Anal. Chem. 2001, 73, 5992-5999
A High-Efficiency Cross-Flow Micronebulizer Interface for Capillary Electrophoresis and Inductively Coupled Plasma Mass Spectrometry Jinxiang Li, Tomonari Umemura, Tamao Odake, and Kin-ichi Tsunoda*
Department of Chemistry, Faculty of Engineering, Gunma University, 1-5-1 Tenjin-cho, Kiryu, Gunma 376-8515, Japan
A pneumatic nebulizer interface for capillary electrophoresis (CE) and inductively coupled plasma mass spectrometry (ICPMS) is reported. The interface is constructed using a high-efficiency cross-flow micronebulizer (HECFMN) and has the following features. (1) Makeup solutions can be fed to the interface by nebulizer selfaspiration and liquid gravity pressurization. (2) The liquid dead volume of the interface is ∼65 nL, much smaller than those (200-2500 nL) reported for other interfaces. (3) The interface can be stably operated at a liquid flow rate down to 5 µL/min with a high analyte transport efficiency up to 95% to the plasma and (4) does not induce noticeable laminar flow in the CE capillary at typical nebulizer gas flow rates of 0.8-1.2 L/min. Because of these features, baseline resolution of 10 lanthanides with a CE-ICPMS system using the HECFMN interface is achieved, and detection limits and peak asymmetry are 0.05-1 µg/L and 0.93-1.23, respectively, improved significantly over those reported previously for a CEICPMS system using a high-efficiency nebulizer interface. Peak precision for the 10 lanthanides is in the range of 6.2-12.3% RSD (N ) 5). Peak widths are from 9.1 s for 139La to 17.9 s for 175Lu. The effects of nebulizer gas flow rate, makeup solution flow rate, and spray chamber volume on CE-ICPMS signal intensity and separation are also evaluated for the HECFMN interface by the separation of Cr3+ and Cr2O72-. The coupling of capillary electrophoresis (CE) with inductively coupled plasma mass spectrometry (ICPMS) for trace element speciation measurements was first reported in 1995.1 Since that time, this technique has received considerable attention because of the potential benefits offered by the combination of these methodologies. In comparison with other chromatographic techniques, CE offers high separation efficiencies and relatively rapid separations and is expected to exert minimal disturbance on the existing equilibrium between different species. As a detection technique, ICPMS provides low detection limits, multielement detection, and element- and isotope-specific detection. The coupling of CE with ICPMS, however, is not as straightforward as * Corresponding author. Fax: +81 277 30 1251 E-mail: tsunoda@chem. gunma-u.ac.jp. (1) Olesik, J. W.; Kinzer, J. A.; Olesik, S. V. Anal. Chem. 1995, 67, 1-12.
5992 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
the coupling of chromatographic techniques with ICPMS and requires a specially designed interface that should be efficient in sample introduction to the plasma and not degrade the resolution achieved in the CE capillary.2,3 A variety of nebulizers have been used for the interfaces between CE and ICPMS. They include pneumatic devices such as the standard concentric nebulizer,1,4-7,12 the cross-flow nebulizer,8-10,14 the direct injection nebulizer,11 the high-efficiency nebulizer,6,7 the microconcetric nebulizer,9,12-18 the MicroMist nebulizer,18,19 the oscillating capillary nebulizer,7,8,20 the directinjection high-efficiency nebulizer,10 and the glass frit nebulizer.21 Also, the electrically driven, ultrasonic nebulizer was investigated.22,23 Among the nebulizers mentioned above, the pneumatic devices are popular, partly because of their simplicity and low cost. However, when being used for the coupling of CE and ICPMS, nearly all the currently commercially available (2) Olesik, J. W.; Kinzer, J. A.; Grunwald, E. J.; Thaxton, K. K.; Olesik, S. V. Spectrochim. Acta, Part B 1998, 53, 239-251. (3) Sutton, K. L.; Caruso, J. A. LC-GC 1999, 17, 36-45. (4) Michalke, B.; Schramel, P. Fresenius J. Anal. Chem. 1997, 357, 594599. (5) Lu, Q.; Bird, S. M.; Barnes, R. M. Anal. Chem. 1995, 67, 2949-2956. (6) Kinzer, J. A.; Olesik, J. W.; Olesik, S. V. Anal. Chem. 1996, 68, 32503257. (7) Sutton, K. L.; B’Hymer, C.; Caruso, J. A. J. Anal. At. Spectrom. 