Three-Dimensional Motional Stabilization in the Trapping Field of an

Three-dimensional motional stabilization of radial trajec- tories of low-mass organic ions in an open cell using only a dc trapping field is applied t...
0 downloads 0 Views 168KB Size
Download PDF
Anal. Chem. 1996, 68, 1321-1327

Three-Dimensional Motional Stabilization in the Trapping Field of an Open-Ended Trapped-Ion Cell: Application to the Remeasurement Experiment in Fourier Transform Ion Cyclotron Resonance Mass Spectrometry Victor H. Vartanian and David A. Laude*

Department of Chemistry and Biochemistry, University of Texas at Austin, Austin, Texas 78712

Three-dimensional motional stabilization of radial trajectories of low-mass organic ions in an open cell using only a dc trapping field is applied to the FT-ICR remeasurement experiment. More than 300 remeasurement cycles are observed with 99.59% remeasurement efficiency for benzene (m/z 78) using a high-pressure helium buffer gas. The enhancement in remeasurement efficiency is due to collisional stabilization of the guiding center of ion motion by dynamic motional averaging in the axial position-dependent radial electric field. Such dc-induced radial stabilization is in contrast to the stability produced by application of radio frequency fields characteristic of quadrupolar axialization or rf-only mode operation. The same effect is produced because ions experience a radial “pseudopotential” during axial oscillation as in timevarying fields. Trajectory simulations for ions oscillating in an open cell trapping well above a z-amplitude critical threshold energy of 0.60 eV (in a potential well of 0.84 V) indicate that radially stabilized trapping motion is achieved because the outward-directed radial electric field existing near the cell center line is compensated by an opposing inward-directed radial electric field at extended z-amplitude. Sufficient axial kinetic energy permits ion penetration into the inward-directed radial electric field regions, enabling >50% residence time of each trapping oscillation period in regions inducing radial stability, thereby inhibiting magnetron radius growth. A highpressure, low-mass buffer gas such as helium provides the requisite increase in the axial amplitude of the ion cloud, similar to the mechanism observed for axial excitation of low-mass ions observed in collision-induced dissociation. The result is radial stability at high pressure, even after multiple remeasurement cycles. An optimized excitation radius of 12.5% of the cell radius yields maximum remeasurement efficiency with a 500 ms relaxation delay between excitation events. Summed signal intensity decreases with increased trap potential due to the greater radial electric field and reduced axial expansion of the ion cloud and also decreases with buffer gas mass in response to greater radial scattering. 0003-2700/96/0368-1321$12.00/0

© 1996 American Chemical Society

The inherent analytical advantages of the open-ended trappedion cell geometry1-3 have been applied by several research groups over the last several years to improve FT-ICR performance.3-7 Following are the results of a fundamental comparison between the open and closed cell, specifically related to the axially dependent radial electric field. Three-dimensional motional stability is achieved in an open cell using only the dc trapping field and is applied to the remeasurement of low-mass ions as a demonstration of the fundamental difference in performance between closed and open cells. Closed trapped-ion cells may be designed to exhibit reduced radial electric field as in the elongated,8,9 screened,10-12 or compensated cells.13,14 The open cell is ideal for external source applications due to greater particle injection efficiency15 and resistance to electrode charging and contamination.3 However, the open cell suffers from increased radial electric field magni(1) Gabrielse, G.; Haarsma, L.; Rolston, S. L. Int. J. Mass Spectrom. Ion Processes 1989, 88, 319-332. (2) Jhe, W.; Phillips, D.; Haarsma, L.; Tan, J.; Gabrielse, G. Phys. Scr. 1992, 46, 264-267. (3) Beu, S. C.; Laude, D. A., Jr. Int. J. Mass Spectrom. Ion Processes 1992, 112, 215-230. (4) Beu, S. C.; Senko, M. W.; Quinn, J. P.; McLafferty, F. W. J. Am. Soc. Mass Spectrom. 1993, 4, 190-192. (5) Little, D. P.; Speir, J. P.; Senko, M. W.; O’Connor, P. B.; McLafferty, F. W. Anal. Chem. 1994, 66, 2809-2815. (6) Chu, I.-H.; Zhang, H.; Dearden, D. V. J. Am. Chem. Soc. 1993, 115, 57365744. (7) Marto, J. A.; Guan, S.; Marshall, A. G. Rapid Commun. Mass Spectrom. 1994, 8, 615-620. (8) Hunter, R. L.; Sherman, M. G.; McIver, R. T., Jr. Int. J. Mass Spectrom. Ion Phys. 1983, 50, 259-273. (9) Grosshans, P. B.; Wang, M.; Marshall, A. G. Proceedings of the 36th ASMS Conference Mass Spectrometry and Allied Topics, San Francisco, CA, June 5-10, 1988; pp 592-593. (10) Wang, M.; Marshall, A. G. Anal. Chem. 1989, 61, 1288-1293. (11) Castoro, J. A.; Koster, C.; Wilkins, C. L. Anal. Chem. 1993, 65, 784-788. (12) Zeller, L. C.; Kennady, J. M.; Kentta¨maa, H.; Campana, J. E. Anal. Chem. 1993, 65, 2116-2118. (13) Naito, Y.; Fujiwara, M.; Inoue, M. Int. J. Mass Spectrom. Ion Processes 1992, 120, 179-192. (14) Hanson, C. D.; Castro, M. E.; Kerley, E. L.; Russell, D. H. Anal. Chem. 1990, 62, 520-526. (15) Gabrielse, G.; Haarsma, L.; Rolston, S. L. Int. J. Mass Spectrom. Ion Processes 1989, 88, 319-332.

