Photodissociation of N-(Triphenylmethyl) anilines: A Laser Flash Photolysis, ESR, and Product Analysis Study

Article abstract of DOI:10.1021/jo9719261

N-(Triphenylmethyl)anilines (Ph₃C-NHPh-R, R = H, o-Me, m-Me, p-Me, p-CPh₃), upon photolysis with 248- and 308-nm laser light, as well as by lamp irradiation with 254-nm light in acetonitrile and hexane solutions undergo exclusively C-N bond homolysis to give the triphenylmethyl radical in a monophotonic process. Product analysis gives as main products Ph₃CH, 9-Ph-fluorene, and PhNH₂. The quantum yields for the formation of Ph₃C are high (0.6-0.8, 248-nm excitation) and independent of the solvent. This effective homolytic dissociation results from the low electronegativity difference between the carbon and nitrogen atoms constituting the bond to be broken, the low bond dissociation enthalpy of the C-N bond, the high excitation energy of the local chromophore (aniline), and probably from a favorable alignment of the C-N bond in a plane perpendicular to the anilino chromophore (due to the large steric requirements of the trityl group), thus enabling an effective hyperconjugative interaction with it. The above dissociation competes effectively with heterolytic cleavage, which is the pathway dominating, e.g., in the photolysis of Ph₃C-Cl in MeCN under the same conditions. At high pulse intensities the trityl radicals formed above are excited by a second photon leading to either electrocyclization to 4a,4b-dihydro-9-phenylfluorenyl radical (DHPF·), or photoionization to Ph₃C⁺, the latter only in MeCN and only on 248-nm photolysis. A new intermediate (9-Ph-4aH-fluorene) on the way to the final product 9-Ph-fluorene is identified resulting from electrocyclization, supporting a mechanism proposed earlier. The optical measurements are supported by ESR studies (irradiation with 254-nm light).

Full text of DOI:10.1021/jo9719261

J . Org. Chem. 1998, 63, 3251-3259
3251
P h otod issocia tion of N-(Tr ip h en ylm eth yl)a n ilin es: A La ser F la sh
P h otolysis, ESR, a n d P r od u ct An a lysis Stu d y1
†
,†
‡
†
†
M. G. Siskos, A. K. Zarkadis,* S. Steenken, N. Karakostas, and S. K. Garas
Department of Chemistry, University of Ioannina, 451 10 Ioannina, Greece, and Max-Planck-Institut f u¨ r
Strahlenchemie, M u¨ lheim, FRG
Received October 17, 1997
N-(Triphenylmethyl)anilines (Ph
3
3
C-NHPh-R, R ) H, o-Me, m-Me, p-Me, p-CPh ), upon photolysis
with 248- and 308-nm laser light, as well as by lamp irradiation with 254-nm light in acetonitrile
and hexane solutions, undergo exclusively C-N bond homolysis to give the triphenylmethyl radical
in a monophotonic process. Product analysis gives as main products Ph
2 3
PhNH . The quantum yields for the formation of Ph C are high (0.6-0.8, 248-nm excitation) and
3
CH, 9-Ph-fluorene, and
•
independent of the solvent. This effective homolytic dissociation results from the low electrone-
gativity difference between the carbon and nitrogen atoms constituting the bond to be broken, the
low bond dissociation enthalpy of the C-N bond, the high excitation energy of the local chromophore
(aniline), and probably from a favorable alignment of the C-N bond in a plane perpendicular to
the anilino chromophore (due to the large steric requirements of the trityl group), thus enabling
an effective hyperconjugative interaction with it. The above dissociation competes effectively with
heterolytic cleavage, which is the pathway dominating, e.g., in the photolysis of Ph C-Cl in MeCN
3
under the same conditions. At high pulse intensities the trityl radicals formed above are excited
by a second photon leading to either electrocyclization to 4a,4b-dihydro-9-phenylfluorenyl radical
•
+
(
3
DHPF ), or photoionization to Ph C , the latter only in MeCN and only on 248-nm photolysis. A
new intermediate (9-Ph-4aH-fluorene) on the way to the final product 9-Ph-fluorene is identified
resulting from electrocyclization, supporting a mechanism proposed earlier. The optical measure-
ments are supported by ESR studies (irradiation with 254-nm light).
In tr od u ction
of the atoms making up the bond to be broken, with
respect to the energy and the shape of the first excited
The photodissociation of the benzylic carbon-heteroa-
singlet (S
almost unchanged as δAB increases, the energy and shape
of S rapidly approaches that of T and the two differ very
1 1 0 1
) and triplet (T ) state; while S and T remains
+
-
+
2 3
tom bond (C-X, X ) Hal, OCOR, OR, SR , PO H , -
NR
3
) of molecules in solution has been the focus of
1
1
extensive product and time-resolved spectroscopic stud-
ies, as well as of theoretical considerations.
little and merge in the fully dissociated limit. In the
special case of a tritopic dissociation of a σ bond (like the
2
3,4,5b
Despite
the many-sided consequences in mechanistic and applied
photochemistry (photochromic materials, dyes, photo-
deprotection techniques, etc.), the fundamental question
of the responsible excited states, the nature of the
primary fragments (radicals, ions), and their intercon-
version via electron transfer (ET) is far from being
settled.5
C-N bond in CH
3
2 1 1
-NH ), both S and T states acquire
3
a
3
similar, dissociative states. Furthermore, Michl et al.
formulated conditions which facilitate the homolytic
photodissociation of a benzylic bond: e.g., (1) a large local
excitation energy, (2) a weak σ bond, (3) excitation into
the triplet state, or efficient intersystem crossing, (4) σ
bond perpendicular to the plane of the local excited
system.
,6b,7,8
We have decided to concentrate our studies on the
photochemistry (248- and 308-nm laser light) of the
(3) (a) Michl, J .; Bonacic-Koutecky, V. Electronic Aspects of Organic
Photochemistry; Wiley-Intrescience, 1990, pp 138, 292, 374. (b) Michl,
J .; Balaji, V. In Computational Advances in Organic Chemistry:
Molecular Structure and Reactivity; O¨ gretir, C., Csizmadia, I. G., Lang,
E. A., Eds.; Kluwer: Dordrecht, 1991; NATO ASI Series, Vol. 330, p
323; (c) Michl, J . Acc. Chem. Res. 1990, 23, 127.
3
Especially Michl et al. pointed out the essential role
(
4) Larson, J . R.; Epiotis, N. D.; McMurchie, L. E.; Shaik, S. S. J .
Org. Chem. 1980, 45, 1388.
5) (a) McGowan, W. M.; Hilinski, E. F. J . Am. Chem. Soc., 1995,
which is played by the electronegativity difference (δAB
)
(
*
Corresponding author. Tel: (+30 651) 98 379. 98398, Fax: (+30
117, 9019. (b) Zimmerman, H. E. J . Am. Chem. Soc. 1995, 117, 8988.
(c) Pincock, J . A. Acc. Chem. Res. 1997, 30, 43.
6
51) 45840. E-mail: azarkad@cc.uoi.gr.
†
University of Ioannina.
(6) (a) Faria, J . L.; Steenken, S. J . Am. Chem. Soc. 1990, 112, 1277.
(b) Bartl, J .; Steenken, S. Mayr, H.; McClleland, R. A. J . Am. Chem.
Soc. 1990, 112, 6918. (c) J agannadham, V.; Steenken, S. J . Am. Chem.
Soc., 1988, 110, 2188. (d) Faria, J . L.; Steenken, S. J . Phys. Chem.
1993, 97, 1924. (e) McClelland, R. A.; Kanagasabapathy, V. M.;
Steenken, S. J . Am. Chem. Soc. 1988, 110, 6913.
‡
Max-Planck-Institut f u¨ r Strahlenchemie.
(
1) Zarkadis, A. K.; Siskos, M. G.; Georgakilas, V.; Steenken, S.;
th
Karakostas, N.; Garas, S. K. Partly presented at the XVIII Interna-
tional Conference on Photochemistry, Warsaw, August 3-8, 1997.
(2) (a) Cristol, S. J .; Bindel, T. H. Org. Photochem. 1983, 6, 327. (b)
Kropp, P. J . Acc. Chem. Res. 1984, 17, 131. (c) Smirnov, V. A.;
Plotnikov, V. G. Russ. Chem. Rev. 1986, 55, 929. (d) Das, P. K. Chem.
Rev. 1993, 93, 119. (e) J ohnston, L. J . Chem. Rev. 1993, 93, 25. (f)
Wilson, R. M.; Schnapp, K. A. Chem. Rev. 1993, 93, 223. (g) Scaiano,
J . C.; J ohnston, L. J . Org. Photochem. 1989, 10, 309.
(7) Gaillard, E.; Fox, A. M.; Wan, P. J . Am. Chem. Soc. 1989, 111,
2180.
(8) (a) Manring, L. E.; Peters, K. S. J . Phys. Chem. 1984, 88, 3516.
(b) Peters, K. S.; Li, B. J . Phys. Chem. 1994, 98, 401. (c) Dreyer, J .;
Peters, K. S. J . Phys. Chem. 1996, 100, 15156.
