Hypervalent pyrrolidinium radicals by neutralization—reionization mass spectrometry: Metastability and radical leaving group abilities

Hypervalent Pyrrolidinium Radicals by Neutralization-Reionization Mass Spectrometry: Metastability and Radical Leaving Group Abilities Lars Frosig* and Frantigek Tureeek Department of Chemistry, University of Washington, Seattle, Washington, USA

Neutralization-reionization mass spectrometry was used to generate hypervalent radicals pyrrolidinium (1H'), N-methylpyrrolidinium (2H), N-ethylpyrrolidinium (3H'), N-phenylpyrrolidirdum (4H'), N,N-dimethylpyrrolidinium (5"), N-methyl-N-ethylpyrrolidinium (6), and their deuterium-labeled derivatives and to study their dissociations in the gas phase. Isotopomers of pyrrolidinium and N-phenylpyrrolidinium showed small fractions of stable radicals of microsecond lifetimes that were detected following collisional reionization. The leaving group abilities in radical dissociations were established as H" >> C2H 5 ~- C6H 5 > CIq 3. The hydrogen atom was the best leaving group in secondary and tertiary pyrrolidinium radicals 1 H ' - 4 H , whereas losses of ethyl, phenyl, and ring openings by N~2 bond cleavages were less facile. Methyl was the worst leaving group among those studied. Ring cleavages dominated the dissociations of quaternary pyrrolidinium radicals 5" and 6", whereas losses of alkvl substituents were less efficient. The electronic properties of hypervalent a m m o n i u m rad'ica Is are discussed to rationalize the experimental leaving group abilities of hydrogen atom, alkyl, and phenyl radicals. (J Am Soc Mass Spectrom 1998, 9, 242-254) © 1998 American Society for Mass Spectrometry

ne-electron reduction of organic a m m o n i u m cations by femtosecond collisional electron transfer in the gas phase results in the formation of transient radicals [1, 2] that show unusual chemical properties, as studied by neutralization-reionization mass spectrometry (NRMS) [3-8]. A m m o n i u m radicals formally have nine valence electrons on the tetracoordinated nitrogen atom and are therefore viewed as hypervalent (9-N-4) species according to the nomenclature introduced by Perkins et al. [9]. Hypervalent radicals have been invoked as intermediates in electrical discharge in gases [10], electrochemical reduction o~! a m m o n i u m cations in solution [11], and radical substitution reactions [91, although direct evidence for their existence has been scant. Previous mass spectrometric studies have mainly focused on detecting hypervalent radicals that did not dissociate in a few microseconds between the neutralization and reionization collisions. Such species can be denoted as metastables, because their microsecond lifetimes are of kinetic nature due to potential energy barriers to exothermic dissociations [12-14]. Deuterium isotopomers of a m m o n i u m [1, 15a-c], methylammo-

O

Address reprint requests to Professor Frantigek Ture~ek, Department of Chemistry, Bagley Hall, Box 351700, University of Washington, Seattle, WA 98195-1700. E-mail: turecek@macmail'chemwashingt°n'edu * Present address: Department of Organic Chemistry, University of Copenhagen, [)enmark.

nium [1, 15b], dimethylammonium [12], and trimethyla m m o n i u m [13] have been found to be metastable on the time scale of several microseconds, such that stable "survivor" ions were detected following collisional reionization to cations. Methylated a m m o n i u m radicals showed unusual deuterium isotope effects on their metastability, both direct and inverse [12, 13], that were attributed to the properties of excited electronic states formed by electron capture [12]. Ab initio calculations [12-14], laser photoionization [12, 16], and photoexcitation [12, 16] have been used to explore the ground and excited electronic states of small a m m o n i u m radicals and to explain their unimolecular dissociations by N - H and N-CH3 bond cleavages. Larger organic a m m o n i u m radicals have also been studied by NRMS [17-21]. Benzylammonium radicals dissociate extensively by N - H and N--CH2aryl bond cleavages that depend on the substituents at the nitrogen atom and in the benzene ring [19]. Leaving group abilities in the dissociations by N - H and N - C bond cleavages were also studied for alkyl and alkenylamm o n i u m radicals [17, 21] and found to depend on the nature of the substituents. A particularly intriguing feature of these dissociations was the inefficient loss of methyl groups that was observed for several systems [17-211. In the present work we investigate the properties of hypervalent radicals derived from cyclic a m m o n i u m

© 1998 American Society for Mass Spectrometry. Published by Elsevier Science Inc. 1044-0305/98/$19.00 Pll $1044-0305(97)00249-3

Received July 23, 1997 Revised October 21, 1997 Accepted October 21, 1997

J Am Soc Mass Spectrom 1998, 9, 242-254

HYPERVALENT PYRROLIDINIUM RADICALS BY NRMS

? / / CH3SSCH3

ring opening

N~.H2

-I-. CH3SSGH

3

? Scheme I ions. Cyclic ions have a special place in mass spectrometry in that their dissociations require cleavage of at least two bonds, a feature that usually stabilizes the ion kinetically [22]. For cyclic ammonium radicals one can envision that a ring cleavage will lead to an open-ring alkyl radical, which, although being a transient species, could have a microsecond lifetime. Competitive losses of the dangling substituents at the nitrogen atom (R ~ or R2, Scheme I) will result in a dissociation to form stable pyrrolidine molecules. This may allow for a better distinction of competitive losses of R ~ and R2 and evaluation of their leaving group abilities in dissociations of hypervalent ammonium radicals. The cyclic systems investigated here include pyrrolidinium (1H), N-methylpyrrolidinium (2H), N-ethylpyrrolidinium (3H), N-phenylpyrrolidinium (4H), N,N-dimethylpyrrolidinium (5), N-methyl-N-ethylpyrrolidinium (6), and several derivatives labeled with deuterium in the ammonium group (1D, laD, 2D, 3D, 4D), pyrrolidine ring (lbH) or substituents (2all, 2aD, 3all, 3aD).

Experimental

Methods Neutralization-reionization mass spectra were obtained on a tandem quadrupole acceleration-decelera-

{> 1:

la: 2;

2a: 3:

3a: 4;

R=H R = D R = OH 3 R = CD s R = C2H 5 R = OD2CH 3 R = C6H 5

lb

1H*:

R1,R 2 = H

1D*:

R 1 - H, R z - D,

laD*:

R1,Rz = D

2H+:

R 1 = C H 3, R a = H

2D+: 2all+: 2aD*:

R 1 - O H 3, R 2 = D

3H*:

R 1 = C 2 H 5, R 2 = H

3 D +1

R 1 - C2H!i, R 2 - D

R I = C D a, R 2 = H R 1 = O H 3, R 2 = D

3aH*:

R 1 = CDaCH3, R 2 - H

3aD*:

R ~ - C D 2 C H ~, R z = D

4H+:

R 1 - C,6H5, R 2 = H

4D*: 5+:

R 1 = C6H5, R z = D R 1, R 2 = C H : ,

6+;

