3.1.3. Optimization of extraction conditions by an orthogonal array experimental design
Apart from above two critical factors, the remaining variables that could potentially affect extraction efficiency included sample quantity, acid solution volume, extraction temperature and extrac- tion time. These four variables were systematically investigated by means of an orthogonal array experimental design L9 (34) to optimize the operation conditions. All the experiments were conducted in triplicate by detecting the above spiked cosmetic samples. The level set values of four factors (A, B, C and D) and the mean recoveries were shown in Table 1. Numerical analysis of the experimental values led to the ANOVA results listed in Table 2. The R factor was used for evaluating the influential variable. The variable with the largest value of R was the most important factor that influenced extraction efficiency. Compared with all the R values, extraction temperature (C) was the significant factor for four analytes including MEA, DEA, DMA and TEA, while samplequantity (A) was only significant factor for one analyte (DIEA). These two variables appeared to have a positive and a negative effect on the extraction efficiency. The other two variables (B and D) had less influence on recovery. The further analysis of variance of recoveries indicated that extraction temperature had statisti- cally significant influence on extraction efficiency (F 1⁄4 3.412 lager than critical value F0.05 3.110) while the other three variables (A, B and D) had not marked influences (F-value less than 3.110). This is due to the fact that cosmetic sample matrices were complex colloids and had the worse diffusion ability in solution, and a high extraction temperature could decrease viscosity of cosmetic solu- tion and enhance velocity of matrix diffusion and mass transfer of analytes, which led to the improvement of extraction efficiency in short equilibrium time. According to the largest donating rule, the largest value of K which affects the extraction efficiency should be the selected value. In Table 2, the K1–K3 values were average recoveries at three levels of the studied variables, respectively. Considering time saving, the most favorable extraction condition was established as follows: extraction solvent, 10 mL of 2.5 mM MSA; extraction time 10 min; and extraction temperature 70 oC. It indicated that the extraction efficiency could be enhanced by using