o-Fluoranil: Stereochemistry and mechanism of its Diels-Alder reactions

Article abstract of DOI:10.1021/jo9015562

(Chemical Equation Presented) In the absence of significant steric effects, Diels-Alder reactions of the title quinone generally take place with preservation of configuration, and are therefore probably concerted. However, hybrid density functional calcul

Full text of DOI:10.1021/jo9015562

pubs.acs.org/joc
o-Fluoranil: Stereochemistry and Mechanism of Its Diels-Alder
Reactions
David M. Lemal,* Dayong Sang, and Sudharsanam Ramanathan
Department of Chemistry, Dartmouth College, Hanover, New Hampshire 03755
david.m.lemal@dartmouth.edu
Received July 18, 2009
In the absence of significant steric effects, Diels-Alder reactions of the title quinone generally take place
with preservation of configuration, and are therefore probably concerted. However, hybrid density
functional calculations indicate that often these reactions are highly asynchronous. Steric hindrance can
result in reaction at quinone oxygen instead of carbon. Preference for endo over exo cycloaddition is
observed, and is reinforced by a repulsive secondary orbital interaction in exo transition states.
Introduction
o-Fluoranil (tetrafluoro-o-benzoquinone, 1) undergoes
4 þ 2] cycloaddition with alkenes and acetylenes either by
[
Diels-Alder reaction on the ring diene to give an R-diketone
(
2) or by hetero-Diels-Alder addition on the carbonyl
1,2
oxygens to give a dioxene (3) (or dioxin). Despite a strong
thermodynamic bias favoring the latter mode of reaction,
normal Diels-Alder addition occurs with many alkenes and
acetylenes, both electron-rich and electron-deficient. This is
illustrated in Figure 1 with calculated pathways for reaction
Results and Discussion
Symmetrical Dienophiles. The transition states for the ob-
served and hypothetical [4 þ 2] cycloadditions of ethylene are
both predicted to be perfectly symmetrical; i.e., the reactions
should not only be concerted, but also synchronous. The
question of concert was addressed experimentally with stereo-
chemically labeled alkenes. Cis isomers were employed because
loss of configuration in a stepwise reaction should be faster and
more extensive than with the corresponding trans isomers
because of nonbonded repulsion. Again theory predicted nearly
symmetrical transition states for endo and exo Diels-Alder
addition to the quinone of cis-2-butene, as shown in Table 1
where the calculated lengths of the newly forming C-Cσbonds
are recorded. In the event, cis-2-butene added to the quinone at
3
of the quinone with the archetypal dienophile ethylene, and
the predicted formation of the normal Diels-Alder adduct 4
to the exclusion of its hetero counterpart 5 has now been
verified experimentally. Likely reasons for the kinetic bias
favoring normal Diels-Alder reaction were discussed in our
2
earlier report. Whether normal Diels-Alder reactions take
place concertedly or stepwise, and the closely related ques-
tion whether dienophile stereochemistry is preserved in the
reaction, have not been addressed. In the present report these
questions and that of endo vs. exo preference are explored
both computationally and experimentally with selected
examples.
70 °C to give endo, cis (6) and exo, cis (7) Diels-Alder adducts
(
1) Shteingarts, V. D.; Budnik, A. G. Izv. Sib. Otd. Akad. Nauk SSSR,
Ser. Khim. Nauk 1967, 124.
2) Lemal, D. M.; Ramanathan, S.; Shellito, J. J. Org. Chem. 2008, 73,
392.
3) Jaguar, version 7.0; Schrodinger, LLC: New York, 2007. Frequency
(
3
(
calculations carried out on all transition states found a single negative
frequency corresponding to motion along the reaction coordinate. Except
where a larger basis set is indicated, calculations were performed at the
B3LYP/6-31G** level of theory.
but no trans adduct, in harmony with theory. Since the
carbonyls of o-fluoranil’s Diels-Alder adducts generally
7
804 J. Org. Chem. 2009, 74, 7804–7811
Published on Web 09/18/2009
DOI: 10.1021/jo9015562
r 2009 American Chemical Society

Lemal et al.
JOCArticle
FIGURE 1. Enthalpy changes in the reactions of ethylene with o-fluoranil.
a
TABLE 1. Bond Lengths in Diels-Alder Transition States
maleonitrile is nearly symmetrical (Table 1), and for maleal-
dehyde a similar transition state was found with newly forming
C-C σ bond lengths of 2.162 and 2.202 A (Figure 2A).
b
˚
cycloaddend
r
C-C (A)
˚
ethylene
cis-2-butene
cis-2-butene (exo)
maleonitrile
malealdehyde
dimethyl maleate
butadiene
butadiene (exo)
styrene
2.211, 2.211
2.237, 2.275
2.202, 2.211
2.178, 2.184
1.943, 2.697
2.014, 2.518
1.951, 2.739
2.056, 2.461
1.951, 2.804
1.948, 2.900
1.956, 2.624
However, a second transition state (Figure 2B) was located
for the aldehyde that was very unsymmetrical (Table 1), and it
is both lower in enthalpy than A by 3.14 kcal/mol and higher in
entropy by 5.37 cal/(mol K). Intrinsic reaction coordinate
calculations confirmed that both transition structures connect
with starting materials and product, and show that A arises
from the s-trans, s-trans conformation of the aldehyde while B
originates with the s-cis, s-trans conformation. In both A and B
the carbonyl groups are aligned to overlap well with the
developing C-C bonds.
styrene (exo)
methyl acrylate
a
Endo unless otherwise noted; calculated at the B3LYP/6-311G**þ
b
level of theory. Newly forming carbon-carbon σ bonds.