R 1 = CH3, R 2 _ C2PI 5

lbH +

243

tion mass spectrometer, as described previously [23]. A m m o n i u m cations were generated by chemical ionization (CI) in a tight ion source at 180-200°C and at reagent gas pressures ensuring efficient protonation and minimizing the formation of amine cation radicals. This precaution was important in order to prevent contamination of the (M + H)- ions with the ~3C and 15N isotope satellites of M ~ at the same nominal m / z values. The optimized reagent gas pressures in the ion source were estimated at 0.1-(I.2 torr. Tile [M + H] ~/ [M +'] ratios for the pyrrolidines were >20 in most instances. Methane (99.97%), isobutane (99.5°), ammonia (99.99%, all Matheson), and the pyrrolidines themselves in self-CI were used as reagent gases. Gas-phase deuteronations were conducted with ND3-CI (ND3, Cambridge Isotope Laboratories, 99% D3) or acetoned6-CI (acetone-d 6, Cambridge Isotope Laboratories, 99.5% D) as a less expensive substitute [24]. The cations were extracted from the ion source, passed through a quadrupole mass filter operated in an rf-only mode, accelerated to 8200 eV and neutralized by collisions with gaseous dimethyldisulfide, which was admitted to a floated collision cell at pressures allowing 70% transmittance of the precursor ion beam. The residual ions were reflected by a special cylindrical lens, and the neutral intermediates were reionized to cations by collisions with oxygen in a downbeam collision cell. The oxygen pressure was adjusted to achieve 70% transmittance of the precursor ion beam. Neutral lifetimes corresponding to the drift time between the neutralization and reionization cells were 4.0-4.1/xs for 1H'-lbH', 4.4-4.5 /~s for 2 H - 2 a D , 4.8 /~s for 3H" - 3 a D a n d 5', 5.1/zs for 6, and 5.8/~s for 4 H and 4D'. The ions exiting the reionization cell were decelerated to 75-80 eV kinetic energy by a special multielement lens [23], energy filtered by an asymmetric chicane lens [23], and mass analyzed by a quadrupole mass filter, which was operated at unit mass resolution. The quadrupole ac and dc voltages were scanned in J~ink with the deceleration lens potentials, which achieved mass selection of both the precursor and product ions [7, 23]. Typically 25-50 repetitive scans were collected and averaged at 75 data points per mass unit. Variable-time NRMS [25, 26] was used to monitor neutral dissociations. In these experiments, oxygen was admitted to the segmented neutral drift region and the segments were floated at combinations of fixed +250 V and variable negative kV potentials to create observation windows for the products of neutral and post-reionization dissociations, as described in detail previously [27]. Collisionally activated dissociation (CAD) mass spectra were obtained as kinetic energy scans on the Copenhagen JEOL HX-110/HX-110 four-sector tandem mass spectrometer of an EBEB geometry. 8-keV precursor cations were selected by the first two sectors (E~B~)at a mass resolution of ~-1000 and dissociated in a collision cell located in the 3rd field-free region between B1 and E2. Oxygen was used as a collision gas at pressures allowing 70% transmittance of the precursor ion beam.

244

FROSIG A N D TURE~EK

J A m Soc Mass Spectrorn 1998, 9, 242-254

Table 1. Protonation energetics for 1~ Acid

mine, and (iii) Eschweiler-Clark modification of the Wallach-Leuckart reductive methylation with formaldehyde and formic acid to give 7 [29]. N-(Methyl-dg)pyrrolidine (2a) was prepared by reduction of N-ethoxycarbonylpyrrolidine with LiA1D 4 in ether. N-(Ethyl-l,l-d2)pyrrolidine was prepared by reduction of N-acetylpyrrolidine with LiA1D4. Pyrrolidine2,2,5,5-d 4 (lb) was prepared by reduction of succinimide (Aldrich) with LiA1D4 in refluxing ether. The products were purified by distillation and gave satisfactory 1H-NMR and mass spectral data.

- &Hr.29~

CH~ t-C4H9~ (CH3)2C-OH ÷

391 (397) b 140 b 131 b 89 ~-25 ~

NH~ 1 + "(self-CI)

aFrom the corresponding proton affinities (kJ mol 1) [47]. bproton affinities f r o m Szulejko and M c M a h o n [48]. °Estimated f r o m an analogous self-CI of d i m e t h y l a m i n e [12].

Materials Pyrrolidine (1), N-methylpyrrolidine (2), and N,N,N',N'tetramethylbutane-l,4-diamine were purchased from Aldrich and purified by several freeze-pump-thaw cycles before use. N-Ethylpyrrolidine (3) was synthesized from pyrrolidine by acetylation (acetyl chloride and triethylamine in ether, 72% yield) followed by LiA1H4 reduction in ether (52% yield), and purified by distillation. Product 3 was characterized by its 1H-NMR spectrum and showed a single peak by GC-MS. NPheny]pyrrolidine (4) was prepared by cyclization of aniline with 1,4-dibromobutane (Aldrich) in ethanol. The product was purified by vacuum distillation and showed satisfactory 1H-NMR and GC-MS. N,N'-Dimethy}-N,N'-diethylbutane-l,4-diamine (7) was prepared from butane-l,4-diamine in three steps: (i) acetylation to N,N'-diacetylbutane-l,4-diamine [28], (ii) reductxon with LiAIH4 to N,N'-diethylbutane-l,4-dia-

Results To characterize neutral dissociations by NRMS, the intervening ion dissociations must be distinguished. In particular, CAD of precursor ions concomitant with neutralization can produce neutral fragments that are then reionized and detected in the NR spectra [30]. CAD spectra are therefore needed for reference to identify neutral fragments originating by ion dissociations. Collisional reionization often results in ion dissociations that are superimposed on neutral dissociations and complicate the spectra [26]. It is therefore important to find signatures for the presumed neutral products; this can be achieved through survivor or fragment ions that are unique for the NR mass spectrum of the given neutral product [19]. The ion dissociations are therefore discussed first.

1°°1

3O

100

H~H 8

32

c

43 50~

) A

10

30

50

nt/z

70

45

10

10

30

50

70

32

lOO7

D

C

30

m/z

73

50

70

Figure 1. Collisionally activated dissociation mass spectra of (a) pyrrolidinium (1H+), (b) pyrrolidinium-N,N-d:: (laD+), and (c) pyrrolidinium-2,2,5,5-d4. Oxygen at 80% ion beam transmittance was used as the collision gas.

J Am Soc Mass Spectrom 1998, 9, 242-254

Ion Formation and Dissociations Protonations of I were carried out with gas-phase acids ranging from the most energetic CH~ to the least energetic 1 ÷. produced under self-CI conditions, as shown in Table 1. The CAD spectrum of 1H ÷ from self-CI protonation showed consecutive losses of H" and H 2 (rtl/g 71, 70, 68), elimination of ammonia (m/z 55), loss of ethyl (m/z 43), and the formation of CH2NH ~ at m / z 30 (Figure la). Doubly charged precursor ions and their dehydrogenation products also appeared in the spectrum at m / z 36, 35.5, 35, and 34.5, respectively. The CH~--protonated ion 1H* gave a similar CAD spectrum (not shown). The largest difference was in the relative intensities of the doubly charged ions, which were about three times greater for the more energetic 1H * from protonation with CH~. The CH2NHd and (1H-NHz) + ions also appeared in the metastable-ion spectrum of 1H +. Although the relative intensities of metastably formed ions may depend on the precursor ion energy [31], those of CH2NH~ and (1H-NH3) ~ from 1H ' prepared by protonations of vastly different exothermicities showed little sensitivity to internal energy effects. The CAD spectrum of the labeled ion laD~ showed losses of H', HD and (HD + D), which were more abundant then losses of D" or H 2, [laD-HI +/ [ l a D - D ] *" = 7 (Figure lb). The other dissociations, which were analogous to those in the CAD spectrum of 1H +, were the clean elimination of ND2H that gave rise to the fragment ion at m/z 55, elimination of C2H 5 and C2H4D forming the ions at m/z 45 and 44, and formation of CH2ND ~ at m / z 32. The CAD spectrum of l b H + showed loss of H " and HD, [ l b H - H ] + ' / [ l b H D ] * = 4 (Figure lc). These relative abundances can be fitted roughly by assuming a random loss of H from all positions in 1H + and an isotope effect, kH/k D = 2.2, favoring the loss of light hydrogen from the labeled ions. The CAD spectrum of l b H + showed a clean formation of CD2NH ~ at m/z 32. However, the elimination of ethyl resulted in the formation of ions at m/z 47 (retaining four deuterium atoms) and m/z 44 (retaining one deuterium atom) in a 2:3 ratio. The former ion points to the elimination of a C-3q2-4 fragment with transfer of a hydrogen atom from the ammonium group. The formation m / z 44 ion must involve loss of a C-2-C-3 fragment with transfer of D from C-5. The ion dissociations and the neutral fragments formed thereby are summarized in Scheme II. CAD of the N-methylpyrrolidinium ion 2H ~ resulted in eliminations of H (m/z 85), H 2 (m/z 84), CH 4 (m/z 70), C2H 5 (//'//7, 57), and C3H 6 (m/z 44) (Figure 2a). The CAD spectrum of the N-D-labeled ion 2D* showed mass shifts corresponding to eliminations of H" (m/z 86), H 2 (m/z 85), CH 4 (11l/2 71), CRH5 (m/z 58), and C3H6 (m/z 45) (Figure 2b). The ammonium deuteron was cleanly retained in the nitrogen-containing fragments and did not participate in the hydrogen transfers leading to the eliminations of CH 4 and C2H 5.

