 This work is focused on the correlation of the visual assessment of cellulose nitrate films and the molecular assessment of the same supports. So we want to understand if there is a clear correlation of the visual signs of what we see in the surface of the films and what it is happening in the film support. So when we are thinking about cellulose nitrate, we think about a material that is highly flammable with a characteristic chemical instability that results in the rise of acidity and the decay of the film support. Basically we can see that the film becomes brittle and sticky. But it also occurs the degradation of the gelatin due to the oxidative action of the cell of the nitrate acid that is produced during the degradation process. This is an autocatalytic and irreversible process. However, we have some opposite situations where we can find cellulose nitrate good as new with no signs of degradation, with flexibility, transparency and dimensional stability. Taking these aspects in consideration, we find intriguing why some film supports are as good as new while others decay severely. Is it because of the type of composition or the film format or because of the past enclosures? Maybe because of some interventions made by the photographers like these type of masks or even because they were kept on carriers? It seems that there are a lot of questions about it, but no clear answers for it and a huge amount of variables that we have to consider. So in order to understand a little bit more about the chemical and physical decay of the cellulose nitrate we establish a study in which the main goal is to correlate the visual signs as I said and the molecular assessment of the films. For that we have selected five Portuguese collections. These collections were selected according to the past use and condition. All the films have been identified and characterized by infrared spectroscopy. A little bit more about the cellulose nitrate. Fernando told a lot, so I'm going to be quick on this. Cellulose nitrate is a semi-synthetic polymer obtained by the chemical treatment of cellulose with nitric acid and sulfuric acid that acted like a catalyst. Cellulose is obtained on wood or cotton. Cotton leaders were preferred due to the high quality of the cotton material. Cellulose is a macro-molecule with repetitive units of an hydrochromocos that contain three hydroxyl groups. They are esterified and substituted by the nitrate groups here. The chemical and physical properties, like Fernando said, depend on the degree of the substitution of the hydroxyl groups. So we have cellulose nitrates, films that have a percentage of 11 to 12% of nitrogen. In order to achieve the desired chemical and mechanical properties, such as flexibility and dimensional stability, plasticizes were added. Fernando said and very well that Comfort was one of them. However, in the literature also seems to be a reference to the use of 3-phenylphosphate as a plane retardant for cellulose nitrate. In this slide we can see a cellulose nitrate spectrum. Here in these three bands, the presence of the nitrate groups. In this region for the cellulose ring. And here the peak for the camphor. Well, the most common degradation causes for the cellulose nitrates are usually pointed as the high temperature, high relative humidity, UV radiation, but also the biological contamination. In the literature we also find that the presence of the sulfates and water used for the manufacture of the films may cause degradation. Three degradation steps for the mechanism are proposed. The first is the cleavage here of the ester group and the abstraction of the hydrogen and the formation of a radical. The second step is the homolytic season of the linkage here of the nitrate group and the formation of products of degradation such as the carbonyl and hydroxyl groups that correspond to the arising of the pH. This basically can be correlated with the first two steps of the degradation chart. The third step for the degradation is a little bit more complex and proposed the ring disintegration ester formation depending on the acidity and also in the radical formation. In this step we have the autocatalytic points achieved and we follow these three to five steps and get the total loss of the film. So in order to find these signs of degradation and correlate the molecular degradation of the films we have selected five Portuguese collections. The first, the Almanquini Costa, is the result of an ethnographical survey done between 1935 and 1939 in Angola. It is the survey of the documentation made in several tribes in Angola much of them to do some registration of the costumes and rituals for different tribes. The collection is all in six per six black and white film negatives in a total of 8,700 negatives. They were cut individually and they were put in these small albums that we can see stored. This collection is particularly interesting. We found some negatives with yellow, pink and green issue that we believe that are intensifications of the images performed by the photographer. It was a collection that was very well preserved overall with no signs of degradation and it was also found a diary of the photographer. A very important report of what he had done during that period. He reports that he was aware of the high humidity in Angola and the effect of the high relative humidity in the supports. He kept all the films in metal box with calcium carbonate. He also said that it was important the quality of the water for the processing of the films. He only did the processing on a restricted area that he said that it was Musamich in Angola where he knew that he could rely on the water. It is an interesting report from a photographer that we have never seen in other collections. The second collection that we have selected is Direção Geral de Itifícios Imigrantes Nacionales. It is an architectural survey of several monuments and classified Portuguese heritage. It is a huge collection but unfortunately the project has stopped and we have only 18,000 black and white film-based negatives described, treated and digitized. Sorry, I was saying Portuguese digitized. This collection or in this archive the decision is to keep all the cellulose nitrates frozen. So we only had access to a restrict set of negatives. From that set we have selected negatives with different formats and also with different signs of degradation what we didn't have in the previous collection. A third collection, because we needed more samples of larger formats was the Sampaio collection. Sampaio was a photographer, that Portuguese photographer also that was mainly dedicated to portraits. He worked between 1920s and 1950s and from 1700 and 17,580 black negatives it was thought that there were all cellulose nitrates there was no clue about cellulose nitrates. This collection has not been treated so it was like open a box with lots of pearls and good things to dip in. So I presented you three collections however none of these collections had one severe degraded film, not one. So we had to look for two other collections to find or at least to have something to compare with the others that were better. We found these two collections, the Orasinovais and Cinema Teca collections. The Orasinovais, this