Charge-transfer-induced photoreduction of azoalkanes by amines

Article abstract of DOI:10.1021/ja9706460

The unique DBH-type azoalkanes 1, which exhibit high intersystem crossing quantum yields, have made possible the exploration of the bimolecular photoreduction of the n,π* triplet-excited azo chromophore. In the laser-flash photolysis, amines were found to

Full text of DOI:10.1021/ja9706460

J. Am. Chem. Soc. 1997, 119, 6749-6756
6749
Charge-Transfer-Induced Photoreduction of Azoalkanes by
Amines
Waldemar Adam,*,† Jarugu N. Moorthy,† Werner M. Nau,‡,§ and J. C. Scaiano*,‡
Contribution from the Institut fu¨r Organische Chemie, UniVersita¨t Wu¨rzburg, Am Hubland,
D-97074 Wu¨rzburg, Germany, and the Department of Chemistry, UniVersity of Ottawa,
10 Marie Curie, Ottawa, Ontario K1N 6N5, Canada
ReceiVed February 28, 1997X
Abstract: The unique DBH-type azoalkanes 1, which exhibit high intersystem crossing quantum yields, have made
possible the exploration of the bimolecular photoreduction of the n,π* triplet-excited azo chromophore. In the laser-
flash photolysis, amines were found to quench the triplet azoalkane 1a with high rate constants (kq ca. 108 M-1 s-1).
Steady-state photolysis of the azoalkanes 1a and 1b (φISC ca. 0.5) in the presence of primary, secondary, and tertiary
aliphatic amines gave high chemical yields of the corresponding hydrazines 4a and 4b in competition with the
unimolecular products, namely the housanes 2 and the aziranes 3. In contrast, the azoalkane 1c undergoes appreciable
photoreduction only in neat amines, while the azoalkane 1d (φISC ca. 0.10) is not reduced even under such conditions.
Except for N,N-dimethylbenzylamine, the amine oxidation products of the azoalkane photoreduction are analogous
to those obtained from the reactions of amines with triplet benzophenone. In marked contrast, the absolute quantum
yields of photoreduction for azoalkanes 1 are substantially lower (0.01-0.06) than for benzophenone (0.3-1.0).
Efficient deactivation of the triplet-excited states by charge-transfer (kCq T), which competes with hydrogen atom
abstraction (kCHT), is postulated to account for the low quantum yields. The efficiencies of photoreduction follow the
trend primary ≈ tertiary . secondary amines observed with benzophenone, for which secondary amines also display
the poorest efficiency. Electron transfer to triplet-excited azoalkanes, analogous to benzophenone, is observed for
amines with low oxidation potentials. Thus, when triphenylamine (Eox ) 0.85 V Versus SCE) is used, the formation
of its radical cation can be readily detected by laser-flash photolysis.
Introduction
The DBH-type azoalkanes 1 (Scheme 1) are unique in that
the geometric features intrinsic to the bicyclic skeleton promote
efficient ISC and, thus, represent the first examples for which
the elusive triplet states have been recently spectroscopically
characterized.5 For example, the cyclic azoalkane 1a undergoes
efficient intersystem crossing to the triplet state with φISC ca.
0.5, whose lifetime has been determined from laser-flash
spectroscopic studies to be 0.63 µs. Therefore, the azoalkanes
1 offer at last the opportunity to explore the bimolecular
reactivity of the n,π* triplet-excited azo chromophore. Com-
parison with the established n,π* triplet photochemistry of the
carbonyl chromophore would enhance our knowledge of the
triplet-state reactivity.
Bicyclic azoalkanes are useful sources of biradicals and even
polyradicals.1 Their photoreactivity is markedly influenced by
the geometrical constraints imposed on the azo chromophore
through the bicyclic ring skeleton in which the NdN linkage is
incorporated. This is dramatically revealed by the two repre-
sentative homologous bicyclic azoalkanes, namely 2,3-diaza-
bicyclo[2.2.1]hept-2-ene (DBH) and 2,3-diazabicyclo[2.2.2]oct-
2-ene (DBO). In general, whereas the DBH derivatives extrude
molecular nitrogen with a very high efficiency (φN ca. 1 for
2
the parent DBH),1a,2 the DBO derivatives display significantly
lower reactivity (φN ca. 0.01 for the parent DBO). The latter
2
dissipate their excitation energy predominantly by fluorescence
or radiationless deactivation.3 Thus, deazatation, fluorescence,
and radiationless deactivation compete efficiently with inter-
system crossing (ISC) from the singlet-excited state to the triplet
manifold. As a result, the demonstration of bimolecular
reactivity of the n,π* triplet-excited azo chromophore, as has
been extensively documented for n,π* carbonyl triplets, has
remained heretofore a formidable task. To date only one
bimolecular reaction of triplet-excited azoalkanes, generated by
sensitization, has been reported.4
The photoreduction of ketones by amines has attracted
considerable mechanistic interest.6 Subsequent to the seminal
investigations by Cohen and co-workers, a comprehensive
picture on the dynamics of ketone photoreduction was provided
by Peters et al. through laser-flash photolysis studies.7 Ac-
cordingly, electron transfer from the amine to the benzophenone
(4) Engel, P. S.; Keys, D. E.; Kitamura, A. J. Am. Chem. Soc. 1985,
107, 4964.
(5) (a) Adam, W.; Nau, W. M.; Sendelbach, J.; Wirz, J. J. Am. Chem.
Soc. 1993, 115, 12571. (b) Adam, W.; Nau, W. M.; Sendelbach, J. J. Am.
Chem. Soc. 1994, 116, 7049. (c) Adam, W.; Fragale, G.; Klapstein, D.;
Nau, W. M.; Wirz, J. J. Am. Chem. Soc. 1995, 117, 12578.