Endo addition of dimethyl maleate is more complicated
because the methoxycarbonyl groups are enough bulkier
than formyl groups that only one carbonyl group is aligned
with a developing C-C bond in a transition structure. There
are four ways that one carbonyl group can be so aligned with
the other carbonyl twisted roughly perpendicular to it. We
found four transition structures, each corresponding to one
of these orientations, the lowest energy of which appears in
Table 1 and Figure 3. All four are comparably unsymme-
trical, and lie within 3 kcal/mol of each other. As would be
expected, in each case it is the longer, weaker developing
C-C bond that is aligned with a carbonyl.
hydrate readily and often voraciously, these and other
adducts were not isolated as such, but were transformed
with o-phenylenediamine into and characterized as quinox-
alines (eq 1). Because reduction of the adducts by the
diamine competed with quinoxaline formation, water was
added at the outset to protect them as hydrates from that side
reaction. The basis for assignment of adduct configurations
is discussed below.
1
,2
Whether the reaction of dimethyl maleate with o-fluoranil
is concerted was thus an open question. Most commercial
dimethyl maleate contains a few percent of the fumarate
ester, so maleate free of its isomer was prepared from the acid
with diazomethane. Because of its electron-deficiency, vig-
orous conditions (a day at ∼100 °C) were required for the
To learn whether electron-deficient alkenes would behave
similarly to the electron-rich 2-butene, transition states were
calculated for endo Diels-Alder addition to o-fluoranil of
maleonitrile, malealdehyde, and dimethyl maleate. That of
1
9
maleate-quinone reaction. F NMR analysis of the reac-
tion mixture revealed the endo, cis Diels-Alder adduct (8) as
a major product, but there was no more than a trace of the
J. Org. Chem. Vol. 74, No. 20, 2009 7805

Lemal et al.
JOCArticle
o-fluoranil reacted with cis-stilbene to give cis-dioxene 11
with at most a far smaller amount of cis Diels-Alder
5
product. With its phenyl rings twisted 43° out-of-plane,
cis-stilbene brings considerable steric hindrance to cycload-
dition, so its strong preference for attack on the periphery of
the quinone molecule is understandable.
FIGURE 2. Calculated transition structures for endo addition of
malealdehyde to o-fluoranil (B3LYP/6-311G**þ).
The dominance of dioxene 9 over 10 in the product from
trans-stilbene is probably also attributable to steric hin-
6
drance even though this stilbene isomer is planar. See below
for the contrasting behavior of styrene.
Unsymmetrical Dienophiles. As shown in Figure 4, buta-
diene was predicted to prefer Diels-Alder over hetero-
Diels-Alder cycloaddition with o-fluoranil despite the very
strong thermodynamic bias favoring the latter reaction
course. The prediction was borne out when we found that
butadiene reacted readily at room temperature to yield a
mixture of endo (12) and exo (13) Diels-Alder adducts but
almost no dioxene. Both Diels-Alder transition structures
were predicted to be highly unsymmetrical (Figure 5
and Table 1). At room temperature styrene also gave endo
FIGURE 3. Two views of the lowest energy transition structure for
endo addition of dimethyl maleate to o-fluoranil.
trans (fumarate) adduct. So this cycloaddition too was
completely stereoselective, implying concerted reaction.
(
no dioxene, and again calculations generated very unsym-
14) and a small amount of exo (15) Diels-Alder adduct but
7
metrical transition structures (Table 1). These results
suggested that styrene might well react via a biradical inter-
mediate.
cis-Stilbene reacted with o-fluoranil quite differently from
the other symmetrical dienophiles. Unless the quinone was
carefully purified, the products obtained were the dioxene 9
and the Diels-Alder adduct 10 in a 5:1 ratio, both with the
trans configuration. It was found that the cis alkene isomerized
to the trans form under the reaction conditions (even though
calcium carbonate was present to destroy any adventitious
HF), and that the products arose from the trans isomer.
Starting with trans-stilbene, the same products were obtained
in essentially the same ratio. That Diels-Alder adduct 10 has
the trans configuration is obvious from the chemical shift
inequivalence of the bridgehead fluorines and of the vinyl
fluorines, but the dioxene would possess 2-fold symmetry
whether it were cis or trans. Our finding that the reaction of
dl-hydrobenzoin with hexafluorobenzene effected by sodium
hydride in DMSO yielded the same compound provided
The stepwise vs. concerted question was addressed with
trans-β-deuteriostyrene. Endo adducts from both the unla-
beled and labeled styrene were isolated and purified as their
quinoxaline derivatives. The three aliphatic hydrogens
of the unlabeled quinoxaline were well separated in the
1
500 MHz H NMR spectrum, and the highest field of
4
further confirmation that the configuration of 9 is trans (eq 2).
these signals corresponding to the exo methylene hydrogen
was virtually absent in the labeled quinoxaline spectrum.
Thus, within experimental error, the adduct obtained from
(
5) Traetteberg, M.; Frantsen, E. B. J. Mol. Struct. 1975, 26, 69. Molina,
V.; Merchan, M.; Roos, B. O. Spectrochem. Acta, Part A 1999, 55, 433.
