Homolytic vs Heterolytic Paths in the Photochemistry of Haloanilines

Article abstract of DOI:10.1021/ja036000r

The photochemistry of 4-haloanilines and 4-halo-N,N-dimethylanilines has been studied in apolar, polar aprotic, and protic solvents. Photophysical and flash photolysis experiments show that the reaction proceeds in any case from the triplet state. It is rather unreactive in apolar media, the highest value being Π = 0.05 for the iodoanilines in cyclohexane. Changing the solvent has little effect for iodoanilines and for the poorly reacting bromo analogue, while it leads to a variation of over 2 orders of magnitude in the quantum yield for the chloro and fluoro derivatives. The triplets have been characterized at the UB3LYP/6-31G(d) level of theory, evidencing a deformation and an elongation (except for C-F) of the C-X bond. Homolytic fragmentation is in every case endothermic, but calculations in acetonitrile solution show that heterolytic cleavage of C-Cl and C-Br is exothermic. Experimentally, the occurrence of heterolytic fragmentation has been monitored through selective trapping of the resulting phenyl cation by allyltrimethylsilane. Heterolytic dechlorination occurs efficiently in polar media (e.g., Π = 0.77 in MeCN), while debromination remains ineffective due to the short lifetime of the triplet. Heterolytic defluorination is efficient only in protic solvents (Π = 0.48 in MeOH), in accord with calculations showing that in the presence of an ancillary molecule of water fragmentation is exothermic due to the formation of the strong H-F bond. The energy profile for both homo- and heterolytic dissociation paths has been mapped along the reaction coordinates in the gas phase and in acetonitrile. The conditions determining the efficiency and mode of dehalogenation have been defined. This is significant for devising synthetic methods via photogenerated phenyl cations and for rationalizing the photodegradation of halogenated aromatic pollutants and the phototoxic effect of some fluorinated drugs.

Full text of DOI:10.1021/ja036000r

Published on Web 10/03/2003
Homolytic vs Heterolytic Paths in the Photochemistry of
Haloanilines
Mauro Freccero, Maurizio Fagnoni, and Angelo Albini*
Contribution from the Department of Organic Chemistry, UniVersity of PaVia,
Via Taramelli 10, 27100 PaVia, Italy
Received May 7, 2003; E-mail: angelo.albini@unipv.it
Abstract: The photochemistry of 4-haloanilines and 4-halo-N,N-dimethylanilines has been studied in apolar,
polar aprotic, and protic solvents. Photophysical and flash photolysis experiments show that the reaction
proceeds in any case from the triplet state. It is rather unreactive in apolar media, the highest value being
Φ ) 0.05 for the iodoanilines in cyclohexane. Changing the solvent has little effect for iodoanilines and for
the poorly reacting bromo analogue, while it leads to a variation of over 2 orders of magnitude in the quantum
yield for the chloro and fluoro derivatives. The triplets have been characterized at the UB3LYP/6-31G(d)
level of theory, evidencing a deformation and an elongation (except for C-F) of the C-X bond. Homolytic
fragmentation is in every case endothermic, but calculations in acetonitrile solution show that heterolytic
cleavage of C-Cl and C-Br is exothermic. Experimentally, the occurrence of heterolytic fragmentation
has been monitored through selective trapping of the resulting phenyl cation by allyltrimethylsilane. Heterolytic
dechlorination occurs efficiently in polar media (e.g., Φ ) 0.77 in MeCN), while debromination remains
ineffective due to the short lifetime of the triplet. Heterolytic defluorination is efficient only in protic solvents
(Φ ) 0.48 in MeOH), in accord with calculations showing that in the presence of an ancillary molecule of
water fragmentation is exothermic due to the formation of the strong H-F bond. The energy profile for
both homo- and heterolytic dissociation paths has been mapped along the reaction coordinates in the gas
phase and in acetonitrile. The conditions determining the efficiency and mode of dehalogenation have
been defined. This is significant for devising synthetic methods via photogenerated phenyl cations and for
rationalizing the photodegradation of halogenated aromatic pollutants and the phototoxic effect of some
fluorinated drugs.
Scheme 1
Introduction
The photochemistry of aryl halides has been the subject of
extensive investigation and has revealed many mechanistically
interesting facets,1 besides the environmental application for the
degradation of persistent pollutants of this class. Different paths
have been envisaged for the observed photoreactions and are
indicated in Scheme 1.2
Homolysis (path a) has some synthetic potential, for the
preparation of biaryls, in particular in the intramolecular version
for the construction of polycyclic systems.2b,3 The viability of
this path appears to be strictly determined by thermochemistry.
C-X bond cleavage in the triplet state4 is exothermic for iodo-
and bromobenzenes and is probably moderately endothermic
for chlorobenzenes, though there has been some controversy
about the correct value of the Ph-Cl bond dissociation energy.5
On the contrary, the largely endothermic homolysis of the aryl-
fluorine bond does not take place. The decrease of the triplet
energy in naphthalenes and in higher arenes makes cleavage
unlikely. A variation of this path is cage electron transfer within
the primarily formed radical pair leading to the ion pair in an
indirect way (path a′).6
(1) (a) Bunce, N. J. In Handbook of Photochemistry and Photobiology;
Horspool, W. M., Song, P. S., Eds.; CRC Press: Boca Raton, FL, 1995; p
1181. (b) Grimshaw, J.; de Silva, A. P. Chem. Soc. ReV. 1981, 10, 181. (c)
Lodder, G.; Cornelisse, J. In The Chemistry of Halides, Pseudo-halides
and Azides; Patai, S., Rappoport, Z., Eds.; Wiley: New York, 1995; Suppl.
D2, p 861. (d) Davidson, R. S.; Goodwin, J. W.; Kemp, G. AdV. Phys.
Org. Chem. 1984, 20, 191. (e) Mangion, D.; Arnold, D. R. Can. J. Chem.
1999, 77, 1655 and references therein. (f) Klan, P.; Ansorgova, A.; Del
Favero, D.; Holoubek, I. Tetrahedron Lett. 2000, 41, 7785. (g) For
analoguous reactions from heteroaryl halides, see: D’Auria, M.; De Luca,
B.; Mauriello, G.; Racioppi, R. J. Chem. Soc., Perkin Trans. 1 1998, 271
and references therein.
In other cases, the reaction proceeds from a (hetero)excimer
(path b)4b,5,7 or from an exciplex (e.g., with an aliphatic amine8
(4) (a) The singlet state may also be involved, at least with heavily halogenated
compounds; see, e.g., ref 4b. (b) Freeman, P. K.; Ramnath, N.; Richardson,
A. D. J. Org. Chem. 1991, 56, 3643.
(5) Bunce, N. J.; Bergsma, M. D.; De Graaf, W.; Kumar, Y.; Ravanal, L. J.
Org. Chem. 1980, 45, 3708.
(6) Orvis, J.; Weiss, J.; Pagni, R. M. J. Org. Chem. 1991, 56, 1851.
(2) (a) Dulin, D.; Drossman, H.; Mill, T. EnViron. Sci. Technol. 1986, 20, 72.
(b) Choudhry, G. G.; Webster, G. R. B.; Hutzinger, O. Toxicol. EnViron.
Chem. 1986, 13, 27. (c) Bunce, N. J.; Landers, J. P.; Langshaw, J. A.;
Nakai, J. S. EnViron. Sci. Technol. 1989, 23, 213.
(3) Kessar, S. V.; Mankotia, A. K. S. In ref 1a, p 1218.
