diffuse reflectance hybrid spectrometer for analysis of inorganic pigments

diffuse reflectance hybrid spectrometer for analysis of inorganic pigments

Accepted Manuscript Portable hybrid Laser-Induced Breakdown Spectroscopy-diffuse reflectance spectrometer for spectroscopic analysis of inorganic pigm...

1MB Sizes 0 Downloads 55 Views

Accepted Manuscript Portable hybrid Laser-Induced Breakdown Spectroscopy-diffuse reflectance spectrometer for spectroscopic analysis of inorganic pigments

P. Siozos, A. Philippidis, D. Anglos PII: DOI: Reference:

S0584-8547(17)30015-0 doi:10.1016/j.sab.2017.09.005 SAB 5298

To appear in:

Spectrochimica Acta Part B: Atomic Spectroscopy

Received date: Revised date: Accepted date:

14 February 2017 31 August 2017 10 September 2017

Please cite this article as: P. Siozos, A. Philippidis, D. Anglos , Portable hybrid LaserInduced Breakdown Spectroscopy-diffuse reflectance spectrometer for spectroscopic analysis of inorganic pigments. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Sab(2017), doi:10.1016/ j.sab.2017.09.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Portable hybrid laser induced breakdown spectroscopy-diffuse reflectance spectrometer for spectroscopic analysis of inorganic pigments P. Siozos1, A. Philippidis1, D. Anglos1,2 1

Institute of Electronic Structure and Laser, Foundation for Research and Technology-Hellas (IESL-

Department of Chemistry, University of Crete, P.O. Box 2208, GR 710 03, Heraklion, Crete, Greece

Corresponding author e-mail address: [email protected]

RI

2

PT

FORTH), P.O. Box 1385, GR 711 10, Heraklion, Crete, Greece

SC

Keywords: LIBS, diffuse reflectance, portable instrument, hybrid spectrometer,

NU

pigments

MA

Abstract

A novel, portable spectrometer, combining two analytical techniques, laser-induced

D

breakdown spectroscopy (LIBS) and diffuse reflectance spectroscopy, was developed

PT E

with the aim to provide an enhanced instrumental and methodological approach with regard to the analysis of pigments in objects of cultural heritage. Technical details

CE

about the hybrid spectrometer and its operation are presented and examples are given relevant to the analysis of paint materials. Both LIBS and diffuse reflectance spectra

AC

of several samples of neat mineral pigments were recorded and the complementary information was used to effectively distinguish different types of pigments even if they had similar colour or elemental composition. The spectrometer was also employed in the analysis of different paints on the surface of an ancient pottery sherd demonstrating the capabilities of the proposed hybrid diagnostic approach. Despite its instrumental simplicity and compact size, the spectrometer is capable of supporting analytical campaigns relevant to archaeological, historical or art historical

1

ACCEPTED MANUSCRIPT investigations, particularly when quick data acquisition is required in the context of surveys of large numbers of objects and samples. Introduction Characterization and identification of pigments in cultural heritage artefacts is of key importance for understanding what materials have been used and which painting

PT

techniques have been employed. The information, one derives from such an analysis,

RI

reveals components and technologies that provide insights related to provenance,

SC

dating or authenticity of painted works of art and archaeological findings or monuments [1,2]. Furthermore, such knowledge allows informed decisions to be

NU

made, which are necessary for designing and conducting proper conservation /restoration actions on valuable art or historical objects and monuments [3].

MA

A broad spectrum of advanced laboratory techniques for materials analysis is nowadays available and widely used for pigment characterization and analysis in

D

archaeological research and conservation science [3,4,5]. However, there is a growing

PT E

demand for exploiting modern analytical methods in order to perform analytical investigations on-site and in real time, for example, during excavation and survey

CE

campaigns or in subsequent conservation work, and this demand generates a constant driving force for developing or adapting mobile instrumentation towards this purpose.

AC

In fact, recent technological advances in relation to light sources, spectrometers and detectors enable the development of such compact analytical instruments opening up ways for establishing versatile, non-destructive methodologies that would avoid sampling and enable analytical studies to be performed rapidly and routinely, in the museum or conservation laboratory [6,7]. Several examples describing the application of mobile instruments and methods in heritage science have been published over the past decade with the majority referring to techniques such as Fiber-optic reflectance

2

ACCEPTED MANUSCRIPT spectroscopy (FORS) and multi-spectral imaging [8,9,10,11,12], X-ray Fluorescence (XRF) [7,13,14,15], Raman microspectrometry [15,16,17], Infrared spectroscopy (FTIR) [18,19], Fluorescence spectroscopy [12,20] or Laser-Induced Breakdown Spectroscopy (LIBS) [14,21,22,23,24,25,26]. However, despite various efforts, the development of compact instruments and efficient methodologies remains still an

PT

important challenge and in this context the aim of the research presented in this paper

SC

spectroscopy in a hybrid instrument configuration.

RI

is to investigate the combined use of two mobile techniques, LIBS and reflectance

Laser-Induced Breakdown Spectroscopy (LIBS) has been extensively used in the

NU

study of materials in objects of archaeological and artistic importance [27,28,29] yet it has to be pointed out that LIBS is still far from being a widely popular or routine

MA

technique. It provides information about the elemental composition of materials on the basis of the optical emission spectrum recorded from a transient micro-plasma that

D

is generated by means of focusing a nanosecond laser pulse on the surface of the

PT E

object/sample investigated. One of the important features of LIBS is that neither sampling nor sample preparation is required. The analysis is carried out directly on

CE

the object. The technique is almost non-destructive, actually classified as microdestructive, since the diameter of the spot that is probed is on the order of 100-200 μm

AC

or even smaller when proper laser objective lenses are employed. In the case of wall paintings, the depth of the crater formed does not exceed a few micrometers. In recent years, several research groups have demonstrated that LIBS analysis can indeed be performed using compact equipment, and as a result portable LIBS units have emerged and used in the context of cultural heritage analysis [14,21,22,23,24,25,26]. As explained, LIBS is an elemental analysis method, hence it does not provide direct information about the molecular structure of the material analyzed and as a

