Rapid identification and quantification of 27 primary aromatic amines in kitchen utensils


Parts-per-trillion (ppt) sensitivity using the SCIEX 7500 system

Sabarinathan1 , Sashank Pillai1 , Jessica Smith2 , Jack Steed2 and Jianru Stahl-Zeng3
1SCIEX, India; 2SCIEX, UK; 3SCIEX, Germany 

Abstract


In this technical note, a method was developed to analyze 27 primary aromatic amines (PAAs) in food contact materials (FCMs). The method achieved LOQs of 0.001-0.50 ng/mL, which correspond to detection limits several orders of magnitude lower than EU regulation mandate. In addition, the high sensitivity of the SCIEX 7500 system permitted dilution of the sample extracts to reduce matrix effects and improve data quality, as shown by the high accuracy and precision (%CV <15%) of spiked matrix samples (Figure 1).

RUO-MKT-02-15224-A1

Introduction


PAAs are widely used in the production of certain colorants and azo pigments used in a wide range of products such as kitchenware, paper napkins, and printed packaging. PAAs have raised concern because they can migrate FCMs into food 1 and be ingested. Most PAAs are considered safe and might not impact human health but some PAAs have been identified as possible carcinogens.

Due to rapidly changing global regulations, the detection and quantification of PAAs in FCMs is critical to consumer safety. Specific migration limits have been established by EU regulations for PAAs in FCMs. Therefore, a detectable limit of 10 ng/mL (EU 2020/1245 amending and correcting regulation EU No 10/2011) in food or a food simulant is applied to the sum of PAAs released.2  

Figure 1. Representative extracted ion chromatogram (XIC) of 3,3’- dimethylbenzidine with ion ratio tolerance lines overlaid for the quantifier and qualifier ions. Compound identification was based on ion ratio calculation. The tolerance levels were set to ±30% for the quantifier and qualifier ions. This calculation was performed at the LOQ of 0.0025 ng/mL in solvent (left) and at 1 ng/mL in spiked spoon sample extract (10 ng/mL in-sample, 10x diluted) (right).

Key features of the SCIEX 7500 system for the analysis of PAAs 
 

  • Simple, robust, reproducible and rapid sample preparation

  • Good chromatographic separation with a fast run time of 12 minutes

  • Due to the excellent sensitivity of the SCIEX 7500 system, LOQs achieved in diluent were at sub-ng/mL levels

  • A single standard calibration curve in diluent was used for 2 different FCM matrices. The majority of the compounds analyzed were within acceptable limits for accuracy (70- 130%) and precision (%CV <15%)

  • All quantification results used 2 transitions and were confirmed using both quantifier and qualifier ions

Methods


Standard preparation: Individual standard stock solutions were prepared by dissolving 1 mg of neat standard into 1 mL of the appropriate solvent before vortexing for 1 min (Table 4). The individual stock solutions were used to prepare a mixed solution of 27 PAAs at 200 ng/mL in methanol, which was diluted with 70:30 water/methanol (diluent, Table 1) to cover a linear range of 0.001-10 ng/mL.

Sample preparation: A 6 g sample of spoon or cake mold was weighed into a 50 mL glass centrifuge tube and 12 mL of 3% acetic acid in water (simulant solution, Table 1) was added. The solution was vortexed for 2 minutes and then incubated for 2 hours at 80°C. After incubation, the samples were filtered using a 0.22 µm PTFE syringe filter (Phenomenex part #: AF8-6710- 12) prior to analysis.

Post-spiked sample preparation: Samples were extracted following the sample preparation protocol described above. After filtration, 0.950 mL of the filtered sample and 0.050 mL of the 200 ng/mL stock solution were used to prepare a 10 ng/mL solution. The samples were vortexed and transferred into autosampler vials for the analysis. To evaluate the effect of sample dilution on recovery and accuracy, the 10 ng/mL postspiked sample was diluted 1:9 ratio with methanol to obtain a 1.0 ng/mL solution (i.e., 10x dilution). The diluted samples were vortexed prior to analysis. 