Analytical Chemistry, Vol. 68, No. 8, April 15, 1996 1321

tude compared to a closed cell of the same aspect ratio3,16 because the trap electrodes are positioned closer together. Although the open cell exhibits an approximate quadrupolar potential17 at the center of the cell (where magnetic and electric inhomogeneity is minimal18 ) that can be solved analytically using simple equations,19 a more complex function is needed to describe the potential at increased axial displacement. (By convention, a coordinate system is designated in which ions are constrained radially in the xy-dimension by the magnetic field Lorentz force and axially in the z-dimension in a potential well formed by a static voltage applied to the trap electrodes.) Due to the collinear electrode geometry, ions may oscillate axially past the physical dimensions of the trap electrodes into regions where the radial electric field becomes inward-directed. It is then possible, by means of motional averaging in the axial position-dependent radial electric field, to reaxialize ions to reduce magnetron radius, increase residence time in the cell, and thereby increase effective cyclotron radius. In principle, such a capability would produce some of the same analytical benefits as quadrupolar axialization20-24 or rf-only25,26 or combined mode27,28 trapping, all of which are used to maintain radial stability. In the first demonstration of three-dimensional ion motional stability using only the dc trapping field, ions are maintained in stable radial trajectories along the z-axis as long as they attain sufficient axial amplitude to penetrate the regions of the cell where the radial electric field reorientation occurs. This phenomenon has recently been demonstrated to compensate for the radial electric field-induced downward shift in cyclotron frequency by increasing the ion z-axis amplitude through application of a suspended trapping event or implementation of axial excitation.29 In this demonstration, the ability to remeasure low-mass ions hundreds of times under high-pressure conditions is further evidence of the analytical advantages of the open cell. MOTIONAL STABILITY IN TIME- AND POSITION-VARYING FIELDS Motional stability induced by time-varying fields is the basis for the operation of the Paul trap. In ICR, an alternative to the reduction in the radial electric field by cell geometry alteration (as with the screened cell or elongated cell) is quadrupolar(16) Vartanian, V. H.; Laude, D. A. Int. J. Mass Spectrom. Ion Processes 1995, 141, 189-200. (17) Brown, L. S.; Gabrielse, G. Rev. Mod. Phys. 1986, 58, 233-311. (18) Marshall, A. G.; Comisarow, M. B.; Parisod, G. J. Chem. Phys. 1979, 71, 4344-4444. (19) Heinzen, D. J.; Bollinger, J. J.; Moore, F. L.; Itano, W. M.; Wineland, D. J. Phys. Rev. Lett. 1991, 66, 2080-2083. (20) Wineland, D. J.; Dehmelt, H. G. Int. J. Mass Spectrom. Ion Phys. 1974, 16, 338-442. (21) Becker, S.; Bollen, G.; Kern, F.; Kluge, H.-J.; Moore, R. B.; Savard, G.; Schweikard, L.; Stolzenberg, H.; ISOLDE Collaboration. Int. J. Mass Spectrom. Ion Processes 1990, 99, 53-77. (22) Savard, G.; Becker, S.; Bollen, G.; Kluge, H.-J.; Moore, R. B.; Otto, T.; Schweikard, L.; Stolzenberg, H.; Weiss, U. Phys. Lett. A 1991, 158, 247252. (23) Schweikard, L.; Guan, S.; Marshall, A. G. Int. J. Mass Spectrom. Ion Processes 1992, 120, 71-83. (24) Bruce, J. E.; Anderson, G. A.; Hofstadler, S. A.; Van Orden, S. L.; Sherman, M. S.; Rockwood, A. L.; Smith, R. D. Rapid. Commun. Mass Spectrom. 1993, 7, 914-919. (25) March, R. E.; Hughes, R. J. Quadrupole Storage Mass Spectrometry; Wiley: New York, 1989. (26) Rempel, D. L.; Gross, M. L. J. Am. Soc. Mass Spectrom. 1992, 3, 590-594. (27) Major, F. G.; Dehmelt, H. G. Phys. Rev. 1968, 170, 91-107. (28) Li, G.-Z.; Werth, G. Phys. Scr. 1992, 46, 587-592. (29) Vartanian, V. H.; Laude, D. A. J. Am. Soc. Mass Spectrom., in preparation.

