Copper(I)-catalyzed [2 + 2] photocycloadditions with tethered linkers: Synthesis of syn-photodimers of dicyclopentadienes

Article abstract of DOI:10.1021/jo0055990

Cu(I)-catalyzed intramolecular photocycloadditions of diesters made of endo-dicyclopentadiene derivatives linked by the ester bonds with tethers are highly regio- and stereoselective and complete within hours, and the tethers can be easily cleaved afterward upon reduction with LiAlH₄. Irradiation of the diesters afforded a 1:1 mixture of the heretofore unknown exo-cis-exo dimer, originating from the (R,S/S,R) diastereomer of the diester and the exo-trans-exo, deriving frond the (R,R/S,S) diastereomer. The intermolecular photodimerization yielded, instead, only exo-transexo isomers and side products after irradiation for several days. The role of the tether's length and structure on the course of the photocycloadditions was investigated, and it was observed that short tethers introduce considerable strain in the products' framework. Adamantyl-containing tethers the shortest reaction times and highest yields. X-ray diffraction analysis of an exo-cis-exo provided stereoisomer containing adamantane in the tether exhibited an unusually close approach between H atoms on the methylene bridges and a long C-C distance in the cyclobutane ring. A rearrangement induced by X-ray irradiation was observed in this molecule.

Full text of DOI:10.1021/jo0055990

1
62
J . Org. Chem. 2001, 66, 162-168
Cop p er (I)-Ca ta lyzed [2 + 2] P h otocycloa d d ition s w ith Teth er ed
Lin k er s: Syn th esis of Syn -P h otod im er s of Dicyclop en ta d ien es
,
†
†
‡
†
Elena Galoppini,* Ravikrishna Chebolu, Richard Gilardi, and Wei Zhang
Department of Chemistry, Rutgers University, 73 Warren Street, Newark, New J ersey 07102, and
Laboratory for the Structure of Matter, The Naval Research Laboratory, Washington DC 20375
galoppin@andromeda.rutgers.edu
Received August 7, 2000
Cu(I)-catalyzed intramolecular photocycloadditions of diesters made of endo-dicyclopentadiene
derivatives linked by the ester bonds with tethers are highly regio- and stereoselective and complete
within hours, and the tethers can be easily cleaved afterward upon reduction with LiAlH .
4
Irradiation of the diesters afforded a 1:1 mixture of the heretofore unknown exo-cis-exo dimer,
originating from the (R,S/S,R) diastereomer of the diester and the exo-trans-exo, deriving from
the (R,R/S,S) diastereomer. The intermolecular photodimerization yielded, instead, only exo-trans-
exo isomers and side products after irradiation for several days. The role of the tether’s length and
structure on the course of the photocycloadditions was investigated, and it was observed that short
tethers introduce considerable strain in the products’ framework. Adamantyl-containing tethers
provided the shortest reaction times and highest yields. X-ray diffraction analysis of an exo-cis-exo
stereoisomer containing adamantane in the tether exhibited an unusually close approach between
H atoms on the methylene bridges and a long C-C distance in the cyclobutane ring. A
rearrangement induced by X-ray irradiation was observed in this molecule.
In tr od u ction
Polycarbocyclic cage molecules, a class of organic
compounds that has interested chemists for decades,
might be deconstructed in numerous different retrosyn-
thetic pathways. One of the simplest, that can be applied
to many cage frameworks, is cleavage into two cyclic
dienes “halves”.1
be accomplished with the aid of a tether. A cage-like
structure could then be formed either in a single step or
following a series of transformations, as illustrated in
Scheme 1B.
This strategy was employed by Gleiter and Karcher
for the synthesis of a cubane derivative: irradiation of a
syn-tricyclo[4.2.0.02 ]octa-3,7-diene bridged by C
,5
units
3
led to a photocyclic ring closure not observed in the
5
absence of the propane bridges. The use of tethers to
link two molecules and perform a reaction intramolecu-
larly is a strategy that is increasingly employed in a wide
variety of organic reactions, including copper(I)-catalyzed
[
2 + 2] photocycloadditions.6 This methodology is espe-
,7
cially practical when it uses “disposable” or “removable”
tethers, often silicon- or oxygen-based, that can be cleaved
after the reaction.8 The reaction rates are generally
higher, because of the decreased entropic demands, and
the reduction in the degrees of freedom of the unimo-
Dimerization of such cyclic dienes via syn-stereoselec-
tive [2 + 2] photocycloadditions may therefore open a
direct route to a variety of new cage and ring molecules.
The attempted syntheses of important cage molecules by
2
2a,3
dimerization of norbornenes, norbornadienes,
dicy-
(2) (a) Arnold, D. R.; Trecker, D. J .; Whipple, E. B. J . Am. Chem.
clopentadienes, triquinacenes,4 and other polycyclic
2b
Soc. 1965, 87, 2596. (b) Salomon, R. G.; Kochi, J . K. J . Am. Chem.
Soc. 1974, 96, 1137.
bridged alkenes1
a,b,e
were all based on this concept. These
(
3) (a) Marchand, A. P.; Earlywine, A. D.; Heeg, M. J . J . Org. Chem.
efforts, however, involved intermolecular reactions that
produced anti-fused dimers, Scheme 1A. A syn photocy-
cloaddition would require the face-to-face alignment of
the two reacting “halves” which, at least in theory, may
1
986, 51, 4096. (b) Acton, N.; Roth, R. J .; Katz, T. J .; Frank, J . K.;
Maier, C. A.; Paul, I. C. J . Am. Chem. Soc. 1972, 94, 5446. (c) Chow,
T. J .; Chao, Y.-S.; Liu, L.-K. J . Am. Chem. Soc. 1987, 109, 797. (d)
Chow, T. J .; Liu, L.-K.; Chao, Y.-S. J . Chem. Soc., Chem. Commun.
1
985, 700.
4) Codding, P. W.; Kerr, K. A.; Oudeman, A.; Soresen, T. S. J .
(
†
Rutgers University.
The Naval Research Laboratory.
Organomet. Chem. 1982, 232, 193.
(5) Gleiter, R.; Karcher, M. Angew. Chem., Int. Ed. Engl. 1988, 27,
840.
