- Email: [email protected]
Laser cleaning and laser-induced breakdown spectroscopy applied in removing and characterizing black crusts from limestones of Castello Svevo, Bari, Italy: A case study
Microchemical Journal 124 (2016) 296–305
Contents lists available at ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
Laser cleaning and laser-induced breakdown spectroscopy applied in removing and characterizing black crusts from limestones of Castello Svevo, Bari, Italy: A case study☆ G.S. Senesi a,⁎, I. Carrara b, G. Nicolodelli c, D.M.B.P. Milori c, O. De Pascale a a b c
Istituto di Nanotecnologia (NANOTEC), CNR, 70126 Bari, Italy Impresa Ing. Antonio Resta e C. Srl, 70124 Bari, Italy Embrapa Instrumentation, 13560-970 Sao Carlos, SP, Brazil
a r t i c l e
i n f o
Article history: Received 15 July 2015 Received in revised form 11 September 2015 Accepted 15 September 2015 Available online 21 September 2015 Keywords: Monument restoration Laser cleaning LIBS Black crusts Limestones Depth profile
a b s t r a c t The laser cleaning process combined with laser-induced breakdown spectroscopy (LIBS) were applied to restore and characterize altered limestones of the ancient jamb of the historic entrance gate of Castello Svevo, Bari, Italy. This area of the masonry blocks of the limestone castle was chosen because of its evident degradation with an apparent deposit of black crusts. The combination of a Q-switched Nd:YAG pulsed laser with the diagnostics typical of the LIBS technique was shown to be very effective for monitoring, controlling and characterizing the laser cleaning process of limestone. The different elemental compositions of the black encrustations covering the stone surface and the underlying stone allowed to evaluate and avoid over-cleaning and/or under-cleaning. Further, coupling LIBS to the cleaning process provided important information about the optimal experimental conditions to be used for evaluating the conservation status and determining the most proper cleaning restoration procedure before operating the consolidation of the blocks. Thus a sufficient removal of unwanted layers could be achieved without modifying the surface underneath and ameliorating the effectiveness of traditional cleaning techniques. In this work, the elemental composition of the ablated black crust and the underlying stone were determined by the spectroscopic study of plasma emitted from either a single pulse (SP) or a double pulse (DP) LIBS configuration. With respect to SP LIBS, a marked enhancement of the signal emission was observed by DP-LIBS used after a previous stratigraphic DP-LIBS assessment of the cleaning depth. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Laser radiation applied to the cleaning of decayed stones was first investigated on several Carrara marble sculptures in Florence and Venice in a pioneering work by Asmus et al. [1]. Since then laser based techniques have been applied widely for cleaning art objects of various natures, including stones, marbles, metal sculptures, stained glasses, paintings, and icons [2–5]. The cleaning process aims at the controlled removal of contaminant layers, e.g., crusts, soiling patinas, and stains, by mechanical and thermal ablation using nanosecond laser pulses focused on the sample surface. These deposits consist of microparticles of various natures, including fly ashes and fine soots originating from urban traffic and industrial activities, pollens, spores, windborne seeds, generally cemented by gypsum. Thus the material properties, e.g., its absorptivity, roughness, mechanical properties, affect markedly the process [2]. ☆ Selected papers presented at TECHNART 2015 Conference, Catania (Italy), April 27–30, 2015. ⁎ Corresponding author. Fax: +39 0805929505. E-mail address: [email protected] (G.S. Senesi).
http://dx.doi.org/10.1016/j.microc.2015.09.011 0026-265X/© 2015 Elsevier B.V. All rights reserved.
During the cleaning process the high energy laser pulses are absorbed significantly by the dark-colored surface encrustations, whereas the underlying stone, generally white to yellow-colored, is prone to reflect the incident radiation. The very fast processes, of the order of a few billionths of second, occurring during the laser–material interaction depend markedly from irradiation parameters, e.g., wavelength, energy density, and pulse duration [2,6,7]. In particular, short laser pulses in the nanosecond range are likely to produce mechanical damages, microfragmentation and increased porosity of the material surface, whereas longer duration pulses, up to the millisecond range, are expected to induce larger thermal modifications of the surface [6, 8]. For example, the temperature rise of crusts and underlying material impinged by pulsed laser irradiation can be predicted by a few quantitative models [2]. Even though the mechanisms involved are not fully understood, the removal of crusts is ascertained to result from a combination of thermal and mechanical interactions between the absorbed radiation and the irradiated surface [2]. In particular, thermal interaction has been shown to produce a fast, very localized and brief rise of temperature on the material surface, which determines its melting and vaporization. This is followed by the formation of plasma at a temperature of several
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
297
thousand degrees, whereas no significant transfer of heat to the stone underneath, i.e., no high-temperature mineral phase transition, has been observed [9]. The dynamic expansion of the plasma formed then generates mechanical shock waves that break down the material dispersing particles of various sizes. Further, the laser induces the explosive vaporization of the water possibly present in the crusts, thus intensifying the mechanical effects and increasing the particle dispersal rate [9]. Several factors, such as the pulse duration, the nature and optical, mechanical, thermal and chemical properties of the material and the presence of water, affect the extent of the thermal versus mechanical effects, and thus the effectiveness of the ablation process [9]. In general, laser cleaning is a very precise and progressive process able to remove layers of few microns for each laser pulse, i.e., it follows the microstratigraphy of the altered layers and can be interrupted at any defined stratigraphic level. A critical aspect to obtain a successful cleaning by any type of cleaning methodology is to determine at which surface depth the process, i.e., the thermal, mechanical and chemical effects operating for the undesired layers removal, must be interrupted. Thus, very weak and highly altered surfaces can be treated successfully. Besides the actual surface cleaning performance, the laser technique coupled with a spectroscopic system, i.e., laser induced breakdown spectroscopy (LIBS), is able to perform also the evaluation of the extent of contaminant layers removal. This by measuring the distinct differences between the emission spectra of the contaminant layers and those on the sample surface beneath, thus defining properly the end point of cleaning [10–12]. This specific performance adds to the well known advantages of LIBS, which include no sample preparation, destruction of sample areas of the order of few hundreds of microns, automation, selectivity, versatility and a high degree of precision. In particular, despite the higher technological complexity and costs, multiwavelength Q-switched Nd:YAG laser systems have been used for approaching and solving the problem of the yellow appearance in cleaning whitish substrates [13–21]. One of the most serious problems facing conservation is the deterioration of carbonate stones used in construction, as a consequence of the constant exposure to decay mechanisms. Where stone is severely weakened, some form of consolidation may be necessary to reduce the rate of decay and to reinforce the stone's cohesion. However, even if the use of consolidant products such as exfoliants, and sprays, might stop the degradation, on the other hand it could favor a subsequent fixation of even more dirt on the stones, thus compromising the achievement of an optimal cleaning. The main aim of this study was to confirm the effectiveness of laser cleaning and LIBS microdestructive stratigraphic analyses by applying an innovative approach for dirt removal before performing the preconsolidation by adhesion and cohesion of the calcareous stones and quoins of the left jamb of the southern entrance gate to the courtyard of Castello Svevo, Bari, Italy. A qualitative chemical analysis of the bulk samples and their black encrustations was carried out using single pulse (SP) and double pulse (DP) LIBS configurations. In particular, DP LIBS depth profile analyses were performed on the limestone sample covered by black crusts in order to identify the gradual decrease with depth of some specific elements strictly related to the patina, including Fe, Co, Ti, Al and Si.
followed by several successive building phases from the twelfth to the sixteenth century. The area selected showed a very complex surface and structural degradation of limestone blocks (Fig. 1). In particular, the surface degradation featured a layer of black crust and incoherent and coherent surface deposits, whereas the structural degradation was characterized by several cracks with solutions of continuity in the quoins and widespread spalling with irregular shapes and consistent thickness. Some spallings showed the total detachment of the stone material with subsequent fall and loss, whereas others showed a partial detachment which needed proper consolidation. Several areas of disintegration and erosion appeared, especially where the black crust experienced a structural change with the passage of calcium carbonate to gypsum which resulted in the pulverization of the constituent material of the building stone. The experimental samples collected in the central area of the masonry showed apparent fractures and detachment of material, which resulted from the heavy load crushing probably accentuated by the mechanical properties of stone blocks. The masonry featured different types of limestone, thus the various quoins showed different compactness and porosity. Officials of the Superintendence of Fine Art and Landscape for the Provinces of Bari, Foggia Barletta–Andria–Trani have suggested the possibility that an old protective treatment (“scialbaturas”) would have been present under the thick layer of the black crust. In order to decide the degree of cleaning and the procedures to be used for conservation, four areas of the blocks showing a macroscopic different texture and firmness were sampled for analysis (Fig. 2).
