Sensitive determination of polybrominated diphenyl ethers (PBDEs) in water and milk by effervescence-enhanced switchable hydrophilic-hydrophobic solvent-based salting-out microextraction and high-performance liquid chromatography with photodiode array detection (HPLC-PDA)

Abstract Here is reported effervescent reaction-enhanced switchable hydrophilic-hydrophobic solvent-based salting-out microextraction (ERSSM) with high-performance liquid chromatography with photodiode array detection (HPLC-PDA) for the preconcentration, extraction, and trace-level determination of polybrominated diphenyl ethers (PBDEs) in milk and water. The important variables were rigorously optimized to be 150 µL of hexanoic acid as the extraction solvent, 400 µL of Na2CO3 as the alkaline source, 350 µL of H2SO4 solution as the acid source, and NaCl as the salting-out agent. The optimized conditions provided a linear range from 0.2 to 50 µg L−1 with limits of detection and quantitation from 0.026 to 0.085 µg L−1 and 0.087 to 0.28 µg L−1 for six PBDEs. The intra-day and inter-day precision values were from 2.48% to 6.75%, and the fortified recoveries for PBDEs were between 78.06% and 103.07% in water and milk samples. The developed procedure avoids the use of a traditional organic disperser and mechanical mixing allowing the determination of PBDEs outdoors by synchronous extraction method and high efficiency for the analytes. These advantages suggest a high potential for the routine monitoring of PBDEs in liquid food and environmental samples.


Introduction
Polybrominated diphenyl ethers (PBDEs) are an important class of flame retardants, which are of concern owing to their toxicity to aquatic organisms and human health (Zhang et al. 2020).The recycling of discarded electrical apparatuses releases large quantities of PBDEs (Paliya et al. 2022;Ryota et al. 2022).Some PBDE congeners are prohibited from use by the European Union and the United States (Mizukawa et al. 2021).Although some PBDEs have been banned, their environment or food remains due to their low biodegradability (Shi et al. 2019).
Currently, PBDEs are ubiquitously detected in various environmental and food matrices around the world and affect biological development in the immune, nervous, and reproductive systems (Wu et al. 2020).Therefore, the development of simple, rapid, and convenient methodologies is crucial for determining PBDEs in environmental and food matrices.
As an alternative, effervescent reaction-enhanced microextraction (EREM) was developed by Lasarte-Aragon es et al. (2011).This method rapidly disperses the extraction solvent into the sample by vigorous CO 2 bubbles from an acid-base reaction (Xu et al. 2020).The superiority of EREM is that no additional electrical power is required for the dispersion and extraction processes, such as ultrasound, shaking, vortexing, and centrifugation.The magnetic effervescent tablet provides rapid dispersion, extraction, separation, and collection into one synchronous step (Lasarte-Aragon es, Lucena, and C ardenas 2020).
Switchable hydrophlic-hydrophbic solvents (SHSs) are green with superior extraction performance and environmental friendliness, and to date have been widely used to replace conventional organic solvents for separation and enrichment of organic pollutants (Bozyi git et al. 2020;Ren et al. 2022).SHSs reversibly switch between two forms to promote rapid separation and eliminate the tedious/time-consuming steps in extraction (Musarurwa and Tavengwa 2020).Compared to traditional solvents, the primary advantage of SHSs is no requirement for distillation, separation, or centrifugation (Liao et al. 2022), leading to their widespread applications (Di et al. 2021;Zhang, Wang, et al. 2022).
Our group developed switchable fatty acid-based microextraction based upon the solidification of floating organic droplets for determining fluoroquinolones in seawater (Gao et al. 2018).Torbati's group integrated the salt/pH-induced homogeneous liquidliquid microextraction with a laboratory-constructed extraction device for pyrethroid pesticides in fruit juice (Torbati et al. 2018).Both microextraction approaches require cumbersome low-temperature solidification and room-temperature dissolution.Jing and coworkers engineered effervescence-assisted dual microextraction for determining polyaromatic hydrocarbons in edible oils (Jing, Chen, et al. 2021).
As green solvents, the advantages of medium-chain fatty acids (MFAs) are their hydrophobic-to-hydrophilic convertibility (Lebedinets et al. 2020).When the pH is > pKa, MFAs are dissolved in the aqueous phase due to the surfactant properties of their ionization forms.Conversely, when the pH is < pKa, the MFAs are in molecular states with water-insoluble, producing a clear interface layer with the aqueous phase (Gao et al. 2018).Hence, MFAs are pH-dependent solvents.The hydrophobic-to-hydrophilic conversion of MFAs may be performed by an acid-base effervescent reaction (Jing, Huang, et al. 2021).Hence, the combination of MFAs with effervescent reaction may offer environmentally benign, convenient, and rapid preconcentration/extraction of PBDEs.
Here is reported a novel effervescent reaction-enhanced, SHSs-based salting-out microextraction (ERSSM).This pretreatment was integrated with high-performance liquid chromatography with photodiode array detection (HPLC-PDA) for the trace determination of PBDEs in the milk and water.In the ERSSM procedure, MFAs were employed as extraction solvents, which switched from hydrophilic to hydrophobic using an effervescent reaction.Simultaneously, the salting-out effect promoted the stratification phenomenon without controlling the temperature.ERSSM/HPLC-PDA offers no requirements for traditional organic dispersive solvents and physical mixing, allowing outdoor determination of PBDEs in remote environments with synchronous microextraction to provide enhanced efficiency for the analytes.Thus, the present approach provides an alternative to conventional monitoring of trace-level PBDEs.
The stock solution (1 mL) of each PBDEs' congener was treated by an N 2 flow to remove isooctane.Acetonitrile was employed to dissolve PBDEs for preparing 10 mL of 10 mg L À1 stock solution.The working solution was obtained by appropriately diluting the stock solution using ultrapure water.For the fortified recovery experiments, the working solution was added to the pretreated water or milk samples at 1, 10, and 40 mg L À1 .