3

ACCEPTED MANUSCRIPT result it often becomes difficult for one to distinguish between pigments or mixtures of pigments consisting of the same elements. To overcome this limitation, significant efforts have been directed towards the development of hybrid instruments combining complementary optical techniques, for example, LIBS and Raman, or LIBS and LIF spectroscopy [30,31,32,33,34]. The main instrumental components of LIBS, Raman

PT

or LIF spectrometers include a laser light source, appropriate optics, that deliver the

RI

excitation light to the sample and collect the emitted (scattered) light, and finally the

SC

spectrometer. In principle, the same components can be used for all three methods, although LIBS requires nanosecond laser pulses in conjunction with high resolution

NU

spectrographs coupled to fast triggered and gated detectors, while Raman and LIF rely on the use of low noise, high sensitivity detectors, with Raman typically employing

MA

continuous wave (cw) laser sources for excitation. In the case of LIBS the laser wavelength is not highly critical. In contrast Raman and LIF spectra are significantly

D

dependent on excitation wavelength with fluorescence requiring sources operating in

PT E

the UV in order to ensure efficient excitation of organic chromophores. Even though the specifications of the parts required for putting together a spectrometer (excitation

CE

source, spectrograph, detector) have, in general, significantly improved as a result of technological advances, still, the combination of these parts into a single hybrid

AC

system compromises the overall result by the fact that requirements for optimum performance can be contradictory. Consequently, limitations in the sensitivity and the efficiency of such hybrid systems cannot be avoided. At this stage, it is noted that when performing LIBS analysis of paints, the appearance, namely the color, of the paint is evaluated along with the elemental composition data for identifying as accurately and reliably as possible the pigment or pigments present in the paint. As a relevant example, two iron-based pigments, yellow

4

ACCEPTED MANUSCRIPT ochre (FeO(OH)) and red iron oxide (Fe2O3), are known to yield very similar LIBS spectra, dominated by emission lines due to Fe, but on the basis of the paint color one could argue in favor of the presence of one or the other. An objective way to assess the color of a paint is to record its absorbance in the visible and the near infrared part of the spectrum. In fact, given that paints are non-transparent and highly scattering,

PT

their absorbance is recorded via measurements of the corresponding diffuse (non-

RI

specular) reflectance spectrum. Even though the spectral features corresponding to

SC

pigment materials are rather broad, diffuse reflectance spectroscopy is a rather established non-destructive method that permits quick surveying of painted artefacts

NU

aiding pigment characterization, particularly in combination with imaging [8,9,10,11,35,36,37,38,39,40].

MA

In a diffuse reflectance measurement, a broadband white light beam illuminates the paint surface where it is transmitted, scattered and absorbed throughout the composite

D

paint material. Part of the light that is not absorbed by the paint, is back-scattered,

PT E

being what is termed as diffuse reflection. Due to the simplicity of the method, a broadband light source and a low resolution spectrometer can be used for the

CE

implementation of a reflectance setup capable of providing consistent spectra of materials. As mentioned, diffuse reflectance spectroscopy has been widely used for

AC

the characterization of pigments, while significant effort has been made in order to improve the capabilities of the method in the analysis of pigment mixtures [38,39]. To this end, coupling LIBS with diffuse reflectance spectroscopy appears as a logical choice, in-line with the approach combining elemental information with paint color, as discussed above. In fact, the combined application of LIBS along with multispectral reflectance imaging, for the analysis of pigments, by use of two different instruments, has already been demonstrated [41]. In the present work a novel hybrid

5

ACCEPTED MANUSCRIPT portable LIBS - diffuse reflectance spectrometer is presented. The performance of the instrument is evaluated using neat pigments in powder form and the capabilities of the combined use of LIBS and diffuse reflectance spectroscopy for the identification of inorganic pigments is discussed based on the analysis of a painted pottery sherd.

PT

Experimental Setup

RI

Portable LIBS spectrometer: The mobile LIBS spectrometer (LMNTII+), developed

SC

and constructed at IESL-FORTH (Fig. 1a), has been described previously [24,25,26]. Briefly, the instrument employs a compact, passively Q-switched Nd:YAG laser

NU

emitting pulses at 1064 nm (10 mJ/pulse, 10 ns). The laser beam is focused by means of plano-convex lens (f = + 75 mm) on the surface of the object generating a transient

MA

micro-plasma. The emission from the plasma is collected through the same lens and transmitted via a bifurcated optical fiber into a dual spectrometer unit (Avaspec-2048-

D

2-USB2, Avantes) that records spectra across a wavelength range extending from 250

PT E

to 660 nm, with resolution in the range of 0.2-0.3 nm. The laser along with the necessary optics and a miniature CCD camera are integrated in the optical probe head.

CE

The camera offers a magnified view of the object surface during analysis and permits accurate aiming of the laser beam with the aid of a cross-hair indicator superimposed

AC

on the image. All LIBS spectra were collected with a time-delay of 1.3 μs with respect to the laser pulse and for an integration time of 1 ms. The instrument fits in a compact case (dimensions: 46 x 33 x 17 cm3) and weighs less than 9 kg. Diffuse Reflectance module: The optical probe head and optical fibers of the LIBS module are also used for measuring the diffuse reflectance from the sample surface (Fig. 1b). An external halogen tungsten lamp (Osram Decostar 35, 10 W), attached on the front end of the LIBS optical head, is used for illuminating the object. The diffuse

6

ACCEPTED MANUSCRIPT reflectance from the object surface, collected through the LIBS optical detection line, is fed into the spectrometer via the optical fiber. To achieve a broader spectral coverage, a low resolution spectrometer is utilized, covering the range from 200 1100 nm with a resolution of approximately 1.4 nm (Avaspec-2048L-USB2, Avantes). Calibration of the y-axis, for defining the 100% on the reflectance scale,

PT

was performed using a pellet of compressed BaSO4 powder, known to display very

RI

high and relatively flat reflectance over the visible and near IR. The reflectance

SC

spectrum is calculated based on Equation 1, with Isample and IBaSO4 being the intensity of the light reflected by the sample surface and the white reference and Idark the signal

NU

captured by the detector with the white light source on and no sample at the sample

MA

position.