2. Material and methods
2.4. Laser induced breakdown spectroscopy
2.1. Sample description and sampling procedure
The LIBS technique was used to analyze the concentrations of key elements related to the formation of alteration layers at various depths and in the limestone underneath. The LIBS system used consisted of two Nd:YAG lasers operating at different wavelengths, i.e., 1064 nm (IR) and 532 nm (VIS) (Fig. 3b). The IR pulse was generated by a Nd:YAG Q-switch Ultra (Quantel) at a maximum energy of 75 mJ, and a width of 6 ns. The VIS pulse was generated by a Nd:YAG Q-switch Brillant
The sample studied was a selected area of the limestone masonry blocks of the left jamb of the southern entrance gate to the courtyard of Castello Svevo, Bari, Italy. The castle is an historic multilayered building featuring a quadrangular shape and square towers built originally by Roger the Norman in 1131 on the remains of a Byzantine structure, and
2.2. Optical and scanning electron microscopy and X-ray diffraction analysis In order to characterize the sequence of layers, the encrustation/ limestone interface and the textures of each sample, thin polished cross sections of the four samples were prepared and examined by an optical polarizing microscope (ZEISS Axioskop 40 POL) equipped with a micrometer particle size device and a digital camera. The depth crater image was obtained by a JEOL (JSM-6510, Thermo Scientific) scanning electron microscope (SEM). X-ray diffraction analyses were conducted on micro-samples using a Philips powder diffractometer PANalytical Pro PDM (CuK, 40 kV, 40 mA) equipped with an X'Celerator detector. Sampling of layers showing a coarse texture was performed using the finer portion enriched in binder. Diffraction spectra were acquired in the angular continuous scanning mode from 3° to 70° of 2θ at a speed of 0.07° 2θ/s. 2.3. Laser cleaning The cleaning was performed by a portable pulsed Nd:YAG Q-switched laser mod. Thunder Art of Quanta System at a pulse width of 8 ns (Fig. 3a) is equipped with a multi-articulated arm with seven folding mirrors that enable the laser beam to travel into an aluminum tube. The choice of the optimal experimental parameters for the laser was made after comparison of results obtained using two different wavelengths, i.e., 1064 nm and 532 nm, different pulse repetition frequencies (from 10 Hz to a maximum of 20 Hz), and varying the energy (maximum energy per pulse, 900 mJ at the wavelength 1064 nm and 400 mJ at 532 nm).
298
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
Fig. 1. Surface and structural degradation of the limestone blocks of castle masonry.
(Quantel) coupled with a second harmonic generator module at a maximum energy of 180 mJ and a width of 4 ns. A 400-Butterfly Aryelle system was used to detect and select the wavelengths. The spectrometer, equipped with an intensified charge-coupled device (ICCD) camera with 1024 × 1024 pixels, operated in two spectral bands, 175–330 nm and 275–750 nm, with a spectral resolution of 13–24 pm and 29–80 pm, respectively. The beams from the two lasers were directed to the target sample by dichroic mirrors at appropriate wavelengths (532 nm and 1064 nm). Lens of focal length of 100 mm and anti-reflecting coating (532 nm and 1064 nm) were used at the best of laser optical energy and placed between the sample and the tip
Fig. 2. The position of samples examined on the limestone block of the castle masonry.
Fig. 3. The laser cleaning system (a) and LIBS set-up (b).
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
299
Fig. 4. Limestone samples analyzed by optical microscopy: sample 1, sample 2, sample 3 and sample 4 (a; b; c; d).
of the fiber for efficient collection of the emitted plasma. The sample support was placed in a micro-controlled xy stage for an easy and fast scanning of the impinging laser beam. A pulse generator with eight channels (Quantum Composers Manufacturer, model 9618) was used
to synchronize the delay time between pulses and the delay detection during the experiments. The collinear geometry assembly with two laser beams each with energy of 45 mJ was used for the acquisition of DP LIBS spectra. The
Fig. 5. The black crusts of a thickness ranging from 0.1 to 0.8 mm, in contact with the limestone surface. Sample 1, sample 2, sample 3 and sample 4 (a; b; c; d).
300
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
Fig. 6. The black crusts consisting of a fine and opaque particulate component combined with quartz silt and bound by crypto- to micro-crystalline gypsum mixed with calcite. Sample 1, sample 2, sample 3 and sample 4 (a; b; c; d).
beams were focused and aligned to hit the sample in the overlapping mode with a delay between them. The temporal parameters used in this experiment were optimized for the best LIBS signal, i.e., the delay time was 500 ns, the gate time 10 μs and the interpulse 500 ns. The 532-nm laser of pulse energy fixed at 90 mJ was used for the acquisition of SP LIBS spectra. Two distinct experiments were performed, the first one aiming to compare the performance of SP and DP LIBS configurations, and the other one to study the crater profile/depth. In the first experiment, a total of 20 measurements, 10 with SP LIBS and 10 with DP LIBS, were performed on the cleaned surface of limestone. Each measure was performed by exploring the sample surface at different positions. In the second experiment two profiles of the black crust surface of the limestone were analyzed with DP LIBS by accumulating 28 shots at each different position of sample, and obtaining one spectrum for each shot. The emission background of the LIBS spectra was corrected by subtracting the average noise region near the element emission line. Results were based on the height of the peaks obtained from the Lorentzian fit for one peak in each spectrum, and then by averaging the heights. 3. Results and discussion 3.1. Petrographic and mineralogical analyses The purpose of the preliminary petrographic, mineralogical, textural and chemical analyses was to recognize the petrographic nature, the type of degradation and the possible presence of “scialbaturas” of the masonry before operating the cleaning process. The four representative limestone samples (Fig. 2) analyzed by optical microscopy show similar petrographic characteristics, although some compositional and textural differences are distinguishable. Sample 1 is a limestone with a packstone texture consisting mainly of benthic foraminifera and peloids embedded in a micritic cement (Fig. 4a). Sample 2 shows the most abundant allochemical peloids,
followed by benthic foraminifera (Fig. 4b), thus it can be classified as a pelbiomicritic packstone. The texture of the original rock of sample 3 (Fig. 4c) was apparently canceled by a dolomitization process, as shown by the typical lozenge form of dolomite crystals that were then replaced by recrystallization from pseudomorfa microcrystalline calcite (dedolomitization). Sample 4 shows petrographic features similar to sample 1 (Fig. 4d), and can be classified as a biopelmicritic packstone. Chemical and textural alterations are evident on all analyzed samples. Phenomena of dissolution and recrystallization of calcite are apparent on the surface exposed to the atmosphere (Fig. 5a–d). Further, the presence of black crusts with a thickness ranging from 0.1 to 0.8 mm can be recognized in contact with the stone surface (Fig. 5a–d). The black crusts consist of a fine and opaque particulate component combined with quartz silt (Fig. 6a–d) and bound by crypto- to microcrystalline gypsum mixed with calcite, together with deposits of soot and dust. No presence of any kind of plastering (“scialbaturas”) can be detected. In conclusion, petrographic and mineralogical analyses allow to recognize three types of limestone rocks, i.e., biopelmicritic packstone for samples 1 and 4, pelbiomicritic packstone for sample 2 and dedolomitizated limestone for sample 3, with no layers of protective treatment in any case, and the presence of black crusts in all cases. All evidences from petrographic and mineralogical analyses indicate that the types of limestone rocks used to build the castle wall were
Table 1 Semi-quantitative mineralogical content of the samples analyzed by XRPD. Sample
Gypsum
Calcite
Quartz
1 2 3 4
xx xxxx xxxx xxxx
xxx xx xxx xx
tr tr tr
Legend: x–xxxx = relative amounts; tr = trace.
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
301
Fig. 7. XRD patterns of black crusts from the limestone samples 1, 2, 3 and 4 (a; b; c; d).
caved from the about 150-m thick Mesozoic succession that represents a portion of the Formation named “Calcare di Bari”, and consists of biopeloidal and peloidal wackestones/packstones alternated with stromatolitic bindstones with frequent intercalations of dolomitic limestones and gray dolostones [22]. The X-ray diffraction spectra measured on the powders of the alteration patina on the sample surfaces exposed to the atmosphere show in all samples the presence of gypsum and secondary calcite with traces of quartz as part of the particulate cemented by gypsum (Table 1) (Fig. 7a–d). These results confirm that the formation of the black crust was mainly due to air pollution that caused black soiling and loss of structural integrity due to the attack by atmospheric sulfur and subsequent formation of gypsum [23].
(10 Hz to a maximum of 20 Hz) and varying the energy (maximum energy per pulse, 900 mJ at the wavelength 1064 nm, 400 mJ at 532 nm). After these tests, the officials of the Superintendence of Fine Art and Landscape for the Provinces of Bari, Foggia Barletta–Andria–Trani preferred to use the less invasive result/effect of cleaning, and as far as possible with a color/aspect of cleaned surface similar to the old original limestone. Thus the approach based on test 2, i.e., irradiation at 1064 nm by a Q-switched Nd:YAG laser, was chosen for removing the black encrustations resulting from atmospheric pollution from limestone
3.2. Laser cleaning tests Three different procedures were applied to test the level of cleaning achieved. The first test was performed at the wavelength of 532 nm (zone 1), the second one at the wavelength of 1064 nm (zone 2), and a third one by applying two consecutive wavelengths, first at 1064 nm and then at 532 nm (zone 3) (Fig. 8). The three tests show a different level of cleaning with an increasing of whitening when two wavelengths are used. The choice of parameters for performing the cleaning process was based on the comparison of results obtained from the use of the three different approaches using different pulse repetition frequencies
Fig. 8. Comparative cleaning tests on a masonry block using different wavelengths: 532 nm (1); 1064 nm (2), and first 1064 nm and then 532 nm (3).
