Conical intersection control of heterocyclic photochemical bond scission

Article abstract of DOI:10.1021/ja061851v

The photochemistry of the heterocycle 5,6-dihydro-1-methyl-5,5- diphenylpyridin-2(1H)-one (compound 1 in the text) leads to two competitive reactions (the reactions are depicted in the Introduction to the article). These arise from fission of bond a, betw

Full text of DOI:10.1021/ja061851v

Published on Web 09/12/2006
Conical Intersection Control of Heterocyclic Photochemical
Bond Scission1,2
Howard E. Zimmerman* and Oleg D. Mitkin
Contribution from the Department of Chemistry, UniVersity of WisconsinsMadison,
Madison, Wisconsin 53706
Received March 17, 2006; E-mail: zimmerman@chem.wisc.edu
Abstract: The photochemistry of the heterocycle 5,6-dihydro-1-methyl-5,5-diphenylpyridin-2(1H)-one
(compound 1 in the text) leads to two competitive reactions (the reactions are depicted in the Introduction
to the article). These arise from fission of bond a, between the nitrogen (N-6) and C-1, and bond b, between
C-4, with the two phenyl substituents, and C-5, adjacent to the nitrogen. Scission of bond a alone leads to
a zwitterionic intermediate which can be trapped by nucleophiles, while cleavage of bonds a and b together
affords two fragmentssa ketene and an imine. The ketene could be intercepted with nucleophiles and the
imine trimerized. Computation reveals little weakening of bonds a and b. But as stretching begins, conical
intersections are encountered, leading to ground-state products.
Scheme 1. Mechanism of the Type B Rearrangement
Introduction
Most recently we have encountered some intriguing new
photochemistry.
In contrast, most of our past photochemical research has
concentrated on carbocyclic systems, and heavily on their
rearrangements.3 In the present study we turned to the chemistry
of heterocyclic counterparts. The usual reaction of 4-aryl-
substituted cyclohexenones is the ubiquitous “Type-B enone
rearrangement”,4 in which an aryl group at C-4 migrates to C-34
(Scheme 1).
Less common, but fascinating, is the diversion of the reaction
course when carbon-5 is appropriately substituted. Thus, if there
is substitution at C-5 which is capable of stabilizing odd-electron
density at this carbon, then commonly5 scission of bond 4-5
results, as shown in eq 1, where an electron at C-5 is stabilized
by two phenyl groups. More recent research has provided some
aza-steroid examples which might be interpreted similarly.5c This
research has involved heterocyclic photochemistry, and the
question arose whether a nitrogen bonded to C-5 might be
playing a similar role. Our previous observation6 suggests that
possibility.
(1) This is Paper 282 of our General Series.
Results
(2) Zimmerman, H. E.; Cheng, J. J. Org. Chem. 2006, 71, 873-882.
(3) For recent references, see: (a) Zimmerman, H. E.; Nesterov, E. E. J. Am.
Chem. Soc. 2003, 125, 5422-5430. (b) Zimmerman, H. E.; Sereda, G. A.
J. Org. Chem. 2003, 68, 283-292. (c) Zimmerman, H. E.; Novak, T. J.
Org. Chem. 2003, 68, 5056-5066.
We began with a study of monocyclic heterocycles having
nitrogen substitution as part of the ring structure. The synthesis
is outlined in Scheme 2. The desired six-membered heterocyclic
ring closure was obtained nicely by an intramolecular Horner-
Emmons cyclization of phosphonate 16.
(4) (a) Zimmerman, H. E.; Wilson, J. W. J. Am. Chem. Soc. 1964, 86, 4036-
4042. (b) Zimmerman, H. E.; Rieke, R. D.; Scheffer, J. R. J. Am. Chem.
Soc. 1967, 89, 2033-2047. (c) Zimmerman, H. E.; Hancock, K. G. J. Am.
Chem. Soc. 1968, 90, 3749-3760.
(5) (a) Zimmerman, H. E.; Solomon, R. D. J. Am. Chem. Soc. 1986, 108, 6276-
6289. (b) Zimmerman, H. E.; Sam, D. J. J. Am. Chem. Soc. 1966, 88, 4114-
4116. (c) Canovas, A.; Fronrodona, J.; Bonet, J.-J. HelV. Chim. Acta 1980,
63, 2380-2389. (d) Zimmerman, H. E.; Morse, R. L. J. Am. Chem. Soc.
1968, 90, 954-966.
Two photoproducts, 20a and 21, were observed upon direct
photolysis of 1a in methanol (Scheme 3). Parallel photolyses
(6) Zimmerman, H. E.; Pushechnikov. A. Org. Lett. 2004, 6, 3779-3780.
9
10.1021/ja061851v CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 12743-12749
12743

A R T I C L E S
Zimmerman and Mitkin
Scheme 2. Synthesis of Heterocyclic Reactants
Scheme 3. Photochemistry of the Three Heterocycles
Scheme 4. Photolyses as a Function of Reaction Conditions
were run in tert-butyl alcohol with similar results. No identifiable
products were observed with benzene as a solvent. We note
that photoproduct 20a has arisen from fission of bond 4-5, a
result we encountered in one earlier study.6 However, we
observe that photoproduct 21 arises from fission of bond 1-6
in addition to 4-5. To ensure the isolation of all the products
of this reaction, compound 1b, having a bulky 4-tert-butylbenzyl
substituent at nitrogen, was prepared and photolyzed. Indeed,
the isolation of the trimeric product 22b proved possible.