(1)

In addition, a Perkin Elmer (PE) Lambda 950 spectrophotometer, equipped with an

PT E

D

integrating sphere, was used to record spectra of pigment powders for comparison with the diffuse reflectance spectra obtained on the hybrid instrument. The spectrometer makes use of a halogen tungsten lamp for recording spectra in the range

CE

of 360 to 2500 nm.

AC

Pigment samples: Several samples of neat inorganic pigments (Kremer Pigmente GmbH and Co.) were used in the form of pellets. CaCO3 and BaSO4 in powder form (99.9 % purity) were purchased from Sigma-Aldrich. Pellets were produced after pressing the pigment powders under 10 atm for approximately 10 minutes in an appropriate sample holder.

Results and discussion

7

ACCEPTED MANUSCRIPT The portable LIBS instrument has been previously used for analysis of pigments in several types of heritage objects such as icons, easel and wall paintings, plaster or painted pottery and results have already been published [24,25,26]. Therefore, reference to the operation and performance of the LIBS spectrometer will be limited, in the context of the present paper, and only necessary LIBS spectra that support

PT

arguments relevant to the combined implementation of LIBS with diffuse reflectance

RI

spectroscopy will be presented.

SC

Initial tests were focused on evaluating the performance of the diffuse reflectance module (light source, optics and spectrometer). The emission of the tungsten-halogen

NU

lamp was recorded directly through the LIBS optical detection line in order to measure the overall spectral throughput characteristics of the optics and spectrograph-

MA

detector system and thus determine the effective spectral range that can be used for obtaining reliable reflectance spectra. In Figure 2, the emission of the halogen-

D

tungsten lamp recorded on the hybrid spectrometer is shown in comparison with the

PT E

spectral output of the lamp provided by the manufacturer and the black body radiation emission corresponding to a temperature of 4200 K, as indicated in the lamp

CE

specifications [42]. Clearly, the measured spectrum is significantly different reflecting a convolution of the lamp emission with the spectral profile of the CCD detector

AC

sensitivity and the diffraction efficiency of the grating. In fact, as seen in Figure 2 (by comparing curves c and d), the spectrum measured is quite similar to the detector sensitivity curve reported by the spectrometer manufacturer [43]. As a result, the usable wavelength range of the spectrometer extends from 380 to 950 nm. Thus the entire visible range is covered along with a small part of the near IR. It is noted that these measurements were performed by setting an integration time of 10-12 ms on the CCD, which resulted in spectra with good S/N. The useful spectral range can be

8

ACCEPTED MANUSCRIPT expanded, if needed, by about 20-40 nm on either side of the spectrum, just by increasing the integration time on the CCD detector, albeit at the expense of data in the middle part of the spectrum, around 600 nm, where the detector will reach saturation. All reflectance measurements were performed under ambient laboratory light conditions with no indication of any substantial interference of stray light with

PT

the recorded spectra. This is an important feature that makes working with the

RI

reflectance module quite straightforward. It is mainly due to the short integration time

SC

set on the detector in combination with the adequate lamp output in the visible and the near IR. In addition, the optical geometry of the system is an important factor because

NU

it captures the reflectance only from a small spot, no more than 2 mm in diameter, on the object surface.

MA

Further to these initial tests, reflectance spectra of several pigments were recorded for investigating the performance and capabilities of the diffuse reflectance module of

D

the hybrid instrument and were compared against those recorded by use of the PE

PT E

laboratory spectrophotometer (Fig. 3). It is clear that spectra collected on the mobile unit are quite similar, almost identical, with those recorded on the laboratory

CE

spectrophotometer.

Next the performance of the hybrid instrument as a whole, namely both as a LIBS

AC

and as a reflectance spectrometer, was evaluated through analyses of several pigment samples. As already mentioned the identification of certain pigments having very similar elemental composition profile cannot be unambiguous based only on the LIBS spectra collected. Therefore, additional spectroscopic information is required to verify the presence of a specific pigment over another one of similar elemental composition. In this context, two sets of pigments, each characterized by a different element, were selected for investigation.

9

ACCEPTED MANUSCRIPT In the first set, three different copper-based pigments were examined: azurite (2CuCO3  Cu(OH)2, blue), malachite (CuCO3  Cu(OH)2, green) and Egyptian blue (CaCuSi4O10, blue). The LIBS spectra of azurite and malachite are very similar (Fig. 4a-b) with the emission lines of Cu being the dominant ones. Emissions from Ca and Al also present in the spectra of azurite and malachite represent most likely

PT

extraneous impurities and cannot be trusted for differentiating the two pigments. In

RI

the case of Egyptian blue the combined presence of Cu, Si and Ca lines in the LIBS

SC

spectrum (Fig. 4c), provides good evidence for confirming the identity of the pigment. It is noted however, that in the case of an actual painting study, one would not be in a

NU

position to easily rule out the possibility that Ca and Si might have been coming from other materials present in the paint such as calcium carbonate or silica.

MA

In comparison with the LIBS data, the diffuse reflectance spectra of the three copper pigments do show distinct differences (Fig. 4d). Egyptian blue displays two

D

characteristic absorption bands corresponding to minima in the reflectance spectrum

PT E

at 640 nm and 800 nm due to the well-known d-d transitions of Cu2+ in the silicate crystal field [44]. The reflectance of azurite shows a rather similar spectral profile

CE

with Egyptian blue in the blue part of the spectrum with reflectance maxima at 460 nm and 440 nm for the two pigments respectively. In addition, it is evident that the

AC

azurite spectrum exhibits stronger absorbance all across the spectrum and in contrast to Egyptian blue shows no distinct bands in the red and the near IR region of the spectrum. Malachite, a green pigment, is clearly differentiated on the basis of its characteristic reflectance maximum at 530 nm, in the green range of the optical spectrum. In all, it is obvious that combining data from LIBS and reflectance spectroscopy strengthens the ability one has to discriminate among these pigments and arrive at a reliable determination of what might be present in the paint examined.

