|
Chapter 2.3 Laser Cleaning of Paintings – Principles and Case Studies Vassilis
Zafiropulos Department of Human Nutrition and
Dietetics, Technological Educational Institute (TEI) of and Institute
of Electronic Structure & Laser, Contents 2.3.2 Optimization of Laser Parameters: General Considerations 2.3.3 Ablation Efficiency Studies 2.3.4 Light Transmission Studies 2.3.5 Chemical Alteration of
the Substrate 2.3.6 Laser Divestment of Aged Resin Layers from Paintings 2.3.7 Er:YAG Laser in the Removal of Surface Layers 2.3.9
References Among
other laser cleaning applications, pulsed lasers have also been used for the
removal of surface encrustations from marble sculptures and stonework [1-7]
or aged varnish from paintings [8-14]. The successful implementation of these
techniques has been based on the close collaboration of conservators with
laser scientists for defining the optimum process end point. Extensive
literature on laser-based conservation techniques exists in LACONA Conference
proceedings [15-18]. Cleaning
of artworks and antiquities by lasers provides the opportunity for the
selective removal of undesired surface layers. In principle, it is possible
to leave the original delicate surface or substrate entirely unaffected. For
the on-line monitoring and in situ control of laser cleaning the LIBS (Laser
Induced Breakdown Spectroscopy) technique has been proved to be a valuable
tool [4-6, 11, 19-21] (see also Chapter 4.2). In
Chapter 2.1
the presentation of laser divestment applications has been based on the nature
of the layer to be removed. Here in Chapter 2.3, we consider the complex
polymeric substrates where mainly UV lasers of nanosecond or picosecond pulse
duration are necessary. Such materials include oxidized resins - with or
without inorganic inclusions - that have been used as protective layers on
top of precious painted substrates. More complicated situations may be
encountered when the delicate substrate has been covered with a number of
overpaint layers. In such cases, the organic binding medium is usually oil
(e.g. linseed oil) that has polymerized with time. In contrast to inorganic
encrustation/surfaces, where laser divestment mechanisms are characterized by
the self-limiting character of the ablation process, polymerized layers are
usually difficult to treat without processing a detailed know-how and
experience. Here
we will try to provide an overview of the critical factors playing an
important role in the successful laser cleaning of aged resins. Chapter 2.3
has been divided into six more sections starting with some general concerns
related to the optimization of laser parameters. Next a review of recent
studies on ablation efficiency, light transmission and chemical alterations
is presented. The reader may find next some test case studies of laser
cleaning applications on painted surfaces. Finally, some conclusive remarks
are presented.
2.3.2
Optimization of Laser Parameters: General Considerations The
ablation of polymers by short UV laser pulses has been studied due to the
plethora of technological applications [22-24]. A comprehensive review of the
subject can be found in the books of Bäuerle [25] and Luk’yanchuk [26]. Let
us review first some important findings in the laser ablation of polymers,
which resemble the laser ablation of aged resins. In the UV laser ablation of
polymers, the ablation products typically consist of small fragments (atoms,
small molecules and their ions) and medium sized fragments, such as monomers
as well as large portions of them [27-29]. These fragments are created by the
breaking of covalent bonds by direct photodissociation or thermal
decomposition. Photochemical and photothermal ablation could be considered as
limiting cases of the photophysical mechanism, where the removal of
electronically excited species from the surface is taken into account
[30-33]. The
photochemical modifications induced by UV laser ablation of doped PMMA films
was investigated by Georgiou et al. (1998) [10] and Athanassiou et al. (2000)
[34], in simulating the ablation processes in natural resins. However, we
should mention that polymerized materials found in paintings are rather
complex. For example aged polymerized resins may consist of hundreds of
different organic molecules irregularly cross-linked in a 3D network. There
are also different salts in the form of clusters, crystals or particulates
embedded into the polymeric network. Recent studies [13,35,36] have revealed
the presence of a gradient in polymerization and oxidation across the film
thickness giving an additional complexity to the ablation process (see
insert). Here
an attempt will be made to draw the basic guidelines for choosing the proper
laser parameters for safe laser ablation of unwanted surface layers without
affecting the underlying, usually valuable, substrate. First we make the
assumption that the adequate laser wavelength for the ablation of these
materials is in the UV up to the limiting laser wavelength λL = 248 nm
(photon energy of 5 eV) or in some cases 266 nm. At these UV laser
wavelengths the optical penetration depth equals lα= α-1 << dp < dtot
where α is the linear absorption coefficient and dp
is the thickness of the polymerized surface layer to be removed, while dtot
is the total thickness of the resin. A typical cross-sectional structure of
the painting is shown in Fig. 2.3.1. The above condition assures the safety
of the substrate (pigment / paint) that is under the remaining resin. For
highly polymerized natural resins lα is of the order
of 0.1 µm for λL = 193 nm and 1 µm for λL =
248 nm, while dtot is usually 10-100 µm. For
the pulse duration, we assume the use of the easily available nanosecond
lasers without ignoring the ultrashort UV pulse lasers, which are not yet
handy for mass applications. Another point that must be clearly noted here is
the usual choice of the KrF laser over the ArF excimer laser. Although the
193 nm photons offer a better result than the 248 nm photons owing to the
much lower lα of the former, the overall efficiency of ArF excimer laser
is an order of magnitude lower than the efficiency of KrF excimer laser
systems. This is owing to both the commercially available power levels and
the much lower ablation rates of the former. Therefore, from the application
point of view we concentrate on λL = 248 nm with the proviso
that for more delicate interventions the choice will be a shorter wavelength.
Apart
from the wavelength and the pulse duration of the laser, the most important
parameter is the laser energy density or fluence. According to the special
requirements of the specific applications, it is of major interest to
quantify the transmission of laser light through the resin layer as a
function of incident fluence. Aged varnish usually consists of unidentified
complex polymerized material with random inorganic inclusions. An
experimental methodology has been developed for deriving the optimal/best
fluence for each particular application. Three different approaches have been
considered: ablation efficiency studies, light transmission studies, and
chemical alteration studies. 2.3.3
Ablation Efficiency Studies Common
problems encountered in painting conservation have to do with ageing
associated with polymer phase formation, oxidation, cross-linking and
photochemical degradation of the varnish layer, accumulation of various kinds
of pollutants on the surface that can be further polymerized and cross-linked
to the resin, or even over-paints on the original painting that must be
removed. Removal of these surface layers must be accomplished in such a way
that the integrity of the original work is guaranteed. Laser
cleaning is based on the removal of a well-defined surface layer under fully
controlled conditions. In the case of homogeneous resin layers, e.g.
considering constant physical-chemical properties in the top layers of the
aged varnish, the ablation rate is constant for the surface layer to be
processed provided that the laser fluence is kept constant. Therefore, a
step-by-step scanning of the laser beam over the surface results in a layer
removal of controlled thickness (see Fig. 2.3.1). It has been found that,
when processing the deep resin layers (depths > dp), the
ablation rate changes, resembling that of fresh (non-aged) resin [35].