S0022-3263(97)01926-9 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/30/1998

3
252 J . Org. Chem., Vol. 63, No. 10, 1998
Siskos et al.
Ta ble 1. Sp ectr oscop ic Absor p tion Da ta of 1a -e in
Certainly, with the aniline chromophore there is the
risk of introducing additional complications into the
system due to the well-known tendency of the aniline
function to photoionize in polar solvents, usually water,
a
MeCN
1
1,15
or to break the N-H bond in nonpolar solvents.
R
H
o-Me
m-Me
p-Me
p-Ph3C
hν
hν
-
•+
•
•
1
a
b
c
d
e
e + PhNH
7
PhNH2
8 PhNH + H
2
water
hexane
λmax (nm)
log ꢀ
M
236
236
240
240
265
4.23
4.21
4.12
4.16
4.08
-
1
cm-1
)
We present here a laser flash photolysis (LFP) and
electron spin resonance (ESR) study, supported by prod-
uct analysis, on the aniline derivatives 1. Our data
indicate that compounds 1 undergo clean homolytic
photodissociation to trityl radical, in both polar (aceto-
nitrile) and nonpolar solvents (hexane), with high quan-
tum yields, competing effectively with other possible
pathways, such as heterolysis or photoionization. This
photobehavior follows nicely Michl’s predictions. In
(
a
They show an additional shoulder at 300-310 nm (log ꢀ ∼
-
1
-1
3
.3 M cm ).
N-(triphenylmethyl)aniline derivatives 1a -e, keeping in
mind Michl’s “spirit”: a low δCN (nitrogen is the closest
of the first row elements to carbon), a weak central C-N
9
σ bond (∼39 kcal/mol), and a local chromophore
aniline)8
a,10a
(
with high excitation energy (Esinglet > 90
kcal/mol, Etriplet > 70 kcal/mol).10 While the absorption
sharp contrast, we obtained (see Figure 1c) under the
spectra of 1a -e are the sum of the methylene-isolated
phenyl and anilino chromophores (see Table 1), the
fluorescence spectra which we obtained in hexane are
characteristic of the anilino chromophore (emission at
•
same conditions (MeCN/248-nm laser) a ratio [Ph
3
C ]:
+
[
Ph
very different than in diarylmethyl chlorides;
noted here is the large δCCl and the strengthening of the
3
C ] = 1:1.4 by the photodissociation of Ph
3
C-Cl, not
16a
to be
1
0b
λ
max ≈ 330 nm).
our system to Michl’s approach would presumably be the
nonpure” ππ* character of the lowest excited states (S
), due to the perturbing effect of the lone pair on
The main justified objection to relating
1
6b
C-Cl bond (∼60 kcal/mol) compared to C-N.
“
T
1
,
Exp er im en ta l Section
1
Ma ter ia ls. The synthesis of N-(triphenylmethyl)aniline
1
1
nitrogen. However, in aniline this contribution seems
17,18
derivatives 1a -e has been described previously.
They
to be rather small (∼22%), and S
1 2
and probably S remain
were purified by recrystallization (twice) from methanol or a
in fact ππ* states.1
1a,12
mixure of methanol and CHCl . The triphenylmethyl chloride
3
The photodissociation of the C-N bond in benzylic
(Merck) was recrystallized from hexane by adding drops of
acetyl chloride. Ph CH (Fluka) was recrystallized from ligroin,
3
compounds has recently attracted much attention, but
most studies concern ammonium salts,2a,d where the C-N
9-Ph-fluorene (9-Ph-flH) was prepared from the corresponding
carbinol,19 and 9-Ph-fl-NHPh from the 9-Ph-fl-Cl and aniline.
Acetonitrile (MeCN), hexane, cyclohexane, ethanol, tetrachlo-
romethane (Merck), and n-butyl chloride (Fluka) were spec-
troscopic grade and used as received.
20
bond already possesses considerable ground-state car-
+
13
bocation-dipole character (R
3
C ‚‚‚NR′
3
)
due to the
increased electronegativity of the ammonium group.
There do exist also some product studies apart from the
classical papers by Lewis and Lipkin
1
4
In str u m en ts. Fluorescence spectra (uncorrected) were
obtained on a Spex Fluorolog. Electron spin resonance spectra
were obtained on a Varian E-109 spectrometer. Gas chro-
matographic analyses and separations were conducted on a
1
4a
and Porter and
Stracham.1
4b,15b
(
9) On the basis of the known BDH of the C-N bond in H C-NHPh
3
9a
(
68.9 kcal/mol) and the calculated difference of ∼30 kcal/mol in the
(12) The spectrum is recorded 18 µs after the pulse to avoid T-T
9
b
9c,47
12a
BDH’s between H
3
C-H (105 kcal/mol) and Ph
3
C-H (75 kcal/mol),
absorptions of the triplet state of aniline (lifetime of the triplet state
1
2b
we arrive at a value ∼39 kcal/mol for the BDH in Ph
3
C-NHPh, not
is 1.2 µs in benzene and 4.3 µs in dioxane; we found 2.8 µs in MeCN).
(a) Malkin, Ya. N.; Ruziev, Sh.; Kuz'min, V. A. J . Gen. Chem. USSR
1987, 57, 560. (b) Shimamori, H.; Sato, A. J . Phys. Chem. 1994, 98,
13481.
considering, however, possible steric effects which should decrease the
BDH even more. (a) Colussi, A. J .; Benson, S. W. Int. J . Chem. Kinet.
978, 10, 1139; For a critical review, see: Batt, L.; Robinson, G. N.
1
The Chemistry of the Functional Groups, Suppl. F; Patai, S., Ed.; Wiley-
Intersience, Chichester, 1982; Part 2, p 1035. (b) McMillen, D. F.;
Golden, D. M. Annu. Rev. Phys. Chem. 1982, 33, 493. (c) Breslow, R.;
Grant, J . L. J . Am. Chem. Soc. 1977, 99, 7745; see for a discussion of
the trityl stabilization energy: (d) Beckhaus, H.-D.; Dogan, B.;
Schuetzer, J .; Hellman, S.; R u¨ chardt, C. Chem. Ber. 1990, 123, 137.
(13) Boyd, S. L.; Boyd, R. J .; Bessonette, P. W.; Kerdraon, D. I.;
Aucoin, N. T. J . Am. Chem. Soc. 1995, 117, 8816.
(14) (a) Lewis, G. N.; Lipkin, D. J . Am. Chem. Soc. 1942, 64, 2801.
(b) Porter, G.; Stracham, E. Trans. Faraday Soc. 1958, 54, 1595. (c)
Ogata, Y.; Takagi, K. J . Org. Chem. 1970, 35, 1642. (d) Ratclieff, M.
A., Kochi, J . K. J . Org. Chem. 1972, 37, 3268. (e) Bekowies, P. J .;
Albrecht, A. C. J . Phys. Chem. 1971, 75, 431. (f) Shi, M.; Okamoto, Y.;
Takamuku, S. J . Chem. Res.(S). 1990, 346.
(15) (a) Qin, L.; Tripathi, N. R.; Schuler, R. H. Z. Naturforsch. 1985,
40a, 1026. (b) Land, E. J .; Porter, G. Trans. Faraday Soc. 1963, 59,
2027. (c) Malkin, Ya. N.; Ruziev, Sh.; Pigorov, N. O.; Kuz’min, V. A.
Bull. Akad. USSR, Ser. Chem, 1987, 51. (d) Leyva, E.; Platz, M. S.;
Niu, B.; Wirz, J . J . Phys. Chem. 1987, 91, 2293. (e) Wagner, B. D.;
Ruel, G.; Lusztyk, J . J . Am. Chem. Soc. 1996, 118, 511.
(
9
b
9b
H
C-H (105 kcal/mol) and PhCH
-NHPh ∼56 kcal/mol.
10) The energy of the lowest excited state of various (triarylmethyl)-
anilino derivatives is reported to be localized on the anilino
BDH of the C-N bond in PhCH
chromophore.8
solvents), 91.8 kcal/mol (polar) and ETriplet ) 71.0 kcal/mol (nonpolar),
7
8
(16) (a) A ratio 1: 1.6 was reported for Ph CH-Cl by Steenken and
2
6
b
•
L. J . Am. Chem. Soc. 1975, 97, 6777. Cremers, D. A.; Cremers, T. L.
Chem. Phys. Lett. 1983, 94, 102. Spears, K. G.; Gray, T. H.; Huang, D.
J . Phys. Chem.. 1986, 90, 779. Geiger, M. W.; Turro, N. J .; Waddel,
W. H. Photochem. Photobiol. 1977, 25, 15. (b) We measured the
co-workers. (b) From ∆H°
f
(Ph
(Ph
kcal/mol is obtained for the BDH of the C-Cl bond in Ph
is reasonable in view of the reported values of 64 kcal/mol for Ph
3
C-Cl)gas ) 53.3 kcal/mol, ∆H°
f
(Cl )gas
C )gas ) 84.5 kcal/mol, a value of 61
C-Cl. This
1
6c
•
9d
) 29.9 kcal/mol, and ∆H°
f
3
3
2
-
6
b
9b
fluorescence spectra of the compounds PhNHMe, PhNHCH
NHCHPh , and PhNHCPh (1a ) (0.05 mM in hexane, excitation at 240
and 300 nm) and found emissions at λmax 328, 332, 327, and 327 nm,
2
Ph, Ph-
2
CH-Cl, and 72.2 kcal/mol for PhCH -Cl. (c) NIST Standard
Reference Database 25; Lias, S. G., Liebman, J ., Lewin, R. D., Kafafi,
S. A., Eds.; 1994; Version 2.0.
2
3
1
0a.
respectively, typical values for the anilino chromophore.
