Identification of hazardous organic compounds in e-waste plastic using non-target and suspect screening approaches

Chemosphere 356 (2024) 141946 Available online 9 April 2024 0045-6535/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/bync/4.0/). Identification of hazardous organic compounds in e-waste plastic using non-target and suspect screening approaches G. Castro * , M. Cobo , I. Rodríguez Department of Analytical Chemistry, Nutrition and Food Sciences, Institute for Research in Chemical and Biological Analysis (IAQBUS), Universidade de Santiago de Compostela, 15782, Santiago de Compostela, Spain HIGHLIGHTS GRAPHICAL ABSTRACT •Non-target screening analysis for the identification of hazardous substances in WEEE. •More than 300 volatile and semi-volatile substances were identified in WEEE plastic. •Flame retardants, UV filters and PAHs exhibited the highest detection frequencies. •TBBPA and TPhP displayed the highest concentrations. ARTICLE INFO Handling Editor: Myrto Petreas Keywords: Waste of electrical and electronic equipment Hazardous organic contaminants Flame retardants Non-target screening Suspect screening Recycling ABSTRACT End-of-life electric and electronic devices stand as one of the fastest growing wastes in the world and, therefore, a rapidly escalating global concern. A relevant fraction of these wastes corresponds to polymeric materials containing a plethora of chemical additives. Some of those additives fall within the category of hazardous organic compounds (HOCs). Despite the significant advances in the capabilities of analytical methods, the comprehensive characterization of WEEE plastic remains as a challenge. This research strives to identify the primary additives within WEEE polymers by implementing a non-target and suspect screening approach. Gas chromatography coupled to time-of-flight mass spectrometry (GC-QTOF-MS), using electron ionization (EI), was applied for the detection and identification of more than 300 substances in this matrix. A preliminary comparison was carried out with nominal resolution EI-MS spectra contained in the NIST17 library. BPA, flame retardants, UV-filters, PAHs, and preservatives were among the compounds detected. Fifty-one out of 300 compounds were confirmed by comparison with authentic standards. The study establishes a comprehensive database containing m/z ratios and accurate mass spectra of characteristic compounds, encompassing HOCs. Semi-quantification of the predominant additives was conducted across 48 WEEE samples collected from handling and dismantling facilities in Galicia. ABS plastic demonstrated the highest median concentrations, ranging from 0.154 to 4456 μ g g −1 , being brominated flame retardants and UV filters, the families presenting the highest concentrations. Internet router devices revealed the highest concentrations, containing a myriad of HOCs, such as tetrabromobisphenol A (TBBPA), tribromophenol (TBrP), triphenylphosphate (TPhP), tinuvin P and bisphenol A (BPA), most of which are restricted in Europe. * Corresponding author. E-mail address: [email protected] (G. Castro). Contents lists available at ScienceDirect Chemosphere journal homepage: www.elsevier.com/locate/chemosphere https://doi.org/10.1016/j.chemosphere.2024.141946 Received 2 February 2024; Received in revised form 5 April 2024; Accepted 6 April 2024 Chemosphere 356 (2024) 141946 2 1. Introduction Electrical and electronic equipment (EEE) have become essential in our daily lives. The production of EEE remains as one of the fastest growing global manufacturing activities (Buekens and Yang, 2014), mainly fuelled by rapid technological advances and accelerated obsolescence. This translates into a rapid growth of waste of electrical and electronic equipment (WEEE), also known as e-waste. WEEE is defined as a complex mixture of plastic, metals and metalloid materials (Runde et al., 2022), which contains both valuable and hazardous substances that require special handling and recycling to avoid environmental contamination and detrimental effects on human health (Robinson, 2009). Mobile phones, televisions, laptops, or any equipment with a battery or an electrical power supply is considered as EEE, so they are likely to swell the mountain of WEEE at the end of their useful life. In 2019, 53.6 million tons of WEEE were generated worldwide (Global E-waste Statistics Partnership, 2020), of which only 17.4% was collected and recycled responsibly, while the rest was sent and, illegally dumped in lowor middle-socioeconomic countries (Global E-waste Statistics Partnership, 2020). To date, recycling to recover reusable components as Cu and precious metals from WEEE is the ultimate goal (European-- Parliament, 2020). However, due to the limited number of facilities and high costs, landfilling, incineration, and exportation to poor countries remain as the main alternatives. These practices carry an associated environmental and human risk due to the presence of toxic substances (Das et al., 2021; Zeng et al., 2021). Thus, current European Directive on WEEE is under revision to ensure the protection of the environment by preventing, or reducing, the adverse impacts of the generation and the management of WEEE and a new Directive is planned to be released this year 2024. In the meantime, the scientific community efforts are focused on deepening the knowledge of plastic chemicals to bring together relevant information to finally establish an informed policy in the context of plastics-related processes. Thus, a new report on this topic has recently been published (Wagner et al., 2024). Present European Union (EU) Directives require an important reduction of the WEEE plastic (WEEEP) landfilled, as this material is estimated to constitute 20.6 % of total European WEEE (Buekens and Yang, 2014). Nevertheless, one of the main limitations to achieve an actual sustainable circular economy is the lack of information regarding the chemical additives used in the different plastics and applications (Wagner et al., 2024). Among the different plastic polymers, the most commonly used in WEEE are acrylonitrile butadiene styrene (ABS), polystyrene (PS) (or high impact PSHIPS), polycarbonate (PC), polypropylene (PP), polyethylene (PE), polyvinylchloride (PVC) and combinations of them (Buekens and Yang, 2014; Jia et al., 2022). As for the additives, many substances are added as pigments, polymer fillers, plasticizers, antistatic agents, flame retardants, stabilizers as thermic or UV filters, and reinforcing glass or carbon fibres (Buekens and Yang, 2014). Thus, containing many hazardous organic compounds (HOCs), many of them included in the restriction of organic hazardous substances (ROHs) directive, such as polybrominated biphenyls (PBB) and polybrominated diphenyl ethers (PBDEs) (Bill et al., 2022; European Parliament, 2011). In particular, great attention has been paid to halogenated flame retardants (HFRs), such as tetrabromobisphenol A (TBBPA) and hexabromocyclodecane (HBCD) (Kajiwara et al., 2011; Kousaiti et al., 2020; Sindiku et al., 2015; Yu et al., 2017). Information on the occurrence of other families of potentially HOCs is still limited, which is a key for improving reutilization. Analytical approaches for the determination of metals in WEEEP are mainly based on the use of spectrometric techniques such as inductively coupled plasma optical emission spectroscopy (ICP-OES), ICP coupled to mass spectrometry (ICP-MS) and atomic absorption spectroscopy (AAS) (Andrade et al., 2022; ´ Angel Aguirre et al., 2013; Dimitrakakis et al., 2009; Duarte et al., 2010; Guimar˜ aes Araújo et al., 2012; Shi et al., 2021; Zeng et al., 2021), however methodologies focused on the comprehensive identification of organic additives remain limited. WEEEP is a complex matrix, generally composed of different polymers, so sample preparation and determination techniques oriented to non-target screening analysis are challenging. Due to this variety, certified reference materials are scarce, which further complicates the validation of the proposed methodologies. Common analytical strategies applied to complex solid matrix, such as liquid-solid extraction (LSE), Soxhlet, and ultrasonic-assisted extraction (UAE) are usually employed as extraction techniques for plastic materials, using solvents that swell and/or dissolve both the target analytes and the plastic, followed by the precipitation of the polymeric matrix (Ballesteros-G´ omez et al., 2014; Cobo-Golpe et al., 2024; Silva et al., 2006). At this step of the analytical process, the selection of an appropriate solvent is the key step to successful results, due to transformation or degradation risks of some compounds during extraction, non-quantitative recoveries and interferences with other species generated from partial degradation of the parent polymers skeleton. High resolution/accurate mass spectrometry (HRMS), either based on time of flight (TOF) or Orbitrap mass analysers are essential for nontarget screening analysis, as they provide spectral information of any substance present in the sample amenable to chromatographic separation and ionization steps. When it comes to the analysis of volatile and semi-volatile compounds, such as phenols or halogenated flame retardants, gas chromatography -electron ionization (GC-EI) coupled to HRMS constitutes the preferred technique, as it provides reproducible and robust fragmentations (Hollender et al., 2017). Generally, accurate m/z values for fragment and molecular ion (when observed), are specific enough for the unambiguous identification of a given compound and no MS/MS data is required. Identification of unknowns is based on a database containing spectra for a high number of substances. NIST contains low (nominal) resolution spectra for more than 15,000 compounds (NIST v17). To date, the main disadvantage of this technique is the lack of accurate EI-MS spectral libraries available (Castro et al., 2019; Hollender et al., 2017), however as EI-MS spectra are more reproducible than MS/MS and libraries contains a higher number of spectra, spectral comparison with EI-MS, especially when including retention index information, constitute an excellent approach to reach tentative identifications (Hollender et al., 2017). Herein, the aims of the present study are to 1) investigate the possibilities of a non-target screening approach, based on accurate electron ionization mass spectrometry (EI-MS) spectra, for the identification of volatile and semi-volatile compounds present in WEEEP collected from different handling facilities; 2) to create a database containing the m/z and spectra of the common chemical additives present in WEEEP and 3) to evaluate the concentrations of the identified substances, with particular attention to those considered as hazardous species based on their environmental toxicity, persistence and/or mobility. To the best of the authors’ knowledge this is the first study aiming a non-target screening of organic additives in WEEEP samples and presenting an in-depth analysis of their occurrence. 2. Materials and methods 2.1. Chemicals and materials LC-MS grade methanol (MeOH, 99.8%), dichloromethane (DCM, 99.8%) and iso-octane (i-Octane, 99%) were purchased from Merck (Darmstadt, Germany). Ammonium fluoride (NH 4 F) was acquired from Sigma Aldrich (Milwaukee, WI, USA). Ultrapure water (18.2 mΩ cm −1 ) was obtained from a Milli-Q system by Millipore (Billerica, MA, USA). Standards of tentatively identified analytes in WEEEP samples, were acquired from Sigma-Aldrich and TCI Europe (Zwijndrecht, Belgium). Isotopically labelled standards (either deuterated or 13 C species) were obtained from Sigma Aldrich and Toronto Research Chemicals (North York, ON, Canada), Table S1. Individual stock solutions of each compound were prepared in MeOH and further dilutions were made in the same solvent. These compounds were first employed to assess the G. Castro et al. Chemosphere 356 (2024) 141946 3 performance of the analytical methodology and later as surrogate standards (SSs) to improve the reliability of the quantitative data provided for a selection of additives in WEEEP. 