(6) (a) Cohen, S. G.; Parola, A.; Parsons, G. H. Chem. ReV. 1973, 73,
141. (b) Inbar, S.; Linsschitz, H.; Cohen, S. G. J. Am. Chem. Soc. 1981,
103, 1048. (c) Davidson, R. S.; Steiner, P. R. J. Chem. Soc., Perkin Trans.
1 1972, 1375. (d) Pienta, N. J.; McKimmy, J. E. J. Am. Chem. Soc. 1982,
104, 5501. (e) Dunn, D. A.; Schuster, D. I.; Bonneau, R. J. Am. Chem.
Soc. 1985, 107, 2802. (f) Xu, W.; Jeon, Y. T.; Hasegawa, E.; Yoon, U. C.;
Mariano, P. S. J. Am. Chem. Soc. 1989, 111, 406. (g) Inbar, S.; Linsschitz,
H.; Cohen, S. G. J. Am. Chem. Soc. 1980, 102, 1419. (h) Davidson, R. S.;
Oroton, S. P. J. Chem. Soc., Chem. Commun. 1974, 209. (i) Griller, D.;
Howard, J. A.; Marriott, P. R.; Scaiano, J. C. J. Am. Chem. Soc. 1981, 103,
619. (j) von Raumer, M.; Suppan, P.; Haselbach, E. Chem. Phys. Lett. 1996,
253, 263.
† Universita¨t Wu¨rzburg.
‡ University of Ottawa.
§ Present address: Institute of Physical Chemistry, University of Basel,
Klingelbergstrasse 80, CH-4056 Basel, Switzerland.
X Abstract published in AdVance ACS Abstracts, July 1, 1997.
(1) (a) Engel, P. S. Chem. ReV. 1980, 80, 99. (b) Dougherty, D. Acc.
Chem. Res. 1991, 24, 88. (c) Adam, W.; Barneveld, C.; Bottle, S. E.; Engert,
H.; Hanson, G. R.; Harrer, H. M.; Heim, C.; Nau, W. M.; Wang, D. J. Am.
Chem. Soc. 1996, 118, 3974.
(2) Clark, W. D. K.; Steel, C. J. Am. Chem. Soc. 1971, 93, 6347.
(3) (a) Berson, J. A.; Davis, R. F. J. Am. Chem. Soc. 1972, 94, 3658. (b)
Engel, P. S.; Kitamura, A.; Keys, D. E. J. Am. Chem. Soc. 1985, 107, 4964.
(c) Engel, P. S.; Culotta, A. M. J. Am. Chem. Soc. 1991, 113, 2686.
S0002-7863(97)00646-X CCC: $14.00 © 1997 American Chemical Society

6750 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Adam et al.
Scheme 1
triplet forms the radical anion and amine radical cation as a
triplet contact ion pair (CIP). The latter may subsequently
undergo solvation to the solvent-separated ion radical pair (SSIP)
and proton transfer to form the neutral radicals or ISC followed
by electron back-transfer to regenerate the starting materials.
Fluorescence quenching of DBO and of a DBH derivative
by amines is precedented.8 In a preliminary study9 we have
found that the triplet absorption of the azoalkane 1a, in analogy
to that of ketones, is quenched by a variety of amines with
relatively high rate constants. In view of the high rate constants
with which amines quench the triplet state of the azoalkane 1a,
we were motivated to examine if these triplet states undergo,
in analogy to the ketones, bimolecular photoreduction in the
presence of amines.9 We herein demonstrate that amine
quenching of the azoalkanes 1 affords efficiently the corre-
sponding hydrazines, in competition to deazatation. The
mechanism of photoreduction and the relative efficiencies for
primary, secondary, and tertiary amines are contrasted with those
of the extensively investigated prototypal carbonyl chromophore,
namely benzophenone.
of the characteristic 1H NMR signals with those observed upon
photolysis of azoalkanes 1a-c in the presence of an efficient
hydrogen donor such as 1,4-cyclohexadiene. The photolysis
of azoalkanes 1 in 1,4-cyclohexadiene afforded quantitatively
the hydrazoalkanes 4, which for isolation and characterization
were trapped as the derivatives 5 (eq 1) by treatment with ethyl
chloroformate and triethylamine. Furthermore, exposure of the
photolysates to air over a period of 2-3 days regenerated the
azoalkanes 1 by oxidation of the hydrazines 4, as monitored by
1H NMR analysis.10 The product distributions for the photolysis
of the azoalkanes 1 in the presence of a variety of amines are
summarized in Table 1.
The azoalkanes 1a and 1b afforded the hydrazines 4a and
4b in moderate to high yields when photolyzed in the presence
of primary, secondary, and tertiary amines (Table 1). In
contrast, at similar amine concentrations (up to 1.0 M), the
azoalkane 1c produced the hydrazine 4c only in low yields
(<10%, data not shown); however, photolysis in neat amines
resulted in much higher yields of the hydrazine (entries 26-
28). Similarly, a substantial enhancement in the hydrazine
yields was found when azoalkane 1b was photolyzed in neat
amines (entries 13, 17, and 23). The azoalkane 1d gave the
housane 2d exclusively even at high concentrations of the
amines (entries 29-31). The poor solubility of azoalkane 1d
precluded its photolysis in neat amines. Most significantly, tert-
butylamine and diphenylamine were found to be ineffective in
the photoreduction of azoalkanes 1a and 1b (entries 3, 6, 15,
and 19). It is noteworthy that no coupling products between
the azoalkanes and amines were observed.
Results
Photoreduction by Amines. The photochemistry of the
azoalkanes 1 has been reported previously.5a,b Whereas the
direct photoexcitation of these azoalkanes gave a mixture of
housanes 2 and aziranes 3 (CN and CC bond cleavage products,
Scheme 1), the housanes 2 were exclusively formed in the
photolysis of the azoalkanes 1c and 1d. The sensitized
photolysis has been reported to afford exclusively the aziranes
3.5a,b
Photoexcitation of the nitrogen-purged solutions of the
azoalkanes 1a-c (ca. 0.1 M) and amines (0.5-1.0 M) led to
hydrazines 4 together with the unimolecular products 2 and 3,
as monitored by 1H NMR spectroscopy (Scheme 1). The
identity of the hydrazoalkanes 4 was established by comparison
(7) (a) Simon, J. D.; Peters, K. S. J. Am. Chem. Soc. 1982, 104, 6542.