302
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
materials, which has the advantages of not damaging the underlying stone and possibility of control the degree of the cleaning action. This type of laser provides pulses of typically few ns of 1064 nm radiation, which are preferentially absorbed by the stone soiled layers, thus producing a rapid increase of the surface temperature of the black crust which allows its removal. However, further irradiation with the second harmonic (532 nm) of the Q-switched Nd:YAG laser of the areas previously cleaned with the fundamental radiation results in a more complete removal of dust particles.
3.3. Laser induced breakdown spectroscopy (LIBS) Sample 4 was chosen for the following LIBS analysis based on results of petrographic and mineralogical analyses, and because the limestone masonry blocks under study are constituted mainly of biopelmicritic packstone. The surface of sample 4 was divided in three areas that were subjected to different treatments and analyses. Areas 1 and 2 (Fig. 9a) were first cleaned by removing the black crust using either one wavelength (1064 nm) or two successive wavelengths (1064 nm and then 532 nm), respectively. Then the cleaned limestone surfaces were analyzed along two lines (squares in Fig. 9a) either in the SP mode (1064 nm) (area 1) or in the DP mode (1064 nm and then 532 nm) (area 2). Area 3, i.e., the limestone sample covered by black crusts (Fig. 9a), was subjected to depth profile analyses using the DP configuration in collinear LIBS mode in order to identify the disappearance, or decrease with depth, of some specific elements typical of the patina. The crater depth after 28 shots was ~ 310 μm, i.e., 11 μm per shot (Fig. 9b). The emission lines of the elements detected (Al, Ba, C, Ca, Co, Fe, K, Li, Mg, Mn, Na, Si, Sr, and Ti) by qualitative LIBS analysis of the three sample areas 1, 2, and 3, are listed in Table 2. The emission line intensities using the DP configuration were about 5 times higher than those obtained by SP LIBS (Fig. 10), which can be attributed to a combination of increased
Fig. 9. SP and DP LIBS characterization (area 1 and area 2) and DP depth profile (area 3), (a); SEM image of the crater depth after 28 shots (b).
Table 2 Spectroscopic atomic emission lines (NIST, KURUCZ) of samples examined. Element Wavelength (nm) Co II Fe II Si I Al I Mn I Mn II KI CI Sr I Ti II Mg I Mg II Ca I
Li I Ba II Na I
233.79; 235.11; 240.47 234.33; 234.41; 234.52; 234.81; 235.91; 259.93; 261.18; 273.95; 274.91; 275.57 288.17; 250.69; 251.43; 251.61; 251.92; 252.41; 252.85; 390.55 308.21; 309.27 403.08; 403.31; 403.45; 403.57; 404.14 257.61; 259.37 766.49; 769.89 193.09; 247.85 460.73 323.45; 323.65; 323.90; 324.19; 336.12; 337.27; 338.37 280.98; 285.21; 383.23; 383.83; 516.73; 517.27; 518.36 279.55; 280.27 299.73; 300.09; 300.68; 300.94; 428.30; 428.94; 429.90; 430.78; 431.87; 442.55; 534.95; 558.22; 558.89; 559.02; 559.47; 559.87; 560.16; 560.29; 585.74; 671.72; 720.22 670.77 455.40 588.99; 589.59
laser ablation and plume re-heating. Thus the DP LIBS mode should be preferred to the SP mode in these studies. Fig. 11 shows the LIBS spectra obtained for the encrustation (area 3, Fig. 9a) at three different pulse numbers/depth (1, 10 and 20) which correspond to three depths from the surface. In the spectral range employed (250–330 nm), the emission lines detected confirm the presence of Ca and Mg as the major constituents of limestone and Si, Al, Fe, Co, Mn, Ti as the main contaminant elements that disappear with increasing the pulse number, i.e., the depth. The relative intensities of Ca I (534.95 nm), Mg I (285.21 nm), Si I (288.21 nm), Al I (308.21 nm), Fe II (234.41 nm), Ti II (323.65 nm), Sr I (460.73 nm), Ba II (455.40 nm), Li I (670.77 nm), Na I (588.99 nm), K I (769.89 nm), C I (247.85 nm) and Co II (240.47 nm) detected by the in-depth LIBS profile for the black crust with increasing the number of pulses are shown in Fig. 12. The trends of LIBS profiles of Si, Al, Fe, Ti, Ba, Li and Co, show a marked decrease after the first five–ten pulses (around 60–110 μm depth) (Fig. 12a). After ten pulses Fe, Co, Ti and Li almost disappear, i.e., are almost undetectable, whereas the intensities of Si, Al and Ba emissions remain low and almost constant down to 200 μm depth (Fig. 12a). LIBS intensities of Ca, Mg, K, Na, C and Sr show different trends, i.e., those of Ca and Sr increase with increasing depth, whereas those of Mg, K and C first decrease up to five–ten pulses, than increase and remain almost constant until the bottom, and that of K remains constant always (Fig. 12b). In particular, the increase shown by C with respect to Mg up to 5 pulses suggests an additional possible
Fig. 10. SP and DP LIBS emission spectra in the region 232–247 nm of the black crust from sample 4.
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
303
origin of C from organic contaminants in the upper layers of the black crust. Limestone and dolomite contain the highest Na concentrations, up to 5.400 mg kg−1 Na, of all sedimentary rocks, which is attributed to the effect of sea water during their formation and the presence of skeletal material [24]. In particular salt weathering in coastal areas of Apulia can be considered the most important cause of stone decay, especially where historic monuments and buildings are close to the coastline and partially or completely submerged by sea water. The calcarenites are soft and porous, and suffer much from deterioration as a result of weathering [25,26]. The type and rate of weathering of this building stone depend on the petrographic and physical properties (fabric, porosity, etc.), position within a monument or building, geographical location and air pollution level [27]. The relative intensities of atomic LIBS emission lines of Ca I (534.95 nm) and Mg I (285.25 nm) are shown in the inset of Fig. 12b. The results suggest that Ca, Mg, Na, Sr, K and C in the black crusts are mainly related to the composition of the limestone underneath, Si, Al, Ba and Li may be ascribed mostly to the soil–dust deposition, whereas Fe, Co and Ti would originate mainly from other sources, such as atmospheric pollution. The size of the Sr2+ ion (118 pm) is intermediate between those of 2+ Ca (100 pm) and K+ (138 pm), which may be replaced in a variety of rock forming minerals including K-feldspar, gypsum, plagioclase and, especially, calcite and dolomite. Enrichment of Sr up to concentrations of ca. 1000 mg kg−1 is, therefore, common in limestone and evaporates, although the Sr/Ca ratio in most types of limestone is less than 1:1000 [28]. Sr is easily mobilized during weathering, especially in oxidizing acid environments.
Fig. 12. Relative intensities of atomic and ionic LIBS emission lines of Si I 288.17 nm, Al I 308.21, nm, Fe II 234.41 nm, Ba II 455.40 nm, Li I 670.77 nm, Co II 240.47 nm and Ti II 323.65 nm (a) and Mg I 285.21 nm, Sr I 460.73 nm, K I 769.89 nm, Na I 588.99 nm, Ca I 534.95 nm and C I 247.85 (b) vs. number of pulses/depth. The relative intensities of atomic LIBS emission lines of limestone of Ca I (534.95 nm) and Mg I (285.25 nm) are shown in the inset of Fig. 12b.
Due to the high concentration of Ca and Mg across the encrustation relatively to that of the other elements detected, their emission intensities have been normalized with respect to the intensity of the Ca I line at 534.95 nm and Mg I line at 285.21 nm (Fig. 13a, b). With increasing the number of pulses, i.e., the depth, the emission of all elements decreases continuously with respect to Ca whereas with respect to Mg the emissions increase and then decrease, except for Ca. The decreasing trend of the Mg/Ca ratio with increasing depth (Fig. 14) suggests the increasing dedolomitazion of the ablated material [29,30].
4. Conclusions
Fig. 11. LIBS spectra of the ablated material from a black crust at different pulses (1, 10, 20).
The laser cleaning process was confirmed to be appropriate and efficient for achieving the removal of unwanted layers from the surface of limestone without modifying the surface morphology of the rock underneath. The optimal irradiation conditions causing minimal damaging effects were defined by both surface analysis and optimization of the Nd:YAG laser parameters (wavelength, pulse repetition frequencies, energy, etc.). The use of laser cleaning before the pre-consolidation operations allowed to prevent the entrapping of the dirty layers in the limestone rocks without any further possibility to remove it further on. LIBS was confirmed to be a powerful diagnostic technique able to monitor and control the laser cleaning process of limestone, thus allowing to replace traditional invasive laboratory analyses and provide a prompt compositional response in situ.
