issues to reconsider are:

Point 1: The term “monitoring” which is misused; all experiments were performed off-line to the cleaning procedure and thus the term “monitoring” is rather misleading. Alternative terms recommended are “evaluation”, “assessment”, “validation”, terms that refer to systematic studies to characterize the cleaning level and result, but they do not include the concept of on-line control which is clearly implied by the term “monitoring”. Thus the term “monitoring” must be revised through the manuscript including the title and the keywords.

Response 1: According to the Reviewer’s suggestion, the term “monitoring” has been replaced in the title and the entire text.

Point 2: Another key issue refers to “the details of the cleaning procedure”. The authors must provide the details of their irradiation tests as they are of importance e.g. F-values, number of pulses, wetting –if it was applied-, and so on. As an example one could say that “treatment b” is effective because the total F applied is higher than the F applied for “treatment a” and in the condition that areas irradiated using the “treatment a” -only with 1064nm- at higher F values (equal to F₍₁₀₆₄₎ +F₍₅₃₂₎) the result would be equivalent. Actually in several areas treated with “treatment b” one could say that excessive cleaning was performed and thus the “cleaning degree” discussed within the paper is debateable. So it is very important to give analytical information, by adding to the manuscript a relevant table, and also specify the areas that were under-cleaned, over-cleaned and ideally treated.

Response 2: As requested laser parameters and more details on the cleaning procedure have been added in the revised version of the manuscript in subsection 2.2. Laser cleaning procedure. F values used were: 4 J/cm² at 1064 nm with a pulse energy of 800 mJ, and 6 J/cm² at 532 nm with a pulse energy of 400mJ. The number of pulses used was 300 pulses/min. No wetting procedure was applied to the stone surface.

Regarding the laser treatments, treatment b is more effective than treatment a only because two different wavelengths with different optical properties were used. The effectiveness of the 532 nm is not due to a thermic interaction but to an optical interaction, which in turn becomes useful once the 1064 nm wavelength starts to operate. The differently cleaned areas were clarified in the revised version.

Point 3: Figure 1b must be revised as it is rather dark, not focused (the actual details of laser cleaning, which is a very important aspect of this paper, cannot be read easily), while there is a lot of dark background image that does not contribute to the figure. As already mentioned the detailed laser cleaning parameters must be given and a photo with higher resolution must be provided. Similar applies to Figures 3a and 6b where it seems that the black background was added digitally (as one can see from the top of the image, above the stone surface …)

Response 3: Figure 1b has been replaced by a higher quality photo.

With regards to the initial Figures 3a and 6b, we explained in the text that these images are not true photos, but the result of processed HSI images that were used to extract the classification map (initial Fig. 6a) and the corresponding RGB image, which allow to visualise the corresponding areas on the true sample. The colour of background cannot be corrected as these images are the results of HSI acquisition that was made in a dark room, as required, by scanning the illuminated strip on the target. Furthermore, HSI measurements were performed by optimising the S/N ratio of reflectance signals, which not necessarily result in a focused RGB reconstructed image.

For these reasons, we believe that the initial Figures 3a and 6b (now 4a and 7b) should be kept as they are, being the results of HSI data elaboration, whose aim was not to provide high quality RGB. The captions of new Figures 4a and 7b have been revised to illustrate these aspects.

Point 4: The authors must further discuss and argue on “which is the information revealed using the two analytical techniques”. Figures 5 and 6 do not clearly indicate which is the additional knowledge that was revealed through this study. The question to be answered is "Why a conservator would follow this assessment methodology, given that the studied areas are clearly discernible also through visible imaging?"

The authors must indicate which is the different information shown in Figure 6a when compared to the info offered by Figure 6b. This is a very important point and must be well discussed. Possibly this methodology works better in another application of cleaning but in this specific case/sample it is not clear which is the additional info revealed while the concept of monitoring cannot be discussed. Specifically in the text: “ As the HSI spectra were qualitatively comparable with those measured with FORS, both Vis and NIR data cubes were processed by PCA which enabled to extract new images and maps that allowed to visualize the materials distributions and details on the surface that could not be detected by naked eye.” [lines 243-6] the authors are encouraged to explain further and in details which are the materials that cannot be detected by naked eye.

Response 4: The authors thank the reviewer for raising this crucial question and acknowledge that this point should be better clarified in the text.