(b) Schaefer, C. G.; Peters, K. S. J. Am. Chem. Soc. 1980, 102, 1566.
(8) (a) Evans, T. R. J. Am. Chem. Soc. 1971, 93, 2081. (b) Engel, P. S.;
Kitamura, A.; Keys, D. E. J. Org. Chem. 1987, 52, 5015. (c) Ramamurthy,
V.; Turro, N. J. Ind. J. Chem. 1979, 18 B, 72.
Oxidation Products of the Amines. 1H NMR analysis of
the photolysates derived from the azoalkanes 1a and 1b with
(9) Adam, W.; Moorthy, J. N.; Nau, W. M.; Scaiano, J. C. J. Org. Chem.
1996, 61, 8722.
(10) Newbold, B. T. In The Chemistry of Hydrazo, Azo and Azoxy
Groups; Patai, S., Ed.; Wiley: Chichester, 1975; p 541.

Photoreduction of Azoalkanes by Amines
J. Am. Chem. Soc., Vol. 119, No. 29, 1997 6751
Table 1. Product Studiesa for the Photolysis of Azoalkanes 1 in the Presence of Amines
product distributionb
entry
azo
amine
concn (M)
solvent
conversion (%)
mass balance (%)
2 + 3
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1a
n-butylamine
1.0
0.9
1.0
1.0
0.7
0.3
0.6
0.6
0.6
0.6
0.6
1.0
10.1
0.9
1.0
1.0
9.6
0.7
0.3
0.6
0.6
0.6
7.2
0.6
0.6
10.1
9.7
7.2
7.0
7.0
5.0
C6D6
C6D6
C6D6
C6D6
C6D6
CD3CN
C6D6
CD3CN
CD3OD
C6D6
CD3CN
C6D6
neat
C6D6
C6D6
C6D6
neat
C6D6
CD3CN
C6D6
CD3CN
CD3OD
neat
71
72
100
76
66
25
76
73
97
89
100
88
84
95
100
94
76
82
40
95
90
95
90
94
90
95
60
80
75
90
90
93
90
90
90
95
90
90
95
75
90
89
90
90
90
>90
>90
>90
90
90
>90
29
37
100
20
38
100
40
33
34
18
11
62
10
55
100
28
10
45
100
61
41
39
10
30
28
50
25
31
100
100
100
71
63
cyclohexylamine
tert-butylamine
diethylamine
diisopropylamine
diphenylamine
triethylamine
80
62
60
67
66
82
89
38
90
45
Me2NCH2Ph
1b
n-butylamine
cyclohexylamine
tert-butylamine
diethylamine
72
90
55
diisopropylamine
diphenylamine
triethylamine
39
59
61
90
70
72
50
75
69
87
99
84
83
Me2NCH2Ph
C6D6
CD3CN
neat
neat
neat
C6H6
C6H6
C6H6
80
1c
n-butylamine
diethylamine
triethylamine
n-butylamine
diethylamine
triethylamine
100
100
100
100
100
100
1d
a A ca. 0.1 M solution of the azoalkane 1 was irradiated (λ ) 350 nm) in the Rayonet photoreactor at 8 °C for 15 h, except for entries 4 and 15
1
(ca. 40-44 h). b Analysis by H NMR spectroscopy; relative yields are normalized to 100%, error limit (6% of the stated values.
primary and secondary amines revealed that the oxidation
products were similar to those formed in the photoreduction of
benzophenone by amines,6a,11 as confirmed by comparison with
the authentic materials. The stoichiometry of the photoreduc-
tions (eqs 2 and 3) were deduced from the relative areas of the
the hydrolysis of the enamine (eq 4) by adventitious water.
Because of the volatile nature of diethylamine, the measured
yields constitute a lower limit. The photolysis in the presence
of N,N-dimethylbenzylamine (Table 1) yielded the dimers as
the sole products according to the stoichiometry in eq 5, as de-
duced from 1H NMR analysis. Thus, the pathway for the forma-
tion of hydrazine appears to be the disproportionation of the
intermediate hydrazinyl radicals, while the corresponding (N,N-
dimethylamino)benzyl radicals undergo dimerization (cf. Discus-
sion). It is noteworthy in this context that aminobenzyl radicals
undergo such dimerization in the photoreduction of aryl imines.12
Quantum Yields and Relative Efficiencies. The determi-
nation of quantum yields was accomplished by isolating the
333-nm line of the argon-ion laser and by employing 2,3-
diazabicyclo[2.2.1]heptane (DBH) as an actinometer, whose
decomposition quantum yield has been reported to be unity.13
The products were quantified by GC analysis with diphenyl ether
as an internal standard. The GC response factors for the
characteristic NMR signals of the hydrazoalkanes 4 and the
oxidation products of the primary and secondary amines.
Accordingly, two molecules of the primary amine are consumed
to produce one molecule of the hydrazine and one molecule of
the imine, while for secondary amines, only one molecule of
the amine leads to one molecule of the hydrazine and the
imine.
The oxidation products derived from the photolysis of
azoalkanes 1a and 1b in the presence of triethylamine were more
complex than for the primary and secondary amines. GC
analysis of the photolysates indicated diethylamine (60% relative
to the yield of the hydrazine), which presumably is formed from
1
hydrazines were determined by combined H NMR and GC
analyses. To assess the relative order in the efficiency with
which the amines (primary, secondary, and tertiary) reduce the
(11) (a) Cohen, S. G.; Stein, N. M. J. Am. Chem. Soc. 1971, 93, 6542.
