RAPID COMMUNICATIONS IN MASS SPECTROMETRY, VOL. 6 , 239-241 (1992)
Matrix-assisted Laser DesorptionIIonization of High-mass Molecules by Fourier-transform Mass Spectrometry John A. Castro, Claus Koster and Charles Wilkins* Department of Chemistry, University of California Riverside, Riverside, CA 92521, USA SPONSOR REFEREE: Professor R. J. Cotter, Department of Pharmacology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA
Following the first demonstrations of high-mass analysis using time-of-flight matrix-assisted laser desorptionl ionization (MALDI) techniques by Hillenkamp,’.2 Tanaka3 and their co-workers, there have been significant efforts in a number of laboratories to adapt the new methodology to Fourier-transform mass spectrometry (FTMS). The motivation for this research is obvious. Namely, it would be desirable to couple the unparalleled high mass resolution of FTMS4 with the extended mass range provided by MALDI, particularly for analysis of polymers and biomolecules. Unfortunately, prior to the present work, attempts to mate FTMS and MALDI have met with limited success. The highest mass matrix-assisted laser-desorption-FTMS result previously obtained appears to be the unpublished low kesolution spectrum of bovine insulin recently reported by Russell and co-workers.’ We,6 Campana and c o - ~ o r k e r s ,and ~ Hettich and B~chanan’~’have had some success with MALDI-FTMS of biomolecules with masses lower than 3000 Da, including melittin,7 a variety of lower mass peptides,698and oligonucleotides with masses lower than 1800Da.’ Furthermore, with the single exception of Campana’s report of obtaining mass resolution of 5000 for the molecular ion of me lit ti^^,^ such spectra have not displayed high resolution. Here, we report successful development of MALDI-FTMS, demonstrated with spectra obtained from a variety of high-mass polymer and biomolecule samples, using 355 nm radiation from an excimer-pumped dye laser for desorptionlionization and sinapinic acid as matrix. Some of these spectra are of much higher mass resolution than is possible with current time-of flight mass spectrometers.
EXPERIMENTAL Apparatus Experiments were performed using a Nicolet FTMS-2000 dual cell Fourier-transform mass spectrometer (Nicolet Analytical Instruments, Madison, WI, USA) equipped with a 7.2 T superconducting magnet, with differentially pumped 4.76cm cubic source and analyser cells separated by a 2 mm conductance limit, and an automatic solids probe. For laser desorption, a Lambda Physik EMG-201 MSC excimer laser (Lambda Physik, Gottingen, Germany) (operating at 308 nm, 180 mJ/28 ns pulse) was used to pump a Lambda Physik FL-2001 dye laser. 355 nm ultraviolet radiation was produced by pumping the dye laser cell containing a 0.60g/L dioxane solution of 2,2”’ dimethyl-pquaterphenyl (BMQ, Lambda Physik), resulting in a maximum output energy of 5 mJ/pulse. Figure 1 shows the experimental arrangement for performing MALDI-FTMS. 355 nm laser light attenuated by an iris diaphragm enters the mass spectrometer through a fused silica window and is focused onto a probe tip by a 12.5cm fused silica lens. The lens is mounted on a rotating lens assembly attached to the analyser flange, allowing it to be rotated out of the way when electron ionization or chemical ionization measurements are made. The sample probe is positioned exterior to the source cell, approximately 1-2 mm from the front trap plate. The distance between the lens and the probe tip was adjusted to obtain a power density of 106-107W/cm2, which can be fine-tuned by attenuating the laser beam. Author to whom correspondence should be addressed. 0951-4198/92/040239-03 $05.00 @ 1992 I by John Wiley & Sons, Ltd.
Sample preparation Sinapinic acid (3.4-dimethoxy-4-hydroxy-cinnamic acid, Aldrich Chemical Co. , Milwaukee, WI, USA) was used as matrix to obtain all spectra. Samples were prepared by mixing a 0.5 mmol/L methanol sample suspension or solution with a 50mmol/L methanol matrix solution to which 1 drop of trifluoroacetic acid (Mallinckrodt, Inc, St Louis, MO, USA) was added. The volumes used were adjusted to give a 1:lOOO analyte :matrix molar ratio. Homogeneous samples were deposited upon the probe tip by spraying an aerosol of the resulting sample matrix solution onto a rotzting probe tip. The sample compounds examined were melittin (from honey bee venom), insulin (from bovine pan-
7 T superconducting magnet
Figure 1. Block diagram of the FTMS instrument configuration.
