Mock-Ups in Plastic Conservation Research: Processing and Aging of 3D Celluloid Specimens Simulating Historical Objects

The preparation of mock-ups in heritage science studies represents a valid alternative for investigation purposes, avoiding extensive sampling of cultural heritage objects. This work presents for the first time the successful preparation of three dimensional (3D) mock-ups made of celluloid, considering a combination of historical industrial production strategies and small-scale lab facilities. Prefabricated transparent celluloid sheets were acquired and then shaped through compression molding for creating mock-ups with 3D geometries. These reflected common and representative shapes encountered in the collection of the Deutsches Museum. Visual inspection of the mock-ups allowed determining the best compression molding conditions. Attenuated total reflection Fourier-transform infrared spectroscopy (ATR-FTIR) confirmed the absence of molecular heterogeneity due to the processing method. Artificial aging of the mock-ups was conducted to reach degradation states comparable with naturally aged objects. ATR-FTIR investigation offered first insights into the induced artificial degradation. Ion chromatography (IC) and gel permeation chromatography (GPC) analyses allowed to assess the extent of the artificial aging of the celluloid mock-ups and confirmed the occurrence of loss of camphor, denitration, and main chain polymer scission, the latter being the predominant decay path. The comparison with historical objects highlighted that the mock-ups are representative of moderately aged artifacts. As such, this study paves the way for implementing moderately aged celluloid 3D mock-ups in heritage science research, enabling in-depth testing for the scope of conservation.


Introduction
Celluloid counts as one of the milestones in the history of polymers, since it marked the starting point for the commercial use of plastics at the end of the 19th century [1] and the beginning of our plastic age [2]. Celluloid is a semisynthetic plastic made of cellulose nitrate (CN) plasticized with terpenoid camphor. Thanks to its versatile properties, celluloid was used in many applications to produce cinematographic films, photographs, textile fibers, and molded goods, but also renowned constructivists of the 20th century appreciated this material for their artworks. Therefore, three-dimensional (3D) artifacts made of celluloid are present in various collections, including design, science, technology, and modern and contemporary art. Because of its inherent instability, celluloid is considered one of the malignant plastics and is challenging to preserve in collections. The comprehensive

Mock-Ups: Instrumentation for the Shaping Process
To gain the geometries shown in Table 1, the sheets were cut from the prefabricated sheets into the desired dimensions and were ready for usage without further manipulation. For the production of tines and cylinders, the prefabricated sheets were cut and then shaped under a heatable slab press (PCS II, COLLIN) for compression molding. Compression molding with heat and pressure was employed because it was one of the past's stand-

Mock-Ups: Instrumentation for the Shaping Process
To gain the geometries shown in Table 1, the sheets were cut from the prefabricated sheets into the desired dimensions and were ready for usage without further manipulation. For the production of tines and cylinders, the prefabricated sheets were cut and then shaped under a heatable slab press (PCS II, COLLIN) for compression molding. Compression molding with heat and pressure was employed because it was one of the past's stand-

Mock-Ups: Instrumentation for the Shaping Process
To gain the geometries shown in Table 1, the sheets were cut from the prefabricated sheets into the desired dimensions and were ready for usage without further manipulation. For the production of tines and cylinders, the prefabricated sheets were cut and then shaped under a heatable slab press (PCS II, COLLIN) for compression molding. Compression molding with heat and pressure was employed because it was one of the past's stand-

Mock-Ups: Instrumentation for the Shaping Process
To gain the geometries shown in Table 1, the sheets were cut from the prefabricated sheets into the desired dimensions and were ready for usage without further manipulation. For the production of tines and cylinders, the prefabricated sheets were cut and then shaped under a heatable slab press (PCS II, COLLIN) for compression molding. Compression molding with heat and pressure was employed because it was one of the past's standard methods for producing combs and spectacle arms [25,26,28]. The molds were designed with CAD and manufactured in aluminum (tines) and steel (cylinders).

Selection of Compression Molding Parameters for Processing Tines and Cylinders
Because the prefabricated 0.5 mm celluloid sheet did not match the thickness required for the tines and cylinders, the following adjustments were made. The prefabricated sheets were cut in geometries following the contour of tines or cylinders, or cut into tiny pieces to simulate industrial granulates, and then layered or pilled in the mold to produce thick 3D shapes (Figure 1). The precut geometries or granulates were softened with solvent to ensure better cohesion between layers. Acetone (≥ 99.5% for synthesis, Carl Roth) was chosen as a solvent instead of ethanol because it evaporates faster and dissolves celluloid more effectively [21,26,30].

Mock-Ups: Instrumentation for the Shaping Process
To gain the geometries shown in Table 1, the sheets were cut from the prefabricated sheets into the desired dimensions and were ready for usage without further manipulation. For the production of tines and cylinders, the prefabricated sheets were cut and then shaped under a heatable slab press (PCS II, COLLIN) for compression molding. Compression molding with heat and pressure was employed because it was one of the past's standard methods for producing combs and spectacle arms [25,26,28]. The molds were designed with CAD and manufactured in aluminum (tines) and steel (cylinders).

Selection of Compression Molding Parameters for Processing Tines and Cylinders
Because the prefabricated 0.5 mm celluloid sheet did not match the thickness required for the tines and cylinders, the following adjustments were made. The prefabricated sheets were cut in geometries following the contour of tines or cylinders, or cut into tiny pieces to simulate industrial granulates, and then layered or pilled in the mold to produce thick 3D shapes ( Figure 1). The precut geometries or granulates were softened with solvent to ensure better cohesion between layers. Acetone (≥ 99.5% for synthesis, Carl Roth) was chosen as a solvent instead of ethanol because it evaporates faster and dissolves celluloid more effectively [21,26,30]. As the processing greatly influenced the final-product properties, the following parameters were considered: • raw material (granules and sheets); • time between softening and pressing (4 h, 1 day, 2 days, 6 days); • set temperature of the slab press (90, 100, 120 °C); • preheating time (for tines: 3, 5 and 7 min; for cylinders: 10, 15, 20 min); As the processing greatly influenced the final-product properties, the following parameters were considered: • raw material (granules and sheets); • time between softening and pressing (4 h, 1 day, 2 days, 6 days); • set temperature of the slab press (90, 100, 120 • C); • preheating time (for tines: 3, 5 and 7 min; for cylinders: 10, 15, 20 min); • pressure (0 and 25 bar); • pressing time (5, 8, 10 min); • cooling method (after pressing, the mold was either opened directly (no cooling), cooled by the cooling function of the press (slow cooling) or cooled by putting it in a water (fast cooling)).
These were tested in different combinations to assess the best compression molding procedure. Only one parameter was changed per production test, and the raw material, preheating time, pressure, and cooling method were the first to be defined, being the other parameters tested in every possible combination. The set temperature of the press and pressure were chosen following historical sources [25,26,28]. The preheating time, pressing time, and cooling method were derived from DIN EN ISO 293 "Plastics-Compression moulding of test specimens of thermoplastic materials" [31].
The quality of the produced mock-ups was assessed visually after production. To select the best combination of processing conditions, both time efficiency and the occurrence of minimal or no defects (e.g., blowholes, bubbles, deformation by demolding, discoloration, incomplete filling of the mold, irregularities on the surface, cavities and sunken areas at the surface, warpage, and shrinkage, deviation of the metal rod from the cylinders' centers) were taken into consideration.