The text in initial lines 243-6 has been reworded and clarified in the revised version (lines 537-547) with a clearer interpretation of the maps shown in new Figs. 6 and 7 and their capacity to assess the effectiveness of the laser treatment in removing gypsum. The soundness of visual assessment should be proved by compositional data as those provided in this work by NIR, and hence by elaborated maps. Actually, the correspondence between the visible colour of the cleaned surface and the absence of gypsum, or a residual presence of it, is not evident visually, whereas the methodology proposed enables to visualise it.

Point 5: [lines 202-204]. The approach given to explain the yellowing is rather superficial and must be revised so to properly address (briefly, in one sentence) the yellowing reason. Yellowing is not caused due to rapid temperature increase as this would also happened upon 532nm laser ablation. Furthermore, the 532nm does not remove the “dust” particles but instead the remaining gypsum matrix. This paragraph must be significantly revised to properly discuss the actual situation.

Response 5: The yellowing effect generally appears more pronounced at 1064 nm, and less or non-existent at the second (532 nm) and third harmonics. Thus, yellowing can be ascribed to thermal effects. In the case of stones, yellowing was ascribed to thermal changes in iron compounds within the stone [Siano et al., Journal of Cultural Heritage 1 (2000) S47–S53; Vergès-Belmin, Dignard, Journal of Cultural Heritage 4 (2003) 238s–244s].

Point 6: [lines 208-209] Could the authors explain how they “assess the cleaning accuracy”? How do they define the “cleaning accuracy” and how this is assessed (qualitatively or quantitatively) through this paper?

Response 6: As referred in section 3. Results and Discussion, the methodology proposed enabled an immediate visualization of the areas of the surface where gypsum has been totally and uniformly removed from those apparently cleaned (without black crusts), but still presenting gypsum residues. The term “qualitatively” has been added to specify that the procedure was not proposed as a quantitative method.

In our case of study, the officials of the Authority of Fine Arts and Landscape for the Provinces of Bari, Foggia Barletta–Andria–Trani prefer to evaluate the effectiveness of cleaning by using the less invasive effect achieved (at 1064 nm) with a color/aspect of cleaned surface as far as possible similar to the original limestone (yellow-brown). Thus, the “cleaning accuracy” concept is variable and dictated by the preference of the relevant officials. The combined use of these techniques can help to evaluate better the preferred degree of cleaning.

Point 7: [lines 214-225] why 13 sets of FORS spectra were chosen? Which are the four groups? Please describe them. What does the term “bulk” mentioned in the figure 2 caption refers to; crust or un-weathered stone?

Response 7: The text has been revised extensively attempting to clarify the aim of the preliminary FORS analysis, and the criteria followed to group the spectra. The caption of initial Fig. 2 (now Fig. 3) has been expanded with a better description of the four types of spectra.

The term “bulk” refers to the unaltered stone, as clarified in the revised text.

Point 8: A reference spectrum from gypsum or clear indications to the gypsum peaks will enhance Figure 2

Response 8: The reference spectrum of gypsum has been added as a new Fig. 3b.

Point 9: Discussion as regards the contribution of crust remains and the micro-relief that can be seen on this stone fragment, which appear not cleaned upon certain laser cleaned areas (i.e. 6 or 11) will benefit this manuscript. How do they affect your measurements?

Response 9: Black crust residues appear at the boundaries of laser cleaned areas, a feature that is revealed by FORS better than by visual inspection, and clearly identify the laser cleaning homogeneity. If the gypsum crust was totally removed, part of the original stone could be destroyed with possible damage to painting or decoration. Further, the areas not perfectly cleaned are clearly apparent in the false colour maps that show less uniform colours in correspondence of these zones.

Point 10: [lines 296-298] “… clearly demonstrated the efficiency of FORS-HSI to monitor the effects of laser cleaning procedures, allowing to choose the best parameters for cleaning by performing additional tests to compare the effect of ultraviolet and infrared Q-switched pulses…” How the images enabled the choice of the best cleaning parameters? "Ultraviolet"??

Response 10: The sentence has been modified according to the Reviewer’s comment. However, the FORS-HSI images help to distinguish the degree of cleaning achieved and enable to choose the best laser cleaning procedure (see initial Figs. 5 and 6). “Ultraviolet” is not a matter of this paper, it was cited as a general option.