Characteristic
ablation rate curves as a function of laser fluence obtained from model and
real samples using the KrF excimer laser (emitting at 248 nm, pulse duration
of 25 ns) are shown in Fig. 2.3.2a. The
four representative polymerized materials presented here are two types of
artificially aged varnish (dammar and gum lac), a multi-layer polymerized
linseed oil-based over-painting and a black polyurethane paint film. In
short, artificially aged resin films may be made [13] by applying the film on
a quartz plate via spin-coating and then by exposing the sample to UV
radiation and elevated temperature. The end point of the exposure is reached
when the absorbance stabilizes. The way ablation rate measurements are made,
as well as the dependence of ablation rate on laser wavelength and pulse
duration are described elsewhere (e.g. see [35,37]) and are outside the scope
of this presentation. The data in Fig. 2.3.2a show the mean removed depth per
laser pulse as a function of laser energy fluence in units of J/cm2.
It can be seen that the resolution attained by the KrF excimer laser can be
as high as 0.1 µm per pulse, which is simply impossible using other
mechanical or chemical techniques. Fig. 2.3.2a also demonstrates the attainable
control of varnish removal. It is possible to completely adjust the depth of
each layer removal by choosing the laser fluence and/or the number of pulses.
The abscissa has intentionally a logarithmic scale to demonstrate the linear
response of ablation rate to the logarithm of the laser energy fluence, at
low fluence values excluding the so called Arrhenius tail [25,27]. For
every polymerized material this linearity is always strictly localized within
a narrow range of laser fluence. The
ablation efficiency curves (Fig. 2.3.2b) provide the optimum laser
fluence for processing the material. Here the word ‘optimum’ refers to the
fluence where the ablation becomes most efficient, and it should not be
confused with the fluence corresponding to the highest step-resolution (i.e.
just above the ablation threshold). In the next section it is shown that at
optimum fluence the transmission becomes minimal. The efficiency is a quite
significant parameter in cases where large surfaces or thick varnish layers
are encountered, and therefore, the goal is to remove the maximum possible
volume of material per available photon (units Å3/photon). The
data presented in Fig. 2.3.2b indicate for instance, that the optimum laser
fluence for the ablation of the particular samples of artificially aged
dammar and gum lac is 0.3 and 0.2 J/cm2 respectively, while for
the specific polymerized overpaint it is about 0.6 J/cm2. Another
important observation in Fig. 2.3.2b is the similar behaviour of experimental
trends. As the energy fluence increases, the efficiency rises until it
reaches a maximum. This maximum corresponds to the point where saturation is
reached in the bond breakage within the first layers of the surface and/or
where the screening effect starts to occur (part of the laser pulse is
absorbed and/or scattered by the generated photofragments). This may be
supported by the onset of plume generation at fluence values at or just above
the optimal fluence. At this optimum laser fluence, the excited chromophores
[38] - light absorbing sites within the 3D macromolecular lattice of the
polymerized phase - have reached a maximum density. For even higher fluence
values, the efficiency drops. Especially in the case of the artificially aged
varnishes as we further increase the energy fluence and above a certain value
(0.9 and 0.35 J/cm2 for dammar and gum lac, respectively), the
ablation efficiency starts to increase for a second time. This phenomenon may
be attributed to the fact that, for such laser fluence values the energy is
sufficient for additional loss of material through vaporization. At this
fluence range photomechanical effects, for example thermoelastic stresses and
shock wave formation, may also be important during laser ablation. Their
direct contribution in laser ablation becomes significant when the ablated
polymerized material includes inorganic particulates (e.g. overpaintings)
and/or when there is delamination that can lead to detachment/ejection of
flakes. In such cases a large percentage of the material leaves the surface
as small particles or flakes. When removing uncontaminated not-fragile aged
varnish though, the shock wave produced seems to have a negligible effect
especially when working near the optimum laser fluence. Of course, this
applies to the contribution of the photomechanical effects in the removal
process itself. The possible long-term consequence of the shock wave is a
different issue that is currently under investigation (e.g. see the work by
Tornari et al. [39]). Another
phenomenon that cannot be overlooked is the so called incubation. The word
‘incubation’ is fluently used in laser ablation terminology and it stands for
the effect of the laser on the surface following the elapse of the laser
pulse. The incubation generally results in non-stable ablation rates for the
first laser pulse or pulses interacting with a surface. In general,
incubation phenomena are very important especially for fluence values near
the ablation threshold [25]. Although a detailed analysis is essential to
study the effects occurring in this low energy regime, it finally turns out
that in the applications under consideration the optimal/best fluence is well
above the ablation threshold, where incubation was found to be negligible.
Here
it is worthwhile to present some results obtained using ultrashort laser
pulses. Figure 2.3.3 shows the ablation rate and efficiency data of
artificially aged dammar resin (same as in Fig. 2.3.2) using laser pulses of
500 fs duration and λL = 248 nm. The nanosecond data is also
plotted in the same graph for comparison. The maximum ablation efficiency for
both pulse durations is reached at about the same laser fluence (~0.3 J/cm2).
For this particular resin and for the encountered low degree of polymer
formation (5-7 monomers) it seems that the ablation rate depends on laser
fluence and not flux. One of the many different reasons for such a behaviour
may be the long relaxation times of the excited sites/bonds of the low
polymer complexes in the particular material.
Finally,
for a highly polymerized naturally aged varnish the situation is expected to
be different. In such a case the sub-nanosecond data has been found to be
clearly different compared to the nanosecond data. Fig. 2.3.4 shows the
results for three different pulse durations (500 fs, 5 ps and 25 ns) at
λL = 248 nm obtained from an unknown highly polymerized
varnish, which has undergone a natural aging of more than 60 years. Here the
corresponding relaxation times must be of the order of picoseconds owing to
the dense polymer network. Such short relaxation times have been measured in
polymers [25,26]. As a result, it is expected that the operative mechanisms
are non-thermal when a sub-picosecond UV laser pulse interacts with a well
polymerized material such as a naturally aged resin. This may be seen at SEM
images of the same resin used to acquire the data of Fig. 2.3.4, after laser
ablation using the three different pulse durations. Fig.
2.3.5a shows the SEM image of the reference surface, while Figs. 2.3.5b-d
show the resin after removal of ~1 µm film thickness with pulses of different
durations (500 fs, 5 ps, and 25 ns) at λL = 248 nm. The laser
ablation was carried out under scanning mode operation with overlapping laser
spots. The white particles in the images are salts mainly containing Ca and
Na, usually present in the natural resins (also see section 2.3.6). The
surface left after using nanosecond pulses has a very smooth texture, like a
surface which has undergone melting (Fig. 2.3.5d). Additionally, the inorganic
particles have also undergone a melting-freezing transformation, resulting in
the sharp peaks seen in Fig. 2.3.5d. On the other hand, the surface of the
resin and the morphology of the inorganic particles remain as in the
reference after ablation by sub-picosecond or picosecond pulses (Figs. 2.3.5b
and c, respectively). This is the main advantage of short UV laser pulses, in
addition to the higher resolution compared with the nanosecond pulses. In
contrast, the nanosecond UV laser has been proved to be much more efficient
and in addition, it is easily available; therefore it is the best choice for
‘regular laser interventions’, where the ~1 µm order of magnitude resolution
in depth-step is adequate.