(b) Cohen, S. G.; Davis, G. A.; Clark, W. D. K. J. Am. Chem. Soc. 1972,
94, 869.
(12) Padwa, A. Chem. ReV. 1977, 77, 37.
(13) Clark, W. D. K.; Steel, C. J. Am. Chem. Soc. 1971, 93, 6347.

6752 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Adam et al.
Table 2. Quantum Yieldsa for the Photolysis of Azoalkane 1b
(0.05 M) in the Presence of Various Hydrogen Donorsb
hydrogen donor
φrc
φ(2b)d
φ(4b)e
none
0.550
0.070
0.090
0.020
0.037
0.070
0.089
0.039
0.031
0.034
0.008
0.018
0.023
0.029
n-butylamine
cyclohexylamine
diethylamine
diisopropylamine
triethylamine
Me2NCH2Ph
0.037
0.049
0.011
0.026
0.040 (0.46)f
0.063
a At 333 nm in benzene; GC analysis, error (10%, conversion
5-20%. b The concentration was 1.0 M. c Quantum yields for azoalkane
disappearance. d Quantum yield for 2b formation. e Quantum yield for
hydrazine 4b formation. f Value in parentheses refers to acetonitrile.
Figure 2. Difference transient absorption spectrum of azoalkane 1a
in the presence of triphenylamine (ca. 1.5 mM) in benzene obtained
by subtraction of the transient absorption spectra at 1.0- and 0.2-µs
time delay.
error (15-30%) in the rate data is due to a number of limitations,
which comprise low signal intensity of the visible absorption
(∆OD450 ≈ 0.01), absorption of the additives in the monitored
spectral region (at ca. 300 nm), photochemistry of the aromatic
amines upon 355-nm irradiation, and the formation of additional
(transient) photoproducts. Hence, the rate data could only be
obtained from changes of the weak absorption in the visible
region, since photoproducts interfered in the relevant UV region
(supposedly aminoalkyl and hydrazinyl radicals). The experi-
mental problems encountered during the determination of the
quenching rate constants of the amines discouraged the exami-
nation of a more comprehensive set of amines, because the
variations in the rate constants for different amines were likely
to be smaller than the experimental error. In some cases, the
influence of the solvent was examined, e.g., in benzene, carbon
tetrachloride, methylene chloride, and acetonitrile, but the
differences were within the rather large experimental error.
Laser-flash photolysis of a solution of azoalkane 1a in
benzene or acetonitrile in the presence of triphenylamine (ca.
1.5 mM) produced a time-resolved growth at ca. 670 nm,
assigned to the triphenylamine radical cation,14 which was
accompanied by a concomitant decay in the absorption of 31a.
In such case, the spectra of the growing and decaying transient
can be readily extracted by recording the transient absorption
spectra at two different delay times (e.g., at 0.2 and 1.0 µs) and
subsequent spectral subtraction. This procedure afforded the
difference spectrum in Figure 2, in which negative ∆OD values
belong to the decaying transient and positive ones indicate a
transient growth (note that the spectral region below 550 nm
Figure 1. Transient absorption spectrum of the triplet-excited azoalkane
1a obtained upon laser-flash photolysis in benzene and a representative
decay trace recorded at 450 nm (inset).
Table 3. Quenching Rate Constants (kq)a for Triplet Azoalkane 1a
amine
E1/2 (ox) (Versus SCEb) kq (108 M-1 s-1
)
triethylamine
diphenylamine
triphenylamine
1,1-dimethylhydrazine
0.96
0.84
0.85
1.5
13
6.6
14
a Determined by laser-flash photolysis (λexc ) 355 nm) by using the
triplet absorption of the azoalkane 1a in benzene, error (15%. b In
acetonitrile from ref 22.
azoalkanes, quantum yields for the hydrazine 4b formation from
the azoalkane 1b were determined at 1 M amine concentration.
At this amine concentration, >95% of the azoalkane triplets
are quenched since the triplet lifetime of azoalkane 1b is 440
ns5c and a quenching rate constant of the order of 108 M-1 s-1
applies. In Table 2 are summarized the results for the quantum
yields of azoalkane decomposition (φr), housane 2b formation
[φ(2b)], and the hydrazine 4b formation [φ(4b)]. The relative
efficiencies of the amines to photoreduce the triplet azoalkanes
follows the order primary ≈ tertiary . secondary. Thus, the
poorest efficiency is displayed by secondary amines (Table 2).
Rate Constants for Amine Quenching. The T-T absorp-
tion of azoalkane 1a in benzene shows one maximum in the
UV region at ca. 300 nm and a second one in the visible region
at ca. 420 nm (Figure 1). The slight shift of the absorption
maxima as compared to the previously reported data5c is
supposedly due to different spectral recording techniques. The
transient triplet lifetimes were obtained by single-exponential
fitting to the decay traces, typically recorded at 450 nm (Figure
1).
3
represents virtually the mirror image of the 1a absorption in
the absence of amine, cf. Figure 1). The assignment of the
growth to the amine radical cation was based on the comparison
with its formation in the photolysis of benzophenone in the
presence of triphenylamine.14 This provides evidence that at
3
least a fraction of 1a undergoes electron transfer (eq 6).
Discussion
The azoalkanes 1 represent a unique category of azo
compounds in which intersystem crossing from the n,π* singlet-
excited state competes efficiently with R C-N bond cleavage
(k1CN, Scheme 1). Some of the relevant photophysical para-
The triplet lifetimes of azoalkane 1a or its CdC hydrogenated
derivative were plotted (τ0/τ ) 1 + kqτ0[amine]) Versus the
amine concentration (typically five data points) to obtain the
bimolecular quenching onstants (Table 3). The quenching rate
constant for 1,1-dimethylhydrazine, a model for the hydrazine
photoproducts 4, was also determined for comparison. The large
(14) Bartholomew, R. F.; Davidson, R. S.; Lambeth, P. F.; McKellar, J.