Point 11: [lines 299-303] “The results obtained in this work confirmed those obtained in previous studies [36,42] in that the black crust deposit consisted mainly of cryptocrystalline gypsum mixed with calcite combined with quartz silt and soot and dust deposits. The formation of a gypsum crust and related stone surface modification could be attributed to particulate (dust) deposition especially in areas sheltered from rain and rain-wash, as it is the case of the monument stone fragment studied in this work”. Could you please revise this sentence and specify how the images resulted from your research confirmed the black crust composition (how did you identify the quartz, the soot and the dust deposits?] and how you reached the result that dust deposition in sheltered areas caused the gypsum crust?

Response 11: The sentence has been revised.

Point 12: The introductive section is long and rather rambling. All the info on limestone, its weathering and patinas, is rather not lengthy for the concept of the paper. A considerable revision is recommended to keep only the information that is important for the discussion of this paper. Also line 69 [portable..?] necessitates revision, while there are more than the 2 studies [Ref 34-35 in lines 93-96,] that have been published as regards assessment of laser cleaning that could help the authors to further argument on their work.

Response 12: According to the Reviewer’s suggestion the Introduction section has been shortened and revised extensively. Only References 34 and 35 (now 33 and 34) have evaluated laser cleaning by HSI.

Point 13: [lines 124-126] The authors could further give information on the black crust. Is there any biological agent within the crust? Judging by the figures one could say that biological species may also be part of this crust.

Response 13: The black crusts nature is well known and generally ascribed to a sulfation reaction (SO₂ reacts with calcite to form gypsum (CaSO₄∙2H₂O)) and the entrapment of particles (especially soot), causing its blackening [Camuffo et al. Water, Air, and Soil Pollution 19 (1983) 351-359; Sabbioni and Zappia, Science of The Total Environment 126 (1992) 35-48; La Russa et al. Science of The Total Environment 593-594 (2017) 297-309]. Further, sulphated crusts were shown to be composed of several layers, i.e. an upper dark layer containing several atmospheric particles, such as fly ash and soot, and an internal layer that is less opaque and composed almost exclusively of gypsum [Weber et al. Materials Characterization 58 (2007) 1042-1051]. Although, it can be expected that in the upper layer some micro-organisms are present, the identification of biological species was not in the aim of this work.

Point 14: Author contribution section must be filled properly.

Response 14: Done.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper proposed a potential method to monitor and control the laser cleaning process on stone samples. It is well organized and explained. The PCA and SAM data treatments are well adapted in this application. However, some points have to be clarified. 

1 In the section 2.2, how the treatments a and b have been done? Such as the number of laser shot at one spot, the energy of 1064 nm laser for each case is the same? 

2 In the section 2.3, the fiber is taken in the hand at measurement. How the authors kept the same distance between the output of the fiber and the surface of sample, and the same incident angle? Because these two elements are crucial for a reflectance measurement. 

3 In the section 2.3 and 2.4, for the measurement hyper spectral imaging, how much time it took for acquiring one image and also for the data treatment? 

4 In the Figure 2, b and c have a different peak/valley near 1400 nm, and b and d have the same peak in this area although the authors told the b was well cleaned. 

Generally, the authors need to check again the English spelling: in the title and line 11, it should be "fiber optics"; line 37 "natural ageing" should be aging; line 138 "a whither surface" should be "whiter" I gesse. 

Author Response

Responses to Reviewer 2 Comments

The paper proposed a potential method to monitor and control the laser cleaning process on stone samples. It is well organized and explained. The PCA and SAM data treatments are well adapted in this application. However, some points have to be clarified. 

Point 1: In the section 2.2, how the treatments a and b have been done? Such as the number of laser shot at one spot, the energy of 1064 nm laser for each case is the same? 

Response 1: As requested more details on laser parameters adopted for the cleaning procedure have been added in the revised version in the subsection 2.2. Laser cleaning procedure. The number of laser pulses used was 300 pulses/min; a pulse energy of 800 mJ was used at 1064 nm and of 400 mJ at 532 nm.

Point 2: In the section 2.3, the fiber is taken in the hand at measurement. How the authors kept the same distance between the output of the fiber and the surface of sample, and the same incident angle? Because these two elements are crucial for a reflectance measurement. 

