On the Use of OCT to Examine the

Varnish Layer of Paintings Cleaned with an Er:YAG Laser

Adele DeCruz

, Joseph A. Izatt

and Derek Nankivil

Department of Biomedical Engineering and the Fitzpatrick Center for Photonics,

Duke University, Durham, NC 27708, U.S.A.

Departments of Chemistry, Biology and the Fitzpatrick Center for Photonics,

Duke University, Durham, NC 27708, U.S.A.

Keywords: OCT, Varnish, Paintings, Er:YAG Laser, Conservation.

Abstract: Optical Coherence Tomography (OCT) was used to visualize and quantify the varnish layer of paintings.

And to verify the success of efforts by conservators to remove the varnish using laser ablation. An Er:YAG

laser (λ = 2.94μm) with a repetition rate of 15Hz and an optical power of 1mW was used to remove the

varnish. A spectral domain OCT system with a Michelson topology was constructed using a broadband

super-luminescent diode (λ = 840±25nm). The OCT system provided an 8µm resolution, and a field of view

of 5x5mm. Samples, including pigment, varnish and substrate, approximately 1mm

in size, were removed

from the oil painting, San Giorgio Maggiore by Martin Rico (1833-1908). Varnish thickness obtained from

OCT was validated by similar measurements obtained from SEM analysis. Other paintings, including a late

century landscape, signed Thomas Gainsborough, were imaged with OCT to compare neighboring

regions before and after laser treatment and to examine the layering of the artist’s signature in an effort to

determine its authenticity. In conclusion, the non-invasive OCT technique is an efficient tool for

measurement of varnish layer thickness, for imaging over-paint under varnish layers, and for assessing the

effectiveness of laser assisted varnish removal.

1 PURPOSE

The purpose of this study is to demonstrate that

Optical Coherence Tomography (OCT) can be used

to:

1) Measure the varnish layer thickness of paintings,

and

2) Verify that the varnish layer has been removed

after laser ablation-based conservation without

causing changes to the paint layers.

2 INTRODUCTION

To restore the original intent of the artist, art

conservation is moving towards an increased use of

laser ablation to remove varnish layers, which have

become encrusted with contaminants or have been

otherwise altered over the years

(Asmus, 1986);

Maravelaki et al., 1997); Georgious et al., 1998);

Kalaitzaki et al., 1999); Scholten et al., 2000);

(Klein et al., 2000); (De Cruz et al., 2000); (Branco

et al., 2003); (Pouli et al., 2008). It should be

possible to guide the restoration process with

imaging modalities that provide information about

the varnish layer. In paintings where the encrustation

has rendered the varnish completely opaque, OCT

has the potential to provide details about the

structure and thickness of the varnish layer in a non-

invasive manner

(liang et al., 2005); (Gora et al.,

2006).

3 BACKGROUND

Optical coherence tomography enables fast,

noninvasive, high resolution, three-dimensional

imaging of the internal microstructure of weakly

scattering objects. Conventional OCT systems are

coherence-gated interferometers wherein the optical

measurement technique known as low coherence

interferometry is used to measure the magnitude and

echo time delay of backscattered light. In its

simplest manifestation, time-domain OCT

(TDOCT), the illumination is split and sent to both a

DeCruz, A., Izatt, J. and Nankivil, D.

On the Use of OCT to Examine the Varnish Layer of Paintings Cleaned with an Er :YAG Laser.

DOI: 10.5220/0005729801070112

In Proceedings of the 4th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2016), pages 109-114

ISBN: 978-989-758-174-8

109

reference arm and to the sample. Light returning

from the sample interferes with light returning from

the reference arm, and interference fringes are

observed provided that the reference and sample

path lengths are matched to within the coherence

length of the source. Scanning the reference path

length results in a series of interference fringes that

correspond to different depths in the sample. The

photo-detector signal is demodulated to reconstruct

each A-scan (Figure 1).

Figure 1: Schematic of a fiber optic TDOCT system. Light

blue lines represent fiber optic paths, red lines represent

free space optical paths, and thin black lines represent

electronic signal paths.

Further, prior work (Figure 2) has shown that using

lasers for art conservation offer advantages over

conventional methods with solvents and scalpels

(Asmus, 1986); Maravelaki et al., 1997); Georgious

et al., 1998); Kalaitzaki et al., 1999); Scholten et al.,

2000); (Klein et al., 2000); (De Cruz et al., 2000);

(Branco et al., 2003); (Pouli et al., 2008). Some

contaminants and encrustations require very strong

solvents or cannot be removed with a solvent

without removing some of the paint itself. Moreover,

solvent may saturate the substrate, causing it to

swell, materials (from the substrate) may leach into

the solvent, and the conservator may be exposed to

toxic fumes from the solvent. A brief explanation of

the laser cleaning mechanism follows.

