The recent SARS-CoV-2 pandemic has shown the importance of developing RNA vaccines as well as the importance of cost-effective and timely vaccine production to deliver a rapid response to disease outbreaks. Having an effective RNA-based medicine requires a delivery vehicle to bring the nucleic acid into the cytoplasm and protect it from degradation, however, obtaining an efficient delivery system to achieve the full potency of RNA therapeutics remains a key challenge.
Quantitating the presence of leachables after downstream processing of lipid nanoparticle (LNP) drug product is critical to ensure the quality of the drug product. In this study we demonstrate an effective and complete downstream process for manufacturing of LNPs. We detected and measured the leachables from the NanoAssemblr™ GMP system in the LNP formulation step and demonstrated effective clearance of leachables in the tangential flow filtration (TFF) step. Excellent clearance (> 96%) of organic and elemental PERLs was observed. The top five abundant PERLs were silicone compounds as precusors/degradants of silicone tubing, and they were cleared after 4.1 diafiltration volumes (DV). Four elemental process equipment related leachables (PERLs: Li, Ni, Ba, and Os) were seen in processing fluid during lipid nanoparticle formation and more than 98% of the elements were efficiently cleared via UF/DF.
Introduction
Significant progress and innovation in bioprocessing have accelerated the implementation and integration of single-use technologies (SUT) into biomanufacturing infrastructure offering end-to-end platforms for therapeutic modalities, such as, LNP-encapsulated mRNA. As single-use systems (SUS) are typically constructed with polymeric materials containing additives, antioxidants, and processing aids, safety concerns associated with the persistence of substances migrating from SUS (also known as PERLs) into the pharmaceutical process stream remain a key concern. PERLs that find their way into the final product are called leachables, may be organic or inorganic in nature and are subject to strict regulations by health authorities as unsafe exposure levels could potentially harm patients.
Safety risk assessment is typically carried out on product-contact components using extractables profiles aligned to BioPhorum and USP <665> (United States Pharmacopeia, 2021) requirements provided by the SUT suppliers and/or data from simulated leachables testing performed by the end-users (1). In this study, we operated a complete downstream process for manufacturing of LNPs using a SUS, while tracking and quantitating in-process leachable throughout the manufacturing process.
Various downstream processing steps have been shown to attenuate leachables risk, particularly ultrafiltration/diafiltration (UF/DF) via TFF operation, which is widely used as the final purification step prior to the final sterile filtration and/or formulation and filling step (2). Several studies showed successes of UF/DF in clearing small molecules from recombinant proteins. Hunter et al. performed ultraviolet–visible spectroscopy (UV-Vis) experiments and computer simulations to demonstrate up-to 70-fold clearance of small-molecule impurities via 2-stage diafiltration experiments, albeit using an antibody-dye model system (3). Magarian et al. showed post-UF/DF reduction up to 1000-fold of seven water-soluble PERLs by 1H NMR studies (4).
A COVID-19 vaccine candidate (PNI-003) and optimized the vaccine production has been developed using microfluidic-based manufacturing system known as the NanoAssemblr™. The NanoAssemblr™ technology ensures that nanoparticle formation conditions are uniform, controlled, and deliver reproducible product at various scales. The clinical-scale sample (50 mg) used for this batch was manufactured using the NanoAssemblr™ GMP System.
Following LNP assembly, tangential flow filtration (TFF) was used to remove the ethanol from the LNP and exchange the formulation buffer for a relevant storage buffer which provided a neutral environment (pH 7.4), which imparts a net neutral surface charge to the LNP. The PEG-Lipid, DMG-PEG2000, stabilizes the LNP during the formulation process and prevents aggregation during TFF and sterile filtration.
TFF is considered an excellent bioprocess step for clearance of PERLs, including silicone-related polymer degradants typically originating from large tubing contact areas. We have previously demonstrated > 99.9% clearance of Trimethylsilanol PERLs by TFF in a representative mAb (5). Additionally, Menzel et al. showed that > 98% of Dimethylsilanediol was removed after 4 DV in TFF step (6).