1322

Analytical Chemistry, Vol. 68, No. 8, April 15, 1996

assisted buffer gas cooling for control of magnetron motion. Motional stability can also be induced by operating a trap in an rf-only mode25,26 or as a combined trap27,28 having both magnetic and radio frequency electric fields. In both cases, ions are confined in stable trajectories by time-varying electric fields with no net radial electric field. Motional stability can also be induced by position-varying fields. Motionally induced radial stability in an electrostatic ion guide has recently been proposed by Guan and Marshall to efficiently introduce externally generated ions into a trapped-ion cell.30 By alternating the sign of the potentials applied to individual electrodes in a stacked-ring assembly and introducing ions axially through the guide, ions oscillate as if in a static three-dimensional well, or “pseudopotential”, yet remain focused in the central region, where the field is weakest. The open cell, by virtue of the electrostatic axial positiondependent radial electric field, also has the intrinsic ability to axialize ions without the external application of any time-varying fields. Given sufficient axial amplitude, ions experience a pseudopotential which localizes them radially along the z-axis of the cell. The following is a demonstration of high-efficiency remeasurement of low-mass ions in a Penning trap using only the dc trapping field to induce radial stability. EXPERIMENTAL SECTION Remeasurement experiments were performed in a 75 mm long, 25 mm i.d. open-ended cell appended to the analyzer side of a dual-cell assembly within a 3.0 T, 150 mm warm bore superconducting magnet controlled by a Nicolet (now Finnigan FT/MS, Madison, WI) 1280 data station executing FTMS software version 6.0.1. The cell was constructed of 0.76 mm (0.030 in.) thick oxygen-free high-conductivity copper with the excitation/detection region 23 mm long, the trap electrodes 17 mm long, and the compensation electrodes 4 mm long with 2 mm air gaps separating all electrode segments. The cell was located about 5 mm from the conductance limit separating the source from the analyzer side of the chambers. Helium buffer gas introduced to the analyzer side of the vacuum chamber through a precision leak valve (Varian 951-5106, Lexington, MA) was maintained at 2 × 10-4 Torr (pressures corrected according to ion gauge sensitivity factor31). Neutral benzene was maintained at a static pressure of 6 × 10-6 Torr in the source side of the vacuum chamber, where it effused through the conductance limit to a pressure of 1 × 10-8 Torr in the analyzer side. A remeasurement pulse sequence of the form I-Dz-[ExA-Dxy]n-Q32 was applied with the sequence event designation as follows. An initial ionization event, I, generated ions with a 5 µA, 5 ms, -15 eV electron beam, and +2 V trap potential. A 1 s axialization delay, Dz, was implemented to allow collisional cooling of axial motion prior to excitation. Multiple applications (n) of the three-event cycle [Ex-A-Dxy] resulted in ion remeasurement. Excitation (Ex) was performed using a single 100 µs on-resonance pulse, and the amplitude was systematically varied to optimize the signal magnitude for various experimental conditions. Broad(30) Guan, S.; Marshall, A. G. Presented at the 43rd ASMS Conference Mass Spectrometry and Allied Topics, Atlanta, GA, May 22-26, 1995. (31) Moore, J. H.; Davis, C. C.; Coplan, M. A. Building Scientific Apparatus: A Practical Guide to Design and Construction, 2nd ed; Addison-Wesley Publishing Co.: New York, 1989; p 83. (32) Guan, Z.; Hofstadler, SA.; Laude, D. A., Jr. Anal. Chem. 1993, 65, 15881593.

