Where art meets science: Restoring masterpieces by Nelli, Rembrandt, Pollock and Van Gogh

Most of the art you see in galleries and museums has survived because of restoration and conservation work. This work is absolutely necessary because a work of art is subject to quite complex chemical, physical and mechanical processes that degrade the materials from which it is made. In this article, Dr Nick Gordon looks at the art of restoration, from the substantial intervention on Plautilla Nelli’s Last Supper, to insights into Jackson Pollock revealed by his Alchemy, the fading of Van Gogh’s Sunflowers, and some mysteries discovered in Rembrandt’s The Jewish Bride.

Restoration in action

When we think of restoration, one of the things that comes to mind is ‘cleaning’ of old layers of varnish that have darkened. It is true that some types of varnish darken over time: varnish is usually just a protective layer that absorbs soot, smoke, sweat and various particulate matter from the atmosphere.

Removing a layer of darkened varnish with a solvent requires a lot of care, so that you don’t damage or remove the layers of paint underneath it, or accidentally remove layers of varnish that contain pigments put there by the artist. To watch the process of cleaning, you can watch this work being cleaned by Chicago’s Julian Baumgartner Restoration.

For a more in-depth narration of a similar restoration, watch this short video produced by The Art Institute of Chicago about the restoration of Gustave Caillebotte’s Paris Street, rainy day.

But many of the agents that darken varnish also attack the paint layers underneath, making restoration a complex job. In addition, many pigments are vulnerable to light, heat, cold, humidity and acidity. This is to say nothing of the damage done to a painting, accidentally or deliberately, by moving it, attacking it with a knife, cutting bits off to make it fit somewhere else, throwing acid on it, storing it improperly, or by previous restorers. This may seem like an odd list, but each of these things has happened to Rembrandt’s The Night Watch. More on it later, but for a short documentary on restoring a work after deliberate damage you might be interested in watching this one from the Tate, which follows the conservation team as they repair a Mark Rothko vandalised by a visitor.

In some cases, a work has become so damaged that it requires significant intervention to make it worth looking at again. One such recently restored work is Plautilla Nelli’s Last Supper, in Florence’s Santa Maria Novella. You can watch this short video produced by the Advancing Women Artists Foundation on the restoration of Nelli’s masterpiece, which took four years and was mostly crowd-funded.

The difference between the before and after is extraordinary, but it required the addition of significant amounts of new paint. The general rule of thumb among restorers and conservators is first ‘do no further damage’, and secondly to make sure that a viewer can tell the difference between ‘original’ and ‘restored’ sections.

Before and after – a four year restoration on Plautilla Nelli’s The Last Supper on permanent display in Florence.

But in works such as Nelli’s Last Supper, it is quite difficult to tell what is new and what is not. Future restorers will be able to tell the difference because of the materials used in the restoration, and with major restorations such as Nelli’s, the documentation of the project. But for the general viewer, curators and restorers approach a complex philosophical conundrum, sometimes called the Ship of Theseus. In this thought experiment, Theseus’s ship has its nails replaced, then its rigging, then its sails, its caulking, its fittings and the timbers of its hull. At what point in this process is the vessel still the ship of Theseus?

The Nelli restoration project was quite clear about its position on this – addressing the social and historical factors that have made women artists almost invisible required the painting to appear closer to what it would have done in Nelli’s lifetime, rather than what it looked like at the turn of the 21st century. The most powerful way of renovating the artist’s reputation is by seeing what her worked like.

For the most part, restorers don’t make such significant interventions on the work. The common practice is to make the differences between the ‘original’ and the ‘restored’ sections clear. In the case of ancient vases, frescoes and mosaics this can be quite straightforward: the new parts are a different colour. For a look at the sorts of damage that may happen to antiquities and how restorers approach their work, you might like this short video on the conservation of a Greek Kylix by the Toledo Museum of Art.

You could also watch this longer video on the restoration of antiquities at the Art Institute of Chicago.