11) (a) Malkin, J . Photochemical and Photophysical Properties of
(17) Siskos, M. G.; Tzerpos, N. I.; Zarkadis, A. K. Bull. Chem. Soc.
Belg. 1996, 105, 759.
(
Aromatic Compounds; CRC, Boca Raton, FL, 1992; p 117, 201. (b)
Malkin, Ya. N.; Kuz'min, V. A. Russ. Chem. Rev. 1985, 54, 1041. (c)
Saito, F.; Tobita, S.; Shizuka, H. J . Photochem. Photobiol. A, 1997,
(18) Siskos, M. G.; Garas, S. K.; Zarkadis, A. K.; Bokaris, E. P. Chem.
Ber. 1992, 125, 2477.
(19) Schuler, R. H.; Hartzell, A, L.; Behar, B. J . Phys. Chem. 1981,
85, 192.
1
06, 119. (d) Saito, F.; Tobita, S.; Shizuka, H. J . Chem. Soc., Faraday
Trans. 1996, 92, 4177.
(20) Kliegl, A. Ber. Dtsch. Chem. Ges. 1905, 38, 284.

Photodissociation of N-(Triphenylmethyl)anilines
J . Org. Chem., Vol. 63, No. 10, 1998 3253
nm light (KrF*) from a Lambda Physik EMG 103MSC excimer
laser or 308-nm light (XeCl*) from a Lambda Physik EMG150E
6
a,b,d
laser.
Qu a n tu m Yield s. These were determined21 by measuring
•
the initial absorbance of Ph
3
C at λmax ) 334 nm. The initial
absorbance of the hydrated electron at 600 nm produced by
photolysis of an aqueous solution of KI under identical
conditions (the same optical density at 248 nm, OD ) 1.0/cm)
was used as the reference. The measurements were carried
out by varying the laser intensity (0-10 mJ /pulse) and plotting
2
1,22b
23
the absorbance versus the laser dose.
We used values of
4
-1
-1
•
ꢀ
e-(aq) ) 1.33 × 10 M cm at 600 nm and ꢀ(Ph
3
C ) ) 3.6 ×
04
M
-1
cm
-1
at 334 nm.
6a,b
The quantum yield for the
1
formation of the hydrated electron was taken as 0.29, a value
2
4
obtained by product analysis following 254-nm irradiation.
The accuracy of this method is not better than (0.1.
P u lse Ra d iolysis Exp er im en ts. A 3-MeV Van der Graaf
electron accelerator was used as radiation source. Dosimetry
was performed with N
2
O-saturated 10 mM KSCN aqueous
solution taking G(OH) ) 6.0 and ꢀ(SCN)2- ) 7600 dm mol
3
-1
-
1
6c
cm at 480 nm.
Resu lts a n d Discu ssion
. P r od u ct Stu d ies. (a) In Aceton itr ile. Product
1
studies were performed in MeCN solution under argon
in quartz cells (lamp irradiation, see Experimental Sec-
tion). The photoconversion of 1a (colorless) is very
effective (about 90% in 6 min) and the solution turns
bright red. At ca. 10% consumption of 1a the major
3
products were found to be Ph CH, 9-Ph-fluorene, and
aniline in an approximately 1:1:2 ratio, using gas chro-
matographic and GC-MS analyses (comparison with
authentic samples). These products inform at first
instance of a homolytic C-N bond cleavage (see also LFP
experiments later) leading to formation of the radicals
•
•
Ph
Actually, the first two products (Ph
fluorene) have been identified since the early studies of
3
C (appearance of bright yellow color) and PhNH .
3
CH and 9-Ph-
trityl radical chemistry2
5-27
(see, however, ref 28) in terms
of the well-known photoproducts of the trityl radical (eq
1). Aniline could be generated from the anilino radical
through hydrogen atom abstraction, see later.
F igu r e 1. Absorption spectra obtained on 248-nm LFP in
MeCN: (a) 0.3 mM 1a in oxygen-saturated solution recorded
at b, 85 ns; O, 170 ns; and 2, 350 ns after the pulse. Inset:
yield of trityl radical (∆A338) as a function of the laser energy
hν
hν
(
under argon); see the text. (b) 0.1 mM 1d under argon
•
•
Ph C-NHPh
98 [Ph C + NHPh] 98 Ph CH +
3 3
3
recorded at 70 ns. Inset: yield of trityl cation as a function of
2
9-Ph-fluorene + PhNH2 (1)
the laser power I (full circles) and I (open circles); see text.
c) 0.7 mM Ph C-Cl under argon: b, 200 ns; 4, 14 µs.
(
3
Under oxygen atmosphere the product that was identi-
fied almost quantitatively was Ph CdO. Benzophenone
Siechromat or Hewlett-Packard 5890, Series II or Varian 3700,
FID gas chromatograph with an OV-1 (10 m) or an SE-54 (35
m) or an OV-1701 (15 m) and RTX-1701 (15 m) capillary
column (injector, 200 °C; detector, 300 °C; column temperature,
2
(21) Wintgens, V.; J onson, L.; Scaiano, J . C. J . Am. Chem. Soc. 1988,
10, 511. Cozens, F. L.; Mathivanan, N.; Mcclelland, R. A.; Steenken,
1
S. J . Chem. Soc., Perkin Trans.2, 1992, 2083.
22) (a) McClelland, R. A.; Banait, N.; Steenken, S. J . Am. Chem.
7
0-280 °C, 8 or 10 °C/min). The hydrogen production was
measured on a Varian 1400 WLD gas chromatograph with a
.5 m charcoal column at 70 °C using argon as carrier gas (42
(
Soc. 1986, 108, 1757. (b) McClelland, R. A.; Banait, N.; Steenken, S. J .
Am. Chem. Soc. 1989, 111, 2929.
3
(
23) Hug, G. L. Natl. Stand. Ref. Data Ser. 1981, 69, 1.
mL/min). GC-MS analyses were performed on a SSQ 700,
EI instrument, using the same conditions and columns.
(24) J ortner, J .; Ottolenghi, M.; Stein, G. J . Phys. Chem. 1962, 66,
2
029, 2037, 2042; 1964, 68, 247.
P h otoch em ica l Exp er im en ts (P r od u ct An a lyses). Pho-
tochemical reactions were run with a Philips HPK-125 W
medium-pressure mercury vapor lamp at 20 °C in special
quartz cuvettes. The photolysis solutions (concentration ∼0.1
mM) were purged with argon or oxygen before they were
irradiated. The same quartz cuvettes were used also for the
reaction with the 248-nm (KF*) or 308-nm (XeCl*) excimer
laser. The photolyses at 254-nm were performed with a
Haraeus TNN-8 W low-pressure mercury lamp.
(25) (a) Gomberg, M.; Cone Ber. Dtsch. Chem. Ges. 1904, 37, 3545;
906, 39, 1469, 2967. (b) Gomberg, M. Chem. Rev. 1924, 1, 91.
1
(26) (a) Schlenk W.; Herzenstein A., Ber. Dtsch. Chem. Ges. 1911,
4
1
4, 3541. (b) Schmidlin J .; Garcia-Banus A., Ber. Dtsch. Chem. Ges.
912, 45, 1344. (c) Bowden S. T.; J ones J . J . J . Chem. Soc. 1928, 1149.
(d) Berlin K. D.; Sturm G. P. J . Chem. Soc. 1964, 2275. (e) Lewis, H.
G.; Owen, E. D. J . Chem. Soc. 1967, 422.
(27) (a) Letsinger R. L., Collat R., Magnusson M. J . Am. Chem. Soc.
1
954, 76, 4185. (b) Wilson, R. M.; Schnapp, K. A.; Hannemann, K.;
Ho, D. M.; Memarian, H. R.; Azadnia, A.; Pinchas, A. R.; Figley, T. M.
Spectrochim. Acta, 1990, 46A, 551. (c) Engel, P. S.; Chen, Y.; Wang,
C. J . Org. Chem. 1991, 56, 3073.
La ser F la sh P h otolysis Exp er im en ts. Solutions of the
corresponding substrates (A/cm ∼1.0) were deoxygenated by
bubbling with argon and photolyzed at 20 °C in a flow system
2
6,27a
(
28) Actually, they
photolyzed concentrated solutions of the R,
•
•
p-dimer of the Ph
component.
3
C , where Ph
(a) Lankamp, H.; Nauta, W. Th.; MacLean, C.