1998, 13, 885-891. (8) Majidi, V.; Miller-Ihli, N. J. Analyst 1998, 123, 803-808. (9) Baker, S. A.; Miller-Ihli, N. J. Appl. Spectrosc. 1999, 53, 471-478. (10) Majidi, V.; Qvarnstrom, J.; Tu, Q.; Frech, W.; Thomassen, Y. J. Anal. At. Spectrom. 1999, 14, 1933-1935. (11) Liu, Y.; Lopez-Avila, V.; Zhu, J. J.; Wiederin, D. R.; Beckert, W. F. Anal. Chem. 1995, 67, 2020-2025. (12) Mei, E.; Ichihashi, H.; Gu, W.; Yamasaki, S.-i. Anal. Chem. 1997, 69, 21872192. (13) Taylor, K. A.; Sharp, B. L.; Lewis, D. J.; Crews, H. M. J. Anal. At. Spectrom. 1998, 13, 1095-1100. (14) Tu, Q.; Qvarnstrom, J.; Frech, W. Analyst 2000, 125, 705-710. (15) Day, J. A.; Sutton, K. L.; Soman, R. S.; Caruso, J. A. Analyst 2000, 125, 819-823. (16) Tangen, A.; Lund, W. J. Chromatogr., A 2000, 891, 129-138. (17) Schaumloffel, D.; Prange, A. Fresenius J. Anal. Chem. 1999, 364, 452456. (18) Holderbeke, M. V.; Zhao, Y.; Vanhaecke, F.; Moens, L.; Dams, R.; Sandra, P. J. Anal. At. Spectrom. 1999, 14, 229-234. (19) B’Hymer, C.; Day, J. A.; Caruso, J. A. Appl. Spectrosc. 2000, 54, 10401046. (20) Kirlew, P. W.; Caruso, J. A. Appl. Spectrosc. 1998, 52, 770-772. (21) Tomlinson, M. J.; Lin, L.; Caruso, J. A. Analyst 1995, 120, 583-589. (22) Lu, Q.; Barnes, R. M. Microchim. J. 1996, 54, 129-143. (23) Kirlew, P. W.; Castillano, M. T. M.; Caruso, J. A. Spectrochim. Acta, Part B 1998, 53, 221-237. 10.1021/ac010595w CCC: $20.00
© 2001 American Chemical Society Published on Web 11/17/2001
pneumatic nebulizers suffer from one or more of the following deficiencies. First, at the typical sample uptake rates of 0.5-2 mL/min, conventional pneumatic nebulizers including the concentric and cross-flow nebulizers have a poor analyte transport efficiency (95% to the plasma. Additionally, the cross-flow design and the use of a smaller sample uptake capillary significantly alleviate the suction of the HECFMN compared to concentric nebulizers or to conventional cross-flow nebulizers. In the present study, a new interface for CE-ICPMS was constructed using the HECFMN with a special effort to overcome the above problems with other pneumatic nebulizer interfaces. This report treats construction of the HECFMN interface and evaluation of it in terms of electrical connection and nebulization stabilities, nebulizer suction effect, and its utility for the coupling of CE and ICPMS. EXPERIMENTAL SECTION CE. The CE system was fabricated in-house, as illustrated in Figure 1. A 75-µm-i.d. by 150-µm-o.d. by 75-cm-length polyimidecoated fused-silica capillary (GL Sciences) was used. The total volume of the capillary is 3.3 nL. The two ends of the capillary were positioned on an identical level to eliminate siphonic effect. Electrophoresis through the CE capillary was driven by an HCZE30PN0.25 high-voltage dc power supply (Matsusada Precision Devices, Inc.) which can be operated in a voltage-controlled (30 kV maximum) or current-controlled (250 µA maximum) mode. Electrical connections between the power supply and the electrolytes were maintained using platinum electrodes. The inlet end of the capillary was held at a positive potential while the outlet end was grounded. No separate cooling measures were taken on the CE system, but the temperature of the laboratory was regulated at 25 °C by a thermostat and an air conditioning system. (25) Li, J.; Umemura, T.; Odake, T.; Tsunoda, K.-i. Anal. Chem. 2001, 73, 14161424.
