14 mins
DOWN TO THE GROUND
While much attention has been paid to the varnish used by Antonio Stradivari, there has been comparatively little research into the initial preparation layers. Giacomo Fiocco presents the results of a study using innovative spectroscopic techniques to examine two of the master’s finest instruments
FIGURE 1 The 1690 ‘Toscano’ (left) and 1718 ‘San Lorenzo’ (right) violins by Antonio Stradivari. The sampling areas were selected under the tailpieces of the two violins.
It is widely recognised that both the aesthetic and the acoustic qualities of Antonio Stradivari’s masterpieces are influenced by various stages of the manufacturing process, with particular emphasis on the finishing treatments. Of these, the varnish layers and ground coats play a crucial role in shaping the vibro-mechanical and acoustic behaviour of violins. Consequently, scientific research in the field of musical instruments, both in the past and the present, has focused on characterising the materials used in the finishing processes and their distribution within the coating system.
While significant attention has been given to studying the varnishes, which enhance the instrument’s aesthetics and provide protection against moisture, damage and biological growth, there has been relatively less emphasis on the treatments applied before varnishing, such as inorganic fillers, sealer or preparation layers, which serve to fill the wood’s surface pores.
The use of a preparation is a very common procedure in Cremonese violin making, and is probably as customary today as it was in the classical period. Some studies have already highlighted the presence of protein-based materials in these layers; in particular, animal glue has been found on the wood surface of two cellos from the early 18th century, while casein has been detected on a violin dating back to 1730. Regarding Stradivari’s instruments, Jean-Philippe Échard conducted some studies at the Musée de la Musique in Paris, which detected only a layer of drying oil at the interface between the varnish and the wood. Others, such as Brigitte Brandmair, have identified the possible presence of proteinaceous materials on the wood surface through chemical staining and Gas Chromatography– Mass Spectrometry (GC–MS) analysis. In the last few years, the research group of the University of Pavia’s Arvedi Laboratory of Non-Invasive Diagnostics, housed in the Museo del Violino in Cremona, have provided additional evidence of applied treatments beneath the varnish of Stradivari’s instruments. This research utilised non-invasive Fourier Transform Infra-Red (FTIR) spectroscopy, which further confirms the use of proteinaceous materials in most of the instruments analysed over the last decade and preserved at the Museo del Violino. Despite this scientific evidence, there is an ongoing and in-depth debate among experts regarding the organic materials used in this wood preparation.
The Arvedi Laboratory research team has now collaborated with the Elettra Sincrotrone Trieste facility in Italy to conduct a new study using Synchrotron Radiation (SR). The aim was to obtain further robust evidence regarding the materials used by Stradivari in the construction of his violins, through the application of highly accurate instrumentation. Our results were published in the journal Spectrochimica Acta Part A in 2021 (Fiocco et al) and, in 2022, in Analytical Chemistry (Stani et al), a world-renowned reference in the field of analytical sciences. The analyses were conducted on two cross-sectioned samples taken from the top plates of Stradivari violins: the 1690 ‘Toscano’ and the 1718 ‘San Lorenzo’ (figure 1).
TECHNIQUES
The use of synchrotron radiation is particularly appropriate for the investigation of complex layered micro-structures, as it can dramatically improve both the lateral resolution of the resulting images, and the ‘signal-to-noise ratio’, i.e. the level of interfering
Preliminary observation of the ‘San Lorenzo’ violin through the stereomicroscope aimed at the selection of the sampling area
ALL PHOTOS COURTESY ARVEDI LABORATORY, CISRIC, UNIVERSITY OF PAVIA
THE AIM WAS TO OBTAIN ROBUST EVIDENCE REGARDING THE MATERIALS USED BY STRADIVARI IN THE CONSTRUCTION OF HIS VIOLINS
FIGURE 2 UV images of the cross-sectioned samples detached from the ‘Toscano’ (left) and the ‘San Lorenzo’ (right). The layers identified are highlighted by capital letters in red circles. The yellow grids mark the areas considered for s-SNOM analyses.