Artificial Aging
Artificial aging can help to speed up deterioration processes for studying their reactions in the laboratory [32]. In this study thermohygrometric artificial aging was applied to accelerate the deterioration processes in the newly processed mock-ups following the strategies presented in Table 2. The artificial aging aimed at inducing a moderately degraded condition. Therefore, the mock-ups were monitored by visual assessment and mass loss. The duration of the artificial aging was determined when mock-ups were visibly in good condition but off-gassing (acid detection). A fan-assisted dynamic climate chamber (MKF 115, Binder, Tuttlingen, Germany) was used to age sheets, tines, and cylinders. Unaged mock-ups for each geometry were kept in the dark at room temperature as control.
Sheets. Eighty sheets were artificially aged at 70 • C and 75% relative humidity (RH), following Quye et al. [12] and Steward [36,37]. A hole was drilled in one of the corners to allow their hanging on a glass rod, permitting air circulation during aging in the chamber. Every working day, eight sheets were weighed, and during this measuring time (around 20 min), the door of the climate chamber was left open to allow a complete exchange of air.
Tines. The conditions of the artificial aging were 60 • C and 75% RH. Forty tines were placed in petri dishes without being in contact with each other. Following the same strategy as for sheets, eight tines were weighed following the same procedure. The tines were weekly turned upside down on the petri dishes to homogeneously expose their surfaces.
Cylinders. The cylinders were aged together with the tines at 60 • C and 75% RH sharing the same environment. As for the tines, 40 cylinders were placed in petri dishes without touching each other and weekly turned during the aging. The cylinders were weighted as for sheets and tines.
More information about the rationale behind lowering the temperature in the aging of tines and cylinders is reported in Section 3.3 and the Supplementary Materials ( Figure S1 and related paragraph). artificial aging, samples of each mock-up geometry were measured, and the percentage mass losses during time were calculated.

Off-Gassing Control with A-D Strips
A-D strips (Danchek AD Strips), an acid-base indicator made of paper strips dyecoated with bromocresol green indicator developed by the Image Permanence Institute (IPI), were used to determine the presence of acids associated with off-gassing from the mock-ups after artificial aging. The indicator will change from blue, via green, to finally yellow in the presence of acids depending on the pH. Even though such A-D strips were designed for acetate films [42], they are also applied to detect celluloids' nitric acid emission [43,44]. Four samples of each mock-up geometry were bagged individually in a 7 × 10 cm PE-LD zipper bag (WeltiesSmartTools Premium) with an A-D strip close but not in contact with the samples. The PE-LD bags were then stored in the dark in a lab environment, and the color of the A-D strips was visually evaluated after 24 h of exposure.

Attenuated Total Reflection Fourier-Transform Infrared Spectroscopy (ATR-FTIR)
ATR-FTIR analysis was applied to (a) assess if the processing methods of the mock-ups in this study induced unwanted heterogeneity at molecular level of the tines and cylinders, and (b) estimate their degradation caused by artificial aging.
For both aims, the mock-ups were characterized with a Bruker Alpha ATR-FTIR spectrometer equipped with a diamond crystal. The collected spectral region was 4000-400 cm −1 using 64 scans with a spectral resolution of 4 cm −1 . Opus 8.1 software (Bruker Optics) was used to acquire and elaborate the spectra.
In order to check the possible occurrence of heterogeneities in the tines and cylinders due to their processing, the standard deviation (SD) of the absorbance intensities of a selection of ATR-FTIR bands was calculated following Ballany [45] and Shashoua and Johansen [46]. The unaged sheets, neither manipulated by the shaping process nor the aging, were used to measure the instrumental and sampling errors that would set the SD ranges. The instrumental errors were determined by running ten spectra at the same spot (experiment A). Sampling errors were considered variations in the spectra resulting from ten different positions on one sheet (experiment B) and ten different sheets (experiment C) ( Table 3). All acquired spectra were normalized by using the peak of C-O-C in the backbone of anhydroglucose units at 1051 cm −1 as reference (A/A1051cm −1 ). This peak is supposed to be less affected during degradation and is considered unchanged in the early stages of degradation [12,36,[47][48][49]. After normalization, the SDs of the absorbance intensities at 1730 (C=O), 1634 (NO 2 ), 1273 (NO 2 ) and 826 cm −1 (NO) were calculated. The minimum and maximum values of SDs achieved from experiments A to C gave the ranges of allowable variations for each peak. Heterogeneity due to the processing would be evident if the SD values of each absorbance intensity for the tines and cylinder were higher than the maximum allowable SD value calculated from the sheets. allowable variations for each peak. Heterogeneity due to the processing would be evident if the SD values of each absorbance intensity for the tines and cylinder were higher than the maximum allowable SD value calculated from the sheets. Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, Polymers 2023, 15, x FOR PEER REVIEW 7 of 23 allowable variations for each peak. Heterogeneity due to the processing would be evident if the SD values of each absorbance intensity for the tines and cylinder were higher than the maximum allowable SD value calculated from the sheets. Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, Polymers 2023, 15, x FOR PEER REVIEW 7 of 23 allowable variations for each peak. Heterogeneity due to the processing would be evident if the SD values of each absorbance intensity for the tines and cylinder were higher than the maximum allowable SD value calculated from the sheets. Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, tines and cylinders were sampled with a surgical scalpel at six sampling spots under the stereomicroscope, and the sampled chips were subjected to ATR-FTIR analysis straight after the sampling. Each chip was handled under magnification lenses to ensure the measurement of its outer surface. One ATR-FTIR measurement was performed for each location, and four samples per mock-up geometry were considered to guarantee representativeness. All acquired spectra were normalized (A/A1051cm −1 ). Absorbance intensities at 1634, 1273 and 826 cm −1 associated with nitro groups and at 1730 cm −1 originating from carbonyl stretching were selected as probe bands to estimate the degradation due to artificial aging. Table 4. Measuring and sampling strategy implemented for the ATR-FTIR investigation of the mock-ups after production, before and after artificial aging.