2.2. Waste sampling Electronic and electrical waste samples were collected from 3 different waste handling facilities located in Galicia (Northwest of Spain) during winter 2023. The facilities included specialized sorting based on the different WEEE categories according the WEEE Directive 2012/19/EU (Council of the European Union, 2012) (Table 1). Samples were collected manually by representatively subsampling large piles of specific waste type using gloves. The selection of WEEE samples included small appliances containing different types of plastics. Once in the laboratory, samples were disassembled to recover plastic components (casings, fans, keyboards, etc). These components were rinsed with ultrapure water, allowed to dry at room temperature and finally ground below 2 mm using a mill (Retsch® SM100, Biometa) to reach their homogeneity. Those samples that were not possible to grind in the mill due to melting or rigidity, were broken down using a conic drill. The homogenized materials were stored in glass jars in a refrigerator at 4 ◦C until analysis. More information about the samples is presented in Table 1. 2.3. Sample preparation A representative sample of WEEEP (≈0.3 g of ground plastic) was weight out and placed into a 15 mL polypropylene tube (PP). The sample was fortified with 40 μ L of a mixture of SSs (50 μ g mL −1 ). Samples were extracted through ultrasound assisted extraction (UAE) with 8 mL of DCM (30 min, 35 ◦C). After extraction, the polymeric matrix was precipitated with 2 mL of MeOH (Yao et al., 2023) and centrifuged for 10 min at 4000 rpm. The obtained supernatant was collected and transferred to a clean PP tube (≈8 mL). Thereafter, 1 mL of extract was filtrated (hydrophobic filter 0.45 μ m) and concentrated below 0.2 mL under a gentle nitrogen stream (N 2 ). Finally, samples were reconstituted in 1 mL of i-octane and ready for GC-MS analysis, or in 1 mL of MeOH, in case of TBBPA quantification by UPLC-MS. 2.4. Instrumental analysis 2.4.1. GC-QTOF-MS WEEEP extracts were analysed using a gas chromatography hybrid quadrupole time-of-flight (GC-QTOF) MS instrument, purchased from Agilent (Wilmington, DE, USA), which consist of a 7890A gas chromatograph and a 7200 QTOF MS spectrometer, furnished with an EI source. The system was operated in 2 GHz mode (mass resolution (FWHM) of 6500 at m/z 131 Da). In case of non-target screening (NTS) and suspected target approaches, the MS spectra were recorded in profile mode (required for spectral deconvolution) at 2.5 Hz (5430 transients per spectrum), in the range of m/z values from 50 to 700 Da, while in target experiments, MS spectra were recorded in centroid mode. The m/z axis was automatically re-calibrated by infusion of perfluorotributyl amine (PFTBA) every 6 injections. Chromatographic separation was carried out in a HP-5 MS capillary column (30 m ×250 μ m i.d., 0.25 μ m film thickness) acquired also from Agilent. Helium was employed as carrier gas at a constant flow of 1.2 mL min −1 . The temperature of the oven was programmed as: 90 ◦C (1 min), rated at 8 ◦C min −1 to 300 ◦C (15 min). Standards and sample extracts (2 μ L) were injected in the pulsed splitless mode (25 psi, 1 min), with the injector temperature set at 300 ◦C. The split flow and the splitless time were 60 mL min −1 and 1 min, respectively. The transfer line and the EI source temperatures were 280 ◦C and 230 ◦C, respectively. The temperature of the quadrupole (Q) MS analyser was 150 ◦C and the system was employed in the MS mode, with the Q serving as ion guide to the TOF. Table 1 Overview of the samples analysed in this study. ID Type of plastic WEEE category b Description AC-001 ABS a IT and telecommunication equipment Headset AC-002 ABS a IT and telecommunication equipment Headset AC-003 PS IT and telecommunication equipment Computer speaker AC-004 ABS-PVC a IT and telecommunication equipment Headset BAS-001 PS Small household appliances Scale CAL-001 ABS a IT and telecommunication equipment Desk calculator CB-001 PVC rigid + phthtalate a Cable plastic Cable plastic CB-002 PVC rigid + phthtalate a Cable plastic Cable plastic CB-003 PVC rigid + phthtalate a Cable plastic Cable plastic CB-004 PVC rigid + phthtalate a Cable plastic Cable plastic COND001 Silicon a IT and telecommunication equipment Conductivimeter DEP-001 ABS Small household appliances Hair remover DIS-002 ABS a IT and telecommunication equipment Floppy disk DIS-004 PS a IT and telecommunication equipment Floppy disk DNI-001 ABS a IT and telecommunication equipment ID reader IMP-001 ABS IT and telecommunication equipment Printer IMP-003 PS-HI IT and telecommunication equipment Printer IMP-005 ABS-FR IT and telecommunication equipment Printer INT-001 PC-ABS IT and telecommunication equipment Ethernet connection station LU-001 PC a Lighting equipment Emergency lights LU-002 PC a Lighting equipment Emergency lights MAN001 ABS IT and telecommunication equipment Controller MAN002 Silicon a IT and telecommunication equipment Controller PC-001 ABS IT and telecommunication equipment Computer PC-002 PC a IT and telecommunication equipment Laptop PL-001 PP a Small household appliances Iron machine PL-002 ABS-nylon a Small household appliances Hair straightener RAD-001 ABS Large household appliances Electric radiator RAD-002 PP a Large household appliances Electric radiator RAT-001 PC-ABS a IT and telecommunication equipment Computer mouse RAT-004 ABS IT and telecommunication equipment Computer mouse RD-001 SB IT and telecommunication equipment Radio RD-002 ABS-PVC a IT and telecommunication equipment Radio RD-003 ABS a IT and telecommunication equipment Radio ROU-001 ABS y PC-ABS IT and telecommunication equipment Router ROU-002 ABS y PC-ABS IT and telecommunication equipment Router ROU-003 ABS a IT and telecommunication equipment Router ROU-004 ABS IT and telecommunication equipment Router (continued on next page) G. Castro et al. Chemosphere 356 (2024) 141946 4 2.4.2. UHPLC-QqQ-MS Quantification of TBBPA was performed by ultraperformance liquid chromatography tandem mass spectrometry (UPLC-MS/MS), using an Acquity UPLC system connected to a Xevo TQD triple quadrupole mass spectrometer, furnished with a Z spray ESI source, both acquired from Waters (Milford, MA, USA). Chromatographic separation was carried out in a Zorbax Eclipse Plus C 18 Rapid Resolution HD column (2.1 ×50 mm, 1.8 μ m) connected to a C 18 2.1 mm i.d. Security Guard™ ultracartridge, supplied by Agilent and Phenomenex (Torrance, CA, USA), respectively. Mobile phases were ultrapure water (A) and MeOH (B), both containing 1 mM NH 4 F, at a constant flow rate of 0.4 mL min −1 . The gradient of mobile phase was programmed as follows: 0–1.30 min, 2% B; 1.31–2.50 min, 50% B; 2.51–6.00 min, 100% B; and 6.10–7.00 min, 2% B. Column and pre-column were maintained at 40 ◦C. The injected volume for solvent standards and sample extracts was 0.5 μ L. TBBPA was ionized under ESI negative mode. Two transitions were monitored per compound (multiple reaction monitoring mode, MRM) considering a time window of 60 s around the retention time (RT), being the quantification transition (Q1) 542.8 → 447.8 (CE 34V) and qualification transition (Q2) 542.8 → 419.8 (38 V) for TBBPA; and Q1 552.8 → 454.8 (35 V) and Q2 552.8 → 80.9 (50 V) for TBBPA-d 10 . The capillary and cone voltage were optimized as 3.50 kV and 27 V, respectively. 