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5993
Table 1. Makeup Solution Flow Rates (MSFRs) Measured at Different (a) Makeup Solution Levels (MSLs) and (b) Nebulizer Gas Flow Rates (NGFRs) aa
Table 2. CE-ICPMS Operating Conditionsa
capillary
bb
MSL (cm)
MSFR (µL/min)
NGFR (L/min)
MSFR (µL/min)
-30 -10 0 5 10 15
7.7 8.3 8.5 9.2 9.9 10.4
0.8 0.9 1 1.1 1.2
7.5 7.9 8.5 10.1 12.3
a
At a nebulizer gas flow rate of 1.0 L/min. b At a makeup solution level of zero.
ICPMS. The ICPMS system incorporated a JMS-PLASMAX1 high-resolution inductively coupled plasma spectrometer and a personal computer equipped with a Hueline software (JEOL Co.). A standard torch for this instrument was used. The used sampler cone (1.0-mm-diameter orifice) and skimmer cone (1.2-mmdiameter orifice) were made of copper. Interface. The HECFMN interface for CE and ICPMS is illustrated schematically in Figure 1. A 1/16-in. PEEK cross fitting was used to connect the makeup solution delivery PTFE tubing (1.0-mm i.d., 1.58-mm o.d.), the Pt electrode, and the CE capillary to the HECFMN sample delivery tubing. The CE capillary was threaded through a section of microbore Tygon tube (250-µm i.d., 1.58-mm o.d.) and the cross fitting and was inserted into the HECFMN sample delivery tubing. The outlet end of the CE capillary was positioned 100-150 µm from the inlet end of the HECFMN capillary (50-µm i.d., 150-µm o.d., ∼30 mm long). The makeup solution was fed to the interface by nebulizer selfaspiration and liquid gravity pressurization. This was done by adjusting the makeup solution reservoir position up or down along an iron bar having scales with the zero point being on the level of the CE capillary ends. The makeup solution flow rates at different makeup solution levels and nebulizer gas flow rates are presented in Table 1. The CE capillary outlet end was grounded through the makeup solution sheath flow. The spray chamber was fabricated from a section of PP tube (15-mm i.d., 16.5-mm o.d.). The spray chamber volume was varied through using the tubes of different lengths. No exit for drain was prepared on the spray chamber. Reagents. Two standard solutions were used for CE separations with one containing Cr3+ and Cr2O72- at 20 µg/L Cr per species and the other 10 lanthanides (La, Nd, Sm, Tb, Dy, Ho, Er, Tm, Yb, Lu) each at 10 µg/L. The solution containing the two Cr species was prepared from Cr(NO3)3‚9H2O and K2C2O7, and the solution of the 10 lanthanides was made from their respective commercial standard stock solutions (1000 mg/L in 0.1 M HNO3) of nitrate salts. The run buffers used were 20 mM CH3COONa (pH 7.2) and 4 mM 2-hydroxyisobutyric acid (HIBA, pH 4.2) for the separations of the Cr species and the lanthanides, respectively. A solution of 0.1% HNO3 was used as the makeup solution throughout. Y of 1 µg/L and Pb of 20 µg/L were added to the makeup solution and the buffers, respectively, to initially adjust the ICPMS system daily and to monitor the stability of the CEICPMS system during separations. All the reagents used were of 5994 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
voltage room temperature sample injection nebulizer capillary spray chamber volume nebulizer gas flow rate makeup solution flow rate forward power radio frequency coolant gas auxiliary gas mass resolution ion beam adjustment detection mode magnetic field dwell time sweep width sweep rate smoothing points isotopes monitored
a
CE polyimide-coated, 75-µm i.d., 150-µm o.d., 75 cm long 20 kV 25 °C electrokinetic at 20 kV for 20 s Interface 50-µm i.d., 150-µm o.d. 15 mL 1.0 L/min 8.5 µL/min ICPMSb 1100 W 40.68 MHz 14 L/min 0.8 L/min 500 optimized daily on 89Y selected ion monitoring (SIM) 100 ms 6000 ppm 100 ms 0 53Cr, 208Pb, 139La, 146Nd, 149Sm, 159Tb,162Dy, 165Ho, 166Er, 169Tm, 174Yb, 175Lu
Unless otherwise noted. bRefer to ref 25 for the other parameters.
