|
Chapter 4.4 Fluorescence Lifetime Imaging and Spectroscopy Daniela Comelli1, Gianluca
Valentini1, Rinaldo Cubeddu1,
Lucia Toniolo2 1CNR-INFM and CNR-IFN, Politecnico di Milano -
Dipartimento di Fisica, Milan,
Italy 2Istituto per la Conservazione e la Valorizzazione
dei Beni Culturali – CNR, Milan,
Italy
Contents
4.4.4.1 Measurement
on Frescoes
4.4.4.2 Measurements
on Marble Sculptures Generally speaking, to correctly plan a
conservation activity, it is necessary to perform a careful diagnosis
about the state of conservation of the work of art. This requires the
knowledge of all the materials present in the artefact, i.e. original
materials, new formation materials deriving from deterioration
processes and restoration materials that have been applied on the
surfaces throughout the centuries. Such a deep investigation can be
presently achieved through a wide micro-destructive sampling, followed
by extensive laboratory analyses. Although effective, this procedure
has the drawback of relying on materials taken from the artefact. This
is particularly undesirable for artworks of great artistic importance
or for very large surfaces, as mural paintings, which require a great
number of samples to get a significant knowledge of the constituent
materials. Even though a limited micro-sampling is
generally unavoidable, non-destructive analyses that can be performed
in situ gives an invaluable support to the diagnostic process. For this
reasons, in recent years, a great effort has been devoted to transform
laboratory techniques in portable equipment for in
situ measurements. Most of the investigation techniques that
can be applied in situ share the use of electromagnetic radiation for
non-contact analysis of the artefacts. The study of the intensity of
the radiation as a function of the wavelength is commonly indicated
with the very general word of “spectroscopy”,
which, for the very specific case of the visible radiation, corresponds
to the analysis of the colour of the surface. Other parts of the
electromagnetic spectrum, beyond the visible one, can be profitably
applied for the investigation of works of art, leading to techniques as
Raman and infrared spectroscopy, which use radiation at frequency below
that of visible light (infrared light), and X-ray spectroscopy, which
employ the same high frequency radiation used for medical diagnosis.
Most of the previous techniques provide information at molecular and
atomic level, yet their application is limited to single points in the
artwork. Imaging methods are certainly more effective
for a precise and quick investigation of complex and large artefacts.
In this case, the most popular in situ techniques operate in the
visible part of the electromagnetic spectrum and are: diffuse
reflectance imaging, which study the colour of the light backscattered
from the analysed surface, and fluorescence imaging, which consider the
light remitted by the surface after the absorption of ultraviolet
radiation. Diffuse reflectance is mainly intended for the investigation
of pigments, while fluorescence, which is the main topic of this
chapter, is one of the most interesting techniques to study the binders
made of organic compounds. In fact, it is well known that many organic
materials show a fluorescence emission after being excited by an
ultraviolet radiation. Even if optical imaging techniques, including
reflectance and spectroscopy, are easily applied and allow one to
investigate a wide area in a short time, the extraction of the
diagnostic information from reflectance or fluorescence images is not
straightforward. For what concerns fluorescence, the emission is
typically due to a mixture of materials and fluorescence spectroscopy
cannot provide alone an exhaustive identification of a sample. Yet,
other features of fluorescence beyond its spectrum supply information
that can assist the discrimination and identification of organic
compounds present in an artefact. To this purpose, one of the most
interesting parameters is the fluorescence lifetime, which is in the
range of nanoseconds, and represents the average time the fluorescence
lasts after excitation with a very short light pulse. Presently, UV fluorescence examination during
conservation works is usually performed in-situ
with a low pressure Mercury lamp having a dark screen that absorbs the
visible light (Wood lamp). The emitted UV radiation is relatively
broadband. In fact, the wavelength ranges between 340 and 380 nm. A
simple visual inspection of the emission allows one to reveal the
presence of fluorescent compounds on the surfaces and provides only a
rough discrimination between them, based on the different colour of the
emission. Thus, it cannot be considered a measurement procedure and its
effectiveness relies completely on the skill and experience of
restorers. The photographic recording of the fluorescence emission
carried out with a film camera or, more recently, with a digital camera
is a step forward with respect to visual inspection. Yet, no great
improvement is achieved for what concerns the identification and
quantification of organic compounds. More objective parameters of
fluorescence, like the intensity, the spectrum or the lifetime and a
well-designed laboratory analysis on micro-samples are really required
to this purpose. As a general comment, it is worth noting that no one
technique can face alone the complex problems encountered in cultural
heritage analysis, while the synergic combination of
in situ measurements and laboratory techniques allows one to
gather the information required for a well-designed restoration. This chapter provides two examples of
application of integrated in situ and laboratory measurements for the
assessment of the conservation status of works of art.
