Tandem Mass Spectrometry in an Electrostatic Linear Ion Trap

Aug 11, 2014 - Fourier-Transform MS and Closed-Path Multireflection Time-of-Flight MS Using an Electrostatic Linear Ion Trap. Eric T. Dziekonski , Jos...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Tandem Mass Spectrometry in an Electrostatic Linear Ion Trap Modified for Surface-Induced Dissociation Ryan T. Hilger, Robert E. Santini, and Scott A. McLuckey* Department of Chemistry, Purdue University, West Lafayette, Indiana 47907-2084, United States ABSTRACT: A variety of ion traps are used in mass spectrometry. A key feature shared by most of them is the ability to perform tandem mass spectrometry (MS/MS). The Orbitrap is perhaps the most notable ion trap in which MS/ MS has yet to be performed. An electrostatic linear ion trap (ELIT) is analogous to an orbitrap in that ions are trapped using solely electrostatic fields. However, the relatively simple ion motion within an ELIT facilitates analysis of fragment ions produced within the device. In this report, we describe an ELIT to which we have added a target for surface induced dissociation (SID). When combined with our previously described method for isolating a precursor ion trapped in an ELIT,1 this apparatus enables MS/MS to be performed. Measurement of product ion m/z is facilitated by the fact that the ELIT is isochronous over the energy range of 1850−2000 eV so that changes to ion energy during SID do not cause major m/z shifts. We demonstrate MS/MS by isolating and dissociating each compound in a four component mixture of tetraalkylphosphonium cations. We also discuss the optimization of collision energy and the length of time that the SID target is available for collision, two parameters that are important in the performance of these experiments.

T

andem mass spectrometry2 (MS/MS) is the process of isolating precursor ions of a specific mass-to-charge ratio (m/z), subjecting those ions to a physiochemical process that results in changes in mass, charge, or both, and then measuring the m/z of the ions that are produced. Tandem mass spectrometry is used extensively to identify compounds as well as to derive information about the physiochemical properties of their related ions. Precursor ion isolation and processing are usually performed using a quadrupole mass filter or a quadrupole ion trap. Product ions may be analyzed by other means that offer higher resolution, such as time-of-flight (TOF), Orbitrap, or Fourier transform ion cyclotron resonance (FTICR) MS. Many analyzers, such as the quadrupole ion trap, TOF (in the form of a TOF/TOF instrument), and FTICR permit isolation, processing, and product ion analysis to all be performed within the same device. Notably absent from this list is the Orbitrap. While precursor ion isolation has been performed using an Orbitrap,3,4 ion activation and product ion analysis have not, presumably because of the difficulty inherent in recapturing the product ions in the electrostatic field in a manner that produces coherent motion. It might be possible to eject the isolated precursor ions out of the Orbitrap to an external device for processing and reinjection, but this has yet to be demonstrated. An electrostatic linear ion trap (ELIT) is an analyzer consisting of a field free region with a reflectron on either side.5−7 Ions are injected into the device and “bounce” backand-forth between the reflectrons while an image charge pickup electrode in the center of the trap records the time domain signal. The signal is then Fourier transformed to determine the frequencies present, which are proportional to the inverse square root of m/z. Like the Orbitrap, m/z analysis is achieved using electrostatic fields; however, in an ELIT, the ion velocities © 2014 American Chemical Society

