Characterization of Ultraviolet-Degraded Polyethylene Terephthalate Film Using a Complementary Approach: Reactive Pyrolysis–Gas Chromatography–Mass Spectrometry and Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry

Abstract

Polyethylene terephthalate (PET) is widely used across various industries owing to its versatility and favorable properties, including application in beverage bottles, food containers, textile fibers, engineering resins, films, and sheets. However, polymer materials are susceptible to degradation from factors such as light, oxygen, and heat. Therefore, it is crucial to understand the structural changes that occur during degradation and the extent of these changes. This report investigates the structural alterations in PET films resulting from ultraviolet (UV) irradiation utilizing pyrolysis–gas chromatography time-of-flight mass spectrometry (Py-GC-TOFMS) and matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry (MALDI-TOFMS). Using the reactive Py-GC-TOFMS, we estimated the composition of the pyrolysis products resulting from UV degradation through electron ionization, soft ionization, and exact mass measurements. Additionally, artificial intelligence (AI)-based structure analysis was performed to evaluate these compounds’ structures. Notably, most degradation products were not found in the National Institute of Standards and Technology database, underscoring the effectiveness of our approach. Using MALDI-TOFMS analysis, we determine the changes in the end groups before and after UV irradiation. This analysis confirmed the generation of a series of carboxylic acid end groups as a result of degradation, a polymer series not detected by reactive pyrolysis GC-MS. We also explored degradation in the depth direction, demonstrating that degradation progresses gradually to depths of several micrometers. Our findings highlight the importance of employing mass spectrometry techniques for a comprehensive analysis of polymer degradation.

1. INTRODUCTION

Polyethylene terephthalate (PET) is widely used across various industries owing to its versatility and favorable properties, including application in beverage bottles, food containers, textile fibers, engineering resins, films, and sheets. However, polymer materials are susceptible to degradation from factors such as light, oxygen, and heat. Therefore, it is crucial to understand the structural changes that occur during degradation and the extent of these changes. Conducting degradation analysis on PET products is vital for ensuring quality control and maintaining long-term performance. Additionally, such analysis plays a key role in assessing environmental impact and promoting recyclability, thereby contributing to sustainable product development efforts.

Pyrolysis–gas chromatography–mass spectrometry (Py-GC-MS) and matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry (MALDI-TOFMS) are effective methods for polymer characterization using mass spectrometry. In Py-GC-MS, the majority of pyrolysis products are monomers or dimers of the repeating units, facilitating the identification of polymer species and enabling effective analysis of structural changes in the polymer backbone.¹⁾ MALDI, a widely used soft ionization method, efficiently ionizes intact polymer molecules. Unlike the electrospray method, which often generates multiply charged ions, MALDI predominantly produces singly charged ions even for high molecular weight compounds. This characteristic simplifies the interpretation of synthetic polymers, as the m/z value of singly charged ions corresponds directly to the mass of the polymer ions, allowing for straightforward analysis of molecular weight distribution. With high mass resolution, MALDI-TOFMS can identify polymer series based on variations in repeating units and end groups by estimating elemental composition through accurate mass analysis. Because Py-GC-MS and MALDI-TOFMS employ different approaches to polymer characterization, complementary analysis is beneficial for the detailed structural characterization of polymers.²⁾