9
13182
J. AM. CHEM. SOC. 2003, 125, 13182-13190
10.1021/ja036000r CCC: $25.00 © 2003 American Chemical Society

Photochemistry of Haloanilines
A R T I C L E S
or with an aniline,9 path c) rather than directly from the excited
state.
Scheme 2
Interaction of an excited aryl halide with a nucleophile leads
either to substitution via the SN2Ar* mechanism (path d) or to
electron transfer (path e).10 In the latter case, the resulting radical
anion may fragment (path e′) to give the aryl radical, which in
turn may initiate a chain process (SRN1 mechanism).11
In contrast, unimolecular heterolytic cleavage (path f) has
been rarely invoked in the photochemistry of aromatics.
Actually, until recently heterolytic dehalogenation leading to
nucleophilic photosubstitution via an SN1Ar* mechanism had
been proposed only in a couple of cases.12,13 In the last few
years, however, considerable evidence about the photoheterolysis
of halogenated phenols,14,15 anisoles,16 and anilines17,18 in water
as well as of 2-propylchlorobenzene in trifluoroethanol19 has
appeared. Our group found that with chloroanilines the reaction
is efficient in polar organic solvents and affords a smooth entry
to aryl cations.20 This adds a synthetic perspective, because these
otherwise hardly accessible intermediates21 can be exploited in
a variety of arylation reactions, in particular of alkenes and
arenes, which offer a convenient access to functionalized
anilines.22
Furthermore, it is not known to what extent homolytic cleavage
competes in different haloanilines and in different media. These
facts motivated a comprehensive examination of the photo-
chemistry of haloanilines, based on the determination of product
distribution and photophysical experiments in various solvents,
as well as on computational studies both in the gas phase and
in acetonitrile solution in the frame of DFT theory and the
C-PCM (conductor version of PCM) solvation model.
The available evidence, mainly obtained with 4-chloroaniline,
suggests that heterolytic cleavage proceeds efficiently from the
triplet state and forms the triplet phenyl cation. However, under
suitable conditions also 4-fluoroaniline cleaves,20a,22d which is
intriguing in view of the high energy of the aryl-fluorine bond.
(7) Freeman, P. K.; Jang, J. S.; Ramnath, N. J. Org. Chem. 1991, 56, 6072.
(b) Freeman, P. K.; Jang, J.; Haugen, C. M. Tetrahedron 1996, 52, 8397.
(8) (a) Bunce, N. J. J. Org. Chem. 1982, 47, 1948. (b) Beecroft, R. A.;
Davidson, R. S.; Goodwin, D. C. Tetrahedron Lett. 1983, 24, 5673. (c)
Tanaka, Y.; Uryu, T.; Ohashi, M.; Tsujimoto, K. J. Chem. Soc., Chem.
Commun. 1987, 1703.
Results
Photochemistry. The halides examined were the fluoro-,
chloro-, bromo-, and iodoanilines (1-4), as both N,N-dimethyl
derivatives (1a-4a) and nonmethylated derivatives (1b-4b).
It was previously found that chloroanilines 2a,b are poorly
photoreactive in apolar solvents while their irradiation in a polar
solvent such as acetonitrile gives diphenyldiamines 6 ac-
companied by some anilines 7. These products were rationalized
as arising from cation 5 via electrophilic attack of the starting
compound or, respectively, reduction. A product of structure 6
was formed also from fluoroaniline 1a. However, in the presence
of alkenes the reaction was diverted to arylation of the latter
substrates. This reaction was particularly efficient when using
allyltrimethylsilane, in which case allylanilines 8 became by
far the major product (81% 8a and 4% 7a from 2a in MeCN).22a
On the other hand, anilines 7 may result from homolytic
cleavage via radical 9. Therefore, we decided to use the reaction
with the allylsilane for the specific “titration” of the aryl cation
and to explore the medium dependence of the reaction.
The irradiation of anilines 1-4 was carried out in a series of
solvents of differing polarity and proticity (cyclohexane, dichlo-
romethane, acetonitrile, methanol, and trifluoroethanol), and the
experiments were repeated in the same solvents in the presence
of 1 M allyltrimethylsilane. The product distribution and
quantum yield of the reaction were determined (see Scheme 2
and Tables 1 and 2).
(9) (a) Pac, C.; Tosa, T.; Sakurai, H. Bull. Chem. Soc. Jpn. 1972, 45, 1169.
(b) Grodowski, M.; Latowski, T. Tetrahedron 1974, 30, 767. (c) Bunce,
N. J.; Gallacher, J. C. J. Org. Chem. 1982, 47, 1955.
(10) (a) Cornelisse, J. In ref 1a, p 250. (b) Cornelisse, J.; Lodder, G.; Havinga,
E. ReV. Chem. Intermed. 1979, 2, 31.
(11) (a) Bunnett, J. F.; Sundeberg, J. E. Acc. Chem. Res. 1978, 11, 413 (b)
Rossi, R. A.; De Rossi, R. H. Aromatic Substitution by the SRN1
mechanism; American Chemical Society: Washington, DC, 1983.
(12) Groen, M. B.; Havinga, E. Mol. Photochem. 1974, 6, 9.
(13) Yang, N. C.; Huang, A.; Yang, D. H. J. Am. Chem. Soc. 1989, 111, 8060.
(14) (a) Grabner, G.; Richard, C.; Koehler, G. J. Am. Chem. Soc. 1994, 116,
11470. (b) Lipczynska-Kochany, E.; Bolton, J. R. J. Photochem. Photobiol.
1991, 58, 315. (c) Oudjehani, K.; Boule, P. J. Photochem. Photobiol. 1992,
68, 383. (d) Ouardaoui, A.; Steren, C. A.; van Willigen, H.; Yang, C. J.
Am. Chem. Soc. 1995, 117, 6803. (e) Durand, A. P. Y.; Brown, R. G.;
Worrall, D.; Wilkinson, F. J. Chem. Soc., Perkin Trans. 2 1998, 365. (f)
Bonnichon, F.; Grabner, G.; Guyot, G.; Richard, C. J. Chem. Soc., Perkin
Trans. 2 1999, 1203.
(15) (a) Guyon, C.; Boule, P.; Lemaire, J. Tetrahedron Lett. 1982, 23, 1581.
(b) Bonnichon, F.; Richard, C.; Grabner, G. Chem. Commun. 2001, 73.
(16) Zhang, G.; Wan, P. J. Chem. Soc., Chem. Commun. 1994, 19.
(17) (a) Othmen, K.; Boule, P.; Szczepanik, B., Rotkiewicz, K.; Grabner, G. J.
Phys. Chem. 2000, 104, 9525. (b) Szczepanik, B.; Latowski, T. Pol. J.
Chem. 1997, 71, 807.
(18) (a) Othmen, K.; Boule, P. Bull. EnViron. Contam. Toxicol. 1997, 59, 924.
(b) David, B.; Lhote, M.; Faure, V.; Boule, B. Water Res. 1998, 32, 2451.
(c) Othmen, K.; Boule, P.; Richard, P. New J. Chem. 1999, 23, 857. (d)
Othmen, K.; Boule, P. J. Photochem. Photobiol., A 2000, 136, 79.
(19) Hori, K.; Sonoda, T.; Harada, M.; Yamazaki-Nishida, S. Tetrahedron 2000,
56, 1429.
(20) (a) Fagnoni, M.; Mella, M.; Albini, A. Org. Lett. 1999, 1, 1299. (b)
Guizzardi, B.; Mella, M.; Fagnoni, M.; Freccero, M.; Albini, A. J. Org.