Spectra acquisition
Direct measurement on the mock-up

Measurement of a sampled chip Measurement of a sampled chip
Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, tines and cylinders were sampled with a surgical scalpel at six sampling spots under the stereomicroscope, and the sampled chips were subjected to ATR-FTIR analysis straight after the sampling. Each chip was handled under magnification lenses to ensure the measurement of its outer surface. One ATR-FTIR measurement was performed for each location, and four samples per mock-up geometry were considered to guarantee representativeness. All acquired spectra were normalized (A/A1051cm -1 ). Absorbance intensities at 1634, 1273 and 826 cm -1 associated with nitro groups and at 1730 cm -1 originating from carbonyl stretching were selected as probe bands to estimate the degradation due to artificial aging. Table 4. Measuring and sampling strategy implemented for the ATR-FTIR investigation of the mock-ups after production, before and after artificial aging.

Spectra acquisition
Direct measurement on the mock-up

Measurement of a sampled chip
Location of the measuring and sampling points Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, tines and cylinders were sampled with a surgical scalpel at six sampling spots under the stereomicroscope, and the sampled chips were subjected to ATR-FTIR analysis straight after the sampling. Each chip was handled under magnification lenses to ensure the measurement of its outer surface. One ATR-FTIR measurement was performed for each location, and four samples per mock-up geometry were considered to guarantee representativeness. All acquired spectra were normalized (A/A1051cm -1 ). Absorbance intensities at 1634, 1273 and 826 cm -1 associated with nitro groups and at 1730 cm -1 originating from carbonyl stretching were selected as probe bands to estimate the degradation due to artificial aging. Table 4. Measuring and sampling strategy implemented for the ATR-FTIR investigation of the mock-ups after production, before and after artificial aging.

Spectra acquisition
Direct measurement on the mock-up

Measurement of a sampled chip
Location of the measuring and sampling points Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, tines and cylinders were sampled with a surgical scalpel at six sampling spots under the stereomicroscope, and the sampled chips were subjected to ATR-FTIR analysis straight after the sampling. Each chip was handled under magnification lenses to ensure the measurement of its outer surface. One ATR-FTIR measurement was performed for each location, and four samples per mock-up geometry were considered to guarantee representativeness. All acquired spectra were normalized (A/A1051cm -1 ). Absorbance intensities at 1634, 1273 and 826 cm -1 associated with nitro groups and at 1730 cm -1 originating from carbonyl stretching were selected as probe bands to estimate the degradation due to artificial aging. Table 4. Measuring and sampling strategy implemented for the ATR-FTIR investigation of the mock-ups after production, before and after artificial aging.

Spectra acquisition
Direct measurement on the mock-up

Measurement of a sampled chip
Location of the measuring and sampling points

Location of the measuring and sampling points
Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, tines and cylinders were sampled with a surgical scalpel at six sampling spots under the stereomicroscope, and the sampled chips were subjected to ATR-FTIR analysis straight after the sampling. Each chip was handled under magnification lenses to ensure the measurement of its outer surface. One ATR-FTIR measurement was performed for each location, and four samples per mock-up geometry were considered to guarantee representativeness. All acquired spectra were normalized (A/A1051cm -1 ). Absorbance intensities at 1634, 1273 and 826 cm -1 associated with nitro groups and at 1730 cm -1 originating from carbonyl stretching were selected as probe bands to estimate the degradation due to artificial aging. Table 4. Measuring and sampling strategy implemented for the ATR-FTIR investigation of the mock-ups after production, before and after artificial aging.

Spectra acquisition
Direct measurement on the mock-up

Measurement of a sampled chip
Location of the measuring and sampling points Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, tines and cylinders were sampled with a surgical scalpel at six sampling spots under the stereomicroscope, and the sampled chips were subjected to ATR-FTIR analysis straight after the sampling. Each chip was handled under magnification lenses to ensure the measurement of its outer surface. One ATR-FTIR measurement was performed for each location, and four samples per mock-up geometry were considered to guarantee representativeness. All acquired spectra were normalized (A/A1051cm -1 ). Absorbance intensities at 1634, 1273 and 826 cm -1 associated with nitro groups and at 1730 cm -1 originating from carbonyl stretching were selected as probe bands to estimate the degradation due to artificial aging. Table 4. Measuring and sampling strategy implemented for the ATR-FTIR investigation of the mock-ups after production, before and after artificial aging.

Spectra acquisition
Direct measurement on the mock-up

Measurement of a sampled chip
Location of the measuring and sampling points Note: X measuring spot on one side of the sheets; X measuring spot on the other side of the sheets. Table 4 depicts the measuring and sampling strategy used to assess the possible heterogeneity after processing and the degradation after artificial aging. While measurements of the sheets could be performed directly on five locations at the surface, the irregular and rounded shapes of the tines and cylinders inhibited comparable band intensities and quality of the IR spectra due to unequal contact with the diamond crystal. Therefore, tines and cylinders were sampled with a surgical scalpel at six sampling spots under the stereomicroscope, and the sampled chips were subjected to ATR-FTIR analysis straight after the sampling. Each chip was handled under magnification lenses to ensure the measurement of its outer surface. One ATR-FTIR measurement was performed for each location, and four samples per mock-up geometry were considered to guarantee representativeness. All acquired spectra were normalized (A/A1051cm -1 ). Absorbance intensities at 1634, 1273 and 826 cm -1 associated with nitro groups and at 1730 cm -1 originating from carbonyl stretching were selected as probe bands to estimate the degradation due to artificial aging. Table 4. Measuring and sampling strategy implemented for the ATR-FTIR investigation of the mock-ups after production, before and after artificial aging.

Mock-Up Geometry Sheet Tine Cylinder
Spectra acquisition Celluloid samples were taken as displayed in Figure 2. IC and GPC were performed in two measuring campaigns between September 2021 and May 2022. Four different samples were measured for each mock-up geometry before and after aging, considering at least one microsample of four mock-ups for each geometry for covering a statistically representative set. Celluloid samples were taken as displayed in Figure 2. IC and GPC were performed in two measuring campaigns between September 2021 and May 2022. Four different samples were measured for each mock-up geometry before and after aging, considering at least one microsample of four mock-ups for each geometry for covering a statistically representative set. Following the sample preparation proposed by Mazurek et al. [23], 5 mg of celluloid sample was dissolved in 1 mL acetone (Acetone ≥ 99.5% for synthesis, Roth) at 50 °C for one hour. The naturally aged samples (from objects P and G in Section 2.6.) did not dissolve in pure acetone; therefore, these samples were dissolved in 1 mL 7/3 (acetone:water)  Following the sample preparation proposed by Mazurek et al. [23], 5 mg of celluloid sample was dissolved in 1 mL acetone (Acetone ≥ 99.5% for synthesis, Roth) at 50 • C for one hour. The naturally aged samples (from objects P and G in Section 2.6.) did not dissolve in pure acetone; therefore, these samples were dissolved in 1 mL 7/3 (acetone:water) mixture at ca. 50 • C for two hours. The process of alkaline hydrolyses was conducted by adding 2 mL 1 N NaOH (Merck) to the previously dissolved samples and heated for 2 h at 60 • C. Finally, 17 mL water was added for a final volume of 20 mL. Before analysis, the sample solution was 100-fold diluted (0.5 mL sample solution added to 50 mL water) and then measured by a The theory/calculation summarized by Mazurek et al. [23] was used to calculate the nitrogen content. In contrast to [23], no previous plasticizer extraction was made in this research. Therefore, for each 5 mg celluloid sample, the respective amount of camphor measured by GPC was subtracted.