2.3.4
Light Transmission Studies During
the laser ablation of polymerized resins part of the laser light may be
transmitted through the resin, finally reaching the underlying substrate.
Therefore, quantification of the optical penetration depth is significant.
The goal here is to measure the transmitted photon energy and find the
optimal laser parameters for minimizing the transmitted light. The
methodology presented here permits the measurement of the transmission of
laser photons through model resin samples on quartz. A similar study on
polymers, e.g. polyimide, has been reported by Pettit et al. [40]. During
laser ablation, the energy of laser light penetrating the sample is measured
for each successive laser shot. This provides the energy of UV transmitted
light as a function of depth. If
the intensity of the laser beam that irradiates the surface of the film is I0,
the intensity of the transmitted light is Itrans and the intensity
of the dissipated light (in ablation, reflection, scattering etc) is I1,
the measured absorbance, A, of the remaining resin film is given by:
where
α is the effective linear absorption coefficient and d the film
thickness left after every laser pulse. Here for simplicity we assume that
α is constant over the film thickness, although this is not quite true
in aged resins [13,35,36]. Eq. (1) results in:
According
to Eq. (2) the results may be plotted as log(Itrans/I0)
versus d. Such a graph is presented in Fig. 2.3.6 for an artificially
aged mastic film prepared on a quartz plate. For this, the laser beam is
appropriately shaped and directed on the sample. A sensitive energy meter
measures the energy of transmitted light for each consecutive laser pulse.
For each laser fluence value used, a series of laser pulses were fired until
the whole mastic film was removed. The d values are calculated using the mean
ablation rate data and the total film thickness. The light reflected by the
surface is taken into account by normalizing the Itrans
measurements, comparing them with the transmission after the total removal of
the film. The slope of the graph represents the effective absorption
coefficient for the laser light. This graph provides the following
information: —
The total energy density of transmitted light. This can be calculated if we
know the initial film thickness, the fluence and ablation rate, and the
number of laser pulses fired. Knowledge of the total transmitted energy
density is crucial when the substrate is sensitive to the laser light. —
The optimal laser fluence can be identified, where ‘optimal’ in this case is
defined in terms of least transmitted light at a certain d value. Although
here ‘optimal’ is defined differently than in section 2.3.3, it turns out
that the two different definitions give the same value as it is shown below.
Here
it is interesting to compare the transmission data with the ablation efficiency
data obtained from the same samples. Figs. 2.3.7a and b show the results of
ablation rate and ablation efficiency studies, respectively, of the same aged
mastic sample as in Fig. 2.3.6. The six different fluence values that were
used in obtaining the transmission data presented in Fig. 2.3.6 are marked in
Fig. 2.3.7b for comparison. At the low fluence of 0.2 J/cm2 just
above the ablation threshold the transmission of laser light is rather high.
For slightly higher fluence values we observe a sharp drop of the
transmission, reaching a minimum at a fluence, which corresponds to the
highest ablation efficiency. For higher fluence values the transmission
curves tend to coincide, possibly owing to plume absorption and screening.
Fig. 2.3.7c shows the transmission as a function of fluence when the
remaining mastic resin is 3 µm. A comparison between Figs. 2.3.7b and c shows
that the optimum fluence defined in terms of transmission coincides with the
optimum fluence defined in terms of efficiency. This has been found to be a
general rule for a series of different polymerized resins [35]. The results
are in general in accordance with the three-level chromophore model of Pettit
et al. [40].
The
implications of these observations are essential for the safe use of
UV-pulsed lasers in the removal of superficial resin layers from paintings or
from other delicate substrates. It is also important to note that in such
applications the approach cannot be empirical and certainly cannot be the
same as in the removal of inorganic encrustation. In the latter case the end
user starts from low fluence values progressively increasing the energy density
until a “cleaning threshold” is reached. In the case of polymerized resins,
however, such an approach could endanger photosensitive substrates.
2.3.5
Chemical Alteration of the Substrate Most
known pigments used in paintings consist of inorganic salts mixed in a
binding medium, usually linseed oil or egg tempera. On top of the pigments, a
resin layer has usually been applied, mainly for protecting the pigments from
the accumulation of dirt. It has been found [41] that resins also protect
pigments from the ambient UV light that induces free radical formation and
subsequent oxidation reactions both in the resin and in the binding medium.
At the same time most salts are sensitive to direct laser irradiation
undergoing chemical and/or phase transformation [42-44]. As described in the
next section, the top layers of the aged resin must be removed from the surface
of the pigment. During laser-assisted divestment, the transmitted light may
lead to long-term deterioration of the original surface of the painting. To
address the issue, laser ablation was combined with Gas Chromatography
coupled with Mass Spectroscopy (GC-MS) analysis. The goal was to detect
possible oxidation products induced by laser radiation [12]. The
process of degradation of the linseed oil medium under oxidation conditions
involves a simultaneous polymerization and de-polymerization reaction. In
general, unsaturated fatty carbon chains undergo various oxidation reactions
resulting in unstable peroxides. These finally result in conjugated products,
of up to nine carbon atoms. At the same time, 3D cross-linking occurs as a
basic consequence of the presence of radicals. ‘Oxidative cleavage’ (bond
breakage accompanied with oxidation) and production of low molecular weight
volatile products as the end-result of the possible oxidation processes, is
routinely investigated by gas chromatography [45]. After trans-esterification
of all possible glyceryl esters with hydrochloric methanol, a mixture of
lower molecular weight methyl esters of all present fatty acids is produced
and analyzed. Such a gas chromatogram e.g. from an aged linseed oil medium
reveals that: —
long unsaturated fatty chains (e.g. oleate) are decreased due to oxidation, —
saturated fatty chains are basically unchanged and —
a number of oxidation products, such as dimethyl esters of azelaic
(nonadioic) acid are formed. The
quantities of esterified fatty acid residues such as methyl palmitate
(C16:0), stearate (C18:0), oleate (C18:1), as well as monomethyl and dimethyl
azelate (C9) have been previously used as a measure of various oxidative
routes in oils [45]. Here the two numbers correspond to the number of C atoms
and the number of double bonds, respectively. Oleate is slowly oxidized
(decrease in concentration), while azelate is produced as an oxidation
product, thus enabling an evaluation of the aging process. Therefore, low
C18:1/C18:0 or C18:1/C16:0 and high C9/C16:0 or C9/C18:0 ratios suggest a
strongly oxidized medium. Based
on the above, model samples of cinnabar pigment in linseed oil covered with
dammar varnish (~20 µm thickness) were made and artificially aged using a UV
lamp and elevated temperature. Then for each sample a predetermined thickness
of varnish was removed by choosing the value of laser fluence (KrF excimer)
based on the ablation curves of dammar (see Fig. 2.3.2). Note that the same
dammar samples were used to obtain the ablation rate data and the GC-MS data.
