1 Introduction

Laser techniques have been widely and successfully applied in the cleaning of works of art. Laser cleaning is based on knowledge of the thermal, photochemical and photomechanical mechanisms related to different laser ablation parameters, and in particular the possible instantaneous or long-term damage to the original substrate [1].

Since works of art are complex multisystem and multilayered materials, the interaction of their organic and inorganic constituents with the laser beam and the subsequent removal mechanism are rather complex.

Several types of laser are available for cleaning and how they act and how effective/safe they depend on the wavelength of the radiation emitted, pulse duration, fluences and other parameters [2–14].

The Er:YAG laser was introduced as an alternative to traditional methods for cleaning easel and mural paintings [4, 15–18], paper and parchment [19], and different crust types from stones [20]. The compounds commonly present on the surfaces of art objects have absorption bands in mid-infrared and near ultraviolet, which lead to vibrational and electronic excitations, respectively. The Er:YAG (2,940 nm) wavelength is absorbed by a large variety of compounds present in different materials.

The absorption of IR increases the temperature on the irradiated area. Through water vaporization, gas expansion and micro-distillation, the unwanted material is then removed from the surface.

Studies based on noninvasive µ-profilometry analysis, executed before, during and after laser cleaning of easel painting, have enabled the depth of laser ablation to be quantified as 3–5 µm [21]. Effective and gradual cleaning can also be achieved by applying appropriate wetting agents to the surface, which absorb the radiation and thus reduce its penetration.

Most of the criticism of the use of Er:YAG lasers in restoration relates to the heating effects produced on the surface material, which must be preserved. Although the absorption of the laser is confined to a surface depth of a few microns, the side effects due to thermal overheating need to be investigated.

There are a few studies investigating the temperature increment induced by Er:YAG laser irradiation, in medical and industrial applications. The physical parameters due to temperature changes were measured with thermocouple assemblies for a simulated dental pulp irradiation [22] and during the welding of double-layer pieces of steel [23].

For cultural heritage studies, the investigation of the thermal effect of laser cleaning has been done only through laboratory tests on model systems mainly irradiating metal surfaces with a Nd: YAG (1,064 nm) and evaluating the one-dimensional temperature rise by means of numerical estimations [9, 24].

Our aim was to evaluate the temperature increase induced by a Er:YAG laser at different fluences on samples of works of art containing various commonly encountered materials.

We built a device to make the irradiation process reproducible in terms of the distance of the end of the optical fiber from the surface and the angle of incidence of the laser beam. Type T (copper-constantan) thermocouples were purpose built for each temperature measurement and were sufficiently small in order to minimize any perturbation in temperature and to be suitable for the case of really small surface tested real art work samples. Another peculiarity of this device is the possibility of performing the tests in an inert atmosphere.

Moreover, during Er:YAG laser cleaning of the organic material, intense bright flashes of light have been noticed. In this study, we have investigated this phenomenon, performing flash intensity measurements with a photomultiplier tube directly coupled to our device and collecting spectra by means of a USB2000 spectroradiometer.

For the evaluation of the thermal effect induced by the Er:YAG laser in and around the irradiated area, standard sample surfaces were prepared containing graphite black carbon, which simulates an organic material with a high IR absorption.

2 Apparatus

We used a free-running Er:YAG laser at 2.94μm manufactured by MonaLaser (LLC., Orlando, Fl, USA), with a pulse duration of 250 µs. In order to measure the thermal effect induced by the Er:YAG laser in and around the irradiated spot, we built our own device (Fig. 1). It consists of a plastic case (nylon) (A), covered with a transparent lid (D), containing a sample holder (B) with an x–y movement system (C). In the center of the cover, there is a gastight quartz window that houses the photomultiplier tube in a position as close as possible to the sample surface (G). Near the windows, there is an insert to attach the end of the optical fiber from the laser (E). The insert maintains the end of the fiber at a fixed distance of 1 cm from the surface of the sample with an inclination of 45° to the bearing plane of the sample. A spot of 1.3 × 0.8 mm was obtained.

