Rectilinear Ion Trap Mass Spectrometer with Atmospheric Pressure

Qinghao Wu , Yuan Tian , Ailin Li , Derek Andrews , Aaron R. Hawkins , Daniel E. Austin. Journal of The American Society for Mass Spectrometry 2017 28...
0 downloads 0 Views 318KB Size
Download PDF
Anal. Chem. 2006, 78, 718-725

Rectilinear Ion Trap Mass Spectrometer with Atmospheric Pressure Interface and Electrospray Ionization Source Qingyu Song,† Sameer Kothari,† Michael A. Senko,‡ Jae C. Schwartz,‡ Jonathan W. Amy,† George C. Stafford,† R. Graham Cooks,*,† and Zheng Ouyang*,†

Department of Chemistry, Purdue University, West Lafayette, Indiana 47907, and Mass Spectrometry Division, Thermo Electron Corp., 355 River Oaks Parkway, San Jose, California 95134

A rectilinear ion trap (RIT) mass analyzer was incorporated into a mass spectrometer fitted with an electrospray ionization source and an atmospheric pressure interface. The RIT mass spectrometer, which was assembled in two different configurations, was used for the study of biological compounds, for which performance data are given. A variety of techniques, including the use of a balanced rf, elevated background gas pressure, automatic gain control, and resonance ejection waveforms with dynamically adjusted amplitude, were applied to enhance performance. The capabilities of the instrument were characterized using proteins, peptides, and pharmaceutical drugs. Unit resolution and an accuracy of better than m/z 0.2 was achieved for mass-to-charge (m/z) ratios up to 2000 Th at a scan rate of ∼3000 amu/(charge‚s) while reduced scan rates gave greater resolution and peak widths of less than m/z 0.5 over the same range. The mass discrimination in trapping externally generated ions was characterized over the range m/z 190-2000 and an optimized low mass cutoff value of m/z 120-140 was found to give equal trapping efficiencies over the entire range. The radial detection efficiency was measured as a function of m/z ratio and found to rise from 35% at low m/z values to more than 90% for ions of m/z 1800. The way in which the ion trapping capacity depends on the dc trapping potential was investigated by measuring the mass shift due to space charge effects, and it was shown that low trapping potentials minimize space charge effects by increasing the useful volume of the device. The collision-induced dissociation (CID) capabilities of the RIT instrument were evaluated by measuring isolation efficiency as a function of mass resolution as well as measuring peptide CID efficiencies. Overall CID efficiencies of more than 60% were easily reached, while isolation of an ion with unit resolution at m/z 524 was achieved with high rejection (>95%) of the adjacent ions. The overall analytical capabilities of the ESI-RIT instrument were demonstrated with the analysis of a mixture of pharmaceutical compounds using multiple-stage mass spectrometry. 718 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