one. We found large format films severely degraded and from the National Archive, the Cinema National Archive that were very generous with us gave us these metal blocks with all these cinema rolls most of them severely degraded too. So shortly to see how we did the selection we had five collections, Alman and Cunha and they had some difference between them mostly in past condition and present condition. Alman and Sampaia were visually identified as cellulose acetate by the conservators in the archives. Alman and Direção Geral de Edificios Imunimentos Nacionais have already been treated, digitized and stored in climatic, climatized storing areas. From all collections only the Alman collection is completely reported and described. Sampaia, Marinovais and Cinema Teca collections have not been treated or digitized. From the five collections we have selected a total of 172 samples according with the color U, silver mirroring, silver fading, nitric acid, odor, stickiness and brittleness. All the film supports have been identified and characterized by infrared spectroscopy. Well, we have treated the results from the visual assessment and also from the molecular assessment and we have seen that the results for the visual assessment of the collections there was a percentage of samples selected as fair, here, good, poor and severe. A small percentage of samples classified as very good. It was clearly not what we were expecting but at the end with all the evaluation of the different levels of degradation that's what we have got. However, when we treated the results from the infrared analysis, we saw that we have clearly a higher percentage of samples that are very good or good and a smaller percentage of fair and poor. This small percentage of severe cannot be correlated with this one here because we have only analyzed a small sample that we have taken from this group. So in conclusion, it seems that we found there is a poor correlation between the visual assessment and the molecular results that can be related to a subjective selection based on the color U or, however, there was no doubt that the negatives that have never been evaluated as the negatives that have been evaluated as very good or severe stage were clearly identified. The problem is when we try to identify or evaluate the negatives that are in these possible three stages of degradation. We have selected five specters that could represent the five stages of the degradation and what we can see here from black to red is that the black is a specter for a very good evaluated, very good negative, brown for good, brown or gray, brown for fair and blue for poor. We can see in this infrared spectra that in this region, the hydroxyl group region, there is a slow progression of the degradation and suddenly this speeds up and we get the total loss of the film. This is obvious here also in the area where we have the degradation products such as haldehydes. We also see that there is an increasing of the peak or the band for camphor and a broadening and intensification of the peak for the cell loss ring. However, for the nitrate groups, we can see here in this area a decrease of the groups. This means that we have the typical formation of the hydroxyl groups and the loss of the nitrate groups. We have also found from the infrared results that we have 16 spectra. They are different from the other ones. Mostly in the green spectra here, we can see different peaks and new bands here, a shoulder here, a new band here, two more here. It's not so clear here, but there are two peaks more intense and also a shoulder here. We have not found a clear answer for this. We think that it can be a plasticizer. We have checked these three plasticizers, but there is not a clear match. It can be a composition, another composition. Like Fernanda said, maybe it can be a mixture of nitrate and acetate. We have to see. Maybe also a degradation path. However, it's different from all the other spectra that we have seen even in the degradation. We have treated qualitatively all the spectra from the collection. However, we wanted to perform a quantification of the areas of the infrared brands. That could give us the formation of the degradation products. That we did by the evaluation of the broadening of the areas of the bands. And also the decreasing of the intensity of the bands that is correlated with the loss of the nitrate groups. For that, we have applied the Gauss and Lorentz functions. We have applied the quantification before to have the same criteria for the evaluation of the degradation. We have selected this nitrate band to do the normalization. For the quantification, we have used this band, this one here, and also the cell loss band. What we have found is that this band here with the Lorentz function give us a minor error and can give us more clues for the correlation between the degradation signs and the molecular results. However, these functions are time-consuming since we have to apply each function to parts of the spectra and to each spectra at each time. We try to find another tool that could leave us an analysis of the full spectra and all spectra, like this 172 spectra, in order to find variables that could somehow clarify the path and give us clues for the most degraded or less degraded films. We have selected the principal component analysis. That is an exploratory analysis based on a mathematical procedure that allowed to develop and predict models based on the variables found in the results obtaining the analysis of samples. The number of principal components depends on the number of variables present in the samples. It always depends on something. We can find the results in the plots that we have done. We have applied the PCA for the full spectra, but there are so many variables that we don't have a clear identification of the samples. By restricting the area of the spectra, we have here a clear identification of the severe degraded sample, a cluster for samples classified as good, good and poor. However, we have not a clear result for samples classified as fair that we see in blue. In order to have these results, we have to restrict the set samples for 51. That way we restricted also the variables. In conclusion, there is not a clear correlation between the visual signs of degradation and the molecular evaluation, especially for the samples evaluated as fair. It seems that the U changes can lead us to misinterpretation and to incorrect evaluation of the degradation stage. In the overall collections, the small formats negatives reveal to be less degraded. In opposition, the cinema role and the sheet films have more or shown more severe degradation. According to the results achieved for Elm and Quinicoste collection, past conditioning and storage seem to have a great influence on the preservation of the films. There is some work to do about the different specter that we found. However, we found that the PCA is a useful tool for micro-FTRI analysis, especially when we are dealing with a great amount of data and variables. For the future, we think that it would be necessary to find non-destructive techniques and analytical techniques that somehow could allow us to identify and characterize film supports. Thank you for your attention.