Tailored Ion Radius Distribution for Increased Dynamic Range in FT

Nov 29, 2012 - Department of Chemistry and Biochemistry, Florida State University, 95 Chieftain Way, Tallahassee, Florida 32306, United States. ABSTRA...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/ac

Tailored Ion Radius Distribution for Increased Dynamic Range in FTICR Mass Analysis of Complex Mixtures Nathan K. Kaiser,*,† Amy M. McKenna,† Joshua J. Savory,† Christopher L. Hendrickson,†,‡ and Alan G. Marshall†,‡ †

Ion Cyclotron Resonance Program, National High Magnetic Field Laboratory, Florida State University, 1800 East Paul Dirac Drive, Tallahassee, Florida 32310-4005, United States ‡ Department of Chemistry and Biochemistry, Florida State University, 95 Chieftain Way, Tallahassee, Florida 32306, United States ABSTRACT: Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) typically utilizes an m/zindependent excitation magnitude to excite all ions to the same cyclotron radius, so that the detected signal magnitude is directly proportional to the relative ion abundance. However, deleterious space charge interaction between ion clouds is maximized for clouds of equal radius. To minimize ion cloud interactions, we induce an m/z-dependent ion radius distribution (30%−45% of the maximum cell radius) that results in a 3-fold increase in mass spectral dynamic range for complex mixtures, consistent with increased ion cloud lifetime for less-abundant ion clouds. Further, broadband frequencysweep (chirp) excitation that contains the second and/or third harmonic frequency of an excited ion cloud swept from low-to-high frequency produces systematic variations in accurate mass measurement not observed when the sweep direction is reversed. The ion cyclotron radius distribution induces an m/zdependent frequency shift that can be corrected to provide a root-mean-square (rms) mass measurement error of 30 000 peaks per mass spectrum with >80 peaks per nominal mass can now be resolved and identified in a single mass spectrum.7 Importantly, the spectral complexity of petroleum samples provides an excellent diagnostic tool to improve FT-ICR MS performance parameters through visualization of perturbations in ion behavior.8 The exceptional performance characteristics of FT-ICR MS enabled through nondestructive image charge detection of coherent ion motion approach the theoretical limit only if all ion packets maintain phase coherence throughout the entire data acquisition time period. Electric and magnetic field inhomogeneities, ion-neutral collisions, and Coulombic interactions between ion clouds disrupt ion packet phase coherence and limit transient lifetime.9−11 Ion packets for more-abundant species contain more ions of a given m/z ratio, which are less affected by shear forces that disrupt ion phase coherence than less-abundant species, which typically have shorter transient lifetimes than do more abundant species in the same mass spectrum.12,13 However, the charge capacity of the ICR trap is limited, so improved transient lifetime for low-abundance ion clouds is needed.14,15 © 2012 American Chemical Society

Received: September 14, 2012 Accepted: November 29, 2012 Published: November 29, 2012 265

dx.doi.org/10.1021/ac302678v | Anal. Chem. 2013, 85, 265−272

Analytical Chemistry

Article

Figure 1. Magnitude-mode excitation spectra (obtained by Fourier transform of the detected signal during excitation) applied to vary the ion cyclotron radius distribution for ions of m/z 200−1200: (A) broadband “chirp” excitation, whose variation in radio-frequency (rf) magnitude across the frequency bandwidth results in m/z-dependent post-excitation ion cyclotron radius; (B) SWIFT excitation for excitation of ions of all m/z values to the same post-excitation radius; (C) SWIFT excitation to yield a post-excitation ion radius distribution that varies linearly with mass over a 15% range; and (D) SWIFT excitation to generate a post-excitation ion radius distribution tailored to create a uniform ion charge density for a particular complex mixture.

function of ion radius,27 so that frequency sweep excitation from low frequency to high frequency excites ω+ to a large initial radius prior to resonance at 2ω+ and 3ω+, which, in turn, triggers a larger response from resonances at 2ω+ and 3ω+ than a frequency sweep from high frequency to low frequency, during which the higher-order resonances are excited from essentially zero initial ion radius. Here, we optimize broadband ion cyclotron frequency excitation for the analysis of petroleum crude oil. Excitation of ions to the same radius maximizes space charge interaction between ion clouds and limits transient lifetime, particularly for less-abundant species. Instead, we impose an m/z-dependent variation in post-excitation ion cyclotron radius that distributes ion charge over a greater ICR cell volume, thereby reducing Coulombic interaction and increasing transient lifetime of the less-abundant species. Spectral dynamic range increases 3-fold, with 2-fold increase in the number of resolved and identified components. Furthermore, the predicted radial frequency response at the second and third harmonics during the lowto-high frequency sweep is manifested as a systematic mass measurement error increase, and we demonstrate improved performance for the high-to-low frequency sweep, which requires axially linearized excitation electric field in an open ICR cell.28 Although most critical for complex spectra, we predict that all ICR spectra will improve by use of ICR postexcitation ion radius distribution and a high-to-low frequency sweep.