304
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
Fig. 13. Ratio of the intensities of atomic and ionic LIBS emission lines of Fe II 234.41 nm, Co II 240.47 nm, Si I 288.17 nm, Al I 308.21 nm, Mg I 280.98 nm, C I 193.09 nm and Ti II 323.65 nm and the intensity of Ca I line at 534.95 nm (a); and of Fe II 234.41 nm, Co II 240.47 nm, Si I 288.17 nm, Al I 308.21 nm, C I 193.09 nm, Ca I 534.95 nm and Ti II 323.65 nm and the intensity of Mg I line at 285.21 nm (b) vs. number of pulses/depth.
In particular, LIBS provided an evaluation of the black crust by confirming the absence of any protective layers (“scialbatures”) on the quoins of the masonry of the Castello Svevo, Bari, Italy. A comparison of SP and the DP LIBS configurations showed that the emission line intensities obtained by DP LIBS were about 5 times higher than those obtained by SP LIBS, which might be attributed to a combination of increased laser ablation and plume re-heating. The DP LIBS configuration stratigraphy was also used in order to assess the optimal cleaning depth, i.e., the number of pulses that should be applied to avoid under-cleaning and/or over-cleaning. The black crust layer (approximately 150-μm deep) on the limestone analyzed was completely removed after about 10 double pulses, which added to the common well-known advantages of laser-based cleaning methods, including automatization, selectivity, versatility, and a high degree of precision.
Acknowledgments The authors kindly acknowledge the financial funding received under the project “Il restauro delle grandi opere in Puglia: l'innovazione attraverso le nanotecnologie e metodologie diagnostiche avanzate”, P.O. Puglia FESR 2007–2013, Bando “Aiuti a Sostegno dei Partenariati Regionali per l'Innovazione” (3Z3VZ46).
Fig. 14. Ratio of the intensities of atomic LIBS emission lines of Mg I line at 285.21 nm and Ca I line at 534.95 nm vs. number of pulses/depth starting from the black crust surface (squares), and from underlying stone (dots).
References [1] J.F. Asmus, M. Seracini, M.J. Zetler, Surface morphology of laser-cleaned stone, Lithoclastia 1 (1976) 23–46. [2] M. Cooper, Laser Cleaning in Conservation: An Introduction, ButterworthHeinemann, Oxford, 1998. [3] R. Pini, R. Salimbeni, Tecniche e sistemi laser per il restauro dei beni culturali, Cardini Editore, Firenze, 2001. [4] C. Fotakis, D. Anglos, V. Zafiropulos, S. Georgiou, V. Tornari, Lasers in the preservation of cultural heritage: principles and applications, Series in Optics and Optoelectronics, CRC Press, Taylor & Francis Group, Boca Raton, 2007. [5] S. Siano, J. Agresti, I. Cacciari, D. Ciofini, M. Mascalchi, I. Osticioli, A.A. Mencaglia, Laser cleaning in conservation of stone, metal, and painted artifacts: state of the art and new insights on the use of the Nd:YAG lasers, Appl. Phys. A 106 (2012) 419–446. [6] L. Bartoli, P. Pouli, C. Fotakis, S. Siano, R. Salimbeni, Characterization of stone cleaning by Nd:YAG lasers with different pulse duration, Laser Chem. 2006 (2006) 1–6 (Article ID 81750). [7] S. Siano, M. Giamello, L. Bartoli, A. Mencaglia, V. Parfenov, R. Salimbeni, Laser cleaning of stone by different laser pulse duration and wavelength, Laser Phys. 18 (2008) 27–36. [8] C. Rodriguez-Navarro, K. Elert, E. Sebastian, R.M. Esbert, C.M. Grossi, A. Rojo, F.J. Alonso, M. Montoto, J. Ordaz, Laser cleaning of stone materials: an overview of current research, Rev. Conserv. 4 (2003) 65–82. [9] S. Siano, Principles of laser cleaning in conservation, in: M. Schreiner, M. Strlic (Eds.),Handbook on the Use of Lasers in Conservation and Conservation Science, COST G7 2007, pp. 1–26. [10] I. Gobernado-Mitre, A.C. Prieto, V. Zafiropulos, Y. Spetsidou, C. Fotakis, On-line monitoring of laser cleaning of limestone by laser-induced breakdown spectroscopy and laser-induced fluorescence, Appl. Spectrosc. 51 (1997) 1125–1129. [11] P. Maravelaki-Kalaitzaki, V. Zafiropulos, C. Fotakis, Excimer laser cleaning of encrustation on Pentelic marble: procedure and evaluation of the effects, Appl. Surf. Sci. 148 (1999) 92–104. [12] P. Maravelaki-Kalaitzaki, D. Anglos, V. Kilikoglou, V. Zafiropulos, Compositional characterization of encrustation on marble with laser induced breakdown spectroscopy, Spectrochim. Acta B 56 (2001) 887–903. [13] P. Bromblet, M. Labouré, G. Orial, Diversity of the cleaning procedures including laser for the restoration of carved portals in France over the last 10 years, J. Cult. Herit. 4 (2003) 17–26. [14] S. Siano, A. Casciani, A. Giusti, M. Matteini, R. Pini, S. Porcinai, R. Salimbeni, The Santi Quattro Coronati: cleaning of the gilded decorations, J. Cult. Herit. 4 (2003) 123–128. [15] V. Vergès-Belmin, C. Dignard, Laser yellowing: myth or reality? J. Cult. Herit. 4 (2003) 238–244. [16] S. Siano, A. Giusti, D. Pinna, S. Porcinai, M. Giamello, G. Sabatini, R. Salimbeni, The conservation intervention on the Porta della Mandorla, in: K. Dickmann, C. Fotakis, J.F. Asmus (Eds.), Lacona V Proceedings, Springer-Verlag, Berlin 2005, pp. 171–178. [17] J. Kolar, M. Strlič, D. Müller-Hess, A. Gruber, K. Troschke, S. Pentzien, W. Kautek, Laser cleaning of paper using Nd:YAG laser running at 532 nm, J. Cult. Herit. 4 (2003) 185–187. [18] K. Ochocińska, A. Kamińska, G. Śliwiński, Experimental investigations of stained paper documents cleaned by Nd:YAG laser pulses, J. Cult. Herit. 1 (2000) 188–193. [19] M.C. Gaetani, U. Santamaria, The laser cleaning of wall paintings, J. Cult. Herit. 1 (2000) S199–S207.
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305 [20] V. Zafiropulos, P. Pouli, V. Kylikoglou, P. Maravelaki-Kalaitzaki, B.S. Lukyanchuk, A. Gogariu, Synchronous use of IR an UV laser pulses in the removal of encrustation: mechanistic aspects, discoloration phenomena and benefits, in: K. Dickmann, C. Fotakis, J.F. Asmus (Eds.), Lacona V Proceedings, Springer-Verlag, Berlin 2005, pp. 311–318. [21] P. Pouli, K. Frantzikinaki, E. Papakonstantinou, V. Zafiropulos, Pollution encrustation removal by means of combined ultraviolet and infrared laser radiation: the application of this innovative methodology on the surface of the Parthenon west frieze, in: K. Dickmann, C. Fotakis, J.F. Asmus (Eds.), Lacona V Proceedings, Springer-Verlag, Berlin 2005, pp. 333–340. [22] P. Pieri, L. Sabato, L. Spalluto, M. Tropeano, Notes to the geological map of the urban area of Bari (southern Italy) 1:25,000 scale, Rend. Online Soc. Geol. Ital. 14 (2011) 26–36. [23] S.S. Potgieter-Vermaak, R.H.M. Godoi, R. Van Grieken, J.H. Potgieter, M. Oujja, M. Castillejo, Micro-structural characterization of black crust and laser cleaning of building stones by micro-Raman and SEM techniques, Spectrochim. Acta A 61 (2005) 2460–2467.
305
[24] G.K. Billings, P.C. Ragland, Geochemistry and mineralogy of the Recent reef and lagoonal sediments south of Belize (British Honduras), Chem. Geol. 3 (1968) 135–153. [25] G.F. Andriani, N. Walsh, Physical properties and textural parameters of calcarenite rocks: qualitative and quantitative evaluations, Eng. Geol. 67 (2002) 5–15. [26] G.F. Andriani, N. Walsh, Fabric, porosity and water permeability of calcarenites from Apulia (SE Italy) used as a building and ornamental stone, Bull. Eng. Geol. Environ. 62 (2003) 77–84. [27] S. Ordonez, R. Fort, M.A. Garcia del Cura, Pore size distribution and the durability of a porous limestone, Q. J. Eng. Geol. 30 (1997) 221–230. [28] J.L. Kulp, K. Turekian, D.W. Boyd, Strontium content of limestones and fossils, Geol. Soc. Am. Bull. 63 (1952) 701–716. [29] R.L. Folk, The natural history of crystalline calcium carbonate: effect of magnesium content and salinity, J. Sediment. Petrol. 44 (1974) 40–53. [30] R.L. Folk, L.S. Land, Mg/Ca ratio and salinity: two controls over crystallization of dolomite, Am. Assoc. Pet. Geol. Bull. 59 (1975) 60–68.