Response 2: We thank the Reviewer for this comment. We added further details in the description of experimental procedure [lines 305-309]. As clarified in the revised text, the probe-head was used in the contact mode according to a procedure that ensured the reproducibility of repositioning. The FORS technique is often performed with hand-held contact probes designed to be used in-field, and the procedure adopted is reproducible and well-established.

Point 3: In the section 2.3 and 2.4, for the measurement hyper spectral imaging, how much time it took for acquiring one image and also for the data treatment? 

Response 3: Technical specifications on the scan speed and data-sets typical dimensions have been added in the revised text, so that the corresponding times can be assessed. In the specific case, times were very fast due to the small dimensions of the target. The acquisition took few minutes, whereas data-processing times depended on the type of algorithm, PC specifications, etc. In the present case, the HSI data sets were of order of few GB, thus data-processing was fast as well (few minutes to run PCA).

Point 4: In the Figure 2, b and c have a different peak/valley near 1400 nm, and b and d have the same peak in this area although the authors told the b was well cleaned. 

Response 4: The comments to initial Figure 2 have been expanded and clarified in the revised text. As illustrated, the meaningful spectral markers to assess the presence/absence of gypsum are the triplet bands in the 1450-1550 nm range and the band around 1900 nm. These signals are comparable for spectra b and c. The small signal at 1400 nm cannot be considered as a meaningful compositional marker and has not been discussed.

Point 5: Generally, the authors need to check again the English spelling: in the title and line 11, it should be "fiber optics"; line 37 "natural ageing" should be aging; line 138 "a whither surface" should be "whiter" I gesse. 

Response 5: English spelling has been checked and corrected throughout the entire text.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The revised version of Manuscript Minerals-966097 has undergone a significant revision, but there are still some major issues that have not been approached effectively and to my personal opinion they necessitate a second revision. Specifically:

1. The term “monitoring” has been effectively replaced along the manuscript.
2. The “details of cleaning procedure” have been, partially, updated.
   My suggestion was “to give analytical information, by adding to the manuscript a relevant table, and also specify the areas that were under-cleaned, over-cleaned and ideally treated” as this would enhance the understanding of the examined surfaces. There is some extra information, which to my point needs further update and also there is a kind of mapping given in the new Fig 4a, as regards the areas treated, which though needs further explanation. Specifically:
   1. cleaning parameters: it is not clear the exact laser cleaning methodology as regards the number of pulses that were applied. You mention “300 pulses/min”, does that means that when working at 20 Hz, a total of 60 pulses were applied on each point/area or a total of 300 pulses was applied?
      And does this applies for “treatment a” or also for “treatment b”?
      In the latter case you have applied the total number of pulses (60 or 300) using the IR laser and then a second irradiation took place using another 60 or 300 pulses at 532nm?
      This is possibly an excessive cleaning condition for a standard black crust and possibly the surface has undergone some degree of damage (one could possibly see some damage at the left part of the sample).

Have you performed any preliminary study to determine the damage thresholds of your surfaces/materials?

b. the areas treated. As one can see in Figure 4a there are areas treated with “treatment b” , as for example areas close to spots #8,9 (in the left part of the image) and # 11 (in the lower-right part of the image), that although have been cleaned using the “treatment b” they appear different.
Have they be cleaned with the same conditions (i.e. number of pulses)?

I insist on this information as the final stage of the surface, after cleaning, is very important for the evaluation of the treatment and the final conclusions of your study. To my understanding it is vital to show which areas have been treated with treatment “a” and “b” and whether all parameters were the same or different cleaning levels were reached (i.e. under-, over-, optimum). This is the core of your results and must be clearly shown, because in case all treatments “b” have been performed with the same parameters/conditions then you results as shown in Figures 6 and 7, can be in doubt.

As regards the authors’ comment, that “treatment b” is more effective than “treatment a”, we in agreement because 2 different wavelengths were used, but I do not understand how this becomes “useful once 19064nm starts to operate”??

In “treatment b” you apply the 2 beams consecutively, that means that after having finished the irradiation with 60 or 300 pulses at 1064 nm, and possibly you have acquired a “yellow-ish” surface you have applied consecutively the 60 or 300 pulses of 532nm.. Is this correct?

This is rather confusing and definitely needs further discussion as this is actually the main aspect of your study and any uncertainties/misunderstanding as regards this matter jeopardize the whole study and its result.

