Dye-sensitized photooxygenation of the C=N bond. 5. Substituent effects on the cleavage of the C=N bond of C-aryl-N-aryl-N-methylhydrazones

Article abstract of DOI:10.1021/jo050014t

The title compounds are cleaved cleanly at the C=N bond by singlet oxygen (¹O₂, ¹Δ_g) yielding arylaldehydes and N-aryl-N-methylnitrosamines. These reactions take place more rapidly at -78 °C than at room temperature. The effects of substituent variation at both the C-aryl and N-aryl groups were studied using a competitive method. Good correlations of the resulting rate ratios with substituent constants (σ^- or σ⁺) were obtained yielding small to very small ρ values indicative of small to very small changes in charge distribution between the reactant and the rate determining transition state. Electron withdrawing groups on the C-aryl moiety retard reaction somewhat by preferential stabilization of the hydrazone. Electron donors on the other hand slightly stabilize the rate determining transition state. Substituents on the N-aryl group have almost no effect. Inhibition by 3,5-di-tert-butylphenol was not observed showing that free (uncaged) radical intermediates are not involved in the mechanism. We postulate a mechanism in which the initial event is exothermic electron transfer from the hydrazone to ¹O₂ leading to an ion-radical caged pair. Subsequent covalent bond formation between the hydrazone carbon and an oxygen atom is rate controlling. The transition state for this step also has a lower enthalpy than the starting reactants, but the free energy of activation is dominated by a large negative TΔS? term leading to the negative temperature dependence. Direct formation of a C-O bond in the initial step is not unambiguously ruled out. Subsequent steps lead to C-N cleavage.

Full text of DOI:10.1021/jo050014t

Dye-Sensitized Photooxygenation of the CdN Bond. 5. Substituent
Effects on the Cleavage of the CdN Bond of
C-Aryl-N-aryl-N-methylhydrazones1
Ihsan Erden,* Pinar Ergonenc Alscher, James R. Keeffe,* and Colin Mercer
Department of Chemistry and Biochemistry, San Francisco State University, 1600 Holloway Avenue,
San Francisco, California 94132
ierden@sfsu.edu; keeffe@sfsu.edu
Received January 3, 2005
1
The title compounds are cleaved cleanly at the CdN bond by singlet oxygen (1O2, ∆g) yielding
arylaldehydes and N-aryl-N-methylnitrosamines. These reactions take place more rapidly at -78
°C than at room temperature. The effects of substituent variation at both the C-aryl and N-aryl
groups were studied using a competitive method. Good correlations of the resulting rate ratios
with substituent constants (σ- or σ+) were obtained yielding small to very small F values indicative
of small to very small changes in charge distribution between the reactant and the rate determining
transition state. Electron withdrawing groups on the C-aryl moiety retard reaction somewhat by
preferential stabilization of the hydrazone. Electron donors on the other hand slightly stabilize
the rate determining transition state. Substituents on the N-aryl group have almost no effect.
Inhibition by 3,5-di-tert-butylphenol was not observed showing that free (uncaged) radical
intermediates are not involved in the mechanism. We postulate a mechanism in which the initial
1
event is exothermic electron transfer from the hydrazone to O2 leading to an ion-radical caged
pair. Subsequent covalent bond formation between the hydrazone carbon and an oxygen atom is
rate controlling. The transition state for this step also has a lower enthalpy than the starting
reactants, but the free energy of activation is dominated by a large negative T∆Sq term leading to
the negative temperature dependence. Direct formation of a C-O bond in the initial step is not
unambiguously ruled out. Subsequent steps lead to C-N cleavage.