Contents lists available at ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
Laser cleaning and laser-induced breakdown spectroscopy applied in removing and characterizing black crusts from limestones of Castello Svevo, Bari, Italy: A case study☆ G.S. Senesi a,⁎, I. Carrara b, G. Nicolodelli c, D.M.B.P. Milori c, O. De Pascale a a b c
Istituto di Nanotecnologia (NANOTEC), CNR, 70126 Bari, Italy Impresa Ing. Antonio Resta e C. Srl, 70124 Bari, Italy Embrapa Instrumentation, 13560-970 Sao Carlos, SP, Brazil
a r t i c l e
i n f o
Article history: Received 15 July 2015 Received in revised form 11 September 2015 Accepted 15 September 2015 Available online 21 September 2015 Keywords: Monument restoration Laser cleaning LIBS Black crusts Limestones Depth profile
a b s t r a c t The laser cleaning process combined with laser-induced breakdown spectroscopy (LIBS) were applied to restore and characterize altered limestones of the ancient jamb of the historic entrance gate of Castello Svevo, Bari, Italy. This area of the masonry blocks of the limestone castle was chosen because of its evident degradation with an apparent deposit of black crusts. The combination of a Q-switched Nd:YAG pulsed laser with the diagnostics typical of the LIBS technique was shown to be very effective for monitoring, controlling and characterizing the laser cleaning process of limestone. The different elemental compositions of the black encrustations covering the stone surface and the underlying stone allowed to evaluate and avoid over-cleaning and/or under-cleaning. Further, coupling LIBS to the cleaning process provided important information about the optimal experimental conditions to be used for evaluating the conservation status and determining the most proper cleaning restoration procedure before operating the consolidation of the blocks. Thus a sufficient removal of unwanted layers could be achieved without modifying the surface underneath and ameliorating the effectiveness of traditional cleaning techniques. In this work, the elemental composition of the ablated black crust and the underlying stone were determined by the spectroscopic study of plasma emitted from either a single pulse (SP) or a double pulse (DP) LIBS configuration. With respect to SP LIBS, a marked enhancement of the signal emission was observed by DP-LIBS used after a previous stratigraphic DP-LIBS assessment of the cleaning depth. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Laser radiation applied to the cleaning of decayed stones was first investigated on several Carrara marble sculptures in Florence and Venice in a pioneering work by Asmus et al. [1]. Since then laser based techniques have been applied widely for cleaning art objects of various natures, including stones, marbles, metal sculptures, stained glasses, paintings, and icons [2–5]. The cleaning process aims at the controlled removal of contaminant layers, e.g., crusts, soiling patinas, and stains, by mechanical and thermal ablation using nanosecond laser pulses focused on the sample surface. These deposits consist of microparticles of various natures, including fly ashes and fine soots originating from urban traffic and industrial activities, pollens, spores, windborne seeds, generally cemented by gypsum. Thus the material properties, e.g., its absorptivity, roughness, mechanical properties, affect markedly the process [2]. ☆ Selected papers presented at TECHNART 2015 Conference, Catania (Italy), April 27–30, 2015. ⁎ Corresponding author. Fax: +39 0805929505. E-mail address: [email protected] (G.S. Senesi).
http://dx.doi.org/10.1016/j.microc.2015.09.011 0026-265X/© 2015 Elsevier B.V. All rights reserved.
During the cleaning process the high energy laser pulses are absorbed significantly by the dark-colored surface encrustations, whereas the underlying stone, generally white to yellow-colored, is prone to reflect the incident radiation. The very fast processes, of the order of a few billionths of second, occurring during the laser–material interaction depend markedly from irradiation parameters, e.g., wavelength, energy density, and pulse duration [2,6,7]. In particular, short laser pulses in the nanosecond range are likely to produce mechanical damages, microfragmentation and increased porosity of the material surface, whereas longer duration pulses, up to the millisecond range, are expected to induce larger thermal modifications of the surface [6, 8]. For example, the temperature rise of crusts and underlying material impinged by pulsed laser irradiation can be predicted by a few quantitative models [2]. Even though the mechanisms involved are not fully understood, the removal of crusts is ascertained to result from a combination of thermal and mechanical interactions between the absorbed radiation and the irradiated surface [2]. In particular, thermal interaction has been shown to produce a fast, very localized and brief rise of temperature on the material surface, which determines its melting and vaporization. This is followed by the formation of plasma at a temperature of several
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
297
thousand degrees, whereas no significant transfer of heat to the stone underneath, i.e., no high-temperature mineral phase transition, has been observed [9]. The dynamic expansion of the plasma formed then generates mechanical shock waves that break down the material dispersing particles of various sizes. Further, the laser induces the explosive vaporization of the water possibly present in the crusts, thus intensifying the mechanical effects and increasing the particle dispersal rate [9]. Several factors, such as the pulse duration, the nature and optical, mechanical, thermal and chemical properties of the material and the presence of water, affect the extent of the thermal versus mechanical effects, and thus the effectiveness of the ablation process [9]. In general, laser cleaning is a very precise and progressive process able to remove layers of few microns for each laser pulse, i.e., it follows the microstratigraphy of the altered layers and can be interrupted at any defined stratigraphic level. A critical aspect to obtain a successful cleaning by any type of cleaning methodology is to determine at which surface depth the process, i.e., the thermal, mechanical and chemical effects operating for the undesired layers removal, must be interrupted. Thus, very weak and highly altered surfaces can be treated successfully. Besides the actual surface cleaning performance, the laser technique coupled with a spectroscopic system, i.e., laser induced breakdown spectroscopy (LIBS), is able to perform also the evaluation of the extent of contaminant layers removal. This by measuring the distinct differences between the emission spectra of the contaminant layers and those on the sample surface beneath, thus defining properly the end point of cleaning [10–12]. This specific performance adds to the well known advantages of LIBS, which include no sample preparation, destruction of sample areas of the order of few hundreds of microns, automation, selectivity, versatility and a high degree of precision. In particular, despite the higher technological complexity and costs, multiwavelength Q-switched Nd:YAG laser systems have been used for approaching and solving the problem of the yellow appearance in cleaning whitish substrates [13–21]. One of the most serious problems facing conservation is the deterioration of carbonate stones used in construction, as a consequence of the constant exposure to decay mechanisms. Where stone is severely weakened, some form of consolidation may be necessary to reduce the rate of decay and to reinforce the stone's cohesion. However, even if the use of consolidant products such as exfoliants, and sprays, might stop the degradation, on the other hand it could favor a subsequent fixation of even more dirt on the stones, thus compromising the achievement of an optimal cleaning. The main aim of this study was to confirm the effectiveness of laser cleaning and LIBS microdestructive stratigraphic analyses by applying an innovative approach for dirt removal before performing the preconsolidation by adhesion and cohesion of the calcareous stones and quoins of the left jamb of the southern entrance gate to the courtyard of Castello Svevo, Bari, Italy. A qualitative chemical analysis of the bulk samples and their black encrustations was carried out using single pulse (SP) and double pulse (DP) LIBS configurations. In particular, DP LIBS depth profile analyses were performed on the limestone sample covered by black crusts in order to identify the gradual decrease with depth of some specific elements strictly related to the patina, including Fe, Co, Ti, Al and Si.
followed by several successive building phases from the twelfth to the sixteenth century. The area selected showed a very complex surface and structural degradation of limestone blocks (Fig. 1). In particular, the surface degradation featured a layer of black crust and incoherent and coherent surface deposits, whereas the structural degradation was characterized by several cracks with solutions of continuity in the quoins and widespread spalling with irregular shapes and consistent thickness. Some spallings showed the total detachment of the stone material with subsequent fall and loss, whereas others showed a partial detachment which needed proper consolidation. Several areas of disintegration and erosion appeared, especially where the black crust experienced a structural change with the passage of calcium carbonate to gypsum which resulted in the pulverization of the constituent material of the building stone. The experimental samples collected in the central area of the masonry showed apparent fractures and detachment of material, which resulted from the heavy load crushing probably accentuated by the mechanical properties of stone blocks. The masonry featured different types of limestone, thus the various quoins showed different compactness and porosity. Officials of the Superintendence of Fine Art and Landscape for the Provinces of Bari, Foggia Barletta–Andria–Trani have suggested the possibility that an old protective treatment (“scialbaturas”) would have been present under the thick layer of the black crust. In order to decide the degree of cleaning and the procedures to be used for conservation, four areas of the blocks showing a macroscopic different texture and firmness were sampled for analysis (Fig. 2).