3. Although figure 1b has been revised but, I am afraid, this is not a good image. The photo is saturated due to the flash-light and is definitely not acceptable for the purposes of this study. As a result of its bad quality, it is not possible to see and discriminate the details of the laser cleaned surfaces. Although not clearly discernible from figure 1b there are areas that appear damaged, or they are under-cleaned, and this is exactly the point that I wanted to emphasize.

It is clear that Figures 3a and 6b (now 4a and 7b) are the result of processing a number of recorded images, but they must be compared with a real image so to be able to show to the reader the actual and measurable differences (see previous comment). Your argument is not clear with the current presentation.
Actually, to my personal opinion and aesthetics, figure 4a in page 12, in which the different areas are marked in unclear blue and yellow dashed-lines and arrows, can be definitely enhanced. Each area should be uniquely marked and the exact laser irradiation parameters must be given also accompanied with remarks on the degree of cleaning.

4. The main question is how one can discriminate not only treatments “a” and “b” but also areas treated with “treatment b” but using different quantitative conditions (i.e. number of pulses as I can figure out from the current images).

For example, the area on the left part of the sample (treatment b) when compared with areas treated with the same methodology (on the low-right part of the sample) they appear different; in figure 6a (page 14) they appear both purple but with different intensity, while in figure 4b (page 14) one appears yellow and the other one red –slightly lighted to the area treated with “treatment a” (low-centre of the sample) and finally in figure 7a they appear yellow and red respectively. It is rather confusing.

5. Yellowing is a multifaceted and difficult to uniquely explain issue. It is true that for certain cases thermal changes due to the presence of iron compounds may be the reason for the yellowing, nevertheless your case seems not to be such. Actually have you determine the percentage of iron in the black crust and whether its alteration upon laser ablation may cause the yellowing? Although there are no close magnifications , as asked, to clearly see and judge the situation, it appears that the areas cleaned with the “treatment b” are rather over-cleaned.
   
   
6. We totally agree that the Authorities that are responsible for the preservation of each object/monument are the ones to set the appropriate cleaning methodology and set the cleaning limits. Actually in the case of the sample studied in this study the personal point of view is in total agreement with the fact that the 1064nm beam alone results to a more agreeable result, as from the figures given the areas treated with the “treatment b” appear rather over-cleaned.
   
   But this is a concept completely different from the use of the term “cleaning accuracy”. The term accuracy implies a quantitative character and must be defined in such a way. It is suggested to be revised with another term, connected with qualitative aspect in order to avoid misunderstandings; possibly with the “cleaning level” or “cleaning degree”.
   
   
7. Thank you for this revision.
   Is there an area of reference stone surface –before the crust formation- “unaltered stone” shown on the specific surface of the studied sample?
   
   

8. OK, this enhances significantly the reading of the graph

9. We do not agree that a complete removal of black crust may endanger the original surface, no matter if the treated surface is rough or perfectly smooth. This is the most important advantage of laser cleaning versus the other techniques and relies on its self-limiting character due to the differential absorption of crust/stone material to the ablative laser beam!!! There have been developed conditions and methodologies (i.e. to shape the beam etc.) that enable the conservator to treat such surfaces. This actually is perfectly followed in case of laser cleaning of heavily weathered and encrusted marble objects.

The question was different (do the crust remains affect your measurements?) and I am afraid, it was not answered. Actually the presence of the crust remains (under-cleaned surfaces), or inversely of damaged areas (over-cleaned surfaces), are particularly important in the process of investigating the applicability of an assessment technique and may give highly valuable information. I would encourage the authors to reconsider this question which possibly may require the revision of their experiments.

10. I guess ultraviolet was a mistake instead of green@532nm... It was effectively revised in the manuscript.

But, yet, the question “how the images enabled the choice of the best cleaning parameters?” has not been answered to my point of view. For the specific case/sample which would be the “best cleaning parameters” and how do you arrive to this decision/conclusion?

11. I am afraid the question was not answered. Definitely it is true the argument that “The formation of gypsum on the surface of the monument stone fragment studied in this work could be attributed to the effect of rain and rain-wash” but the question is not what caused the formation of the black crust. It was stated that you identified the crust composition on the basis of the acquired images and the question was “how did you succeed so?” and “how did you reached to the result that dust deposition in sheltered areas caused the gypsum crust on the basis of this research?”