‡
(1) For reviews mentioning this approach, see: (a) Mehta, G.; Padma
S. In Carbocyclic Cage Compounds; Osawa E., Yonemitsu, O., Eds.;
VCH: New York, 1994; pp 183-215. (b) Mehta, G. In Strain and its
Implications in Organic Synthesis; de Meijere, A., Blechert, S., Eds.;
NATO ASI Ser., 273, 1989; pp 269-281. (c) Hopf, H. Classics in
Hydrocarbon Chemistry; Wiley-VCH: Wienheim, 2000; Chapter 4, pp
(6) The classification “Cu(I)-catalyzed [2 + 2] photocycloaddition”
is commonly used, but it does not necessarily imply that these are
concerted reactions which strictly follow the Woodward-Hoffmann
orbital symmetry rules.
(7) For reviews on this class of reactions, see: (a) Langer K.; Mattay,
J . In CRC Handbook of Organic Photochemistry and Photobiology;
Horspool, W. M., Song, P.-S., Eds.; CRC Press: Boca Raton, 1995; pp
84-101. (b) Salomon, R. G. Tetrahedron 1986, 42, 5753-5839.
(8) For a recent review, see: Gauthier, D. R.; Zandi, K. S.; Shea, K.
J . Tetrahedron 1998, 54, 2289.
4
1-51 and Chapter 5, pp 53-80. (d) Marchand, A. P. In Second
nd
Supplements to the 2
Edition of Rodd’s Chemistry of Carbon
Compounds; Sainsbury, M., Ed.; Elsevier Science: Amsterdam, 1994;
Vol. II, pp 277-329. (e) Forman, M. A.; Dailey, W. P. Org. Prep. Proc.
Int. 1994, 26, 291.
1
0.1021/jo0055990 CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/16/2000

Synthesis of Syn-Photodimers of Dicyclopentadienes
J . Org. Chem., Vol. 66, No. 1, 2001 163
Sch em e 1
The length of the alkyl chain tethers was varied
incrementally to study the effect on the reaction’s rate
and stereochemistry, while the purpose of the adamantyl
1
3
spacer was to make the linker more rigid. Since 1 was
prepared in racemic form,1
4,15
each diester was a mixture
of two diastereomeric pairs, (R,R/S,S) and (R,S/S,R). The
photocycloadditions were performed on this mixture,
except for 3d and 3e, where in each case the (R,R/S,S)
pair was isolated in pure form.
As the Cu(I) source we used CuOTf, which is often
preferred to copper halide salts in this type of reactions
because it is soluble in organic solvents and is photo-
stable. Cu(I) can coordinate to alkenes to form 1:1 or 1:2
Cu(I)-olefin complexes that exhibit UV absorption bands
(metal-to-ligand or ligand-to-metal charge-transfer bands)
in the range of 250-300 nm. According to the most
generally accepted mechanism, excitation of these bands
in the 1:2 complexes is the first step of a process that
7
leads to the photocycloaddition products. Depending on
lecular transition state results in higher regio- and
stereoselectivity. Isomers that are not observed in the
corresponding intermolecular reactions can often be
synthesized only with this method. For instance, the [2
the structure of the alkene and the copper salt employed,
the reaction can proceed through different mechanisms:
the photocyclization can occur via concerted, orbital-
symmetry allowed photochemical cycloadditions as well
as sequential pathways involving radical or cationic
copper alkyl intermediates.7
In ter m olecu lar P h otocycloaddition . For a compari-
son, we first studied the intermolecular photocycloaddi-
tion of untethered 4, Scheme 3.10 As had been observed
with other polycyclic bridged alkenes, including endo-
+
2] photocycloaddition of dihydropyridones, coumarins,
maleimides and numerous other molecules occurred with
syn stereoselectivity when performed intramolecularly
9
using tethered linkers. In view of these results, we have
applied this methodology to polycyclic bridged alkenes,
aimed at the synthesis of syn-fused dimers. In a prelimi-
nary study, we showed that linking two endo-dicyclopen-
2
b
dicyclopentadiene, irradiation of 4 in THF in the pres-
ence of a catalytic amount of CuOTf 16 for several days
produced two exo-trans-exo dimers, 5 and 6, in 62% yield
along with traces of other dimers (GC/MS) and decom-
position products. Dimers 5 and 6 are diastereomers
derived from the photocycloaddition between two opposite
and identical enantiomers of 4, respectively, and were
formed in 1:1 ratio. The configuration of each isolated
dimer was deduced through single-crystal X-ray structure
determinations (see the Supporting Information).
In tr a m olecu la r P h otocycloa d d ition s. The intramo-
lecular [2 + 2] photocycloadditions between tethered
diesters 3a -f were also performed in THF and in the
presence of a catalytic amount of CuOTf.16 In most cases,
these photocycloadditions were complete in a few hours,
whereas the corresponding intermolecular reactions re-
quired days. More interestingly, the photocycloaddition
afforded a completely different set of products because
of the constraints imposed by the tethers. For instance,
stereoisomers having the alcohol groups on opposite
sides, as in 5, did not form. In all cases, only two isomers
were recovered, the exo-trans-exo-7 and the exo-cis-exo-8
in 1:1 ratio and 70-80% yields, Scheme 4.17 The synthesis
of the exo-cis-exo-dicyclopentadiene dimers was unprec-
edented and demonstrates the usefulness of the tethers.
The photocycloaddition is highly stereospecific: the
exo-trans-exo isomers 7 are formed from the (R,R/S,S)
tadiene derivatives with a tether increased the efficiency
of the photocycloaddition reaction.10 Herein, we report a
comparison between several substrates with different
tethers and expand the study of the reaction stereose-
lectivity.
Resu lts a n d Discu ssion
Syn th esis of Teth er ed Diester s. Diesters 3a -f
shown in Scheme 2 served as the prototypes to test this
strategy. Each diester consists of two endo-dicyclopen-
tadiene moieties linked by a tether containing either an
alkyl chain or an adamantyl spacer. To reduce the
number of photocycloaddition products, and to simplify
the analysis of the stereoselectivity of the reaction, the
cyclopentene double bond in each moiety was reduced and
only the norbornyl double bond was left available for the
photocycloaddition.11 Diesters 3a -f were obtained in 70-
1
2
8
2
5% yield in one step from alcohol 1 and diacyl chlorides
a -f.
