Research methodology

Proposed methodology includes well established spectroscopic methods within the Cultural Heritage field such as – Attenuated Total Reflection Fourier Transform Infrared Spectroscopy (ATR-FTIR) and X-ray fluorescence (XRF), as well as state-of-the-art techniques – Hyperspectral Imaging (HSI). This complementary approach can provide a large set of significant data and overcomes the intrinsic limitations of each specific method, thus gaining a high selectivity in the identification and mapping of materials. Moreover, the interplay between highly-specific spectroscopic methods and macro-imaging methods can offer exciting insights and information hitherto inaccessible on regard the production process, painting technique and style of the artist, while at the same time limiting as much as possible the sampling.

FTIR spectroscopy is a powerful molecular fingerprinting technique widely used for the investigation of cultural heritage materials that has the advantage of providing information on both organic and inorganic compounds. Due to its sensitivity and versatility in relation to the possibilities of investigating/measuring samples, FTIR has proved very effective within the conservation science field, giving accurate and reliable results even with extremely small amount of sample. Although a powerful technique, FTIR has its limitations when examining complex multilayer samples, or when various aged organic binders are present. Interpretation of FTIR spectra of such complex matrices is often hindered, broad overlapping absorption bands hiding key analytical information that could aid in material identification. To overcome such problems spectrum refinement methods can be used in order to achieve a signal enhancement, while application of multivariate analysis on IR data sets can highlight small chemical variations that further make it possible to detect or distinguish the presence of certain compounds. In terms of the capabilities to characterize and identify constituent materials in paint layers, another drawback of FTIR is given by the fact that a wide range of inorganic pigments don’t show any characteristic absorption bands within the mid-infrared region making thus impossible to detect certain groups of pigments.

X-ray fluorescence has been extensively used for paintings investigations as this technique can easily identify pigments (and other inorganic materials) by showing their elemental signature. Although ideal when analyzing metal-based pigments (lead white, realgar, cinnabar, chromium or copper pigments), XRF shows clear limitations when investigating materials composed of lighter elements, making the identification of common salts for example (carbonates, sulfates) impossible by XRF alone. Earth based pigments are also difficult to identify by using only their elemental composition, but the analysis of variation of the major, minor, and trace-element signature can indicate chemical patterns or fingerprints that are characteristic of a particular geological source. Trace element characterization can also be used to locate ancient sources, understand ancient technologies and map ancient trade routes. Another key information that can be extracted from XRF data is identification of impurities and so gain insight within the production and manufacturing process of the pigments. Based on the fact the X-rays can penetrate the object, XRF spectra will contain information within several paint layers. These facts, along with the layered structure of the investigated object/sample have to be taken into account as they can complicate the data evaluation.

Hyperspectral Imaging (HSI) is a non-destructive technique than can extract key information from paintings such as color, pigment composition, line features from existing underdrawings, damage characteristics or painting technique. The shortwave infrared region (SWIR), 1000 – 2500 nm, was found to facilitate the extraction of such features even using a single band image. Initially focused on the field of remote sensing, HSI has proven to have a great potential in cultural heritage applications. The imaging spectrum data acquired by SWIR hyperspectral systems have high spatial and spectral resolution being able to provide a deeper and more intuitive understanding of the materials and painting techniques used to create a painting. More specific, the spectrum information obtained can be used for pigment identification, while the image highlights the pigments’ spatial distribution within the painting, especially when employed together with complementary analytical techniques. Hidden or apparently lost information can also be highlighted – such as underneath inscriptions, illegible information or previous restoration treatments.

Combining multiple point analysis (FTIR/XRF) with macroscopic imaging (HSI), a variety of complementary information may be collected and integrated to obtain an in-depth assessment of the various problems and questions faced when investigating complex multilayered works of art. Giving the fact that each art object has specific characteristics and a unique history, a key goal of the research would be to identify and extract targeted information by using cost effective services while at the same time limit the amount of sampling. By integrating the various types of information that may be obtained with the proposed methodology, specific questions could be answered on regard the painting materials and techniques, and ultimately on regard the history of the object under investigation. A comprehensive examination and identification of painting materials can offer exciting insights on how artists and workshops used to create their artworks, how painted surfaces change/degrade over time, or how anachronistic use of materials can be associated with either fakes/forgeries, or past restorations.

Working plan

STAGE 1

Duration: 12 months

Specific objectives: (O1) Development of an integrated spectral library of artists’ and cultural heritage materials. (O2) Workflow for material identification and discrimination (with a focus on historical pigments)

Planned tasks:

Task 1.1 Documentation on historical painting techniques

Task 1.2 Preparation of reference paint samples. Data collection

Task 1.3 Development of integrated spectral database

Task 1.4 Workflow for material identification/discrimination

Estimated deliverables: (D1) Scientific and management report; (D2) Database; (D3) Technical study; (D4) International conference presentation (x1); (D5) ISI paper (x1); (D6) Project website

STAGE 2

Duration: 12 months

Specific objectives: (O3) Algorithm design for automatic classification and identification of painting materials using multivariate analysis. (O4) Development of integrated pigment-mapping techniques

Planned tasks:

Task 2.1 Algorithm design for rapid classification and identification of paint materials

Task 2.2 Testing and validation of developed algorithm on complex mixtures and artworks

Task 2.3 Development of semi-quantitative maps of pigment distributions

Task 2.4 Depth profiling and sub-surface imaging investigations of multilayered works of art

Estimated deliverables: (D7) Scientific and management report; (D8) Algorithm design; (D9) Portfolio of case studies; (D10) International conference presentation (x1); (D11) ISI paper (x1); (D12) Final report of the project