Retro Diels-Alder Reactions of 5,6-Disubstituted-7-oxabicyclo[2.2.1]hept-2-enes: Experimental and Density Functional Theory Studies

Article abstract of DOI:10.1021/jo0003091

Several 5,6-disubstituted-7-oxabicyclo[2,2.1]hept-2-enes (1-4) were synthesized on ≥0.1 mol scale. The heat-induced retro Diels-Alder (rDA) decomposition of these derivatives was studied by thermal analysis, and the kinetics of the rDA were measured for 4. First-order rate constants (k = 1.91-14.2 x 10^-5 s^-1), measured at four temperatures between 124 and 150°C, were used to calculate Arrhenius activation parameters E_a (34.5 ± 0.5 kcal/mol) and In A (1.77 ± 0.03 x 10⁴). The observed activation energy was significantly larger (by 9.5 kcal/mol) than that previously measured for the maleic anhydride adduct 1, and this was attributed to the difference in LUMO energies for the two dienophiles. Modeling of the activation parameters found for 4 with density functional theory (DFT) calculations for similar compounds 5 and 6 gave close quantitative correlations for ΔH^?, ΔG^?, ΔS^?. The rDA reactions studied were found to be entropy-driven.

Full text of DOI:10.1021/jo0003091

5202
J . Org. Chem. 2000, 65, 5202-5206
Retr o Diels-Ald er Rea ction s of
5,6-Disu bstitu ted -7-oxa bicyclo[2.2.1]h ep t-2-en es: Exp er im en ta l a n d
Den sity F u n ction a l Th eor y Stu d ies
J ason T. Manka,† Andrew G. Douglass,†,‡ Piotr Kaszynski,‡ and Andrienne C. Friedli*,†,‡
Department of Chemistry, Middle Tennessee State University, Murfreesboro, Tennessee 37132
afriedli@mtsu.edu
Received March 6, 2000
Several 5,6-disubstituted-7-oxabicyclo[2.2.1]hept-2-enes (1-4) were synthesized on g0.1 mol scale.
The heat-induced retro Diels-Alder (rDA) decomposition of these derivatives was studied by thermal
analysis, and the kinetics of the rDA were measured for 4. First-order rate constants (k ) 1.91-
14.2 × 10-5 s-1), measured at four temperatures between 124 and 150 °C, were used to calculate
Arrhenius activation parameters Ea (34.5 ( 0.5 kcal/mol) and ln A (1.77 ( 0.03 × 104). The observed
activation energy was significantly larger (by 9.5 kcal/mol) than that previously measured for the
maleic anhydride adduct 1, and this was attributed to the difference in LUMO energies for the two
dienophiles. Modeling of the activation parameters found for 4 with density functional theory (DFT)
calculations for similar compounds 5 and 6 gave close quantitative correlations for ∆Hq, ∆Gq, ∆Sq.
The rDA reactions studied were found to be entropy-driven.
In tr od u ction
mental or theroretical investigations of their thermal
stability aside from our initial observations.7 In fact,
there are very few experimental1,10 and theoretical11
studies of DA adducts of furan aside from 1. Compound
1 is the most extensively characterized derivative of
7-oxabicyclo[2.2.1]hept-2-ene, with X-ray structural data12
and theoretical13 as well as kinetic and thermodynamic
studies at atmospheric14,15 and elevated16,17 pressures
reported.
Therefore, we were interested in gaining a detailed
understanding of the kinetics and thermodynamics of the
rDA for 2-4 and comparing the results to those for 1.
Here we describe the convenient large scale preparation
and isolation of compounds 2-4 and the kinetics of the
rDA reaction for a representative compound, 4. The
experimental results are compared with those for 1 and
discussed in the context of DFT calculations for 1 and
model compounds 5 and 6.