3
C
itself constitutes the minor
28a,42e
(Suprasil quartz cell) using 20-ns pulses (0.3-100 mJ ) of 248
Tetrahedron Lett. 1968, 249.

3
254 J . Org. Chem., Vol. 63, No. 10, 1998
Siskos et al.
is known as a decomposition product (thermal or
absorptions at 337 and 486 nm were affected by purging
photochemical)29c,d of trityl peroxide or trityl hydroper-
the solution with oxygen (Figure 1a) and they decayed
•
35
oxide (oxidation products of Ph
3
C ). Small amounts
in saturated solutions ([O ] ) 9.1 mM) with rate
2
8
8
-1
(
<3%) of PhOH, PhCHdO, and PhNO
2
were also pro-
constants k ) 6.8 ( 0.2 × 10 and 1.3 ( 0.4 × 10 M
-1
duced.
s
, respectively, typical values for scavenging of benzyl
2
9
Analysis of the photochemical products formed by
radicals by oxygen. In contrast, the signals observed
irradiation of 1a with the 248- and 308-nm laser light
at 400-440 nm show no difference in the decay rate in
5
-1
(
50-200 shots, under argon) in MeCN solution indicated
the same products: PhNH , Ph CH, and 9-Ph-fluorene.
b) In n -Hexa n e. The mixture of photoproducts
argon and in oxygen atmosphere (k = 2 × 10 s ).
In a water/MeCN mixture (4:1) of 1a , only the decay
of the absorption at 400-440 nm was accelerated (reac-
2
3
(
obtained by lamp irradiation (λ > 290 nm) of 1a in hexane
has the same composition as in the case of the photolysis
in MeCN, the only difference being the formation of the
2
tion with H O), while with the strong nucleophile azide
-
9
-1 -1
N
3
a rate constant of 3.2 × 10 M
s
was found. We
assign the transient at 400-440 nm to the trityl cation
CdNPh (traces).30
+
imine Ph
At very low conversion (∼3%), the oxygenated products
Ph CdO, Ph COH, PhOH, and aniline are the major
2
Ph
3
C
on the basis of the similarity of its absorption
6
,22
spectrum to that reported in the literature
(to compare
see also Figure 1c) and its nonreactivity with oxygen.
2
3
components identified (GC), in line with the assumption
This assignment is further supported by the reactivity
-
of a primary photohomolysis of the C-N bond and
of this species with the strong nucleophile N
3
; a rate
•
9
-1
-1
formation of the radical Ph
3
C . Obviously, under this
constant k ) 4.1 × 10
M
s
was measured previ-
ously6 for the reaction of Ph
,22
C with the same nucleo-
+
low conversion the small concentration of the trityl
radicals formed does not allow for an effective photocon-
version to the final products according to eq 1, so only
oxidation products, obtained on exposing the samples to
the air, were obtained.
3
phile.
The transient absorbing at 486 nm must be a radical
species because of the 10-fold acceleration of its decay
rate in the presence of oxygen, and on the basis of the
By pulsed irradiation (248 and 308 nm) no imine was
detected, the other products being the same as in MeCN.
By the 308-nm photolysis an increased amount of 9-Ph-
fluorene (GC) was measured and its characteristic ab-
sorption spectrum with λmax at 266, 293, and 303 nm was
observed, identical with that of an authentic sample. It
is interesting to note here (308 nm, 0.1 mM) that no new
photoproducts are formed even after ∼90% consumption
of the starting material 1a . Under oxygen atmosphere
identity of its absorption spectrum with the literature
spectrum,6
a,31b,f
it is assigned to the 4a,4b-dihydro-9-
•
6a,31b,f
phenylfluorenyl radical (DHPF ), the proposed
cy-
•
clization product of the photoexcited Ph
3
C radical (see
Scheme 1). Similar conclusions were drawn also in a
recent product study on the photolysis of (triphen-
ylmethyl)diethylamine.1
4f
The dependence of the intensities of the above absorp-
tions was studied by varying the laser doses in the range
∼0-10 mJ /pulse (248 nm) with the use of filters placed
in the beam, using deoxygenated solutions. A linear
2
Ph CdO is again the main product.
These results show the formation30 of hydrogen gas to
be insignificant (<2%) for low conversions of 1a (<10%).
dependence was found for the absorption at 337 nm,
•
2
.
Id en tifica tion of th e P h otoch em ica lly P r o-
d u ced Tr a n sien ts by La ser F la sh P h otolysis (LF P ).
a ) Aceton itr ile. On 248-nm laser photolyses in deoxy-
genated and oxygenated acetonitrile solution (OD248
.0/cm), N-tritylaniline 1a undergoes at room tempera-
ture very efficient homolysis (Φ ) 0.65), leading to the
indicating that the formation of the radical Ph
3
C is a
monophotonic process (Figure 1a, Scheme 1a), and a
quadratic dependence was found for the signals at 400-
440 and 486 nm, showing a biphotonic pathway leading
(
=
•
1
to the trityl cation (Scheme 1e, or 1i,j) and DHPF radical
(Scheme 1b), respectively. The biphotonic processes are
particularly clear in the cases of the photolysis of 1c and
1d , where relatively strong absorptions in the range of
400-500 nm were obtained (see, for example, Figure 1b).
In the other cases, the signals are too weak (see, for
example, Figure 1a) to make an unambiguous assign-
ment, the other disturbing factor being here the overlap-
formation of the well-known6
max )337 nm); see Figure 1a. In addition, very weak
absorption bands were apparent at λmax ) 400, 440, and
86 nm. Irradiating the other aniline derivatives 1b-e
,31-34
triphenylmethyl radical
(
λ
4
under the same conditions gave almost identical tran-
sient spectra. In Figure 1b is shown the absorption
spectrum obtained by photolysis of the aniline derivative
1
d (R ) p-Me), which is characterized by increased
(31) (a) Bromberg, A.; Schmidt, K. H.; Meisel, D. J . Am. Chem. Soc.
1
984, 106, 3056. (b) Bromberg, A.; Schmidt, K. H.; Meisel, D. J . Am.
Chem. Soc. 1985, 107, 83. (c) Bromberg, A.; Meisel, D. J . Phys. Chem.,
985, 89, 2507. (d) Meisel, D.; Das, P. K.; Hug, G. L.; Bhattacharyya,
absorption maxima at 400, 440, and 486 nm. The
1
8
29a
and 3.16 × 10 M s-1 29b
8
-1
(
29) The rate constants k ) 2.36 × 10
K.; Fessenden, R. W. J . Am.Chem. Soc. 1986, 108, 4706. (f) Schmidt,
J . A.; Hilinski, E. F. J . Am. Chem. Soc. 1988, 110, 4036.
(32) (a) Luckhurst, G. R.; Ockwell, J . N. Tetrahedron Lett. 1968,
4123. (b) Fox, M. A.; Gaillard, E.; Chen, C.-C. J . Am. Chem. Soc. 1987,
109, 7088. (c) Ruberu, S. R.; Fox, M. A. J . Phys. Chem. 1993, 97, 143.
(d) Fox, M. A.; Dulay, M. T.; Krosley, K. J . Am. Chem. Soc. 1994, 116,
10992; (e) J ohnston, L. J .; Scaiano, J . C. J . Am. Chem. Soc. 1986, 108,
2349.
(33) (a) Grellmann, K.-H.; Sherman, G. M.; Linschitz, H. J . Am.
Chem. Soc. 1963, 85, 1881. (b) Linschitz, H.; Grellmann, K.-H. J . Am.
Chem. Soc. 1964, 86, 303. (c) Lopez, D.; Boule, P.; Lemaire, J . Nouv.
J . Chem. 1980, 4, 615. (d) Gawinecki, R.; Boszczyk, W.; Rasala, D.;
Bak, T. J . Photochem. Photobiol. A; Chem. 1993, 71, 133.
(34) (a) Lewis, G. N.; Lipkin, D.; Magel, T. T. J . Am. Chem. Soc.
1944, 66, 1579. (b) Ting, L. C.; Weissmann, S. I. J . Chem. Phys. 1954,
22, 21.
•
•
are given for the reaction of O
respectively. (a) Maillard, B.; Ingold, K. U.; Scaiano, J . C. J . Am. Chem.
Soc. 1983, 105, 5095. (b) Howard, J . A.; Chenier, J . H. B.; Yamada, T.
2
with PhCH
2
and Ph
2
MeC radicals,
7
-1 -1
Can. J . Chem. 1982, 60, 2566. (c) A lower limit of 4.8 × 10
M
s
•
3 2
was found for the rate constant of the reaction of the Ph C with O :
Bodner, G. S.; Gladysz, J . A.; Nielsen, F. M.; Parker , V. D. J . Am.
Chem. Soc., 1987, 109, 7023. (d) Neckers, D. C.; Linden, S.-M.;
Williams, B. L.; Zakrzewski, A. J . Org. Chem., 1989, 54, 131. (e) Akaba,
R.; Kamata, M., Itoh, H.; Nakao, A.; Goto, S.; Saito, K.; Negishi, A.;
Saguragi, H.; Tokumaru, K. Tetrahedron Lett. 1992, 7011.
