Dynamics of anilinium radical α-heterolytic fragmentation processes. Electrofugal group, substituent, and medium effects on desilylation, decarboxylation, and retro-aldol cleavage pathways

Article abstract of DOI:10.1021/ja981541f

A single electron transfer (SET) photosensitization technique in conjunction with time-resolved, laser spectroscopy has been employed to generate and kinetically analyze decay processes of anilinium radicals derived by one-electron oxidation of α-anilinocarboxylates, β- anilinoalcohols, and α-anilinosilanes. In this manner, the rates of unimolecular decarboxylation of aniliniumcarboxylate radicals were determined to be in the range 10⁶ - 10⁷ s^-1 and dependent upon solvent polarity, the nature of the metal cation, and substituents on the aniline ring, nitrogen, and α-carbon. In addition, kinetic analysis of base-induced retro-aldol fragmentations of cation radicals arising by SET oxidation of β- anilinoalcohols has shown that they occur with bimolecular rate constants which vary from 10⁴ to 10⁵ M^-1 s¹. These values are close to those for α-deprotonation reactions of related N,N-dialkylanilinium radicals. The retro-aldol fragmentation rates, like those for α-decarboxylation, also vary in a patterned way with changes in arene ring, nitrogen, and α- and β- carbon substituents. An investigation of the dynamics of methanol-promoted reactions of α-(trimethylsilyl)methyl-substituted anilinium radicals, has demonstrated that a change in the nitrogen substituent from alkyl to acyl causes an ca. 10-fold increase in the desilylation rate. Parallel photochemical studies have been conducted to gain chemical evidence to support assignment of the anilinium radical decay pathways in the LFP experiments and to demonstrate the preparative consequences of the kinetic results. First, clean formation of products derived by coupling of the (N- methylanilino)methyl radical in photochemical reactions of 1,4- dicyanobenzene with either tetra-n-butylammonium N-methyl-N-phenylglycinate or β-(N-methyl-N-phenyl)aminoethanol shows that the respective decarboxylation and retro-aldol cleavage processes occur with exceptionally high efficiencies. Second, in accord with the high rates observed for aminium radical decarboxylation and base-induced retro-aldol fragmentation, tethered cyclohexenone - α-aminocarboxylates and - β-aminoethanols undergo high- yielding SET-promoted photocyclization reactions under both direct and SET- sensitized conditions. Last, results which depict how the rates of aminium radical α-fragmentation correlate with quantum efficiencies of SET-promoted reactions of tertiary amines and amides have come from a study of photocyclization reactions of N-(aminoethyl)- and (amidoethyl)phthalimides. The quantum yields for these SET-promoted processes are observed to vary with the electrofugal group and nitrogen substituent in the manner predicted on the basis of the LFP-determined fragmentation rates.

Full text of DOI:10.1021/ja981541f

10676
J. Am. Chem. Soc. 1998, 120, 10676-10686
Dynamics of Anilinium Radical R-Heterolytic Fragmentation
Processes. Electrofugal Group, Substituent, and Medium Effects on
Desilylation, Decarboxylation, and Retro-Aldol Cleavage Pathways
Zhuoyi Su,† Patrick S. Mariano,*,†,1 Daniel E. Falvey,‡ Ung Chan Yoon,§ and Sun Wha Oh§
Contribution from the Department of Chemistry, UniVersity of New Mexico,
Albuquerque, New Mexico 87111, Department of Chemistry and Biochemistry, UniVersity of Maryland,
College Park, Maryland 20742, and Department of Chemistry, College of Natural Sciences,
Pusan National UniVersity, Pusan 609-735, Korea
ReceiVed May 4, 1998
Abstract: A single electron transfer (SET) photosensitization technique in conjunction with time-resolved,
laser spectroscopy has been employed to generate and kinetically analyze decay processes of anilinium radicals
derived by one-electron oxidation of R-anilinocarboxylates, â-anilinoalcohols, and R-anilinosilanes. In this
manner, the rates of unimolecular decarboxylation of aniliniumcarboxylate radicals were determined to be in
the range 106-107 s-1 and dependent upon solvent polarity, the nature of the metal cation, and substituents on
the aniline ring, nitrogen, and R-carbon. In addition, kinetic analysis of base-induced retro-aldol fragmentations
of cation radicals arising by SET oxidation of â-anilinoalcohols has shown that they occur with bimolecular
rate constants which vary from 104 to 105 M-1 s1. These values are close to those for R-deprotonation reactions
of related N,N-dialkylanilinium radicals. The retro-aldol fragmentation rates, like those for R-decarboxylation,
also vary in a patterned way with changes in arene ring, nitrogen, and R- and â-carbon substituents. An
investigation of the dynamics of methanol-promoted reactions of R-(trimethylsilyl)methyl-substituted anilinium
radicals, has demonstrated that a change in the nitrogen substituent from alkyl to acyl causes an ca. 10-fold
increase in the desilylation rate. Parallel photochemical studies have been conducted to gain chemical evidence
to support assignment of the anilinium radical decay pathways in the LFP experiments and to demonstrate the
preparative consequences of the kinetic results. First, clean formation of products derived by coupling of the
(N-methylanilino)methyl radical in photochemical reactions of 1,4-dicyanobenzene with either tetra-n-
butylammonium N-methyl-N-phenylglycinate or â-(N-methyl-N-phenyl)aminoethanol shows that the respective
decarboxylation and retro-aldol cleavage processes occur with exceptionally high efficiencies. Second, in
accord with the high rates observed for aminium radical decarboxylation and base-induced retro-aldol
fragmentation, tethered cyclohexenone-R-aminocarboxylates and -â-aminoethanols undergo high-yielding
SET-promoted photocyclization reactions under both direct and SET-sensitized conditions. Last, results which
depict how the rates of aminium radical R-fragmentation correlate with quantum efficiencies of SET-promoted
reactions of tertiary amines and amides have come from a study of photocyclization reactions of N-(aminoethyl)-
and (amidoethyl)phthalimides. The quantum yields for these SET-promoted processes are observed to vary
with the electrofugal group and nitrogen substituent in the manner predicted on the basis of the LFP-determined
fragmentation rates.
Introduction
associated with the charge and odd electron characters of ion
radical intermediates can foster the rational design of new and,
in some cases, synthetically important redox processes.3 Al-
though the current level of knowledge about the chemistry of
anion and cation radicals is high,4 much less is known about
the dynamics of even the most common reactions of these
intermediates despite the fact that these data are of central
importance to predicting and controlling the efficiencies of SET-
promoted reactions. The relevance of ion radical reaction
dynamics to this issue is readily demonstrated by use of the
hypothetical photoinduced SET process given in Scheme 1. It
can be seen that the rates of the two cation radical reactions (k1
Ion radicals serve as key intermediates in a large number of
ground-state and excited-state redox reactions. As a result,
knowledge about the chemical properties of these short-lived
species leads to a greater understanding of the detailed mech-
anisms of those redox reactions which are initiated by single
electron transfer (SET).2 Also, information about the properties
† University of New Mexico.
‡ University of Maryland.
§ Pusan National University.
(1) A major portion of the studies reported in this paper was conducted
as part of the doctoral work of Z.S. at the University of Maryland, and a
preliminary communication reporting some of the results of the laser flash
photolysis study has appeared (Su, Z.; Falvey, D. E.; Yoon, U. C.; Mariano,
P. S. J. Am Chem. Soc. 1997, 119, 5261).
(2) (a) Mann, C. K.; Barnes, K. K. Electrochemical Reactions in Non-
Aqueous Systems; Marcel-Dekker: New York, 1970. (b) Fox, M. A.,
Chanon, M., Eds. Photoinduced Electron Transfer; Elsevier: New York,
1988.
(3) Mariano, P. S. Tetrahedron 1983, 39, 3845. Mariano, P. S. Synthetic
Organic Photochemistry; Horspool, W. M., Ed.; Plenum Press: London,
1984.
(4) (a) Chanon, M.; Rajzmann, M.; Chanon, F. Tetrahedron 1990, 46,
6193. (b) Schmittel, M.; Burghart, A. Angew Chem., Int. Ed. Engl. 1997,
36, 2550.
10.1021/ja981541f CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/01/1998

Dynamics of Anilinium Radical Fragmentation Processes
J. Am. Chem. Soc., Vol. 120, No. 41, 1998 10677
Scheme 1
Scheme 2
and k2) govern both the relative yields of product 1 and product
2 (1/2 k1/k2) as well as the quantum efficiency (φ) for product
formation (φ [k1 + k2]/[k1 + k2 + kBSET]).
Amine cation radicals (aminium radicals) are among the most
interesting members of the charged radical family owing to the
role they play as intermediates in oxidation reactions of amines.
The chemical reactivity of these transients governs the nature
and efficiencies of a variety of processes initiated by SET
photochemical,5 metal cation,6 electrochemical,7 and enzymatic8
oxidation.
