Inﬂuence of the Laser Wavelength on Harmful E ﬀ ects on Granite Due to Bioﬁlm Removal

: The colonization of stone-built monuments by di ﬀ erent organisms (algae, fungi, lichens, bacteria, and cyanobacteria) can lead to biodeterioration of the stone, negatively a ﬀ ecting the artistic value of the heritage. To address this issue, laser cleaning has been widely investigated in recent years, due to the advantages it o ﬀ ers over traditional mechanical and chemical methods: it is gradual, selective, contactless, and environmentally friendly. That said, the laser parameters should be optimized in order to avoid any by-e ﬀ ects on the surface as a result of overcleaning. However, as the adjustment of each parameter to clean polymineralic stones is a di ﬃ cult task, it would be useful to know the e ﬀ ect of overcleaning on the di ﬀ erent forming minerals depending on the wavelength used. In this paper, three di ﬀ erent wavelengths (355 nm, 532 nm, and 1064 nm) of a Q-Switch neodymium-doped yttrium aluminum garnet (Nd:Y 3 Al 5 O 12 ) laser, commonly known as QS Nd:YAG laser were applied to extract a naturally developed sub-aerial bioﬁlm from Vilach á n granite, commonly used in monuments in the Northwest (NW)Iberian Peninsula. In addition to the removal rate of the bioﬁlm, the by-e ﬀ ects induced for ﬂuences higher than the damage threshold of the stone were evaluated using stereomicroscopy, color spectrophotometry, and scanning electron microscopy with energy-dispersive x-ray spectroscopy. The results showed that di ﬀ erent removal rates were obtained depending on the wavelength used and 532 nm obtained the highest removal level. In terms of by-e ﬀ ects, biotite melting was registered on all surfaces regardless of the wavelength. In addition, 532 nm seemed to be the most aggressive laser system, inducing the greatest change in appearance as a result of extracting the kaolinite crackled coating and the segregations rich in Fe, which are a result of natural weathering. These changes were translated into colorimetric changes visible to the human eye. The surfaces treated with 355 nm and 1064 nm showed lower surface changes.


Introduction
Laser cleaning is a technique that is currently being fine-tuned in order to clean cultural heritage stones [1][2][3]. The advantages of this technology lie in its selectivity and graduality (precise removal of thin layers), not forgetting the fact that it does not come into contact with the surface to be cleaned and it is environmentally friendly [1,4]. Furthermore, the possibility of automation is an enticing option that is still under investigation.
Since laser was first used in stone cleaning in the 1970s-to clean incrustations from Venetian marble [5]-there has been a considerable amount of scientific research focused on the optimization of lasers to maximize cleaning while minimizing damage caused to the forming minerals. A neodymium-doped yttrium aluminum garnet (Nd:YAG) laser, a neodymium-doped yttrium orthovanadate (Nd:YVO 4 ) laser, and an erbium-doped yttrium aluminium garnet (Er:YAG) laser have been used to extract graffiti, as well as sulphated and lichenic black crusts from different rocks with different texture and mineralogy, primarily limestone, marble, and granite [3,[6][7][8][9][10][11].
In the Northwest (NW)Iberian Peninsula, cultural heritage is built with granite, which is a bioreceptive stone [12]. Considering the climate of this region (a humid climate with rainy winters), the stone's durability would be compromised by the colonization of monuments and facades with organisms [12]. Sub-aerial biofilms (SAB), composed mainly of phototrophs (algae and cyanobacteria), are the initial colonization stage on granite. Consequently, the artistic and historical value of the monument will be negatively affected. Therefore, removal of this initial biocolonization from monuments should be carried out urgently in order to avoid greater damage from hyphae penetration and mineralogical transformations, or neoformations due to the acids generated by the organisms [13,14]. Different procedures have been implemented by professionals to extract biological patinas and crusts, mostly involving chemical methods and the mechanical scalpel [15,16]. Although scientific publications devoted to the optimization of biofilm removal from granite are scarcer than for carbonate stones, these traditional procedures have nonetheless exhibited some drawbacks in this area, making a case for the use of other less invasive techniques in granite cleaning. Pozo et al. [17] have evaluated chemical products (ethanol, benzalkonium chloride, hydrochloric acid solutions, and the commercial biocides Hyvar X ® and LimpiaFachadas1 ® ) used by professionals to extract a biofilm composed of filamentous green algae belonging to the genus Trebouxia and cyanobacteria belonging to the genera Gleocapsa and Choococcus from coarse-grained granite. Despite the satisfactory results achieved by the commercial biocides, salts were precipitated on the surface. Moreover, the same research evaluated a low-pressure (0.5-1.5 bar) air rotation system that used a mixture of air-water-silicate grains as an abrasive. Due to the formation of cracks and grain extraction, surface roughness was increased. Therefore, in order to avoid such issues that pose a risk to manually carved cultural heritage elements, laser treatment is increasingly being implemented. However, to the best of our knowledge, there is insufficient scientific information relating to the laser cleaning of SAB, with this alteration being the initial stage in the colonization of granitic monumental heritage. López et al. [18] found that by means of an Nd:YVO 4 laser at 355 nm, a patina composed of filamentous green algae was successfully removed from a fine-grained granitic surface. In the cleaning of a biological colonization from marble, the third harmonic (355 nm) of the Nd:YAG system showed more advantages when compared with the fundamental wavelength (1064 nm); also, the yellowing effect observed when using the latter was avoided [19]. When treating a more complex crust composed of lichen, Sanz et al. [11] found that the optimal conditions for laser removal were highly dependent on the lichen species and, to a lesser extent, on the type of stone. They applied a Q-Switched (QS) Nd:YAG laser at 1064 nm (IR), 355 nm and 266 nm (UV), as a single wavelength and in dual sequential irradiation (IR+UV). What we learn from the research conducted so far is that every study case needs its optimization. Therefore, before to perform a real cleaning of biological colonization on a heritage stone, it is necessary to perform preliminary tests in small areas with the laser systems available, playing with the different laser parameters such as frequency, velocity, fluence, etc.