Fig. 1
figure 1

Design (a) and photo (b) of the device used in the experiments. A—nylon box; B—sampler holder; C—micrometer screw for the x–y movement; D—transparent lid; E—optical fiber; F—sample; G—photomultiplier; H—connections for vacuum or gas or liquid; I—thermometer; and L—hole for the thermocouple junction

Full size image

The whole device is gastight; samples were irradiated in a modified atmosphere introducing (through the pipe connections H), for example, an inert gas such as nitrogen or carbon dioxide.

Two micrometer screws (C) control the x–y movement of a plastic support for the sample inside the cell. The possibility of moving the support not only enables a better position for the laser spot over the sample but also simulates the movement of the laser tip during a normal cleaning procedure.

The temperature of the laser-treated surfaces was measured using a Type T (copper-constantan) thermocouple (Figs. 1b, L 2), periodically positioned below the irradiated surface.

Fig. 2
figure 2

Holder for the reference material: cardboard (a) with a hole in the middle (b) in which the thermocouple was positioned and the PVA glue/graphite dough was placed over the junction to fill the hole (c)

Full size image

This type of temperature sensor is easy to make, is suitable for temperatures up to 300 °C and has a very quick response. To reduce the size of the temperature probe, the thermocouple was made out of copper and constantan wires with a diameter of 0.05 mm sufficiently small to minimize the heat dispersion through the wire. Moreover, the small dimension is also suitable in terms of the small surface tested (about 1 mm2) as in the case of real art work samples.

The cold junction of the thermocouple was soldered onto a small adhesive copper strip, which was wrapped around a thermometer bulb during the measurements. The difference in temperature between the irradiated surface and the inside of the plastic container was thus recorded.

The amount of radiant energy emitted as a flash immediately after the laser pulse was measured by an 11-stage photomultiplier tube (Philips 50AVP) positioned on the quartz window over of the irradiation point. The photomultiplier was supplied with a total voltage of −450 V, and a 10-kΩ resistance was used to convert the current into a voltage signal. The signal produced by the photomultiplier was collected by an oscilloscope, and the maximum intensity was measured in volts. Figure 3 shows the typical shape collected during a laser pulse.

Fig. 3
figure 3

The profile of the photomultiplier signal when a flash was produced

Full size image

3 Experimental setup and material

Graphite black carbon was used as reference material for the absorption of the laser radiation. Graphite was selected because it absorbs IR radiation well and could be considered as a good approximation of a black body. Graphite is also a good conductor of heat; thus, a layer of about 0.5 mm of this material can be considered as being homogeneous in temperature during irradiation.

Preliminary tests showed that the carbon dust is a good material for absorption of the laser IR radiation, showing a very similar intense flash emission to the emission observed, for example, during the irradiation of the organic matter. Unfortunately, preliminary experiments showed that consecutive laser shots in the same position above the carbon powder dispersed the irradiated material, with a significant change in the surface characteristics. In order to prevent a significant variation in the layer thickness above the hot junction of the thermocouple, the graphite was mixed with PVA glue to obtain a dough (Fig. 2) with a 30 % content of black carbon.

A holder for the reference material was made using two small rectangles (1 cm × 3 cm) of cardboard (Fig. 2). The two rectangles were fixed on top of each other with double-sided adhesive (Fig. 2a). The upper rectangle had a hole in the middle of about 0.5 mm of diameter; the thermocouple was positioned between the two rectangles with the hot junction in the middle of the hole (Fig. 2b).

The PVA glue/graphite layer covered the junction (Fig. 2c), and the PVA glue was allowed to dry for one night. The hot junction of the thermocouple was immersed in the material less than 0.5 mm from the surface. The holder was fixed onto the surface of the support inside the cell using a double-sided tape. The cold junction of the thermocouple was placed in contact with a thermometer bulb placed inside the cell. The temperature of the thermometer bulb was taken as room temperature, which was checked before each set of measurements and was 22 ± 0.5 °C.