The most common reaction pathway followed by aminium
radicals involves base-promoted deprotonation at either nitrogen
(for primary (1°) and secondary (2°) aminium radicals) or the
R-carbon (for tertiary (3°) aminium radicals).5 The kinetics of
these acid-base reactions were investigated originally by using
product distribution techniques.5,9 More recently, reliable data
on the rates of R-CH deprotonation reactions of 3° aminium
radicals (eq 1) have come from direct measurements by use of
Included in this general family are the R-decarboxylation,16
R-desilylation,15 and R-retro-aldol17 cleavage reactions shown
in eqs 2-4. However, little18 is known about the dynamics of
these reactions and, equally significant, how the rates of these
fragmentations are governed by factors (e.g., solvents, additives,
N-blocking groups) which can be experimentally manipulated.
pulse radiolysis,10 stopped-flow,11 electrochemical,12 and laser
flash photolysis (LFP) techniques.13 In addition, these studies
have yielded information about how the rates of R-CH depro-
tonation are affected by base strength, R-isotopic substitution,
and substituents at the nitrogen and R-carbon centers.
From a synthetic perspective, SET-oxidations of 3° amines
can be used as a method to generate both R-amino radicals and
iminium cations (Scheme 2). These intermediates, having
opposite charge affinity profiles, participate in interesting C-C
bond-forming reactions with respective electron-poor and
electron-rich alkenes.14 The yields of these processes when
conducted on complex molecular systems can be severely
compromised by (1) the lack of regioselectivity associated with
an unselective aminium radical deprotonation step and (2) the
comparable rates of oxidation of the reactant and product
amines. Fortunately, there exist other types of R-fragmentation
reactions of 3° aminium radicals which can serve as potentially
more selective surrogates for the R-deprotonation process.9,14,15
The aim of the investigation described below was to
determine the rates of the common aminium radical fragmenta-
tion reactions, to delineate how the rates vary with solvents,
additives, and substituents, and to assess the significance of the
rate data in terms of the design of selective and efficient
photoinduced SET oxidation reactions of amines.1
Results
Laser Flash Photolysis Studies. Earlier,13 we showed that
laser flash excitation of tertiary anilines in the presence of the
electron acceptor 1,4-dicyanobenzene (DCB) leads to production
of spectroscopically detectable anilinium radicals whose decay
kinetics could be quantitatively analyzed. This SET sensitization
technique is employed in the current study to generate tertiary
anilinium radicals from designed series of R-anilinocarboxylates
1-7, â-anilinoethanols 8-15, and R-anilinotrimethylsilanes 16-
18. The decay kinetics of anilinium radicals derived from each
of these substances has been measured and analyzed in terms
of respective decarboxylation, retro-aldol cleavage, and R-de-
silylation reactions which result in formation of corresponding
(5) Lewis, F. D. Acc. Chem. Res. 1986, 19, 401. See also: Pienta, N. J.
Photoinduced Electron Transfer; Fox, M. A., Chanon, M., Eds.; Elsevier:
New York, 1988; Part C.
(6) Zhang, X. M.; Mariano, P. S.; Fox, M. A.; Martin, P. S.; Merkert, J.
Tetrahedron Lett. 1993, 34, 5239. Castro, P.; Overman, L. E.; Zhang, X.
M.; Mariano, P. S. Tetrahedron Lett. 1993, 34, 5243.
(7) Yoshida, K. Electrooxidations in Organic Chemistry; Wiley: New
York, 1984.
(8) Van Houten, K. A.; Kim, J. M.; Bogdan, M.; Ferri, D. C.; Mariano,
P. S. J. Am. Chem. Soc. 1998, 120, 5864.
(9) Xu, W.; Mariano, P. S. J. Am. Chem. Soc. 1991, 113, 1431. Xu, W.;
Zhang, X. M.; Mariano, P. S. J. Am. Chem. Soc. 1991, 113, 8863.
(10) Das, S.; von Sonntag, C. Z. Naturforsch. 1986, 416, 505.
(11) Dinnocenzo, J. P.; Banach, T. E. J. Am. Chem. Soc. 1989, 111,
8646.
(12) Parker, V. D.; Tilset, M. J. Am. Chem. Soc. 1991, 113, 8778.
(13) Zhang, X.; Yeh, S. R.; Hong, S.; Freccero, M.; Albini, A.; Falvey,
D. E.; Mariano, P. S. J. Am. Chem. Soc. 1994, 116, 4211.
(14) Jeon, Y. T.; Lee, C. P.; Mariano, P. S. J. Am. Chem. Soc. 1991,
113, 8863.
(16) Davidson, R. S.; Steiner, P. R. J. Chem. Soc. (C) 1971, 1682.
Davidons, R. S.; Harrison, K.; Steiner, P. R. J. Chem. Soc. (C) 1971, 3480.
Davidson, R. S.; Orton, S. P. J. Chem. Soc. Chem. Communm. 1974, 209.
See also: Battacharyya, S. M.; Das, P. K. J. Chem. Soc., Faraday Trans.
1984, 80, 1107.
(17) Gaillard, E. R.; Whitten, D. G. Acc. Chem. Res. 1996, 29, 292.
(18) (a) This is not true for the retro-aldol cleavage process which has
been subjected to thorough kinetic studies by Schanze and co-workers (ref
18b) and to a lesser extent the desilylation process which has been the subject
of preparative photochemical studies (refs 9, 14, and 15). (b) Burton, R.
D.; Bartberger, M. D.; Zhang, Y.; Eyler, J. R.; Schanze, K. S. J. Am. Chem.
Soc. 1996, 118, 5655. Lucia, L. A.; Burton, R. D.; Schanze, K. S. J. Phys.
Chem. 1993, 97, 9078.
(15) Yoon, U. C.; Mariano, P. S. Acc. Chem. Res. 1992, 25, 233.

10678 J. Am. Chem. Soc., Vol. 120, No. 41, 1998
Su et al.
Table 1. Second-Order Rate Constants (kBSET) for Decay of the Respective Cation and Anion Radicals Derived from the Anilines 22-24 and
DCB at 25 °C
a
kBSET × 10-10 (M-1 s-1
)
solvent
viscosity (η)
22
DCB
23
DCB
24
DCB
EtOH
108
55
35
3.5 ( 0.2
4.7 ( 0.1
6.0 ( 0.1
5.6 ( 0.1
4.3 ( 0.1
4.8 ( 0.3
6.5 ( 0.4
6.2 ( 0.2
1.1 ( 0.1
1.2 ( 0.1
1.5 ( 0.1
1.3 ( 0.1
1.0 ( 0.1
1.1 ( 0.1
1.6 ( 0.1
1.5 ( 0.1
5.6 ( 0.2
6.1 ( 0.1
6.5 ( 0.1
6.2 ( 0.2
5.4 ( 0.2
6.2 ( 0.1
7.3 ( 0.4
6.9 ( 0.3
MeOH
MeCN
mixtureb
a Error limits were obtained by use of three to five independent experiments each. b MeCN-MeOH, 60:40 v/v.
Scheme 3
R-anilino radicals (eq 5). The aniline derivatives used in this
study were prepared by the straightforward procedures described
in Supporting Information.
Table 2. Second-Order Rate Constants (kdep) for TBAA-Induced
R-CH Deprotonation of the Cation Radicals Derived from Anilines
19-24 in MeCN at 25 °C
aniline
kdep (M-1 s-1
)
aniline
kdep (M-1 s-1
)
19
20
21
2.0 × 105
8.0 × 104
8.9 × 105
22
23
24
(1.3 ( 0.2) × 105
2.7 × 105
2.2 × 104
Experimental Section) and known extinction coefficients19.20
gives the rate constants for bimolecular decay of these transients
(representative data in Table 1).
In each case, decay rates for the anilinium radicals and DCB
anion radical partners are nearly equivalent and dependent on
solvent viscosity. These observations are consistent with the
assignment of diffusion governed back electron transfer as the
major pathway for decay of the transients in the absence of
added base. In addition, para substituents appear to influence
the back-SET rates in the order OMe < H < CF3.
An alternate mode of anilinium radical decay, involving R-CH
deprotonation, is introduced when the base, (nBu)4NOAc
(TBAA), is present in the solution subjected to LFP. The
dependence of the 460-nm transient decay rate on TBAA
concentration is used to determine the anilinium radical second-
order deprotonation rate constants. Accordingly, flash irradia-
tion of solutions containing each of the anilines 19-24, DCB,
and TBAA along with (nBu)4NClO4 (TBAP) to maintain a
constant ionic strength of 0.5 M gives rise to transient anilinium
radicals whose decays are kinetically analyzed by using the
mixed second-order (BSET)-pseudo-first-order (acetate-pro-
moted deprotonation) mechanistic sequence shown in Scheme
3 (see eq 7 in the Experimental Section). This treatment gives
pseudo-first-order rate constants (kobs) for TBAA-induced decay.
Plots of kobs vs TBAA concentration yields slopes from which
the second-order rate constants (kdep) for TBAA-induced
R-deprotonation of the anilinium radicals are calculated (Table
2).