The aim and practice of restoring Old Masters is well summed up by Valeria Merlini and Daniela Storti, two of Rome’s leading restorers who have recently restored Caravaggio’s works in situ. The restorer’s role is to protect the work and to help bring you as close to what the artist saw as possible.

The research behind the scenes

Watching restorers at work can be mesmerising – their skill and precision while working is matched by a calmness and confidence that few of us could conjure while applying solvent to a masterpiece.

But behind the restorer is an extensive team of researchers.

Conservational research is required because artists have used a huge range of materials and exactly which materials were used is not always known at the start of a project. Different materials behave differently over time – some fade in light, some react badly to different chemicals, some go brittle, some exude metal salts. To make it more complicated, artists might use very different combinations of materials to make similar looking colours and these materials also react differently with each other over time. Knowing what materials have been used is one of the first steps in figuring out exactly what work needs to be done.

Identifying the materials in a painting is a tricky business – the most sophisticated chemical analysis available cannot yet identify some of the media used by Jan van Eyck, Rembrandt or Turner. Identifying pigments might seem more straight-forward, but colours are made from different combinations of chemicals whose identity is often only revealed through the use of optical and scanning-electron microscopy. For a brief introduction, you might take a look at how the science department of the National Gallery, London identifies what was used in purple-looking paint in Renaissance paintings, and the problems associated with these different pigments.

To take a closer look at the different sorts of scanning processes conservators use, you might like this video on the restoration of Van Gogh’s Bedroom from the Art Institute of Chicago.

But pigments and media are only a small portion of what a painting is materially speaking. Recently, the Smithsonian took on the research for a very particular problem: a large number of mid-20th century paintings in major collections have started to crack badly. In some cases, the paint is delaminating (quite literally falling off the canvas). Traditional approaches based on chemistry hadn’t revealed enough, so they turned to mechanical engineers. How do different materials expand and contract with heat and humidity? Exactly what sort of stresses and strains do very thin layers of paint endure over time? Which materials in the support layers (canvas, rabbit skin glue, acrylic or oil-based primers) help protect the paint from mechanical strain? The chief culprits were zinc oxide and titanium dioxide, white pigments introduced to replace lead white (basic lead carbonate). You can read their results here: insider.si.edu/researcher-solves-puzzle-of-cracks-in-mid-century-art/

Jackson Pollock’s Alchemy

The teams of people behind the restorers are extraordinary and they can also be very large. The restoration of Jackson Pollock’s Alchemy, undertaken in Florence’s Opificio delle Pietre Dure, involved a team of about 50 people. Their task was especially complex: no-one had been asked to restore old house paint on a tablecloth before.

But the job was considered essential – Alchemy was Pollock’s first drip painting. It had made a splash when Peggy showed it off during the Venice Biennale, consolidating Pollock’s reputation and helping Europeans see America as a country with a culture (beyond tanks, planes and the Marshall Plan). It went on to have a profound influence on post-war painting on both sides of the Atlantic. But Alchemy had been exposed to 50 years of cigar smoke, sweat, fine particulate matter and god knows what else inside Peggy Guggenheim’s palace on the Grand Canal. When confronted by the brownish-grey painting, it was hard to see what the fuss had been about.

Before and after – Jackson Pollock’s Alchemy (1947), a celebrated icon of postwar abstraction that Peggy Guggenheim acquired through her financial support of the artist

The results of the restoration were spectacular: it was colourful, with 19 different sorts of paint including bright yellows, blues and reds. But the significance of the project extends beyond making the painting look like it used to. The restoration included analysis of its composition and how each area of paint was applied. There is a clear, rational plan underpinning the painting and Pollock seems to have been following a process. He was not purely spontaneous or random in his approach to his art, as attractive as that idea may be. The New York Times compared his process to that of a conductor leading an orchestra. You can read more about the painting and its restoration here www.conceptualfinearts.com or take a closer look at the painting, which underwent state-of-the-art 3D scanning by Italy’s Visual Computing Lab.