Chem. 2001, 66, 6353.
(21) (a) Stang, P. J. In Dicordinated Carbocations; Rappoport, Z., Stang, P. J.,
Eds.; Wiley: New York, 1997; p 451. (b) Hanack, M.; Subramanlan, L.
R. In Methoden der Organischen Chemie; Hanack, M., Ed.; Thieme:
Stuttgart, Germany, 1990; Vol. E19C, p 249. (c) Ambroz, H. B.; Kemp, T.
J. Chem. Soc. ReV. 1979, 8, 353. (d) Steenken, S.; Askokkuna, M.;
Maruthamuthu, P.; McClelland, R. A. J. Am. Chem. Soc. 1998, 120, 11925.
(22) (a) Mella, M.; Coppo, P.; Guizzardi, B.; Fagnoni, M.; Freccero, M.; Albini,
A. J. Org. Chem. 2001, 66, 6344. (b) Guizzardi, B.; Mella, M.; Fagnoni,
M.; Albini, A. Tetrahedron 2000, 56, 9383. (c) Coppo, P.; Fagnoni, M.;
Albini, A. Tetrahedron Lett. 2001, 42, 4271. (d) Guizzardi, B.; Mella, M.;
Fagnoni, M.; Albini, A. J. Org. Chem. 2003, 68, 1067.
Fluoroaniline 1a reacted quite sluggishly in cyclohexane or
dichloromethane and only slightly more efficiently in acetoni-
trile; in the last case diphenyldiamine 6a and amine 7a were
the main products. The reaction was much faster in alcohols,
9
J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003 13183

A R T I C L E S
Freccero et al.
Table 1. Photochemistry of Some Haloanilines in Various
Solvents
Table 3. Selected Photophysical Parameters for Haloanilines 1-4
in Various Solvents
yield of
6, %
yield of
7, %
τF,
ns
τT,
µs
substrate
X
R
solvent
C6H12
CH2Cl2
MeCN
Φ-1 or 2
substrate
X
R
solvent
ΦF
1a
F
Me
<0.005
0.01
0.05
80
65
32
82
23
80
65
38
87
40
1a
F
Me
C6H12
MeCN
MeOH
C6H12
MeCN
MeOH
C6H12
MeCN
MeOH
C6H12
MeCN
MeOH
0.14
0.23
0.20
0.16
0.19
0.19
0.024
0.026
0.026
0.012
0.019
0.019
2.8
5
3.9
2.1
2.7
2.8
0.7 0.3
0.8
0.7
1
65
15
55
MeOH
CF3CH2OH
C6H12
CH2Cl2
MeCN
MeOH
0.55
1b
2a
2b
F
H
0.30
2a
Cl
Me
0.003a
0.2
15
58
8
Cl Me
0.87a
0.95a
0.50a
CF3CH2OH
46
Cl
H
0.6
0.5
0.5
a From ref 20b.
3a,b
4a,b
Br Me, H C6H12, MeCN, MeOH <0.001 <0.1
I Me, H C6H12, MeCN, MeOH <0.001 <0.1
Table 2. Photochemistry of Some Haloanilines in Various
Solvents in the Presence of Allyltrimethylsilane (1 M)
a, R ) Me
Φ7
b, R ) H
products. Only in trifluoroethanol the proportion slightly shifted
in favor of allylated 8 (see Table 2).
solvent
C6H12
CH2Cl2
MeCN
MeOH
CF3CH2OH
C6H12
CH2Cl2
MeCN
MeOH
CF3CH2OH
C6H12
CH2Cl2
MeCN
MeOH
CF3CH2OH
C6H12
CH2Cl2
MeCN
MeOH
Φ8
Φ8
Φ7
1, X ) F
2, X ) Cl
3, X ) Br
4, X ) I
0.003
0.01
<0.005
0.06
Emission Properties. The emission of the above haloanilines
was examined in fluid solution in apolar, polar, and protic media.
The fluoro derivatives 1 exhibited a moderate fluorescence, with
ca. 20% efficiency and a lifetime of some nanoseconds (see
Table 3), with a limited effect of the nature of the solvent. With
the chloroanilines 2 both ΦF and τF decreased by up to 1 order
of magnitude with respect to those of the fluorinated analogues.23
Fluoro- and chloroanilines also exhibited phosphorescence
in a glassy matrix at 77 K, both polar (EPA) and apolar (3-
methylpentane), with λmax values around 425-435 nm.24 The
triplet energies (ET) evaluated from the edge of the phospho-
rescence spectrum are reported in Table 5.
0.04
0.48
0.28
<0.01
0.24
0.22
0.02
0.02
0.03
0.04
0.18
0.02
0.01
0.04
0.10
0.01
0.001
0.001
0.001
0.01
0.16
0.77
0.68
0.55
0.001
0.002
0.003
0.02
0.05
0.01
0.02
0.02
0.03
0.03
0.08
0.39
0.36
0.22
0.001
0.003
0.001
0.004
<0.001
0.04
0.04
0.03
0.04
0.02
0.001
<0.001
0.001
0.01
0.03
0.01
<0.001
0.03
0.04
0.03
0.03
On the other hand, with both bromo- and iodoanilines 3 and
4 both fluorescence at room temperature and phosphorescence
in a glassy matrix were very weak.
0.01
0.02
0.02
0.025
CF3CH2OH
0.02
Transients. Laser excitation of a cyclohexane solution of
fluoroanilines 1a,b produced a transient with a concentration-
dependent lifetime (1 µs or below in a 10-4 M solution), which
was quenched in oxygen-saturated solutions. All of these
characteristics closely matched those of the parent anilines
7a,b,25 and the transient was confidently assigned to the triplet
state (Figure 1, Table 3). In acetonitrile the shape and lifetime
of the transient changed little, although in the case of 1a a
fraction (<20%) of the signal around 460 nm persisted for a
longer time.
with reduction to amine 7a predominating in methanol and
formation of diamine 6a predominating in trifluoroethanol
(Table 1). In acetonitrile as well as in the alcohols allylaniline
8a became the major product in the presence of allyltrimeth-
ylsilane, though the overall quantum yield of the starting material
decomposition remained the same as in the neat solvents (Table
2). Fluoroaniline 1b followed a parallel pattern, with a slightly
lower reaction quantum yield than 1a under the same conditions.
In alcohols, on the other hand, the above transient was
completely superseded by a different, intense absorption. With
1 × 10-4 M 1a in MeOH, a persistent (>10 µs) absorption
with maxima at 430, 450, and 480 nm was observed (Figure
2a). The transient was similar in i-PrOH. An additional feature
in TFE was a further absorption with maxima at 400 and 420
nm (Figure 2b).
Chloroanilines 2a,b were again almost photostable in cyclo-
hexane, but significantly photoreactive in dichloromethane, and
reached the maximal value of photoreaction in acetonitrile, with
diamines 6 as the main products, and in methanol (in the latter
case, reduction to 7 predominated, Table 1). Also in this case,
the overall quantum yield of photodecomposition changed little
in the presence of the silane, where the main products became
allylanilines 8 (Table 2).
The results were quite similar to those of the other fluorinated
aniline 1b with the difference that the transients were blue-
With bromoanilines 3a,b the quantum yield remained low
under all of the conditions examined, with a moderate increase
from some parts per thousand in cyclohexane to some percent
in the alcohols and further some increase in the yield of
allylanilines 8 (see Table 2).
(23) The present results compare well with literature data obtained under some
of the above conditions; see: Koutek, B.; Musil, L.; Soucek, M. Collect.