Table 1. Simulant, dilution solvent, and diluent for PAA

Chromatography: An ExionLC AD system was used with a Phenomenex Kinetex F5 analytical column (2.6 µm, 100 x 3.0 mm). Table 2 outlines the gradient conditions that were used with a flow rate of 0.300 mL/min. A 5 µL injection volume was used and the column oven temperature was set to 40°C.

Table 2. Gradient program for the analysis of 27 PAA compounds

Mass spectrometry: The SCIEX 7500 system was used with electrospray ionization operating in positive mode. Data were acquired using scheduled MRM mode (sMRM) to optimize dwell time and the number of data points collected across the chromatographic peak. The source and compound-specific parameters used are presented in Tables 3 and 4, respectively. Individual Q0D values were individually optimized for all 27 analytes (Table 4) to reduce the background noise and to improve sensitivity and selectivity.

Data processing: All data were processed using SCIEX OS software, version 2.1.6.

Table 3. Optimized source parameters for the analysis of 27 PAA compounds.

Table 4. Stock solution diluent, MRM conditions and compound-dependent parameters for 27 PAAs tested. Due to the variability of compound solubility, the 1 mg/mL stock was prepared with individual solvents. 

Chromatographic separation of 27 PAAs using a 12 min LC gradient


The chromatographic gradient used resulted in good separation of the 27 PAAs in 12 min. The suite of PAAs covered a moderate range of compound polarity, requiring careful LC gradient development. For example, the method achieved the baseline separation of a pair of PAA isomers, 2,4- and 2,6- dimethylaniline (Figure 2). Although not all isomer groups were separated, this is acceptable, according to EU regulations. Further, the separation of the 27 PAA compounds was obtained while running at HPLC pressures due to the 2.6 µm particle size and relatively low flow rate.

Linear dynamic range and sensitivity


Calibration curves for all 27 PAA compounds showed r2 values >0.99 with accuracies ranging between 80% and 110% (Table 5). The calibration curves generated for 4-aminophenylthioether and 3,3’-dimethylbenzidine are shown in Figure 3 and display a linear range of 0.0025 to 10.00 ng/mL. The SCIEX 7500 system showed excellent sensitivity for the entire suite of 27 PAA compounds with LOQs ranging between 0.001 and 0.5 ng/mL for standards prepared in the 70:30 water/ methanol diluent. The LOQ value was selected based on 2 selective MRM transitions, S/N ratio >10 for quantifier and qualifier of calibration standard, accuracy ±30%, %CV <15% and ion ratio tolerance ±30%. Overall, the observed sensitivity was several orders of magnitude greater than 10 ng/mL, as currently required by the EU regulation. Figure 4 shows the quantifier transition XIC for 4-aminophenylthioether at the LOQ concentration (0.0025 ng/mL). For comparison, the overlaid quantifier and qualifier ion transitions at 1 ng/mL in spoon extract matrix (post-spiked at 10 ng/mL and diluted 10x) are shown in the right panel. The ion ratio tolerance was within ±30%, which met the acceptance criteria.

Figure 2. XIC of 27 PAAs at 1 ng/mL from the quantifier transitions. Good chromatographic peak-to-peak separation was achieved using the Phenomenex Kinetex F5 column. Highlighted traces indicate 1) Anisidine, 2) o-Toluidine, 3) 2-Aminonapthalene, 4) 4,4'-Methylene-bis(2-chloroaniline) and 5.) 4-Amino-2',3-dimethylazobenzene. 

Table 5. Calibration curve correlation coefficient (r2 ) and accuracy range across the calibration curve for quantifier ion. 

Figure 3. Representative calibration curve for 4-aminophenylthioether (left) and 3,3’-dimethylbenzidine (right). The 1/x2 weighting factor was used. A linear range of 0.0025-10 ng/mL with an r2 value of >0.99 was achieved. 