In addition, the photolysis of heterocycle 1c, with an electron-
withdrawing group on nitrogen, was run in methanol. In this
case, beside the expected products 20c and 21, the bicyclic
compound 23c was identified, and it was the only product when
the reaction was run in benzene. The structure of product 23c
was established by X-ray analysis. We note that this is a
common Type B rearrangement product (Scheme 1). Such a
diversity in reactivity urged us to investigate the reaction
multiplicity. For this purpose, the photolysis of 1a in the
presence of acetophenone as a sensitizer was run in tert-butyl
alcohol. Bicyclic compound 23a was isolated then as the sole
product. Since it was an oil, its structure was established by
synthesis involving saponification and subsequent methylation
of compound 23c. Thus, we conclude that products 20, 21, and
22 derive from a singlet (i.e., S1) reaction, and product 23 comes
from a triplet.
reaction course, we turned to ab initio computations. We
employed CASSCF(6,6) with a 6-31g(d) basis set in Gaussian
2003.7 For practicality, two methyl groups were placed at C-4
rather than phenyl substituents, and these corresponding com-
pounds are named with an “M” appended. It is worth noting
that an active space of (6,6) (MOs 35-42) proved possible by
inspection of the MOs weighted appreciably in each reaction
step. Thus, although a total of eight electrons is involved
throughout the entire reaction sequence (C1dO7 and C2dC3
are lost, C1dC2 is formed, and C3sC4 is broken, as is C1s
N6), no more than six electrons are redistributed at any one
stage of the reaction. The validity was confirmed by inspection
of the bond weightings using the Weinhold8 NBO analysis as
well as by a natural orbital analysis of each species. Smooth
convergence resulted in all cases.
With the intent of trapping intermediates and helping to
establish the reaction mechanism, the photolysis of 1a in
methanol with the addition of potassium cyanide was performed.
Photoproduct 4 was formed and clearly had arisen from cyanide
attack in competition with methanol, as shown in Scheme 4.
The mechanism remains to be discussed (vide infra).
Results
(7) Frisch, M. J.; et al. Gaussian 03, Revision C. 02; Gaussian, Inc.:
Wallingford, CT, 2004.
(8) Weinhold, F. Natural Bond Orbital Methods. In Encyclopedia of Compu-
tational Chemistry; Schleyer, P. v. R., Ed.; Wiley: New York, 1998; Vol.
3, p 1792.
Computational Aspects and Methodology. In parallel with
our experimental efforts, and in order to better understand the
9
12744 J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006

Heterocyclic Photochemical Bond Scission
A R T I C L E S
Table 1. Computational Results
entry
molecular species
energy, hartrees
energy, kcala
S0 Species
1
2
3
4
5
6
conical intersection A
S0 geometry-optimized parent heterocycle
S0 with geometry of optimized S1, parent heterocycle
S0 with S1 transition-state geometry
conical intersection B
-439.68533
-439.89660
-439.88467
-439.81295
-439.72588
-439.82941
132.57
0.00
7.48
52.49
107.13
42.16
product pair S0 geometry optimized
S1 Species
conical intersection A, S1 surface
S1 with geometry of optimized S0 of the parent heterocycle
S1 optimized S1, parent heterocycle
S1 transition-state geometry
conical intersection B, S1 surface
7
8
9
10
11
12
-439.68829
-439.72236
-439.75189
-439.70678
-439.72588
-439.49641
130.71
109.33
90.81
119.11
107.13
251.12
product S1 with S0 optimized geometry
Further Conical Intersections
conical intersection C
conical intersection D
13
14
-439.70940
-439.67042
117.57
141.93
Relaxed Conicals B, C, and D as S0 (B3LYP)
15
16
17
relaxed conical B
relaxed conical C
relaxed conical D
-442.57828
-442.65740
-442.57790
Relaxed Conical B (CASSCF S0 - Local Minimum)
relaxed conical B (local minimum)
18
-439.8990
a Energies in kcal/mol are relative to reactant ground state (S0).
Scheme 5. The Role of Conical Intersections
For the present, we merely outline the computations done
and our reasoning. The conclusions derived are detailed below
in the Discussion section. Details are included in the Supporting
Information.
The first item was geometry optimization of the ground state,
S0, of the reactant heterocycle 1M (entry 2 in Table 1). We
then obtained the Franck-Condon S1 excited state (entry 8 in
Table 1), with an energy 109.34 kcal/mol above the S0 ground
state.
We next wanted the relaxed singlet S1. We note that CAS
S1 optimizations often prove difficult. Thus, a unique approach
involving “state averaging” was used, with the ratio of S1 to
S0 being increased incrementally from 1:1 until S1 was reached
(Table 1, entry 9), giving an 18.53 kcal/mol relaxation energy.
Interestingly, the S1 structure was similar in geometry to that
of the ground state S0, with bond lengths being somewhat
perturbed but bond 1-6 twisted by 15°.
Having the S1 excited-state wavefunction in hand, we looked
for evidence of weakening of bonds 1-6 and/or 4-5. For this,
the NBO8 analysis was employed with an NHO (hybrid orbital)
basis set. However, the electron density in the two bonds of
interest was normal and similar to those of the other molecular
bonds, but the system had developed a 1-2 π-bond, and lone
pairs were found at C-3 and the oxygen π-orbital. The NBO
analysis revealed the excited state 1M* to be π-π*.
further twisted by an additional 65° but, again, no indication of
weakening of bonds 1-6 and/or 4-5.
Hence, we proceeded to try stretching the two bonds,
separately and together, prior to further computation. On
stretching bond 4-5 of S1 to 1.8 Å and doing a conical
optimization, the geometry returned the same conical intersection
A. However, on further stretching, to 2.0 Å, a new conical
intersection was found with a severed bond 4-5 and an energy
of -439.7259 hartrees (107.13 kcal/mol above S0). This is
termed conical intersection B (entries 5 and 11 in Table 1),
which can lead to product or revert to reactant ground state (vide
infra; see Scheme 5).