Czech. Chem. Commun. 1985, 50, 1753.
(24) For exemplificative results on halobenzenes and anilines, see: (a) Loeff,
I.; Lutz, H.; Lindquist, L. Isr. J. Chem. 1970, 8, 141. (b) Perichet, G.;
Chapelon, R.; Pouyet, B. J. Photochem. 1980, 13, 67. (c) Sarkar, S. K.;
Maiti, A.; Kastha, G. S. Chem. Phys. Lett. 1984, 105, 355. (d) Huppert,
D.; Rand, S. D.; Reynolds, A. M.; Rentzepis, P. M. J. Phys. Chem. 1982,
77, 1214. (e) Ermolaev, V. L.; Svitachev, K. K. Opt. Spectrosc. 1959, 7,
393. (f) Grieser, F.; Thomas, J. K. J. Chem. Phys. 1980, 73, 2115.
(25) Previtali, C. M. J. Photochem. 1985, 31, 233.
Iodoanilines 4a,b were more reactive than the bromo deriva-
tives with a quantum yield of decomposition of ca. 5% under
all of the conditions examined and anilines 6 as the main
9
13184 J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003

Photochemistry of Haloanilines
A R T I C L E S
Figure 3. Triplet state of haloanilines 1b-3b. Bond lengths (Å) at
UB3LYP/6-31G(d) and UB3LYP-C-PCM/6-31G(d) (in parentheses) are
reported.
Figure 1. Transient absorption upon flashing fluoroaniline 1a in cyclo-
hexane ([) and acetonitrile (9). Inset: decay of the absorption at 450 nm
in cyclohexane, time unit 1 µs.
As for the bromo- and iodoanilines, only weak signals were
observed in the region 370-410 nm, with little difference in
the solvent tested.
Irradiation in a frozen solution was also explored, with
negative results. As an example 1 h of irradiation of the glass
obtained from a 1 × 10-4 M solution of chloroaniline 2a in
EtOH at 90 K led neither to a significant change of the
absorption spectrum nor to the formation of a permanent
photoproduct as determined by HPLC analysis.
Thermochemistry of the Homolytic and Heterolytic Dis-
sociation Processes in the Gas Phase and Acetonitrile
Solution. It was vital for the rationalization of the above data
that the viability of homolytic or heterolytic fragmentation of
the haloanilines was evaluated. Cleavage of the weak carbon-
iodine bond (63.7 kcal/mol in iodobenzene)26 was expected, but
in the case of the other halides a prediction of the feasibility
and of the mode of fragmentation was not obvious. As is
illustrated in the Discussion, the photochemical reaction pro-
ceeded from the triplet state. Accordingly, the structure and the
reactivity of triplet haloanilines were examined at the UB3LYP/
6-31G(d) level of theory. This method had been previously
shown to be well suited for triplets and when applied to the
haloanilines actually reproduced the experimental C-X bond
dissociation energy data (see the Supporting Information). The
optimized triplet states in the gas phase and acetonitrile solution
of compounds 1b-3b (by the C-PCM solvation model) were
characterized by the out-of-plane deformation of the C4 atom
and of the C4-X bond (see Figure 3).
Figure 2. Transient absorption 1 µs after flashing fluoroaniline 1a in
methanol (a) and trifluoroethanol (b). Inset: decay of the absorption at 450
nm in TFE, time unit 10 µs.
As it appears from Table 4, the C-F bond in triplet 1b3*
was only slightly elongated (2.5% in the gas phase, 3.5% in
MeCN). The elongation of the C-Cl bond in 2b3* was more
marked in the gas phase (11.3%) and increased strongly in
MeCN bulk (32.1%), while C-Br in 3b3* was severely stretched
already in the gas phase (19.6%; 27.1% in MeCN).
Furthermore, the C4-F bond in 1b3* had a polar character
with roughly equal opposite charges (0.25-0.30) on fluorine
and C4 atoms, and the charges did not vary in MeCN with
respect to the gas phase. On the contrary, the negative charge
on the chloro atom increased markedly in 2b3* on going from
the gas phase to MeCN (up to -0.505) and was not counterbal-
anced by a charge on C4 (-0.06 and +0.03 in the two cases),
most of the positive charge being localized on the amino group.
The bromoaniline triplet was similar to the chloro analogue in
this respect.
shifted by ca. 20 nm (e.g., λmax ) 425 nm in methanol) and the
persistent absorption in alcohols was even longer lived (τ )
ca. 100 µs).
With the chloroanilines 2a,b flash photolysis experiments had
been previously carried out in various solvents both by Othmen
et al.17a and by us.20b They revealed the triplet in cyclohexane
(see Table 3) and a persistent absorption closely matching that
observed with 1a,b in methanol. Flashing 2a,b in TFE likewise
gave results similar to those of the fluoro derivatives. A
difference between the chloro- and fluoroanilines was that with
the former derivatives the end-of-pulse absorption had a
maximum at 430 nm and evolved into a persistent absorption
extending from 340 to 510 nm, similar to that observed in
alcohols, whereas with 1a,b the triplet was observed. Experi-
ments in the presence of 1 M allyltrimethylsilane were then
carried out and showed that the long-lived transient was
suppressed with both chloro- and fluoroanilines.
(26) Konimar, R. J.; Price, S. J. Can. J. Chem. 1976, 54, 2981.
9
J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003 13185

A R T I C L E S
Freccero et al.
Table 4. Triplet State of Haloanilines 1b-3b [at UB3LYP/6-31G(d)]
a
b
dC-X
,
ET,
dC-X,
aniline
Å
kcal/mol
Å
charge(C4)b
charge(X)b
1a, H2NC6H4F
1.355
1.762
1.915
78.0
72.2
68.0
1.392 (1.406)
1.961 (2.327)
2.290 (2.434)
1.421c
+0.258 (+0.238)
-0.06 (+0.03)
+0.013 (+0.032)
+0.237c
-0.290 (-0.315)
-0.197 (-0.505)
-0.303 (-0.458)
-0.300c
1b, H2NC6H4Cl
1c, H2NC6H4Br
1a, H2NC6H4F‚H4O
a In the ground state. b In the gas phase; values in parentheses were taken in MeCN. c In the presence of one molecule of water in MeCN.
Table 5. Computed (B3LYP) Electronic Energies (kcal/mol) for the Triplet States of Haloanilines and Homo- and Heterolytic Fragmentationa
ET
∆E(vacuum)
∆E(MeCN)
∆E(vacuum)
∆E(MeCN)
1b, H2NC6H4F
78.0
72.2
68.0
78.0
H2NC6H4• + F•
51.5
21.3
18.3
39.0
52.0
23.4
18.8
37.0
H2NC 6H4+ + F-
149.1
81.1
74.5
9.1
-31.4
-34.6
-4.6
2b, H2NC6H4Cl
3b, H2NC6H4Br
1b, H2NC6H4F‚ H2O
H2NC6H4• + Cl•
H2NC6H4• + Br•
H2NC6H4• + F• + H2O
H2NC6H4+ + Cl-
H2NC6H4+ + Br-
H2NC6H4+ + HF + OH-
103.7
a Solvent effect calculation; see the Experimental Section.
The thermochemistry of both homolytic and heterolytic
fragmentations (eqs 1 and 2) was then evaluated by calculating
the potential energy change associated with these processes.