Finally, the ensemble {pigment, oil medium, remaining varnish} of each sample
was analyzed by GC-MS (for details see [12]).
In
general, the results show no increase of oxidation products when at least a
thin varnish layer is left over the pigment and at the same time the optimum
laser fluence is used. On the other hand, when a high laser fluence is used
the oxidation products increase. In more detail, Fig. 2.3.8 presents the
C18:1/C16:0 and C9/C18:0 ratios of the GC-MS peaks for different values of
removed resin layer. Based on the results described in the previous sections,
the optimal fluence for the ablation of the particular aged dammar layer is
0.3 J/cm2, while a fluence of 0.8 J/cm2 may be
considered rather high for this specific resin/sample (see Fig. 2.3.2b).
Figures 2.3.8a and b correspond to the results using these two values of
energy fluence, respectively. Figure 2.3.8a shows that there is no evidence
of any oxidation change compared to the reference sample, even when the
remaining varnish layer is only 2 µm thick. On the contrary, when using a
laser fluence of 0.8 J/cm2 the concentration of the oleic
component which is sensitive to oxidation (C18:1) is decreased as the
thickness of the removed layer increases (see Fig. 2.3.8b). At the same time
the concentration of the C9 oxidation product increases - the exception to
the last point may be due to the further oxidation of C9 to lower molecular
weight products. Therefore, it is concluded that the chemical alteration of
the binding medium can be avoided when using the optimal/best laser fluence
that is found to be different for each resin and its polymerization
parameters. At the same time a portion of the resin must be left for
“filtering out” the transmitted light. An additional reason for letting a
portion of the varnish on top of the pigment is the following: It has been
recently proven [35] that owing to the existing gradient in resin degradation
[13], the lowest resin layers (the ones next to the pigment) have a chemical
structure almost identical to the unaged resin. This is due to the lack of
oxygen and ambient UV-photons in these ‘deep layers’. The
combination of several analytical techniques has been used to characterize
and quantify the chemical and physical effects induced by KrF excimer laser
cleaning of tempera paint systems [44]. An understanding of the effects of the
UV laser radiation reaching the paint layers is necessary in order to define
the consequences that could result from the total removal of a protective
varnish or over-paint layer in a worst case scenario of laser cleaning. The
direct interaction of the laser at high fluence can result in irreversible
unwanted changes, although these effects are constrained to the surface of
the sample and are strongly dependent on the nature and chemical
characteristics of the paint. It was found that organic tempera paint systems
are very stable under intense laser treatment. In contrast, direct laser
irradiation of inorganic paint systems induces various types and degrees of
alterations in the properties of both the pigment and the binding medium,
which are not a mere addition of effects on individual components in the
mixture. The observed discoloration could be due to a modification of the
chemical composition of the pigment through a change in its state of
oxidation or due to a change in the pigment crystalline phase. On the other
hand, the ablation of organic material from the substrate leads to the
formation of a thin layer of charred material that covers the original paint
surface. This layer is not produced in the absence of the inorganic pigment
in the paint, indicating that pigment particles mediate the charring through
an energy-transfer mechanism. Clearly, the actual paint/pigment composition
will determine which of the two effects, change of pigment chemical
composition or charring of the outermost paint layer, is dominant. Finally,
it was found that modifications induced by laser irradiation at laser fluence
values below the ablation threshold in both organic and inorganic paint
systems were consistently less severe or undetectable. In conclusion, the
results obtained by Castillejo et al. [44] further confirm the adequate
strategy for laser cleaning consisting of a controlled partial removal of the
outermost varnish layer, leaving a protective coating that prevents
discoloration or other chemically harmful effects.
2.3.6
Laser Divestment of Aged Resin From Paintings The
application of varnish layers plays a significant role in protecting painted
surfaces. During aging, the varnish degrades through oxidation,
auto-oxidation, and cross-linking processes, catalyzed by the absorption of
light [46,47]. In particular, the absorption of UV light leads to the
formation of products playing a significant role in the polymer phase
formation (cross-linking). For aged natural resins it has been found
[13,36,48] that certain optical properties related to polymerization and/or
oxidation (e.g. UV or IR absorption, concentration of carbonyls etc.) are
scalar with depth from the surface. This is owing to the exponential decrease
of ambient light as it propagates into the resin. Note that UV light
catalyzes the formation of radicals and therefore the polymerization
processes. The gradient of oxidation into the film is directly related to an
equivalent gradient of the solubility throughout the thickness of varnish
layer [36,48]. Consequently, the deeper one goes into the varnish layer the
more dilute the solution (of the appropriate chemical agent/solvent) required
to remove the remaining layer. As a general rule, the required concentration
drops by a factor of two when going from the surface to a depth of about 8-10
µm. Due
to preservation reasons the artwork must be relieved of the degraded varnish
that can affect the inner layers [49]. Chemical cleaning is a complicated
procedure that employs the action of organic solvents potentially able to
initiate a process of swelling and leaching of the film [50]. The outcome may
be a significant alteration of the constitution of the image layers of
paintings. It has been found that laser-assisted removal of the top varnish
layers (~8-10 µm) is not only very helpful but also unique, since it cannot
be matched by any other cleaning method [51]. Moreover, it offers the choice
of either leaving the rest of the resin intact (preserving the “touch of
time”) [11,35] or further removing it, using less aggressive and toxic
solvents than in a regular chemical treatment [11,35,36]. The
laser-assisted removal of the outermost aged varnish layers must be performed
at the optimal fluence, as described in the previous sections. When working
on real paintings, it is not trivial how to determine the optimal fluence.
The only established methodology [20] is to go through an ablation
rate/efficiency study using a small part of the aged varnish (typically at
the edge of the painting) for generating the ablation spots. An area of 5 x
50 mm2 is enough, where a number (usually 7-10) of ablation spots are
created, using a different fluence value for each. Depending on the roughness
of the surface and the varnish thickness, each ablated spot should
practically have a depth of the order of 5-10 µm in order to be measured with
minimal error and, at the same time, not to penetrate into the deep varnish
layers where the resin is basically unaged (see insert). Profilometry is
applied across the Gausian direction of the rectangular spot and its
integrated area is divided by the spot size along the same direction. This
methodology provides the mean ablation rate across the whole beam profile.
The procedure is repeated for at least three parallel directions spanning the
top-hat direction of the excimer beam and a final mean ablation depth is
derived, which is in turn divided by the number of laser pulses fired. In
this way the mean ablation rate is obtained. All ablation rated presented in
this chapter are mean ablation rates measured following the methodology
described above. Finally, the ablation efficiency is calculated from the mean
ablation rate. A description of such measurements is given elsewhere [35,37].