This report investigates the structural alterations in PET film resulting from ultraviolet (UV) irradiation. Analyzing condensation polymers with ester bonds in the main chain, such as PET, poses challenges when using Py-GC-MS, as the highly polar pyrolysis products often produce broad peaks in the pyrogram.³⁾ To address this issue, reactive Py-GC-MS is frequently employed, utilizing pyrolysis in the presence of an organic alkali, such as tetramethylammonium hydroxide ((CH₃)₄NOH, TMAH). This method is effective for analyzing condensation polymers, as it simultaneously hydrolyzes ester bonds and converts them into methyl derivatives, thereby minimizing side reactions.^4–6) In GC-MS analysis, compound identification is typically performed through database (DB) searches using mass spectra obtained from electron ionization (EI).⁷⁾ The latest EI mass spectral DB published by the National Institute of Standards and Technology (NIST), NIST23, contains mass spectra for 347,100 compounds.⁸⁾ However, many pyrolysis products obtained from Py-GC-MS and reactive Py-GC-MS lack standards and are not included in the NIST DB. We refer to these compounds as “unknown compounds.” To address this, we propose an “integrated analysis” approach that estimates the elemental composition of unknown compounds using mass spectra obtained from both EI and soft ionization, along with their exact masses.⁹⁾ Additionally, we have developed an “artificial intelligence (AI) structure analysis method” that employs machine learning to predict EI mass spectra based on compound structures.¹⁰⁾ The machine learning model was trained on the relationships between structures and mass spectral patterns in the NIST DB, resulting in a predicted EI mass spectra library for over 100 million compounds listed in PubChem. Combining the “integrated analysis” and “AI structure analysis method” makes it possible to estimate the structure of unknown compounds. First, the elemental composition of the unknown compound is estimated from the accurate mass of the peaks in the mass spectra using EI and soft ionization. Next, only structural isomers with this elemental composition are extracted from the predicted EI mass spectra library. Finally, the structure is estimated by calculating the similarity, hereafter called AI score, between the predicted EI spectra of the isomers and the observed EI mass spectra.

Using MALDI-TOFMS, we investigated the structural changes in PET resulting from UV irradiation. A previous study¹¹⁾ demonstrated that degradation primarily occurs within 1.3 μm of the sample surface, as shown by attenuated total reflection–Fourier transform infrared (ATR-FT-IR) analysis. The degree of degradation at the surface was assessed by measuring the amount of carboxylic acid generated during the degradation process. In this study, we utilized MALDI-TOFMS to compare the changes in PET end groups in relation to UV irradiation time, focusing on a bulk sample. Additionally, we measured the sections at depths of 0–1 μm and 2–4 μm to investigate the progression of degradation in the depth direction.

2. EXPERIMENTAL

2.1 Sample preparation

A commercially available PET film (thickness: 25 μm) was used as the test sample. This film was irradiated with UV light using a UV curing device (HLR100T-2, SEN LIGHTS Corporation, Osaka, Japan; primary wavelength: 365 nm, intensity: 170 mW/cm²) under ambient conditions. For GC-MS measurements, samples irradiated for 0, 1, 2, 6, and 8 h were analyzed. For MALDI-TOFMS measurements, the bulk analysis included samples irradiated for 0, 8, and 54 h, while depth direction analysis focused on samples irradiated for 0 and 4 h. In the depth direction analysis, samples were sectioned in the 0–1-μm and 2–4-μm regions using an ultramicrotome (EM UC7, Leica, Wetzlar, Germany) equipped with a glass knife, with a cutting thickness of 0.5 μm.

2.2 GC-TOFMS measurement

Sample measurements were conducted using a gas chromatograph–time-of-flight mass spectrometer (JMS-T2000GC AccuTOF^TM GC-Alpha, JEOL Ltd., Tokyo, Japan), combined with a pyrolyzer (EGA/PY-3030D, Frontier Laboratories, Fukushima, Japan). In this study, GC-TOFMS measurements were performed using both EI and field ionization (FI) with a combined EI/FI ion source. For the measurements, 0.2 mg of the sample for EI and 0.5 mg for FI were combined with 10 μL of a 25% TMAH aqueous solution (Thermo Fisher Scientific, Waltham, MA, USA) in a deactivated stainless-steel sample cup (Frontier Laboratories). The mixture was pyrolyzed at 400°C under a flow of helium carrier gas. A relatively low pyrolysis temperature of 400°C was employed to prevent the formation of side products resulting from the pyrolysis of polymer chains at higher temperatures. The carrier gas flow rate of 100 mL/min at the pyrolyzer was reduced to 1 mL/min at the capillary column (DB-5MS UI, 30 m × 0.25 mm, 0.25 μm, Agilent Technologies, Santa Clara, CA, USA) using a splitter. The temperature of both the Py/GC interface and the GC injection port was maintained at 320°C. The column temperature was held at 40°C for 2 min, then increased to 320°C at a rate of 20°C/min, with the final temperature held for an additional 10 min. The ionization energy was set to 70 eV for EI, while the cathode voltage was set to −10 kV for FI. Qualitative data processing was performed using msFineAnalysis AI version 2 (JEOL Ltd.). Following data acquisition, a single-point or multi-point external mass drift compensation was performed for accurate mass calibration. In EI mode, a column bleed peak corresponding to C₇H₂₁O₄Si₄⁺ (calculated mass: 281.05114 u) was employed. In FI mode, octamethylcyclotetrasiloxane (Tokyo Chemical Industry, Tokyo, Japan) was introduced into the mass spectrometer at approximately 15-minute intervals, and its fragment ion, C₇H₂₁O₄Si₄+ (calculated mass: 281.05114 u) was employed.