245

HYPERVALENT PYRROLIDINIUM RADICALS BY NRMS

.÷

H

H

c

N•H2 H

H

-C3H6

+

CHzNH z m/z

30

-....

1H + I .NH3

I

04H7.4-

C-~- C

rn/z

I

55

-C2Hs

-C2H5

m/z

43

m/z

43

Scheme II

Note that the doubly charged ion 2D 2+" appears in the CAD spectrum at m / z 43.5 and is clearly distinguished from the singly charged fragment ions (Figure 2b). The CAD spectrum of the CD3-1abeled ion 2all ÷ showed clean eliminations of CDBH (m / z 70), C:H 5 (m / z 60), and C g H 6 (11I/2 47) (Figure 2c). Hence, the latter two dissociations must involve the ring carbon atoms only. Finally, CID of 2aD ÷ resulted in elimination of CD3H (m/z 71), C2H5 (m/z 61), and C3H 6 (m/z 48) (Figure 2d), which were in keeping with the dissociations of the other isotopomers. CAD of the N-ethylpyrrolidinium ion 3H ÷ resulted in eliminations of H (m/z 99), CH 4 (/*//:~ 84), C2H 4 ~, (m/z 72-70), and CBH6 (lll/z 58) (Figure 3a). Deuterium labeling in 3D * resulted in mass shifts that indicated losses of H" (m/z 100), CH 4 (m/2; 85), C2H4_.6 (m/z 73-71), and C3H6 (m/z 59, Figure 3b). Hence, the ammonium deuteron remained in the nitrogen-containing fragment ions. CAD of ion 3all + showed loss of H " (m/z 101), CH4, CH3D , and CH2D 2 (m/z 8 6 - 8 4 ) , C2H3D2 ( / / / / 7 71), C 2 H 4 D 2 (m/z 70), and CBH 6 (m/z 60) (Figure 3c). The elimination of the C2H 4 (~ neutral fragments thus involved mostly the N--ethyl group. Finally, CAD of ion 3aD + showed eliminations of CH3D, C2H3D 2, C2H4D2, and C3H6, which were in keeping with those in 3all + and 3D + and indicated that the elimination of ethane involved a transfer of a hydrogen atom from the ring methylenes but not from the ammonium group. CAD of the N-phenylpyrrolidinium ion 4H ÷ showed a major loss of H" (m/z 147), and eliminations of C2H5_6 (rn/z 117-119, poorly resolved in the MIKE s p e c t r u m ) , C 3 H 6 (m/z 106), the phenyl group (m/z 71) and benzene (m/z 70). Aromatic series fragments at m/z 51, 77, and 91 were also formed (Figure 4a). Charge stripping gave rise to an abundant doubly charged ion 4H 2+" at m / z 74. Deuterium labeling in the ammonium group in 4D ÷ resulted in the e~pected mass shifts in the CAD spectrum thai: indicated losses of H ' (m/z 148), C2H 5 6 (m/z 119-;120), C3H,s (m/z 107), C6H 5 (m/z 72), and C6H 6 (m/z 71, Figure 4b). The

246

FRf~SIG A N D TURECEK

1007

J A m Soc Mass Spectrom 1998, 9, 242-254

47

100

44

85

.

n 88

so~

50 57

6O 70

28

28 15

A

10

I 50

30

rn~

C

30

50

70 m/z

45

©

100

70

10

48

o" 'c.~

'°°l

86

d

]

89

50 42 58 71

61

29

71

29 15

B

10

30

50

70

,,~

D

10

30

50

70 m/z

Figure 2. Collisionally activated dissociation mass spectra of (a) N-methylpyrrolidinium (2H ~), (b) N-methylpyrrolidiniurn-N-d (2D+), (c) N-(methyl-dg)pyrrolidinium (2all+), and (d) N-(methyld3)pyrrolidinium-N-d (2aD+). Collision conditions as in Figure 1.

clean loss of the unlabeled aromatic ring in the latter two dissociations, as well as the formation of C6H~(m/z 77, Figure 4b) indicated that deuteronation occurred predominantly at the pyrrolidine nitrogen atom and not in the benzene ring. The quaternary pyrrolidinium ions 5 + and 6 + were prepared by losses of N,N-dimethyl and N-methyl-Nethyl groups from ionized N,N,N',N'-tetramethylbutane-l,4-diamine and N,N'-dimethyl-N,N'-diethylbutane-l,4-diamine, respectively. Cleavages of the N - C bonds in amine cation-radicals with charge retention in the alkyl group are unusual and indicate intramolecular participation by the other dialkylamino group to form a ring structure [32]. The cyclic structure for 5 + was inferred from the following arguments. The CAD spectrum of 5 + differed from those of 12 other C6H13N+ isomers [33] in that it showed elimination of methane in addition to forming (CH3)2N==CH ~- at m/z 58 as the most abundant fragment. The latter i m m o n i u m ion is typical of the ring fragmentations in pyrrolidinium ions as observed for 1H+-3H + (vide supra). In addition, dissociation of the N,N,N'N'-tetramethyl-l,l,4,4-D 4butane-l,4-diamine cation-radical [34] gave cleanly 5-D4~ (m/z 104), indicating that the loss of the N(CH3) 2 group occurred without intramolecular H / D exchange. The CAD spectrum of 5-D~- gave mostly CD2~

N(CH3) f (m/z 60) in keeping with structure 5 +. Structure 6 ~ was assigned by analogy with the lower homolog 5 +. In summarizing the ion dissociations, it may be pointed out that they mostly involved eliminations of C2H5 and C3H6 by fragmentations of the pyrrolidine ring, and simple losses of N-substituents as radicals or their eliminations as molecules accompanied by hydrogen transfer from the ring methylene groups.

Neu tralization-Reionization Bond dissociations in hypervalent radicals formed by neutralization of pyrrolidinium ions are expected to produce inter alia N-substituted pyrrolidines. NR spectra of pyrrolidine ions 1 ~'-4 +. were therefore obtained for reference and are briefly discussed together with the spectra of the hypervalent radicals.

