Anal. Chem. 2001, 73, 4748-4753
Pulsed Electrospray for Mass Spectrometry Yu Lu,†,‡ Feng Zhou,† Wenqing Shui,† Liping Bian,† Yinlong Guo,§ and Pengyuan Yang*,†
Department of Chemistry, Fudan University, Shanghai 200433, China, and Shanghai Institute of Organic Chemistry, Chinese Academy of Science, Shanghai 200032, China
Pulsed electrospray has been developed and been reported for the first time. This new technique is based on the principle of pulsed ion sources, combined with the conventional electrospray. The pulsed ion spray was realized by a homemade pulsed HV circuit and was monitored by a digital microscope and an oscilloscope. Results show that the pulsed ESI device can be operated under proper conditions for a clear on-and-off spray process and that the device was kept in good electric contact for electrospray when pulsed HV was on. A pulsed ion current and pulsed mass spectra can be achieved with this pulsed ESI device. Furthermore, it has been noted that, under the same conditions (i.e., shape and size of sprayer tip, distance from sprayer tip to sampling nozzle, and other parameters for mass spectrometer), stable electrospray could be obtained for lower flow rates with a pulsed spray device. This experimental fact indicates the possible reduction in the total sample consumed could be realized by exploiting this novel design. Since electrospray ionization (ESI) was successfully applied to mass spectrometry (MS), it has exerted a revolutionary impact on the analysis of biomolecules.1 The development of micro- and nanoelectrospray techniques makes it possible to couple microseparation techniques with ESI-MS.2,3 The combinations of both micro-high-performance liquid chromatography (micro-HPLC)2 and capillary electrophoresis (CE)3 with ESI-MS have attracted much attention in protein/peptide analysis recently. To apply these hyphenated techniques, it seems more preferable to use sprayers of smaller diameter (or more technically, nanosprayers), which could improve sensitivity and minimize sample consumption.4,5 Nevertheless, the effectiveness of hyphenated techniques has been restricted by the noncontinuous mode of modern mass spectrometers. Normally, the quadrupole MS, ion trap MS, and time-of-flight (TOF) MS are utilized in analysis, and they are all working at a noncontinuous or pulsed mode. Thus, only a portion of electrosprayed ions is sent to the detector while the rest of * Corresponding author: (phone) +86 21 65642009; (fax) +86 21 65641740; (e-mail)
[email protected]. † Fudan University. ‡ Current address: Department of Chemistry, University of Washington, Seattle, WA 98195. § Shanghai Institute of Organic Chemistry. (1) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Mass Spectrom. Rev. 1990, 9, 37-70. (2) Cappiello, A.; Famiglini, B.; Berloni, A. J. Chromatogr., A 1997, 768, 215222. (3) Barnidge, D. R.; Nilsson, S.; Markides, K. E. Anal. Chem. 1999, 71, 41154118.
4748 Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
them are wasted. To exploit all ions produced in electrospray can be extremely critical in the case of micro-HPLC or a CE-MS platform. This is because such techniques have rather low flow rates of microliters to submicroliters per minute.2-5 Even if all the analytes in the effluent were ionized, the sensitivity of ESIMS could still be a problem mainly due to the wasting of ions as mentioned above, such as in capillary isoelectric focusing (CIEF)MS, which has been studied as a promising new technique for proteomics recently.6 Therefore, to generate a pulsed ion beam with little wasting of ions has attracted a number of research efforts.7-12 Accumulation of ions during an ion beam modulation before entering a mass analyzer is naturally conceived to realize the noncontinuous or pulsed mode. The construction of an ion trap (IT)-TOF-MS has been investigated for ion storage and beam modulation.7 Prior to entering the TOF-MS, the ions are expected to be accumulated and then repelled out for TOF-MS analysis by applying the ion