Benzodifurantrione: A stable phenylogous enol

Article abstract of DOI:10.1021/jo9022155

(Chemical Equation Presented) The first example of a stable phenylogous enol, resulting from an extended keto-enol tautomerization across a benzene ring, is described. The enol has been isolated, and its structure was proven by X-ray crystallography. The equilibrium between the keto- and enol-tautomers has been extensively studied and quantified in solution by NMR and UV-vis spectroscopy. The position of equilibrium showed a linear correlation to the Kamlet-Taft solvatochromic scale for solvent H-bond acceptor strength (β_OH), and the equilibrium was proven to be fully dynamic, obeying first-order equilibrium kinetics. To attempt to explain why enolization occurs, at what surprisingly appears to be the expense of aromatic resonance stabilization, various structural features have been considered and explored further with the aid of MO calculations. Nucleus independent chemical shift (NICS) index of aromaticity calculations for each of the rings comprising both tautomers showed that while the central benzene ring loses aromaticity on enolization, the R-keto-lactone ring showed an unexpected and significant antiaromaticity in the keto-tautomer, which is by no means intuitive. The loss of stabilization energy associated with the central benzene ring is, therefore, to a certain degree compensated by removal of the antiaromatic destabilization of the R-keto-lactone ring rendering the two structures much closer in energy than would otherwise be expected.

Full text of DOI:10.1021/jo9022155

pubs.acs.org/joc
Benzodifurantrione: A Stable Phenylogous Enol
Anthony J. Lawrence,*,† Michael G. Hutchings,*,† Alan R. Kennedy,‡ and
Joseph J. W. McDouall§
†DyStarUK Ltd., School of Chemistry, University of Manchester, Manchester M13 9PL, U.K., ‡Department
of Pure and Applied Chemistry, University of Strathclyde, Glasgow G1 1XL, U.K., and §School of Chemistry,
University of Manchester, Manchester M13 9PL, U.K.
anthony.lawrence@manchester.ac.uk; mike.hutchings@manchester.ac.uk
Received October 15, 2009
The first example of a stable phenylogous enol, resulting from an extended keto-enol tautomeriza-
tion across a benzene ring, is described. The enol has been isolated, and its structure was proven by
X-ray crystallography. The equilibrium between the keto- and enol-tautomers has been extensively
studied and quantified in solution by NMR and UV-vis spectroscopy. The position of equilibrium
showed a linear correlation to the Kamlet-Taft solvatochromic scale for solvent H-bond acceptor
strength (βΟH), and the equilibrium was proven to be fully dynamic, obeying first-order equilibrium
kinetics. To attempt to explain why enolization occurs, at what surprisingly appears to be the expense
of aromatic resonance stabilization, various structural features have been considered and explored
further with the aid of MO calculations. Nucleus independent chemical shift (NICS) index of
aromaticity calculations for each of the rings comprising both tautomers showed that while the
central benzene ring loses aromaticity on enolization, the R-keto-lactone ring showed an unexpected
and significant antiaromaticity in the keto-tautomer, which is by no means intuitive. The loss of
stabilization energy associated with the central benzene ring is, therefore, to a certain degree
compensated by removal of the antiaromatic destabilization of the R-keto-lactone ring rendering
the two structures much closer in energy than would otherwise be expected.
Introduction
developed to synthesize the enols directly.3a,5,6 Although
thermodynamically unstable, many have sufficient kinetic
stability in certain media to have allowed their structures to
be studied extensively.3b,c,5,7 Stabilized enol systems, such as
1,3-dicarbonyl,3d,8,9 1,5-dicarbonyl,10-13 R-nitro, R-cyano,
The equilibrium between keto- and enol-tautomers for
simple carbonyl compounds is normally strongly biased
toward the more thermodynamically stable keto form.1,2 As
such, much research has been targeted at the study of the
elusive enol entity, and its equilibrium with its keto-tautomer
has been a prime topic of study in organic chemistry for many
years.1,3 In non-stabilized systems, the so-called “simple enols”
of which vinyl alcohol is the simplest,4 methods have been
(5) (a) Capon, B.; Zucco, C. J. Am. Chem. Soc. 1982, 104, 7567–7572. (b)
Capon, B.; Rycroft, D. S.; Watson, T. W.; Zucco, C. J. Am. Chem. Soc. 1981,
103, 1761–1765.
(6) Capon, B; Siddhanta, A. K.; Zucco, C. J. Org. Chem. 1985, 50, 3580–
3584.
(7) Capon, B; Siddhanta, A. K. J. Org. Chem. 1984, 49, 255–257.
ꢀ
(8) Forsen, S.; Nilsson, M. In The Chemistry of the Carbonyl Group
(1) Toullec, J. Adv. Phys. Org. Chem. 1982, 18, 1–77.
(2) Keeffe, J. R.; Kresge, A. J.; Schepp, N. P. J. Am. Chem. Soc. 1988, 110,
1993–1995.
(Chemistry of Functional Groups Series); Zabicky, J., Ed; Interscience-Wiley:
London, 1970; Vol. 2, pp 198-226.
(9) Hesse, G. In Methoden der Organischen Chemie (Houben-Weyl);
€
(3) Rappoport, Z. The Chemistry of Enols (Chemistry of Functional
Groups Series); John Wiley & Sons Ltd: Chichester, 1990. (a) Chapters 5
and 8 of this book. (b) Chapters 4 and 8 of this book. (c) Stable “simple enols” are
reviewed extensively in Chapter 8, parts C and D of this book. (d) Chapter 6, pp
353-378 of this book. (e) Chapter 4 of this book. (f) Aliphatic and alicyclic
R-dicarbonyls are reviewed in Chapter 6, part V, pp 378-380 of this book.
(4) Hart, H. Chem. Rev. 1979, 79, 515–528.
Muller, E., Bayer, O., Eds.; Thieme: Stuttgart, 1978; Vol. 6-1d; pp 9-216.
(10) Fuson, R. C.; Melby, L. R. J. Am. Chem. Soc. 1953, 75, 5402–5405.
(11) Bayly, R. C.; Barbour, M. G. In Microbiological Degradation of
Organic Compounds; Gibson, D. T., Ed.; Marcel Dekker: New York, 1984;
pp 253-294.
(12) Whitman, C. P.; Aird, B. A.; Gillespie, W. R.; Stolowich, N. J. J. Am.
Chem. Soc. 1991, 113, 3154–3162.
690 J. Org. Chem. 2010, 75, 690–701
Published on Web 01/07/2010
DOI: 10.1021/jo9022155
r
2010 American Chemical Society