band acquisition (A) was performed in direct mode over a 1 MHz bandwidth using 32K data points, sinebell apodization, padding with another 32K zeroes, and Fourier transformation to yield magnitude mode frequency domain spectra. A pressure-dependent collisional cooling delay (Dxy) of varied duration facilitated relaxation of the cyclotron motion prior to the next remeasurement cycle. After the desired number of cycles was performed, all residual charged particles were removed from the cell by applying a quench pulse (Q) to the trap electrodes. Parameters affecting remeasurement efficiency (RE) were altered, including excitation energy, relaxation time, buffer gas pressure, trap voltage, and buffer gas mass. RESULTS AND DISCUSSION Limitations of the Remeasurement Experiment. Image current detection in FT-ICR is characteristically a nondestructive technique, permitting the remeasurement of the same population of ions. Signal-to-noise ratio is increased by the recurrent application of excitation, detection, and collisional damping of the cyclotron motion and coaddition of the time domain scans to signal average prior to Fourier transformation.32-34 The principal difficulty in this technique is the inability to retain ions in the cell long enough to perform a sufficient number of remeasurements. Consecutive excitation events are performed in the presence of a radial electric field that causes continual radial diffusion of the ion cloud and prohibits relaxation to the original radial position in the cell. The growth in the magnetron radius reduces the maximum attainable cyclotron radius, reduces the ion population in the cell, and limits remeasurement efficiency. In addition, as the magnetron radius changes, magnetron frequency shifts according to the following equation:35

ω- ) Er/(rmB)

(1)

where ω- is the magnetron frequency, Er is the radial electric field at rm, rm is the magnetron radius, and B is the magnetic field strength. Consequently, the reduced cyclotron frequency (ω+) must also shift, as the total radial energy must remain constant (ωc ) ω+ + ω-).35,36 Thus, both magnetron radius growth and ion loss during the remeasurement experiment hinder the acquisition of high-resolution spectra.37 A high-pressure buffer gas used to damp the cyclotron motion for subsequent reexcitation further induces radial diffusion due to the high number of collisions.35,38-41 As the mass of the ion decreases with respect to the mass of the buffer gas, the radial scattering angle following a collision increases, and the guiding center of ion motion shifts to greater (33) Williams, E. R.; Henry, K. D.; McLafferty, F. W. J. Am. Chem. Soc. 1990, 112, 6157-6162. (34) Speir, J. P.; Gorman, G. S.; Pitsenberger, C. C.; Turner, C. A.; Wang, P. P.; Amster, I. J. Anal. Chem. 1993, 65, 1746-1752. (35) Dunbar, R. C. Int. J. Mass Spectrom. Ion Processes 1984, 56, 1-9. (36) Bollen, G.; Moore, R. B.; Savard, G.; Stolzenberg, H. J. Appl. Phys. 1990, 68, 4355-4374. (37) Campbell, V. L.; Guan, Z.; Laude, D. A., Jr. J. Am. Soc. Mass Spectrom. 1995, 6, 564-570. (38) Dunbar, R. C.; Chen, J. H.; Hays, J. D. Int. J. Mass Spectrom. Ion Processes 1984, 57, 39-56. (39) Francl, T. J.; Fukada, E. K.; McIver, R. T., Jr. Int. J. Mass Spectrom. Ion Phys. 1983, 50, 151-167. (40) Sharp, T. E.; Eyler, J. R.; Li, E. Int. J. Mass Spectrom. Ion Phys. 1972, 9, 421-439. (41) Ridge, D. P.; Beauchamp, J. L. J. Chem. Phys. 1976, 64, 2735-2746.

Figure 1. (a) Plot of radial electric field at 1 mm radial displacement from the cell center line as a function of axial position for a closed elongated cell. The radial electric field is outward-directed (positive) at every position along the z-axis. (b) Plot of radial electric field at 1 mm radial displacement from the cell center line as a function of axial position for an open elongated cell. The radial electric field is outwarddirected (positive) near the center of the cell but becomes inwarddirected (negative) at increased axial displacement.

radius following a collision.42 The result is even more rapid radial transport and ion loss. This is the principal limitation to the remeasurement experiment, and if the scattering angle is high per collision, then radial ion diffusion is accelerated.40 Because low-mass ions scatter at larger radial angles following a collision with neutrals, remeasurement has been performed using only heavy biomolecules with low radial scattering angles, with rhodamine 6G (m/z 443) being the smallest ion remeasured at high efficiency thus far.34 To limit the mass-dependent collisioninduced radial diffusion, early remeasurement experiments were performed in closed elongated cells32,41 or at low amplitude in open43 or closed cubic cells.34 Effect of the Open Cell Axially Dependent Radial Electric Field on Ion Motion. The radial electric field in all trapped-ion cells is a hindrance to high remeasurement efficiency, and experiments heretofore have relied on cell geometries having minimal radial electric field or the application of quadrupolar axialization to achieve superior remeasurement efficiency. A (42) Chen, F. F. Plasma Physics and Controlled Fusion, 2nd ed.; Plenum Press: New York, 1984; Chapter 5. (43) Campbell, V. L.; Guan, Z.; Vartanian, V. H.; Laude, D. A. Anal. Chem. 1995, 67, 420-425.