A second methodology for obtaining the optimal fluence, which is still under
investigation, is based on the onset of plume generation. Plume starts to
appear at fluence values near or just above the optimal fluence. The
plume observation has also been associated with the well known monitoring
technique, LIBS, or Laser Induced Breakdown Spectroscopy, being a
micro-destructive technique. Here, however, the usually observed plume emission
during laser ablation is a useful byproduct that has been used for on-line
controlling the exact depth advancement. [4-6,11,19,20,52]. During laser
ablation the emission produced is characteristic of the composition of the
ablated material. In most cases, the layer composition changes in a predicted
way as the material removal advances towards the final step/depth, and it is
exactly this spectrum alteration that can be used to guide and control the
laser ablation process. There are a number of interesting examples of the use
of laser ablation methods described in the literature. The aim of the
following paragraphs is not to provide a literature review but rather to
highlight some interesting applications on different materials.
For
automatic control during the laser cleaning process, the light emission
spectra are continuously recorded in a pre-selected spectral region. The data
acquisition processor is programmed to calculate certain peak intensity
ratios that are then used as inputs in an algorithm, which produces an output
value after each laser shot. When this value falls outside certain
boundaries, the laser stops firing and the motorized x-y-stage moves the
painting or the laser beam to a new position where the cleaning process
starts again. At this new position, laser pulses are delivered until the
output value moves again out of the chosen confines. The distance that the
sample or beam is moved in every step is determined by an overlapping
protocol for the sequential sample movement in both the x and y directions. A
schematic of the experimental setup is given in Fig. 2.3.9. Besides
the plum analysis methodology described above, it has been recently found
that the detection of the plum intensity can be used as well to monitor the
endpoint of laser ablation [53]. The methodology is based on an automated
closed-loop laser cleaning where a reference value of plume signal being
necessary for comparison of the required and actual signal must be defined by
the operator. This can be found either based on restorers experience or by a
teaching procedure on test samples. During the laser cleaning process the
plume signal of each laser pulse is compared with the reference signal. Laser
removal continues until the reference signal is reached. In a next step, the
x-y control moves the sample to a new position. At
this point it is worthwhile presenting another piece of evidence for the
existing scalar properties in the resin, which has been used for on-line
control during the laser-assisted removal of aged varnish. Apart from polymer
phase formation [35], there is also natural oxidation [46] which has taken
place especially at the outer layers [35]. In the case of a single resin
layer, the elemental composition throughout a cross-section of the layer does
not change with aging. This fact could be prohibitive in using LIBS as a
monitoring technique, since LIBS is in general an atomic analytical method
(see Chapter 4.2).
However, in LIBS experiments that were performed on naturally aged resin
samples we found that there is a change in the spectrum of the produced plume
- as the varnish removal progresses towards the lower varnish layers -
originating from produced excited dimers. Fig. 2.3.10 shows a typical series
of LIBS spectra observed when successive laser pulses ablate aged varnish.
The
four spectra correspond to the 2nd, 3rd, 4th,
and 5th pulse respectively. A comparison of the four spectra
reveals the gradual decrease of the CO B1Σ-A1Π
bands (Ångström bands) and the simultaneous increase and then stabilization
of the two most intense C2 A3Πg-X’3Πu
(Swan bands). The CO bands originate from the carbonyl groups (products of
oxidation, e.g. see [35,46]) whose density is expected to be higher in the
outer varnish layers than the inner ones. This is verified by observing the
micro-FTIR spectra (reflection mode) of the surface after each consecutive
laser pulse.
Figure
2.3.11 shows a representative FTIR spectrum [54]. The ratio of the peak
attributed to carbonyl groups (1650-1730 cm-1, marked as A) versus
the characteristic C-H bending of methyl and methylene groups (1400-1470
cm-1, marked as B) was found to decrease after each laser shot (see small
graph enclosed in Fig. 2.3.11). This is in accordance with the LIBS
observations shown in Fig. 10. In this particular varnish system, the
limiting value of A/B was found to be 1.4 ± 0.1. Here
we will present a few test case studies to demonstrate the correct
application of laser-assisted outermost-layer removal in real paintings. The
research has been carried out in collaboration with Michael Doulgeridis,
director of the Conservation Department of the National Gallery of Athens,
who has provided the paintings and offered his guidance from the
conservator’s point of view. The first example was an 18th century
Flemish tempera painting on a wooden panel of size 27 x 37 cm². It suffered
from an unsuccessful restoration some decades ago, and the varnish that was
used to protect the painting had undergone severe polymerization and was
consequently hard as glass. The optimal fluence was found to be 0.38 J/cm2
with an ablation rate of 0.25 µm/pulse (λL = 248 nm, 25 ns
pulse duration). The final fluence used was 0.42 J/cm2 while LIBS
on-line control was employed as described above with the mean thickness of
removed varnish being 9 µm. Fig. 2.3.12a shows an initial stage of processing
where a small area (left, towards the bottom) has been laser-cleaned, while
in Fig. 2.3.12b the two thirds of the painting (top) have been additionally
processed. At the very bottom of Fig. 2.3.12b and just below the small
initially laser-cleaned area, we can also see two small vertical zones that
have been cleaned by the chief restorer using a mixture of solvents, for
comparison. The complete exposure of the pigments in the latter case can be compared
with the delicate outer-layer, laser-assisted, removal. Fig. 2.3.12c presents
the interface between the original surface of the oxidized resin (top) and
the uncovered layer of the less oxidized resin (bottom). Fig. 2.3.12d shows
the final result after completion of the laser processing and the subsequent
removal of the exposed soft varnish using a mild solvent [51].
The
second example is a 19th century oil painting of size 49 cm x 77
cm that has been severely damaged by an inhomogeneous layer of lime mortar
over the surface of the aged varnish (Fig. 2.3.13a].
The
varnish was only a few microns thick (~7 µm) and evenly spread over the
original pigment. Any attempt to remove the lime particles and/or the aged
varnish using the traditional technique of a cotton swab wet by the
appropriate solvent would have been catastrophic, owing to the transformation
of the cotton/particles system into a form of sandpaper. The strategy of KrF
excimer laser-assisted divestment in this case was to ablate the outermost
aged resin layers resulting in the ejection of lime particles together with
the resin photofragments. In this particular case, a fluence value of 0.30
J/cm2 (slightly below the optimal fluence of 0.33 J/cm2)
was used in order to minimize the etching step. The thickness of removed
varnish was ~2 µm with the synchronous removal of the superficial lime
particles. The final result of the restoration [51] is presented in Fig.
2.3.13b. In
the case of evenly applied resins, we can detect evidence of the gradient of
polymerization with depth, in terms of crack formation. Figure 2.3.14 shows
part of a tempera icon painting covered with an unknown varnish aged over a
period of more than 100 years. A KrF excimer laser (fluence 0.54 J/cm2)
has been used to remove the resin at ten different zones. The zones are
indicated by increasing numbers that represent an increased removal of
varnish [36]. In zones 1-5 the increase of depth-step is 3.3 µm, i.e. zone 1
corresponds to 3.3 µm depth from the original surface, while zone 5
corresponds to 16.5 µm. In zones 6-10 the depth-step intervals are larger,
with a maximum depth of 82 µm at zone 10, where the varnish remaining is ~20
µm. The observed cracks in the resin are due to the stiffness of the aged
resin near the surface. The gradual decrease of cracks going from zone 1 to
10 is owing to the polymer phase gradient as a function of depth from the
surface. The more material that is removed the fresher the remaining resin is
and therefore the less cracks are revealed. In this case, the end user can
choose the endpoint according to the desired appearance and/or solubility of
the remaining varnish.
