Online Coupling of Capillary Electrophoresis with Direct Analysis in

Nov 29, 2012 - ... Alexander C. Johnson , Megan M. Weisenberger , and Michael T. .... Jennifer J. Pittman , María Eugenia Monge , Guilong Cheng , She...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/ac

Online Coupling of Capillary Electrophoresis with Direct Analysis in Real Time Mass Spectrometry Cuilan Chang, Gege Xu, Yu Bai,* Chengsen Zhang, Xianjiang Li, Min Li, Yi Liu, and Huwei Liu* Beijing National Laboratory for Molecular Sciences, Key Laboratory of Bioorganic Chemistry and Molecular Engineering of Ministry of Education, Institute of Analytical Chemistry, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China ABSTRACT: The online coupling of capillary electrophoresis with ambient direct analysis in real time mass spectrometry (DART-MS) was realized by a coaxial tip interface. The analytes eluted from capillary electrophoresis (CE) were directly ionized by the metastable helium flux produced by DART and transferred into MS for the detection, with which the online separation and simultaneous detection were achieved. The CE-DART-MS can tolerate higher concentrations of detergents and salts than traditional CE-electrospray ionization (ESI)-MS and avoided the difficulties of collecting CE effluent and cleaning the interface, which simplified the experimental procedures and shortened the analysis time. The performance of the technique was successfully verified by capillary zone electrophoresis (CZE) and micellar electrokinetic chromatography (MEKC) using a mixture of 4-aminoantipyrine, zolmitriptan, and quinine. This online technique showed good repeatability with the relative standard deviations (RSDs; n = 5) of 0.56−1.23% for the retention times and 2.01−7.41% for the peak areas. The quantitative analysis of 4-aminoantipyrine was accomplished in the range of 0.01−0.50 mg/mL with the linear correlation coefficient of 0.9995 and limit of detection of 14.7 fmol. Compared with CE-ESI-MS, the ion suppression effects of nonvolatile salts and detergents were efficiently minimized. The signal intensity remained constant when the concentrations reached 100 mM for sodium borate and 30 mM for SDS (in 30 mM sodium borate buffer). In addition, the proposed method was successfully applied to the detection of the endogenous caffeine in Chinese white tea.

C

apillary electrophoresis (CE)1 has been accepted as a powerful separation approach due to its high analysis throughput, high separation efficiency, and minimal sample consumption. It offers higher resolution than conventional pressure-driven flow systems because of the flat flow pattern resulting from electroosmotic driving. In addition, various separation modes could be conveniently tuned by altering the composition of the running buffer with detergents and/or organic solvents. Among the CE modes, capillary zone electrophoresis (CZE)1−3 and micellar electrokinetic chromatography (MEKC)4,5 are two widely used modes. In most cases, the effluents of CE could be detected by various optical detectors, such as ultraviolet−visible (UV−vis) spectroscopy1,6 and laser induced fluorescence (LIF).7,8 UV−vis is the most commonly used detector, but its sensitivity is relatively low, whereas LIF could provide higher sensitivity but it is only adequate for species with fluorescent absorption or amenable to derivatization with fluorescing or absorbing chromophores. Thus, an ideal detector for CE with the features of universal detection, high sensitivity, and without degradation of separation efficiency is highly desired. In the past decade, mass spectrometry (MS) was increasingly used as an attractive detector for CE because of its generality, sensitivity, and molecular structure information. Great success has been achieved in online coupling of CE to MS (CE-MS) using electrospray ionization (ESI).9−12 However, conventional CE-ESI-MS was limited by the incompatible buffer, such as © 2012 American Chemical Society

