Available via license: CC BY-NC-ND 3.0
Content may be subject to copyright.
ISSN:1369 7021 © Elsevier Ltd 2008
NOVEMBER 2008 | VOLUME 11 | NUMBER 11
40
Looking over the
artist’s shoulder
American artist Winslow Homer’s watercolor painted in 1887,
For to be a Farmer’s Boy (Fig. 1a), shows a young boy in a field of
pumpkins. We can almost feel the heat of the sun beating down
on him as he labors in the field. But the work of researchers at
Northwestern University and the Art Institute of Chicago reveals
that the artist may have had a rather different impression in mind.
The color of light
Here is the materials science connection. Understanding how pigments
such as the chrome yellow and pink madder used by Homer in
his watercolor respond to light and/or time can only be achieved
using sensitive – preferably non- or minimally destructive – analysis
techniques familiar to any materials scientist.
Chemical analysis of the watercolor using surface enhanced Raman
spectroscopy (SERS) shows that the red and yellow pigments used
for the sky have significantly faded over time. A simulation of the
original colors casts a rather different hue onto the scene – instead of a
bleached out sky, there is a soft orange glow of sunset (Fig. 1b).
Raman spectroscopy is a stalwart in the investigation of artworks.
The technique is based on the fact that when an incident photon in
a laser beam is scattered inelastically by a material, its frequency
changes. This unique chemical fingerprint of a material is known as its
Raman spectrum1. However, natural organic dyes are easily damaged
by light and fluoresce when probed with Raman spectroscopy. Natural
organic dyes only started to be replaced with synthetic dyes in the late
1800s and 1900s, so this makes the analysis of many artworks using
Raman somewhat limited.
SERS, which is a variant of Raman spectroscopy, is not plagued
by this problem and takes this kind of analysis to the next level.
Pioneered by Richard Van Duyne at Northwestern among others,
What could materials science possibly have to do with revealing an
artist’s intentions? Intriguingly, art institutions and museums around
the world are using a panoply of analysis techniques familiar to any
materials scientist to reveal vital and previously hidden information about
works of art. We often assume that a painting or statue we are looking
at in a gallery or museum is just as it was when the artist created it.
But thanks to the rigors of time, artifacts and paintings may deteriorate
in subtle or not so subtle ways that can alter our perception of those
works, as well as our understanding of their place in history. Or an
artist may have reworked or even painted over a piece.
Cordelia Sealy
Oxford Science Writing Ltd., Oxford, UK
Email: Cordelia.Sealy@Googlemail.com
mt1111.p40_45.indd 40 12/11/2008 12:23:40
Looking over the artist’s shoulder INSIGHT
NOVEMBER 2008 | VOLUME 11 | NUMBER 11 41
SERS uses a roughened metal film to enhance Raman scattering from
sample molecules. Until recently, a nanoscale layer of Ag was the
most commonly used material, but now arrays of metal nanoparticles
are proving to have advantages. By effectively amplifying the Raman
scattering signal and quenching fluorescence, SERS is able to analyze
very small amounts of small molecules – such as organic dyes.
Art Institute conservation scientist Francesca Casadio and Van
Duyne turned to SERS when X-ray fluorescence spectrometry and
optical microscopy revealed that Homer had used red and yellow
pigments in his watercolor. “Typically, when we have questions about
the palette of a painting we do a survey with X-ray fluorescence,”
says Casadio. “This gives us an overview of the materials present
– the elemental composition but not the chemistry.” Using SERS,
however, enables the researchers to discover remaining traces of
pigments that have degraded over time. Even more tellingly, they can
use this information to create a digital simulation of how the original
watercolor would have looked when the artist painted it (Fig. 1b).