The rectilinear ion trap (RIT)1 is a simplified version of the linear ion trap (LIT),2 a modern high-performance version of the radio frequency (rf) trapping mass spectrometer originally introduced by Paul.3 An RIT mass analyzer, constructed in the most basic configuration,4 was characterized previously in an existing mass spectrometer using volatile organic compounds ionized by electron impact ionization, showing increased ion trapping capacity and tandem mass spectrometric capabilities. Methods of optimizing the mass resolution were developed, and unit or better mass resolution was demonstrated at a scan speed above 15 000 amu/ (charge‚s).4 Given the demonstrated performance of this mass analyzer, it is of great interest to continue the development of mass spectrometers using RIT mass analyzers and to evaluate their performance for the analysis of biologically interesting compounds. Given adequate performance characteristics in applications in the life science, the simple configuration and ease of fabrication of the RIT would make this device a good candidate for future development of ion trap mass spectrometers. This is especially the case for instruments with arrays of analyzers as in high-throughput mass spectrometers,5-7 where fabrication of the array of analyzers is expected to become a challenge. The criteria used to evaluate complete mass spectrometer systems are usually significantly different from those used to evaluate the mass analyzer itself.4 For mass spectrometers used in biological studies, an electrospray ionization (ESI) or an atmospheric pressure chemical ionization source is usually used, and an atmospheric pressure interface needs to be developed to couple the ionization source to the mass analyzer. To achieve the * Corresponding authors. E-mail: [email protected]. Tel: (765) 494-5262. Fax: (765) 494-9421. E-mail: [email protected]. Tel: (765) 496-1539. Fax: (765) 494-9421. † Purdue University. ‡ Thermo Electron Corp. (1) Ouyang, Z.; Cooks, R. G. U.S. Patent 6,838,666, 2005. (2) Douglas D. J.; Frank Aaron J.; Dunmin, M. Mass Spectrom. Rev. 2005, 24, 1-29. (3) March, R. E., Todd, J. F. J., Eds. Practical Aspects of Ion Trap Mass Spectrometry, Vol. I. Fundamentals of Ion Trap Mass Spectrometry; CRC Press: Boca Raton, FL, 1995. (4) Ouyang, Z.; Wu, G.; Song, Y.; Li, H.; Plass, W. R.; Cooks, R. G. Anal. Chem. 2004, 76, 4595-4605. (5) Misharin, A. S.; Laughlin, B. C.; Vilkov, A.; Takats, Z.; Ouyang, Z.; Cooks, R. G. Anal. Chem. 2005, 77, 459-470. (6) Tabert, A. M.; Griep-Raming, J.; Guymon, A. J.; Cooks, R. G. Anal. Chem. 2003, 75, 5656-5664. (7) Tabert, A. M.; Misharin, A. S.; Cooks, R. G. Analyst 2004, 129, 323-330. 10.1021/ac0512709 CCC: $33.50

© 2006 American Chemical Society Published on Web 12/27/2005

Figure 1. Configuration of the rectilinear ion trap with a sealed trapping volume. Half-distance between x electrodes x0 ) 5.0 mm, half-distance between y electrodes y0 ) 4.0 mm, length in z direction z0 ) 50.0 mm, length of slits 29.0 mm, and width 1.0 mm.

best performance using mass analyzers such as linear ion traps, especially for applications where the range of analyte concentrations is large and the m/z range of interest is wide, a variety of techniques and methods have been developed.8,9 Some of these technologies, which are detailed below, were adapted from conventional Paul (3D) ion trap instruments, while others were developed specifically for traps with 2D trapping fields and elongated configurations. As in 3D ion trap instruments, buffer gas is introduced into a sealed volume within the mass analyzer of 2D ion trap instruments to allow a significantly higher pressure inside the trap8,10 in order to increase the number of collisions during the trapping, cooling, and CID periods.11 To increase the ion trapping efficiency, instead of using a single-phase rf as is done for 3D instruments, doublephase rf inputs with balanced amplitudes are usually applied on the rf electrode pairs in 2D instruments.8,9 Moreover, the waveforms used for ion isolation and excitation have to be coupled with the high-voltage rf8,9,12 requiring more complicated rf circuitry. Linear ion traps are also known to have less discrimination, in comparison with 3D traps, in the trapping efficiency for ions with different m/z values, due to the fact that the ions are injected in a direction that is orthogonal to the rf field. A single low mass cutoff (LMCO) operating value can be found that maximizes the trapping efficiency for ions covering a large m/z range.8 While linear ion traps have higher trapping capacities,8,9 their space charge limits8,13,14 remain to be characterized8 and techniques15,16 like automatic gain control (AGC) can be implemented to keep the number of ions trapped constant for optimum performance. (8) Schwartz, J. C.; Senko, M. W.; Syka, J. E. P. J. Am. Soc. Mass Spectrom. 2002, 13, 659-669. (9) Hager, J. W. Rapid Commun. Mass Spectrom. 2002, 16, 512-526. (10) Stafford, G. C.; Kelley, P. E.; Syka, J. E. P.; Reynolds, W. E.; Todd, J. F. J. Int. J. Mass Spectrom. Ion Processes 1984, 60, 85-98. (11) Vachet, R. W.; Glish, G. L. J. Am. Soc. Mass Spectrom. 1996, 7, 11941202. (12) Campbell, J. M. Rapid Commun. Mass Spectrom. 1998, 12, 1463-1474. (13) Schwartz, J. C. 9th Sanibel Conference on Mass Spectrometry, Sanibel Island, FL, 1997.