are overwhelmingly axial. This facilitates ejection of ions for isolation and processing in external devices, as has been shown previously by ourselves8 as well as by other groups.9 Methods also exist for isolating ions residing within an ELIT.1,10 However, previous attempts at processing ions inside an ELIT and analyzing the products have been relatively crude. Dugourd and co-workers irradiated single ions trapped in an ELIT with an infrared laser and were able to observe dissociation events.11−13 However, they were not able to directly measure the m/z of the products because the kinetic energy loss associated with the fragmentation event was unknown. Additionally, the product ions exhibited limited stability probably as a result of a significant kinetic energy change. Fragmentation and product ion analysis have been performed within an electrostatic storage ring.14 In these experiments, precursor ions were selected by a magnetic sector external to the storage ring and photodissociated while inside the storage ring using a laser pulse. However, in an electrostatic storage ring, the precursor ion’s kinetic energy is always high, so the kinetic energy partitioning that occurs upon fragmentation yields product ions with m/z dependent kinetic energies that are very different from that of the precursor ion. Therefore, storage of product ions with a particular m/z required adjustments to the potentials applied to the ring’s various optics. Recording a product ion spectrum therefore required that these potentials be scanned. In order to avoid large and unpredictable changes to ion kinetic energy, we have engineered our ELIT to process ions Received: June 10, 2014 Accepted: August 11, 2014 Published: August 11, 2014 8822

dx.doi.org/10.1021/ac502143p | Anal. Chem. 2014, 86, 8822−8828

Analytical Chemistry

Article

using surface induced dissociation (SID).15 By positioning the target surface at the back of a reflectron, the ion-surface collision occurs after most of the ion’s kinetic energy has been converted to potential energy by the reflectron’s field. Therefore, the total energy (kinetic plus potential) of the product ions formed in the vicinity of the surface via a prompt dissociation, which would be consistent with the shattering mechanism,16,17 is primarily determined by the electrostatic potential near the surface and to a much lesser extent by the energy partitioning that occurs during the collision. It should be noted that an energy change still occurs upon collision, as prior work has indicated that fragment ions retain only about 10% of the precursor ion’s kinetic energy.18−20 Therefore, correct m/z measurement of product ions requires that the ELIT be isochronous over the energy range exhibited by the product ions, a feature possessed by our ELIT, as discussed below. It should also be noted that SID products that do not form near the surface probably undergo large energy changes that result in unstable trajectories and are therefore not observed, which occurs in the storage ring experiments when the potentials are not changed.14 Using our apparatus, we are able to measure the m/z of all product ions simultaneously using the same potentials that are used for precursor ion analysis. In this report, we have combined our previously reported method for isolating ions in an ELIT1 with SID and subsequent analysis of product ions in order to perform true MS/MS in an ELIT. We describe our apparatus and use ion-motion simulations to demonstrate that our ELIT is isochronous over the required energy range. We demonstrate the instrument’s capabilities by performing MS/MS on each component of a mixture of tetraalkylphosphonium cations. We also discuss factors that must be considered when performing these experiments including the collision energy and the timing of the voltage pulse applied to the SID surface that makes it available for collision. To our knowledge, this is the first report of MS/MS being performed entirely within an ELIT.

Figure 1. Cartoon (not to scale) depicting the ion path of the mass spectrometer; additional details are available elsewhere.8

acquired for 5−6 ms after the collision, whereas for m/z analysis without a collision, signal was acquired for 30 ms. These acquisition times were sufficient for the signal to decay into the noise. The time domain signals were zero filled to double their original length before being fast Fourier transformed. All of the spectra presented below are frequency domain averages of 100 acquisitions. The SID surface consisted of a polished gold disk (5 mm outer diameter, 4 mm thick, Pine Research Instrumentation, Durham, NC). The surface was prepared by first rinsing it with ethanol followed by sonication for 10 min in ultrasonic cleaning solution and then sonication for 10 min in ethanol. The surface was then rinsed with water. Nanoscale polishing of the surface was performed by placing the surface in 1 mM Fe2+, 10 mM H2O2, and 1 mM EDTA in 10 mM acetate buffer (pH = 4.7) for 1 h followed by rinsing with water.21 Oxides on the surface were then reduced by placing the surface in 0.5 M NaBH4 in 1:1 ethanol/water for 10 min followed by rinsing with water then ethanol.22 Finally, a self-assembled monolayer (SAM) was added to the surface by placing the surface in a 1 mM solution of 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro-1-decanethiol in ethanol for 24 h. The surface was then sonicated in ethanol for 5 min three times. The surface was then allowed to air-dry before being mounted 2 mm behind the final electrode of the rear reflectron (see Figure 1) to which it is electrically connected using a wire. Ions were produced by nanoelectrospray and accumulated in a collision cell. The ions were then pulsed out of the collision cell through an electrostatic deflector. A voltage pulse with a width of 2 μs was applied to the deflector in order to limit the width of the ion pulse entering the ELIT. Precursor ions of the desired m/z were then selected by applying a 600 V square wave with frequency and phase matched to the m/z of interest for about 500 μs as described previously.1 Then, the potential on the SID surface was reduced using a fast, high voltage switch (PVX-4140, Directed Energy, Fort Collins, CO) in order to allow the ions to collide with the surface. Since the ions are injected with a nominal energy of 2000 eV, the collision energy was taken to be the difference between 2000 eV and the voltage applied to the SID surface during the collision. The timing of the voltage pulse to the SID surface as well as the selection of

