Special issue research article Received: 29 April 2014 Revised: 1 August 2014 Accepted: 19 September 2014 Published online in Wiley Online Library: 21 October 2014 (wileyonlinelibrary.com) DOI 10.1002/mrc.4164 Characterization of aging and solvent treatments of painted surfaces using single-sided NMR † Gwendoline R. Fife, a Bascha Stabik, a Allison E. Kelley, b Jared N. King, b Bernhard Blümich, c René Hoppenbrouwers a and Tyler Meldrum b * Typical experiments conducted with single-sided NMR are incapable of unique chemical identification and, thus, often rely on comparative measurements in scientific study. However, cultural heritage objects have unique natures and histories, making a genuine ‘control’ sample a rarity and complicating many scientific investigations. In this paper, we present some compar- ative results enabled by such a rare, control sample. Two paintings, The Dinner and The Dance from the 1616 set Pipenpoyse Wedding, were made by the same artist with indistinguishable materials and techniques. However, despite their shared his- tory, The Dinner has undergone varnishing and subsequent varnish removal multiple times, whereas The Dance has not. NMR measurements on these two paintings show the effect of organic-solvent-based treatments on the stiffness of the paintings as measured by T 2,eff , supporting visual and tactile observations that The Dinner is stiffer throughout its thickness than The Dance, probably due to ingress of natural resins and organic solvents into the paint and ground layers. In addition to a com- parative analysis of these two paintings, initial experiments to compare solvent penetration with different varnish removal methods are described. Model canvas painting samples were treated with solvent in two ways—with free solvent on a swab and with cellulose gel thickened solvent in a tissue. Both treatment methods cause a measurable change in T 2,eff ; however, the thickened-solvent method affects a narrower region of the model than does the free solvent. Copyright © 2014 John Wiley & Sons, Ltd. Keywords: NMR; 1 H; single-sided; low-field; cultural heritage; paintings; solvent treatment Introduction The development of single-sided NMR devices has changed our perception of which materials can be studied with NMR. Studies of oil wells, [1,2] foods, [3–5] and manufacturing processes [6,7] have benefited from NMR measurements made with devices of uncon- ventional geometry. Many objects of cultural heritage—including wall paintings, [8,9] musical instruments, [10] ceramics, [11,12] paper, [13,14] and building materials [15] —have been examined with single-sided NMR, giving insight into the composition, degra- dation, and condition of objects for which conventional NMR techniques and sample removal are inappropriate. Easel paintings have also been studied with single-sided NMR, primarily as subjects of NMR profilometry. [16] Because single-sided magnets are typically small and portable, they can be mounted to a mechanical lift, with which the entire mag- net assembly can be moved relative to a painting, acquiring data at different depths within the layer structure of a painting; the spatial resolution achievable with this approach is on the order of the thickness of ground and paint layers. [16,17] Using the Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence, [18,19] sig- nal intensities (A) and relaxation times (T 2,eff ) can be determined, both of which characterize the sample. Signal intensities are commonly used to produce a depth profile, indicating the num- ber of layers and the thickness of each, whereas T 2,eff suggests the stiffness of a paint layer—smaller values of T 2,eff indicate a material with stronger intermolecular couplings, suggestive of a stiffer material. [20] Owing to its dependence on experimental parameters, T 2,eff cannot provide absolute information on a given object. How- ever, T 2,eff can be used in comparative studies between a control and a test sample, or between before and after treatments on one sample, if experimental parameters are kept constant. Thus, comparative measurements of A and T 2,eff between two samples produce a data set that can be used to characterize some of the effects of treatments and other chemical processes that are relevant to cultural heritage and its conservation. In this paper, we present research into cultural heritage tech- niques and processes by way of comparative analyses on two different sample sets. The first is a pair of early 17th-century Dutch paintings, The Dinner and The Dance from the Pipenpoyse Wedding * Correspondence to: Tyler Meldrum, Department of Chemistry, The College of William and Mary, 540 Landrum Drive, Williamsburg, VA 23185, USA. E-mail: tkmeldrum@wm.edu † This article is published in Magnetic Resonance in Chemistry as a special issue on the NMR in Cultural Heritage by Donatella Capitani and Noemi Proietti (Istituto di Metodologie Chimiche, CNR Area della Ricerca di Roma, Monterotondo (Rome) Italy). a Stichting Restauratie Atelier Limburg, Avenue Ceramique 224, 6221 KX, Maastricht, the Netherlands b Department of Chemistry, The College of William and Mary, 540 Landrum Drive, Williamsburg, VA 23185, USA c Rheinisch-Westfälische Technische Hochschule, Institut für Technische und Makromolekulare Chemie, Worringerweg 1, 52074 Aachen, Germany 58 Magn. Reson. Chem. 2015, 53, 58–63 Copyright © 2014 John Wiley & Sons, Ltd.