Starting from these considerations, in this
paper we propose the application of a Fluorescence Lifetime Imaging
(FLIM) system [1] for the investigation of surfaces of artistic
interest. FLIM is based on the measurement of the temporal properties
of the fluorescent emission in every point of a sample, thus allowing
the reconstruction of the lifetime map of the analysed region. FLIM
enhances the capabilities of fluorescence imaging. In fact, lifetime
provides effective discrimination among different fluorescent
substances, also called fluorophores.
A FLIM apparatus capable of ns (10-9
s) temporal resolution has been recently used to analyse Renaissance
frescoes and marble sculptures. The apparatus has been combined with a
portable spectrometer, in order to record fluorescence spectra (i.e.
the colour of the emitted light after UV excitation) in points of
interest. In fact, spectra provide complementary information to
lifetime measurements, thus enhancing the discrimination capability of
the FLIM technique. The information obtained with the FLIM system
is combined with measurements performed in laboratory on micro-samples
taken from the painting. Optical microscopy, scanning electron
microscopy equipped with X-ray spectrometry and Fourier Transform
Infrared Spectroscopy are applied. Such analytical measurements give a
synergic effect when performed together with FLIM: in fact, the precise
chemical identification of the materials provided by analytical
techniques can be transferred to the whole extension of the artwork
thanks to the FLIM maps, without the need for an extensive sampling. FLIM is mainly intended to give "imaging
capabilities" to material investigations, carried out through
laboratory measurements. Dealing with the large surfaces typical of
wall paintings, many important results can be achieved with FLIM: the
need of micro-sampling is strongly reduced; fluorescent spots or areas,
corresponding to organic contaminants or intentionally added
treatments, can be identified just on the basis of fluorescence
measurement and the care of restorers can be addressed toward
fluorescent (i.e. anomalous) details. Furthermore, FLIM can support a
monitoring activity, which is crucial to guarantee a good control of
the restoration procedures.
The fluorescence imaging system used for this
study is based on a time gated CCD camera provided with a light
intensifier that amplifies the light and has an electronic shutter
exhibiting a minimum exposure time of 10 ns. A sequence of images is
acquired by activating the shutter at different delays with respect to
excitation pulses. In this way, the temporal behaviour of the
fluorescence emitted by each pixel is recorded. Then, by applying a
suitable fitting procedure, the fluorescence lifetime map of the field
of view is reconstructed. The UV (λ
= 337 nm) excitation
light is provided by a Nitrogen laser that generates 1 ns long pulses,
synchronous with the gated camera, at a repetition rate of 50 Hz. The
excitation beam is coupled to an optical silica fibre having a core
diameter of 600 μm
and delivered to the artwork in a circular area of about 20 cm in
diameter. A portable spectrometer completes the
experimental apparatus. It measures fluorescence spectra from 400 to
800 nm (blue to near infrared). The excitation light is provided by a
second Nitrogen laser. The laser beam is coupled to a silica fibre
bundle, which is put in gentle contact with the artwork through a
metallic spacer covered with a Teflon ring.
In a typical measurement performed with the
FLIM system several images of the fluorescence emission are acquired.