2.3 MALDI-TOFMS measurement

In the bulk analysis using MALDI-TOFMS, PET films were dissolved in hexafluoro-2-propanol (HFIP, Kanto Chemical, Tokyo, Japan) at a concentration of 10 mg/mL. For the depth direction analysis, sections of PET films from the sample surfaces at 0–1 μm and 2–4 μm were prepared using an ultramicrotome and placed into vials. To each vial, 20 μL of HFIP was added to dissolve the samples. For the matrix, 2,4,6-trihydroxyacetophenone (THAP, Sigma-Aldrich, St. Louis, MO, USA) was dissolved in tetrahydrofuran (THF, Fujifilm Wako Pure Chemical Corp., Osaka, Japan) at a concentration of 10 mg/mL. Sodium trifluoroacetate (NaTFA, Sigma-Aldrich) in THF was used as the cationizing reagent at a concentration of 1 mg/mL. The THAP, sample, and NaTFA were mixed in a ratio of 10/1/1 (v/v/v). Approximately 1 μL of the mixed solution was deposited on the target plate and allowed to air dry. Mass spectra were acquired using the MALDI time-of-flight mass spectrometer JMS-S3000 “SpiralTOF^TM-plus 3.0” (JEOL Ltd.).¹²⁾ In the mass calibration, the cyclic oligomers in the mass spectrum before degradation were assigned using polymethyl methacrylate. Then, each mass spectrum was calibrated using the mass of the cyclic oligomer.

3. RESULTS AND DISCUSSION

3.1 GC-TOFMS measurement

To investigate the structural changes caused by UV irradiation, we compared the Py-GC-TOFMS data of PET before and after 10 h of UV irradiation. Figure 1 displays the total ion current chromatogram (TICC) obtained from EI measurements, with the blue representing data before UV irradiation and the red TICC representing data after irradiation. Three common compounds (peaks A, B, and C) were prominently detected before and after UV irradiation. A search in the NIST DB identified these compounds as follows: (A) 2-methoxyethanol (match factor: 926), (B) 1,2-dimethoxyethane (match factor: 762), and (C) dimethyl benzene-1,4-dicarboxylate (match factor: 851). Peaks A and B correspond to methyl derivatives from ethylene glycol, while peak C is a methyl derivative from terephthalic acid, a component of PET.¹³⁾

Fig. 1. TICC obtained through EI measurement. The blue trace represents the sample before UV irradiation, while the red trace indicates the sample after UV irradiation. EI, electron ionization; TICC, total ion current chromatogram; UV, ultraviolet.