2. Material and methods
2.4. Laser induced breakdown spectroscopy
2.1. Sample description and sampling procedure
The LIBS technique was used to analyze the concentrations of key elements related to the formation of alteration layers at various depths and in the limestone underneath. The LIBS system used consisted of two Nd:YAG lasers operating at different wavelengths, i.e., 1064 nm (IR) and 532 nm (VIS) (Fig. 3b). The IR pulse was generated by a Nd:YAG Q-switch Ultra (Quantel) at a maximum energy of 75 mJ, and a width of 6 ns. The VIS pulse was generated by a Nd:YAG Q-switch Brillant
The sample studied was a selected area of the limestone masonry blocks of the left jamb of the southern entrance gate to the courtyard of Castello Svevo, Bari, Italy. The castle is an historic multilayered building featuring a quadrangular shape and square towers built originally by Roger the Norman in 1131 on the remains of a Byzantine structure, and
2.2. Optical and scanning electron microscopy and X-ray diffraction analysis In order to characterize the sequence of layers, the encrustation/ limestone interface and the textures of each sample, thin polished cross sections of the four samples were prepared and examined by an optical polarizing microscope (ZEISS Axioskop 40 POL) equipped with a micrometer particle size device and a digital camera. The depth crater image was obtained by a JEOL (JSM-6510, Thermo Scientific) scanning electron microscope (SEM). X-ray diffraction analyses were conducted on micro-samples using a Philips powder diffractometer PANalytical Pro PDM (CuK, 40 kV, 40 mA) equipped with an X'Celerator detector. Sampling of layers showing a coarse texture was performed using the finer portion enriched in binder. Diffraction spectra were acquired in the angular continuous scanning mode from 3° to 70° of 2θ at a speed of 0.07° 2θ/s. 2.3. Laser cleaning The cleaning was performed by a portable pulsed Nd:YAG Q-switched laser mod. Thunder Art of Quanta System at a pulse width of 8 ns (Fig. 3a) is equipped with a multi-articulated arm with seven folding mirrors that enable the laser beam to travel into an aluminum tube. The choice of the optimal experimental parameters for the laser was made after comparison of results obtained using two different wavelengths, i.e., 1064 nm and 532 nm, different pulse repetition frequencies (from 10 Hz to a maximum of 20 Hz), and varying the energy (maximum energy per pulse, 900 mJ at the wavelength 1064 nm and 400 mJ at 532 nm).
298
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
Fig. 1. Surface and structural degradation of the limestone blocks of castle masonry.
(Quantel) coupled with a second harmonic generator module at a maximum energy of 180 mJ and a width of 4 ns. A 400-Butterfly Aryelle system was used to detect and select the wavelengths. The spectrometer, equipped with an intensified charge-coupled device (ICCD) camera with 1024 × 1024 pixels, operated in two spectral bands, 175–330 nm and 275–750 nm, with a spectral resolution of 13–24 pm and 29–80 pm, respectively. The beams from the two lasers were directed to the target sample by dichroic mirrors at appropriate wavelengths (532 nm and 1064 nm). Lens of focal length of 100 mm and anti-reflecting coating (532 nm and 1064 nm) were used at the best of laser optical energy and placed between the sample and the tip
Fig. 2. The position of samples examined on the limestone block of the castle masonry.
Fig. 3. The laser cleaning system (a) and LIBS set-up (b).
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
299
Fig. 4. Limestone samples analyzed by optical microscopy: sample 1, sample 2, sample 3 and sample 4 (a; b; c; d).
of the fiber for efficient collection of the emitted plasma. The sample support was placed in a micro-controlled xy stage for an easy and fast scanning of the impinging laser beam. A pulse generator with eight channels (Quantum Composers Manufacturer, model 9618) was used
to synchronize the delay time between pulses and the delay detection during the experiments. The collinear geometry assembly with two laser beams each with energy of 45 mJ was used for the acquisition of DP LIBS spectra. The
Fig. 5. The black crusts of a thickness ranging from 0.1 to 0.8 mm, in contact with the limestone surface. Sample 1, sample 2, sample 3 and sample 4 (a; b; c; d).
300
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
Fig. 6. The black crusts consisting of a fine and opaque particulate component combined with quartz silt and bound by crypto- to micro-crystalline gypsum mixed with calcite. Sample 1, sample 2, sample 3 and sample 4 (a; b; c; d).
beams were focused and aligned to hit the sample in the overlapping mode with a delay between them. The temporal parameters used in this experiment were optimized for the best LIBS signal, i.e., the delay time was 500 ns, the gate time 10 μs and the interpulse 500 ns. The 532-nm laser of pulse energy fixed at 90 mJ was used for the acquisition of SP LIBS spectra. Two distinct experiments were performed, the first one aiming to compare the performance of SP and DP LIBS configurations, and the other one to study the crater profile/depth. In the first experiment, a total of 20 measurements, 10 with SP LIBS and 10 with DP LIBS, were performed on the cleaned surface of limestone. Each measure was performed by exploring the sample surface at different positions. In the second experiment two profiles of the black crust surface of the limestone were analyzed with DP LIBS by accumulating 28 shots at each different position of sample, and obtaining one spectrum for each shot. The emission background of the LIBS spectra was corrected by subtracting the average noise region near the element emission line. Results were based on the height of the peaks obtained from the Lorentzian fit for one peak in each spectrum, and then by averaging the heights. 3. Results and discussion 3.1. Petrographic and mineralogical analyses The purpose of the preliminary petrographic, mineralogical, textural and chemical analyses was to recognize the petrographic nature, the type of degradation and the possible presence of “scialbaturas” of the masonry before operating the cleaning process. The four representative limestone samples (Fig. 2) analyzed by optical microscopy show similar petrographic characteristics, although some compositional and textural differences are distinguishable. Sample 1 is a limestone with a packstone texture consisting mainly of benthic foraminifera and peloids embedded in a micritic cement (Fig. 4a). Sample 2 shows the most abundant allochemical peloids,
followed by benthic foraminifera (Fig. 4b), thus it can be classified as a pelbiomicritic packstone. The texture of the original rock of sample 3 (Fig. 4c) was apparently canceled by a dolomitization process, as shown by the typical lozenge form of dolomite crystals that were then replaced by recrystallization from pseudomorfa microcrystalline calcite (dedolomitization). Sample 4 shows petrographic features similar to sample 1 (Fig. 4d), and can be classified as a biopelmicritic packstone. Chemical and textural alterations are evident on all analyzed samples. Phenomena of dissolution and recrystallization of calcite are apparent on the surface exposed to the atmosphere (Fig. 5a–d). Further, the presence of black crusts with a thickness ranging from 0.1 to 0.8 mm can be recognized in contact with the stone surface (Fig. 5a–d). The black crusts consist of a fine and opaque particulate component combined with quartz silt (Fig. 6a–d) and bound by crypto- to microcrystalline gypsum mixed with calcite, together with deposits of soot and dust. No presence of any kind of plastering (“scialbaturas”) can be detected. In conclusion, petrographic and mineralogical analyses allow to recognize three types of limestone rocks, i.e., biopelmicritic packstone for samples 1 and 4, pelbiomicritic packstone for sample 2 and dedolomitizated limestone for sample 3, with no layers of protective treatment in any case, and the presence of black crusts in all cases. All evidences from petrographic and mineralogical analyses indicate that the types of limestone rocks used to build the castle wall were
Table 1 Semi-quantitative mineralogical content of the samples analyzed by XRPD. Sample
Gypsum
Calcite
Quartz
1 2 3 4
xx xxxx xxxx xxxx
xxx xx xxx xx
tr tr tr
Legend: x–xxxx = relative amounts; tr = trace.
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
301
Fig. 7. XRD patterns of black crusts from the limestone samples 1, 2, 3 and 4 (a; b; c; d).
caved from the about 150-m thick Mesozoic succession that represents a portion of the Formation named “Calcare di Bari”, and consists of biopeloidal and peloidal wackestones/packstones alternated with stromatolitic bindstones with frequent intercalations of dolomitic limestones and gray dolostones [22]. The X-ray diffraction spectra measured on the powders of the alteration patina on the sample surfaces exposed to the atmosphere show in all samples the presence of gypsum and secondary calcite with traces of quartz as part of the particulate cemented by gypsum (Table 1) (Fig. 7a–d). These results confirm that the formation of the black crust was mainly due to air pollution that caused black soiling and loss of structural integrity due to the attack by atmospheric sulfur and subsequent formation of gypsum [23].
(10 Hz to a maximum of 20 Hz) and varying the energy (maximum energy per pulse, 900 mJ at the wavelength 1064 nm, 400 mJ at 532 nm). After these tests, the officials of the Superintendence of Fine Art and Landscape for the Provinces of Bari, Foggia Barletta–Andria–Trani preferred to use the less invasive result/effect of cleaning, and as far as possible with a color/aspect of cleaned surface similar to the old original limestone. Thus the approach based on test 2, i.e., irradiation at 1064 nm by a Q-switched Nd:YAG laser, was chosen for removing the black encrustations resulting from atmospheric pollution from limestone
3.2. Laser cleaning tests Three different procedures were applied to test the level of cleaning achieved. The first test was performed at the wavelength of 532 nm (zone 1), the second one at the wavelength of 1064 nm (zone 2), and a third one by applying two consecutive wavelengths, first at 1064 nm and then at 532 nm (zone 3) (Fig. 8). The three tests show a different level of cleaning with an increasing of whitening when two wavelengths are used. The choice of parameters for performing the cleaning process was based on the comparison of results obtained from the use of the three different approaches using different pulse repetition frequencies
Fig. 8. Comparative cleaning tests on a masonry block using different wavelengths: 532 nm (1); 1064 nm (2), and first 1064 nm and then 532 nm (3).