2.5. Qualitative and quantitative data analysis The performance of the employed methodology was first validated in terms of extraction and determination for a selection of 18 isotopically labelled standards, belonging to different chemical families recognized as common additives in plastic from WEEE. A pooled WEEEP sample was prepared by mixing 7 different samples containing different plastic types. The recoveries (R%) of the sample preparation for the labelled compounds were estimated by dividing the response (peak area for each analyte) obtained in the pooled sample spiked before the treatment (preextraction) by the corresponding response of the same analyte obtained in a sample extract fortified at the same level after the sample treatment (post-extraction) and multiplied by 100. Potential matrix effects (MEs, %) during GC-MS analysis were assessed and estimated as the ratio between the analyte response obtained in a sample extract fortified after the treatment (post-extraction) and the response of the same compound for a solvent (i-octane)-based standard of the same concentration and multiplied by 100. Therefore, values below 100 % indicate signal suppression, while values above 100% indicate enhancement of the instrumental sensitivity during analysis. Ratios with values in proximity to 100% suggest the absence of ME (Raposo and Barcel´ o, 2021). The instrumental limits of quantification (iLOQ) were calculated for each analyte as the concentration of the lowest calibration standard concentration providing a signal-to-noise (S/N) ratio of 10 or, alternatively, 10 times the standard deviation of the instrumental blanks divided by the slope of the calibration curve if the analyte was detected in them. The NTS approach was applied to a set of 20 samples. Data mining was performed following a workflow previously described by the authors (Castro et al., 2019) with minor modifications (Fig. 1). Briefly, raw GC-EI-QTOF-MS files, acquired in profile mode, were loaded into Unknowns Analysis software and peak picking was performed with the SureMass algorithm, limiting the filtration to entities with an absolute peak height above 2000 counts. The deconvolution of the chromatograms was based on four different retention time windows (25, 50, 100 and 200 %) around the average peak width in each record. Spectra of deconvoluted features were compared to those compiled in the NIST17 library to determine the similarity to known spectra. The comparison algorithm was a combination of forward-reverse search (Blum et al., 2019), values of 0 and 1 corresponding to pure reverse and forward modes, respectively. Herein, a weight factor of 0.7 was set. A PCDL library, including linear retention index (LRI) and experimental accurate EI-MS spectra of all the injected standard solutions was created, independently of whether compounds were finally confirmed in the samples, or not. As a result, a list of tentatively identified compounds was created and a set of 51 potentially hazardous compounds was selected for quantification. Quantification was performed using matrix-matched standards, containing the same concentrations of isotopically labelled SSs as those added to WEEEP samples before the extraction (Castro et al., 2019). 2.6. Quality assurance and quality control (QA/QC) QA/QC proceedings were implemented during sample preparation and analysis to avoid potential contamination problems and to reduce the number of false positives and false negatives. Thus, all PP material was discarded after use. In addition, one procedural blank (without sample) was analysed per batch of 10 samples. Any signal in the blanks were subtracted from the samples as it could be assumed to be contamination valid for all the samples. SSs were spiked to the samples to account for any signal losses during sample preparation and instrumental analysis, and ten samples were analysed (including extraction) in duplicate for analytical method performance purposes. To monitor instrumental sensitivity deviations and carry-over effects during analysis, solvent blanks and standards (50 ng mL −1 ) were analysed every 10 samples and a mass calibration was performed after 6 injections. 2.7. Software and spectral libraries GC-QTOF-MS data was acquired with Agilent enhanced Mass Hunter GC-MS acquisition software, while peak inspection of raw chromatograms and quantification was performed with Mass Hunter Qualitative software (vB.08.00) and Mass Hunter Quantitative (v10.2), respectively. Spectral deconvolution was achieved with Unknown Analysis software (based on the SureMass algorithm), integrated in the Mass Hunter Quantitative software (version B.08.00). The creation of a custom-made spectral library (PCDL) was managed using also dedicated functions in the Mass Hunter software. Preliminary tentative identifications of the deconvoluted components in the GC-QTOF-MS files, were obtained using the NIST17 EI-MS library. The MS Search (v. 2.3) software was employed to manage spectra compiled in this library, and to calculate the theoretical m/z ratios of fragment ions with known structures in the NIST17 library. UPLC-MS/MS data were acquired with the MassLynx v4.1 software, and quantification of TBBPA in plastic extracts was performed with TargetLynx (Waters, Milford, U.S.). Excel (Microsoft 365) Table 1 (continued) ID Type of plastic WEEE category b Description SEC-001 PC-nylon a IT and telecommunication equipment Hair dryer SEC_002 ABS-PC a Small household appliances Hair dryer TEL-001 ABS IT and telecommunication equipment Phone TEL-002 ABS IT and telecommunication equipment Phone TEL-003 ABS a IT and telecommunication equipment Phone TEL-004 ABS IT and telecommunication equipment Phone TEL-005 ABS-PVC a IT and telecommunication equipment Phone TEL-006 ABS IT and telecommunication equipment Phone TV-001 PS a IT and telecommunication equipment TV VENT001 PC a IT and telecommunication equipment Computer fan ABS (acrylonitrile butadiene styrene); PS (polystyrene); PVC (polyvinyl chloride); PC (polycarbonate); PP (polypropylene). a Plastic analysed through ATR analysis. b Categories according the WEEE Directive 2012/19/EU. G. Castro et al. Chemosphere 356 (2024) 141946 5 was used to perform data analysis. 