*3.3. Method Validation*

For the spiking experiments, the ENR standard was spiked in blank samples at 2, 6, 54, 486, 1458, and 4373 μg/kg, respectively, with the further addition of ENR-*d5* (50 ng) and ENR-*d3* (1000 ng) as isotope standards. Sample preparation was performed according to the above-mentioned procedure. Afterward, two solutions (solution 1 and solution 2) for each sample were measured by HPLC–MS/MS. Compared with the reported methods, our sample preparation process is obviously easier to operate without any enrichment [1] or clean up using SPE material [41], and/or hexane [42]. Furthermore, the sample can be directly diluted using the dual internal calibration method and then quantified using the isotope surrogate with a high concentration, which is feasible for accurate quantification of high residue samples without repeated sample preparation. Thus, it is time-efficient and requires less reagent and fewer consumables. Notably, linearity of the calibration curves was achieved regardless of whether ENR-*d5* was used at low or high levels of concentration. Similar results were observed for ENR-*d3*. In other words, the concentrations of ENR-*d5* and ENR-*d3* can be interchanged with a similar good linearity of the calibration. The spiked

samples at 1458 and 4374 μg/kg were quantified using solution 2 with ENR-*d3* as the isotope standard. The residue at MQL of 2 μg/kg was obtained with ENR-*d5,* and the CCα and CCβ were calculated to be 0.5 μg/kg and 1.5 μg/kg, respectively. As shown in Figure 5a, the blank sample shows a clean background signal with this method, and an excellent sensitivity is obtained (Figure 5b). The spiking experiment at 4374 μg/kg does not show a saturation response on the chromatogram (Figure 5c). Recovery ranging from 97.1% to 106% was observed at all levels of spiking samples (Table S3). Precision values evaluated by the RSD of the measurements of three positive samples in different matrices were found to be below 6.90% and 6.49% for intra- and inter-precision, respectively (Table S4 and S5). These results demonstrate that the new method was efficient during extraction, and fully satisfied the requirements of quantification for a broad ENR residue range. It is difficult to find a fish muscle CRM with high ENR residue levels. However, we conducted a test with our dual isotope surrogate method on a newly obtained CRM. The specified ENR residue level of the CRM was 62.5 ± 6.3 μg/kg. The residue level of this CRM could be quantified both with ENR-*d5* and ENR-*d3*. By using the dual surrogate method, residue levels of 58.7 ± 2.3 μg/kg and 64.1 ± 1.2 μg/kg were obtained with the calibration curve prepared using ENR-*d5* and ENR-*d3*, respectively. Moreover, we performed a spiked experiment on this positive sample with a high spiking amount, which was quantified using the dual isotope surrogate method. The result also shows a higher accuracy and stability with ENR-*d3* than with ENR-*d5* (Table 2). This demonstrates the good accuracy of the developed method. Furthermore, this method was compared with previous methods using mass spectrometry for testing for ENR residues in aquatic products, as shown in Table 3. Our method has comparable sensitivity and stability, but a much wider linear range through the use of dual isotope surrogates, which demonstrates the advantages of our method in the sample test for high ENR residue levels.

**Figure 5.** HPLC−MS/MS chromatograms: (**a**) blank sample (EIC of ENR); (**b**) blank samples spiked with ENR at an MQL of 2 μg/kg (EIC of ENR); (**c**) blank samples spiked with ENR at 4374 μg/kg (EIC of ENR); (**d**) positive sample, common carp (EIC of ENR); (**e**) EIC of ENR-*d5*; (**f**) EIC of ENR-*d3* (samples in (**c**,**f**) were diluted at a ratio of 1:20).


**Table 2.** Recovery and RSDs of a spiking experiment in a positive sample with the dual isotope surrogates method for the determination of ENR (n = 3).

**Table 3.** Comparison of the developed method with previous reports of ENR residue tests using mass spectrometry.