(
9) (a) Comins, D. L.; Lee, Y. S.; Boyle, P. D. Tetrahedron Lett. 1998,
3
1
9, 187. (b) Hertel, R.; Mattay J .; Runsink, J . J . Am. Chem. Soc. 1991,
13, 657. (c) Garner, P.; Cox, P. B.; Anderson, J . T.; Protasiewicz J .;
Zaniewski, R. J . Org. Chem. 1997, 62, 493. (d) De Schryver, F. C.;
Boens N.; Put, J . In Advances in Photochemistry; Pitts, J . N., Hammond
G. R., Collnick, K., Eds.; J ohn Wiley & Sons: New York, 1977; Vol.
1
0, pp 359-464.
10) Chebolu, R.; Zhang, W.; Galoppini, E.; Gilardi, R. Tetrahedron
Lett. 2000, 41, 2831.
11) The strained “norbornyl” double bond is the most reactive and
(
(
usually the “cyclopentenyl” double bond does not participate in the
photocycloaddition. However, depending on the reaction conditions,
there are two competing reactions that involve the cyclopentene
moiety: (a) Photopolymerization reactions, which can be avoided by
working with highly diluted solutions; (b) The intramolecular [2 + 2]
photocycloaddition between the norbornyl and the cyclopentenyl double
bond within each dicyclopentadiene moiety, leading to the formation
of bishomocubane. Although this competing reaction does not occur in
(13) In addition, the adamantane frame is transparent in the region
of the UV absorption spectrum where Cu(I)-olefin complexes absorb.
(14) The (R) and (S) designations refer to the alcohol stereocenter
of 1.
(15) The resolution of 1 was not performed, because it would result
in the diastereoselective synthesis of the (R,R) or the (S,S) diester,
and each is the precursor of the less interesting exo-trans-exo product.
(16) The photocycloaddition did not occur in the absence of CuOTf.
(17) Products having endo-cis-endo geometry were not observed. Ab
initio calculations using Gaussian 98 performed at 6-31G** level on
structures optimized using PM3 (as well as simple molecular models)
clearly show that these molecules would be highly strained.
2
b
THF in the presence of Cu(I), we did isolate a small amount of
bishomocubane products when the photocycloaddition was slow and
the dicyclopentadiene diesters were irradiated for many hours.
(12) Dilling W. L.; Plepys, R. A. J . Org. Chem. 1970, 35, 2971.

1
64 J . Org. Chem., Vol. 66, No. 1, 2001
Galoppini et al.
Sch em e 2
Sch em e 3
Sch em e 4
F igu r e 1. X-ray structure of 8f. The cis-fused ring system in
f is distorted, displaying an overall torsion that removes left-
8
right symmetry, and a long C12-C15 bond in the cyclobutane
ring (see text).
disrupt the crystal lattice and leads to a second structural
2
1
isomer in a fraction of the unit cells of the crystal.
Electron density maps show a spatial average over all
unit cells, so we see a superposition of the two isomers,
with 8f as the dominant species.
diastereomers of 3, and the exo-cis-exo isomers 8 derive
from the (R,S/S,R). Therefore, either anti- or, more inter-
estingly, syn-fused products can be selectively obtained
by starting with the appropriate diastereomer of the
diester.15 We found, however, that it is easier to separate
the mixture of 7 and 8 than a diastereomeric mixture of
Role of th e Str u ctu r e of th e Teth er . To understand
how the structure of the tether influences the course of
the photocycloaddition we used several linkers differing
in length and rigidity. In general, we observed that the
efficiency of the reaction depended dramatically on these
two parameters, while the stereoselectivity did not
change. The effect of the tether’s length is shown in Table
1, 3a -d : reaction rates increase as the aliphatic chains
become longer. The yields also improved with shorter
reaction times, as prolonged irradiation leads to decom-
position products. A possible explanation for the trend
shown in Table 1 may be found in the strain energy
content of the products. A correlation between calculated
strain energy values with the number of methylenic units
3
. The configuration of each isolated stereoisomer was
determined by single-crystal X-ray analysis or NMR.
The crystal structure of exo-trans-exo-7d 1 showed
diastereomer (R,R) and two molecules of (S,S) differing
in their torsion angles along the C6 tether-chain. The
crystal structure of exo-cis-exo 8f, shown in Figure 1,
exhibited several unusual features, including a very close
approach (1.74 Å) between H atoms on the methylene
bridges,18 C10 and C17, and a long C12-C15 distance
in the cyclobutane ring (1.68(1) Å). The distortion of the
central cyclobutane ring (C13-C14 ) 1.534(9) Å; C12-
C15 ) 1.677(7) Å; C13-C12 ) 1.547(8) Å; C14-C15 )
0
19,20
2 n
(CH ) (n ) 0-6) in the alkyl chain-tethered exo-trans-
(
18) The H atoms in 8f were placed in ideal sp3 locations. A similar
nonbonded distance, 1.61 Å, was reported in a polycycle by: Warner,
P. M.; LaRose, R. C.; Palmer, R. F.; Lee, C.-M.; Ross, D. O.; Clardy, J .
C. J . Am. Chem. Soc. 1975, 97, 5507.
1
.530(8) Å) suggests that the short semirigid tether and
(19) This result may be affected by the presence of 9 but it is not
the very close approach between the methylene bridges
introduce a considerable amount of strain in the mol-
ecules’ framework.
The collection and analysis of the X-ray data from 8f
show the occurrence of an X-ray-initiated rearrangement
of the strained ring system. This rearrangement does not
unprecedented: a cyclobutane ring distance of 1.720(4) Å is reported
in: Toda, F.; Tanaka, K.; Stein, Z.; Goldberg, I. Acta Crystallogr. 1996,
C52, 177.
(
20) For a review on long C-C single bonds in strained molecules:
Osawa, E.; Kanematsu, K. In Molecular Structure and Energetics;
Liebman, J . F., Greenberg, A., Eds.; VCH: Deerfield Beach, FL, 1986;
Vol. 3, pp 329-369.