3. Results and discussion 3.1. WEEEP characterization Samples were categorized according to the Directive 2012/19/UE (Council of the European Union, 2012) as i) large household appliances; ii) small household appliances; iii) IT and telecommunication equipment; iv) consumer equipment and photovoltaic panels; v) lighting equipment; vi) electrical and electronic tools; vii) toys, leisure and sports equipment; viii) medical devices; ix) monitoring and control instruments; and x) automatic dispensers (Table 1). The type of plastic material was determined by attenuated total reflectance (ATR) with infrared (IR) spectroscopy and Raman IR (laser wavelength 532 and 785 nm), when it was not specified in the plastic cover of the product. Information on the instrument and main features of the methodology are summarized in Text S1. Reference ATR-IR and RAMAN-IR spectra were found in the IRUG spectral database available online (IRUG). The most characteristic absorption bands at the different wavelengths were detected in the experimental spectra and compared with those compiled in the database in order to elucidate the type of polymers present. Examples of experimental and database IR spectra leading to the identification of the polymers ABS and silicon are presented in Fig. S1 (A and B). Thus, the presence of ABS, PVC, HIPS, PC, silicon, and combinations of different types of plastic (copolymers), were verified in the selected samples (Table 1), being ABS the main component in most of the WEEEP analysed. As for the case of copolymers, they were classified in the category of the predominant plastic. According to Jia et al. ABS accounts for 30 % of the WEEP applications, followed by HIPS, which accounts for 25 %, while the rest are usually PC (10 %), ABS/PC blends (9 %) and PP (8 %) (Jia et al., 2022). Among the main characteristics of ABS, this polymer possesses great toughness and chemical resistance by a low prize, which might explain their widespread use (Olivera et al., 2016). This variety and combination of polymers difficult sample preparation, since each polymer presents a different behaviour in contact with the organic solvents tested during extraction of organic additives and HOCs. Table S2 summarizes the stability of different polymers vortexed for 1 min and ultrasonicated during 30 min, at 35 ◦C, as function of the extraction solvent. In those conditions, ABS and PS were completely dissolved in acetone, ethyl acetate, and DCM. On the other hand, PP, PVC, and silicon were more resistant to the organic solvents, being not affected under the abovementioned conditions. From these data, DCM and MeOH were selected for extraction and polymer re-precipitation, respectively. In this way, bonded and non-bonded additives existing in the polymers were transfer to the organic extract. 3.2. Characterization of the sample preparation methodology Sample preparation methodology was evaluated in terms of extraction efficiency and matrix effects for a set of 18 isotopically labelled compounds belonging to different chemical families (Table S1). R % for isotopically labelled compounds were evaluated at one concentration (200 ng g −1 ) in triplicate, and ranged between 67 % and 84 %, with RSD<41 %, while values of ME % indicates enhancement of the signal for most of the considered chemicals (Table S1). As previously described by Meng et al. (2020), during plastic analyses many components are often co-eluting, leading to poor quantitation accuracy. However, herein signal enhancement is likely related to the injection processes, rather than to variations in the efficiency of compounds ionization at the EI source. Anyway, the large variability in the extraction efficiency observed for some of the isotopically labelled model compounds might be also related to non-quantitative extractions and losses during polymer re-precipitation and solvent-exchange. These MEs were especially problematic in ABS and PVC plastic, thus quantitation with matrix-matched calibration is highly recommended (Raposo and Barcel´ o, 2021). Instrumental repeatability (n =3) and reproducibility (n =3 days) were assessed by analysing replicates of 200 ng mL −1 standard solvent solution; and the obtained results demonstrated RSDs <14 and 36 %, respectively. 3.3. Non-target screening analysis of WEEEP samples A set of 20 samples of WEEEP products were extracted and analysed via GC-QTOF-MS (section 2.5.). A total of 305 candidates presented high spectra similarity, displaying normalized scores (0–100) above 60, which can be considered either level 2 (probable structure) or level 3 (tentatively identified) (Schymanski et al., 2014). Among these candidates, those presenting mass errors below 5 mDa, detection frequencies (DF%) above 5% and that were not detected, or detected with higher intensities than in procedural blanks, were tentatively identified. Compounds commonly known for blank contamination are bisphenol A (BPA), phthalates and the organophosphate flame retardant (OPFRs), triphenylphosphate (TPhP) (Brandsma et al., 2022; Castro et al., 2019). In the present study, blank contamination was carefully investigated. BPA, TPhP and tinuvin 329 were detected in the blanks, although presenting significantly lower signals than in the organic extracts from plastic items. Commercial standards of 51 out of 305 candidates, together with other compounds belonging to the same families or plastic additives