According to the literature, the polymineralic composition of granite hinders the effectiveness of laser cleaning. Therefore, it is a challenge to perform a satisfactory level of cleaning without damaging some mineral grains. Biotite is the most critical granite-forming mineral [20,21]. Moreover, tests applying an Nd:YAG laser at the fundamental wavelength (1064 nm) and at the third harmonic (355 nm) on a uncoated pinkish granite have reported that the characteristic coloration became paler as a result of the physical elimination of the ZnFe 2 O 4 particles from the feldspars due to thermal effects [22]. However, these color changes were not detected on grey granites and whitish limestones [23,24]. On carbonate stones, a yellowish effect was detected on marble when it was irradiated with the Coatings 2020, 10,196 3 of 17 fundamental of a QS Nd:YAG laser in order to extract black crusts [7,25]. However, UV radiation using the third harmonic of a QS Nd:YAG laser induced a grey discoloration on these surfaces [7].
From a practical point of view, Nd:YAG lasers are used as portable equipment to perform cleaning in situ. Currently, these portable systems offer the possibility to experiment with different wavelengths (266 nm, 355 nm, 532 nm, and 1064 nm). Considering the findings from the references cited above, as well as the current incorporation of new wavelengths in portable laser equipment to clean cultural heritage objects in situ, it is vital to ascertain, in addition to the removal level, the effect produced by the different wavelengths on the stone when overcleaning takes place. This is because, during in situ cleaning, fluence variations could occur and lead to unintended damage in the forming minerals. Therefore, before any laser cleaning is performed, the extent of potential damage will be a decisive factor in the choice of wavelength. In this study, three wavelengths (355 nm, 532 nm, and 1064 nm) of a QS Nd:YAG laser system were used to extract a SAB from a pre-Hercynian granite commonly used in cultural heritage in the NW Iberian Peninsula. Particular attention was paid to the by-effects on the different granite-forming minerals in order to find different behaviours of the granite and each forming mineral under different wavelengths. The findings of this research will be used by conservator-restorers in order to select an appropriate wavelength to perform the cleaning of pre-Hercynian granites, characterized by their yellowish-brown coloration. The cleaning removal rate and the damage caused to the different forming minerals were evaluated by stereomicroscopy, spectrophotometry, Fourier transform infrared spectroscopy (FTIR), pyrolysis-gas chromatography/mass spectrometry (PY-GC/MS), and scanning electron microscopy (SEM).

Granite and Sub-Aerial Biofilm (SAB)
A natural exposed granite slab with a disc cutting finish surface of (approx.) 20 cm × 20 cm × 2 cm with an intense biological colonization (a sub-aerial biofilm-SAB) was taken from a stone workshop in the NW Iberian Peninsula ( Figure 1A). The SAB homogenously covered the surface ( Figure 1A). Under stereomicroscopy (SMZ800 NIKON®, Tokyo, Japan), it was found that the SAB was composed of filamentous green algae belonging to the genus Trebouxia, with a minor presence of cyanobacteria belonging to the genera Gleocapsa and Choococcus. A C-coated 1 cm × 1 cm × 0.5 cm-scale of the SAB on the granite was visualized under SEM in secondary electron (SE) and backscattered electron (BSE) modes using a Philips XL30 with energy-dispersive x-ray spectroscopy (EDS). Samples were visualized at an accelerating potential of 15-20 kV, a working distance of 9-11 mm and a specimen current of approximately 60 mA. SEM micrographs of the SAB showed a dense crackled layer covering the stone, with silicate grains entrapped within the C-rich matrix ( Figure 1B).