nonvolatile salts/detergents, which are usually used in the background electrolyte (BGE) of CZE and MEKC for optimal separations but bring serious ion suppression and contamination of the mass spectrometer inlet.13,14 One possible solution to this problem is changing the conditions of CE separations, such as using volatile salts and detergents or applying partial-filling technique.15,16 However, these approaches may complicate the separations and decrease the resolution, repeatability, and universality. Another solution is utilizing alternative ionization techniques which could tolerate nonvolatile salts and detergents, such as matrix-assisted laser desorption/ionization (MALDI),17 atmospheric pressure chemical ionization (APCI),18 and atmospheric pressure photoionization (APPI).19−21 Although great success has been achieved for the coupling of CE-MS when the above ionization techniques were used, the design and construction of the interface is relatively complex as the ionization must be performed in a closed environment. Besides, source contamination came from the nonvolatile salts/detergents is another problem that should not be ignored. Ambient ionization MS,22,23 which is operated in open air conditions and just requires minimal or no sample preparation, is one of the most important breakthroughs in modern MS Received: August 19, 2012 Accepted: November 29, 2012 Published: November 29, 2012 170

dx.doi.org/10.1021/ac303450v | Anal. Chem. 2013, 85, 170−176

Analytical Chemistry

Article

was provided by Hangzhou Wahaha Group (Hangzhou, Zhejiang, China). Capillary Electrophoresis. All CE separations were performed on an Agilent 7100 CE system with air-cooling and a diode array detector (Agilent Technologies, Palp Alto, CA, USA). An 80 cm fused silica capillary with 75 μm i.d. and 360 μm o.d. (Sino Sumtech, Handan City, Hebei, China) was utilized for CE separations. The new capillary was washed with 1.0 M sodium hydroxide (15 min), water (15 min), and the running electrolyte (10 min) in turn. Before each analysis, the capillary was flushed with 0.1 M sodium hydroxide (1 min), water (2 min), and the running electrolyte (3 min) sequentially to guarantee good reproducibility. About 5 mm of the exterior polyimide protection coating was removed before the capillary was reinserted into the buffer vial. The sample solution containing 0.2 mg/mL of each model compound (4-aminoantipyrine, zolmitriptan, and quinine) was injected at 50 mbar for 10 s. The maximum injection volume was calculated to be 100 nL according to the column capacity, which yielded approximately 60−100 pmol of each analyte as the molecular weights of the analytes are 203.1059, 287.1624, and 324.1838, respectively. UV detection wavelength was set at 230 nm with the capillary temperature of 25 °C and the separation voltage of 25 kV. Fifteen mM sodium borate was used for CZE, and for MEKC, the separation buffer consisted of 15 mM sodium borate, 15 mM SDS, and 18% acetonitrile. CE-DART-MS Interface. For generation of the new CEDART-MS interface, it is important to introduce additional liquid at the outlet of the CE capillary to stabilize the electric current. The commercial Agilent sheath liquid tip (Palo Alto, CA, USA)37−40 matched our thought with good quality. The tip was grounded without high voltage; therefore, no electrospray will be created. The liquid was provided by an additional LC pump, and the nebulizing gas was turned on to carry the analytes and stabilize the mass spectrometric signal. The sprayer tip was attached on the mass spectrometer through a homemade holder, and the position could be adjusted horizontally and vertically. Similarly to the reports34,35 and our previous work,36 the signal was the best when the end of the tip was placed at the same height as the DART outlet and MS inlet. The distance between the DART outlet and the end of tip was adjusted to 2 mm, and that between the MS inlet and the end of tip was 3 mm (Scheme 1) to gain high ionization