ALL PHOTOS COURTESY ARVEDI LABORATORY, CISRIC, UNIVERSITY OF PAVIA
THE S-SNOM TECHNIQUE HAS NEVER BEFORE BEEN USED TO STUDY THESE TYPES OF OBJECTS AND MATERIALS
background noise is greatly reduced. At the research centre in Trieste we applied two different analytical techniques, namely FTIR microscopy in reflection geometry, and the infra-red scattering type Scanning Near-field Optical Microscopy (s-SNOM). The reflection geometry avoids contact between the probe and the sample – contrary to other set-up conditions, where the pressure applied by the crystal can seriously alter the cross-sectioned sample. It could therefore be considered as micro-destructive for these very brittle samples.
Infra-red s-SNOM combines an Atomic Force Microscope (AFM) with an infra-red interferometer enabling the simultaneous collection of sample topography and the chemical information from the material making up the stratigraphic layers. It is worth noting that, differently from the SR-FTIR microscopy, this technique achieves a lateral resolution of a few tens of nanometres, comparable to the radius of the metallic tip used for scanning (around 20 nanometres). The s-SNOM technique has never before been used to study these types of objects and materials, and holds great potential due to its exceptional analysis resolution, which enables the isolation of very small areas for analysis and yields highly precise results.
Both violins have been carefully preserved, respectively at the Accademia di Santa Cecilia in Rome, and as a valuable part of the Munetsugu collection in Japan. Figure 1 also shows the two sampling areas (‘SL’ and ‘Tos’) from which the micro-samples were removed. To preserve such brittle samples and their multi-layer systems, in particular the sequence of layers, the interfaces and the thickness, they were embedded into small 1cm cubes of epoxy resin. Subsequently, the sample surface of interest was exposed and dry-polished with fine sandpapers.
RESULTS
Images of the instrument cross-sections taken under ultraviolet (UV) light (figure 2) reveal three distinct layers: a varnish layer (V), a ground layer (G), and the wood substrate (W). The coating system observed in these samples is extremely thin (just a few micrometres) and difficult to differentiate owing to material mixing at the interface between varnish, the ground coat and the wood. This complexity presents a challenging scenario for researchers exploring the secrets of the renowned Cremonese masters’ wood preparation.
To begin the analysis, we defined two ‘regions of interest’ (ROIs) on each cross-section, and investigated them using SR-FTIR microscopy in reflection mode. These regions encompassed the transition between the three layers, in particular the interface between the wood and the ground coat (G in figure 2) where the presence of a proteinaceous preparation was expected.
The results revealed the diagnostic bands of proteins, namely Amide I and Amide II. Chemical maps were generated, hinting at the potential distribution of proteins within the ground layer, penetrating deep into the first rows of wood cells (W in figure 2). Deviations in the shapes and proportions of these bands challenged the definitive detection of proteins, especially in the case of the ‘Toscano’ violin, and were likely influenced by spectral contributions from water, wood, and epoxy resin. For this reason, the application of a more precise and accurate technique such as Infra-red s-SNOM, with an analysis spot of a few tens of nanometres, has become mandatory.
As for these analyses, two areas designated as Tos-A and Tos-B (shown in figure 2 in yellow frames) were selected for the ‘Toscano’ sample, while SL-A and SL-B were chosen on the ‘San Lorenzo’ cross-section (figures 2c and 2d). The A-labelled areas (Tos-A and SL-A) were situated within the previously identified ROIs, characterised by the highest intensity in the chemical maps discussed above. The Tos-B and SL-B areas were selected at the interface between the B and C layers, guided by UV-light microscopy images.
The application of infrared s-SNOM provided unprecedented insights into the sample morphology, revealing nanoscopic details such as depressions, scratches, dispersed particles and holes. The measured areas of the ‘San Lorenzo’ sample exhibited deep hollows (blue regions in Figure 3) with an average width ranging from 6 to 9µm and a depth of around 1µm. These depressions likely originated during the initial steps of cross-section polishing using sandpapers with a particle size of around 9µm (e.g. 4000-grit), which aligns with the width of the observed ‘valleys’. The measured areas of the ‘Toscano’ sample (figures 3a and 3b) displayed surface depressions similar in extent and depth to those observed in the ‘San Lorenzo’ sample, despite the overall surface of the cross-section appearing smoother under the s-SNOM microscope. Notably, the ‘Toscano’ sample exhibited a consistent pattern of parallel scratches measuring approximately 2.5 to 1µm wide and 50 to 100nm deep, likely resulting from the use of finer grits (8000- grit with a particle size of 3µm) during the final polishing of the cross-section with micro-mesh abrasive sandpaper. This morphology of the cross-section surfaces is expected, as these samples are manually prepared, making perfect reproducibility at the nanometre scale unrealistic. However, it highlights the importance of surface details when employing a nanometrescale approach for the analysis of polished cross-sections.
