*3.4. Statistical Analysis of the Physicochemical Parameters and Removal Efficiency of Ten Pesticides*

According to the correlation analysis (Table 5), the dominant physicochemical parameter affecting the reduction of pesticide residues during thermal processing (blanching and boiling) was log *P* (negative correlation) and that for the running water washing method was water solubility (negative correlation). PCA was performed to better understand the correlation between various physicochemical parameters of pesticides and pesticide reduction in each of the nine removal methods. Figure 3 shows the PCA results of the nine methods and their characteristic results. Of note, the relationship between the molecular weight, polarity (log *P*), water solubility, melting point, Henry's constant of the pesticides, and reduction is discussed below for each method of pesticide residue removal. Score plots were obtained for each removal method using PCA. The first and second principal components (PC1 and PC2, respectively) were selected according to the Kaiser's rule of selecting principal components with eigenvalues greater than 1 (Figure 3a–i1). PC1 and PC2 were further analyzed.


**Table 5.** Correlation analysis between nine washing methods and properties of pesticide.

\* *p* < 0.05; \*\* *p* < 0.01.

**Figure 3.** The plots of PCA: The scree (**a**–**i1**), score (**a**–**i2**), and loading (**a**–**i3**) plots. A: Azoxystrobin, B: Chlorantraniliprole, C: Chlorfenapyr, D: Diniconazole, E: Fludioxonil, F: Imidacloprid, G: Indoxacarb, H: Lufenuron, I: Pyraclostrobin, and J: Thiamethoxam.

Score plots and loading plots revealed that the PC1 and PC2 accounted for more than 68% of the total variance in PCA. The two significant principal components accounted for a proportion of the total variation: running water: 67.93% (PC1: 46.97% and PC2: 20.96%), boiling: 70.3% (PC1: 50.08% and PC2: 20.26%), detergent: 69.14% (PC1: 44.52% and PC2: 24.63%), alkaline water: 70.3% (PC1: 45.73% and PC2: 24.58%), blanching: 71.0% (PC1: 50.69% and PC2: 20.31%), vinegar: 72.14% (PC1: 47.70% and PC2: 24.44%), stagnant water: 67.76% (PC1: 45.14% and PC2: 22.61%), ultrasonic cleaning: 69.64% (PC1: 44.66% and PC2: 24.98%), and NaHCO3: 71.12% (PC1: 48.03% and PC2: 23.10%). The pesticides were categorized based on the scores and loadings and grouped into three clusters, as shown in Figure 3a–i2. Each group consisted of the following pesticides: first pesticides group (imidacloprid and thiamethoxam), second pesticides group (chlorantraniliprole, diniconazole, fludioxonil, and lufenuron), and third pesticides group (azoxystrobin, chlorfenapyr, indoxacarb, and pyraclostrobin). Imidacloprid and thiamethoxam, classified into the first group, displayed similar reduction patterns in all methods based on Figures 2 and 3. The first group had log *P* less than 1 and high water solubility, and the second and third groups had log *P* greater than 1 and low water solubility.