The Northwestern-Art Institute collaboration is an example of how
work can span across the two main categories of scientific activity
typically undertaken by museums and galleries. On the one hand,
explains David Saunders of the British Museum, activities are directed
at the conservation and restoration of works of art. This includes
gathering an understanding – through scientific means – of how the
materials making up an artwork are deteriorating and how these
processes can be slowed down. In some works, such as with the Homer
watercolor, this relates to the pigments used, while in others it might
relate to the corrosion products of a bronze statue, for example. This
complements the other aspect of scientific activity – understanding
the making and meaning of objects or works of art. The authentication
of works is rarer than one might imagine, says Saunders, but more
often there are questions of provenance or authorship that scientific
investigation can elucidate. Furthermore, these techniques and the
information they provide can be used to reveal the original purpose or
intention of an artist or the maker of an object – in effect, an insight
into the artist’s mind.
DORIS beams on van Gogh
An even more startling example of what the analysis of a painting
can reveal about the working methods of an artist has just been
demonstrated by a team of European researchers2. Many painters reuse
canvases – but van Gogh was particularly prolific and often painted
over old, abandoned paintings.
“At the height of his productivity, van Gogh was producing
a painting every three days and up to 20% of his oeuvre shows
overpainting,” says Joris Dik, one of the researchers at Delft University
of Technology who led the project.
Traditional radiographic analysis of one of van Gogh’s paintings,
Patch of Grass (Fig. 2a), had revealed the presence of a ghostly head.
However, traditional X-ray radiography looks through the whole
painting in one go and is mainly sensitive to heavy elements. In the
analysis of a painting, this can mean that the heavy metal components
of a pigment, such as the lead in lead white or the mercury in
vermillion, can swamp the contrast from other elements. “It can give a
partial and colorblind view,” says Dik. Traditional techniques could only
give a hint of the ghostly face in van Gogh’s painting.
So Dik and coworkers from the University of Antwerp in Belgium
and the Centre of Research and Restoration of the French Museums
in Paris convinced the Kröller-Müller Museum in the Netherlands to
subject the painting to a more rigorous examination using a new
technique – synchrotron radiation induced X-ray fluorescence (XRF)
spectroscopy.
XRF records the intensity of secondary radiation emitted when
Fig. 1 (a) Winslow Homer’s For to be a Farmer’s Boy (1887). Watercolor with touches of scraping, over graphite, on cream wove paper. 355 x 509 mm. The Art
Institute of Chicago, Gift of Mrs. George T. Langhorne in memory of Edward Carson Waller. (b) Digital simulation of Homer’s watercolor For to be a Farmer’s Boy
providing an approximation of the work’s original appearance, in which Homer created an orange sky with washes of chrome yellow, pink madder, and vermillion,
now faded.
(b)(a)
mt1111.p40_45.indd 41 12/11/2008 12:23:51
INSIGHT Looking over the artist’s shoulder
NOVEMBER 2008 | VOLUME 11 | NUMBER 11
42
a beam of high energy X-rays bombards the surface of an object.
The technique is commonly used for the point analysis of pigments
in paintings, but the researchers have extended the scope of the
technique to a whole new level. By taking the painting to the DORIS-III
synchrotron light source at Deutsches Elektronen-Synchrotron (DESY)
in Hamburg, Germany, they were able to scan an area of the van Gogh
painting in front of a pencil beam of high intensity X-rays (Fig. 3).
The researchers took 90,000 individual measurements over a
17.5cm x 17.5cm area of the painting where they believed the hidden
head would be found. Even though the dwell time of the beam at each
pixel was only 2s, the total scan took two days.
“For every pixel scanned, we record the fluorescence spectra to get
the full elemental analysis,” explains Dik. “From this we can deduce
which pigments were used and create a three-dimensional model of
the painting. We can then peel off the layers one by one until we get
to the layer of interest.”
The decimeter-scale XRF intensity maps reveal the distribution of
specific elements in an individual paint layer, in particular antimony
and mercury, which are used in red and light tones, respectively, for
flesh colors. The much more detailed, approximate-color reconstruction
of the hidden face (Fig. 2b) rendered in this way proved to be the
missing link in understanding its presence under Patch of Grass.
