Surface-induced dissociation by Fourier transform mass spectrometry

ciation (SID) of Ions In a dual-cell Fourier transform mass spectrometer Is .... substance P (m/z 1348) consisting primarily of fragments with m/z < 6...
0 downloads 0 Views 658KB Size
Anal. Chem. 1990, 62, 1295-1299 (47) March, R. E.; Hughes, R. J. Ouahpok Stofage Mass Spectrometry; John W k y 8 Sons: New York, 1989. Vedel, F.; Andre, J. Phys. Rev. A : Gen. Phys. 1984, 2 9 , 2098. Vedel, F.; Andre, J. In!. J. Mass Spectrom. Ion Processes 1985, 65, 1. (50) Allli, A.; Andre, J.; Vedel, F. Phys. Scr. 1988, T22, 325. (51) McLean, M. A.; Freas, R. B. Anel. Chem. 1989, 61, 2054. (52) Wong, S. F.; Meng, C. K.; Fenn, J. B. J. Phys. Chem. 1988, 92, 546. (53) Flscher, E. 2.Phys. 1959. 156, 1. (54) Rettinghaus, V. von G. 2.Angew. Phys. 1967, 2 2 , 321. (55) Lehman, T. A.; Bursey, M. M. Ion Cyclotron Resonance Spectrometry; John Wlley 8 Sons: New York, 1976. (58) Syka. J. E. P.; Fies, W. J., Jr. Proceedings of the 35th ASMS Conference on Mass Spectrometry and Allied Topics, Denver, CO, 1987; p 767.

1295

(57) Roepstorff, P.; Fohlman, J. Biomed. h4ass Spectrom. 1984, 1 1 , 601. (58) Busch, K. L.; Gllsh, G. L.; McLuckey, S. A. Mass SpectrometrylMass Spectrometry : Techniques and Applications of Tandem Mass Spec trometry; VCH Publishers: New York, 1988. (59) McLuckey, S. A.; Glish, G. L.; Van Berkel, G. J., Oak Rkige National Laboratory, unpubllshed results, 1989.

-

RECEIVED for review January 25,1990. Accepted March 22, 1990. This research was sponsored by the United States Department of Energy Office of Basic Energy Sciences under Contract DE-AC05-840R21400 with Martin Marietta Energy Systems, Inc.

Surface- Induced Dissociation by Fourier Transform Mass Spectrometry Carl F. Ijames and Charles L. Wilkins* Department of Chemistry, University of California, Riverside, California 92521

A detalled procedure for performlng surface-Induced dlssoclatlon ( S I D ) of Ions In a dual-cell Fourier transform mass spectrometer Is described. It Is shown that the technlque Is appllcable to both electron Ionization and laser desorptlon measurements. S I D spectra of perfluorotrl-n-butylamine, anthracene, (5,10,15,20-tetraphenyl-21 H,23H-porphlnato)Iron( I I I ) chloride, and [5,10,15,20-tetrakis( 2,6-dlbromophenyl)-2lH,23HporpMnato]kon( 111) CMOrkle are presented. Converslon efflclencles of molecular Ions between 1% and 30% are obtained. It Is concluded the method hdds promise for dlssoclatlon of high mass laser-desorbed Ions.

INTRODUCTION A recent and very promising addition to the list of ion fragmentation techniques is surface-induced dissociation (SID),introduced by Cooks and co-workers in 1985 (I). The technique consists of accelerating parent ions into a metal target and then mass analyzing the inelastically scattered daughter ions. In a series of papers, these workers have described implementation of SID using hybrid magnetic sector-quadrupole (I), tandem quadrupole (2), and tandem time-of-flight mass spectrometers (3). These instruments provide collision energies up to 300 eV for electron ionization-generated parent ions and up to unit daughter ion mass resolution. Conversion efficiencies, defined as the sum of daughter ion abundance9 divided by the parent ion abundance in the absence of SID, range from 2% to 15%. It should be noted that this common definition for multiple analyzer instruments exagerates the SID efficiency because it neglects substantial losses of parent ions in the second and subsequent analyzers. The stricter definition, which would use the parent ion abundance arriving at the SID site, gives less than 1% SID efficiency. Here, surface-induced dissociation with comparable efficiency using a Fourier transform mass spectrometer (FTMS), where there is no such distinction, is described. From an analytical standpoint, a major advantage of SID is its ability to convert a relatively high percentage of translational energy into internal energy. For example, 25 eV SID of iron pen0003-2700/90/0362-1295$02.50/0