commonly reported in literature due to their occurrence, persistence and toxicity (i.e., organophosphate flame retardants, phthalates, …) were analysed in the GC-EI-TOF-MS for spectral and retention time comparison with those corresponding to the obtained deconvoluted spectra (maximum allowable differences equal to 5 mDa and 0.1 min, respectively), and incorporated in a customized library, also containing their CAS numbers, retention times and LRI values. Fig. 1. Workflow followed during the non-target screening of HOCs in WEEEP (Figure adapted from (Castro et al., 2019)). G. Castro et al. Chemosphere 356 (2024) 141946 6 3.4. Suspect screening and semi-quantification of organic additives and HOCs by GC-EI-TOF-MS A customized EI-MS library containing 339 compounds with their corresponding accurate spectra, which included 51 of the tentatively identified substances, was applied for the suspect screening of 48 WEEEP samples. The accurate EI-MS library is available in Zenodo repository (https://zenodo.org/; https://doi.org/10.5281/zenodo.10045 296). The abovementioned strategy applied during the NTS was also followed in the suspect screening analysis. Thus, deconvolution and library search of the obtained spectra were carried out. Compound identification was achieved by matching all the peaks in the experimental spectra with the spectra in the customized EI-MS library, limiting the identification results to a minimum match factor of 60 %, accurate mass tolerance below 50 ppm and a chromatographic retention time window of 40 s. The identification of tribromophenol (TBrP) in ABS plastic using the accurate EI-MS library is illustrated in Fig. 2. Extracted ion chromatogram and mass spectrum of TBrP (M + 329.7682 and retention time 12.436 min) are presented in Fig. 2A and B, respectively. Mass spectrum was compared to the spectra contained in the EI-MS library (Fig. 2C). Thus, differences between calculated and experimental m/z for the most intense fragment ions were calculated and stayed below 2.3 mDa (Table 2). The accurate experimental spectrum for a spiked pool of WEEEP confirmed the identity of the candidate as TBrP. Following this workflow, a total of 79 compounds were tentatively identified in the studied samples (Table 2). Table 2 compiles a list of compounds detected in WEEEP following the abovementioned methodology, including their use, detection frequency (DF %), calculated mass errors (mDa) and confirmation with commercial standards. Most of the candidates can be categorized as flame retardants (FR), PAHs, personal care products (PCPs), pesticides, plasticizers, and UV filters (Table 2). Information on the chemical composition of WEEEP is still scarce, only a few authors have previously performed suspect screening of plastic consumer products and plastic from WEEE toys (Lowe et al., 2021; Meng et al., 2020; Wagner et al., 2024), reaching the same conclusions as those presented herein: PAHs, pesticides, bisphenols and FR (OPFRs and BFRs) are the families displaying the highest detection frequencies (DF %) in this type of matrix (Table 2). Other authors have addressed the occurrence of FRs, such as polychlorinated biphenyl substances (PCBs) (Yu et al., 2017), new brominated flame retardants (NBFRs) and organophosphorus flame retardants (OPFRs) (Ballesteros-G´ omez et al., 2014; Bill et al., 2022; Roth et al., 2012), and other contaminants of emerging concern (CECs) such as bisphenols, benzophenone UV filters (Runde et al., 2022), and perand polyfluoroalkyl substances (PFAS) (Tansel, 2022). The presence of NBFRs (Lan et al., 2023), PFAS (Zhang et al., 2020), phthalates (Li et al., 2023), bisphenols (Wei et al., 2023), fluorinated biphenyls and analogues (Zhu et al., 2021), benzothiazoles and benzotriazoles (Li et al., 2020) was also reported in dust and soils from WEEE dismantling areas, pointing out that this waste might constitute an emission source of the abovementioned chemicals. As for the semi-quantification, a total of 51 analytes were considered (Table S3). Most of the studied compounds presented corrected R % between 60 % and 100 %, with RSD <28 %. In the case of ME %, following the same tendency as their deuterated analogues, enhancement of the signal was observed for most of the considered analytes (Table S3). The linear response range for every analyte was assessed by the injection of solvent-based standards at seven concentrations ranging from 20 ng mL −1 to 500 ng mL −1 . Within this interval, the obtained calibration curves fitted a linear model, presenting determination coefficients (R 2 )>0.99 for the majority of the analytes. The iLOQs were also calculated displaying values between 0.70 and 567 ng mL −1 (Table S3). 3.5. Determination of TBBPA by UPLC-QqQ-MS The quantification of TBBPA constitutes a challenge due to its medium polarity (logK ow 5.9) and presence of two phenolic moieties in the molecule (Liu et al., 2016; Sunday et al., 2022). GC-MS analysis requires a previous derivatization step, i.e. using methyl chloroformate, to improve the detectability and the reproducibility of the analysis (Covaci et al., 2009; Sunday et al., 2022). The obvious result is an increase in the complexity of the analytical procedure. Herein, target analysis of TBBPA was accomplished by UPLC-QqQ-MS, due to this technique provides higher sensitivity and lower iLOQs. The suitability of the proposed methodology was demonstrated through the evaluation of R % and ME % (Table S3), providing values of 74 % (RSD 19 %) and 80 % (RSD 13 %), respectively. Obtained iLOQ of TBBPA was 2.04 ng mL −1 . 