The granite was an ornamental granite (called Vilachán) of great commercial value, traditionally used in the construction of historical buildings and sculptures in this region ( Figure 1C,D). This granite is classified as a fine-grained panallotriomorphic heterogranular granite. It is composed of quartz (47%), muscovite (18%), potassium feldspar (10%), sodium plagioclase (15%), and biotite (7%) as the main minerals, and includes apatite and zircon as accessory minerals [26]. Grain size for the different minerals oscillates between 2 mm and 0.3 mm. The open porosity of the rock was 2.82% following [27].
In addition to the colonized surface, an uncoated slab (approx. 10 cm × 10 cm × 2 cm) subjected to the same atmospheric conditions was collected to be used as a reference surface. This granite is characterized by its yellowish-brown color ( Figure 1C-D), associated with the existence of iron forms filling the fissures of the rock [26,28]. Therefore, this yellow color must be considered as a characteristic textural feature. A C-coated 1 cm × 1 cm × 0.5 cm-scale of the uncoated granite was visualized under SEM-EDS. Due to the moderately weathered state of this granite, intact minerals were detected ( Figure 1E-G), but also some weathering symptoms such as crackled coatings rich in Al and Si, precursors of kaolinite, and segregations rich in Fe on the surface and fissures ( Figure 1H and EDS spectra).

Laser Cleaning
The equipment used to carry out the laser cleaning on the SAB was composed of a Q-Switch neodymium-doped yttrium aluminum garnet (Nd:Y 3 Al 5 O 12 ), commonly known as QS Nd:YAG laser (Quanta Ray, INDI) with a 6 ns pulse duration that can deliver the fundamental 1064 nm beam (IR radiation), the second harmonic (532 nm-green Vis radiation) and the third harmonic (355 nm-UV radiation). The pulse repetition rate could be selected from single-shot to 10 Hz with energy per pulse of around 0.01 J, 0.03 J and 0.15 J for 355 nm, 532 nm, and 1064 nm respectively. The laser beam reached the surface using a spherical plane-convex lens (NewPort) with a focal length of 250 mm.
As the damage threshold for this kind of granite is set in the 0.7-1.5 J·cm −2 range for an Nd:YAG laser [20,29], and because the laser effect depends on the patina or coating [11], fluences of 2 J·cm −2 and 5 J·cm −2 were used in order to exert damage on the stone. In previous studies, the damage threshold for the same granite as that used in this research was 1.1 J·cm −2 for 1064 nm and 0.7 J·cm −2 for 355 nm [29,30]. Since in this research, the objective was to evaluate the influence of the wavelength used on the by-effects in the different granite forming minerals due to an overcleaning, the fluences applied were higher than the damage thresholds of the granite. Note that the damage threshold of the granite corresponds to the highest fluence that causes no alterations to biotite. As consequence of the material evaporation or sublimation due to the heating by the absorbed laser energy, several damages can be detected in the biotite grains, such as melting, opening of the exfoliation planes or extraction of grains), because this forming mineral is the most sensitive to laser beam radiation [20,21]. The standard cleaning methodology applied during the laser cleaning of rocks formed by a single mineral (generally marble or limestone) is usually based on the performance of single spots affecting a small area quickly and comfortably to evaluate [7,19,31]. However, as was reported in a previous work [21], in the case of fine grained granites, such as the one used in this paper, an individual spot may not cover only one mineral of the four main silicates (quartz, K-feldspar, plagioclase and biotite) due to the higher size of the spot diameter comparatively to the mineral grain size. Therefore, the fine grain size of this granite ( Figure 1A), made it unfeasible to determine the ablation threshold through single-spot irradiation. For this reason, setting a distance of 0.075 mm between adjacent scans and a scan speed of 50 mm·s −1 , areas of 3 cm × 3 cm were irradiated using 1-2 overlapping laser pulses. Values of the spot diameter were adjusted for each fluence and wavelength, and the overlap was calculated ( Table 1). The fixed beam impinged perpendicularly onto the target surface of the specimen placed vertically on a motorized XYZ-translation stage. The sample was precisely positioned at the beam waist of the focused beam using the Z-direction. Customized software was used to control the XY movement and to synchronize the laser with specimen displacement.

Analytical Techniques
In order to compare the surfaces treated with the three different wavelengths and to state the SAB removal levels and the damage caused to the different forming minerals, the following techniques were applied: • Stereomicroscopy using an SMZ800 NIKON ® was firstly used to characterize the surfaces treated with the different conditions (Table 1).

