Because microscopy is the first need of most museum research laboratories and perhaps the only possibility for many museums, this part of the work is described in rather greater detail.
Optical and chemical microscopy has been a feature of the Department's work since the institution of a chemical laboratory in 1949. At that time the space allotted to the subject was virtually a cupboard under the stairs, it occupied one member of staff half-time and throughout the 1950s and early '60s most of the work was done on an 1895 Leitz brass microscope (which we still have and sometimes use, its optical system as good as ever). Photomicrography was done by balancing a box camera precariously over the tube of the monocular microscope, waiting for the vibration to die down, then guessing the exposure. In the new department microcopy is spread over three rooms, occupies two persons full time and, over the years, as the need has arisen and as funds have become available, a comprehensive range of apparatus has been systematically acquired.
By means of a combination of microscopical examination and chemical tests (carried out either under the microscope or as 'spot tests' on drops of the solution of the pigment on filter paper or porcelain spot plates) the majority of artists' pigments can be identified in very small samples. Some mineral pigments can be identified with certainty, given only a few grains, solely from their particle characteristics and optical properties (e.g. refractive indices, birefringence). The optical microscope is also used for identifying species of wood of panel paintings and textile fibres of canvases. We have found though, that its greatest value lies, in the case of paintings, in the elucidation of the layer structure. For this purpose a minute flake of paint, no more than about 0.5 mm diameter, is removed from the picture and after a preliminary microscopical examination is embedded in a block of transparent plastic. One face of the block is ground down until the edge of the paint flake is exposed then grinding and polishing is continued until the surface is sufficiently smooth and flat to focus under the microscope. The sequence of layers – the ground (or layer of preparation), the drawing and undermodelling, the various paint layers, glazes and final varnish and, sometimes later overpaint – are clearly seen at magnifications about 100–300 X. In practice such samples are taken only when a picture is undergoing cleaning and restoration and if microscopical examination or chemical analysis of samples might help to solve some of the problems of diagnosis and treatment. Even then sampling is confined to areas of existing damage or to edges of the picture. While the picture is in the conservation studio, however, under ideal conditions for examination, opportunity is taken to look at the materials, condition and technique of the painting as a whole. Afterwards the prepared paint cross-sections and other samples are stored away with their accompanying data and have gradually been built up into a reference library of painters' materials and techniques. It is now possible, for example, to take and compare paint cross-sections from more than twenty different paintings by Rembrandt or Tintoretto.
The new ‘Microscope Room’ is a comparatively spacious squarish room with benches and storage cabinets running round the walls. A narrower, more compact room, like the Microchemical Laboratory might have been preferred but the size and the shape of the present room is dictated by the need to accommodate from time to time as many as twenty people for talks and demonstrations and also to house the extensive collection of reference samples mentioned above. The largest and most-used microscope, a Leitz 'Ortholux' research model, occupies the most prominent position (Fig.3). The height of the bench on which it stands (76 cm) is less than that of the standard laboratory bench in order that the operator may sit comfortably at the microscope, feet firmly on the floor. Microscope and camera accessories are in drawers and cupboards within arm's reach. A narrow shelf attached to the wall above the main microscope bench is provided for boxes of electrical controls and other much-used accessories. The 'Ortholux' microscope has built-in transmitted and incident light sources, separately adjustable as well as capable of being used together (often a help in studying paint and pigment samples). For incident light, the Leitz 'Ultropak' illuminator and objectives were chosen. In this system a hollow cone of light is directed from around the objective lens onto the sample. The resultant illumination shows up the structure of semi-opaque materials like paints and pigments far better than the vertical illumination customarily used for metal specimens. On the 'Orthomat' polarizing facilities have been restricted deliberately to a slide-in analyser and a rotating, but not graduated, polarizer, sufficient for detection of birefringence. A fully-equipped polarizing microscope is also available in the Department (see below and Fig.5). Photomicrographic equipment was soon added, first the 'Orthomat' 35 mm automatic camera then a large-format (12.5 X 10cmsheet film or plate) bellows camera, also with automatic control. Although high quality photomicrographs can be obtained using an ordinary reflex