The classical Diels-Alder (DA) adduct 7-oxabicyclo-
[2.2.1]hept-5-ene-2,3-dicarboxylic anhydride (1) is a key
starting material for a variety of biologically relevant
compounds1 and polymers.2-5 Among the polymers are
polyelectrolytes, which we have pursued6,7 for applica-
tions in solid-state batteries. Polyethers are particularly
well-suited for ion conduction8 and can be synthesized
from diol 2, monoether 3, and diether 4 via 1.
One important issue in the preparation of polymers is
the purity of the monomers. This is a concern in the case
of monomers 1-4, which can undergo retro Diels-Alder
(rDA) reactions upon heating. The rDA decomposition of
1 does not pose a serious obstacle to its synthesis.
However, the limited thermal stability of 1-4 does
intefere with purification, since for large scale prepara-
tions of these products, recrystallization (for 1) or vacuum
distillation (for 2-4) is preferable to chromatography.
Although the syntheses of 2-4 have been described in
the literature on various scales,3,4,9 there are no experi-
Resu lts a n d Discu ssion
Syn th esis. Syntheses of 2-4 were carried out accord-
ing to Scheme 1. The classical Diels-Alder reaction of
furan (7) and maleic anhydride (8) resulted in a mixture
of crystalline 1-exo contaminated with 1-en do in amounts
determined by reaction time.18 Without purification, the
* Author for correspondence: Phone/fax: 615-898-2071/5182.
† Department of Chemistry, Middle Tennessee State University,
Murfreesboro, TN 37132.
‡ Department of Chemistry, Vanderbilt University, Nashville, TN
37235.
(1) Vogel, P.; Cossy, J .; Plumet, J .; Arjona, O. Tetrahedron 1999,
55, 13521-13642.
(9) Das, J .; Vu, T.; Harris, D. N.; Ogletree, M. L. J . Med. Chem.
1988, 31, 930-935.
(2) Lu, S.-Y.; Quayle, P.; Heatley, F.; Booth, C.; Yeates, S. G.; Padget,
J . C. Macromolecules 1992, 25, 2692-2697.
(10) Cook, M. J .; Cracknell, S. J . Tetrahedron 1994, 50, 12125-
(3) Lu, S.-Y.; Armass, J . M.; Majid, N.; Glennon, D.; Byerley, A.;
Heatley, F.; Quayle, P.; Booth, C.; Yeates, S. G.; Padget, J . C.
Macromol. Chem. Phys. 1994, 195, 1273-1288.
(4) Novak, B. M.; Grubbs, R. H. J . Am. Chem. Soc. 1988, 110, 960-
961.
12132.
(11) Bachrach, S. M. J . Org. Chem. 1995, 60, 4395-4398.
(12) Baggio, S.; Barriola, A.; de Perazzo, P. K. J . Chem. Soc., Perkin
Trans. 2 1972, 934-937.
(13) Calvo-Losada, S.; Sua´rez, D. J . Am. Chem. Soc. 2000, 122, 390-
(5) Manning, D. D.; Hu, X.; Beck, P.; Kiessling, L. L. J . Am. Chem.
Soc. 1997, 119, 3161-3162.
391.
(14) Dewar, M. J . S.; Pierini, A. B. J . Am. Chem. Soc. 1984, 106,
203-208.
(15) Lee, M. W.; Herndon, W. C. J . Org. Chem. 1978, 43, 518.
(16) Zhulin, V. M.; Bogdanov, V. S.; Kel'tseva, M. V.; Kabotyanskaya,
E. B.; Koreshkov, Y. D. Bull. Acad. Sci. Russ. 1989, 38, 2303-2308.
(17) Zhulin, V. M.; Kel'tseva, M. V.; Bogdanov, V. S.; Koreshkov, Y.
D.; Kabotyanskaya, E. B. Bull. Acad. Sci. Russ. 1990, 39, 456-459.
(6) Zheng, M.; Manka, J . T.; Friedli, A. C. In NASA Microgravity
Materials Science Conference; NASA Conf. Publ.: Huntsville, AL, 1998;
pp 225-228.