Multispectral
imaging techniques may provide a powerful tool for optimizing the divestment
process and achieving a homogeneous cleaning result. A specialized detection
system appropriate for spectral imaging in the region of 0.35-1.6 µm has been
developed [55] and integrated together with user-friendly software in a
single unit. It has been used to quantitatively monitor the cleaning of
paintings. A
characteristic test case example of the use of the multi-spectral imaging
camera is shown in Fig. 2.3.15. The 18th century oil painting on
canvas (a French duke, M. Doulgeridis’ collection) was subjected to KrF
excimer laser cleaning. The painting had suffered loss of paint at many
points of the surface. A restoration was attempted some decades ago without
first removing the old varnish and dirt. On the contrary, in order to cover
the poor workmanship, a stiff after drying varnish had been applied. In an
attempt to remove this new varnish, the restorers met great difficulties
since it had formed a rigid and resistant layer. For a proper removal of this
resin using traditional chemical means, the application of chemicals had to
be prolonged, while the results were of questionable quality. On the other
hand, the laser cleaned region presented several advantages: —
The surface appearance was excellent. —
In the undamaged areas where there was no previous restoration, a thin clean
layer of the old varnish was left. —
In the sections where the defective restoration had been attempted, the new
varnish was removed down to the original damaged layers, therefore, with good
retouching results. —
The use of hazardous strong solvents, both for the restorer and the artwork,
was avoided. —
The whole process was totally controllable with the use of the multi-spectral
camera and the fact that the laser parameters chosen allowed for a 0.5 µm
layer removal for each laser scan over the surface. Fig.
2.3.15a shows a visible image (14 x 10 cm²) in the forehead eye brow region
cleaned by laser, while Fig. 2.3.15b shows a 20% expansion of this visible
image. Fig. 2.3.15c shows the UV fluorescence image from the same region as
in Fig. 2.3.15b, but here the distinct marks of the originally damaged areas
are enhanced. For the laser cleaning, though, the main part of the process
concentrated on the top resin layers. Therefore, the UV reflectance image
(Fig. 2.3.15d) is very helpful in displaying the homogeneity of the remaining
resin after the partial removal of the top resin layers. The latter image
presented at a pseudo-colour display (Fig. 2.3.15e) helps to locate the
over-cleaned areas, which appear red in colour.
2.3.7
Er:YAG Laser in the Removal of Surface Layers Up
to now we referred to laser ablation in terms of UV short (several nanoseconds
or shorter) laser pulses that are able to photodissociate the covalent bonds
of a polymeric matrix and remove molecular fragments from the surface. This
dry process allows for a high depth resolution, i.e. in the order of sub
micrometers, owing to the strong absorption of UV radiation by polymeric
matrixes. An alternative cleaning methodology that also relies on the strong
absorption of laser radiation by the surface layers, primarily those
containing OH bonds, is using the Er:YAG laser. The Er:YAG laser emits a
train of pulses at 2.94 μm. The duration of each pulse is of the order
of 1-2 μs, while the whole train has duration of about 250 μs. The
2.94 μm wavelength is strongly absorbed by water, alcohols and other OH
containing molecules that are present or sprayed onto the surface prior laser
cleaning. The ablation mechanisms operating at this wavelength are
considerably different to each other: direct mostly electronic excitation
followed by dissociation for the former; explosive vaporization of the OH
containing molecules followed by parallel ejection of the contaminant
species, for the latter [14,56]. Another important difference between the two
processes is that the former is a clearly dry method, while the latter
usually requires the application of a liquid. Apart from the possible
advantages and disadvantages of the one methodology over the other, wetting
may prove to be a crucial consideration factor when choosing which method to
use in various conservation problems. Here
let us consider more closely the energy deposition and thermal diffusion
phases in Er:YAG laser cleaning. The pattern of energy deposition in material
depends on absorption and scattering. Both of these vary in each material
type with wavelength, and together they define the limit for minimizing the
volume of energy deposition and the succeeding thermal diffusion. In Er:YAG
laser process at 2.94 μm, the important absorber as mentioned above is
the OH bond in water or other molecules applied on the surface. Its absorption
and scattering properties determine the pattern of energy deposition [57].
Pure water has an absorption peak at 2.94 μrn and an absorption
coefficient of 104 cm-1, corresponding to an absorption
length of 1 μm [14]. At this wavelength, and for the absorption depths
in question, scattering is insignificant. Thus, the pattern of energy
deposition is determined by the distribution of molecules containing OH
bonds. The
removal of superficial layers and at the same time the minimization of
thermal damage to adjacent layers is limited by the material absorption
characteristics, the thermal conductivity, and the time for the heat
diffusion in the surface layers. The input energy needed for the phase change
from water to steam without appreciable thermal diffusion into adjacent
materials, is reached only when the pulse length is shorter than the thermal
relaxation time [14]. For the specific pulse duration of the Er:YAG laser the
thermal relaxation criterion is essentially satisfied. Accordingly, an
efficient ablation with the least surrounding thermal damage could be
expected to be approached by a pulsed Er:YAG laser. With an exposure duration
of 1-2 μs (each pulse in the train), the absorbed energy available for
heating will be essentially confined almost entirely in the volume where it
is absorbed (in the top few micrometers) long enough to heat and vaporize
just the desired portion of the material. This is in reality the case when
the OH containing liquid homogeneously penetrates into the surface layers. On
the contrary, if this does not happen (e.g. owing to absence of cracks, not
absorbing or hydrophobic surface, vaporization by previous pulses etc), it is
possible that the heated volume is much larger than a few microns. Testing
and evaluation of Er:YAG laser cleaning [58] revealed different results
depending on the type of painted surface. Working on laboratory models, some
drawbacks were noticed, probably due to the incomplete drying of some organic
materials. Generally, the results were much better when working on old
paintings. It was found that Er:YAG laser cleaning can be considered
effective, selective and safe when used at proper fluence levels, which are
fixed for each category of surface material. Another important parameter was
found to be the wetting of the surface with an OH containing liquid, usually
a dilution of alcohols or glycols. No significant chemical change in the
surface materials was observed during testing. Finally, it was realized that
a 'combined method' (laser exposure followed by solvent or scalpel cleaning)
facilitates the removal of hard and resistant overpaints and old varnishes
due to physical transformations induced by laser. The Er:YAG laser pulses
cause the ablation of a portion of the top-layer components and disaggregate
the remaining part, which can then be easily treated using mechanical tools
and mild solvents. Er:YAG laser systems for the cleaning of painted surfaces,
incorporating articulated arms for laser beam manipulation are commercially
available nowadays, and easy to use.