10

ACCEPTED MANUSCRIPT The second set includes pigments based on iron minerals: red iron oxide (Fe2O3), yellow ochre (FeO(OH)), green earth (K[(Al,FeIII),(FeII,Mg](AlSi3,Si4)O10(OH)2) and mars black (Fe3O4). Emission lines due to Fe cover densely significant parts of the LIBS spectra (Fig. 5a) and obviously no conclusive results can be derived solely from the LIBS spectra about the chemical structure of these pigments. In contrast, the

PT

diffuse reflectance spectra of each one of these iron based pigments appear to be

RI

clearly different from one another. In particular, the red iron oxide and yellow ochre

SC

present the typical S-shape profile characterized by a sharp increase of the reflectance at 550-600 nm and 500-580 nm respectively. Green earth shows a broad band with a

NU

reflectance maximum at 560 nm approximately while mars black exhibits very low reflectance across the entire spectral range. These differences, combined with the

MA

elemental composition from LIBS, discriminate the chemical identity of each pigment.

D

In order to test how the hybrid system works when actual artefacts are examined,

PT E

we studied an ancient (Minoan) pottery sherd bearing red, black and white slip decorations on its surface. The LIBS spectra of the three coloured areas can be seen in

CE

Figure 6a and b. In the black and red areas, the similarity of the spectra is rather impressive. Intense Fe emission peaks were recorded in both indicating strongly the

AC

use of iron based pigments most likely black and red iron oxides respectively. On the white area intense calcium emission peaks were recorded indicating most probably the use of a lime-based paint (CaCO3), commonly used in antiquity. Furthermore, weak Fe and Al peaks were detected in the white area, which are related either to the presence of impurities in the paint, expected since it must be coming from a natural source, or to contribution from the underlying clay. Concerning Ca, based on our LIBS data, in the range of 250-660 nm, we cannot exclude the presence CaSO4 even

11

ACCEPTED MANUSCRIPT though historically its use in Minoan painted pottery has not been documented. It is noted that with LIBS spectra extending in the near IR [45,46], where sulfur emits (in the range of 921-924 nm), or possibly following a double-pulse LIBS approach [46,47] one could prove or disprove the presence of sulfur and relate this to the potential use of CaSO4.

PT

Further to LIBS analysis, diffuse reflectance spectra were also recorded from the

RI

three colored areas of the pottery sherd (Fig. 6c). The spectrum collected from the red

SC

paint appears to have the typical S-shaped profile similar to that obtained with the red ochre pigment, characterized by a transition from low to medium reflectance between

NU

550 and 600 nm and a reflectance maximum close to 50% near 740 nm. The first derivative of the spectrum was also calculated (Fig. 7), identifying an inflexion point

MA

between 570 nm and 580 nm, in close similarity to the corresponding derivative spectrum of the neat red iron oxide pigment [8,37]. In combination with the recording

D

of intense Fe emissions in the LIBS spectrum the spectral features of the reflectance

PT E

analysis confirm with rather high certainty the presence of red ochre in the red paint on the pottery. However, it is noted that overall the reflectance of the red paint on the

CE

sherd is rather higher if compared against that of the neat pigment measured. This can be understood in part on the basis of a slightly lower density of the red paint on the

AC

sherd or the possible loss of parts of it during its burial time or as a result of surface cleaning during conservation. But furthermore, comparing carefully the two spectra, one observes that the reflectance curve of the red paint has a different slope in the blue-green part of the spectrum with respect to that of the red ochre pigment. In addition, a weak absorption feature in the near-infrared (~870 nm), clearly observed in the spectrum of the neat pigment [8,37], is hardly visible in the spectrum of the red area of the sherd. These features are indicative of the presence of an additional

12

ACCEPTED MANUSCRIPT material or pigment, absorbing in the blue-green area. Considering the LIBS data, one could assume that an iron-based material, such as for example yellow ochre, is present in combination with the red. This scenario will be revisited further below, considering the possibility to express the observed reflectance spectrum, coming from a pigment mixture, as a superposition of the corresponding neat pigment spectra.

PT

Significant similarities were also evident between the spectrum recorded on the

RI

sherd’s black slip and that corresponding to the black iron oxide pigment, both

SC

showing low reflectance values across the whole spectral range. Even though this similarity in reflectance does not constitute a solid proof for the identity of the

NU

pigment, the complementary information from the LIBS and reflectance data does provide strong evidence for the use of mars black.

MA

Finally, concerning the white area of the decoration, the diffuse reflectance spectrum is characterized by high values in a broad spectral range, which, however,

D

decrease progressively from the orange down to the blue part of the spectrum with

PT E

respect to those observed in the spectrum of pure CaCO3. This observed difference can be attributed to the presence of an additional pigment or substance along with the

CE

white paint. Like in the case of the red paint, a potential candidate could be a yellow pigment, with the LIBS analysis, indicating the possible presence of yellow ochre,

AC

given the weak but clear emissions due to Fe observed in the spectrum (Fig. 6a). Moreover, no additional information can be derived in order to verify the presence of either CaCO3 or CaSO4 since in this case both materials present similar reflectance spectra in the visible and the near IR. In order to further investigate the presence of other materials and specifically yellow ochre in the red and white colored areas of the sherd, the Kubelca-Munk (KM) theory was used as a tool to describe the diffuse reflectance of paint mixtures [48,49].

13

ACCEPTED MANUSCRIPT The KM theory provides a simple approach to the complex problem of radiative transfer in the presence of both absorption and scattering that relates the diffuse reflectance of a layer, without interfaces, to the effective absorption (K) and scattering coefficients (S). When hiding is complete (namely, for an optically thick paint layer),

PT

according to the KM theory, the ratio K/S at each wavelength λ can be expressed as: (2)

RI

where R∞ is the spectral reflectance of a layer with infinite optical thickness.