4-H2NC6H4-X3* f 4-H2NC6H4• + X•
4-H2NC6H4-X3* f 4-H2NC6H4+ + X-
(1)
(2)
The key results are summarized in Table 5. With bromoaniline,
homolytic fragmentation is moderately endothermic in the gas
phase (18.3 kcal/mol), while the heterolytic process is largely
so (74.5 kcal/mol). Shifting the calculation to an acetonitrile
solution (with optimization of both reactants and adducts in
acetonitrile bulk) left the homolytic process unchanged, while
making heterolysis a markedly exothermic reaction (-34.6 kcal/
mol). Essentially the same situation was encountered with
chloroaniline. On the other hand, the strong carbon-fluorine
bond was not as easily cleaved. The homolytic reaction was
strongly endothermic in the gas phase (51.5 kcal/mol) as well
as in acetonitrile (52.0 kcal/mol). Heterolytic fragmentation was
extremely endothermic in the gas phase, and although dramati-
cally dropping from 149.1 to 9.1 kcal/mol, it remained unfavor-
able in a polar aprotic solvent such as acetonitrile.
The experimentally noted effect of protic solvents on the
photochemistry of fluoroanilines suggested also taking into
consideration proticity. To model the specific interaction with
the triplet excited state of fluoroaniline, we added an ancillary
water molecule H-bonded to the fluorine atom. As it appears
from Table 4 (last entry) and Figure 3, the characteristics of
triplet 1b3* were only scarcely affected by the presence of water.
However, in this case additional cleavage modes involving a
molecule of water could be considered (eqs 3 and 4).
Figure 4. Geometries, Mulliken charges (with hydrogen charges summed
into heavy atoms), and spin densities of triplet chloroaniline 2b at a 3.4 Å
C-Cl bond length in the gas phase (left) and in acetonitrile (right). Bond
lengths are in angstroms and energies in kilocalories per mole.
Energetics of the Fragmentation Processes along the
C‚‚‚X Reaction Coordinate. A detailed mapping of the
fragmentation process was carried out both in the gas phase
and in MeCN by 0.1 Å increments of the C-X bond length
and optimization at each point. Starting with the chloroaniline
triplet, for which homolytic cleavage is slightly less endothermic
in the gas phase (21.3 kcal/mol vs 23.4 kcal/mol for heterolysis),
the scan showed a rather uniform increase in the potential
surface.
The homolytic mode of the reaction could be recognized
through the structure of the intermediates. At 10.51 kcal/mol,
the bond was essentially broken (dC-Cl ) 3.461 Å; see Figure
4, left) and the chlorine atom had slightly increased the charge
(-0.336) and strongly the spin (0.663), while the charge on C4
was unchanged (+0.05).
The result was quite different in MeCN bulk. In this case,
heterolytic cleavage occurred at a comparable elongation of the
C-Cl bond as in the gas-phase dissociation process above (dC-Cl
) 3.427 Å; see Figure 4, right). The energy had increased by
8.98 kcal/mol, and this involved localization of the charge
(-0.638) rather than of the spin (0.357) at the chlorine atom,
while the organic moiety had a geometry quite similar to that
we previously calculated for triplet 4-aminophenyl cation,22a
with an almost planar structure. The charge was mainly located
on the C1-NH2 moiety, not on C4 (+0.089).
The dissociation process of the fluoroaniline triplet presented
a more complex energy profile. In the gas phase, strongly
endothermic (51.5 kcal/mol) dissociation proceeded via a
transition state (30.7 kcal/mol, TS-1b in Figure 5) with the
fluorine atom between C4 and C3 (C4-F ) 2.106 Å, C3-F )
4-H2NC6H4-F3*‚‚‚H2O f 4-H2NC6H4• + HF + HO• (3)
4-H2NC6H4-F3*‚‚‚H2O f 4-H2NC6H4+ + F-‚‚‚H2O (4)
Table 5 shows that the homolytic cleavage according to eq 3
remained strongly endothermic both in the gas phase (39.0 kcal/
mol) and in acetonitrile solution (37.2 kcal/mol), but heterolysis
according to eq 4 became slightly but clearly exothermic in
acetonitrile (-4.6 kcal/mol), due to the determining contribution
of the strong H-bonding involving fluorine anion and water.
9
13186 J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003

Photochemistry of Haloanilines
A R T I C L E S
Scheme 3
Table 6. Some Kinetic Parameters for the Excited States of
Anilines 7 and Haloanilines 1-4
k
ISC, s-1
kd, s-1
khet, s-1
khomo, s-1
7, X ) H 2.5 × 108 a 7 × 105
1, X ) F 2.5 × 108 b 1 × 106
<1 × 105 (MeCN)
1 × 106 (MeOH)
<1 × 105
3 × 107 (MeCN, MeOH) <1 × 105
>5 × 106
2, X ) Cl 3 × 109 b
3 × 107
3, X ) Br >1 × 1010 >1 × 108
4, X ) I
>1 × 1010 >1 × 108
a From ref 24c. b From ref 23 [see also the reported value for PhCl, kISC
) ca. 1 × 109 s-1 (Ichimura, T.; Mori, Y. J. Chem. Phys. 1973, 58, 288)].
Figure 5. TS and intermediate involved in C-F bond cleavage of triplet
fluoroaniline 1b (top) and in the complex with one molecule of water
(bottom) in the gas phase at the UB3LYP/6-31G(d) level. Stationary points
optimized in MeCN bulk (in parentheses) have been optimized at the
UB3LYP-C-PCM/6-31G(d) level of theory. Bond lengths are in angstroms
and energies in kilocalories per mole.
Previous studies established that in fluid solution the main
decay path for the singlet is intersystem crossing (ISC) to the
triplet state, both for parent aniline (ca. 80%)25 and for the chloro
derivative (ca. 95%),17a,20b and for the fluoroanilines both ΦF
and τF are close to those of the parent compound. Thus, the
main path of decay from the singlet is ISC, and kISC (see Scheme
3) is evaluated as (2-5) × 108 s-1 for the fluoroanilines, (1.3-
2) × 109 s-1 for the chloroanilines, and >1 × 1010 s-1 for the
bromo and iodo analogues, where fluorescence is too short lived
to be detected (Table 6).
The triplet is clearly revealed by the T-T absorption in flash
photolysis (Figure 1) for the chloro- and fluoroanilines, with a
lifetime of 1 µs or less for dilute solutions in apolar solvents
(Table 3). A similar transient is not detected with either the
bromo or the iodo analogues. Since there is no reason to expect
a dramatic decrease of the absorbance, this indicates efficient
depopulation of the triplet due to a heavy atom effect (T1 to S0,
k′ISC = kd more than an order of magnitude larger than in the
previous cases).
1.96 Å). Such a TS is not directly connected to free products,
but along the reaction coordinate there is an additional minimum
(12.8 kcal/mol, i-1b) with the fluorine atom bonded at C3 (C4-F
) 2.377 Å, C3-F ) 1.421 Å). The presence of a molecule of
water only marginally affected the path followed toward
homolytic dissociation (39.0 kcal/mol) through a transition state
(24.5 kcal/mol, C4-F ) 2.205 Å, C3-F ) 1.995 Å, F-H )
1.739 Å, TS-1b + H2O) and a minimum (C4-F ) 2.380 Å,
C3-F ) 1.453 Å, H-F ) 1.901 Å, i-1b + H2O).
In neat MeCN both transition-state and minimum geometries
for fluorine transfer from C4 to C3 were similar (respectively,
27.39 kcal/mol, C4-F ) 2.167 Å, and C3-F ) 1.964 Å and
12.87 kcal/mol, C4-F ) 2.374 Å, C3-F ) 1.435 Å; see Figure
5), but now the thermochemistry of heterolysis was only slightly
endothermic (by 9.1 kcal/mol).