The back of the ‘Toscano’ is made from beautifully figured maple
DEVIATIONS IN THE SHAPES AND PROPORTIONS OF THESE BANDS CHALLENGED THE DEFINITIVE DETECTION OF PROTEINS
THE ‘TOSCANO’ STRADIVARI
The 1690 ‘Toscano’ is the sole remaining violin of Antonio Stradivari’s so-called Medici Quintet, a project originally commissioned by the Marquis Bartolomeo Ariberti as a gift for Grand Prince Ferdinando de’ Medici. The significantly later 1718 ‘Medici’ violin would seem to have been a replacement for a lost second violin from the original set, although this is disputed. Like all the surviving instruments from that set, the violin’s back is made from beautifully figured maple while the top is made from fine-grained spruce. According to an inventory made in 1716 by Bartolomeo Cristofori, the violin might originally have been decorated: the fingerboard had ‘a carved decoration of mother-of-pearl reproducing the Medici’s coat of arms’. The original tailpiece has also been lost.
The violin remained part of the Medici collection until the death of the last Grand Duke, Gian Gastone de’ Medici, in 1737. His sister Anna Maria Luisa then bequeathed the Medicean art collection to the Tuscan state, with the proviso that each part of it must stay within the Grand Duchy. However, in 1794 the director of the Tuscan court, Giovanni Felice Mosell, sold the instrument to an Irishman, David Ker of Donaghadee. After Ker died in 1811, the violin was missing for more than 30 years, even though it remained somewhere in his house. The Hills report: ‘About the year 1845, Portavo House, the residence of David Ker (son of the original purchaser) was destroyed by fire, and the family took refuge at Bangor Castle, Mr Ward’s seat. Among the chattels rescued from the fire and conveyed to the Castle was found the longsought fiddle. It was enclosed in its original old leather-covered case, studded with brass nails and opening at one end; but was quite out of order, with no soundpost.’ It was then restored at the workshop of J.B. Vuillaume in Paris.
After going through a succession of owners, the ‘Toscano’ was bought by the Hills, who were effusive in their praise: ‘Of all the violins made in the period previous to 1704, the “Tuscan” represents most perfectly the greatness of his ability. The bold, original style, perfect technical workmanship, and splendid all-round tone, completely convince one for the first time of his great skill and originality.’ They also recalled showing the instrument to Joseph Joachim, who exclaimed: ‘How pure and brilliant! I was led to believe that the tone was unsatisfactory. How is it possible for anybody to have come to such a conclusion? Surely they could not have tried it.’
FIGURE 3 3D Atomic Force Microscope (AFM) images collected by s-SNOM on the areas Tos_A and Tos_B (a,b) of the ‘Toscano’ and SL_A and SL_B (c,d) of the ‘San Lorenzo’
Regarding the chemical information obtained from infrared s-SNOM, the nano-resolved measurements conducted on the ‘San Lorenzo’ sample unequivocally confirmed the presence of proteins within the stratigraphy. The spectra collected at the interface between ground layer (G) and wood (W) exhibited the Amide I and II bands within the coating layer, in close proximity to the first row of wood cells. This observation confirms the tendency of the treatment to penetrate deeply into the wood surface, as only partially suggested by SR-FTIR microscopy analysis. Furthermore, clear proteinaceous hotspots were detected in the A and B layers of the stratigraphy (as shown in figure 2). It is worth noting that the presence of this proteinaceous material, likely animal or casein glue, in the V layer suggests the possibility of particle displacement between layers, due to the polishing treatments. This could also open up avenues for investigating the unique construction strategies employed by the Cremonese masters, such as the potential use of proteins within the varnish layers. On this crucial point, we prefer to leave the floor to experienced luthiers!