F.; Turner, P. H. J. Chem. Soc., Perkin Trans. 2 1972, 577.

Photoreduction of Azoalkanes by Amines
J. Am. Chem. Soc., Vol. 119, No. 29, 1997 6753
Table 4. Photophysical Parameters for Azoalkanes 1a
basic aspects of our mechanistic interpretation are independent
of spin multiplicity.
b
azoalkane
φ(ISC)
3τ (ns)c
φrd
ETe (kcal/mol)
The photoreduction of azoalkanes 1a-c may, in principle,
occur by a mechanism which involves direct hydrogen atom
abstraction from the R carbon atoms of the amines. This
mechanism, however, cannot be reconciled for the following
two reasons: Firstly, despite rather comparable bond dissocia-
tion energies for the R C-H bonds of amines (e.g., 94, 87 and
84 kcal mol-1 for H2NCH2-H, MeNHCH2-H, and Me2NCH2-
H)16a and a typical hydrogen donor such as 2-propanol (88 kcal
mol-1),16b the rate constants for quenching of triplet azoalkane
1a by amines, e.g., triethylamine (kq ) 1.5 × 108 M-1 s-1, Table
3), are 2 orders of magnitude higher than for 2-propanol (kq )
6.4 × 105 M-1 s-1).15 This is in contrast to what would be
expected from the dependence of the rate of hydrogen abstrac-
tion based on bond dissociation energies.17 Secondly, the
quantum yield for decomposition of azoalkane 1b in pure
benzene is 0.55 (Table 2), while that in the presence of 1.0 M
amines decreases to typically ca. 0.08, which points to sub-
stantial deactivation of the excited states.
A perusal of the results of laser-flash photolysis experiments
in Table 3 shows that the quenching rate constants determined
for the triplet azoalkane 1a, a representative case, are relatively
high. Furthermore, the dependence of quenching rate constants
on the oxidation potentials of the amines, e.g., the rate constant
for diphenylamine is 1 order of magnitude higher than for
triethylamine, clearly suggests the importance of charge-transfer
interactions. It should be noted that the fluorescence quenching
of DBO8a,b by olefins and of a DBH derivative by olefins and
amines8c has also been interpreted in terms of charge transfer.
To rationalize the results of the bimolecular photochemistry
in Table 1, let us consider the various mechanistic pathways in
Scheme 2 to produce the primary hydrazinyl radical, for which
on the basis of the aforementioned we propose that n,π* triplet-
excited states serve as the precursors. The initial interaction of
the triplet azoalkanes 1 with the ground state amines may lead
either (i) to a partial charge transfer (k3q) complex (or exci-
plex) or (ii) to electron transfer (k3ET) to produce the radical ion
pair; in polar solvents, the latter may also be formed by ionic
dissociation of the polar charge transfer (kCETT) complex.18
Hydrogen atom transfer within the charge-transfer complex
would generate the hydrazinyl radical (kCHT); alternatively, the
latter may be produced by proton transfer from the amine radical
cation to the radical anion of the azoalkane (kEHT). Which of
these two pathways prevails should be governed by the redox
properties of the donor/acceptor pair and the polarity of the
medium.
1a
1b
1c
1d
0.5
0.5
630 ( 20
440 ( 20
0.59
0.60
62.2
61.5
e440 ( 40f
200 ( 10
0.1
0.88
61.6
a Reference 5c. b Quantum yield for intersystem crossing in benzene.
c Triplet lifetimes. d Quantum yield for azoalkane disappearance, error
(10%, except for 1a (5%). e Triplet energy, (0.1 kcal mol-1 f This
.
work in benzene (λexc ) 351 nm); the value constitutes an upper limit
since it cannot be decided whether the observed transient (λmon ) 315
nm) is the triplet azoalkane or the triplet biradical derived therefrom
(cf. ref 5c).
meters of the azoalkanes 1 are given in Table 4. The direct
photolysis of azoalkanes 1a and 1b affords not only the expected
housanes 2a and 2b but also the aziranes 3a and 3b derived
from â C-C bond cleavage (k3CC).5a,b From triplet-quenching
experiments with piperylene, the intersystem-crossing quantum
yields for the azoalkanes 1a and 1b have been estimated to be
as high as 0.5.5a,c Furthermore, from the temperature depen-
dence of the photoproduct distribution, the activation energies
for R C-N bond cleavage from the singlet state (k1CN) and R
C-N (k3CN) and â C-C (kC3 C) bond cleavages from the triplet
state have been estimated5b for the parent azoalkane 1a as ER
1
> 3.2, ER > 10.5 and Eâ > 7.8 kcal mol-1. Particularly
noteworthy is the influence of incipient radical stabilization:
the better the carbon radical site in the intermediate diazenyl
biradical is stabilized, i.e., Ph (1c and 1d) . Me (1b) > H
(1a), the more R C-N cleavage to the housane occurs. Thus,
the phenyl substitution in 1c and 1d is so effective that
denitrogenation from the singlet-excited state dominates with
marginal ISC to the triplet manifold (φISC ca. 0.10).5c Further-
more, the influence of the radical-stabilizing ability of the
substituents is also reflected in the triplet lifetimes (Table 4),
which follow the order 1a (630 ns) > 1b (440 ns) g 1c (e440
ns) > 1d (200 ns). A manifestation of these two factors, namely
shortened triplet lifetimes and lower ISC quantum yields, is
clearly revealed in the results of photoreduction by the amines
collected in Table 1. Whereas the photolysis of azoalkanes 1a
and 1b at moderate concentrations of amines affords high yields
of the hydrazines 4a and 4b (Table 1), that of 1c leads to major
amounts of the hydrazine 4c only in neat amines (entries 26-
28). The azoalkane 1d undergoes no photoreduction at all, even
at high concentrations of amines (entries 29-31).