Received 19 February 1992 Accepted 19 February 1992
240
HIGH MASS MATRIX-ASSISTED LASER DESORPTION/IONIZATION BY FTMS
creas), obtained from Fluka Chemical (Buchs, Switzerland), cytochrome c (from horse heart, type III), myoglobin (from equine skeletal muscle, type I), trypsinogen (from bovine pancreas), poly(ethyleneg1ycol) 8000, and poly(ethyleneglyco1) 10 000 (Sigma Chemical Co., St Louis, MO, USA). All samples were used without further purification. Experiment sequence To measure MALDI spectra, source-cell observation was used for all experiments. Prior to firing the desorption/ionization laser pulse, the front trap plate of the source cell is set to ground potential and the rear trap plate to 9 V . The laser is then triggered and following a variable delay of between 100 and 200ps (optimized for each sample), the potentials of the front and rear trapping plates are adjusted to 3 V for trapping. Following a variable delay (between 0 and several seconds) spectra are obtained using a 200V peak-topeak excitation sweep from 0 to 200 kHz at 180 Hz/ps sweep rate, followed by detection. All spectra result from 3 co-added time domain scans of 8192 data points (unless otherwise noted) augmented by 8192 zeros, and are baseline corrected prior to magnitude-mode Fourier transform to produce mass spectra. No apodization was used. The resolution is estimated from the ratio of peak position to peak width at half-height. RESULTS AND DISCUSSION One potential problem with adapting MALDI to FTMS is the requirement that the laser-desorbed ion species be trapped prior to analysis. Recent experimental timeof-flight observations using 354 nm radiation and a sinapinic acid matrix, showed that desorbed peptide ions appeared to have common average velocities (approximately 750 m/s) and similar velocity distributions, independent of their mass," leading to the prediction that ions with masses as low as 1000 u would be trapped inefficiently in an FTMS and higher-mass peptides would not be trapped at all. On the other hand, Pang and Cotter recently reported T O F studies confirming the previous results, but suggesting that velocities of matrix-desorbed peptides may not scale so linearly with mass. Thus, although the pessimistic predictions of Beavis and Chait are not unreasonable, it appears from the present results that the extrapolation from microsecond-scale time-of-flight measurements to millisecond-time-scale FTMS conditions may not be entirely valid. Specifically, MALDI-FTMS spectra of polymers, peptides, and peptide dimers with masses as high as 34000 can be obtained and are reported here for the first time. Because of the expected difficulties in ion trapping, when the new procedure reported here was tested, two poly(ethyleneglyco1) (PEG) samples were analysed in order to establish whether the expected mass discrimination effects would be evident. One of these oligomer mixtures (PEG 8000) previously had been characterized by low resolution direct infrared laser desorption (1RLD)-FTMS4 and is known to produce stable and abundant cationized molecular ion species. Thus, it serves as a useful standard and probe of the characteristics of the MALDI technique. For this mixture, enhanced mass resolution over the previous IRLD
1oboo
11000
13600
12000
14000
15000
m/z
Figure 2. Sinapinic acid matrix-assisted laser desorptionlionization FTMS spectrum of PEG 10000.
measurements was obtained, with an average resolution ( m / A m )of 3200. Also, the distribution of sodium and potassium cationized oligomers was in excellent agreement with the previous results. Equally good MALDI spectra of a series of PEG mixtures ranging from PEG 1000 to PEG 6000 were also obtained. For PEG 10000, although resolution is much lower, it is still sufficiently good to resolve the oligomers (repeating unit mass 44). Figure 2 is the MALDI-FTMS spectrum, where masses as high as 14 000 u are clearly discernible. Comparison of this spectrum with lessresolved spectra obtained using two commercial MALDI-TOF instruments showed similar ion distributions and average molecular weight, emphasizing the fact that the predicted mass discrimination for the FTMS measurements is not observed. Equally significant, there is no evidence of matrix adduct ions. For these polymer measurements, as well as for the peptides investigated, matrix ions were not ejected, as the experimental protocol does discriminate against lowmass ions. This results from the choice of delay time following laser desorption, but before establishing trapping conditions (i.e., the change from the 0 V and 9 V to the 3 V and 3 V trapping potential configuration).
!
