2193

J. Sep. Sci. 2015, 38, 2193–2200

Lihong Gao1,2 Yali Shi2 Wenhui Li1 Wenli Ren3 Jiemin Liu1 Yaqi Cai2 1 School

of Chemistry and Biological Engineering, University of Science and Technology Beijing, Beijing, China 2 State Key Laboratory of Environmental Chemistry and Ecotoxicology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, China 3 College of Science, Northwest A&F University, Yangling, Shaanxi, China Received February 24, 2015 Revised March 23, 2015 Accepted April 2, 2015

Research Article

Determination of organophosphate esters in water samples by mixed-mode liquid chromatography and tandem mass spectrometry A simple and sensitive method for the determination of organophosphate esters in water samples by mixed-mode liquid chromatography with electrospray ionization tandem mass spectrometry coupled with solid-phase extraction is developed. Using seven alkyl phosphates, three chlorinated alkyl phosphates, and four aryl phosphates as the targets, the developed method was systematically evaluated on the basis of the influence of the solid-phase extraction cartridge, eluting solvent, sample-loading volume, mobile phase condition, and the separation of reversed-phase chromatography and mixed-mode liquid chromatography. Under the optimal conditions, these organophosphate esters can be extracted by ENVI-18 cartridge, eluted by 6 mL of 25% dichloromethane in acetonitrile, and then qualified and quantified by mixed-mode liquid chromatography with tandem mass spectrometry in the multiple reaction-monitoring mode. The application of mixed-mode liquid chromatography endows the separation with reasonable retention for both hydrophilic and hydrophobic organophosphate esters regardless of their polarity, which is hardly achieved by reversedphase chromatography. Good linearity (from 0.9877 to 0.9969), low quantification limits (1–35 ng/L after extraction of 100 mL of river water), and acceptable recovery rates (58.6– 116.2%, with the relative standard deviation 76.92 (>60)

(0.78) 0.36 (1.06) 0.72 (1.34) 2.36 (2.62) 9.69 (8.34) 9.01 (7.81) 1.34 (1.81) 9.92 (8.52) 3.19 (3.27) 14.94 (12.43) 54.97 (43.66) 33.46 (26.88) 14.10 (11.78) 8.67 (7.54) >75.92 (>60)

(0.99) 1.05 (2.03) 1.16 (2.14) 1.34 (2.32) 1.60 (2.57) 1.60 (2.57) 1.16 (2.14) 1.62 (2.59) 1.28 (2.26) 1.71 (2.68) 2.11 (3.08) 1.94 (2.91) 1.61 (2.58) 1.41 (2.39) 4.83 (5.77)

(0.77) 0.53 (1.18) 0.71 (1.32) 1.27 (1.75) 2.79 (2.92) 2.79 (2.92) 0.88 (1.45) 2.57 (2.75) 1.29 (1.76) 3.34 (3.34) 8.19 (7.08) 5.73 (5.18) 3.27 (3.29) 2.04 (2.34) >76.92 (>60)

(0.78) 0.33 (1.04) 0.49 (1.16) 0.96 (1.53) 2.29 (2.57) 2.29 (2.57) 0.60 (1.25) 1.99 (2.33) 1 (1.56) 2.63 (2.83) 6.96 (6.21) 4.64 (4.4) 2.83 (2.99) 1.62 (2.04) >75.92 (>60)

Figure 1. Chromatogram of a 50 ␮g/L standard solution with mixed-mode HILIC-1 column. (1) TMP; (2) TEP; (3) TCEP; (4) TPrP; (5) TCPP; (6) TiBP; (7) TnBP; (8) TDCP; (9) TBEP; (10) TPhP; (11) CDPP; (12) TCrP; (13) EHDPP; (14) TEHP.

and preserved at 4⬚C, and treated by SPE within two days. Before SPE procedure, water samples were filtered through GF/C glass fiber filters (1.2 ␮m, Whatman, UK) to remove suspended particulate matter.