As for the photochemical reaction, this is efficient (Φ g 0.5)
only in polar solvents with the chloroanilines and only with
protic solvents with the fluoro derivatives (see Table 2). With
these substrates, no parallel change occurs in the photophysical
parameters of the singlet excited state (which if anything lives
longer in polar or protic solvents; see Table 3). Thus, the singlet
is not involved in the photoreaction, which proceeds only from
the triplet state. Indeed, the triplet becomes too short lived to
be detected in the media where these substrates react. A fortiori,
any photochemical reaction involves the triplet state with the
bromo and iodo derivatives, for which S1-T1 intersystem
crossing is even faster.
Flash photolysis data support this representation. Under
conditions where the haloanilines react efficiently, the triplet
(Figure 1) is substituted by a long-lived transient (Figure 2).
The long-wavelength part of such a transient is, in accord with
previous reports,17a,20b attributed to radical cation 7•+. The short-
wavelength part is attributed to the adduct cation 6-H+ (Scheme
2).20b,21d The latter part is more important in TFE (and from 2
also in MeCN), where indeed 6 is the main product.
The geometry of the intermediates was similar also when a
water molecule was added (see Figure 5, bottom), but heterolytic
fragmentation according to eq 4 was now exothermic (by 4.6
kcal/mol). We were unable to locate a further TS between the
minimum (i-1b + H2O) and a separated triplet aniline cation-
F-‚‚‚H2O complex. However, it was clear that a molecule of
water assisted C-F fragmentation, as shown by the increased
C
4-F (2.305 Å) and decreased H-F (1.651 Å) distances both
in transition state TS-1b + H2O and in intermediate i-1b +
H2O (C4-F ) 2.380 Å, C3-F ) 1.466 Å, H-F ) 1.852 Å; see
Figure 5). Furthermore, the path was now more favorable and
both structures had now a lower energy (TS-1b + H2O, 14.4
kcal/mol; i-1b + H2O, 11.71 kcal/mol).
Discussion
Excited State Involved. In the case of aniline, as with many
other aromatic substrates, the sum of the fluorescence quantum
yield and phosphorescence quantum yield in the glass at low
temperature (ΦF + ΦP) comes close to unity. In fluid solution,
the fluorescence lifetime is much shorter than the natural
lifetime.
Among the presently considered compounds (Table 1),
fluoroanilines 1a,b both fluoresce and (in the glassy state)
phosphoresce with similar intensity as parent anilines 7a,b.
Quantitative Determination of Heterolytic Cleavage. We
previously supplied chemical trapping evidence that dechlori-
nation of 4-chloroanilines 2a,b gives triplet 4-aminophenyl
cation 5.20b,22a The selective electrophilic reactivity of this
9
J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003 13187

A R T I C L E S
Freccero et al.
intermediate supports its role, although direct detection fails,
apparently due to spectroscopic characteristics. This species adds
efficiently to π nucleophiles such as alkenes and (hetero)-
aromatics, but not to σ nucleophiles such as water or methanol,
as has been computationally and experimentally documented.20a
The cation is in part reduced through hydrogen transfer from
the solvent via 7•+ and in part adds to the starting material
(Scheme 2).
the medium show no important change with respect to those
referred to the gas phase.
Experimentally, an indication for the homolytic path can be
obtained from the photolysis in cyclohexane. With chloroben-
zene it has been found that homolysis is efficient in this solvent,
with a limiting value of Φ ) 0.5 at infinite dilution, though the
actual values at a reasonable concentration are lower due to the
chemically unproductive formation of an exciplex.28 The reac-
tion has been attributed to internal conversion of the ππ* triplet
to a close-lying, repulsive πσ* triplet,29 as supported by ab initio
UHF calculations.30
In the presence of a sufficient amount of a π nucleophile,
the cation is trapped. This approach has been employed in this
work for titrating the cation formed from haloanilines 1-4 under
various conditions. Allyltrimethylsilane has been chosen, since
elimination of the electrofugal Me3Si+ group leads to a single
product, allylaniline 8, which is easily gas chromatographically
determined. As shown in Table 2, product 8 is formed efficiently
from chloroanilines in polar solvents and from the fluoro
derivatives in protic solvents. Under such conditions, the long-
lived transients are not observed, apparently due to quenching
of their precursor, phenyl cation 5 (see Scheme 2). On the
contrary, the inefficient dechlorination of 2a,b in cyclohexane
is not affected in the presence of the alkene and the products
remain anilines 7a,b. Thus, under these conditions the homolytic
path via 9 is followed both for chloroanilines and for analo-
gously behaving bromo- and iodoanilines (see Table 2).
Structure and Homolysis of Haloaniline Triplets. The
lowest lying triplet of haloanilines 1b-3b is in every case
characterized by a strong out of plane deformation of the halogen
atom. Previously, an out of plane shift of the fluorine atom in
triplet fluorobenzene was indicated by the SINDO1 method.27
Our calculations at the UB3LYP/6-31G(d) and UB3LYP-C-
PCM/6-31G(d) levels of theory (for the gas phase and aceto-
nitrile solution, respectively) confirm such evidence. In addition,
our data suggest that such a deformation is accompanied by
elongation of the C-X bond. This is small in fluoroaniline (from
1.355 Å for the singlet ground state to 1.392 and 1.406 Å for
the triplet state in the gas phase and in acetonitrile solution,
respectively), while it is conspicuous for both cloroaniline (2.327
Å) and bromoaniline (2.425 Å) in acetonitrile (see Table 4).
Noteworthy is the bulk solvent effect on C-X elongation
for the cloroaniline triplet. The bond distance increases by 0.20
Å in the gas phase and by 0.56 Å in acetonitrile solution. The
deformation initiates the path toward homolysis, although the
barrier on the potential energy surface (PES) remains significant
for both chloroaniline and bromoaniline triplets (21.3 and 18.3
kcal/mol in the gas phase, essentially unchanged in MeCN).
Cleavage is expected to be easier in iodoaniline in view of the
much lower (by ca. 15 kcal/mol) bond strength. On the contrary,
with the fluoro analogue elongation of the bond is small and
does not increase in acetonitrile, homolysis remaining highly
unfavorable in both cases (Table 4).
With the present haloanilines, homolysis is, as expected,
negligible with the fluoro derivative and inefficient with the
other derivatives, since the triplet energy is lower than that of
halobenzenes. This is in accord with the straightforward
correlation between efficiency and the difference between
excited-state energy and bond strength generally observed with
haloarenes and with the calculated endothermicity of the process
(Table 5). In the event, Φhomo decreases from chloro (0.02) to
bromo (0.002) derivatives and increases to iodo derivatives
(0.05). This is due to the balance between competing processes
from T1, homolysis and heavy-atom-induced intersystem cross-
ing to S0, the rate of both increasing with atomic weight. In
every case khomo < k′ISC (Scheme 3). With iodoanilines, khomo
= 0.05kd; thus, >5 × 106 s-1, with that of the bromo derivatives
at least an order of magnitute lower.
Triplet Heterolysis. Calculations at the UB3LYP/6-31G(d)
level show that triplet heterolysis in the gas phase is a strongly
endothermic process for fluoro-, chloro-, and bromoanilines, and
the barrier on the PES for such a process is quite high (149.1,
81.1, and 74.5 kcal mol-1, respectively). However, evaluation
by C-PCM stationary point optimization of the acetonitrile bulk
effect evidences a strong change in both triplet chloroaniline
structure (as discussed above) and PES profile. In detail, the
optimized triplet in MeCN shows not only a more elongated
C-Cl bond, but also an increased anionic character of the
chlorine atom with respect to that in the gas phase (from -0.197
to -0.505). The corresponding positive charge resides es-
sentially on the amino group.