•
The color of the reference granite surface and the laser-treated surfaces was measured with a Minolta CM-700d spectrophotometer. Color was characterized in the CIELAB and CIELCH color spaces [32,33]. A total of 15 random measurements were taken for each surface in order to obtain statistically representative results [34]. Measurements were taken in specular component included (SCI) mode, with a spot diameter of 8 mm, using illuminant D65 at a viewing angle of 10 • . The color parameters measured were L*, lightness which varies from 0 (absolute black) to 100 (absolute white); a*, showing color changes in the red-green range (+a*: red and −a*: green); and b*, related to changes in the yellow-blue range (+b*: yellow and −b*: blue). Moreover, 2 angular parameters were measured: C* ab , chroma or saturation, related to the intensity of color; and h ab , hue or tone of color, which refers to the dominant wavelength indicating redness, yellowness, greenness or blueness on a circular scale. In order to evaluate the cleaning effectiveness, color data were processed as color differences (∆L*, ∆a*, ∆b*, ∆C* ab , ∆H*) and global color change (∆E* ab ) between the reference uncoated granite and the laser-treated surfaces. Global color change (∆E* ab ) was computed as follows [33]: • Fourier transform infrared spectroscopy (FTIR) data in micro mode were collected from the reference granite, the SAB and the laser-treated surfaces using a FTIR Thermo Nicolet®Continuµm. The FTIR spectra were recorded in reflectance mode in the 4000-400 cm −1 region, with 4 cm −1 resolution.

•
The organic material before and after laser ablation was investigated with pyrolysis-gas chromatography/mass spectrometry (PY-GC/MS) using a Multi-Shot Pyrolyzer EGA/PY-3030D micro-furnace (Frontier Lab), coupled on-line with an Agilent Technologies (USA) 6890/5973 Gas Chromatography/Mass Selective Detector system. The experiments were conducted by inserting the sample (approximately 0.5 mg) into a stainless steel cup, and after adding 5 µL of hexamethyldisilazane as derivatization agent, performing the pyrolysis with a furnace temperature of 550 • C (for 60 s) and an interface temperature of 180 • C. The conditions for the gas chromatograph and the mass spectrometer are reported in literature [35]. • Then, C-coated 1 cm × 1 cm × 0.5 cm-scales of the laser-treated surfaces were visualized under SEM in SE and BSE modes using a Philips XL30 with EDS. Optimal conditions were the same as those reported for the granite characterization.

Results
Stereomicroscopy identified differences regarding the SAB removal rate and the by-effects on the surfaces irradiated with the different wavelengths. At low magnifications, the surfaces treated with 355 nm seemed to be satisfactorily cleaned (Figure 2A,C) compared with the reference surface ( Figure 1C), but at higher magnifications, organic remains with a yellowish-brown coloration similar to the color of the granite were found ( Figure 2B,D).
Surfaces treated with 532 nm showed a decrease in the yellowish-brown coloration ( Figure 2E-H), which is an intrinsic characteristic of this granite [26,28]. On the surface treated with 2 J·cm −2 ( Figure 2E,F), organic remains were found as was the case with the surfaces after 355 nm ( Figure 2B, D), but the remains were paler and less extensive with the second harmonic wavelength. On the surface treated with 532 nm and 5 J·cm −2 , no organic remains were found ( Figure 2H).
Surfaces treated at 1064 nm ( Figure 2I-L) showed more effective cleaning levels than those cleaned at 355 nm, and less variation from the original color than those treated at 532 nm. As for the fluences used, as expected, 5 J·cm −2 achieved a greater removal of SAB than 2 J·cm −2 ( Figure 2I,K). In this case, the remains showed a dark color (similar to the SAB) in contrast to the paler remains found on the surfaces treated with the other wavelengths. Moreover, regarding the by-effects, treatment with 5 J·cm −2 ( Figure 2K,L) resulted in a slight paling effect, as was found on the surface treated with 532 nm.
Coatings 2020, 9, x FOR PEER REVIEW 7 of 18 Therefore, the first evaluation performed with stereomicroscopy identified the conditions that did not cause a visible loss of the yellowish-brown coloration of the granite: 355 nm at both fluences, and 1064 nm at 2 J•cm −2 . However, some organic remains were found on those surfaces; these were darker on the surface treated with 1064 nm. Paler surfaces were detected after treatment at 532 nm, regardless of the presence of organic remains. In order to obtain a general evaluation of the treated surfaces, color spectrophotometry was performed. Colorimetric differences (ΔL*, Δa*, Δb*, ΔC*ab, and ΔH*) and global color changes (ΔE*ab) were computed considering the color of the uncoated granite as a reference (Table 2 and Figure 3). Therefore, the first evaluation performed with stereomicroscopy identified the conditions that did not cause a visible loss of the yellowish-brown coloration of the granite: 355 nm at both fluences, and 1064 nm at 2 J·cm −2 . However, some organic remains were found on those surfaces; these were Coatings 2020, 10,196 8 of 17 darker on the surface treated with 1064 nm. Paler surfaces were detected after treatment at 532 nm, regardless of the presence of organic remains.