6) The fact that the thermodynamic bias favoring dioxene formation
(
over Diels-Alder addition is nearly the same for styrene and trans-stilbene
(ΔΔH = 14.0 vs. 13.4 kcal/mol, respectively) supports the conclusion that
steric effects are responsible for the dioxene from trans-stilbene. The thermo-
dynamic bias is considerably greater for cis-stilbene (ΔH = 18.5 kcal/mol), a
reflection of steric hindrance in the (endo) Diels-Alder adduct, not just in the
transition structure leading to it.
Stilbene isomerization occurred with quinone that ap-
1
peared to be pure by F and H NMR, but truly pure
9
1
(
4) An analogous reaction of hexafluorobenzene with ethylene glycol was
carried out, see: Burdon, J.; Damodaran, V. A.; Tatlow, J. C. J. Chem. Soc.
964, 763.
(7) An irc calculation for the endo transition structure smoothly con-
nected starting materials with the adduct.
1
7
806 J. Org. Chem. Vol. 74, No. 20, 2009

Lemal et al.
JOCArticle
FIGURE 4. Enthalpy changes in reactions of butadiene with o-fluoranil.
o-fluoranil to yield cis-dioxene with much less, if any,
Diels-Alder adduct.
Cycloaddition Preference: Endo vs. Exo. In light of the
quinone’s planarity, there can be little steric influence on the
endo/exo preference. For Diels-Alder reactions in general,
stabilizing secondary orbital interactions involving the fron-
tier orbitals are present in endo, but not exo, transition
8
states. This is true as well for Diels-Alder reactions of
o-fluoranil, the HOMO and LUMO of which are depicted in
Figure 7.
Both the quinone LUMO-dienophile HOMO and qui-
none HOMO-dienophile LUMO interactions have a sec-
ondary component that is bonding, as represented in
Figure 8. Because the carbonyl groups are part of the
quinone’s conjugated system, secondary orbital interaction
can play a role in exo transition states as well, also illustrated
in Figure 8. The node at the carbonyl carbons in the quinone
HOMO precludes a secondary interaction with that orbital,
but there is one involving the quinone LUMO and dieno-
phile HOMO. This interaction is antibonding, and it there-
fore reinforces the endo preference in o-fluoranil’s
Diels-Alder reactions.
FIGURE 5. Calculated transition structure for endo addition of
butadiene to o-fluoranil (B3LYP/6-311G**þ).
trans-β-deuteriostyrene was exclusively trans (16), consistent
with concerted cycloaddition.
In Table 2 are presented calculated endo/exo transi-
tion state enhalpy differences, together with experimental
endo/exo ratios. For cis-2-butene and butadiene the
calculated values are misleading, but the deviations from
experiment are <1 kcal/mol. Endo stereochemistry domi-
nates in all cases except cis-2-butene. Why exo is slightly
preferred here is not clear, but this is the one dienophile for
which secondary orbital interactions should be unimpor-
In contrast to butadiene, trans,trans-2,4-hexadiene sur-
prisingly reacted with the quinone to give dioxene 17 instead
of Diels-Alder adducts, in agreement with theoretical pre-
diction (Figure 6). Since the thermodynamic bias in favor of
dioxene formation is similar for butadiene and hexadiene
9
tant.
We finally address the question of how configurations are
assigned to the Diels-Alder adducts. Which isomer is which
sometimes becomes apparent when an adduct mixture is
treated with water. Owing to steric hindrance, hydration of a
carbonyl group is often slower in the exo than in the endo
(
ΔΔH = 12.5 vs. 14.5 kcal/mol, respectively), nonbonded
repulsions involving the methyl group on the reacting double
bond may play a role in determining a different course of
reaction for the hexadiene. Again, cycloaddition at the
oxygens entails less nonbonded interaction in the transition
state than attack at carbon. A further example of the
difference a methyl group can make is provided by cis-
β-methylstyrene. In sharp contrast to styrene, it reacts with
(8) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. Engl. 1969, 8,
81.
9) For a discussion of other factors that may control endo/exo selectivity,
7
(
see: Lahiri, S.; Yadav, S.; Banerjee, S.; Patil, M. P.; Sunoj, R. B. J. Org.
Chem. 2008, 73, 435.
J. Org. Chem. Vol. 74, No. 20, 2009 7807

Lemal et al.
JOCArticle
FIGURE 6. Enthalpy changes in reactions of trans,trans-2,4-hexadiene with o-fluoranil.
TABLE 2. Comparison of Endo and Exo Cycloadditions and Derived
Quinoxalines
b
quinoxaline
a
adduct
calcd
q
exptl
vinyl
F δ
vinyl
C δ
cycloaddend configuration ΔΔH exo-endo endo
ethylene
cis-2-butene
-153.2
0.6:1 -151.4
-154.2
137.2
endo
exo
0.40
136.1
137.7
137.3,
135.9
137.9,
137.6
135.7,
137.5
butadiene
endo
-0.18
2:1 -150.5,
-
154.1
-152.8,
153.4
7:1 -149.0,
154.0
-152.4,
153.4
FIGURE 7. HOMO (left) and LUMO (right) of o-fluoranil
B3LYP/6-311G**þ).
exo
(
-
styrene
methyl
endo
exo
1.13
-
-
c
endo
2.77
3.81
-149.4,
-152.6
137.0,
135.3
2
acrylate
exo
c
dimethyl
maleate
a
endo
-149.2
135.9
Transition state enthalpy difference in kcal/mol, calculated at the
b
B3LYP/6-311G**þ level of theory. Chemical shifts relative to internal
c
CC1 F (for F) and TMS (for C). Exo isomer not unambiguously
identified.