Results
Multiples factors were taken into consideration when developing the downstream processing step: Two types of TFF technologies were investigated (hollow fiber versus cassette). Different materials of construction (Polysulfone, Polyethersulfone, and regenerated cellulose) and various molecular weight cut-offs were also investigated (30 to 500 Mr).
The T-series TFF cassettes with Delta regenerated cellulose (RC) membrane with a 30 000 molecular weight cut-off (MWCO) was chosen for this study. The membrane performance was optimized by flux excursion where an optimal transmembrane pressure (TMP) was defined at a given feed flux. After the TFF step, sterilizing grade filtration was accomplished with the Supor™ EX ECV membrane in a Mini Kleenpak™ capsule.
Figure 1 shows the permeate flux as a function of volumetric concentration factor for the LNP concentration step on a semi-log plot. The feed flux was set to 7.5 LMM and the transmembrane pressure (TMP) was maintained at 0.14 MPa (20 psig, 1.34 bar). The average permeate flux during the concentration step was 140 LMH, while achieving a volume concentration factor (VCF) of 5.5-fold in 19 min.
Fig 1. TFF concentration versus permeate flux as a function of volumetric concentration factor (VCF).
Figure 2 shows the relationship of the process flux as a function of the number of DVs for the diafiltration. The feed flux was set to 7.5 LMM and the TMP was maintained at 1.34 bar (0.14 MPa, 20 psig). The average permeate flux during the diafiltration step was 110 LMH, while achieving 4.1 DV in 22 min. The data was interrupted during the switch from concentration to diafiltration modes as the scale was disturbed.
Fig 2. TFF diafiltration versus permeate flux as a function of diafiltration volumes.
Figure 3 shows the entire process flow as a function of time including the final concentration which achieved 3.3-fold VCF. The total average process flux was 121 LMH with an active processing time of 45 min. The data was interrupted during the switch from concentration to diafiltration modes as the scale was disturbed.
Fig 3. Permeate flow as a function of time.
Product quality analytics
Encapsulation efficiency as it relates to self-amplifying RNA concentration is shown in Table 1. There was high saRNA encapsulation during all the formulation steps. The effects of post encapsulation steps on the LNP characteristics were minimal. These results were within the typical specifications for the NanoAssemblr™ GMP System.
Table 1. Particle characteristics
| Size (nm) | PDI | % encapsulation | saRNA concentration (µg/mL) | |
| 3× dilution |
70 |
0.207 | 99 | 45 |
| 8× dilution | 65 | 0.172 | 99 | 10 |
| TFF concentrated 6.0 to 6.5× | 73 | 0.256 | 99 | 42 |
| TFF dilution buffer (4 DV) | 83 | 0.278 | 98 | 39 |
| TFF final collected | 92 | 0.343 | 98 | 94 |
| After filtration on Supor™ EX ECV | 77 | 0.171 | 98 | 35 |
| After freeze-thaw | 94 | 0.195 | 98 | 48 |
Fig 4. In vitro potency of final product in BHK cells.
Potency values of the final product after production produced spike protein expression with an EC50 of 25 ng/mL (Table 2).
Table 2. Potency results for SARS-CoV-2 protein expression dose response following PNI-003 transfection in BHK 570 cells
| EC50 (ng/mL) | r2 | |
| Assay reference control | 251 | 0.9953 |
| Final LNP product | 25 | 0.9897 |
Leachables testing
Aliquots of samples were collected into glass bottles at six time points (0 s, 20 s, 60 s, 120 s, 180 s, and 230 s) during 4 min of the lipid nanoparticle formation run, along with after the TFF step, and at the final product. Then the samples were analyzed by headspace gas chromatography-mass spectrometry (GC/MS), direct injection GC/MS, liquid chromatography/photodiode array/mass spectrometry (LC/PDA/MS), and inductively coupled plasma MS (ICP/MS). The amount of PERLs were measured accordingly. Representative HS GC/MS, DI GC/MS, and LC/MS chromatograms showing depletion of organic PERLs after TFF diafiltration steps are presented in Figure 5.
Fig 5. Representative overlaid (a) HS GC/MS, (b) DI GC/MS, (c) LC/MS (ES+, m/z 114.1) showing clearance of PERLs compounds in TFF step.