In order to obtain a general evaluation of the treated surfaces, color spectrophotometry was performed. Colorimetric differences (∆L*, ∆a*, ∆b*, ∆C* ab , and ∆H*) and global color changes (∆E* ab ) were computed considering the color of the uncoated granite as a reference (Table 2 and Figure 3). On all of the treated surfaces, it was observed that the color change mainly affected the b* coordinate, except in the case of the cleaning with 355 nm and 2 J·cm −2 , which had a higher change in coordinate a*. Coordinate b* decreased in all cases, showing a lessening of the yellowish coloration and this decrease was more pronounced in the surfaces treated with 5 J·cm −2 for all three wavelengths. The highest b* decreases were detected in the color of the surfaces treated with 532 nm, while the surfaces cleaned with 355 nm showed the lowest values. Likewise, a* also decreased (although to a lesser extent than b*), causing a decrease in the reddish coloration. Note that the original color of this granite is yellowish-brown. Conversely, the L* parameter showed increases in most of the cases (only the surface cleaned with 1064 nm at 5 J·cm −2 showed a slight L*decrease). As a result, the chroma (∆C* ab ) decreased in all cases, primarily on the surfaces treated with 532 nm and the surfaces treated with 355 nm exhibited the lowest ∆C* ab . The hue experienced low changes and surfaces cleaned with 355 nm and 532 nm showed increases, while surfaces cleaned at 1064 nm showed decreases.  FTIR spectroscopy was used to identify any organic remains on the surfaces. Figure 4 depicts the FTIR spectra of the reference granite, the SAB on the granite, and the surfaces treated with different wavelengths and fluences. In the FTIR spectra of the treated surfaces, characteristic bands  In general terms, it would be fair to say that, after laser treatments, the surfaces became lighter and less yellow, especially the surfaces treated with 532 nm.
Global color changes ( Figure 3) were higher in the surfaces cleaned with 532 nm (∆E* ab > 6 CIELAB units). The ∆E* ab for the surfaces cleaned at 1064 nm with both fluences were higher than 3 CIELAB units. The lowest ∆E* ab were found in the 355 nm-treated surfaces (the ∆E* ab of the 2 J·cm −2 -treated surface was 1.38 CIELAB units; the ∆E* ab of the 5 J·cm −2 -treated surface was 3.16 CIELAB units).
Note that according to Berns [36], 3 CIELAB units is the threshold from which a color change is perceivable to the human eye. Therefore, only the 355 nm system at 2 J·cm −2 , caused an undetectable color change in the granite surface. However, referring to García and Malaga [37], who reported that a color change is visible when the ∆E* ab is higher than 5 CIELAB units, then 355 nm at a fluence of 5 J·cm −2 and 1064 nm at a fluence of 2 J·cm −2 also did not provoke any detectable color changes.
FTIR spectroscopy was used to identify any organic remains on the surfaces. Figure 4 depicts the FTIR spectra of the reference granite, the SAB on the granite, and the surfaces treated with different wavelengths and fluences. In the FTIR spectra of the treated surfaces, characteristic bands of silicates were found in the range 500-1180 cm −1 , as would be expected [38]. However, bands assigned to the SAB were inexistent or less intense. In the SAB FTIR spectra, the bands corresponding to organic matter were mainly the broad shoulder at 3600-3100 cm −1 , centred at 3400 cm −1 , assigned to the O-H stretching vibration of water, and the double band at 2950 cm −1 and 2850 cm −1 , corresponding to the C-H asymmetric stretching vibration of alkanes [38]. The absorption band at 1880 cm −1 is assigned to esters of unsaturated aliphatic fatty acids [38]. In addition, different bands in the range 1700-1300 cm −1 were detected: (i) a band at 1660 cm −1 suggested the presence of ester, aldehyde, and ketone groups; (ii) a band at 1550 cm −1 was assigned to amide II; (iii) a band at 1440 cm −1 was assigned the C=C group indicative of chromophores of chlorophyll; and (iv) a band around 1300 cm −1 indicated the presence of C=N groups [38,39].
After cleaning, very low intense bands assigned to organic matter (the broad shoulder at 3600-3100 cm −1 ) were found only on the surfaces cleaned with 355 nm mainly at 2 J·cm −2 (Figure 4). This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation as well as the experimental conclusions that can be drawn.