tacarbonyl ions yields an average internal energy of approximately 4 eV while collisions at the same laboratory energy with argon gas deposit less than 2 eV, on average (4). The average internal energy increases with translational energy, reaching 8 eV at a collision energy of 140 eV. This corresponds approximately to the energy deposited by a 7-keV collision with an argon atom (4). It should also be noted that similar high efficiencies can be obtained under higher pressure, multiple collision conditions (e.g. collision induced dissociation (CID) using a triple quadrupole or ion trap mass spectrometer). A major limitation of conventional collisional activation tandem mass spectrometry (MS/MS) is that the yield of secondary ions decreases as the parent mass increases (5). This is a result of the increasing mismatch in parent and target masses and the greater number of vibrational modes available in the larger molecules, facilitating internal distribution of the deposited energy, rather than dissociation. SID has the potential to increase internal energy deposition and thus to increase the efficiency of MS/MS of high mass ions. Evidence for this is found in Aberth’s recent report of SID of parent ions produced by secondary ion mass spectrometry using a tandem Wien filter instrument (6). A microchannel plate in a grazing incidence configuration was employed as the target and collision energies extended up to lo00 eV. Both positive and negative spectra of leucine-enkephalin (molecular weight 555) were reported, with fragmentation similar to that produced by high-energy CID. Also, (CSI)~&S+was fragmented to yield (CsI),Cs+ ions from n = 0 to 23. These results show that SID is feasible with high mass ions at collision energies as low as 1 keV. Fourier transform mass spectrometry is a technique that offers several advantages over other types of mass analyzers. These advantages, which are well-known, include its extremely high mass resolution, the ability to acquire a complete spectrum following a single pulsed ionization event, and the capability of performing MS/MS analysis without adding additional mass analyzers (7).However, to achieve maximum resolution using FTMS, analyzer cell pressures of less than 1X Torr are required, while pressures of (0.1-1.0) X lo4 Torr are necessary for CID. Also, the maximum cyclotron 0 1990 American Chemical Society

1298

ANALYTICAL CHEMISTRY, VOL. 62, NO. 13, JULY 1, 1990

T R A P VOLTAGE IN

SID PULSE GENERATOR

I DIVIDE B Y 25

TRAP VOLTAGE C L -

t ANALOG MULTIPLEXER

HIGH VOLTAGE SET

MULTIPLY E Y

25

LOW VOLTAGE SET I

1

I

COUNTERS FOR HIGH VOLTAGE

%%?sA?~

1

I

CLOCKAND CLOCK AND TIMING rnhlTFlnl CONTROL

COUNTERSFOR COUNTERS FOR LOW VOLTAGE

1 Figure 2.

t PULSE WIDTH

TRIGGER IN

SOURCE CELL

CONDUCTANCE LIMIT

ANALYZER CELL

Ion trajectory to impact for surface-induced dissociation.

based on a 4-MHz clock oscillator, providing 250-11s resolution for voltage changes. The control circuit waits for the next clock cycle after the asynchronoustrigger input, resulting in jitter before application of the SID pulse but allowing precise timing of the pulse duration. Also, provision is made for a second voltage step after the SID pulse, before returning to the normal trap voltage. This allows for extra flexibility in deceleration of the daughter ions. However, this feature was not utilized in the experiments described here. Simulations. Electrostatic fields in the FTMS cell were modeled using SIMION V4.0 (available from Dave Dahl, Idaho National Engineering Laboratory, EG&G Idaho, Inc., Idaho Falls, ID). The cell was modeled as a cylinder of length 50 mm and diameter 50 mm, with a mesh spacing of 1mm. Calculations were performed with an IBM PC/AT computer. Procedure. Ions were formed in the source cell by either electron ionization or laser desorption, transferred to the analyzer cell through the 2-mm conductance limit, and stored for 100 ms (EI) or 8 s (LD) to allow for relaxation to the center of the analyzer cell and to allow laser-desorbed neutrals to be pumped away. Parent ions were isolated by an ejection sweep to remove fragment ions lower in mass. The cyclotron motion of the chosen parent ion was then excited by application of a pulse of radio frequency energy at the cyclotron frequency. After a 3-ms delay, a +25-V pulse was applied to the analyzer trap plate, causing the ions to strike the conductance limit and fragment (Figure 2). The duration of this pulse was set to 1.5 times the calculated flight time of the parent ion from the center of the cell to the conductance limit. The average z-axis translational energy of ions at impact was 1.3 eV (independent of mass). Ten milliseconds later the product ions were excited and detected in the usual manner. Typically, 1000 to 5000 scans were coadded for electron ionization spectra and 4 to 16 scans for laser desorption spectra.