Mapping the cleavage shows no transition state, but a smooth
increase of energy accompanied by localization of the charge
on the chloro atom. When the C-Cl distance is stretched to
3.4 Å, the organic moiety attains a planar configuration very
close to that of the final product, phenyl cation. The moderate
barrier encountered (ca. 9 kcal/mol) and the strong exothermicity
of the overall process (-31 kcal/mol) clearly indicate the
feasibility of the heterolytic process. This is supported both by
the increase of the reaction efficiency with solvent polarity (for
2a, Φ ) 0.2 in CH2Cl2, 0.85 in MeCN, 0.97 in MeOH, and 0.5
in CF3CH2OH, with a decrease in the last solvent presumably
because H-bonding by the acidic alcohol limits the electron
Mapping the surface for the chloroaniline triplet in the gas
phase shows no transition state, but a smooth increase of the
energy, with increasing spin, not charge, localization on the
chloro atom (Figure 4). With fluoroaniline, the strong endot-
hermicity of C-F homolysis (51.5 kcal/mol) leads to an
interesting change in the path, and the fluorine atom migrates
to position 3 and reaches, after passing a transition state, a
minimum with a carbene configuration (see the structure i-1b
in Figure 5). In this case, calculations referred to acetonitrile as
(28) Bunce, N. J.; Bergsma, J. P.; Bergsma, M. D.; De Graaf, W.; Kumar, Y.;
Ravanal, L. J. Org. Chem. 1980, 45, 378. Arnold, D. R.; Wong, P. C. J.
Am. Chem. Soc. 1977, 99, 3363.
(29) Kadi, M.; Davidsson, J.; Tarnovsky, A. N.; Rasmusson, M.; Akesson, E.
Chem. Phys. Lett. 2001, 350, 93. Okutsu, T.; Kounose, N.; Tsuchiya, J.;
Suzuki, T.; Ichimura, T.; Hiratsuka, H. Bull. Chem. Soc. Jpn. 2000, 73,
1763. Ichimura, T.; Mori, Y.; Shinohara, H.; Nishi, N. Chem. Phys. 1994,
189, 117. Previtali, C. M.; Ebbesen, T. W. J. Photochem. 1985, 30, 259.
Wang, G. J.; Zhu, R. S.; Zhang, H.; Han, K. L.; He, G. Z.; Lou, N. Q.
Chem. Phys. Lett. 1998, 288, 429.
(30) Nagaoka, S.; Takemura, T.; Baba, H.; Koga, N.; Morokuma, K. J. Phys.
Chem. 1986, 90, 759.
(27) Malar, E. J. P.; Jug, K. J. Phys. Chem. 1984, 88, 3508.
9
13188 J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003

Photochemistry of Haloanilines
A R T I C L E S
donation from the amino group) and by the essentially complete
trapping of the cation by the allylsilane to produce 8 (except in
good hydrogen donating MeOH, where reduction still competes
with a 20% proportion). The role of solvation in allowing
detachment of the oppositely charged fragments is also indicated
by nonreactivity of 2a in rigid ethanol glass, where the ions
apparently recombine. The dramatic change of Φ-2 in polar
solvents while the endothermicity of homolysis remains un-
changed (Table 5) ensures that the process occurring is
heterolytic fragmentation directly from the triplet, not electron
transfer after homolysis (path f, not path a′, in Scheme 1).
Path a′ is probably involved with iodoanilines (Φ-4 low and
solvent independent, though the proportion of product 8 grows
somewhat in polar solvents), while the tendency of bromoa-
nilines to heterolysis in polar solvents is barely detectable since
the fast decay of the triplet cuts photochemistry to low values
(Φ-3 ) 0.002 to 0.02; see Table 2).
On the other hand, examination of fluoroanilines shows no
change parallel to that of the chloro analogues in MeCN (and
actually the quantum yield remains low). In fact, calculations
support that the fragments, triplet phenyl cation and fluoride
anion, are greatly stabilized in comparison to those in the gas
phase, but remain at a higher energy on the PES with respect
to the bonded triplet state (9.1 kcal mol-1). However, when both
bulk (acetonitrile) and specific interactions with a protic
medium, which has been computationally modeled by an
ancillary water molecule added to the reactive system (H2-
NC6H4F‚H2O), are taken into account, calculations do show a
noteworthy change. A transition state still remains, but the
activation energy is lowered (14.4 kcal/mol), and although the
path still goes over a minimum with the fluorine atom essentially
bonded to C3 (i-1b + H2O), the final products, viz., the phenyl
cation and the fluoride anion hydrogen bonded to the water
molecule, are stabilized with respect to the triplet state by -4.6
kcal/mol. Thus, the experimentally observed efficient defluori-
nation in alcohols (up to Φ ) 0.5) is not merely due to the
increased polarity of the medium, but is the result of synergism
with the strong H-bonding interaction (eq 4). Once again,
trapping by allylsilane is unambiguous evidence for the involve-
ment of phenyl cation in protic media.
via photoheterolysis of halides complements the formation of
such intermediates from diazonium salts and, differently from
that case,31 leads exclusively to the triplet state, the reactions
of which are characterized by a high selectivity, in contrast to
the unselective reactions of singlet phenyl cations. Photohet-
erolysis of the aryl-Cl and aryl-F bonds appears to be a
convenient method for the smooth generation of aryl cations
and allows exploitation of the chemistry of these as yet scarcely
investigated intermediates.21 Addition of aryl cations to alkenes,
aromatics, and heterocycles is a synthetically promising
method.20,22
Furthermore, the rationalization of the structure and reactivity
of triplet haloanilines is significant for applications such as the
photochemical decontamination of aryl halides as pollutants in
water,32a for phototochemical applications in nanotechnology,32b
and for understanding the efficient (Φ ) 0.5) heterolytic
defluorination of several fluoroquinolone drugs in water,33 which
is probably the cause of the disturbing phototoxic side effect
of those antimicrobic agents.
Experimental Section
General Procedures. The haloanilines either were commercial
samples or were prepared by N-methylation34 or ring iodination35
according to published procedures. The other reagents (of commercial
origin) were distilled or recrystallized before use. For the irradiations,
spectroscopic grade solvents were used as received.
Photochemical Reactions. Irradiations were carried out by using 5
mL portions of the haloaniline solutions in quartz tubes. The tubes
were capped after being flushed with argon for 15 min and externally
irradiated by means of six 15 W phosphor-coated lamps (center of
emission 310 nm) for an appropriate time in a merry-go-round
apparatus. The conversion was limited to 20%. Product formation was
assessed by GC and HPLC on the basis of calibration curves with
dodecane as the internal standard. Allylanilines 8 have been previously
characterized.22a The light flux for quantum yield determination was
measured by ferrioxalate actinometry. Irradiations in a glassy matrix
were carried out in a quartz cuvette immersed in an Oxford DN104
liquid nitrogen cryostat kept at 90 K.
Photophysical Measurements. Fluorescence spectra and lifetimes
were measured by means of an Edinburgh Instruments F900 single-
photon-counter instrument. Phosphorescence spectra were measured at
77 K by means of an Aminco Bowman SPF instrument fitted with a
rotating phosphoroscope.