12. Indeed, the Introduction section has been significantly revised as requested.

There are more works on the use of Spectral imaging in general, and specifically of multi- and hyper-spectral imaging, to assess the cleaning (not only laser) of stonework or other materials. As for example:

• https://doi.org/10.1007/s00339-011-6689-1
• DOI: 10.1016/j.ndteint.2006.07.012
• https://doi.org/10.1016/j.culher.2009.10.007
• https://doi.org/10.1179/sic.2006.51.Supplement-1.3 (which is actually your reference 25)
• https://doi.org/10.1117/12.445655
• https://doi.org/10.1117/12.2063777

13. the question on the biological component of this crust, didn’t mean to ask for reference information on the black crusts but its aim was to initiate discussion on whether the fluorescing component of biological formations may be also useful for its identification within the crust and also for the assessment of the cleaning efficiency/level.

14. ok

Author Response

Comments and Suggestions for Authors

The revised version of Manuscript Minerals-966097 has undergone a significant revision, but there are still some major issues that have not been approached effectively and to my personal opinion they necessitate a second revision. Specifically:

1   The term “monitoring” has been effectively replaced along the manuscript.

Point 2    The “details of cleaning procedure” have been, partially, updated.

    My suggestion was “to give analytical information, by adding to the manuscript a relevant table, and also specify the areas that were under-cleaned, over-cleaned and ideally treated” as this would enhance the understanding of the examined surfaces. There is some extra information, which to my point needs further update and also there is a kind of mapping given in the new Fig 4a, as regards the areas treated, which though needs further explanation. Specifically:

        cleaning parameters: it is not clear the exact laser cleaning methodology as regards the number of pulses that were applied. You mention “300 pulses/min”, does that means that when working at 20 Hz, a total of 60 pulses were applied on each point/area or a total of 300 pulses was applied?

Response 2: We have clarified in the previous revised version that 300 pulses/min were applied and the laser was impinging on the sample for a period lasting two minutes. We believe that available details on the cleaning procedure were provided in the revised manuscript.

We do not understand which kind of Table should be added to improve the manuscript. We believe that all relevant available analytical informations are provided in the revised text. As well, we believe that the differently cleaned areas are well differentiated through out the revised text.

And does this applies for “treatment a” or also for “treatment b”?

Response 2: The response above applies to both treatments.

        In the latter case you have applied the total number of pulses (60 or 300) using the IR laser and then a second irradiation took place using another 60 or 300 pulses at 532nm?

Response 2: At both consecutive wavelengths we have applied the procedure: 300 pulses/min over two minutes for IR and 300 pulses/min over two minutes for the Vis.

        This is possibly an excessive cleaning condition for a standard black crust and possibly the surface has undergone some degree of damage (one could possibly see some damage at the left part of the sample).

Have you performed any preliminary study to determine the damage thresholds of your surfaces/materials?

Response 2: The procedure was tested previously and then approved as not producing excessive cleaning and surface damage by the officials of the Authority of Fine Arts and Landscape.

1. the areas treated. As one can see in Figure 4a there are areas treated with “treatment b” , as for example areas close to spots #8,9 (in the left part of the image) and # 11 (in the lower-right part of the image), that although have been cleaned using the “treatment b” they appear different.

Have they be cleaned with the same conditions (i.e. number of pulses)?

Response 2: Yes.

I insist on this information as the final stage of the surface, after cleaning, is very important for the evaluation of the treatment and the final conclusions of your study. To my understanding it is vital to show which areas have been treated with treatment “a” and “b” and whether all parameters were the same or different cleaning levels were reached (i.e. under-, over-, optimum). This is the core of your results and must be clearly shown, because in case all treatments “b” have been performed with the same parameters/conditions then you results as shown in Figures 6 and 7, can be in doubt.

As regards the authors’ comment, that “treatment b” is more effective than “treatment a”, we in agreement because 2 different wavelengths were used, but I do not understand how this becomes “useful once 19064nm starts to operate”??

Response 2: We never assessed that “treatment b” is more effective than “treatment a”.

In “treatment b” you apply the 2 beams consecutively, that means that after having finished the irradiation with 60 or 300 pulses at 1064 nm, and possibly you have acquired a “yellow-ish” surface you have applied consecutively the 60 or 300 pulses of 532nm.. Is this correct?