, PULSE WIDTH

Figure 1. Block diagram of the surface-induced dissociation pulse

generator. radius to which an ion can be excited is fixed by the cell radius and, at constant radius, an ion's kinetic energy is inversely proportional to the mass. Thus, for FTMS, the efficiency of CID decreases with increasing mass more rapidly than for conventional instruments. Therefore, a technique such as SID, which imposes no additional gas load, appears to offer the greatest promise as a tool for MS/MS of large molecules using FTMS. To date, there have been two reports of SID using FTMS detection of daughter ions. McLafferty and co-workers, employing a dual cell FTMS, stated that "Results for toluene molecular ions ... are similar to those reported by Cooks"(8). McLafferty and co-workers also reported SID using an external quadrupole ion source to inject ions into an FTMS, to collide with a trap plate, producing daughter ion spectra from substance P (mlz 1348) consisting primarily of fragments with m / z < 600 (9). However, it is not entirely clear that SID was responsible for the daughter ions, because a higher absolute abundance of daughter ions than original parent ions was observed.

EXPERIMENTAL SECTION Mass Spectrometers. Two different Fourier transform mass spectrometers were utilized in this study. Electron ionization (EI) experiments were performed with a FTMS-1000 FT mass spectrometer (NicoletAnalytical Instruments, Madison, WI), upgraded with a standard Nicolet FTMS-2OOO dual cell (4.76 cm cubic cells) and chamber assembly, and operated at a magnetic field strength of 2.93 T. Laser desorption (LD) experiments employed a modified Nicolet FTMS-2000 system equipped with 4.76 cm cubic cells, C02laser (Tachisto, Boston, MA) and Autoprobe accessory for sample insertion and operated at a magnetic field strength of 7 T. Sample Introduction. Samples for electron ionization were introduced through either the batch inlet or direct insertion probe, to a pressure of (0.5-5.0) X lo-' Torr in the source region and less than 8 X lo4 Torr in the analyzer region. Electron energies of 10-20 eV were utilized to minimize fragmentation, and electron emission current was optimized for each sample to maximize the abundances of parent ions. Samples for laser desorption were prepared by dissolving approximately 1 mg of sample in 200 pL of a suitable solvent. Next, 10 pL of this solution was deposited on a stainless steel probe tip and the solvent evaporated before introduction into the mass spectrometer. Pressures were less than Torr, respectively, in the source and 5.0 X lo-* and 5 X analyzer regions. SID Electronics. A block diagram of a device that controls application of high-voltage pulses to the analyzer trap plate is shown in Figure 1. The key feature is the output amplifier which provides voltage changes of up to 285 V with a worst-case slew plus settling time of 3 ps. For the 25-V pulses used in these experiments, the total settling time is 1.5 ps. Timing control is