3.6. Occurrence of organic additives and HOCs in WEEEP Forty-eight out of 51 targeted compounds were detected in the different samples. Most of them are compounds of environmental concern due to their persistence, toxicity, mobility (i.e. benzotriazole) Fig. 2. Extracted ion chromatogram (EIC) of TBrP (A) for a non-spiked sample (RAT-001). Spectra of the deconvoluted component (B) and the experimental accurate spectra of TBrP (C). G. Castro et al. Chemosphere 356 (2024) 141946 7 Table 2 Tentatively identified compounds in the extracts from plastic samples by GC-QTOF-MS. Compound Other ions Use/Type Retention time (min) CAS number Database m/z Experimental m/z Mass error (mDa) DF (%) Ion 2 Ion 3 Confirmed with standard BHT Antioxidant 19.48 128-370 205.16 205.1595 −0.50 35% 177.1285 ✓ 3,5-Di-tert-butyl-4hydroxybenzoic acid (BHTCOOH) Antioxidant 23.62 142149-4 235.1682 235.1717 3.50 5% 250.1918 ✓ 2,6-Di-tert-butyl-4-hydroxy-4methyl-2,5-cyclohexadien-1one (BHT-OH) Antioxidant 25.25 1039680-2 219.1443 219.1385 −5.80 20% 191.1132 ✓ Antioxidant 2246 Antioxidant 37.17 119-471 340.2404 340.2408 0.40 10% 177.1282 Pentachloroanisole Bactericide 24.25 182521-4 264.8338 264.8369 3.10 15% 236.8373 ✓ Triclosan Bactericide 31.91 338034-5 287.9494 287.9525 3.10 5% 218.0123 145.9685 ✓ Nicotine Drug 15.37 54-11-5 161.1073 161.1038 −3.50 5% 84.0808 119.0612 ✓ Octanoic acid Fatty acid 10.64 124-072 73.0292 73.027 −2.20 30% 101.0605 60.0215 Nonanoic acid Fatty acid 13.15 112-050 73.0293 73.027 −2.30 15% 115.0762 dodecanoic acid Fatty acid 20.56 143-077 73.0291 73.0271 −2.00 5% 129.092 TCEP Flame retardant 25.13 115-968 248.9853 248.9877 2.40 30% 204.9587 142.9661 ✓ TCPP Flame retardant 25.93 1367484-5 125.0001 125.0006 0.50 10% 201.0083 277.0165 ✓ Triphenyl phosphate (TPhP) Flame retardant 36.85 115-866 325.0723 325.0638 −8.50 50% 215.0275 169.0658 ✓ Cresyl diphenylphosphate (CDP) Flame retardant 38.20 2644449-5 340.0871 340.0876 0.50 5% 165.0709 ✓ 2,3 ′ ,4 ′ ,6-Tetrabromodiphenyl ether (BDE 71) Flame retardant 38.71 18908462-6 485.7114 485.7111 −0.30 5% 325.877 ✓ EHDPP Flame retardant 38.91 124194-7 251.0481 251.0482 0.10 15% ✓ BDE-99 Flame retardant 42.59 6034860-9 403.7863 403.7877 1.40 5% 563.6223 ✓ Tetrabromobishenol A (TBBPA) Flame retardant 46.05 79-94-7 528.7309 528.7286 −2.30 10% 447.8116 272.8718 ✓ Methyl anthranilate Flavouring Agent 15.21 134-203 119.0377 119.0383 0.60 5% 151.0642 Vainillin Flavouring agent 16.65 121-335 151.0395 151.0383 −1.20 35% 123.0443 ✓ 1,1 ′ -Biphenyl Flavouring agent or adjuvant 16.17 92-52-4 154.0772 154.0758 −1.40 25% 76.0309 Benzaldehyde, 4-methoxyFlavouring Agents 12.94 123-115 135.0444 135.0407 −3.70 10% 107.0497 a-Methylstyrene Flavouring agent 6.05 98-83-9 118.0772 118.0754 −1.80 30% 103.0542 alpha-amylcinnamaldehydeisomer2 Flavouring agent 22.60 122-407 129.0694 129.07 0.60 5% 115.0539 202.1347 alpha-hexylcinnamaldehyde Flavouring agent 24.78 101-860 129.0699 129.0678 −2.10 10% 115.0542 216.1503 alphahexylcinamaldehydeisomer2 Flavouring agent 24.78 101-860 129.0695 129.0705 1.00 15% 115.0544 216.1501 Diphenyl ether Flavouring agent and fragrance 16.75 101-848 170.0725 170.0721 −0.40 5% 141.0701 Cinnamal Flavouring agent and Pesticide 13.36 104-552 131.0491 131.0488 −0.30 15% 103.0539 Longifolene Fragrance 16.94 475-207 161.1343 161.1324 −1.90 40% 91.0551 Hedione Fragrance 22.73 2485198-7 83.0501 83.0479 −2.20 10% 153.1286 α -Guaiene Natural product 16.84 369112-1 105.0718 105.0682 −3.60 15% 91.0559 Caffeine Natural product 26.82 58-08-2 194.0809 194.0805 −0.40 5% ✓ N-Nitrosodiphenylamine Nitrosamine 21.99 86-30-6 169.0885 169.0854 −3.10 5% 83.5364 Naphthalene PAH 11.09 91-20-3 128.0622 128.0593 −2.90 20% 99.0461 ✓ Acenaphthene PAH 18.79 83-32-9 153.0701 153.0684 −1.70 5% ✓ (continued on next page) G. Castro et al. Chemosphere 356 (2024) 141946 8 Table 2 (continued) Compound Other ions Use/Type Retention time (min) CAS number Database m/z Experimental m/z Mass error (mDa) DF (%) Ion 2 Ion 3 Confirmed with standard Fluorene PAH 21.11 86-73-7 166.0776 166.0775 −0.10 5% 139.0542 ✓ 2,6-Diisopropylnapthalene PAH 24.28 2415781-1 197.1316 197.1317 0.10 40% 212.1558 155.0853 ✓ Phenanthrene PAH 25.44 85-01-8 178.0781 178.0777 −0.40 15% 152.0619 ✓ Pyrene PAH 31.05 129-000 202.0781 202.0802 2.10 5% ✓ Coumarin PCP 17.67 91-64-5 146.0362 146.036 −0.20 5% 118.0412 Methyl paraben PCP 18.04 99-76-3 121.0299 121.0272 −2.70 5% 152.0467 ✓ Ethyl paraben PCP 19.76 120-478 121.0301 121.0243 −5.80 20% 166.0623 ✓ Propyl paraben PCP 21.98 94-13-3 121.0294 121.0255 −3.90 10% 138.0312 ✓ Ethyl-4dimethylaminobenzoate PCP 23.96 1028753-3 193.1092 193.1078 −1.40 55% 148.0753 164.0702 Galaxolide PCP 27.01 122205-5 243.1747 243.1764 1.70 5% 213.1633 ✓ Tonalide PCP 27.28 2114577-7 243.1753 243.1772 1.90 5% 187.1127 ✓ 2,6-dimethylphenol Pesticide 9.00 576-261 122.0724 122.0709 −1.50 15% 107.0489 2,5-Dichlorotoluene Pesticide 9.42 1939861-9 159.9872 159.9832 −4.00 5% 125.018 2,4-Dichlorophenol Pesticide 10.60 120-832 161.9652 161.9625 −2.70 10% 97.9941 2,3,5-Trichlorophenol Pesticide 15.46 933-788 195.9298 195.9246 −5.20 15% 159.9526 96.9879 ✓ 2,4,6-Tribromophenol Plasticizer 22.32 118-796 329.7685 329.7708 2.30 15% 140.9277 ✓ Benzothiazole Plasticizer 12.16 95-16-9 135.0145 135.013 −1.50 60% 108.0034 ✓ Phthalic anhydride Plasticizer 14.52 85-44-9 104.0227 104.0246 1.90 20% 76.0282 ✓ Dimethyl phthalate Plasticizer 18.06 131-113 163.0422 163.0388 −3.40 15% 133.0306 ✓ Diethyl phthalate Plasticizer 21.39 84-66-2 149.0234 149.0241 0.70 70% 121.0283 ✓ 2-Hydroxy-benzothiazole Plasticizer 22.91 934-349 151.009 151.0092 0.20 20% 96.0027 123.0138 ✓ N-butyl benzenesulfonamide Plasticizer 25.52 362284-2 170.0276 170.0299 2.30 15% 141.001 Dibutyl phthalate Plasticizer 29.10 84-74-2 149.0234 149.0218 −1.60 25% 121.0309 ✓ Bisphenol A Plasticizer 33.49 80-05-7 213.0944 213.0908 −3.60 10% 119.0517 ✓ Tributyl Acetylcitrate Plasticizer 34.57 77-90-7 185.0769 185.0816 4.70 15% 129.015 Benzyl butyl phthalate Plasticizer 36.01 85-68-7 149.0232 149.0228 −0.40 5% 91.0537 ✓ Bis(2-ehtylhexyl) adipate Plasticizer 36.77 103-231 129.0545 129.052 −2.50 5% 111.0438 70.0767 Dicyclohexyl phthalate Plasticizer 39.09 84-61-7 149.0264 149.0226 −3.80 40% 167.0385 Di-(2-ethylhexyl) terephthalate Plasticizer 41.98 642286-2 261.1533 261.1533 0.00 35% 149.027 167.0378 ✓ Di-isononyl phthalate Plasticizer 42.76 2054862-3 149.0284 149.0219 −6.50 10% 293.1807 ✓ Diisodecyl phthalate Plasticizer 44.55 89-16-7 149.0285 149.0248 −3.70 20% 307.1969 ✓ Benzophenone UV filter 22.18 119-619 182.0728 182.0706 −2.20 50% 105.0338 77.0385 ✓ Homosalate UV filter 27.66 118-569 120.0218 120.0186 −3.20 5% 138.0324 ✓ Benzophenone 3 UV filter 30.43 131-577 227.0696 227.0713 1.70 5% 151.0383 ✓ Tinuvin P (drometrizole) UV filter 31.09 244022-4 225.0913 225.0865 −4.80 5% 168.0827 ✓ EHMC UV filter 35.50 546677-3 178.0639 178.0659 2.00 15% 161.0609 ✓ Tinuvin 350 UV filter 38.51 3643737-3 308.1797 308.1762 −3.50 5% 238.0996 ✓ Tinuvin 326 UV filter 39.17 729335 300.0925 300.0921 −0.40 5% 272.0609 ✓ Tinuvin 329 UV filter 39.52 5218876-8 252.113 252.1144 1.40 5% 133.0608 ✓ Tinuvin 328 UV filter 40.75 2597355-1 322.1946 322.1924 −2.20 20% 252.1152 ✓ Tinuvin 327 UV filter 40.88 386499-1 342.138 342.1364 −1.60 20% 286.0759 ✓ Octocrylene UV filter 41.00 619730-4 248.0713 248.0731 1.80 35% 204.0837 360.1972 ✓ 4,7-Methanoazulene, 1,2,3,4,5,6,7,8-octahydro1,4,9,9-tetramethyl-, [1S-(1. alpha.,4.alpha.,7.alpha.)]