The scientific investigation of works of art has an essential role in understanding museum collections and is fundamental in establishing successful conservation and restoration strategies. In the multidisciplinary environment of museums, scientists work with conservators and curators not only to more profoundly understand works of art but to better preserve them, and this often involves using analytical techniques borrowed from different disciplines of Earth sciences (e.g. mineralogy, geochemistry, and petrology). Two case studies – the stones of Angkor (Cambodia) and a blue paint mineral pigment – demonstrate how the Earth sciences are helping to identify, determine provenance, and conserve a broad spectrum of works of art. The impact on archaeological and art historical scholarship is substantial
Of all the different objects that can be investigated in an art museum, stone artefacts are those most easily associated with, and directly linked to, the world of the Earth sciences, which should predictably spark the interest of geologists. Indeed, it didn't take too long for Earth scientists, particularly petrologists and mineralogists, to get involved in the study of stone artefacts of the human past, supporting archaeologists and art historians in provenance as well as in preservation studies (Clark and Merrill 1888). After almost 130 years, characterizing stone works of art in museum collections continues to benefit from the application of different techniques and methodologies borrowed (and often optimized) from Earth science subdisciplines: for example, Lazzarini and Marconi (2014) petrologically analyzed marble sculptures and Harlow et al. (2006) did provenance work on jadeite. But stone represents only a small fraction of the geomaterials that constitute the physical and tangible essence of the works of art that are collected, exhibited, and preserved in museums worldwide. The number and typology of art and archaeological materials with a geological origin are vast and range from unprocessed ores to highly engineered materials. These include simple mono- or polymineralic assemblages that have been used, for instance, as pigments; or mineral aggregates used for a variety of purposes, such as casting cores, investment molds, ceramics, or gypsum and lime plasters (Lombardi 2002). This also includes complex compounds, mixtures, and alloys that require an elaborate chain of physical and chemical processes in order to be successfully manufactured—Kakoulli (2009), for example, describes the technology behind the synthesis of the pigment Egyptian blue, and Rehren and Freestone (2015) review ancient glass production. Weathering and corrosion products—as well as later additions, substitutions, and restoration materials—can affect the original structure and composition of the studied objects, and these factors can complicate the situation further (Frame et al. 2013).
The scientific investigation of archaeological objects and works of art can, thus, be a challenging undertaking, which often requires combining analytical tools and protocols from different disciplines within the natural and physical sciences. For this reason, when scientific research and conservation departments are present within a museum's walls they tend to include a variety of staff members with different backgrounds, including chemistry, biology, physics, engineering, and Earth sciences.
Scientists contribute to art historical studies and cooperate with curators or visiting scholars most often when exhibitions are in preparation or when new acquisitions are made. They do so by identifying constituent materials and manufacturing techniques, by researching issues of provenance and authenticity, and by helping with conservation studies. They also support conservators in short- and long-term conservation projects of museum collections by characterizing the products and mechanisms of weathering and by evaluating treatments. Museum scientists also act as principle investigators in research projects jointly funded by museums and by national and international agencies, such as the National Science Foundation (in the USA) or the European Research Council. Such investigations—which are usually collaborative and involve several institutions, such as other museums or universities—focus either on the development of analytical techniques that are tailored to museum studies or on in-depth characterization of specific materials, artistic productions, or degradation processes.
Analytical challenges in museum scientific studies are also related to the difficulty of collecting representative samples from unique and delicate works of art without permanently damaging them. Allowed samples size is, in this field, considerably smaller than that which Earth scientists are normally used to, rarely exceeding a few millimeters and frequently much less. Most of the time, sampling is just not permitted.
These technical constrains are forcing scientific research in the field of cultural heritage in two main directions. On the one hand, museum scientists can rely on ultrasensitive analytical techniques that require a minimal amount of material, such as surface-enhanced Raman spectroscopy (SERS) in different sampling configurations (Londero et al. 2013), or high-brightness microbeam sources, such as synchrotron radiation beamlines (Bertrand et al. 2012). On the other hand, it is possible to circumvent sample collection by turning to noninvasive analytical techniques: these include certain types of X-ray fluorescence or ultraviolet–visible–infrared spectrometers, which can now be portable (Hunt and Speakman 2015); or, when possible, by placing small objects into variable pressure scanning electron microscopes and/or X-ray microdiffraction units (Frantz et al. 2009).