(7) Douglass, A. G.; Zheng, M.; Friedli, A. C.; Kaszynski, P. Polym.
Prepr. 1996, 37, 634.
(8) Gray, F. M. Solid Polymer Electrolytes; VCH: New York, 1991.
10.1021/jo0003091 CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/01/2000

Retro Diels-Alder Reactions
Sch em e 1
J . Org. Chem., Vol. 65, No. 17, 2000 5203
Sch em e 2
F igu r e 2. 1H NMR spectra showing progress of rDA reaction
of 4 at 124 °C.
F igu r e 1. Thermal gravimetric analysis of compounds 1-4
heated at 20 °C/min under N2.
product was reduced with lithium aluminum hydride to
afford crude diol 2. Williamson ether synthesis using
carefully dried 2 and stoichiometric NaH in the presence
of excess iodomethane produced monoether 3 in up to
69% yield. Product 3 was conveniently separated from
byproducts 2 and 4 using an alumina column and a
gradient of ethyl acetate and hexanes, which was effec-
tive even on a 0.1 mole scale. This represents a significant
improvement over previous purification of small amounts
of 3 using silica gel.3 The double Williamson reaction of
2 with excess NaH followed by iodomethane resulted in
an 86% yield of diether 4 after purification. A lower yield
of 4 was obtained when higher temperatures were used
during distillation.
F igu r e 3. Plot of ln(x + 1) versus time for 4 at 124, 130, 140,
and 150 °C, where x is the mole ratio of 4 to 9. Rate constants
were derived from the line slopes.
region of the spectrum (Figure 2). Integration of these
peaks allowed the calculation of mole ratios of starting
material and products. Application of the first-order
kinetic formula kt ) ln(x + 1), where x is the mole ratio
of 4 to 9 present at a given time t and k is the rate
constant, resulted in the plots shown in Figure 3. An
Arrhenius plot of the kinetic data (included in the
Supporting Information) allowed the calculation of acti-
vation energy Ea of 35.0 ( 0.5 kcal/mol and ln A ) 1.77
((0.03) × 104 from the slope and intercept, respectively.
The enthalpy and entropy of activation were calculated
to be ∆Hq ) 34.5 ( 0.5 kcal/mol and ∆Sq ) 5.0 ( 0.1
cal/mol K. These values are higher than those measured
for 1-exo14 by 9.5 kcal for ∆Hq and 4.8 cal/mol K for ∆Sq
(Table 1).
Th er m a l Decom p osition . Thermogravimetric Analy-
sis (TGA) of 1-4 under a nitrogen atmosphere showed
rapid and complete (>95%) weight losses between ∼200
and 260 °C (Figure 1). The lowest temperature for onset
of decomposition was observed for 1 (147 °C) and highest
for 2 (217 °C). The thermolysis of 4 gave two products
1
which were identified by H NMR analysis as furan (7)
and cis-1,4-dimethoxy-2-butene (9), the expected products
of the retro Diels-Alder reaction. (Scheme 2).
The kinetic behavior of the thermal decomposition of
4 was investigated at four temperatures using solution
1H NMR spectroscopy. As the reaction proceeded, signals
at 4.75 ppm in 4 disappeared while peaks consistent with
919 appeared at 3.85 ppm in an otherwise uncomplicated
Unfortunately, independent measurement of the ther-
modynamic values by differential scanning calorimetry
(DSC) was difficult due to the superposition of several
peaks. An endotherm (onset at 161 °C), presumably due
to the vaporization of the rDA products 7 (bp 32 °C) and
9 (bp 141 °C), obscured the onset of the exotherm due to
(19) Malanga, C.; Mannucci, S.; Lardicci, L. Tetrahedron 1998, 54,
1021-1028.
(18) Anet, F. A. L. Tetrahedron Lett. 1962, 25, 1219-1222.

5204 J . Org. Chem., Vol. 65, No. 17, 2000
Manka et al.
