Every potential conservation treatment and every technical study of a particular painting begins with a thorough visual examination of the artwork in ordinary light. If the surface is then illuminated by a spotlight held at an acute angle, in the same way that the setting sun causes objects to cast long shadows, the unevenness of the paint surface is thrown into sharp relief. This simple technique can reveal a great deal about the way the paint layers have been built up.
Next, a binocular microscope is used. This piece of equipment has two eye pieces resulting in a three-dimensional view, as shown in Fig.1 Flexible fibre-optic guides provide raking or diffuse light, and a magnification range of x7-100 allows a highly detailed examination of the surface. At x7 magnification it becomes easier to see the texture of the canvas or paper support, brush-strokes of paint or watercolour, and any cracks and their accumulated contents. The impression of a pen or a tool on paper, a graphite pencil mark showing through the paint on top, slight unsteadiness of hand, the overlay of one colour on another and the handling and flow of the paint when wet (known as its rheological properties) all become apparent as the magnification is zoomed up. At x40 the sequence of paint application can no longer be grasped, but it is possible to see how the artist combined one pigment with another; for example where a blue and a yellow have been mixed to make a green. At x40-100 it may even be possible to identify a few of the pigments that give the paint its surface colour.
Technical photography is a useful non-invasive tool for conservation scientists, and is brought into play next. Many artists' materials exhibit ultra-violet (UV) fluorescence: that is, they emit visible light when illuminated with ultra-violet light. This is exactly the same effect as seen on dance floors when brightening agents in clothing fluoresce white under UV lights. The human eye cannot detect UV but some materials absorb it and then emit it as visible light. Fig.2 illustrates an example of the use of photography in ultra-violet light to detect a traditional red pigment called madder which fluoresces a bright orange colour. This distinguishes it from Alizarin Crimson, a closely matching synthetic red paint.
Infra red (IR) is another form of radiation that the eye cannot detect. It is familiar in many applications, such as night-time photography which provides a visible image based on the normally invisible IR night-time view. IR has the property of penetrating paint films so that under-layers can be viewed. Some drawing materials in particular absorb IR well, such as pencil or charcoal. At Tate, conservators have been able to detect dramatic changes in composition using IR equipment during the Tudor Stuart cataloguing project. Fig.3 shows Portrait of a lady, thought to be Dionesse Cullum, wife of Robert Colman (c. 1685) by Harman Verelst (Tate, T07241) in which the sitter is depicted looking to her right; in Fig.4 IR photography reveals that in an earlier version she looked the other way. Illumination is by a light source which provides infrared light - and most lights do - causing the paint surface to be heated very slightly. This technique is called infrared reflectography, when it is applied to paintings.
Another important technique for all conservators and conservation scientists is X-radiography. From our familiarity with x-radiography in hospitals, we know that x-rays penetrate through structures. Bones and teeth absorb X-rays most and show as light areas on the resulting films. This same principle applies to paintings, but it is only in the case of oil paintings where the artist has used lead white that the resulting x-radiographs look like the painted image. This traditional white pigment strongly absorbs x-rays, so the distribution of white lead matches the distribution of white on the X-radiograph. If the modern pigment titanium white has been used instead, a meaningless X-ray image results which will not resemble the painted image. Since pale-coloured paint is often used for modelling forms before further paint is applied, earlier versions and alterations in lead white-based paint often show up on X-radiographs.
All of the techniques described above do not alter the painting in any way; they are non-destructive. However, if small samples can be taken from the paint, a number of analytical techniques can also be applied. For sampling to be feasible, there usually have to be areas of the painting which are damaged and where a small additional loss will not be noticed. Otherwise the maximum size of a sample is limited by the need for its removal to be invisible under all normal viewing conditions, and its minimum size depends on its intended use. The size of samples varies from a pin-point to a pin-head.
For a cross-section, the sample must be taken through more than one layer of paint and remain coherent. It is placed on a solid block of transparent synthetic resin in a mould and further liquid resin is added to embed the sample. Once set, this is ground down carefully at right angles to the paint surface, until the sample is exposed edge-on. It is then polished, avoiding contact with water if it is feared that any layer might be water-sensitive, and viewed in reflected light under an optical microscope at x 100-500, as shown in Fig.3. Layers of varnish, paint and ground can be examined, and the fuzzy or sharp interface between layers indicates whether they were applied wet-in-wet or after a period of drying.
