Anal. Chem. 1982, 5 4 , 7437-7433
showed no evidence of yttrium-90. However, repeated washings of the organic layer yielded no further activity. Perhaps the yttrium is bound by an impurity in the organic layer that renders it nonextractable. This method was dleveloped for low-level activity urine samples in which the maximum total added activity of all the radionuclides was 39 WCi (1.4 MBq) and the corresponding mass was 3 nmol. It w i therefore not known whether the analysis schemes presented herein are applicable to levels greatly in excess of this figure. The limits of detection for the radionuclides are estimated to be 0.6 pCi strontium-90, 0.7 pCi yttrium-90, 1.0 pCi promethium-147, and 0.8 pCi cerium-144 (1 pCi = 37 mBq). The detection limits are calculated a t the 95% confidence level, taking into account the following: 70 recovery of radionuclide, ambient counter background, counter efficiency, and a counting time off 60 min. LITEZATURE CITED (1) Kramer, G. H.; Davies, J. Report AECL-7227, Atomic Energy of Canada Limited, 1981.
1431
(2) Johnson, J. R.; Stewart, D. G.; Carver, M. 8. Report AECL-6540. Atomic Energy of Canada Llmited, 1979. (3) Peppard, D. F.; Mason, G. W.; Maler, J. L.; Driscoll, W. S.J . Inorg. Nucl. Chem. 1957, 4 , 334. (4) Peppard, D. F.; Moline, G. W.; Mollne, S.W. J. Inorg. Nucl. Chem. 1957, 5 , 141. (5) Ludwlck, J. D. Anal. Chem. 1964, 3 6 , 1104. (6) Johnson, W. C.; Campbell, M. H. USEAC Report HW-SA-3549, 1964. (7) Hampson, B. L. Ana/yst(London) 1964, 89, 651. (8) McCown, J. J.; Larson, R. P. Anal. Chem. 1960, 3 2 , 597. (9) Healy, T. V. Radlochlm. Acta 1963, 2 , 52. (10) Petrow, H. G. Anal. Chem. 1965, 3 7 , 584. (11) Davis, S.; Arnold, A. J . Assoc. Off. Anal. Chem. 1964, 47, 580; 1965, 48, 5. (12) Testa, C.; Santorla, G. Energ. Nucl. (Milan) 1970, 17, 320. (13) Bogen, D. C. Health Phys. 1968, 14, 131. (14) Purkayastha, B. C . ; Bhattacharyya, S. N. J. Inorg. Nucl. Chem. 1959, 70, 103. (15) Matsui, M. Bull. Chem. Soc. Jpn. 1966, 3 9 , 1114. (16) Kramer, G. H. Report AECL-6879, Atomic Energy of Canada Limited, 1980.
RECEIVED for review January 11,1982. Accepted April 5,1982. This is paper AECL-7756.