3
3
Strong evidence has been brought forward in our preliminary
study9 that the triplet states of azoalkanes 1 are responsible for
the observed photoreduction. For example, the fact that only
azoalkanes with high quantum yields of intersystem crossing
(φISC ca. 0.5, Table 4) undergo photoreduction implies that the
observed bimolecular photochemistry results from relatively
long-lived (3τ ca. 0.5 µs) triplet states. Early fluorescence
measurements8c by Ramamurthy and Turro have shown that the
singlet-excited azoalkanes (also of DBH-type) are also quenched
From the known excitation energies and redox potentials of
the donor/acceptor pair, the thermodynamic feasibility of
electron transfer may be predicted by the Rehm-Weller
2
equation,19 ∆Gel ) 23.06[Eox - Ered - e0 /aꢀ] - ET. Although
the triplet energies of azoalkanes 1a and 1b are fortunately
available from a previous study,5c the irreversible electrochemi-
cal reduction of the azoalkanes 1a-d (estimated in the range
-2.4 to -2.7 V Versus SCE in acetonitrile)20 makes a prediction
by aliphatic amines with rate constants as high as 108 M-1 s-1
,
and since the singlet lifetimes of azoalkanes 1a and 1b are
relatively long (1τ ca. 2-3 ns),5c bimolecular quenching of the
azoalkane singlet states is expected (<20%) at the typically
employed preparative concentrations of the amines (0.5-1.0
M). Hence, contributions due to singlet quenching cannot be
entirely ruled out in the preparative experiments, in particular
in neat amines as solvents. The efficiency for singlet photore-
duction must be low, however, since the previous fluorescence
quenching studies did not lead to bimolecular photoproducts
with the amines.8c Regardless of the possible involvement of
singlet quenching, it should be noted that our data can be
satisfactorily rationalized by triplet photoreduction and that the
(15) Nau, W. M.; Cozens, F. L.; Scaiano, J. C. J. Am. Chem. Soc. 1996,
118, 2275.
(16) (a) Griller, D.; Lossing, F. P. J. Am. Chem. Soc. 1981, 103, 1586.
(b) Jacques, P. J. Photochem. Photobiol. A 1991, 56, 159.
(17) (a) Wagner, P. J. Top. Curr. Chem. 1976, 66, 1. (b) Turro, N. J.
Modern Molecular Photochemistry; The Benjamin/Cummings Publishing
Company, Inc.: Menlo Park, CA, 1978; p 364.
(18) Levin, P. P.; Kuzmin, V. A. Russ. Chem. ReV. 1987, 56, 2527.
(19) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259.
(20) The previously reported values for related bicyclic azoalkanes (cf.
Ess, C. H.; Gerson, F.; Adam, W. HelV. Chim. Acta. 1991, 74, 2078) appear
to be in error (F. Gerson, personal communication).

6754 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Adam et al.
Scheme 2
of the free energy for electron transfer from the Weller relation
unreliable. However, even by assuming -2.0 V as a limiting
lower value for the reduction potential of the azoalkanes 1, one
calculates positive free enthalpies of electron transfer for the
trialkylamines, e.g., triethylamine (∆Gel ) 4-5 kcal mol-1),
constants. Detailed mechanistic studies have shown that the
ketones, e.g., trifluoroacetophenone, are reduced by irreversible,
rate-determining charge-transfer complexation with varying
degrees of charge separation.24 It has been shown that charge-
transfer complexation precedes the hydrogen atom transfer step,
although direct hydrogen atom abstraction without prior com-
plexation may be competitive. The photoreduction of DBO by
1,3-cyclohexadiene has similarly been postulated to proceed
through a charge-transfer complex with subsequent hydrogen
atom transfer.8b
The photoreduction of azoalkanes 1a and 1b by N,N-
dimethylbenzylamine provides additional evidence in favor of
the proposed charge-transfer mechanism (k3q, Scheme 2). In an
electron-transfer mechanism (k3ET), competition would be ex-
pected for deprotonation from the methyl and the benzylic sites
of the N,N-dimethylbenzylamine radical cation (eq 7). Indeed,
2
an estimate in which the Coulomb interaction term (e0 /aꢀ) has
been ignored in the polar solvent.
This unfavorable trend in the electron-transfer energetics is
also reflected in a purely qualitative comparison of the quench-
ing rate constants of triethylamine for the triplet azoalkane 1a
with that for benzophenone; the latter possesses a higher triplet
energy (by ca. 7 kcal mol-1) and more favorable reduction
potential (-1.8 V Versus SCE in acetonitrile).21 Whereas the
rate constant for quenching of the triplet benzophenone is close
to the diffusion limit (3.0 × 109 M-1 s-1),6b that for the
azoalkane 1a is 1 order of magnitude lower (Table 3). While
marginally positive for triethylamine (Eox ) 0.96 V Versus
SCE),22 the free-energy change of electron transfer becomes
progressively more positive for secondary and primary amines
because of their higher oxidation potentials. Thus, to account
for the high quenching rate constants but low quantum yields
of photoreduction by the aliphatic amines employed in the
present study, we favor the pathway with a charge-transfer
complex (k3q) in which the extent of charge separation is
determined by the redox properties of the donor/acceptor pair.
