the ‘hybrid cell’: a new compensated infinity cell for larger radius ion excitation in fourier...
TRANSCRIPT
RAPID COMMUNICATIONS IN MASS SPECTROMETRY
Rapid Commun. Mass Spectrom. 2008; 22: 1423–1429
) DOI: 10.1002/rcm.3516
Published online in Wiley InterScience (www.interscience.wiley.comThe ‘hybrid cell’: a new compensated infinity cell for
larger radius ion excitation in Fourier transform ion
cyclotron resonance mass spectrometry
Sunghwan Kim1, Myoung Choul Choi1, Manhoi Hur1, Hyun Sik Kim1, Jong Shin Yoo1*,
Christopher L. Hendrickson2 and Alan G. Marshall2
1Korea Basic Science Institute, 804-1 Yangcheong-Ri, Ochang-Myun, Cheongwon-Gun, Chungcheongbuk-Do, 363-883, Republic of Korea2Ion Cyclotron Resonance Program, National High Magnetic Field Laboratory, Florida State University, 1800 E. Paul Dirac Drive,
Tallahassee, FL 32310-4005, USA
Received 14 December 2007; Revised 25 February 2008; Accepted 26 February 2008
*CorrespoYangcheocheongbuE-mail: joContract/FT-ICR inContract/SpectromContract/Contract/Laborato
A new ‘hybrid’ ion cyclotron resonance (ICR) trap is proposed and analyzed by computer simu-
lations. The trap is basically a hybrid of a segmented end cap (Infinity) and capacitively coupled
cylindrical cell, with additional electrodes placed at the center of each end cap. The new trap
produces an on-axis electric field z-profile similar to that of the Infinity cell or capacitively coupled
open cylindrical cell during ion excitation. Simion simulations demonstrate that, during detection,
appropriate changes of the potentials applied to the two new sets of electrodes produce a radial
electric field z-profile that more closely approaches that for an ideal axial three-dimensional
quadrupolar potential at high post-excitation ICR orbital radius, for improved signal-to-noise ratio
and resolving power, and minimal m/z-discrimination. Copyright # 2008 John Wiley & Sons, Ltd.
Since its inception,1 Fourier transform ion cyclotron reson-
ance mass spectrometry (FT-ICR MS) has become the
ultimate standard for high-resolution broadband mass
analysis.2 A resolving power of 3 300 000 for a �1 kDa
peptide has been reportedwith this technique.3 Furthermore,
the mass measurement accuracy is sufficiently high that for
molecules up to �500Da in mass (with appropriate
elemental constraints), a unique elemental composition for
molecules in even the most complex natural mixtures
(namely, petroleum crude oil) can be calculated.4 The
high-resolution capability of FT-ICR MS has had a high
impact on protein and natural organic mixture analysis.5–12
The resolving power in FT-ICR MS is limited by the
duration of the time domain ICR signal.2 The time duration is
critically related to ion motion. Ideally, a Penning trap13
confines and stores ions by combination of a spatially
uniform static magnetic field and a three-dimensional axial
quadrupolar electrostatic field. Ions in such a trap exhibit
three periodic motions (cyclotron rotation, magnetron
rotation, and axial ‘trapping’ oscillation).14 Ion stability
derives from these motions. Cyclotron rotation results from
the Lorentz force on an ion of mass,m, and charge, q, moving
ndence to: J. S. Yoo, Korea Basic Science Institute, 804-1ng-Ri, Ochang-Myun, Cheongwon-Gun, Chung-k-Do, 363-883, Republic of [email protected] sponsor: KBSI project ‘‘Development of anstrument’’.grant sponsor: NSF National High Field FT-ICR Massetry Facility; contract/grant number: DMR-06-54118.grant sponsor: Florida State University.grant sponsor: The National High Magnetic Fieldry in Tallahassee, FL, USA.