Calibration of the LFP Systems. Initial studies were
conducted with the aniline derivatives 19-24 to allow com-
parison of the results of this effort with those obtained from
our earlier work.13 Laser excitation (308 nm, 6 ns, 50-60 mJ)
of deoxygenated solutions of each of these aniline derivatives
in polar solvents including MeOH, EtOH, and MeCN containing
DCB (50 mM) leads to generation of two transients character-
ized as the corresponding anilinium radical (430-480 nm)19
and DCB anion radical20 (330 nm). Analysis of the decay
profiles by use of a second-order treatment (see eq 6 in the
(19) Arimitsu, A.; Masuhara, N.; Mataga N. J. Phys. Chem. 1975, 79,
1255.
(20) Robinson, E. A.; Schulte-Frohlinde, D. J. Chem. Soc., Faraday
Trans. 1973, 69, 707.
The R-CH deprotonation rate constants for the anilinium
radicals derived from 19-21 are close to those previously
determined,13 thus establishing the integrity of the current

Dynamics of Anilinium Radical Fragmentation Processes
J. Am. Chem. Soc., Vol. 120, No. 41, 1998 10679
Table 3. Second-Order Rate Constants (kBSET) for Decay of the
Ion Radicals from LFP of the â-Anilinoethanols 8-10 and DCB at
25 °C
Table 4. Second-Order Rate Constants (kra) for TBAA-Promoted
Retro-Aldol Cleavage of the Cation Radicals Derived from 8-15 in
MeCN at 25 °C
a
a
a
kBSET × 10-10 (M-1 s-1
)
anilinoethanol
kra (M-1 s-1
)
anilinoethanol
kra (M-1 s-1
)
solvent
8
DCB
9
DCB
10
DCB
8
9
10
11
(4.1 ( 0.6) × 104
(2.8 ( 0.3) × 104
3.1 × 105
12
13
14
15
3.1 × 105
2.0 × 105
3.3 × 104
EtOH
1.2
1.6
2.3
2.2
1.2
1.3
2.3
2.1
0.9
1.2
2.0
1.4
1.1
1.2
1.8
1.4
1.2
1.7
2.1
1.8
1.3
1.8
2.8
2.6
MeOH
MeCN
mixtureb
3.2 × 104
(3.8 ( 0.7) × 105
a Error limits were obtained by use of three to five independent
experiments.
a Maximum error limits of (0.1 × 1010 were determined by use of
three to five independent experiments. b MeCN-MeOH, 60:40 v/v.
Figure 1. Plot of the observed rate constant for decay of the
â-anilinoethanol 8 cation radiacal vs TBAA concentration in 60:40
MeOH-MeCN at 25 °C.
Figure 2. Transient absorption spectra following (1, 2, 5, 10, and 40
µs) 308-nm excitation of and MeCN (25 °C) solution of 1 (0.5 mM)
and DCB (50 mM).
experimental system. The small differences (e.g., for 19•+ 2.0
× 105 M-1 s-1 (current) vs 3.1 × 105 M-1 s-1 (previous))
between the current and previous data are a result of the use of
a mathematical model comprised of a mixed second-order-
pseudo-first-order decay treatment (Scheme 3) in this work as
compared to an inferior one which was used earlier.13
concentration gives the small CD3 isotope effect of 1.3. Both
observations are inconsistent with a mechanism involving R-CH
deprotonation of the anilinoethanol cation radicals since a near-
zero OD isotope effect and large R-CD3 isotope effect13 would
have been expected (see below).
Anilinium Radical Retro-Aldol Cleavage Kinetics. The
LFP methodology discussed above is employed to determine
the decay kinetics of cation radicals derived by SET-sensitized
photolysis of the â-anilinoethanol derivatives 8-15. In each
case, decay of the anilinium radical intermediates in the absence
of base occurs by back electron transfer with bimolecular rate
constants which again depend on the solvent viscosity and para
substituents (Table 3).
The rates of decay of the cation radicals derived by one-
electron oxidation of the â-anilinoethanols depend in a linear
fashion on the concentration of the base, TBAA. The observed
pseudo first-order rate constants for acetate-promoted decay of
these transients at various TBAA concentrations and constant
ionic strength of 0.5 M (maintained with TBAP) are determined
by use of the competitive second-order (BSET)-pseudo-first-
order (acetate-promoted deprotonation) decay sequence like that
shown in Scheme 3 (see eq 7 in the Experimental Section). Plots
of kobs vs TBAA concentration (e.g., Figure 1) yield the second-
order rate constants (kra) listed in Table 4 for TBAA-promoted
decay of the â-anilinoethanol cation radicals.
The new pathway for TBBA-promoted decay of the anilino-
ethanol cation radicals is attributed to retro-aldol cleavage. This
conclusion is based on the results of deuterium isotope labeling
and preparative photochemical studies. The rate constant for
bimolecular decay of the OD isotopomer of 8•+ (measured in
60:40 MeOD-MeCN) is 9.0 × 103 M-1 s-1, which corresponds
to an OD isotope effect of 2.2. In addition, the CD3 derivative
11 was prepared and subjected to LFP study. Analysis of the
dependence of kobs for decay of 11•+ as a function of TBAA
Anilinium Radical r-Decarboxylation Kinetics. An initial
concern in our studies with R-anilinocarboxylates 1-7 is the
question of whether SET-photosensitized oxidation of these
substrates occurs at the tertiary amine center rather than at the
carboxylate moiety. Two factors suggested that anilinium rather
than carboxy radical formation would be favored in these
systems. In the LFP experiment with excitation at 308 nm,
the primary light absorbing species is the anilinocarboxylate
and not the electron acceptor, DCB. Thus, the pathway for ion
radical production involves SET from the excited aniline
chromophore to the ground state of DCB. In addition, the
oxidation potentials of carboxylates are generally in the range
1.2-1.5 V,21 while those of tertiary anilines are ca. 0.3-0.8
V.22 Thus, even if equilibration of the anilinium and carboxy
radicals occurs following the initial SET event, the former
species would be thermodynamically favored.
As expected, LFP excitation (308 nm) of an MeCN (25 °C)
solution of anilinocarboxylate 1 (0.5 mM) containing 50 mM
DCB leads to generation of transients absorbing at 340 and
450-480 nm, characteristic of the respective ion radicals of
DCB•- and 1•+ (e.g., Figure 2). The decay profiles of the
intermediates (e.g., Figure 3) show that the anilinium radicals
disappear at much greater rates than that of DCB•-. Thus, in
(21) Eberson, L. Chemistry of the Carboxyl Group; Patai, S., Ed.;
Wiley: London, 1969; Chapter 2. Torii, S.; Tanaka, H. Organic electro-
chemistry; Lund, H., Baiter, M. M., Eds.; Marcel Dekker: New York, 1991;
Chapter 14.
(22) Cf. Miller, L. L.; Nordblum, G. D.; Mayeda, E. A. J. Org. Chem.
1972, 37, 916. Yoshida, J.; Isoe, S. Tetrahedron Lett. 1987, 28, 6621.
Cooper, B. E.; Owen, W. J. Organomet. Chem. 1971, 29, 33.

10680 J. Am. Chem. Soc., Vol. 120, No. 41, 1998
Su et al.
Table 6. Solvent Polarity Effects on the Decarboxylation Rates of
the Anilinium Radical of 1
solvent
ET(30)a
kdec (s-1) (25 °C)b
MeOH
EtOH
MeCN
55.5
51.9
46.7
2.8 × 106
2.5 × 106
(1.7 ( 0.2) × 106
60:40 MeOH-MeCN
1.9 × 106
a Dimroth, K.; Reichardt, C.; Siepmann, T.; Bohlmann, F. Ann.
Chem. 1963, 661, 1. b Error limits were determined by use of three
independent experiments.
Table 7. Metal Cation Effects on the Decarboxylation Rates of
the Anilinium Radical of 1
kdec × 10-6
kdec × 10-6
concn (s-1) (25 °C,
concn (s-1) (25 °C,
salt
(mM)
MeCN)
salt
(mM)
MeCN)
nBu4NClO4
nBu4NClO4
1
10
2.4
1.2
0.5
1.1
1.9
2.5
2.8
CsClO4
CsClO4
CsClO4
Mg(ClO4)2
Mg(ClO4)2
Ca(ClO4)2
Ca(ClO4)2
1
10
100
1
10
1
4.1
2.9
0.3
0.15
0.12
0.15
0.08
nBu4NClO4 100
LiClO4
NaClO4
KClO4
RbClO4
10
10
10
10
10
Figure 3. Decay profiles for the ion radicals of 1 and DCB obtained
from the LFP experiment described in Figure 2.
Scheme 4
Table 5. First-Order Rate Constants (kdec) for Decarboxylation of
the Cation Radicals Derived from the R-Anilinocarboxylates 1-7 in
MeCN at 25 °C
Figure 4. Plot of the observed rate constant for decay of the cation
radical of anilinosilane 16 vs MeOH concentration in MeCN at 25 °C.