We turned next to searching for conical intersections starting
with the excited singlet (i.e., S1) geometry. This led to a species,
conical A (entry 1 in Table 1), whose geometry was different
from that of the S1 excited state itself, with bond 1-6 being
A third approach involved stretching bond 1-6 of S1 to 1.8
Å. Computation then led to a new conical intersection with bond
1-6 severed. This we termed conical intersection C (-439.7094
9
J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006 12745

A R T I C L E S
Zimmerman and Mitkin
Scheme 6. Reaction Hypersurface
hartrees, 117.47 kcal/mol above S0, entry 13 in Table 1; see
Scheme 5 again).
intact (entry 18 in Table 1). With more common and larger step
sizes in geometry optimization of conical intersection B,
complete fragmentation resulted, as noted above. It is this
species, 31M, which undergoes nucleophilic attack by methanol
and by cyanide anion.
Discussion. We see that our original report on the effect of
stabilization of odd-electron density on carbon-5, leading to
scission of bond 4-5 of cyclohexenones, has proven a more
general phenomenon than originally thought. In the early work,
this stabilization was invariably achieved by virtue of phenyl
substitution. However, we now see that a ring nitrogen operates
in the same fashion.
Interestingly, in research by Canovas,5c ring-opening of
dihydropyridones with fission of bond 4-5 had been encoun-
tered but was ascribed to an electrocyclic process. In that
research, there was some evidence for ketene formation, which
was described as not subject to a mechanistic rationale.
Very recently, we have observed still another example of 4-5
bond fission6 (see eq 2). The commonality of fission of bond
A fourth computation began with both bonds 1-6 and 4-5
stretched to 1.8 Å. Interestingly, only bond 4-5 remained
severed, while bond 1-6 collapsed to give conical intersection
D (Table 1, entry 14, 141.93 kcal/mol above S0). Thus, a cluster
of conical intersectionssA, B, C, and Dshad been encountered,
depending on the initial molecular deformation. This is sum-
marized structurally in Scheme 5.
Next, it was of considerable interest to consider the S0 surface
for molecular geometries corresponding to the four conical
intersections A, B, C, and D. This promised to lead to an
understanding of the fate of molecules utilizing these conical
intersections. Ground-state S0 geometry optimization starting
with conformation C returned the geometry to that of the
ground-state heterocycle 1M. Dramatically, parallel ground-state
(i.e., S0) geometry optimization of structures B and D led the
molecule to completely fragment to species 3M and 32M, in
accord with experiment and the mechanistic interpretation in
Scheme 7 (below). Often, geometry optimization of unstable
species with an S0 configuration leads to an experimentally
observed product. This then permits us to consider a hypersur-
face, as depicted in Scheme 6.
In order to obtain a better understanding of the reaction
mechanism, we proceeded to a QST2 computation of the S1
reaction surface, beginning with 1M* (S1) and proceeding
onward to conical intersection B. To our knowledge, QST2
computations have hitherto been limited to ground-state surfaces.
With the geometry-optimized reactant S1 and conical intersec-
tion B as end points, an energy maximum (TS) was found,
interposed with a 28 kcal/mol activation energy from the
minimum S1. It is seen that vertical excitation of reactant 1M
(S0) leads to S1 near a conical intersection (conic A), 39.91
kcal/mol above the S1 minimum, and 21.38 kcal/mol above the
Franck-Condon S1. Passage of S1 via the 28 kcal/mol TS leads
to conical intersection B, which leads, via S0, to two of the
main initial photoproducts, ketene 3M and imine 32M. Scheme
6 thus shows a cross section of the hypersurfaces. However,
one species was still lacking, namely zwitterion 31M. Extensive
S0 computations with the geometry of conical intersection B
as a starting point but with a minimum geometry step size led
to zwitterion 31M as a local minimum, with bond 1-2 still
4-5 is striking. The evidence suggests that the initial step is
generally cleavage of this bond. This clearly is involved in the
route to product 20. Scheme 7 suggests that scission of bond
1-6, leading to ketene 3 and imine 32, occurs only after the
initial 4-5 fission via zwitterion 31. However, to some extent,
loss of bonds 1-6 and 4-5 may occur concertedly. Thus,
conical intersection B funnels to a critical point on the ground-
state surface. As noted in Scheme 6, that surface leads to
fragmentation to afford the ketene 3M and the imine 32M.
Analogously, stretching of bond 1-6 leads to the conical
intersection C, which provides another route to the ground-state
surface. In this case, S0 electronics leads the molecule back to
reactant 1M.
Turning now to the triplet photochemistry, we note that the
reaction affording 23a (cf. Scheme 4) is basically a Type B
process, occurring by bridging of a C-4 phenyl group. The triplet
9
12746 J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006

Heterocyclic Photochemical Bond Scission
A R T I C L E S
Scheme 7. Reaction Mechanisms
Experimental Section
General Procedures. Melting points were determined in open
capillaries and are uncorrected. Column chromatography was performed
on silica gel (60-80 mesh) mixed with 1% of a Sylvania 2282 green
phosphor and slurry-packed into quartz columns to allow monitoring
1
of zones with a hand-held UV lamp. H and 13C NMR spectra were
recorded at 300 and 75.4 MHz, respectively, with TMS as an internal
standard; 31P NMR sprectra were recorded at 121.4 MHz with 85%
H3PO4 as an external standard. Photolysis experiments were carried
out with a 400-W medium-pressure mercury lamp in a quartz immersion
well with 0.2 M copper sulfate filter solution.
2-Chloro-N-(3-hydroxy-2,2-diphenylpropyl)acetamide (14). A
solution of 0.30 mL (3.77 mmol) of chloroacetyl chloride in 5.0 mL
of dichloromethane was added dropwise to a stirred solution of 855
mg (3.77 mmol) of amino alcohol 1310 and 0.55 mL (3.9 mmol) of
triethylamine in 10 mL of dichloromethane at -10 to -5 °C. The
mixture was stirred for 4 h, quenched with water, extracted with
dichloromethane, washed with diluted HCl and brine, dried, concen-
trated in vacuo, and treated with ether to give 0.895 g (78%) of the
crystalline 2-chloro-N-(3-hydroxy-2,2-diphenylpropyl)acetamide (14),
mp 128-129 °C, after filtration and recrystallization from ethyl acetate.