The ease of operation and affordable cost of portable, noninvasive instruments are making them a must-have in museum laboratories, even though analytical accuracy and precision are often limited by the instruments' geometry and by the surface characteristics of the object under examination. Given the wide range of materials and techniques involved, it is not possible in this article to give an exhaustive review of the contribution of Earth sciences to museum studies. Instead, by presenting two selected case studies, we highlight the intimate connection between geological materials and works of art, and we illustrate how it is possible to improve the understanding of our cultural heritage by applying and adapting analytical approaches and tools typical of the Earth sciences. In the first case study, basic petrographic and geochemical techniques proved essential in reconstructing the building and sculptural traditions of the greatest civilization of Southeast Asia. In the second, petrologic studies are helping to identify the complex processes behind the production of a unique synthetic blue pigment and to understand the mechanisms of its degradation.
UNDERSTANDING SCULPTURAL TRADITIONS IN SOUTHEAST ASIA: THE CONTRIBUTION OF THE EARTH SCIENCES
Research on the archaeological heritage of Cambodia, and of mainland Southeast Asia generally, exemplifies the synergy between the various fields of the Earth sciences and our cultural heritage. Such research is symptomatic of how the Earth sciences have contributed to the study and preservation of archaeological objects and works of art. Since the beginning of the first explorations of the Angkor region of Cambodia by western expeditions in the mid-nineteenth century, geologists were participating at the side of archaeologists, geographers, and botanists, and were describing the various types of stone employed in the local sculptures and buildings, suggesting their possible origin within the Cambodian landscape (Garnier et al. 1873). Technical studies of Khmer stone materials, by means of standard petrographic analysis, followed the arrival of the first samples of Cambodian stones to Europe in the first half of the twentieth century (Fromaget and Bonelli 1932). Somewhat later, these studies expanded to include thorough petrographic, chemical, physical, and mechanical characterization of temple and sculpture stone materials (Delvert 1963). The involvement of Earth scientists at an early stage in the archaeological research of Southeast Asia is indicative of the importance of stone in the history of these ancient societies, primarily as a medium to create the sacred buildings and the images of gods that are installed in them. In addition, stones were also vehicles to the kingdoms' development, as material for roads, bridges and other public works. Evidence of the major use of stone dates from the early kingdoms that settled around the lower and middle Mekong River (fifth to ninth centuries) and continued during the rise of the centralized Khmer Empire and of its main capital, Angkor, north of Tonle Sap Lake (ninth to fifteenth centuries). This unique heritage carved in stone represents a precious source of information about the civilization that created it, and the Earth sciences are actively contributing, in conjunction with other disciplines, to its understanding and preservation.
Analyses of the Cambodian stone materials have double purpose. First, scientific analyses are used to identify which stone materials were used and where they were quarried and distributed: this was to ultimately understand what the different material choices can tell us about social, cultural, and economic aspects of Khmer kingdoms. Second, information on petrophysical characteristics is essential for the preservation of stone artefacts, particularly in the harsh tropical climate of Southeast Asia, as indicated by the numerous research papers published on this topic (e.g. André et al. 2011).
The variety of lithotypes used by the Khmers might look relatively limited when compared to other ancient civilizations, such as those in Egypt or in south Asia. Khmer sculptures are carved almost exclusively from sedimentary rocks, mainly sandstone; sedimentary rocks also predominate in architectural production. The first petrographic and chemical analyses of Khmer sandstones identified three main lithologies: (1) a feldspathic arenite, referred to as green or grey sandstone; (2) a feldspatho-lithic arenite, also called a green or grey sandstone, or sometimes a greywacke; (3) a quartz arenite, often referred to as a pink or red sandstone (Delvert 1963). This classification, although approximate, includes the main lithologies used by the Khmers, particularly in the Angkor region, and it is still widely used as a reference in more recent preservation studies (André et al. 2011).