302
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
materials, which has the advantages of not damaging the underlying stone and possibility of control the degree of the cleaning action. This type of laser provides pulses of typically few ns of 1064 nm radiation, which are preferentially absorbed by the stone soiled layers, thus producing a rapid increase of the surface temperature of the black crust which allows its removal. However, further irradiation with the second harmonic (532 nm) of the Q-switched Nd:YAG laser of the areas previously cleaned with the fundamental radiation results in a more complete removal of dust particles.
3.3. Laser induced breakdown spectroscopy (LIBS) Sample 4 was chosen for the following LIBS analysis based on results of petrographic and mineralogical analyses, and because the limestone masonry blocks under study are constituted mainly of biopelmicritic packstone. The surface of sample 4 was divided in three areas that were subjected to different treatments and analyses. Areas 1 and 2 (Fig. 9a) were first cleaned by removing the black crust using either one wavelength (1064 nm) or two successive wavelengths (1064 nm and then 532 nm), respectively. Then the cleaned limestone surfaces were analyzed along two lines (squares in Fig. 9a) either in the SP mode (1064 nm) (area 1) or in the DP mode (1064 nm and then 532 nm) (area 2). Area 3, i.e., the limestone sample covered by black crusts (Fig. 9a), was subjected to depth profile analyses using the DP configuration in collinear LIBS mode in order to identify the disappearance, or decrease with depth, of some specific elements typical of the patina. The crater depth after 28 shots was ~ 310 μm, i.e., 11 μm per shot (Fig. 9b). The emission lines of the elements detected (Al, Ba, C, Ca, Co, Fe, K, Li, Mg, Mn, Na, Si, Sr, and Ti) by qualitative LIBS analysis of the three sample areas 1, 2, and 3, are listed in Table 2. The emission line intensities using the DP configuration were about 5 times higher than those obtained by SP LIBS (Fig. 10), which can be attributed to a combination of increased
Fig. 9. SP and DP LIBS characterization (area 1 and area 2) and DP depth profile (area 3), (a); SEM image of the crater depth after 28 shots (b).
Table 2 Spectroscopic atomic emission lines (NIST, KURUCZ) of samples examined. Element Wavelength (nm) Co II Fe II Si I Al I Mn I Mn II KI CI Sr I Ti II Mg I Mg II Ca I
Li I Ba II Na I
233.79; 235.11; 240.47 234.33; 234.41; 234.52; 234.81; 235.91; 259.93; 261.18; 273.95; 274.91; 275.57 288.17; 250.69; 251.43; 251.61; 251.92; 252.41; 252.85; 390.55 308.21; 309.27 403.08; 403.31; 403.45; 403.57; 404.14 257.61; 259.37 766.49; 769.89 193.09; 247.85 460.73 323.45; 323.65; 323.90; 324.19; 336.12; 337.27; 338.37 280.98; 285.21; 383.23; 383.83; 516.73; 517.27; 518.36 279.55; 280.27 299.73; 300.09; 300.68; 300.94; 428.30; 428.94; 429.90; 430.78; 431.87; 442.55; 534.95; 558.22; 558.89; 559.02; 559.47; 559.87; 560.16; 560.29; 585.74; 671.72; 720.22 670.77 455.40 588.99; 589.59
laser ablation and plume re-heating. Thus the DP LIBS mode should be preferred to the SP mode in these studies. Fig. 11 shows the LIBS spectra obtained for the encrustation (area 3, Fig. 9a) at three different pulse numbers/depth (1, 10 and 20) which correspond to three depths from the surface. In the spectral range employed (250–330 nm), the emission lines detected confirm the presence of Ca and Mg as the major constituents of limestone and Si, Al, Fe, Co, Mn, Ti as the main contaminant elements that disappear with increasing the pulse number, i.e., the depth. The relative intensities of Ca I (534.95 nm), Mg I (285.21 nm), Si I (288.21 nm), Al I (308.21 nm), Fe II (234.41 nm), Ti II (323.65 nm), Sr I (460.73 nm), Ba II (455.40 nm), Li I (670.77 nm), Na I (588.99 nm), K I (769.89 nm), C I (247.85 nm) and Co II (240.47 nm) detected by the in-depth LIBS profile for the black crust with increasing the number of pulses are shown in Fig. 12. The trends of LIBS profiles of Si, Al, Fe, Ti, Ba, Li and Co, show a marked decrease after the first five–ten pulses (around 60–110 μm depth) (Fig. 12a). After ten pulses Fe, Co, Ti and Li almost disappear, i.e., are almost undetectable, whereas the intensities of Si, Al and Ba emissions remain low and almost constant down to 200 μm depth (Fig. 12a). LIBS intensities of Ca, Mg, K, Na, C and Sr show different trends, i.e., those of Ca and Sr increase with increasing depth, whereas those of Mg, K and C first decrease up to five–ten pulses, than increase and remain almost constant until the bottom, and that of K remains constant always (Fig. 12b). In particular, the increase shown by C with respect to Mg up to 5 pulses suggests an additional possible
Fig. 10. SP and DP LIBS emission spectra in the region 232–247 nm of the black crust from sample 4.
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
303
origin of C from organic contaminants in the upper layers of the black crust. Limestone and dolomite contain the highest Na concentrations, up to 5.400 mg kg−1 Na, of all sedimentary rocks, which is attributed to the effect of sea water during their formation and the presence of skeletal material [24]. In particular salt weathering in coastal areas of Apulia can be considered the most important cause of stone decay, especially where historic monuments and buildings are close to the coastline and partially or completely submerged by sea water. The calcarenites are soft and porous, and suffer much from deterioration as a result of weathering [25,26]. The type and rate of weathering of this building stone depend on the petrographic and physical properties (fabric, porosity, etc.), position within a monument or building, geographical location and air pollution level [27]. The relative intensities of atomic LIBS emission lines of Ca I (534.95 nm) and Mg I (285.25 nm) are shown in the inset of Fig. 12b. The results suggest that Ca, Mg, Na, Sr, K and C in the black crusts are mainly related to the composition of the limestone underneath, Si, Al, Ba and Li may be ascribed mostly to the soil–dust deposition, whereas Fe, Co and Ti would originate mainly from other sources, such as atmospheric pollution. The size of the Sr2+ ion (118 pm) is intermediate between those of 2+ Ca (100 pm) and K+ (138 pm), which may be replaced in a variety of rock forming minerals including K-feldspar, gypsum, plagioclase and, especially, calcite and dolomite. Enrichment of Sr up to concentrations of ca. 1000 mg kg−1 is, therefore, common in limestone and evaporates, although the Sr/Ca ratio in most types of limestone is less than 1:1000 [28]. Sr is easily mobilized during weathering, especially in oxidizing acid environments.
Fig. 12. Relative intensities of atomic and ionic LIBS emission lines of Si I 288.17 nm, Al I 308.21, nm, Fe II 234.41 nm, Ba II 455.40 nm, Li I 670.77 nm, Co II 240.47 nm and Ti II 323.65 nm (a) and Mg I 285.21 nm, Sr I 460.73 nm, K I 769.89 nm, Na I 588.99 nm, Ca I 534.95 nm and C I 247.85 (b) vs. number of pulses/depth. The relative intensities of atomic LIBS emission lines of limestone of Ca I (534.95 nm) and Mg I (285.25 nm) are shown in the inset of Fig. 12b.
Due to the high concentration of Ca and Mg across the encrustation relatively to that of the other elements detected, their emission intensities have been normalized with respect to the intensity of the Ca I line at 534.95 nm and Mg I line at 285.21 nm (Fig. 13a, b). With increasing the number of pulses, i.e., the depth, the emission of all elements decreases continuously with respect to Ca whereas with respect to Mg the emissions increase and then decrease, except for Ca. The decreasing trend of the Mg/Ca ratio with increasing depth (Fig. 14) suggests the increasing dedolomitazion of the ablated material [29,30].
4. Conclusions
Fig. 11. LIBS spectra of the ablated material from a black crust at different pulses (1, 10, 20).
The laser cleaning process was confirmed to be appropriate and efficient for achieving the removal of unwanted layers from the surface of limestone without modifying the surface morphology of the rock underneath. The optimal irradiation conditions causing minimal damaging effects were defined by both surface analysis and optimization of the Nd:YAG laser parameters (wavelength, pulse repetition frequencies, energy, etc.). The use of laser cleaning before the pre-consolidation operations allowed to prevent the entrapping of the dirty layers in the limestone rocks without any further possibility to remove it further on. LIBS was confirmed to be a powerful diagnostic technique able to monitor and control the laser cleaning process of limestone, thus allowing to replace traditional invasive laboratory analyses and provide a prompt compositional response in situ.