Introduction
In this study we report the reaction of singlet oxygen
with C-aryl-N-aryl-N-methylhydrazones in which para
substituents on the C-aryl and N-aryl rings have been
systematically varied. Competition for singlet oxygen
between different hydrazones was designed to allow
determination of relative reaction rates for the differently
substituted substrates, and these results permitted analy-
sis with use of established substituent constant values.4
The effect of temperature on reaction efficiency was also
determined. To the best of our knowledge this is the first
systematic study of purely electronic substituent effects
In previous studies we have shown that reaction of
singlet oxygen, O2 (1∆g), with the CdN bond requires
1
the presence of good to very good electron donating
groups attached to the doubly bonded nitrogen. For
example, aryloximes and aryloxime ethers react slug-
gishly or not at all, but oximate anions react readily, as
do nitronate anions, O-trimethylsilylnitronates, C-aryl-
N,N-dimethylhydrazones, and nitrones.2 For the reac-
1
tions of O2 with nitrones and C-aryl-N,N-dimethylhy-
drazones we have observed peroxidic intermediates by
low-temperature NMR spectroscopy. The accumulated
observations permitted mechanistic proposals to be made
about these reactions which do not depend solely on
identification of the products.3
1
on the reactivity of O2 with a single substrate type.
Results
Four types of relative rate comparisons were made. In
Set I hydrazones with X ) variable and Y ) H [see eq 1]
were compared with the parent hydrazone for which X
(1) Taken partly from the M.S. thesis of Pinar Mara Ergonenc
(Alscher), San Francisco State University, San Francisco, CA, June
1993. For Part 4 see: Ocal, N.; Yano, L. M.; Erden, I. Photooxygenation
of the CdN bond: A Mild New Method for Oxidative C-C Cleavage.
Tetrahedron Lett. 2003, 44, 6947.
(2) Castro, C.; Dixon, M.; Erden, I.; Ergonenc, P.; Keeffe, J. R.;
Sukhovitsky, A. J. Org. Chem. 1989, 54, 3732.
(3) Erden, I.; Griffin, A.; Keeffe, J. R.; Brinck-Kohn, V. Tetrahedron
Lett. 1993, 34, 793.
(4) Hine, J. Structural Effects on Equilibria in Organic Chemistry;
Wiley-Interscience: New York, 1975; Chapter 3.
10.1021/jo050014t CCC: $30.25 © 2005 American Chemical Society
Published on Web 04/23/2005
J. Org. Chem. 2005, 70, 4389-4392
4389

Erden et al.
TABLE 1. Product Ratios (Relative Rates) for
Compeititve Reaction with Singlet Oxygen, Set I:
4-X-C6H4-CHdN-N(CH3)Ph vs Ph-CHdN-N(CH3)Ph(-78 °C,
CH2Cl2 solution)
product
ratioa
standard deviation
of the meanb
X
NO2
0.115
0.752
0.93
0.003
0.007
0.001
0.0012
0.002
Cl
N(CH3)2
CH3
OCH3
1.21
) Y ) H. In Set II compounds with X ) CH3 and Y )
variable competed with X ) CH3, Y ) H. In Set III X )
Cl and Y was again varied. These were compared with X
) Cl and Y ) H. Set IV, a smaller group, allowed both X
and Y to vary and compete with X ) Y ) H. In all cases
the reaction temperature was -78° C, initial subtrate
concentrations were 0.05 M (in CH2Cl2), and the photo-
sensitizer, TPP, was at 10-4 M. Reaction progress was
monitored to no greater than 10% conversion to ensure
approximately zero-order kinetics. See the Experimental
Section for details. Relative rates were determined by 1H
NMR analysis of integrated areas of one or more char-
acteristic product signals. Ten to fifteen replicates were
run for each competition. The relative rate ratios for each
set are given in Tables 1-4 along with the standard
deviation of the mean.
In addition to the competition experiments just de-
scribed, the reactions of the hydrazones with X ) Y ) H
and X ) CH3O, Y ) H were compared at two tempera-
tures, -78 and ∼20 °C. In these experiments the stirring
time prior to illumination was varied in order to test
whether oxygen saturation of the reaction solutions had
occurred. At short reaction times, just long enough to
detect product formation for the slower reaction (10 min),
the reactions at -78 °C were uniformly faster by an
amount independent of prior stirring time. The ratio of
reactivities at the two temperatures was found to be at
least 45:1. We emphasize that this is a minimum ratio
because by the time product formation could be detected
for the slower reaction (∼1% conversion), product forma-
tion at -78° C had reached 40-70%, thus substrate
concentration in that solution had fallen significantly.