Pyrrolidinium The NR mass spectrum of 1+ showed a survivor ion at m/z 71, a major (M-H) + fragment at m/z 70, and the C2HsN +" ion at m/z 43 due to loss of ethylene (Figure 5a). The NR spectrum of the N - D labeled pyrrolidine ion l a +" also showed loss of H' and formation of

J A m Soc Mass S p e c t r o m 1998, 9, 242-254

H Y P E R V A L E N T P Y R R O L I D I N I U M R A D I C A L S BY NRMS

58

100

247

60

,°° t ~2Hs

F~D2CH3

i

a

C

70

71

50

101 28

43

84 43

25

A

50

0

75

m/z

C

25

I

50

86

75

m/~'

100

61

59

,°°l

72

72

100

50~

5(]

i

102

4 85

29

31

42

B

2s

so

~

7s

~oo

D

25

44

50

m/z

75

100

Figure 3. CollisionaUyactivated dissociation mass spectra of (a) N-ethylpyrrolidinium (3H+), (b) N-ethylpyrrolidinium-N-d, (c) N-(ethyl-l,l-d2)pyrrolidinium (3all+), and (d) N-(ethyl-l,1d2)pyrrolidinium-N-d. Collision conditions as in Figure 1.

C2H4DN +" by loss of ethylene, l b ÷" lost mostly D' from the a,a'-positions, [ l b - D ] ' / [ l b - H ] + > 20, and formed a 3:1 doublet of C2H2D3N +" and C2HD4N-" ions at m / z 47 and 48 due to eliminations of C2H3D and C2H4, respectively. These products indicated that the elimination of ethylene involved hydrogen exchange between the ring methylene groups. Similar eliminations of C2H3 D and C2H 4 were observed in the 70-eV electron ionization mass spectrum of l b and were attributed to ion dissociations following reionization. The NR spectrum of the C2HsN-" ion (m/z 43) from 1--" was very similar to that of the CH2~NH-CH2"' distonic ion, but differed from those of aziridine and N-methylmethyleneimine ions, as described previously [34]. The ions at m/z 71, 70, and 43 provided signatures for I in the NR spectra. Neutralization of 1H + resulted in extensive dissociation, such that no survivor ions were detected at m / z 72 following reionization (Figure 5b). The signature ions at m/z 71, 70, and 43 indicated formation of 1 by loss of H" from hypervalent radical 1H. In addition, the spectra showed a peak of C3H ~" at rn/z 42, whose relative abundance could not be accounted by a dissociation of reionized 1 +. (Figure 5a). The formation of C3H 6 can be due in part to CAD of 1H + concomitant with collisional neutralization, or result from a ring fragmentation in 1H' following N - C bond cleavage. The

latter neutral dissociation would also form the stable CHaNH2 radical [35], which would appear at m / z 30 following reionization. However, the NR spectra of 1H + showed CH2NH2~ relative intensities comparable to those in the NR spectrum of 1 +', indicating that CH2NH 2 was not formed efficiently from 1H'. It should be noted that the NR spectra of 1H + .generally displayed greater relative intensities of low mass fragments than did the spectrum of 1 +.. This can be explained by internal energy effects. Dissociations of hypervalent a m m o n i u m radicals are typic:ally 20-100 kJ mo1-1 exothermic [12], so that the pyrrolidine molecules formed must contain a portion of the available energy in the form of vibrational excitation. This vibrational energy is carried over to reionized 1 +. and, in combination with Franck~Condon energy acquired on reionization, drives its dissociations. Internal energy effects in NR mass spectrometry have been discussed recently [36]. This interpretation is also consistent with the dependence of the NR spectra on the internal energy of the 1H + precursor ions, as determined by the protonation exothermicity (Table 1). Ions 1H ÷ generated by the most exothermic protonation with t-C4H~ showed the largest extent of fragmentation on NR and gave low relative intensities of 1 +. and (l-H) +. Further information was obtained from the NR spectra of labeled ions 1 D ' , l a D *, and l b H ÷ (Figure 6a-c).

248

FRfDSIG A N D

TURE(~EK

J ,~tl]] Soc M a s s S p e c t r o m 1998, 9, 2 4 2 - 2 5 4

147

'°°1 !

28

70

H/ x126H5

"

106

43 H

74 71

91

25

A

118

77

-

50

75

71

14 125

100

m/z

1007

1H +

39

G

148

72

50 ¸ 107

K10 1

74.5

B

25

50

75

-

120

72177 1OO

125

m/z

Figure 4. Collisionally activated dissociation mass spectra of (a) N-phenylpyrrotidinium (4H') and (b) N-phenylpyrrolidiniumN-d (4D *). Collision conditions as in Figure 1.

In contrast to 1H', small fractions of m e t a s t a b l e radicals w e r e f o r m e d b y n e u t r a l i z a t i o n of l a D - a n d l b H ' that a p p e a r e d as s u r v i v o r ions in the NR spectra at m / z 74 a n d 76, respectively (Figure 6b,c). The s u r v i v o r ion a b u n d a n c e s , relative to the s u m s of N R ion intensities, °./o~.[NR , w e r e 0.5 a n d 0.23% for l a D 4 a n d l b H + , respectively. The p r e s e n c e of s u r v i v o r ions of l a D + w a s further corroborated b y the signature ions of ( l a D - H ) + , (laD-CRH4) ~, a n d CH2ND2 ~ at m/z 73, 45, and 32 in the NR s p e c t r u m (cf. the C A D s p e c t r u m of l a D - in Figure lb). It s h o u l d be noted that the 2 u m a s s shift upon quantitative deuteronation with NDd, la -~ l a D ~, effectively p r e c l u d e d c o n t a m i n a t i o n b y 13C2 and 13C15N i s o t o p e satellites of l a - , w h i c h c o u l d a m o u n t to <0.002% of l a D + p e a k intensity. Interestingly h o w e v e r , the NR s p e c t r u m of l a D * f r o m self-CI d e u t e r o n a t i o n s h o w e d m u c h smaller s u r v i v o r ion a n d its s i g n a t u r e ion p r o d u c t s . This s u g g e s t e d that the m e t a s t a b l e fractions l a D ' d e c r e a s e d w i t h the d e c r e a s i n g internal energy' of the p r e c u r s o r ion, as o b s e r v e d p r e v i o u s l y for several h y p e r v a l e n t o x o n i u m a n d a m m o n i u m radicals [12, 37-

401. The s u r v i v o r ion of l b H ~ w a s p r o n e to c o n t a m i n a tion b y a 0.2°/,, fraction of ~3C-lb ~', as j u d g e d from the r e s p e c t i v e ion relative intensities in the CI s p e c t r u m , [m/z 75]/[m/z 76] = 0.04 a n d from the 4.8°/,, comb i n e d fraction of 13C a n d LSN isotope satellites at m/z

lr0

Ii i

'

30

71

510

I

70 m,Z

Figure 5. Neutratization-reionization (CH~SSCH3, 70"&T/02, 70%'F) mass spectra of (a) pyrrolidine (1 ~) and (b) pyrrolidinium (1tt),

76. This w a s c h e c k e d b y the s u r v i v o r ion s p e c t r u m [41, 42] of the p r e c u r s o r ions at m/z 75 a n d 76, w h i c h g a v e relative intensities of [m/z 76]/[m/z 75] = 0.067. This i n d i c a t e d that a 28';4, fraction, (6.7 4.8),/6.7 = 0.28, of the p e a k at m/z 76 c o r r e s p o n d e d to r e i o n i z e d l b H ~, c o n f i r m i n g that a fraction of h y p e r v a l e n t l b H w e r e metastable. The existence of s u r v i v o r l b H ~ w a s also consistent w i t h the. p r e s e n c e in the s p e c l r u m of signature f r a g m e n t ions at m/z 74, 47, 44, and 32 (cf. Figure 6c), a l t h o u g h s o m e o v e r l a p with the f r a g m e n t s from dissociations of l b H a n d reionized l b ~ was possible. The NR s p e c t r u m of the N(H,D)-labeled ion 1D ~, p r e p a r e d b y t-C4H. ~ p r o t o n a t i o n of l a , s h o w e d a small fraction of s u r v i v o r ions at m / z 75 (0.12% w ~1NR, Figure 6a). H o w e v e r , this c o u l d be w h o l l y a c c o u n t e d for b y the r e s i d u a l ~3C and lSN satellites of l a " , w h i c h a m o u n t e d to 0.9% of 1D- in intensity in the CI s p e c t r u m (Figure 6a, inset). Hence, the existence of m e t a s l a b l e 1D " w a s not s u p p o r t e d by the N R spectra. The NR s p e c t r u m of 1D a l l o w e d one to e s t i m a t e the kinetic i s o t o p e effect on the c o m p e t i t i v e loss of H or D from radical 1 D . After corrections for the consecutive losses of F{ and D from the p r i m a r y dissociation p r o d u c t s , the f o r m a t i o n of l a b y loss of H was f o u n d to be p r e f e r r e d by a factor of 2.3, w h i c h g a v e a r o u g h m e a s u r e of the i n t r a m o l e c u l a r kinetic i s o t o p e effect [43].