Gel Permeation Chromatography for Determining Molecular Weight (M w )
Based on the method of Kavda et al. [24], 2 mg of sample was dissolved in 1 mL of N,N-dimethylacetamide (DMAc 99.5% HPLC grade, Alfa Aesar) with 0.5% w/v lithium chloride (LiCl ACS 99%, Alfa Aesar) and left to dissolve overnight. Before injection, samples were filtered through a poly(tetrafluoroethylene) filter (PTFE, pore size 0.2 µm, diameter 25 mm) to remove any insoluble material and impurities. Then, 100 µL of sample solution was manually injected into a SECcurity2-GPC system (PSS). This system was equipped with a set of PSS GRAM columns (precolumn 10 µm 8 × 50 mm, 1 column 10 µm 30 Å 8 × 300 mm, and 2 columns 10 µm 1000 Å 8 × 300 mm) all sustained at 60 • C. The eluent DMAc/LiCl was maintained at a flow rate of 1 mL/min with an isocratic pump. An RID detector was set at 40 • C. The calibration with PSS ReadyCal-kit poly(methyl methacrylate) and PSS WinGPC UniChrom ® Software 8.3SR2 was used for the Mw estimation.

Gel Permeation Chromatography for Determining Camphor Content
Of the sample, 5 mg was dissolved in 1 mL of tetrahydrofuran (dried and fresh distilled from technical THF) and left to dissolve for 24 h in a shaker. Before injection, samples were filtered through a poly(tetrafluoroethylene) filter (PTFE, pore size 1.0 µm, diameter 25 mm, VWR) to remove any insoluble material and impurities. Then, 20 µL of sample solution was injected by an autosampler (SECurity2 standard autosampler (404-2104), PSS) into a SECcurity2-GPC system (PSS). This system was equipped with a set of PSS SDV columns (precolumn 3 µm 8 × 50 mm, 1 column 3 µm 10,000 Å 8 × 300 mm, and 2 columns 3 µm 1000 Å 8 × 300 mm) all maintained at 25 • C. The eluent THF was maintained at a flow rate of 1 mL/min with an isocratic pump. An RID detector was set at 25 • C. The calibration with pure Camphor (Sigma Aldrich, 99+%) was done by a 5-point concentration calibration.

Naturally Aged Transparent Objects
In this study, two collections of transparent naturally aged celluloid were considered (Figure 3), which were also already part of previous investigations by the authors Kavda et al. [24] and Elsässer et al. [16]. These materials were in extremely poor condition and fragmented, allowing extensive sampling for research purposes. throughout, became severely cracked, crazed, and fallen apart in many pieces.
Temples of eyeglasses (G). Fragments of transparent glasses. The production date of the glasses is unknown, but their form and shape resemble designs of the 1950s/1960s. Until their deaccession, they belonged to the Deutsches Museum optics collection. These temples were cracked, crazed, yellowed, and presented a greenish corrosion of metal elements. Samples for analysis were taken in locations where no greenish corrosion products were visible to avoid the influence of those corrosion salts in the degradation stage of the samples and make them more comparable with the plates (P) that lacked metal elements.

Shaping Tines and Cylinders
Softening trials made with granulates proved ineffective because acetone was lost too fast, reducing the working time to fill the mold and leading to undesirable defects. In contrast, the layering of several celluloid sheets led to the processing of satisfactory mock- Temples of eyeglasses (G). Fragments of transparent glasses. The production date of the glasses is unknown, but their form and shape resemble designs of the 1950s/1960s. Until their deaccession, they belonged to the Deutsches Museum optics collection. These temples were cracked, crazed, yellowed, and presented a greenish corrosion of metal elements. Samples for analysis were taken in locations where no greenish corrosion products were visible to avoid the influence of those corrosion salts in the degradation stage of the samples and make them more comparable with the plates (P) that lacked metal elements.

Shaping Tines and Cylinders
Softening trials made with granulates proved ineffective because acetone was lost too fast, reducing the working time to fill the mold and leading to undesirable defects. In contrast, the layering of several celluloid sheets led to the processing of satisfactory mock-ups. Figures 4 and 5 illustrate nine steps required for making tines and cylinders. Table 5 reports the different conditions tested during the preparation of tines and cylinders.  Table 5 reports the different conditions tested during the preparation of tines and cylinders.

Tines
Steps 1, 2, and 3. The contour was carved on the prefabricated celluloid sheets selected as raw materials, then cut with scissors.
Step 4. Five precut geometries were layered over each other, exceeding the desired

Cylinders
Steps 1 and 2. As for the tines, the contour was carved on the prefabricated celluloid sheets selected as raw materials, then cut with scissors.
Step 3. The geometries were layered over each other, forming the two halves of the cylinder diameter, as displayed in Figure 1. Following the same strategy as for the tines, acetone was added between layers to soften the celluloid in a ratio of approximately 1:1. After softening, a metal rod was inserted between the two halves. The softened cylinders with the metal rod were then placed in a glass tube and covered with Parafilm ® to slow the evaporation of the solvent while being transported from the laboratory to the workshop (time between softening and pressing).
Steps 4 and 5. At the workshop, the strategies used to preheat the metal mold to the set temperature in the slap press, fill the mold, and insert it in the slap press (preheating time, pressing time, and pressure) were the same as for the tines (steps 6 and 7).
Steps 6, 7 and 8. Three methods of cooling, as for the tines, were tested. After removal from the mold, the flash marks around the mock-ups were detached, and shorter cylinders of 3 cm in length were cut with a saw.
Step 9. The surface of the cylinders was polished with different grades of Micro-Mesh ® sanding sheets to create a texture comparable to historical objects [26].
Based on the visual assessment of the produced mock-ups, the optimal working parameters for compression molding were selected ( Table 5). The best conditions for producing tines and cylinders were achieved by using cut sheets, softening them for one day, preheating metal mold in the slap press for 5 (tines) or 15 (cylinders) minutes, and pressing at 25 bar for 8 min at 90 °C (Table 5).  Table 5. Working parameters tested for producing tines and cylinders. The optimal working parameters are highlighted in bold.