SC

Predicting the reflectance of a mixture of paints can be accomplished if the ratio K/S corresponding to the overall mixture can be represented as a function of the

NU

individual K/S values corresponding to the different paint components, according to

MA

the following expression:

(3)

D

For a mixture of pigments, Ki and Si represent the effective absorption and

PT E

scattering coefficients for each individual pigment respectively and ci stands for the concentration (molar fraction) of the ith pigment. Obviously, this approach, in

CE

particular the second equality in Eq. 3, is rigorously valid only if the scattering coefficients of all pigments in the mixture show identical dependence on wavelength

AC

over the whole spectral range investigated. Even though this is not the case it is still possible to express the total K/S of the mixture as a linear combination of the individual (Ki/Si) values with the coefficients C’i as parameters. Despite this simplification the method appears to be quite effective in providing realistic predictions for spectra of pigment mixtures. This is carried out by setting C’i values as unbounded parameters and calculating the best positive least squares fit of the predicted spectrum to the experimentally measured one for different combinations of reference pigments [9,50]. Typically, best results are obtained when a white pigment 14

ACCEPTED MANUSCRIPT is contained in the mixture, because this provides a strong scattering contribution that supports the uniform scattering assumption underlying the proposed approach. It is noted, however, that, Liang et al. [9] have successfully applied the method even in pigment mixtures, which do not contain a white pigment except in cases in which pigments with high absorption or very low scattering were involved.

PT

In Figure 8a the composite spectrum computed using the individual reflectance

RI

spectra of three pigments (CaCO3, red ochre and yellow ochre) is presented in

SC

comparison with the diffuse reflectance spectrum recorded from the red slip on the sherd. The agreement between the computed and the actual spectrum is remarkably

NU

good suggesting that the presence of yellow ochre mixed with red ochre and white in the red area of the sherd is a realistic scenario. Moreover, in Figure 8b, the reflectance

MA

spectrum recorded from the white decoration of the sherd and the computed spectrum derived using a combination of CaCO3 and yellow ochre are compared and the

PT E

present in the white area.

D

agreement between the two is significant indicating that yellow ochre could also be

In an attempt to compare the conclusions from the LIBS – diffuse reflectance

CE

approach against a non-destructive method, widely used for the characterization of cultural heritage artefacts, micro-Raman analysis [51] was carried out on the Minoan

AC

sherd. Unfortunately, even though the sherd was examined under two different excitation wavelengths (λexc = 532 nm and 785 nm), the colour decorated areas did not yield any meaningful Raman spectra. Instead, broad features were observed due to fluorescence from impurities or organic constituents. An effort was also made to employ, FT-IR analysis [52], yet in that case, too, meaningless spectra were recorded due to the fact that the sherd surface was very rough and therefore prevented the sample from coming into close contact with the instrument’s crystal during

15

ACCEPTED MANUSCRIPT measurements performed in ATR mode. As a result, it was not possible to confirm the conclusions derived from the complementary LIBS-diffuse reflectance approach in the analysis of the archaeological artefact by a comparative study with other spectroscopic techniques. However, this result does point out to the capability of the simple, portable, hybrid instrument and moreover the complementary use of LIBS and

PT

diffuse reflectance methods to provide information about the chemical composition of

RI

materials and artefacts in cases in which other more sophisticated and advanced

SC

techniques might fail due to various experimental constrains.

NU

Conclusions

A hybrid instrument combining diffuse reflectance spectroscopy with LIBS has

MA

been implemented, extending the potential of an existing portable LIBS spectrometer, concerning analysis of materials in cultural heritage objects. The performance of the

D

hybrid instrument has been evaluated with the diffuse reflectance module covering the

PT E

spectral range 380-950 nm and the LIBS unit operating in the range of 250-660 nm. The complementary information, provided by the LIBS and diffuse reflectance

CE

spectra of neat pigment samples, was used to discriminate effectively and identify different types of pigments even if they had the same colour and/or similar elemental

AC

composition. The capabilities of the instrument have been demonstrated in the characterization of paint on a Minoan pottery sherd. Combining data, obtained from the two modules of the hybrid spectrometer, clearly aided analysis leading to more reliable pigment identification if compared to using results from each technique individually. This proves not only the strength of combining the two complementary analytical methods but also the benefit from the simultaneous application based on a simple, straightforward instrumental arrangement.

16

ACCEPTED MANUSCRIPT Further work is underway to achieve recording of spectra on the same spectrometers both for LIBS and diffuse reflectance measurements. In particular, an additional spectrometer will be installed to record the LIBS spectrum in the NIR range and all three spectrometers combined will record the LIBS and the diffuse reflectance spectra of the materials in the UV-visible-near IR range with increased

PT

spectral resolution. Moreover, the advanced capabilities of the combination of LIBS

RI

and diffuse reflectance techniques in a single portable instrument will be evaluated as

SC

regards the analysis of various types of materials in the field of cultural heritage as well as other fields of research such as environmental chemistry and geological

NU

surveys.

MA

Acknowledgments

The authors would like to gratefully acknowledge K. C. Vasilopoulos and G.

D

Kenanakis for providing access to the FT-IR spectrometer and the Raman microscope

PT E

as well as G. Tserevelakis and S. Sotiropoulou for helpful discussions. The work presented in this paper has been supported in part by the IPERION CH project,

AC

654028.

CE

funded by the European Commission, H2020-INFRAIA-2014-2015, under Grant No.

17

ACCEPTED MANUSCRIPT References

1.

Price TD, Burton JH (2011) An introduction to archaeological chemistry. Springer, New York

2.

Uda M, Demortier G, Nakai I eds (2005) X-rays for Archaeology, Springer,

3.

PT

London Stuart BH (2007) Analytical techniques in art conservation. Wiley, West

Creagh DC, Bradley DA (2000) Radiation in Art and Archaeometry. Elsevier

SC

4.

RI

Sussex, UK

Science, New York

Nevin A, Spoto G, Anglos D (2012) Laser spectroscopies for elemental and

NU

5.

6.