J Am Soc Mass Spectrom 1998, 9, 242-254

HYPERVALENTPYRROLIDINIUMRADICALSBY NRMS

249

26 42

15

xlo

85 xlO

2H +

laD + ~'~H 3

L_

vvwL_ .

L

x5

.....

dL

'd__G

4O

5

2aD +

o318

lbH +

,o ~

~xlO

x,o

~-.xlO

lb

I

3'0

510

J rn/z

70

Figure 6. Neutralization-reionization (CH3SSCH3, 70%T/O2, 70%T) mass spectra of (a) pyrrolidinium-N-d (1D÷), (b) pyrrolidinium-N,N-d2 (laD+), and (c) pyrrolidinium-2,2,5,5-d4 (lbH+).

In summarizing this part it can be concluded that hypervalent pyrrolidinium radicals form small fractions of metastable species of microsecond lifetimes for Nand C-deuterated isotopomers. The main dissociation is loss of the ammonium hydrogen, which shows a moderate isotope effect.

N-Methylpyrrolidinium N-Methylpyrrolidine (2) is one of the expected dissociation products of N-methylpyrrolidinium radical (2H'), and its identification through NR mass spectra was therefore studied first. The NR spectrum of 2 ÷. showed a survivor ion at m / z 85, a (M-H) + ion at m / z 84, and a ring fragment due to loss of ethylene at m / z 57 (Figure 7a). In addition, a fragment ion at m / z 42 appeared in the NR mass spectrum, which showed a clean mass shift to m / z 45 when formed from 2a ÷'. Hence, this fragment ion must contain the original N-methyl group and likely corresponds to C H g N = CH ÷. The m / z 85, 84, 57, and 42 ions and their deuterated analogs were used as signatures for the formation of 2 from radical 2H'. Interestingly, neutralization of 2 +. formed a fraction of neutral 2 that dissociated on the microsecond time scale of the measurement. This was investigated by variable-time

10

I

30

I

I

50

I

m/z

I

70

I

9b

Figure 7. Neutralization-reionization (CH3SSCH3, 70%T/O2, 70%T) mass spectra of (a) N-methylpyrrolidine (2÷'), (b) Nmethylpyrrolidinium (2H+), and (c) N-(methyl-d:0pyrrolidiniumN-d (2D+).

measurements, which showed the [2 +'] and [2-H] + relative abundances decreasing from 15% to 9% for 2 lifetimes increasing from 0.4 to 2.0/zs. At the same time, the [2-H] ÷/[2 +'] ratio, which is indicative of ion dissociations of reionized 2 +', decreased from 7.2 to 5.5. NR of the N-methylpyrrolidinium ion 2H + resulted in complete dissociation, such that no survivor ions were detected at m / z 86 (Figure 7b). The NR spectrum showed signature ions for 2 and abundant low mass fragments of the H0_2CN and C2H1_4 groups ( m / z 25-28), whereas the m / z 70 and m / z 43 signature ions for 1 were weak. Very similar patterns were found for the dissociations of the labeled analogs 2D', 2all', and 2aD" (Figure 7c), which gave no detectable survivor ions in their NR mass spectra. Weak peaks (~0.2%~/NR) at m / z 89 corresponding by mass to 2all-- were occasionally observed in the NR spectra of 2all', which, however, correlated with the relative intensity of residual 2a +" and were most likely due to 13C isotopomers. Hence, the NR spectra did not support the existence of 2H" or its isotopomers of ~s lifetimes. The NR spectra allowed us to estimate the leaving group abilities for H', D, CH3, and CD3 in dissociations of hypervalent N-methylpyrrolidinium radicals. To quantify the neutral products, the neutral fluxes ([neutral]) were estimated from the relative abundances of

250

FRIEISIGAND TUREOEK

J Am Soc Mass Spectrom 1998, 9, 242-254

26

unique signature ions in the NR spectrum of the pyrrolidinium cation (INR/~INR), their fractions in the reference NR spectrum (fn,~f - In,'~JXIn~f), and the estimated reionization efficiency (aR) , as exemplified for 1H + and 1 in eq 1:

[1] ~

INR(1)

1

EINR(1H) ~R(1) f~r~(1)

(1)

The corresponding combined fractions 0Cnref) for [M +] and [M-H] + from the reference NR spectra of 1, 2, la, and 2a were 0.142, 0.131, 0.167, and 0.180, respectively. The relative reionization efficiencies a R were estimated by using the atomic increments of Fitch and Sauter [44]. The latter scheme provides good (_+10%) absolute estimates for electron-impact ionization cross sections of aliphatic molecules [45] and represents a zero-order approximation for estimating relative reionization efficiencies [26]. Equation 1 holds as long as the reference NR spectra are representative of the neutral products formed from hypervalent ammonium radicals. However, ,dnce NR spectra are sensitive to neutral and ion internal energies [36], eq 1 can be applied to provide only relative fluxes of neutral products, e.g., [1]/[2], under the assumption that the adverse internal energy effects and errors in the cross section estimates cancel out. Nevertheless, it is the relative neutral product abundances that are needed for establishing the competitive leaving group abilities, and so eq 1 retains its validity to this end. The NR spectra indicated that loss of H" from 2H" was greatly preferred over loss of CH 3 (Figure 7b). Quantitative evaluation using eq 1 gave 98/2 ratio for the lo,;s of H" versus loss of CH 3, which did not depend appreciably (_+2%) on the internal energy of the 2H + precursor ion when prepared by protonations with t-C4H,;, NH2, or self-CI (Table 1). Interestingly, deuterium isotope effects slightly suppressed loss of H relative to CD 3 (91/9), but not D relative to CD 3 (90/1(I). Quantification or even distinction of ringcleavage dissociations in 2 H was difficult because of the lack of unique fragments and reference spectra. Possible signatures for ring cleavage in 2 H were indicated by the weak peaks of the C H 2 = N H C H 3 immonium ion at m/z 44 and the fragment ion due to loss of C::H, (m/z 58) (Figure 7b). The former peak showed a mass :shift to m/z 48 in the NR spectrum of 2aD~, which indicated retention of both the ammonium deuteriurr and the methyl group (Figure 7c). Likewise, a mass s.hift revealed loss of C2H 5 (or C2H3D) from 2aD " (m/z 61), which indicated ring cleavage in the radical. Note that these peaks are free from interferences from post-reionization dissociations of l a * or 2a ~. The NR spectra thus clearly indicated that ring cleavage in 2 H did occur, but only as a minor dissociation pathway. The competition between ring cleavage and loss of methyl was further studied with the N,N-dimethylpyrrolidinium radical 5", which lacks the labile ammonium

39 42 H3C4 ~H 3

71

15

14

L I

10

58

44

H3Gd ~2H5

I

81 xlO

39

F

t

f

30

~

I

'

2 81 J. ~ x

[

F

50

1

70

' ~

l I

m/z

90

'