Here we focus on the organic and elemental PERLs which migrated into the processing fluid LNP formation and track clearance during UF/DF in Table 6. The majority of organic PERLs from inline dilution are silicone precursor/degradant, along with residual solvent, and polyamide related compound. Four elemental PERLs were found in the processing fluid during lipid nanoparticle formation, and they are resin related.
Table 6. PERLs during lipid nanoparticle formation and their clearance in TFF steps
| Compound name | Source | CAS number | Detection method | Clearance in TFF step (%) |
| Ethoxytrimethylsilane | Silicone precursor/degradant | 1825-62-3 | HS GC/MS | > 99 |
| Ethyl acetate | Residual solvent | 141-78-6 | HS GC/MS | > 99 |
| Dimethylformamide | Residual solvent | 68-12-2 | HS GC/MS | > 99 |
| Octamethylcyclotetrasiloxane | Silicone precursor/degradant | 556-67-2 | DI GC/MS | 96 |
| Decamethylcyclopentasiloxane | Silicone precursor/degradant | 541-02-6 | DI GC/MS | > 99 |
| Dodecamethylcyclohexasiloxane | Silicone precursor/degradant | 540-97-6 | DI GC/MS | > 99 |
| tetradecamethylcycloheptasiloxane | Silicone precursor/degradant | 107-50-6 | DI GC/MS | 98 |
| hexadecamethylcyclooctasiloxane | Silicone precursor/degradant | 556-68-3 | DI GC/MS | 98 |
| Octadecamethylcyclononasiloxane | Silicone precursor/degradant | 556-71-8 | DI GC/MS | 98 |
| Eicosamethylcyclodecasiloxane | Silicone precursor/degradant | 18772-36-6 | DI GC/MS | 98 |
| Docosamethylcycloundecasiloxane | Silicone precursor/degradant | 18766-38-6 | DI GC/MS | 98 |
| Tetracosamethylcyclododecasiloxane | Silicone precursor/degradant | 18919-94-3 | DI GC/MS | 96 |
| Hexacosamethylcyclotridecasiloxane | Silicone precursor/degradant | 23732-94-7 | DI GC/MS | 96 |
| Octacosamethylcyclotetradecasiloxane | Silicone precursor/degradant | 149050-40-8 | DI GC/MS | 96 |
| Triacontamethylcyclopentadecasiloxane | Silicone precursor/degradant | 23523-14-0 | DI GC/MS | 99 |
| Dotricontamethylcyclohexadecasiloxane | Silicone precursor/degradant | 150026-95-2 | DI GC/MS | > 99 |
| Tetratriacontamethylcycloheptadecasiloxane | Silicone precursor/degradant | 150026-96-3 | DI GC/MS | > 99 |
| ɛ-caprolactam | Polyamide related compound | 105-60-2 | LC/MS | 99 |
| Li* | Resin related | N/A | ICP/MS | 98 |
| Ni | Resin related | N/A | ICP/MS | 99 |
| Ba | Resin related | N/A | ICP/MS | > 99 |
| Os | Resin related | N/A | ICP/MS | 98 |
* The tested elements included: As, Ag, Ru, Mo, Cd, Au, Se, Sb, Hg, Ir, Tl, Sn, Pb, Os, Ba, Co, Pd, Cr, Ni, Pt, Cu, V, Rh, Li
The clearance of each PERL following UF/DF operations was calculated according to Equation 1.
Eq. 1
where Cload is the amount of PERLs in the load solution after inline dilution (mg) and Cretentate is the amount of PERLs in the retentate at the end of TFF step (mg).
Sieving coefficients (S) are a solute-dependent parameter. It was derived from the plot of ln(CN/C0) (y-axis) against number of diafiltration volumes, equation shown below:
Eq. 2
where CN is the concentration of leachables in the retentate after N diavolume, C0 is the concentration of leachables in the product before TFF step (i.e., C at DV = 0), and S = -(slope of the linear curve). The sieving coefficients of three organic PERLs compounds were generated from the linear regression analyses are listed in Table 7 and show S values close to 1. These results are consistent with the ideal clearance trend observed for these PERLs and further demonstrate that these compounds permeate freely through the TFF membrane.