FIGURE 8. Frontier orbital interactions in Diels-Alder reactions
of o-fluoranil: primary indicated in blue, secondary in green. Q =
quinone, D = dienophile.
3
Conclusions
form. The same is true for reaction of the adducts with
o-phenylenediamine to form quinoxalines. NMR che-
mical shifts of the vinyl CF groups of the quinoxalines are
useful for configurational assignment (Table 2). In the
quinoxaline from ethylene, the endo groups are unequivo-
cally hydrogens, so this molecule serves as a reference for
In the absence of substantial steric hindrance, o-fluoranil
normally undergoes Diels-Alder reactions with both electron-
rich and electron-poor dienophiles with preservation of stereo-
chemistry, and thus probably in concert. Hybrid density func-
tional calculations indicate that some, though not all, sym-
metrically substituted dienophiles react not only concertedly,
but also synchronously or nearly so. According to the theory,
unsymmetrical substitution on the dienophile generally results in
highly asynchronous reactions, but we have found that this need
not cause loss of stereochemical control. Steric effects can greatly
alter the course of events, however, resulting in reaction at
oxygen instead of carbon. There is a preference for endo addition
in the Diels-Alder chemistry of o-fluoranil, reinforced by a
repulsive secondary orbital interaction in exo transition states.
19
13
F and C chemical shifts. In the quinoxalines we have
studied, an endo hydrogen gives rise to a fluorine δ value less
than -152 ppm, but an endo substituent results in a value
greater than -152 ppm. Similarly, an endo hydrogen gives
rise to a vinyl carbon δ value greater than 137 ppm, but for an
endo substituent it is less than 137 ppm. Table 2 also shows
that when both endo groups are hydrogens, the vinyl fluorine
chemical shifts are closer together than when one group is
a substituent.
7
808 J. Org. Chem. Vol. 74, No. 20, 2009

Lemal et al.
JOCArticle
1
9
Experimental Section
solution, which showed a number of new F signals, contained
much unreacted exo adduct and a very small amount of
the endo.
A solution of o-phenylenediamine (617 mg, 5.71 mmol) in
2 2
0 mL of CH Cl was added, and much dark solid formed
In some of the following experiments, o-fluoranil was intro-
duced as a readily synthesized mixture with its much less reactive
para isomer (ratio 59:41),
Other experiments employed the ortho isomer alone, synthe-
sized by a variation to be published on the tetrafluorocatechol
preparation of Barthel and Bustrich followed by the catechol
oxidation of Shteingarts et al. Quinoxalines are designated by
the number of the corresponding Diels-Alder adduct followed
by a Q.
2
,10
which was removed in the workup.
1
immediately. The rapid formation of dark oxidation products
from the diamine indicated that much reduction of adducts
occurred in competition with the desired condensation reaction.
1
1
1
19
The mixture was refluxed and progress was monitored by
F
NMR. Because completion of quinoxaline formation from the
exo adduct was very slow, refluxing was continued for 3.5 h. The
mixture was dried over MgSO , filtered, and evaporated to leave
dark brown syrup plus solid, which was dissolved as much as
possible in hot benzene and placed on a column of silica gel
25 g). Elution with 30% ethyl acetate in hexanes yielded 653 mg
38%) of crude quinoxalines. The exo adduct eluted somewhat
faster than the endo, but separation was far from complete.
Recrystallization from methanol of a later fraction gave pure
1,4-Dihydro-1,2,3,4-tetrafluoro-1,4-ethanophenazine (4Q). A
4
heavy-walled Pyrex tube (12 mm ꢀ 200 mm) was charged with a
solution of 1.20 g of o,p-fluoranil mixture (3.9 mmol of ortho) in
benzene (11 mL). It was briefly evacuated and cooled in liquid
(
(
2
N . Ethylene (ca. 200 mL) contained in a balloon was vacuum
transferred to the reaction tube, which was sealed with a flame.
Placed in a pipe heater, the tube was maintained at 80-85 °C for
2
and the F NMR spectrum (CDCl
1 h. A small sample of the reaction mixture was evaporated,
) of the residue showed
19
endo quinoxaline, mp 191-192 °C. F NMR (CDCl ) δ -151.4
19
3
3
1
(
(
s, 2F), -203.9 (s, 2F); H NMR (CDCl
3
) δ 8.20 (m, 2H), 7.84
) δ 149.2, 140.2,
signals for the unreacted p-quinone at δ -142.0 and the ethylene
Diels-Alder adduct at δ -146.8 (s, 2F, vinyl) and -193.0 (s, 2F,
bridgehead). H NMR: δ 2.40 (m, 4H). The reaction solution
13
m, 2H), 2.54 (2H), 1.29 (6H); C NMR (CDCl
3
1
36.1 ( J =294 Hz), 130.7, 129.3, 91.5 ( JCF=219 Hz), 39.0,
1
1
1
1
CF
0.5. Anal. Calcd for C H F N : C, 62.33; H, 3.93; N, 9.09.
2
1
6
12
4
was concentrated to ca. 5 mL, a few drops of water was added,
and the mixture was shaken vigorously for several minutes. A
solution of o-phenylenediamine (0.75 g, 6.9 mmol) in 5 mL of
hot benzene was added. The resulting black mixture of liquid
and solid was boiled for several minutes, then the hot liquid was
transferred to a 25 g column of silica gel and followed with a
benzene wash of the residual solid. The column was eluted with
Found: C, 62.40; H, 3.99; N, 9.11.
