Effects of Reprocessing on Acrylonitrile–Butadiene–Styrene and Additives

Acrylonitrile–butadiene–styrene (ABS) is one of the most extensively used engineering polymers. It is necessary to study the recycling of ABS because of environmental, economic and energy reasons. In this study, an ABS resin was processed using a torque rheometer at different temperatures and for different numbers of cycles. Pyrolysis gas chromatography mass spectrometry (Py-GC/MS) was used to study the effects of the processing parameters on additives. Fourier transform infrared spectroscopy, hydrogen nuclear magnetic resonance spectroscopy, and gel permeation chromatography (GPC) were used to analyse the structural changes in the resin. GPC results showed that after processing at 290 °C using the torque rheometer, large size soluble polymeric components increased. The increase in the large size soluble polymeric components after processing at 290 °C was probably related to the crosslinking reactions in the grafted polybutadiene. Furthermore, chemical analysis of the ABS resin samples after multiple extrusion cycles in a twin-screw extruder indicated that reprocessing considerably affected the ABS resin.


Introduction
Acrylonitrile-butadiene-styrene (ABS) is one of the most successful engineering thermoplastics. It comprises a dispersed phase containing polybutadiene (PB) rubber and a continuous rigid phase containing styrene-acrylonitrile (SAN) copolymer. The ABS resins used in this study were produced through the following process: Styrene and acrylonitrile monomers are grafted onto cross-linked PB latex particles to prepare a PB-g-SAN copolymer, and then the PB-g-SAN copolymer was blended with SAN to produce ABS. ABS resins usually contain antioxidants such as octadecyl 3-(3,5-di-tert-butyl-4-hydroxyphenyl)propionate (antioxidant 1076) and tris(2,4-di-tert-butylphenyl) phosphite (antioxidant 168), and lubricants such as N,Nʹethylenebis(stearamide) (EBS) and metal stearate.
ABS is widely used in automobiles, communication instruments, and commodities, etc. It is therefore necessary to study the recycling of ABS because of environmental, economic and energy reasons. Extensive research has been conducted on ABS mechanical recycling [1,2]. Casale et al. studied ABS degradation during processing as well as during ageing under natural and artificial light. They concluded that the toughness decreased mainly because of the degradation of the PB phase [3]. Eguiazábal and Nazábal studied the effects of repeated injection at 260 °C on the properties of polycarbonate (PC)/ABS blends. They found that the rubber phase of ABS was cross-linked and oxidised after five processing cycles [4]. Kim and Kang compared the effects of reprocessing on three types of ABS resins with different PB contents. They found that the impact strength of the ABS with the highest PB content decreased by about 27%, and reduced faster than those of the others [5]. Boldizar and Möller investigated the effects of combined extrusion and thermo-oxidative ageing on ABS. They observed that elongation at break increased after repeated extrusion, and the percentage of increase was about 90% [6]. Bai et al. used a torque rheometer to reprocess an ABS plastic at 230 °C for four cycles, impact strength decreased by about 81%. They attributed the reduction in impact strength to the degradation of rubber phase and the loss of small molecules [7]. Salari and Ranjbar studied the properties of ABS during reprocessing and thermo-oxidative ageing. They found that reprocessing and thermo-oxidative ageing had a much more severe effect on the impact strength than on the tensile strength [8]. Karahaliou and Tarantili studied the effects of reprocessing on ABS and ABS/montmorillonite nanocomposites. They found elongation at break of ABS increased by 51% in the third cycle and increased by 44% in the fourth cycle, but decreased significantly in the fifth cycle [9,10]. Pérez et al. found that the impact strength slightly decreased after reprocessing and that the number of reprocessing cycles had an obvious influence on the tensile strength during ultraviolet irradiation [11]. Peydro et al. reprocessed ABS at 220 and 260 °C and then mixed the recycled ABS with styrene-ethylene/butylene-styrene copolymers. Their differential scanning calorimetry (DSC) results showed that the crosslinking reactions in the rubber phase caused a decrease in the crosslinking enthalpy, and this decrease was more apparent in the samples reprocessed at 260 °C [12]. Scaffaro et al. studied the effect of the recycled ABS content and number of reprocessing cycles on the properties of ABS/recycled ABS blends [13]. Wang et al. added an epoxy-based chain extender (0.7 wt%) to recycled ABS. The chain extender could react with the carbonyl groups of the recycled ABS, impact strength and tensile strength were obviously increased [14]. Li et al. added pyromellitic dianhydride as a chain extender to recycled ABS. When the content of pyromellitic dianhydride was 0.9 wt%, impact strength and elongation at break were greatly increased. The compatibility between PB and SAN was improved [15]. Zhan et al. used styrene-ethylene-butylene-styrene block maleic anhydride as a compatibilizer. The hydroxyl groups of recycled ABS could react with the compatibilizer. The recycled ABS with the compatibilizer had excellent impact strength [16].
In previous studies, it was found that rubber phase of ABS was cross-linked and oxidised during reprocessing. Mechanical properties, especially impact strength, decreased after reprocessing due to the degradation of rubber phase. Extenders and compatibilizers were used to modify recycled ABS to achieve high quality recycled materials. However, the degradation of grafted PB during reprocessing had not been systemically studied. In this study, we focused on the degradation of grafted PB and the effects of degradation of grafted PB during reprocessing.
The degradation mechanisms of ABS during reprocessing had been explained in previous studies. ABS has α-hydrogens of the carbon-carbon double bonds in butadiene, which are susceptible to oxidation due to the dissociation energy of α-hydrogen is relatively low. During processing, the free radicals are produced and then the hydroperoxides are formed. Subsequently, the hydroperoxides decompose, and all kinds of oxidation products are produced or further crosslinking reactions occur. In addition, the butadiene grafting sites contain tertiary carbons, they can be oxidised during processing and then chain scission occurs [13,14]. During processing and ageing, the carbon-carbon double bonds in the butadiene phase produced oxygen containing groups such as carbonyl groups and hydroxyl groups. The hydroxyl groups formed in the degradation process have potential to react with pyromellitic dianhydride and styreneethylene-butylene-styrene block maleic anhydride [15,16].
Many characterisation techniques are employed to examine the structures of polymers. Gas chromatography mass spectrometry (GC/MS) has previously been employed to study the effects of reprocessing on additives in ABS plastics [17]. Pyrolysis gas chromatography mass spectrometry (Py-GC/MS) has also been used in various aspects of polymer research including microstructure determination, degradation studies, and additive analysis [18]. Pentimalli et al. used hydrogen nuclear magnetic resonance ( 1 H NMR) spectroscopy to investigate the influence of γ irradiation on ABS [19]. In this study, we concentrated on analysing the effects of reprocessing on ABS using chemical analysis methods. We used Py-GC/MS to study the effects of processing on additives. 1 H NMR was used to examine the changes in the chemical structure of ABS during processing.
The Gel permeation chromatography (GPC) results obtained in our earlier work showed that the proportion of large size soluble polymeric components increased after processing using an torque rheometer at high temperature of 270 °C, but the reason for that was not studied [7]. In another of our earlier works, we studied the mechanical properties of ABS during repeated extrusion [20], but we did not comprehensively analyse the chemical structure changes. To overcome these problems, the chemical changes in the structure of ABS during processing were further studied using several characterisation techniques in this work. The aim of this work was to provide more information about degradation of ABS and changes in additives during reprocessing, and to help protect the environment.