Table 7. Sieving coefficients of three leachable compounds from linear regression analysis
| Leachable compounds | Sieving coefficient |
| Octamethylcyclotetrasiloxane | 1.1 |
| Decamethylcyclopentasiloxane | 1.0 |
| Dodecamethylcyclohexasiloxane | 1.1 |
Excellent clearance (> 96%) of organic and elemental PERLs was observed after 4.1 DV with 3.3-fold VCF. During UF/DF, clearance of PERLs takes place both during the ultrafiltration and buffer exchange steps. In ultrafiltration, clearance is measured by the amount of PERLs remaining in the retentate compared to the initial concentration. The top five abundant PERLs are silicone compounds as precursor/degradant of silicone tubing, and they were cleared after 4.1 DV (Fig 6). Elemental analysis was performed by ICP/MS tests on the collected samples targeting twenty-four elements defined in industry guidelines. Four elemental PERLs (Li, Ni, Ba, and Os) were seen in processing fluid during LNP formation. Ni and Os are Class II elements and Li and Ba are Class III elements according to USP <232>. More than 98% of the elements were efficiently cleared via UF/DF.
Fig 6. Amount of organic and element PERLs ln(m) detected from inline dilution, after TFF step, and final product.
Safety risk assessment
Chemical safety risk assessment is the process by which the potential adverse patient safety impact of leachables is determined and Quantitated. A patient’s exposure to leachable is expressed as a total daily intake (TDI) based on a typical worst case application scenario. Based on the leachable’s toxicity, a permitted daily exposure (PDE) is established (7). Margin of safety (MoS) is the ratio of an accepted PDE or risk index level to the TDI for a given route of administration. If the MoS is much greater than one, then the leachable may be assumed to exhibit negligible impact. Table 8 shows the three highest amount of PERLs in the final product, and their MoS. The MoS of the three PERLs were well above 1000, thus the safety risks of PERLs from downstream process for manufacturing of LNP were negligible.
Table 8. Safety assessment of three highest amount of organic PERLs
| Compound name | Amount in final product (µg) | Total daily intake (µg/d)* | Margin of safety |
| Decamethylcyclopentasiloxane | 12.2 | 0.0029 | > 1000 |
| Dodecamethylcyclohexasiloxane | 17.2 | 0.0041 | > 1000 |
| Tetradecamethylcycloheptasiloxane | 46.2 | 0.011 | > 1000 |
* Volume of final product was 420 mL and dose volume would be 0.1 mL
Conclusions
Quantitating the presence of leachables following downstream processing of a LNP is critical to ensure the quality of a drug product. We have demonstrated:
- A complete and effective downstream process for manufacturing of an saRNA-LNP.
- saRNA-LNP formulation with the NanoAssemblr™ GMP system, UF/DF on Delta RC cassettes, and sterilizing grade filtration with Supor™ EX ECV.
- The critical quality attributes of LNP size, PDI, encapsulation efficiency, and potency were within the typical specifications.
- Additionally, we have detected and measured the leachables from the NanoAssemblr™ GMP system in LNP formulation step and observed good clearance of in process leachables (or PERLs) in the TFF step.
References
- Scott B, Ullsten S, Wang P, et al. Biophorum best practices guide for extractables testing of polymeric single-use components used in biopharmaceutical manufacturing. BioPhorum Operations Group Ltd. April 2020. https://www.biophorum.com/download/extractables-testing-of-polymeric-single-use-components-used-in-biopharmaceutical-manufacturing/
- Brose DJ, Dosmar M, Jornitz MW. Membrane filtration. Pharm Biotechnology. 2002;14:213-79. doi: 10.1007/978-1-4615-0549-5_5.
- Hunter A, Fulton A, Pabst T, Savery J. Clearance of persistent small-molecule impurities. Bioprocess Int. 2016l;14(5):20-27.
- Magarian N, Lee K, Nagpal K, Skidmore K, Mahajan E. Clearance of extractables and leachables from single-use technologies via ultrafiltration/diafiltration operations. Biotechnol. Prog. 2016;32(3):718-724.