MA

molecular analysis in art and archaeology. Appl. Phys. A 106: 339-361 Miliani C, Rosi F, Brunetti BG, Sgamellotti A (2010) In situ noninvasive study

Shugar AN, Mass JL, Eds (2012) Handheld XRF for Art and Archaeology.

PT E

7.

D

of artworks: the MOLAB multitechnique approach. Acc Chem Res. 43:728-38.

Leuven University Press, Leuven 8.

Cheilakou E, Troullinos M, Koui M (2014) Identification of pigments on

CE

Byzantine wall paintings from Crete (14th century AD) using non-invasive Fiber

9.

AC

Optics Diffuse Reflectance Spectroscopy (FORS). J Archaeol Sci 41:541-555 Liang H (2012) Advances in multispectral and hyperspectral imaging for archaeology and art conservation. Appl Phys A 106:309-323 10.

Kogou S, Neate S, Coveney C, Miles A, Boocock D, Burgio L, Cheung CS, Liang H (2016) The origins of the Selden map of China: scientific analysis of the painting materials and techniques using a holistic approach. Herit Sci 4:28

11.

Delaney JK, Trumpy G, Didier M, Ricciardi P, Dooley KA (2017) A high sensitivity, low noise and high spatial resolution multi-band infrared 18

ACCEPTED MANUSCRIPT reflectography camera for the study of paintings and works on paper. Herit Sci 5:32 12.

Comelli D, Valentini G, Nevin A, Farina A, Toniolo L, Cubeddu R (2008) A portable UV-fluorescence multispectral imaging system for the analysis of painted surfaces. Rev Sci Instrum 79:086112 Kantarelou V, Karydas AG, Sokaras D, Mahfouz L, Qurdab A, Al-Saadi M,

PT

13.

RI

Giannoulaki M, Argyropoulos V (2015) In situ scanning micro-XRF analyses of

SC

gilded bronze figurines at the National Museum of Damascus. J Anal At Spectrom 30: 1787-1798.

Ferretti M, Cristoforetti G, Legnaioli S, Palleschi V, Salvetti A, Tognoni E,

NU

14.

Console E, Palaia P (2007) In situ study of the Porticello Bronzes by portable

MA

X-ray fluorescence and laser-induced breakdown spectroscopy. Spectrochim Acta B 62: 1512-1518

Madariaga JM, Maguregui M, De Vallejuelo SFO, Knuutinen U, Castro K,

D

15.

PT E

Martínez-Arkarazo I, Giakoumaki A, Pitarch A. (2014) In situ analysis with portable Raman and ED-XRF spectrometers for the diagnosis of the formation

CE

of efflorescence on walls and wall paintings of the Insula IX 3 (Pompeii, Italy). J Raman Spectrosc 45:1059–67. Vandenabeele P, Weis TL, Grant ER, Moens LJ (2004) A new instrument

AC

16.

adapted to in situ Raman analysis of objects of art. Anal Bioanal Chem 379:137–142 17.

Colomban P. (2012) The on-site/remote Raman analysis with mobile instruments: a review of drawbacks and success in cultural heritage studies and other associated fields. J Raman Spectrosc 43:1529–35.

19

ACCEPTED MANUSCRIPT 18.

Miliani C, Rosi F, Daveri A, Brunetti BG (2012) Reflection infrared spectroscopy for the non-invasive in situ study of artists’ pigments. Appl Phys A 106: 295-307

19.

Arrizabalaga I, Gómez-Laserna O, Aramendia J, Arana G, Madariaga JM (2014) Applicability of a Diffuse Reflectance Infrared Fourier Transform

PT

handheld spectrometer to perform in situ analyses on Cultural Heritage

20.

RI

materials. Spectrochim Acta A 129: 259-267

Colao F, Fantoni R, Fiorani L, Palucci A, Gomoiu I (2005), Compact scanning

SC

lidar fluorosensor for investigations of biodegradation on ancient painted

21.

NU

surfaces. Journal of Optoelectronics and Advanced Materials 7: 3197-3208 Cunat J, Palanco S, Carrasco F, Simon MD, Laserna JJ (2005) Portable

MA

instrument and analytical method using laser-induced breakdown spectrometry for in situ characterization of speleothems in Karstic caves. J Anal At Spectrom

Fortes FJ, Cunat J, Cabalin LM, Laserna JJ (2007) In situ analytical assessment

PT E

22.

D

20:295–300

and chemical imaging of historical buildings using a man-portable laser system.

23.

CE

Appl Spectrosc 61:558–564.

Agresti J, Mencaglia AA, Siano S (2009) Development and application of a

AC

portable LIPS system for characterising copper alloy artefacts. Anal Bioanal Chem 395:2255-2262 24.

Westlake P, Siozos P, Philippidis A, Apostolaki C, Derham B, Terlixi A, Perdikatsis V, Jones R, Anglos D (2012) Studying pigments on painted plaster in Minoan, Roman and Early Byzantine Crete. A multi-analytical technique approach, Anal Bioanal Chem 402: 1413-1432

20

ACCEPTED MANUSCRIPT 25.

Papliaka ZE, Philippidis A, Siozos P, Vakondiou M, Melessanaki K, Anglos D (2016) A multi-technique approach, based on mobile/portable laser instruments, for the in situ pigment characterization of stone sculptures on the island of Crete dating from Venetian and Ottoman period, Herit Sci 4:15

26.

Brysbaert A, Siozos P, Vetters M, Philippidis A, Anglos D, (2017) Materials

PT

analyses of pyrotechnological objects from LBA Tiryns, Greece, by means of

RI

Laser-Induced Breakdown Spectroscopy (LIBS): Results and a critical

SC

assessment of the method, J Archaeol Sci 83 49–61; doi: 10.1016/j.jas.2017.06.007

Anglos D (2001) Laser-induced breakdown spectroscopy in art and

NU

27.

archaeology. Appl Spectrosc 55:186A–205A

Giakoumaki A, Melessanaki K, Anglos D (2007) Laser induced breakdown

MA

28.

spectroscopy (LIBS) in archaeological science-applications and prospects. Anal

Anglos D, Detalle V (2014) Cultural heritage applications of LIBS. In: Musazzi

PT E

29.