Since the temperature observed depends on the point of irradiation on the thermocouple junction below the surface, in order to position the spot of the laser on the graphite surface as near as possible to the hot junction, the lower pulse repetition rates was set and, with the aid of the x–y movement system, the maximum of temperature increase was found. In these conditions, the minimal perturbation to the surface was induced and when the hot junction reached room temperature again, the laser was set to the operative pulse frequency and irradiation was started.

For each measurement, the maximum temperature variation and the mean signal from the photomultiplier were registered.

Before each irradiation test, the optical fiber was removed from the cell and the mean power emitted at the operative pulse frequency was measured by a laser power meter (PowerMax500A—Molectron, Portland, OR, USA).

A USB2000 spectroradiometer (Ocean Optics Inc. Dundein, FL) coupled with a 1,000-μm-diameter fiber optic cable has been used for the acquisition of the bright flash of light occurring during laser ablation.

4 Results and discussion

4.1 Results on the reference material

Several replicate PVA glue/graphite samples were subjected to laser irradiation in order to measure the temperature increase and the light emitted, varying the power and the pulse repetition rate of the pulse. The average power from 15 to 300 mW (corresponding to a fluence ranging from 0.1 to 2.6 J/cm2) and three different pulse repetition rates (low 3.75, medium 7.5, high 15 pulse/s) have been tested. For each laser parameter, the highest temperature reached by the thermocouple after consecutive pulses on the same position was recorded.

Figure 4 shows the plot of temperature increases versus the mean power of the laser beam when the maximum temperature of the surface was reached. The symbols in the plot refer to different conditions of reference sample irradiation, obtained by varying the pulse repetition rate and the mean power of the laser beam. The linearity of the data demonstrates that the temperature increase was proportional to the average irradiation power in spite of the use of different combinations of pulse repetition rate and laser intensity. Moreover, the choice of the PVA glue/graphite reference material was good for understanding the laser energy absorption, because a good reproducibility of the results was reached.

Fig. 4
figure 4

Plot of the temperature increase in PVA glue/graphite surface versus the power of the laser beam: different symbols refer to various reference samples used in each experimental set. + and × symbols refer to measurements in presence of isopropanol and white spirit on the surface, respectively

Full size image

When the laser irradiation begins, the temperature does not immediately reach its maximum value; a stable temperature on the surface was observed only after almost 20 s of consecutive pulses on the same position of the beam spot.

As the linearity between the increase in temperature and laser power is clear, it is also possible to evaluate the effect of the different solvents used to wet the surface. Temperature measurements were taken using PVA glue/graphite samples wetted with isopropanol or a white spirit mixture. With the solvents, the surface temperature was stabilized by evaporation and the increase in temperature was drastically reduced until the surface dried. If the surface was continuously wetted with isopropanol or white spirit using a syringe during the irradiation, the mean decrease in surface temperature was more than 40 % (dotted lines in Fig. 4).

4.2 Results on the samples from works of art

To investigate the thermal effect induced by the laser, four works of art were selected on the basis of the composition of the unwanted material to be removed, and on the typology of the substrate: a panel painting, a bronze coin, a Roman urn and lichens.