(
30) The amount of the imine increases on prolonged irradiation
(
conversion of 1a >30%, lamp irradiation) and evolution of hydrogen
gas was then observed, which continues (increasing linearly with the
irradiation time) even after all the starting material has been
consumed. It was also found that irradiation of the imine alone under
the same conditions also gives H
interesting phenomenon.
2
. We are currently investigating this
(35) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photo-
chemistry, 2nd ed.; Marcel Dekker: New York, 1993.

Photodissociation of N-(Triphenylmethyl)anilines
J . Org. Chem., Vol. 63, No. 10, 1998 3255
Sch em e 1
F igu r e 2. Absorption spectrum obtained on 308-nm LFP of
1
.1 mM 1a in MeCN under argon recorded at 280 ns after the
•
pulse. Inset: yield of DHPF , as a function of the laser power
I (full circles) and I (open circles); see the text.
2
ping absorptions due to anilino radical (λmax ≈ 400 nm,
•
see discussion below) and DHPF (absorption at 420-
•
•
5
20, λmax ) 486 nm). To remove some PhNH and DHPF ,
F igu r e 3. Absorption spectrum obtained on 308-nm LFP of
we measured in saturated oxygen solution (1d , MeCN)
the dosis dependence of the signal intensity of Ph
1
.7 mM 1a in hexane under argon b, 50 ns; O, 5 µs; and 2, 80•
+
3
C at
µs after the pulse. Insets are as follows: decay of DHPF
radical at 486 nm, and buildup of transient 2 at 377 nm.
4
38 nm (delay times of 62 and 370 ns after the pulse)
and found again a biphotonic process. In general we
observe saturation to take place starting with doses of
ca. 10 mJ /pulse.36
characteristics: first, the increased intensities at 377 and
4
86 nm and, second, the isosbestic points at 350 and 414
nm. The process leading to the absorptions at 337 and
86 nm was found (in both the 248- and 308-nm excita-
tions) to be clearly mono- and biphotonic, respectively.
A slightly different picture is obtained by irradiating
1
a with 308-nm laser light. While we observe again the
4
absorptions at 337 and 486 nm previously assigned to
•
•
Ph
3
C and the DHPF radicals, respectively, the broad
6
-1
The decay at 486 nm (kobs ) 1.4 ( 0.3 × 10
s ) is
+
signal at 400-440 nm corresponding to Ph
3
C is absent;
accompanied by the increase at 377 nm (kobs ) 1.7 ( 0.3
instead a very weak absorption at 377 nm appears (see
discussion later); in Figure 2 a representative spectrum
obtained by the photolysis of 1a is displayed. None of
6
-1
×
10 s ), both processes showing similar (dosis-depend-
ent) rate constants (laser dosis = 35 mJ /pulse, OD308
)
•
1
.03/cm) indicating conversion of DHPF radical to a
the decays of the three absorbances was affected by
species absorbing at 377 nm, see insets in Figure 3. The
77 nm absorption remains unaffected by saturation with
(regarding only its growth kinetics, while the yield is
-
addition of the strong nucleophile N
3
, indicating absence
3
•
3
of carbocations. The pathways leading to Ph C and
O
2
•
DHPF are again mono- and biphotonic processes, re-
spectively. In contrast to the 248-nm photolysis (see
above), the increased signal intensity at 486 nm (owing
to the stronger absorption of the trityl radical at 308 nm
than at 248 nm) facilitates the biphotonic processes.
•
clearly reduced as expected due to the reaction of DHPF
with O
2
) and, as mentioned before (in MeCN), also with
-
N
3
, which means that the unknown species is not of
cationic or radical nature. This intermediate is not the
final product 9-Ph-fluorene (see product analysis), which
absorbs at shorter wavelengths, but must be a neutral,
(
b) In n -Hexa n e a s Solven t. Laser excitation (248
nm) of 1a in hexane leads to an absorption spectrum
•
nonradical species arising by way of the DHPF radical
almost identical with that of Figure 2, indicating the
to the end product 9-Ph-fluorene.
Letsinger et al.27a discussed the possible involvement
•
•
formation of Ph C (λmax ) 337 nm), DHPF (λmax ) 486
3
nm), and the unknown intermediate with the very weak
absorption at ca. 377 nm.
of the cyclic polyene 9-Ph-4aH-fluorene 2 in the above
•
transformation, produced by H-abstraction from DHPF
Irradiation with 308-nm light produces the spectrum
displayed in Figure 3. It is similar to that obtained in
MeCN (see Figure 2), showing, however, two new
and subsequent aromatization (Scheme 1c), and Fox et
3
2b,c
al.
made similar assumptions for the case of the
photocyclization of perchlorinated analogues (perchlorot-
riphenylmethyl radical), not providing, however, any
definitive spectroscopic identification. The only com-
pound structurally related to 2 that we found in the
literature is the unstable hydrocarbon 4a-methyl-4aH-
(
36) For an analysis of the intensity dependence on product forma-
tion in laser flash photolysis experiments see: (a) Lachish, U.;
Schafferman, A.; Stein, G. J . Chem. Phys. 1976, 64, 4205. (b) Nikogo-
syan, D. N.; Angelov, D. A.; Oraevsky, A. A. Photochem. Photobiol.
1
982, 35, 627.

3
256 J . Org. Chem., Vol. 63, No. 10, 1998
Siskos et al.
F igu r e 5. Absorption spectrum obtained on 248-nm LFP of
2
.8 mM 9-Ph-fl-NHPh in MeCN under argon 170 ns after the
F igu r e 4. Absorption spectrum obtained on 248-nm LFP of
.3 mM PhNH in MeCN under argon recorded at 18 µs after
the pulse.
pulse.
0
2
•
25a
3
the early days of the history of Ph C its decoloration
in direct sunlight is reported, followed by its decomposi-
fluorene 33 (see Scheme 1) having λma ) 371 nm and
7a
26
tion to Ph
dimer of 9-Ph-fl ). The fact
radical leads photochemically to perchlorinated 9-Ph-fl ,
makes a closer investigation necessary.
We generated for this reason 9-Ph-fl by irradiating
3
CH and 9,9′-diphenyl-9,9′-difluorenyl (the
1
-1
ꢀ
371 ) 4900 M- cm in ethanol.
Meisel and his co-workers calculated for DHPF ꢀ480
•
32
32b,c
that also perchlorotrityl
3
1b
•
•
4
-1
-1
)
(1.5 ( 0.1) × 10 M cm . On the basis of this value
•
•
and assuming a clean transformation of DHPF to the
species absorbing at 377 nm, not unreasonable in view
of the observed isosbestic points at 350 and 414 nm, we
9-Ph-fl-NHPh in MeCN with 248-nm light, see Figure 5,
which we interpret in terms of reaction 2 occurring.
-
1
-1
arrive at a value of ꢀ377 ) 5100 ( 300 M cm for the
unknown species, very close to that of 3. We therefore
suggest that the intermediate absorbing at 377 nm is
compound 2.3 An allylic shift of a hydrogen atom of 2
accompanied by aromatization can explain the conversion
to the final product 9-Ph-fluorene (Scheme 1d).
A species which is missing in the above presentation
is the anilino radical PhNH . It must be the second
fragment generated by the photolytic C-N bond homoly-
sis, as also the product analysis (formation of PhNH , eq
2
hν
•
•
9
-Ph-fl-NHPh
9
8 9-Ph-fl + PhNH
(2)
7b
•
-
As expected, 9-Ph-fl is unaffected by N
surprise, also by O .
2
3
but, to our
3
8
Furthermore, we obtained the
same spectrum in the nonpolar solvent hexane, thus
eliminating the possibility of having produced the cor-
•
+
39
responding cation 9-Ph-fl , whose spectrum resembles
•
that of 9-Ph-fl , at least in the characteristic region of
4
50-500 nm. Moreover, the spectrum in Figure 5 is very
•
1
) indicates. PhNH is a well-known radical having
similar to the literature spectrum of the 9-mesitylfluo-
-1
-1
absorption maxima at λmax ) 308 nm (ꢀ ) 3500 M cm
)
4
0
renyl radical and to a spectrum assigned earlier by
-1
-1 15a
and 401 nm (ꢀ ) 1250 M cm
)
in water, not different
Lewis3 to 9-Ph-fl . Comparing the spectrum of 9-Ph-fl
4a
•
•
1
5b
15c
from those in hexane, cyclohexane,
heptane,
or
(
(
Figure 5) with that obtained by the photolysis of 1a
benzene/di-tert-butyl peroxide.1 In Figure 4 is shown
5d
•
Figures 1a and 2), formation of 9-Ph-fl by the photolysis
the absorption spectrum of the anilino radical obtained
•
of 1a can be excluded. The 9-Ph-fl absorptions in the
region 450-500 nm are ca. 10 nm red-shifted with
respect to those of DHPF and possess additional absorp-
by 248-nm laser photolysis12 of PhNH
in MeCN. It
possesses λmax ) 309 and 398 nm and shows an intensity
2
•
ratio in agreement with the mentioned reports.1
2,15
As-
tions below 300 nm (at 272 and 287 nm, a small
•
suming an equimolar formation of the radicals PhNH
and Ph C produced immediately after the laser pulse,
3
we must expect a signal intensity ratio 1:12 at 309 nm
and 1:32 at 398 nm with respect to the absorption at 337
nm, which corresponds to Ph
absorptions of the anilino radical at 309 and 398 nm are
very weak and overlap with those of the trityl radical at
37 nm and the trityl cation at 400-440 nm, respectively.