The Er:YAG laser, with a wavelength of 2.94



coincides with a strong absorption peak in the

infrared spectra of OH- or NH-containing organic

molecules. The energy of photons at this wavelength

excites a bond vibrational stretching mode. Any

substance containing a high concentration of OH

bonds at its surface has a strong affinity for photons

at 2.94



m, and confines the absorption of these

photons to a surface depth of no more than a few

microns. A painting 's organic contaminant, which

either contains the OH bonds or has been treated

with a thin liquid film (water, alcohol, NH

, OH

)

immediately before lasing, acts as a stain of

relatively high concentration and very high

absorption, providing a natural barrier to energy

penetration into underlying layers.

The energy per photon of the Er:YAG radiation

is not sufficient to break bonds. The energy required

for OH bond dissociation in most organic molecules

ranges from 3.4 to 4.5 eV/molecule, while photons

of 2.94



m wavelength have an energy of only 0.4

eV. Furthermore, the irradiance of the laser is below

the level required to generate multi-photon effects,

which might provide the necessary dissociation

energy. The temperature rise due to cleaning with

the Er:YAG is principally limited to the affected

contaminant volume and reaches its vaporization

maximum of under a 100 °C for a few milliseconds

at most. The bulk of the laser energy goes into the

ejection of the heated contaminant from the paint

surface. The temperature rise in the underlying paint

layer is therefore small, and not sufficient to cause

thermal decomposition of most materials. As a

comparison, surface consolidation or the lining

processes often involve heating of the paint layer too

much higher temperatures for much longer time

periods.

The process as used (with adjustable moderate

pulse energies at a 15 Hz repetition rate) volatilizes

greases with high vapor pressure and can thus be

pictured as a type of steam distillation. Because of

the strong absorption, the photon energy is deposited

in a layer that is only a couple of microns thick on

the targeted surface. This energy goes into near-

instantaneous heating of the absorbing contaminant

through the vaporization of water or grease. The

rapid attendant rise in local pressure causes the

affected volume to be ejected forcefully from the

surface, taking much of the heat with it. As a result,

the underlying non-absorbent film does not undergo

significant heating, as estimated in a previous paper

(De Cruz et al., 2000); (DeCruz et al., 2014).

Figure 2: Micro-photo of the signature from an oil

painting, The Turkish Noble, by Charles Bargue (1825-

1883), date: 1859, showing part of the painter's signature

before and after laser treatment.

PHOTOPTICS 2016 - 4th International Conference on Photonics, Optics and Laser Technology

110

4 METHODS

In this study, OCT was used to visualize and

quantify the varnish layer of paintings. And to verify

the success of efforts by conservators to remove the

varnish layer using laser ablation. A free-running

Er:YAG (MonaLaser, Orlando, Florida) laser with a

central wavelength of 2.94 μm, a repetition rate of

15 kHz and an optical power of 1 mW was used to

remove the varnish. The IR light couples directly

into a 1-mm bore hollow glass fiber about 1-m long.

The other end of the fiber can be manipulated like a

pen.

A spectral domain OCT (SDOCT) system with a

Michelson topology was constructed using a

broadband super-luminescent diode (SLD-371,

Superlum, Carrigtwohill, Ireland) with a central

wavelength of 840 nm and a 50 nm bandwidth and a

line scan CMOS sensor (AViiVA, e2v Inc., Milpitas

CA) with a 20 kHz line rate. The sample arm design

utilized a 4f relay between the first and second

galvanometer (scanner) and a telecentric beam

delivery system to minimize optical distortions. The

OCT system provided an 8.5 µm axial and 7.5 µm

lateral resolution, a sensitivity of 105 dB, an

imaging range of 0.8 mm (6dB fall off) and a field

of view of 5 x 5 mm (Figure 3).

Figure 3: SDOCT system with Michelson topology.

SDOCT permits faster image acquisition and higher

signal-to-noise ratio than the predecessor TDOCT

technology. Through the use of spectral interferometry,

depth information is collected without movement of the

reference mirror. Backscattered light from the sample and

reference interfere, and the broadband interference pattern

is measured with the inverse Fourier transform of the

broadband interferogram. Thus, measurements of the

power spectral density of the interferogram as a function

of wavelength are obtained. Since the temporal coherence

function (also called the autocorrelation function) and the

power spectral density form a Fourier transform pair

(Wiener-Khinchin theorem), the A-scan can be

reconstructed by merely taking the inverse Fourier

transform of the broadband interferogram.

Samples, including pigment, varnish and substrate,

approximately 1 mm

in size, were removed from an

oil painting on panel (San Giorgio Maggiore) by

Martin Rico (1833-1908) and imaged using

Environmental Scanning Electron Microscopy

(ESEM). Varnish thickness obtained from OCT was

validated by similar measurements obtained from

ESEM.