The nature and provenance of the feldspathic arenite are well established, thanks to several studies that have been published in scientific journals since the early 1990s (Carò and Im 2012). This sandstone is consistent with the upper portion of the Lower–Middle Jurassic Terrain Rouge Formation of Cambodia. Clear traces of its organized exploitation are scattered all along the eastern edge of the sacred Kulen Mountain, north of Angkor, and form the main quarrying district that served the capital of the Khmer empire for at least four centuries. Available data suggest that this Mesozoic sandstone was one of the favorite building materials of mainland Southeast Asia, together with the quartz arenites that belong to the younger Cretaceous formations (Uchida et al. 2010). Hard, fine-grained quartz arenites were also favored for the carving of intricate architectural decorative elements, such as lintels and colonnettes. Furthermore, it is becoming evident that, since the early Khmer kingdoms, and at least until the thirteenth century, sculptures of divinities were typically carved from sandstones different from those used for architectural and decorative elements: it is almost as if certain aesthetic attributes, such as color, texture, or capacity to attain a polish, were reserved for the gods installed inside the shrines. Systematic petrographic and geochemical characterization of sculptural and architectural materials is helping to prove that such distinction was deliberate, and that different sandstones were selectively employed within the same temples according to their destined use.
An example of a lithotype reserved for sculptures, and specific to a royal production (the reign of Jayavarman VII, who ruled 1182/3–circa 1218), is an immature sandstone of Triassic age exposed in central Cambodia. This stone was recently characterized as being markedly lithovolcanic, i.e. quartz-poor and extremely rich in microlithic volcanic fragments, albitized feldspar, and hornblende grains (Fig. 1). The narrow compositional variation of the studied sculptures carved under Jayavarman VII's reign strengthens the hypothesis that a localized source of stone was deliberately exploited for the production of royal sculptures (Carò and Douglas 2013). Such findings, besides enriching our knowledge of sculptural practices, have direct implications on authenticity studies, as discrepancies in well-established and characterized material trends are used to unmask later productions mimicking earlier iconographies. This knowledge is also particularly useful in case of sculptures that arrived in museum collections without a detailed provenance history.
Thanks to the development of fully portable instruments, recent research on Southeast Asian stone is taking advantage of on-site noninvasive analytical techniques that allow the analysis of a large number of objects in provincial museums and from archaeological sites without the need for sampling. For instance, the combination of ultraviolet–visible–near infrared reflectance spectrometry and XRF spectrometry has been successfully used by a team of archaeologists and geologists, to characterize Ayutthayan sandstone sculptures (fifteenth to sixteenth centuries), providing new insights into the cultural exchanges between medieval Cambodia and Thailand (Polkinghorne et al. 2013).
Noninvasive surface magnetic susceptibility measurements, which can be performed with lightweight and inexpensive contact probes, have been extensively used to characterize the sandstone blocks of several Khmer temples and to deduce temple construction history (Uchida et al. 2014). Sandstone's magnetic behavior, which varies according to the amount of ferrimagnetic minerals (predominantly magnetite) present within the probed volume of rock, has also been used to pinpoint the geographical provenance of building stones and to infer the location and distribution of the quarries of origin. In these cases, this magnetic data should be supported and validated by complementary geological, mineralogical, and geochemical information in order to build strong, and statistically meaningful, conclusions.
COBALT BLUE PIGMENTS: PROVENANCE AND CONSERVATION
The availability of coloring pigments has played, and continues to play, a key role in the cultural and artistic development of our society. In prehistoric times, the only mineral pigments known were the so-called earth pigments, mostly iron and manganese oxides and hydroxides (i.e. sienna, umber, and ocher), clay minerals (e.g. green earth and white clay), and chalk. These mineral pigments provided a limited palette of colors that ranged from light yellow to brown and black, depending on the relative proportion of the mineral constituents. The color blue was not available to prehistoric humans.