The case of the fading Van Gogh’s

But sophisticated scanning technology doesn’t necessarily produce instant knowledge. Results require interpretation, cross-referencing with results from others types of scanning, and often further research conducted by scientists, curators and art historians. A recent example of the type of problem that can emerge in scanning is in the research into exactly why some of the yellows in Van Gogh’s Sunflowers and other paintings, especially from his Arles period, are losing their colour and vibrancy. One of the initial suggestions was that he had mixed lead white with his chrome yellow, leading the Guardian to declare “Van Gogh doomed his Sunflowers by adding white pigments to yellow paints”. It’s as if the artist had wanted his work to deteriorate.

Van Gogh’s Sunflowers are among his most famous, painted in Arles, in the south of France, in 1888 and 1889.

Lead white – basic lead carbonate (2PbCO3·Pb(OH)2) – is, however, quite a stable material and has been used for a couple of thousand years. (Indeed, it is manufactured today in more or less the same way as is described in the late 14th century by Cennino Cennini – it involves a coil of lead plate and a bucket of animal dung mixed with vinegar.) That it could cause a chemical reaction with chrome yellow – lead chromate (PbCro4), a pigment first manufactured in 1816 – raised some eyebrows, so off went some Van Gogh’s for scanning.

Scanning-electron microscopy is usually used to ‘read’ chemical signatures from highly localised distribution of energy following the material’s bombardment with X-rays. But the presence of an element in one compound may mask the presence of that element in a different compound. This seems to have led to some complications in the research into the fading Van Goghs, with a general agreement that it had something to do with the light sensitivity of lead chromate and the presence of sulphur.

In 2019, new research by Muriel Geldof, Inez Dorothé van der Werf and Ralph Haswell used a different method. Rather than scanning for the distribution of energy as X-rays bounced off different elements, they scanned for changes in wavelength. The results were somewhat curious – there seemed to be more sulphur in some sections than there would normally be.

But their results also allowed them to determine that there were at least three and possibly four different types of “Chrome yellow” paint used in Van Gogh’s paintings from Arles: a “lemon” version (PbCr1xSxO4), a medium chrome yellow (PbCr1), and a Chrome yellow-orange (Pb2CrO5). In his letters to his brother, Theo, Van Gogh requested he buy these colours (‘Chrome 1’, ‘Chrome 2’ and ‘Chrome 3’) from his supplier in Paris.

They also found, however, higher concentrations of barium sulphate than is typically found in Van Gogh’s paintings. Van Gogh’s letters let us know that he was not impressed by the paints he could buy locally in Arles (thus he asked his brother to get them for him in Paris). Cheaper or less reputable paints often contain ‘binders’ and ‘extenders’ – things that give the impression of a lot of paint in the tube, but with less of the expensive essentials. One of the most common extenders was barium sulphate, which, when exposed to light, degrades quickly. The ions it releases during its degradation accelerate the degradation of lead chromate compounds, especially the paler sulphurous version.

It seems quite possible that Van Gogh had used a cheap paint – such as those from Arles that he complained about to his brother – or was extending his own paint with barium sulphate and calcium carbonate, both of which react badly with lead-chromates. The next step in the research process is to figure out how to arrest the speed of degradation. In the meantime, the lights have been dimmed in that room of the museum.

Rembrandt’s surprises

That Van Gogh may have used a cheaper paint in some paintings is only one of the recent surprises arising from technical analysis conducted for the purpose of restoration. Analysis of Rembrandt’s The Jewish Bride, for example, has revealed a number of surprises.

Rembrandt’s The Jewish Bride, painted around 1665‒1669, now held in the Rijksmuseum, Amsterdam

Firstly, the colour of some regions of the painting has changed significantly over time. This is because Rembrandt had used smalt as an ingredient in his glazes. Smalt, which is made by firing potassium glass with cobalt ore and grinding it into a powder, was one of the most common substances to make greens and blues before the 19th century. It is, however, vulnerable to light and consequently tends to fade or become transparent. The background of The Jewish Bride contains substantial traces of degraded smalt. What looks red-brown to us was probably a combination of deeper blues and browns, similar to the backgrounds in Rembrandt’s earlier work.