The SAB residue was investigated using PY-GC/MS analysis. The same amount (0.5 mg) was scraped from the surfaces after laser treatment at the three different wavelengths. The SAB was characterized studying the saccharide content of the cyanobacteria. As an example, Figure 5 depicts the extract ion pyrograms of the ion at m/z 217 (which is characteristic of saccharide material) for the material scraped from the SAB, and from the surface treated with 2 J·cm −2 at 1064 nm. The analysis revealed that after the laser treatment, the amount of biofilm on the surface was markedly reduced. However, it should be noted that a quantitative analysis was not possible because the presence of the markers of the biofilm was too low in the scraped samples after laser application.
A more detailed analysis via SEM-EDS helped identify organic remains and determine the by-effects of overcleaning on the treated surfaces. For the surfaces treated with 355 nm (Figure 6), SEM confirmed the presence of organic remains previously detected with the high magnification-visualizations during stereomicroscopy and FTIR. SAB remains were clearly identified as low contrast C-rich deposits on the surface cleaned at the lower fluence ( Figure 6A). A smaller amount of organic remains was found on the surface cleaned at 5 J·cm −2 ( Figure 6B).
Regarding damage caused to the forming minerals at this wavelength (355 nm), biotite melting was found to have occurred on the surfaces cleaned with both fluences (compare the treated biotite in Figure 6C,D with the untreated biotite in Figure 1G). No melting or fracturing of the other forming minerals (muscovite, K-feldspar, plagioclase and quartz) was detected, even at the highest fluence ( Figure 6C-F). Moreover, the Al-and Si-rich crackled coating found on the surface of the grains in the reference granite ( Figure 1H) remained on the surface after treatment ( Figure 6F). However, on the surfaces treated with the other wavelengths, the weathered coating was extracted (Figures 7 and 8). Coatings 2020, 9, x FOR PEER REVIEW 10 of 18 The SAB residue was investigated using PY-GC/MS analysis. The same amount (0.5 mg) was scraped from the surfaces after laser treatment at the three different wavelengths. The SAB was characterized studying the saccharide content of the cyanobacteria. As an example, Figure 5 depicts the extract ion pyrograms of the ion at m/z 217 (which is characteristic of saccharide material) for the material scraped from the SAB, and from the surface treated with 2 J•cm −2 at 1064 nm. The analysis revealed that after the laser treatment, the amount of biofilm on the surface was markedly reduced. However, it should be noted that a quantitative analysis was not possible because the presence of the markers of the biofilm was too low in the scraped samples after laser application. A more detailed analysis via SEM-EDS helped identify organic remains and determine the byeffects of overcleaning on the treated surfaces. For the surfaces treated with 355 nm (Figure 6), SEM confirmed the presence of organic remains previously detected with the high magnificationvisualizations during stereomicroscopy and FTIR. SAB remains were clearly identified as low contrast C-rich deposits on the surface cleaned at the lower fluence ( Figure 6A). A smaller amount of organic remains was found on the surface cleaned at 5 J•cm −2 ( Figure 6B).
Regarding damage caused to the forming minerals at this wavelength (355 nm), biotite melting was found to have occurred on the surfaces cleaned with both fluences (compare the treated biotite in Figure 6C,D with the untreated biotite in Figure 1G). No melting or fracturing of the other forming minerals (muscovite, K-feldspar, plagioclase and quartz) was detected, even at the highest fluence ( Figure 6C-F). Moreover, the Al-and Si-rich crackled coating found on the surface of the grains in the reference granite ( Figure 1H) remained on the surface after treatment ( Figure 6F). However, on the surfaces treated with the other wavelengths, the weathered coating was extracted (Figures 7 and  8).
In the same way as the treatment at 355 nm, 532 nm ( Figure 7) produced better results at a fluence of 5 J•cm −2 than at 2 J•cm −2 in terms of the SAB removal rate ( Figure 7A,B). However, an important difference was found compared with the 355 nm-treated surfaces: in addition to biotite melting at both fluences ( Figure 7C,F), muscovite exfoliation planes showed slight melting and fracturing at the higher fluence of 5 J•cm −2 ( Figure 7E,H), while the other minerals (k-feldspar, plagioclase, and quartz) did not exhibit any surface changes ( Figure 7D,G).