2.3 SPE SPE was performed with a SPE vacuum manifold (BESEP, China). Cartridges were conditioned with 5 mL of acetonitrile and 5 mL of ultrapure water sequentially. Then, 100 mL of water samples were passed through at a flow rate of  C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

4 mL/min. The cartridges were rinsed with 10 mL of ultrapure water, dried for 30 min, and finally eluted with 6 mL of 25% DCM in acetonitrile. The eluent was finally concentrated down to approximately 0.4 mL at 37⬚C under a stream of N2 and diluted to a final volume of 1 mL with ultrapure water. Four commercial available cartridges were tested: Oasis HLB cartridges (6 mL, 200 mg; Waters, USA), ENVI-18 cartridges (6 mL, 500 mg; Supelco), C8 cartridges (6 mL, 200 mg; ThermoFisher, USA), and PEP cartridges (6 mL, 200 mg; ThermoFisher). The PEP packing material is a high purity and highly porous polystyrene divinylbenzene material www.jss-journal.com

2196

J. Sep. Sci. 2015, 38, 2193–2200

L. Gao et al.

Table 2. Linearity and precision of LC-MS detection, limits of detection and limits of quantification

Instrument

Method

Analytes

Calibrationa) (␮g/L)

Linearity (R)

RSDb) (%)

LODsc) (ng/L)

LOQsd) (ng/L)

LODse) (ng/L)

LOQsf) (ng/L)

TMP TEP TPrP TnBP TiBP TCEP TPhP TCPP CDPP EHDPP TCrP TBEP TDCP TEHP

0.05–1000 0.2–1000 0.05–1000 0.2–200 0.2–200 1.0–1000 0.2–1000 1.0–1000 0.2–1000 2.0–1000 0.05–1000 0.2–200 2.0-500 0.05-50

0.9961 0.9950 0.9912 0.9908 0.9922 0.9963 0.9945 0.9964 0.9969 0.9877 0.9926 0.9920 0.9910 0.9929

3.5 1.9 2.9 2.2 2.7 8.3 2.6 2.8 5.8 19.5 6.1 4.2 6.5 12.6

20 50 30 50 50 400 100 400 50 500 20 50 1000 20

50 200 50 200 200 1000 200 1000 200 2000 50 200 2000 50

1 0.6 0.3 0.6 0.6 5 2 6 0.7 9 0.3 0.6 16 0.4

3 2 1 2 2 11 4 14 3 35 1 2 32 1

a)Direct injection of standards prepared in acetonitrile/water (4/6). b)Fifteen microliters injection of a 20 ␮g/L standard (n =10). c)S/N ࣙ 3, direct injection of 15 ␮L standards. d)S/N ࣙ 10, direct injection of 15 ␮L standards. e)LODs (method) = LODs (instrument)/[100 × Recovery (without IS correction)]. f)Method LOQs (method) = LOQs (instrument)/[100 × Recovery (without IS correction)]. Recoveries (without IS correction) of the analytes were obtained by SPE of 100 mL river water at spiked level of 200 ng/L.

modified with urea functional groups to provide the balanced retention of polar and nonpolar analytes. The extraction efficiency of SPE cartridges was calculated as follows: extraction efficiency = (As – Ab )/(Ase – Ab ) × 100%, where As is the peak area measured in the extract from a river water sample spiked before SPE; Ase is the peak area measured in the extract from the same water sample spiked after SPE; Ab is the peak area measured in the extract of a nonspiked fraction from the same water sample.

2.4 LC–MS/MS analysis Analytes were separated using an HPLC system equipped with a P680 pump and Ultimate 3000 autosampler (ThermoFisher). A triple-quadrupole mass spectrometer (API 3200; Applied Biosystems/MDS SCIEX, USA) was connected with HPLC for the determination of analytes. The ESI was operated in the positive-ion mode under multiple-reaction monitoring with the ion spray voltage of 5.0 kV, a source temperature of 600⬚C, a collision gas pressure of 0.02 MPa, and curtain gas of 0.14 MPa. Individual MS/MS parameters for each compound are shown in Supporting Information Table S2. Three chromatographic columns were tested: XterraMS C18 (2.1 × 100 mm, 3.5 ␮m; Waters), Accucore RP-MS (2.1 × 100 mm, 2.6 ␮m; ThermoFisher ), and Acclaim mixed-mode HILIC-1 (2.1 × 150 mm, 5.0 ␮m; ThermoFisher). The stationary phase of Acclaim mixed-mode HILIC-1 column is based on high purity and spherical silica with the functional  C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