Synthesis of Syn-Photodimers of Dicyclopentadienes
J . Org. Chem., Vol. 66, No. 1, 2001 165
Ta ble 1. Rea ction Tim esa a n d Yield sb for th e
Ta ble 2. Exp er im en ta la a n d Ca lcu la ted b (Sh ow n in
Ita lics) Bon d Len gth s (Å) in th e Cyclobu ta n e Rin g of
Exo-Tr a n s-Exo P r od u cts
P h otocycloa d d ition Rea ction s of 3a -f
a
1
Time required for disappearance of olefinic protons in H NMR
b
c
28
spectra. Isolated yield of 7 and 8. Isolated yield of 7.
exo product series indicates that the strain decreases as
n increases.22,23 Compound 6, which has no tether (n )
0
), is the least strained photocycloaddition product.
An additional indication that short tethers introduce
tension in the molecules’ frameworks derives from ex-
perimental and calculated values of C-C bond distances
in the central cyclobutane ring of exo-trans-exo products,
Table 2.24 These data show that the central cyclobutane
ring is not symmetrical in the tethered molecules. Specif-
ically, C-C bond distances obtained from the X-ray
a
From X-ray crystallographic data. b From semiempirical cal-
culations (AM1and PM3).24
Finally, the rigidity of the tether introduced by the
adamantyl groups also plays a role in this reaction, as
suggested by the difference in rates of products formation
between 3b and 3f, Table 1. Both diesters have the two
reacting moieties separated by a three-carbon atom
spacer, but the rate of photocycloaddition for 3f, in which
the spacer incorporates a rigid adamantane frame, is nine
times faster than that of 3b. One possible explanation is
that conformational restrictions introduced by the
semirigid 1,3-adamantyl spacer favor the formation of a
Cu(I) complex bridging the two alkenes in the transi-
tion state. CuOTf shows a strong tendency to form 1:2
Cu(I)-olefin complexes,7 where the metal bridges two
olefins. In particular, this has been observed for dicyclo-
pentadiene, where only the more strained “norbornyl”
1
0
crystallographic data analysis show that in 7d and 7e
see the Supporting Information) the cyclobutane bond
(
on the side of the tether (labeled as a in Table 2) is
slightly shorter than the one on the opposite side (bond
b) while in 5 and 6, which do not contain a tether, a and
b are identical. We examined only exo-trans-exo products
to correlate the tethers’ length with the strain energy
because this association can be established more clearly
in the trans compounds, which lack the additional strain
due to close contacts between the methylene bridges as
seen in the exo-cis-exo products.
a
(21) A possible structure for the rearrangement product is 9. This
was suggested by the final difference maps taken from a crystal that
had been irradiated for 4 days, showing the appearance of two peaks
25
double bond coordinates with Cu(I), and for norbornene.
In our case, however, it is not straightforward to assume
the same kind of structure for the Cu(I)-diester 3 com-
plexes. The ester group could play a role in coordinating
with the metal and the bulky tethers may introduce steric
factors not present in unsubstituted dicyclopentadiene.
Although our observations on the role of the linker are
limited to experiments with one kind of polycyclic alkene,
we believe that the use of adamantyl tethers will be
useful in future studies. In addition to more efficient
photocycloadditions, the isolation and characterization
of all adamantyl-containing compounds were consider-
ably easier than that of compounds with aliphatic
chains: the separation of 7 and 8 on silica gel was
(
C15′ and C16′) in the vicinity of C15 and C16, and by the observation
that atoms C12 to C23 have counterparts in both 8f and the rearranged
product.
2
6
increased; 8f and 7e formed crystals of good quality,
and the aliphatic region in the 1H NMR spectra was
somewhat simplified.
Additional features in favor of 9 are the observation that the rear-
ranged product is a structural isomer of 8f, the fact only two bonds
Clea va ge of Teth er . Finally, it was necessary to
verify that the tethers are really “disposable” and that
they can be cleaved after the intramolecular reaction.
This was readily accomplished by reduction with
(C12-C15 and C16-C20) need be “disconnected” and “reconnected”
22
to obtain 9 from 8f, and preliminary calculations, indicating that the
replacement of the four-membered ring with a cyclopentane would
relieve strain in the molecule. However, we currently do not have any
additional experimental evidence supporting 9.
2
7
(22) Strain energy values for 8f (176 kcal/mol) and 9 (122 kcal/mol)
LiAlH
4
.
Thus, 7 and 8 afforded diols 10 and 11 in high
and for the alkyl chain-tethered exo-trans-exo product series were
calculated using single-point molecular mechanics calculations (SYBYL)
using PC Spartan Pro, Wavefunction, Inc.
yield, respectively, Scheme 5.
(
23) Ab initio calculations of the final photocycloaddition products
(25) Salomon, R. G.; Kochi, J . K. Am. Chem. Soc. 1973, 95, 1889.
(26) In general, the compounds with an alkyl chain as the tether
did not form crystals. Only 7d was crystallized and the crystals were
extremely small.
using Gaussian 98 performed at 6-31G** level on structures optimized
using PM3 are in progress.
(24) The calculated bond lengths reported in Table 2 were obtained
from structures minimized using semiempirical calculations (AM1 and
PM3) using PC Spartan Pro, Wavefunction, Inc.
(27) Standard acidic or basic hydrolysis of the esters led to decom-
position products.

1
66 J . Org. Chem., Vol. 66, No. 1, 2001
Galoppini et al.
solution of alcohol 112 (3.3 mmol) in benzene (15 mL) and
pyridine (1 mL) under nitrogen atmosphere. After the addition
was complete, the mixture was heated to reflux, with stirring,
for about 10 h. The reaction mixture was cooled to room
temperature and diluted with benzene. The organic layer was
washed with 10% HCl (aq), water, and brine and then dried
Sch em e 5
2 4
over Na SO . The solvent was removed in vacuo to afford a
crude oil. Column chromatography (silica gel, hexane/ethyl
acetate, 100-95/0-5 v/v depending on the product) afforded
pure 3 as a (R,R/S,S) and (R,S/S,R) diastereomeric mixture
1
in 75-80% yields. 3a : colorless oil; H NMR δ 6. 16 (m, 4H),
5
.09 (m, 2H), 3.00 (m, 2H), 2.84 (m, 4H), 2.69 (m, 6H), 1.81
1
3
(m, 2H), 1.53-1.64 (m, 4H), 1.32-1.40 (m, 6H). C NMR δ
1
2
3
2
2
72.1, 137.5, 134.1, 77.6, 52.0, 49.1, 46.7, 45.7, 44.5, 32.6, 29.3,
Con clu sion s. In summary, we have described a
+
4.3; HRMS (FAB) calcd for C24
30 4
H O 382.4710, found (MH )
synthetic methodology for the formation of photodimers
of endo-dicyclopentadienes, notably the previously un-
known syn-fused isomers. The reaction is highly stereo-
and regioselective and the configuration of the cyclobu-
tane rings (either syn or anti) can be controlled. Finally,
adamantyl-containing tethers provided the best results
and may prove to be particularly useful linkers in future
studies. We are now extending this study to different
polycyclic alkenes and dienes to establish whether these
results can be generalized.