in a static magnetic field, B0, and prevents ions from escaping
in directions perpendicular to B0. In the absence of an electric
field the ion cyclotron angular frequency, vc, is given by:
vc ¼qB0
mð‘Unperturbed
;cyclotron frequency; S:I:unitsÞ (1)
In a spatially uniform direct current (dc), magnetic field,
and a three-dimensional quadrupolar electrostatic trapping
potential (e.g. near the center of an ICR ion trap), the ion
radial and axial motions are uncoupled in energy and the ion
cyclotron, magnetron, and axial frequencies are each
independent of ion position.15 Ions can be confined for a
long period of time without significant loss (the record is
10 months for a single electron in a Penning trap16). However,
collisions with neutrals,17 deviation from quadrupole
electrostatic trapping potential due to truncated, apertured,
or otherwise imperfect trapping electrodes, and Coulombic
charge interactions18 can destabilize ions radially and result
in damping and distortion of the time-domain ICR signal.
Mathematically, higher order terms must be added to the
quadrupolar electrostatic potential, and the three ion
motions then depend on ion position in the trap. Various
trap configurations have been designed to ‘shim’ the
radio-frequence (rf) and/or dc potentials toward their ideal
shape.19–27
Application of a spatially uniform transverse (to the
magnetic field) resonant rf electric field coherently excites
ions to a detectably large ion cyclotron orbital radius.2
However, the non-uniform rf excitation electric field in an
actual finite-size ICR cell can result in axial ejection of ions
(particularly at low m/z).28 Axial ejection may be reduced by
distributing the ion excitation voltage onto appropriately
Copyright # 2008 John Wiley & Sons, Ltd.
Figure 1. Sagittal cross-sections of a schematic hybrid
cylindrical ICR cell, showing capacitive coupling between
compensation and excitation electrodes.
1424 S. Kim et al.
segmented end cap electrodes (e.g. the Infinity cell19 or the
capacitively coupled open cell28) to flatten the rf excitation
electric field isopotential surfaces.
In a three-dimensional axial quadrupolar dc potential,
F(r,z), the radial electric field, dF/dr, increases linearly with
increasing r but is independent of z.29 A grounded screen
mesh30 or a trap with interlacing ‘comb’ wires as rf end cap
electrodes20 both effectively produce near-zero (and thus
constant) radial electric field as a function of z. Physically or
electronically segmented trap electrodes have been pro-
posed.22 Othermulti-electrode configurations can also flatten
the radial electric field.23–25,31, 32
It is known that ICR detection sensitivity, measurement
accuracy, and/or isotopic distributions can be improved by
optimizing post-excitation ICR orbital radius.2,33,34 More
Figure 2. Simulated potential (top), radial electric field (middle), and radial electric field
z-gradient (bottom) at 33% (left column) and 60% (right column) of the trap radius in a closed,
cylindrical ICR trap. The end cap potentials are 1V. Red dashes indicate the ideal (quad-
rupolar) distributions of dF/dr and d2F/drdz. This figure is available in color online at
www.interscience.wiley.com/journal/rcm.
Copyright # 2008 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2008; 22: 1423–1429
DOI: 10.1002/rcm
Hybrid FT-ICR MS cell for larger radius excitation 1425
recently, it has been shown that altering the dc electric
potential field during detection can lengthen ICR time-
domain signal duration, for improved FT-ICR mass spectral
resolving power and peak shape.29,35
Here, we propose a new ICR trap that adds two types of
electrodes to an Infinity19 trap, to modify the trapping
potential at a given ion post-excitation ICR orbital radius.
The new trap produces an on-axis electric field distribution
similar to that of the Infinity cell during rf excitation. After
excitation, the field distribution of the new trap ismodified to
approach more closely the ideal quadrupolar electric field at
the post-excitation ion cyclotron radius in the ICR trap.
EXPERIMENTAL
The proposed hybrid cell (Fig. 1) is similar to the Infinity
cell,19 with two types of additional electrodes. One set of
‘compensation’ electrodes is located between the excitation/
detection and end cap electrodes, and consists of four
Figure 3. Simulated potential (top), radial e
field z-gradient (bottom) at 33% (left column) a
in an open cylindrical ICR trap. The end cap po
ideal distributions of dF/dr and d2F/drdz. T
www.interscience.wiley.com/journal/rcm.