Next, to investigate the minor changes induced by UV irradiation, we performed a statistical difference analysis¹⁴⁾ on the data collected before and after UV irradiation. This analysis utilized EI data from 3 trials for samples before and after 10 h of UV irradiation. Peaks in the chromatogram with relative heights of 0.20% or more compared to peak C, that is, peak height 4.0 × 10⁻⁵ or more in Fig. 1, were included in the difference analysis. Compounds exhibiting double ion intensity ratios after UV irradiation compared to those before UV irradiation along with p-values from t-test variance analysis results of less than 0.05 were identified as characteristic compounds of UV irradiation. Thus, 11 characteristic compounds, ID: 001–011 in Table 1, were detected after UV irradiation. The peaks of ID: 004–011 were also indicated in the expanded TICC in Fig. 1. Table 1 shows the results of the NIST DB search and AI structure analysis estimation that scored high, with the method used being displayed in the “Library” column. The scores of the NIST DB search and AI structure analysis are called “match factor” and “AI score,” respectively. The NIST DB search considers a match factor of 900 or greater to be an excellent match; 800–900 is a good match; and 700–800 is a fair match.¹⁵⁾ Only 4 compounds (IDs: 002, 003, 006, and 010) had match factors exceeding 700 according to the NIST DB search. Consequently, the remaining 7 compounds were classified as unknowns. For compound ID: 008, which had the highest intensity among the characteristic compounds after UV irradiation at a retention time of 11.47 min, we estimated its structure and analyzed the change in ion intensity in relation to the irradiation time.

Table 1. Characteristic compounds identified after UV irradiation.

ID	RT (min)	Relative height (%)	Compound name	CAS#/PubChem CID	Library	Match factor or AI score	Formula	DBE	Measured molecular ion m/z	Calculated m/z	Mass error (mDa)	Ion intensity ratios	p-Value
001	3.83	0.55	2-(Methoxymethoxy)acetaldehyde	10931385	AI	873	C4 H8 O3	1.0	104.04650	104.04680	–0.30	>16	0.007
002	6.68	0.33	Butanedioic acid, dimethyl ester	106–65–0	NIST (replib)	741	C6 H10 O4	2.0	146.05611	146.05736	–1.25	>16	0.027
003	9.78	0.38	4-Methoxymethylbenzoic acid methyl ester	1719–82–0	NIST (mainlib)	919	C10 H12 O3	5.0	180.07787	180.07810	–0.23	>16	0.026
004	10.62	0.21	Methyl 4-formyl-3,5-dimethylbenzoate	89836469	AI	883	C11 H12 O3	6.0	192.07821	192.07810	0.11	>16	0.024
005	10.73	0.23	Methyl 4-(2-methoxyacetyl)benzoate	134494725	AI	889	C11 H12 O4	6.0	208.07340	208.07301	0.39	>16	0.013
006	10.79	0.34	2-Methylbenzene-1,4-dicarboxylic acid dimethyl ester	14186-60–8	NIST (mainlib)	940	C11 H12 O4	6.0	208.07294	208.07301	–0.07	>16	0.017
007	10.94	0.28	Methyl 4-butanoylbenzoate	23143286	AI	905	C12 H14 O3	6.0	206.09353	206.09375	–0.21	>16	0.005
008	11.47	2.11	Dimethyl 2-methoxybenzene-1,4-dicarboxylate	18677354	AI	832	C11 H12 O5	6.0	224.06781	224.06792	–0.11	>16	0.017
009	11.80	0.38	5-Ethyl-2-methoxy-3-methoxycarbonylbenzoic acid	117100460	AI	832	C12 H14 O5	6.0	238.08321	238.08357	–0.36	>16	0.021
010	12.15	0.44	1,2,4-Benzenetricarboxylic acid, trimethyl ester	2459–10–1	NIST (replib)	901	C12 H12 O6	7.0	252.06262	252.06284	–0.22	>16	0.027
011	12.23	0.57	Ethyl 4-(4-hydroxy-3-oxobutyl)benzoate	10609833	AI	691	C13 H16 O4	6.0	236.10388	236.10431	–0.43	>16	0.045

Figure 2 displays the EI and FI mass spectra for compound ID: 008. A search in the NIST DB using the EI mass spectrum identified ID: 008 as methyl 4,6-dihydroxy-2,3-dimethylbenzoate, dimethyl ether (C₁₂H₁₆O₄), with a match factor of 691. This indicates that compound ID: 008 is not registered in the NIST DB. The peak at m/z 224.06781 was determined to be the molecular ion based on the FI measurement. The elemental composition of this molecular ion was estimated to be C₁₁H₁₂O₅⁺ (calculated mass: 224.06792 u), with a mass error of −0.00011 u. Although both the top hit in the NIST DB search C₁₂H₁₆O₄ (calculated mass: 224.10431 u) and C₁₁H₁₂O₅ have integer masses of 224 u, the mass difference of 0.036 u was identified using accurate mass analysis from the FI mass spectrum. This confirmed that the top result, methyl 4,6-dihydroxy-2,3-dimethylbenzoate, dimethyl ether, in the NIST DB search was incorrect. Therefore, this study underscores the importance of accurate mass estimation in determining structural changes associated with oxidation.