An early fraction from the chromatogram was recrystallized
from methanol with decolorizing charcoal treatment and sub-
19
limed to give the exo quinoxaline, mp 172-173 °C. F NMR
1
(
8
(
(
CDCl
3
) δ -154.2 (s, 2F), -203.8 (s, 2F); H NMR (CDCl
3
) δ
.22 (m, 2H), 7.85 (m, 2H), 2.95 (2H), 0.83 (6H); C NMR
13
1
CDCl ) δ 147.9, 140.2, 137.7 ( JCF=299 Hz), 130.7, 129.4, 91.1
3
1
CF = 213 Hz), 39.0, 10.1; HRMS calcd for C16
08.0937, found 308.0935.
Dimethyl endo,cis-1,4-Dihydro-1,2,3,4-tetrafluoro-1,4-etha-
2
0% ethyl acetate/hexanes, and crystalline fractions rich in the
J
12 4 2
H F N
quinoxaline were combined in 30 mL of benzene. This solution
was extracted with 10% aqueous Na CO (10 mL) to remove
3
2
3
tetrafluorohydroquinone formed by reduction of the p-quinone.
After a wash with 5 mL of water, the benzene solution was dried
over MgSO4 with decolorizing charcoal present. Filtration
through filter-cel and evaporation of the filtrate left a pale
yellow solid (704 mg, 2.51 mmol, 64% yield). Treatment of the
product with a few milliliters of hot benzene left, after cooling, a
white solid that was recrystallized from benzene/hexanes to give
nophenazine-11,12-dicarboxylate (8Q). In a 25 mL round-bot-
tomed flask were placed commercial dimethyl maleate (2.5 g, 17
mmol), 2.1 g of o,p-fluoranil mixture (6.9 mmol ortho), some
CaCO , and 5 mL of toluene. The mixture was heated from
3
9
5 °C to the reflux temperature for 23 h. After cooling, the
mixture was treated with water (200 μL, 11 mmol) and thor-
oughly shaken. A hot solution of o-phenylenediamine (1.3 g,
1
9
pure quinoxaline, mp 189.5-190.5 °C. F NMR (CDCl
3
) δ
1
2 mmol) in 10 mL of benzene was added, and the resulting
1
-
153.2 (s, 2F), -198.2 (s, 2F); H NMR (CDCl
.85 (m, 2H), 2.58 (m, 2H), 2.26 (m, 2H); C NMR (CDCl
3
) δ 8.20 (m, 2H),
black mixture was boiled for several minutes. Residue from
evaporation of the toluene was transferred to a column of silica
gel (30 g), which was eluted with 20% ethyl acetate/hexanes.
Fractions containing the Diels-Alder adduct were combined in
1
3
7
1
2
1
3
) δ
=
CF
1
48.9, 140.1, 137.2 ( JCF = 294 Hz), 130.9, 129.3, 89.0 ( J
1
19 Hz), 29.6. Anal. Calcd for C14
0.00. Found: C, 60.06; H, 2.84; N, 10.04.
endo,cis- and exo,cis-1,4-Dihydro-1,2,3,4-tetrafluoro-11,12-
8 4 2
H F N : C, 60.00; H, 2.88; N,
2 2
CH Cl and treated with decolorizing charcoal. The mixture
was filtered and evaporated, then the residue was recrystallized
from methanol to give pure endo quinoxaline, mp 222.5-
dimethyl-1,4-ethanophenazine (6Q and 7Q). In a heavy-walled
Pyrex tube (12 mm ꢀ 200 mm) was placed a solution of
19
23.5 °C. F NMR (CDCl ) δ -149.2 (s, 2F), -201.1 (s, 2F);
2
3
o-fluoranil (1.00 g, 5.56 mmol) in 8 mL of CH
2
Cl
2
. The tube
1
H NMR (CDCl
3
) δ 8.21 (m, 2H), 7.89 (m, 2H), 3.83 (s, 6H),
was attached by way of a stopcock to a transfer bridge with a
cylinder of cis-2-butene connected via a graduated Y-tube to the
other end. After brief evacuation, the reaction vessel was cooled
in liquid nitrogen. The whole system was evacuated and filled
with nitrogen, then the Y-tube was cooled in a dry ice-
isopropanol bath. cis-2-Butene (1.1 mL, ca. 100% excess) was
condensed into the Y-tube, then allowed to transfer to the
reaction vessel. Sealed with a flame, the vessel was mounted in
13
3
.63 (s, 2H); C NMR (CDCl ) δ 167.3, 147.5, 140.3, 135.9
3
1
1
(
JCF=293 Hz), 131.5, 129.5, 89.1 ( JCF=225 Hz), 533.3, 49.9.
Anal. Calcd for C18 : C, 54.55; H, 3.05; N, 7.07.
Found: C, 54.56; H, 2.96; N, 7.08.