4.2.1 Panel painting

The cleaning of painted surface, wall and easel painting is one of the most critical procedures in the conservation of works of art. Innovative cleaning procedures, and among them the laser ablation, have been developed in order to safely remove unwanted material, without damaging any external thin original varnishes or old over painting, which the conservator wishes to retain. A masterpiece by Lluís Borrassà, a Catalan painter from the Gothic period, was in the process of being conserved in the laboratory at the North Caroline Museum. The easel painting presented a thick and dark patina and preliminary laser test demonstrated the efficacy of the Er:YAG laser in removing the patina. The application of laser on this easel painting is an important case study because GC/MS analysis highlighted the presence of egg and siccative oil in the dark patina. These two components have been widely used as binders or varnishes, so it was therefore important to study the interaction of the laser radiation with these materials, and the effect on the original paint layer. In order to evaluate the temperature increase, a 1-mm fragment was sampled using a scalpel from an area close to the frame, and glued onto the thermocouple and inserted into the nylon box. In order to perform a safe removal of patina and old varnishes, precedent studies have highlighted the importance of a progressive cleaning by applying subsequent laser applications at low fluences, rather than of a single laser pulse at higher one [3, 5, 17, 21].

The increase in surface temperature (Table 1) was limited when isopropanol was used as a wetting agent and the ablation was very efficient. The original paint layer was not affected by the laser ablation, even when the drying process was complete. Clearly, if the ablation is performed without solvents, the increase in surface temperature is higher (Table 1), and though the ablation is effective, it is not considered suitable for the constituent of the original paint layer.

4.2.2 Bronze coin

The bronze coin came from the collection of the American Academy in Rome and dates back to the eighth-century C.E. Bronze works of art corrode leading to the formation of a patina consisting of inorganic compounds (such as copper carbonate, malachite salts and silicates), which encrust the surface.

To assess the effect of laser radiation and the subsequent temperature increase, the thermocouples were inserted into a crack in the metal, as shown in Fig. 5. The crack made it possible to house the thermocouple below the surface.

Fig. 5
figure 5

Example of the Copper–Constantan positioned on the coin surface. In this case, the wire is inserted into a crack in the metal

Full size image

Once positioned in the hot junction, the crack was filled with a dough consisting of copper carbonate hydrate, one of the common components of patinas of copper objects [25].

A few tests on the coin were performed with a laser power ranging from 150 to 300 mW, in conjunction with an alcohol/water (1:1) mixture as wetting agents. The laser power, selected on the basis of the preliminary tests for the temperature measurement, was of 150 mW. The temperature increase at the surface, just below the hot junction, was 45°C, but the presence of the solvent limits the increase to 15°. Laser ablation dehydrates the patina, which then breaks up, and this facilitates the subsequent mechanical removal. In the case of bronze, heating the surface does not melt the metal; the temperature measured was well below the melting point. Figure 6 shows the results of the laser cleaning of the bronze coin surface in the conditions reported above; the copper oxide patina was safely and completely removed.

Fig. 6
figure 6

An area of the bronze coin before and after laser cleaning

Full size image

4.2.3 Cinerary urn

The Roman Cinerary urn was under restoration at the conservation laboratory of Duke University (collection of the St. Louis Art Museum). It is encrusted with an intractable calcite material, which covered the decorative area of the marble surface. In addition, in a number of areas on the surface and in the cavity of the marble, round or oval black fungi were found, which were responsible for a bright white flash of black body light occurring during irradiation. Before beginning the tests, the laser was used to sample the organic material in the superficial encrustation. A microscope glass coverslip (15 × 15 mm) was placed on the surface of the urn, and after laser impact, the removed material condensed on the glass [3] was analyzed using GC/MS [26].

The chromatographic and spectroscopic analyses highlighted the presence of a proteinaceous material (most probably egg) and of a nondrying oil such as palm or olive oil. Finally, GC/MS analysis ruled out the presence of beeswax or saccharide material, while revealing the presence of oxalic acid, which was identified by Raman analysis as calcium oxalate.

Over time, the organic material applied to the surface of the urn had decomposed. In fact, protein and oils undergo a photo-oxidation and biological process partially turning them into oxalic acid, which in the presence of marble, CaCO3 yields calcium oxalate. Moreover, the presence of microorganisms such as blue algae, fungi and lichens, which provide oxalic acid as metabolic product, could have contributed to the formation of such thick patina [27, 28].

Again, a fragment containing both the marble substrate and the patina was sampled using a scalpel, glued onto the thermocouple and inserted into the nylon box for the tests.