1.52
a Reactions were run to no greater than 10% conversion of the
more reactive substrate. Product ratios therefore measure relative
reaction rates to a good approximation. b Ten to fifteen replicate
compeititions were done for each pair.
TABLE 2. Product Ratios (Relative Rates) for
Competitive Reaction with Singlet Oxygen, Set II:
4-CH3-C6H4-CHdN-N(CH3)C6H4-4-Y vs
4-CH3-C6H4-CHdN-N(CH3)Ph (-78 °C, CH2Cl2 solution)
product
ratioa
standard deviation
of the meanb
Y
NO2
Cl
CH3
OCH3
1.27
1.07
0.94
0.75
0.03
0.02
0.03
0.04
a Reactions were run to no greater than 10% conversion of the
more reactive substrate. Product ratios therefore measure relative
reaction rates to a good approximation. b Ten to fifteen replicate
competitions were done for each pair.
TABLE 3. Product Ratios (Relative Rates) for
Competitive Reaction with Singlet Oxygen, Set III:
4-Cl-C6H4-CHdN-N(CH3)C6H4-4-Y vs
4-Cl-C6H4-CHdN-N(CH3)Ph (-78 °C, CH2Cl2 solution)
product
ratioa
standard deviation
of the meanb
Y
Cl
1.24
0.89
0.65
0.51
0.06
0.047
CH3
OCH3
a Reactions were run to no greater than 10% conversion of the
more reactive substrate. Product ratios therefore measure relative
reaction rates to a good approximation. b Ten to fifteen replicate
competitions were done for each pair.
Discussion
TABLE 4. Product Ratios (Relative Rates) for
Competitive Reaction with Singlet Oxygen, Set IV:
4-X-C6H4-CHdN-N(CH3)C6H4-4-Y vs Ph-CHdN-N(CH3)Ph
(-78 °C, CH2Cl2 solution)
Structure of the Hydrazones. The preferred geom-
etry about the CdN bond of C-aryl-N-aryl-N-methylhy-
drazones is taken as syn (E) rather than anti (Z) by
analogy with the observations of Karabatsos5 and He-
garty6 and their respective co-workers. In this form the
geometry is virtually planar about the CdN bond, and
the amino nitrogen lone pair is able to conjugate more
effectively with the CdN π bond provided the groups
attached to the CdN carbon (H in our case) and the
amino nitrogen are not large. Hegarty and co-workers
also obtained convincing evidence that the syn to anti
isomerization barrier is about 23 kcal/mol at room
temperature for a hydrazone of our type, namely, Ph-
CHdN-N(CH3)Ar, where Ar ) 4-nitrophenyl.7 Our
semiempirical molecular orbital calculations (AM1) sup-
product
ratioal
standard deviation
of the meanb
X, Y
NO2, H
NO2, OCH3
OCH3, H
0.115
0.91
1.52
0.003
0.002
0.002
a Reactions were run to no greater than 10% conversion of the
more reactive substrate. Product ratios therefore measure relative
reaction rates to a good approximation. b Ten to fifteen replicate
competitions were done for each pair.
port these conclusions, placing the E isomer 1.1 kcal/mol
below the Z isomer, and showing the amino nitrogen
better situated for conjugation.8
Inverse Temperature Dependence. An inverse
(5) (a) Karabatsos, G. J.; Graham, J. D.; Vane, F. M. J. Am. Chem.
Soc. 1962, 84, 753. (b) Karabatsos, G. J.; Taller, R. A.; Vane, F. M.
Tetrahedron Lett. 1964, 18, 1081.
(6) Hegarty, A. F.; Scott, F. L. J. Org. Chem. 1968, 33, 753.
(7) Scott, F. L.; Groeger, F. A.; Hegarty, A. F. J. Chem. Soc. B 1971,
1411.
temperature dependence on the rate of singlet oxygen
(8) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.
J. Am. Chem. Soc. 1985, 107, 3902.
4390 J. Org. Chem., Vol. 70, No. 11, 2005

Dye-Sensitized Photooxygenation of the CdN Bond
reactions with hydrazones2,9 and other substrates10-12 has
been noted previously. Several factors, physical and
chemical, could play a role in determining this difference.