Assuming a mono-exponential behaviour of the fluorescent emission:
the outcome of the FLIM analysis are two 2D
arrays τ(x,
y) and A(x, y) that represent the
spatial maps of the fluorescence lifetime and amplitude of the sample
in the field of view of the gated camera. The first map reveals areas
with different chemical composition, while the second gives information
on the relative abundance of the fluorescent materials in the field of
view. By merging the two maps, a third one, named HSV map, is created.
This map is based on the HSV ( The spectral measurements result in a
wavelength dependent fluorescence amplitude A(λ)
that gives further information to identify organic materials.
The painted surfaces are first analysed with a
Wood lamp, in order to easily find out the regions where fluorescent
materials are mostly present. In each of these regions a FLIM
measurement is performed; the typical area imaged by the intensified
camera has a diameter of 20 cm. The exposure time of the image detector
is set to 100 ns and some images are recorded after different delays
with respect to the excitation pulses. Typical delays are: 0, 2, 5, 10,
15, 20, and 30 ns. From the reconstructed FLIM maps it is easy to
locate the regions on the painted surfaces where organic compounds are
present. It is also possible to distinguish regions showing similar
characteristics. In correspondence of each homogeneous region, a
fluorescence spectrum is recorded and, whenever possible, a
micro-sample is taken as well. Collected samples are studied in the
laboratory through an optical microscope, a scanning electron
microscope and a Fourier Transform Infrared spectrometer.
4.4.4.1
Measurements
on Frescoes The FLIM technique was applied to analyse mural
paintings located in the vault of a quite famous Renaissance church:
the Collegiata of Castiglione Olona, near The first result achieved was a wide area
assessment of the decay state of the painted walls: the whole surface
is permeated with a mixture of calcium caseinate,
gypsum and poly-vinylacetate
(PVA). Fig. 4.4.1 shows a picture of a damaged surface. As the picture
reveals, large areas of painted layers are lacking and an oblique
stucco joint is clearly visible. In Figs. 4.4.2a and 4.4.2b, the
fluorescence amplitude and lifetime maps are shown. In Fig. 4.4.2c, the
information provided by the two maps is combined in the HSV map.
From an analysis of these maps, the presence of
different regions in the field of view is evident. The areas where some
painting layers are still present appear to be dark in the amplitude
map, revealing the inorganic nature of the pigments, which simply act
as absorbers of the UV radiation. On the contrary, areas where the
painting layer is lacking show a strong fluorescence due to the
plaster. Actually, plasters are typically made of a mixture of lime and
sand, materials that are not supposed to show any fluorescence. Thus,
this unexpected emission reveals the presence of an organic compound
permeating the painted surface. More interestingly, the lifetime map allows one
to identify three different fluorescent materials: the plaster,
characterized by the shortest decay time (near 8 ns); the oblique
joint, with a longer lifetime (9.7 ns); circular spots, not visible in
white illumination, characterized by the longest lifetime (10.5 ns).
The shape of these fluorescent spots, as well as their position along
the joint, lead us to attribute their presence to restorers’
activity, who probably used nails to fix the panels to the rigid
support and masked them with stucco made of organic and inorganic
compounds. The fluorescence spectra of the plaster and of
a circular spot confirm their different nature (Fig. 4.4.3) being the
former spectrum peaked at 460 nm and the latter at 550 nm.
The identification of the materials
constituting this portion of the artwork was obtained through FTIR
spectroscopy on collected micro-samples. Fig. 4.4.4 shows the FTIR
spectra of a plaster sample as it was (a) and after acid attack (b),
which removes the inorganic fraction. Curve (b) reveals that the
organic behaviour of the plaster is due to the presence of calcium caseinate and PVA, while in
curve (a) the presence of gypsum can be recognized. Gypsum can be
related to the sulphation
process of the calcium carbonate, which is the binding phase in the
plaster. Spectra of the joint and of circular spots reveal the same
molecular composition, with a prevalence of gypsum.It
is worth noting that the nature of the analysed surface is quite
complex. The plaster, the circular spots and the joint are made of a
mixture of the same organic and inorganic materials, but in different
proportion. These differences are well outlined by fluorescence
measurements.