R-anilino-
carboxylate
R-anilino-
carboxylate
a
a
k
dec (s-1
)
k )
dec (s-1
1
2
3
4
(1.7 ( 0.2) × 106
(8.2 ( 0.3) × 105
1.3 × 107
5
6
7
3.6 × 107
1.3 × 106
2.6 × 106
1•+. In addition, metal perchlorate salts also influence the
kinetics of this process. LFP experiments were conducted with
MeCN (25 °C) solutions of 1 (0.5 mM) and DCB (50 mM)
containing varying concentrations of the salts listed in Table 7.
Two trends appear to be associated with the data. First, kdec
decreases as the concentration of the added salt increases and
the magnitude of the effect increase in proceeding from salts
with small, monovalent metal cations to those with large,
divalent metal cations.
2.8 × 107
a Error limits were determined by use of three to five independent
experiments.
this system, back electron transfer is not a major reaction
pathway involved in the decay of the anilinium radicals. Thus,
kinetic treatments of the decay of the anilinium radicals by use
of either the competitive first-order (-CO2)-second-order
(BET) sequence shown in Scheme 4 (see eq 7 in the Experi-
mental Section) or a first-order process (correcting for compa-
rably slow BET, see eq 8 in the Experimental Section) both
give excellent fits of the data and equivalent (within experi-
mental error) first-order decay rate constants (kdec) attributed
to unimolecular R-decarboxylation (see below). This experi-
mental method and the ensuing kinetic treatments when applied
to a series of para-, nitrogen-, and R-carbon-substituted R-ani-
linocarboxylates give the decarboxylation rate data included in
Table 5.
Anilinium Radical r-Desilylation Kinetics. In an earlier
effort,13 we determined the second-order rate constants for
MeOH-, H2O-, and nBu4NF-promoted R-desilylation of the
anilinium radical arising by SET-sensitized laser excitation of
the R-anilinosilane 16. As a consequence of the effects of the
N-substituent on the rates of anilinium radical retro-aldol
cleavage and decarboxylation seen in this study (see Tables 4
and 5), the desilylation process was reinvestigated. Irradiation
(308 nm) of MeCN (25 °C) solutions of 16-18 (0.5 mM) and
DCB (50 mM) containing varying concentrations of MeOH
followed by employment of a competitive second-order (BSET)-
pseudo-first-order (MeOH promoted desilylation) kinetic treat-
ment (see eq 7 in the Experimental Section) leads to the
observed desilylation rate constants. A plot of kobs vs MeOH
concentration (e.g., Figure 4) yields the second-order rate
Additional information about the anilinium radical decar-
boxylation process is gained from studies probing the effects
of solvent and added salts. As seen by inspecting the data given
in Table 6, solvent polarity has a small but patterned effect on
the rate of the decarboxylation reaction of the anilinium radical,

Dynamics of Anilinium Radical Fragmentation Processes
J. Am. Chem. Soc., Vol. 120, No. 41, 1998 10681
Table 8. Second-Order Rate Constants (kDes) for MeOH-Assisted
Desilylation of R-Anilinosilane 16-18 Derived Cation Radicals in
MeCN at 25 °C
Scheme 7
a
a
R-anilinosilane kdes (M-1 s-1
)
R-anilinosilane kdes (M-1 s-1
)
16
17
7.0 × 105
18
6.0 ( 1.0 × 107
1.8 × 107
a The error limit was determined by use of three independent
experiments.
Scheme 8
Scheme 5
cyclization reactions of aminoalkyl-substituted, conjugated
ketones and esters which follow either direct (diradical coupling)
or SET-sensitized (radical addition) pathways. The results
highlighted the importance of aminium radical R-heterolytic
fragmentation in governing both the efficiencies and regio-
selectivities of the cyclization processes. For example, photo-
reactions of R-silylaminoenones 28 (Scheme 7) lead to selective
and high-yielding formation of non-silicon-containing products
27 owing to the larger rates of MeOH-induced desilylation vs
intramolecular deprotonation of intermediate aminium radicals.
In contrast, irradiation of these same substrates in MeCN yields
silicon-containing products 29 which arise by R-deprotonation
of the nitrogen-centered cation radical intermediates.
The results of the LFP studies described above suggest that
SET-promoted photocyclization reactions of R-aminocarboxylate
tethered conjugated ketones in MeCN should favor pathways
in which decarboxylation rather than deprotonation of interme-
diate aminium radicals occurs. Likewise, conjugated ketones
containing aminoethanol substitution should also undergo pho-
tocyclization reactions in the presence of the base, TBAA, which
are characterized by selective retro-aldol fragmentation of amine
cation radical intermediates. These predictions are fully sup-
ported by observations made in studies of the preparative
photochemistry of cyclohexenones 30 and 31 (Scheme 8). For
example, irradiation of the aminocarboxylate 30 in MeCN or
MeOH under either direct or 9,10-dicyanoanthracene (DCA)
sensitized conditions leads to efficient formation of the sprio-
cyclic ketone 32. In a similar manner, 32 is formed in the direct
and SET-sensitized photoreactions of aminoethanol 31. Finally,
DCA-sensitized irradiation of the analogous carbamidocarbox-
ylates 33 and 35 promotes efficient formation of the respective
bicyclic ketones 34 and 36 via pathways involving formation
and decarboxylation of intermediate carbamate cation radicals
(Scheme 9).
Scheme 6
constants for MeOH-promoted desilylation (kdes) of the R-silyl-
anilinium radicals (Table 8).
Preparative Photochemical Studies. Photoreactions of
r-Anilinocarboxylate 1 and â-Anilinoethanol 8 with DCB.
To connect the observations made in the LFP effort to real
intermediates and decay pathways, preparative photochemical
reactions of the R-anilinocarboxylate 1 and â-anilinoethanol 8
with DCB were conducted under conditions which mimic, as
best as possible, those existing in the laser excitation experi-
ments. Accordingly, an MeCN solution of 1 (2.5 mM) and DCB
(2.5 mM) is irradiated with Pyrex filtered light (λ > 290 mm).
An exceptionally clean reaction ensues when the reactant
conversion is ca. 50%. Chromatographic separation affords the
diamine 25 (30%) and cyanobenzylaniline 26 (27%) along with
recovered DCB (55%) (Scheme 5).
In a similar fashion, low-conversion irradiation of 8 (2.4 mM)
in an MeCN solution containing both DCB (2.4 mM) and TBAA
(0.3 M) followed by chromatographic separation gives the
cyanobenzylaniline 26 (32%) along with recovered DCB (63%)
and 8 (41%) (Scheme 6).23
The outcomes of these preparative photochemical processes
are in full accord with the conclusion that the major if not
exclusive mechanistic pathways followed in the SET photo-
reactions of the R-anilinocarboxylates and â-anilinoalcohols
involve respective decarboxylation and retro-aldol cleavage of
intermediate cation radicals. Both R-fragmentation reactions
lead to production of R-anilinomethyl radicals which undergo
either self-coupling to form diamine 25 or addition to the DCB
anion radical to provide the cyanobenzylaniline 26. The
differences between the 25 and 26 product ratios in these
photoprocesses is interesting and is discussed below.
Quantum Efficiencies of N-(Aminoethyl)phthalimide Pho-
tocyclization Reactions. The rates of aminium radical R-het-
erolytic fragmentation also can govern the quantum efficiencies
of SET-promoted photoreactions of amine substrates. In these
processes, fragmentation in the direction of product formation
competes with energy-wasting back electron transfer to regener-
ate the ground-state starting substrate. Thus, aminium radical
fragmentation rates should correlate with the quantum yields
of SET photoreactions conducted under identical conditions
within a regular series of amine substrates. Results which
demonstrates this relationship and show how electrofugal group
and N-substituent selection can be used to optimize quantum
Photocyclization Reactions of Aminocarboxylate- and
Aminoethanol-Substituted Cyclohexenones. Earlier,8,9,14 we
demonstrated the synthetic potential of SET-promoted photo-
(23) (a) Schuster, I. I. J. Org. Chem. 1985, 50, 1656. (b) Robinson, E.
A.; Schulte-Frohlinde, D. J. Chem. Soc., Faraday Trans. 1 1973, 707. (c)
Albini, A.; Fasani, E.; Oberti Tetrahedron 1982, 38, 1034. Lewis, F. D.;
Petisce, J. R. Tetrahedron 1986, 42, 6207.

10682 J. Am. Chem. Soc., Vol. 120, No. 41, 1998
Su et al.
Table 9. Quantum Efficiencies for SET-Promoted
Photocyclizations of the (Aminoethyl)- and
(Amidoethyl)phthalimides 37-42
Scheme 9
phthalimide
R
E
additive
φa
37
38
39
40
41
42
Me
Ac
Me
Ac
Me
Ac
H
H
TMS
TMS
CO2NBu4
CO2NBu4
TBAA
TBAA
TBAP
TBAP
TBAP
TBAP
0.003
0.025
0.038
0.12
0.32
0.40
Scheme 10
a All photoreactions were conducted on MeOH solutions containing
added TBAA (0.3 M) or TBAP (0.3 M) at 25 °C. Conversions are
between 5 and 20%. Quantum efficiencies are reported as averages of
three independent measurements and errors range from (0.0001 to
(0.007.