50:50 (v:v) toluene/methanol and 1% formic acid for positiveion electrospray analysis. All reagents were obtained from Sigma−Aldrich (St. Louis, MO). Ions are generated by direct infusion into a microelectrospray source29 at a flow rate of 500 nL/min by a syringe pump. 9.4-T FT-ICR MS. All experiments are performed with a custom-built 9.4-T FT-ICR mass spectrometer.30 Ions are accumulated (0.25−2.0 s) in a radio-frequency (rf) octopole (2.0 MHz, 200 Vp‑p) modified with tilted-wire extraction electrodes for improved ion axial ejection.31 Ions are then transferred to a compensated open cylindrical Penning trap via an octopole ion guide.32,33 Ions are excited to ∼40% of the maximum cell radius by frequency-sweep excitation from high frequency to low frequency (720 kHz to 120 kHz) at 50 Hz/μs, unless otherwise stated, and are observed by broadband image charge detection (5.6 s data acquisition period, 8 M word data points, 746 kHz bandwidth (lowest m/z = 193)). Multiple (50−200) time-domain acquisitions are averaged to obtain each stored transient, which is Hanning-apodized and zero-filled once before fast Fourier transform (FFT) and conversion to m/z. Instrument control and data analysis is performed by a modular ICR data acquisition and analysis system.34 RF Excitation Magnitude Variation. The open cylindrical ICR cell is capacitively coupled for linearization of the excitation electric field.28 The specified ion cyclotron excitation waveform is produced by an arbitrary waveform generator, amplified by a high-power rf amplifier (ENI, Rochester, NY), and coupled through a center-tapped transformer ((50−1200 Ω balun (0904LB)), Northhills, Syosett, NY) that splits the rf into 0° and 180° phases and provides a 3−5 fold voltage increase. The capacitive load imposed by the excitation circuitry induces a frequency-dependent rf voltage, whose magnitude is reduced at higher frequency (low m/z), as shown by the excitation magnitude spectra (obtained by Fourier transform of the signal (in the absence of ions) acquired during excitation) in Figure 1. The waveform in Figure 1A displays the radial



EXPERIMENTAL METHODS Sample Preparation. Athabasca bitumen heavy vacuum gas oil (HVGO) or North American crude oil (∼5 mg) is diluted with 5 mL of toluene (HPLC grade, Sigma−Aldrich Chemical Co., St. Louis, MO) to make a stock solution. The stock solution is further diluted to a final concentration of 0.5 mg/mL with 50:50 (v:v) toluene/methanol and 1% ammonium hydroxide solution for negative-ion electrospray analysis, or 266

dx.doi.org/10.1021/ac302678v | Anal. Chem. 2013, 85, 265−272

Analytical Chemistry

Article

Figure 2. Broadband negative-ion electrospray ionization 9.4 T FT-ICR mass spectra of Athabasca bitumen HVGO, for a distribution in postexcitation cyclotron (left) and for ions all excited to the same cyclotron radius (right). The mass scale-expanded insets show a significant increase for low-abundance species, for ions excited to a different cyclotron radius.

were internally calibrated with respect to a highly abundant homologous alkylation series containing either two 16O atoms for negative ions or one 14N atom for positive ions. Ions with relative abundance greater than six standard deviations of baseline rms noise (6σ) were used for analysis. Mass spectra were further calibrated by use of a “walking” calibration that applies separate calibration coefficients over small segments of the m/z range (see below).39,40