CORRESPONDENCE High-Resolutilon Detection of Collision-Induced Dissociation Fragments by Fourier Transform Mass Spectrometry Sir: Collision-induced dissociation (CID) is the most convenient and popular method for ion structure determination currently available to mass spectrometrisk. In particular, the use of collision-induced dissociation for complex mixture analysis in the MS/MS technique (1-3) and the commercial availability of triple quadrupole mass spectrometers designed for such experiments (4)have contributed to the latest surge of interest in the CID technique. One limitation noticeable for most MS/MS instrumentation, however, is the resolution obtainable for the detection of CID fragment ions. This is often termed MS-I1 or “back-end” resolution in order to distinguish it from MS-I or “front-end” resolution, which is the resolution available for selecting the parent ion to undergo CID. Reverse geometry instruments sufffer from poor (often less than unit mass) “back-end” resolution due to kinetic energy release. Quadrupole instruments are capable of somewhat better (unit) iresolution in most cases, but increased mass resolution is limited by precise machining tolerances. Linked-scanning techniques have been used to improve the “back-end” mass resolution of double-focusing instruments (5,6). Improved resolution in the daughter ions is provided a t the expense of resolution in the parent ions. Boerboom and co-workers have improved resolution in their tandem mass spectrometer by utiliziing postcollision acceleration of fragments to reduce the relative energy spiread due to kinetic energy release (7).Of particular note is a unique tandem instrument presently being modified by McLafferty and coworkers at Cornell. This instrument, which combines two double-focusing mass spectrometers in series, is predicted to be capable of a “back-end” resolution of >10000 and has a demonstrated usable “front-end” resolution of 50 000 (8, 9). Recently, we reported the first observations of collisioninduced dissociation in a Nicolet prototype Fourier transform mass spectrometer (FTMS) ( 1 0 , I I ) . Since that time, we have 0003-2700/62/0354-1431$01.25/0
applied the FTMS collision-induced dissociation technique to the study of proton-bound alcohol dimers and metal-ion bound alcohol dimers (12) and have successfully observed sequential collision-induced dissociations by operating the instrument in an MS/MS/MS fashion to examine the collision-induced dissociation of ions that are themselves the result of CID fragmentation (13). We have also begun to compare the FTMS low energy CID processes to those observed in multiple sector quadrupole instruments (BQQ and QQQ) (14). The capability of Fourier transform mass spectrometers to perform ultrahigh resolution, exact mass measurements is well-known (15-18) and suggested to us that increased resolution should be possible for CID spectra as well by making use of the heterodyne mode of operation (19) and maintaining collision gas pressures as low as possible to avoid broadening of peaks due to collisional damping of the transient (20). In this paper we present the first results of these experiments and demonstrate the capability of the Fourier transform mass spectrometer to determine elemental composition of CID fragment ions by high-resolution, exact mass measurement.
EXPERIMENTAL SECTION The experiments described here were performed with a Nicolet prototype Fourier transform mass spectrometer. The techniques and instrumentation used for obtaining CID spectra have previously been described in detail (11). All spectra presented in this paper were obtained by operating the instrument in the heterodyne mode. Exponential apodization of the transient was used to improve peak shapes and improve the signal to noise ratio. Although some loss of resolution can be expected from such damping of the transient, this was not seen for the spectra displayed here, possible due to the resulting improvement in peak shape. A 4K or 8K transient was acquired with one or two zero fills employed to increase the number of points on the peaks. The number of data points taken was limited by the length of the 0 1982 American Chemical Society
1432
ANALYTICAL CHEMISTRY, VOL. 54, NO. 8, JULY 1982
430
43 I 43.2 Mass ( o r n u )
430
Flgure 1. High-resolution scan of m l z 43 region for a mixture of acetone at 1 X IO-' torr and argon at 4.6 X lo-' torr. The top scale shows the frequency in Hz. Ions of m l z 43 were first ejected by a double resonance pulse after which the molecular ion ( m / z 58)was excited with a 1 . 5 4 rf pulse for 100 ks at 234.741 kHz. This corresponds to a translational energy of 7.4 eV. Detectlon occurred after a 60-ms delay. 8K data points were taken with one zero fill.
transient. Sample pressures were (1-3) X torr with argon added as a collision gas to give a total pressure of no more than 5 X lo4 torr. The choice of argon as a collision gas (rather than helium or nitrogen) resulted from our observation that collisional energy transfer increases with increasing target gas molecular weight (14). The same trend has been observed for other low energy CID processes (21)and is due to center of mass considerations. The magnetic field was maintained at 0.9 T, with a trapping voltage of 1-1.44 V. Ions of m / z 43 produced directly by electron impact were first removed from the cell by double resonance ejection, followed by irradiation of the acetone and 2-chloropropane molecular ions for 100 ps each at their resonance frequencies (234.730kHz and 174.272 kHz, respectively). A delay of 60-200 ms prior to detection was necessary to allow time for collisions to occur. Prior to adding the collision gas, the instrument was calibrated for exact mass determination by using CH3CHC1CH3+.(m/z 78.023 628), (CH3)&O+.( m / z 58.041865), and C3H7+( m / z 43.054 775) as calibrant ions.