development. Among the various ambient MS ionization sources, desorption electrospray ionization (DESI)24 and direct analysis in real time (DART)25 have been widely used. Many reports demonstrated that DESI can be used to directly analyze salt-containing sample and did not require “make-up” solvents/ acids to be doped in a sample prior to ionization.26−28 Recently, off-line coupling of CE to DESI has been constructed by depositing CE effluents on a rotating Teflon disk covered with paper for further DESI-MS analysis,29 demonstrating that the negative effects of detergents and salts on the MS analysis could be minimized. However, one problem associated with the above off-line method was the difficulty of collecting the small volume CE fractions, which not only prolonged the analysis time but also reduced the repeatability. Consequently, it is attractive to develop an online coupling technique for CE and ambient MS. DART, with helium or nitrogen as the working gas, can produce heated metastable plasma in the glow discharge chamber and transfer the protons to analytes by the reactive species in the ambient atmosphere. It has been widely used in many fields, such as direct analysis of pheromones from live drosophila,30 fast screening of counterfeit drugs,31 reaction monitoring in drug discoveries,32 quantitative detection of warfare agents,33 and coupling to liquid chromatography (LC).34,35 Previous reports have shown that DART has high salt tolerance. For example, DART-MS was able to tolerate HPLC eluents containing 120 mM phosphate buffer at the flow rate of 1.0 mL/min without contamination of the mass spectrometer or ion suppression,34 and an online coupling of normal phase HPLC with DART-MS for chiral separation and detection were reported.36 The excellent salt tolerance makes DART an attractive ionization technique for feasible combination of CE with MS. This work presents an interface that enables the online coupling of CE to DART-MS, showing outstanding tolerability of nonvolatile salts and detergents used in CE without source contamination or ion suppression. The key of the online interface lies in the coaxial tip, which mixes the CE effluent with the sheath liquid to stabilize the electric current. Carried by the nebulizing gas, the analytes are subsequently ionized by DART and delivered into MS for analysis. The convenience and availability of online CE-DART-MS have been fully testified by the experimental results of CZE-DART-MS and MEKCDART-MS using three model compounds (4-aminoantipyrine, zolmitriptan, and quinine). Compared with CE-ESI-MS, no negative effects of nonvolatile salts and detergents were observed, demonstrating the feasibility and efficiency of CEDART-MS. At the same time, this online coupling technique avoided the difficulties of collecting CE effluent and cleaning the interface, simplified the procedures, and shortened the analysis time. In addition, the proposed method was successfully applied to determine the endogenous caffeine in Chinese white tea.

Scheme 1. Interface for Online Coupling of CE to Ambient MS Using DART



EXPERIMENTAL SECTION Chemicals. 4-Aminoantipyrine and quinine were purchased from Alfa Aesar (Ward Hill, MA, USA), and zolmitriptan was kindly donated by Orient Lüyuan Co., Ltd. (Beijing, China). Sodium dodecyl sulfate (SDS) was purchased from Fluka (Buchs, Saint Gall, Switzerland). Sodium borate, disodium hydrogen phosphate, and phosphoric acid were obtained from Beijing Chemical Company (Beijing, China). HPLC grade methanol, acetone, and acetonitrile were purchased from Dikma Technology (Richmond, VA, USA). Purified water 171

dx.doi.org/10.1021/ac303450v | Anal. Chem. 2013, 85, 170−176

Analytical Chemistry

Article

by directly infusing samples dissolved in pure water and respective buffers. Figure 1 showed the results obtained for a

