Vibrational spectroscopy as a tool for the investigation of polymer bases in motion picture films: A comparison between mid-infrared, near-infrared and Raman techniques

The goal of this study is to evaluate various vibrational spectroscopy techniques—FTIR (MIR/NIR), Raman with SSE™, and NIR with PLS regression—for analyzing the polymeric support materials of motion picture films. The films examined span a wide chronological range and include different base types such as cellulose nitrate, cellulose acetate, PET, and cellophane.

A key aim is to compare the informational value of each method regarding molecular specificity and penetration depth, and to estimate the degree of substitution in cellulose acetate films, which reflects both production period and possible degradation. The findings provide insights for film preservation and material identification.

The original article

Vibrational spectroscopy as a tool for the investigation of polymer bases in motion picture films: A comparison between mid-infrared, near-infrared and Raman techniques

Alessia Buttarelli, Margherita Longoni, Valentina Rossetto, Silvia Bruni 

Vibrational Spectroscopy, Volume 139, 2025, 103818

https://doi.org/10.1016/j.vibspec.2025.103818

licensed under CC-BY 4.0

Selected sections from the article follow. Formats and hyperlinks were adapted from the original.

In the present work, the polymer supports of numerous cinematographic film samples of various brands and time periods—ranging from approximately 1895 to the first decade of the 21st century— were studied using a comprehensive approach based on vibrational spectroscopic techniques. Among the materials analyzed were Ozaphan films, which had never been studied before from the perspective of materials science. Introduced for the first time in the 1920s in France and subsequently adopted in Germany for small formats such as 16 mm, their production was interrupted during and immediately after World War II, but was then resumed until 1958. Their peculiarity lies in the absence of an emulsion layer, as the image was printed directly onto a cellophane support using the diazotype printing process [12].

The first objective of the work was to investigate the films using techniques that allow a completely non-invasive and, in principle, “in situ” analysis by means of portable instrumentation. This approach is especially valuable for the application in those contexts where films are preserved, such as archives and film libraries. For this reason, FTIR spectroscopy was applied in external reflection (ER) mode, not requiring any contact with the sample, both in the mid-infrared (MIR) range and in the longer-wavelength portion of the near-infrared (NIR) region. For the same reason, Raman spectroscopy was also used in this study, as its use has been reported in the literature for the characterization of cellulose-based heritage objects [13], [14]. However, its application to motion picture films has so far been rather limited, with the exception of a study on the degradation of cellulose nitrate materials [15]. In particular, the Sequentially Shifted Excitation (SSE™) patented technology was used here to mitigate the issue of fluorescence emission, which is often associated with organic materials. This technology represents the up-to-date development of the so-called shifted excitation methods for extracting Raman information from an interfering background. These methods are based on the fact that, by varying the excitation laser wavelength during the acquisition of the Raman spectrum, a change occurs in the position of the Raman bands in the spectral space, while unwanted spectral components such as those corresponding to fluorescence remain unchanged in the same space. In the simplest version, just two excitation wavelengths are used and the final Raman spectrum without the fluorescence contribution is obtained as the difference between the two measurements. In the SSE technology, diode lasers are used that operate at different temperatures and therefore deliver several slightly shifted wavelengths in the NIR region; a specific algorithm, which is part of the patent, allows the extraction of the spectral data in the Raman space [16].

The study, based on the significant number of film samples examined and on their representativeness in terms of different materials, made it possible to compare the aforementioned techniques from the point of view of the type of information they can provide. It is worth highlighting that the techniques considered differ first of all in the depth of penetration into the material. This depth is more limited and is around a few tens of micrometers for MIR radiation [17], while it is greater for NIR radiation, even reaching a few hundred micrometers [18]. This last fact applies to both ER-NIR measurements and Raman measurements with NIR excitation. Another important difference between the ER-MIR and ER-NIR techniques lies in the intensity of the absorptions observed in the two spectral regions. In fact, in the near infrared, overtone and combination vibrational bands are observed, with an intensity lower than the fundamental ones, and therefore the specular reflection spectra are less subject to distortions due to the combination of reflection and absorption contributions compared to those obtained in the mid-infrared region [19]. Finally, it is worth mentioning that IR and Raman spectroscopies are based on different selection rules, so they are usually complementary from the point of view of sensitivity towards different molecular structures [20].

Subsequently, in the second part of the work, using spectra from historical cellulose acetate films and reference standards acquired through diffuse reflection spectroscopy with benchtop instrumentation across the full NIR range (800–2500 nm, 12,500–400 cm⁻¹), a preliminary statistical model based on partial least squares (PLS) regression was developed to estimate the DS of film supports. The results obtained were then compared with those derived from ATR-FTIR spectra according to the procedure suggested in the literature [4].