It
has been shown that using a short-pulse UV laser or a long pulse Er:YAG laser
it is possible to divest delicate substrates such as paintings or composite
structures from polymeric layers (e.g. network of polymerized resin or
polymeric paints, respectively) when bearing in mind the following
fundamental canons: —
The first layer over the delicate substrate – usually a layer of a few
micrometers – must be always left intact, acting as a light filter for the
transmitted photons. In addition, this layer has been found to be chemically
unaged. —
The optimal laser parameters such as laser energy flux must be found for
succeeding the minimal transmission of laser light towards the substrate. To
this end, the shortest available UV-laser wavelength is preferable both in
terms of short optical penetration depth and in terms of high resolution of
the ablation depth. —
On-line and/or in situ monitoring/control techniques must be used for
identifying the end point of the intervention. Such techniques include Laser
Induced Breakdown Spectroscopy (LIBS) and multispectral imaging.
[1]
Asmus, J.F., More light for art
conservation. IEEE Circuits and Devices Magazine, 1986, March: p. 6. [2]
Watkins K. G., Larson J. H.,
Emmony D. C., Steen W. M., Laser cleaning in art restoration: a review. in
Proceedings of the NATO Advanced Study Institute on Laser Processing: Surface
Treatment and Film Deposition. 1995: Kluwer Academic Publishers. p. 907 [3]
Cooper, M., Cleaning in
Conservation: an Introduction. 1998, Oxford: Butterworth Heinemann. [4]
Maravelaki P.V., Zafiropulos V.,
Kylikoglou V., Kalaitzaki M., Fotakis C., Laser Induced Breakdown
Spectroscopy as a Diagnostic Technique for the Laser Cleaning of Marble.
Spectrochimica Acta B, 1997. 52: 41. [5]
Maravelaki-Kalaitzaki, P., V.
Zafiropulos, and C. Fotakis, excimer laser cleaning of encrustation on
Pentelic Marble: procedure and evaluation of the effects. Appl. Surface Sci.,
1999. 148: 92. [6]
Gobernado-Mitre I., Prieto A.C.,
Zafiropulos V., Spetsidou Y., Fotakis C., On-line monitoring of laser
cleaning of limestone by laser induced breakdown spectroscopy. Appl.
Spectrosc., 1997. 51: 1125. [7]
Klein S., Stratoudaki T., Marakis
Y., Zafiropulos V., Dickmann K., Comparative study of different wavelengths
from IR to UV applied to clean sandstone. Appl. Surface Sci., 2000. 157: 1. [8]
Zergioti I., Petrakis A.,
Zafiropulos V., Fotakis C., Fostiridou A., and Doulgeridis M., Laser
Applications in Painting Conservation. in Restauratorenblätter, Sonderband –
Lacona I, Ed. by E. König and W. Kautek, Vienna: Verlag Mayer & Comp.
p.57 [9]
Fotakis C., Zafiropulos V., Anglos
D., Georgiou S., Maravelaki P. V., Fostiridou A., Doulgeridis M., Lasers in
art conservation. The Interface between Science and Conservation, 1997 (S.
Bradley Ed., The Trustees of the British Museum): p. 83. [10]
Georgiou S., Zafiropulos V., Anglos D., Balas C.,
Tornari V., Fotakis C., Excimer laser restoration of painted artworks:
Procedures, Mechanisms and Effects. Appl. Surface Sci., 1998. 127-129: 738. [11]
Zafiropulos, V. and C. Fotakis, Laser in
Conservation: an Introduction, Chapter 6: Lasers in the Conservation of
painted Artworks. 1998, Oxford: M. Cooper Ed., Butterworth Heineman. P. 79. [12]
Zafiropulos V., Galyfianali A., Boyatzis S.,
Fostiridou A., Ioakimoglou E., UV-laser Ablation of Polymerized Resin Layers
and Possible Oxidation Process in Oil-based painting media. Optics and Lasers
in Biomedicine and Culture - Series of the International Society on Optics
within Life Sciences, ed. G.v. Bally. Vol. V. 2000, Berlin: Springer-Verlag.
p. 115. [13]
Zafiropulos V., Manousaki A., Kaminari A., and Boyatzis
S., Laser Ablation of aged resin layers: A means of uncovering the scalar
degree of aging. in ROMOPTO 2000: Sixth Conference on Optics, V. I. Vlad Ed.,
SPIE Proceedings 4430, 2001. Washington. p. 181 [14]
De Cruz, A., M.L. Wolbarsht, and S.A. Hauger, Laser
removal of contaminants from painted surfaces. J. Cult. Heritage, 2000. 1: p.
s173. [15]
König, E. and Kautek W. Eds., Restauratorenblätter,
Sonderband – Lacona I, Proceedings of the First International Conference
LACONA I - Lasers in the Conservation of Artworks. 1995, 4-6 October.
Heraklion, Crete, Greece: Verlag Mayer & Comp., Vienna. [16]
Salimbeni, R. and G. Bonsanti. Eds., Proceedings of
the International Conference LACONA III - Lasers in the Conservation of
Artworks. 1999, April 26-29. Florence: J. Cult. Heritage. [17]
Verges-Belmin, V. Ed., Lasers in the Conservation of
Artworks IV. in Proceedings of the International Conference LACONA IV. 2003:
J. Cult. Heritage. [18]
Dickmann, K., Fotakis C., and Asmus J.F. Eds., Lasers
in the Conservation of Artworks - LACONA V Proceedings. in Springer
Proceedings in Physics. September 15-18, 2003. Osnabrueck, Germany:
Springer-Verlag. [19]
Klein S., Stratoudaki T., Zafiropulos V., Hildenhagen
J., Dickmann K., Lehmkuhl T., Laser-induced breakdown spectroscopy for
on-line control of laser cleaning of sandstone and stained glass. Appl. Phys.
A, 1999. 69: p. 441. [20]
Zafiropulos, V., Chapter 8: Laser ablation in
cleaning of artworks. In Laser Cleaning, ed. B. Luk'yanchuk. 2002, Singapore,
New Jersey, London, Hong Kong: World Scientific. p. 343. [21]
Scholten J.H., Teule J.M., Zafiropulos V., Heeren
R.M.A., Controlled Laser cleaning of painted artworks using accurate beam
manipulation and on-line LIBS-detection. J. Cult. Heritage, 2000. 1: p. s215 [22]
Fogarassy, E. and Lazare S.. Laser ablation of
electronic materials. - Basic mechanisms and applications. in Proc. European
Materials Research Society Monographs. 1992. Amsterdam: North-Holland
Elsevier. [23]
Dyer, P.E., Photochemical Processing of Electronic
Materials. 1992, London: Academic. [24]
Miller, J.C., Laser Ablation - Principles and
Applications. Springer Ser. in Materials Science 28. 1994, Berlin: Springer-Verlag. [25] Bäuerle, D., Laser Processing and
Chemistry. 2000, Berlin, Heidelberg: Springer. [26]
Luk'yanchuk, B., Laser Cleaning. 2002, Singapore,
New Jersey, London, Hong Kong: World Scientific. [27]
Srinivasan, R. and Braren B., Ultraviolet Laser
ablation of organic polymers. Chem. Rev., 1989. 89: 1303. [28]
Srinivasan, R., Ablation of polyimide (kapton) films
by pulsed (ns) ultraviolet and infrared (9.17 µm) lasers; a comparative
study. Appl. Phys. A, 1993. 56: 417. [29]
Srinivasan, R., Interaction of laser radiation with
organic polymers. In Laser Ablation, 1994.