Figure 2. Schematic diagrams of two configurations of the ESIRIT instrument (a) ESI-RIT I adapted from an older 3D ion trap mass spectrometer and (b) ESI-RIT II in which an identical RIT analyzer is fitted into a commercial LTQ instrument. A, heated capillary; B, tube lens; C, square quadrupole ion guide; D, intermultipole lens; E, octapole ion guide; F, rectilinear ion trap; G, electron multipliers; H, rough pump; I, turbomolecular pump; J, split flow turbopump; K, dynode-multiplier assembly; L, buffer gas inlet.

Figure 3. Mass spectra recorded using the RIT in configuration I for (a) cytochrome c and (b) myoglobin/MRFA mixture at a scan rate of 5555 amu/(charge‚s).

The RIT is a simplified version of linear ion trap, and the techniques described above are expected to be applicable to this device and hence to enhance the analytical performance of mass spectrometers that use an RIT as the mass analyzer. In this work, an RIT was designed with a semisealed collision gas chamber (14) March, R. E.; Hughes, R. J. Quadrupole Storage Mass Spectrometry; John Wiley and Sons: New York, 1989.

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

719

Figure 4. Mass spectra for calibration mixture recorded using the RIT in configuration II, showing unit resolution up to m/z 2000. Scan rate 2778 amu/(charge‚s).

and it was used to construct a mass spectrometer fitted with an ESI source and an atmospheric pressure interface. In addition to the use of an elevated internal buffer gas pressure in the sealed RIT, the ESI-RIT instrument developed in this work employs a balanced double-phase rf potential, waveforms with dynamically adjustable amplitudes, and AGC.15 The performance of the instrument was characterized with emphasis on the aspects discussed above, using biological samples such as peptides and proteins. EXPERIMENTAL SECTION The RIT analyzer used in this work has a similar geometry (Figure 1) to one described previously.4 However, instead of using ceramic holders for alignment of the electrodes, four commercially available ceramic rods (Advanced Ceramics Co., Cleveland, OH) with a diameter of 4.775 mm were used to align the rf electrodes and to seal the gaps between them. Four flat x, y rf electrodes were machined from stainless steel with a tolerance of 0.050 mm. The assembly consisting of the rf electrodes and ceramic rods was held together by the compression forces applied through eight 4-40 machine screws located on two polycarbonate holders. Two flat z electrodes, made from stainless steel plates of 0.5-mm thickness and each with a center hole of 3.0-mm diameter, were mounted on the polycarbonate holders. The length of the RIT is 50.0 mm, and the half distance between the rf electrodes is 5.0 mm for the x pair and 4.0 mm for the y pair. A slit of 1.0-mm width and 29.0-mm length was center-located in each x electrode. This optimized RIT geometry has been demonstrated to provide an optimized electric field and mass resolution.4 This RIT was tested on two instrument platforms, a home-built ESI-RIT I (Figure 2a) mass spectrometer at Purdue University and a modified LTQ mass spectrometer ESI-RIT II (Figure 2b) at Thermo Electron Corp. (San Jose, CA). We found no significant (15) Stafford, G. C.; Taylor, D. M.; Bradshaw, S. C.; Syka, J. E. P., Proceedings of the 35th ASMS Conference on Mass Spectrometry and Allied Topics, Denver, CO, 1987. (16) Franzen, J.; Brekenfeld, A. U.S. Patent 6,600,154, 2003.