However, looking at the transient absorption spectra in
Figures 1a,b, 2, and 3 we see in every case a shoulder at
•
additional peak at 310 nm should belong to the PhNH
•
radical produced by the C-N bond homolysis, eq 2).
•
Moreover, the decay of DHPF increases by ca. 1 order of
3
5
2 2
magnitude in the presence of O ([O ] ) 9.1 mM), while
•
3
C . This implies that the
•
that of 9-Ph-fl remains unchanged. Additional support
for these observations arises from the fact that upon
•
irradiation of Ph
3
C in the cavity of an ESR spectrometer
3
(
see ESR studies below) its spectrum disappears, but is
•
not replaced by that of 9-Ph-fl .
. Qu a n tu m Yield s. In Table 2 are presented the
quantum yields for the formation of Ph C radical upon
3
4
3
09 nm which becomes clearer under oxygen (Figure 1a).
•
•
•
3 2
While Ph C reacts rapidly with O (see above), PhNH
2
48-nm irradiation of 1a -e in MeCN and other solvents.
1
5e
as an aminyl radical is less reactive
observable at longer delay times.
and becomes
21
These values were determined by measuring the initial
absorbance at λmax ) 334 nm (see Experimental Section).
•
3
. 9-P h -fl P a r ticip a tion ? A further point which
should be addressed with respect to the photoconversion
of the trityl radical is whether the 9-phenyl-9-fluorenyl
radical (9-Ph-fl ) and its dimer are involved. Even since
•
40
(
2
38) 9-Ph-fl reacts slowly with O and obviously in a time scale
longer than 100 µs used in our measurements.
(39) (a) Allen, D. M.; Owen, E. D. J . Chem. Soc. Chem. Commun.
971, 848. (b) McClelland, R. A.; Mathivanan, N.; Steenken, S. J . Am.
•
1
Chem. Soc. 1990, 112, 4857. (c) Cozens, F.; Robert, J . L.; McClelland,
R. A.; Steenken, S. Angew. Chem. Internat. Ed. 1992,104, 753; Van
Tamelen, E. E.; Cole, T. M., J r. J . Am. Chem. Soc. 1971, 93, 6158.
(40) Chandross, E. A.; Sheley J r, C. F. J . Am. Chem. Soc. 1968, 90,
4345.
(
37) (a) Neuhaus, D.; Rees, W. C. J . Chem. Soc. Chem. Commun.
983, 318. (b) An additional phenyl group at the 9-fluorenyl position
has no any appreciable effect on the UV absorption spectrum, because
1
4
0
it is substantially twisted out of the fluorene ring plane.

Photodissociation of N-(Triphenylmethyl)anilines
J . Org. Chem., Vol. 63, No. 10, 1998 3257
Ta ble 2. Qu a n tu m Yield s for th e F or m a tion of th e
nonpolar (hexane) solvents. In a second step trityl
radical is excited and forms the DHPF radical (in a
•
Ra d ica l P h 3C (u n d er Ar gon )
•
1
solvent
Φ(Ph
3
C•)
1
solvent
Φ(Ph
3
C•)
biphotonic process relative to the parent 1a , which is
more effective with 308- than with 248-nm light), in
agreement with the suggestion by Meisel and co-
a
a
a
b
MeCN
hexane
EtOH
0.65
0.71
0.61
0.67
c
d
e
MeCN
MeCN
MeCN
0.80
0.64 (0.61)a
0.57
31a,b
+
workers.
3
Moreover, we identified Ph C only in the
MeCN
polar medium MeCN and only by photolysis with 248-
nm laser light (biphotonic process relative to the 1a ).
To explain the results we propose the mechanistic
scheme outlined in Scheme 1. The easy photochemical
C-N bond homolysis in compounds 1a -e is not surpris-
a
Under oxygen.
The high quantum yields (Φ ≈ 0.6-0.8) indicate an
efficient homolytic photodissociation of the C-N bond of
the aniline compounds 1. The quantum yield for the
formation of Ph
in MeCN solution saturated with O
smaller value of 0.61 was found, not different, however,
than that under argon (0.64) within the experimental
error. This indicates that the dissociation proceeds faster
than the quenching of the photoexcited state of 1d with
ing if we consider the high excitation energy localized
•
3
C (photolysis of 1d ) was measured also
on the anilino chromophore
8a,10
either in its singlet (>90
2
, and only a slightly
kcal/mol) or triplet (>70 kcal/mol) state and the low
bond dissociation enthalpy (BDH) of the corresponding
10
10
9
bond C-N, which is estimated to be ∼39 kcal/mol. The
weakness of this bond dictates in fact the high efficiency
of the C-N bond homolysis and competes effectively with
4
1
O
2
.
It is noteworthy that the N-H photodissociation
in PhNH
proceeds to 80% via the triplet state.11
. ESR Stu d ies. To provide additional evidence for
the very effective C-N bond rupture and formation of
11
the common photochemical property of aromatic amines,
2
which is the dissociation of the N-H bond. The BDH of
the N-H bond in N-methylaniline, Ph(Me)N-H, which
5
9a
should be not very different from that in 1, is 83.3 kcal,
•
3
the radical Ph C upon irradiation of the N-tritylaniline
more than twice that of the corresponding C-N bond.
Thus, the bond scission of the N-H bond in 1a -e
derivatives 1, ESR spectroscopy was applied. Upon
irradiating a solution of 1 in hexane in the cavity of the
ESR spectrometer with a low-pressure mercury lamp (see
Experimental Section, main emission at 254 nm), we
(
Scheme 1 g) should be very unfavorable and was not
43
observed in the LFP experiments, a fact corroborated
by the ESR studies and the product analysis (absence
43
H
observe a spectrum [g ) 2.0026, a ortho(6H) ) 2.54 G,
of hydrogen gas evolution in hexane).
H
H
a
meta(6H) ) 1.13 G, a para(3H) ) 2.77 G] identical with
4
The high quantum yields for trityl radical formation
ranging from 0.57 to 0.8 (Table 2) do not reflect any
significant differentiation in the photodissociation abili-
ties of the various derivatives 1a -e, if one considers the
accuracy (10%) of the measurements. Although aniline-
ring substitution does not seem to affect the C-N bond
2b-e
•
the well-known
3
spectrum of the Ph C radical. It is
interesting that no signal was detected if we irradiated
with a high-pressure mercury lamp, which has strong
emissions at the wavelength of the maximum absorbance
of the trityl radical (337 nm). Obviously, a fast destruc-
tion of the radical takes place due to the above-described
photocyclization to 9-Ph-fluorene such that the steady-
state concentration is insufficient for an observable ESR
signal. A further confirmation of this conclusion was the
fact that upon producing Ph
nm light and subsequently changing the light source to
the high-pressure mercury lamp, the signal disappeared
rapidly. Here it is worth to note that we do not observe
44
strength, a small steric effect may be present due to
bulkiness of the trityl group in 1. However, the “propel-
ler”-shaped trityl group forces the C-N bond to approach
a perpendicular plane with respect to the aniline’s plane
•
3
C in the cavity with 254-
(as space-filling models show), and this is more obvious
in the case of the o-Me and m-Me substituents. The
latter has as a consequence a better hyperconjugative
interaction of the bond being broken (C-N) with the
•
any change of the signal to that of 9-Ph-fl , in line with
the conclusions (no formation of 9-Ph-fl ) reached above
by the LFP studies.
3
excited π-system and therefore (see introduction) a more
•
effective dissociation. The last point is related once again
3
to Michl’s ideas and the role of the small electronega-
These results represent a clear demonstration of how
easily erroneous conclusions can be drawn about the
involvement of free radicals in photochemical processes
when the photochemical behavior of the radicals them-
selves is unknown and when taking into account that
photolysis is one of the most common methods for
generating free radicals.
tivity difference (δCN) in promoting homolysis (in 1a -e,
Figure 1a,b), as contrasted by Ph C-Cl (Figure 1c),
3
where heterolysis of the C-Cl bond is the dominant
process (increased δCCl).
We believe the data do not allow one to consider the
12a
effect described by Malkin,
where o-Me, m-Me, and
p-Me aniline derivatives have a more effective isc (S
2
f
Here we note that 1 dissociates thermally (decalin/205
T
8
) and a higher quantum yield on photodissociation of
•
17,18
°
C) to Ph
3
C (ESR).
their N-H bond with respect to unsubstituted aniline,
as related to our results; the deviations in the measured
quantum yields (Table 2) are too insignificant to account
for such an effect.