RESULTS AND DISCUSSION The molecular ions of acetone and isopropyl chloride can both undergo collision-induced dissociation to give isobaric ions of a nominal m / z 43: CID
(CHJ&O+*
mlz 58.041865
CH,CO+
-+
-
+ CH3.
(1)
+ C1*
(2)
mlz 43.018390
CID
CH&HClCH3+* m / z 18.023628
431
432
Mass (ornu)
C3H7+
m l z 43.054775
These CID reactions also correspond to the major fragmentations observed in the mass spectra of these compounds. The first experiment was performed by using a mixture of acetone and argon. Figure 1 shows a high-resolution scan of the m / z 43 region which clearly shows CH3CO+resulting from CID reaction 1. The resolution (fwhm) for this peak is 2700 with the peak center at 43.0108 amu. The 175 ppm mass error for this spectrum resulted from a rough calibration in the direct mode rather than in the heterodyne mode. Since this result exceeded the resolution of 1190 or greater required to separate CH3CO+from C3H7+,a mixture of acetone, 2-chloropropane, and argon was added to the instrument. The expectation was that alternately exciting (CHJ2CO+. and CH3CHC1CH3+.on successive scans and averaging the transients would produce a transform showing a doublet consisting of CH,CO+ and C3H7+, Successive CID of both molecular ions in the same scan is also possible; the choice of the former approach was made by considering current software limitations. After double resonance ejection of m / z 43 (but prior to performing the CID experiment), an ion-molecule reaction was found to
Flgure 2. High-resolution scan of m l z 43 region for a mixture of acetone at 4.6 X IO-' torr, 2-chloropropane at 2 X lo-' torr, and argon at 4.6 X torr. Ions of m l z 43 formed directly by electron impact were first ejected by a double resonance pulse after which the acetone molecular ion ( m l z 58) was excited to 7.4 eV with a 1.5 V rf pulse for 100 ks at 234.741 kHz. Detection occurred after a 60 ms delay. 4K data points were taken with two zero fills.
43.0 43.1 432 Mass (ornu)
Flgure 3. High-resolution scan of the m l z 43 region for the same mixture as Figure 2. Acetone molecular ions were accelerated to 18.5 eV with a 2 . 4 4 rf pulse for 100 k s at 234.741 kHz, while 2-chloropropane molecular ions were excited to 8.6 eV with a 1.9-V rf pulse for 100 ~s at 174.272 kHz on alternate pulses. The remaining conditions were the same as in Figure 2.
result in the regeneration of an ion at this mass. Since double resonance ejection (22) showed the ion to be coming from a reaction of CH3CHC1CH3+.(mlz 78), the product ion was first thought to be CH3CO+, arising from dissociative charge transfer from the isopropyl chloride molecular ion to acetone. A high-resolution scan, however, yielded an exact mass of 43.054 rather than 43.018, indicating that the product was in fact C3H7+. That this was not an error in exact mass determination was shown by using double resonance to excite (CH3)&O+-,causing it to undergo reaction 1. Figure 2 shows the result: The C3H7+a t m / z 43.054 forms a doublet with the CH3CO+ a t m / z 43.015 formed by CID of the acetone molecular ion. The resolution achieved in this spectrum (full width at half height) is 3250 for CH3CO+and 2760 for C3H7+. The error in exact mass determination is -85 ppm for CH3CO+ and -11.7 ppm for C3H7+. By noting the increase in C3H7+produced by CID over the amount of C3H7+produced by ion-molecule reaction, the originally intended experiment could still be performed. The resulk me seen in Figure 3. The increase in signal from Figure 2, visible from the decrease in the relative size of the noise in the base line, occurred due to an increase in the rf excitation energy. Although difficult to see from a comparison of these spectra (which are normalized to the largest peak in each case), the C3H7+peak in Figure 3 is 1.68 times larger than in Figure 2, while the CH3CO+peak is larger by a factor of 3.74 than in Figure 2. The somewhat greater increase for CH,CO+ is due to the use of a slightly higher rf level for irradiating the