2840
2860
nu2
I
lob0
lsbo
2sbo
mh
Figure 3. Sinapinic acid matrix-assisted laser desorptionlionization FTMS spectrum of melittin. Inset shows expanded molecular ion region.
H I G H MASS MATRIX-ASSISTED LASER DESORPTION/IONIZATION BY FTMS
24 I
above 50ms. This may result from a combination of metastable decompositon (previously observed in TOF studies of both compounds by Spengler and co-workers") and ion losses due to trapping inefficiencies. A more detailed study presently in progress will address these issues. In any event, it is possible to obtain low resolution MALDI-FTMS spectra for a variety of representative biomolecules. Figure 4 shows spectra obtained for bovine insulin, cytochrome c, myoglobin, and trypsinogen. For each.of these except trypsinogen, in addition to the molecular ion, dimer ions are also observed. Thus, in the present study, peptide ions with masses as high as 34 000 u (myoglobin dimer) are unambiguously observed by MALDI-FTMS.
(a)
5000
10000
15OOO
20000
25000
30000
35000
40000
m/z
&,,,\.":yvv+l
+*h& 5000
10000
15000
20000
25000
30000
35000
40000
m/z
CONCLUSION It has been shown that matrix-assisted laser-desorption/ ionization spectra of a number of polymers and peptides with masses as high as 34 000 Da can be obtained by FTMS under appropriate conditions. Furthermore, at least some of these spectra show mass resolution greater than that which can be obtained with current TOF mass analysers. Because the phenomenological constraints on FTMS resolution are different from those governing time-of-flight mass analysis, there is some reason to believe that the high mass resolution hoped for in the MALDI-FTMS measurments may be achieved, once optimized experimental procedures are developed.
Acknowledgements
d 5000
Support from the National Institutes of Health (GM-44606) and the National Science Foundation (CHE-89-11685) is gratefully acknowledged.
d 10000
l''"~"''I"'"'"'I''''"'"I'''"""1''''"'''I
15000
20000
25000
30000
35000
40000
nvz
Figure 4. Sinapinic acid matrix-assisted laser desorption/ionization FTMS spectra of (a) bovine insulin; (b) cytochrome c ; (c) myoglobin; (d) t rypsinogen.
For the delay times used, the higher-mass ions are trapped, while the matrix ions do not appear to be. Figure 3 is the mass spectrum of melittin, with a molecular ion species [M + HI' of mlz 2847. This is a relatively high resolution spectrum (ml Am = 3300), sufficient to resolve the molecular ion region. This is the highest resolution we have yet obtained for the more intractible peptide samples examined after completing the polymer studies. Other measurements of both melittin and bovine insulin, where the trapping time (after trapping, but prior to spectral detection) was varied from 0 to about 50 ms, showed that molecular ions were not detectable for trapping periods much
REFERENCES 1. M. Karas, D . Bachmann. U . Bahr and F. Hillenkamp, Inr. J . Muss Spectrom. Ion Processes 78, 53 (1987). 2. M. Karas and F Hillenkamp. Anal. Chem. 60. 2288 (1988). 3. K. Tanaka, H. Waki, Y. Ido. S. Akita. Y. Yoshida and T. Yoshida. Rupid Commun. Muss Spenrom. 2, 151 (1988). 4. C. F. Ijames and C. L. Wilkins, J . Am. Chem. Soc. 110. 2687 (1988). 5 . T. Solouki, L. R. Beeler and D. H . Russell, presenred (11 f H r h Annual FACSSrneering. Anaheim, C A , Abstract N o . 327 (IYYI). 6. L. M. Nuwaysir and C . L. Wilkins, SPIE Proc. No. 14.?7. SPIE. Bellingham. WA. pp. 112-123 (1901). 7. J . Campana. Z. Liang and J . Covey, presented ut 4th Suriihel Conference on Mass Specrrome/ry, Sanibel Island, FL (1992). 8. R. L. Hettich and M. V. Buchanan, J . Am. Soc. Muss S p e c t r ~ m . 2, 22 (1991). 9. R. L. Hettich and M. V. Buchanan. J . Am. SOL.. Mus.~Specrrom 2, 402 (1991). 10. R. C . Beavis and B. T. Chait, Chem. Phys. Left. 181,474, (1901). 11. Y. Pang and R . J . Cotter, Org. Muss. Spectrom. 27. 3 (1992). 12. B. Spengler, D . Kirsch and R. Kaufmann. Rapid Commun. Mass Spectrom. 5 . 198 (1991).