O I

112

I ~

110

Figure 8. NeutraliTation-reionization (CH3SSCH3, 70%T/O2, 70%T) mass spectra ,of (a) N,N-dimethylpyrroIidinium (5*) and (b) N-methyl-N-ethy[pyrrolidinium (6').

hydrogen atom. Radical 5" was generated from the N,N-dimethylpyrrolidinium ion 5 ~. The NR spectrum of 5 ~ showed nitrogen-containing fragments at m/z 58 and 44 (Figure 8a), whereas the signature ions for 2 (m/z 85, 84, and 57) were very weak or absent. The spectrum showed no survivor ions at m / z 100 attesting to the low stability of the intermediate quaternary ammonium radical 5" [18]. The formation of the stable CH2N(CH3)2 [34l and N(CH~) 2 146] radicals can be readily envisioned by ring cleavage in 5 followed by elimination of C3H 6 or C4Hs, respectively (Scheme liB. In keeping with this interpretation, the NR spectrum a ~'N\

•

H2

H3d"~ H 3

/

HsC

X"CH3

5

z~

C4H8

H3C; NGH3

Scheme llI

if_ _~ -C~Ht

H3C/ ''oH3

J Am Soc Mass Spectrom 1998, 9, 242-254

HYPERVALENT PYRROLIDINIUM ]CADICALS BY NRMS

251

27

~5

84

39

3H"

a ] -CsHe

d ,," -CH a ", ¢

15

_,:=,_x,oj

.2 °

'1

(7

ZJ'

/

C3HaN +

m/z 58

84

3D+]

'

JL

V' 84

I

I

2O

I

4 0I

e

d

' u0

C5HlIN

+"

m/z 85

SchemeIV '~hqWV~ x l O

cf %.~

"? 9-9

'

i

m/z

8'0

~

1 0'0

Figure

9. Neutralization-reionization (CH3SSCH 3, 70%T/O 2, 70%T) mass spectra of (a) N-ethylpyrrolidine (3+'), (b) N-ethylpyrrolidinium (3H+), and (c) N-ethylpyrrolidinium-N-d(3D ~).

showed the complementary C 4 H d" ion at m/z 56 (Figure 8a).

N-Ethylpyrrolidinium Continuing with the approach outlined above, we first examined the NR spectra of N-ethylpyrrolidine (3) and its labeled derivative 3a to establish signature ions for neutral product identification. NR of 3 +. gave rise to a weak survivor ion at m / z 99 and products of a-cleavage dissociations at m/z 98 (loss of ring H), and m/z 84 (loss of methyl) (Figure 9a). In addition, weak fragment ions at m/z 71 and 70 due to losses of C 2 H 4 and C2H5, respectively, were formed, which were isobaric with the signature ions from 1. Deuterium labeling (3a) revealed that the ethylene molecule originated from both the pyrrolidine ring and the ethyl group in a 60/40 ratio, suggesting a mixture of product ion structures. The NR spectrum of the m / z 71 ion from 3 showed a weak survivor ion and loss of CH~ which were significantly different from the NR fragmentations of 1 +'. The ions at m/z 99, 98, and 84 were used as signatures for 3, and their combined relative intensities (]NR/~INR) w e r e used to quantify 3 in mixture analysis according to eq 1. NR of N-ethylpyrrolidinium ( 3 H ' ) resulted in complete dissociation, such that no survivor ions were observed at m / z 100 (Figure 9b). The NR spectrum

showed weak signature peaks of 3 at m / z 99, 98, and a more abundant peak at 84. Very weak peaks also appeared at m/z 71 and 70 indicating formation of 1. In addition, fragment ions were observed at m / z 85 (loss of CH3) and 58 (loss of C3H6) ,. which were not due to dissociations of reionized 1 ~" or 3 +. and must have corresponded to ring cleavage in 3H'. Equation 1 was used to quantify the competitive losses of H and C2H~s to establish the relative leaving group abilities. The figures for 3H" generated from 3H + of different internal energies (by protonation with t-C4H~, NH~, and selfCI) showed consistently 80/20 ratios for loss of H and C2H~, respectively. Hence loss of H" was preferred. Since the fraction for the C 2 H 5 loss could be affected by the overlap at m/z 70 and 71 of the signature ions for 1 and the fragment ions from reionized 3 +', measurements were also made for 3D" (Figure 9c) and 3aD" (spectrum not shown), where the fragments appeared at different mass-to-charge ratio values. The branching ratios for loss of D" and C 2 H 5 o r CH3CD2, 83/17 and 77/23 for 3 D and 3aD', respectively, were similar to those for dissociations of 3 H and indicated that the interferences in the NR spectra of the latter were small. The fragments due to the ring dissociation showed the expected mass shifts upon deuterium labeling. Thus, shifts of m / z 58 to m/z 59, 60, and 61l were observed for 3D', 3all', and 3aD, respectively, indicating retention of the a m m o n i u m hydrogen and the exocyclic methylene group in the ion. The loss of CH3 leading to the fragment at m/z 85, which is very weak in the NR spectrum in Figure 9c, increased markedly with the increasing energy of the precursor ion. Deuterium labeling showed complete retention of the exocyclic methylene in the (3H-CH3) ~' ion, whereas --50% of the a m m o n i u m hydrogen was lost. The results of the deuterium labeling experiments are summarized in Scheme IV, which depicts the proposed mechanisms for the ring cleavage fragmentations of 3H'. Quantification of the ring cleavage dissociations was difficult because of the lack of reference spectra for the ring-open intermediates and the final products. Nevertheless, it appeared from the NR spectra that ring cleavage dissociations of 3H " were substantially less abundant than the loss of the a m m o n i u m hydrogen atom.

252

FROSIG A N D TURECEK

J A m Soc Mass Spectrom 1998, 9, 242-254

N-Methyl-N-ethylpyrrolidinium Radical 6 was generated from the N-methyl-N-ethylpyrrolidinium cation (6+), which was in turn produced by dissociative ionization of N,N'-dimethylN,N'-diethylbutane-l,4-diamine in a fashion analogous to the generation of ion 5 + (vide supra). Ion 6 + underwent complete dissociation upon NR, such that no surviw~r ions were present in the spectrum at m / z 114 (Figure 8b). Signature ions for the formation of 2 by loss of C2H ~ (m/z 85, 84, 57) and 3 by loss of CH3 (m/z 99, 98, 85) were extremely weak or absent, indicating that 6 mostly decomposed by ring cleavage dissociations. These were indicated directly by losses of C2H 4 (m/z 86), and C3H 6 (m/z 72), which were analogous to the fragments observed for the lower homolog 5. The conspicuous C2H6N + ion at m/z 44 was formed by elimination of ethylene from C4H10N ~ (m/z 72), which is typical of i m m o n i u m ion dissociations [33]. It m a y be concluded that in the absence of the weak N - H bond, the quaternary a m m o n i u m radicals dissociate primarily by cleavage of the ring N - C bonds.