Copyright # 2008 John Wiley & Sons, Ltd.
cylindrical electrode quadrants of the same radius as the
excitation/detection electrodes. The compensation electro-
des adjacent to the excitation electrodes are capacitively
coupled to the excitation electrodes. The compensation
electrodes are physically separated from the trap and
excitation/detection electrodes. The placement of the
compensation electrodes is similar to that for the compen-
sated trap.16,31 The electrostatic potential field was simulated
for compensation electrodes of different lengths. The choice
of length is not critical because the change in potential could
be compensated for by optimizing the applied voltages.
The hybrid cell has additional electrodes compared with
the compensated trap: namely, the ‘center compensation’
electrodes located in a circular hole in the middle of each end
cap. In the conventional Infinity cell, so-called ‘sidekick’
electrodes corresponding to the ‘center compensation’
electrode used here are installed at one end of the cell.
The Hybrid cell has center compensation electrodes at both
ends of the cell for geometrical symmetry. The end caps
lectric field (middle), and radial electric
nd 60% (right column) of the trap radius
tentials are 1V. Red dashes indicate the
his figure is available in color online at
Rapid Commun. Mass Spectrom. 2008; 22: 1423–1429
DOI: 10.1002/rcm
1426 S. Kim et al.
themselves are the same as for the Infinity cell,19 namely, a
pattern of electrically isolated segments designed to achieve
the excitation profile of an infinitively long trap. The center
compensation electrode diameter was chosen to match that
of a typical transfer octopole (5mm i.d.).
Three types of cells were simulated. The open cylindrical
cell had 60mm length and 60mm i.d. of detection and end
trapping electrodes. The same length and i.d. detection elec-
trodes were used for the new ‘hybrid’ and closed cylindrical
cells. Model cells were divided into small grid spacings
(0.5mm/grid unit), and were refined to a convergence level
of 10�5. Use of smaller grid spacing and better convergence
levels did not significantly change the results. Computer
simulations were performed with SIMION 3D36 version 7
running on a 3.1GHz Pentium 4 PC with 2 GByte RAM.
Figure 4. Simulated potential (top), radial elec
z-gradient (bottom) at 33% (left column) and 6
proposed hybrid ICR trap. Potentials on the compe
electrodes are indicated. Red dashes indicate th
and d2F/drdz. This figure is available in color on
rcm.
Copyright # 2008 John Wiley & Sons, Ltd.
RESULTS AND DISCUSSION
Electrostatic potential field during detection inconventional ICR trapsIn a three-dimensional quadrupolar electrostatic dc potential
and spatially uniform rf excitation and detection electric
field, the image charge induced on ICR cell detection
electrodes increases linearly from zero (ions on-axis) with
increasing ion cyclotron orbital radius. However, the
trapping potential deviates increasingly from quadrupolar
with increasing ICR radius.37 The magnitudes of ‘harmonic’
signals at odd-integer multiples of the ‘fundamental’ ICR
frequency also increase,37 and the magnetic field homogen-
eity decreases. In addition, there are complex effects due to
Coulomb repulsions between ions.38 As a result, optimal
tric field (middle), and radial electric field
0% (right column) of the trap radius in a
nsation, end cap, and center compensation
e ideal (quadrupolar) distributions of dF/dr
line at www.interscience.wiley.com/journal/
Rapid Commun. Mass Spectrom. 2008; 22: 1423–1429
DOI: 10.1002/rcm
Hybrid FT-ICR MS cell for larger radius excitation 1427
detection is typically achieved at an ICR orbital radius of
30–60% of the cell radius.34 For the present simulations, we
have chosen to evaluate the electrostatic potential inside the
ICR trap at a radial displacement corresponding to ions
excited to 33% and 60% of trap radius (Figs. 2–5) (We
calculate a typical excitation radius of 33% in our Bruker
Apex FT-ICR instrument.) Achieving optimal post-excitation
radius is important to improve sensitivity and accuracy in
representing isotopic distributions.33
In a perfectly quadrupolar electrostatic trapping potential,
the radial electric field increases linearly with r but is
independent of z, so that d2F/dzdr is zero throughout the
trap. Plots ofF, dF/dr, and d2F/dzdr as a function of z thus
provide a simple graphical basis for judging departures from
Figure 5. Simulated potential (top), radial elec
z-gradient (bottom) at 33% (left column) and 6
proposed Hybrid ICR trap. Potentials on the comp
electrodes are indicated. Red dashes indicate th
and d2F/drdz. This figure is available in color on
rcm.