analytical reagent grade and from Wako Pure Chemical Industries, and all the dissolvations and dilutions were performed using 18 MΩ‚cm distilled, deionized water. All the solutions were filtered through 0.45-µm membrane filters (13P, GL Sciences) and degassed by vacuum prior to use. CE-ICPMS Measurements. The CE capillary was conditioned daily by purging with 0.1 M NaOH for 20 min and 18 MΩ‚ cm distilled, deionized water for 10 min and finally equilibrated with the run buffer for 20 min. Between each run, the CE capillary was flushed with the run buffer for 10 min. These were performed by forcing the cleaning agents into the CE capillary from the inlet end at a flow rate of 5 µL/min using a syringe pump (model 100, KD Scientific Inc.) equipped with a 1-mL syringe. The ICPMS ion beam was adjusted daily for the optimal signal of 89Y in the makeup solution. Sample injections were electrokinetic at 20 kV for 20 s without any effect of pressure. Data were acquired in selected ion monitoring (SIM) mode. The operating conditions used for the CE-ICPMS system were all typical, as presented in Table 2. The effects of nebulizer gas flow rate, makeup solution flow rate, and spray chamber volume on measurements were evaluated for the CE-ICPMS system in terms of signal-to-noise ratio, temporal peak width at the baseline, migration time, peak asymmetry, and electrophoretic resolution by separating Cr3+ and Cr2O72- using 20 mM CH3COONa as the run buffer. Separation of 10 lanthanides was performed using 4 mM HIBA as the run buffer under the selected operating conditions listed in Table 2, and several analytical figures of merit, including detection limits, temporal peak widths at the baseline, peak asymmetry, and signal precision, were acquired for the 10 lanthanides. Data Processing. Data were acquired after peaks were smoothed with five smoothing points. Signal-to-noise ratios, detection limits, and signal precision values were calculated based on peak height. The detection limit was defined as the concentra-
tion giving a signal equivalent to 3 times the standard deviation of noise signals. The noise standard deviation was calculated from seven successive noise signals adjacent to and in front of the peak. For Cr2O72-, the seven noise signals were taken from the rear of the peak. Resolution (R) was calculated with the following equation:
R ) 2∆t/(WCr3+ + WCr2O72-)
where ∆t is the difference in migration time and WCr3+ and WCr2O72are the peak widths at baseline for Cr3+ and Cr2O72-, respectively. Peak asymmetry (As) was calculated using the following formula:
As ) B/A
where A is the distance from the peak front to the center of the peak at 10% of the peak height and B is the distance from the peak tail to the center of the peak. RESULTS AND DISCUSSION Structural Considerations. CE capillary dimensions have an effect on CE separation in several respects such as migration time, sensitivity, and separation efficiency. When ICPMS with a pneumatic nebulizer interface is used as the detector for CE, the CE capillary dimensions also influence the magnitude of the suction effect created by the pneumatic nebulizer. Generally, the CE capillaries with a smaller inside diameter offer higher separation efficiencies and suffer from weaker nebulizer suction effect, but give poorer detection limits than those with a larger inside diameter.16 In the present study, a 75-µm-i.d. by 75-cm-length capillary was intentionally selected for evaluation of the HECFMN utility in CE-ICPMS because these dimensions are very typical in CE. As for outside diameter of the CE capillary, 150 µm was chosen because it matches the inside diameter of the HECFMN sample delivery tubing (250-µm i.d.) into which the CE capillary was inserted (see