In addition, other paintings, including a late 18

century landscape, signed Thomas Gainsborough,

were imaged with OCT to compare neighboring

regions before and after laser treatment and to

examine the layering of the artist’s signature in an

effort to determine its authenticity.

Two Bioptigen Envisu SDOIS (Spectral Domain

Ophthalmic Imaging System) systems were also

used: 1) R3500 2) R2300. The two systems utilize

different sources and spectrometer designs, but both

permitted visualization of varnish and paint, at

different depths. The 2300 has a lower imaging

depth but higher axial resolution, whereas the 3500

can image deeper with a slightly lower resolution. In

addition, the 2300 system has an 840 nm SLED,

whereas the 3500 has a 1064 nm SLED.

5 RESULTS

The painting San Giorgio Maggiore, Venice, by

Martin Rico (1833- 1908) with the signature, Rico in

the painting's lower left corner is covered by

discolored varnish (Figure 4). In the lighter areas,

the discolored yellow-orange varnish has been

removed with laser ablation.

Figure 4: San Giorgio Maggiore, Venice. Martin Rico

(1833-1908), oil on panel, 6x12” circa 1890; a) varnished

surface before cleaning; b) partially removed varnish; c)

surface with varnish completely removed; d) the signature

Rico covered by varnish.

The shallow penetration of the laser pulse enables

the conservator to remove varnish layers gradually

and this difference of thickness is illustrated in area

b in Figure 4: the varnish layer, microns thin, can be

imaged and measured by OCT (Figure 6).

In figure 5, the OCT scan of the painting layers

with near-infrared light creates a cross-sectional

On the Use of OCT to Examine the Varnish Layer of Paintings Cleaned with an Er:YAG Laser

111

view as well as an en face projection of the surface.

The structure of the paint surface and the varnish

layer are clearly visible with the Envisu SDOIS

R3500. The first bright layer is the varnish-air

interface. The second bright layer is the varnish-

pigment interface. The black space in between is the

varnish itself, measured in one location using on

screen calipers as 65μm. The regions of differing

varnish thickness correspond to the variable texture

of the paint. Basically, in areas where the texture of

the paint made a void or valley, more varnish is

present (i.e. the varnish layer is thicker) and vice

versa in regions where the texture of the paint made

a mound or hill (Figure 7).

Figure 5: OCT B-scan (left) and volume intensity

projection (VIP) of San Giorgio Maggiore acquired with

the 3 μm axial resolution SDOIS R3500.

Figure 6: 3D rendering of the San Giorgio Maggiore

illustrating the underlying brush strokes and varnish

surface texture.

The 3D view (Figure 6), corresponding to the

images shown in Figure 5, confirms distinct

differences between varnish and paint layers.

The thinning of the varnish layer is confirmed in

Figure 7. The thickest layer of varnish is 64.3 m. In

the thinnest regions, the varnish is only 9.6 m

thick.

Interesting results were obtained of the signature

Rico in Figure 8 located in the painting's lower left

corner, which is covered with a thick layer of

discolored varnish.

Figure 7: Detail, Rico painting, area b of Figure 4. Left is

the thinned partially removed varnish layer (after laser

cleaning) and right is the varnish surface before laser

ablation.

The shallow depth of penetration of the Er:YAG

laser pulse at 2.94m is demonstrated in figure 9

The texture of the paint is preserved after lasing,

with no visible damage to individual brush strokes.

To the left in figure 9 is the varnish layer over the

paint. The right is the painting surface with the

varnish removed by laser ablation.

Figure 8: Images of the signature on an oil painting on

wood 19

century Venetian landscape panting by Rico.

Microscope images of the R in the signature acquired at 4x

magnification (A), close-up of R indicated by the light

blue box (B), OCT SVP of the same region shown in B

(C), OCT cross-sectional images show the layer of paint

under the varnish layer and over the painting’s surface (D

& E) acquired along the dotted lines shown in C.

Figure 9: Layers of darker pigmented paint can be

distinguished from subsurface paint layers (different

optical scattering and absorption properties), as well as the

presence of a distinct varnish layer. SDCIS R2300.

PHOTOPTICS 2016 - 4th International Conference on Photonics, Optics and Laser Technology

112

Figure 10: Detail of the varnish and paint surface of Rico

painting. Image taken with Bioptigen Envisu SDCIS

R2300.

SEM images before and after laser ablation are

reported in Figure 11. They were taken to verify the

accuracy of the OCT measurements as well as the

ablation efficiency.