D

Bioanal Chem 387:749–760

S, Perini U, Eds, Laser-induced breakdown spectroscopy - Theory and

531-553

Marquardt BJ, Stratis DN, Cremers DA, Angel SM (1998) Novel probe for

AC

30.

CE

applications. Springer Series in Optical Sciences, vol. 182, Springer, Berlin, pp

Laser induced breakdown spectroscopy and Raman measurements using an imaging optical fiber, Appl Spectrosc 52:1148-1153 31.

Telle HH, Beddows DCS, Morris GW, Samek O (2001) Sensitive and selective spectrochemical analysis of metallic samples: The combination of laser induced breakdown spectroscopy and laser induced fluorescence spectroscopy, Spectrochim Acta B 56:947-960.

21

ACCEPTED MANUSCRIPT 32.

Wiens RC, Sharma SK, Thompson J, Misra A, Lucey PG (2004) Joint analyses by laser-induced breakdown spectroscopy (LIBS) and Raman spectroscopy at stand-off distances, Spectrochim Acta A 61:2324–2334

33.

Giakoumaki A, Osticioli I, Anglos D (2006) Spectroscopic analysis using a hybrid LIBS-Raman system, Appl Phys A 83:537-541 Osticioli I, Mendes NFC, Porcinai S, Cagnini A, Castellucci E (2009)

PT

34.

RI

Spectroscopic analysis of works of art using a single LIBS and pulsed Raman

35.

SC

setup. Anal Bioanal Chem 394:1033–41

Kogou S, Lucian A, Bellesia S, Burgio L, Bailey K, Brooks C, Liang H (2015)

NU

A holistic multimodal approach to the non-invasive analysis of watercolour paintings. Appl Phys A 121: 999-1014

Bacci M (1995) Fiber optics applications to works of art. Sens Actuat B 29:190-

MA

36.

196

Picollo M, Bacci M, Casini A, Lotti F, Porcinai S, Radicati B, Stefani L (2000)

D

37.

PT E

Fiber optics reflectance spectroscopy: a non-destructive technique for the analysis of works of art In: Martellucci S, Chester AN, Mignani AG, Eds,

259-265

Dupuis G, Elias M, Simonot L (2002) Pigments identification by fiber optics

AC

38.

CE

Optical sensors and microsystems. Kluwer Academic Publisher, New York, pp

diffuse reflectance spectroscopy. Appl Spectrosc 56:1329-1336 39.

Johnston-Feller R (2001) Color science in the examination of museum objects. The J Paul Getty Trust, Los Angeles

40.

Dyer J, Sotiropoulou S (2017) A technical step forward in the integration of visibleinduced luminescence imaging methods for the study of ancient polychromy. Herit Sci 5:24

22

ACCEPTED MANUSCRIPT 41.

Melessanaki K, Papadakis V, Balas C, Anglos D (2001) Laser-induced breakdown spectroscopy and hyper-spectral imaging analysis of pigments on an illuminated manuscript. Spectrochim Acta B 56:2337-2346

42.

https://www.osram.com/ds/products/lamps/halogen-lamps/decostar/decostar35/index.jsp (last access 29/08/2017) https://www.avantes.com/support/theoretical-background/spectrometers/958-

PT

43.

Pozza G, Ajò D, Chiari G, De Zuane F, Favaro M (2000) Photoluminescence of

SC

44.

RI

introduction-spectrometers/437-sensitivity (last access 29/08/2017)

the inorganic pigments Egyptian blue, Han blue and Han purple. J Cult Herit

45.

NU

1:393-398

Siozos P, Philippidis A, Hadjistefanou M, Gounarakis C, Anglos D (2013)

MA

Chemical analysis of industrial scale deposits by combined use of correlation coefficients with emission line detection of laser induced breakdown

Gaft M, Nagli L, Fasaki I, Kompitsas M, Wilsch G, (2009) Laser-induced

PT E

46.

D

spectroscopy spectra, Spectroch Acta B, 2013 87:86-91

breakdown spectroscopy for on-line sulfur analyses of minerals in ambient

47.

CE

conditions, Spectrochim Acta B 64:1098-1104 Lazic V, Colao F, Fantoni R, Spizzichino V, Teppo E (2007) Online Monitoring

AC

of the Laser Cleaning of Marbles by LIBS Sulphur Detection. In: Nimmrichter J, Kautek W, Schreiner M, eds, Lasers in the Conservation of Artworks , LACONA VI Proceedings, Vienna, Austria, Sept. 21-25, 2005. Springer Proceedings in Physics, vol 115, Springer, Berlin, pp 377-382 48.

Kubelka P, Munk F (1931) Ein Beitrag zur Optik der Farbanstriche. Z Tech Phys 12:593–601

23

ACCEPTED MANUSCRIPT 49.

Kubelka P (1948) New contributions to the optics of intensely light scattering materials. Part I. J Opt Soc Am 38:448-457

50.

Liang H, Keita K, Peric B, Vajzovic V (2008) Pigment identification with optical coherence tomography and multispectral imaging. In Proc. OSAV2008, The 2nd Int. Topical Meeting on Optical Sensing and Artificial Vision, St.

Raman measurements were carried out on mobile Raman microspectrometer

RI

51.

PT

Petersburg, pp 33

SC

(JY Horiba, HE 785) with excitation at 785 nm and on a laboratory Raman microscope (Horiba LabRAM HR Evolution) with excitation at 532 nm

NU

Measurements were performed on a Bruker Vertex 70v FT-IR spectrometer, equipped with a A225/Q Platinum ATR unit with a single reflection diamond

CE

PT E

D

MA

crystal, operating in the range of 7500 – 350 cm-1.

AC

52.

24

ACCEPTED MANUSCRIPT Figure captions

Figure 1: Schematic diagram of the LIBS-diffuse reflectance hybrid spectrometer in the a) LIBS and b) diffuse reflectance measurement configurations.