Apparatus
A Leici PHB-4 pH meter, produced by the Inesa Scientific (Shanghai, China), was used for measuring the solution pH.The centrifugation and blending steps were conducted by a TDL-5 centrifuge and a XW-80A vortex mixer (Jiangsu Jintan Medical Instrument, Jintan, China).A mill was applied for pulverizing Na 2 CO 3 and NaHCO 3 (Wenzhou, China).A Milli-Q system (Bedford, MA, USA) was used to produce ultrapure water.

Detection of PBDEs by HPLC-PDA
The PBDE concentrations were determined by a Shimadzu LC-20AT HPLC-PDA.A Shim-Pack GIST C 18 column (250 mm Â 4.6 mm, 5-lm particle diameter) was used for separating the analytes.The mobile phase consisted of acetonitrile and water at 85%:15% (v/v) at a flow rate of 1.0 mL min À1 .The detection wavelength and injection volumes were 226 nm and 20 lL.

Collection and preparation of the water and milk
Water from local rivers was passed through 0.45-lm membranes to remove particles and impurities and stored at 4 C for further use.
Low-fat milk (high calcium and low fat, Yili, China) and whole-fat milk (high-calcium, Yili, China) were obtained from a local supermarket.An aliquot of the milk (6 mL) and 20% acetic acid (200 lL) were treated with 2.8 mL water, ultrasonically dispersed for 1 min, and centrifuged at 3000 rpm for 15 min.The supernatant was passed through a 0.22-lm membrane filter and stored at 4 C for subsequent ERSSM analysis.

ERSSM
Figure 1 is a schematic of the experimental procedure for ERSSM.Briefly, 150 lL of MFAs were added to 5 mL of the sample.After vortexing for 1-min, clear stratification occurred between MFAs and the aqueous phase.When 400 lL of saturated Na 2 CO 3 were added, the solution became homogeneous/clear due to conversion of MFAs into their salt forms.Subsequently, 350 lL of H 2 SO 4 (95.0%) were added and vortexed for 1 min, and a large volume of CO 2 bubbles were produced, promoting the diffusion of MFAs in aqueous solution.When the effervescent reaction was complete, MFAs switched from their miscible to water-insoluble states, and thus were removed from an aqueous phase.Excess NaCl was added to remove more MFAs by the salting-out effect, which further enhanced the efficiency of the analytes.The upper layer was collected and 20 lL aliquots were directly injected into the HPLC-PDA for analysis.