Ultraviolet fluorescence microscopy reveals if the varnish is actually comprised of multiple layers and provides clues as to whether the paint medium in each layer is oil-based, protein-based, or a modern synthetic resin. This is invaluable when the medium is analysed in more detail later, as it helps determine which techniques are most appropriate for the analysis. To further assist in distinguishing the different layers the cross-section will then be stained with stains specific for proteins [Fig. 5], or examined using thermomicroscopy. This is the controlled heating of a paint fragment while its sudden or gradual melting is observed with an optical microscope, since many materials in paintings have specific, known melting points.
Examining the optical properties of the inorganic pigments present allows a skilled and experienced microscopist to identify most of the commonly used, traditional pigments, even in quite complex mixtures. Virtually all historic pigments can be identified, because they were ground by hand and tend to have large particles with distinctive properties. The identity of pigments in a tiny fleck of paint, or in a layer within a cross-section, can be confirmed by analysing the elements that are present, and subsequently deducing which combination of pigments gave rise to the range of elements detected. Most scanning electron microscopes (SEMs) are equipped to carry out such analysis, which is known as EDX energy-dispersive X-ray analysis. This technique is in principle similar to UV rays being emitted as visible light: the electron beam reacts with the sample to generate an x-ray that characterises a given element present in it. Fig.6 shows the sample chamber of an SEM microscope.
All of the techniques described produce a 'finger print' for a specific material and most of the techniques involve removing a sample from a work of art, hence they are destructive. The fingerprint produced is identified by referring to reference standards and libraries, but if an artist has used a very unusual material it is possible there will not be reference materials available. This is therefore an area of continuing research. Sometimes Tate's conservation scientists make up paint according to historic recipes, in order to create reference materials.
The binding medium of paints can also be complex. Analytical techniques for binding medium analysis have developed rapidly over past decades to reach a level of sophistication that allows most materials in a work of art to be identified. Drying oils, waxes, egg, casein, animal glue, as well as modern paint media such as acrylic or alkyd, and plastics and common varnishes or coatings (natural and synthetic resins, waxes etc.) are all regularly analysed and characterised today. The two techniques commonly employed are Fourier transform infra-red spectroscopy (FTIR) [Fig.7] and gas chromatography-mass spectrometry (GC-MS), shown in [Fig.8].
FTIR analysis can provide information on varnishes or coatings, binding media, some white pigments known as extenders or fillers, organic pigments and some inorganic pigments present in a painting. It works on the principle that particular chemical bonds absorb infra-red radiation of characteristic energy, and each sample produces a 'spectrum' that reflects the combined absorption of the materials present. As paint samples are often complex and consist of many materials, the FTIR spectra of each component present may be superimposed on one other, so identifying all of the components in a complex mixture such as a paint film can be quite an art! Often the characteristic fingerprint of one material is masked by that of another and the identification of some of the materials, particularly those present in smaller amounts, can consequently be difficult.
The other problem encountered with FTIR is that it is not always the best technique for distinguishing between the different types of proteins, for example egg and animal glue, or the different types of oils, such as linseed and poppy. FTIR is best used for the general identification of 'oil' or 'protein' in a sample rather than the specific types. If the exact identification of oils/proteins etc. in a work of art is required for technical art history purposes, or to help with a conservation treatment, GC-MS is usually employed.
GC-MS has been used for the detailed identification of oligomeric and polymeric materials such as drying oils, plant gums, proteins, waxes and natural resins for decades. This technique has more specificity and can distinguish between different types of oils (linseed/poppy/walnut/castor oil), acrylics (emulsion/solution), gums (Arabic/tragacanth) and proteins (egg/animal glue/casein). It has also been used to explore the deterioration processes of artists' and conservation materials as well as monitoring conservation treatments.
With all forms of chromatography, the identification of any polymeric material is made by breaking down the polymeric material - using chemicals or heat - into its smallest possible unique units, called monomers. Each polymer type has a unique set of monomers that can then be used to identify the 'unknown' polymer. For example, the monomers of proteins are amino acids, and amino acid compositions are a common way of identifying unknown proteins.
Once the polymers are broken down into a 'soup' of monomers, they are then separated into single monomers by a complex process through the GC column, and each monomer type leaves the column at a different time (called the retention time), with the lower molecular weight compounds generally passing through first. Once through the column, each monomer is detected by the MS part of the GCMS instrument, and the monomer appears as a peak on the final results of the analysis, called a 'chromatogram'. The MS part of the instrument also provides information on the molecular weight of each of the monomers and, along with the characteristic retention time of the monomer leaving the GC column, this helps to precisely identify the monomer. Once all the monomers are known, and sometimes also their quantities, the polymer can be identified as linseed oil, poppy oil, castor oil, egg, animal glue, casein etc.
All of these techniques except thermomicroscopy and EDX are available in Tate's Science section. The analytical results obtained greatly inform technical art history and the treatment of works in Tate's collection.
February 2007