Ta ble 1. Ca lcu la ted (B3LYP /6-31G*)a a n d Exp er im en ta lb
Th er m och em ica l P a r a m eter s for th e Retr o Diels-Ald er
Rea ction
calculated ∆Hq is about 2.6 kcal/mol lower than the
298
experimental value, while the ∆Sq is larger by 2.8 cal/
mol K. The resulting ∆Gq of 3.3 kcal/mol reflects, to some
degree, the difference between the environments used for
calculations (gas-phase) and experiment (solution) in
addition to systematic errors of the quantum-mechanical
method.
reaction overall
activation parameters
Although the accurate quantitative reproduction of
experimental values was not necessarily expected, the
relative energies of the transition states should be
modeled well because they are not affected much by
solvent.13,14 The effect of temperature upon the results
was also expected to be small21 so no attempt was made
to correct to experimental temperature.
compd
R, R′
∆H
∆S
∆G
∆Hq
∆Sq
∆Gq
1-exo -COOCO-
+3.44 +45.9 -10.2
22.4
2.95
21.6
(25.0)c (0.17)c (24.9)
5-a n ti -CH2OCH2- +3.07 +48.2 -11.3
31.8
(34.5)d (5.0)d (33.0)
31.2 3.1 30.3
4.11
30.5
6
H
+5.3 +43.5
-7.6
(-15.2)e (36.7)e (-6.7)e
a
∆G and ∆H are in units of kcal/mol; ∆S is in units of eu. The
The tricyclic compound 5, used as a model for dimethyl
ether 4, has two conformational minima corresponding
to structures with oxygens on either the opposite (5-a n ti)
or same (5-syn ) side of the bicyclic ring. The former
conformer is the global minimum on the potential energy
surface (PES) and is separated from the less stable 5-syn
form (∆G ) 0.40 kcal/mol) by an interconversion barrier
of 2.1 kcal/mol. The transition state for the rDA reaction
was identified as 5TS-syn , and the analogous 5TS-a n ti
could not be located on the PES. The calculated thermo-
dynamic parameters for 5TS-syn differ from those
measured experimentally for the diether 4 by about 2.5
kcal/mol, a coincidentally similar margin of error to that
found for 1TS-exo.
A comparison of the rDA reactions for the anhydride
1-exo with its di-deoxy analogue 5 shows that both are
entropically driven with almost identical overall ther-
modynamics. The small reaction endotherm (about 3 kcal/
mol) is offset by a large entropy gain (about 46 cal/mol
K) to give a moderate free energy gain of about -11 kcal/
mol. The most significant difference between the two
reactions is their TS energies. When compared to 1TS-
exo, the ∆Gq of 5TS is higher by 8.9 kcal/mol, a value
which is very close to the 8.1 kcal/mol difference between
the experimental results (Table 1). The two PES's are
graphically compared in Figure 5.
The calculated thermodynamic values for the rDA
reaction of the parent 7-oxabicyclo[2.2.1]hept-2-ene (6)
are similar to those for 5, indicating that the perturbation
of 6 by the oxymethyl substituents in 2-5 is small,
almost negligible.
The observed difference in the TS energies between
rDA reactions resulting in a conjugated dienophile (such
as 8) and the two with isolated double bonds (10 and 11)
are reflected in the geometries of the TS state structures
shown in Figure 4. The transition state 1TS-exo is early
relative to 5TS (and 6TS), with breaking C-C σ bonds
at 2.117 Å, while in the latter, bond cleavage is further
along, with C-C separations of about 2.158 Å (2.150 Å).
The same trends are observed in the bond lengths of the
furan ring and are consistent with an early transition
state for the anhydride 1 decomposition.
temperature for the gas-phase calculated values is 298 K. Full
computational results are listed in the Supporting Information.
b
d
Experimental data is in parentheses. c Reference 14. Diether
4. e Reference 1.
the rDA reaction. The peak shape and position measured
with DSC were highly dependent upon sample size and
conditions. An example of a DSC thermogram of 4
appears in the Supporting Information.