Material Processing
In this work, ABS is a commercially available material known as PA-757, produced by Chi Mei Corp. (Taiwan).
Melt processing of the resin was conducted using a torque rheometer (RM-200A Mixed Torque Rheometer, Harbin Hapro Electrical Technology Co. Ltd., China). The ABS raw material (sample code ABS0) was dried in an oven at 1 3 80 °C for 2 h before processing. When the temperature in the cavity of the torque rheometer reached 230 °C, 60 g of the ABS raw material was placed in the cavity, and a time of 3 min was allowed for the cavity temperature to stabilise. Then, the material was processed at 30 rpm for 10 min. After processing, the blades were stopped, the cavity was opened at 230 °C, and approximately 10 g of material was removed (sample code ABS-L1). Subsequently, the cavity was closed, the remaining material was processed at 30 rpm for another 10 min at 230 °C; the cavity was opened, and approximately 10 g of ABS was again removed from the cavity (sample code ABS-L2). The cavity was closed again, the remaining material was processed for another 10 min at 230 °C, and all the processed material was then removed from the cavity (sample ABS-L3). Another batch of the ABS resin was processed at 290 °C for one, two, and three cycles using the same method, and the corresponding samples were designated as ABS-H1, ABS-H2, and ABS-H3, respectively. Two ABS samples were also processed at 260 °C and 290 for one cycle (samples ABS-M and ABS-H); the materials were removed when the temperature of the cavity had reduced to 230 °C. Table 1 presents the number of processing cycles and processing temperatures for each sample processed using the torque rheometer.
Another set of samples was prepared from the ABS raw material, which was processed using a twin-screw extruder over a period of time. In March 2009, the ABS was reprocessed using a twin-screw extruder (SJSH-30, Nanjing Rubber & Plastics Machinery Plant Co. Ltd., China). The following extruder temperature profile was used for reprocessing ABS: 190, 205, 205, 205, 205, 205, and 205 °C. After each extrusion cycle, the extrudate was cooled using water and pelletised. Some of the pellets were moulded into samples using an injection moulding machine (CJ50E-II, Nanjing Rubber & Plastics Machinery Plant Co. Ltd., China) at 190-210 °C, and the other pellets were processed again in the next extrusion cycle. Before extrusion and injection moulding, the material was heated at 80 °C for approximately 2 h in an air oven. Sample ABS-E0 was prepared by moulding the ABS raw material in the injection moulding machine without subjecting it to extrusion.
In March 2019, following preservation in the dark at room temperature for 10 years, some injection-moulded samples were aged in a thermal ageing test chamber at 110 °C for 21 days. Before thermo-oxidative ageing, the surfaces of these samples were ground using sandpaper. Table 2 mentions the number of extrusion cycles and ageing times for each representative sample prepared through extrusion and injection moulding.