1H NMR (CDCl3, 300 MHz): δ 7.17-7.37 (m, 10H), 6.69 (t, broad,
1H), 4.18 (d, J ) 6.3 Hz, 2H), 4.15 (d, J ) 6.3 Hz, 2H), 4.00 (s, 2H),
2.72 (t, J ) 6.3 Hz, 1H). 13C NMR (CDCl3, 75.4 MHz): δ 167.21,
143.47, 128.80, 127.97, 127.18, 67.26, 52.25, 45.50, 42.64. HRMS:
calcd for C17H18ClNO2, 326.0926 (M + Na); found, 326.0913 (M +
Na).
Scheme 8. Singlet-Triplet Reactivity Differences
2-Chloro-N-(2-formyl-2,2-diphenylethyl)acetamide (15). To a
solution of 2.5 mL (28.7 mmol) of oxalyl chloride in 50 mL of
dichloromethane at -70 °C under nitrogen was added 2.7 mL (38
mmol) of DMSO. The mixture was stirred for 15 min, and then 5.69
g (18.7 mmol) of amido alcohol 14 in 60 mL of dichloromethane was
added. The mixture was stirred at -70 °C for 20 min and treated with
10.5 mL (75.3 mmol) of triethylamine. The mixture was allowed to
warm to room temperature, washed with water, dried, and concentrated
in vacuo to give the product 15 in literally quantitative yield as a thick
1
oil which slowly crystallized; mp 64-65 °C. H NMR (CDCl3, 300
MHz): δ 9.89 (s, 1H), 7.33-7.44 (m, 6H), 7.17-7.21 (m, 4H), 6.83
(t, broad, 1H), 4.23 (d, J ) 6.0 Hz, 2H), 3.84 (s, 2H). 13C NMR (CDCl3,
75.4 MHz): δ 199.04, 165.62, 137.83, 129.30, 128.87, 128.33, 64.08,
42.85, 42.61. HRMS: calcd for C17H16ClNO2, 324.0767 (M + Na);
found, 324.0775 (M + Na).
Diethyl (2-Formyl-2,2-diphenylethylcarbamoyl)methylphosphonate
(16). A mixture of 0.81 g (2.7 mmol) of formylamide 15, 0.50 mL
(2.9 mmol) of triethylphosphite, and 0.403 g (2.7 mmol) of sodium
iodide was refluxed in 20 mL of acetonitrile for 6 h, diluted with water,
extracted with ether, washed with brine, dried, concentrated in vacuo,
and subjected to chromatography on a silica gel column (12.0 cm ×
2.5 cm, eluent hexane/ethyl acetate 3:1, then ethyl acetate). Fraction 2
gave 0.643 g (59%) of the product 16 as an oil. 1H NMR (CDCl3, 300
MHz): δ 9.90 (s, 1H), 7.30-7.42 (m, 6H), 7.19-7.23 (m, 4H), 6.82
(t, J ) 5.4 Hz, 1H), 4.22 (d, J ) 5.7 Hz, 2H), 4.01 (dq, J1 ) 7.2 Hz,
J2 ) 7.2 Hz, 4H), 2.69 (d, J ) 20.4 Hz, 2H), 1.25 (t, J ) 7.2 Hz, 6H).
13C NMR (CDCl3, 75.4 MHz): δ 198.93, 164.02, 163.97, 138.08,
129.18, 128.89, 128.07, 63.83, 62.76, 62.68, 43.24, 36.05, 34.31, 16.41,
16.34. 31P NMR (CDCl3, 121.4 MHz): δ 22.73. HRMS: calcd for
C21H26NO5P, 404.1627 (M + H); found, 404.1617 (M + H).
5,6-Dihydro-5,5-diphenylpyridin-2(1H)-one (17). A solution of 634
mg (1.57 mmol) of the phosphonate 16 in 5.0 mL of methanol was
added to a solution of sodium methoxide (from 72 mg (3.14 mmol) of
sodium in 10 mL of methanol), and the mixture was stirred overnight
at ambient temperature. The solvent was evaporated in vacuo, and the
residue was treated with water, extracted with benzene, washed, dried,
concentrated in vacuo, and subjected to chromatography on a silica
vs singlet reactivity is understood in terms of the large K-small
K (exchange integral) concept9 (Scheme 8). With an energy
separation of 2 K between S1 and T1, large-K species are
preferred by the triplet, while the small-K species are preferred
by the singlet. The Type B process of Scheme 1 does lead to a
first intermediate with diradical character and odd-electron
densities; in contrast, the ring fragmentation process has
zwitterionic character (Scheme 8a).
In conclusion, we see that there is a cooperative parallelism
between electron-pushing mechanistic treatment of photochemi-
cal reactions and an ab initio computational treatment. A most
interesting point in the present study is that the energy barriers
between the Franck-Condon excited state and the point where
the molecule reaches ground-state product would not have been
observed without the search for conical intersections.
(9) (a) Zimmerman, H. E.; Armesto, D.; Amezua, M. G.; Gannett, T. P.;
Johnson, R. P. J. Am. Chem. Soc. 1979, 101, 6367-6383. (b) Zimmerman,
H. E.; Factor, R. E. Tetrahedron 1981, 37, Suppl. 1, 125-141. (c)
Zimmerman, H. E.; Penn, J. H.; Johnson, M. R. Proc. Natl. Acad. Sci.
U.S.A. 1981, 78, 2021-2025. (d) The present situation is in accord with
the previously published examples.
(10) Meschino, J. A.; Bond, C. H. J. Org. Chem. 1963, 28, 3129-3134.