Copyright # 2008 John Wiley & Sons, Ltd.
the ideal quadrupolar trapping potential. The ideal distri-
butions of dF/dr and d2F/drdz are noted with red dashes to
make the point graphically. Figures 2 and 3 show such plots
for two conventional trap configurations (closed right
circular cylinder of 1:1 aspect ratio and three equal-length
open cylinders), for an experimentally typical end cap
potential of 1V. For both conventional traps, the radial
electric field deviates more from the ideal quadrupolar form
as the distance from the trap central axis increases. Note the
much larger maximal magnitude of d2F/dzdr at 60% of the
cell radius than at 33%. That deviation helps to account for
reduced ion stability (and thus greater FT-ICR mass spectral
peak width and lower resolving power) at higher post-
excitation ICR orbital radius.39,40
tric field (middle), and radial electric field
0% (right column) of the trap radius in a
ensation, end cap, and center compensation
e ideal (quadrupolar) distributions of dF/dr
line at www.interscience.wiley.com/journal/
Rapid Commun. Mass Spectrom. 2008; 22: 1423–1429
DOI: 10.1002/rcm
1428 S. Kim et al.
Electrostatic potential field during detection inthe hybrid cellThe trap potential well depth of the open cylindrical cell
(Fig. 3) is different from that of the closed cell (Fig. 2). In
general, the trapping potential is smaller in the center of the
cell for the open cylindrical cell than for the closed cell.28
Therefore, for the most direct comparison between hybrid
and conventional traps, we adjusted the voltages applied to
the outer and center compensation electrodes so that the
depth of the potential well of the hybrid cell matches that of
the closed cell (Fig. 4) or the open cylindrical cell (Fig. 5), for
33% and 60% radius, while maintaining the end cap voltage
at 1V. Although the radial potential gradient (dF/dr) near
the trap axial center (�10mm< z <10mm) and 33% of the
cell radius is as flat as those of the conventional traps
(compare middle left panels in Figs. 2 and 4, or 3 and 5),
the radial electric field z-gradient at 60% of the cell
radius is much flatter for the hybrid trap than for the
conventional traps (compare middle right panels in Figs. 2
and 4, or 3 and 5). These simulations clearly show that
the radial potential distribution at large post-excitation
ICR orbital radius can be improved with the hybrid cell
configuration.
Figure 6. Isopotential contours (left) and radi
of maximal radius for a hybrid trap, for each of
center compensation, and compensation ele
60% trap diameter are presented to indicate ex
color online at www.interscience.wiley.com/jo
Copyright # 2008 John Wiley & Sons, Ltd.
Radial electric field contribution for eachelectrode simulated from the principle ofsuperpositionBy the principle of superposition, the overall potential field
inside an ICR cell is the sum of the potential fields for each of
the individual electrodes. For example, for the hybrid cell,
the potential field at 60% of maximum cell radius (Fig. 5, top
right) is obtained as the sum of the potential fields for (a) 3V
on the compensation electrodes, 0V on the end cap
electrodes, and 0V on the center compensation electrodes
(Fig. 6, top left), (b) 0V on the compensation electrodes, 1V
on the end caps, and 0V on the center compensation
electrodes (Fig. 6, middle left), and (c) 0V on the
compensation electrodes, 0V on the end caps, and –70V
on the center compensation electrodes (Fig. 6, bottom left).
The arrows in the figure represent a 60% excitation radius.
Ideally, d2F/dzdr should be zero at all values of z inside
the trap. Separating the potential (and radial electric field and
radial electric field z-gradient) components for each paired
set of electrodes helps in deciding how to scale the various
electrode voltages so as to minimize d2F/dzdr at a given
post-excitation ICR orbital radius. The compensation elec-
trodes of the hybrid trap generate a radial electric field
al electric field z-gradients (right) at 60%
three indicated combinations of end cap,
ctrode voltages. The arrows indicating
citation radius. This figure is available in
urnal/rcm.