Figure 1). If a CE capillary with a larger outside diameter is used, the inside diameter of the HECFMN sample delivery tubing should be widened properly in order to provide enough annular space for makeup solution to pass through. Otherwise laminar flow might be caused in the CE capillary because of nebilizer suction. As described earlier, dead volumes of an interface, including liquid and gas dead volumes, contribute to band broadening and peak asymmetry, thereby degrading electrophoretic resolution. Therefore, the reduction in the dead volumes should be taken into consideration. For the HECFMN interface, the liquid dead volume between the CE capillary outlet end and the outlet end of the HECFMN capillary (50-µm i.d., ∼30 mm long) was calculated to be ∼65 nL. In comparison with the liquid dead volumes (2002500 nL) of other interfaces reported previously,6,8,9 the 65-nL liquid dead volume is much smaller; thus, one should expect an improvement in peak widths and symmetry with the HECFMN interface. On the gas dead volume of the HECFMN interface, namely, the volume of the spray chamber, the discussion will be presented in a subsequent section. As introduced earlier, for the interfaces reported previously, the most commonly used method for makeup solution introduction
Figure 2. Electropherograms obtained for Cr3+ and Cr2O72- at different nebulizer gas flow rates (NGFRs). Each species was present at 20 µg /L Cr. Buffer, 20 mM CH3COONa (pH 7.2); electrophoretic current, 37-39 µA.
was driving the liquid using a pump. In these cases, besides the purpose of establishment of the CE electrical circuit, the makeup solution also served the purposes of supplement of the low liquid flow from the CE capillary for stable nebulization and control of nebuizer suction. The HECFMN can efficiently and stably nebulize the liquid flows down to 5 µL/min25 and does not induce noticeable laminar flow in the CE capillary (see a subsequent section). Therefore, in the present study, the makeup solution was fed to the interface using nebulizer self-aspiration and gravity pressurization, considering that this technique has the advantages described earlier. Effect of Nebulizer Gas Flow Rate. A higher nebulizer gas flow rate tends to induce a faster laminar flow in the CE capillary,1,6-9,18 while a lower nebulizer gas flow rate results in a lower analyte transport efficiency to the plasma6,25 and a longer spray chamber washout time.1,8 Figure 2 and Table 3 show the effect of nebulizer gas flow rate on the Cr signal intensity and the separation of Cr3+ and Cr2O72- at a makeup solution level of zero. The maximal signalto-noise ratios are found at 1.0 L/min of the nebulizer gas flow rate. This is consistent with that observed when the HECFMN was used for direct introduction of sample to the plasma at a flow rate of 5 µL/min in an our previous study.25 Generally, peak widths, peak asymmetry, and electrophoretic resolution are not Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5995
Table 3. Effect of Nebulizer Gas Flow Rate on Signal Intensity and Separationa nebulizer gas flow rate (L/min) (7.5)b
0.8 0.9 (7.9)b 1 (8.5)b 1.1 (10.1)b 1.2 (12.3)b
S/N Cr3+
Cr2O72-
Cr3+
14.7 21.9 30.5 16.1 12.6
8.9 13.3 18.4 9.7 7.1
17.4 10.2 8.7 7.8 8.9
W (s) Cr2O7213.2 7.8 7.1 7.8 7.4
As Cr3+
Cr2O72-
Cr3+
1.46 1.21 1.07 0.95 1.1
1.2 0.95 1.11 1.14 0.94
6.32 6.36 6.38 6.41 6.19
T (min) Cr2O727.7 7.54 7.57 7.55 7.29
R 5.4 7.9 9 8.8 8.1
a S/N, signal-to-noise ratio; W, temporal peak width at the base line; A , peak asymmetry; T, migration time; R, resolution; Sample and buffer s conditions, same as those in Figure 2. bThe values in parentheses represent the makeup solution flow rates (in µL/min) corresponding to the nebulizer gas flow rates.