Flash Photolysis. The laser flash photolysis studies were carried
out by using the fourth (266 nm) harmonics of a Q-switched Nd:YAG
laser (model HY 200, JK Laser Ltd. Lumonics). This delivered 3 mJ
pulses with a duration of ca. 10 ns. The monitor system, arranged in
cross-beam configuration, consisted of a laser kinetic spectrophotometer
(model K 347, Applied Photophysics) fitted with a 200 W Xe arc lamp,
an F/3.4 monochromator, and a five-stage photomultiplier. The signals
were captured by a Hewlett-Packard 54510A digitizing oscilloscope,
and the data were processed on a 286-based computer system using
software developed by Prof. C. Long (Dublin).
Conclusion
The photochemistry of haloanilines proceeds via the triplet
state and exhibits a varied pattern. The course of the reaction
changes from the inefficient (ca. 5%), solvent-independent
homolytic cleavage of iodoaniline to the efficient (ca. 50% or
more) heterolytic cleavage observed in polar solvents with
chloroaniline and in protic solvents with fluoroaniline. Formation
of the strong H-F bond plays a key role in the last case, as
supported by both experimental and computational evidence.
On the other hand, the heavy atom effect decreases the triplet
lifetime and thus limits the efficiency of homolytic fragmentation
also for relatively weak C-Br and C-I bonds.
Heterolytic fragmentation and SN1Ar* reactions are quite
efficient with chloroanilines and fluoroanilines under appropriate
conditions. Literature reports13,14,16 and ongoing experiments
suggest that conditions can be found for extending such reactions
to further aryl halides, at least with electron-donating substituted
derivatives, and their scope may be larger than it now appears
from the literature. The smooth generation of a phenyl cation
Computational Details. The B3LYP method was chosen due to its
viability for computing PESs for organic reactions involving open shell
(31) (a) Zollinger, H. Diazochemistry I; VCH: New York, 1995. (b) Gasper, S.
M.; Devadoss, C.; Schuster, G. B. J. Am. Chem. Soc. 1995, 117, 5206. (c)
Milanesi, M.; Fagnoni, M.; Albini, A. Chem. Commun. 2003, 216.
(32) (a) Boule, P.; Othmen, K.; Richard, C.; Szczepanik, B.; Grabner, G. Int. J.
Photoenerg. 1999, 1, 49. (b) Bermudez, V. M. Surf. Sci. 2002, 519, 173.
(33) Fasani, E.; Barberis Negra, F. F.; Mella, M.; Monti, S.; Albini, A. J. Org.
Chem. 1999, 64, 5282. Monti, S.; Sortino, S.; Fasani, E.; Albini, A. Chem.s
Eur. J. 2001, 7, 391.
(34) Giumanini, A. G.; Chiavari, G.; Musani, M. M.; Rossi, P. Synthesis 1980,
743.
(35) Fabbrini, M.; Galli, C.; Gentili, P.; Macchitella, D.; Pellide, H. J. Chem.
Soc., Perkin Trans. 2 2001, 1516.
9
J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003 13189

A R T I C L E S
Freccero et al.
species, including radical cations, and has been successfully tested.36
UDFT methods and in particular UDFT with hybrid functionals (such
as B3LYP) have been proved to provide a basis for a reasonable
description of biradical species,37,38 or systems with high biradical
character. Furthermore, the suitability of DFT for reliably describing
hydrogen-bonded systems has been the subject of many investigations,39
and such methods have proven useful for studying hydrogen-bonded
complexes.40 The B3LYP functional in particular is highly effective,
at least as long as an appropriate basis set is used.41 Therefore, the
level of theory chosen for the optimization of all the stationary points
in our investigation [UB3LYP/6-31G(d)] appears to be a proper choice.
All calculations were carried out using the Gaussian 9442 and Gaussian
9843 program packages.
The contribution of bulk solvent effects to the potential energy was
calculated via the self-consistent reaction field (SCRF) method using
C-PCM44 employing the HF parametrization by Barone’s united atom
topological model (UAHF),45 as implemented in Gaussian 98. Such a
model includes the nonelectrostatic terms (cavitation, dispersion, and
repulsion energy) in addition to the classical electrostatic contribution.
The former terms, unlike the latter one, always give a positive
contribution to the solvation energy, but this is much less important
than the electrostatic term. For all PCM-UAHF calculations, the number
of initial tesserae per atomic sphere was set to 60 as in the default.
Acknowledgment. Dr. A. Ricci and Mrs. S. Esposti are
thanked for help with the measucements. Partial support of this
work by MIUR, Rome, is gratefully acknowledged.
To confirm the nature of the stationary points, vibrational frequencies
(in the harmonic approximation) were calculated for all the optimized
structures and used, unscaled, to compute the zero-point energies. The
computed relative (to the lowest triplet state) electronic energies for
transition structures and products are listed in Table 5.
Supporting Information Available: This material is available
free of charge via the Internet at http://pubs.acs.org.
(36) (a) Haberl, U.; Wiest, O.; Steckhan E. J. Am. Chem. Soc. 1999, 121, 6730-
6736 (b) Hofmann, M.; Schaefer, H. F., III. J. Am. Chem. Soc. 1999, 121,
6719-6729. (c) Hrouda, V.; Carsky, P.; Ingr, M.; Chval, Z.; Sastry, G.,
N.; Bally, T. J. Phys. Chem. A 1998, 102, 9297-9307.
JA036000R
(37) Gra¨fenstein, J.; Hjerpe, M. A.; Kraka, E.; Cremer D. J. Phys. Chem. A
2000, 1748-1761.
(43) Gaussian 98, Revision A.7: Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.;
Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.;
Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.;
Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.;
Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.;
Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui,
Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B.
B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin,
R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,
A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle,
E. S.; Pople, J. A., Gaussian, Inc., Pittsburgh, PA, 1998.
(38) Freccero, M.; Gandolfi, R.; Sarzi-Amade`, M.; Rastelli, A. J. Org. Chem.
2003, 68, 811-823.
(39) Guo, H.; Sirois, S.; Proynov, E. I.; Salaub, D. R. In Theoretical Treatment
of Hydrogen Bonding; Hadzy, D., Ed.; Wiley: New York, 1997.
(40) (a) Di Valentin, C.; Freccero, M.; Zanaletti, R.; Sarzi-Amade`, M. J. Am.
Chem. Soc. 2001, 123, 8366-8377. (b) Di Valentin, C.; Freccero, M.; Sarzi-
Amade`, M. J. Am. Chem. Soc. 2003, 125, 3544-3553.
(41) (a) Del Bene, J. E.; Person, W. B. Szczepaniak, K. J. Phys. Chem. 1995,
99, 10705. (b) Kim, K.; Jordan, K. D. J. Phys. Chem. 1994, 98, 10089.
(42) Gaussian 94, Revision E.3: Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.;
Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.;
Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M.
A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.;
Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala,
P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts,
R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart,
J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A., Gaussian, Inc.,
Pittsburgh, PA, 1995.
(44) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995. For the original
COSMO model see: (b) Klamt, A.; Schu¨rmann J. Chem. Soc., Perkins
Trans. 1993, 2. (c) Klamt, A.; Jonas, V.; Bu¨rger, T.; Lohrenz, J. C. W. J.
Phys. Chem. A 1998, 102, 5074. (d) For a more comprehensive treatment
of solvation models see: (e) Cramer, C. J.; Truhlar, D. G. Chem ReV. 1999,
99, 2161.
(45) Barone, V.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1997, 107, 3210.
9
13190 J. AM. CHEM. SOC. VOL. 125, NO. 43, 2003