Rapid Commun. Mass Spectrom. 2008; 22: 1423–1429
DOI: 10.1002/rcm
Hybrid FT-ICR MS cell for larger radius excitation 1429
z-profile similar to that of the open cylindrical or closed cell
(compare top right in Fig. 6 with bottom left in Figs. 2 and 3).
The overall design of the end cap electrodes with grounded
compensation electrodes (middle left in Fig. 6) is similar to a
closed elongated cell.41 Therefore, it is not surprising that the
end cap electrodes provide a radial electric field z-profile
with the same sign as the closed cell (middle right in Fig. 6).
The negative voltage at the center compensation electrodes
plays a crucial role in reducing radial electric field
z-gradients by providing the z-profile with the opposite
sign (bottom in Fig. 6). It cancels out the combined radial
electric field z-gradients provided by other electrodes. The
role of negative voltage at the center compensation electrode
can partially explain what was observed in the previous
studies.29,35 A beam of electrons at the center of a trap or
negative voltage at the sidekick electrodes can reduce the
radial electric field z-gradients and thus idealize the trapping
potential as is presented in Fig. 6.
In this study, a relatively large negative voltage (�70V in
Fig. 6) at the center compensation electrodes is required to
idealize the trapping potential at 60% excitation radius. The
negative voltage can reduce the trapping potential well
depth, which could result in ion loss. The reduced trapping
field is in turn compensated for by a combination of
potentials at the compensation and end cap electrodes.
The combination of the three sets of electrodes changes the
overall potential field to more closely approach the desired
three-dimensional axial quadrupolar potential at large ion
cyclotron post-excitation radius, without reducing the depth
of the axial trapping potential well at that radius.
CONCLUSIONS
A new cell (‘hybrid cell’) for FT-ICR MS is presented and
characterized by Simion simulation. The hybrid cell
combines the designs of the Infinity,19 compensated,31 and
capacitively coupled open cylindrical cell28 to achieve a
uniform rf excitation electric field. Comparison with
conventional cells with the same end cap potential shows
that a more ideal trapping potential can be achieved with the
hybrid cell at larger excitation radius, because three degrees
of freedom are available in the hybrid cell to achieve ideal
trapping potential compared to just one for the conventional
cells. Analysis from the principle of superposition shows that
the extra degrees of freedom provided by compensation
electrodes are critical to achieve a better trapping potential
distribution. The voltages applied to the hybrid cell in
Figs. 4–6 were chosen empirically by trial and error.
However, a more systematic method could be developed
to further optimize the result. Future efforts will be directed
at building and testing the hybrid cell experimentally, with
particular emphasis on maximizing the post-excitation ICR
orbital and its impact on detection sensitivity, mass accuracy,
and m/z discrimination.
AcknowledgementsThis work was supported by KBSI project ‘‘Development of
an FT-ICR instrument’’, the NSF National High Field FT-ICR
Mass Spectrometry Facility (DMR-06-54118), Florida State
University, and the National High Magnetic Field Labora-
tory in Tallahassee, FL, USA.
Copyright # 2008 John Wiley & Sons, Ltd.