The cleaning at 1064 nm did not completely extract the SAB, because remains were found on the surfaces treated at both fluences ( Figure 8A,B). Regarding damage caused to the mineral grains, as with the cleaning at 532 nm, in addition to the melting of the biotite planes ( Figure 8C,F), the muscovite grains after 5 J•cm −2 showed fracturing of the planes, but no fusion was observed ( Figure  8E,H). The rest of the forming minerals seemed to be unaltered by the laser beam ( Figure 8D,G).      In the same way as the treatment at 355 nm, 532 nm ( Figure 7) produced better results at a fluence of 5 J·cm −2 than at 2 J·cm −2 in terms of the SAB removal rate ( Figure 7A,B). However, an important difference was found compared with the 355 nm-treated surfaces: in addition to biotite melting at both fluences ( Figure 7C,F), muscovite exfoliation planes showed slight melting and fracturing at the higher fluence of 5 J·cm −2 ( Figure 7E,H), while the other minerals (k-feldspar, plagioclase, and quartz) did not exhibit any surface changes ( Figure 7D,G).
The cleaning at 1064 nm did not completely extract the SAB, because remains were found on the surfaces treated at both fluences ( Figure 8A,B). Regarding damage caused to the mineral grains, as with the cleaning at 532 nm, in addition to the melting of the biotite planes ( Figure 8C,F), the muscovite grains after 5 J·cm −2 showed fracturing of the planes, but no fusion was observed ( Figure 8E,H). The rest of the forming minerals seemed to be unaltered by the laser beam ( Figure 8D,G).

Discussion
This paper evaluated the influence of three different wavelengths (355 nm-third harmonic; 532 nm-second harmonic; and 1064 nm-fundamental radiation), in terms of damage caused to the granite-forming minerals, during laser cleaning by a QS Nd:YAG to extract a SAB. A naturally biocolonized pre-Hercynian Vilachán granite with a certain level of natural weathering, indicated by a crackled coating rich in Al and Si identified as kaolinite [40,41] and Fe-rich segregations on the surface and fissures, was selected. Note that the presence of this Fe-rich segregation is responsible for the yellowish coloration of this granite [26,28]. Therefore, it has to be considered as an intrinsic property of this stone and it should be safeguarded in any cleaning campaign.
As stated by Rodrigues and Grossi [42] and Revez and Delgado Rodrigues [43], laser cleaning at any wavelength poses a high risk, given the vulnerability of the treated surface, the aggressiveness of the cleaning method, the synergistic effects due to the laser-weathered granite combination, the possible uniqueness of the cultural heritage object, and the global color changes. Therefore, as the application of any of the tested wavelengths represents a complex intervention, a detailed characterization of the by-effects due to overcleaning should always be performed, in order to successfully clean while safeguarding the substrate.
In order to ensure damage to the substrate, fluences higher than the damage threshold for this type of granite [20,30] were applied in only one scan. Therefore, 2 J·cm −2 and 5 J·cm −2 were irradiated for each wavelength. Any organic remains belonging to the SAB after cleaning were also described. Stereomicroscopy, FTIR, and SEM identified the wavelength of 532 nm at a fluence of 5 J·cm −2 as the condition that achieved the best results in terms of SAB removal. On the rest of the surfaces, biological remains were detected: (i) on the surfaces treated with 355 nm (both fluences) and 532 nm at 2 J·cm −2 , burnt remains (brownish color) were observed on the surfaces; (ii) on the 1064 nm-treated surface at 2 J·cm −2 , dark remains belonging to the patina were found by stereomicroscopy, and (iii) on the surface treated with 1064 nm at 5 J·cm −2 , organic remains were detected using SEM. In a previous study focusing on the removal of lichen from Carrara marble using the fundamental wavelength (1064 nm) and the second harmonic (532 nm) of an Nd:YAG, it was reported that the latter provided significant advantages over 1064 nm, because of the low optical absorption of the lichen at the IR wavelength [44]. The phycocyanins and phycoerythrins that are accessory pigments to chlorophyll exhibit a significant optical absorption at 555 nm and 545 nm, respectively, and could enhance the removal effectiveness at 532 nm [44]. Note that, in this study, FTIR identified chromophores of chlorophyll in the treated SAB.
Turning now to the by-effects, stereomicroscopy, color spectrophotometry and SEM were used to identify changes detected on the surfaces irradiated by different wavelengths. The first two analytical techniques indicated that the 532 nm wavelength, regardless of the fluence, induced a visible modification in the color, as was identified by the ∆E* ab above 3 CIELAB units, which is often considered as the threshold for a color difference to be perceivable to the human eye [36]. A color modification also occurred in the 1064 nm-treated surface with the highest fluence (5 J·cm −2 ), although it was lower than the changes found on the 532 nm-treated surfaces. A lessening of the characteristic yellowish color of this stone was registered, in accordance with decreases in the b* color coordinate. As for the remaining conditions, a wavelength of 355 nm (regardless of the fluence) and of 1064 nm at 2 J·cm −2 did not cause any visible coloration modifications; in the surface treated with 1064 nm and 2 J·cm −2 , the modification was barely visible, and this was attributed to dark SAB remains on the surface after laser cleaning, as detected by stereomicroscopy and SEM. Therefore, just as SAB removal with 532 nm was higher in relation to the other two wavelengths, color modification of the surface was also higher. Therefore, the highest SAB extraction rate at 532 nm might be related to the highest absorption at this wavelength, comparatively to the other two (355 nm-ultraviolet radiation and 1064 nm-infrared radiation). In fact, 532 nm is close to the typical wavelength at which chlorophyll a absorbs (675 nm) [45,46]. Carrère et al. [47] assigned absorption features at 590 and 635 nm to absorbing pigments such as chlorophyll c. At 665~670 nm, intense absorption was registered in the absorbance spectra of granite samples inoculated with green algae and cyanobacteria [48].