720 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

difference in any of the performance characteristics reported on in this paper for these two configurations. For the first configuration, ESI-RIT I, the manifold, pumping system, and atmospheric interface of a TSQ 7000 (Thermo Electron Corp.) were modified and used with an additional octapole ion guide from a TSQ 700 (Thermo Electron Corp.) atmospheric pressure interface, which was added between the square quadrupole and the RIT. Two K&M TX7505 electron multipliers (K and M Electronics, Inc., Springfield, MA) were used to detect the ions ejected from the RIT. An rf coil was constructed to apply balanced two-phase rf voltages to the x and y electrode pairs. Modified LCQ Duo electronics and ion trap control language (ITCL) (Thermo Electron Corp.) programs were used to control ESI-RIT I. For control of the second configuration, ESI-RIT II, the original linear trap mass analyzer of an LTQ (Thermo Electron Corp.), was replaced by the same RIT as used for ESI-RIT I and the ITCL programs were modified to operate the LTQ fitted with the RIT (Figure 2b). In both instruments, helium buffer gas was directly introduced into the RIT through a hole in one of the polycarbonate holders to achieve a higher pressure inside the RIT. During the measurement and tuning of the RIT inside pressure, a MKS 925C Micro Pirani pressure transducer (MKS) was connected to another hole in the other RIT holder using 1.6-mm-diameter Teflon tubing. An 18fold pressure difference between the inside and outside of the RIT was maintained with the semisealed configuration. An outsideRIT pressure of 5.0 × 10-5 Torr was used in all experiments. An rf frequency of 825 kHz was used for ESI-RIT I and 1030 kHz was used for ESI-RIT II. Waveforms were coupled to the rf applied to the x electrodes to allow ion isolation, excitation, and resonance ejection.17 The q values used in this study were 0.3 for CID and 0.83 for both isolation and resonance ejection. A 100-V dc voltage was applied to the z electrodes during the trapping period of all experiments except for the experiment in which dc potential effects were investigated. Almost identical results were obtained (17) Louris, J. N.; Taylor, D. M. U.S. Patent 5,324,939, 1993.

in the characterization experiments reported in this paper when the RIT was inserted into either instrumental configuration. RESULTS AND DISCUSSION Various samples, including proteins, peptides, and a calibration mixture, were used to test the ESI-RIT instrument. Mass spectra were recorded using the mass-selective instability scan with resonance ejection at a nonlinear resonance point q ) 0.83. The mass spectrum for cytochrome c (MW 12 360, 5 pmol/µL in methanol/water, 1:1 v/v with 1% acetic acid) recorded using configuration I is shown in Figure 3a. Charge states from +8 to +19 are observed in the range m/z 600-1600. The distribution of charge states is similar to that seen in spectra acquired using other instruments under similar conditions.18 In another experiment, spectra were acquired for a mixture of apomyoglobin (MW 16 951, 5 pmol/µL) and the peptide MRFA (MW 524, 20 pmol/ µL) in methanol/water solution (1:1 v/v with 0.5% acetic acid). Charge states, +1 to +2 for MRFA and +9 to +26 for apomyoglobin, appear in the range m/z 200-2000, and a characteristic charge-state distribution is observed for apomyoglobin (Figure 3b). Mass Resolution and Accuracy. The mass resolution achieved using the RIT was characterized using a calibration solution that contains 500 ng/µL Ultramark 1621 (PCR, Inc., Gainesville, FL),19 6 ng/µL MRFA, and 40 ng/µL caffeine in 50% acetonitrile, 25% water, 25% methanol, and 0.1% acetic acid. A spectrum recorded using an AGC target of 5000 for total ion current and a scan speed of 2778 amu/(charge‚s) is shown in Figure 4. The amplitude of the dipolar auxiliary ac potential(305.3 kHz) was ramped from 2.25 to 29.6 V during the rf scan to facilitate efficient ejection of ions20 covering an m/z range from 150 to 2000. Peaks with fwhm less than 0.7 Th were obtained throughout the available m/z range of 2000, and a resolution of better than 2700 was achieved for the peak m/z 1922. For comparison, spectra were collected for the same calibration solution using an LTQ and an LCQ operated at the same scan rate using the same q-ejection value (0.83) and showed peak fwhm values of less than m/z 0.3 and 0.5 for the LTQ and LCQ, respectively. The mass resolution and accuracy for the selective ion monitoring mode was also characterized: five peaks at m/z 195.1, 524.3, 1222.0, 1522.0, and 1822.0 were used to cover the range from m/z 190 to 2000. The ions with each of these mass/charge values were first isolated using a window of 40 Th and then scanned out of the RIT in order to record the spectrum. The auxiliary ac amplitude was optimized to enhance the resolution. An AGC target of 5000 was used at a compromise value between higher values that give peak broadening due to the space charge effects and lower values that give lower signal-to-noise ratios. A resolution similar to that achieved in full scan spectra was obtained in the selective ion monitoring scan mode, and peak widths were found to be almost independent of mass-to-charge ratios. A deviation of less than m/z 0.15 occurred throughout the range up to m/z 2000, which is very similar to the performance of the 2D quadruple ion trap with hyperbolic electrodes in a LTQ