Springer Ser. Mat. Sci. 28 (J.C. Miller Ed., Springer): p. 107. [30]
Luk'yanchuk B., Bityurin N., Anisimov S., and
Bäuerle D., The role of excited species in UV-laser materials ablation; Part
I: Photophysical ablation of organic polymers. Appl. Phys. A, 1993. 57: 367. [31]
Luk'yanchuk B., Bityurin N., Anisimov S., and
Bäuerle D., The role of excited species in UV-laser materials ablation; Part
II: The stability of the ablation front. Appl. Phys. A, 1993. 57: 449. [32]
Luk'yanchuk B., Bityurin N., Anisimov S., and
Bäuerle D., Photophysical ablation of organic polymers. Excimer Lasers, 1994 (L.D. Laude Ed., NATO ASI Ser. E). [33]
Luk'yanchuk B., Bityurin N., Anisimov S., Arnold N.,
and Bäuerle D., The role of excited species in ultraviolet-laser materials
ablation; III. Non-stationary ablation of organic polymers. Appl. Phys. A,
1996. 62: 397. [34]
Athanassiou A., Lassithiotaki M., Anglos D.,
Georgiou S., Fotakis C., A comparative study of the photochemical
modifications effected in the UV laser ablation of doped polymer substrates.
Appl. Surface Sci., 2000. 154-155: 89. [35]
Theodorakopoulos, C., Ph.D. Thesis, 2005, London,
U.K.: Royal College of Arts. [36]
Theodorakopoulos, C. and Zafiropulos V., Uncovering
of scalar oxidation within a naturally aged varnish layer. J. Cult. Heritage,
2003. 4: 216. [37]
Zafiropulos, V., Petrakis J., and Fotakis C.,
Photoablation of polyurethane films using UV laser pulses. Optical and
Quantum Electronics, 1995. 27: 1359. [38]
Pettit, G.H. and Sauerbrey R., Pulsed Ultraviolet
laser ablation. Appl. Phys. A, 1993. 56: 51. [39]
Tornari V., Fantidou D., Zafiropulos V., Vainos N.
A., Fotakis C., Photomechanical effects of laser cleaning: A long-term
nondestructive holographic interferometric investigation on painted artworks.
in Third International Conference on Vibration Measurements by laser
techniques: Advances and Applications, SPIE. 1998. Washington: SPIE - The
International Society for Optical Engineering, SPIE Proceedings 3411, p. 420 [40]
Pettit G. H., Ediger M. N., Hahn D. W., Brinson B.
E., and Sauerbrey R., Transmission of polyimide during pulsed ultraviolet
laser irradiation. Appl. Phys. A, 1994. 58: 573. [41]
De la Rie, E.R., The Influence of Varnishes on the
Appearance of Paintings. Studies in Conservation, 1987. 32: 1. [42]
Zafiropulos V., Stratoudaki T., Manousaki A.,
Melesanaki K., Orial G., Discoloration of pigments induced by laser
irradiation. Surface Engineering, 2001. 17: 249. [43]
Luk'yanchuk, B.S. and Zafiropulos V.. The model for
discoloration effect in pigments at cleaning of artworks by laser ablation.
in Second International Symposium on Laser Precision Microfabrication, SPIE.
2002. Washington: The International Society for Optical Engineering. 4426, p.
326 [44]
Castillejo M., Martin M., Oujja M., Silva D., Torres
R., Manousaki A., Zafiropulos V., van den Brink O. F., Heeren R. M. A., Teule
R., Silva A., and Gouveia H., Analytical Study of the Chemical and Physical
changes induced by KrF Laser cleaning of Tempera Paints. Analytical
Chemistry, 2002. 74: 4662. [45]
Mills, J.S. and R. White, The Organic Chemistry of
Museum Objects and references therein. 1987, London: Butterworths. [46]
De la Rie, E.R., Photochemical and thermal
degradation of films of dammar resin. Studies in Conservation, 1988. 33: 53. [47]
De la Rie, E.R., Old Master Paintings: A Study of
the Varnish Problem. Analytical Chemistry, 1989. 61: 1228. [48]
C. Theodorakopoulos, Zafiropulos V., Fotakis C.,
Boon J.J., Horst J. v.d., Dickmann K., Knapp D., A study on the oxidative
gradient of aged traditional triterpenoid resins using "optimum"
photoablation parameters. Lasers in the Conservation of Artworks -. in LACONA
V Proceedings. 2005: Springer-Verlag, 100, p. 255 [49]
Stout, G.L., The Care of Pictures. 1975, New York:
Dover Publications, INC. [50]
Sutherland, K., Solvent extractable components of
oil paint films. Ph.D. thesis. 2001: University of Amsterdam. [51]
Doulgeridis, M., private communication. 1997. [52]
Hildenhagen, J. and Dickmann K., Low-cost sensor
system for online monitoring during laser cleaning. J. Cult. Heritage, 2003.
4: 343s. [53]
Lentjes, M., Klomp D., and Dickmann K.. Sensor
concept for controlled laser cleaning via photodiode. Lasers in the
Conservation of Artworks -. in LACONA V Proceedings. 2005: Springer
Proceedings in Physics, p. 427 [54]
Chryssoulakis, Y., S. Sotiropoulou, and V.
Zafiropulos, unpublished results. 1996. [55]
Balas, C., An imaging colorimeter for the non
contact color mapping. IEEE-Transactions on Biomedical Engineering, 1997. 44:
468. [56]
DeCruz, A., Hauger, S., Wolbarsht, M. L., The role
of lasers in fine arts conservation and restoration. Opt. Photon News, 1999.
10: p. 36. [57]
Wolbarsht, M.L., Laser Surgery: CO2 or HF. IEEE J. Quant. Electron. QE, 1984. 20: 1427 [58]
Bracco, P., Lanterna, G., Matteini, M., Nakahara,
K., Sartiani, O., de Cruz, A., Wolbarsht, M. L., Adamkiewicz, E., Colombini,
M. P., Er:YAG laser: an innovative tool for controlled cleaning of old
paintings: testing and evaluation. J. Cult. Heritage, 2003. 4: 202s.
Vassilis Zafiropulos Department of Human Nutrition and Dietetics Technological Educational Institute (TEI) of Ioannou Kondylaki 46 GR - 72300 Sitia Greece and affiliated with the Institute of Electronic Structure & Laser Heraklion Crete Greece |
,
(1)
,
(2)