trap. On-beam compression has also been used to accumulate ions during the beam modulation.8 With a proper setting of ion optics in the beam trajectory, incoming ions can be compressed and accumulated and then be gated to the TOF-MS. The on-beam compression and modulation seems to be a reasonable way to produce a noncontinuous ion beam, except that it is rather complicated and therefore inconvenient to realize in the IT-TOF-MS. Alternatively, an ion beam in pulsed mode can also be yielded externally, i.e., in the ion source, though this technique has not been explored in ESI-MS yet. The pulsed ion sources have been well acknowledged for matrix-assisted laser deposition/ionization (MALDI) MS.9 Since the introduction of pulsed glow discharge mass spectrometry (GD-MS),10,11 one of our authors has purposely studied the pulsed Grimm-type GD-MS, which shows a number of merits such as high ion current and low sample consumption.12 In this study, the pulsed mode in an electrospray ion source is (4) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1-8. (5) Juraschek, R.; Du ¨ lcks, T.; Karas, M. J. Am. Soc. Mass Spectrom. 1999, 10, 300-308. (6) Gao, H.; Shen, Y.; Veenstra, T. D.; Harkewicz, R.; Anderson, G. A.; Bruce, J. E.; Pasa-Tolic, L.; Smith, R. D. J. Microcolumn Sep. 2000, 12, 383-390. (7) Chien, B. M.; Michael, S. M.; Lubman, D. M. Anal. Chem. 1993, 65, 19161924. (8) Piel, N.; Becker, H. W.; Meijer, J.; Schulte, W. H.; Rolfs, C. Nucl. Instrum. Methods A 1999, 437, 521-530. (9) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. (10) McLuckey, S. A.; Glish, G. L.; Duckworth, D. C.; Marcus R. K. Anal. Chem. 1992, 64, 1606-1609. (11) Duckworth, D. C.; Smith, D. H.; McLuckey, S. A. J. Anal. At. Spectrom. 1997, 12, 43-48. (12) Su, Y.; Yang, P.; Chen, D.; Zhang, Z.; Zhou, Z.; Wang, X.; Huang, B. J. Anal. At. Spectrom. 1997, 12, 817-822. 10.1021/ac0103118 CCC: $20.00
© 2001 American Chemical Society Published on Web 08/25/2001
Figure 1. Sprayer tip and droplet at tip under different electrospray potentials: (a) 0, (b) 1000, and (c) 2000 V. Distance between the sprayer tip and sampling orifice, 0.35 mm; capillary tip, 15-µm i.d. and 100-µm o.d.
demonstrated and the pulsed ESI phenomenon is presented and discussed. The factors and parameters in operating a pulsed ESI are also investigated and the potential application of pulsed ESI is predicted. EXPERIMENTAL SECTION Mass Spectrometer and Reagent. The Mariner ESI-TOFMS spectrometer (PE Corp.) was utilized in all experiments. The instrument calibration was made with a PEG-400 solution. The testing sample was clarithromycin (1 µg/µL) dissolved in a methanol-water (7:3) mixture with 5% acetic acid added. Electrospray Needle. A 10-cm-long fused-silica capillary was pulled slowly by hand under a hydrogen flame, to get the inner diameter and outer diameter of about 15 and 100 µm, respectively. The capillary used for the electrospray needle was one of 375-µm o.d. with 25-µm i.d. (Yongnian Optical Fiber Co., Hebei, China). The orifice of the sharpened capillary was then ground with fine sandpaper. The sprayer tip was thereafter coated with gold in an ion-sputtering coating device (Eiko IB-3 model, Hitachi Inc.). A proper sprayer tip was selected according to its shape as discerned under an Intel Play 200X USB portable digital microscope (Intel Inc.) and by surface conductance with a digital meter. The final sprayer tip with a surface resistance of ∼800 Ω is shown in Figure 1. Apparatus Setup. Figure 2 shows the overall schematic of the experimental setup. A 50-µL syringe on the micropump (Beijing Institute of Chemistry) was connected to the capillary sprayer tip through a stainless steel union. The pump and one stainless steel plate (acting as the opposite electrode) were
Figure 2. Experimental setup. The syringe was connected to the capillary sprayer tip; the syringe pump and the sampling plate were mounted on two mini-type 3D translational stages separately; and a microscope (200× zoom) was settled above the sprayer tip.