- Sun B, Hadidi M, Nuñez JS, et al. Efficiency of ultrafiltration/diafiltration in removing organic and elemental process equipment related leachables from biological therapeutics. Biotechnology Progress. 2023;e3400.
- Menzel R, Korzun A, Golz C, Maier T, Pahl I, Hauk A. Dimethylsilanediol from silicone elastomers: Analysis, release from biopharmaceutical process equipment, and clearance studies. Int J Pharm. 2023;646:123441.
- Jenke D, Carlson T. A compilation of safety impact information for extractables associated with materials used in pharmaceutical packaging, delivery, administration, and manufacturing systems. PDA J Pharm Sci Technol. 2014;407-455.
- Duhen R, Beymer M, Jensen SM, et al. OX40 agonist stimulation increases and sustains humoral and cell-mediated responses to SARS-CoV-2 protein and saRNA vaccines. Front Immunol. 2022;13:896310.
- Geall, AJ, Verma A, Otten, GR, et al. Nonviral delivery of self-amplifying RNA vaccines. Proc Natl Acad Sci. 2012;109(36):14604-9.
Materials and methods
RNA production
About 300 mg of nCoV PNI A5 saRNA was synthesized as per the method previously described by Duhen R et al. with modifications (8). Briefly, the codon-optimized gene, encoding the full-length prefusion spike protein of SARS-CoV-2 with a single dominant mutation – D614G, was synthesized and cloned into our proprietary custom self-amplifying mRNA cloning vector. The vector incorporates non-structural protein encoding an attenuated alphavirus replicases (Venezuelan Equine Encephalitis Virus) TC83 strain and strong sub-genomic promoter with an engineered multiple cloning site. The cloned codon-optimized genes were synthetically constructed and amplified in Escherichia coli and purified using a Plasmid Maxi kit (Qiagen).
Then the saRNA was synthesized. The circular plasmid DNA was linearized by restriction digest at the 3’ end of the saRNA sequence. Next, the linearized DNA templates were transcribed into RNA using the cell-free in vitro transcription and enzymatic method (9). The capped RNA was polished and purified using Cytiva's chromatographic and tangential flow filtration products to achieve a target concentration between 0.4 to 0.6 mg/mL in RNA storage buffer (Cytiva). The final product was sterilized using a 0.2 µm filter unit and stored at -80°C until use.
Lipid encapsulation with the NanoAssemblr™ GMP
PNI-003 was manufactured using NanoAssemblr™ GMP system with the high-capacity microfluidic architecture allowing LNP assembly at 200 mL/min. Prior to product formulation, the aqueous and organic solutions were prepared. To prepare the organic solution, defined weights of the four lipids were dissolved in a known weight of ethanol. The solution was heated by immersion in a water bath at 55°C until the lipids were dissolved. Once fully dissolved, the four lipid solutions were combined with manual mixing. To prepare the aqueous solution, the concentrated saRNA substance was thawed at room temperature, then mixed with 100 mM sodium acetate buffer at the appropriate concentration (0.084 µg/mL). The LNP were at a nitrogen/phosphate ratio of 8:1 and an aqueous to ethanol ratio of 3:1 followed by a 3× dilution in LNP dilution buffer.
The flow rate on the aqueous pump was set to 150 mL/min, the organic pump operated at 50 mL/min, and the dilution pump operated at 400 mL/min. During the first 25 s of the formulation run, the outlet was diverted into a waste bag, after which the outlet was collected in the product collection bag. The program was allowed to run for a total of 4 min, or until either the aqueous or organic inlet bags were emptied as verified by visual inspection.
The product is added to a bulk volume of additional phosphate-buffered saline solution, to bring the product to a total dilution of at least 8-fold.
