Motion picture film has always been a precarious medium. Early cellulose nitrate reels were highly flammable, and their successor – cellulose acetate “safety film” – proved chemically unstable, prone to the dreaded “vinegar syndrome.” Even polyester, introduced in the 1960s, is not immune to time. For archives and film libraries charged with safeguarding cinema’s legacy, knowing exactly what polymer base a film carries is critical, since the degradation of one reel can spread to others nearby.
In a study published in Vibrational Spectroscopy, researchers from the University of Milan and the Cineteca Nazionale in Rome systematically compared three non-invasive methods – external reflection FTIR in the mid- and near-IR ranges, and Raman spectroscopy – across more than a century’s worth of film samples. Each technique offered unique advantages: mid-IR for pinpointing cellulose-based supports, near-IR for penetrating coatings and quantifying acetylation in acetate films, and Raman for revealing additives such as plasticizers and azo dyes. Together, the results establish a versatile toolkit that can be deployed directly in archives using portable instruments.
To learn more about the motivations behind the research, the surprising findings, and the future of spectroscopy in cultural heritage science, we spoke with Silvia Bruni, Associate Professor of Analytical Chemistry at the University of Milan and senior author of the study.
What inspired this work – was it the analytical possibilities, specific challenges in film preservation, or perhaps an interest in film?
I would have to say all three. In fact, I have been a cinephile since I was a child when my parents had to spend an entire afternoon at the cinema (there were no DVDs back then) to allow me to watch Walt Disney cartoons two or three times in a row.
Today, I teach courses for the degree in Conservation Sciences for Cultural Heritage and one of my colleagues, Alessandro Rizzi, is a specialist in the digital restoration of cinematographic films. First we started studying the dyes in chromogenic films and then, through him and his collaborators, I met Valentina Rossetto of the Cineteca Nazionale di Roma (National Film Library), co-author of the present study. I then learned that the identification of the polymeric support of the films, so important for conservation purposes, is often not as immediate as one might expect – especially for those institutions that deal with the custody and conservation of the films.
This need could in fact be satisfied, as already suggested by previous work in the scientific literature, through spectroscopic techniques – particularly vibrational, which combine specificity and non-invasiveness, as well as the availability of portable instrumentation. Part of my activity in analytical chemistry for cultural heritage was already dedicated to the development of spectroscopy-based protocols for the “in situ” identification of materials, particularly organic ones such as binders in paintings or dyes in textiles. At the same time, infrared and Raman spectroscopy are among the most frequently used techniques in my laboratory.
What are the main reasons for film deterioration?
The main problems concern cellulose nitrate and acetate films. The latter in particular are subject to the so-called “vinegar syndrome”: the hydrolysis of cellulose acetate with the release of acetic acid – which causes the characteristic vinegar smell. The reaction is autocatalytic and is favored when the films are stored in warm, humid climates. Its effects are the shrinkage and deformation of the films, which can be permanently destroyed due to the separation of the emulsion from the base, and the fading of the dyes. Furthermore, the released acid can attack other films from the same collection. For this reason, conservator guidelines recommend storing the films at lower temperature and humidity and separating films for which vinegar is detected from the rest of the collection.
What motivated you to compare mid-infrared, near-infrared, and Raman spectroscopy?
Previous work in the scientific literature focused on one or another technique, and a subset of the polymer materials used historically for motion picture films. It is worth remembering that cellulose nitrate was initially used, then gradually replaced starting around 1950, due to its flammability, with the “safety film” made of cellulose acetate. Finally, when even acetate films proved to lack chemical and dimensional stability, polyethylene terephthalate began to be used in the 1960s. Interestingly, in some cases, other less common materials were also adopted: cellulose acetate butyrate for brands other than Kodak, for example, or cellophane for the Ozaphan films, first introduced in the 1920s in France and subsequently adopted in Germany until the 1950s for small formats such as 16 mm. So, the idea was to compare the performance of all non-invasive vibrational spectroscopic techniques (i.e., external reflection FTIR in both mid- and near-IR and Raman) for the identification and characterization of all these materials.
Were there any findings that genuinely surprised you – either in terms of technical performance or the condition of certain film types?
From the point of view of technical performance, I should say that the results obtained were truly paradigmatic and therefore rewarding for an analytical chemist engaged in the use of spectroscopic techniques. Each technique gave its best in terms of information, so it was possible to appreciate the different penetration depth for mid- and near-IR radiation, or the lower degree of distortion of the reflection spectra in the near-IR compared to those in the mid-IR, or again the greater sensitivity of Raman spectroscopy towards aromatic molecular structures.
In this way, for example, the cellulose nitrate surface coating of Ozaphan films (possibly present according to notes on the manufacturing technique) could be selectively detected in respect to the cellophane support, by comparing the spectral signals in the two ranges of the IR region. This is something we had already exploited for contemporary paintings, where for example we can thus distinguish a primer from the color layer.
Raman spectroscopy has confirmed itself as indicative especially for additives used in plastics based on cellulose esters, and it is important to remember that the migration of these compounds can cause part of the degradation phenomena. A curiosity concerns the identification of the true polymer support of a film labeled “Gevaert Safety AB”, suggesting that it was made of cellulose acetate butyrate, while the IR reflection spectrum indicated that it was indeed cellulose acetate: this is not an uncommon occurrence, because the code of the original negative (in this case of acetate butyrate) can persist in the positive print on a film with a different polymer support. This is interesting – and important – because the code possibly present on the edge of a film is the first criterion suggested in conservation practice to identify the material.
Based on your results, which technique would you recommend as the most practical starting point for heritage institutions that may have limited access to advanced analytical tools?
I think one cannot ignore the economic issue and the fact that purchasing equipment for an institution facing conservation problems can be challenging, so perhaps the best compromise between costs and information one can obtain is a portable NIR spectrometer, operating in the 1000 to 2500 nm region. The spectra obtainable in this region are characteristic of the different polymers and furthermore, as demonstrated by us and other researchers, allow us also to determine the degree of acetylation of cellulose acetate. This is a factor that in films depends on the chronology, since cellulose diacetate was introduced before triacetate, and on the conservation state of the film itself.
Do you see similar potential for vibrational spectroscopy in other areas of cultural heritage, beyond motion picture films?
Well, Raman spectroscopy has long played a prevalent role in the non-invasive identification of inorganic pigments in ancient paintings and illuminations and of synthetic organic pigments in contemporary paintings. FTIR spectroscopy was less established in this area – emerging more slowly largely due to the fact that it is commonly used in transmission and, therefore, in a destructive way. But the spread of reflection analyses, both in the mid- and near-IR, has helped to overcome this obstacle. FTIR opens the door to “in-situ” analysis, which can be of great help in the identification of original materials and also degradation products in all forms of artwork – from painting to sculpture, architecture, and crafts – especially organic materials such as natural and synthetic binders, varnishes, resins, adhesives, textiles, and so on.
What’s next for your team?
Regarding cinema, we are completing a study about the spectroscopic identification of materials responsible for colors in early cinema films, but we are also more generally interested in the recognition of the first synthetic dyes also in textiles using alternative methods to chromatographic ones. Then, with a completely different target but still using a variant of vibrational spectroscopy, we are working on the application of surface-enhanced Raman spectroscopy to the analysis of food products.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