Fig. 2. Mass spectra of compound ID: 008, which exhibited the highest intensity among the characteristic compounds after UV irradiation: (A) EI mass spectrum and (B) FI mass spectrum. EI, electron ionization; FI, field ionization; UV, ultraviolet.

Figure 3 presents the estimated structure of compound ID: 008. Figure 3A illustrates the results obtained from an AI structure analysis method that uses machine learning to predict the EI mass spectrum based on compound structure. The analysis includes the top 10 candidates ranked by AI scores, calculated using cosine similarity.¹⁰⁾ The structural estimation yielded candidates with benzene rings and methyl ester groups. Given the experimental conditions, it can be concluded that methyl derivatization with TMAH occurs, suggesting that carboxylic acid or hydroxyl groups are likely absent. Additionally, because the original polymer is PET, it is reasonable to assume that the compound has a methyl ester group or a similar structure at the para position. Therefore, the structure formula of ID: 008 estimated candidate 7, dimethyl 2-methoxybenzene-1,4-dicarboxylate. Previous research has indicated that aromatic hydroxylated species are generated through photooxidative reactions caused by UV irradiation of PET (Fig. 3B).¹⁶⁾ Thus, candidate 7 is presumed to be a reaction pyrolysis product of this compound. This compound may serve as a marker for degradation, allowing for the investigation of degradation levels in relation to UV irradiation time. Figure 4 shows the variation in peak area values for ID: 008 from the extracted ion chromatogram at m/z 193.0495 ± 0.01 u. The peak area increased linearly with UV irradiation time, allowing for the detection of changes as early as 1-hour post-irradiation. This highlights the advantage of the reactive Py-GC-MS method, which enables quantitative evaluation of the early stages of UV degradation through bulk sample analyses.

Fig. 3. Structure estimated result of ID: 008. (A) The structure analysis results by AI structure analysis method. Candidate 7 will be the reasonable structure from the structure change in the main chain by UV irradiation shown in reference 16. (B) shows aromatic hydroxylated species produced by photooxidation of PET by UV irradiation, as shown by Fechine et al.¹⁶⁾ AI, artificial intelligence; PET, polyethylene terephthalate; UV, ultraviolet.

Fig. 4. Variation in the peak area of compound ID: 008 from the extracted ion chromatogram at m/z 193.0495 ± 0.01 u. The bars represent the mean values ± SDs (n = 3).

3.2 MALDI-TOFMS measurement

Figures 5A–5C illustrates the changes in the mass spectrum of the entire sample before UV irradiation and after 8 and 54 h of UV irradiation, as analyzed by MALDI-TOFMS. In Fig. 5A, a series with 192-u intervals is observed, with the composition estimated by accurate mass to be C₁₀H₈O₄, representing the repeating unit of PET. The end-group composition analysis before UV irradiation revealed that the main polymeric series consisted of cyclic oligomers, as detected as [M+Na]⁺. In Fig. 5C, after 54 h of UV irradiation, multiple series with 192-u intervals were noted, indicating a shift mass due to changes in the end groups from cyclic oligomers. Accurate mass measurement suggested that the 3 primary polymer series (series B, C, and D) contained carboxylic acids at both end groups, observed as [M+Na]⁺. These polymeric series were identified as 3 distinct series in the mass spectra due to combinations of carboxylic acid end groups: COOH/COOH, COOH/COONa, and COONa/COONa. This series with carboxylic acid end groups likely results from the fragmentation of the PET main chain via a Norrish type II reaction.¹⁷⁾ Notably, these polymer series cannot be identified using reactive Py-GC-MS, as that method produces reactive pyrolysis products similar to the main chain structure prior to degradation. This highlights the effectiveness of end-group analysis via MALDI-TOFMS. Conversely, in Fig. 5B, after 8 h of UV irradiation, the series of carboxylic acids at both ends was not significantly observed, making it challenging to detect degradation in the early stages of UV degradation using bulk analysis. This capability is better suited for GC-TOFMS. In summary, reactive pyrolysis GC-TOFMS is effective for confirming the overall degradation amount of the sample.