39 .50

77

L 77 4H1 ~6N5

.,,dJ

ULK

L4_.¢

"~sH5

._

77

Z

I

L/

N-Phenylpyrrolidinium Neutralization of the N-phenylpyrrolidinium ion 4H ÷ formed radical 4H" whose dissociations were investigated. N-phenylpyrrolidine (4), which was one of the possible dissociation products, was characterized by its NR mass spectrum (Figure 10a). The spectrum showed the survivor ion of 4 +. at m / z 147 and fragments due to loss of H (m/z 146), H and C2H 6 (m/z 117), C3H 7 (m/z 104), the phenyl ion at m / z 77, and low-mass fragments of the aromatic series. A very similar NR spectrum was obtained for the (M-H) + ion at m/z 146 (spectrum not shown). NR of ion 4H ÷ prepared by protonation with N H 4 gave the spectrum shown in Figure 10b. A very similar spectrum was obtained from 4H + prepared by a more exothermic protonation with t-C4H~. Both spectra showed a survivor ion at m / z 148 and fragments at m / z 147, 146, 117, 104, and 77, which were characteristic of 4 formed by loss of the a m m o n i u m hydrogen atom, and fragment ions at m / z 70 and 43, which characterized 1 produced by loss of the phenyl group. Quantification of the losses of H" and C6H 5 according to eq i gave a 90/10 ratio favoring the former dissociation. The spectrum gave very little indication of dissociations by the pyrrolidinium ring cleavage in 4H. The increased relative abundance of the ions at m / z 105 and 106 due to losses of C3H 6 and C3H7, respectively, m a y indicate a ring opening in 4H, by analogy with the dissociations of the aliphatic homologs 2H', 3H', 5, and 6" (vide supra). The identity of the survivor ion of 4H + could not be established unambiguously from the NR spectra of 4H+ alone. Because 4H + has l0 carbon atoms, the contribution of the 13C and ~5N isotope satellites of residual 4 ~ at m / z 148 m a y be significant even under

I

50

~

'

'

'

I

100

'

'

~

m/z

'

I

150

Figure 10. Neutralization-reionization (CH3SSCH:~, 70%T/O2, 70%T) mass spectra of (a) N-phenylpyrrolidine (4+'), (b) Nphenylpyrrolidinium (4H*), and (c) N-phenylpyrrolidinium-N-d (4D+).

conditions of efficient (>95%) protonation, as achieved in the present measurements. The NR mass spectrum of the N-deuterated ion 4D + showed a small peak of survivor 4D ÷ and fragments due to losses of both hydrogen and deuterium atoms. The former dissociation, which must have occurred in the ion, clearly indicated that a small fraction of 4H" survived to be reionized. However, the spectrum did not indicate whether the surviving species was the hypervalent radical 4H', or an isomeric radical due to ring opening a n d / o r rearrangement of the pyrrolidine ring. The fragments characteristic of loss of the phenyl group showed clean mass shifts in the NR spectrum of 4D', e.g., m / z 70 to m / z 71, and m / z 43 to m / z 44, which indicated retention of the a m m o n i u m deuterium in its original position. The spectrum also showed partial mass shifts of m / z 91 to m / z 92 and m / z 117 to m / z 118, which could not originate from dissociations of 4*' and thus were indicative of a cleavage of the pyrrolidine ring and retention of the a m m o n i u m deuterium atom.

Discussion The experimental results for dissociations of hypervalent radicals 1H" through 6" showed some general fea-

j Am Soc Mass Spectrom 1998, 9, 242-254

HYPERVALENT PYRROL1DINIUM RADICALS BY NRMS

tures that are n o w discussed. The radicals d e r i v e d from p y r r o l i d i n e a n d N - m o n o s u b s t i t u t e d p y r r o l i d i n e s inv a r i a n t l y u n d e r w e n t p r e f e r e n t i a l losses of a m m o n i u m h y d r o g e n atoms. Thus, h y d r o g e n a t o m w a s the b e s t l e a v i n g g r o u p in these dissociations. This conclusion w a s also in line w i t h the results of several p r e v i o u s studies of alkyl [12, 13, 17], a l k e n y l [21], a n d b e n z y l a m m o n i u m radicals [19] d e r i v e d from tertiary amines. The reasons for the facile N - H b o n d c l e a v a g e in h y p e r v a l e n t a m m o n i u m radicals are b a s e d on b o t h t h e r m o c h e m i s t r y a n d kinetics. As d e d u c e d from ab initio calculations for s m a l l e r s y s t e m s [12, 13], dissociations of N - H b o n d s in s e c o n d a r y a n d t e r t i a r y h y p e r v a l e n t a m m o n i u m radicals are ~ 2 3 a n d 70 kJ m o l 1 exothermic, respectively, a n d thus t h e r m o d y n a m i c a l l y possible. H o w e v e r , dissociations of N - C b o n d s are even m o r e e x o t h e r m i c (95-105 kJ tool 1 [12-14]), a n d so the p r e f e r e n c e for the N - H b o n d dissociations m u s t be kinetic in nature. P r e v i o u s calculations i n d i c a t e d that the N - H b o n d dissociation in tertiary a m m o n i u m radicals h a d a v e r y small p o t e n tial e n e r g y b a r r i e r for the g r o u n d electronic state, so that the radicals w e r e only w e a k l y b o u n d or u n b o u n d w i t h r e s p e c t to loss of H . These conclusions a p p e a r to also a p p l y to the dissociations of 1H" t h r o u g h 4H', which s h o w small or no fractions of m e t a s t a b l e radicals a n d p r e d o m i n a n t H losses. In contrast, cleavages of N - C b o n d s have substantial activation b a r r i e r s ( 5 0 - 6 0 kJ tool ~) a n d c o m p e t e p o o r l y w i t h the h y d r o g e n loss. This conclusion h o l d s even for losses of the larger ethyl a n d p h e n y l substituents, w h i c h were less efficient than loss of H', b u t m o r e efficient than loss of methyl. The chemical rationale for the facile cleavage of the N - H b o n d s in h y p e r v a l e n t a m m o n i u m radicals stems from their electronic structure [12, 13]. In the g r o u n d electronic state the o d d electron enters a diffuse orbital (SOMO), w h i c h is d e l o c a l i z e d over the a m m o n i u m n i t r o g e n a n d h y d r o g e n atoms, w h i c h thus c a r r y substantial n e g a t i v e charge a n d spin density. In a d d i t i o n , the S O M O is a n t i b o n d i n g along the N - H b o n d , w h i c h contributes to its lability. The m u c h subtler differences in the l e a v i n g g r o u p abilities for m e t h y l , ethyl, a n d p h e n y l are m o r e difficult to interpret, b e c a u s e they m a y involve dissociations from excited electronic states as recently e s t a b l i s h e d for d i m e t h y l a m m o n i u m [12]. The dissociations of q u a t e r n a r y a m m o n i u m radicals 5" a n d 6, w h i c h i n d i c a t e d preferential ring c l e a v a g e dissociations, b e l o n g to the s a m e category. It m a y be noted, h o w e v e r , that recent orbital analysis of organic a m m o n i u m radicals in their g r o u n d electronic states i n d i c a t e d that the S O M O w a s d e l o c a l i z e d over the s u b s t i t u e n t s d u e to m i x i n g w i t h w-type orbitals [121]. This m a y s u g g e s t that the -a'(CH2) orbitals of the ethyl g r o u p a n d the p y r r o l i d i n e ring a n d the p h e n y l vr s y s t e m interact w i t h the diffuse s a n d p atomic orbitals on the a m m o n i u m nitrogen. These three-electron interactions give rise to m o l e c u l a r orbitals, w h i c h are a n t i b o n d i n g a l o n g the N - C a n d / o r C - C b o n d s [21], a n d c o u l d p o s s i b l y l o w e r the activation barriers to N - C b o n d cleavages a n d facilitate e x o t h e r m i c dissociations.