REFERENCES
1. Comisarow MB, Marshall AG. Chem. Phys. Lett. 1974; 25:282.
2. Marshall AG, Hendrickson CL, Jackson GS. Mass Spectrom.Rev. 1998; 17: 1.
3. He F, Hendrickson CL, Marshall AG. Anal. Chem. 2001; 73:647.
4. Kim S, Rodgers RP, Marshall AG. Int. J. Mass Spectrom. 2006;251: 260.
5. Zhang LK, Rempel D, Pramanik BN, Gross ML. Mass Spec-trom. Rev. 2005; 24: 286.
6. Kelleher NL. Anal. Chem. 2004; 76: 196A.7. He F, Emmett MR, Hakansson K, Hendrickson CL, Marshall
AG. J. Protein Res. 2004; 3: 61.8. McLafferty FW. Int. J. Mass Spectrom. 2001; 212: 81.9. Masselon C, Anderson GA, Harkewicz R, Bruce JE, Pasa-
Tolic L, Smith RD. Anal. Chem. 2000; 72: 1918.10. Hughey CA, Rodgers RP, Marshall AG.Anal. Chem. 2002; 74:
4145.11. Kim S, Kramer RW, Hatcher PG. Anal. Chem. 2003; 75:
5336.12. Kim S, Kaplan LA, Hatcher PG. Limnol. Oceanogr. 2006; 51:
1054.13. Penning FM. Physica 1936; 3: 873.14. Brown LS, Gabrielse G. Rev. Mod. Phys. 1986; 58: 233.15. Horvath GZK, Hernandez-Pozos JL, Dholakia K, Rink J,
Segal DM, Thompson RC. Phys. Rev. A 1998; 57: 1944.16. Gabrielse G, Dehmelt H, Kells W. Phys. Rev. Lett. 1985; 54:
537.17. Guan SH, Li GZ, Marshall AG. Int. J. Mass Spectrom. 1997;
167: 185.18. Mitchell DW, Smith RD. Phys. Rev. E 1995; 52: 4366.19. Caravatti P, Allemann M. Org. Mass Spectrom. 1991; 26:
514.20. Franzen J, Nikolaev E. US Patent Application 20060027743,
2006.21. Gillig KJ, Bluhm BK, Russell DH. Int. J. Mass Spectrom. Ion
Processes 1996; 158: 129.22. Guan SH, Marshall AG. Int. J. Mass Spectrom. Ion Processes
1995; 146: 261.23. Gooden JK, Rempel DL, GrossML. J. Am. Soc. Mass Spectrom.
2004; 15: 1109.24. Naito Y, Fujiwara M, Inoue M. Int. J. Mass Spectrom. Ion
Processes 1992; 120: 179.25. Vartanian VH, Hadjarab F, Laude DA. Int. J. Mass Spectrom.
1995; 151: 175.26. Gabrielse G. Phys. Rev. A 1983; 27: 2277.27. Misharin AS, Zubarev RA. Rapid Commun. Mass Spectrom.
2006; 20: 3223.28. Beu SC, David A, Laude J. Anal. Chem. 1992; 64: 177.29. Kim S, Choi MC, Kim S, Hur M, Kim HS, Yoo JS, Blakney
GT, Hendrickson CL, Marshall AG. Anal. Chem. 2007; 79:3575.
30. Wang M, Marshall AG. Anal. Chem. 1989; 61: 1288.31. Gabrielse G, Mackintosh FC. Int. J. Mass Spectrom. Ion Pro-
cesses 1984; 57: 1.32. Rempel DL, Grese RP, Gross ML. Int. J. Mass Spectrom. Ion
Processes 1990; 100: 381.33. Hawkridge AM, Nepomuceno AI, Lovik SL, Mason CJ,
Muddiman DC. Rapid Commun. Mass Spectrom. 2005; 19:915.
34. Gorshkov MV, Marshall AG, Nikolaev EN. J. Am. Soc. MassSpectrom. 1993; 4: 855.
35. Kaiser NK, Bruce JE. Anal. Chem. 2005; 77: 5973.36. SIMION v.7.0. Scientific Instrument Services Inc.: Ringoes,
NJ, USA.37. Grosshans PB, Shields PJ, Marshall AG. J. Am. Chem. Soc.
1990; 112: 1275.38. Nikolaev EN, Heeren RMA, Popov AM, Pozdneev AV,
Chingin KS. Rapid Commun. Mass Spectrom. 2007; 21:3527.
39. Paasche P, Angelescu C, Ananthamurthy S, Biswas D,Valenzuela T, Werth G. Eur. Phys. J. D 2003; 22: 183.
40. Hubner K, Klein H, Lichtenberg C, Marx G, Werth G.Europhysics Lett. 1997; 37: 459.
41. Campbell VL, Guan ZQ, Vartanian VH, Laude DA. Anal.Chem. 1995; 67: 420.
Rapid Commun. Mass Spectrom. 2008; 22: 1423–1429
DOI: 10.1002/rcm