efficiency and suitable operation space. The exit tip of the capillary from 7100 CE instrument was orthogonally inserted into the sprayer interface. To maintain stable current, 5 mm polyimide coating was removed from the capillary end and 0.2 mm of the capillary was pulled out of the stainless-steel needle. The analytes separated by CE were ionized by DART and detected by the downstream MS analyzer. The effect of the coaxial sheath liquid composition on the CE-DART-MS signal was investigated using binary and ternary mixture of water, methanol, acetone, and isopropanol. Interestingly, the experimental results showed that addition of acetone into the mixture of methanol/water can improve the signal. When adding 0.1−5% formic acid to the sheath liquid, no enhancement of the signal was observed for the three model compounds. Therefore, the sheath liquid consisting of methanol/acetone/water (25:25:50 v/v/v) was delivered at 1.01 mL/min by an isocratic pump (Agilent, Palo Alto, CA, USA) through a splitter set at 1:100, thus keeping a flow rate of 10 μL/min into the tip. The nebulizing gas was also optimized. When it was lower than 1 psig, the nebulization was imperfect with residual liquid; when the nebulizing gas was higher than 3 psig, the target molecules were blowing off and the signal was notably reduced. As a result, the coaxial nebulizing gas was set at 2 psig. The amplified inset in Scheme 1 shows the fine structure of the interface. Direct Analysis in Real Time Mass Spectrometry. The DART-SVP ion source (IonSense, Saugus, MA, USA) was coupled to an Agilent 6530 accurate-mass quadrupole time-offlight mass spectrometer (QTOF-MS) (Agilent, Palo Alto, CA, USA) after removing the original Agilent Jet Stream electrospray ionization source. The DART ion source was operated with helium for ionization. Temperature and flow rate of helium were the two important factors for DART ionization process. The temperature was systematically optimized between 250 and 450 °C, and the helium flow rate was between 0.08 and 0.14 m3/h. After optimization, the temperature and flow rate were set at 350 °C and 0.12 m3/h. The QTOF-MS was tuned and calibrated with Tuning Mix (Agilent P/N G1969-85000) before being used. The MS parameters were as follows: capillary voltage at −3500 V, fragmentor at 175 V, skimmer at 65 V, and drying gas temperature at 300 °C. The mass spectra were acquired in the range of m/z 50−800 with the scan rate of 1.0 spectrum/s. In the targeted MS/MS mode, the MS/MS information was collected with the collision energy of 10, 20, 25, and 25 V for 4-aminoantipyrine, zolmitriptan, quinine, and caffeine, respectively. Iso. width of precursor ion was set as narrow (∼1.3 m/z). All of the MS and MS/MS data were collected by MassHunter Data Acquisition B.02.00 (Agilent, CA, USA) and analyzed with MassHunter Qualitative Analysis B.02.00 (Agilent Technologies, CA, USA). Preparation of Chinese White Tea Sample. In total, 2 g of Chinese white tea leaves was ultrasonically extracted for 20 min with 10 mL of methanol/H2O (95/5, v/v). Then, the extract was filtered and injected for online CE-DART-MS analysis.

Figure 1. (a) Relative MS signals of 4-aminoantipyrine (100 μg/mL) in various buffers: (A) water; (B) 15 mM sodium borate; (C) 30 mM sodium borate; (D) 50 mM sodium borate; (E) 100 mM sodium borate; (F) 15 mM SDS in 15 mM sodium borate; (G) 30 mM SDS in 30 mM sodium borate; (b) mass spectrum of analytes dissolved in pure water; (c) mass spectrum of analytes dissolved in 15 mM sodium borate buffer containing 15 mM SDS.

100 μg/mL 4-aminoantipyrine solution. As expected, nonvolatile sodium borate and SDS did not show ion suppression on the signal in comparison to the sample in water. Remarkably, the signal intensity remained constant even when the concentrations reached 100 mM for sodium borate and 30 mM for SDS (in 30 mM sodium borate). Moreover, about 75% of the signal intensity was remaining when the concentration of SDS increased to 50 mM (in 50 mM sodium borate). Two typical mass spectra, one for the sample dissolved in pure water and another for the sample dissolved in 15 mM sodium borate buffer containing 15 mM SDS, were presented in Figure 1b,c. For each sample, three peaks corresponding to the three compounds without adducts or contaminants were found. Therefore, it was concluded that nonvolatile salts or detergents in this system had no obvious adverse impacts on the signal intensity of MS when DART was applied as the ionization source. CZE and MEKC, two frequently used CE separation modes, were investigated in order to find out whether DART-MS could be regarded as a suitable online detector for CE separations using nonvolatile salts and detergents in the BGE. When the