Scheme 11
silylation-promoted generation and secondary reaction of R-ami-
no radical intermediates (see Scheme 7).
efficiencies have come from quantitative studies with the
N-(aminoethyl)phthalimides 37-42.
Earlier observations made by Davidson16 indicate that
aminium radicals produced by SET oxidation of R-aminocar-
boxylates and â-aminoalcohols decay by decarboxylation and
retro-aldol fragmentation routes. Later studies of the â-amino-
alcohol cation radical fragmentation process have been con-
ducted by Whitten and co-workers.17 The results of our current
work provide further support for these proposals in the context
of the specific systems probed in the LFP effort. Accordingly,
SET-promoted preparative photoreactions of the anilinocar-
boxylate 1 and anilinoethanol 8 with DCB both occur in near
quantitative yield to generate products of direct (25) and cross
(26) coupling of the R-anilinomethyl radical intermediate.
Therefore, R-decarboxylation and retro-aldol cleavage must be
the exclusive processes responsible for anilinium radical decay
in the mechanistic sequences for production of 25 and 26, from
1 and 8 (Scheme 11).
Photoreactions of phthalimides 37-42 occur by pathways
involving initial SET from the amine-nitrogen center to the
excited phthalimide followed by aminium radical fragmentation
to generate the diradical precursor of the cyclic products
(Scheme 10). As can be seen by viewing the data given in
Table 9, quantum yields for photocyclization product formation
vary in a regular manner with the LFP measured rates of
N-alkyl- and N-acylaminium radical decay by desilylation,
decarboxylation, and deprotonation.
Discussion
Anilinium Radical Decay Pathways. A general problem
associated with the interpretation of LFP experiments that rely
on observations of transient decay rates is the unambiguous
assignments of chemical pathways to the decay processes. In
the current effort, we have attempted to accumulate supporting
evidence to enable firm assignments of R-decarboxylation and
retro-aldol cleavage as the major, if not exclusive, mechanisms
responsible for decay of anilinium radicals derived from the
respective R-anilinocarboxylates and â-hydroxyanilines.
Preparative photochemistry often serves as a useful method
to correlate transient decay routes with chemical mechanisms.
An example is found in our earlier studies with R-silylamines.
The fact that desilylation is the major pathway for LFP-detected
R-silylanilinium radical decay is suggested by the observations
that the decay rates show first-order dependences on the
concentration of nucleophiles (e.g., ROH and F-) and a direct
dependence on the silophilicity of the nucleophiles (e.g., F- >
H2O > MeOH > MeCN). This conclusion is strongly supported
by results emanating from preparative photochemical studies14,15
which demonstrate that the major products of R-silylamine SET
photoreactions occurring in silophilic solvents arise by de-
A number of factors could be responsible for the different
25:26 product ratios observed in these reactions. For example,
TBAA-promoted retro-aldol fragmentation of 8•+ produces
acetic acid as a consequence of hydroxyl proton transfer to
acetate. Consequently, protonation or hydrogen bonding23a of
HOAc to the base DCB•- 23b is likely in this case. In this case,
the 25:26 product ratio is then governed by the relative rates of
R-anilino radical coupling with itself vs the dicyanohydro-
phenyl radical DCB-H• or H-bonded DCB•-. In contrast,
decarboxylation of 1•+ does not produce an acid, and as a result,
the 25:26 distribution is controlled by the rates of R-anilino
radical self-coupling vs its addition to DCB•-. It should be
mentioned in this regard that Albini and Lewis23c have
independently observed effects of added trifluoroacetic acid and
trifluoroethanol on the ratios of products formed in SET-
promoted photoreactions of toluene derivatives with poly-
cyanoarenes and have rationalized these results in terms of
hydrogen bonding to or protonation of the intermediate cy-
anoarene anion radicals.23

Dynamics of Anilinium Radical Fragmentation Processes
J. Am. Chem. Soc., Vol. 120, No. 41, 1998 10683
Equally relevant to this issue are the photocyclization
reactions of tethered aminoenone systems (Schemes 8 and 9)
which are driven by the same aminium radical fragmentation
processes. For example, the DCA-photosensitized transforma-
tions of aminocarboxylate 30 and aminoalcohol 31 to spirocyclic
amino ketone 32 occur by routes in which selective decarbox-
ylation and retro-aldol cleavage of the aminium radical inter-
mediates sets the stage for the key R-amino radical cyclization
steps.
Relative Rates of Retro-Aldol vs r-Deprotonation. In-
spection of the rate data included in Tables 2 and 4 reveals a
curious, yet understandable, feature of the kinetics of TBAA-
promoted R-CH deprotonation and retro-aldol cleavage proc-
esses. Accordingly, bimolecular R-CH deprotonation rates for
the N,N-dialkylanilium radicals derived from 19-24 are larger
than those for bimolecular retro-aldol cleavage of comparably
substituted â-hydroxyanilinium radicals [8-15•+]. This com-
parison leads one to question (1) if retro-aldol cleavage is really
the process involved in the LFP-determined transient decay
processes of [8-15•+], and (2) if so, why retro-aldol fragmenta-
tion is faster than R-CH deprotonation in the anilinium radicals
which bear a â-hydroxyl group.
Additional information was accumulated to substantiate the
operation of TBAA-induced retro-aldol fragmentation rather
than R-CH deprotonation pathways for decay of the transients
arising by SET-photosensitized irradiation of the â-hydroxy-
anilines 8-15. If retro-aldol cleavage is the major or ex-
clusive pathway for anilinium radical decay in these systems, a
large primary OD isotope effect (ca. 2)24 would be expected.
Also, since anilinium radical R-CH deprotonation is kinetically
preferred at R-CH3 vs R-CH2-R centers,13,25 a large primary
R-CD3 isotope effect (ca. 3)13 would be anticipated if this
mechanism is responsible for transient decay. The data recorded
for the OD analogue of 8•+ and CD3 analogue 11•+ demon-
strate that TBAA-induced decay of the â-hydroxyanilinium
radicals is associated with large (2.2) OD and small (1.3) R-CD3
isotope effect. Thus, the preparative photochemical and kinetic
isotope effect results leave little doubt that retro-aldol cleavage
is faster than R-CH deprotonation in â-hydroxyanilinium
radicals.
On the basis of these results, it is reasonable to conclude that
â-hydroxyanilinium radicals are intrinsically less reactive than
their non-hydroxyl-bearing analogues. Relevant to this issue
are observations which show that cation radicals derived from
properly structured tertiary diamines exist as through-space,
three-electron N-N-bonded structures26a and that “tied-back”
tertiary aminium radicals form three-electron-bonded dimers
with neutral tertiary amines.26b Thus, it may be generally true
that tertiary aminium radicals can be stabilized by weak,
through-space, three-electron-bond interactions with neighboring
groups which bear nonbonded electrons. Interactions of this
type between the aminium radical and â-hydroxyl centers (e.g.,
45) would lead to stabilization of â-hydroxyanilinium radicals
Figure 5. Linear-free energy plot of the log of the rate constants for
decarboxylation (circles), deprotonation (squares), and retro-aldol
cleavage (diamonds) reactions of anilinium radicals vs σp values of
the p-phenyl substituents, MeO, H, and CF3.
promoted R-CH deprotonation and retro-aldol cleavage reac-
tions. It should be noted that an explanation of this type is in
full accord with the suggestion made by Schanze18b that
â-hydroxyl stabilization is responsible for the smaller rate of
pyridine induced retro-aldol fragmentation of threo vs erthryo
â-hydroxy-R,â-diphenylethylanilinium radicals.
Electrofugal Group and Substituent Effects on Aminium
Radical Fragmentation Rates. In our earlier efforts,13 we
showed that p-phenyl substituents play a significant role in
determining the rates of TBAA-induced anilinium radical R-CH
deprotonation. The current data indicates that parallel substitu-
ent effects exist for both the decarboxylation and retro-aldol
cleavage reactions of anilinium radicals. A simple linear free
energy treatment of the data (Figure 5) yields F values for
deprotonation, decarboxylation, and retro-aldol cleavage of 1.5,
1.8, and 1.6, respectively. The near equality of these values
demonstrates that the transition states of all three fragmentation
processes have similar charge and odd-electron distributions.
As we13 and others27 have discussed, rates of the anilinium
and related cation radical R-CH deprotonation reactions parallel
thermodynamic acidities. Thus, the rates are directly propor-
tional to the oxidation potentials28 of the amine precursors or,
said in another way, inversely related to the thermodynamic
stabilities of the aminium radicals. Since N-acyl substituents
have a large effect on the oxidation potentials of amines, one
would predict that the rates of R-heterolytic fragmentation of
amide-derived cation radicals would be greater than those of
their amine analogues. Indeed, the large increases in the rates
of anilinium radical decay caused by changes from N-alkyl to
N-acyl substituents are observed for all of the fragmentation
processes investigated.