dispersion created when an ideal excitation waveform (frequency-independent magnitude) is input into the excitation circuit. Extension of the excitation m/z range beyond that shown in Figure 1A continues to follow the same frequencydependent amplitude modulation with flatter response at high m/z and further reduction in amplitude at lower m/z. Therefore, ions of different m/z are excited to different postexcitation cyclotron radius. To further control the ion cyclotron radius distribution, many “stored waveform inverse Fourier transform” (SWIFT) excitation waveforms were generated to create various ion cyclotron radius distributions,35 which should track the SWIFT magnitude spectra shown in Figure 1. The excitation profile across a defined m/z range is created by adjusting the amplitude of the input waveform into the excitation rf amplifier across a defined frequency range. Simulations. SIMION 8.0 was used to calculate the postexcitation ion cyclotron radius distribution for either frequencysweep direction. The ICR cell model was based on the design of the current cell in our 9.4 T FT-ICR instrument (capacitively coupled open cylindrical with an internal diameter of 94 mm). The ICR celled was modeled with 1 mm per SIMION grid unit. Simulations were performed to replicate typical experimental broadband excitation conditions (720−120 kHz), magnetic field strength of 9.4 T, and frequency sweep rate of 50 Hz/μs. The ions were placed at the center of the ICR cell and then subjected to frequency-sweep excitation (either high-to-low or low-to-high) to ∼40% cell radius, after which the ion orbital trajectory was recorded. A single ion of each m/z was flown without consideration of ion−ion interactions in the SIMION simulation. Mass Calibration and Data Analysis. Internal calibration36,37 is based on a homologous ion series whose members differ in mass by integer multiples of CH2 (14.01565 Da) across the molecular weight distribution.38 ESI FT-ICR mass spectra



RESULTS AND DISCUSSION Ion Cyclotron Radius Distribution. The FT-ICR signal magnitude is proportional to the ion cyclotron radius. Therefore, it is preferable to excite all ions to the same radius to produce a mass spectrum whose peak heights accurately represent the relative abundances of ions in the ICR cell. However, the excitation of all ions to the same radius creates a thin cylinder of high charge density that has a deleterious impact on ion cloud lifetime (note that each species is detected as a separate, coherent ion cloud and that the lifetime of each cloud may be different). Therefore, we impose an m/zdependent radius distribution to spread the ions over a larger volume to reduce the interaction between different m/z clouds. Figure 2 shows broadband mass spectra for two different excitation waveforms (a range of post-excitation radius versus no spread in radius). The excitation waveform shown in Figure 1A produces a cyclotron radius of ∼45% of the cell radius for ions of m/z 1200 and ∼30% of the cell radius for ions of m/z 200, whereas the flat excitation waveform (Figure 1B) excites all ions to ∼40% of the cell radius. The post-excitation ion cylotron radius for each waveform was independently optimized to provide the largest number of identified peaks over the widest m/z range. Both excitation conditions produce similar molecular weight distributions and signal magnitude when viewed over the entire mass range. However, a mass-scale

267

dx.doi.org/10.1021/ac302678v | Anal. Chem. 2013, 85, 265−272

Analytical Chemistry

Article

when all ions are excited to the same radius for different numbers of ions in the ICR cell. For a low number of trapped ions (i.e., 0.3 s accumulation period), there is a smooth distribution of assigned elemental compositions with mass. However, as more ions are added to the trap with increased ion accumulation period, a significant reduction in the number of assigned elemental compositions occurs in the middle of the distribution (mass range 400−600 Da). That mass range includes the most-abundant ion clouds. The premature loss of phase coherence reduces the signal magnitude to such an extent that the less-abundant species no longer exceed the signal-tonoise threshold. Figure 3 (right) shows the spectral complexity of the same sample for a distribution in cyclotron radius (Figure 1A). The results indicate that the most significant dephasing of low-abundance ion clouds occurs when they are spatially proximal to the most abundant ion clouds. Therefore, the largest variation in cyclotron radius should be imposed on the m/z range containing the most-abundant ion clouds. The highest charge density occurs if all ions are excited to the same radius, resulting in the most destructive Coulombic interactions. A series of excitation waveforms that impose different linear variation in the radius distributions was applied to systematically vary the charge density. The results shown in Table 2 indicate that the highest dynamic range (largest number of peaks) is obtained by use of the excitation waveform that produces the greatest difference in cyclotron radius. Coulombic interactions between ion clouds decrease with greater dispersion in cyclotron radius, and result in reduced shear forces that induce dephasing of low-abundance ion packets. We obtained the best results by varying the ion radius between 30% and 45% of the ICR cell radius. Increase in cyclotron radius beyond 45% of the cell radius reduces ion cloud lifetime, presumably because of significant electric and/or magnetic field imperfections. Ion radius below 30% of the cell radius also limits ion cloud lifetime, presumably due to increased Coulombic interaction at low ion cyclotron radius. The negative slope indicated in Table 2 corresponds to the low mass ions excited to the larger radius. That type of radius distribution is unfavorable for these types of complex samples, presumably because the number of species increases with mass, so that the least-complex part of the mass range would be excited to a high radius and the more-complex part of the mass range would be excited to a lower radius, thereby further exacerbating the destructive Coulombic interactions. We created a tailored SWIFT excitation waveform to produce the most favorable ion density distribution. The tailored SWIFT excitation waveform (shown in Figure 1D) contains three distinct regions, with a shallow slope at low m/z (because the sample is less complex at lower mass). The middle section has the highest slope to spread out the more-abundant ion clouds in the middle of the spectrum to the greatest extent. Finally, there is a shallow slope again at higher mass, to limit the maximum radius to which the ions are excited. The results of this tailored SWIFT waveform and for an uncorrected “chirp” excitation waveform, for each of three different ion accumulation periods, are shown in Figure 4. For the tailored SWIFT waveform, the number of assigned elemental compositions for different ion accumulation periods is closer to the desired response, in that the number of assigned peaks increases with increase in accumulation period. For the uncorrected waveform, the number of assigned peaks in the most complex part of the spectrum maximizes at a much lower ion number. Furthermore,