In antiquity, the color palette was expanded to include a variety of yellows, reds, greens, and blues, in connection with the development of mining techniques and ore processing, which increased the availability of colored rocks and minerals, both natural and synthetic colors. The blue pigment ultramarine was, for instance, derived from the painstaking refinement of lapis lazuli rock, while smalt blue was obtained by grinding cobalt-doped glass to an ultrafine powder (Fig. 2C–D). Cobalt itself is a fascinating chromophore. The only way to obtain a cobalt-based blue pigment, at least until the introduction in 1802 of Thénard's cobalt blue (a synthetic cobalt aluminum oxide) was through the use of Co-doped silicate glass, which allowed Co2+ ions to be stabilized by partly filling tetrahedral sites (Colomban 2013; Delamare 2013). In museum objects, cobalt can be detected as a colorant in glasses, ceramic glazes, paintings (as smalt blue), and faience.
Cobalt occurs in nature as Co2+ and Co3+, mostly in arsenates, arsenides, sulfarsenides (Fig. 2A–B), and in minor amounts in carbonates and sulfides. None of these ore minerals are blue in color, and all are usually subordinate to one or more abundant associated gangue minerals. The different geological processes driving crystallization of Co-bearing minerals influence the overall chemical composition of the ores, which, in turn, influences the processed pigment. Cobalt ores form in several ways, geologically: by magmatic differentiation (from low or very low silica magmas), which concentrates cobalt, nickel, and chromium; by hydrothermal activity (fissure fillings or replacement of the country rock), which is typically related to silver mineralization, with associated cobalt and nickel; and by chemical weathering, particularly lateritization, that concentrates cobalt, iron, nickel, and manganese. This variability, along with the contamination by country rocks surrounding the ore veins (e.g. the gangue) control the presence of other metals such as indium and bismuth, and these can individually characterize different cobalt ores. Ideally, trace elements can thus be used to identify the original cobalt source rock, their ratios in Co-based pigments may be considered as a fingerprint of the ore provenance. An example is the Freiberg district, in the Erzegebirge metallogenetic province of Germany, where indium enrichment in silver veins, associated with cobalt ores, acts as a unique geochemical fingerprint (Seifert and Sandmann 2006) and can be used by researchers to determine the source of cobalt occurring in a variety of glass and glaze productions (Gratuze et al. 1996; Porter 1997; Delamare 2013). In contrast, the cobalt ores of cobaltite and erythrite from Qamsar (Iran), an area known as an historical source of cobalt, formed in magnetite veins resulting from hydrothermal mineralization due to fluids circulating in skarn and igneous rocks (Mohammaddoost et al. 2010). Consequently, a particular association of cobalt, arsenic, and iron might be expected in blue glassy artefacts from this region.
Transformations occurring during the roasting and processing of mineral ores cause further variations in the relative concentrations of characteristic chemical elements. In fact, after mining, the cobalt ores were first separated from impurities, then roasted. The calcined ore was then mixed with the raw materials used for glass-making (quartz and plant ashes). The addition of such components contributes to the bulk composition and will have its own geochemistry. For these reasons, directly correlating artefacts colored with a cobalt blue pigment with their potential original sources has been challenging: until now, the provenance of a cobalt pigment has mainly been inferred by studying the elemental compositions of the pigments and comparing them with literature data rather than with potential cobalt ores (Porter 1997; Colomban 2013). Several authors have speculated on the origin of the cobalt minerals (e.g. Dayton 1993; Gratuze et al. 1996; Porter 1997), and Delamare (2013) has summarized the compositional groups and the attributed provenances reported thus far. These groups are based on trace element compositions, which may strongly characterize a single object. To our knowledge, the only paper dealing with the possible composition of any original cobalt-bearing raw materials is by Shortland et al. (2006). These authors suggest a source from the cobaltiferous alums of the Western Oasis (between the Dakhla and Kharga oases) in Egypt, interpreting the association of Co–Ni–Zn–Mg–Al as the fingerprint of the Co-alum deposits from the eastern Sahara desert. The deposits were sampled thanks to the collaboration with the Egyptian Geological Survey, and were analyzed by means of geochemical techniques, showing the compositional variability of each alum deposit. This work also demonstrates how a systematic geological survey is a prerequisite when trying to provenance ancient raw material sources.