Analytical Chemistry, Vol. 68, No. 8, April 15, 1996

1323

SIMION44 plot of the radial electric field adjacent to the closed elongated cell center line as a function of z-axis position is represented in Figure 1a, indicating that the radial electric field is outward-directed (positive) at any point along the z-axis. An ion sustaining a collision with a neutral molecule undergoes magnetron destabilization anywhere along the z-axis. Because the radial electric field increases with axial displacement, it is advantageous to maintain ions in the center of a closed cell so that magnetron growth can be minimized. Ions are, therefore, commonly damped to the center of a closed cell to reduce radial diffusion. The closed cell radial electric field, which increases with axial amplitude, is in contrast to the open cell in which ions encounter reduced average radial electric field with increased axial amplitude. The axially dependent radial electric field in an open cell may be utilized to counteract collisionally induced radial diffusion. In Figure 1b, the radial electric field adjacent to the cell center line as a function of z-axis position is outward-directed (positive) near the center of the cell (z ) 0) but becomes increasingly inwarddirected (negative) at greater axial amplitude. An ion of low axial oscillation amplitude that sustains a collision with a neutral molecule near the center of the open cell scatters radially outward, resulting in magnetron destabilization. However, an ion of greater axial amplitude that sustains a collision near the perimeter of the trapping well scatters radially inward, resulting in magnetron stabilization. Thus, the effect of the open cell radial electric field depends on the ion axial amplitude. Rapid radial diffusion from the cell occurs with decreased axial amplitude, while ions undergoing large-amplitude trapping motion undergo minimal radial diffusion. Because the field orientation changes from outward to inward, the average radial electric field asymptotically approaches 0 V/mm at the maximum in well potential (but cannot become less than zero because there would be a radial repulsive force and no axial restoring force to confine ions). If the ion trajectory following a collision results in increased axial amplitude, the radial expansion of the ion cloud attributed to the radial electric field at the center of the cell can be counteracted. Collisions with background neutrals during the excitation event fortuitously provide the energy necessary for increased axial scattering of low-mass ions. Because the potential well depth is only a few volts, a collision of a low-mass ion with a neutral target molecule (e.g., helium) at high pressure imparts increased z-motion in addition to radial motion. An increase in axial amplitude compared to radial amplitude is expected because ions are not required to cross magnetic field lines in the axial dimension as they must in the radial dimension. Therefore, relatively little collision energy must be partitioned into the axial mode to increase axial amplitude. In an open cell, the increase in axial amplitude results in longer ion cell residence time compared to a closed cell, where an increase in axial amplitude actually accelerates ion loss because the radial electric field increases with axial displacement. The hypothesis for the requisite increase in z-amplitude is consistent with observations made in the collision-induced dissociation (CID) experiment. In CID, a high-pressure buffer gas such as argon is used to affect fragmentation when ions are accelerated to several kiloelectronvolts. In such experiments, ion loss has been observed due to axial excitation, not to radial

diffusion.45 Collision-mediated axial ejection is a function of pressure, excitation energy, ion-to-neutral mass ratio, and trap potential. Axial ejection increases with target neutral pressure and excitation energy due to ion cloud expansion. Similarly, axial ejection increases as the ion-to-neutral mass ratio decreases. Ion loss decreases as trap potential increases due to ion cloud compression to the center of the cell. This is in contrast to radial ion diffusion. If the CID-induced axial expansion of the ion cloud is responsible for the ability to perform remeasurement in the open cell, changes in these parameters should affect RE. As will be shown, variation in target neutral pressure, excitation energy, ion-to-neutral mass ratio, and trap voltage are consistent with the RE expected in the CID model. Control of the ion cloud axial amplitude by collisions with a low-mass buffer gas enables the remeasurement experiment to occur at relatively high efficiency. A typical open cell remeasurement spectrum for benzene molecular ion is shown in Figure 2 as a comparison of the signal-to-noise ratio enhancement from 1 measurement (a) to 10 measurements (b). A 2.5-fold increase in signal-to-noise ratio occurs (compared to the theoretical improvement of 3.2(101/2) for no ion loss. Some deterioration in spectral mass resolving power is observed from m/∆m of 2613 to 2353 (fwhm) because ions encounter either electric field inhomogeneity or variation in the space-charge environment during the experi-

(44) Dahl, D. A. SIMION 3D, Version 6.0; INEL-95/0403; Idaho National Engineering Laboratory: Idaho Falls, ID, 1995.