expanded zoom inset exposes significant differences in observation of less-abundant ions. Nearly twice as many species are detected at 6 times the baseline rms noise level when ions are excited to a different cyclotron radius. The inset clearly shows that greater radius dispersion increases the signal magnitude for low-abundance species; the appearance of previously undetectable low-abundance species thus reflects an increase in dynamic range. The results suggest that the ion cyclotron radius distribution minimizes destructive Coulombic interactions, particularly for the less-abundant ion clouds. The more-abundant ion clouds that are inherently more stable are not severely affected by decreased radial dispersion, as evident from their similar spectral magnitude. A similar effect has been observed at the edges of a protein isotope distribution, for which the less-abundant isotopic ion packets dephase faster.12,13 When the post-excited cyclotron radius is varied, the signal magnitude is no longer proportional to the relative ion abundances over a broad m/z range. However, because ICR signal magnitude is proportional to the magnitude spectrum of the excitation waveform, the relative ion abundances can be corrected accordingly. Ion Charge Density. Space-charge conditions in the ICR cell depend on ion density, which depends on the occupied volume of the trapped ions and on the number of ions. We vary the ion volume by imposing an m/z-dependent ion radius distribution and also vary the number of trapped ions by variation of the external ion accumulation period. Experiments were performed for different numbers of trapped ions, with all ions at the same radius and with a distribution of cyclotron radius to observe differences in spectral complexity. The number of peaks detected for an HVGO at each of four different ion accumulation periods is shown in Table 1. At low Table 1. Number of Peaks Identified from Negative-Ion Electrospray Ionization 9.4 T FT-ICR Mass Spectra of Athabasca Bitumen HVGO Obtained with Four Different Ion Accumulation Periods and Either of Two Excitation Waveformsa Number of Identified Elemental Compositions ion accumulation period (s)

ion radius distribution (“chirp” excitation)

all ions excited to a common radius

0.4 0.7 1.0 1.3

1518 2573 2879 2678

1523 1458 919 677

a

Ions were excited with either no radius distribution or with a radius distribution corresponding to Figure 1A. Each spectrum represents a signal average from 50 data acquisition periods

ion accumulation period, the number of species detected is approximately the same for both excitation conditions. However, as the number of trapped ions increases, the number of peaks detected decreases when ions are excited to the same cyclotron radius, due to rapid dephasing of the less-abundant ion clouds. However, if ions are excited to a different cyclotron radius, the number of peaks observed nearly doubled for a 3fold increase in the ion accumulation period. Optimal Radius Distribution. Variation in ion cyclotron radius across the excitation bandwidth increases transient lifetime of the less-abundant ion clouds, compared to excitation of all ions to the same radius. Figure 3 (left) illustrates where the loss of spectral complexity occurs within the mass spectra 268

dx.doi.org/10.1021/ac302678v | Anal. Chem. 2013, 85, 265−272

Analytical Chemistry

Article

Figure 3. Number of assigned elemental compositions for singly charged positive-ion mass spectra of a North American crude oil, binned every 10 Da. The listed ion accumulation period is varied for each of two different excitation conditions. (Left) All ions are excited to the same cyclotron radius. (Right) Ions are “chirp”-excited (Figure 1A) to yield a distribution in ion radius. Note the decrease in number of identified species with an increase in the accumulation period, if all ions are excited to the same radius.