instrument although the scan speed used was twice as great.8 A similar mass resolution was achieved with the previous version of the RIT at a considerably higher scan rate with a higher rf frequency.4 The configuration of the current RIT (as used in both configurations I and II) provides for ease of fabrication and assembly as well as providing a better seal of the buffer gas; however, the overall mechanical tolerance for the device is estimated to be 100 µm, which is worse than that of the previous version.4 Scans with lower scan speeds (“zoom scans”) have been used to enhance the mass resolution of ion trap instruments.8 A higher resolution is usually expected when using a reduced scan rate due to the increased number of increments of the rf voltage in a given mass range and the increased time allowed for ions with adjacent m/z values to be ejected at the threshold of their instability.21-23 The effect of the scan speed on mass resolution was characterized. Spectra for caffeine (molecular mass 194.2) were collected at various scan rates from 278 to 2778 amu/ (charge‚s), and the peak width plotted as a function of the scan

(18) Chowdhury, S. K.; Katta V.; Chait, B. T. J. Am. Chem. Soc. 1990, 112, 90129013. (19) Moini, M. Rapid Commun. Mass Spectrom. 1994, 8, 711-714. (20) Dobson, G.; Murrell, J.; Despeyroux, D.; Wind, F.; Tabet, J. C. J. Mass Spectrom. 2004, 39, 1295-1304.

(21) Schwartz, J. C.; Syka, J. E. P.; Jardine, I. J. Am. Soc. Mass Spectrom. 1991, 2, 198-204. (22) Williams, J. D.; Cox, K. A.; Cooks, R. G.; Kaiser, R. E.; Schwartz, J. C. J. Am. Soc. Mass Spectrom. 1991, 5, 327. (23) Makarov, A. A. Anal. Chem. 1996, 68, 4257-4263.

Figure 5. Full width at half-maximum (fwhm) of caffeine peak (m/z 195.09) recorded using the RIT mass spectrometer (configuration II) as a function of scan rate.

Figure 6. Relative trapping efficiency as a function of LMCO during ion injection. (Experiments performed using configuration II.)