Tines
Steps 1, 2, and 3. The contour was carved on the prefabricated celluloid sheets selected as raw materials, then cut with scissors.
Step 4. Five precut geometries were layered over each other, exceeding the desired thickness by ca. 20 vol.% to ensure complete filling of the mold after pressing. Acetone was added between layers in a ratio of approximately 1:1 (1 g celluloid to 1 mL acetone) to soften the celluloid precut geometries.
Step 5. The softened geometries were placed in petri dishes and covered with Parafilm ® to slow the evaporation of the solvent while being transported from the laboratory to the workshop (time between softening and pressing).
Steps 6 and 7. At the workshop, the metal mold was preheated to the set temperature in the slap press. When the desired temperature was reached, the mold was removed, and the softened geometries were inserted into the preheated mold. Then, the mold was re-replaced into the slap press. The mold stayed slightly open during preheating without pressure to allow almost complete solvent evaporation. Finally, the mold was closed for a certain pressing time under the pressure of 25 bar.
Step 8. Three main methods of cooling were available (Table 5).
Step 8 in Figure 4 displays the cooling in water at approximately 20 • C.
Step 9. After cooling, the mold was opened, and the mock-ups were removed. In the final step, the excess material around the mock-ups (also called flash [50]) was removed by hand or with scissors, and the surface was polished to create a texture comparable to historical objects [26].
Cylinders Steps 1 and 2. As for the tines, the contour was carved on the prefabricated celluloid sheets selected as raw materials, then cut with scissors.
Step 3. The geometries were layered over each other, forming the two halves of the cylinder diameter, as displayed in Figure 1. Following the same strategy as for the tines, acetone was added between layers to soften the celluloid in a ratio of approximately 1:1. After softening, a metal rod was inserted between the two halves. The softened cylinders with the metal rod were then placed in a glass tube and covered with Parafilm ® to slow the evaporation of the solvent while being transported from the laboratory to the workshop (time between softening and pressing).
Steps 4 and 5. At the workshop, the strategies used to preheat the metal mold to the set temperature in the slap press, fill the mold, and insert it in the slap press (preheating time, pressing time, and pressure) were the same as for the tines (steps 6 and 7).
Steps 6, 7 and 8. Three methods of cooling, as for the tines, were tested. After removal from the mold, the flash marks around the mock-ups were detached, and shorter cylinders of 3 cm in length were cut with a saw.
Step 9. The surface of the cylinders was polished with different grades of Micro-Mesh ® sanding sheets to create a texture comparable to historical objects [26].
Based on the visual assessment of the produced mock-ups, the optimal working parameters for compression molding were selected ( Table 5). The best conditions for producing tines and cylinders were achieved by using cut sheets, softening them for one day, preheating metal mold in the slap press for 5 (tines) or 15 (cylinders) minutes, and pressing at 25 bar for 8 min at 90 • C (Table 5).
In sum, 100 tines and 100 cylinders were processed, and 70% were successfully made with nearly no visible defects (Table S1). From those, 40 mock-ups per geometry were later artificially aged. The remnant unaged 30% mock-ups with high and intermediate defects were rejected. Details of the defects are presented as Supplementary Materials (Table S2).

Assessing Whether the Processing Methods Induced Molecular Changes in Tines and Cylinders
The unaged tines and cylinders were investigated by ATR-FTIR to understand whether the processing could have induced unwanted heterogeneity. The values for the sheets depicted in Table 6 give the maximum allowable SD values of the absorbance intensities from the instrumental and sampling errors resulting from experiments A-C. Table 6. Normalized absorbance intensities of the ATR-FTIR probe bands of sheets, tines, and cylinders. The sheets, neither manipulated by the shaping process nor the aging, set the maximum allowable SD value for normalized absorbance intensities for each ATR-FTIR probe band calculated from experiments A-C. On one hand, all SD values attributed to the nitro groups (1634, 1273, 826 cm −1 ) were lower than the maximum allowable SD value, suggesting that no variation was detected by ATR-FTIR due to the processing of the tines and cylinders (Table 6). In addition, no significant differences were observed in the mock-ups considering the sampling locations, as the values from the top to the base for the tines or middle to edges for the cylinders were comparable.
On the other hand, the SD values of the carbonyl band (1730 cm −1 ) in the cylinders exceeded the maximum allowable limit. Steward [36] already recognized in his experiments that the carbonyl peak attributed to camphor generally shows poorer precision than the nitrate peaks. He also stated that one reason for the carbonyl band's high variability could be the volatile nature of camphor. As such, the handling of the samples (e.g., variation in the duration of sampling and pressure during measurement on the crystal) very likely played a significant role. From these signs, the carbonyl peak could not be considered an adequate probe band to assess the possible influences due to the processing of the mock-ups.