“We can now see that the head appears to dovetail with a series of
other portraits of Dutch peasants painted between October 1884 and
May 1885 before the artist moved to Antwerp,” says Dik. The portraits
appear to be studies for his opus magnum, the Potato Eaters, suggests
Dik.
From van Gogh’s letters, we can surmise that the original painting
under Patch of Grass is likely to have been one of a number of canvases
of heads that he sent to his brother Theo in Paris. When van Gogh
joined his bother in Paris a few years later, the original woman’s head
may have seemed “hopelessly old-fashioned” and ripe for overpainting
with a colorful, Parisian-style floral landscape, suggest the researchers.
The new approach literally opens up new vistas in the
nondestructive analysis of hidden paint layers, says Dik. “Revealing
overpainting allows us to look over the artist’s shoulder and understand
their artistic journey.”
Restored but not forgotten
Analysis of artworks can not only reveal how materials have
deteriorated or degraded over time, and the presence of underpainting,
but also what kind of restoration has been undertaken on an artwork.
Until relatively recently, the care and handling of artworks has not been
well documented, if at all. This can make it hard to determine which
features of a piece were bestowed by the original artist and which have
been added later to restore or ‘improve’ the piece.
Professor of materials science at Northwestern, Katherine Faber,
describes an example of this kind of investigation undertaken in her lab.
The Art Institute of Chicago has a statue of a kneeling figure from the
Shang Dynasty (1600-1045 BC) in its ancient Chinese jade collection
that is so dark as to be almost black – not the vibrant green one
traditionally associates with jade pieces. Faber worked with Casadio
and a student in the department to determine whether the unusual
Fig. 2 (a) Patch of Grass by Vincent van Gogh, Paris, Apr-June 1887, oil on
canvas, 30 cm x 40 cm, Kröller-Müller Museum, Otterlo, The Netherlands
(KM 105.264; F583/JH1263). (b) Patch of Grass with insert revealing tritonal
reconstruction of antimony (yellowish white) and mercury (red) representing
the flesh tones of the hidden face.
(b)
(a)
mt1111.p40_45.indd 42 12/11/2008 12:23:55
Looking over the artist’s shoulder INSIGHT
NOVEMBER 2008 | VOLUME 11 | NUMBER 11 43
color of the 3000-year-old statue (Fig. 4) was a result of the mineral
used to make it, a coating, or the result of some kind of burning
process.
While the answers aren’t definitive, says Faber, a combination
of Raman and Fourier transform infrared (FTIR) spectroscopy, X-ray
diffraction studies, and scanning electron microscopy of the surface
indicates that the unusual color of the statue is likely to be the result
of two factors. Heating of the piece – either during its carving or as
part of ritual ceremonies – and subsequent treatment with Japanese
wax, which is favored by dealers to improve the appearance of jade
or other hard stone items, or a combination of the two, are the likely
culprits.
Letting the cat out of the bag
Revelations about past restoration or repair work can also lead to an
insight into the original intentions of an artist, as recent work at the
British Museum on one of its most famous pieces demonstrates. No
trip to the museum would be complete without seeing one of the
highlights of the Egyptian collection – the Gayer-Anderson cat (Fig.5).
This fine bronze sculpture of a rather majestic feline dating from
around 600 BC came into the museum’s possession some time in the
1930s after being uncovered in an antiquities shop in Cairo. But beyond
that, little is known about the piece.
A team at the British Museum recently undertook an investigation
of the statue with some surprising results3. X-ray investigation
indicated that the piece had been made using a lost-wax technique
consistent with the period it was believed to have dated from, but it
also revealed serious damage. At some point in the past, it appears
that the statue had broken in two. Further analysis with Raman
Fig. 3 The experimental set up showing Patch of Grass at the synchrotron set up DESY in Hamburg. (Courtesy of DESY).