Long and Kearns13 have determined that in CHCl3
solvent the lifetime of 1O2 is decreased by only about 50%
with a 100 °C increase in temperature. The temperature
dependence of the solubility of oxygen in dichloromethane
was not found in the literature by us. However, studies
of oxygen solubility as a function of temperature in other
solvents have been reported.14 The results show a good
deal of variability. For example, the solubility of O2 in
CCl4 changes hardly at all over the range 0 to 60 °C. In
chlorobenzene the solubility of oxygen steadily increases
between 0 and 80 °C. In acetone the oxygen solubility is
almost invariant with temperature between -78 and 40
°C, and in diethyl ether the solubility declines by a factor
of 1.8 over the range -78 to 20 °C. Without information
for dichloromethane a firm statement cannot be made,
but it appears likely that the influences of increased
lifetime of singlet oxygen and possibly increased solubility
of oxygen at the lower temperature, countered by in-
creased viscosity, hence a smaller diffusion rate at -78
°C, cannot account for the ca. 50-fold greater reactivity
we see at the lower temperature
Other circumstances can lead to an inverse dependence
of rate on temperature. Braderic and Leigh have dis-
cussed possible causes noting that this observation is not
uncommon, especially for bimolecular reactions of reac-
tive intermediates.15 They write that there are two
common mechanistic explanations for such behavior. In
one case there is an initial exothermic formation of a
complex that may revert to reactants or proceed to
products, the product-forming path having a higher free
energy barrier but a lower enthalpy barrier than forma-
tion of the complex or its reverse. The negative temper-
ature dependence is then due to a dominant negative
entropy of activation. To satisfy these requirements the
initially formed complex would be relatively loose com-
pared with the rate-determining transition state on the
product-forming path. In fact, singlet oxygen reactions
often do have near-zero enthalpies of activation and
entropies of activation ranging from -25 to -40 cal/(deg‚
mol).10-12 A second possibility is that the reaction is
concerted, again with domination of ∆Gq by a large,
negative T∆Sq term.16,17 Large negative entropies of
activation are expected whenever a bimolecular process
occurs, but especially when that process involves separa-
tion of charge in a relatively nonpolar medium.18
FIGURE 1. Plot of log(product ratio) vs σ- for Set I. The
diamond-shaped point is for the competition between 4-Me2N-
C6H4-CHdN-N(Me)Ph and Ph-CHdN-N(Me)Ph.
did not provide a good linear correlation with any of the
data sets. Set I (C-ring substituents) were very well
correlated with σ- values, see eq 2. In this correlation
the point for the p-dimethylamino group was 0.64 log unit
below the correlation line established by the other
substituents, see Figure 1. Tertiary amines (DABCO for
example) are known to quench singlet oxygen, and we
ascribe the apparent retardation for this compound to
quenching by the substituent. Sets II and III (N-ring
substituents) were best correlated with σ+ values, see eqs
3 and 4.
Set I:
log(product ratio)I )
-0.74σ- - 0.008; r2 ) 0.994, n ) 5 (2)
Set II:
Set III:
log(product ratio)II )
0.14σ+ + 0.002; r2 ) 0.975, n ) 5 (3)
log(product ratio)III
)
0.28σ+ + 0.34; r2 ) 0.952, n ) 4 (4)
The most prominent result of these correlations is that
the slopes (F values) are either small (Set I) or very small
(Sets II and III). Set IV, a small set in which substituents
on both rings were changed, gave only a fair correlation
with (σ-C + σ+N). A rough value for that slope, F ≈ -0.5,
is in qualitative agreement with the difference between
F (Set I) and F (Set II or III).
The small F values indicate that charge development
in the rate determining transition state is not extensive;
it is likely that the transition state is an early one. We
interpret the result for Set I to signify that the reactant
hydrazone is somewhat stabilized by electron withdraw-
ing groups on the C-ring, especially those capable of a
resonance interaction with the hydrazine function, as
illustrated by structures 1a and 1b. Such stabilization
is decreased in the rate-controlling transition state. Sets
II and III show that reaction is slightly retarded by
electron donors on the N-ring. The effect is so small that
a convincing explanation is lacking.