12 4 2 4
H F N O
Dimethyl maleate free from the fumarate was prepared by
treatment of a methanol solution of maleic acid with ethereal
diazomethane. In a 5 mL round-bottomed flask were placed
this ester (164 mg, 1.14 mmol), sublimed o-fluoranil (60 mg,
1
9
a pipe heater and maintained at 70 °C for 22 h. The F spectrum
CH Cl ) of the product showed singlets at δ -146.5 and -198.6
0
.33 mmol, mp 66.5-67.5 °C), some CaCO , and toluene
3
(
2
2
(
ca. 0.8 mL). When the mixture had been refluxed under
1
9
for the endo Diels-Alder adduct and at δ -147.9 and -198.3
for the exo adduct, in the ratio 0.6:1. Water (400 μL, 300%
excess) was added to the reaction mixture, which was shaken
for a few minutes. Some hydrate crystallized out, but the
nitrogen for 10 h, its F NMR spectrum showed the endo, cis
adduct, unreacted quinone, and at best a trace of the fumarate
adduct. Refluxing was continued for another 22 h to complete
the reaction, and again the amount of fumarate adduct was
1
9
negligible. F NMR (toluene) for the maleate adduct: δ -145.8
1
9
(
s, 2F), -196.5 (s, 2F). The F NMR (toluene) signals of an
(
10) Shteingarts, V. D.; Budnik, A. G.; Yakobson, G. G.; Vorozhtsov,
N. N., Jr. Zh. Obsh. Khim. 1967, 37, 1537.
11) Barthel, J.; Bustrich, R. German patent DE 19633027, 1988.
authentic sample of the fumarate adduct appeared at δ -143.3
(s, 1F), -149.9 (s, 1F), -195.0 (s, 1F), and -195.9 (s, 1F).
(
J. Org. Chem. Vol. 74, No. 20, 2009 7809

Lemal et al.
JOCArticle
trans-5,6,7,8-Tetrafluoro-2,3-diphenyl-2,3-dihydrobenzo[1,4]-
dioxin (9). In a 25 mL round-bottomed flask were placed cis-
stilbene (1.01 g, 5.6 mmol), 0.86 g o,p-fluoranil mixture (2.8
mmol ortho), and 4 mL of toluene. The black mixture was
refluxed under nitrogen for 23 h. The trans-dioxene and trans-
Diels-Alder adduct were present in the ratio 5:1, together with
quinoxaline-containing fractions consisted of yellow crystals
totaling 1.172 g (3.83 mmol, 47% crude yield).
Recrystallizations from hexanes and from methanol of an
endo-rich fraction (487 mg) failed to free it of the less-soluble exo
isomer (ratio 8.5:1), but it was pure otherwise. Mp 114-120 °C;
1
9
F NMR (CDCl
(s, 1F), -201.5 (s, 1F); H NMR (CDCl
7.86 (m, 2H), 5.90 (m, 1H), 5.43 (m, 2H), 3.05 (m, 1H), 2.52 (m,
3
) endo δ -150.5 (s, 1F), -154.1 (s, 1F), -199.0
1
9
1
unreacted p-quinone. F NMR (toluene) for the dioxene: δ
3
) endo δ 8.21 (m, 2H),
1
9
-
164.2 (m, 2F), -170.1 (m, 2F). F NMR (toluene) for the
Diels-Alder adduct: δ -144.4 (s, 1F), -149.0 (s, 1F), -194.6 (s,
F), -195.2 (s, 1F). The mixture was diluted with 15 mL of
CH Cl and extracted with 10 mL of 10% aqueous NaHSO to
1
3
1H), 2.39 (m, 1H); C NMR (CDCl ) endo δ 149.1, 148.8,
3
1
140.2, 140.1, 137.3 ( JCF=294 Hz), 135.9 ( JCF=294 Hz), 132.9,
1
1
1
130.9, 130.8, 129.4, 129.3, 121.5, 90.6 ( J = 220 Hz), 88.5
2
2
3
CF
1
( J = 220 Hz), 44.6, 37.0. Anal. Calcd for C H F N : C,
62.74; H, 3.29; N, 9.15. Found: C, 62.88; H, 3.20; N, 9.18.
remove the p-quinone. Phase separation was poor, so the
aqueous phase was washed with another 5 mL of CH Cl . The
combined organic phase was dried over MgSO with decoloriz-
CF
16 10
4
2
2
2
4
A fraction from the chromatogram containing slightly more
exo than endo quinoxaline (466 mg) was enriched in the exo
form by recrystallization from hexanes and from methanol/
water, but the endo isomer could not be completely eliminated.
Thus, the isomer mixture was recombined and chromato-
graphed on 30 g of silica gel with CH Cl as eluent. A fraction
ing charcoal added. After filtration through filter-cel, the solu-
tion was evaporated to leave a dark brown oil that was
chromatographed on 20 g of silica gel with 20% ethyl acetate/
hexanes as eluent. Fractions containing trans-stilbene and the
slightly slower moving dioxene were combined and rechroma-
tographed, again on 20 g of silica gel, but with 1% ethyl acetate/
hexanes as eluent. This time the stilbene and dioxene were
cleanly separated, and fractions bearing the latter were com-
bined and evaporated to give yellow crystals. These were
recrystallized from hexanes to yield a crop of thin, colorless
2
2
rich in the exo form was recrystallized from hexanes and then
from methanol/water to yield exo quinoxaline containing just
1
9
19
2.6% of the endo isomer by F NMR. Mp 154-156 °C;
F
NMR (CDCl ) exo δ -152.8 (s, 1F), -153.4 (s, 1F), -199.1 (s,
3
1
3
1F), -201.4 (s, 1F); H NMR (CDCl ) exo δ 8.21 (m, 2H), 7.86
1
9
plates (208 mg, 21%). Mp 152.5-153 °C; F NMR (CDCl
3
)
(m, 2H), 5.21 (m, 3H), 3.44 (m, 1H), 2.88 (m, 1H), 2.04 (m, 1H);