This is rather confusing and definitely needs further discussion as this is actually the main aspect of your study and any uncertainties/misunderstanding as regards this matter jeopardize the whole study and its result.

Response 2: Yes, it is correct. At both consecutive wavelengths we have applied the procedure: 300 pulses/min over two minutes for IR and 300 pulses/min over two minutes for the Vis.

3    Although figure 1b has been revised but, I am afraid, this is not a good image. The photo is saturated due to the flash-light and is definitely not acceptable for the purposes of this study. As a result of its bad quality, it is not possible to see and discriminate the details of the laser cleaned surfaces. Although not clearly discernible from figure 1b there are areas that appear damaged, or they are under-cleaned, and this is exactly the point that I wanted to emphasize.

Response 3: Concerning Figure 1b we are sorry but we cannot provide a much better image than the new one now replacing the previous one, which although still a bit saturated by the flash light, allows to distinguish treated or not treated areas.

It is clear that Figures 3a and 6b (now 4a and 7b) are the result of processing a number of recorded images, but they must be compared with a real image so to be able to show to the reader the actual and measurable differences (see previous comment). Your argument is not clear with the current presentation.

Actually, to my personal opinion and aesthetics, figure 4a in page 12, in which the different areas are marked in unclear blue and yellow dashed-lines and arrows, can be definitely enhanced. Each area should be uniquely marked and the exact laser irradiation parameters must be given also accompanied with remarks on the degree of cleaning.

Response 3: The definition of the different areas in Figure 4a has been improved at our best using only one color to highlight the different areas. The laser parameters are reported in the text, and the labels “treatment a” or “treatment b” have been associated to each area of the figure.

4    The main question is how one can discriminate not only treatments “a” and “b” but also areas treated with “treatment b” but using different quantitative conditions (i.e. number of pulses as I can figure out from the current images).

For example, the area on the left part of the sample (treatment b) when compared with areas treated with the same methodology (on the low-right part of the sample) they appear different; in figure 6a (page 14) they appear both purple but with different intensity, while in figure 4b (page 14) one appears yellow and the other one red –slightly lighted to the area treated with “treatment a” (low-centre of the sample) and finally in figure 7a they appear yellow and red respectively. It is rather confusing.

Response 4: As described in the manuscript at page 14, the maps the level of efficiency of the treatment in restoring the level surface. The maps are not used to highlight the type of laser treatment, but only the cleaning level reached in various areas. The Vis map shows the treated-b area in purplish, while the NIR map highlights where the plaster has been completely removed, which does not necessarily coincide with all areas subjected to treatment b. The red horizontal part of the NIR map shows that there was no complete removal of the gypsum there, even though treatment b was used.

Further, the laser cleaning efficiency, i.e. the degree of removal, depends, besides on laser parameters such as energy, wavelength, duration and number of pulses, and others, on a large number of factors related to the samples which include surface roughness, porosity, presence of cavities, moisture content, chemical/physical properties and thickness of black crust.

For example, surface roughness and the presence of cavities affect the laser cleaning process of our sample that is not a compact smooth marble. The limestone surface is porous in some areas with deep holes, so that the surface is cleaned less efficiently, whereas in other areas the rock fracture the surface is smoother thus allowing a higher level of cleaning.

5    Yellowing is a multifaceted and difficult to uniquely explain issue. It is true that for certain cases thermal changes due to the presence of iron compounds may be the reason for the yellowing, nevertheless your case seems not to be such. Actually have you determine the percentage of iron in the black crust and whether its alteration upon laser ablation may cause the yellowing? Although there are no close magnifications , as asked, to clearly see and judge the situation, it appears that the areas cleaned with the “treatment b” are rather over-cleaned.

Response 5: We disagree with the reviewer because in our previous studies (see ref. 18, 35 and 41) we detected and quantified Fe only in the superficial layers of the black crust. We can confirm that this treated area does not appear over-cleaned.

6    We totally agree that the Authorities that are responsible for the preservation of each object/monument are the ones to set the appropriate cleaning methodology and set the cleaning limits. Actually in the case of the sample studied in this study the personal point of view is in total agreement with the fact that the 1064nm beam alone results to a more agreeable result, as from the figures given the areas treated with the “treatment b” appear rather over-cleaned.