camera on top of the microscope, the process is rather time- and film-consuming, particularly with colour work. The 35 mm camera is particularly useful for rapidly recording series of samples and for photography of transient phenomena like chemical reactions, the sample being under continuous observation while the photograph is taken. Both cameras have two types of exposure reading, one giving an average reading for the whole field, the other a reading on a small selected area, the latter indispensible for small samples in the centre of a dark or bright field. A tungsten-halogen lamp can be inserted in a separate lamp-housing to replace or supplement the built-in tungsten lamp for incident-light study of dark specimens or for accurate focusing of the image on the ground-glass screen of the large-format camera. This lamp is interchangeable with a high-pressure mercury vapour lamp to provide illumination in the blue and ultra-violet range for fluorescence microscopy. Although a few pigments fluoresce in UV, it is the paint media and varnishes which more commonly exhibit fluorescence. Such organic materials cannot be specifically identified by their fluorescence in UV, but they exhibit different types and degrees of fluorescence depending not only on differences in their chemical composition initially, but also on differences in age and degree of chemical degradation of the same material. It is often possible to distinguish by this means layers of old varnish between original paint and repaint, or to discriminate between two paint layers in a cross-section which look identical under visible light and which may even have identical pigment but different media. The method is a help, like staining tests for media (see p.56), in establishing whether more than one sort of medium is present before embarking on instrumental methods for media analysis, such as gas or thin-layer chromatography. In fact it might here be stressed that optical microscopy is a useful and often necessary preliminary to the further examination or analysis of samples by any other method. For dissecting samples under the microscope prior to separate treatment of the various layers and components, there is a pair of Singer micromanipulators. Unlike the more usual type of micromanipulator which moves along coordinates, this type simply scales down the natural movements of the hand by a factor of four (in the case of our particular model) and enables small tools to be used under the microscope in the manner of a knife and fork to manipulate or dissect the sample.
The ‘Microchemical Laboratory’, for want of a more apt name, is a long narrow room with benches on three sides. One long bench, with a small sink and fume-hood at one end, serves for chemical microscopy, spot tests and other small-scale chemical reactions. A comprehensive range of chemical reagents, particularly organic reagents used in spot tests, is in glass-fronted cupboards above the bench. The rest of the bench space in this room is used for preparation of cross-sections of samples. The more frequently-prepared type of cross-section, i.e. the thick, opaque sections for viewing by incident light as described above, is ground and polished on a 'Metaserv' low-speed polishing wheel with interchangeable heads for the various grades of abrasives used. Thin sections for viewing by transmitted light may be cut from even smaller paint samples using the LKB 'Pyramitome' glass knife microtome (Fig.4). Old paint is not the ideal subject for thin-section microtomy for it often contains large particles of hard minerals, such as haematite, likely to damage any knife blade. Steel microtome knives are expensive to resharpen and the cost of a diamond knife of adequate cutting width would be prohibitive. Hence the choice of a glass-knife microtome. The triangular-shaped knives are literally pieces of broken glass. Nowadays, instead of being made by breaking glass with two pairs of pliers, they are rapidly and inexpensively made from strips of plate glass with a special knife-making machine (Fig.4) which gives a choice of angle for the cutting edge and exactly reproducible knives. The paint flake to be sectioned is embedded in a synthetic resin as before (though an epoxy resin is usually preferred to the polyester used for thick sections). The trimmed block of resin is placed in the sample holder and thin serial sections are cut, knife blade and sample simultaneously observed through a binocular magnifier. If one knife-blade is seen to be damaged another of three on a rotating mount can be moved into place. The thickness of the sections to be cut can be chosen between I-IOµ. Even at this degree of thinness paint layers containing lead white and similarly dense pigments arc opaque by transmitted light when viewed under the microscope. The principal use for these thin sections is the study of glaze layers containing relatively transparent pigments such as lakes of organic dyestuffs. There are also advantages in using thin sections rather than thick opaque ones in staining tests for media (see p.56). A programme of work on the identification of lake pigments of organic dyestuffs and of other glazing pigments using thin sections is being carried out on a Leitz comparison microscope equipped with a photomultiplier and variable-wavelength interference filter (see pp.35–44).