1
83.5221. 3b: colorless oil; HNMR δ 6.10 (m, 4H), 5.02 (m,
H), 2.95 (m, 2H), 2.79 (m, 2H), 2.74 (m, 2H), 2.63 (m, 2H),
.39 (t, 2H, J ) 7.5 Hz), 1.98 (m, 2H), 1.76 (m, 2H), 1.42-1.62
13
(
m, 6H), 1.28-1.37 (m, 6H); C NMR δ 172.6, 137.2, 134.0,
77.2, 51.9, 49.0, 46.8, 45.6, 44.4, 33.4, 32.6, 24.2, 20.2; HRMS
+
(FAB) calcd for C25
H
32
O
4
396.4976, found (MH ) 397.2392.
c: colorless oil; H NMR δ 6.11 (m, 4H), 5.06 (m, 2H), 2.95
m, 2H), 2.82 (m, 2H), 2.76 (m, 2H), 2.65 (m, 2H), 2.38 (m,
1
3
(
4
1
4
H), 1.80 (m, 2H), 1.72 (m, 4H), 1.50 (m, 4H), 1.39 (m, 4H),
.30 (m, 2H); 13C NMR δ 172.5, 137.4, 134.1, 77.2, 52.0, 49.1,
7.0, 45.7, 44.5, 34.2, 32.7, 24.5, 24.3; HRMS (FAB) calcd for
+
C
26
H
34
O
4
410.5242, found (MH ) 411.2554. 3d : (R,R/S,S)
1
Exp er im en ta l Section
diastereomer; H NMR δ 6.06 (m, 4H), 4.97 (m, 2H), 2.92 (m,
2
7
H), 2.76 (m, 2H), 2.69 (m, 2H), 2.59 (m, 2H), 2.26 (t, 4H, J )
.5 Hz), 1.72 (m, 2H), 1.41-1.62 (m, 8H), 1.24-1.36 (m, 12H);
Gen er a l Meth od s. NMR spectra were obtained on a Varian
1
INOVA 500 spectrometer operating at 499.90 MHz for H and
13
C NMR δ 173.3, 137.3, 134.0, 77.1, 52.0, 49.1, 46.9, 45.7, 44.5,
4.3, 32.7, 28.7, 24.8, 24.3; HRMS (FAB) calcd for C28
1
3
1
24.98 MHz for C and collected at ambient probe tempera-
3
4
7
38 4
H O
H O : C,
38 4
1
ture in CDCl
3
, unless otherwise specified. H spectra were
+
38.5773, found (MH ) 439.2834. Anal. Calcd for C28
1
3
referenced to internal tetramethylsilane and C spectra to the
central line of the solvent. Chemical shifts are reported to a
precision of ( 0.01 ppm for proton and (0.1 ppm for carbon.
Proton coupling constants (J ) are reported to a precision of
6.68; H, 8.73. Found: C, 76.61; H, 8.75. 3e: (R,R/S,S)
1
diastereomer: mp 145-146 °C; H NMR δ 6.12 (m, 4H), 5.02
(
(
4
4
m, 2H), 2.93 (m, 2H), 2.79 (m, 4H), 2.64 (m, 2H), 2.07-2.12
m, 6H), 1.78 (m, 2H), 1.50-1.62 (m, 16H), 1.30-1.39 (m,
(
0.1 Hz. High-resolution mass spectral data were obtained
13
H); C NMR δ 171.5, 137.5, 134.1, 76.9, 52.1, 49.1, 48.6,
from a commercial facility (the Michigan State University
Mass Spectrometry Facility). Elemental microanalytical data
have been obtained from a commercial facility (Robertson
Microlit Laboratories, Inc., Madison, NJ ). Low-resolution mass
spectra were obtained on a GC/MS HP6890 connected to a
HP5973 MSD detector (EI); major ions are reported to unit
mass and their intensities are reported parenthetically as a
percentage of the strongest peak. Melting points were deter-
mined by DSC and the thermograms were performed on a
Perkin-Elmer Pyris 1 (scan rate ) 10 °C/min). 1,3-Adamantane
diacetic acid, 1,3-Adamantane dicarboxylic acid, Copper(I)-
7.4, 47.0, 45.9, 44.7, 41.4, 35.7, 33.3, 32.8, 28.8, 24.4; HRMS
(
FAB) calcd for C32
H
44
O
4
516.3240, found (MH+) 517.3314.
Anal. Calcd for C32
H
44
O
4
: C, 79.03; H, 8.58. Found: C, 78.24;
1
H, 8.48. 3f: mp 172-173 °C; H NMR δ 6.16 (m, 4H), 5.06
(
m, 2H), 3.0 (m, 2H), 2.86 (m, 2H), 2.79 (m, 2H), 2.55 (m,
2
2
H), 2.23 (m, 2H), 2.10 (m, 2H), 1.95 (m, 8H), 1.80 (m,
H), 1.75 (m, 2H), 1.48 (m, 4H), 1.49 (m, 4H), 1.31 (m, 2H);
C NMR δ 176.7, 137.5, 134.3, 77.2, 51.8, 48.9, 47.1, 45.7,
4.3, 40.9, 40.1, 38.1, 35.4, 32.5, 27.9, 24.2; HRMS (FAB)
1
3
4
calcd for C32
Calcd for C32
.31.
In ter m olecu la r P h otocycloa d d ition of 4. A 50 mg
H
H
40
O
4
488.2927, found (MH+) 489.2991. Anal.
