Anal. Chem. 1996, 68, 883-887
Isotope Ratio Measurements with Elemental-Mode Electrospray Mass Spectrometry Michael E. Ketterer* and John P. Guzowski, Jr.†
Department of Chemistry, John Carroll University, University Heights, Ohio 44118
The isotope ratio capabilities of an electrospray ionization source interfaced to a quadrupole mass spectrometer are described. With the instrument operated in the metal ion mode, isotope measurements of Ag, Tl, and Pb are conducted using elemental ions produced from 1 × 10-4 M solutions of metal nitrates or acetates in methanol. For Ag and Tl, spray conditions are identified that produce spectra free of MH+ ions. Unbiased Ag and Tl ratio measurements with precisions of ∼0.2% RSD are readily attained. Further improvement in relative precision appears to be limited by temporal drift in the degree of mass discrimination imparted to the measurements by the mass spectrometer. Isotopic analysis of Pb is greatly complicated by significant yields of PbH+ polyatomic ions. Recent advances in electrospray mass spectrometry (ESMS) by Horlick and co-workers have established its potential as an ionization source for elemental mass spectrometry.1,2 It therefore follows that ESMS should offer the capability of isotopic measurements for applications such as isotope dilution, isotope tracer studies, and stable isotope geochemistry. The well-established solid source technique of thermal ionization mass spectrometry (TIMS) has long been applied to isotope measurements in analytical chemistry, medicine, and the earth sciences. While TIMS produces isotope ratio measurements with outstanding accuracy (∼0.005 relative percent), sample preparation is laborious, analyses are time consuming, and sample throughput is therefore low. Isotope ratio measurements using inductively coupled plasma mass spectrometry (ICPMS) have been studied extensively, and the technique offers important advantages over TIMS in terms of sample preparation requirements and sample throughput.3 With multicollector, magnetic sector mass analyzers, ICPMS can produce isotope ratio data with relative accuracies on the order of 0.02-0.05%.4-7 The general capabilities of various techniques for atomic mass spectrometry have been recently reviewed.8 † Present address: Department of Chemistry, Indiana University, Bloomington, IN 47405. (1) Agnes, G. R.; Horlick, G. Appl. Spectrosc. 1992, 46, 401-406. (2) Agnes, G. R.; Horlick, G. Appl. Spectrosc. 1994, 48, 649-661. (3) Jarvis, K. E.; Gray, A. L.; Houk, R. S. Inductively Coupled Plasma Mass Spectrometry; Blackie: Glasgow, 1992. (4) Walder, A. J.; Furuta, N. Anal. Sci. 1993, 9, 675-680. (5) Walder, A. J.; Abell, I. D.; Platzner, I.; Freedman, P. A. Spectrochim. Acta 1993, 48B, 397-402. (6) Walder, A. J.; Platzner, I.; Freedman, P. A. J. Anal. At. Spectrom. 1993, 8, 19-23. (7) Walder, A. J.; Koller, D.; Reed, N. M.; Hutton, R. C.; Freedman, P. A. J. Anal. At. Spectrom. 1993, 8, 1037-1041. (8) Colodner, D.; Salters, V.; Duckworth, D. C. Anal. Chem. 1994, 66, 1079A1089A.
0003-2700/96/0368-0883$12.00/0
© 1996 American Chemical Society
Isotope measurements with an electrospray source configured for optimal production of elemental ions might be expected to offer performance capabilities similar to those of ICPMS, with some potential advantages such as lower capital investment and operating costs. Electrospray ion sources have been readily coupled to separation techniques such as microcolumn liquid chromatography and capillary electrophoresis.9-11 The electrospray source is intrinsically suited for handling low sample flow rates of about 1-10 µL/min. In this report, results are presented that indicate the potential of ESMS as a technique for isotope ratio measurements. Our results have been obtained by interfacing an electrospray ionization source (fabricated in-house) to a commercially available quadrupole mass spectrometer designed for use in conjunction with an ICP ion source. This work emphasizes the capabilities of the ionization method since quadrupole mass analyzers have well-known limitations in isotopic analysis (i.e., high levels of mass discrimination as well as nonsimultaneous detection of ions of differing m/z). The ES ion source used herein is based on the design of Agnes and Horlick.1 It has been operated in the metal ion mode by appropriate adjustment of dc potentials. We present findings for isotopic analyses of Ag, Tl, and Pb. EXPERIMENTAL SECTION Ion Source Construction and Operation. An electrospray ionization source was constructed in-house using the design of Agnes and Horlick.1 Preliminary experiments with syringe pumping of the sample solution revealed poor signal stability. Consequently, a pneumatic sample delivery device was constructed and is depicted in Figure 1. The sample solution is placed into a 10 or 25 mL Erlenmeyer flask, whereby a 1/16 in. o.d. stainless steel dip tube transfers solution to the electrospray tip. Nitrogen or argon is used to pressurize the vessel to 20 psig, subsequently forcing liquid through the transfer lines. Two GCtype flow controllers (Model 202-3(3)-1 VICI Condyne, Duarte, CA; Model 8601C, Brooks Instrument Co., Hatfield, PA) are configured in series and used in conjunction with a Valco stainless steel tee to deliver ∼5 µL/min of sample to the tip. The electrospray tip is prepared by inserting a 25 mm length of 0.2 mm o.d. stainless steel capillary tubing (Hamilton) into a length of 1/16 in. stainless steel HPLC tubing and securing the pieces together with polymeric resin (Loctite 271). The tip is electrically isolated from the flow controllers by 1/16 in. PEEK tubing and (9) Blades, A. T.; Ikonomou, M. G.; Kebarle, P. Anal. Chem. 1990, 62, 957967. (10) Smith, R. D.; Barinaga, C. J.; Udseth, R. Anal. Chem. 1988, 60, 19481952. (11) Olivares, J. A.; Nguyen, N. T.; Yonker, C. R.; Smith, R. D. Anal. Chem. 1987, 59, 1232-1236.