4.4.4.2
Measurements
on Marble Sculptures
Extensive FLIM measurements were carried out on
Michelangelo’s The FLIM images were processed and grouped
according to the identified fluorescent characteristics of the
different areas. Only a few images and spectra are reported and
discussed here, with the aim of illustrating and summarising the most
relevant results. First of all, David’s surface shows a
generally intense fluorescence emission. Actually, the
David’s marble surface is extensively covered with or
permeated by extraneous materials, which have their own emission
properties. In fact, the calcium carbonate itself, which constitute the
marble stone, is characterised only by a very faint fluorescence
emission after UV excitation. Three main types of overlaid materials were
identified: wax residues, concentrated in small drops or permeated into
the marble surface; salt deposits, mainly composed of gypsum, calcium
oxalates and particulate matter; organic contaminants (not precisely
identified), concentrated in small areas or spots. In regions showing a light-blue or violet
emission in the Wood's lamp images (Fig. 4.4.6b, red ellipse) the
fluorescence emission is intense (Fig. 4.4.6d) and characterised by a
lifetime value of 5.5 - 6 ns (Fig. 4.4.6c); these features are
common to areas that seems to be rather clean or unaltered in visible
light (Fig. 4.4.6a).
Some small fluorescent spots, sporadically
present on the surface (e.g. on the right forearm) also show a lifetime
value of approximately 5.5 - 6 ns (Fig. 4.4.7c). They show a light-blue
emission of high intensity, as it is clear from the UV pictures (Fig.
4.4.7b) while they are not visible in white light (Fig. 4.4.7a).
The FTIR analysis of a micro-sample collected
in correspondence of the light-blue fluorescent spot (Fig. 4.4.8a),
allowed the identification of wax residues (generally adsorbed in the
porous structure of the stone substrate). Through the lifetime maps
collected all over the marble surface, it was possible to localize wax
remains also on the back, on the shoulders and on the trunk. Since the
fluorescence lifetime of unaltered areas is similar to that of wax (see
Fig. 4.4.6, red ellipse) we can hypothesize that, once, the whole
surface had been treated with beeswax.
In the regions appearing yellow in white light
and dark ochre or brown in the Wood's lamp images, the fluorescence
emission is generally characterised by a low amplitude and a rather
short lifetime value around 4-5 ns. Those areas have been correlated to
regions showing a deposit of salts and particulate matter. This is
clearly visible in Fig. 4.4.6c (blue rectangle), where short lifetime
values correspond to the surface covered by the dark deposit (Fig.
4.4.6a). FTIR analysis of the thick brownish deposits, well visible in
white light, assessed that they are generally composed of gypsum along
with calcium oxalate and sometimes nitrates, carbonates, quartz (Fig.
4.4.8b). Most likely, the fluorescence emission (due to the underlying
wax residues on the marble surface) has been quenched (lifetime
dumping) by inorganic salts and by the presence of metals like iron,
potassium, sodium, calcium and lead. In Fig. 4.4.9, an example of the longitudinal
deposits that are hardly visible in white light but are dark brown in
UV light is shown. The area is located on the back, left to the sling:
signs of deposits, due to the rainfall, are evident. Short lifetime
values (blue areas in Fig. 4.4.9c) are precisely correlated with the
vertical deposits, while longer lifetime values correspond to the
“clean” part of the surface. This surface
alteration should be correlated to the outdoor exposition of David
sculpture. In any case, under the running lengthwise deposits, the wax
seems to permeate the whole marble surface, since its characteristic
light-blue fluorescence and long lifetime values are evident in all
images.
The FLIM apparatus was also applied to compare
different cleaning methods applied to small test areas on the statue.