Matrix-assisted Laser DesorptionIIonization of High-mass Molecules by Fourier-transform Mass Spectrometry John A. Castro, Claus Koster and Charles Wilkins* Department of Chemistry, University of California Riverside, Riverside, CA 92521, USA SPONSOR REFEREE: Professor R. J. Cotter, Department of Pharmacology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA
Following the first demonstrations of high-mass analysis using time-of-flight matrix-assisted laser desorptionl ionization (MALDI) techniques by Hillenkamp,’.2 Tanaka3 and their co-workers, there have been significant efforts in a number of laboratories to adapt the new methodology to Fourier-transform mass spectrometry (FTMS). The motivation for this research is obvious. Namely, it would be desirable to couple the unparalleled high mass resolution of FTMS4 with the extended mass range provided by MALDI, particularly for analysis of polymers and biomolecules. Unfortunately, prior to the present work, attempts to mate FTMS and MALDI have met with limited success. The highest mass matrix-assisted laser-desorption-FTMS result previously obtained appears to be the unpublished low kesolution spectrum of bovine insulin recently reported by Russell and co-workers.’ We,6 Campana and c o - ~ o r k e r s ,and ~ Hettich and B~chanan’~’have had some success with MALDI-FTMS of biomolecules with masses lower than 3000 Da, including melittin,7 a variety of lower mass peptides,698and oligonucleotides with masses lower than 1800Da.’ Furthermore, with the single exception of Campana’s report of obtaining mass resolution of 5000 for the molecular ion of me lit ti^^,^ such spectra have not displayed high resolution. Here, we report successful development of MALDI-FTMS, demonstrated with spectra obtained from a variety of high-mass polymer and biomolecule samples, using 355 nm radiation from an excimer-pumped dye laser for desorptionlionization and sinapinic acid as matrix. Some of these spectra are of much higher mass resolution than is possible with current time-of flight mass spectrometers.
EXPERIMENTAL Apparatus Experiments were performed using a Nicolet FTMS-2000 dual cell Fourier-transform mass spectrometer (Nicolet Analytical Instruments, Madison, WI, USA) equipped with a 7.2 T superconducting magnet, with differentially pumped 4.76cm cubic source and analyser cells separated by a 2 mm conductance limit, and an automatic solids probe. For laser desorption, a Lambda Physik EMG-201 MSC excimer laser (Lambda Physik, Gottingen, Germany) (operating at 308 nm, 180 mJ/28 ns pulse) was used to pump a Lambda Physik FL-2001 dye laser. 355 nm ultraviolet radiation was produced by pumping the dye laser cell containing a 0.60g/L dioxane solution of 2,2”’ dimethyl-pquaterphenyl (BMQ, Lambda Physik), resulting in a maximum output energy of 5 mJ/pulse. Figure 1 shows the experimental arrangement for performing MALDI-FTMS. 355 nm laser light attenuated by an iris diaphragm enters the mass spectrometer through a fused silica window and is focused onto a probe tip by a 12.5cm fused silica lens. The lens is mounted on a rotating lens assembly attached to the analyser flange, allowing it to be rotated out of the way when electron ionization or chemical ionization measurements are made. The sample probe is positioned exterior to the source cell, approximately 1-2 mm from the front trap plate. The distance between the lens and the probe tip was adjusted to obtain a power density of 106-107W/cm2, which can be fine-tuned by attenuating the laser beam. Author to whom correspondence should be addressed. 0951-4198/92/040239-03 $05.00 @ 1992 I by John Wiley & Sons, Ltd.
Sample preparation Sinapinic acid (3.4-dimethoxy-4-hydroxy-cinnamic acid, Aldrich Chemical Co. , Milwaukee, WI, USA) was used as matrix to obtain all spectra. Samples were prepared by mixing a 0.5 mmol/L methanol sample suspension or solution with a 50mmol/L methanol matrix solution to which 1 drop of trifluoroacetic acid (Mallinckrodt, Inc, St Louis, MO, USA) was added. The volumes used were adjusted to give a 1:lOOO analyte :matrix molar ratio. Homogeneous samples were deposited upon the probe tip by spraying an aerosol of the resulting sample matrix solution onto a rotzting probe tip. The sample compounds examined were melittin (from honey bee venom), insulin (from bovine pan-
7 T superconducting magnet
Figure 1. Block diagram of the FTMS instrument configuration.