Th eor etica l An a lysis. The significant differences in
the stability of the anhydride and the diether apparent
from the TGA studies and measured activation energies
for the rDA of 1 and 4 can be attributed20 to the presence
of conjugated carbonyl functionality in dieneophile 8 as
compared to an isolated alkene in 9. To gain more insight
into the electronic factors influencing these reactions, we
sought to compare the calculated energies and transition
state geometries.
The ground and transition state structures for anhy-
dride 1-exo, cyclic ether 5, the parent 7-oxabicyclo[2.2.1]-
hept-2-ene (6), and their respective rDA products: furan
(7), maleic anhydride (8), 2,5-dihydrofuran (10), and
ethylene (11), were calculated at the B3LYP/6-31G* level
of theory using appropriate symmetry constraints. This
level of theory has been successfully used in modeling
concerted and stepwise pathways for the retro Diels-
Alder and related pericyclic reactions,21 as well as to
explain the stereoselectivity of the DA reaction that
produces 1-exo as the major product.13 The molecular
geometries for the reagents, products, and transition
states in the rDA reactions of the exo isomers of 1, 5,
and 6 are shown in Figure 4. The calculated and
experimental thermodynamic parameters for the rDA
reactions are collected in Table 1.
The calculated structure for 1-exo compares well to
its experimental molecular geometry despite significant
distortions in the solid-state environment (Figure 4).
Most calculated bonds are longer than experimental but
within one standard deviation of the X-ray data.12
A
notable exception is the C(2)-C(3) bond, which was
calculated to be 0.06 Å shorter than the experimental
result. The biggest disagreement in bond angles (up to
2°) involved the carbonyl groups. Ring puckering angles
(Supporting Information) were found to be within 0.6°
of the experimental values.
Activation parameters calculated for the rDA reaction
of 1-exo are close the experimental values14 and similar
to those calculated with perturbation methods.13 The
These reactivity differences can be traced to the
relative energies of the FMO’s. For the forward DA
reaction, the HOMO of furan (A2, -6.11 eV) can be
expected to interact with the same symmetry LUMO of
the dienophile according to classical FMO theory.20
According to our calculations, the LUMO (A2, 0.50 eV) of
the isolated CdC bond in dihydrofuran (10) is signifi-
cantly higher in energy than that in maleic anhydride
(8, A2, -3.19 eV). This results in a 3.7 eV difference in
the HOMO-LUMO gap for 7 and 8 relative to that
(20) Fleming, I. A. Pericyclic Reactions; Oxford Science Publica-
tions: London, 1999.
(21) Tian, J .; Houk, K. N.; Kla¨rner, F. G. J . Phys. Chem. 1998, 102,
7662-7667.

Retro Diels-Alder Reactions
J . Org. Chem., Vol. 65, No. 17, 2000 5205
F igu r e 4. B3LYP/6-31G* geometries for 1-exo, 5-a n ti, and 6, the transition states in the rDA reaction 1TS-exo, 5TS-syn , and
6TS, and the respective products 7-11. Bond lengths are in angstroms. Those in parentheses are experimental (ref 12).
tent with expectations from qualitative FMO analysis.
Diol 2 and related ethers 3 and 4 underwent the rDA
reaction at relatively low temperatures, but higher than
that required for the anhydride 1. Kinetic data for the
rDA reaction for 4 were closely comparable with calcu-
lated thermodynamic and kinetic parameters and ex-
perimental data on 1 from the literature. Generally, for
the solution reactions described here, the kinetic and
thermodynamic data compared unusually well quanti-
tatively to the calculated (gas phase) data. Computational
analysis of the transition states at the B3LYP/6-31G*
F igu r e 5. Reaction coordinate for the B3LYP/6-31G* Gibbs
free energies calculated for 5, 5TS, and products 7 and 10
plotted with those for 1-exo, 1TS-exo, and products 7 and 8.
level confirmed that the Ea in 1 is lowered by 9.5 kcal/
mol relative to that of model compound 5, a value almost
identical to the difference in the experimental ∆Hq for
1-exo and 4. The faster rate of the rDA in 1 as compared
to 4 is likely due to the energetic differences in the
LUMO’s of 8 and 9.