Figure 4. Chromatogram of 4-aminophenylthioether. Left) Representative XIC of 4-aminophenylthioether at the LOQ level of 0.0025 ng/mL from the quantifier transition (m/z: 217.1/ 124.0). Excellent peak shape and sensitivity were achieved. Right) Representative XIC for 4-aminophenylthioether with the ion ratio overlaid for the quantifier and qualifier ions at 1 ng/mL (post-spiked 10ng/mL solution diluted with methanol) in spoon sample.

Post-extraction spiked samples


A 10 ng/mL mixed standard solution was spiked into the processed spoon and cake mold samples and diluted 10x with methanol (1 ng/mL final concentration) to evaluate the analyte recovery in spoon and cake mold samples (n=2 per matrix). The accuracy data were calculated from 6 injections of quality control samples (1 ng/mL) and the calibration curves generated for all PAAs diluated in 70:30 water/methanol. The accuracy range achieved between 70-130% for most of the analytes and %CV <15% for all compounds analyzed (Table 6). In addition, the 10x diluted samples showed improved accuracy, compared to the undiluted samples, presumably due to reduced matrix effects. The enhanced sensitivity of the SCIEX 7500 system permitted sample dilution while maintaining low detection limits. These results highlight an additional benefit of the SCIEX 7500 system sensitivity, improved data quality from reducing potential matrix effects through sample dilution.

Table 6. Average accuracy and %CV (n=6) for all 27 PAAs in undiluted and diluted spoon and cake mold samples. All recovery measurements (n=6) were performed at 10 ng/mL (post-spiked) and 1 ng/mL (post-spiked 10 ng/mL, diluted with methanol in 1:9 ratio) against the single external calibration curve. The average accuracy and %CV were calculated from the quantifier ion.

Unspiked spoon and cake mold samples


Spoon and cake mold samples were processed using the sample preparation procedure described and diluted 10x with methanol before analysis to reduce potential matrix effects. The unspiked samples were analyzed against a single external calibration curve in solvent for the presence of the 27 PAA compounds. The values obtained in the unspiked samples were below the LOQ level for the 2 analytes detected (Table 7).

Table 7. Analysis of unspiked spoon and cake mold samples. Unspiked samples showed detection of 2 PAAs.

Conclusion
 

  • The SCIEX 7500 system showed excellent sensitivity, linearity and reproducibility for the quantification of PAAs in kitchen utensils

  • Using a calibration curve prepared in 70:30, water/methanol diluent, excellent accuracy (80-110%) and precision (%CV <15%) was achieved for all PAAs analyzed. LOQs were excellent, ranging between 0.001 and 0.50 ng/mL.

  • When quantifying the spiked samples against an external standard calibration curve, the recovery was between 70% and 130%, which meets the acceptance criteria for most PAAs

  • The linear range was assessed across a concentration range of 0.001-10.00 ng/mL with an r2 value > 0.99 achieved for all compounds

  • Good chromatographic separation was achieved with a fast run time of 12 minutes and an isomer pair was separated 

References
 

  1. Rapid Determination of 41 Primary Aromatic Amines in Water. SCIEX technical note, RUO-MKT-02-10038-ZH-A.

  2. Commission Regulation (EU) 2020/1245 of 2 September 2020 amending and correcting Regulation (EU) No 10/2011 on plastic materials and articles intended to come into contact with food

  3. Rapid Determination of 33 Primary Aromatic Amines in Kitchen Utensils, Textiles Dyes and Food Packaging Materials. SCIEX technical note, RUO-MKT-02-7950-B.

  4. Patrick Kämpfer, Stéphanie Crettaz, Susanne Nussbaumer, Michael Scherer, Scott Krepich, Otmar Deflorin. Quantitative determination of 58 aromatic amines and positional isomers in textiles by high-performance liquid chromatography with electrospray ionization tandem mass spectrometry. J Chromatogr A. 2019 May 10;1592:71-81 
Request information