9
J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006 12747

A R T I C L E S
Zimmerman and Mitkin
gel column (4.0 cm × 2.5 cm, eluent hexane/ethyl acetate 1:2) to afford
238 mg (61%) of product 17; mp 166-167 °C after recrystallization
from butyl ether. 1H NMR (CDCl3, 300 MHz): δ 7.20-7.36 (m, 10H),
7.10 (d, J ) 10.2 Hz, 1H), 6.08 (dd, J1 ) 10.2 Hz, J2 ) 2.0 Hz, 1H),
5.82 (s broad, 1H), 3.97 (d, J ) 2.4 Hz, 2H). 13C NMR (CDCl3, 75.4
MHz): δ 165.94, 148.86, 144.32, 128.76, 127.67, 127.27, 123.94, 51.11,
48.39. HRMS: calcd for C15H17NO, 250.1232 (M + H); found,
250.1239 (M + H).
Alkylation of 5,6-Dihydro-5,5-diphenylpyridin-2(1H)-one (17):
General Procedure. A 71 mg (1.77 mmol) portion of sodium hydride
(60% in mineral oil) was washed free of oil with hexane and suspended
in 5.0 mL of DMF, and then 400 mg (1.61 mmol) of lactam 17 in 10
mL of DMF was added with external ice cooling and stirring. After
the evolution of hydrogen was complete, the appropriate halide (1.7
mmol) was added and the stirring continued overnight. The mixture
was diluted with water, extracted with benzene, washed, dried,
concentrated in vacuo, and subjected to chromatography or crystal-
lization to afford one of the following:
MHz): δ 172.52, 172.08, 144.18, 143.79, 141.76, 141.70, 139.36,
139.30, 129.65, 129.58, 128.23, 128.15, 128.05, 127.95, 127.58, 127.26,
125.90, 125.46, 121.71, 121.25, 80.98, 77.63, 55.65, 54.88, 34.89, 34.49,
33.49, 33.41. HRMS: calcd for C19H21NO2, 318.1470 (M + Na); found,
318.1456 (M + Na).
Fraction 3: N-Methyl-4,4-diphenylbut-3-enamide. A product of
hydrolysis of N-(methoxymethyl)-N-methyl-4,4-diphenylbut-3-enamide
on the column. Yield: 80 mg (6%); mp 128-129 °C. 1H NMR (CDCl3,
300 MHz): δ 7.15-7.42 (m, 10H), 6.26 (t, J ) 7.7 Hz, 1H), 5.56
(broad, 1H), 3.05 (d, J ) 7.7 Hz, 2H), 2.80 (d, J ) 4.8 Hz, 3H). 13C
NMR (CDCl3, 75.4 MHz): δ 171.71, 145.55, 141.88, 139.23, 129.79,
128.61, 128.34, 127.69, 127.66, 127.52, 121.29, 37.75, 26.57. HRMS:
calcd for C17H17NO, 252.1388 (M + H); found, 252.1387 (M + H).
Irradiation of 5,6-Dihydro-1-methyl-5,5-diphenylpyridin-2(1H)-
one (1a) in tert-Butanol. A solution of 400 mg (1.52 mmol) of
heterocycle 1a in 250 mL of tert-butanol was purged with nitrogen for
1 h. Irradiation was carried out in an enclosed box with a 0.2 M CuSO4
filter solution for 2 h. TLC indicated three new spots and practically
complete conversion of the starting material. The mixture was separated
by chromatography (12.0 cm × 2.5 cm column, eluent hexane/ether
9:1, then hexane/ethyl acetate 3:1, 1:1) to give the following products:
Fraction 1: tert-Butyl 4,4-Diphenylbut-3-enoate (25).12 Yield: 121
5,6-Dihydro-1-methyl-5,5-diphenylpyridin-2(1H)-one
(1a).
Yield: 361 mg (82%); mp 121-122 °C after recrystallization from
1
heptane. H NMR (CDCl3, 300 MHz): δ 7.24-7.35 (m, 6H), 7.15-
7.18 (m, 4H), 7.00 (d, J ) 9.9 Hz, 1H), 6.07 (d, J ) 9.9 Hz, 1H), 3.89
(s, 2H), 3.04 (s, 3H). 13C NMR (CDCl3, 75.4 MHz): δ 164.12, 146.52,
144.39, 128.78, 127.47, 127.28, 124.48, 58.92, 49.09, 34.37. HRMS:
calcd for C18H17NO, 264.1388 (M + H); found, 264.1392 (M + H).
1-(4-tert-Butylbenzyl)-5,6-dihydro-5,5-diphenylpyridin-2(1H)-
one (1b). Yield: 577 mg (91%); mp 162-163.5 °C after recrystalli-
1
mg (27%); mp 109-110 °C after recrystallization from hexane. H
NMR (CDCl3, 300 MHz): δ 7.17-7.40 (m, 10H), 6.24 (t, J ) 7.3 Hz,
1H), 3.07 (d, J ) 7.3 Hz, 2H), 1.45 (s, 9H). 13C NMR (CDCl3, 75.4
MHz): δ 171.39, 144.46, 142.36, 139.64, 130.07, 128.49, 128.30,
127.67, 127.49, 121.38, 80.91, 36.87, 28.31.
1
Fraction 2: N-(tert-Butoxymethyl)-N-methyl-4,4-diphenylbut-3-
enamide (24). Yield: 110 mg (21.5%) as an oil, mixture of isomers.
1H NMR (CDCl3, 300 MHz): δ 7.18-7.40 (m, 10H), 6.32, 6.37 (2t,
J ) 7.5 Hz, 1H), 4.47, 4.87 (2s, 2H), 3.16, 3.28 (2d, J ) 7.5 Hz, 2H),
2.84, 2.96 (2s, 3H), 1.08, 1.23 (2s, 9H). 13C NMR (CDCl3, 75.4
MHz): δ 172.39, 171.50, 144.25, 143.68, 142.21, 142.17, 139.81,
139.73, 130.02, 129.94, 128.51, 128.41, 128.39, 128.20, 127.99, 127.60,
127.58, 127.50, 127.43, 127.34, 122.54, 121.79, 74.11, 73.76, 72.92,
70.66, 35.41, 34.55, 33.65, 33.32, 28.11, 27.70. HRMS: calcd for
C22H27NO2, 360.1939 (M + Na); found, 360.1949 (M + Na).