RESULTS AND DISCUSSION This online coupling CE-DART-MS shows outstanding tolerability of nonvolatile salts and detergents without source contamination or ion suppression. In order to investigate its tolerability of nonvolatile salts and detergents, a contrast experiment has been conducted. The influence of nonvolatile sodium borate and SDS at different concentrations was studied 172

dx.doi.org/10.1021/ac303450v | Anal. Chem. 2013, 85, 170−176

Analytical Chemistry

Article

Figure 2. CZE-DART-MS analysis of a mixture of three analytes using sodium borate buffer. See Experimental Section for other conditions. (a) Electropherogram measured by UV at 230 nm; (b) EIC for ion at m/z 288.1707; (c) EIC for ion at m/z 325.1911; (d) EIC for ion at m/z 204.1131; (e) mass spectrum at 4.701 min; (f) mass spectrum at 5.460 min; (g) mass spectrum at 5.575 min; (h) targeted MS/MS spectrum of the precursor ion at m/z 288.1707; (i) targeted MS/MS spectrum of the precursor ion at m/z 325.1911; (j) targeted MS/MS spectrum of the precursor ion at m/z 204.1131.

spectra across each peak in the EICs were shown in Figure 2e,f,g. Figure 2e presented a single peak at m/z 288.1707 that was identified as zolmitriptan. Due to the coelution, both of the latter two spectra consisted of ions at m/z 204.1131 and 325.1911, but they can be easily distinguished by their migration times and the MS/MS spectra (Figure 2h,i,j). Moreover, the repeatability of CZE-DART-MS was verified using the mixed sample at the concentration of 0.2 mg/mL. For zolmitripta, quinine, and 4-aminoantipyrine, the relative standard deviations (RSDs) of migration times were 0.81%,

mixed sample of 4-aminoantipyrine, zolmitriptan, and quinine was separated in CZE mode using 15 mM sodium borate buffer, three peaks were observed in the electropherogram measured by UV at 230 nm (Figure 2a). The first peak at 4.814 min was zolmitriptan, and the latter two peaks with partial overlap were corresponding to quinine (5.664 min) and 4aminoantipyrine (5.843 min). Extracted ion chromatograms (EICs) at m/z 288.1707, 325.1911, and 204.1131 were presented in Figure 2b,c,d, corresponding to zolmitriptan, quinine, and 4-aminoantipyrine, respectively. Average mass 173

dx.doi.org/10.1021/ac303450v | Anal. Chem. 2013, 85, 170−176

Analytical Chemistry

Article

Figure 3. MEKC-DART-MS analysis of a mixture of three analytes. See Experimental Section for other conditions. (a) Chromatogram measured by UV at 230 nm; (b) EIC for ion at m/z 204.1131; (c) EIC for ion at m/z288.1707; (d)EIC for ion at m/z 325.1911; (e) mass spectrum at 7.898 min; (f) mass spectrum at 9.118 min; (g) mass spectrum at 10.536 min.

operation, proper inspection of the mass spectrometer and the ion source revealed no decrease in signal intensity and critical contamination. Similar to CZE-DART-MS, MEKCDART-MS also showed good repeatability. The RSDs of the retention times were 0.62%, 0.96%, and 1.23% (n = 5) for zolmitriptan, quinine, and 4-aminoantipyrine, respectively, and those of the peak areas were 2.01%, 7.41%, and 3.68% (n = 5). The LODs for the three analytes were 3, 50, and 10 ng, two of which were relatively lower than those in CZE mode. It was worth mentioning that the MEKC-DART-MS method had a linear range of 0.01−0.5 mg/mL with the linear correlation coefficient (R2) of 0.9995 for 4-aminoantipyrine, demonstrating the feasibility of CE-DART-MS for quantitative analysis. In contrast, when the same MEKC system was online coupled to ESI-MS, the serious inlet contamination and ion suppression prevented us from further experiments. It was noted that, in above analyses, there were some differences of the peak widths and migration/retention times between the EICs and the corresponding UV electropherograms. Peak broadening in the EICs could be partly attributed to the suction effect of the nebulizing gas and the lack of temperature control,41,42 and another possible reason was the different microenvironment of the capillary outlet. For instance, the UV electropherograms were obtained by traditional CE separations, and both the inlet and outlet of the capillary were immersed in the BGE, while the EICs were obtained by online coupled CE-DART-MS, in which the outlet of the capillary was kept in ambient conditions. Compared to off-line techniques, the disturbance of the interface to CE separation has already been reduced to some extent.26 Thus, this online CE-DARTMS can meet the requirements of high speed separations and simultaneous detections without source contamination or ionization suppression. Finally, the proposed method was applied to determine the endogenous caffeine in Chinese white tea to test the