silyl ligand consisting of both hydrophilic and hydrophobic functional groups. This new packing material can be used in either HILIC mode (in high organic conditions) or RP mode (in high aqueous conditions). In our study, the mixed-mode HILIC-1 column was used in RP mode. Different compositions of mobile phase were tested for the selected columns as follows: pure water (A) and acetonitrile (B); water containing 0.1% formic acid (A) and acetonitrile (B); water containing 100 mmol ammonium acetate (A) and acetonitrile (B). The selected column was finally performed at a flow rate of 0.25 mL/min and the gradient as follows: initial 40% B with the maintenance for 1.0 min, sequential linear increase to 60% B in 4 min; following increase to 100% B in 3 min and hold for 7 min; then, the gradient return to the initial conditions of 40% B in 0.2 min and hold for 6.8 min to allow for equilibration. The column temperature was kept at 25⬚C and an injection volume of 15 ␮L was used.

2.5 ME and recovery River water samples were used for the study of ME occurring in the ESI source, and it was calculated as follows: ME = [(Ase – Ab )/At ] × 100%, where Ase is the peak area measured for a targeted compound in the extract from water sample spiked after SPE; Ab is the peak area measured in the extract of a nonspiked fraction of the same water sample; At is the peak area for a standard solution containing the same spiked analyte. The values of ME greater or less www.jss-journal.com

Liquid Chromatography

J. Sep. Sci. 2015, 38, 2193–2200

2197

Table 3. Recovery (IS correction) of OPEs from different sample matrices at different spike levels and the MEs

Recovery (IS correction) (%, RSD, n = 6)

ME s (%)

OPEs

Pure water (1 ng/L)

Pure water (5 ng/L)

Pure water (50 ng/L)

Pure water (500 ng/L)

River water (200 ng/L)

River water

TMP TEP TPrP TnBP TiBP TCEP TPhP TCPP CDPP EHDPP TCrP TBEP TDCP TEHP

a a 68.3 (11.7) a a a a a a a 86.8 (16.5) a a 81.0 (15.1)

111.8 (4.6) 85.0 (12.8) 104.5 (2.9) 93.1 (10.1) 115.7 (8.6) a 88.3 (3.9) a 79.8 (10.6) a 63.7 (10.7) 115.1 (8.1) a 94.3 (26.6)

110.7 (1.9) 105.8 (1.6) 97.9 (2.2) 103.2 (3.8) 103.7 (7.9) 99.9 (8.9) 99.3 (1.7) 102.9 (5.9) 72.9 (7.0) 91.0 (26.4) 66.8 (16.3) 126.5 (6.6) 64.9 (15.6) 85.9 (24.1)

120.6 (1.5) 109.3 (1.8) 105.0 (3.8) 112.1 (3.9) 107.4 (4.1) 115.0 (9.9) 110.5 (3.5) 109.6 (2.0) 56.6 (6.9) 80.0 (9.0) 73.1 (6.2) 135.2 (2.2) 65.4 (8.9) 108.1 (22.3)

98.8 (5.0) 98.9 (1.3) 96.5 (1.5) 104.6 (4.1) 100.2 (4.9) 104.1 (9.1) 100.9 (2.3) 99.9 (6.7) 58.6 (3.6) 104.5 (9.9) 64.5 (8.0) 115.0 (3.3) 64.5 (10.1) 116.2 (18.0)

43.0 70.8 81.2 87.0 84.2 66.5 69.4 71.3 50.0 96.1 51.9 86.0 60.4 89.1

a)No recovery determined because the spike level were lower than the LOQs of the analytes.