The results of this study also provide some information about
the effects of R-C substituents on the rates of the R-heterolytic
fragmentation reactions. In anilinium radical R-CH deproton-
ation,13 retro-aldol cleavage, and decarboxylation reactions,
(25) Lewis, F. D.; Ho, T. I.; Simpson, J. T. J. Am. Chem. Soc. 1986,
104, 1924. Lewis, F. D.; Ho, T. I.; Simpson, J. T. J. Org. Chem. 1981, 46,
1077.
(26) (a) Kirste, B.; Alder, R. W.; Sessions, R. B.; Bock, M.; Kurreck,
H.; Nelsen, S. F. J. Am. Chem. Soc. 1985, 107, 2635. (b) Dinnocenzo, J.
P.; Banach, T. E. J. Am. Chem. Soc. 1988, 110, 971.
(27) (a) Anne, A.; Fraoua, S.; Grass, V.; Moiroux, J.; Saveant, J.-M. J.
Am. Chem. Soc. 1998, 120, 2951 and references therein. (b) Schlesener, C.
J.; Amatore, C.; Kochi, J. K. J. Am. Chem. Soc. 1984, 106, 3567, 7472.
Masanovi, J. M.; Sankaraman, S.; Kochi, J. K. J. Am. Chem. Soc. 1989,
111, 2263.
and, consequently (see below), to reduced rates of their TBAA-
(24) Schanze and co-workers have reported an OD isotope effect of 2.0
for pyridine-promoted retro-aldol fragmentation of the anilinium radical
derived from the 2-(N-phenylamino)-1,2-diphenylethanols (ref 18b).
(28) Nicholas, A. P. M.; Arnold, D. R. Can. J. Chem. 1982, 60, 2166.
Nicholas, A. P. M.; Boyd, R. J.; Arnold, D. R. Can. J. Chem. 1982, 60,
3011.

10684 J. Am. Chem. Soc., Vol. 120, No. 41, 1998
Su et al.
Figure 7. Plot of the rate constant for decarboxylation of 1•+ in MeCN
at 25 °C containing 10 mM MClO4 (M ) Li, Na, K, Cs, and Rb) vs
the ionic radius of M+ (Keer, J. A. Chem. ReV. 1966, 66, 465. Janz, G.
J. Thermodynamic Properties of Organic Compounds; Academic
Press: New York, 1993).
Figure 6. Plot of the rate of unimolecular decarboylation of 1•+ in
MeCN at 25 °C vs the reciprocal of Ca(ClO4)2 concentration.
R-methyl substitution leads to decreased fragmentation rates
whereas R-phenyl substitution leads to increased rates. It
appears that the rationale put forth by Lewis25 for R-substituent
effects on the rates of aminium radical R-CH deprotonations
(i.e., that they are governed by a combination of steric/
Scheme 12
-
•+
steroelectronic (σC-E pN overlap) and electronic (R-amino
radical stabilization) holds for all related processes in which an
electrofugal group is lost or transferred from an R-position.
From the perspective of preparative photochemistry, the
substituent effects on the rates of aminium radical fragmentations
can be used advantageously in the design of more highly
quantum efficient processes operating via SET pathways.
Observations made in studies of the (aminoethyl)phthalimide
photocyclization reactions exemplify this feature. Specifically,
the results demonstrate that the quantum yields are qualitatively
related to the LFP measured rates of aminium radical fragmen-
tation. The preparative utility of the LFP rate data is further
exemplified by the observation that the inefficient photo-
cyclization of phthalimide 37 can be transformed into excep-
tionally efficient photoreaction by simply changing the elec-
trofugal group from H to CO2NBu4 and the N-substituent from
methyl to acetyl.
Medium and Metal Cation Effects on the Rates of
r-Decarboxylation of Anilinium Carboxylates. Owing to the
potential relevance of the anilinium radical decarboxylation
process to the design of fast unimolecular processes for R-amino
radical initiated polymerizations, we probed this reaction in
greater detail. The results provide useful information about how
medium and countercation influence the rates of these processes.
As seen by viewing the data in Table 6, the rates of decarbox-
ylation of the ammonium carboxylate 1•+ vary slightly, yet in
a patterned way (log kdec ET) with changes in the polarity of
the solvent. Also, the presence of metal perchlorate salts results
in a reduction of the 1•+ decarboxylation rates and this effect
displays well-behaved metal cation concentration (Figure 6),
ionic radius (Figure 7), and charge/radius dependencies.
These observations can be explained by use of the simple
mechanistic model shown in Scheme 12. The anilinium
carboxylate in solution can exist as two limiting structures,
represented by the free zwitterion 46 and ion paired form 47.
Which of these structural extremes best represents the major
anilinium carboxylate produced in the SET step should be
governed by solvent polarity and oxophilicity of the metal cation
(e.g., 46 should dominate in more highly polar solvents and
when M+ is a large monovalent cation). In addition, decar-
boxylation of the free zwitterion 46 should occur at a faster
rate than that of the ion-paired intermediate 47. Contributing
to this is the fact that loss of CO2 in 47 requires rupture of the
metal-oxygen bond and a larger reorganization of the ion
pairing and solvation interactions.
Conclusions
The issues associated with aminium radical R-heterolytic
fragmentation processes discussed above have been addressed
in this study by using a combination of photophysical and
photochemical methods. An SET photosensitization technique
in conjunction with time-resolved laser spectroscopy has been
employed to generate and kinetically analyze the decay proc-
esses of anilinium radicals derived by one-electron oxidation
of R-anilinocarboxylates, â-anilinoalcohols, and R-anilinosilanes.
In this manner, we determined that the rates of unimolecular
decarboxylation of anilinium carboxylate radicals are in the
range 106-107 s-1 and are dependent upon solvent polarity,
the metal cation, and substituents on the aniline ring, nitrogen,
and R-carbon. In addition, kinetic analysis of base-induced
retro-aldol fragmentations of cation radicals arising by SET
oxidation of â-anilinoalcohols has shown that bimolecular rate
constants vary from 104 to 105 M-1 s1 and are close to those
for R-deprotonation reactions of related N,N-dialkylanilinium
radicals. The retro-aldol fragmentation rates such as those for
R-decarboxylation reactions also vary in a patterned way with
changes in arene ring, nitrogen, and R- and â-carbon substit-
uents. In an extension of an earlier investigation of the dynamics
of methanol-promoted reactions of R-(trimethylsilyl)methyl
anilinium radicals, we have found that changes in the nitrogen
substituent have a profound effect on R-desilylation rates.
Parallel photochemical studies have been conducted both to
gain chemical evidence to support assignment of the anilinium
radical decay pathways and to demonstrate the preparative
consequences of the kinetic results. First, clean formation of
products derived by coupling of the (N-methylanilino)methyl

Dynamics of Anilinium Radical Fragmentation Processes
J. Am. Chem. Soc., Vol. 120, No. 41, 1998 10685
1
26: H NMR (CD3CN) 3.04 (s, 3H, NCH3), 4.60 (s, 2H, NCH2),
6.60-6.71 (m, 3H, aromatic), 7.11-7.20 (m, 2H, aromatic), 7.36 and
7.65 (ABq, J ) 8.3, 4H, aromatic); 13C NMR (CD3CN) 39.4 (NCH3),
56.7 (NCH2), 113.2, 117.5, 119.7, 128.5, 130.1, 133.3, 146.6 and 150.2
(aromatic), 180.1 (CN); IR 2919, 2360, 1599, 1506, 1347, 1251, 934,
749, 691; MS 223 (15), 222 (86), 221 (24), 120 (100), 116 (23), 106
(24), 105 (12), 91 (14), 77(35); HRMS calcd for C15H14N2 222.1157,
found 222.1159.
radical in photochemical reactions of 1,4-dicyanobenzene with
either tetra-n-butylammonium N-methyl-N-phenylglycinate or
â-N-methyl-N-phenylaminoethanol shows that the respective
decarboxylation and retro-aldol cleavage processes occur with
exceptionally high efficiencies. Second, in accord with the high
rates observed for aminium radical decarboxylation, tethered
cyclohexenone-R-aminocarboxylates undergo high-yielding
SET-promoted photocyclization reactions under both direct and
SET-sensitized conditions. Last, we have demonstrated how
the rates of aminium radical R-fragmentation correlate with
quantum efficiencies of SET-promoted reactions of tertiary
amines and amides. A study of photocyclization reactions of
N-(aminoethyl)- and (amidoethyl)phthalimides show that the
quantum yields for these SET-promoted processes vary with
the electrofugal group and nitrogen substituent in the manner
predicted on the basis of the LFP-determined R-fragmentation
rates.