Hydrogen Nuclear Magnetic Resonance Spectroscopy
ABS samples (0.6 g) were added to 10 mL of CDCl 3 containing tetramethylsilane (TMS) as the internal reference (Aldrich) and kept for one day to dissolve the materials; the mixtures were then filtered through a 0.2-μm syringe filter. The suspensions before filtering and the solutions after filtering were analysed at 25 °C using an NMR system (AVANCE III HD 600 M Hz, Bruker Company, Switzerland). Figure 1 shows photograph of the some samples. The suspensions of ABS are on the right, and their filtrates are on the left.

Fourier Transform Infrared Spectroscopy
ABS samples (0.6 g) were added to 10 mL of CDCl 3 and kept for one day to dissolve the ABS. Then, the formed suspensions were filtered through a 0.2-μm syringe filter; the suspensions before filtering and the solutions after filtering 10-year preservation ABS-E5N 5 10-year preservation ABS-E0NA 0 10-year preservation and 21-day accelerated ageing ABS-E5NA 5 10-year preservation and 21-day accelerated ageing were formed into films and examined using an FTIR system (VERTEX 70, Bruker, Germany). Powder samples were extracted from some of the injection-moulded samples using a file, mixed with KBr, compressed into thin disc-shaped pellets, and then subjected to FTIR spectroscopy.

Gel Permeation Chromatography
GPC analysis was conducted using an Agilent 110 instrument (Agilent, USA). The ABS samples (30 mg) were added to tetrahydrofuran (THF, 15 ml) to prepare sample solutions. The mixtures were kept for one day to dissolve the ABS. The solutions thus obtained were filtered through a 0.2-μm syringe filter prior to conducting chromatography. The molecular weight distribution was determined at 30 °C using a PLgel column (Mixed-C, 300 × 7.5 mm, 5-μm particle size) using a refractive index detector. THF was used as the eluent at a flow rate of 1.0 mL/min. The GPC system was calibrated with polystyrene (PS) standards.

Pyrolysis Gas Chromatography Mass Spectrometry
For each run, the sample was placed in a pyrolyser (EGA/ PY-3030D, Frontier Lab, Japan) and heated at 300 °C for 1 min. The volatiles were carried by He gas (0.8 mL/min) through a column (Ultra-5, 30 m × 0.25 mm × 0.25 μm) of a gas chromatograph-mass spectrometer (QP2010-Ultra, Frontier Lab, Japan). The oven temperature was held at 40 °C for 1 min, then increased at 8 °C/min to 300 °C, and held for 10 min. The split/splitless injector was used in the split mode with a split ratio of 5:1. Each component was analysed using an MS detector with an electron impact source and a mass range of m/z = 29-1000. The interface and source temperatures were 280 and 200 °C, respectively. The NIST 05 Library of Mass Spectra was used for identifying the compounds.
Some samples were analysed using another Py-GC/MS system. For each run, the sample was placed in a pyrolyser (CDS5250, USA) and heated at 300 °C for 1 min. The volatiles were carried by He gas (1 mL/min) through a column (HP-5MS, 30 m × 0.25 mm × 0.25 μm) of a gas chromatograph-mass spectrometer (HP7890/5975, Agilent Technologies, USA). The oven temperature was held at 40 °C for 1 min, then increased at 8 °C/min to 300 °C, and held for 10 min. The split/splitless injector was used in the split mode with a split ratio of 100:1. The interface and source temperatures were 280 and 230 °C, respectively. The NIST 14 Library of Mass Spectra was used for identifying the compounds.