Anal. Chem. 1982. 54, 1433-1435
acetone molecular ion. The resolution achieved for this spectrum is 2975 for CIH3CO+and 2100 for C3H7+.The error in the exact mass determination is -21.4 ppm for CH3CO+and -17.5 ppm for C3H7+. These examples serve to demonstrate that high-resolution, exact mass detection of CID fragments is possible with Fourier transform mass spectrometry. The resolution obtained in these spectra, although adequate for the separation of CH3CO+ and C3H7+would not be sufficient to separate CH3NHCH2+ (mlz 43.042 199) from C3H7+(mlz 43.054 775), for which a resolution of 3419 would be required. These measurements, however, were taken at a magnetic field strength of only 0.9 T. Since the resolution obtainable from a Fourier transform mass spectrometer is dlirectly proportional to magnetic field strength (ZO), the use of currently available superconducting magnets to provide higher magnetic fields, for example 8.5 T, should provide resolutions >20000 at m/z 50 at these pressures, but decreasing inversely with mass. Another consideration is that since resolution varies inversely with pressure, the maximum obtainable resolution will be found at the lowest possible collision gas pressures. At the pressures employed in these experiments, ion-molecule reactions were observed within the delay time necessary to allow collisions to occur between the riample ions and the target gas. This fact complicates the spectra somewhat and is definitely undesirable. One solutioin is to decrease the partial pressure of the sample gas, but this will also result in the decrease of the size of the signal. Another possibility under investigation in our laboratory is the use of a solenoid valve to introduce the collision gas in a “bur,st” for the CID experiment and then allow the target gas to be pumped out of the cell before ion detection occurs. Thiri should increase the obtainable resolution due to the decrease in collisional damping of the transient, but the delay between the CID event and ion detection will be dependent on the pumpiing speed and sample ion-molecule reactions could remain a problem. Although high-resolution mass spectra of parent ions, Le., “front-end” resolution, may readily be achieved by using FTMS (15, 16), it is presently difficult to selectively excite one ion in the presence of another for CID due to the spectral bandwidth of the excitation pulse. At present, unit “front-end” resolution appears to be possible with appropriate choice of rf levels and excitatioin times (6). The fact that the maw measurement accuracy reported for these spectra is somewhat less than has been previously reported (18) may in part be due to a perturbation of the trapping cell electric fiield due to a 1/4 in. hole placed in one of the receiver plates to allow laser ionlization experiments (23-25). Such a perturbation might affect the model used for exact mass calibration. Nevertheless, although exact mass determination of CID products was not t,he original intention of these experiments, the exact mass measurements were accurate enough to aesign elemental compositions to the product ions observed here. Lastly, it must be noted that both resolution and the maximum translationall energy that can be imparted to an ion before ejection are inversely proportional to ionic mass (10,
1433
11,ZO). Increasing the magnetic field extends both the energy and mass range, while utilization of a larger trapping cell would increase the maximum amount of translational energy available for collision. Such measures will be necessary if CID studies of high mass ions are to be undertaken with FTMS. There is no doubt, however, that high-resolution CID capabilities will add to the analytical utility of FTMS.