A darker, richer hued background also makes sense aesthetically – it would have amplified the contrast with the red and gold of Isaac and Rebecca in the foreground. But these regions too have revealed some interesting insights into Rembrandt and his times. Isaac’s sleeve, for example, has an exquisite range of colour and tone, constructed by a complex combination of thickly impastoed regions with very fine glazes of colour. In amongst these colours, van Loon et al discovered what appears to be an “artificial orpiment” (As2S3), a new colour in Rembrandt’s palette that in this painting helps him bridge the gap in colours between the lead-tin yellow (lead stannate – Pb2SnO4) and the rich red of cinnabar (mercuric sulphide – HgS). Naturally occurring orpiment is a relatively stable compound in a painting and is virtually indistinguishable from synthetically produced orpiment.

Unless, that is, you take a very close look through a microscope. Natural orpiment has dark lines inside caused by the intense and irregular refraction of light inside the crystal. When it is produced synthetically – by sublimation in the controlled conditions of a laboratory – orpiment crystals loose their irregularity. Through a microscope, Rembrandt’s orpiment looks to be the synthetic variety. The synthetic production of pigments, however, is traditionally thought to have commenced after Rembrandt’s lifetime (in 1706, to be precise), and artificial orpiment isn’t known to have been produced until the 1800s. Someone seems to have been doing it in the 1600s – and this suggests that his contemporaries’ knowledge of chemistry was more sophisticated than we thought.

A final thing hidden in The Jewish Bride was found by looking at the cross section of Rembrandt’s impasto. To make paint thicker – so it can be moulded – it has long been thought that artists such as Rembrandt and Velazquez added chalk (basic calcium carbonate) or marble dust. Rembrandt’s impasto, however, is built up slowly with layer after layer of lead white. This process is unusual, time consuming and does not allow for much spontaneity. Having built up an area using lead white, Rembrandt would then glaze over it to give it the desired colour.

Analysis of the lower layers – ie the ones hidden beneath later layers of paint – has revealed the presence of a rare compound: plumbonacrite (Pb5O(OH)2(CO3)3). This compound is rarely found in oil paintings because it doesn’t usually survive the acidity of oil-based media. Its presence could be explained if Rembrandt had mixed his lead carbonate with litharge – an oil media made from heated, refined linseed oil and lead monoxide. The lead monoxide reduced the acidity of the oil, locking in the plumbonacrite – perhaps a byproduct of a reaction between lead carbonate and lithagre, for someone to discover later one. Litharge is used by oil painters to reduce drying time; it seems Rembrandt didn’t like watching paint dry.

More on Rembrandt is sure to come. In 2018, the Night Watch restoration program began. For an introduction to the project, and what the conservators have discovered so far, you can watch this 30 minute video from the Rijksmuseum www.rijksmuseum.nl/en/live-updates-of-operation-night-watch or follow live updates on the Instagram TV account www.instagram.com/tv/CAK_EUSB_kG/

The restoration of one of the most magnificent paintings from the Dutch Golden Age is taking place inside a purpose built glass cube in the Rijksmuseum so that visitors can come and watch the restorers in action. It’s a big job  – the painting is just under 4 metres high, about 5 metres wide and weighs 337kgs. The painting was cut to fit a new location in 1715, slashed 5 times in 1911, stored in a salt mine near Maastrecht during WWII. It has also been restored several times before, most recently in 1975 after it was slashed 12 times in 1970. Luckily this last restoration included a strong protective layer: in 1991 a disgruntled visitor threw acid on the painting, but it barely got through the restorer’s protective coat of varnish.

Restoration commencing on Rembrandt’s masterpiece The Night Watch, in a purpose built glass cube in the Rijksmuseum

Dr Nick Gordon

Dr Nick Gordon is a cultural historian and artist, with over 10 years of experience leading tours to Europe. He has strong interests in art, history, philosophy and architecture, from the ancient world to the present. Nick is also a practicing painter and brings this passion to the visual arts. He holds a University Medal and PhD in history from the University of Sydney.


Sign up to our mailing list for tour updates and latest news.