Enrichment factor and extraction recovery
The enrichment factor (EF) and extraction recovery (ER) were determined by where C C and C 0 are the concentration of the analyte in the organic phase and the initial concentration of the analyte in the sample and V c and V aq are the volumes of the organic phase and sample.

Optimization of the important parameters in the ERSSM method
To obtain the high average recoveries for the PBDEs, the type and volume of MFAs, the volume of Na 2 CO 3 , the volume of H 2 SO 4 and the type of salt were optimized.Each experiment was conducted in triplicate (n ¼ 3) and reported as the mean ± standard deviation.

Type and volume of MFAs
To improve extraction efficiency, an appropriate solvent is crucial in ERSSM (Amini, Khandaghi, and Mogaddam 2018).The solvent should offer low volatility, strong hydrophobicity, good extraction capacity for the analytes, and low chromatographic response (Li et al. 2019).Six MFAs were selected as possible extraction solvents based upon their alkyl-chain lengths.The alkyl-chain length in the MFAs may influence the amphiphilic characteristics of these chemicals.Acidic solvents were employed because acidic protons of the solvents and PBDEs may interact and enhance the solubility of PBDEs.Torbati et al. (2018) previously developed salt and pH-induced homogeneous liquid-liquid microextraction for the preconcentration of pyrethroid insecticides in fruit juice.
Valeric acid, hexanoic acid, heptanoic acid, octanoic acid, pelargouic acid, and decanoic acid were investigated as MFAs in this study.Decanoic acid was in a solid state at room temperature and was unsuitable for ERSSM.Hexanoic acid provided the highest average recoveries for the PBDEs (Figure 2) and was employed in subsequent measurements.
The volume of extraction solvent is another important variable affecting the extraction efficiency.To evaluate the influence of hexanoic acid volume upon the recovery, values between 75 and 200 lL were investigated.An insufficiently low volume led to poor extraction efficiency.Conversely, excess hexanoic acid consumed more Na 2 CO 3 and H 2 SO 4 and decreased the enrichment factor and sensitivity.Figure 3 shows the recoveries were the highest using 150 lL that was selected to be the optimum volume of hexanoic acid.

Effect of alkaline source
In ERSSM, Na 2 CO 3 serves as the alkaline source in the effervescent reaction and ionexchange salt to convert hydrophobic hexanoic acid into its water-soluble salt (Shishov et al. 2020).Consequently, the quantity of Na 2 CO 3 should convert all hexanoic acid into its water-soluble state to transfer the emulsification.However, excess Na 2 CO 3 is required to neutralize H 2 SO 4 and produce the water-insoluble form.The highest average recovery for the PBDEs was obtained using 400 lL that was deemed to be optimum (Figure 4).

Effect of neutralizer
As a divalent proton donor, H 2 SO 4 supplied acid for the effervescent reaction with Na 2 CO 3 (Di et al. 2019).H 2 SO 4 also neutralized Na 2 CO 3 to convert sodium hexanoate into its neutral acid form.Thus, the volume of H 2 SO 4 solution influenced the transformation of MFAs and the subsequent recoveries for PBDEs.The extraction efficiency increased with volume from 200 to 350 lL but decreased at higher values (Figure 5).Consequently, 350 lL of H 2 SO 4 was employed in subsequent experiments.

Salting-out reagent
The salting-out effect reduces the solubility of hydrophilic compounds in an aqueous solution by accelerating the transfer of the analytes into the organic phase (Jing, Xue, et al. 2021;Kaur et al. 2022).In the present study, NaCl, Na 2 SO 4 , NaH 2 PO 4 , K 2 SO 4 , and KCl  were investigated as salting-out agents.Figure 6 shows NaCl provided the highest average recovery for the PBDEs, followed by NaH 2 PO 4 , while KCl produced the lowest value.Therefore, NaCl was selected to be the salting-out agent in subsequent measurements.