0.61%, and 0.56% (n = 5), and the RSDs of peak areas were 5.21%, 4.58%, and 5.44% (n = 5), respectively. The limits of detection (LODs) for the three analytes were 50, 50, and 20 ng, respectively. These results demonstrated that DART-MS could serve as a suitable detector for CZE and prevent the ion suppression and contamination induced from the nonvolatile salts in the BGE. The comparison of DRAT and traditional ESI using the same model compounds were also carried out, demonstrating that the ion suppression in the case of ESI-MS was so serious that the signal of [M + H]+ for 4aminoantipyrine was not detected. The signal of zolmitriptan and quinine could be detected with the LOD of 10 ng, which was in the same magnitude with DART. However, the residual nonvolatile salts caused severe inlet contamination and interfered with subsequent experiments. Therefore, it is believed that DART could be regarded as a more suitable ion source for online coupling of CE-MS when nonvolatile salts were used in the running buffer. Detergents like SDS are usually used in MEKC while it is a mortal factor to suppress the signal of analytes in traditional MS. In the following work, MEKC-DART-MS was selected as the separation mode for further demonstration of its tolerability of detergents like SDS. Under the MEKC mode, three model compounds were baseline separated as shown by UV eletropherogram (Figure 3a). The elution order was 4aminoantipyrine (7.898 min), zolmitriptan (9.118 min), and quinine (10.536 min), which was different from that of CZE mode. The EICs detected by DART-MS for ions at m/z 204.1131, 288.1707, and 325.1911 were presented in Figure 3b,c,d, corresponding to 4-aminoantipyrine, zolmitriptan, and quinine, respectively. Average mass spectra across each peak in the EICs were shown in Figure 3e,f,g, all of which showed a single peak and avoided the coelution. In addition, the mass spectra did not contain any adduct or contaminant peaks from the nonvolatile salts or detergents. After a two-month 174

dx.doi.org/10.1021/ac303450v | Anal. Chem. 2013, 85, 170−176

Analytical Chemistry

Article

up the development of CE-MS and broaden its application in various fields. It should be noted that the detection sensitivity of CE-DART-MS at present is not very satisfactory as we expected, and the possibility for improving sensitivity may include the use of laser desorption43 or/and surface-assistant techniques44 in DART-MS, that is in progress in our lab.

applicability of online coupling of CE-DART-MS for the detection of target molecule in real sample. Separation of the sample was fulfilled on CE using 15 mM sodium borate as the running buffer. The EIC and mass spectrum of m/z 195.0880 were shown in Figure 4a,b, respectively, which was correspond-



AUTHOR INFORMATION

Corresponding Author

*Phone: +86-10-62754976 (H.L.). E-mail: [email protected] (H.L.); [email protected] (Y.B.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was financially supported by the National Natural Science Foundation of China (Grant Nos. 21027012, 21275012, and 21175005); Agilent Technologies was thanked for providing the 7100 CE system.