304
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305
Fig. 13. Ratio of the intensities of atomic and ionic LIBS emission lines of Fe II 234.41 nm, Co II 240.47 nm, Si I 288.17 nm, Al I 308.21 nm, Mg I 280.98 nm, C I 193.09 nm and Ti II 323.65 nm and the intensity of Ca I line at 534.95 nm (a); and of Fe II 234.41 nm, Co II 240.47 nm, Si I 288.17 nm, Al I 308.21 nm, C I 193.09 nm, Ca I 534.95 nm and Ti II 323.65 nm and the intensity of Mg I line at 285.21 nm (b) vs. number of pulses/depth.
In particular, LIBS provided an evaluation of the black crust by confirming the absence of any protective layers (“scialbatures”) on the quoins of the masonry of the Castello Svevo, Bari, Italy. A comparison of SP and the DP LIBS configurations showed that the emission line intensities obtained by DP LIBS were about 5 times higher than those obtained by SP LIBS, which might be attributed to a combination of increased laser ablation and plume re-heating. The DP LIBS configuration stratigraphy was also used in order to assess the optimal cleaning depth, i.e., the number of pulses that should be applied to avoid under-cleaning and/or over-cleaning. The black crust layer (approximately 150-μm deep) on the limestone analyzed was completely removed after about 10 double pulses, which added to the common well-known advantages of laser-based cleaning methods, including automatization, selectivity, versatility, and a high degree of precision.
Acknowledgments The authors kindly acknowledge the financial funding received under the project “Il restauro delle grandi opere in Puglia: l'innovazione attraverso le nanotecnologie e metodologie diagnostiche avanzate”, P.O. Puglia FESR 2007–2013, Bando “Aiuti a Sostegno dei Partenariati Regionali per l'Innovazione” (3Z3VZ46).
Fig. 14. Ratio of the intensities of atomic LIBS emission lines of Mg I line at 285.21 nm and Ca I line at 534.95 nm vs. number of pulses/depth starting from the black crust surface (squares), and from underlying stone (dots).
References [1] J.F. Asmus, M. Seracini, M.J. Zetler, Surface morphology of laser-cleaned stone, Lithoclastia 1 (1976) 23–46. [2] M. Cooper, Laser Cleaning in Conservation: An Introduction, ButterworthHeinemann, Oxford, 1998. [3] R. Pini, R. Salimbeni, Tecniche e sistemi laser per il restauro dei beni culturali, Cardini Editore, Firenze, 2001. [4] C. Fotakis, D. Anglos, V. Zafiropulos, S. Georgiou, V. Tornari, Lasers in the preservation of cultural heritage: principles and applications, Series in Optics and Optoelectronics, CRC Press, Taylor & Francis Group, Boca Raton, 2007. [5] S. Siano, J. Agresti, I. Cacciari, D. Ciofini, M. Mascalchi, I. Osticioli, A.A. Mencaglia, Laser cleaning in conservation of stone, metal, and painted artifacts: state of the art and new insights on the use of the Nd:YAG lasers, Appl. Phys. A 106 (2012) 419–446. [6] L. Bartoli, P. Pouli, C. Fotakis, S. Siano, R. Salimbeni, Characterization of stone cleaning by Nd:YAG lasers with different pulse duration, Laser Chem. 2006 (2006) 1–6 (Article ID 81750). [7] S. Siano, M. Giamello, L. Bartoli, A. Mencaglia, V. Parfenov, R. Salimbeni, Laser cleaning of stone by different laser pulse duration and wavelength, Laser Phys. 18 (2008) 27–36. [8] C. Rodriguez-Navarro, K. Elert, E. Sebastian, R.M. Esbert, C.M. Grossi, A. Rojo, F.J. Alonso, M. Montoto, J. Ordaz, Laser cleaning of stone materials: an overview of current research, Rev. Conserv. 4 (2003) 65–82. [9] S. Siano, Principles of laser cleaning in conservation, in: M. Schreiner, M. Strlic (Eds.),Handbook on the Use of Lasers in Conservation and Conservation Science, COST G7 2007, pp. 1–26. [10] I. Gobernado-Mitre, A.C. Prieto, V. Zafiropulos, Y. Spetsidou, C. Fotakis, On-line monitoring of laser cleaning of limestone by laser-induced breakdown spectroscopy and laser-induced fluorescence, Appl. Spectrosc. 51 (1997) 1125–1129. [11] P. Maravelaki-Kalaitzaki, V. Zafiropulos, C. Fotakis, Excimer laser cleaning of encrustation on Pentelic marble: procedure and evaluation of the effects, Appl. Surf. Sci. 148 (1999) 92–104. [12] P. Maravelaki-Kalaitzaki, D. Anglos, V. Kilikoglou, V. Zafiropulos, Compositional characterization of encrustation on marble with laser induced breakdown spectroscopy, Spectrochim. Acta B 56 (2001) 887–903. [13] P. Bromblet, M. Labouré, G. Orial, Diversity of the cleaning procedures including laser for the restoration of carved portals in France over the last 10 years, J. Cult. Herit. 4 (2003) 17–26. [14] S. Siano, A. Casciani, A. Giusti, M. Matteini, R. Pini, S. Porcinai, R. Salimbeni, The Santi Quattro Coronati: cleaning of the gilded decorations, J. Cult. Herit. 4 (2003) 123–128. [15] V. Vergès-Belmin, C. Dignard, Laser yellowing: myth or reality? J. Cult. Herit. 4 (2003) 238–244. [16] S. Siano, A. Giusti, D. Pinna, S. Porcinai, M. Giamello, G. Sabatini, R. Salimbeni, The conservation intervention on the Porta della Mandorla, in: K. Dickmann, C. Fotakis, J.F. Asmus (Eds.), Lacona V Proceedings, Springer-Verlag, Berlin 2005, pp. 171–178. [17] J. Kolar, M. Strlič, D. Müller-Hess, A. Gruber, K. Troschke, S. Pentzien, W. Kautek, Laser cleaning of paper using Nd:YAG laser running at 532 nm, J. Cult. Herit. 4 (2003) 185–187. [18] K. Ochocińska, A. Kamińska, G. Śliwiński, Experimental investigations of stained paper documents cleaned by Nd:YAG laser pulses, J. Cult. Herit. 1 (2000) 188–193. [19] M.C. Gaetani, U. Santamaria, The laser cleaning of wall paintings, J. Cult. Herit. 1 (2000) S199–S207.
G.S. Senesi et al. / Microchemical Journal 124 (2016) 296–305 [20] V. Zafiropulos, P. Pouli, V. Kylikoglou, P. Maravelaki-Kalaitzaki, B.S. Lukyanchuk, A. Gogariu, Synchronous use of IR an UV laser pulses in the removal of encrustation: mechanistic aspects, discoloration phenomena and benefits, in: K. Dickmann, C. Fotakis, J.F. Asmus (Eds.), Lacona V Proceedings, Springer-Verlag, Berlin 2005, pp. 311–318. [21] P. Pouli, K. Frantzikinaki, E. Papakonstantinou, V. Zafiropulos, Pollution encrustation removal by means of combined ultraviolet and infrared laser radiation: the application of this innovative methodology on the surface of the Parthenon west frieze, in: K. Dickmann, C. Fotakis, J.F. Asmus (Eds.), Lacona V Proceedings, Springer-Verlag, Berlin 2005, pp. 333–340. [22] P. Pieri, L. Sabato, L. Spalluto, M. Tropeano, Notes to the geological map of the urban area of Bari (southern Italy) 1:25,000 scale, Rend. Online Soc. Geol. Ital. 14 (2011) 26–36. [23] S.S. Potgieter-Vermaak, R.H.M. Godoi, R. Van Grieken, J.H. Potgieter, M. Oujja, M. Castillejo, Micro-structural characterization of black crust and laser cleaning of building stones by micro-Raman and SEM techniques, Spectrochim. Acta A 61 (2005) 2460–2467.
305
[24] G.K. Billings, P.C. Ragland, Geochemistry and mineralogy of the Recent reef and lagoonal sediments south of Belize (British Honduras), Chem. Geol. 3 (1968) 135–153. [25] G.F. Andriani, N. Walsh, Physical properties and textural parameters of calcarenite rocks: qualitative and quantitative evaluations, Eng. Geol. 67 (2002) 5–15. [26] G.F. Andriani, N. Walsh, Fabric, porosity and water permeability of calcarenites from Apulia (SE Italy) used as a building and ornamental stone, Bull. Eng. Geol. Environ. 62 (2003) 77–84. [27] S. Ordonez, R. Fort, M.A. Garcia del Cura, Pore size distribution and the durability of a porous limestone, Q. J. Eng. Geol. 30 (1997) 221–230. [28] J.L. Kulp, K. Turekian, D.W. Boyd, Strontium content of limestones and fossils, Geol. Soc. Am. Bull. 63 (1952) 701–716. [29] R.L. Folk, The natural history of crystalline calcium carbonate: effect of magnesium content and salinity, J. Sediment. Petrol. 44 (1974) 40–53. [30] R.L. Folk, L.S. Land, Mg/Ca ratio and salinity: two controls over crystallization of dolomite, Am. Assoc. Pet. Geol. Bull. 59 (1975) 60–68.