3.1 Selection of conditions for MSPD
Key influencing factors include the ratio of samples and adsorbent, the types of adsorbent, and the types and volume of elution solvent in the MSPD process.
3.1.1 Optimization of sample to the adsorbent ratio (dispersion ratio)
According to the previous reports, C18 was the common adsorbents with better adsorption effect of PBDEs. Firstly, the weight of the vegetable was 0.2 g, and the weight of C18 was 0.2, 0.4, 0.6, 0.8, 1.0 g, respectively (i.e., the dispersion ratio was 1:1; 1:2; 1:3; 1:4. 1:5). Then, PBDEs were eluted with 10 mL the mixture of n-hexane-DCM (1:1, v:v), collected, concentrated near dry with nitrogen gas, and re-solved 500 μL with ACN before sample analysis.
Figure 1a shows that the ER of PBDEs first gradually increased with the increase of the proportion of adsorbents. The reasonable explanation is that increasing the amount of adsorbent can increase the contact area between vegetable and adsorbent. The increase of interaction between the two will lead to a rise in extraction rate. However, when the dispersion ratio exceeds 1:4, the ER of PBDEs decrease. Because of the excessive amount of adsorbents also adsorbed more interfering substances from vegetable, and led to the reduction the recoveries of PBDEs. Therefore, a mass ratio of 1:4 between the sample and adsorbent was selected as the optimal condition for extraction PBDEs, with an ER between 76.5% and 96.2%.
3.1.2 Selection of adsorbents
In this paper, three commonly used adsorbents (C18, PSA, Florisil) were optimized with a sample mass of 0.2 g and a dispersion ratio of 1:4 and the same conditions as 3.1.1. As can be seen from Figure 1b, first of all, the recovery rates of low bromated BDE (BDE-28 and 47) employing MSPD based on single C18 and single PSA as adsorbents are both with no noticeable significant difference, but the recovery of highly bromated BDE (BDE-183) based on MSPD with single C18 as adsorbent higher than that on PSA. The extraction effect of all PBDEs based on a single Florisil as adsorbent is not ideal. Since almost all the commercial solid phase extraction columns are single adsorbents on the market, we also studied the adsorption effect of the composite adsorbent on the PBDEs. The high recoveries of all targets (except for BDE-47), especially BDE-153 and BDE-183, were observed using MSPD with the combination of C18 and PSA. Due to the more strong absorption capacity of PSA on the pigments in the vegetables than the elution solvent, resulted in the elimination of some interfering substance and the more apparent elution solution. In general, the recoveries of all PBDEs ranged from 81.6% to 98.2% in MSPD with the combined C18 with PSA as adsorbents.
3.1.3 Selection of elution solvent
Because the hydrophobicity of PBDE congeners increases with the increase of their molecular weight, non-polar to medium polarity extraction solvents should be selected (Pietron & Malagocki, 2017). According to the previous literature reports on the extraction solvents for PBDEs, six common solvents included n-hexane, DCM, acetone, the mixture of n-hexane-DCM (1:1, v/v), n-hexane-acetone (1:1, v/v), and DCM-acetone (1:1, v/v) (Babalola & Adeyi, 2018; Jagic, Dvorscak, Juric, Safner, & Klincic, 2021; Smielowska & Zabiegala, 2018). The results are shown in Figure 1c. It can be seen that the recoveries of all PBDEs were lowest used single acetone as the elution solvent, and the color of elution solution was black and green due to it dissolving more pigments, and causing interference for subsequent analysis. The ER of BDE-183 using the mixture of n-hexane-DCM (1:1, v/v) as elution solvent was significantly higher than that of n-hexane. Therefore, the ratio of n-hexane to DCM is further verified in this paper, and the results that good recoveries is shown at the volume ratio of n-hexane to DCM of 1:1 for all PBDEs (Figure 1d). In conclusion, the ER of 7 PBDE congeners were 86.1 to 101.1% when n-hexane-DCM (1:1, v/v) was selected as the elution solvent.
3.1.4 Volume of elution solvents
A staged collection method was used to optimize the volume of eluent, and namely eluent was collected every 2.0 mL during the elution process and measured separately. As shown in Figure 1e, the recoveries of all PBDEs in the first 2.0 mL of elution solvent accounted for almost 40%. Virtually all the target substances are eluted when the volume of eluent increased to 8.0 mL, and the recoveries of all PBDEs were less than 1% at the last 2.0 mL of the volume of the eluent. The volume of eluent was finally selected as 8.0 mL according to environmental friendliness and economy.