The varnish layer thickness, on the left of Figure

11, was measured in three locations with the

following results: 13.09 μm, 11.87 μm and 13.24

μm. In the image acquired after Er:YAG laser

varnish removal, on the right of Figure 11, the

surface of the paint appears devoid of varnish, and

the texture of the surface of the painting appears

fluid-like with pockets of smoothed areas

Figure 11: ESEM images before (left) and after (right)

varnish removal. Note, ESEM imaging required removal

of a small (approximately 1 mm

) portion of the painting.

Visible under the varnish layer before varnish removal is

the paint layer.

Varnish layer thickness was 10.8 ± 3.8μm and 12.7

± 0.7μm measured by OCT and ESEM respectively.

Complete varnish layer removal was observed in

several regions of the paintings after laser treatment

with occasional residual varnish in regions of

significant surface topological variation.

Additionally, the presence of over-paint and

differences in penetration depth were observed in the

OCT cross-sections (Figure 7).

In Figure 12, the summed voxel projection

(SVP) (top) shows ablation craters in the varnish

layer on the left. In addition, the corresponding OCT

cross-sections taken at locations 1-3 (bottom) are

shown. Notice that there is less varnish with each

retreating edge of the varnish layer (from 1 to 3),

which has been removed with the Er:YAG laser in

each cross-section.

Figure 12: A late 18th century landscape, signed Thomas

Gainsborough, imaged in a region with and without

varnish.

6 CONCLUSIONS

We believe that this is the first demonstration of the

application of OCT to show that the varnish is

removed by Er:YAG laser treatment. Given the

apparently preserved texture of the underlying

pigment after laser treatment, these results suggest

that the laser radiation does not penetrate

significantly into the paint layer.

In conclusion, we demonstrated that OCT may

provide a non-invasive technique that offers

measures of the varnish layer thickness and

verification of its removal after laser ablation-based

conservation efforts.

ACKNOWLEDGEMENTS

This project was funded by the Fitzpatrick

Foundation (Scholar - DN), the Michael J.

Boberschmidt Fund, and the Ottmar Foundation.

REFERENCES

Asmus JF, More light for art conservation, IEEE Circuits

and Devices, March: p. 6,1986.

Maravelaki PV, Zafiropulos V, Kylikoglou V, Kalaitzaki

M, Fotakis C. Laser induced breakdown spectroscopy

as a diagnostic technique for the laser cleaning of

On the Use of OCT to Examine the Varnish Layer of Paintings Cleaned with an Er:YAG Laser

113

marble. Spectrochim. Acta B, 52:41, 1997.

Georgiou S, Zafiropulos V, Anglos D, Balas 3. C, Tornari

V, Fotakis C. Excimer laser restoration of painted

artworks: procedures, mechanisms and effects. Appl.

Surf. Sci., 127–129,:738, 1998.

Kalaitzaki P, Zafiropulos V, Fotakis C. Excimer laser

cleaning of encrustation on pentelic marble: procedure

and evaluation of the effects. Appl. Surf. Sci., 148:92,

1999.

Scholten JH, Teule JM, Zafiropulos V, Heeren RMA.

Controlled Laser cleaning of painted artworks using

accurate beam manipulation and on-line LIBS-

detection, J. Cult. Heritage, 1:S215, 2000.

Klein S, Stratoudalsi T, Marakis Y, Zafiropulos V,

Dickmann K. Comparative study of different

wavelengths from IR to UV applied to clean

sandstone. Appl. Surf. Sci., 157:1, 2000.

De Cruz A, Wolbarsht ML, Hauger SA. Laser removal of

contaminants from painted surfaces, J. Cult. Heritage,

1:S173, 2000.

Bracco P, Lanterna G, Matteini M, Nakahara K, Sartiani

O, De Cruz A, Wolbarsht ML, Adamkiewicz E, and

Colombini MP. Er:YAG laser: an innovative tool for

controlled cleaning of old paintings: testing and

evaluation. J. Cult. Heritage, 4:202s-208s, 2003.

Pouli P, Paun I-A, Bounos G, Georgiou S, Fotakis C. The

potential of UV femtosecond laser ablation for varnish

removal in the restoration of painted works of art.

Appl. Surf. Sci., 254:6875-6879, 2008.

Liang H, Cid MG, Cucu RG, Dobre GM, Podoleanu AG,

Pedro J, Saunders D. En- face optical coherence

tomography – a novel application of non-invasive

imaging to art conservation. Optics Express,

13;16:6133-6144, 2005.

Gora M, Targowski P, Rycyk A, Marczak J. Varnish

ablation control by optical coherence tomography.

Laser Chemistry, 2006.

DeCruz A, Andreotti A, Ceccarini A.Colombini MP.

Laser cleaning of works of art: evaluation of the thermal

stress induced by Er:YAG laser. Applied Physics B,

2014.

PHOTOPTICS 2016 - 4th International Conference on Photonics, Optics and Laser Technology

114