The polarizing microscope mentioned above (Fig.5) is housed in the ‘Laser Room’ adjacent to the Micro-chemical Laboratory, which enables it to be used when the Microscope Room is occupied. A Leitz 'Ortholux-Pol', it is fully-equipped with rotatable analyser having scale and vernier reading, polarizing condenser, rotating graduated stage, Bertrand lens for conoscopic observations and compensating plates. There is built-in transmitted light with interchangeable tungsten or sodium lamps. The polarizing microscope is used for examinations at higher magnifications by transmitted light, for measurement under the microscope and for determination of optical properties such as refractive indices and birefringence. A recently-acquired accessory is a heating stage (Fig. 5) which may be rapidly and easily attached to the rotating stage. The sample is placed between a heat-proof glass slide and cover slip in a central heating chamber and can be heated at a controlled rate to 350°C. An eyepiece attachment enables sample and thermometer scale to be observed at the same time. Applications have been in the study of effect of heat on pigments and media (the presence of waxes, for example, is readily detected) and accurate determination of the melting point of organic compounds. It is hoped to carry out tests on a series of lining adhesives in conjunction with the Conservation Department.
The ‘Laser Room’ takes its name from its principal occupant, a Zeiss (DDR) Laser Microspectral Analyser (Fig.6). The apparatus is ranged along one side of the room on a slate bench supported by brick piers. When first acquired it was laid out on sturdy wooden benches but each time a change of relative humidity occurred in the atmosphere the wood moved and the very long light path had to be painstakingly realigned. A touch of tradition is that the slate for the bench top was salvaged from the floor of one of the old galleries! The central feature of the apparatus is a binocular microscope above which is situated the laser head containing a neodymium (Nd3+) glass laser resonator (laser wavelength 1060nm) and flash tube. The laser head is connected to the microscope by an optical adaptor with variable diaphragm. The sample to be analysed is placed on the microscope stage. It can be of any shape or size which can be accommodated on the stage and requires no previous preparation. The surface of the sample is focused under the microscope and the intersection of the eyepiece cross- wires centred on the exact spot to be analysed. The laser beam strikes the sample at this spot and in doing so vaporizes a minute quantity of it leaving a small crater in the surface. The diameter and depth of the crater may be controlled by varying the capacitance and inductance values and the flash lamp voltage at the supply unit and additionally by means of the optical adaptor. Situated between the microscope objective and the sample surface is a pair of carbon electrodes. When the laser beam strikes the sample a minute cloud of vapour is formed between the electrodes and causes them to discharge (the voltage having been preset to just below that for discharge). The light produced, the wavelengths of which characterize the chemical elements present in the sample, is directed into the slit of a conventional prism spectrograph, in our case the Zeiss Q24 UV Spectrograph and the spectrum produced recorded on a photographic plate. By a happy coincidence the range of diameter of crater produced, approximately 10–100µ, is about the same as the range of thickness of the layers found in cross-sections of samples from pictures. This makes it possible to analyse paint sections layer by layer or even to analyse single large pigment particles. By modifying operating conditions it is also possible to choose whether to analyse for principal constituents only or for minor impurities as well. The laser microspectral analyser was chosen in preference to the more usual type of emission spectrograph (in which the sample is inserted into one of the electrodes and is destroyed in the process of analysis) because we wanted to do further analysis on many of our large and irreplaceable collection of paint cross-sections from pictures without destroying their value for optical microscopy. The slight damage to the surface of the section caused by the laser beam can be removed by a little further polishing if required, but a neatly-formed crater is sometimes a useful record of the exact site of analysis. The equipment is completed by a spectrum projector for reading the resultant spectrographic plates. At present the instrument is used for qualitative and semi-quantitative analyses. Certain difficulties arise with quantitative analysis as compared with conventional emission spectrography. It is doubtful, however, to what extent quantitative analysis is applicable to the heterogeneous mixtures so often found in paint samples from old pictures and it is a problem we have not yet tackled. It should be pointed out that laser micro-spectrochemical analysis, like many other methods of analysis, indicates only the chemical elements present in the sample. To interpret these results in terms of chemical compounds or actual pigments present the method needs to be used in conjunction with optical and chemical microscopy and/or X-ray diffraction.