REFERENCES

(1) Jorgenson, J. W.; Lukacs, K. D. Science 1983, 222, 266−272. (2) Lauer, H. H.; McManigill, D. Anal. Chem. 1986, 58, 166−170. (3) Jorgenson, J. W.; Lukacs, K. D. Anal. Chem. 1981, 53, 1298− 1302. (4) Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Anal. Chem. 1984, 56, 111−113. (5) Wiedmer, S. K.; Jumppanen, J. H.; Haario, H.; Riekkola, M. L. Electrophoresis 1996, 17, 1931−1937. (6) Swinney, K.; Bornhop, D. J. Electrophoresis 2000, 21, 1239−1250. (7) Gassmann, E.; Kuo, J. E.; Zare, R. N. Science 1985, 230, 813−814. (8) Cheng, Y. F.; Dovichi, N. J. Science 1988, 242, 562−564. (9) Olivares, J. A.; Nguyen, N. T.; Yonker, C. R.; Smith, R. D. Anal. Chem. 1987, 59, 1230−1232. (10) Smith, R. D.; Udseth, H. R. Nature 1988, 331, 639−640. (11) Smith, R. D.; Olivares, J. A.; Nguyen, N. T.; Udseth, H. R. Anal. Chem. 1988, 60, 436−441. (12) Han, Y. H.; Bai, Y.; Xiao, Y. L.; Du, F. Y.; Liang, Y.; Tan, Z. J.; Zhao, M. P.; Liu, H. W. Electrophoresis 2011, 32, 2693−2699. (13) Lu, W. Z.; Poon, G. K.; Carmichael, P. L.; Cole, R. B. Anal. Chem. 1996, 68, 668−674. (14) Varghese, J.; Cole, R. B. J. Chromatogr., A 1993, 652, 369−376. (15) Nelson, W. M.; Tang, Q.; Harrata, A. K.; Lee, C. S. J. Chromatogr., A 1996, 749, 219−226. (16) Nelson, W. M.; Lee, C. S. Anal. Chem. 1996, 68, 3265−3269. (17) Johnson, T.; Bergquist, J.; Ekman, R.; Nordhoff, E.; Schurenberg, M.; Kloppel, K. D.; Muller, M.; Lehrach, H.; Gobom, J. Anal. Chem. 2001, 73, 1670. (18) Takada, Y.; Sakairi, M.; Koizumi, H. Anal. Chem. 1995, 67, 1474. (19) Nilsson, S. L.; Andersson, C.; Sjoberg, P. J. R.; Bylund, D.; Petersson, P.; Jornten-Karlsson, M.; Markides, K. E. Rapid Commun. Mass Spectrom. 2003, 17, 2267. (20) Mol, R.; de Jong, G. J.; Somsen, G. W. Electrophoresis 2005, 26, 146. (21) Schappler, J.; Guillarme, D.; Prat, J.; Veuthey, J. L.; Rudaz, S. Electrophoresis 2007, 28, 3078. (22) Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M. Science 2006, 311, 1566−1570. (23) Harris, G. A.; Galhena, A. S.; Fernandez, F. M. Anal. Chem. 2011, 83, 4508−4538. (24) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471−473. (25) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297−2302. (26) Miao, Z. X.; Chen, H. J. Am. Soc. Mass Spectrom. 2009, 20, 10. (27) Li, J. W.; Dewald, H. D.; Chen, H. Anal. Chem. 2009, 81, 9716.

Figure 4. Online CE-DART-MS analysis of endogenous caffeine in Chinese white tea. See Experimental Section for other conditions. (a) EIC for ion at m/z 195.0877; (b) mass spectrum at 5.761 min; (c) targeted MS/MS spectrum of the precursor ion at m/z 195.0877.

ing to caffeine. The result was further confirmed by the MS/MS spectrum. For example, the peak of m/z 138.0672 shown in Figure 4c was attributed to the loss of the CH3NCO group from caffeine. In summary, online coupling of CE-DART-MS provided a rapid and successful screening approach for the detection of endogenous caffeine in Chinese white tea.