The paling effect due to the decrease in the yellowish coloration on the surfaces treated with 532 nm could be related to the extraction of the kaolinite crackled coating and colorimetric changes associated to the Fe-rich segregations. Working with a granite with different texture and color than the Vilachán granite, Urones-Garrote et al. [22] with the pinkish Rosa Porriño granite irradiated with an Nd:YAG at 355 nm, found that this characteristic color disappeared due to the physical elimination of the ZnFe 2 O 4 particles from the feldspars as a result of thermal effects. Therefore, color changes seem to be associated to chemical changes affecting the Fe forms [22,24]. As it is well known, laser radiation can modify the substrate by thermal shock [2,3], inducing mineral phase transformation of iron forms as consequence of the quick temperature increases [49,50]. Chakrabarti et al. [49] found that color change in sandstones is mainly due to the dehydration of iron forms after heating. Hajpál and Török [50] working with sandstones exposed to different temperatures, found that the color changes were related to the transformation of iron-bearing mineral phases. Therefore, color change of the iron segregations or even their ablation would be behind the color change suffered for the granitic stone after the laser treatment. Moreover, kaolinite deposits covering feldspar grains should undoubtedly be affected by the laser; in the current research, kaolinite deposits were eliminated with the laser irradiation. In [44], where 532 nm achieved better results than 1064 nm in the removal of lichen from Carrara marble, this paling effect was not found due to the original white color of the marble without any neoformed layer as a result of weathering. Considering the results obtained in our research, the mineralogy of the stones after being subjected to weathering are key parameters to understand the results of laser cleaning. In addition to the melting of biotite planes found in all the treated surfaces, the 532 nm-treated surfaces also showed a slight melting of the muscovite exfoliation planes. As reported in previous research [20,21,30], biotite is the first mineral to melt due to its high optical absorption along with its relatively low melting point of 650 • C while 1110 • C for plagioclase (albite), 1250 • C for potassium feldspar, and 1710 • C for quartz [51,51].

Conclusions
This research found that different wavelengths (355 nm-third harmonic; 532 nm-second harmonic; and 1064 nm-fundamental radiation) produced different by-effects on a piece of fine-grained, low-weathered granite after laser cleaning with an ns QS Nd:YAG to remove a SAB. In order to ensure damage to the substrate, fluences above the damage thresholds were applied.
Despite showing satisfactory SAB removal rates, a wavelength of 532 nm (second harmonic) induced the highest color modifications in the granite due to the extraction of the kaolinite crackled layer and the Fe-rich segregations, which are a result of the natural weathering process. In addition to biotite melting, found in all of the treated surfaces, regardless of wavelength and fluence, treatment with 532 nm also induced a slight melting and fracturing of the muscovite exfoliation planes.
The third harmonic (355 nm), at both fluences, did not cause visible changes to the surface. However, burnt organic remains were detected on the surface, jeopardizing the cleaning effectiveness of this wavelength. In addition, these organic remains could induce recolonization. Further studies should therefore be performed to examine the tertiary bioreceptivity of these surfaces.
The 1064 nm-treated surfaces did not show coloration modifications in the granite at the lowest fluence used, despite being above the damage threshold. However, the dark color of the remains could mean that a second laser scan would enhance cleaning effectiveness.
In conclusion, when designing a programme to laser-clean slightly weathered granite that has the kaolinite crackled layer and Fe-rich segregations, laser parameters should be optimized in advance of the actual cleaning of the stone objects. It would be highly inadvisable to transpose the results obtained from the cleaning of granite that has the same mineralogy and texture, but has a different degree of deterioration. However, considering that granites similar to the used in this research are commonly found in the ancient architectural heritage (Romanesque, Baroque, and Gothic) of the Western (W) Europe due to their greater ease of carving and dimensioning, the methodology of this paper would be potentially used to characterize the impact of the different wavelengths before to proceed with the cleaning.