Conclusions

253

N R M S a n a l y s i s of the dissociations of h y p e r v a l e n t a m m o n i u m radicals d e r i v e d f r o m p y r r o l i d i n e a n d N s u b s t i t u t e d p y r r o l i d i n e s allows us to arrive at the f o l l o w i n g conclusions. The h y d r o g e n a t o m is the best l e a v i n g g r o u p in the dissociations of h y p e r v a l e n t radicals d e r i v e d from s e c o n d a r y a n d tertiary amines. This conclusion is c o r r o b o r a t e d b y p r e v i o u s studies of other t y p e s of h y p e r v a l e n t organic a m m o n i u m radicals a n d a p p e a r s to be g e n e r a l l y valid. H y p e r v a l e n t radicals d e r i v e d from q u a t e r n a r y a m m o n i u m ions d i s s o c i a t e d m a i n l y b y fission of the p y r r o l i d i n e ring. These dissociations can be r a t i o n a l i z e d q u a l i t a t i v e l y b y c o n s i d e r i n g the electronic p r o p e r t i e s of the frontier m o l e c u l a r orbitals in the radicals. H o w e v e r , further theoretical studies of h y p e r v a l e n t a m m o n i u m radicals are n e e d e d to prov i d e e n e r g y data, e l u c i d a t e the isotope effects on metastability, a n d r a t i o n a l i z e the u n u s u a l c h e m i s t r y of these transient species.

Acknowledgments Support of this work by a grant from the National Science Foundation (CHE-9412774) is gratefully acknowledged. LF thanks the Denmark-America Foundation, the Fullbright Commission, and the Anders Maanssons Foundation for fellowships during which tenure this work was carried out. Thanks are also due to Jill Wolken, Aaron Frank, Dr. Viet Q. Nguyen, and Dr. Martin Sadilek for technical assistance.

References 1. Gellene, G. I.; Cleary, D. A.; Porter, R. F. ]. Chem. Phys. 1982, 77, 3471. 2. Gellene, G. I.; Porter, R. F. Ace. Chem. Res. 1983, 16, 200. 3. Wesdemiotis, C.; McLafferty, F. W. Chem. Rev. 1987, 87, 485. 4. Terlouw, J. K.; Schwarz, H. Angew. Chem. h~t. Ed. Engl. 1987, 26, 805. 5. Holmes, J. L. Mass Spectrom. Rei~. 1989, 8, 513. 6. McLafferty, F. W. Science, 1990, 247, 925. 7. Turecek, F. Org. Mass Spectrom. 1992, 27, 1087. 8. Goldberg, N.; Schwarz, H. Acc. Chem. Res. 1994, 27, 347. 9. Perkins, C. W.; Martin, J. C.; Arduengo, A. J.; Lau, W.; Alegria, A.; Kochi, J. K. J. Am. Chem. Soc. 1980, 102, 17753. 10. Herzberg, G. FartTday Discuss. R. Soc. Chem. 1981, 71, 165. 11. (a) Gedye, R. N.; Sadana, Y. N. J. Or[¢. Chem. 1980, 45, 3721. (b) Kariv-Miller, E.; Nanjudiah, C.; Eaton, J.; Swenson, K. E. J. Electroanal. Chem. 1984, 167, 141 12. Nguyen, V. Q.; Sadilek, M.; Frank, A. J.; Ferrier, J. G.; Turecek, F. J. Phys. Chem. /t, 1997, 101, 3789. 13. Shaffer, S. A.; Tureeek, F. J. Altl. Chem. Soc. 1994, 116, 8647. 14. Boldyrev, A. I.; SJmons, J. J. Chem. Phys. 1992, 97, 6621. 15. (a) Williams, B. W.; Porter, R. F. J. Chem. Phys. 1980, 73, 5598. (b) Jeon, S.-J.; Raksit, A. B.; Gellene, G. I.; Porter, R. F. J. Ant. Chem. Soc. 1985, 107, 4129. (c) Gellene, G. I.; Porter, R. F. Int. J. Mass Spectrom. Ion Processes 1985, 64, 55. 16. Sad/lek, M.; Tureeek, F. Chem. Phys. Lett. 19q6, 263, 203. 17. Shaffer, S. A.; Tureeek, F. J. Am. Soc. Mass :~pectrom. 1995, 6, 1004. 18. Beranova, S.; Wesdemiotis, C. l,~t. J. Mass 5~pectrom. hm Processcs 1994, 134, 83. 19. Shaffer, S. A.; Sadflek, M.; Turei!ek, F. I. Or)y. Chem. 1996, 61, 5234.

254

FROSIG AND TURECEK

20. Wolken, J. K.; Nguyen, V. Q.; Turecek, F. ]. Mass Spectrom. 1997, 32, 1162-1169. 21. Shaffer, S. A.; Wolken, J. K.; Turecek, F. J. Ant. Soc. Mass Spect~'om. 1997, 8, 1111-1123. 22. McLafferty, F. W.; Turecek, F. Interpretation of Mass Spectra, 4th ed.; Uniw~rsity Science Books: Mill Valley, CA, 1993, pp 68, 178. 23. Ture~!ek, F.; Gu, M.; Shaffer, S. A. J. Am. Soc. Mass Spectrom. 1992, 3, 493. 24. Vairamani, M.; Mirza, U. A.; Srinivas, R. Mass Spectrom. Rev. 1990, 9, 235. 25. Kuhr~s, D. W.; Tran, T. B.; Shaffer, S. A.; Ture~ek, F. J. Phys. Chem. 1994, 98, 4845. 26. KuhrLs, D. W.; Ture~ek, F. Org. Mass Spectrom. 1994, 29, 463. 27. Sadfl,.~k, M.; Ture~ek, F. J. Phys. Chem. 1996, 100, 224. 28. Dickey, F. H.; Fickett, W.; Lucas, H. J. J. Am. Chem. Soc. 1952, 74, 944. 29. Icke, R. N.; Wisegarver, B.; Alles, G. A. Organic Synthesis Coll. Vol. HI, Wiley: New York, 1955, p 723. 30. Polce, M. J.; Beranova, S.; Nold, M. J.; Wesdemiotis, C. J. Mass Spectrom. 1996, 31, 1073. 31. Proctor, C. J.; McLafferty, F. W. Org. Mass Spectrom. 1983, 18, 193. 32. McLafferty, F. W.; Turecek, F. Interpretation of Mass Spectra, 4th ed.; LIniversity Science Books: Mill Valley, 1993, pp 214-215.

J Am Soc Mass Spectrorn 1998, 9, 242-254

33. Bowen, R. D. Mass Spectrom. Rev. 1991, 10, 225. 34. Shaffer, S. A.; Turecek, F.; Cerny, R. L. L Am. Chem. Soc. 1993, 115, 12117. 35. Griller, D.; Lossing, F. P. J. Am. Chem. Soc. 1981, 103, 1586. 36. Nguyen, V. Q.; Turecek, F. J. Mass Spectrom. 1996, 31,843. 37. Holmes, J. L.; Sirois, M. Org. Mass Spectrom. 1990, 25, 481. 38. Wesdemiotis, C.; Fura, A.; McLafferty, F. W. ]. Am. Soc. Mass Spectrom. 1991, 2, 459. 39. Sirois, M.; George, M.; Holmes, J. L. Orq,. Mass Spectrom. 1994, 29, 11. 40. Sadilek, M.; Turecek, F. J. Phys. Chem. 1996, ~00, 15027. 41. Turecek, F.; Gu, M. Org. Mass Spectrom. 1992, 27, 1335. 42. Turecek, F.; Gu, M. J. Mass Spectrom. 1995, 30, 144. 43. Derrick, P. J. Mass Spectrom. Rev. 1983, 2, 285. 44. Fitch, W. L.; Sauter, A. D. Anal. Chem. 1983, 55, 832. 45. Turecek, F.; Brabec, L.; Hanus, V~; Zima, V.; Pytela, O. Int. J. Mass Spectrom. lon Processes 1990, 97, 117. 46. McMillen, D. F.; Golden, D. M. Annu. Rev. Fhys. Chem. 1982, 33, 493. 47. Lias, S. G.; Liebman, J. F.; Levin, R. D. J. Phys. Chem. ReJq Data 1984, 13, 695. 48. Szulejko, J. E.; McMahon, T. B. ]. Am. Chem. Soc. 1993, 115, 7839.