■

EXPERIMENTAL SECTION Materials. Ethylenediaminetetraacetic acid (EDTA), 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro-1-decanethiol, sodium borohydride, ammonium iron(II) sulfate hexahydrate, tetraphenylphosphonium bromide, benzyltriethylphosphonium bromide, and tetrabutylphosphonium bromide were purchased from Sigma-Aldrich (St. Louis, MO). Glacial acetic acid, sodium hydroxide, hydrogen peroxide (30% w/w), and methanol were purchased from Macron Fine Chemicals (Center Valley, PA). Ethanol was purchased from Decon Laboratories (King of Prussia, PA). Tetraethylphosphonium bromide was purchased from Tokyo Chemical Industry (Tokyo, Japan). Ultrasonic cleaning solution was purchased from Fisher Scientific (Pittsburgh, PA). Water was purified with a water purifier (D8961, Barnstead, Dubuque, IA) prior to use. Mass Spectrometry. The mass spectrometer used herein is shown in Figure 1 and has been described previously8 but has undergone slight modifications. The image charge detection circuit described in the previous publication was replaced with a new circuit based on the PC250 test board for the A250 charge sensitive preamplifier (Amptek, Bedford, MA). The input JFET was moved inside the vacuum chamber next to the pickup electrode, and the test board was attached to the atmospheric side of the vacuum flange. The output of the test board was sampled at 10 MHz by the digitizer card which was AC coupled with an input impedance of 1 MΩ. For SID spectra, signal was 8823

dx.doi.org/10.1021/ac502143p | Anal. Chem. 2014, 86, 8822−8828

Analytical Chemistry

Article

the collision energy are discussed further below. After the collision, the voltage on the SID surface was returned to the typical value for m/z analysis. The potentials used on the reflectron electrodes were 0, −2070, −1914, 0, 970, 1700, and 2445 V going from the field free region to the outermost electrode of the reflectron. Ion Motion Simulations. A model of the ELIT including the SID surface was constructed in SIMION 8.0 (Scientific Instrument Services, Ringoes, NJ). Simulations of product ion motion were conducted by setting the potentials of all electrodes to the values present when a singly charged precursor ion collides with the surface with a nominal kinetic energy of 150 eV. Product ions (mass = 100 Da, charge = +1e where e = the electron’s charge) were initially placed 0.00625 in. in front of the surface and heading toward the center of the ELIT. The ion’s kinetic energy was varied between 0 and 150 eV in order to represent an energy loss of between 0% and 100%. Ion motion was allowed to proceed until the ion had reached the center of the field free region, at which time the potentials on the SID surface and the adjacent reflectron electrode were changed to the typical value for ion trapping and analysis. Ion motion was then allowed to continue for a total of 5 ms, and a simulated time domain signal was generated by monitoring the time periods during which the ion was present within a 2 cm detection volume in the center of the field free region. A total of 100 ions was simulated at each kinetic energy value. The initial position of each ion was randomly located within a disk (radius of 0.5 mm) located 0.00625 in. in front of the SID surface. The signals from all 100 ions were then summed, and the result was Fourier transformed to create a simulated frequency spectrum. A similar simulation of ion motion in a typical analysis (no surface collision) was conducted by starting the ions in the center of the field free region with a total energy of 2000 eV, the nominal energy of ions injected into our ELIT.