mounted on two mini-type 3D translational stages (Fulian Inc., Shanghai, China) separately, with the microscope settled above the sprayer tip. The 3D translational stages made it easy and accurate to adjust the relative positions among the sprayer tip, the opposite electrode, and the microscope. The USB microscope used here could transfer microscopic images and video (∼12 frames/s) into a PC computer. The zooming factor set in the observation was 200. The homemade circuit applied in the experiments can generate a dc and a pulsed high voltage up to 2 kV. The pulse frequency and width can be adjusted within 10 Hz and 100 ms. With the aid of a HV probe (Shanghai No. 4 Ammeter Factory), the pulsed HV signal was monitored with an oscilloscope (TDS 220, Tektronix Inc.). The spectra obtained by the oscilloscope were sent to an IBM-compatible computer. Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
4749
Figure 3. Series of frames with second labeled in one cycle of pulsed ESI, taken by the digital microscope. Spray conditions: potential, from 1000 to 2000 V; pulse frequency, 0.8 Hz; pulse width, 50 ms; distance between sprayer tip and sampling plate, 0.35 mm; flow rate, 0.16 µL/min.
RESULTS AND DISCUSSION Factors for Stable Electrospray. For the sake of getting a stable pulsed electrospray, experiments were designed to find the proper factors and parameters for the stability of the electrospray. Those parameters under inspection were flow rate, applied voltage, and distance between sprayer tip and opposite electrode. As is well known, a continuous electrospray will appear when there is a steady Taylor cone at the sprayer tip13,14 so that the impact of those factors could be partly estimated by observing the droplet shape at the sprayer tip. When the needle potential was set to 0, 1000, and 2000 V, different droplet shapes at the tip could be seen in Figure 1A-C, respectively. The flow rate was fixed to 0.08 µL/min, and the distance between the two electrodes was regulated to ∼0.35 mm. As indicated in Figure 1, the droplet shape not only implies the existence of electrospray but also means the possible realization of pulsed electrospray, which will be discussed later. After measuring the induced ESI current, which was collected on one sampling resistor connected into the ground end (0 V) in the HV circuit, factors for stable continuous electrospray have been optimized to be the following: potential, ∼2000 V; distance from sprayer tip to sampling nozzle, between 0.3 and 1 mm; flow rate, less than or equal to ∼0.2 µL/min. It is worth noting that such conditions may be optimal only for the particular sprayer tip mentioned in the Experimental Section. Formation of Stable Pulsed Electrospray. As described previously, only the potential around 2000 V can result in an observable electrospray. In an earlier experiment, a device was designed to create a pulse on (2000 V) and off (0 V) to control the electrospray. The flow rate was set to 80 nL/min, and the (13) Taylor, G. I. Proc. R. Soc. London A 1964, 280, 383. (14) Kebarle, P.; Peschke, M. Anal. Chim. Acta 2000, 406, 11-35.
4750
Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
distance between the two electrodes was adjusted to be 0.35 mm. After the pulse was turned on, it was observed that there was an aerosol formed at the ESI tip toward the opposite electrode, and a Taylor cone could be seen at the tip. However, a droplet was growing up and fell downward when the pulse was off. This badly shaped droplet resulted in a relatively longer time to establish a normal Taylor cone again during the next pulse and thus can cause an unstable pulsed electrospray. Therefore, the use of a 0-2000-V pulse seems to be troublesome for a stable pulsed spray process. To obtain a stable pulsed electrospray, a dc 800-1000-V bias was introduced into the pulse waveform. As is well understood, there are two competitive forces existing in the process of electrospray: the surface tension and the Coulomb force due to the applied electrical field.14 The surface tension sustains a droplet on the tip, while the Coulomb force attracts the droplet toward the opposite electrode and produces an explosion inside the droplet.14 The ESI phenomenon is the result of such a balance between these two forces. In the case of the earlier experiment for pulsed ESI, 0 V was applied when pulse was off such that there was no counter force to balance the surface tension. As a result, the droplet would grow up and fall downward mainly due to gravity. For the introduced dc 800-1000-V bias, it was not sufficient enough to produce an electrospray during the pulse-off period. However, it did provide a proper Coulomb force to prevent the droplet from falling downward and to attract the elliptical droplet ready for a next pulsed electrospray process. As a consequence, the formation time for a normal Taylor cone during the next pulse was shortened, and the pulsed electrospray became much more stable eventually.