- 16.05 514-512 105.0716 105.0684 −3.20 25% 161.1353 G. Castro et al. Chemosphere 356 (2024) 141946 9 and ubiquitous character. Identifications and concentrations in duplicate and reinjected samples were self-cross checked, presenting RSD <36 % (n =10 duplicates). Total mean concentrations in the different types of plastics ranged from 13.9 μ g g −1 to 7218 μ g g −1 , being PP and ABS, the polymers containing the lowest and the highest concentrations for this set of hazardous semi-volatile compounds, respectively (Table 3 and Fig. 3A). Among the different families, the plasticizer bisphenol A (BPA) presented the highest concentrations in PC and PP, accounting for the 74 % and 83 % of the total concentration, respectively (Fig. 3B). UV filters and BFRs were predominant in ABS and silicon plastic (Fig. 3B). 3.6.1. BPA BPA is one of the most abundant synthetic chemicals in the world (Usman and Ahmad, 2016). This substance is commonly used as monomer during the synthesis of PC plastics and epoxy resins. Runde et al. previously reported the concentrations of BPA in WEEE samples collected in Norway, reporting median concentrations of 22 μ g g −1 (Runde et al., 2022) and similar median concentrations (71 μ g g −1 ) were also previously reported by Arp et al., in 2017 (Arp et al., 2017) for WEEE samples. In the present study, median concentrations of 34.5 μ g g −1 were measured for BPA. The complete set of sample concentrations is presented in Table S4. 3.6.2. HFRs BFRs, both PBDEs and NBFRs, were detected in the samples, being TBrP and TBBPA those showing the higher prevalence, with DF of 21 % and 38%, respectively (Table S4). TBrP is commonly used as pesticide and flame retardant in thermoplastics and epoxy resins, ABS and PS (Norwegian Environmental Agency, 2016). Herein, TBrP was mostly detected in ABS plastic and copolymers in median concentrations of 11.6 μ g g −1 . As for TBBPA, 70–90 % of the produced substance is used as a reactive flame retardant in epoxy, PC and phenolic resins in printed circuit boards, while the rest is commonly used as additive in ABS plastic and phenolic resins (Fj¨ ader et al., 2022; Liu et al., 2016). In the latter ones, TBBPA is not chemically bound to the material, so it can be easily released to the environment. Many authors reported concentrations in WEEE and dust from dismantling areas. Yu et al. studied the occurrence of BFRs in housing plastics from different types of WEEE, reporting concentrations from n.d. to 34.0 μ g g −1 (Yu et al., 2017). Lower concentrations were presented by Kousaiti et al., indicating that TBBPA concentrations vary significantly within the same plastic type, whose values stayed between <LOQ and 3.11 μ g g −1 (Kousaiti et al., 2020). Herein, TBBPA was detected in ABS and PC plastic, presenting higher concentrations than reported in literature and ranging from 23.8 ng g −1 to 20.7 mg g −1 (median 269 ng g −1 ). Routers from different production years and brands were found to contain the highest TBBPA concentrations, ranging from 0.16 μ g g −1 for the newest acquired item (year 2021) to 20.7 mg g −1 for the oldest (year 2002), while PBDEs remained non detected. This trend was previously reported by Bill et al., who compared PBDE and TBBPA median concentrations in EEE samples Table 3 Summary of the main descriptive statistics (detection frequency (DF, %), mean, median, min and max) of the concentrations (ng g −1 ) obtained for the different HOCs families in WEEE samples (n =48 samples). Type of plastic Descriptive statistics ∑PAHs ∑Preservatives ∑benzothiazoles ∑BFRs BPA ∑OPFRs ∑Fragrances ∑UV filters ∑Others ∑TOTAL ABS (n =24) DF (%) 88% 25% 4% 58% 75% 83% 4% 83% 4% 96% mean 1106 597 808 3861249 1588439 72084 308 4455753 154 7218092 median 420 811 808 1194 48401 1238 308 13983 154 78765 min 61 63 808 24 5455 140 308 228 154 0 max 5143 923 808 20749247 27826282 1283173 308 88278244 154 117387859 ABSPVC (n = 3) DF (%) 100% 0% 33% 33% 100% 100% 67% 100% 0% 100% mean 642 – 1927 56369 12434 11048 525 17832 – 61737 median 600 – 1927 56369 5938 5357 525 2054 – 13456 min 535 0 1927 56369 5510 511 521 803 0 8373 max 790 0 1927 56369 25853 27276 529 50639 0 163383 PC (n = 7) DF (%) 86% 29% 14% 43% 86% 71% 29% 100% 0% 100% mean 7594 630 823 15159 559865 116331 1043 47117 – 623697 median 4008 630 823 7319 147990 74882 1043 28074 – 179212 min 542 269 823 68 69260 2460 879 3778 0 114592 max 18021 990 823 38089 2496584 306930 1206 123744 0 2826710 PP (n = 2) DF (%) 100% 0% 0% 0% 100% 0% 50% 100% 0% 100% mean 269 – – – 11877 – 608 1547 – 13997 median 269 – – – 11877 – 608 1547 – 13997 min 222 0 0 0 8346 0 608 179 0 12185 max 317 0 0 0 15409 0 608 2915 0 15810 PS (n = 5) DF (%) 100% 0% 20% 60% 40% 60% 20% 100% 0% 100% mean 1846 – 9031 13446 11916 176 308 2171 – 18825 median 1884 – 9031 75 11916 115 308 1668 – 4139 min 590 0 9031 27 9237 94 308 616 0 2475 max 3523 0 9031 40235 14595 318 308 4211 0 70290 PVC (n =4) DF (%) 100% 50% 75% 25% 25% 100% 100% 100% 100% 100% mean 10915 9672 53555 319 18581 21834 27321 34333 615 144744 median 7389 9672 390 319 18581 23345 11894 25943 733 104893 min 2363 189 281 319 18581 13857 1355 5470 45 67439 max 26518 19155 159995 319 18581 26787 84141 79975 949 301750 SB (n = 1) DF (%) 100% 100% 0% 0% 0% 100% 0% 100% 0% 100% mean 1366 41651 – – – 23 – 628 – 43668 median 1366 41651 – – – 23 – 628 – 43668 min 1366 41651 0 0 0 23 0 628 0 43668 max 1366 41651 0 0 0 23 0 628 0 43668 Silicon (n = 2) DF (%) 100% 0% 0% 0% 50% 100% 100% 100% 100% 100% mean 291 – – – 34549 1309 3229 32047 624 54776 median 291 – – – 34549 1309 3229 32047 624 54776 min 43 0 0 0 34549 382 805 7268 341 8838 max 540 0 0 0 34549 2237 5654 56827 907 100714 G. Castro et al.