Received 19 February 1992 Accepted 19 February 1992
240
HIGH MASS MATRIX-ASSISTED LASER DESORPTION/IONIZATION BY FTMS
creas), obtained from Fluka Chemical (Buchs, Switzerland), cytochrome c (from horse heart, type III), myoglobin (from equine skeletal muscle, type I), trypsinogen (from bovine pancreas), poly(ethyleneg1ycol) 8000, and poly(ethyleneglyco1) 10 000 (Sigma Chemical Co., St Louis, MO, USA). All samples were used without further purification. Experiment sequence To measure MALDI spectra, source-cell observation was used for all experiments. Prior to firing the desorption/ionization laser pulse, the front trap plate of the source cell is set to ground potential and the rear trap plate to 9 V . The laser is then triggered and following a variable delay of between 100 and 200ps (optimized for each sample), the potentials of the front and rear trapping plates are adjusted to 3 V for trapping. Following a variable delay (between 0 and several seconds) spectra are obtained using a 200V peak-topeak excitation sweep from 0 to 200 kHz at 180 Hz/ps sweep rate, followed by detection. All spectra result from 3 co-added time domain scans of 8192 data points (unless otherwise noted) augmented by 8192 zeros, and are baseline corrected prior to magnitude-mode Fourier transform to produce mass spectra. No apodization was used. The resolution is estimated from the ratio of peak position to peak width at half-height. RESULTS AND DISCUSSION One potential problem with adapting MALDI to FTMS is the requirement that the laser-desorbed ion species be trapped prior to analysis. Recent experimental timeof-flight observations using 354 nm radiation and a sinapinic acid matrix, showed that desorbed peptide ions appeared to have common average velocities (approximately 750 m/s) and similar velocity distributions, independent of their mass," leading to the prediction that ions with masses as low as 1000 u would be trapped inefficiently in an FTMS and higher-mass peptides would not be trapped at all. On the other hand, Pang and Cotter recently reported T O F studies confirming the previous results, but suggesting that velocities of matrix-desorbed peptides may not scale so linearly with mass. Thus, although the pessimistic predictions of Beavis and Chait are not unreasonable, it appears from the present results that the extrapolation from microsecond-scale time-of-flight measurements to millisecond-time-scale FTMS conditions may not be entirely valid. Specifically, MALDI-FTMS spectra of polymers, peptides, and peptide dimers with masses as high as 34000 can be obtained and are reported here for the first time. Because of the expected difficulties in ion trapping, when the new procedure reported here was tested, two poly(ethyleneglyco1) (PEG) samples were analysed in order to establish whether the expected mass discrimination effects would be evident. One of these oligomer mixtures (PEG 8000) previously had been characterized by low resolution direct infrared laser desorption (1RLD)-FTMS4 and is known to produce stable and abundant cationized molecular ion species. Thus, it serves as a useful standard and probe of the characteristics of the MALDI technique. For this mixture, enhanced mass resolution over the previous IRLD
1oboo
11000
13600
12000
14000
15000
m/z
Figure 2. Sinapinic acid matrix-assisted laser desorptionlionization FTMS spectrum of PEG 10000.
measurements was obtained, with an average resolution ( m / A m )of 3200. Also, the distribution of sodium and potassium cationized oligomers was in excellent agreement with the previous results. Equally good MALDI spectra of a series of PEG mixtures ranging from PEG 1000 to PEG 6000 were also obtained. For PEG 10000, although resolution is much lower, it is still sufficiently good to resolve the oligomers (repeating unit mass 44). Figure 2 is the MALDI-FTMS spectrum, where masses as high as 14 000 u are clearly discernible. Comparison of this spectrum with lessresolved spectra obtained using two commercial MALDI-TOF instruments showed similar ion distributions and average molecular weight, emphasizing the fact that the predicted mass discrimination for the FTMS measurements is not observed. Equally significant, there is no evidence of matrix adduct ions. For these polymer measurements, as well as for the peptides investigated, matrix ions were not ejected, as the experimental protocol does discriminate against lowmass ions. This results from the choice of delay time following laser desorption, but before establishing trapping conditions (i.e., the change from the 0 V and 9 V to the 3 V and 3 V trapping potential configuration).
!
2840
2860
nu2
I
lob0
lsbo
2sbo
mh
Figure 3. Sinapinic acid matrix-assisted laser desorptionlionization FTMS spectrum of melittin. Inset shows expanded molecular ion region.