1
13
1
δ -163.8 (m, 2F), -169.2 (m, 2F); H NMR (CDCl ) δ 7.26 (m,
C NMR (CDCl ) exo δ 148.9, 147.4, 140.3, 140.2, 137.9 ( J
3
3
CF
1
3
1
=291 Hz), 137.6 ( JCF =291 Hz), 132.1, 131.0, 130.9, 129.5,
6
(
H), 7.05 (m, 4H), 4.98 (s, 2H); C NMR (CDCl
3
) δ 137.4
CF=248 Hz), 135.9 ( JCF=247 Hz), 134.4, 129.2 128.4, 127.6,
1.1. Anal. Calcd for C20 : C, 66.67; H, 3.36; F, 21.09.
1
1
1
1
J
129.4, 121.5, 90.5 ( JCF=219 Hz), 88.4, ( JCF=219 Hz), 45.3,
36.9.
8
12 4 2
H F O
Found: C, 66.72; H, 3.19; F, 21.23.
o,p-Fluoranil mixture (75 mg, 0.25 mmol ortho) that had been
recrystallized from hexanes was combined in a 5 mL round-
endo-1,4-Dihydro-1,2,3,4-tetrafluoro-11-phenyl-1,4-ethanophe-
nazine (14Q). A small sample of o,p-fluoranil mixture and excess
3
styrene were dissolved in CDCl in an NMR tube, and the
bottomed flask with cis-stilbene, CaCO
toluene. The mixture was stirred and refluxed for 16 h. Again
3
, and ca. 1 mL of
solution was allowed to stand for 23 h at room temperature.
The F spectrum revealed the presence of endo and exo
1
9
19
19
the F spectrum (toluene) showed peaks for the trans-dioxene
and Diels-Alder product, but in addition prominent signals at δ
Diels-Alder adducts in the ratio 7:1, respectively. F NMR
(CDCl ) endo δ -144.0 (s, 1F), -148.0 (s, 1F), -193.6 (s, 1F),
3
19
-
164.1 (m, 2F) and -169.1 (m, 2F) corresponding to the cis-
-193.7 (s, 1F); F NMR (CDCl ) exo δ -145.8 (s, 1F), -146.8
3
dioxene. A very small singlet at δ -194.0 may represent the
bridgehead fluorines of a cis Diels-Alder adduct.
(s, 1F), -193.3 (s, 1F), -194.0 (s, 1F). The endo quinoxaline was
prepared, and because no spectral information had been re-
1
1
ported, the following data appear here. F NMR (CDCl
9
endo- and exo-1,4-Dihydro-1,2,3,4-tetrafluoro-11-vinyl-1,4-
ethanophenazine (12Q and 13Q). Into a heavy-walled glass tube
with a sealable gas inlet connection was placed a solution of
3
)
δ -149.0 (m, 1F), -154.0 (s, 1F), -199.3 (d, J=4.5 Hz, 1F),
1
-199.6 (s, 1F); H NMR (CDCl
3
) δ 8.24 (m, 2H), 7.88 (m, 2H),
1
2
CH Cl . The vessel was cooled in liquid nitrogen, and butadiene
.50 g of o,p-fluoranil mixture (8.19 mmol ortho) in 15 mL of
7.44 (m, 5H), 3.60 (m, 1H), 2.83 (m, 2H); H NMR (C D ) δ 8.09
6 6
(m, 2H), 7.28 (m, 2H), 7.05 (m, 5H), 2.86 (m, 1H), 2.27 (m, 1H),
1.96 (m, 1H).
1,4-Dihydro-1,2,3,4-tetrafluoro-11-exo-deuterio-12-endo-phe-
2
2
(
0.70 g, 13 mmol) was introduced by vacuum transfer. The tube
was sealed and the reaction mixture was allowed to stand at
room temperature for 23 h. A sample was taken for NMR
analysis, and the solvent was replaced with CDCl . The princi-
nyl-1,4-ethanophenazine (16Q). trans-β-Deuteriostyrene was
synthesized by the method of Casey and Strotman. By H
1
2
1
3
pal products were the endo and exo Diels-Alder adducts in the
NMR, 2.5% of the styrene contained hydrogen at the trans, β
position. The small signal at that position (δ 5.15, CDCl )
3
comprised a doublet (J=10.8 Hz) superimposed upon a broad
singlet, corresponding to a mixture of undeuterated and
R-deuterated styrene. In a 25 mL round-bottomed flask were
placed 1.526 g of o,p-fluoranil mixture (5.00 mmol ortho),
1
9
ratio 2:1, respectively. F NMR for the endo adduct: δ -144.8
1
9
(
s, 1F), -148.1 (s, 1F), -193.7 (s, 1F), -195.7 (s, 1F). F NMR
for the exo adduct: δ -146.4 (s, 1F), -147.0 (s, 1F), -193.9 (s,
1
1
F), -195.9 (s, 1F). The reaction mixture was transferred to a
00 mL round-bottomed flask and water (300 μL, 16.7 mmol)
was added. After the mixture had been vigorously shaken for a
few minutes, a hot solution of o-phenylenediamine (1.57 g, 14.5
mmol) in 10 mL of benzene was added followed by a benzene
rinse. When the black mixture had been boiled for several
deuterated styrene (ca. 30% excess), CaCO
CH Cl . The mixture was stirred and refluxed for 10 h. Another
5 mL of CH Cl and water (200 μL, 11 mmol) were added, and
3
, and 5 mL of
2
2
2
2
the flask was shaken well for several minutes. Much hydrate
came out of soluton as a fine precipitate. A hot solution of
o-phenylenediamine (0.97 g, 9.0 mmol) in 8 mL of benzene was
added, quickly enough that foaming resulted in some mechan-
19
minutes, F NMR indicated that reaction had gone to comple-
tion. The mixture was concentrated to eliminate CH Cl , then
diluted with benzene to ca. 60 mL. It was filtered hot and the
filtrate was washed with 15 mL of 10% Na CO solution to
2
2
1
9
ical loss. After the mixture was boiled for several minutes,
F
2
3
remove tetrafluorohydroquinone formed by reduction of the
p-fluoranil. The organic layer was washed with water (5 mL) and
dried over MgSO with decolorizing charcoal added. Filtration
4
through filter-cel followed by evaporation left a concentrated
benzene solution that was placed on a column of silica gel (30 g).