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

721

rate is shown in Figure 5. The mass resolution increases as the scan rate decreases to 333 amu/(charge‚s), where it is believed that mechanical tolerances dominate and the effect of scan rate on mass resolution is overwhelmed by these effects. A fwhm as narrow as m/z 0.15 was achieved for protonated caffeine at a scan rate of 333 amu/(charge‚s) using the current version of the RIT. Mass Discrimination and Trapping/Detection Efficiencies. Two of the advantages of using a 2D linear ion trap instead of a 3D ion trap are the significant improvement in trapping efficiency and the smaller mass discrimination for trapping of externally injected ions.2,8 The mass discrimination in a 3D trap is a result of the alternating rf field, which makes ions unstable (in the absence of effective collisions) over much of its phase.24,25 This problem is especially serious when ions covering a wide range of m/z values are to be trapped and analyzed. Multiple rf amplitudes are used during ion injection into 3D traps to optimize the trapping conditions for ions of different m/z values.10 In linear ion traps, the ions are injected in a direction that is orthogonal to the rf field and it has been demonstrated that ions covering the m/z range 190-2000 can be trapped using a single rf amplitude with minimum mass discrimination.8 A similar feature is expected if a balanced double-phase rf is used and ions are injected along the z axis. Experiments were conducted using configuration II to measure the trapping efficiency for ions in the range from m/z 190 to 2000 at various rf amplitudes, as expressed by the LMCO. The helium buffer gas, which was introduced directly into the RIT, was leaked into the manifold through the slits in the x electrodes and the holes in the z electrodes, resulting in a manifold pressure of 3 × 10-5 Torr. Ions from the LTQ calibration mixture were injected into the RIT using an ionization time of 2.0 ms for each LMCO value, and the abundances of the monitored ions were recorded. The normalized abundances are plotted as a function of the LMCO (Figure 6). For LMCOs lower than m/z 120, the trapping efficiency increases but reaches its maximum at different m/z values for ions in the different m/z ranges. To optimize the trapping efficiency for all ions between m/z 190 and 2000, a LMCO between m/z 120 and 140 is recommended for trapping of ions spread over this entire m/z range, as shown in Figure 6. Besides the study of mass discrimination due to the difference in the rf potential, mass discrimination during radial ejection (the normal RIT ejection mode) was also investigated by comparing the peak intensities acquired for the ions ejected through the slits with those for ions ejected along the z axis. Two detectors were installed next to the x electrode slits to detect the radially ejected ions, and the gains were set to 4.0 × 105 using the isotope ratio method, which is based on the correlation between the variation in the ratio of the measured isotopic intensities and the number of trapped ions.26 Another detector was installed along the z axis of the RIT to detect ions ejected axially, and its gain was also adjusted to 4.0 × 105 (Figure 2b). The ions of each monitored m/z value were first isolated with a window of m/z 10 and then ejected either axially, by lowering the dc voltage on the z electrode close to the axial detector, or radially, by scanning the rf amplitude with resonance ejection in the usual way. In the axial ejection (24) Louris, J. N.; Amy, J. W.; Ridley, T. Y.; Cooks, R. G. Int. J. Mass Spectrom. Ion Processes 1989, 88, 97-111. (25) Quarmby, S. T.; Yost, R. A. Int. J. Mass Spectrom. 1999, 190/191, 81-102. (26) Schwartz, J. C. U.S. Patent 20040245451, 2004.

722 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

Figure 7. Mass shifts for the positive ion m/z 1121.99 due to space charge. (Experiments performed with configuration II.)

mode, the bias voltages for the electron multipliers of both the radial detectors were turned off. Since well-balanced rf voltages were applied to the rf electrodes, all the ions were expected to be ejected in the z direction with an efficiency close to 100%, an expectation that was confirmed in this experiment by a follow-up rf scan and radial detection. To record the intensity of the ions ejected radially in terms of peak area, the ions ejected through each x slit were detected separately by turning off the electron multiplier bias voltage on the other detector. To calculate the radial detection efficiency, the sum of the total peak areas acquired by both radial detectors was compared to that from the axial detector. A constant ionization time (2.0 ms) was used to keep the ion abundance constant between experimental cycles. Ions having the same set of m/z values from the calibration mixture were used for this measurement. The radial ejection efficiencies were calculated for ions of m/z 195, 524, 1222, 1522, and 1822. An overall detection efficiency close to 100% was observed for radial detection of ions in the high m/z range; however, the detection efficiency for the ions with low m/z values was relatively low (35% at m/z 190), possibly due to the fact that ions of lower mass were more easily scattered during collisions with the buffer gas molecules. Trapping Capacity and Space Charge Effects. In comparison with 3D ion traps, the elongated z direction of the 2D trap allows a much larger volume for trapping of ions for a trap of given cross-sectional area. During the ion ejection process for 2D traps such as the RIT or the Thermo LIT used in this study, the ion cloud spreads along the z axis,8,12 instead of shrinking into a sphere of ∼1-mm diameter as it does in a 3D trap.27 To obtain a uniform electric field to achieve the desired resolution, the mechanical variation along the z direction of the electrode assembly must be minimized. For a 2D trap with given mechanical tolerances, larger dc voltages can be applied to the end electrodes4 or end sections8 to push the ions toward the center of the trap, resulting in improved resolution. Compression of the ion cloud is expected to accentuate space charge effects, usually indicated by an increase in the mass shift or deterioration in peak shape. The depth of the dc axial potential well of the 2D linear ion trap is usually selected as a compromise between the resolution and (27) Hemberger, P. H.; Nogar, N. S.; Williams, J. D.; Cooks, R. G.; Syka,. J. E. P. Chem. Phys. Lett. 1992, 191, 405-410.