(45) Riegner, D. E.; Laude, D. A., Jr. Int. J. Mass Spectrom. Ion Processes 1992, 120, 103-116.

1324 Analytical Chemistry, Vol. 68, No. 8, April 15, 1996

Figure 2. Comparison open cell remeasurement spectra of benzene molecular ion using a excitation voltage of 2.5 Vp-p. A signal-to-noise ratio increase of 2.5 is observed on comparing (a) a signal-to-noise ratio of 32.1 for the spectrum from the first measurement with (b) a signal-to-noise ratio of 80.7 for the spectrum from 10 measurements. The insets indicate comparable spectral peak width.

Figure 3. (a) SIMION trajectory plot of an ion with 0.02 eV axial kinetic energy in a 50 mm diameter, aspect ratio 2, closed elongated cell with 1 V applied trap potential illustrating the effect of the outward-directed radial electric field on ion trajectory. The effects of the magnetic field and collisions have been removed to illustrate only the effect of the radial electric field on ion trajectory. (b) SIMION trajectory plot of an ion with 0.73 eV axial kinetic energy in a 50 mm diameter, aspect ratio 2, open elongated cell in the absence of a magnetic field and collisions. The ion is trapped indefinitely due to the inward-directed radial electric field at increased axial displacement, which counteracts the effect of the outwarddirected radial electric field at the cell center.

ment as a result of ion loss. In a 25 mm diameter open cell, the radial electric field is significantly greater than that of a cell of greater diameter or length.43 Notwithstanding, the ability to perform successive remeasurements indicates that a significant number of ions can be retained in the cell over a considerable length of time, even in the presence of a high-pressure buffer gas. Open Cell SIMION-Modeled Ion Trajectories. Illustrating the position-dependent dynamic effects of the radial electric field on ion trajectories in two 50 mm diameter, aspect ratio 2, elongated closed and open cells, SIMION trajectory plots are shown in Figure 3, in which ions are given initial z-axis energies and their trajectories monitored. Each simulation is shown in the absence of a magnetic field and in a collisionless environment in order to demonstrate the effect of the radial electric field alone. The presence of a magnetic field would increase the radial stability by reducing the amplitude of radial motion, while collisions would damp the axial motion to the center of the cell. Figure 3a is a plot of a 0.02 eV ion directed along the z-axis in a closed elongated cell with 1 V applied trap potential, indicating a collision with a cell electrode due to the effect of the radial electric field. The presence of a magnetic field would greatly increase the ion residence time in the trap, but the overall drift toward the cell perimeter over time would still occur. In contrast, Figure 3b indicates that a 0.73 eV ion trapped in a 0.84 V deep well formed by a 1 V applied trap potential in an open elongated cell remains trapped indefinitely because the ion motion is dynamically stabilized. The ion trajectory assumes a characteristic saddle shape as the outward-directed radial electric field at the center of the cell repels ions away from the z-axis as ion amplitude increases, while the inward-directed radial electric field at the cell perimeter drives ions inward as ion amplitude decreases. The net effect of both motions produces stabilized

radial motion. Although the radial amplitude of the motion is greatly exaggerated without a magnetic field, the fact that ion motion can be stabilized in three dimensions in a trap using only a static trapping field is significant. The amount of radial stabilization ions experience in the open cell during encounter with the inward-directed radial electric field is a function of time spent in the stabilizing regions as a fraction of the total ion axial oscillation period. Because ions spend more time at the turnaround points in the trapping well, where axial velocity is eventually reduced to zero, ions undergoing largeamplitude trapping motion have more residence time in the cell as the stabilizing field becomes dominant compared to the destabilizing field. The same 0.73 eV ion spends 69.5% of each trapping oscillation period in the peripheral stabilization regions and only 30.5% of each period in the central destabilization region. Because there are two stabilization regions to one destabilization region per trapping oscillation, ion stability is enhanced. The ability of the radial electric field to reaxialize ions is also a function of the distance an ion penetrates into the inwarddirected radial electric field, which increases in magnitude as a function of distance (Figure 1b). Ions having insufficient axial kinetic energy to penetrate into the stability regions remain in the destabilizing region in the center of the cell, where radial diffusion eventually causes ion loss. For example, SIMIONgenerated data of ion lifetime prior to neutralization with a cell electrode as a function of z-axis kinetic energy in a 1.0 V applied trap potential (0.84 V well depth) indicate that a significant increase in cell residence time results if an ion possesses >0.59 eV axial kinetic energy. Sufficient axial penetration into the inwarddirected radial electric field allows ion reaxialization to occur. The relative fraction of time spent in the open cell radially stabilizing regions during one trapping oscillation period compared Analytical Chemistry, Vol. 68, No. 8, April 15, 1996