: C, 78.65; H, 8.25. Found: C, 78.58; H,
40
O
4
trifluoromethanesulfonate benzene complex (CuOTf)
2
6 6
‚C H ,
8
oxalyl chloride and anhydrous pyridine were purchased from
Aldrich. Diacid chlorides 2a -d were directly purchased by
Aldrich or Acros, or prepared from the commercially available
R,ω diacids and oxalyl chloride as indicated below. Anhydrous
THF (Aldrich) was distilled from sodium/benzophenone im-
mediately prior to use. Benzene was dried by removal of water
azeotrope and distillation over sodium. Column chromatogra-
phy was performed using silica gel (Selecto Silica Gel 230-
portion of 4 (0.3 mmol) was dissolved in dry THF (30 mL) and
put in a quartz tube shielded by a Vycor filter. A catalytic
amount (about 5-8 mol %) of (CuOTf)
2
6 6
‚C H was added, and
the tube was closed with a septum and the solution was
irradiated with a 450 W Hanovia medium-pressure Hg lamp
placed in a quartz well. The well was positioned at 3 cm
distance from the quartz tube containing the solution. The
reaction was monitored by GC/MS. After 56h the solvent was
removed under reduced pressure and the products were
separated by silica gel column chromatography (hexane/ethyl
acetate, 98/2) and recrystallized from hexane. The total yield
of 5 and 6 is 62% (31 mg) and the ratio of 5:6 was 1:1. 5: mp
180-182 °C; MS m/z 328 (M+, 25), 197 (55), 164 (46), 132
6
00 mesh). In each case, the crude product was first purified
by elution through a first short column, and the purified
products mixture was then separated using a long (about 30
cm), thin (1.3 cm internal diameter) column. TLC was per-
formed using Whatman silica gel plates, using iodine and/or
a 10% ethanolic solution of phosphomolybdic acid as the
developing agents. All reactions were performed under a
nitrogen atmosphere in rigorously dry glassware (oven-dried
and/or flamed under vacuum).
Gen er a l P r oced u r e for P r ep a r a tion of Diester s 3a -f.
To a stirred solution of the diacid 2a ,b (1.2 mmol) in benzene
2 mL) was added pyridine (0.2 mL). To this solution was
added oxalyl chloride (0.3 mL, 4 mmol) in benzene (1 mL), and
the resulting mixture was heated to reflux for 30 min. Benzene
was distilled off, and any residual oxalyl chloride was removed
in vacuo. The resulting acid chloride was dissolved in benzene
1
(100), 129 (75), 67 (85); HNMR δ 3.74 (m, 2H), 3.36 (s, 6H),
2.36 (m, 2H), 2.28 (m, 2H), 2.21 (d, 2H, J ) 5.5 Hz), 2.15 (d,
2H, J ) 10.0 Hz), 2.06 (d, 2H, J ) 4.0 Hz), 2.01 (d, 2H, J )
5.0 Hz), 1.85 (m, 4H), 1.55 (m, 2H), 1.38 (m, 4H), 1.32 (d, 2H,
J ) 10.0 Hz); 13CNMR δ 84.5, 57.9, 45.2, 44.1, 42.7, 42.0, 40.2,
38.6, 37.4, 32.7, 22.7. 6: mp 125-127 °C; MS m/z 328 (M+,
(
1
26), 197 (75), 164 (43), 132 (95), 129 (100), 67 (96); HNMR δ
3.74 (m, 2H), 3.36 (s, 6H), 2.34 (m, 2H), 2.28 (m, 2H), 2.21 (d,
2H, J ) 5.5 Hz), 2.15 (d, 2H, J ) 10.0 Hz), 2.11 (d, 2H, J )
4.5 Hz), 1.95 (d, 2H, J ) 5.0 Hz), 1.83 (m, 4H), 1.55 (m, 2H),
(
3 mL) and transferred to a flask containing an ice-cold

Synthesis of Syn-Photodimers of Dicyclopentadienes
J . Org. Chem., Vol. 66, No. 1, 2001 167
1
4
.38 (m, 4H), 1.31 (d, 2H, J ) 10.0 Hz); 13CNMR δ 84.5, 57.9,
5.0, 44.1, 42.7, 42.1, 40.8, 38.1, 37.4, 32.7, 22.7.
Gen er a l P r oced u r e for [2 + 2] In tr a m olecu la r P h oto-
room temperature and stirred for another 30 min. Excess
LiAlH
4
was cautiously destroyed by dropwise addition of ethyl
SO satd. solution. The
acetate and 1-2 drops of aqueous Na
reaction mixture was then extracted with ethyl acetate and
washed with water and brine, and the organic layer was dried
2
4
cycloa d d ition s. The photocycloaddition conditions were op-
timized for each diester by first performing the reaction in a
quartz NMR tube with THF-d
1
8
and monitoring by H NMR
2 4
over Na SO . The solvent was removed in vacuo to yield a solid
spectroscopy. Each reaction was then repeated on a larger
scale as follows. Diester 3 (0.1 mmol) was dissolved in dry THF
residue which was purified by column chromatography (hex-
ane/ethyl acetate, 60/40 v/v) to obtain pure 10 (from 7d ,e) or
11 (from 8c,f) (7 mg, 84% yield) as a crystalline solid. 10: mp
(20 mL) and placed in a dry quartz tube shielded with a Vycor
1
filter. A catalytic amount (about 5 mol %) of (CuOTf) ‚C
2
6
H
6
202-204 °C; H NMR δ 4.22 (m, 2H), 3.09 (m, 2H), 2.79 (m,
was added and the solution was irradiated with a 450 W
Hanovia medium-pressure Hg lamp placed in a quartz well.