Analytical performance of ERSSM/HPLC-PDA
The analytical performance of ERSSM/HPLC-PDA was investigated, including the linear range (LR), coefficient of determination (R 2 ), enrichment factors (EFs), limit of  detection (LODs), and the limit of quantification (LQDs).The linear range was from 0.2 to 50 lg L À1 with R 2 values exceeding 0.9990.The enrichment factors were from 52.4 to 61.4.The limits of detection and quantification were 0.026 and 0.085 lg L À1 based upon 3-fold and 10-fold signal-to-noise ratios.These satisfactory figures of merit demonstrate the high enrichment and good sensitivity for PBDEs by the developed method.
The intra-day and inter-day precision were used to evaluate the reproducibility.For the inter-day precision, six parallel experiments were conducted on six continuous days, and the recoveries were used to evaluate the relative standard deviation (RSD).For the intra-day precision, six replicate experiments were carried out in one day at 2 h intervals.Table 1 shows the intra-and inter-day precision were from 2.48% to 4.78% and 3.00% to 6.75%, respectively.These values meet the precision requirements for determining PBDEs at trace levels in food and environmental samples.

Determination of PBDEs in milk and water
ERSSM/HPLC-PDA was employed for the determination of PBDEs in water and milk samples.Figure 7 shows the typical chromatograms for six PBDEs in unspiked and spiked (40 lg L À1 ) samples.Table 2 shows BDE-154 was present at 1.58 ± 3.75 lg L À1 in river water, but undetectable in the tap water and milk samples.
At three fortification levels (1, 10, and 40 lg L À1 ), the average recoveries for PBDEs were between 78.06 and 103.07%.These results demonstrate that the developed method is simple, efficient, and sensitive for the determination of PBDEs in liquid samples.

Greenness assessment of the ERSSM-HPLC/PAD method
Green chemistry aims to minimize the use and production of toxic and harmful substances (Clarke et al. 2018).Currently, ComplexGAPI and analytical Eco-Scale Table (Płotka-Wasylka 2018) are used to evaluate the greenness.ComplexGAPI establishes pictograms to evaluate greenness of analytical methods (Justyna and Wojnowski 2021) on the bases of sample collection, transportation, preservation, storage to sample preparation and analysis based upon the low, medium, and high environmental impacts of each step designated by green, yellow, and red.The Analytical Eco-Scale assigns appropriate values to a series of parameters, such as reagent mass, toxicity, and energy consumption in the analytical processes.These values are used to calculate penalty points (PPs) based upon each parameter compared to the ideal value to obtain a final score (Gałuszka et al. 2012).When the final score exceeds 75, the method meets the greenness standard.Figure 8 and Table 3 compare ERSSM/HPLC-PDA with temperature-controlled airassisted dispersive liquid-liquid microextraction (TC-AA-DLLME-HPLC-UV) (Zhang, Guo, et al. 2022).The greenness scores were 74 for TC-AA-DLLME-HPLC-UV and 86 for ERSSM/HPLC-PDA.The findings from the ComplexGAPI pictograms are in agreement with those using the Eco-Scale Table, which confirm that ERSSM/HPLC-PDA is greener than TC-AA-DLLME-HPLC-UV.
Comparison of the ERSSM/HPLC-PDA with the literature ERSSM/HPLC-PDA was compared with the literature on the bases of limits of detection, extraction time, type and volume of extraction solvent, and type and volume of dispersive solvent.Supplementary Table S1 shows the provided method provides comparable limits of detection to SPME-Orbitrap MS (Liang et al. 2023).However, the ME-Orbitrap MS uses an expensive mass spectrometer.In addition, the developed method offers a shorter extraction period compared d-SPE/GC-MS/MS (Liu et al. 2021), MSPE/ APGC-MS/MS (Zhang et al. 2021), EA-DLLME-SAP/GC-MS-MS (Wang et al. 2021), SPME/GC-ECD (Li et al. 2023), and SPME-Orbitrap MS (Liang et al. 2023).This approach also employs hexanoic acid as the extraction solvent which is less toxic than n-hexane (Liu et al. 2021) and ethyl acetate (Zhang et al. 2021).Moreover, no dispersive or elution solvent is required for ERSSM/HPLC-PDA.Conclusions ERSSM/HPLC-PDA is reported for the determination of PBDEs in water and milk.The MFAs serve as extraction and dispersive solvents using an effervescent acid-base reaction.CO 2 bubbles promote the dispersion of the MFAs which eliminates the need for mechanical mixing.The salting-out reagent minimizes losses of MFAs and increases the extraction efficiency for the analytes.ERSSM/HPLC-PDA offers two significant advantages.First, organic dispersive solvents and mechanical mixing are not required, thereby allowing the monitoring of PBDEs outdoors without electrical power.Second, the synchronous protocol provides enhanced extraction efficiency.The reported protocol offers simple and green operation with the potential for the outdoor determination of environmental water and food samples.Table 3. Penalty points (PPs) and overall score for the TC-AA-DLLME-HPLC-UV (Zhang, Guo, et al. 2022) and ERSSM-HPLC-PDA (this work) microextraction procedures.