Figure 2. Simulated frequency spectra of ions produced at the SID surface with kinetic energy between 0 and 150 eV. Also shown is a simulated frequency spectrum of similar ions that are started in the center of the ELIT and do not undergo a surface collision. The spectra show that variations in kinetic energy result in a frequency shift of no more than 200 Hz or 0.1 m/z. Therefore, the fragment ions are expected to be observed at the correct nominal m/z.

position observed in the absence of a collision. A 200 Hz shift corresponds to about 0.1 m/z at this frequency (m/z = 100). Figure 2 demonstrates that the product ions should be observed at their correct nominal m/z values despite any energy loss that may occur as a result of the surface collision, at least up to a collision energy of 150 eV. MS/MS was performed on a sample containing tetraethylphosphonium (TEP), benzyltriethylphosphonium (BTEP), tetrabutylphosphonium (TBP), and tetraphenylphosphonium (TPP) cations. Figure 3 shows a mass spectrum of the mixture as well as spectra obtained after isolation of each component. It

■

RESULTS AND DISCUSSION As mentioned above, fragment ions produced by SID will have different total energy (kinetic plus potential) than their precursors because of energy partitioning that occurs during the collision and dissociation processes. The question therefore arises as to how these energy changes will affect the measured ion frequencies. Ideally, the frequencies would depend only on m/z and be independent of energy over the relevant range. Failure to achieve this criterion would result in product ion spectra in which the peak shapes reflect the energy distributions of the fragments, or spectra in which peak shape and position are dependent upon the collision energy. This would result in poor resolution as well as the need to calibrate product ion spectra at each collision energy. We simulated the motion of fragment ions with various energies formed at the SID target in order to examine the effect of ion energy on the measured frequency. If an ion has a kinetic energy of 150 eV prior to collision with the surface (similar to the collision energy for most experiments reported herein), the product ion must be formed with a kinetic energy between 0 and 150 eV. Therefore, we simulated the trajectories of product ions imparted with various kinetic energies within this range (100 ions for each kinetic energy). The resulting frequency spectra are shown in Figure 2, along with a simulated frequency spectrum from 100 ions that did not undergo a collision with the surface. Figure 2 demonstrates that the energy change causes the peak to shift by no more than 200 Hz from the

Figure 3. (A) Spectrum of the tetraalkylphosphonium cation mixture on which MS/MS was performed. (B) Spectra obtained after isolation of each component in the mixture. 8824

dx.doi.org/10.1021/ac502143p | Anal. Chem. 2014, 86, 8822−8828

Analytical Chemistry

Article

Figure 4. Spectrum obtained after SID (collision energy of 140 eV) of the mixture of TEP, BTEP, TBP, and TPP. Inelastically scattered precursor ions are identified while fragments are labeled with their nominal m/z. Harmonics of intense peaks are labeled with asterisks.

Figure 5. Spectra obtained after SID (collision energy of 140 eV) of (A) isolated TEP, (B) isolated BTEP, (C) isolated TBP, and (D) isolated TPP. Inelastically scattered precursor ions are identified while fragments are labeled with their nominal m/z. Harmonics of intense peaks are labeled with asterisks. These spectra demonstrate that isolation of each component enables determination of the fragments produced from each precursor.