Photoaddition of DCB to the Aminoalcohol 8. A deoxygenated
MeCN solution (100 mL) containing 8 (36 mg, 0.24 mmol), DCB (31
mg, 0.24 mmol), and tetrabutylammonium acetate (TBAA) (9.04 g,
30.0 mmol) was irradiated with Pyrex glass filtered light for 2 h. The
photolyzate was concentrated under reduced pressure to ca. 20 mL.
The mixture was diluted with water and extracted with ether. The
ethereal extracts were dried and concentrated in vacuo, and the residue
was subjected to column chromatography (silica gel, 20% ether/hexanes)
to yield 17 mg (32%) of 26 and 3 mg of an unknown with 15 mg
(41%) of 8 and 19 mg (63%) of recovered DCB.
Photochemistry of Aminoenone 30. A deoxygenated MeCN
solution (100 mL) containing 30 (150 mg, 0.33 mmol) was irradiated
with uranium glass filtered light for 4 h. The photolyzate was
concentrated to ca. 5 mL, diluted with water, and extracted with ether.
The ethereal extracts were dried and concentrated in vacuo. The residue
was subjected to preparative TLC separation (silica gel, 66% ether/
hexanes) to yield 34 mg (61%) of the spirocyclic amino ketone 32:
1H NMR 1.55-1.64 (m, 2H), 1.69-1.75 (m, 2H), 1.74-1.85 (m, 2H),
2.18-2.31 (m, 5H), 2.25 (s, 3H, NCH3), 2.34 (d, J ) 9.3, 1H), 2.46-
2.54 (m, 2H); 13C NMR 23.3 (C7), 36.9 (C4 or C6), 37.6 (C6 or C4),
41.0 (C8), 42.1 (NCH3), 46.9 (C5), 54.5 (C10), 55.7 (C3), 68.0 (C1),
210.6 (CdO); MS 167 (13), 130 (35), 109 (10), 96 (20), 70 (15), 58
(46), 57 (100); HRMS calcd for C10H17NO 167.1310, found 167.1311.
A MeOH solution (100 mL) containing 30 (150 mg, 0.33 mmol)
was irradiated with uranium glass filtered light for 5 h. Workup and
preparative TLC separation (silica gel, 66% ether/hexanes) furnished
32 mg (58%) of 32.
An MeCN solution (100 mL) containing 30 (150 mg, 0.33 mmol)
and DCA (8 mg, 0.033 mmol) was irradiated with uranium glass filtered
light for 4 h. Workup followed by preparative TLC separation (silica
gel, 66% ether/hexanes) provided 33 mg (59%) of 32.
A MeOH solution (20 mL) containing 30 (25 mg, 0.055 mmol) and
DCA (3 mg, 0.011 mmol) was irradiated with uranium glass filtered
light for 5 h. Workup followed by preparative TLC separation (silica
gel, 66% ether/hexanes) gave 6 mg (63%) of 32.
Photochemistry of Aminoenone 31. An MeCN solution (100 mL)
containing 31 (100 mg, 0.51 mmol) and TBAA (3.0 g, 10 mmol) was
irradiated with uranium glass filtered light for 7 h. The photolyzate
was concentrated under reduced pressure to give a residue which was
subjected to column chromatography (silica, ether). The ethereal eluent
was concentrated under reduced pressure to give a residue which was
subjected to preparative TLC separation (silica gel, 75% ethyl acetate/
hexanes) to yield 40 mg (47%) of 32.
A MeOH solution (20 mL) containing 31 (20 mg, 0.10 mmol) and
TBAA (1.5 g, 5 mmol) was irradiated with uranium glass filtered light
for 10 h. Workup and preparative TLC separation (silica gel, 66%
ether/hexanes) furnished 10 mg (62%) of 32.
An MeCN solution (100 mL) containing 31 (100 mg, 0.51 mmol),
TBAA (3.0 g, 10 mmol), and DCA (12 mg, 0.052 mmol) was irradiated
with uranium glass filtered light for 8 h. Workup followed by
preparative TLC separation (silica gel, 66% ether/hexanes) provided
54 mg (64%) of 32.
A MeOH solution (20 mL) containing 31 (20 mg, 0.10 mmol),
TBAA (1.5 g, 5 mmol), and DCA (2.5 mg, 0.011 mmol) was irradiated
with uranium glass filtered light for 9 h. Workup followed by
preparative TLC separation (silica gel, 66% ether/hexanes) gave 10
mg (58%) of 32.
Experimental Section
General Procedure. All reported reactions were run under a dried
nitrogen atmosphere. Unless otherwise noted, all reagents were
obtained from commercial sources and used without further purification.
Dimethyl sulfoxide and triethylamine were distilled from calcium
hydride. Anhydrous solvents were obtained by distillation from the
indicated reagents: ether (Na, benzophenone ketyl), tetrahydrofuran
(Na, benzophenone ketyl), methylene chloride (P2O5). Column chro-
matography was performed with silica gel (230-400 mesh), Florisil
(100-200 mesh) adsorbents, or alumina (neutral, 80-120 mesh).
Preparative TLC was performed on 20 × 20 cm plates coated with
silical gel. All compounds were isolated as oils and shown to be >90%
1
pure by H and/or 13C NMR unless otherwise noted.
1H NMR and 13C NMR spectra were recorded by using CDCl3
solutions unless otherwise specified, and chemical shifts are reported
1
in parts per million relative to residual CHCl3 at 7.24 ppm (for H
NMR) and 77.0 ppm (for 13C NMR). 13C NMR resonance assignments
were aided by the use of the DEPT technique to determine numbers of
1
attached hydrogens, and H NMR coupling constants (J values) are
reported in hertz. Infrared spectra were obtained neat liquids unless
otherwise specified, and data are reported in units of cm-1. Low (MS)
and high (HRMS) mass spectra, reported as m/z (relative intensity),
were recorded by using electron impact ionization (EI) unless otherwise
specified as chemical ionization (CI) or fast atom bombardment
ionization (FAB).
Photochemical reactions were conducted by using an apparatus
consisting of a 450-W medium-pressure mercury lamp surrounded by
a Pyrex or uranium glass filter and within a quartz, water-cooled well
that was purged with O2-free N2 both before and during irradiation.
Photochemical reaction progress was monitored by gas chromatography,
TLC, or 1H NMR. The solvents used in the photoreactions were
spectrograde: MeCN (Baker) or MeOH (Baker). 1,4-Dicyanobenzene
(DCB) was purchased from Aldrich and recrystallized (CHCl3) prior
to use. 9,10-Dicyanoanthracene (DCA) was purchased from Eastman
Kodak and recrystallized from benzene.
Photoaddition of DCB to the Anilinocarboxylate 1. A deoxy-
genated MeCN solution (20 mL) containing 1 (20 mg, 0.049 mmol)
and DCB (6 mg, 0.049 mmol) was irradiated with Pyrex glass filtered
light for 20 min. The photolyzate was diluted with water and extracted
with ether. The ethereal extracts were dried and concentrated in vacuo,
and the residue was subjected to column chromatography (silica gel,
20% ether/hexanes) to yield 2 mg (30%) of N,N′-dimethyl-N,N′-
diphenylethylene-1,2-diamine (25) and 3 mg (27%) of N-(4-cyanoben-
zyl)-N-methylaniline (26) with 4 mg (55%) of recovered DCB.
1
25: H NMR (CD3CN) 2.89 (s, 6H, 2NCH3), 3.51 (s, 4H, 2NCH2),
DCA-Sensitized Irradiation of 33. An MeCN solution (100 mL)
containing 33 (190 mg, 0.37 mmol) and DCA (21 mg, 0.09 mmol)
was irradiated with uranium glass filtered light for 10 h. The
photolyzate was concentrated under reduced pressure to give a residue.
This residue was dissolved in 10 mL of methanol and filtered. The
filtrate was diluted with water and extracted with ether. The ethereal
6.60-6.71 (m, 6H, aromatic), 7.12-7.21 (m, 4H, aromatic); 13C NMR
(CD3CN) 38.8 (NCH3), 50.1 (NCH2), 112.8, 116.8, 118.9 and 130.1
(aromatic); IR 2957, 2925, 2863, 1733, 1599, 1506, 1374, 747, 692;
MS 241 (3), 240 (15), 149 (22), 120 (100), 105 (12), 91 (5), 77 (10);
HRMS calcd for C16H20N2 240.1626, found 240.1626.