Compared to the ‘San Lorenzo’ sample, the nano-resolved infrared s-SNOM spectra of the ‘Toscano’ violin exhibited more complex profiles. Since the surface of the ‘Toscano’ sample appeared smoother than the ‘San Lorenzo’ surface, spectral quality degradation due to surface roughness can be ruled out, pointing to chemical differences as the underlying reason. At the nanometre scale, the spectra collected from areas Tos-A and Tos-B (figure 3) show the presence of Amide I and II bands, although they are not well resolved, due to overlapping features. In particular, a signal predominantly observed in the spectra from the Tos-A area may indicate the initial formation of degradation products. Additionally, a specific signal mainly seen in the spectra from the Tos-B area could be attributed to the formation of dehydrated protein layers. It is important to note that these small samples may not represent the entire violin surface, and the differences observed between the two samples could suggest variations both in distribution, composition, and thickness of the protein layer, where it is applied on the wood surface or deeply penetrated into the first rows of wood cells.
In summary, the analysis conducted on the 1690 ‘Toscano’ and 1718 ‘San Lorenzo’ violins demonstrates the complexity of the layering structure and the materials used, which may have undergone degradation over time. This necessitates an approach that maximises detail while avoiding contamination in the analysis of surrounding materials. The micrometre-scale resolution offered by SR-FTIR microscopy presents limitations in this context, but these limitations are effectively overcome by infra-red s-SNOM. The data presented provides unprecedented chemical insights into Stradivari’s craftsmanship and undoubtedly encourages further application of this technique, which has been employed for the first time in cultural heritage field, and in particular on priceless musical instruments, for deeper investigations of challenging samples. While the analysis of only two samples is insufficient for generalising the construction techniques employed in the ‘Toscano’ and ‘San Lorenzo’ violins, the information obtained from these samples offers a new point for addressing the long-debated question regarding Stradivari’s use of proteinaceous materials.
The author thanks Lisa Vaccari, Chiaramaria Stani, and Giovanni Birarda of the SISSI – Chemical and Life Sciences branchline at Elettra Sincrotrone Trieste for sharing their experience and ideas during the project and the CERIC-ERIC Consortium for access to experimental facilities and financial support. This work has been partly performed in the framework of the Nano-science Foundry and Fine Analysis (NFFA-MIUR Italy) facility. Special thanks to the Museo del Violino, the Accademia of Santa Cecilia in Rome, Andrea Zanrè, Elisa Scrollavezza, and Gregg Alf. We also thank the Munetsugu Foundation and Sota Nakazawa (Nippon Violins), conservator of the 1718 ‘San Lorenzo’ violin.
THE ‘SAN LORENZO’ STRADIVARI
Made during the so-called ‘golden period’, the 1718 ‘San Lorenzo’ is one of the few Stradivari instruments that still retains some of its decoration. Only faint traces remain of the words ‘Gloria et divitiae’ on one side of the ribs, while on the other are the remains of the words ‘in domo eius’. Translated from Latin, this reads as ‘Wealth and riches shall be in his house,’ a quote from Psalm 112 verse 3. Duane Rosengard has speculated that the instrument was a wedding gift to its first recorded owner, the violinist Mauro D’Alay (1687–1757), who was married in October 1717, although there is no proof of this.
D’Alay kept the instrument his whole life, dying in 1757, and in his will (dates 1754) he left the violin to the Sacred Military Constantinian Order of Saint George, of which he was a Knight. The order sold the ‘San Lorenzo’ the following year, to Ferdinando Tondù-Mangani of Parma. (D’Alay, Tondù-Mangani and the Order were all based in that city.) Upon his death, his widow sold it to the Czech violinist–composer Václav Pichl.
After passing through several other hands, the violin was bought from the British dealer John Betts by Giovanni Battista Viotti. It was Viotti who, in 1823, sold the instrument to the man whose title it now bears: Lorenzo Fernández de Villavicencio Cañas y Portocarrero, third duke of San Lorenzo. After that, the violin’s history becomes obscure: Ernest Doring claimed the duke’s heir sold it to an antiques dealer in Madrid, whereas the Caressa & Francais sales ledger shows a 1903 sale by the Duke of San Lorenzo to a M. Sanz. The violin is recorded as having a Latin inscription, and was therefore probably the ‘San Lorenzo’.
The instrument is now part of the Munetsugu collection in Japan, and in 2018 was chosen to represent the Tokyo Stradivarius Festival in all its posters and publicity material. It was also featured in The Strad Calendar for 2019.