Large entropic contributions to the rDA resulted in a
favorable ∆G for the reactions of 1, 5, and 6, overcoming
a slightly unfavorable enthalpic term.
between 7 and 10. The better overlap between the FMOs
in 7 and 8 lowers the energy of 1TS by 9 kcal/mol below
that of 5TS.
The calculated activation parameters for the model
rDA reactions of 5 and 6 were almost exactly the same:
the two values of ∆Hq were within 0.5 kcal/mol of one
another, and ∆Sq values were within 1.0 cal/mol K,
resulting in ∆Gq values differing by only 0.2 kcal/mol.
Compared to the experimental values for 4, the calculated
∆Gq value for 5 was low by 2.5 kcal/mol (a difference of
2.7 kcal/mol for ∆Hq and 0.9 cal/mol K for ∆Sq). These
results affirm the validity of using 5, a known com-
pound,22 as a model for 4.23 Furthermore, the influence
of the methoxymethyl substitutents at the reaction center
in 4 or 5 is minimal when compared to 6.
Exp er im en ta l Section
Kin etics. Rates of retro Diels-Alder reactions were mea-
1
sured by following the decomposition of 4 by solution H NMR.
For each datapoint, a 20 µL sample was transferred with an
Eppendorf pipet to an NMR tube and 0.6 mL of 1,2-dichlo-
robenzene-d4 added. The solution was freeze-thaw degassed
twice, and the NMR tube sealed under vacuum (∼50 Torr).
Samples were immersed in a Fisher Isotemp 800 constant-
temperature bath heated to 124 °C, 130 °C, 140 °C, or 150 °C
( 0.5 °C for at least six different time intervals corresponding
to between 5 and 95% decomposition. Each decomposition was
taken to at least three half-lives to determine whether the
process was first order. The sample was periodically cooled to
room temperature and the ratio of one product [1H NMR δ
3.14-3.20 (m, max at 3.17, 6H), 3.84-3.87 (m, center at 3.85,
4H), 5.58-5.62 (m, max at 5.60, 2H)] to the starting material
4 [1H NMR δ 1.73-1.80 (m, 2H), 3.07-3.2l (m, max at 3.17,
6H), 3.22-3.35 (m, 2H), 4.72-4.75 (m, max at 4.75, 4H), 6.13
Con clu sion s
The results for thermolysis of 7-oxabicyclo[2.2.1]-
heptenes via the rDA reaction reported here are consis-
(22) Gilchrist, T. L.; Wood, J . E. J . Chem. Soc., Perkin Trans. 1 1992,
9-15.
(23) The relatively high calculated stability of 5 and its similarity
to 1 and 4 is inconsistent with the identification of 5 as an unstable
structure.
1
(m, max at 6.13, 2H)] was monitored by H NMR (200 MHz).

5206 J . Org. Chem., Vol. 65, No. 17, 2000
Manka et al.
The assignment of structures of the products was confirmed
by comparison to a published 1H NMR spectrum of 919 and
NMR of pure 7 in 1,2-dichlorobenzene-d4 [1H NMR δ 6.22 (t,
J ) 1.3 Hz, 2H), 7.27 (t, J ) 1.2 Hz, 2H)]. The enthalpies and
entropies of activation were calculated using standard formu-
las: ∆Hq ) -Ea - RT and ∆Sq ) R[ln A - ln(ekT/h)].
Com p u ta tion a l Meth od s. Quantum-mechanical calcula-
tions were carried out using the Gaussian 98 package24 on an
SGI PowerIndigo2 workstation, and the resulting energies are
listed in the Supporting Information. Geometry optimizations
were undertaken using appropriate symmetry constraints at
the B3LYP/6-31G* level of theory.25 Transition states were
located using the QTS2 procedure. Vibrational frequencies
were used to characterize the nature of the stationary points
and to obtain thermodynamic parameters. Zero-point energy
corrections (ZPEC) were calculated using a 0.9804 scaling
factor.26 Thermodynamic properties were calculated at 298 K
and 1 atm.