than 100% indicated signal enhancement or suppression, respectively. The overall recovery (without IS correction) of the procedure was calculated as follows: Recovery (without IS correction) = [(As – Ab )/At ] × 100%, where As is the peak area measured in the extract from a water sample spiked before SPE; Ab is the peak area measured in the extract of a nonspiked fraction of the same water sample; At is the peak area for a standard solution containing the same spiked analyte. The overall recovery (IS correction) of the procedure was defined as follows: Recovery (IS correction) = [(Cs – Cb )/Ct ] × 100%, where Cs is the concentration measured in the extract from a water sample spiked before SPE; Cb is the concentration measured in the extract of a nonspiked fraction of the same water sample; Ct is the concentration added to the sample. The concentrations of Cs , Cb , and Ct were determined using calibration curves obtained by standard solutions prepared in acetonitrile/water (4:6) with constant IS amount of 5 ng.

3 Results and discussion 3.1 Optimization of SPE conditions 3.1.1 Extraction efficiency of SPE cartridges The extraction efficiency of four different cartridges was evaluated by spiking 100 mL of river water at a concentration of 500 ng/L. Supporting Information Table S3 showed the extraction efficiency of four different SPE cartridges for the targeted OPEs. First, C8 was excluded because of its poor extraction efficiency for TMP (0%) and TEP (45.9%). Then, HLB and PEP cartridges were excluded since they showed  C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

lower extraction efficiency for CDPP, EHDPP, and TCrP than ENVI-18 cartridge. Because of the wide polarity range of OPEs, the hydrophilic compounds, such as TMP and TEP, maybe poorly retained on the SPE cartridges with RP materials, while the hydrophobic compounds, TEHP, EHDPP, TCrP, and CDPP may be difficult to elute from cartridges after extraction. In sum, better extraction efficiency for the 14 OPEs was obtained by ENVI-18, and it was finally selected in our study. 3.1.2 Elution and sample loading volume The ENVI-18 cartridge was eluted with 25% DCM in acetonitrile, and the optimal elution volume was chosen by spiking 100 mL of pure water at a concentration of 500 ng/L. Supporting Information Fig. S1 demonstrated the recovery (without IS correction) of OPEs under different elution volumes. It can be observed that 2 mL of 25% DCM in acetonitrile was sufficient for the elution of TCEP, TCPP and TDCP. However, more volume of eluent was needed for other OPEs such as TEP, TBEP, and TEHP, and 6 mL of 25% DCM in acetonitrile was the most appropriate volume for all OPEs as the recovery reached up to the maximum level. Moreover, loading volume of water samples was also evaluated. Totally 50, 100, 200, and 500 mL of river water spiked with 200 ng/L OPEs were analyzed. Recovery (without IS correction) of OPEs under different loading volumes were shown in Supporting Information Table S4. Satisfactory recovery (without IS correction; >80%) of TPrP, TBEP, TnBP, and TiBP under different loading volumes were achieved. Higher sensitivity could be obtained as the sample volume for extraction increases. However, in terms of TMP, TEP, TCEP, EHDPP, and TEHP, the recovery revealed a decline as the increase of water volume. When changing the loading volume from 50 mL to 500 mL, recovery of TMP and TEHP www.jss-journal.com

2198

L. Gao et al.

dropped to 4.5% from 39.5% and to 27.1% from 55.6%, respectively. Apparently, small sample volume was benefit for the recovery of most OPEs during the SPE procedure, but it also led to lower sensitivity of the established method. In the present study, the sample volume of 100 mL was chosen finally.