The hard encrustation that obscured the delicate carving of the Roman urn was removed only using a laser power of 200 mW. The increase in surface temperature during the laser cleaning in presence of alcohol:water 1:1 as wetting agent was very small (Table 1). The urn was cleaned, and the porosity was not affected since a compact thin layer of calcium oxalate was left on the marble surface. The result is shown in Fig. 7.

Fig. 7
figure 7

The Roman Urn before (a), and after (b), laser ablation

Full size image

4.2.4 Removal of lichens from limestone

The lichen sample is from the Culberson Collection of the Biology Department of Duke University. A limestone fragment [20] was covered with a black thick crust on which a Diploschistes scruposus lichen was identified. Before the ablation tests, the superficial layer was incised with a scalpel and the thermocouple was inserted into the crack, in contact with the limestone surface.

A bright flash of light occurred under ablation without solvents as already observed during laser application in similar substrates. Preliminary measurements using spectroradiometer were performed in order to investigate the nature of this flash. The distribution of the wavelength emitted during the flash is compatible with a black body emission. The color temperature associated to this emission was measured between 2.700 and 3.000 K. Obviously, this temperature cannot be attributed to the temperature of the irradiated surface, in particular with a long pulse duration laser such as the ER:YAG [1, 2].

In fact, the thermocouple probe highlighted that the temperature at the substrate level was low during irradiation (Table 1) and the emitted flash can be explained with the presence on the atmosphere above the surface of a dispersion of carbon particles produced by the combustion of organic matter during the laser treatment. These carbon particles, absorbing the laser energy, can easy reach a sufficiently high temperature for a black body emission.

Table 1 Summary of laser ablation conditions used on the four works of art
Full size table

In the case of lichen, the power increase did not seem to produce a proportional variation in temperature, but this is probably due to the heterogeneity of the three portions of the lichen surface chosen for the experiments. A sufficient laser cleaning efficacy for the removal of the thick lichen encrustation has been reached using a power beam ranging between 100 and 185 mW.

5 Conclusions

We designed a system to measure the temperature increase at the surface of impact of a laser pulse. These experiments were carried out with a free-running Er:YAG laser at 2.94 µm (MonaLaser) and measured using a copper (copper-constantan) thermocouple to measure the increase in temperature at the point of impact when a surface is exposed to laser radiation.

The method was tested on a standard material of graphite particles mixed in PVA glue. Further measurements were taken with original artifacts from a marble Roman urn, antique copper coins, a medieval tempera painting and lichen on stone. The temperature increase in the irradiated surface ranged from 45 to 140 °C depending on the substrates, but the use of a solvent as wetting agent during the application limits the increase to 15/18 °C.

In all instances, the temperature increase produced by irradiation in the most common working setup of the pulsed free-running Er:YAG laser at 2.94 µm was at a level that would not cause any thermal damage to the organic or inorganic compounds in the substrate of the artifact whose encrustation was removed. During the laser cleaning of organic material without any auxiliary wetting agent, a bright flash has been registered. The phenomena can be explained as an emission, as a black body radiator, of the carbon particles present in the atmosphere above the surface treated. The carbon dust dissipates the energy absorbed during consecutive laser pulses application, without involving the cleaned surface. In fact, the experiments revealed a temperature at the substrate level significantly low during irradiation.

This research has demonstrated that, providing suitable optimization, the pulsed free-running Er:YAG laser is a unique tool for restoring artistic objects; it acts on various substrates without damaging them and removing intractable materials.

The apparatus and the measuring procedure adopted in this paper respond to the need of evaluating the photo thermal effect of the Er:YAG laser highlighted in the literature. Moreover, the experimental set-up developed, based on the use of PVA glue/graphite reference samples, can be used also in future experiments to assess the impact of other types of lasers that have long been used in the cleaning of artifacts adding new fundamental data to the laser performances for cleaning purposes.