6
. Mech a n istic Con sid er a tion s. In summary, we
have shown that the aniline derivatives 1a -e undergo
very efficient (see Table 2) homolysis on irradiation with
•
•
2
48- or 308-nm laser light, leading to Ph
3
C and PhNH
(
43) In the case of a N-H photohomolysis we would expect the
in a monophotonic pathway in both polar (MeCN) and
•
formation of a nitrogen-centered radical (Ph
Such radicals should give ESR spectra with different g-values and
hyperfine splittings than Ph C : Siskos, M. G.; Zarkadis, A. K.
3
3
C-N -Ph) (Scheme 1g).
•
(
41) Similar cases are known in the literature, see for example:
Mizuno, K.; Ichinose, N.; Otsuji, Y. J . Am. Chem. Soc. 1985, 107, 5797.
42) (a) Gomberg, M. Ber. Dtsch. Chem. Ges. 1900, 33, 3150. (b)
unpublished results. Moreover, the absorption spectrum of the radical
•
15b,e
(
Me-N -Ph displays λmax at 314 and 400 nm,
different than we
Maki, A. H.; Allendoerfer, R. D.; Danner, J . C.; Keys, R. T. J . Am.
Chem. Soc. 1968, 90, 4225. (c) Sinclair, J .; Kivelson, D.; J . Am. Chem.
Soc. 1968, 90, 5075. (d) Kulkarni, S. V.; Trapp, C. J . Am. Chem. Soc.
observe on the photolysis of 1a -e.
(44) Ring substitution in aniline does not affect its N-H bond
strength: (a) J onsson, M.; Lind, J .; Eriksen, T. E.; Merenyi, G. J . Am.
Chem. Soc. 1994, 116, 1423. (b) J onsson, M.; Lind, J .; Merenyi, G.
Eriksen, T. E. J . Chem. Soc. Perikin Trans.2 1995, 67.
1
970, 92, 4801. (e) Neumann, W. P.; Uzick, W.; Zarkadis, A. K. J . Am.
Chem. Soc. 1986, 108, 3762.

3
258 J . Org. Chem., Vol. 63, No. 10, 1998
Siskos et al.
•
31
The formation of DHPF (Scheme 1b) is rationalized
in terms of an electrocyclic ring closure of the excited
•
Ph
3
C , resulting from an increased twist angle of the
phenyl rings out of the central molecular plane, in
•
contrast to Ph
2
CH , where no intramolecular photochem-
6
b,d,31,45
istry was observed.
to the DHPF that we observed in the LFP experiments
is consistent with this proposal. DHPF decays rapidly
The biphotonic process leading
•
•
to the cyclic polyene hydrocarbon 2, probably through a
•
•
hydrogen atom abstraction by the PhNH or Ph
Scheme 1c). This is probably a very facile process since
the Caliph-H bond strength of the cylohexadienyl radical,
3
C radical
(
•
46
a simplified analogue to DHPF , is only ca. 25 kcal/mol,
F igu r e 6. Absorption spectrum obtained on pulse radiolysis
of 2.8 mM 1a in n-BuCl under argon 600 ns after the pulse.
considerably weaker than the BDH of the aniline N-H
4
4,47
bond (89.1 kcal/mol)
or of the triphenylmethane C-H
bond (75 kcal/mol).8
c,47
This hydrogen atom abstraction
C+ and PhNH- would be a monopho-
3
sis leading to Ph
tonic process.
Second, the photoionization of Ph C to Ph C in water
3 3
is probably the reason for the relatively shorter decay
•
+
time (<10 µs) obtained in our LFP experiments for
•
31b
DHPF versus those of Meisel (>100 µs).
Fox and co-
with 248- or 308-nm laser light is a well-documented
workers3 found for the perchlorinated-DHPF ∼22 µs,
2b
•
6a
process, consuming another photon and making the
•
which is, however, more stable than DHPF . The iden-
overall process biphotonic, as we also observed in MeCN
(Scheme 1e). The fact that we do not see carbocation
formation upon irradiating with 308-nm (4.03 eV) in
MeCN can be attributed to the lower solvation power of
tification of the intermediate 2 (see above) is the first
strong evidence for the electrocyclization mechanism of
•
the excited Ph
3
C . Concomitant rearrangement of 2
+
50
(Scheme 1d) to the final product 9-Ph-flH through an
Ph
3
C
in MeCN than in H
2
O.
In contrast, the 5 eV
allylic shift and aromatization is a well-known reaction
of such cyclohexadienyl systems, catalyzed by acids,
energy provided by the 248-nm photons compensates for
the missing energy and photoionization takes place.
+
bases, or via free radicals.2
7a,c,48
Another conceivable
From this point of view, it is reasonable that Ph C was
3
•
not observed in hexane, neither with 248- nor with 308-
possibility of transforming 2 to 9-Ph-fl (Scheme 1f) could
be excluded (see above), and this radical could not be
identified either by LFP or by ESR.
nm light. Meisel and co-workers3 also failed to see
1b
carbocation formation upon irradiating with 347-nm laser
light (3.57 eV) in MeCN and cyclohexane.
Now, it remains to address the occurrence of the trityl
+
+
A third way to produce Ph
3
C
is, however, the photo-
cation Ph
3
C . There exist at least three possible paths
+
ionization of the parent molecule 1a to the anilinium
3
for its formation; the first one is the formation of Ph C
•
+
radical cation 1a (Scheme 1i) with subsequent frag-
through a direct photoheterolysis of the C-N bond
+
•
Scheme 1h). In the literature2a,5,6b,7,8 there is a long-
mentation to give Ph
3
C and PhNH (Scheme 1j). Actu-
(
ally, photoionization of aromatic amines with concomi-
lasting discussion on the very early processes and the
nature of the formed intermediates after excitation of a
substrate and particularly about the competition between
homolysis and heterolysis. In the case of compounds 1,
although the heterolytic C-N bond dissociation enthalpy
1
1
tant fragmentation is a very common property owing
to their low ionization51 or oxidation potentials. While
direct photoionization is favored in polar media (usually
in water)52 and is less studied, the sensitized photoin-
duced electron transfer in amines is currently a very
44
4
9
is estimated to be lower (∼13 kcal/mol) than the
5
3
active area of research. To clarify the possibility of the
9
homolytic one (∼39 kcal/mol), we found a biphotonic
•
+
participation of 1a , a pulse radiolysis study of 1a was
+
process leading to the formation Ph
3
C ; a direct heteroly-
undertaken and we obtained the absorption spectrum
shown in Figure 6, which coincides with the spectrum of
(
45) Arnold, B. R.; Scaiano, J . C.; McGimpsey, W. G. J . Am. Chem.
+
Ph
3
C . Also it is identical with the spectrum obtained
Soc. 1992, 114, 9978. Scaiano, J . C.; Tanner, M.; Weir, D. J . Am. Chem.
Soc. 1985, 107, 4396.
5
4
by irradiation of 1a in MeCN/CCl
irradiation with 254-nm light (lamp) failed to give an
ESR signal (hexane/CCl ). Apparently, here (pulse ra-
4
(248-nm laser), while
(
46) This derives from the enthalpy liberated (-25 kcal/mol) by the
formation of the cyclohexadienyl radical via hydrogen atom addition
to benzene.9
b,49d
See also: Schaffer, F.; Beckhaus, H.-D.; Rieger, H.-J .;
4
R u¨ chardt, C. Chem. Ber. 1994, 127, 557. Gerst, M.; Morgenthaler, J .;
R u¨ chardt, C. Chem. Ber. 1994, 127, 691.
(50) (a) For a critical discussion of the role of the solvation energy
in heterolytic dissociations and photoionizations, see ref 6a,b and
references therein. (b) Reichardt, C. Solvents and Solvent Effects in
Organic Chemistry; Verlag Chemie, Weinheim, 1990; p 31.
(51) Farrell, P. G.; Newton, J . J . Phys. Chem. 1965, 69, 3506.
Vedeneyev, V. I.; Gurvich, L. V.; Kondrat'yev, V. N.; Medvedev, V. A.;
Frankevich, Ye. L. Bond Energies, Ionization Potentials and Electron
Affinities; Edward Arnold Publ.: London, 1966; p 170.
(52) K o¨ hler, G.; Getoff, N. J . Chem. Soc., Faraday Trans.1 1980,
76, 1576.
(53) (a) Gaillard, E. R.; Whitten, D. G. Acc. Chem. Res. 1996, 29,
292, and references therein. (b) Zhang, X.; Yeh, S.-R.; Hong, S.;
Freccero, M.; Albini, A.; Falvey, D. E.; Mariano, P. S. J . Am. Chem.
Soc. 1994, 116, 4211.
(
47) Increased values for the BDH of aniline N-H (92.0 kcal/mol)
and for C-H (81 kcal/mol) of Ph CH were reported: Bordwell, F. G.;
Cheng, J .-P.; J i, G.-Z.; Satish, A. V.; Zhang, X. J . Am. Chem. Soc. 1991,
13, 9790.