7    But this is a concept completely different from the use of the term “cleaning accuracy”. The term accuracy implies a quantitative character and must be defined in such a way. It is suggested to be revised with another term, connected with qualitative aspect in order to avoid misunderstandings; possibly with the “cleaning level” or “cleaning degree”.

Thank you for this revision.

    Is there an area of reference stone surface –before the crust formation- “unaltered stone” shown on the specific surface of the studied sample?

Response 7: As suggested, we have replaced the term cleaning “accuracy” with cleaning “level” at page 7, line 233. The only unaltered area of the stone is present on the side of the cut.

8   OK, this enhances significantly the reading of the graph

9    We do not agree that a complete removal of black crust may endanger the original surface, no matter if the treated surface is rough or perfectly smooth. This is the most important advantage of laser cleaning versus the other techniques and relies on its self-limiting character due to the differential absorption of crust/stone material to the ablative laser beam!!! There have been developed conditions and methodologies (i.e. to shape the beam etc.) that enable the conservator to treat such surfaces. This actually is perfectly followed in case of laser cleaning of heavily weathered and encrusted marble objects.

Response 9: This is not our case: our stone is not a marble but a limestone so it yields a totally different response due to its physical and chemical features as detailed in previous Response 4. We didn’t shape the beam.

The question was different (do the crust remains affect your measurements?) and I am afraid, it was not answered. Actually the presence of the crust remains (under-cleaned surfaces), or inversely of damaged areas (over-cleaned surfaces), are particularly important in the process of investigating the applicability of an assessment technique and may give highly valuable information. I would encourage the authors to reconsider this question which possibly may require the revision of their experiments.

Response 9: We reconfirm that this work was not aimed to develop and test one type of laser cleaning compared to another, but rather it aimed at presenting hyperspectral technology as a supporting tool for laser cleaning. We agree that further tests need to be done in the future to understand if the method is able to evaluate under or over cleaned surfaces, but this study was only a pilot study to test the feasibility of the hyperspectral methodology as a technique of testing. To this purpose, the study shows that actually it is able to visualize the level of cleaning surface or not, e.g. in gypsum removal.

10    I guess ultraviolet was a mistake instead of green@532nm... It was effectively revised in the manuscript.

But, yet, the question “how the images enabled the choice of the best cleaning parameters?” has not been answered to my point of view. For the specific case/sample which would be the “best cleaning parameters” and how do you arrive to this decision/conclusion?

Response 10: In our case, the study was previously tested and then approved as satisfactory by the officials of the Authority of Fine Arts and Landscape.

11    I am afraid the question was not answered. Definitely it is true the argument that “The formation of gypsum on the surface of the monument stone fragment studied in this work could be attributed to the effect of rain and rain-wash” but the question is not what caused the formation of the black crust. It was stated that you identified the crust composition on the basis of the acquired images and the question was “how did you succeed so?” and “how did you reached to the result that dust deposition in sheltered areas caused the gypsum crust on the basis of this research?”

Response 11: Sorry, but we disagree, we never stated that we identified the crust composition on the basis of the acquired images.

12    Indeed, the Introduction section has been significantly revised as requested.

There are more works on the use of Spectral imaging in general, and specifically of multi- and hyper-spectral imaging, to assess the cleaning (not only laser) of stonework or other materials. As for example:

https://doi.org/10.1007/s00339-011-6689-1

    DOI: 10.1016/j.ndteint.2006.07.012

    https://doi.org/10.1016/j.culher.2009.10.007

    https://doi.org/10.1179/sic.2006.51.Supplement-1.3 (which is actually your reference 25)

    https://doi.org/10.1117/12.445655

    https://doi.org/10.1117/12.2063777

Response 12: We agree and know that there are many works on the use of hyper spectral imaging to assess cleaning of stones, but those we cited are only the few ones that deal with the combined use of hyper-spectral imaging and laser cleaning control.

13    the question on the biological component of this crust, didn’t mean to ask for reference information on the black crusts but its aim was to initiate discussion on whether the fluorescing component of biological formations may be also useful for its identification within the crust and also for the assessment of the cleaning efficiency/level.

Response 13: At this stage, this aspect is out of scope of our work at least but it is interesting and could be considered in further studies.

14    ok

Author Response File: Author Response.docx