10686 J. Am. Chem. Soc., Vol. 120, No. 41, 1998
Su et al.
extracts were dried and concentrated in vacuo. The residue was
subjected to preparative TLC separation (silica gel, 50% hexanes/
acetone) to yield 59 mg (71%) of 34. 1H NMR 1.23 (t, J ) 7.0, 3H,
OCH2CH3), 1.70-1.80 (m, 4H), 1.84-1.89 (m, 2H), 2.30-2.35 (m,
4H), 3.15, 3.22 (s, 2H, H1), 3.39-3.44 (m, 2H), 4.09 (q, J ) 7.0, 2H,
OCH2CH3); 13C NMR 14.8 (OCH2CH3); 23.1, 23.2 (C7), 33.9 (C6 or
C4), 35.8, 35.9 (C4 or C6), 40.1 (C8), 44.1, 44.4 (C10), 45.5, 46.5 (C5),
50.3, 50.4 (C3), 55.8, 56.6 (C1), 61.1 (OCH2CH3), 155.1 (NCO2Et),
209.9 (OdC9); IR 2940, 2872, 1700, 1424, 1381, 1347, 1227, 1115,
770; MS 226 (6), 225 (31), 180 (13), 168 (14), 167 (100), 152 (14),
116 (27), 115 (39), 93 (25), 82 (21), 67 (29); HRMS calcd for C12H19-
NO3 225.1365, found 225.1368.
min prior to the experiment. Sample concentrations were adjusted such
that their optical densities were 1.0-1.5 at the excitation wavelength
(308 nm). The transient absorption spectra were detected in the 300-
800-nm range.
Three mathematical expressions corresponding to the kinetic models
for anilinium radical decay by BET to BDB•- (eq 6), competitive
second-order back BET-first- or pseudo-first-order decarboxylation,
TBAA-induced deprotonation and TBAA-induced retro-aldol cleavage
and MeOH-induced desilylation (eq 7), and corrected (for slow BET
to DCB•-, eq 8) first-order decarboxylation were establishd in Mac II
MATLAB 1.2c and MATLAB 4.2c.1 to analyze the decay processes:
A MeOH solution (100 mL) containing 33 (190 mg, 0.37 mmol)
and DCA (21 mg, 0.09 mmol) was irradiated with uranium glass filtered
light for 5 h. Workup and preparative TLC separation (silica gel, 50%
hexanes/acetone) gave 46 mg (55%) of 34.
DCA-Sensitized Irradiation of 35. An MeCN solution (100 mL)
containing 35 (150 mg, 0.29 mmol) and DCA (17 mg, 0.074 mmol)
was irradiated with uranium glass filtered light for 7 h. The photolyzate
was concentrated under reduced pressure to give a residue. This residue
was dissolved in 5 mL of methanol and filtered. The filtrate was diluted
with water and extracted with ether. The ethereal extracts were dried
and concentrated in vacuo. The residue was subjected to preparative
TLC separation (silica gel, 30% hexanes/ether) to yield 19 mg (41%
based on the conversion) of cis-36 and 24 mg (53% base on the
conversion) of trans-36 with 47 mg (31%) of recovered 35.
Equation 6 was employed to fit the profiles monitored at 340 nm
and/or 460 nm for BET-promoted second-order decay of the ion radicals
derived from aniline derivatives 8-10 and 19-24 and DCB in the
absence of base. This provides the second-order rate constants for both
DCB anion radical decay and anilinium radical decay by BET.
1
Equation 7 was employed to fit the profiles monitored at 460 nm
for respective TBAA-induced retroaldol cleavage and deprotonation
of the aminium radicals derived from 8-11 and 19-24. The second-
order rate constant k2nd-order corresponding to BET was set to be equal
to that obtained for the decay process occurring in the absence of base
by using eq 6. For the respective TBAA- and MeOH-promoted decay
of the aminium radicals from 12-15 and 16-18, k2nd-order corresponding
to back electron transfer between the anilinium radicals and the anion
radical of DCB was set to be equal to the rate constant for diffusion in
MeCN. The pseudo-first-order rate constants, kpseudo-first-order, corre-
sponding to acetate-promoted deprotonation and retro-aldol cleavage,
or MeOH-promoted desilylation were measured at different TBAA and
MeOH concentrations and plots of kpseudo-first-order vs TBAA or MeOH
concentrations gave the respective second-order rate constants for
anilinium radical decay by deprotonation, retro-aldol cleavage, or
desilylation.
cis-36: H NMR 1.18 (t, J ) 7.0, 3H, OCH2CH3), 1.45-2.40 (m,
11H), 3.00-3.20 (m, 2H), 3.54-3.70 (m, 1H), 4.09 (q, J ) 7.0, 2H,
OCH2CH3); 13C NMR 14.7 (OCH2CH3); 25.6, 29.4 (C4, C5), 33.0 (C4a,
C8a), 37.4 (C6), 41.7 (C8), 42.2 (C3), 47.4 (C1), 61.4 (OCH2CH3), 155.9
(NCO2Et), 211.2 (OdC7); IR 2980, 2934, 2865, 1697, 1435, 1239,
1122, 769; MS 225 (7), 196 (5), 180 (5), 168 (13), 167 (100), 154
(16), 152 (14), 116 (14); HRMS calcd for C12H19NO3 225.1365, found
225.1361.
trans-36: 1H NMR 1.18 (t, J ) 7.1, 3H, OCH2CH3), 1.45-2.50 (m,
11H), 3.00-3.20 (m, 2H), 3.58-3.60 (m, 1H), 4.11 (q, J ) 7.1, 2H,
OCH2CH3); 13C NMR 14.6 (OCH2CH3); 25.4, 29.7 (C4, C5), 33.0 (C4a,
C8a), 37.4 (C6), 41.6 (C8), 42.8 (C3), 47.4 (C1), 61.4 (OCH2CH3), 155.9
(NCO2Et), 211.2 (OdC7); IR 2939, 2855, 1697, 1436, 1239, 1123, 769;
MS 226 (6), 225 (12), 168 (14), 167 (100), 154 (18), 152 (14), 122
(10), 116 (44), 95 (12); HRMS calcd for C12H19NO3 225.1365, found
225.1356.
A MeOH solution (100 mL) containing 35 (150 mg, 0.29 mmol)
and DCA (17 mg, 0.074 mmol) was irradiated with uranium glass
filtered light for 6 h. Workup and preparative TLC separation (silica
gel, 30% hexanes/ether) gave 14 mg (26% based on the conversion)
of cis-36, 19 mg (35% base on the conversion) of trans-36 with 27 mg
(18%) of recovered 35.
Equation 7 was also utilized to fit the profiles for decay of the
anilinocarboxylate-derived anilinium radicals 1-7 monitored at 460
nm. The second-order rate constant, k2nd-order, for BET to the anion
radical of DCB was set to be equal to the rate constant for diffusion in
MeCN. The first-order rate constants corresponding to decarboxylation
were then calculated. Equation 8 treats the data as a sum of a fast
first-order decay by decarboxylation corrected for by a term related to
a much slower BET process. Although less rigorously derived, eq 8
was also employed to analyze the decay profiles for the anilinium
radicals. Two rate constants, k1st-order and k′correct, were obtained by
use of eq 8, and the one corresponding to the faster (by ca. 1 order of
magnitude) decay process was ascribed to decarboxylation. The first-
order rate constant for decarboxylation calculated by use of either eq
7 or eq 8, as expected, were equivalent within experimental error.
Quantum Yields for Photoreactions of Phthalimides 37-42.
Quantum yields were measured by using 300-nm light and a 0.10 M
benzophenone-benzhydrol actinomer. Individual methanol solutions
(10 mL) containing phthalimides 37 (31 mg, 14.0 mM) and 38 (17
mg, 6.99 mM), 39 (98 mg, 33 mM) and 40 (107 mg, 34 mM), and 41
(92 mg, 35 mM) and 42 (200 mg, 69 mM) all in sealed tubes were
simultaneously irradiated with 300-nm light to affect ca. 5-20%
conversions. The crude photolyzates were analyzed by UV spectros-
copy at 300 nm to determine phthalimide conversions. Quantum yields
listed in Table 9 are for photoproduct formation, calculated on the basis
of starting phthalimide coversion and photocyclization product yield.
Laser Flash Spectroscopy Experiments. The laser kinetic experi-
ments were performed by using a Questek 2120 excimer laser as the
excitation source. All of the spectra and kinetic runs obtained were
done using XeCl reagent gas which provides UV pulses at 308 nm
with a duration of 6-10 ns and a pulse energy of 30-50 mJ. Transient
UV/vis absorption signals were monitored using a CW 450-W Xe Arc
Lamp beam which was passed through the sample perpendicular to
the excitation beam. Single-wavelength transient wave forms were
digitized using a LeCroy 9420 350-M digital oscilloscope and
transferred to a Macintosh IIcx computer for storage and analysis.
Sample solutions were placed in 10 × 10 mm quartz cuvettes which
were sealed with a rubber serum cap and purged with N2 for 10-15
Acknowledgment. The authors appreciate the financial
support generously provided by the NSF (P.S.M., CHE-9707401
and INT-9796064; D.E.F., CHE-9521601), KOSEF (U.C.Y.,
965-0300-002-2), CBM Postech, and the Korea Ministry for
Education (U.C.Y., BSRI-97-3408).
Supporting Information Available: Experimental descrip-
tions of the synthetic routes used to prepare the substances used
in this study are described (41 pages, print/PDF). See any
current masthead page for ordering information and Web access
instructions.
JA981541F