Substituent Effects. The relative rates compiled in
Tables 1 through 4 were correlated with Hammett
substituent constants4 by using the following relation-
ship, log(product ratio) ) aσ + b. The original σ values
(9) Ito, Y.; Kyono, K.; Matsuura, T. Tetrahedron Lett. 1979, 2253.
(10) Gutman, D. Acc. Chem. Res. 1990, 23, 375.
(11) Mozurkewich, M.; Benson, S. W. J. Phys. Chem. 1984, 88, 6429.
(12) Gorman, A. A.; Gould, I. R.; Hamblett J. Am. Chem. Soc. 1982,
104, 7098.
(13) Long, C. A.; Kearns, D. R. J. Am. Chem. Soc. 1975, 97, 2018.
(14) Battino, R., Ed. Oxygen and Ozone; Solubility Data Series, Vol.
7; IUPAC and Pergamon Press: Oxford, UK, 1981.
(15) Bradaric, C. J.; Leigh, W. J. J. Am. Chem. Soc. 1996, 118, 8971.
(16) Gorman, A. A.; Lovering, G.; Rodgers, M. A. J. J. Am. Chem.
Soc. 1979, 101, 3050.
(17) Houk, K. N.; Rondan, N. G.; Mareda, J. Tetrahedron 1985, 41,
1555.
(18) Frost, A. A.; Pearson, R. G. Kinetics and Mechanism, 2nd ed.;
John Wiley & Sons: New York, 1961; pp 137-139.
Mechanism. Unlike our study of C-aryl-N,N-dimeth-
ylhydrazones,3 we were unable to detect any intermedi-
ates in the reactions of the title compounds of this study.
J. Org. Chem, Vol. 70, No. 11, 2005 4391

Erden et al.
four reaction flasks were subjected to a positive pressure of
O2 and the contents stirred. Two of the flasks, one with TPP,
the other without, were cooled to -78 °C, and the other two
were left at room temperature (∼20 °C). After 30 min of
stirring, irradiation of the four vessels with a 250-W sodium
1
vapor lamp was begun. Irradiation was continued for 4 h. H
NMR analysis of the contents showed that the reaction
solutions without TPP were unchanged. All solutions with TPP
showed that the hydrazones had been partly or entirely cleaved
at -78 °C yielding arylaldehydes and N-aryl-N-methylnitro-
samines as the sole products, as demonstrated by comparison
with the authentic samples available from our synthetic
method. Most of the solutions (13 out of 21) containing TPP
and reacting at room temperature also showed these oxidation
products.
In a separate set of experiments reaction solutions were
prepared as above for a more quantitative study of the
temperature dependence of reaction efficiency. Hydrazones
with X ) Y ) H, and X ) CH3O, Y ) H were compared at -78
°C and at room temperature. The solutions were stirred under
a positive pressure of oxygen for times ranging from 30 to 120
min, then irradiated for only the amount of time necessary
for detection of product formation (∼1%) in the slower reaction
at room temperature. Percent-conversion at -78 °C exceeded
that at room temperature by a factor of at least 45:1 and did
not depend on the stirring time prior to irradiation (see
Results). All the other hydrazones were also compared at the
two temperatures, but, as described above, the irradiation
times were 4 h, and the reactions at -78 °C were almost
complete obviating quantitative comparisons.
However, by analogy with the earlier study, and with the
data gathered in this investigation, we postulate the
mechanistic scheme shown below. Reaction begins with
a bimolecular, exothermic step in which a complex is
formed between hydrazone and singlet oxygen. The
nature of such a complex has not been investigated. Then
either electron transfer from the hydrazone to 1O2 to form
a radical-ion caged pair (not trapped by free radical
scavengers) or direct formation of a bond between the
hydrazone carbon and an oxygen atom occurs. This step
is retarded by C-ring EWGs and accelerated by C-ring
EDGs. If electron transfer is a distinct step, it would be
followed by covalent bond formation. Whether electron
transfer or covalent bond formation is rate controlling,
it occurs with a transition state that is early with respect
to charge development. It is likely that covalent bond
formation, passing through a transition state structure
that is slightly tighter than that of the complex, is rate
controlling. The next step is unimolecular cyclization to
a 3-aza-1,2-dioxetane, which undergoes facile cleavage
to the observed products: an aryl aldehyde and a N-aryl-
N-methylnitrosamine. However, we cannot unambigu-
ously rule out initial complex formation or dioxetane
formation as the rate controlling event.