Dynamic Mechanical Thermal Analysis (DMTA)
A dynamic mechanical thermal analyser (DMA8000, PE Company, USA) was employed to test the samples using a single cantilever-type clamp configuration at a frequency of 1 Hz, with a heating rate of 5 °C/min. The surfaces of the injection-moulded samples were ground before conducting DMTA measurements. The test samples were approximately 10.0-mm long, 7.7-mm wide, and 2.3-mm thick. They were much smaller than the original impact testing samples (10.0-mm wide and 4.0-mm thick), and were extracted from central parts of the original samples.

Melt Flow Rate (MFR) Measurements
The MFR was measured by a melt flow indexer (SRZ-400C, Changchun Intelligent Instrument & Equipment Company, China). The test temperature was 230 °C. The load was 5.0 kg. The MFR is defined as MFR = 600 W/t, where W is the extrudate weight in grams within a time t of 10 s. Therefore, the unit of MFR is g (10 min) −1 .   [21,22], which was relatively stable during processing.

Py-GC/MS Analysis
Some compounds that were confirmed by referring to the NIST 05 Library of Mass Spectra are shown as numbered peaks in Fig. 2. Peak 26 at 37.82 min corresponds to antioxidant 1076. Peak 6 at 23.33 min is related to 1-octadecanol, which can be formed during the decomposition (hydrolysis) of antioxidant 1076 [23]. Interestingly, the size Fig. 2 Total ion chromatograms of volatile compounds originating from virgin ABS, and ABS processed at 230, 260, and 290 °C using the QP2010-Ultra system of the 1-octadecanol peak originating from ABS-M1 was lower than that originating from any of the other samples analysed. Therefore, during processing at 260 °C, the rate of decomposition of 1076 was slower than the rate of loss of 1-octadecanol. Octadecanenitrile (Peak 7 at 23.62 min) and hexadecanenitrile (Peak 2 at 21.20 min) could be the decomposition products of EBS, used as a lubricant [17]. These peaks became much larger in the spectra obtained for the samples processed at a higher temperature, suggesting that EBS could decompose when processing at higher temperatures.
The peaks with retention times in the range 25.6-26.1 min (Peaks 17, 18, and 19) also increased significantly following processing at the highest temperature and Peak 19 is the most abundant of these. The structures of these peaks have not been confirmed. Figure 3 shows the total ion chromatograms of virgin ABS (ABS0), and ABS processed at 230 °C for different numbers of cycles using the torque rheometer. It can be seen that the 1-octadecanol (Peak 6 at 23.33 min) did not increased during reprocessing, the reason for that might be the rate of decomposition of antioxidant 1076 was slower than the rate of loss of 1-octadecanol. After processing at 230 °C for three cycles, the content of octadecanoic acid (Peak 13 at 24.52 min) was reduced after reprocessing, volatilisation was an important reason. However, the content of octadecanoic acid increased after one processing cycle. That indicates octadecanoic acid might originate from metal stearate [17].