Fig. 5. MALDI-TOFMS mass spectra of PET film. (A) Before UV irradiation and after (B) 8 h and (C) 54 h of UV irradiation. (D) Estimated structure of the main series generated by UV irradiation, featuring carboxylic acid groups at both end groups. MALDI-TOFMS, matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry; PET, polyethylene terephthalate; UV, ultraviolet.

Next, we examined the differences in mass spectra from sections at depths of 0–1 μm and 2–4 μm from the sample surface. Fechine et al.¹⁷⁾ reported that degradation occurs primarily within the first 0–1 μm of the sample surface, as indicated by ATR-FT-IR analysis. In this study, we confirmed whether degradation extends to deeper regions beneath the surface. Figure 6 presents the mass spectra before and after 4 h of UV irradiation. The mass spectra at depths of (a) 0–1 μm and (b) 2–4 μm before UV irradiation were nearly identical, primarily showing cyclic oligomer components. After 4 h of UV irradiation, the mass spectra at depths of (c) 0–1 μm and (d) 2–4 μm revealed polymeric series of cyclic oligomers (series A) and carboxylic acids at both end groups (series B, C, and D), similar to those observed in Fig. 5. Upon comparing the ratio of ion intensities of carboxyl acid end groups to cyclic oligomers, we found that the ratio at the 0–1-μm depth was higher than that at 2–4 μm. This suggests that the degree of degradation is more pronounced near the sample surface. Importantly, our findings confirm that degradation occurs even at depths greater than 1 μm, which ATR-FT-IR could not detect, confirming our approach’s effectiveness.

Fig. 6. MALDI-TOFMS mass spectra at varying depths: (A) 0–1 μm and (B) 2–4 μm before UV degradation, (C) 0–1 μm, and (D) 2–4 μm after 4 h of UV irradiation. Before UV degradation, the polymeric series of cyclic oligomers (series A) were predominantly observed. After UV degradation, the presence of carboxylic acids at both end groups (series B, C, and D) increased significantly. MALDI-TOFMS, matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry; UV, ultraviolet.

4. CONCLUSION

Various methods are available for characterizing polymers, each offering unique theoretical advantages. In this study on PET degradation, we utilized reactive Py-GC-TOFMS and an AI structure analysis method that employs machine learning to predict the EI mass spectrum based on compound structure. These methods provided detailed insights into structural changes and enabled discussions regarding the extent of degradation using bulk samples. MALDI-TOFMS revealed fragmentation of the main chain structure through changes in the end groups, a phenomenon not detectable with reactive Py-GC-MS. Furthermore, our analysis confirmed degradation at deeper positions beneath the surface, which ATR-FT-IR could not assess. By understanding the strengths of each method and combining them, we can obtain a more comprehensive understanding of polymer characterization.