CONCLUSIONS This work presents the first online coupling of CE to ambient MS using DART as the versatile ionization technique. This technique takes the advantage of DART’s high salt/detergent tolerance and allows for the use of nonvolatile salts and detergents in CE separations without source contamination and ion suppression in MS. The interface was based on a coaxial tip, with which CE effluents were mixed with the sheath liquid to stabilize the electric current. Carried by the nebulizing gas, the analytes are subsequently ionized by DART and delivered into MS for analysis. The online coupling method avoids the difficulties of collecting the small volume effluent from CE and cleaning the interface. The convenience, efficiency, and salt/ detergent tolerance have been demonstrated by CZE-DARTMS and MEKC-DART-MS analyses of model compounds, showing good reproducibility (RSDs of 0.56−1.23% for migration times and 2.01−7.41% for peak areas) and suitability for quantitative analysis (e.g., for 4-aminoantipyrine analyzed by MEKC-DART-MS, LOD of 14.7 fmol, linear range of 0.01− 0.50 mg/mL with the correlation coefficient of 0.9995). In addition, CE-DART-MS was successfully applied to the analysis of the endogenous caffeine in the Chinese white tea. The remarkable improvement in the tolerance of nonvolatile salts and detergents, as well as the good reproducibility, can speed 175

dx.doi.org/10.1021/ac303450v | Anal. Chem. 2013, 85, 170−176

Analytical Chemistry

Article

(28) Zhang, Y.; Yuan, Z. Q.; Dewald, H. D.; Chen, H. Chem. Commun. 2011, 47, 4171. (29) Barbula, G. K.; Safi, S.; Chingin, K.; Perry, R. H.; Zare, R. N. Anal. Chem. 2011, 83, 1955−1959. (30) Yew, J. Y.; Cody, R. B.; Kravitz, E. A. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 7135−7140. (31) Zhou, Z. G.; Zhang, J. L.; Zhang, W.; Bai, Y.; Liu, H. W. Analyst 2011, 136, 2613−2618. (32) Petucci, C.; Diffendal, J.; Kaufman, D.; Mekonnen, B.; Terefenko, G.; Musselman, B. Anal. Chem. 2007, 79, 5064−5070. (33) Nilles, J. M.; Connell, T. R.; Durst, H. D. Anal. Chem. 2009, 81, 6744−6749. (34) Eberherr, W.; Buchberger, W.; Hertsens, R.; Klampfl, C. W. Anal. Chem. 2010, 82, 5792−5796. (35) Beissmann, S.; Buchberger, W.; Hertsens, R.; Klampfl, C. W. J. Chromatogr., A 2011, 1218, 5180−5186. (36) Chang, C. L.; Zhou, Z. G.; Yang, Y. Y.; Han, Y. H.; Bai, Y.; Zhao, M. P.; Liu, H. W. Electrophoresis 2012, 33, 3387−3393. (37) Groom, C. A.; Hawari, J. Electrophoresis 2007, 28, 353−359. (38) Zhang, Z. X.; Yan, B.; Liu, K. L.; Liao, Y. P.; Liu, H. W. Electrophoresis 2009, 30, 379−387. (39) Fu, X. F.; Liu, Y.; Li, W.; Pang, N. N.; Nie, H. G.; Liu, H. W.; Cai, Z. W. Electrophoresis 2009, 30, 1783−1789. (40) Gao, F.; Zhang, Z. X.; Fu, X. F.; Li, W.; Wang, T.; Liu, H. W. Electrophoresis 2007, 28, 1418−1425. (41) Axen, J.; Axelsson, B. O.; Jornten-Karlsson, M.; Petersson, P.; Sjoberg, P. J. R. Electrophoresis 2007, 28, 3207−3213. (42) Zhong, X. F.; Maxwell, E. J.; Chen, D. D. Y. Anal. Chem. 2011, 83, 4916−4923. (43) Zhang, J. L.; Zhou, Z. G.; Yang, J. W.; Zhang, W.; Bai, Y.; Liu, H. W. Anal. Chem. 2012, 84, 1496−1503. (44) Zhang, J. L.; Li, Z.; Zhang, C. S.; Feng, B. S.; Zhou, Z. G.; Bai, Y.; Liu, H. W. Anal. Chem. 2012, 84, 3296−3301.

176

dx.doi.org/10.1021/ac303450v | Anal. Chem. 2013, 85, 170−176