Variation in mass error with m/z can result from the imposed ion radius distribution, if the calibration equation does not appropriately account for m/z-dependent, space-charge-induced frequency shift.36,41 The greater the distribution of cyclotron radius, the larger will be the observed frequency shift (Table 3). We account for m/z-dependent space charge by dividing the spectrum into small m/z segments (∼14 Da each) and applying an individual calibration term to each segment:39 i.e., assuming that the space-charge-induced frequency shift is different for each segment but constant within each segment. The assumption is that the space-charge condition within each segment is relatively uniform. Use of that “walking” calibration increases the number of assigned peaks by as much as 25%, while reducing the root mean square (rms) mass error by as much as 3-fold. Frequency-Sweep Direction. Ions are excited to a detectable cyclotron radius by frequency sweep through the bandwidth of interest. Consideration of ion trajectory in idealized spatially homogeneous magnetic and excitation electric fields and quadrupolar trapping potential indicates that either frequency-sweep direction accelerates ions to approximately the same post-excitation cyclotron orbit. However, plots of mass measurement error versus m/z shown in Figure 6 reveal significant differences. The frequency sweep from m/z 1200−200 (120−720 kHz) produced pronounced additional measured error, starting at m/z 400 and 600, corresponding to the 2ω+ and 3ω+ integer multiples of the cyclotron frequency corresponding to the lowest excited mass (highest frequency) of the excitation waveform. Mitchell et al. used numerical and approximate analytical solutions to predict additional resonances in a cubic ICR cell, including integer multiples of the ion cyclotron frequency.25 It is not surprising to find that similar resonances exist in a cylindrical open cell. Furthermore, ion response to these resonances depends on ion radius and diminishes at the center of the ICR cell. Consequently, for a low-to-high frequency sweep, highmass ions have already been excited to a large cyclotron orbit by the time that their 2ω+ and 3ω+ harmonic multiples are

Table 2. Number of Spectral Peaks Identified from PositiveIon Electrospray Ionization 9.4 T FT-ICR Mass Spectra from a Whole North American Crude Oil, Obtained for Different Linear Variation in Post-Excitation Ion Cyclotron Radiusa slope of excitation profile (%)

number of assigned elemental compositions

−8.3 −2.9 −1.5 0.0 1.5 2.9 4.4 8.6 11.3

1100 1987 2595 3050 3857 4685 5298 6953 7352

a

A negative slope indicates that low m/z ions are excited to a larger cyclotron radius. Each spectrum represents a signal average from 50 data acquisition periods.

the total number of assigned elemental compositions is 2-fold higher for the tailored radius distribution. Frequency Shift. Ion space charge creates a radial electric field that shifts the observed ion cyclotron frequency. The magnitude of the shift depends on the total number of trapped ions but is approximately independent of m/z if all ions are excited to the same cyclotron radius, as shown in Figure 5 (bottom). However, m/z-dependent excitation magnitude leads to m/z-dependent ion cyclotron radius and induces an m/zdependent, space-charge-induced frequency shift. The spacecharge-induced frequency shift depends on the number of ions that rotate around a smaller (but not larger) cyclotron radius,14 so the frequency shift increases with m/z when the higher m/z ions are excited to larger cyclotron radius (Figure 5 (top)). The reduced cyclotron frequency for each species is scaled to the frequency obtained from calibration for an ion accumulation period of 0.4 s. 269

dx.doi.org/10.1021/ac302678v | Anal. Chem. 2013, 85, 265−272

Analytical Chemistry

Article

Figure 4. Number of assigned elemental compositions for singly charged positive-ion mass spectra from a North American crude oil, binned every 10 Da. The listed ion accumulation period is varied for each of two different excitation conditions. (Left) Ions are excited with a cyclotron radius distribution corresponding to Figure 1A. (Right) Ions are excited to yield a radius distribution tailored to the specific molecular weight distribution (corresponding to Figure 1D).