Figure 8. Tandem mass spectrometry experiments performed using the RIT instrument: (a) product ion MS/MS spectrum of protonated reserpine (m/z 609.3) recorded using ESI-RIT configuration I, at a CID efficiency of 64%; (b) MS/MS product ion spectrum of protonated peptide MRFA, m/z 524.3 recorded using configuration II, at a CID efficiency of 70%.

Figure 9. Isolation resolution and efficiency: (a) isolation of the protonated dipeptide Pro-Val (m/z 215) from a mixture including protonated Pro-Pro (m/z 213) and Val-Val (m/z 217), using the RIT mass spectrometer in configuration I; (b) isolation of the 13C isotope of protonated MRFA (m/z 525.31), using configuration II.

space charge capacity considerations. The 100-V dc potential well used for the RIT allows unit resolution at a scan rate of 2778 amu/ (charge‚s) with a mechanical tolerance of ∼100 µm for the RIT. To better characterize the trapping capacity and the space charge effect for this particular RIT analyzer, the mass shift was measured as a function of the number of ions trapped for axial dc potential well depths of 50, 100, and 200 V. The experiments were done using configuration II, and a LMCO of m/z 922 was used to eliminate ions from caffeine and MRFA. The ions were injected using a series of AGC targets ranging from 2000 to 100 000 ions with an increment of 2500, and the mass shifts for the monoiso topic m/z peak of Ultramark at m/z 1121.99 were measured and

plotted as in Figure 7. The mass shift increases with the increase of the AGC target but becomes constant in the case of 50-V axial dc potential well when the AGC target is larger than 65 000. This indicates that the maximum trapping capacity of the RIT has been reached and that this is ∼65 000 ions. Deeper axial potential wells, such as 100 or 200 V, allow the trapping of more ions; however, the ions are pushed closer to the center of the trap and larger mass shifts are observed with the same number of trapped ions. With a 100-V axial dc potential well depth giving unit resolution, a shift of m/z 0.1 was observed with ∼8500 ions trapped, a number which is approximately half of that for the LIT in a commercial LTQ instrument but 10 times that for a 3D trap in a commercial Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

723

Figure 10. Multiple-stage mass spectrometry data for the analysis of a mixture of drug compounds using the RIT instrument in configuration II. (a) MS spectrum recorded for mixture of albuterol, lidocaine, methamphetamine, and amphetamine; (b-e) MS2 product ion spectra of the protonated molecular ions of the drug compounds; (f, g) MS3 spectra of amphetamine and methamphetamine.