remarkably affected, but indeed they are improved slightly as the nebulizer gas flow rate is increased from 0.8 to 1.2 L/min. Migration times are not changed noticeably as a function of the nebulizer gas flow rate. This will be discussed further below. For the following experiments, a nebulizer gas flow rate of 1.0 L/min was selected. Effect of Makeup Solution Flow Rate. For a CE-ICPMS interface using a pneumatic nebulizer, too low a makeup solution flow rate leads to irregular nebilization7,17 and severe effects of nebulizer suction6,8,13,18 and postcapillary dispersion.7-9,12,16 On the other hand, too high a makeup solution flow rate results in degraded aerosol quality and, thus, a reduced analyte transport efficiency to the plasma.9,12,13,21,25 At a nebulizer gas flow rate of 1.0 L/min, the effect of makeup solution flow rate or makeup solution level on the Cr signal intensity and the separation of Cr3+ and Cr2O72- was examined, and the electropherograms and results obtained are offered in Figure 3 and Table 4, respectively. The makeup solution flow rate exhibits an unremarkable effect on the signal-to-noise ratio, with the flow rates from 8.3 to 8.5 µL/min giving slightly higher signal-to-noise ratios. This is ascribed to the similar analyte transport efficiencies of >90% to the plasma with the HECFMN at all the makeup solution flow rates tested.25 This conclusion is also supported by our careful observation that there was no deposition or collection of liquid on the sides or end of the spray chamber during the measurements. Signal-to-noise ratios are slightly decreased at higher and lower makeup solution flow rates because of peak broadening since the signal-to-noise ratios were calculated based on peak height. Migration times and peak asymmetry are not changed noticeably at makeup solution flow rates less than 8.5 µL/min, viz., at makeup solution levels below the zero position. As the makeup solution flow rate is increased from 8.5 to 9.9 µL/min, i.e., the makeup solution level is raised from 0 to +10 cm, the migration times are prolonged from 6.34 to 13.07 min and from 7.52 to 21.13 min, respectively, for Cr3+ and Cr2O72-, and resolution is improved dramatically despite the peaks being somewhat broadened and distorted. At the makeup solution flow rate of 10.4 µL/min, viz., at the makeup solution level of +15 cm, the Cr3+ peak is eluted at 20.05 min while the Cr2O72- peak does not appear within 30 min of the monitoring period. These observations are attributed to the fact that high liquid back pressures were caused at the outlet end of the CE capillary and, accordingly, retarded, stopped, or even reversed the EOF through CE capillary at makeup solution flow rates above 8.5 µL/min, viz., at makeup solution levels above 5996 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
Figure 3. Electropherograms obtained for Cr3+ and Cr2O72- at different makeup solution flow rate rates (MSFRs). Sample and buffer conditions, as in Figure 2.
the zero position. Kinzer et al.6 and Culbertson and Jorgenson26 demonstrated that migration times and resolution could be increased through generating a laminar flow in a direction away from the detector. Sutton et al.7 also reported that, if the makeup solution flow rate were too high, the back pressure created by (26) Culbertson, C. T.; Jorgenson, J. W. Anal. Chem. 1994, 66, 955-962.