The charge-transfer interaction involves the nitrogen lone pair
of the amine as the donor and the azo chromophore as the
acceptor, which is deficient in electron density by virtue of n,π*
excitation. Such a partial charge transfer weakens the R C-H
bond and facilitates abstraction of the hydrogen atom (kCHT).23
The unproductive radiationless decay of the charge-transfer
complex (kCq T) to the ground state (after intersystem crossing)
accounts for the observed low efficiencies of the photoreduction
(Table 2).
statistical (3:1) deprotonation has been reported in the anodic
oxidation of this amine in methanol.25 Moreover, a stereoelec-
tronic effect has been proposed to operate in the deprotonation
step with overlap between the half-vacant nitrogen p orbital and
the incipient p orbital of the radical cation.26 Consequently,
deprotonation occurs preferentially from the least substituted
methyl site in the photoreduction of benzophenone,6a as well
as the photoaddition reaction with trans-stilbene.26a However,
no oxidation products (N,N-benzylmethylamine and formalde-
hyde) derived from the methylbenzylamino-substituted methyl
The present photoreduction of triplet azoalkanes 1a-c by
aliphatic amines bears a close analogy to the photoreduction of
ketones by alkylbenzenes.24 In this case, the free energy of
electron transfer is endergonic and the quantum yields of
photoreduction are rather low in spite of high quenching rate
1
radical (eq 7) could be detected either by GC or H NMR
analysis in the photoreduction of the azoalkanes 1a and 1b.
Instead, only a mixture of meso and dl dehydro dimers was
found, formed by coupling of the (N,N-dimethylamino)benzyl
radicals (eq 5). Thus, the hydrogen atom is specifically
transferred from the benzylic site to the triplet azo chromophore
in a process akin to direct hydrogen atom abstraction.
(21) CRC Handbook of Chemistry and Physics, Weast, R. C., Lide, D.
R., Astle, M. J., Beyer, W. H., Eds.; CRC Press: Boca Raton, FL, 1990;
Vol. 70.
(22) Arbogast, J. W.; Foote, C. S.; Kao, M. J. Am. Chem. Soc. 1992,
114, 2277. See also: Jonsson, M.; Wayner, D. D. M.; Lusztyk, J. J. Phys.
Chem. 1996, 100, 17539.
Noteworthy is the relatively small solvent effect27 (Table 1),
which is consistent with the proposed charge-transfer mechanism
(23) Tarasyuk, A. Y.; Granchak, V. M.; Dilung, I. I. J. Photochem.
Photobiol. A 1995, 85, 39.
(24) Wagner, P. J.; Truman, R. J.; Puchalski, A. E.; Wake, R. J. Am.
Chem. Soc. 1986, 108, 7727.
(25) Barry, J. E.; Finkelstein, M.; Mayeda, E. A.; Ross, S. D. J. Org.
Chem. 1974, 39, 2695.
(26) (a) Lewis, F. D. Acc. Chem. Res. 1986, 19, 401. (b) Lewis, F. D.;
Ho, T.-I.; Simpson, J. T. J. Am. Chem. Soc. 1982, 104, 1924.

Photoreduction of Azoalkanes by Amines
J. Am. Chem. Soc., Vol. 119, No. 29, 1997 6755
Scheme 3
(k3q) and provides evidence against the electron-transfer-medi-
ated processes (k3ET) in Scheme 2. Thus, for azoalkanes 1a and
1b, a change from benzene (Table 1, entries 7, 10, 20, and 24)
to acetonitrile (entries 8, 11, 21, and 25) or methanol (entries 9
and 22) causes only a marginal enhancement of the yields of
hydrazines. Also, the quenching rate constants in the polar
acetonitrile (data not shown), assessed through laser-flash
photolysis, were found to be comparable within an experimental
error of 15% to the values determined for the nonpolar benzene
(Table 3). In regard to the solvent effects, the high yields of
hydrazines from the photolysis in neat amines are a puzzle
(Table 1, entries 13, 17, and 23). The amine medium may
promote more efficient hydrogen atom transfer in the charge-
transfer complex, but the reasons for this are not evident as yet.
The unfavorable energetics for electron transfer (k3ET,
Scheme 2) from the aliphatic amines to the azoalkanes 1a-c is
remedied by triphenylamine as a better donor, as evidenced by
the fact that the corresponding radical cation is readily detected
in the laser-flash photolysis of azoalkane 1a (eq 6). Hence, an
oxidation potential corresponding to that of triphenylamine (0.85
V Versus SCE in acetonitrile)22 constitutes a higher limit, equal
to or below which the free energy change is negative and the
electron transfer is favored. For such amines the charge-transfer
complex (or triplet exciplex) may undergo dissociation (kCETT) to
the radical ion pair, which may also be formed directly (k3ET) as
in the case of the benzophenone/amine system.
Let us now consider the relative reactivity with which the
charge-transfer complex (k3q) for primary, secondary, and
tertiary aliphatic amines undergoes hydrogen atom transfer.
With the quenching rate constants of the order of 108 M-1 s-1
(Table 3), one calculates that >95% of the triplets are quenched
by the aliphatic amines at 1.0 M concentration. Thus, from
the quantum yields of photoreduction in Table 3 for the
azoalkane 1b, a representative case, in the presence of a variety
of amines (1.0 M), we note that the photoreduction efficiencies
[φ(4b) ca. 0.01-0.06, Table 2] are substantially lower as
compared to those for benzophenone (φ ca. 0.30-1.00).6b
Furthermore, a definite trend in the efficiencies of hydrazine
formation [φ (4b)] for primary, secondary, and tertiary amines
is evident in Table 2. Thus, the primary and tertiary amines
display comparable efficiencies of photoreduction, while those
for secondary amines are definitely lower.
in N-H bond dissociation energies of secondary and primary
amines (96 Versus 103 kcal mol-1),6b a larger proportion of
transfer of the hydrogen atom from the N-H bond to produce
aminyl radicals has been proposed to take place in secondary
Versus primary amines.28 Subsequent disproportionation of the
ketone-derived hydroxydiphenylmethyl and secondary aminyl
radicals to regenerate the ketone and the secondary amine by
hydrogen back-transfer has been advanced as the reason for the
observed lower efficiency of photoreduction.