Fraction 3: N-Methyl-4,4-diphenylbut-3-enamide. A product of
hydrolysis of N-(tert-butoxymethyl)-N-methyl-4,4-diphenylbut-3-en-
zation from isopropanol. H NMR (CDCl3, 300 MHz): δ 7.19-7.29
(m, 8H), 7.01-7.12 (m, 6H), 6.97 (d, J ) 9.7 Hz, 1H), 6.17 (d, J )
9.7 Hz, 1H), 4.59 (s, 2H), 3.82 (s, 2H), 1.32 (s, 9H). 13C NMR (CDCl3,
75.4 MHz): δ 163.88, 150.58, 146.70, 144.278, 133.54, 128.69, 128.66,
127.62, 127.18, 125.66, 124.82, 56.05, 49.76, 48.96, 34.69, 31.56.
HRMS: calcd for C28H29NO, 418.2147 (M + Na); found, 418.2131
(M + Na).
Methyl 5,6-Dihydro-2-oxo-5,5-diphenylpyridine-1(2H)-carboxy-
late (1c). Chromatography (10.0 cm × 2.5 cm column, eluent hexane/
ethyl acetate 3:1) afforded 256 mg (64%) of the starting material
(fraction 2) and 143 mg (29%) of the product (fraction 1); mp 140.5-
142 °C after recrystallization from benzene/heptane. 1H NMR (CDCl3,
300 MHz): δ 7.27-7.36 (m, 7H), 7.16-7.18 (m, 4H), 6.16 (d, J )
9.9 Hz, 1H), 4.47 (d, J ) 0.9 Hz, 2H), 3.88 (s, 3H). 13C NMR (CDCl3,
75.4 MHz): δ 162.68, 154.77, 151.36, 142.82, 128.92, 127.51, 127.46,
124.95, 54.25, 53.94, 49.12. HRMS: calcd for C19H17NO3, 330.1106
(M + Na); found, 330.1097 (M + Na).
1
amide on the column. Yield: 64 mg (16.7%); mp 128-129 °C. H
NMR (CDCl3, 300 MHz): δ 7.15-7.42 (m, 10H), 6.26 (t, J ) 7.7 Hz,
1H), 5.56 (broad, 1H), 3.05 (d, J ) 7.7 Hz, 2H), 2.80 (d, J ) 4.8 Hz,
3H).
Irradiation of 5,6-Dihydro-1-methyl-5,5-diphenylpyridin-2(1H)-
one (1a) in Methanol in the Presence of Potassium Cyanide.
Potassium cyanide (10.0 g) was stirred in 250 mL of methanol for 15
min and the solution filtered. To this solution was added 200 mg (0.76
mmol) of heterocycle 1a, and the solution was purged with nitrogen
for 1 h. Irradiation was carried out in an enclosed box with a 0.2 M
CuSO4 filter solution for 20 min. TLC showed no starting material
left. The resulting solution was concentrated in vacuo, diluted with
water, and extracted with benzene. The organic layer was separated,
washed with brine, dried, concentrated in vacuo, and subjected to
column chromatography (15.0 cm × 2.5 cm column, eluent dichlo-
romethane/ethyl acetate 19:1) to give the following products:
Fraction 1: Methyl 4,4-Diphenylbut-3-enoate (21). Yield: 6.8 mg
(3.5%).
Irradiation of 5,6-Dihydro-1-methyl-5,5-diphenylpyridin-2(1H)-
one (1a) in Methanol. A solution of 400 mg (1.52 mmol) of heterocycle
1a in 250 mL of methanol was first purged with nitrogen for 1 h.
Irradiation was then carried out in an enclosed box with a 0.2 M CuSO4
filter solution for 2 h. TLC indicated three new spots and practically
complete conversion of the starting material. The mixture was separated
by chromatography (12.0 cm × 2.5 cm column, eluent hexane/ether
9:1, then hexane/ethyl acetate 3:1, 1:1) to give the following products:
Fraction 1: Methyl 4,4-Diphenylbut-3-enoate (21).11 Yield: 129
1
mg (34%) as an oil. H NMR (CDCl3, 300 MHz): δ 7.17-7.41 (m,
10H), 6.25 (t, J ) 7.4 Hz, 1H), 3.70 (s, 3H), 3.17 (d, J ) 7.4 Hz, 2H).
13C NMR (CDCl3, 75.4 MHz): δ 172.48, 144.95, 142.12, 139.41,
129.95, 128.58, 128.33, 127.62, 120.50, 52.06, 35.46.
Fraction 2: N-(Methoxymethyl)-N-methyl-4,4-diphenylbut-3-
enamide (20a). Yield: 8.1 mg (3.6%).
Fraction 2: N-(Methoxymethyl)-N-methyl-4,4-diphenylbut-3-
enamide (20a). Yield: 142 mg (38%) as an oil, mixture of isomers.
1H NMR (CDCl3, 300 MHz): δ 7.19-7.42 (m, 10H), 6.32, 6.33 (2t,
J ) 7.4 Hz, 1H), 4.45, 4.80 (2s, 2H), 3.09, 3.28 (2s, 3H), 3.22, 3.28
(2d, J ) 7.0 Hz, 2H), 2.85, 2.98 (2s, 3H). 13C NMR (CDCl3, 75.4
Fraction 3: N-(Cyanomethyl)-N-methyl-4,4-diphenylbut-3-en-
1
amide (4). Yield: 31.5 mg (14.3%) as an oil. H NMR (CDCl3, 300
MHz): δ 7.16-7.43 (m, 10H), 6.27 (t, J ) 7.2 Hz, 1H), 4.33 (s, 2H),
3.21 (d, J ) 7.2 Hz, 2H), 2.95 (s, 3H). 13C NMR (CDCl3, 75.4 MHz):
(11) Frimer, A. A.; Ripstos, S.; Marks, V.; Aljadeff, G.; Hameiri-Buch, J.;
Gilinsky-Sharon, P. Tetrahedron 1991, 47, 8361-8372. Armesto, D.; Ortiz,
M. J.; Agarrabeitia, A. R.; Aparicio-Lara, S. Synthesis 2001, 8, 1149-
1158.