3.2 Optimization of LC–MS/MS conditions 3.2.1 Selection of chromatographic columns The column that can increase the retention of hydrophilic OPEs and weaken the retention of hydrophobic OPEs at the same time will be selected in our study. Under isocratic conditions of 50:50 and 30:70 of pure water (A) and acetonitrile (B), the retention factor (k’ ) was calculated to evaluate the retention capacity of three different columns for OPEs. Besides, the same flow rate of 0.25 mL/min and injection volume of 15 ␮L were used for three columns. Retention factor was calculated as follows: k’ = (tR – t0 )/t0 , where tR and t0 are the retention time of the analyte and the dead time of the column, respectively. Additionally, dead time of the columns was determined by the injection of NO3 – standard in the same conditions of OPEs analysis. The retention time and retention factors of the analytes on three different columns (mixed-mode HILIC-1, Xterra MS C18 , and Accucore RP-MS) under isocratic conditions were listed in Table 1. The retention factors of the analyte indicated that all three columns provided good retention for most of the analytes. For the hydrophilic OPE such as TMP and TEP, however, the RP column (XTerraMS C18 and RP-MS) showed poor retention capacity because the retention factors were in the range of 0.33–1.00. Compared with RP column, the retention factors of TMP and TEP on mixed-mode HILIC-1 column were increased to 1.05–1.45, demonstrating that the retention of hydrophilic OPEs could be enhanced on the mixed-mode HILIC-1 column because of glycol terminal group in the stationary phase of the column. On the other hand, the retention of hydrophobic OPEs (TEHP and EHDPP) was very strong on the RP column with retention factor >76.92. However, the retention of TEHP and EHDPP was weakened on the mixed-mode HILIC-1 column because the retention factor was decreased to 2.11–42.96, which will shorten the analysis time. The stationary phase of the mixed-mode HILIC-1 column consists of a hydrophobic alkyl chain with a glycol terminus. Compared with conventional RP columns, the alkyl chain in the phase of mixed-mode HILIC-1 column provides substantial hydrophobic retention, and also the glycol terminal group provides hydrophilic retention. Therefore, the mixed-mode HILIC-1 column provides good separation performance in a shorter runtime than the conventional RP columns due to the optimal balance between hydrophilic and hydrophobic moieties on the silica surface. In conclusion, the mixed-mode HILIC-1 column was finally selected in our study because

 C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

J. Sep. Sci. 2015, 38, 2193–2200

it realized the purpose to balance the analysis of hydrophilic and hydrophobic OPEs, which will in turn reduce the analysis time of HPLC–MS/MS.

3.2.2 Effect of mobile phase composition In our study, gradient elution was used and the final gradient conditions, as well as flow rates and injection volume for mixed-mode HILIC-1 column are described in Section 2.4. Under the above conditions, the compositions of the mobile phase were optimized. Acetonitrile was used as the organic phase (B) because it displayed lower column pressure than methanol. Generally, appropriate buffer salt helps to improve the peak shape and sensitivity. Three aqueous phases (A) with different acidity, pure water (pH 6.0), 0.1% formic acid (pH 2.7), and 100 mmol ammonium acetate (pH 6.81) aqueous solutions, were investigated to achieve better analytical performance. The comparison of signal intensity of the analytes under different mobile phase compositions was shown in Supporting Information Fig. S2. For mixed-mode HILIC-1 column, pure water (A) and acetonitrile (B) were finally selected as a binary mobile phase because they provided maximum signal intensity for most analytes (Supporting Information Fig. S2), and also provided the simple preparation of mobile phase. Under optimal mixed-mode LC–MS/MS conditions, the chromatogram of a 50 ␮g/L standard solution with mixedmode HILIC-1 column and binary mobile phase of pure water (A) and acetonitrile (B) is shown in Fig. 1.

3.3 Analytical performance 3.3.1 Linearity and LODs and LOQs Under the optimal SPE and mixed-mode LC–MS/MS conditions, the calibration curves were measured for all 14 OPEs with constant amount of IS (5 ng), as indicated in Table 2. A good linearity (R = 0.9877–0.9969) was obtained for analytes in a wide concentration range, and the RSD was

Determination of organophosphate esters in water samples by mixed-mode liquid chromatography and tandem mass spectrometry.

A simple and sensitive method for the determination of organophosphate esters in water samples by mixed-mode liquid chromatography with electrospray i...
2MB Sizes 0 Downloads 8 Views

Recommend Documents