ACKNOWLEDGMENT The authors wish to thank the Nicolet Instrument Co. Mass Spectrometry Group, and in particular R. B. Spencer, for their technical assistance. LITERATURE CITED Kondrat, R. W.; Cooks, R. G. Anal. Chem. 1978, 50, 81A. Yost, R. A,; Enke, C. G., Anal. Chem. 1979, 51, 1251A. Hunt, D. F.; Shabanowitz, J.; Giordani, A. B. Anal. Chem. 1980, 52, 386. Slayback, J. R. B.; Story, M. S. Ind. Res./Dev. 1981, 23, 128-134. Beynon, J. H.; Cooks, R. G. Int. J. Mass Spectrom. Ion Phys. 1976, 19, 107. Weston, A. F.; Jennings, K. R.; Evans, S.; Elliott, R. M. Int. J. Mass Spectrom Ion Phys 1976, 2 0 , 3 17. Louter, 0. J.; Boerboom, A. J. H.; Stalmeier, P. F. M.; Tuithof, H. H.; Kistemaker, J. Int. J. Mass Spectrom. Ion Phys. 1980, 33, 335. McLafferty, F. W. Acc. Chem. Res. 1980, 13,33. McLafferty, F. W.; Todd, P. J.; McGilvery, D. C., Baldwin, M. A. J. Am. Chem. SOC.1980, 702,3360. Cody, R. 8.; Frelser, B. S. Int. J. Mass Spectrom. Ion Phys. 1982, 4 7 , 199. Cody, R. 8.; Burnler, R. C.; Frelser, B. S.Anal. Chem. 1982, 54,96. Burnier, R. C.; Cody, R. B.; Freiser, B. S.,submitted for publication in J. Am. Chem. SOC. Cody, R. B.; Burnier, R. C.; Cassady, C. J.; Freiser, B. S.,submitted for publication in Anal. Chem. McLuckey, S. A,; Verma, S.; Cody, R. B.; Sallans, L.; Burnier, R. C.; Freiser, B. S., submitted for publication in Int. J. Mass Spectrom. Ion Phys Comisarow, M. B.; Marshall, A. G. J. Chem. Phys. 1975, 62, 293. Marshall, A. G.; Comisarow, M. B. Anal. Chem. 1980, 52,463. White, R. L.; Ledford, E. B., Jr.; Ghaderi, S.;Wilkins, C. L.; Gross, M. L. Anal. Chem. 1980, 52, 1525. Ledford, E. B., Jr.; Ghaderi, S.;White, R. L.; Spencer, R. B.; Kulkarni, P. S.;Wllklns, C. L.; Gross, M. L. Anal. Chem. 1980, 52, 463. Comlsarow, M. B. Adv. Mass Spectrom. 1980, 8 , 1698. Marshall, A. G.; Comisarow, M. B.; Parisod, G. J. Chem. Phys. 1979, 7 1 , 4434. McLuckey, S. A.; Glish, G. L.; Cooks, R. G. Int. J . Mass Spectrom. Ion Phys. 1981, 3 9 , 219. Comisarow, M. B.; Grassi, V.; Parlsod, G. Chem. Phys. Lett. 1978, 5 7 , 413. Cody, R. B.; Burnier, R. C.; Reents, W. D., Jr.; Carlin, T. J.; McCrery, D. A.; Lenael. R. K.; Freiser. B. S. Int. J. Mass SDectrom, Ion Phys. 1980, 33,-37. Burnier, R. C.; Byrd, G. D.; Freiser, B. S. Anal. Chem. 1980, 52,
.
.
.
1641
Birnier, R. C.; Byrd, G. 103,4360.
D.;Freiser, B. S. J. Am.
Chem. SOC. 1981,
R. B. Cody B. S. Freiser* Department of Chemistry Purdue University West Lafayette, Indiana 47907
RECEIVED for review February 8, 1982. Accepted April 14, 1982. Support for this research was provided by the Department of Energy (DE-AC02-80ER10689)and the National Science Foundation (CHE-8002685), which provided funds to purchase the F’I’MS.
Direct Precoincentration of Trace Elements in Aqueous Solutions on Macroret ic uIar A c ryIic Ester Resin Sir: The importance of preconcentration methods as complementary adjuncts to instrumental methods for the deter-
mination of trace elements in aqueous solutions has been clearly stated in a recent report by Leyden and Wegscheider
0003-2700/82/0354-1433$01.25/0 0 1982 American Chemical Society