Artificial Aging of Mock-Ups
This study used thermohygrometric artificial aging to accelerate deterioration processes in the newly prepared mock-ups. The cumulative effects of the aging were visually monitored to achieve a moderate deterioration stage of the mock-ups. Supposing that the mock-ups react differently to the artificial aging depending on their shape because of the different surface area to volume ratio and processing method, their results are presented separately.
Sheets. After 32 days of exposure, different degradation phenomena were observed ( Figure S2): 20 just lost their transparency (Figure 6a), 10 sheets were additionally cracked and deformed (Figures 6b,d and S2), while 50 showed severe signs of degradation (Figures 6c,e and S2), including loss of transparency, cracking, deformations, and bubble formation. The formation of bubbles is not a characteristic degradation phenomenon for celluloid, whereas cracks are typical of severe damage [19][20][21]29,44,51]. Aiming at mock-ups in moderate conditions with no cracks and bubbles, the 20 sheets exhibiting loss of transparency were further tested in this investigation.
Tines. After 17 days, all mock-ups appeared slightly yellowed, and incipient bubbles appeared in only one sample of the 40 aged ones. Only two days after, under the same conditions of artificial aging, mock-ups appeared heavily deformed, cracked, expanded and bubbled (Figure 7). Because bubbles are an unrealistic degradation phenomena, only the mock-ups aged for 17 days were considered for further analyses.
6c,e and S2), including loss of transparency, cracking, deformations, an mation. The formation of bubbles is not a characteristic degradation pheno luloid, whereas cracks are typical of severe damage [19][20][21]29,44,51]. Aimin in moderate conditions with no cracks and bubbles, the 20 sheets exhibitin parency were further tested in this investigation.  Tines. After 17 days, all mock-ups appeared slightly yellowed, and incipient bubbles appeared in only one sample of the 40 aged ones. Only two days after, under the same conditions of artificial aging, mock-ups appeared heavily deformed, cracked, expanded and bubbled (Figure 7). Because bubbles are an unrealistic degradation phenomena, only the mock-ups aged for 17 days were considered for further analyses. Cylinders. The mock-ups presented the first signs of degradation after 27 days, such as yellowing and shrinkage (Figure 8). The first bubbling appeared only in one of the 40 mock-ups. A severely degraded stage was reached after 57 days in which some mock-ups appeared cracked, highly deformed, shrunk, and bubbled. Evident corrosion of the metal core was also visible. Following the same rationale as for the sheets and tines, bubbled and cracked samples were discarded, and only the 27-day-aged mock-ups considered to be in moderate condition were further investigated. Cylinders. The mock-ups presented the first signs of degradation after 27 days, such as yellowing and shrinkage (Figure 8). The first bubbling appeared only in one of the 40 mock-ups. A severely degraded stage was reached after 57 days in which some mockups appeared cracked, highly deformed, shrunk, and bubbled. Evident corrosion of the metal core was also visible. Following the same rationale as for the sheets and tines, bubbled and cracked samples were discarded, and only the 27-day-aged mock-ups considered to be in moderate condition were further investigated.
Cylinders. The mock-ups presented the first signs of degradation after 27 days, such as yellowing and shrinkage (Figure 8). The first bubbling appeared only in one of the 40 mock-ups. A severely degraded stage was reached after 57 days in which some mock-ups appeared cracked, highly deformed, shrunk, and bubbled. Evident corrosion of the metal core was also visible. Following the same rationale as for the sheets and tines, bubbled and cracked samples were discarded, and only the 27-day-aged mock-ups considered to be in moderate condition were further investigated. The tested environmental conditions caused bubbling in all mock-up geometries (sheets, tines, cylinders) after longer aging times, which is an unexpected degradation phenomenon. Even though bubbling is not common in celluloid degradation, the occurrence of cracks in the inner sample, is a typical macroscopic alteration of severely naturally aged objects [52]. A preliminary aging experiment yielded intense bubbling after 8 days for tines and 11 days for cylinders at 70 °C and 75% RH (more information for tines is available in Figure S1 and related paragraph). Therefore, the temperature selected to age The tested environmental conditions caused bubbling in all mock-up geometries (sheets, tines, cylinders) after longer aging times, which is an unexpected degradation phenomenon. Even though bubbling is not common in celluloid degradation, the occurrence of cracks in the inner sample, is a typical macroscopic alteration of severely naturally aged objects [52]. A preliminary aging experiment yielded intense bubbling after 8 days for tines and 11 days for cylinders at 70 • C and 75% RH (more information for tines is available in Figure S1 and related paragraph). Therefore, the temperature selected to age the thicker mock-ups (tines and cylinders) was lowered to 60 • C. Nonetheless, unusual bubbles were observed at T = 60 • C (for tines and cylinders) and 70 • C (for sheets) at longer aging times. In a previous study by the authors [52], the industrial manufacturing by Rothko and Frost™ seemed to have strongly influenced the formation of bubbles and extreme volume expansion at the center of the samples made of 1 mm-thickness prefabricated sheets (made of two 0.5 mm blended sheets) due to aging at 70 • C. The 0.5 mm prefabricated sheets used in this study were not industrially layered by the company and the role played by the industrial manufacturing can be excluded. In addition, the processing method of layering used in this investigation (see Sections 2.3. and 3.1.) did not induce volume expansion on the aged samples. Altogether, the combination of high temperatures and long aging times seemed to be the primary cause of bubbling on the mock-ups derived from the 0.5 mm prefabricated sheets, rather than the properties of the material (chemical/physical properties and formulation) or its industrial manufacturing.
To conclude, artificial aging at shorter aging times induced characteristic degradation phenomena of moderate conditions, such as loss of transparency, yellowing and slight deformation, making those mock-ups selectable for further analyses in this study.

Mass Loss
An evident mass loss after artificial aging was detected ( Table 7). The mass loss at different aging times and their trend are available in the Supplementary Materials (Table S3 and Figure S4). It is known that while celluloid degrades, it emits degradation products. In the initial phase, oxides of nitrogen (NO x ) form [12,14,16,23,48,49,[53][54][55], camphor sublimates [16,21,22,39,56], and gaseous decay products resulting from the cellulose matrix evolve [14,16,39]. The observed mass loss was consistent with the acid and camphor smell emission.

Off-Gassing Controlling with A-D Strips
To verify that the mass loss was also due to acidic emissions (Table 7), A-D strips were closely located to the aged mock-ups in closed bags. After 24 h, they turned yellow, confirming that the mock-ups were off-gassing. Their condition state not only fits the requirements of a moderate state of degradation for this investigation but is also common in historic celluloid collections [43,44]. Figures 9 and S4 show a decrease in the nitrate vibrations probe bands' intensities, except for the νNO 826 cm −1 solely for the sheets. The nitrate vibration probe bands for all other mock-ups showed a consistent but slight decrease along aging, which corroborates the assumption that incipient-induced degradation occurred and was likely due to the homolytic scission of the nitro groups (denitration) [12,48,49]. As presented by [52], the monitoring of the nitrate bands in ATR-FTIR can be used as a semiquantitative strategy to assess celluloid degradation, particularly when unaged and severely degraded samples are compared. In this study, ATR-FTIR showed limitations in estimating the degradation extent of the artificially aged mock-ups because of their moderate condition. The carbonyl peak at 1730 cm -1 showed an evident decrease after aging (e.g., mor than 34% for the sheets; Figure 10). This decrease might be partly attributed to the loss o the plasticizer camphor due to sublimation. However, the intensity of the carbonyl peak is not solely affected by the concentration of plasticizer but also by the extent of degrada tion of the polymer chain. Indeed, as cellulose nitrate degrades, carbonyl groups ar formed at the cellulose nitrate chain contributing to absorbance at 1730 cm -1 [33,48,52,54] The intensity of this band can cause a misleading interpretation regarding the loss of plas ticizer. A recent publication by the authors about the aging of similar prefabricated cellu loid samples by ATR-FTIR analysis showed an apparent increase in the C=O contribution due to severe deterioration [52]. Therefore, GPC was used for quantifying camphor in thi study. The carbonyl peak at 1730 cm −1 showed an evident decrease after aging (e.g., more than 34% for the sheets; Figure 10). This decrease might be partly attributed to the loss of the plasticizer camphor due to sublimation. However, the intensity of the carbonyl peak is not solely affected by the concentration of plasticizer but also by the extent of degradation of the polymer chain. Indeed, as cellulose nitrate degrades, carbonyl groups are formed at the cellulose nitrate chain contributing to absorbance at 1730 cm −1 [33,48,52,54]. The intensity of this band can cause a misleading interpretation regarding the loss of plasticizer. A recent publication by the authors about the aging of similar prefabricated celluloid samples by ATR-FTIR analysis showed an apparent increase in the C=O contribution due to severe deterioration [52]. Therefore, GPC was used for quantifying camphor in this study.