Table 9. Formulation process parameters
| NanoAssemblr™ GMP system | |
| Parameter | Value |
| Fluid path | GMP fluid path (NxGen 500) |
| Flow rate ratio | 3:1 (Aq:Org) |
| Total flow rate | 200 mL/min |
| Inline dilution ratio | 3× |
| Final dilution | 8× |
| N/P | 8 |
| Aqueous solution | 0.084 mg/mL saRNA into RNA formulation buffer |
| Aqueous volume | 595 mL (50 mg) |
| Organic solution | 12.5 mM IL-2 lipid mix in ethanol |
| Organic volume | 198 mL |
| Dilution buffer | 1× LNP dilution buffer |
| Dilution volume (in-line) | 1586 mL |
| Dilution volume (bulk) | 3965 mL |
TFF methods
The diluted product was concentrated using a 1000 cm2 Delta RC cassette DC030T12. The Quattroflow 150 SU pump was used to drive the product along a manually constructed flow path with pressure transducers and scales. The system feed flow was set to 7.5 L/m2/min at a TMP of approximately 1.4 bar (0.14 MPa, 20 psi), with a feed pressure of 1.9 bar (0.19 MPa , 27 psig) and a retentate pressure of 0.9 bar (0.09 MPa, 13 psig). A buffer exchange was then performed via continuous diafiltration to transfer the product into a LNP storage buffer. A total of 4 L of LNP storage buffer was used to perform the buffer exchange, equivalent to four diafiltration volumes. The product was further concentrated to a final weight of 320 g in the recirculation reservoir. The system was pumped dry and the product was collected.
Table 10. TFF process parameters
| TFF | |
| Parameter | Value |
| Molecular Weight Cut-Off | 30 000 Mr |
| Size | 1000 cm2 |
| Membrane material | Regenerated cellulose |
| Feed flux | 7.5 LMM |
| Flow rate | 750 mL/min |
| TMP | 20 psi |
| Initial volume | 5400 mL |
| Initial concentration | 5.5× |
| Diafiltration volume | ~ 980 mL (including hold-up) |
| Buffer exchange volume | 4 DV |
| Buffer exchange buffer | LNP storage buffer |
| Final volume | ~ 320 mL (including hold-up) |
| Final concentration | ~ 3× |
| Target saRNA concentration | > 0.05 mg/mL |
Sterilizing filtration
Sterilizing filter tubing was assembled in a BSC, using Platinum-cured silicone tubing, Y connector fitting (Watson-Marlow 520 peristaltic pump), and Supor™ EX ECV filter presterilized by autoclave. The post TFF product was brought into the BSC. The inlet end of the tubing assembly was placed into the product bottle and the tubing assembly outlet downstream of the filter was placed into a glass bottle. The peristaltic pump was started at 20 mL/min and slowly ramped up to 50 mL/min over the first minute where it was maintained until the whole final product was filtered.
Table 10. Sterilizing grade filtration process parameters
| Sterilizing grade filtration | |
| Filter size | 220 cm2 |
| Filter material | PES |
| Filter flow rate | 50 L/min |
Analytical methods for quantitation of formulated saRNA-LNP product
Particle size measurements: The LNP particle size (hydrodynamic diameter of the particles) was determined by Dynamic Light Scattering (DLS) using a ZetaSizer Nano ZS (Malvern Instruments, UK). He/Ne laser at a 633 nm wavelength was used as the light source. Data were measured from the scattered intensity data conducted in backscattering detection mode (measurement angle = 173). Measurements were an average of 10 runs of two cycles each per sample. Z-average size was reported as the particle size and is defined as the harmonic intensity averaged particle diameter, with the polydispersity (PDI) under 0.35.
saRNA quantitation and percent saRNA encapsulation measurement: The total amount of saRNA and the percent of saRNA encapsulated in the LNP was measured by a modified Ribogreen assay (Quanti-iT RiboGreen RNA assay kit, Thermo Fisher Scientific). This assay measures the amount of mRNA in samples with intact LNPs to determine the number of unencapsulated RNA, as well as, in LNP samples disrupted by detergent to measure the total RNA. Encapsulation efficiency % is calculated as the difference between the total RNA and the unencapsulated RNA divided by the total RNA. The LNP were diluted in 1× Tris-EDTA buffer the plate was read at excitation of 485 nm and emission at 528 nm.