LCQ.8 As expected, at a higher dc potential well depth, 200 V, the space charge induced mass shift is larger while a lower well depth helped to reduce the space charge effect. An obvious solution for increasing the trapping capacity of this RIT while retaining the same resolution is to improve the mechanical tolerance of the device so that a lower axial dc potential can be applied to allow the ion cloud to spread more along the z axis. Ion Isolation, Activation, and MSn. The RIT, like all other ion traps, can be operated to perform tandem mass spectrometry, as has been demonstrated previously.4 With the semisealed configuration of the current RIT, the helium buffer gas pressure inside the RIT is higher than that previously used for CID and a higher dissociation efficiency is expected.8,10,11 The tandem mass spectrometry capability of this RIT was investigated, and the efficiencies for ion isolation and CID were characterized using ions from various compounds. In these experiments, the parent ions were isolated at q ) 0.83 using notched waveforms17 and the activation for CID was implemented using resonance excitation at q ) 0.3, which was selected as a compromise between the radial pseudopotential well depth and the LMCO for the fragment ions.8 The MS2 product ion spectra collected for protonated reserpine at m/z 609 and protonated MRFA at 524 Th are shown in Figure 8. An activation time of 110 ms was used and overall CID efficiencies of 64 and 70% were obtained, respectively; these values are comparable to those reported for commercial 2D traps.8 The fragmentation of protonated MRFA at m/z 524 via CID was also done for comparison using a LCQ instrument with a Paul trap. A maximum CID efficiency of 72% was observed for the 3D trap. The isolation efficiency is another important performance parameter that represents the level of selectivity with which the targeted parent ions can be isolated from other ions with adjacent 724 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

mass-to-charge ratios. This parameter is especially important in cases in which complex mixtures are to be analyzed by tandem mass spectrometry. To test the isolation efficiency, a mixture of three dipeptides, Pro-Val (molecular mass 212), Pro-Pro (molecular mass 214), and Val-Val (molecular mass 216), was used to generate three protonated molecules with a difference of m/z 2 between each two. When an isolation window of m/z 2.0 was used, the Pro-Val ion m/z 215 could be isolated from Pro-Pro and Val-Val with an efficiency of ∼100% (Figure 9a). Ions with smaller differences in m/z values can also be isolated using a narrower isolation window. The 13C isotopic ion (m/z 525) of [MRFA + H]+ was isolated from the 12C isotope (m/z 524) and the 13C2 isotope (m/z 526) ions using an isolation window of 1.2 Th with an efficiency of 23.9%, as shown in Figure 9b. As a further test of the RIT capability for analysis using tandem mass spectrometry, a mixture of four drugs, albuterol (molecular mass 239.3), lidocaine (molecular mass 234.34), methamphetamine (molecular mass 149.23), and amphetamine (molecular mass 135.21), each at a concentration of 100 pg/µL for each, was analyzed (Figure 10). The MS2 spectra for albuterol and lidocaine (Figure 10b-e) as well as the MS3 spectra for methamphetamine and amphetamine (Figure 10f, g) were recorded. MSn analysis of the same mixture was carried using the commercial LTQ instrument. Identical fragmentation patterns and comparable signal-tonoise ratios were observed. CONCLUSIONS The performance of the simple RIT mass analyzer, when used in the semisealed configuration described and incorporated into instruments that use ESI ionization, was characterized and capabilities for the analysis of proteins, peptides, and drug

compounds were demonstrated with MS and MSn. Unit resolution has been achieved in the whole m/z range up to m/z 2000 at a scan rate of 2778 amu/(charge‚s), and much higher resolution is available at reduced scan rates. The mass discrimination during axial ion injection was minimized when an appropriate LMCO was selected and the detection efficiency for ions with different massto-charge ratios was characterized. The isolation and fragmentation efficiencies were also characterized. The RIT mass analyzer does not quite reach the levels of performance achieved using a commercial linear ion trap, but the differences are relatively small. The major limitation for mass resolution is probably the mechanical tolerance achieved, especially the tolerance in the z direction, which limits the number of ions that can be trapped. Given the level of performance observed for this device and the simplicity of the trap configuration, the RIT is expected to be good candidate

for mass spectrometers used for biological studies, especially when high-throughput instruments5-7 with multiple analyzers are developed. ACKNOWLEDGMENT We thank Nigel Gore of Thermo Electron Corp. (San Jose, CA) for suggestions and helpful discussions. We acknowledge funding from the NIH (Grant No.1R21DK070290-01), and the Indiana 21st Century Fund through a grant to the Indiana Instrumentation Institute.

Received for review July 18, 2005. Accepted November 22, 2005. AC0512709

Analytical Chemistry, Vol. 78, No. 3, February 1, 2006

725