Extraction of the filtered solids using a Soxhlet apparatus
for 48 h with methylene chloride gave a mixture of isomers
(3.8 g) in a 3:2 ratio. Overlapping NMR signals ascribed to
exo, exo at δ 1.84-1.95 (m, max at 1.89, 2H) were comparable
to literature values9 δ 1.92 (m, 2H) and signals ascribed to
the endo, endo isomer at δ 1.77-1.84 (m, max at 1.80, 2H)
appeared in the same region of the spectrum (see Supporting
Information).
e x o -2 -H y d r o x y m e t h y l -e x o -3 -m e t h o x y m e t h y l -7 -
oxa bicyclo[2.2.1]h ep t-5-en e (3).3 The reaction was done on
0.1 mol to 5 mmol scales according to literature procedure.3
Good quality NaH (1.1 equiv, 60% in oil) was introduced into
a dry round-bottomed flask and THF (1 mL/mmol NaH) was
added to form a slurry. The reaction vessel was cooled to 0 °C
and 2 (1.0 equiv), dissolved in THF (1 mL/mmol 2), was added
dropwise. The mixture was stirred for 30 min at 0 °C, and then
methyl iodide (1.1 eq) added dropwise. The mixture was stirred
for 30 min and then allowed to warm to room temperature
and stir for 3 h. The mixture was filtered through Celite, the
solid washed with THF (0.5 mL/mmol 3), and the solvent
removed. The product was purified by column chromatography
on neutral alumina (80-200 mesh) packed in hexanes using
1:4 ethyl acetate/hexanes eluent, gradually changing the
gradient to 4:1 ethyl acetate:hexanes as products eluted. The
chromography was followed using 1:1 ethyl acetate: hexanes
Rf (2) ∼ 0; Rf (3) ∼ 0.5; Rf (4) ∼1. A single column could be
loaded heavily and used repeatedly without hindering separa-
tion. After separation and solvent removal, 3 was obtained as
1
Syn th esis. Boiling and melting points are uncorrected. H
NMR and 13C NMR spectra were run at 200 and 50 MHz,
respectively. Samples were dissolved in deuteriochloroform
and referenced to CHCl3 unless noted otherwise. IR spectra
were measured as neat samples. Commercial reagents were
used as received. Tetrahydrofuran were dried over benzophe-
none ketyl and distilled before use. Differential scanning
calorimetry measurements were run on a Perkin-Elmer DSC-7
under nitrogen calibrated to an indium standard. Aluminum
pans (50 µL) with a pinhole and sample sizes of 2-5 mg were
used and the heating rate was 20 °C/min. Thermal gravimetric
analysis measurements were done using a Perkin-Elmer
TGA-7 under nitrogen. Samples with weights ranging from
15 to 20 mg were heated from 40 to 350 °C at a rate of 20
°C/min. Compound 127 was synthesized according to a modi-
fication of the method of Furd´ık28 (for details see Supporting
Information).
1
a pale yellow oil, but yield varied from 35 to 69%. H NMR δ
1.83-2.04 (m, max at 1.97, 2H), 3.26 (bs, 1H), 3.26-3.78 (m,
4H), 3.35 (s, 3H), 4.66 (m, 1H), 4.70 (m, 1H), 6.32-6.39 (m,
center at 6.38, 2H); 13C NMR δ 40.2, 42.9, 59.1, 62.7, 73.3, 81.2,
81.5, 135.7, 136.3.