3.2 Selection of conditions for DLLME
The influence factors include type and volume of extractant, type and volume of dispersant, and reaction time in the DLLME procedure.
3.2.1 Selection of type and volume of extractant
Generally speaking, the extraction of target substances was required to more selective and efficient. It also is required to have the characteristics such as the density of the solvent heavier than that of water, lower water solubility, and better chromatographic response to better separate the sedimentary phase after centrifugation. In addition, the volume of precipitated phase increases with the volume of extractant, and thus leading to the decrease of enrichment. According to the above principles, four common extractant (including DCB, CB, TCA, and CTC), and five levels of extractant volume (such as 20, 30, 35, 40, and 50 mL) were carried out. As shown in Figure 2a, the extraction effect of carbon TCA was significantly better than the other three extraction agents for almost all PBDEs, especially for BDE-28 and BDE-153. It could be observed from Figure 2b that the extraction rate of all PBDEs increased with the increase of the volume of extractant, and decreased slightly when the volume exceeded 35 μL. These results indicated that PBDEs were completely extracted through the volume of less than 35 μL. Excessive extractants increased the volume of the deposited phase and resulted in a decrease in the recoveries rate of PBDEs. So the volume of CTC was 35 μL.
3.2.2 Selection of type and volume of dispersant
The effect of the dispersant is mainly to increase the contact area between the organic phase and the target and accelerate the extraction equilibrium. Four dispersants (ACN, ethanol, methanol, and acetone) commonly used in the DLLME process were studied. Figure 2c showed that the recoveries of all PBDEs using ACN as dispersant were slightly better than that of the other three solvents, especially acetone. The volumes of ACN (250, 500, 750, and 1000 mL) were optimized. Figure 2d showed the recoveries of all PBDEs is highest with 500 mL of ACN, especially BDE-183. In summary, 500 μL of ACN was selected considering resource conservation and environmental protection.
3.2.3 Optimization of reaction time
The reaction time was when the emulsion phase was formed by manual shaking after the extraction agent was added. The effect of reaction time (from 1 min to 10 min) on the extraction effect was investigated. The results showed that the extraction efficiency was improved continuously with the increase in reaction time. The optimal extraction efficiency was achieved when the time was 3min, and decreased when the reaction time exceeded 3 min, indicating that the distribution equilibrium between the target and the extractant was achieved within 3 min.
Further reaction time increases may lead to the transformation or decomposition of some targets. As a result, the reaction time was 3 min, and DLLME having outstanding advantages of short reaction time, simple operation, and high efficiency were also verified.
3.3 Method validation
Calibration curves were established for five different concentration levels in the range of 1-1000 ng·g-1. As shown in Table 1, the seven target compounds had a good linearity with correlation coefficients more significant than 0.9992. The LODs and LOQs of the seven PBDE congeners were 0.08-1.0 ng×g-1 and 0.24-3.00 ng×g-1, respectively.
The accuracy (average recoveries) and precision (intra-day and inter-day) of the method were determined by analyzing vegetable samples spiked with PBDEs standard solution of 10, 20 and 100 ng×g-1, and the matrix effects of several vegetable samples were tested. The results are shown in Table 2. The recoveries of all PBDEs ranged from 82.9% to 113.8% (ranging from 58.5% to 82.5% for BDE-183). RSDs for intra-day and inter-day were between 0.9% and 4.3%, and between 1.2% and 5.2%, respectively.
3.4 Real sample analysis
To test the practicability of this method, six kinds of common vegetables including brassica Chinensis, baby cabbage, Chinese cabbage, carrot, sweet potato, and eggplant were analyzed by the developed method under the optimum. The total content of PBDEs in vegetables ranged from 23.7 to 110.3 ng·g-1, and the abundance PBDE congeners in almost all vegetables (except for ) were BDE-28, BDE-47 and BDE-153, accounting for 43.4%-78.5% of the total PBDEs (Table 3). Furthermore, the addition DLLME procedure after MSPD made fewer interference peaks and better chromatographic response of PBDEs than single MSPD (see Figure S1). The GC-MS/MS chromatograms of PBDEs in six vegetable samples were shown in Figure 4, indicating that the developed method was successfully applied to the detection of PBDEs in actual vegetables.
3.5 Method performance comparison
To evaluate the performance of the proposed method, the reported methods for determining PBDEs in complex substrates are listed in Table 4. Compared with other methods, the present process has more advantages in cost-saving, more eco-friendly, common adsorbents, and rapid and convenient operation, without regulating pH value and temperature.