(12) Wong, J.-W.; Natalie, K. J., Jr.; Nwokogu, G. C.; Pisipati, J. S.; Flaherty,
P. T.; Greenwood, T. D.; Wolfe, J. F. J. Org. Chem. 1997, 62, 6152-
6159.
9
12748 J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006

Heterocyclic Photochemical Bond Scission
A R T I C L E S
δ 171.76, 145.11, 141.83, 139.47, 129.85, 128.71, 128.36, 127.78,
127.73, 127.61, 120.56, 115.43, 35.44, 35.40, 34.55. HRMS: calcd
for C19H18N2O, 893.4155 (3M + Na); found, 843.4119 (3M + Na).
Fraction 4: Methyl 2-(N-Methyl-4,4-diphenylbut-3-enamido)-
acetate (24). Yield: 79.3 mg (32.3%) as an oil; mixture of cis and
Fraction 2: Methyl 4,4-Diphenylbut-3-enoylmethoxymethylcar-
bamate (20c). Yield: 16.4 mg (9%) as an oil; one isomer.
1H NMR
(CDCl3, 300 MHz): δ 7.17-7.41 (m, 10H), 6.32 (t, J ) 6.9 Hz, 1H),
5.16 (s, 2H), 3.80 (s, 3H), 3.76 (d, J ) 6.9 Hz, 2H), 3.37 (s, 3H). 13C
NMR (CDCl3, 75.4 MHz): δ 174.69, 154.72, 144.80, 142.41, 139.58,
130.01, 128.52, 128.29, 127.74, 127.57, 127.51, 121.31, 75.26, 57.46,
54.11, 39.20. HRMS: calcd for C20H21NO4, 701.2833 (2M + Na);
found, 701.2827 (2M + Na).
1
trans isomers. H NMR (CDCl3, 300 MHz): δ 7.15-7.43 (m, 10H),
6.26, 6.32 (2t, J ) 7.2 Hz, 1H), 3.84, 4.12 (2s, 2H), 3.62, 3.74 (2s,
3H), 3.14, 3.25 (2d, J ) 7.2 Hz, 2H), 2.92, 2.97 (2s, 3H). 13C NMR
(CDCl3, 75.4 MHz): δ 172.16, 171.73, 169.88, 169.47, 144.34, 144.20,
142.14, 141.77, 139.66, 139.33, 129.94, 128.94, 128.77, 128.54, 128.42,
128.23, 127.97, 127.90, 127.62, 127.54, 127.44, 121.66, 121.61, 52.43,
52.24, 51.44, 49.55, 36.66, 35.51, 35.12, 34.75. HRMS: calcd for
C20H21NO3, 669.2941 (2M + Na); found, 669.2947 (2M + Na).
Irradiation of 5,6-Dihydro-1-methyl-5,5-diphenylpyridin-2-one
(1a) in tert-Butanol with Acetophenone as a Sensitizer. A solution
of 200 mg (0.76 mmol) of 5,6-dihydro-1-methyl-5,5-diphenylpyridin-
2-one (1a) and 1.8 mL (20 equiv) of acetophenone in 250 mL of tert-
butanol was purged with nitrogen for 1 h. Irradiation was then carried
out in an enclosed box with a 0.2 M CuSO4 filter solution for 20 min.
TLC showed no starting material left. The resulting solution was
concentrated in vacuo, and most of the acetophenone was removed in
vacuum with an oil pump. The residue was separated by column
chromatography (10.0 cm × 2.5 cm column, eluent hexane/ethyl acetate
3:1) to afford endo-3-methyl-5,6-diphenyl-3-azabicyclo[3.1.0]hexan-
Fraction 3: The starting material. Yield: 27.2 mg.
Fraction 4: Methyl endo-4-oxo-1,6-diphenyl-3-azabicyclo[3.1.0]-
hexane-3-carboxylate (23c). Yield: 63.5 mg (36.7).
Irradiation of 1-(4-tert-Butylbenzyl)-5,6-dihydro-5,5-diphenylpy-
ridin-2(1H)-one (1b) in Methanol. A solution of 350 mg (0.89 mmol)
of heterocycle 1b in 250 mL of methanol was purged with nitrogen
for 1 h. The irradiation was then carried out in an enclosed box with
a 0.2 M CuSO4 filter solution for 20 min. NMR indicated essentially
complete conversion of the starting material. Chromatography (12.0
cm × 2.5 cm column, eluent hexane/ethyl acetate 9:1, 7:1, 5:1, then
dichloromethane/ethyl acetate 29:1) gave the following products:
Fraction 1: Methyl 4,4-Diphenylbut-3-enoate (21). Yield: 120
mg (53.6%).
Fraction 2: 1,3,5-Tris(4-tert-butylbenzyl)-1,3,5-triazinane (22b).
Yield: 46 mg (29.7%); mp 130.5-131.5 °C after recrystallization from
isopropanol. 1H NMR (CDCl3, 300 MHz): δ 7.26 (q, J ) 8.4 Hz, 12H),
3.64 (s, 6H), 3.42 (broad, 3H), 1.30 (s, 27H). 13C NMR (CDCl3, 75.4
MHz): δ 150.00, 135.69, 128.81, 125.26, 74.10, 56.91, 34.64, 31.62.
HRMS: calcd for C36H51N3, 526.4161 (M + H); found, 526.4176 (M
+ H).
