*3.1. 1HNMR Spectra of ENR-d3*

ENR-*d5* is commonly used as an isotope standard for ENR. With the aim of obtaining a new stable isotope surrogate, ENR-*d3* was synthesized from commercial CIP and 1-bromoethane-2,2,2-*d3* via the SN2 substitution reaction. The structure of ENR-*d3* was firstly characterized by 1H NMR and MS analysis. For 1H NMR (Figure 3), in comparison with CIP (Figure S1), 15 additional protons were observed; these can be attributed to the Et3N salt existing with ENR-*d3* in the spectrum. In detail, the two proton signals in methylene linked to the CD3 group were overlapped with six methylene proton signals of Et3N at 2.51−2.46 ppm. The nine proton signals in the methyl group for Et3N were at 0.99−0.96 ppm. All other signals were in good agreement with the original CIP. The m/z 363 in ENR-*d3* was obtained from product ion mass spectra (Figure S2), and the m/z 102 at the top of this figure shows the protonated Et3N, which was also detected in the ENR-*d3* 1HNMR spectra (Figure 3). Pure m/z 363 without an m/z 360 signal demonstrated the high purity of the synthesized ENR-*d3* without the presence of ENR. Moreover, the HPLC showed that the purity was >95% (Figure S3). Hence, the synthesized ENR-*d3* can be applied for HPLC−MS/MS analysis as an isotope surrogate.

**Figure 3.** 1H NMR spectrum of ENR-*d3*.

### *3.2. Calibration Range and Deviation*

With the aim of detecting ENR with a high variation in concentration in one batch of samples, two calibration curves were individually prepared using ENR-*d5* (5 ng/mL) and ENR-*d3* (100 ng/mL) as isotope surrogates with blank matrix solutions. The two isotope surrogates can compensate for extraction loss during the sample pretreatment and instrument measurement [40]. As shown in Figure 4, the calibration curves were tested for external calibration, and with ENR-*d5* and ENR-*d3* for internal calibration, respectively. It is clear that with such a wide calibration range, the external calibration did not produce a good coefficient of determination (*r2* < 0.99), while internal calibration with ENR-*d5* and ENR-*d3* allowed a satisfactory calibration of *r2* > 0.999. The internal calibration curves exhibited linearity with the ratio of the peak area of the analyte/ isotope standard (y) and the concentration of ENR (x). Firstly, the two calibration curves with different isotope surrogates covered a wide linear range of 1−6561 ng/mL. However, the measured concentration deviating from the specified amount was more than 10% above 243 ng/mL in the calibration curve with ENR-*d5* (Table S1). On the other hand, the deviations between the spiked and measured values were up to 32% and 111% at concentrations of 1 ng/mL and 3 ng/mL, respectively, in the curve with ENR-*d3* (Table S2). The deviations at different concentrations with two isotope surrogates demonstrated the difficulty of accomplishing accuracy when a large concentration difference between the isotope surrogates and the analyte is present. In our study, MQL was set as the first point of the calibration curve using ENR-*d5*, varying from 1 to 243 ng/mL (2−486 μg/kg) with a coefficient of determination of 0.9997 (Table 1). For the calibration curve using ENR-*d3*, the range for ENR was from 27 to 6561 ng/mL (54−13,122 <sup>μ</sup>g/kg) with *<sup>r</sup>*<sup>2</sup> = 0.9996. As a result, the entire calibration curve, covering a wide quantification range from 2 to 13,122 μg/kg, was established for ENR detection.

**Figure 4.** Calibration curves of ENR with the external method (**a**), ENR-d5 (**b**), and ENR-d3 (**c**) for internal calibration.