Similar mechanistic considerations as those above for ben-
zophenone may explain the lower efficiency of reduction of
the triplet azoalkane 1b by secondary amines, except that the
mechanism involves a partial charge (k3q) rather than electron
transfer (k3ET) in Scheme 2. Thus, the reaction of triplet
azoalkanes with secondary amines may result in hydrazinyl and
alkylaminyl radicals (k3NH, Scheme 3) as well as R-aminoalkyl
radicals (k3CH), the latter formed by the transfer of a hydrogen
atom from the R carbon atom of the secondary amine. Whereas
the R-aminoalkyl radicals react further to the hydrazine and
imine (kCN•H), hydrogen back-transfer from the hydrazinyl radi-
cal (kNN•H) to the aminyl radical regenerates the azoalkane and
amine, which lowers the efficiency of photoreduction. Besides
this relative order of efficiencies for the photoreduction of the
triplet azoalkanes by amines, i.e., tertiary ≈ primary >
secondary, one recognizes from Table 2 that the efficiency of
photoreduction is higher for diisopropylamine as compared to
that for diethylamine. Presumably, this results from a lower
bond dissociation energy of the R C-H bond in the isopropyl
Versus ethyl substituents.
As mentioned previously, the N-H bond dissociation energy
of the primary amines is substantially higher as compared to
that of the secondary amines. In view of this, the transfer of a
hydrogen atom from N-H bond of the primary amines to the
triplet-excited azoalkanes does not seem feasible. This is
underscored by the lack of photoreduction in the presence of
amines that do not contain R C-H bonds, namely tert-
butylamine and diphenylamine (Table 1).
Finally, the formation of the hydrazine product from the
hydrazinyl radical in Scheme 2 may occur by the pathways
depicted in eqs 8-10. The first pathway (eq 8) involves
A similar trend in the photoreduction has been reported for
benzoylbenzoic acid by amines in water as the medium.11a It
has been found from laser-flash photolysis studies that the
photoreduction of benzophenone by secondary amines leads to
hydroxydiphenylmethyl radicals with unit efficiency;6b however,
the overall steady-state quantum yields of photoreduction are
significantly lower when compared to those for primary and
tertiary amines. The lower efficiency of photoreduction of
benzophenone for secondary amines has been rationalized in
terms of the possibility of competitive hydrogen atom transfer
from both R C-H and N-H bonds. In view of the difference
(28) The higher kinetic acidity for N-H versus R C-H deprotonation
has been borne out also in the photoreactions between cyanoarenes and
secondary amines; cf. ref 26.
(27) Santamaria, J. In Photoinduced Electron Transfer; Fox, M. A.,
Channon, M., Ed.; Elsevier: Amsterdam, 1988; Part C, p 482.

6756 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Adam et al.
hydrogen transfer from the R-aminoalkyl radical, e.g., derived
from trialkylamines such as triethylamine, to the hydrazinyl
radical to produce the hydrazine and the enamine (eq 4). The
reduction by primary and secondary amines presumably follows
the same pathway (eq 8), except that the second hydrogen atom
is transferred from the N-H bond of the R-aminoalkyl radical
to produce imines (eqs 2 and 3). Alternatively, the R-aminoalkyl
radical may reduce the ground-state azoalkane to produce the
hydrazinyl radical and the enamine (eq 9). The latter reaction
pathway must contribute to photoreduction, since it is well
established that hydroxydiphenylmethyl radicals, which in fact
are less potent reducing agents (Eox ) -0.7 V Versus SCE)30
than R-aminoalkyl radicals (Eox ca. -1.0 V Versus SCE),16a,29
are known to reduce azoalkanes efficiently.30b
In summary, similar to benzophenone, the n,π* excited
azoalkanes are efficiently quenched by amines to afford hydra-
zines as photoreduction products. Electron transfer to azoal-
kanes by amines with low oxidation potentials, e.g., triphenyl-
amine (Eox ca. 0.85 V Versus SCE), to produce radical ion pairs
is analogous to the photoreduction of benzophenone by amines.
In marked contrast to benzophenone, however, the quantum
yields of photoreduction of azoalkanes are significantly lower
for amines with higher oxidation potential, for which efficient
deactivation by charge transfer is postulated. Besides deactiva-
tion of the triplet-excited azoalkanes, charge transfer also
promotes hydrogen atom abstraction and the efficiencies follow
the trend observed for benzophenone, e.g., secondary amines
display the poorest efficiency despite efficient quenching. These
new results encourage further exploration of the bimolecular
photochemistry of azoalkanes.
In the second pathway (eq 10), the disproportionation of two
hydrazinyl radicals affords one molecule of the azoalkane and
one molecule of the hydrazine. Indeed, such a disproportion-
ation of hydrazinyl radicals has been well established in the
literature.31 This case applies for the photoreduction of azoal-
kanes 1a and 1b by N,N-dimethylbenzylamine (eq 5), because
the intermediary R-(dimethylamino)benzyl radical has no R
hydrogen atoms available and, thus, couples to form the dehydro
dimers of the amine.
Acknowledgment. We thank the Deutsche Forschungsge-
meinschaft, the Volkswagen-Stiftung, and the Natural Sciences
and Engineering Research Council of Canada and the Killam
foundation for financial support. J.N.M. is grateful for a
fellowship from the Alexander von Humboldt Foundation and
W.M.N. for a Canada International fellowship and a NATO
fellowship awarded by NSERC and the DAAD.
(29) Wayner, D. D. M.; McPhee, D. J.; Griller, D. J. Am. Chem. Soc.
1988, 110, 132.
(30) (a) Mugnier, Y.; Laviron, E. J. Org. Chem. 1988, 53, 5781. (b)
Engel, P. S.; Wu, W.-X. J. Am. Chem. Soc. 1989, 111, 1830. (c) Holt, P.
F.; Hughes, B. P. J. Chem. Soc. 1955, 98.
Supporting Information Available: Experimental data (6
pages). See any current masthead page for ordering and Internet
access instructions.
(31) Hayon E.; Simic, M. J. Am. Chem. Soc. 1972, 94, 42.
JA9706460