Effects of Reprocessing Using the Torque Rheometer
The ABS samples after processing at 230 °C for one cycle (ABS-L1), 230 °C for three cycles (ABS-L3), and 290 °C for one cycle (ABS-H) were dissolved in CDCl 3 . After filtration through a membrane, the filtrates were all colourless and transparent (see Fig. 1). Figure 4 shows the FTIR spectra of ABS-L1 before and after filtration. The solvent used was CDCl 3 containing a small amount of TMS. The CDCl 3 with TMS were evaporated before the IR spectra were recorded, but the spectrum of residual solvent shows two peaks originating from CDCl 3 (at 745 and 913 cm −1 ). However, these solvent peaks did not affect the study of ABS degradation in this work. The peak at 1028 cm −1 is also attributed to the aromatic rings derived from styrene. The peak corresponding to the out-of-plane bending vibration of the hydrogen atoms in the C-H groups attached to the alkenic carbons in trans-1,4-PB components is observed at 967 cm −1 . The peaks at 762 and 702 cm −1 are related to the out-of-plane bending of the C-H groups in the aromatic rings [24,25]. Comparing the spectrum of ABS-L1 after filtration with that before filtration, three peaks nearly disappeared after filtration, two of which were related to trans-1,4-PB components (967 cm −1 ) and 1,2-PB components (1636 cm −1 ). The third peak was at 3301 cm −1 and may be attributed to hydroxyl groups introduced during the oxidation or hydration of the rubber phase [26][27][28]. The other peaks exhibited minimal changes after filtration. From Fig. 4, the rubber particles were filtered out from the suspension as they were not soluble in CDCl 3 . Figure 5 presents the FTIR spectra of the ABS suspensions after processing for one cycle at 230 and 290 °C, and after processing for three cycles at 230 and 290 °C.  the saturated units (CH 2 and CH groups). In particular, the peaks at 1.9-2.1 ppm are mainly related to the CH 2 protons in the 1, 4-PB units [31,32]. Figure 6a shows that all the peaks of the PB units have become smaller after processing at a higher temperature of 290 °C. Compared to processing at 230 °C for three cycles, processing at 290 °C for one cycle caused a greater reduction in the sizes of the signals at 1.9-2.1 and 4.8-5.5 ppm. This is consistent with the rubber phase of ABS undergoing crosslinking at high processing temperatures [7]. Crosslinking within the rubber phase can reduce the solubility of PB. Figure 6a also shows that the relative intensity of the sharp peak at 1.19 ppm increases after processing at 290 °C. This peak probably arises from antioxidant and lubricants [33], which are compounds with long-chain alkyl groups added to the resin during manufacture. Degradation of these additives could yield molecules that are more easily soluble in CDCl 3 .
The 1 H NMR spectra of the filtrates of ABS-L1, ABS-L3, and ABS-H are shown in Fig. 6b. Here, the peaks at 5.34 and 5.31 ppm, corresponding to the protons in trans-1, 4-PB and cis-1, 4-PB, respectively, are much smaller than the peaks observed in this region prior to filtration, confirming the removal of most of the rubber phase during filtration. After filtration, the sizes of the peaks at 1.9-2.1 ppm also decreased, and two small peaks at 2.01 and 1.96 ppm simultaneously became more apparent. The peaks at 2.01 and 1.96 ppm originated from the -CH-protons in cis-1,4-PB and trans-1,4-PB components, respectively [31], considering that a small amount of PB was originally grafted onto SAN. When the ABS solution was filtered, the grafted PB molecules may have moved into the filtrate even though the PB rubber phase was removed. Figure 6b also reveals that the peaks at 2.01 and 1.96 ppm are reduced after processing at 230 °C for three cycles and at 290 °C for one cycle. One likely reason for this reduction is thermo-oxidative degradation of the grafted PB. Another possibility is that the grafted material itself, i.e. PB, underwent crosslinking. The prominent sharp peak at approximately 1.5 ppm probably originated from a small quantity of water entering the filtrate during filtration.
The 1 H NMR spectra of ABS after processing at 290 °C for different numbers of cycles are shown in Fig. 7. The peaks in the 1.9-2.1 and 4.8-5.5 ppm ranges appear to decrease in intensity during the first processing cycle. With subsequent reprocessing cycles, the intensity of the peaks in these ranges decreased more slowly. Our 1 H NMR results agree with Peydro's findings [12], suggesting that extensive crosslinking reactions could have already occurred within the rubber phase during the early reprocessing cycles.
The 1 H NMR spectra for the filtrates of the ABS samples are also shown in Fig. 7. The peaks attributable to PB (at 1.9-2.1 and 4.8-5.5 ppm) are considerably small after only one processing cycle; moreover, the peaks reduce further ABS samples processed using the torque rheometer were dissolved in THF and were filtered through a 0.2-μm syringe filter. The rubber particles were filtered out as they were not soluble in THF [34]. The molecular weight distribution of soluble components of the filtrate was analysed using GPC. Some of the recorded GPC curves are shown in Fig. 8. The first broad peak represents the higher-molecular-weight polymeric components, and the second peak represents lower-molecular-weight components. The weight-average molecular weights (M w ) of the polymeric (macromolecular) components were calculated considering the section between the two red vertical lines. As shown in Fig. 8, reprocessing at 230 °C for three cycles using the torque rheometer had a minimal influence on the size distribution of the macromolecular portion. From Fig. 6a, it can be known that the crosslinking reactions did not extensively happen in rubber phase during reprocessing at 230 °C. In our earlier work [7], a recycled ABS plastic was reprocessed at 230 °C using a torque rheometer. Dynamic mechanical thermal analysis (DMTA) reflected the occurrence of crosslinking reactions in PB phase. At the same time, GPC results showed that the proportion of large size soluble polymeric components slightly increased. Based on these works, it is suggested that the increase in the proportion of large size soluble polymeric components during reprocessing in our earlier work was possibly caused by crosslinking reactions in the grafted PB.
The GPC curves for the soluble components of ABS0 and ABS-H3 are shown in Fig. 9. It is evident that the large size soluble polymeric components increase after reprocessing at 290 °C for three cycles. Figure 10 presents the M w of the soluble polymeric components determined for the ABS. It is apparent that the M w increases substantially during the first and second processing cycles. In the subsequent processing cycles, the M w increased to a lesser extent. When the number of double bonds of the grafted PB remaining to participate in the crosslinking reactions during the subsequent cycles decreased, the increase in the M w became slow.