can be seen that the isolations are very clean so that product ions are assured to arise from the intended precursor ion. Figure 4 shows the spectrum obtained after performing SID at an energy of 140 eV on the mixture, and Figure 5A−D provides the spectra after performing SID on each isolated component. Inelastically scattered precursor ions are identified, and product ions are labeled with their nominal m/z. Harmonics of intense signals are often visible and are labeled with asterisks. Table 1 lists the observed fragment m/z values along with their assigned elemental compositions. We are confident that the measured m/z values are correct because of the simulation results shown in Figure 2. Additional evidence that the m/z calibration is correct comes from the fact that the precursor ion m/z values are correct and that the product ions overwhelmingly appear within ±0.1 of a nominal m/z value, as expected given the atomic compositions of the precursors. It can be seen that the SID spectrum of the mixture shown in Figure 4 contains peaks from fragments produced from the various precursors. Figure 5 demonstrates a hallmark of MS/ MS: the ability to determine which fragments arise from each precursor ion in a mixture. Unfortunately, TEP gives very weak

Table 1. Assigned Elemental Compositions of the Fragment Ions Labeled in Figures 4 and 5 nominal m/z

elemental composition

nominal m/z

elemental composition

41 48 59 61 62 65 75 76 89 90 91 103 104

C3H5 CH5P C2H4P C2H6P C2H7P C5H5 C3H8P C3H9P C4H10P C4H11P C7H7 C5H12P C5H13P

108 117 118 131 145 173 183 187 201 203 215 261

C6H5P C6H14P C6H15P C7H16P C8H18P C10H22P C12H8P C11H24P C12H26P C12H28P C13H28P C18H14P

product ion signals and fragments with identical m/z are produced by several other components. However, Figure 5B demonstrates that the peak at m/z = 91 is characteristic of 8825

dx.doi.org/10.1021/ac502143p | Anal. Chem. 2014, 86, 8822−8828

Analytical Chemistry

Article

Figure 6. Spectra obtained after performing SID of TBP at various collision energies. In (A), ions are simply reflected leading to an intense precursor ion signal and much weaker signals from contaminants. At 60 eV, fragment ions begin to appear as shown in (B). 140 eV provides the highest signals for fragment ions as shown in (C). Increasing the collision energy to 180 eV simply results in decreased intensities for all ions as shown in (D).

Figure 7. TBP product ion spectra recorded while varying the width of the pulse that makes the surface available for collision (see Figure 6C for peak labels). Similar product ion spectra are produced for pulse widths between 8.5 and 12 μs. Discrimination against high m/z and low m/z products is apparent at lower and higher pulse widths, respectively (see text for details).

produced from multiple precursors, which is expected given the structural similarities of certain precursors. Figure 6 illustrates the effect of varying the collision energy when performing SID of TBP. At 0 eV, large signals from the

BTEP. Figure 5C also contains signals from many fragment ions unique to TBP, and Figure 5D demonstrates that the signals at m/z = 108 and 183 are unique to TPP. Many of the fragments are unique to specific precursors. A few fragments are 8826

dx.doi.org/10.1021/ac502143p | Anal. Chem. 2014, 86, 8822−8828

Analytical Chemistry

Article

discussed phenomenon of the lowest m/z products returning for a second collision with the surface. If the pulse width is increased to 13 μs (Figure 7D), the discrimination against low m/z products becomes more dramatic as ions of increasing m/z are able to return for a second collision. Further increases to the pulse width result in additional signal loss (data not shown).