1325

Table 1 excitation energy (V)

signal intensity (arb units)

4.380 3.900 3.480 3.100

6.152 11.197 12.995 12.159

Table 2 relaxation delay (s)

signal intensity (arb units)

relaxation delay (s)

signal intensity (arb units)

0.010 0.050 0.100 0.200

5.7749 8.8312 10.801 14.449

0.500 1.000 1.500 2.000

15.722 15.372 13.606 11.865

to the time spent in the destabilizing region is also a function of ion z-axis kinetic energy. For example, the fraction of a single trapping oscillation period spent in the stabilizing region with 1.0 V applied trap potential (0.84 V well depth) indicates that, when ions spend >50% of the trapping oscillation period in the regions where the radial electric field is inward-directed, residence time is indefinite. Ions with 0.83 eV are too energetic to be retained in the well. Remeasurement Efficiency of Low-Mass Ions in an Open Cell. To perform repeated measurements of the same ion population with minimal ion loss requires optimization of all the parameters that affect RE. Excitation energy affects RE as ions traverse different regions of the cell as a function of excitation radius so that ions encounter electric field variations that affect relaxation. The duration of collisional relaxation also determines the spatial extent of the ion cloud prior to the next excitation event. Therefore, relaxation duration must be chosen to promote sufficient axial and radial cooling prior to reexcitation. Additionally, trap potential affects RE, as the radial electric field environment of the cell and the spatial characteristics of the ion cloud are a function of trap potential. The axial restoring force of the ion may also prevent ions from penetrating the stabilizing regions of the open cell. The buffer gas used also affects the collisional relaxation efficiency, because increased radial scattering occurs as the target mass increases. Signal intensity as a function of excitation energy is shown in Table 1 for eight measurements, indicating that maximum signal intensity occurs at 3.48 VP-P or at 12.5% of the cell radius (1.6 mm in a 12.5 mm radius cell) based on the quadrupolar approximation.46 The reduced effective cyclotron radius is an indication that the ion cloud has attained substantial magnetron radius. Subsequent excitation increases the axial amplitude of the ion cloud due to collisions and maintains ions in the cell throughout the remeasurement experiment. At lower excitation energies, the signal-to-noise ratio is reduced due to increased radial diffusion and ion loss because lower energy collisions do not partition sufficient energy into the axial mode to permit ion radial stabilization. Increased excitation energies result in excessive axial or radial ion scattering, also resulting in ion loss. Signal intensity versus relaxation delay between successive excitation events in a remeasurement experiment involving eight (46) Grosshans, P. B.; Marshall, A. G. Anal. Chem. 1991, 63, 2057-2061.

1326

Analytical Chemistry, Vol. 68, No. 8, April 15, 1996

Figure 4. (a) Plot of signal intensity versus number of excite-τexcite cycles prior to excitation/detection at the optimized excitation radius of 1.6 mm and 500 ms delay between excitations. The data indicate a final intensity to initial intensity ratio (at second excitation) of 0.7632. For the 20 excitations performed, RE was 98.59%. (b) Plot of signal intensity versus excitation voltage for 20 excite-τ-excite cycles at three different trap potentials (4, 7, 9 V), indicating reduced signal intensity as trap potential increases.

excite-τ-excite cycles followed by detection is shown in Table 2, indicating a maximum signal intensity using a 500 ms relaxation delay. At shorter relaxation delays, ion intensity decreases because collisions have not adequately damped the axial and radial motions back to the center of the cell prior to the next remeasurement cycle. Longer delays allow ions to radially diffuse because they become localized to the center of the cell by collisional damping. A plot of signal intensity versus the number of excite-τ-excite cycles prior to detection is shown in Figure 4a at the optimized radius of 1.6 mm and 500 ms delay between excitations. The data indicate a final abundance-to-initial scan abundance (starting from second excitation) ratio of 0.7632. For the 20 excitations performed, RE is 98.59%, determined from the following equation, where n ) 20. The increase in the signal intensity from the first

last scan abundance (first scan abundance)