In vitro potency assay: The formulation EC50 was determined using a dose response in vitro by testing a decreasing amount of RNA from 1000 to 0.49 ng/mL. BHK 570 cells were plated in a 96-well plate (5 000 cells/well) for 48 h before transfection. Following transfection, the cells were incubated for a further 24 h before being fixed with 4% PFA for 20 min. Cells were then permeabilized in 0.1% detergent and stained with a 1:100 dilution of FAB105403G antibody. Cells were imaged on the Cytation 7 reader.
Analytical methods for leachable testing
Headspace GC/MS
Volatile organic compounds in the samples were analyzed using an Agilent 7890B-5977A GC/MS system. A volume of 500 µL of each sample was injected without any pre-treatment on an Agilent DB-624 MS column (60 m × 0.25 mm × 1.4 µm) using helium as the carrier gas. The oven temperature was programmed from 40°C to 50°C at 5°C/min (held for 5 min), then to 65°C at 5°C/min (held for 5 min), and finally to 200°C at 15°C/min (held for 5 min). The injection port temperature was maintained at 260°C.
Direct Injection GC/MS
Semi-volatile organic compounds in the samples were analyzed using an Agilent 7890B-5977A GC/MS systems. Prior to sample injections, 2 mL methylene chloride was added to the sample for liquid-liquid extraction (LLE), and the organic layer was collected for analysis. The efficiency of LLE sample pre-treatment was determined to be 70% to130% based on recovery studies using 3,4-dimethylbenzaldehyde, decyl decanoate, and octamethylcyclotetrasiloxane. A volume of 1 µL of the organic layer from the pretreated sample solution was injected on an Agilent DB-1 MS column (60 m × 0.25 mm × 0.25 µm) using helium as the carrier gas. The oven temperature was programmed from 50°C (held for 1.5 min) to 135°C at 17°C/min (held for 5.5 min), and finally to 300°C at 12°C/min (held for 6.5 min). The injection port temperature was maintained at 260°C.
LC/PDA/MS
Analysis of non-volatile compounds in the samples was carried out using a Waters Acquity Ultra Performance Liquid Chromatography (UPLC) system equipped with a PDA detector and single-quadrupole MS system. Reversed phase chromatographic separation was accomplished using a Waters Acquity UPLC BEH C18 column (50 mm × 2.1 mm, 1.7 µm) maintained at 60°C column temperature and mobile phases A: water with 0.01% formic acid (v/v) + 3 mM ammonium formate, and B: methanol with 0.01% formic acid (v/v) + 3 mM ammonium formate. The elution gradient program employed was 2% B (0.0 to 1.0 min), 2% B to 100% B (1.0 to 8.0 min), 100% B (8.0 to 10.0 min), 2% B (10 to 10.2 min) at 0.45 mL/min flow rate. MS analyses were performed in the electrospray positive and negative modes with a scan range of m/z 100 to 1400. PDA detection wavelength range was set to 210 to 400 nm. For LC/PDA/MS analysis, replicate injections were performed to ensure consistency of results. Prior to sample injections, the samples were pretreated by adding 8 mL acetone per 2 mL sample, and the resulting supernatant liquid was concentrated under reduced pressure. The efficiency of sample pretreatment was determined to be 70% to 130% based on recovery studies using diphenyl sulfone, bisphenol A, erucamide, and irganox 1010.
ICP/MS
The samples were assayed using Agilent 7900 ICP/MS system. Prior to sample injections, the samples were digested in a microwave oven digester at 200°C for 20 min with 3.0 mL acid mixture (HNO3:HCl, 3:2 v/v) per 1 mL sample. The digested sample was transferred into a 50 mL DigiTUBE and diluted to 100 mL with deionized water. The efficiency of digestion pretreatment was determined to be 70% to 150% based on recovery studies of elements. The limit of detection of the analytes of interest was in the range of 0.1 to 5 ppb. Spiked analytes detected in the samples were quantitated via a 4-point calibration curve using standard solutions (0.5, 5, 10, and 50 ng/mL) prepared in diluent (3% HNO3, 2% HCl in ICP-MS grade water). The helium (He) collision gas flow rate was 4.3 mL/min. The nebulizer pump speed was 0.30 revolutions per second (rps).
CY43917-10Oct24-AN