2. Materials and methods

2.3. Instrumental methods

2.3.1. ER-FTIR spectroscopy

For non-invasive, contactless analyses in reflection mode, a Bruker Alpha FTIR spectrophotometer equipped with a reflection module was used. The instrument features a deuterated triglycine sulfate (DTGS) detector, which operates at room temperature and provides a linear response within the spectral range of 7500–375 cm⁻¹ . The FTIR spectrophotometer collects spectra from a sample area with a diameter of approximately 6 mm at a resolution of 4 cm⁻¹ . An integrated camera allows the operator to precisely select the measurement area.

FTIR spectra were acquired as the sum of 200 scans, following the acquisition of a background spectrum on a gold mirror. Each sample was positioned 1.5 cm from the instrument. Reflection spectra in the MIR region were processed using the Kramers–Kronig transform in the Bruker OPUS software, while those in the NIR range were converted into pseudo-absorbance values using Log(1/R).

2.3.2. ATR-FTIR spectroscopy

Attenuated total reflection (ATR) FTIR spectra were acquired using a Jasco FTIR-470 spectrometer equipped with an ATR accessory featuring a single-reflection ZnSe crystal at a 45° angle of incidence. Each spectrum was obtained by summing 256 accumulations over the spectral range of 4000–700 cm⁻¹ , with a resolution of 4 cm⁻¹ . The contributions of CO₂ and water vapor were compensated by subtracting a baseline spectrum.

2.3.3. Raman spectroscopy

Raman measurements on the films were conducted using a Bruker BRAVO handheld spectrometer. The instrument employs patented SSE™ technology, which minimizes interference from fluorescence emission. Spectra were excited using two diode lasers operating at different temperatures, emitting at 785 nm and 852 nm, respectively. The spectral data were collected in two ranges, 2000–3200 cm⁻¹ and 300–2000 cm⁻¹ , with a dedicated algorithm extracting the final Raman spectra.

The instrument provides an average spectral resolution of approximately 11 cm⁻¹ , with laser power always kept below 100 mW for both lasers. Acquisition time and the number of accumulations were automatically set by the instrument. To prevent potential damage from laser-induced heating, film samples suspected of having cellulose nitrate support were excluded from this analysis due to the lower thermal stability of this polymer compared to other cellulose esters [22].

2.3.4. Diffuse reflection NIR spectroscopy

This analysis was performed exclusively on cellulose acetate-supported films and reference materials using a Jasco UV/VIS/NIR V-570 spectrometer equipped with an integrating sphere internally coated with BaSO₄. Spectra were collected over the range of 800–2500 nm (12,500–4000 cm⁻¹) at an acquisition rate of 200 nm/min and a resolution of 20 nm. A Mylar® polyester target with an assumed reflectance of 100 % was used as the reference.

3. Results and discussion

3.1. Identification of the polymer film base: a comparison of the techniques

3.1.1. ER-FT-MIR spectroscopy

ER-FT-MIR spectroscopy, which unlike the ATR-FTIR technique allows analysis without any direct contact with the sample, first of all makes it possible to separately characterize the polymer support and the emulsion layer of the films, thanks to the above-discussed limited penetration depth of the technique. Fig. 3 presents the IR spectra obtained from the emulsion side of films with different types of polymer support. In all cases, the characteristic amide bond signals were identified at 1636 cm⁻¹ (CO stretching) and 1540 cm⁻¹ (N–H bending and C–N stretching), together with a band at 1440 cm⁻¹ corresponding to CH₂ deformation. These characteristics are consistent with gelatin [23], the primary protein-based component of the emulsion layer in motion picture films. The Ozaphan films were excluded from this comparison, as they lack an emulsion layer, as explained in the Introduction.

Vibrational Spectroscopy, Volume 139, 2025, 103818: Fig. 3. FTIR spectra of the emulsion layer of film samples with different polymeric support: (a) gr1_2A (cellulose nitrate), (b) gr2_1 (cellulose acetate), (c) gr1_6pers (PET).

Notably, in spectrum (a) of Fig. 3, due to a film with a cellulose nitrate support, a band at 840 cm⁻¹ is observed. This peak is not attributed to the emulsion layer but instead to the underlying cellulose nitrate support (as discussed below). The reason why only this band of the polymer is observed when the film is examined from the emulsion side is that it is the only one located in a range where the overlying gelatin layer has no relevant absorptions and is therefore relatively transparent to the radiation.