H I G H MASS MATRIX-ASSISTED LASER DESORPTION/IONIZATION BY FTMS
24 I
above 50ms. This may result from a combination of metastable decompositon (previously observed in TOF studies of both compounds by Spengler and co-workers") and ion losses due to trapping inefficiencies. A more detailed study presently in progress will address these issues. In any event, it is possible to obtain low resolution MALDI-FTMS spectra for a variety of representative biomolecules. Figure 4 shows spectra obtained for bovine insulin, cytochrome c, myoglobin, and trypsinogen. For each.of these except trypsinogen, in addition to the molecular ion, dimer ions are also observed. Thus, in the present study, peptide ions with masses as high as 34 000 u (myoglobin dimer) are unambiguously observed by MALDI-FTMS.
(a)
5000
10000
15OOO
20000
25000
30000
35000
40000
m/z
&,,,\.":yvv+l
+*h& 5000
10000
15000
20000
25000
30000
35000
40000
m/z
CONCLUSION It has been shown that matrix-assisted laser-desorption/ ionization spectra of a number of polymers and peptides with masses as high as 34 000 Da can be obtained by FTMS under appropriate conditions. Furthermore, at least some of these spectra show mass resolution greater than that which can be obtained with current TOF mass analysers. Because the phenomenological constraints on FTMS resolution are different from those governing time-of-flight mass analysis, there is some reason to believe that the high mass resolution hoped for in the MALDI-FTMS measurments may be achieved, once optimized experimental procedures are developed.
Acknowledgements
d 5000
Support from the National Institutes of Health (GM-44606) and the National Science Foundation (CHE-89-11685) is gratefully acknowledged.
d 10000
l''"~"''I"'"'"'I''''"'"I'''"""1''''"'''I
15000
20000
25000
30000
35000
40000
nvz
Figure 4. Sinapinic acid matrix-assisted laser desorption/ionization FTMS spectra of (a) bovine insulin; (b) cytochrome c ; (c) myoglobin; (d) t rypsinogen.
For the delay times used, the higher-mass ions are trapped, while the matrix ions do not appear to be. Figure 3 is the mass spectrum of melittin, with a molecular ion species [M + HI' of mlz 2847. This is a relatively high resolution spectrum (ml Am = 3300), sufficient to resolve the molecular ion region. This is the highest resolution we have yet obtained for the more intractible peptide samples examined after completing the polymer studies. Other measurements of both melittin and bovine insulin, where the trapping time (after trapping, but prior to spectral detection) was varied from 0 to about 50 ms, showed that molecular ions were not detectable for trapping periods much
REFERENCES 1. M. Karas, D . Bachmann. U . Bahr and F. Hillenkamp, Inr. J . Muss Spectrom. Ion Processes 78, 53 (1987). 2. M. Karas and F Hillenkamp. Anal. Chem. 60. 2288 (1988). 3. K. Tanaka, H. Waki, Y. Ido. S. Akita. Y. Yoshida and T. Yoshida. Rupid Commun. Muss Spenrom. 2, 151 (1988). 4. C. F. Ijames and C. L. Wilkins, J . Am. Chem. Soc. 110. 2687 (1988). 5 . T. Solouki, L. R. Beeler and D. H . Russell, presenred (11 f H r h Annual FACSSrneering. Anaheim, C A , Abstract N o . 327 (IYYI). 6. L. M. Nuwaysir and C . L. Wilkins, SPIE Proc. No. 14.?7. SPIE. Bellingham. WA. pp. 112-123 (1901). 7. J . Campana. Z. Liang and J . Covey, presented ut 4th Suriihel Conference on Mass Specrrome/ry, Sanibel Island, FL (1992). 8. R. L. Hettich and M. V. Buchanan, J . Am. Soc. Muss S p e c t r ~ m . 2, 22 (1991). 9. R. L. Hettich and M. V. Buchanan. J . Am. SOL.. Mus.~Specrrom 2, 402 (1991). 10. R. C . Beavis and B. T. Chait, Chem. Phys. Left. 181,474, (1901). 11. Y. Pang and R . J . Cotter, Org. Muss. Spectrom. 27. 3 (1992). 12. B. Spengler, D . Kirsch and R. Kaufmann. Rapid Commun. Mass Spectrom. 5 . 198 (1991).