Figure 1 .
Figure 1.Schematic of effervescent reaction-enhanced switchable hydrophilic-hydrophobic solventbased salting-out microextraction combined with high-performance liquid chromatography equipped with photodiode array detection (ERSSM/HPLC-PDA): (a) switchable hydrophilic-hydrophobic principles of MFAs, (b) ERSSM/HPLC-PDA procedure, and (c) physical phenomena.Steps: (A) sample, (B) 150 lL fatty acid with vortex mixing, (C) addition of Na 2 CO 3 , (D) addition of H 2 SO 4 with vortex mixing for 1 min, (E) addition of NaCl, and (F) separation from the aqueous phase due to the salting-out effect.

Figure 2 .
Figure 2. Optimization of extraction solvent upon the PBDE recovery as the mean ± standard deviation (n ¼ 3).Conditions: volume of the MFAs, 150 lL; fortified concentration of the PBDEs congeners, 40 lg L À1 ; 400 lL of Na 2 CO 3 as the base source; 350 lL of H 2 SO 4 as the neutralizer; and saturated NaCl as the salting-out agent.

Figure 3 .
Figure 3. Optimization of the volume of heaxanoic acid upon the recovery as mean ± standard deviation (n ¼ 3).Conditions: PBDE concentration, 40 lg L À1 ; 400 lL of Na 2 CO 3 as the base source; 350 lL of H 2 SO 4 as the neutralizer; and saturated NaCl as the salting-out agent.

Figure 4 .
Figure 4. Influence of the volume of Na 2 CO 3 upon the recovery as mean ± standard deviation (n ¼ 3).Conditions: 150 lL of hexanoic acid as the extraction solvent; concentration of the PBDEs congeners, 40 lg L À1 ; 400 lL of Na 2 CO 3 as the base source; 350 lL of H 2 SO 4 as the neutralizer; and saturated NaCl as the salting-out agent.

Figure 5 .
Figure 5. Influence of the volume of H 2 SO 4 upon the PBDE recovery as mean ± standard deviation (n ¼ 3).Conditions: volume of MFAs, 150 lL; concentration of the PBDEs congeners, 40 lg L À1 ; 400 lL of Na 2 CO 3 as the base source; and saturated NaCl as the salting-out agent.

Figure 6 .
Figure 6.Influence of the salting-out agent upon the recovery as mean ± standard deviation (n ¼ 3).Conditions: volume of the MFAs, 150 lL; concentration of the PBDEs congeners, 40 lg L À1 ; 400 lL of Na 2 CO 3 as the base source; and 350 lL of H 2 SO 4 as the neutralizer.

Table 2 .
Analytical performance of ERSSM/HPLC-PDA for the analysis of water and milk (n ¼ 3).