3.1.2. ER-FT-NIR spectroscopy

The different support materials of the examined films also exhibit very characteristic spectra in the NIR region (6500 to 4000 cm⁻¹, Fig. 4, box II), which allow for easy distinction between them. Significant bands in the 4500–4000 cm⁻¹ range, which had not been reported in previous studies on motion picture films, were also observed in this investigation. For cellulose acetate and nitrate, these bands can be tentatively assigned to the combination of vibrational modes of the cellulose backbone, specifically C–H stretching and C–C stretching (around 4060 and 4020 cm⁻¹), C–H stretching and CH₂ deformation (around 4250 cm⁻¹), C–H stretching and O–H stretching (around 4330 cm⁻¹), and O–H stretching and C–O stretching (around 4430 cm⁻¹). Furthermore, a band at about 5220 cm⁻¹ is observed, which is attributed to the combination of O–H stretching and deformation of absorbed water molecules [32]. The NIR spectra of the cellulose acetate supports are also characterized by a signal at 4680 cm⁻¹, which is associated with the acetylation of cellulose and attributed to the combination of C–H stretching and CO stretching of acetyl groups [33]. For cellulose nitrate, two broad signals at around 4890 and 4640 cm⁻¹ are observed, which have been previously assigned to the combination of vibrational modes of O–H groups. These are still present in cellulose nitrate, as it does not usually undergo complete substitution of hydroxyl groups with nitrate groups [34].

However, it is in the case of PET-based films that one of the advantages deriving from the use of external reflection of NIR radiation becomes evident. Indeed, as can be deduced from Fig. 4(c), in ER-FTIR this polymeric support is recognized more clearly by its NIR spectrum than by its MIR spectrum, as it does not show any distortion of the bands thanks to the lower extinction coefficients in this spectral region (see Introduction). The NIR spectrum of PET is characterized by a sharp band at 4092 cm⁻¹, typically due to the presence of the aromatic ring in the polymer structure [35], which is also responsible for the three bands at 4691, 4636, and 4581 cm⁻¹.

Finally, another property related to NIR radiation compared to MIR radiation and already mentioned above, i.e. its greater depth of penetration, is demonstrated by the NIR reflection spectra of the Ozaphan films (Fig. 4(d), box II), which correctly show the typical pattern expected for a cellulosic material such as cellophane [32], while the bands in the MIR spectrum were due to the cellulose nitrate surface coating (see Section 3.1.1).

3.1.3. Raman spectroscopy

Raman spectroscopy with SSE™ technology was only applied for cellulose acetate, PET and Ozaphan films, to avoid thermal degradation in the case of cellulose nitrate as mentioned above. Examples of the spectra obtained for each material are shown in Fig. 5 and the spectra acquired for all films examined are reported in Figs. S6–S8 (Supplementary Material).

Vibrational Spectroscopy, Volume 139, 2025, 103818: Fig. 5. SSE™ Raman spectra of motion picture film samples with different polymeric support: (a) gr2_4 (cellulose acetate), (b) gr1_6pers (PET), (c) A8 (Ozaphan film).

4. Conclusions

The present study confirms the effectiveness of vibrational spectroscopy, applied using portable instrumentation, for the non-invasive identification of polymer bases in motion picture films. The use of portable spectrometers makes the techniques suitable for analyses conducted directly where the films are stored, such as in archives and film libraries. Moreover, the spectroscopic approach has proven effective not only for the most common types of support but also for unusual ones, such as Ozaphan films.

ER-FTIR spectroscopy in the MIR region is particularly suitable for identifying polymer supports based on cellulose, such as cellulose acetate and nitrate, due to the characteristic functional groups. Regenerated cellulose, such as cellophane, can also be recognized; however, if a protective layer is present, as in the Ozaphan films, only the latter will be detected in the MIR range. Conversely, the synthetic polymer base PET usually exhibits significantly distorted bands in this spectral region. In contrast, in the NIR region, even limited to its longer-wavelength portion, characteristic spectra are obtained for all supports, even in the presence of a surface coating and without band distortion, thanks to the weaker absorptions typically observed in this spectral range.

Raman spectroscopy is less efficient when cellulose derivatives are involved, as they are rather weak scatterers and thus tend to suffer from fluorescence background interference, even when a dedicated technology is employed, as in the present work. On the other hand, good quality spectra are usually obtained for PET films. Furthermore, the Raman technique enables the identification of additional film components, for example plasticizers such as triphenyl phosphate in cellulose acetate films and azo dyes in Ozaphan films. The rather strong Raman bands associated with molecular structures containing aromatic rings can explain the above observations.

Finally, NIR spectroscopy also proved useful for evaluating the DS value of cellulose acetate films, a parameter linked to the chronology and conservation state of the film. In this case, a benchtop spectrometer was used to access a wider spectral range than portable instrumentation and a method based on a preliminary model constructed via PLS regression was proposed. Good results were obtained even when only portions of films with dark images were available, where broad absorption extending from the visible to the near-infrared region overlaps combination and overtone vibrational bands. This method paves the way for the development of future refined models and further studies on other types of film bases.