Fraction 3: N-(4-tert-Butylbenzyl)-N-(methoxymethyl)-4,4-diphen-
ylbut-3-enamide (20b). Yield: 81 mg (21.3%) as an oil; mixture of
cis and trans isomers. 1H NMR (CDCl3, 300 MHz): δ 7.09-7.40 (m,
13H), 6.93 (d, J ) 7.5 Hz, 1H), 6.32, 6.37 (2t, J ) 7.5 Hz, 1H), 4.62,
4.85 (2s, 2H), 4.38 (s, 2H), 3.23, 3.33 (2d, J ) 7.5 Hz, 2H), 3.05, 3.34
(2s, 3H), 1.30 (s, 9H). 13C NMR (CDCl3, 75.4 MHz): δ 173.31, 172.32,
150.59, 150.45, 144.39, 144.21, 142.07, 139.68, 139.62, 134.59, 133.61,
130.00, 129.89, 128.54, 128.27, 128.16, 128.07, 127.62, 127.57, 127.48,
126.32, 125.84, 125.70, 125.61, 125.41, 122.15, 121.81, 78.81, 76.28,
56.33, 55.25, 48.94, 48.06, 35.21, 34.95, 34.63, 31.51. HRMS: calcd
for C29H33NO2, 450.2409 (M + Na); found, 450.2428 (M + Na).
Fraction 4: N-(4-tert-Butylbenzyl)-4,4-diphenylbut-3-enamide. A
product of hydrolysis of N-(4-tert-butylbenzyl)-N-(methoxymethyl)-
4,4-diphenylbut-3-enamide on the column. Yield: 21 mg (16.7%). 1H
NMR (CDCl3, 300 MHz): δ 7.12-7.38 (m, 14H), 6.29 (t, J ) 7.8 Hz,
1H), 5.77 (broad, 1H), 4.40 (d, J ) 5.7 Hz, 2H), 3.08 (d, J ) 7.8 Hz,
2H), 1.31 (s, 9H). 13C NMR (CDCl3, 75.4 MHz): δ 170.85, 150.80,
145.85, 141.85, 139.27, 135.28, 129.84, 128.68, 128.40, 127.79, 127.73,
127.59, 125.85, 121.12, 43.60, 37.97, 34.72, 31.52. HRMS: calcd for
C27H29NO, 406.2147 (M + Na); found, 406.2164 (M + Na).
1
2-one (23a); 171 mg (85.5%) as a viscous oil. H NMR (CDCl3, 300
MHz): δ 7.26-7.43 (m, 10H), 3.66 (d, J ) 10.8 Hz, 1H), 3.26 (dd, J
) 10.8 Hz, J ) 1.7 Hz, 1H), 2.87 (d, J ) 8.7 Hz, 1H), 2.73 (dd, J )
8.7 Hz, J ) 1.7 Hz, 1H), 2.33 (s, 3H). 13C NMR (CDCl3, 75.4 MHz):
δ 172.12, 140.54, 133.68, 129.45, 129.06, 128.70, 127.51, 127.47,
127.41, 53.19, 33.52, 33.43, 33.11, 28.49. HRMS: calcd for C18H17-
NO, 549.2518 (2M + Na); found, 549.2515 (2M + Na). The structure
was confirmed by basic hydrolysis and methylation of methyl endo-
4-oxo-1,6-diphenyl-3-azabicyclo[3.1.0]hexane-3-carboxylate (23c) to
give an identical sample of endo-3-methyl-5,6-diphenyl-3-azabicyclo-
[3.1.0]hexan-2-one (23a).
Irradiation of Methyl 5,6-Dihydro-2-oxo-5,5-diphenylpyridine-
1(2H)-carboxylate (1c) in Benzene. A solution of 400 mg (1.52 mmol)
of heterocycle 1c in 250 mL of benzene was purged with nitrogen for
1 h. Irradiation was then carried out in an enclosed box with a 0.2 M
CuSO4 filter solution for 20 min. The NMR spectrum showed the
formation of only one new product with ca. 40% conversion. The
irradiation was then continued for a total of 90 min. Concentration in
vacuo and chromatography (10.0 cm × 2.5 cm column, eluent hexane/
ethyl acetate 3:1) afforded methyl endo-4-oxo-1,6-diphenyl-3-azabicyclo-
[3.1.0]hexane-3-carboxylate (23c), 359 mg (89.3%); mp 137-138.5
°C after recrystallization from isopropanol. 1H NMR (CDCl3, 300
MHz): δ 7.26-7.45 (m, 10H), 4.00 (d, J ) 11.7 Hz, 1H), 3.90 (dd, J
) 11.7 Hz, J ) 1.2 Hz, 1H), 3.62 (s, 3H), 3.03 (d, J ) 9.0 Hz, 1H),
2.84 (dd, J ) 9.0 Hz, J ) 1.2 Hz, 1H). 13C NMR (CDCl3, 75.4 MHz):
δ 171.29, 151.23, 139.27, 132.41, 129.28, 129.25, 129.08, 128.00,
127.93, 127.55, 53.52, 49.94, 34.18, 34.00, 33.30. HRMS: calcd for
C18H17NO3, 637.2315 (2M + Na); found, 637.2329 (2M + Na).
Irradiation of Methyl 5,6-Dihydro-2-oxo-5,5-diphenylpyridine-
1(2H)-carboxylate (1c) in Methanol. A solution of 200 mg (0.65
mmol) of heterocycle 1c in 250 mL of methanol was purged with
nitrogen for 1 h. Irradiation was then carried out in an enclosed box
with a 0.2 M CuSO4 filter solution for 40 min. The mixture was
chromatographed (12.0 cm × 2.5 cm column, eluent hexane/ethyl
acetate 7:1, then 4:1 and 2:1) to give the following products:
Fraction 1: Methyl 4,4-Diphenylbut-3-enoate (21). Yield: 11.3
mg (8%).
Acknowledgment. Support of this research by the National
Science Foundation is gratefully acknowledged, with special
appreciation for its support of basic research.
Supporting Information Available: Experimental procedures,
NMR spectra, X-ray data (in CIF format), computational files,
and complete ref 7. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA061851V
9
J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006 12749