Assessment of Degradation via ATR-FTIR
formed at the cellulose nitrate chain contributing to absorbance at 1730 cm -1 [33,48,52,54]. The intensity of this band can cause a misleading interpretation regarding the loss of plasticizer. A recent publication by the authors about the aging of similar prefabricated celluloid samples by ATR-FTIR analysis showed an apparent increase in the C=O contribution due to severe deterioration [52]. Therefore, GPC was used for quantifying camphor in this study.  Table 8 summarizes the results obtained by IC and GPC. All unaged mock-ups are characterized by higher Mw, camphor, and nitrogen contents than the artificial aged ones. These results indicate that homolytic scission of nitro groups (denitration), reduction of the cellulose nitrate Mw (β-glucoside chain cleavage), and loss of the camphor took place.  Table 8 summarizes the results obtained by IC and GPC. All unaged mock-ups are characterized by higher M w , camphor, and nitrogen contents than the artificial aged ones. These results indicate that homolytic scission of nitro groups (denitration), reduction of the cellulose nitrate M w (β-glucoside chain cleavage), and loss of the camphor took place. In general, the M w displayed the most significant differences between unaged and aged mock-ups, with a decline of 59.0% (sheets), 19.8% (tines), and 33.5% (cylinders) compared to the unaged values. Even though it is commonly agreed in the literature that the initial step of CN degradation involves the homolytic scission of the O-NO 2 due to its low bonding energy [57,58], as depicted by Neves et al. [48], the reduced M w values suggest that the mock-ups in moderate condition were subjected to intense β-glucoside chain cleavage. The predominance of the chain cleavage compared to denitration in celluloid degradation has already been observed [23,59].

Comparison of the Mock-Ups with 3D Naturally Aged Objects
Nitrogen contents (used for calculating the degree of substitution-DS), M w (employed for determining the degree of polymerization-DP), and camphor contents of 3D naturally aged objects made of celluloid are seldom reported in the literature. In contrast, extensive literature data are available for 2D cinematographic films and photographs, which is hardly comparable with the aging behavior of 3D mock-ups. In this study, the DS and DP of the artificially aged mock-ups were compared with data of 3D naturally aged celluloid objects: (a) published in [22,23] and (b) measured in eyeglasses and plates from the DM. The DS and DP were calculated according to [23] and plotted in Figure 11. Based on [23,44], four main categories of condition state were adapted and used in this study for the mock-ups and the 3D naturally aged objects: • green color code: visibly in good condition without acid detection • yellow color code: visibly in good condition but with acid detection • red/purple color codes: visibly in poor condition, exhibiting signs of degradation and acids detection. Data plotted in red and purple can derive from samples taken from the same object, displaying the heterogeneous nature of celluloid degradation. Such decay pattern has been already reported in collections [16,19,21,23,27,43,52]. Red corresponds to areas where macroscopically no physical damage appeared, while purple refers to areas that are physically damaged, cracked, or crazed.
Polymers 2023, 15, x FOR PEER REVIEW 18 of 23 Figure 11. Degree of substitution (DS) and degree of polymerization (DP) of the mock-ups (□ sheets, ∆ tines, o cylinders) and 3D naturally aged celluloid objects: from the DM collection (◊), from [23] (x) and from [22] (+). The condition state of the mock-ups and samples from [22,23] were classified in green (good, no off-gassing); yellow (good, off-gassing); red (severe, areas physically not damaged); and purple (severe, areas physically damaged), adapted from [44] and [23]. Purple and red symbols are connected with a solid line when they derive from the same object.
The results indicate that the unaged mock-ups have the highest DP, directly related to a high Mw. It is important to notice that their DP absolute values are hard to corroborate with historical sources from the first half of the 20th century, when celluloid was mainly produced, as size exclusion chromatography was only invented in 1955 and no information on the Mw before that time is available. The raw material, manufacturing, and processing conditions strongly influence the final Mw of cellulose nitrate. Considering literature from 1968, DP can considerably vary from 100 to 1300 [60].
The aged mock-ups show DP values between ca. 500 and 900, which is slightly higher than the DP values of naturally aged objects in moderate conditions (400-700) considering the classification by [23,44]. In comparison, the historical objects classified as being in poor condition are characterized by the lowest DP values.
Considering the DS values, it is hard to recognize a clear trend between moderately (1.5-1.2) and severely aged (1.3-0.4) artifacts. However, Figure 11 clearly emphasizes the heterogeneous nature of severely degraded celluloid within the same object.
Even though the Mw and DS values of the historical objects after production and their extent of decrease due to aging are unknown, a trend in the reduction of both DP and DS can be suggested along the aging, which is underlined by the data plotted from the severely naturally aged objects ( [22,23] and DM collection). If the DP and DS values of all those historical objects in purple and red are representative of highly degraded historical objects, then one can infer that the artificially aged mock-ups are in a moderate condition because the values are ranked between unaged and severely aged celluloid ( Figure 11). and from [22] (+). The condition state of the mock-ups and samples from [22,23] were classified in green (good, no off-gassing); yellow (good, off-gassing); red (severe, areas physically not damaged); and purple (severe, areas physically damaged), adapted from [23,44]. Purple and red symbols are connected with a solid line when they derive from the same object. Based on the previously described categories, the unaged mock-ups were classified as green, the artificially aged ones as yellow, and the naturally aged objects from the DM as purple. The data from [22], described in that article as being collected from very or highly degraded objects (with cracks and disintegrating in several pieces), were here classified as purple.
The results indicate that the unaged mock-ups have the highest DP, directly related to a high M w . It is important to notice that their DP absolute values are hard to corroborate with historical sources from the first half of the 20th century, when celluloid was mainly produced, as size exclusion chromatography was only invented in 1955 and no information on the M w before that time is available. The raw material, manufacturing, and processing conditions strongly influence the final M w of cellulose nitrate. Considering literature from 1968, DP can considerably vary from 100 to 1300 [60].
The aged mock-ups show DP values between ca. 500 and 900, which is slightly higher than the DP values of naturally aged objects in moderate conditions (400-700) considering the classification by [23,44]. In comparison, the historical objects classified as being in poor condition are characterized by the lowest DP values.
Considering the DS values, it is hard to recognize a clear trend between moderately (1.5-1.2) and severely aged (1.3-0.4) artifacts. However, Figure 11 clearly emphasizes the heterogeneous nature of severely degraded celluloid within the same object.
Even though the M w and DS values of the historical objects after production and their extent of decrease due to aging are unknown, a trend in the reduction of both DP and DS can be suggested along the aging, which is underlined by the data plotted from the severely naturally aged objects ( [22,23] and DM collection). If the DP and DS values of all those historical objects in purple and red are representative of highly degraded historical objects, then one can infer that the artificially aged mock-ups are in a moderate condition because the values are ranked between unaged and severely aged celluloid ( Figure 11).
Camphor contents in celluloid have mainly been mainly addressed by IR spectroscopy [13,22,61], which is a not straightforward strategy due to concurrent influences on the carbonyl band vibration. In fact, camphor loss and formation of carbonyl-containing degradation products [52] can both at the same time affect the intensity of this vibration due to aging.
In this study, the camphor content of artificially aged mock-ups and naturally aged objects calculated by a quantitative method (GPC) are compared. The camphor contents of the sheets, tines, and cylinders decreased due to aging 25%, 12%, and 8%, respectively ( Figure 12). The aged sheets displayed the most significant loss. Ca. 16% of camphor was detected in the naturally aged objects from the DM. Bearing in mind that the highest camphor value for 3D objects is around 33% [16,21,62,63], the artificially aged mock-ups would resemble moderately aged objects.