RESULTS AND DISCUSSION To perform SID with FTMS, it is necessary to accelerate ions trapped in the cell into a plate and then trap and analyze the daughter ions. An obvious candidate as a collision target for ions on the z axis in the cell is a trap plate. With a change in the potential of one of the two trap plates while the other is held a t the normal trap potential (1-3 V), ions in the cell can be either attracted toward the biased plate or repelled and pushed toward the other plate. Simulations showed that the penetration of the SID potential into the cell was surprisingly small. For example, using 3 and 250 V on the front and rear trap plates, respectively, the potential falls to 125 V only 10 mm from the plate held at 250 V and is 36 V in the center of the cell. Considering the flight times and energies at impact for positive ions using a bias potential of either +250 or -250 V, it is seen that a positive pulse yields lower energies but more uniform flight times. For example, +250 V yields a final impact energy of 33 eV for a positive ion starting a t the cell center, while -250 V produces 214 eV at impact. Flight times of ions with m / z 500, initially evenly spaced along the z axis, range from 7.2 to 12.3 k s for the first case and from 1.7 to 18.5 ps for the latter. It is expected that the daughter

ANALYTICAL CHEMISTRY, VOL 62, NO. 13, JULY 1, 1990

ions will be emitted from the target with appreciable kinetic energy and so will require deceleration to be trapped. Thus, the positive pulse is to be preferred as it allows much more precise timing control. It is evident that the SID experiment described above closely resembles the quench event normally used to remove all ions from the cell by biasing the trap plates at +10 and -10 V, respectively, for several milliseconds. Note that an ion as massive as one with m / z 10 000 will reach a trap plate in only 150 ps under these conditions. Thus, even if an ion survives its first encounter with the trap plate or produces daughters, they will be reaccelerated into the plate repeatedly and thus removed. For SID, the pulse duration is short enough to allow only a single surface collision. Also, there is a threshold of about 20 eV, below which atomic ions are efficiently neutralized by surface collisions, but above which reflection becomes increasingly efficient up to energies of a few hundreds of electronvolts (IO). If organic ions show similar thresholds, because the energy for the quench event is 10 eV, on average, neutralization is expected to be efficient. Furthermore, typical energies for SID are somewhat greater than 20 eV. Normal Incidence SID. Because McLafferty and coworkers had reported successful implementation of SID of toluene in a dual cell FTMS using normal incidence collision geometry (8),this was the first approach explored in the present research. Initial experiments utilized the m / z 502 fragment ion from perfluorotri-n-butylamine (PFTBA) and a target placed in contact with one trap plate to cover the 4-mm hole in the center of the plate. However, all attempts to detect SID daughter ions by using a wide range of pulse voltages, polarities, and durations, under conditions where products of SID with a conversion efficiency of at least 0.05% would have been detected, were unsuccessful. Other samples, including toluene, acetophenone, and bromobenzene, were also examined unsuccessfully. In none of these cases was there any evidence of daughter ions from SID being trapped. A variation on this protocol, where the ions were produced and trapped in one cell and then accelerated through the conductance limit to strike a target on the final trap plate, was also unsuccessful. This method suffered from the complication that ejection sweeps to remove fragment ions (a necessary control, because they would be indistinguishable from SID-produced daughter ions) greatly lowers the transfer efficiency through the conductance limit. This is a result of parent ions being translated sideways, in the x-y plane, by the ejection sweeps. If these fragments were not removed, no difference in their abundances resulted from SID experiments. Again, this result is in contrast to McLafferty’s report mentioned above (8),which followed this specific procedure for normal incidence SID of toluene. Grazing Incidence SID. A method that would allow incident angles other than normal was conceived to take advantage of the increased reflection coefficient as the incident angle moves toward grazing incidence. This protocol consists of exciting the cyclotron motion (tickling) of the parent ion before applying the SID pulse, resulting in ions following helical trajectories as they move toward the target. The resultant kinetic energy is the vector sum of the two motions and the angle of incidence (macroscopically) relative to the normal is the arctangent of the ratio of the cyclotron to z-axis velocities. For example, an ion with cyclotron translational energy of 210 eV and a z-axis translational energy of 1.3 eV would strike the target at an angle of 4.5O (85.5’ from the normal) with a net kinetic energy of 211.3 eV. SID can be performed simply in the FTMS-2000 by using the central conductance plate as the target. This plate is solid except for the 2-mm hole in the center, and so the minimum excited

1297

*)

MASS IN A.N.U.

Flgure 3. Control and SID Fwrler transform mass spectra of PFTBA: (a) electron ionization spectrum with fragments below mlz 400 ejected; (b) electron ionization spectrum following ejection of ions mlz