Elution was done with 20% ethyl acetate/hexanes. The main
NMR indicated that reaction was complete. Evaporation left a
dark, viscous syrup that was chromatographed on 20 g of silica
gel with 20% ethyl acetate/hexanes as eluent. The exo quinoxa-
line eluted slightly faster than the endo, but the two were not well
(12) Casey, C. P.; Strotman, N. A. J. Am. Chem. Soc. 2004, 126, 1699.
7
810 J. Org. Chem. Vol. 74, No. 20, 2009

Lemal et al.
JOCArticle
separated (total weight 880 mg, 57% crude yield). Major frac-
tions were combined in CHCl and treated with decolorizing
3
(1.0 g, 5.6 mmol) in benzene was added a solution of trans,trans-
hexadiene (0.545 g, 6.0 mmol) in 10 mL of benzene, and the
resulting solution was stirred at room temperature for 4-5 h.
Evaporation of the solvent under reduced pressure left a gum that
was transferred with a minimal amount of hexane to a column of
charcoal. The mixture was filtered hot through filter-cel, then
concentrated to ca. 8 mL on a hot plate. Hexanes (12 mL) were
added, and the solution was cooled in a refrigerator to obtain
1
9
pure crystals of the endo quinoxaline, mp 184.5-185.5 °C.
NMR (CDCl ) δ -149.0 (s, 1F), -154.0 (s, 1F), -199.4 (s, 1F),
199.6 (s, 1F); H NMR (CDCl ) δ 8.23 (m, 2H), 7.88 (s, 2H),
F
2 2
silica gel. Elution was carried out with CH Cl /hexanes, and the
3
dioxene was obtained as a white crystalline solid (0.80 g, 55%
yield). It was then recrystallized from hexanes and sublimed, mp
1
-
3
1
19
7
.44 (m, 5H), 3.60 (d, J = 4.5 Hz, 1H), 2.85 (s, 1H); H NMR
) δ 8.09 (m, 2H), 7.28 (m, 2H), 7.05 (m, 5H), 2.86 (d, J =
74.5-75 °C. F NMR (CDCl
3
) δ -165.0 (m, 1F), -165.2 (m,
) δ 6.05 (m, 1H), 5.50 (m,
1H), 4.18 (t, J = 8 Hz, 1H), 3.96 (m, 1H), 1.83 (d, J = 6.5 Hz, 3H),
1
(
C
D
6 6
1F), -170.3 (m, 2F); H NMR (CDCl
3
1
3
4
1
1
4
.2 Hz, 1H), 2.27 (s, 1H); C NMR (CDCl ) δ 149.2, 140.2,
3
1 1
13
1
37.5 ( J =294 Hz), 135.9, 135.7 ( J =294 Hz), 131.0, 129.4,
1.40 (d, J=6.5 Hz, 3H); C NMR (CDCl ) δ 137.2 ( JCF=246
CF
CF
3
1
1
1
Hz), 135.6 ( JCF=246 Hz), 134.6, 130.0, 124.4, 79.9, 73.8, 18.0,
29.1, 128.9, 128.7, 91.5 ( JCF=221 Hz), 88.7 ( JCF=219 Hz),
5.9, 38.6. Anal. Calcd for C20 11DF
H
N
4 2
: C, 67.22; H þ D, 3.66;
10 4 2
16.9. Anal. Calcd for C12H F O : C, 54.97; H, 3.84; F, 28.99.
N, 7.84. Found: C, 66.90; H þ D, 3.22; N, 7.81.
Found: C, 55.07; H, 3.64; F, 29.15.
To have sufficient resolution in the aliphatic hydrogen region
to answer the question of cycloaddition stereochemistry, it was
1
necessary to run the H spectrum in C
Acknowledgment. The authors thank the National
Science Foundation for support of this work.
6
D
6
at 500 MHz. The peak
at δ 1.96 representing the exo methylene proton was very small.
Its area was consistent with the composition of the styrene
starting material, thereby indicating that the addition had
occurred without loss of stereochemistry.
trans-5,6,7,8-Tetrafluoro-2-methyl-3-(trans-1-propenyl)-2,3-di-
hydrobenzo[1,4]dioxin (17). To a saturated solution of o-fluoranil
Supporting Information Available: Total energies at the
1
13
B3LYP/6-311G**þ level of theory, H and C NMR spectra,
and general experimental information. This material is available
free of charge via the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 74, No. 20, 2009 7811