Table 4. Effect of Makeup Solution Flow Rate or Makeup Solution Level on Signal Intensity and Separationa makeup solution flow rate (µL/min) (-30)b
7.7 8.3 (-10)b 8.5 (0)b 9.2 (5)b 9.9 (10)b 10.4 (15)b
Cr3+ 30.3 32.1 31.7 30.5 28.9 24.5
S/N Cr2O7218 18.3 19.2 18.7 17.4
Cr3+
W (s) Cr2O72-
12 8.4 9 14.4 15.6 19.2
13.8 8.4 6.6 9.2 12
As Cr3+
Cr2O72+
Cr3+
1.07 1 0.96 1.24 1.4 2.32
1.01 0.91 0.93 1.04 1.1
6.73 6.43 6.34 9.91 13.07 20.05
T (min) Cr2O727.82 7.59 7.52 14.11 21.13
R 5.1 8.3 9.1 21.4 35
a Definitions of the symbols, same as those in Table 3; sample and buffer conditions, same as those in Figure 3. bThe values in parentheses represent the makeup solution levels (in cm) corresponding to the flow rates.
the makeup solution at the outlet end of the CE capillary would oppose the direction of the EOF so that peaks were not eluted and remained in the CE capillary. This attribute of the HECFMN interface can be considered to be an advantage considering that it allows easy control of the tradeoffs between resolution and analysis times without significant loss of detection limits. At lower makeup solution flow rates, e.g., 7.7 µL/min, peaks are slightly broadened and resolution is therefore reduced owing to the liquid dead volume of the interface. At higher makeup solution flow rates, peaks are noticeably broadened in comparison to those at lower makeup solution flow rates. Similar peak broadening was also observed by B’Hymer et al.19 and they ascribed such peak broadening to the analyte diffusion in CE capillary. For the Cr3+ peak broadening in the present study, probably there is a second reason that the highly charged cation of Cr3+ interacted with the electric double layer in the CE capillary.12,27 At higher makeup solution flow rates, peak asymmetry also deteriorated, especially for Cr3+. For example, at the makeup solution flow rate of 9.9 µL/min, the peak asymmetry values are 1.4 and 1.1 for Cr3+ and Cr2O72-, respectively. Presumedly this resulted from the interactions of the Cr3+ and the electric double layer. Both the diffusion and the interaction become worse as the migration time is increased.12 For the following experiments, a makeup solution flow rate of 8.5 µL/min was used; i.e., the makeup solution was kept on the zero level with the nebulizer gas flow rate being fixed at 1.0 L/min. Spray Chamber Volume. The spray chamber with a larger internal volume contributes to peak broadening and tailing in CEICPMS,1,5,6,12,16,17,20 while the spray chamber having a smaller volume suffers from inefficiency in introduction of sample to the plasma.13 Figure 4 and Table 5 show the effect of spray chamber volume on the Cr signal intensity and the separation of Cr3+ and Cr2O72-. In general, the spray chamber volume from 5 to 30 mL does not remarkably affect the signal intensity and the separation. Signalto-noise ratios are slightly decreased with the smaller spray chambers. This is because the smaller spray chambers were apt to cause aerosol droplet coagulation on the end of the spray chamber where the deposition of liquid was observed in our study when the 5-mL spray chamber was used. Taylor et al.13 compared two cyclonic spray chambers with the internal volumes of 21.0 and 6.5 mL and found that the smaller one yielded a reduced signal response for 115In in comparison to the larger one. A little peak tailing is observed, especially from the Cr3+ peaks, when (27) Schure, M. R.; Lenhoff, A. M. Anal. Chem. 1993, 65, 3024-3037.
Figure 4. Electropherograms obtained for Cr3+ and Cr2O72- using different spray chamber volumes (SCVs). Sample and buffer conditions, as in Figure 2 or 3.
the larger spray chambers, e.g., the 30-mL spray chamber, are used. Olesik et al.1 reported that the spray chamber did not contribute greatly to peak broadening measured as the full width at half-height, but it might contribute to peak tailing. A spray chamber with a 15-mL internal volume was chosen for the following experiments. Stability. The stability of the CE-ICPMS system was examined by continuously monitoring the signal of Pb+ in the CH3COONa buffer at m/z 208 and the signal of Y+ in the HNO3 makeup solution at m/z 89 for 30 min. The time versus signal response plots obtained for 208Pb and 89Y demonstrated the Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5997
Table 5. Effect of Spray Chamber Volume on Signal Intensity and Separationa S/N
spray chamber volume (mL)
Cr3+
Cr2O72-
Cr3+
5 10 15 20 30
27.2 29.9 32.1 32.4 31.1
16.5 17.8 18.7 18.4 17.1
9.6 9.7 9.8 9.6 14.4
a
W (s) Cr2O726.9 6.8 7.1 7.4 10.8
As Cr3+
Cr2O72-
Cr3+
0.9 0.96 1.03 1.26 1.59
0.95 1.05 1.02 1.16 1.32
6.35 6.38 6.37 6.41 6.44
T (min) Cr2O72+ 7.57 7.58 7.55 7.63 7.58
R 8.9 8.7 8.4 8.6 5.4
Definitions of the symbols, same as those in Table 3 or 4; sample and buffer conditions, as those in Figure 4.