Figure 4. Current profiles induced by pulsed ESI: (a) pulsed HV added on sprayer tip; (b) induced current on grounded resistor R* when syringe pump was on; (c) induced current on grounded resistor R* when syringe pump was off. Pulsed spray condition was the same to that in Figure 3.
Figure 3 shows a series of photos about the pulsed electrospray in one cycle when voltage was switched on (2000 V) and off (1000 V). In this experiment, the other conditions were the same as for the earlier pulsed ESI. The pulse parameters were frequency, 0.8 Hz; and pulsed HV width, 50 ms, respectively. Although the actual exciting animation could not be exhibited here, it is still easy to observe from these pictures that the Taylor cone disappeared very quickly and changed to a hemisphere-shaped drop when the pulse was off (see photos at 1.25 and at 0 s). There was a resting time of ∼1.20 s between the pulse-off and pulse-on for the formation of a Taylor cone. During this period, it could be seen that the droplet was growing gradually. The Taylor Cone came up suddenly at time of ∼1.20 s after pulse-on and lasted for 0.05 s until pulse was off. This phenomenon appeared periodically, which suggested the existence of steady pulsed electrospray. A more convincing fact is the electrospray current recorded in Figure 4. Panels a-c of Figure 4 illustrate the pulsed high voltage, the electrospray ion current induced by the pulsed high voltage, and the background noise, respectively. It can be seen from Figure 4 that the pulse current with acceptable signal-to-
noise ratio was indeed induced by the pulsed electrospray when the pulse was on, although the sampled signal was relatively noisy due mainly to the weak signal and high amplification of the detection system. The electrospray current was measured to be ∼200 nA. It should be pointed out that the pulse duration has not been considerably optimized because the present study is merely an attempt to understand the pulsed ESI phenomenon. The electronic circuit made in this study for pulsed HV has not been optimized for a minimum widening effect, which gave out a pulsed high voltage for the width of ∼55 ms (estimated from Figure 4a), compared with the input pulse width of 50 ms as mentioned above. It is fortunate that this widening in pulse width is still acceptable at the present time for the relatively low frequency used in the experiment. However, an improvement in the pulsed HV circuit can certainly help to reduce the pulse duration. Briefly speaking, the MOSFET used in this pulsed HV circuit works in switch mode with a comparative large current, while the HV source module has only a limit of 1 mA. Thus, either a HV-durable MOSFET with lower current requirement or a more powerful HV source module will greatly enhance the quality of the pulsed HV signal produced by this homemade circuit. Related research is still being undertaken in this laboratory. Establishing Time for Pulsed ESI. A relatively long establishing time required by the pulsed ESI has been experimentally noticed for the tip geometry described in this study. It is speculated that the establishing time for a Taylor cone from pulseoff to pulse-on should be less than 0.05 s (width of the pulsed high voltage) at the present working conditions. However, because the resolution of the USB microscope employed is only 0.08 s, the recorded establishing time for a Taylor cone is just 0.08 s, as can be seen from the photos from 1.17 to 1.25 s in Figure 3. Thus, there seemed to be a notable establishing time for pulsed ESI during each cycle, which should be considered if one tries to use a high running frequency. To have a running frequency of 10 Hz and a off-to-on ratio of at least 5, the pulse duration must be reduced to 0.02 s, although the widening effect caused by electronics and droplet response might be critical if one would like to increase the pulse rate. Pulsed ESI-MS Spectra. The most important proof for pulsed ESI is no doubt the mass spectra obtained in this study. The pulsed electrospray system has been adopted into the PE Mariner TOF-MS instrument to acquire pulsed mass spectra of the test clarithromycin sample (1 µg/µL, dissolved in 7:3 CH3OH/H2O, with 5% acetic acid added). The commercial spray chamber was modified to fit with the pulsed electrospray system. The distance between the sprayer tip and the sampling plate was ∼2 mm. The flow rate was ∼3 µL/min. The internal ESI high voltage provided within the instrument was substituted by the homemade pulsed source. To record the pulsed ESI signal, the data acquisition system was adjusted to a 0.6-s sampling rate. Figure 5 displays the recorded total ion current (TIC, Figure 5a) and the mass spectra for clarithromycin (748.5 Da) at one peak point of the TIC profile (Figure 5b for pulse-on) and at one valley point of TIC profile (Figure 5c for pulse-off). It is interesting in Figure 5 that the peak and valley points in the sawtooth profile happen to correspond to the pulse-on and pulse-off, respectively. It seems that when pulse was off there was no observable mass Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
4751
Figure 5. MS spectra obtained by pulsed ESI-TOF-MS: (a) TIC spectrum; (b) mass spectrum at one peak on TIC for pulse-on; (c) mass spectrum at one valley on TIC for pulse-off. Pulsed voltage used was the same as that in Figure 3.