As an example, cleaning tests were performed in a region located on the
left shin, characterized by the presence of inorganic deposits (mainly
composed of gypsum). Fig. 4.4.10a shows two patches that were treated
with different cleaning procedures: the patch above (G1) was cleaned
with deionised water poultice,
the patch below (G2) was cleaned with ion exchange resin (DES90). The
fluorescence lifetime maps of the two areas taken before (Fig. 4.4.10b)
and after (Fig. 4.4.10c) the cleaning are
also shown. The increase in the fluorescence lifetime that
takes place after the cleaning (red shift of the false colour map) in
both patches indicates that some inorganic deposits have been actually
removed by the cleaning procedures. In fact, the reduction in
fluorescence dumping associated with the very superficial layer let the
long living emission of wax absorbed in the marble to become more
relevant. Moreover, the slightly greater increase in the lifetime shown
by the patch G1 indicates that water poultice is possibly more
effective for cleaning than ion exchange resin.
UV fluorescence is an inspection technique well
appreciated since long time ago by specialists operating in the field
of restoration and conservation of Cultural Heritage. Nevertheless,
recent technological developments allowed us to completely redesign and
improve this technique. The breakthrough that makes FLIM procedure
really different from standard UV imaging by Wood lamp is its
capability to measure objective fluorescence parameters, like lifetime
and spectrum, combined with analytical measurements carried out on
micro-samples taken from the artwork. The synergic combination of
laboratory measurements and in-situ imaging leads to the map of many
organic substances that are relevant for the conservation of the
artwork. These concepts have been successfully applied during an
extensive measurement campaign on Renaissance fresco paintings and on a
marble masterpiece. In particular, the experience with David showed
that FLIM allowed us to identify some of the superimposed materials.
This is especially true for wax, which has been found all over the
statue as a background signal and in well-outlined spots, sometimes
looking like drops. Also most of the inorganic deposits were mapped
thanks to an indirect quenching effect on the underlying fluorescence
emission. The scope of application of the technique is
rapidly expanding. Measurements are currently in progress on other wall
paintings, where the fluorescence analysis is mainly intended to study
the finishing painting details; in this case the fluorescence emission
is due to the organic materials (e.g. tempera glue) used as binders for
pigments. Many other artworks, like oil paintings and
ancient manuscripts, would possible benefits from fluorescence
measurements, thus confirming the great flexibility and ease of
application of optical techniques.
[1]
R. Cubeddu,
D. Comelli, C. D’Andrea, P. Taroni, G. Valentini: Time-resolved
fluorescence imaging in biology and medicine. J. Phys. D: Appl. Phys. 35, R61 (2002). [2]
D. Comelli,
C. D'Andrea, G. Valentini, R. Cubeddu, C. Colombo, L. Toniolo: Fluorescence lifetime
imaging and spectroscopy as a tool for non destructive analysis of
works of art. Appl.
Opt. 43 (2004) 2175. [3]
D. Comelli,
G. Valentini, R. Cubeddu L. Toniolo: Fluorescence lifetime
imaging and FT-IR spectroscopy of Michelangelo’s David. In
press in Appl. Spectrosc. (2005). [4]
L. Toniolo, A. Sansonetti, C. Colombo, R. Cubeddu, G. Valentini, D. Comelli: FLIM – Fluorescence Lifetime Imaging.
In: Exploring David - Diagnostic Tests and State of Conservation. S. Bracci, F. Falletti,
M. Matteini, R. Scopigno (Eds.)
Giunti Editore, Florence- Milan 2004, p.154.
Daniela Comelli CNR-INFM and CNR-IFN Politecnico di Milano - Dipartimento di Fisica Piazza Leonardo da Vinci
32 I-20133 Milan Italy W: http://www.fisi.polimi.it/dip-fisica/
Gianluca Valentini CNR-INFM and CNR-IFN Politecnico di Milano - Dipartimento di Fisica Piazza Leonardo da Vinci
32 I-20133 Milan Italy E: gianluca.valentini@polimi.it W: http://www.fisi.polimi.it/dip-fisica/ Rinaldo Cubeddu CNR-INFM and CNR-IFN Politecnico di Milano - Dipartimento di Fisica Piazza Leonardo da Vinci
32 I-20133 Milan Italy W: http://www.fisi.polimi.it/dip-fisica/ Lucia Toniolo Istituto per la Conservazione e la Valorizzazione
dei Beni Culturali – CNR Piazza Leonardo da Vinci
32 I-20133 Milan Italy
|