Fig. 4 Chinese Shang Dynasty kneeling figure under examination by Raman
micro-spectroscopy. (Courtesy of The Art Institute of Chicago Conservation
Laboratory).
mt1111.p40_45.indd 43 12/11/2008 12:24:10
INSIGHT Looking over the artist’s shoulder
NOVEMBER 2008 | VOLUME 11 | NUMBER 11
44
spectroscopy revealed corrosion products – copper oxide and chloride
– consistent with a bronze piece of this age except in the region of the
fracture, where a mixture of other compounds was detected. In this
region, the analysis reveals materials that bear a strong resemblance to
a pigment introduced in the 19th century known as ‘Brunswick green’.
Most likely the statue was dropped or damaged during excavation,
suggests Saunders, and then carefully repaired and all traces of the
fracture painted over.
Now knowing that the statue was hollow, the scientists used
endoscopy to peer inside the cat. What they found appears to a
mixture of ancient flax fibers and pieces of skin. While it is impossible
to be sure, Saunders suggests that these findings may indicate that the
Gayer-Anderson statue could have been made originally as a repository
for a mummified cat.
The scientists’ analysis also indicates that the original maker of
the piece may not have included the rather gaudy earrings and nose
ring. XRF techniques also reveal that the original cat had a stripy tail –
perhaps in remembrance of the breed of cat it was designed to house.
Art-science collaboration becomes prime
attraction
Behind the scenes at museums and galleries around the world,
conservators and scientists are using a wealth of non- and minimally
destructive techniques to give us an insight into the origins of artworks
and also to ensure that we can continue to enjoy those artworks many
years to come. While the larger world-renown institutes have teams
of experts, smaller museums and galleries often rely on one or two
individuals. In these cases, collaborations with local universities can be
vitally important.
The Art Institute’s collaboration with Northwestern’s Department of
Materials Science and Engineering is just such a case in point. Pioneered
by Faber at Northwestern and Casadio at the Art Institute, the origin
of the relationship between the two organizations was somewhat
serendipitous, says Faber, but they have plenty in common.
“The materials science paradigm of looking at structure property
relationships makes it a natural discipline for art conservation to turn
to,” says Faber.
Casadio agrees that there are many affinities between the two
fields – in particular, the recognition of the importance of applied, real-
world problems. “There is real benefit in faculty looking at a new set of
problems where they have expertise,” says Faber.
As an added bonus, the collaboration with the Art Institute has
proven very attractive to materials science and engineering students
at Northwestern. “It gives them awareness that materials science is so
much a part of everything in life,” says Faber.
In the collaboration between science and art, there could be an
unexpected route to communication with the public about science.
By understanding how scientific method and techniques can reveal
the previously unknown story of a work of art, or an insight into
an artist’s mind, the public can see firsthand, in a familiar (and
nonscientific) context how the scientific process works. “Working with
museums and making science nonthreatening is very valuable work,”
says Casadio.
Its advocates hope that, as well as an insight into the workings of
an artist’s mind, a recognition of this fringe benefit of the interaction
between science and art could provide a stable foundation for the
funding of this kind of work in the future.
Acknowledgments
Many thanks to all those who gave their time and provided valuable insights,
as well as images, for this article. In particular, Francesca Casadio at the Art
Institute of Chicago, Katherine Faber at Northwestern University, Joris Dik at
Delft University of Technology, and David Saunders of the British Museum.
REFERENCES
1. Clark, R. J. H., Appl. Phys. A (2007) 89, 833
2. Dik, J., et al., Anal. Chem. (2008), doi: 10.1021/ac800965g
3. Ambers, J., et al., British Museum Technical Research Bulletin (2008), in press
Fig. 5 The Gayer-Anderson cat (EA 64391). (© Trustees of the British Museum).
mt1111.p40_45.indd 44 12/11/2008 12:24:15