1/n-1

(2)

to the second scan is due to increased collisional relaxation along the z-axis. The effect of trap voltage on RE is shown in Figure 4b as a plot of signal intensity versus excitation energy for 20 excite-τexcite cycles at three different trap potentials. The trend is consistent with reduced axial ejection observed in the CID experiment at higher trap potentials. The increased radial electric

Table 3 signal intensity pressure (Torr)

Ar buffer He buffer gas gas

7.200 × 10-6 0.65976 1.080 × 10-5 1.5211 1.360 × 10-5 2.250 × 10-5 3.0238 2.720 × 10-5

1.2338 3.5661

signal intensity pressure (Torr) 3.600 × 10-5 5.440 × 10-5 8.160 × 10-5 1.777 × 10-4 2.720 × 10-4

Ar buffer He buffer gas gas 2.5836 6.0409 6.6454 9.1830 11.9060

averaging are shown by circles and represent the theoretically derived performance expected for 100% RE. Data represented by squares represent the experimentally derived performance of 99.59% RE for 300 measurement cycles under optimized conditions of 0.500 s relaxation delay and excitation to 12.5% of the cell radius at 4 V applied trap potential and a pressure of 2 × 10-4 Torr of helium buffer gas. As the number of ions decreases with succeeding measurements, the decrease in signal intensity becomes more apparent compared to remeasurement at 100% efficiency. The summed signal intensity was calculated using the following equation, where S is the intensity of the first measurement, n is the total number of measurements and  is the RE.32 n

∑

signal intensity ) S

i-1

(3)

i)1

The primary motivation of the remeasurement experiment is signal-to-noise enhancement, and Figure 5b illustrates the signalto-noise ratio of benzene versus number of measurement cycles. The data shown by circles represent the signal-to-noise ratio enhancement for theoretical unit RE (no ion loss), in which the signal increases as n1/2 measurements. The data shown by squares represent experimental 99.59% RE following 300 measurements. The signal-to-noise ratio improvement drops off rapidly and actually begins to decrease as the number of ions is reduced and noise becomes a more significant factor.

Figure 5. (a) Plot of summed signal intensity for benzene versus number of measurement cycles for theoretically derived performance expected for 100.0% RE (O) and experimentally derived performance of 99.59% RE (0) following 300 measurements with optimal experimental parameters. (b) Plot of signal-to-noise ratio of benzene versus number of measurement cycles for theoretical 100.0% RE (O) and experimental 99.59% RE (0) following 300 measurements with optimal experimental parameters.

field that occurs with trap potential accelerates radial diffusion, reducing RE. Furthermore, increased space-charge results in greater radial expansion of the ion cloud. In addition, as the axial restoring force increases with trap potential, reduced signal intensity occurs due to the inability of the ion cloud to penetrate sufficiently far into the radial stabilizing regions, where reaxialization occurs. The effect of buffer gas mass on signal intensity for CCl4 following 20 remeasurements is illustrated in Table 3 as a function of helium and argon pressure. Increased signal intensity occurs when helium is used as a buffer gas as compared to argon because increased radial scattering occurs as buffer gas mass increases. No signal is observed after the argon pressure is increased to about 4.5 × 10-4 Torr. A plot of summed signal intensity versus the number of measurement cycles is shown for benzene in Figure 5a. Data representing the signal-to-noise ratio enhancement due to signal

CONCLUSION The open cell radial electric field orientation is distinguished from that of closed cells because it becomes inward-directed at increased axial displacement within the trapping well. Radial and axial stability are achieved by dynamically varying the ion position in a position-dependent dc electric field. As a demonstration of three-dimensional motional stability using only a dc trapping field, ions are maintained in stable radial trajectories along the z-axis, as long as they are able to penetrate sufficiently far into the regions of the open cell where the radial electric field orientation shift occurs. The increase in axial amplitude caused by collisional scattering with a high-pressure, low-mass buffer gas allows ions to encounter the inward-directed radial electric field at the trapping perimeter, which results in ion refocusing along the cell z-axis. The open cell is thus used to observe remeasurement of low-mass ions, making use of increased axial scattering consistent with the model for collision-mediated axial ejection observed in CID experiments. ACKNOWLEDGMENT This work was supported by the Welch Foundation (F-1138), the National Institutes of Health, the Texas Advanced Research Program, and the National Science Foundation (CHE9013384 and CHE9057097).

Received for review August 4, 1995. Accepted January 30, 1996.X AC950787P X

Abstract published in Advance ACS Abstracts, March 1, 1996.

Analytical Chemistry, Vol. 68, No. 8, April 15, 1996

1327