The well was positioned at 2-3 cm distance from the quartz
tube containing the solution. The reaction was monitored by
2H), 2.66 (d, 2H, J ) 11 Hz), 2.36 (m, 6H), 2.23 (m, 2H), 1.92
(m, 2H), 1.55-1.66 (m, 6H), 1.41 (m, 2H), 1.36 (d, 2H, J ) 11
Hz); 13C NMR δ 75.1, 49.9, 45.2, 42.5, 42.3, 41.9, 41.1, 40.1,
36.4, 22.6. HRMS (FAB) calcd for C20
28 2
H O 300.2089, found
1
TLC and H NMR spectroscopy (disappearance of olefinic
(M+) 300.2029 (0.3 m); (MH+) 301.2161 (-0.7 m); (M-1)
1
protons). After the reaction was complete, THF was removed
in vacuo and the crude residue purified by silica gel column
chromatography (silica gel, hexane/ethyl acetate, 100-95/0-5
299.2009. 11: mp 190-192 °C; H NMR δ 4.22 (m, 2H), 3.09
(m, 2H), 2.79 (m, 2H), 2.66 (d, 2H, J ) 10.5 Hz), 2.36 (m, 6H),
2.23 (m, 2H), 1.92 (m, 2H), 1.55-1.66 (m, 6H), 1.41 (m, 2H),
2
8
13
v/v depending on the product) to afford pure 7 and 8. 7b:
1.36 (d, 2H, J ) 11 Hz); C NMR δ 75.1, 49.9, 45.2, 42.5, 42.3,
1
mp 116-118 °C; HNMR δ 5.08 (m, 2H), 2.76 (m, 2H), 2.46
41.9, 41.1, 40.1, 36.4, 22.6; HRMS (FAB) calcd for C20
300.2089, found (M ) 300.2097.
H
28
O
2
+
(
m, 4H), 2.44 (m, 6H), 2.28 (m, 4H), 2.07-2.17 (m, 6H), 2.00
1
3
(m, 2H), 1.80 (m, 2H), 1.64 (m, 2H), 1.48 (m, 4H); C NMR δ
Exp er im en ta l Cr ysta l Da ta . Intensity data were mea-
sured on a Bruker diffractometer with CuKR radiation (λ )
1.54178 Å). Structures were solved by direct methods, aided
by program XS, and refined with full-matrix least-squares
1
2
3
2
6
73.4, 79.6, 50.2, 44.3, 42.3, 39.7, 38.4, 37.4, 34.7, 34.6, 21.8,
+
0.8; HRMS (FAB) calcd for C25
32 4
H O 396.4976, found (MH )
1
97.2391. 8b: mp 95-97 °C; HNMR δ 5.00 (m, 2H), 3.23 (m,
H), 2.87 (m, 2H), 2.58 (m, 4H), 2.37 (m, 4H), 2.13-2.34 (m,
H), 1.96-2.06 (m, 8H), 1.57 (m, 2H), 1.41 (m, 4H); 13C NMR
2
9
program XL, from SHELXTL. The XL program minimizes
2
F
differences, and uses all data; i.e., weak data are not
δ 172.6, 78.0, 50.4, 48.3, 44.6, 42.3, 41.5, 39.9, 37.8, 36.0, 35.7,
2
considered “unobserved” during refinement. For 5: C H32O2,
2
2
3.3, 22.5; HRMS (FAB) calcd for C25 396.4976, found
H
32
O
4
FW ) 328.48, triclinic space group P(-1); a ) 6.9036(3) Å, b
) 7.6249(4) Å, c ) 8.9770(7) Å, R ) 87.060(7)°, â ) 70.192(5)°,
+
1
(
2
2
MH ) 397.2393. 7c: mp 120-121 °C; H NMR δ 4.87 (m, 2H),
3
.84 (m, 2H), 2.41 (m, 2H), 2.35 (m, 2H), 2.26 (m, 2H), 2.17 (d,
γ ) 78.226(5)°, V ) 435.15(5) Å , Z ) 1, and D(X-ray) ) 1.253
-
3
H, J ) 9.5 Hz), 2.14 (m, 2H), 2.03-2.09 (m, 6H), 1.75-1.82
mg mm . 1589 data measured to a 2θ max of 115° at T ) 294
K. R ) 0.0359 for 1157 reflections with [I > 2σ(I)], R ) 0.0367,
wR2 ) 0.1017 for all 1192 unique reflections. For 6: C H O ,
1
3
(
m, 6H), 1.58-1.69 (m, 4H), 1.43-1.51 (m, 4H); C NMR δ
1
3
4
74.2, 79.6, 46.8, 43.6, 43.3, 42.9, 39.4, 38.4, 37.0, 36.0, 33.2,
7.5, 28.0, 22.6, 21.7, 14.1; HRMS (FAB) calcd for C26
2
2
32
2
H
34
O
H
4
FW ) 328.48, triclinic space group P(-1); a ) 7.7130(4) Å, b
) 8.9995(7) Å, c ) 13.5820(9) Å, R ) 106.814(6)°, â ) 92.935-
+
1
10.5242, found (MH ) 411.2549. 8c: mp 110-112 °C;
3
NMR δ 4.85 (m, 2H), 2.99 (m, 2H), 2.79 (m, 2H), 2.66 (m, 2H),
(7)°, γ ) 96.154(6)°, V ) 893.9(1) Å , Z ) 2, and D(X-ray) )
-
3
2
2
(
.56 (d, 2H, J ) 11.0 Hz), 2.32-2.39 (series of m, 6H), 2.20 (d,
H, J ) 5.0 Hz), 2.10 (d, 2H, J ) 4.0 Hz), 1.99 (m, 2H), 1.93
1.220 mg mm . 2718 data measured to a 2θ max of 115.5° at
T ) 294 K. R ) 0.0572 for 2320 reflections with [I > 2σ(I)], R
) 0.0586, wR2 ) 0.1521 for all 2444 unique reflections. For
7e: C H O , FW ) 516.69, monoclinic space group C2/c; a )
m, 2H), 1.65-1.73 (m, 4H), 1.55 (m, 2H), 1.41 (m, 4H); 13
C
NMR δ 172.8, 49.8, 47.3, 42.7, 42.4, 41.6, 41.1, 39.0, 35.5, 35.2,
5.7, 22.4; HRMS (FAB) calcd for C26 410.5242, found
MH ) 411.2538. 7d : H NMR δ 4.93 (m, 2H), 2.67 (m, 2H),
.50 (m, 2H), 2.35 (m, 2H), 2.17 (m, 4H), 2.09 (m, 2H), 2.01
3
4
44
4
2
(
H
34
O
4
35.4020(13) Å, b ) 7.7641(4) Å, c ) 19.7431(5) Å, â ) 100.14-
3
+
28
1
-3
(1)°, V ) 5342.0(4) Å , Z ) 8, and D(X-ray) ) 1.285 mg mm .