precursor are present along with some weak signals from contaminants. It should be noted that the precursor ions are probably being reflected rather than grazing the surface because the initial energy of the ions is actually slightly below 2000 eV. At 60 eV, fragment ions at m/z = 62, 76, 90, and 131 begin to appear. At 140 eV, the fragment ions reach their maximum intensity and significant depletion of the precursor ions has occurred. Further increasing the collision energy primarily depletes the signals from all ions, as shown by the spectrum recorded after SID at 180 eV. Similar results were obtained for the other components of the tetraalkylphosphonium mixture, with the exception of TEP which showed slightly higher fragment ion intensities at a collision energy of 70 eV (data not shown). Nonetheless, a collision energy of 140 eV was selected as optimal. The spectrum in Figure 6C is almost identical to that shown in Figure 5C. This is noteworthy because the data in Figure 6 were collected using a sample that did not contain TEP, BTEP, or TPP. Therefore, the isolation of TBP prior to acquiring the data for Figure 5C was highly effective. Furthermore, Figure 6 demonstrates that shifts in peak position as a function of collision energy are indeed small as was suggested by the simulation data shown in Figure 2. Another parameter that must be optimized when performing these experiments is the length of time over which the SID target is made available for collision. The time must be long enough for the entire packet of precursor ions to collide with the surface or a large peak from undissociated precursor ions will result. However, if the time is excessive, fragments that are formed from the first precursors to hit the surface will have time to travel to the opposite end of the ELIT and return for a second collision, which results simply in ion loss for the cases we have studied. Since the fragment ions are typically of low m/ z, they move very fast. For example, an ion with m/z = 75 will complete this journey in about 4 μs. It is for this reason that the deflector is used to limit the width of the precursor ion packet to 2 us. This limits the time required for all of the precursor ions to collide with the surface. Figure 7 shows spectra obtained when the surface is available for collision for various lengths of time after the precursor ions are ejected from the collision cell. It should be noted that these data were collected using a sample of TBP that did not contain TEP, BTEP, or TPP; therefore, no isolation was performed prior to SID. The precursor ions were simply ejected from the collision cell and allowed to collide with the surface, and the resulting product ions were then analyzed. The times shown in the figure represent the elapsed time since the ions were ejected from the collision cell, a significant portion of which the ions spend traveling from the collision cell to the surface. At a pulse width of 8 μs (Figure 7A), fragment ions are detected, but discrimination in favor of lower m/z products is apparent. This discrimination occurs because the higher m/z, slower moving products (along with any remaining incoming precursor ions) are still in the vicinity of the surface when the potential changes from the collision value to the trapping value. This increase in potential dramatically increases the energies of nearby ions, which results in their loss. When the pulse width is increased to 8.5 μs (Figure 7B), the discrimination has disappeared because the higher m/z ions have had time to travel away from the surface before the end of the pulse. The spectrum remains unchanged (data not shown) until the pulse width reaches 12 μs (Figure 7C), when the abundance values of the lowest m/z products start to decrease (the peak at m/z = 65 is a harmonic of the peak at m/z = 259). This results from the previously

■

CONCLUSIONS We have demonstrated the performance of tandem mass spectrometry entirely within an electrostatic linear ion trap. This extends the demonstrated functionality of ELIT technology. Precursor ions were isolated using a high voltage square wave applied to a reflectron lens, as described previously.1 The isolated precursor ions were then subjected to surface induced dissociation using a target consisting of a fluorinated self-assembled monolayer on gold. Product ions recoiled back toward the center of the ELIT enabling their trapping and analysis. Despite variations in product ion energy that result from SID, shifts in peak position were minimal because the ELIT is energy-isochronous over the range of 1850−2000 eV, as demonstrated by simulations and experiments. Performance of MS/MS experiments of this type relies upon proper optimization of both the collision energy and the width of the voltage pulse applied to the SID target that makes it available for collision. For our singly charged tetraalkylphosphonium cations, a collision energy of 140 eV maximized fragment ion abundance. The voltage pulse applied to the surface must be long enough to allow all of the precursor ions to hit the surface and for the product ions to escape the surface’s immediate vicinity. However, the voltage pulse should not be long enough to allow any product ions to return for a second collision.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: (765) 494-5270. Fax: (765) 494-0239. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by AB Sciex. We thank Randy Replogle, Rob Oglesbee, Phil Wyss, Cathy McIntyre, Bob Fagan, and Dr. Hartmut Hedderich of the Jonathan Amy Facility for Chemical Instrumentation for their help with construction of the mass spectrometer. We also acknowledge Mircea Guna and Dr. James W. Hager of AB Sciex for helpful discussions and for providing the collision cell. We also thank Prof. Vicki Wysocki and Dr. Julia Laskin for advice regarding SID.

