1 INTRODUCTION

ANALYTICAL TECHNIQUES used in the characterization of the materials of painting are many. Those used in a particular case are dependent upon available instrumentation, time or financial considerations, amount of sample available (if sampling can be done), and so forth. If possible, a variety of techniques is often used to characterize the material, the particular combination being determined by the general type of material in question.

Infrared spectroscopy is most widely used with organic materials, since carbon-hydrogen, carbon-oxygen, and other types of bonds which are found in compounds of this category have fundamental vibration frequencies in the infrared region, particularly in the “mid-infrared” which extends from about 2.5 to 25 microns (4000 to 400 cm−1).∗ Among the difficulties often encountered when this instrumental technique is employed for pigment analyses are the presence of binding media which may mask or distort the characteristic absorptions of pigments (particularly if the latter are organic compounds), the presence of several pigments which may also mask absorptions or generally make assignments difficult, and sample sizes. Although beam condensors and sample preparation techniques for very small samples are widely available, the quality of spectra obtained from microsamples may often be poor, and weaker absorptions tend to become lost in the instrumental noise. It is often possible to make definite or fairly definite identifications of many painting materials on the basis of the general shapes and intensities of only a few major absorption bands, but in other cases an overall pattern of a large number of peaks is sought, and in these situations the difficulty in separating weaker absorptions from instrumental noise may be a serious problem. These are all factors which limit the use of the infrared technique in studying paint samples.

∗—The spectra in this paper are presented as a function of wavenumber (cm−1). In other publications, spectra and positions of absorption bands may be given in wavelengths (λ) rather than wavenumbers. The relationship between these two numbers is: cm−1 = (104/λ). The units for λ in this formula are microns (10−6 m).

However, infrared spectroscopy has been used successfully for a number of special problems involving painting materials, such as the identification of “copper resinates”1, 2 and organic lake pigments,3, 4, 5 and the characterization of binding media, including varnishes.6, 7, 8 Special note should be made of the pioneering studies by M.J.D. Low with N.S. Baer on the application of the Fourier transform infrared technique in conservation science.9, 10, 11 Publications thus far by these authors in this field have been concerned with “fingerprinting” natural resins and distinguishing between red lakes prepared on different substrates.

Applications to inorganic pigments have been fewer, perhaps because of the widespread availability of other techniques for these types of materials which in general require smaller sample sizes, including X-ray diffraction, and various elemental analytical techniques (emission spectrography, electron beam microprobe, etc.) But the potential of infrared spectroscopy for the study of certain materials in this class has been shown by C. Grissom in her study of green earths,12 and work has also been published on iron oxide-containing earth pigments.13, 14 There is no lack of reference spectra for materials in this class. High-quality spectra of modern inorganic pigments have been published by the paint industry.15, 16 The literature on the infrared study of minerals is voluminous,17 and much of value to the study of pigments is to be found in that literature since many pigments are minerals or synthetic versions thereof.

The purpose of this paper is to review several applications of infrared spectroscopy, particularly in the realm of inorganic pigments.18 The materials chosen for this study were ones for whose characterization infrared spectroscopy may be particularly valuable. Many of the materials studied were taken from reference collections in the Center for Conservation and Technical Studies; Fogg Art Museum.19 Although very large quantities of most of these were available, all samples were purposely kept small, generally on the order of the size of sample that would often be available from an art object, in order that the quality of spectra that may be expected from such samples is more readily evident. Of the infrared spectrometers available to the author, a Fourier transform instrument was found to give the highest quality spectra with the sample sizes here used. The theory of the operation of this type of spectrometer and details of the instrumentation have been well described elsewhere.9, 20