2
4287 data measured to a 2θ max of 116° at T ) 294 K. R )
0.0541 for 2443 reflections with [I > 2σ(I)], R ) 0.088, wR2 )
0.1359 for all 3482 unique reflections. This crystal was twinned
by a rotation about the real c axis. The crystal shape is a thin
lath or sword, with the major face indexed as the 100.
Twinning is caused by stacking of thin lathes that occasionally
differed by 180 degrees in the direction of their +b long axes,
a reasonable stacking fault. The hk0 layers of the two twin
diffraction patterns are exactly superimposed. This was par-
tially corrected by assigning, and refining, a separate scale
factor for the hk0 data. The l ) 3 layer had to be excluded
from the refinement, because the two patterns almost coincide,
and reflections overlap, on that level. Thirteen reflections from
other layers were excluded for likely twin interference; all had
large discrepancies of the F . F type. Several crystals were
(
m, 2H), 1.85-1.93 (m, 6H), 1.60-1.67 (m, 8H), 1.37-1.42 (m,
1
3
8
3
C
H); C NMR δ 173.3, 77.6, 45.3, 43.9, 43.0, 42.1, 40.4, 38.3,
7.2, 36.1, 32.2, 29.6, 26.0, 22.0; HRMS (FAB) calcd for
+
28
29
38 4
H O 438.5773, found (MH ) 439.2834. 7e: mp slow
1
decomposition above 300 °C; H NMR δ 4.72 (m, 2H), 2.76 (m,
2
H), 2.43 (d, 2H, J ) 12.5 Hz), 2.38 (m, 2H), 1.89-2.16 (series
13
of m, 20H), 1.58-1.66 (m, 6H), 1.40-1.49 (m, 10H); C NMR
δ 171.3, 78.7, 47.8, 46.0, 44.5, 44.0, 43.8, 43.7, 40.0, 39.6, 38.7,
3
7.3, 36.3, 34.3, 33.8, 28.6, 21.9; HRMS (FAB) calcd for
34 44 4
C H O 516.3240, found (M+) 517.3214 (-2.6 m), found
(
MH+) 517.3304. 7f: mp: slow decomposition above 300 °C;
1
H NMR δ 5.12 (m, 2H), 2.77 (m, 2H), 2.43 (m, 2H), 2.35 (d,
2
2
8
1
3
H, J ) 5.0 Hz), 2.22 (d, 2H, J ) 10.0 Hz), 2.13-2.18 (m, 4H),
.10 (m, 2H), 1.98-2.06 (m, 8H), 1.88 (m, 4H), 1.65-1.71 (m,
H), 1.50-1.56 (m, 2H), 1.36 (d, 2H, J ) 10.0 Hz); 13C NMR δ
72.3, 79.7, 49.9, 44.5, 43.6, 43.1, 42.1, 40.8, 39.0, 38.9, 37.7,
7.4, 36.6, 35.5, 33.9, 28.3, 21.3; HRMS (FAB) calcd for
o
c
examined and all were twinned but the volume-ratio of the
two components varied. In the best case, reported herein, the
V ratio was about 3:1. For 8f: C H O , FW ) 488.64,
3
2
40
4
C
32
H
40
O
4
488.2927, found (MH+) 489.3005. 8f: mp slow
monoclinic space group P2 /c; a ) 6.4064(4) Å, b ) 18.637(1)
1
1
3
decomposition above 300 °C; H NMR δ 5.03 (m, 2H), 3.18 (m,
Å, c ) 20.683(1) Å, â ) 95.31(1)°, V ) 2458.9(3) Å , Z ) 4, and
-
3
2
1
1
4
H), 2.89 (m, 2H), 2.58 (m, 6H), 2.40 (m, 2H), 1.97-2.22 (m,
0H), 1.91 (m, 2H), 1.71-1.78 (m, 6H), 1.61 (m, 2H), 1.37-
D(X-ray) ) 1.320 mg mm . 7662 data measured to a 2θ max
of 116° at T ) 294 K. R ) 0.0499 for 2754 reflections with [I
> 2σ(I)], R ) 0.0627, wR2 ) 0.1290 for all 3424 unique
reflections.
1
3
.44 (m, 6H); C NMR δ 176.2, 78.0, 50.8, 48.5, 45.1, 42.1,
2.0, 39.9, 38.8, 37.2, 36.8, 35.8, 29.6, 28.4, 28.3, 22.5; HRMS
(
FAB) calcd for C32
Mixture of 7f and 8f. Anal. Calcd for C32
.25. Found: C, 79.43; H, 9.59.
Gen er a l P r oced u r e for Clea va ge of th e Teth er . LiAlH
0.06 mmol) was suspended in THF (1 mL) and cooled with
H
40
O
4
488.2927, found (MH+) 489.3012.
More details, structural results and figures for 5, 6, 7e, and
8f are reported in the Supporting Information and have
been deposited with the Cambridge Crystallographic Data
40 4
H O
: C, 78.65; H,
8
4
(
(28) Compounds 8d and 8e have not been synthesized, because the
an ice bath. To this stirred suspension was added dropwise a
solution of 7d ,e or 8c,f (0.03 mmol) in THF (2 mL). After the
addition was complete, the contents were allowed to warm to
photocycloadditions of 3d and 3e were performed on the isolated (R,
R/S, S) diastereomer, which produces only the exo-trans-exo isomer 7.
(29) Sheldrick, G. SHELXTL96, Acta Crystallogr. A 1990, 46, 467.

1
68 J . Org. Chem., Vol. 66, No. 1, 2001
Galoppini et al.
Centre. Copies can be obtained from Director, CCDC, 12
Union Road, Cambridge CB2 1EZ, UK (e-mail: deposit@ccdc.
cam.ac.uk).
Su p p or tin g In for m a tion Ava ila ble: X-ray structural
information on 5, 6, 7e, 8f including tables of atomic coordi-
nates, bond lengths and angles, anisotropic displacement
parameters, hydrogen coordinates, a description of the ob-
served X-ray-induced rearrangement of 8f and 15 figures. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Ack n ow led gm en t. E.G. thanks Prof. William Dai-
ley, Prof. Kurt Deshayes, and Prof. Andrea Lazzeri for
insightful discussions and the Rutgers University Re-
search Council for financial support. R.G. is grateful for
the support of the Office of Naval Research.
J O0055990