Figure 6. MS data acquired by dc ESI-MS: (a) TIC spectrum; (b) MS at the point indicated in (a) by an arrow. All experimental conditions except the spray potential were the same as those in Figure 5. The spray potential was kept constant at 2000 V.
spectrum for clarithromycin. Thus, the mass spectra clearly indicate that there was a pulsed ion stream to be detected periodically. Merits of Pulsed ESI-MS. Advantages ensured by the characteristics of high off-to-on ratio can be expected from this 4752
Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
pulsed ESI device. The pulsed electrospray has little sample consumption since it is like a conventional electrospray during pulse-on and a nanoflow rate electrospray in average. In the pulsed electrospray mode, by using sprayer tips with properly narrow inner/outer diameters, the average flow rate can be fixed to 10100 nL/min while the temporary flow rate could reach to one to a few microliters per minute. Thus, the pulsed electrospray seems to be a pulsed microspray worked at a continuous nanospray flow rate. Because there is no need for the spray capillary to be tapered down to a jet with an inner diamter of less than 10 µm,4-5 the makeup of a sprayer tip is relatively easy with less blocking problems. The ability to obtain a stable ESI-MS performance at lower flow rate with a pulsed electrospray device has been demonstrated by comparing spectra acquired with the pulsed electrospray device (See Figure 5) and that with the normal electrospray device (See Figure 6) at the same operating conditions, except for the spray potential. With a stainless steel sprayer tip whose inner diameter is ∼100 µm and whose position was kept constant (∼2 mm from sampling orifice of the MS instrument), pulsed spray (spray potential varied from 1000 to 2000 V periodically) could offer stable mass spectra under a flow rate of 3 µL/min (Figure 5a and b), whereas dc spray (spray potential remained at 2000 V) gave out no stable TIC (see Figure 6a). This defect of dc spray is attributed to the relatively low flow rate compared to the spraying rate. Even at the relatively high point in the TIC (indicated by an arrow in Figure 6a), only a quite poor spectrum could be acquired (Figure 6b). CONCLUSION Overall, the pulsed ESI technique would be useful for the performance improvement of the microseparation ESI-MS interface, since it is obviously easy to change different working con-
ditions with the aid of the proposed system. It might be optimistic to apply the pulsed ESI with automated on-line microseparation MS. What is more, it is quite promising to apply the pulsed ESI device to presently popular microfabricated analytical systems since the pulsed ESI design could be conveniently used to achieve an ultralow flow rate for microfluidic chips to hyphenate with mass spectrometer. Also, the pulsed ESI might be connected with TOF-MS. To develop a more effective ESI-TOF-MS is the motivation for pulsed ESI. Although the frequency of pulsed spray has not been raised to the level of TOF-MS, it is really worth much deeper research on the kinetics of the pulsed ESI phenomenon to augment pulse frequency while not impairing the spraying stability. Because the mass spectrometer samples only when the pulse is on, the noise
level would be lowered on average, with no harm to the signal level. Relative studies are currently being carried out in this laboratory. ACKNOWLEDGMENT This research work is supported by the National Nature Science Foundation of China (Contract H29927002) and by the Chinese Ministry of Education. The authors appreciate the great help offered by Yu Geng, Fang Zhang, and Chongtian Yu in the Shanghai Institute of Organic Chemistry. Received for review March 14, 2001. Accepted July 23, 2001. AC0103118
Analytical Chemistry, Vol. 73, No. 19, October 1, 2001
4753