exo,exo-5,6-Dim eth oxym eth yl-7-oxa bicyclo[2.2.1]h ep t-
2-en e (4).29 Sodium hydride (10.0 g, 60% in oil, 0.25 mol) was
weighed into a dry 500 mL round-bottomed flask in a glovebag
and covered with dry THF (125 mL). Diol 2 (15.6 g, 0.100 mol)
in THF (50 mL) was added dropwise via addition funnel, and
the mixture was stirred for 1 h. Methyl iodide (31.8 g, 0.22
mol) was added through the rinsed funnel. After 1 h, wet ether
was added until no bubbling was noted. The solution was dried
over magnesium sulfate and filtered, the solvent was removed,
and the crude product was Kugelrohr distilled to give 15.9 g
(86.1%) of pale yellow oil: bp (lit.29 60-65 °C, 0.001 Torr), 75-
exo,exo-7-Oxa b icyclo[2.2.1]h e p t -5-e n e -2,3-d im e t h a -
n ol (2).9 Following a modification of the method of Novak,29
a
dry 2 L three-necked round-bottomed flask equipped with a
pressure equalizing dropping funnel and reflux condenser was
charged with LiAlH4 (17.1 g, 0.45 mol) and THF (125 mL).
Anhydride 1 (60.0 g, 0.361 mol) was dissolved in warm THF
(850 mL) and added dropwise to the stirred slurry. The
mixture was allowed to stir overnight, and then water (20 mL),
NaOH (20 mL, 15% w/w), and water (55 mL) were then added
slowly in sequence. The resulting white suspension was filtered
through Celite and rinsed with THF (50 mL). The solvent was
removed by rotary evaporation and residue was dried in vacuo
and over P2O5 to give 48.1 g (81.0%) of a colorless, opaque,
viscous oil of pure exo, exo isomer: 1H NMR δ 1.84-1.95 (m,
max at 1.89, 2H), 3.52-3.76 (m, max at 3.66, 4H), 4.51 (bs,
2H), 4.63 (s, 2H), 6.32 (s, 2H).
1
80 °C (0.5 Torr); H NMR δ 1.62-1.68 (m, max at 1.65, 2H),
2.98-3.26 (m, max at 3.09, 4H), 3.11 (s, 6H), 4.57 (t, J ) 0.8
Hz, 2H), 6.09 (t, J ) 0.9 Hz, 2H).
Ack n ow led gm en t. This work was supported by
NASA (HEDS-9528029) and the Petroleum Research
Fund (30202-GB7). J .T.M. is grateful for a scholarship
from the Undergraduate Research Council of MTSU
College of Basic and Applied Sciences. We thank Dr. J .
Howard for use of the heating bath and technical advice.
(24) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; J . A. Montgomery,
J .; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.; Daniels,
A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .; Barone, V.;
Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; C. Adamo; Clifford,
S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma,
K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J . B.;
Cioslowski, J .; Ortiz, J . V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B.; Chen, W.;
Wong, M. W.; Andres, J . L.; Gonzalez, C.; Head-Gordon, M.; Replogle,
E. S.; Pople, J . A. In Gaussian, Inc.: Pittsburgh, PA, 1998.
(25) Becke, A. D. J . Chem. Phys. 1993, 98, 5648-5652.
Su p p or tin g In for m a tion Ava ila ble: Tables of computed
energies for structures discussed in the text, calculated rates,
and puckering angles, a DSC thermogram, NMR spectra
relevant to the thermal decomposition of 4, an 1H NMR
spectrum of the 2 exo and endo isomer mixture, an Arrhenius
plot, and some experimental descriptions are included. This
material is available free of charge via the Internet at
http:/pubs.acs.org.
(26) Scott, A. P.; Radom, L. J . Phys. Chem. 1996, 100, 16502-16513.
(27) Diels, O.; Alder, K. Chem. Ber. 1929, 62, 554-558.
(28) Furd´ık, M.; Dra´bek, J . Acta Fac. Rerum Nat. Univ. Comenianae,
Chim. 1965, 23-24.
(29) Novak, B. M. Ph.D. Thesis, California Institute of Technology,
1989.
J O0003091