Analytical Chemistry, Vol. 68, No. 5, March 1, 1996 883
Figure 1. Pneumatically assisted sample delivery system for delivery of 1-10 µL/min sample solution to the ES spray tip.
reusable PEEK fingertight fittings. Sample solution is changed by removing the tip voltage, exchanging sample reservoirs, and temporarily blocking the split gas outlet. This creates a high flow of gas, flushing out the length of tubing leading to the spray tip with fresh sample. The signal is noted to be strongly dependent upon the physical relationship between the ES tip and front plate orifice. An insulated mechanical arm is placed on top of the ICP torch box (not removed) and used to position and hold the electrospray tip in the vicinity of the front plate. X-Y-Z adjustments to the tip position could be made with the torch box positioning knobs. The arrangement did not offer a quantitative scale for measuring the tip position with respect to the front plate. The tip is placed ∼10 mm in front of the front plate, and several millimeters below the axis of the front plate orifice. The ionization source operating conditions and interface design were not exhaustively studied nor optimized. Parameters were tuned daily to obtain reasonable signal levels for solutions containing analytes of interest. Ion source conditions referred to in subsequent figures and tables are representative examples of parameters identified for optimal elemental ion signal. Typical metal ion mode source conditions are as follows: tip voltage, +3000-7000 V dc; front plate voltage, +1000-2500 V dc; sampling plate voltage, +100400 V dc; curtain gas, 0.5-2.0 slpm nitrogen. Tip voltage was supplied by a Bertan Model 10R 0-10 kV adjustable power supply (Bertan High Voltage, Hicksville, NY). Potentials for the front plate and sampling plate were supplied by 0-3000 and 0-400 V regulated supplies, respectively. In general, somewhat larger potential differences (i.e., between the tip and front plate, and between the front plate and sampling plate) were identified as optimal elemental ion signal conditions for Pb as compared to either Ag or Tl. The quadrupole mass spectrometer is that used in the commercial Sciex Elan Model 500 ICPMS. Two modifications were made to spectrometer: the “shadow stop” was removed, and the channeltron electron multiplier was replaced by an active film multiplier (Model AF 561, ETP Scientific, Auburn, MA). Removal of the shadow stop enhanced the ES signal by ∼3-5 fold. Reagents. Sample solutions were prepared in spectrophotometric or HPLC-grade solvents using lead nitrate, silver nitrate, and thallium acetate. Elemental lead was obtained in the form of NIST 981 (common lead). Lead nitrate was prepared by dissolution of the element in a heated solution of 8 M aqueous nitric acid (Fisher trace metal grade). Pb(NO3)2 readily precipitated from solution after about 1 h. Silver nitrate was obtained from NIST as a certified isotopic standard (NIST 978A). Thallium acetate was obtained from Aldrich Chemical Co. Most studies were performed using methanol as the solvent. A Pb stock solution, ∼0.1 M, was prepared by dissolving Pb(NO3)2 in the 884
Analytical Chemistry, Vol. 68, No. 5, March 1, 1996
minimum volume of water followed by dilution with methanol. Silver and thallium stock solutions were prepared by directly dissolving the soluble nitrate and acetates in methanol. Diluted solutions were prepared by dilution of the ∼0.1 M concentrated stock solutions to ∼1 × 10-4 M for isotopic measurements. Most studies involved use of HPLC-grade methanol as a solvent; a few experiments were also conducted in ethanol, n-propanol, acetonitrile, acetone, and methanol-d4 (Cambridge Isotope Labs) solvents. No attempt was made to exclude water from the starting compounds or polar organic solvents used. Isotopic Measurements. Isotope ratio measurements were obtained in the “peak-hop” mode using scanning and integration parameters analogous to those employed with ICPMS isotopic measurements.12 Parameters investigated include ES tip and plate voltages, ion lens voltages (B, E1, P, and S2), dwell time, and measurement time. Isotope ratio measurements were collected at a single point across each spectral peak. Additionally, total collection times were the same for all masses. Mass spectral resolution was examined by collecting 10 or 20 m/z points across each peak. RESULTS AND DISCUSSION Mass Spectral and Isotope Ratio Characteristics of Pb. Ideally, ESMS could potentially be used in a very straightforward fashion for lead isotopic analysis. As was observed by Agnes and