48) Tzerpos, N. I.; Zarkadis, A. K.; Kreher, R. P.; Repas, L.; Lehnig,
M. J . Chem. Soc. Perikin Trans. 2 1995, 755.
49) On the basis of the calculated C-N homolytic BDH (39 kcal/
3
1
(
(
9
-
mol) for Ph
in DMSO (vs ferrocene/ferrocinium ion)
V in sulfolane (vs ferrocinium/ferrocene couple),
3
C-NHPh, the redox potentials Eox(PhNH ) ) -0.992 V
47,49a +
and E1/2(Ph
3
C ) ) -0.133
and using Wayner
and Parker’s method, we calculate ∼13 kcal/mol for the heterolytic
bond dissociation enthalpy of Ph
4
9b,c
4
9d
3
C-NHPh. One should note here that
this value is not in the solvent used (MeCN); see ref 51b for a comment
about the uncertainities in using redox potentials obtained in DMSO,
4
(54) Photolysis of aromatic amines in CCl leads to primary forma-
5
4a,b
•-
H
2
O, MeCN, sulfolane. (a) Bordwell, F. G. Acc. Chem. Res. 1993, 26,
tion of radical cations on nitrogen,
whereas CCl
4
decomposes in
5
4c
5
10. (b) Arnett, E. M.; Venimadhavan, S.; Amarnath, K. J . Am. Chem.
τ < 18 ps. (a) Budyka, M. F.; Alfimov, M. V. Russ. Chem. Rev. 1995,
64, 705, and literature cited therein. (b) Shimamori, H.; Musasa, H.
J . Phys. Chem. 1995, 99, 14359. (c) Chateauneuf, J .; Lusztyk, J .; Ingold,
K. U. J . Org. Chem. 1990, 55, 1061.
Soc. 1992, 114, 5598. (c) Arnett, E. M.; Flowers III, R. A. Chem. Soc.
Rev. 1993, 9. (d) Wayner, D. D. M.; Parker, V. D. Acc. Chem. Res. 1993,
2
6, 287. (e) Saveant, J .-M. Acc. Chem. Res. 1993, 26, 455.

Photodissociation of N-(Triphenylmethyl)anilines
J . Org. Chem., Vol. 63, No. 10, 1998 3259
MeCN.6d,57 Summarizing, N-benzylaniline can be ionized
even in MeCN with 308-nm photons, and this should hold
also for N-tritylanilines 1 which must have ionization
potentials not higher than N-benzylaniline. This makes
possible, at least it does not exclude, the formation of the
•
+
trityl cation through dissociation of 1a
.
Now, the fact that we do not see any Ph
3
C+ by the
photolysis of 1a under the same conditions (308-nm/
MeCN, see above) indicates probably that, although 1a
•
+
is able to ionize in MeCN (308-nm) to 1a , this is not
+
the way leading to Ph
3
C , but the fast dissociation of the
•
•
F igu r e 7. Absorption spectrum obtained on 248-nm LFP of
3
excited state of 1a produces preferable Ph C and PhNH
0
.1 mM PhCH
2 2
NHPh in H O/2-PrOH (4:1) under argon 300
(
Scheme 1a) and in a second step trityl radical undergoes
ns after the pulse.
+
photoionization to Ph
3
C
(Scheme 1e). This is presum-
ably a consequence of the consecutive biphotonic nature
of the photoionization of these aniline derivatives; the
excited state of 1a is, because of the fast homolysis
•
+
diolysis) the radical cation 1a fragmentates quickly to
give Ph
+
•
3
C and PhNH (Scheme 1j) and can therefore not
be observed in the time scale of our pulse radiolysis
(Scheme 1a), too short living to be ionized by a second
•
+
experiments (g500 ns). The C-N bond in 1a is very
weak (ca. -3 kcal/mol)55 and apparently a very fast
photon, and this in contrast to benzylaniline, where its
triplet state absorption is longer living and observable
in the LFP experiment.58
dissociation makes its detection (if it is formed) impos-
sible. Nevertheless, these two experiments show that if
photoionization of 1 takes place, then formation of the
trityl carbocation occurs.
Electron transfer from Ph C to PhNH• leading to
•
3
+
-
59
Ph C and PhNH , although thermochemically feasible,
3
Direct photoionization of aromatic amines is facilitated
would be a monophotonic process for the formation of
in water.1
1,52
Therefore we irradiated (248 nm) 1a in a
O (3:4) and we recorded the same
+
Ph
3
C
and is thus also eliminated as a possible way.
mixture i-PrOH/H
2
7
. Su m m a r y. The present work describes an effective
spectrum as in MeCN (see Figure 1a,b) with an ad-
ditional weak broad peak at 500-800 nm which can be
and selective homolytic photodissociation of the C-N
bond in N-tritylanilines 1, leading exclusively to the
efficiently removed by O
2
and N
2
O and is therefore
•
assigned6
a,d
-
-
formation of Ph
contrast to Ph
3
C (high quantum yields). This is in
to e aq
.
The formation of e aq obeys a
3
C-Cl, where homolysis and heterolysis
biphotonic pathway, as in all the cases outlined above
+
occur in a ratio 1:1.4, respectively, in line with Michl’s
where Ph
whether the electron is photoejected from Ph
e) or from 1a (Scheme 1i,j). The radical cation 1a is
3
C
was produced, so we still cannot decide
•
theoretical analysis. In addition, a detailed mechanistic
3
C (Scheme
•
•
+
scheme is proposed for the photoionization of Ph C to
1
3
+
Ph
3
C and its photoconversion to 9-Ph-fluorene (MeCN).
also here undetectable, despite the much shorter time
scale of the LFP experiment (g20 ns).
To examine if compounds of the type 1 are able to
undergo photoionization in our LFP experiments, we
A new indermediate (9-Ph-4aH-fluorene 2) going from
•
Ph
3
C to the final product 9-Ph-fluorene is identified for
the first time (hexane) and supports the mechanism
proposed earlier.3
0,34
Furthermore, because the trityl
chose to study N-benzylaniline, PhCH
2
-NHPh, which is
structurally analogous to 1 but possesses a stronger C-N
group belongs to the classical amino-protecting groups
in the synthesis of polyfunctional molecules and particu-
9
e
bond (∼56 kcal/mol), making the corresponding radical
cation (PhCH NHPh) more persistent than 1a . In-
deed, irradiating PhCH -NHPh under the same condi-
tions (248-nm/ i-PrOH/H O), we observed the spectrum
shown in Figure 7. The broad absorption at 440 nm is
assigned to the corresponding radical cation (PhCH
•
+
•+
60
2
larly in peptide synthesis, its potential function as
61,62
2
photodeprotective
group is, after the above results,
2
of great interest.
2
-
•+
Ack n ow led gm en t. We thank Bundesministerium
f u¨ r Forschung und Technologie (Germany) and General
Secretariat for Research and Technology (Greece) for a
grant (Grant No. 6BOA1A) and Dr. K. Hildenbrand and
Dr. C. Borsarelli (MPI f u¨ r Strahlenchemie, M u¨ lheim)
for reading the draft.
NHPh) on the basis of its nonreactivity with oxygen and
5
6
its absorption spectrum (obtained via sensitized pho-
toinduced electron transfer). The strong absorption at
00-800 nm is due to e-aq, which is produced in a
5
biphotonic process. However, when we irradiated PhCH -
2
NHPh in MeCN with 308-nm laser light, we observed
the same spectrum with λmax at 440 nm corresponding
to (PhCH NHPh) , and once again we measured for its
2
J O9719261
•
+
formation a biphotonic process. The absorption spectrum
of the solvated electron e aq cannot be observed in
-
(57) Hirata, Y.; Mataga, N.; Sakata, Y.; Misumi, S. J . Phys. Chem.
983, 87, 1493; Hirata, Y.; Takimoto, M.; Mataga, N. Chem. Phys. Lett.
983, 97, 569.
(58) Siskos, M. G.; Zarkadis, A. K.; Steenken, S. Unpublished
results.
1
1
(
55) The radical cation 1a •+ would be also very instable, at least
5
5a
judging from the ca. -3 kcal/mol BDH of its C-N bond. (a) Taking
approximately the oxidation potential of the N-methylaniline (∼0.7 V
(59) On the basis of the redox potentials of the trityl carbocation
vs SCE),5 which is between that of aniline and dimethylaniline,
5b
44a
and anilino anion, we find an exergonic conversion (ca. -26 kcal/
49
to be the “same” as that of 1a , the calculated BDH of the C-N bond in
mol).
9
49d,e
1
a , and the appropiate thermochemical cycle,
we arrive at a BDH
(60) Zervas, L.; Theodoropoulos, D. M. J . Am. Chem. Soc. 1956, 78,
1359; Barlos, K.; Gatos, D.; Kallitsis, J .; Papaphotiou, G.; Sotiriu, P.;
Wenqing, Y.; Sch a¨ fer, W. Tetrahedron Lett. 1989, 30, 3947.
(61) Schelhaas, M.; Waldmann, H. Angew. Chem. 1996, 108, 2192.
(62) Pillai, V. N. R. Org. Photochem. 1987, 9, 225.
•
+
of ca. -3 kcal/mol for 1a . (b) Galus, Z.; Adams, R. N. J . Phys. Chem.
1
963, 67, 862.
56) Lucian, A. L.; Burton, R. D.; Schanze, K. S. J . Phys. Chem. 1993,
7, 9078.
(
9