■

REFERENCES

(1) Hilger, R. T.; Santini, R. E.; McLuckey, S. A. Int. J. Mass Spectrom. 2014, 362, 1−8. (2) Gross, J. H. Mass Spectrometry: A Textbook, 2nd ed.; Springer: New York, 2011; pp 415−478. (3) Hu, Q.; Makarov, A. A.; Cooks, R. G.; Noll, R. J. J. Phys. Chem. A 2006, 110, 2682−2689. (4) Hu, Q.; Cooks, R. G.; Noll, R. J. J. Am. Soc. Mass Spectrom. 2007, 18, 980−983. (5) Wollnik, H.; Przewloka, M. Int. J. Mass Spectrom. Ion Processes 1990, 96, 267−274. (6) Benner, W. H. Anal. Chem. 1997, 69, 4162−4168. 8827

dx.doi.org/10.1021/ac502143p | Anal. Chem. 2014, 86, 8822−8828

Analytical Chemistry

Article

(7) Zajfman, D.; Heber, O.; Vejby-Christensen, L.; Ben-Itzhak, I.; Rappaport, M.; Fishman, R.; Dahan, M. Phys. Rev. A 1997, 55, R1577− R1580. (8) Hilger, R. T.; Santini, R. E.; McLuckey, S. A. Anal. Chem. 2013, 85, 5226−5232. (9) Plass, W. R.; Dickel, T.; Scheidenberger, C. Int. J. Mass Spectrom. 2013, 349−350, 134−144. (10) Toker, Y.; Altstein, N.; Aviv, O.; Rappaport, M. L.; Heber, O.; Schwalm, D.; Strasser, D.; Zajfman, D. J. Instrum. 2009, 4, P09001. (11) Doussineau, T.; Bao, C. Y.; Clavier, C.; Dagany, X.; Kerleroux, M.; Antoine, R.; Dugourd, P. Rev. Sci. Instrum. 2011, 82, 084104. (12) Doussineau, T.; Antoine, R.; Santacreu, M.; Dugourd, P. Phys. Chem. Lett. 2012, 3, 2141−2145. (13) Antoine, R.; Doussineau, T.; Dugourd, P.; Calvo, F. Phys. Rev. A 2013, 87, 013435. (14) Støchkel, K.; Kadhane, U.; Andersen, J. U.; Holm, A. I. S.; Hvelplund, P.; Kirketerp, M.-B. S.; Larsen, M. K.; Lykkegaard, M. K.; Nielsen, S. B.; Panja, S.; Zettergren, H. Rev. Sci. Instrum. 2008, 79, 023107. (15) Dongré, A. R.; Somogyi, Á .; Wysocki, V. H. J. Mass Spectrom. 1996, 31, 339−350. (16) Raz, T.; Levine, R. D. J. Chem. Phys. 1996, 105, 8097−8102. (17) Laskin, J.; Bailey, T. H.; Futrell, J. H. J. Am. Chem. Soc. 2003, 125, 1625−1632. (18) Wörgötter, R.; Kubišta, J.; Ž abka, J.; Dolejšek, Z.; Märk, T. D.; Herman, Z. Int. J. Mass Spectrom. Ion Process. 1998, 174, 53−62. (19) Rakov, V. S.; Denisov, E. V.; Futrell, J. H.; Ridge, D. P. Int. J. Mass Spectrom. 2002, 213, 25−44. (20) Shukla, A. K.; Futrell, J. H.; Sen, A. D. J. Chem. Phys. 2003, 118, 11217−11225. (21) Nowicka, A. M.; Hasse, U.; Hermes, M.; Scholz, F. Angew. Chem., Int. Ed. 2010, 49, 1061−1063. (22) Yuan, M.; Zhan, S.; Zhou, X.; Liu, Y.; Feng, L.; Lin, Y.; Zhang, Z.; Hu, J. Langmuir 2008, 24, 8707−8710.

8828

dx.doi.org/10.1021/ac502143p | Anal. Chem. 2014, 86, 8822−8828