stabilities of EOF through the CE capillary and nebilization with the HECFMN, respectively. But a regular, downward shift (90% achieved by the HECFMN at the low sample uptake rate of 8.5 µL/min. To minimize the HEN suction effect, Sutton et al.7 used a nonoptimum nebulizer gas flow rate (0.6 L/min) and a high makeup solution flow rate (61 µL/min) in their study. Undoubtedly the nonoptimum nebulizer gas flow rate and the high makeup solution flow rate collectively reduced the analyte transport efficiency to the plasma and, consequently, compromised the detection limits. The peak asymmetry values obtained with the HECFMN interface are from 0.93 to 1.23, compared to 0.383.1 with the HEN interface. The narrow and nearly Gaussianshaped peaks obtained in the present study are attributed to the smaller dead volume and no suction effect for the HECFMN interface. Sutton et al.7 imputed the peak tailing, in part, to the suction effect of the HEN interface in their study. The temporal peak widths at the baseline are from 9.1 to 17.9 s for the 10 lanthanides, generally increasing gradually with the increase in migration times. The time-related and symmetric peak broadening was not observed in similar separation experiments performed by Sutton et al.7 and Shi and Fritz.29 Presumedly, such peak broadening originated from mainly the analyte diffusion in the CE capillary rather than the dead volume of the interface. Since no special cooling measures were taken on the CE system, the band broadening in the CE capillary due to Joule heating is expected in the present study. The signal precision values obtained in the present study are in the range of 6.2-12.3% RSD (N ) 5), comparable to those (∼10% RSD, N ) 5) reported by Sutton et al.7 for the HEN interface.
CONCLUSIONS The HECFMN interface developed in the present study does not induce noticeable laminar flow in the CE capillary at the typical nebulizer gas flow rates of 0.8-1.2 L/min because of the crossflow design and the smaller sample uptake capillary used for the HECFMN; therefore, the nebulizer gas flow rate and the makeup solution flow rate can be optimized independently of nebulizer suction effect. Additionally, the HECFMN interface can be stably operated at a low makeup solution flow rate down to 5 µL/min with a high analyte transport efficiency up to 95% to the plasma.25 These collectively led to lower detection limits with the HECFMN interface (0.05-1 µg/L) than those with the HEN interface (0.340.94 µg/L) for 10 lanthanide ions. The liquid dead volume of the HECFMN interface is ∼65 nL, much smaller than those (2002500 nL) reported previously for other interfaces.6,8,9 The smaller liquid dead volume, together with no nebulizer suction effect, can considerably reduce the post- and in-capillary dispersion. This was demonstrated by the better symmetry of the peaks obtained with the HECFMN interface than those with the HEN interface. The peak widths obtained with the HECFMN interface for 10 lanthanides were in the range from 9.1 to 17.9 s, increasing gradually with the increase in migration times. Such peak broadening was mainly attributed to the heat diffusion in the CE capillary. It is expected that the peak widths for the 10 lanthanides should be improved further if a CE system with a thermostat takes the place of the one used in the present study. All the features of the HECFMN interface suggest a good potential to routinely use the interface in coupling CE with ICPMS for real sample speciation analysis that is currently underway in our laboratory.
Received for review May 29, 2001. Accepted September 23, 2001. AC010595W
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5999