applied to excite lower m/z ions. The previously excited ion clouds absorb additional energy from the harmonic frequencies (2ω+, 3ω+, etc.) and are displaced from the orbit induced by excitation at the primary cyclotron frequency (ω+). Ion response to the higher harmonics alters both the magnetron and cyclotron orbits, as revealed by SIMION. Increase in the ion radial distribution at each harmonic ultimately limits the achievable mass measurement accuracy, because it is difficult to account for that distribution in the mass calibration. The error distribution for the high-to-low frequency sweep is shown in Figure 6 (bottom) with no apparent harmonic response as seen for the low-to-high frequency sweep. The mass error obtained with the high-to-low frequency sweep may be reduced by application of the “walking” calibration, whereas the “walking” calibration does not help the other sweep direction because the harmonic response randomizes the ion’s orbital position and, thus, its frequency shift cannot be corrected by the walking calibration. Therefore, a high-to-low frequency sweep excitation is preferred. The difference in mass error distribution with frequencysweep direction results from the variation in the post-excitation magnetron and cyclotron radius, as shown in Figure 7, which depicts three final ion trajectories (m/z 300, 500, and 700, modeled in SIMION) following low-to-high (left) or high-tolow (right) frequency sweep excitation to ∼40% of the cell radius across an excitation frequency range of m/z 200−1200. The bandwidth is sufficient to cover the fundamental ion cyclotron frequency for m/z 300, 500, and 700. Furthermore, the resonance at 2ω+ is covered for m/z 500 and 700 and 3ω+ is covered for m/z 700. For the high-to-low frequency sweep, the final ion cyclotron and magnetron radius are indistinguishable, resulting in the formation of a thin cylinder of charge created by the excited ion clouds of many different m/z species. For the low-to-high frequency sweep, the different m/z ions attain different final cyclotron and magnetron radius. The cyclotron and magnetron orbits for m/z 500 and 700 are shifted from the center of the cell due to the absorption of radial energy at the 2ω+ and 3ω+ resonances. The orbital radius for ions of m/z 300 is the same as for the high-to-low frequency

Figure 5. Ion cyclotron frequency shift versus m/z for ion accumulation period of 0.4 s (blue), 0.7 s (red), 1.0 s (black), and 1.3 s (green). (Top) An ion radius distribution has been imposed by m/z-dependent ion excitation magnitude similar to that shown in Figure 1A. (Bottom) All ions are excited to the same radius by m/zindependent ion excitation magnitude.

Table 3. Root Mean Square (rms) Mass Error for Different Degree of Cyclotron Radius Variation, With and without the “Walking” Mass Calibration (i.e., Separate Calibration for Each 14 Da Mass Spectral Segment)a slope of excitation profile

single 2-term calibration

walking calibration

2% 4% 8% 11% 15%

248 269 312 363 510

131 101 109 93 86

The mass error without the “walking” calibration increases, whereas the mass error with the “walking” calibration decreases with increasing variation in cyclotron radius. a

270

dx.doi.org/10.1021/ac302678v | Anal. Chem. 2013, 85, 265−272

Analytical Chemistry

Article

Figure 6. (Top) Measured mass error versus m/z for low-to-high frequency sweep excitation from 120 kHz to 720 kHz (m/z 1200−200). Distinct increases in the rms mass measurement error are observed at m/z 400 and 600. The highest excited frequency corresponds with m/z 200, therefore the m/z range from 400−600 absorbs additional energy from the 2nd harmonic and the m/z range from 600−1200 absorbs additional energy from the 2nd and 3rd harmonic. (Bottom) Measured mass error versus m/z for high-to-low frequency sweep excitation from 720 kHz to 120 kHz (m/z 200−1200), showing no changes corresponding to second and third harmonics.

however, the mass measurement accuracy is significantly reduced for the higher m/z species that experience additional resonance effects with the low-to-high frequency sweep excitation.



CONCLUSIONS Excitation of ions of different m/z to different cyclotron radius distributes the ion charge density over a greater volume within the ICR cell. The separation of ion clouds to different cyclotron radius increases the number of ions that can be trapped in the ICR cell without the detrimental effect from Coulombic interaction between ion clouds, thus increasing the dynamic range. The spread of ion density over a larger portion of the ICR cell decreases the space-charge-related factors that contribute to reduce the transient lifetime of the less-abundant ions. Excitation of ions to different cyclotron radius causes a space-charge-induced frequency shift, which is related to the distribution of ion charge density across the mass range. Thus, the space-charge-related frequency shift is m/z-dependent. The change in space-charge frequency shift across the m/z range can be corrected by separate calibration of mass spectral segments, for mass measurement error