Conclusions
This work emphasizes the need and advantages of processing mock-ups in celluloid conservation research with shapes resembling 3D historical artifacts. The 3D celluloid mock-ups, namely, sheets, tines, and cylinders, referring to swatches, combs, and eyeglass temples, were successfully made by combining historical industrial processing strategies and small-scale lab facilities. Different compression molding variables and combinations were tested, which were crucial for determining the best processing conditions, leading to 70% satisfactory results. After processing, tines and cylinders were analyzed by ATR-FTIR, and the nitro group probe bands showed no detectable variation due to shaping, suggesting that the developed protocol (layering, softening, compression molding and cooling) enabled their reproducible creation. The processing protocol developed in this study can be considered adequate for creating celluloid mock-ups in plastic heritage studies, to investigate not only their aging behavior but also to explore the effectiveness of passive and active conservation treatments.
The artificial aging experiments were adequate for inducing aged mock-ups (being yellowed and off-gassing) capable of simulating moderately degraded historical objects, as verified by comparing published and original data from chemical analyses of 3D objects from collections. Besides detecting acidic emissions, IC and GPC confirmed denitration, loss of camphor, and polymer chain scissions. The chain cleavage was predominant com-

Conclusions
This work emphasizes the need and advantages of processing mock-ups in celluloid conservation research with shapes resembling 3D historical artifacts. The 3D celluloid mock-ups, namely, sheets, tines, and cylinders, referring to swatches, combs, and eyeglass temples, were successfully made by combining historical industrial processing strategies and small-scale lab facilities. Different compression molding variables and combinations were tested, which were crucial for determining the best processing conditions, leading to 70% satisfactory results. After processing, tines and cylinders were analyzed by ATR-FTIR, and the nitro group probe bands showed no detectable variation due to shaping, suggesting that the developed protocol (layering, softening, compression molding and cooling) enabled their reproducible creation. The processing protocol developed in this study can be considered adequate for creating celluloid mock-ups in plastic heritage studies, to investigate not only their aging behavior but also to explore the effectiveness of passive and active conservation treatments.
The artificial aging experiments were adequate for inducing aged mock-ups (being yellowed and off-gassing) capable of simulating moderately degraded historical objects, as verified by comparing published and original data from chemical analyses of 3D objects from collections. Besides detecting acidic emissions, IC and GPC confirmed denitration, loss of camphor, and polymer chain scissions. The chain cleavage was predominant compared to denitration. The moderately aged mock-ups are currently being investigated within a research project [15], which is dedicated to systematically assess the effectiveness of cool, cold, and frozen storage conditions for celluloid 3D objects without testing unique and valuable historical artifacts.
This work paves the way for the use of 3D celluloid mock-ups in heritage science studies and can inspire their future application in research on celluloid. In this regard, mock-ups with different material compositions (e.g., different plasticizer concentrations and filler contents) and geometries would support aging and conservation studies. In addition, it opens new avenues for investigating unresolved research questions about the complex decay processes of 3D celluloid. A heterogeneous nature at advanced stages of alteration has been observed for 3D celluloid objects, leading to degradation gradients recognizable by the naked eye [12,15,17,19,35,44,52]. Even though this decay phenomenon is considered typical, it has been studied very little until now. Considering mock-ups as reliable testing materials, research on severely degraded 3D celluloid mock-ups would support a better understanding of the nature of this decay. To this end, mock-ups can be systematically aged under different environmental conditions to simulate historical objects in poor conditions. In doing so, it is advisable to consider the glass-transition temperature (Tg) of celluloid in selecting the temperature of the artificial aging protocol to avoid unwanted degradation phenomena (e.g., bubbling). The chemical characterization of mock-ups in comparison with naturally aged historical objects through spectroscopic and chromatographic methods is essential for verifying their ability to simulate the state of 3D celluloid in collections. The availability of mock-ups simulating different degradation stages will serve to explore conservation treatments, e.g., cleaning, consolidation, planning of proper showcases in exhibitions and tailored storage solutions, towards a better preservation of complex 3D celluloid artifacts for future generations.
Supplementary Materials: The following supporting information can be downloaded at https:// www.mdpi.com/article/10.3390/polym15040852/s1, Figure S1: Preliminary trial of artificial aging of tines at 70 • C and 75 % RH after 6, 7 and 8 days. Bubbling was clearly visible after 7 days and increased excessively after 8 days; Figure S2: Deformations of the sheets after 32 days of exposure to 70 • C and 75 % RH. Deformations were observed at sheets showing: (a) loss of transparency and cracking and (b) loss of transparency, cracking and bubbling; Figure S3: Diagram addressing the time-dependent mass loss of mock-ups; Figure S4: Averaged infrared spectra before and after artificial ageing of (a) sheets (n = 20) (b) tines (n = 24) and (c) cylinders (n = 24). Arrows indicate a decrease in the probe bands after artificial ageing; Table S1: Overview of the rejected and selected mock-ups based on their production defects; Table S2: Production defects that appeared during the preparation of mock-ups; Table S3: Calculated mass loss in % during artificial ageing of the mock-ups shaped as sheets. Tines and cylinders.