We can characterize the dye-sensitized photooxygenation
reactions as shown above in eq 1. The clean product composi-
tion stands in contrast to the reaction of C-aryl-N,N-dimeth-
ylhydrazones with singlet oxygen, which produces arylalde-
hydes, but also aromatic carboxylic acids, minute amounts of
their methyl esters, C-aryl-N-formyl-N-methylhydrazones, and
trace amounts of N-nitroso-N,N-dimethylamine.3
Experimental Section
An experiment was carried out as described above, with TPP
and at -78 °C, but with the singlet oxygen quencher DABCO
(1,4-diazabicyclo[2.2.2]octane) added to the reaction mixture.
No product formation was observed. This control verifies that
singlet oxygen is the actual oxidant in these reactions. In
another test the free radical scavenger 3,5-di-tert-butylphenol
was added. Its presence failed to inhibit reaction, showing that
free (uncaged) radicals do not lie on the pathway to products.
Competition Experiments. Competitive rates of reaction
with singlet oxygen were studied by allowing differently
substituted hydrazones, 0.05 M in CH2Cl2, to compete in a
pairwise fashion at -78 °C. A known concentration of cyclo-
hexane was present as an internal standard in order to assay
percent reaction. In all cases reactions were allowed to proceed
to 10% conversion or less in order to preserve the zero-order
nature of the kinetics. In some cases the two hydrazones were
in the same reaction vessel, in other cases in different vessels,
but run in parallel. Concordant results were obtained with the
two methods.
Materials. All hydrazones, except for two, were prepared
in excellent yields by the method of Yao and Resnick in which
a ring-substituted benzaldehyde is condensed with an N-aryl-
N-methylhydrazine.19 C-(4-Tolyl)-N-methyl-N-(4-nitrophenyl)-
hydrazone and C-(4-chlorophenyl)-N-methyl-N-(4-nitrophenyl)-
hydrazone were made by nitration of the corresponding
N-phenylhydrazones at 0 °C in acetic anhydride, using con-
centrated nitric acid. The necessary N-aryl-N-methylhydra-
zines were prepared by reduction of the corresponding N-aryl-
N-methylnitrosamines with zinc powder in acetic acid,20,21 and
the nitrosamines by treatment of N-aryl-N-methylamines with
nitrous acid.22 Melting points of hydrazones were in good
agreement with literature values where these were found.
Melting points and NMR features for those hydrazones not
found in the literature are given in the Supporting Informa-
tion. All NMR spectra were determined in CDCl3 solution. The
terms “major” and “minor” within the NMR data refer to the
major (syn or E) forms and the minor (anti or Z) forms.
Product Studies. Four solutions of each hydrazone were
made by dissolving the hydrazone (∼2 × 10-4 mol, accurately
weighed) in 4 mL of methylene chloride or in 4 mL of
methylene chloride containing 10-4 M tetraphenylporphyrin
(TPP). This produced 0.05 M solutions of each hydrazone, two
of them with TPP present, and two of them without TPP. All
Acknowledgment. This work was supported in part
by grants from the National Institutes of Health, MBRS
SCORE Program-NIGMS (Grant No. GM52588), and
the National Science Foundation (Grant No. CHE-
972900).
(19) Yao, H. C.; Resnick, P. J. Org. Chem. 1965, 30, 2832.
Supporting Information Available: Physical constants
and 1H and 13C NMR data for all previously unreported
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
(20) Organic Syntheses; Wiley: New York, 1943; Collect. Vol. II, p
418.
(21) Cook, J. W.; Loudon, J. D.; McCloskey, P. J. Chem. Soc. 1951,
1203.
(22) Organic Syntheses; Wiley: New York, 1943; Collect. Vol. II, p
460.
JO050014T
4392 J. Org. Chem., Vol. 70, No. 11, 2005