Effects of Repeated Extrusion and Thermo-Oxidative Ageing
Ten years ago, prior to the present study, we processed virgin ABS through extrusion for five cycles using a twin-screw extruder at 205 °C, and the mechanical properties were measured after the various reprocessing cycles [20]. In the present research, the chemical changes in the structure of ABS during repeated extrusion were further studied.
As shown in Fig. 11, reprocessing at 205 °C for five cycles using the twin-screw extruder had a small influence on the size distribution of the macromolecular portion. The GPC results obtained 10 years ago show that the proportion of large size polymeric components and small size polymeric components slightly increased after reprocessing. Based on the analysis above, it can be suggested that crosslinking reactions occurred in the grafted PB. Our results obtained 10 years ago show a slight increase in MFR (melt flow rate) after repeated extrusion (see Fig. 12). One reason for the increase in MFR could be breaking of the grafts between PB and SAN [6]. Figure 13 shows the FTIR results for the samples without repeated extrusion (ABS-E0) and with repeated extrusion (ABS-E5). The two samples were examined using another FTIR system (EQUINOX 55, Bruker, Germany). Compared to the absorbance of polystyrene components at 1601 cm −1 , no significant change is apparent in the intensity of trans-1,4-PB components at 968 cm −1 .
In our early work, we studied the mechanical properties of ABS during repeated extrusion [20]. Some results are shown in Table 3; most values are the averages of eight samples. From Table 3, the impact strength decreased by 37.7% and the elongation at break increased by 40.1% after repeated   extrusion. The thermo-oxidative degradation of the grafts between PB and SAN could be an important reason for the decrease in the impact strength. Crosslinking reactions in the grafted PB might explain the increase in the elongation at break, because the relation between SAN phase and PB phase became strong.
In the present research, to study the changes in the chemical structure of ABS during thermo-oxidative ageing, some injection-moulded samples were placed in a thermal ageing test chamber at 110 °C for 21 days. Accelerated thermooxidative ageing was performed 10 years after the samples were fabricated. During these 10 years, the samples were preserved at room temperature without light exposure. Although the samples had been stored in sealed plastics bags, the surfaces of the samples were ground with sandpaper before thermo-oxidative ageing to reduce the effects of natural ageing on the test results. The colours of the samples before and after thermo-oxidative ageing can be seen in Fig. 14. ABS-E0N was produced from the material that had not been extruded, whereas ABS-E5N was fabricated from the material that had been extruded for five cycles. After thermo-oxidative ageing, ABS-E0N and ABS-E5N became brown, denoted as ABS-E0NA and ABS-E5NA, respectively. Figure 15 shows the variations in the loss modulus (E″) versus temperature for ABS-E0N and ABS-E5N. The plots of E″ against temperature show that the lower temperature transition due to the rubber phase is slightly shifted to a higher temperature after repeated extrusion, suggesting the occurrence of crosslinking reactions in the rubber phase. The higher temperature transition due to the SAN phase shows no significant change. Furthermore, the DMTA results presented in Fig. 16 prove that reprocessing has a considerable influence on the chemical structure of ABS. The value of the maximum E″ of the rubber phase decreased, and the peak broadened. The increase in the breadth of rubber phase transition indicates the increase in heterogeneity caused by thermo-oxidative degradation. The glass transition temperatures (T g ) are listed in Table 4. Figure 17 presents the FTIR spectra of ABS-E0N, ABS-E0NA, and ABS-E5NA obtained using the KBr pressedpellet method. The difference between the spectra of ABS-E0N and ABS-E0NA is small. Comparing the spectrum of ABS-E5NA with the spectrum of ABS-E0N, the peak attributed to trans-1,4-PB units at 967 cm −1 decreased after repeated extrusion and thermo-oxidative ageing (Y 1 changed from 1.92 to 1.49), whereas the peak attributed to 1,2-PB at 1636 cm −1 was not affected much (Y 2 did not decrease). These results are in accordance with the results of other published works [35,36]. The apparent loss of PB units after the multiple extrusion and thermo-oxidative ageing may reflect the loss of the antioxidant during the extrusion cycles.   Fig. 18. After reprocessing or thermo-oxidative ageing, the intensities of the peaks at 1.9-2.1 and 4.8-5.6 ppm (attributed to PB) decreased, but the decrease was greater after five twin-screw extrusion cycles (ABS-E5N) than after thermo-oxidative ageing at 110 °C for 21 days (ABS-E0NA). Crosslinking within the PB phase is responsible for the loss of the PB signals associated with the presence of double bonds.
After filtration of the suspensions through a membrane, the four filtrates were all colourless and transparent, indicating that the colour was associated with the insoluble rubber phase. The 1 H NMR results for the filtrates are presented in Figure 19. In the spectra of ABS-E0N and ABS-E5N, the intensities of the small peaks corresponding to residual PB at 1.9-2.1 and 4.8-5.6 ppm decreased in the spectra of the filtrate obtained after reprocessing. The peak at 1.19 ppm appeared to have reduced after 5 processing cycles.
The Py-GC/MS results for these samples were also obtained using another Py-GC/MS system (see Fig. 20). From Fig. 20, the contents of some volatile compounds (mainly styrene) were reduced after reprocessing. The concentration of hexadecanoic acid decreased after reprocessing. EBS did not obviously decompose during repeated extrusion at 205 °C. However, we found that N-(2-hydroxyethyl)hexadecanamide formed during thermo-oxidative ageing in another set of experiments (see Fig. 21).