Notes

Mass Spectrom (Tokyo) 2025; 14(1): A0168

REFERENCES

• 1)  S. Tsuge, H. Ohtani. Structural characterization of polymeric materials by pyrolysis-GC/MS. Polym. Degrad. Stabil. 58: 109–130, 1997.
• 2)  H. Ohtani, T. Iura. Complementary characterization of end groups in radically polymerized poly(methyl methacrylate) by matrix-assisted laser desorption/ionization mass spectrometry and pyrolysis-gas chromatography-mass spectrometry. Mass Spectrom. (Tokyo) 3(Special Issue 3): S0041, 2014.
• 3)  I. Coralli, D. Fabbri. Pyrolysis acetylation: A novel on-line Py-GC-MS derivatisation for the characterisation of nitrogen-containing polymers. J. Anal. Appl. Pyrolysis 175: 106178, 2023.
• 4)  J. M. Challinor. Review: The development and applications of thermally assisted hydrolysis and methylation reactions. J. Anal. Appl. Pyrolysis 61: 3–34, 2001.
• 5)  F. Shadkami, R. Helleur. Recent applications in analytical thermochemolysis. J. Anal. Appl. Pyrolysis 89: 2–16, 2010.
• 6)  N. Dimitrov, L. Kratofil Krehula, A. Ptiček Siročić, Z. Hrnjak-Murgić. Analysis of recycled PET bottles products by pyrolysis-gas chromatography. Polym. Degrad. Stabil. 98: 972–979, 2013.
• 7)  S. E. Stein. Estimating probabilities of correct identification from results of mass spectral library searches. J. Am. Soc. Mass Spectrom. 5: 316–323, 1994.
• 8)  W. E. Wallace, A. S. Moorthy. NIST Mass Spectrometry Data Center standard reference libraries and software tools: Application to seized drug analysis. J. Forensic Sci. 68: 1484–1493, 2023.
• 9)  M. Ubukata, A. Kubo, K. Nagatomo, T. Hizume, H. Ishioka, A. J. Dane, R. B. Cody, Y. Ueda. Integrated qualitative analysis of polymer sample by pyrolysis-gas chromatography combined with high-resolution mass spectrometry: Using accurate mass measurement results from both electron ionization and soft ionization. Rapid Commun. Mass Spectrom. 34: e8820, 2020.
• 10)  A. Kubo, A. Kubota, H. Ishioka, T. Hizume, M. Ubukata, K. Nagatomo, T. Satoh, M. Yoshida, F. Uematsu. Construction of a mass spectrum library containing predicted electron ionization mass spectra prepared using a machine learning model and the development of an efficient search method. Mass Spectrom. (Tokyo) 12: A0120, 2023.
• 11)  R. Watanabe, A. Sugahara, H. Shinzawa, S. Yamane, S. Nakamura, H. Sato, H. Hagihara, A. Oishi, J. Mizukado, Y. Ueda, T. Satoh. Photodegradation behavior of polyethylene terephthalate analyzed by MALDI-TOFMS and ATR-FTIR microscopic analysis in combination with two-trace two-dimensional (2T2D) correlation mapping. Polym. Degrad. Stabil. 208: 110246, 2023.
• 12)  T. Satoh, T. Sato, J. Tamura. Development of a high-performance MALDI-TOF mass spectrometer utilizing a spiral ion trajectory. J. Am. Soc. Mass Spectrom. 18: 1318–1323, 2007.
• 13)  S. Tsuge, H. Ohtani, C. Watanabe. Pyrolysis-GC/MS Data Book of Synthetic Polymers: Pyrograms, Thermograms and MS of Pyrolyzates. 1st Ed., Elsevier B.V., Amsterdam, 2011, pp. 356.
• 14)  R. B. Cody, T. Fukudome, M. Ubukata. Integrated data analysis making use of the total information from gas chromatography and high-resolution time-of-flight mass spectrometry to identify qualitative differences between two whisky samples. Rapid Commun. Mass Spectrom. 36: e9225, 2022.
• 15)  S. E. Stein. NIST Standard Reference Database 1A for use with Microsoft^® Windows User's Guide. https://www.nist.gov/system/files/documents/srd/NIST1a11Ver2-0Man.pdf.
• 16)  G. J. M. Fechine, R. M. Souto-Maior, M. S. Rabello. Photodegradation of multilayer films based on PET copolymers. J. Appl. Polym. Sci. 104: 51–57, 2007.
• 17)  G. J. M. Fechine, M. S. Rabello, R. M. Souto Maior, L. H. Catalani. Surface characterization of photodegraded poly(ethylene terephthalate). The effect of ultraviolet absorbers. Polymer (Guildf.) 45: 2303–2308, 2004.