Horlick,2 Pb+ ions could be readily produced using this ES ion source. Pb isotopic analysis, however, was found to be encumbered by two basic problems: (a) a decrease in mass spectral resolution when using this mass analyzer with an ES ionization source and (b) high yields of PbH+ ions. It was noted that the 208Pb peak widths with the ES source were about 1.6 m/z at 10% height under the same mass spectral conditions which produce widths of 1.0-1.1 m/z with use of the ICP source. In the “low” resolution mode, individual isotopes are not well resolved in the ESMS spectrum. By operating in the “high” resolution mode, which produces ICP peak widths of 0.6-0.7 m/z, more suitable ES peak widths of 0.85-0.90 m/z were obtained. The decrease in resolution with the ES ion source (compared to the ICP ion source) is undoubtedly related to differences in ion kinetic energy distributions. Nevertheless, adequate resolution was attainable with ∼30 relative percent signal loss. High-resolution spectra revealed the presence of PbH+ ions; this assignment was clarified by obtaining spectra in methanol-d4 and noting the presence of a PbD+ peak at m/z 210, as shown in Figure 2. Yields of PbH+ species, obtained by measuring the 208Pb1H+/208Pb ratio in non-deuterated methanol, were determined to span the range of 0.02-0.27 and found to be primarily dependent upon the sampling plate voltage and choice of solvent. Adjustments to the curtain gas flow rate as well as the tip, front plate, and sampling plate voltages failed to identify conditions where the 208Pb1H+/208Pb+ ratio could be reduced to negligible levels. Evidently, the PbH+ ions are formed through charge reduction processes which have been previously noted in the ES mass spectra of solvated polyvalent metal ions.13,14 The PbH+ formation mechanism was not investigated further. Switching solvents did not obviate problems with high PbH+ yields; (12) Ketterer, M. E.; Peters, M. J.; Tisdale, P. J. J. Anal. At. Spectrom. 1991, 6, 439-443. (13) Blades, A. T.; Ikonomou, M. G.; Kebarle, P. Int. J. Mass Spectrom. Ion Processes 1990, 102, 251-267. (14) Stewart, I.; Horlick, G. Anal. Chem. 1994, 66, 3983-3993.
Table 1. Electrospray Mass Spectrometric Isotope Ratios of Thallium Acetate Prepared in Methanola
block
no. of replicates
1 2 3
8 10 10
205Tl/203Tl
ratiob,c
2.3909 ( 0.013 2.3775 ( 0.007 2.3894 ( 0.006
precision (%) exptl predd 0.56 0.28 0.24
0.34 0.35 0.21
a Conditions: tip, 5.34 kV; front plate, 2.5 kV; sampler, 220 V; ion lens voltages (V), B, 4.3, E1, -19.41, P, -8.9, S2, -4.7. Blocks 1-3 were obtained under similar experimental conditions on three different days; tmeas ) 7.5 s, tdwell ) 75 ms. b The naturally occurring value for 205Tl/203Tl is 2.3871. c Experimental uncertainties reflect 1 standard deviation. d Relative standard deviations predicted upon counting statistics using the average signal intensities for the block of measurements and the equation RSD ) 100[203N-1 + 205N-1]0.5, where N is the total number of ions counted.
Figure 2. High-resolution mass spectrum of 1 × 10-4 M NIST 981 Pb in methanol-d4. ES conditions: spray tip, +6500 V; front plate, +2500 V; sampling plate, +210 V; curtain nitrogen, 1.0 slpm. Note that the mass calibration is offset by approximately -0.3 m/z.
208Pb1H+/208Pb
ratios similar to those found in methanol were also obtained in acetone, acetonitrile, n-propanol, and ethanol. Since no rigorous attempts to exclude water were made, small quantities of water present in the acetonitrile may be responsible for the persistence of PbH+ ions in the ES mass spectrum obtained from lead in this solvent. Using either low or high mass spectral resolution, lead isotope ratios could be obtained which were reasonably precise (∼0.51.0% RSD) as well as qualitatively in agreement with the expected ratios (biases of ∼5 relative percent). Low-resolution ratios were encumbered by poor abundance sensitivities, which produced significant overlaps of nPb+ on n+1Pb+ and n-1Pb+; high-resolution measurements were still hampered by the remaining problem of isobaric PbH+ ions. While the Pb results do not completely preclude isotopic determinations by ESMS, ion source conditions and solvents could not be identified which enabled the collection of unbiased Pb measurements. Changes in the ion source design as a means to reduce the yield of PbH+ are presently being investigated. Thallium Isotope Ratios. Isotope ratio measurements were obtained for solutions of Tl(I) acetate in methanol. TlH+ ions could not be detected in the mass spectrum; the TlH+/Tl+ was determined to be