The study of cobalt ores and their gangue materials can shed light on the relation between the original raw materials and the pigment. A project currently underway by Elena Basso is focusing on characterization of cobalt ores that were possibly used for cobalt blue productions between the twelfth and sixteenth centuries. On the basis of literature data and ancient documents, ore and rock gangue specimens from Central Europe were specifically selected from different museum collections for analysis. In particular, geological provinces where medieval mines were located were chosen for sampling: Erzgebirge, Rammelsberg, Schneeberg, Saxony, St. Andreasberg, and Harz (Germany); Zinkenwand (Austria); Bohemia (Czech Republic); and Tunaberg (Sweden). Rocks from these regions, such as granites, pegmatites, schists, and metamorphosed shales, were characterized as possible gangue materials surrounding the cobalt ores exploited in the past. Cobalt minerals from some of the key regions were also analyzed: safflorite (Bohemia, Czech Republic), alloclasite (Carasseverin, Romania), erythrite and skutterudite (Saxony, Germany), cobaltite (Sodermanland, Sweden), linnaeite (North Rhine-Westphalia, Germany). Alongside rocks and ores, some smalt samples were taken from important works of art dating to the fifteenth and sixteenth centuries, in order to compare their chemical composition with those of the geological materials above mentioned.
The analytical approach used is traditional for the Earth sciences: a combination of classical mineralogical and petrographic analysis with more innovative geochemical techniques, such as laser ablation inductively coupled plasma mass spectrometry. Once published, the data will contribute to a reference database containing the possible fingerprint elements that could trace the provenance of the cobalt used in pigments.
In the study of cobalt blue pigments, Earth science knowledge and techniques are essential to formulate correct provenance attributions. Furthermore, the innovative analytical techniques routinely used in the study of natural crystalline and noncrystalline solids have been recently applied to investigate pigment degradation. The composition and structure of smalt blue, for example, are affected by degradation, something that has been known since the seventeenth century. Discoloration of this pigment is due to the formation of soaps as a consequence of the reaction between leached potassium from blue glass particles and the fatty acids in the oil-binding medium. This phenomenon has been only partially explained. Nevertheless, Robinet et al. (2011), using synchrotron-based micro-X-ray absorption spectroscopy, have shed light on the mechanism by which Co2+ cations transition from a tetrahedral to an octahedral coordination with oxygen, and, as a consequence, cause the smalt to discolor.
The relationship between works of art and geological materials is strong, intimate, physical, and sometimes concealed behind the esthetic value of the objects. Earth scientists are in a position to generate original and valuable information about materials and technologies of archaeological objects and works of art thanks to their understanding of geological processes, including the genesis, distribution, and properties of rocks, minerals, and ores from which the raw materials originate. Earth scientists are also well versed in the principles of appropriate analytical techniques and, in particular, of the significance of analytical results collected on geomaterials.
The scientific study of museum collections is increasingly growing in number and quality, and, in this multidisciplinary context, the Earth sciences play an important role in understanding and preserving our cultural heritage.
The authors are in debt to two anonymous reviewers and to the editors for their valuable comments and suggestions. The research into the Khmer stone materials was made possible by the cooperation of several institutions, among which are the Ministry of Culture and Fine Arts, Cambodia, the National Museum of Cambodia, the Authority for the Protection and Management of Angkor and the Region of Siem Reap (APSARA), the École Française d'Extrême-Orient (EFEO), the Andrew W. Mellon Foundation, the Forbes Foundation, the Metropolitan Museum of Art, and the Freer Gallery of Art/Arthur M. Sackler Gallery of the Smithsonian Institution, and by the passionate cooperation of numerous scholars, archaeologists, and researchers.
The authors would also like to express their gratitude to the “Fondazione Arvedi-Buschini,” which sustained the research on cobalt blue, and Barbara Berrie, Head of the Scientific Research Department at the National Gallery of Art of Washington DC (USA), for the fruitful exchange of ideas about research into cobalt blue pigments. We are also grateful to the American Museum of Natural History in New York, and in particular to George E. Harlow (Curator, Department of Earth and Planetary Science), Jamie Newman (Senior Scientific Assistant), and Beth Goldoff (Senior Scientific Assistant, Mineral Deposits Collections) for their help in selecting and providing specimens from the Erzgebirge mines.