Conclusion
We comprehensively studied the changes in the chemical structures of an ABS resin using NMR, FTIR, GPC, and Py-GC/MS.    The 1 H NMR spectra of ABS suspensions showed that during processing at 290 °C using a torque rheometer, changes in the structure of the PB phase occurred, the reduction in the intensities of the 1 H NMR peaks corresponding to PB was mainly attributed to the crosslinking reactions within the PB phase. The M w increases substantially during the first and second processing cycles. In the subsequent processing cycles, the M w increased to a lesser extent. According to all the results we obtained, we concluded that the increase in large size soluble polymeric components during processing was probably caused by crosslinking reactions in the grafted PB. In this study, reprocessing at 230 °C using the torque rheometer had a slight influence on PB and SAN of ABS, but the thermo-oxidative degradation of the grafted PB possibly occurred. EBS and antioxidant 1076 tended to decompose during processing at 290 °C.
In our study, elongation at break increased after reprocessing using a twin-screw extruder, the reason for that might be the crosslinking reactions in the grafted PB. However, impact strength decreased after reprocessing, thermo-oxidative degradation of the grafted PB might be an important reason. After repeated extrusion, the rubber phase of ABS was more easily oxidised during thermooxidative ageing. The apparent change of PB units after the multiple extrusion and thermo-oxidative ageing may reflect the loss of the antioxidant during the extrusion cycles. EBS did not obviously decompose during repeated extrusion, but N-(2-hydroxyethyl)hexadecanamide was found formed during thermo-oxidative ageing.
In this study, the effects of reprocessing on the grafted PB were studied. Thermo-oxidative degradation and crosslinking reactions occurred in the grafted PB during reprocessing. The crosslinking reactions in the grafted PB led to the increase in large size soluble polymeric components. The increase in elongation at break during reprocessing was possibly related to the crosslinking reactions in the grafted PB. This study was complementary to previous similar studies, and could help recycle ABS plastics and protect the environment.