PT

Figure 2: a) Black body radiation spectrum, calculated for T = 4200 K (λmax ~700

RI

nm); b) emission spectrum of the tungsten-halogen lamp provided by the

SC

manufacturer (λmax ~700 nm) [42]; c) spectral profile of spectrograph-detector (Avaspec-2048L) sensitivity reported by the spectrometer manufacturer [43] and d)

NU

emission spectrum of the tungsten-halogen lamp recorded by the reflectance module

MA

of the hybrid instrument.

Figure 3: Diffuse reflectance spectra of a) malachite and b) red ochre pigments

PT E

instrument (red).

D

recorded on the Perkin Elmer (PE) spectrophotometer (black) and the hybrid

CE

Figure 4: LIBS spectra of a) azurite, b) malachite, c) Egyptian blue; d) diffuse

AC

reflectance spectra of all three copper-based pigments.

Figure 5: a) LIBS spectra and b) diffuse reflectance spectra of iron-based pigments; red ochre (red), yellow ochre (yellow), mars black (black) and green earth (green).

Figure 6: a) Part of LIBS spectrum recorded on the white area of the sherd, b) Parts of the LIBS spectra recorded on the red (red line) and black (black line) areas of the

25

ACCEPTED MANUSCRIPT sherd and c) diffuse reflectance spectra of the white slip (i), red slip (ii), black slip (iv) on the sherd and of red ochre (iii) and mars black (v).

Figure 7: First derivative of the diffuse reflectance spectra from the red slip on the

PT

sherd (black) and the red ochre pigment (red).

RI

Figure 8: a) Diffuse reflectance spectra recorded on the red slip of the sherd (black)

SC

and computed (red) based on the reflectance spectra of red ochre, yellow ochre and CaCO3 samples; b) diffuse reflectance spectra recorded on the white slip of sherd

NU

(black) and computed (red) based on the reflectance spectra of yellow ochre and

AC

CE

PT E

D

MA

CaCO3 samples.

26

ACCEPTED MANUSCRIPT Figures:

a) LIBS configuration Halogen Lamp Optical head (17x7x3 cm 3 ) Lens

Laser pulse Mirror

LIBS emission

Plasma Sample

Bifurcated optical fiber

LIBS dual spectrometer

RI

Mirror

PT

Nd:YAG Laser

Lens

SC

Diffuse reflectance Spectrometer

NU

Laptop computer

b) Diffuse reflectance configuration

MA

Halogen Lamp

Optical head (17x7x3 cm 3 ) Mirror

D

Lens

Nd:YAG Laser

PT E

Reflectance

Sample

Optical fiber

Mirror

LIBS dual spectrometer

Diffuse reflectance Spectrometer

Laptop computer

AC

CE

Lens

Figure 1

27

ACCEPTED MANUSCRIPT

Intensity (arb.units)

a b

PT

c

600 800 Wavelength (nm)

AC

CE

PT E

D

MA

NU

Figure 2

1000

SC

400

RI

d

28

ACCEPTED MANUSCRIPT

a)

Malachite Hybrid PE

40 30 20

PT

% Reflectance

50

400

500

600

700

800

40

Red Ochre Hybrid PE

30

D

20 10

500

600 700 800 Wavelength (nm)

AC

Figure 3

400

CE

0

PT E

% Reflectance

50

MA

b)

NU

Wavelength (nm)

900

SC

0

RI

10

29

900

ACCEPTED MANUSCRIPT

Ca

320

Cu

Cu

330 390 400 Wavelength (nm)

410

Cu Cu

b)

320

PT

Al Ca

Ca

330 390 400 Wavelength (nm)

410

Ca

c)

Azurite Malachite

Egyptian Blue

NU

% Reflectance

60

Si

Ca

320

330

390

400

410

20

d) 0 400

PT E

Wavelength (nm)

AC

CE

Figure 4

40

MA

Cu

D

Intensity (arb.units)

Egyptian Blue

SC

RI

a)

Cu

Malachite

Cu

Intensity (arb.units)

Intensity (arb.units)

Azurite

30

500

600 700 800 Wavelength (nm)

900

ACCEPTED MANUSCRIPT

Yellow Ochre Green Earth

360

370 380 Wavelength (nm)

NU

a)

D

60

PT E

40 20 0

400

500

600 700 800 Wavelength (nm)

AC

b)

CE

% Reflectance

80

Yellow Ochre Green Earth

MA

100 Mars Black Red Ochre

SC

RI

PT

Intensity (arb.units)

Mars Black Red Ochre

Figure 5

31

900

ACCEPTED MANUSCRIPT

Intensity (arb.units)

Ca

White area

Ca

Al

Ca

390

400

410

420

430

RI

a)

Mg

Ca/Fe

NU

Intensity (arb.units)

Ca/Fe

b)

D

MA

Ca

SC

Wavelength (nm)

Red area Black area

370

380

PT E

360

Wavelength (nm)

CE

60

i ii

AC

% Reflectance

80

40

iii

20

c)

0

iv v

400

500

PT

Fe

Si

600 700 800 Wavelength (nm)

Figure 6

32

900

ACCEPTED MANUSCRIPT

Red area Red Ochre

0.4 0.2 0.0

PT

First derivative

0.6

500

600 700 800 Wavelength (nm)

900

SC

400

RI

-0.2

AC

CE

PT E

D

MA

NU

Figure 7

33

ACCEPTED MANUSCRIPT

Red area Computed

50 40 30 20

PT

% Reflectance

60

500

600 700 800 Wavelength (nm)

MA

60

20

b)

500

White area Computed

600 700 800 Wavelength (nm)

AC

Figure 8

400

PT E

D

40

CE

% Reflectance

80

0

900

SC

400

NU

0

a)

RI

10

34

900

MA

NU

SC

Graphical abstract

RI

PT

ACCEPTED MANUSCRIPT

D

Highlights

A novel portable hybrid LIBS-diffuse reflectance spectrometer is presented



It enables studies of paint materials in heritage objects and monuments



It can support analytical campaigns for archaeological investigations

AC

CE

PT E



35