Handheld XRF for Art and Archaeology

4. Handheld XRF for the examination of paintings: proper use and limitations

Chris McGlinchey 132 the excitation spectra) data will potentially be collected from all paint layers simultaneously. For all these reasons, the effort to report findings in paintings quantitatively should never be attempted. A telltale sign of poorly understood data is when the painting’s identified elements are presented with concentrations that add up to 100%; giving a false sense of completeness and demonstrating a misunderstanding of what X-ray analysis is capable of when examining complex layered structures of potentially organic rich paint. There are numerous questions XRF can help address regarding artists’ pigments. They range from the straight-forward, identifying elements which can lead to pigment identification : is the white pigment titanium dioxide (TiO2 ), zinc white (zinc oxide, ZnO), or white lead (lead carbonate hydroxide, 2PbCO3 . Pb(OH)2 ); is the isolated red pigment iron earth (iron oxide, Fe2 O3 ), cadmium red (cadmium sulpho-selenide, Cd(S,Se)), vermilion (mercuric sulphide,HgS), or red lead Pb3 O4 ; to the more complex: can the pigment creating the faint blue tint in the sky be identified; does the barium identified pigment also contain titanium; is there a little orpiment (arsenic sulphide, As2 S3 ) pigment in a lead white and vermilion sample? The more straightforward type of inquiries may only require examination of a single representative site for a brief period of time. The latter type of investigations, which deal with pigments that might be in low concentration, fluoresce weakly, or have overlapping lines, will require collection of data from multiple sites, careful instrumental setup, standardized collection conditions, and will involve post-acquisition processing and statistical treatment of data. XRF is frequently a good screening tool for the examination of paintings and may be well suited for the identification of elements characteristic of pigments from certain periods. For example, the transition of lead-tin yellow to lead antimonate that occurred in the 18th century; the increased use of zinc oxide, which established prominence in the latter half of the 19th century; the emergence of pure titanium dioxide or selenium containing cadmium pigments, which became common in the early 20th century; to more modern inorganic pigments like molybdate red, which became available in the 1930s. Properly gathered data may help link a painting undergoing technical examination to one that has been previously examined. If an artist is known to have worked with different ground preparations or was active during a time when certain pigments were introduced the method can be exceptionally useful at demarcating which working period the painting belongs to (if it belongs to any at all). The method can also be useful for identification of restoration areas and, since the method does not require sampling or contact, it is a useful way to examine possible acquisitions where sampling restrictions are often imposed. Handheld XRF for the examination of paintings: proper use and limitations 133 Importance of raw data plots Some paintings can be as X-ray transparent as works on paper while others nearly as radio-opaque as some bronze sculptures. In fact, there is often a broad range of X-ray density within a single painting. That property produces informative X-radiographs with regions of contrast originating from areas of high and low X-ray density (Figure 4.1) and it also influences X-ray spectroscopy data. With the spectroscopic examination of a painting, an instrument capable of raw data output in the form of energy vs. X-ray detector counts is the most useful because the intensity of background scatter is revealed. Instruments that only output elemental data as peaks ‘auto-identified’ make certain assumptions about the background scatter and can yield false positives with meaningless quantified values. Figure 4.2 shows why looking at the raw data is important: only one of the three spectra illustrated contains both molybdenum and cadmium but an additional one (the green paint) could be falsely reported if its background intensity was assumed to be the same as the background in the spectrum for the ground preparation. Figure 4.1: X-radiograph detail of Mondrian’s Composition C (1920), 257.1948 The Museum of Modern Art. The strong contrast shows there is a significant variation in radio-opacity. Light areas are radio-opaque and black lines, rich in low Z carbon containing black pigment, are relatively transparent. Conditions: 35 kV 45 μA, 80 second exposure. One method that can be used to resolve this issue is to calculate the net area...