Lawrence et al.
JOCArticle
FIGURE 1. Implied phenylogous enolization across an aromatic
ring.
R-acyl, R-phosphonyl, and R-sulfonyl carbonyl compounds,8,9
are well-known. In these cases, the existence of at least one
canonical resonance form for the enol-tautomer renders its
energy more similar to or even lower than that for the keto-
tautomer.14 In addition, particularly in 1,3-dicarbonyl sys-
tems, an intramolecular H-bond between the enolic hydroxyl
and the carbonyl function can add further stabilization to the
enol-tautomer, especially in non-H-bonding solvents.3d
Although several examples exist for simple and stabilized
enols, to date, we believe there to be no examples of stable
extended phenylogous enols, i.e., enolization through an
aromatic ring to give a hydroxyquinone dimethide as implied
by 1 / 2 in Figure 1. Although such a process is theoretically
feasible, clearly the enol-quinonoid form 2 lacks the aro-
matic resonance stabilization of the keto-benzenoid form 1,
which would be expected to render such an enol much higher
in energy and hence thermodynamically unstable. Therefore,
an absence of stable examples is not too surprising.
FIGURE 2. Proposed generation of a transient phenylogous enol
by reduction of an anthracycline anticancer drug, exemplified with
daunomycin 3.16
FIGURE 3. Proposed solvent-dependent phenylogous keto-enol
tautomerization and enolate formation of BDT 6a-c.
Hydroxyquinone dimethides, i.e., 2, have been proposed
as short-lived intermediates in bioactivation pathways,15
inter alia in the biochemical (and chemical) reduction of
anthracycline anticancer drugs.16-18 One example, shown in
Figure 2, is the tautomer 4 of 7-deoxydaunomycinone 5,
derived from the drug daunomycin 3 by reduction and
elimination.16a-c However, enolization in these cases across
a benzene ring to form a quinone dimethide (2) itself has
not been observed. The enolic intermediates have been
formed indirectly from reduction and elimination and con-
vert rapidly to their thermodynamically more stable keto-
benzenoid tautomers.
that has been of interest to us for some time, not least as a
versatile intermediate to commercial benzodifuranone
dyes.19 Early attempts at structural analyses showed BDT
6a to be more complex than would be expected, where it
appeared to exist in different solvent-dependent forms.19b
For example, in toluene or CH2Cl2, BDT gave a pale yellow
solution that turned brownish red on standing, whereas in
basic and/or H-bonding milieu such as DMF or DMSO,
BDT gave an intense purple solution. These profound color
differences led to the postulate that keto-BDT 6a may in fact
undergo this unusual form of phenylogous tautomerism
allowing it to exist to a greater or lesser extent as the extended
enol-quinonoid tautomer 6b and/or enolate 6c, without
the use of strong base or the need for its generation from
a precursor (Figure 3). However, given the absence of
examples of this phenomenon in the literature, we remained
unconvinced.
7-Phenyl-7H-benzo[1,2-b;4,5-b0]difuran-2,3,6-trione 6a,
known trivially as benzodifurantrione or BDT, is a molecule
(13) Lam, W. W. Y.; Bugg, T. D. H. J. Chem. Soc., Chem. Commun. 1994,
1163–1164.
(14) Waring, A. L. In Comprehensive Organic Chemistry; Stoddart, J. F.,
Ed.; Pergamon: Oxford, 1979; Vol. 1, pp 1024-1026.
In this article, we now show that BDT is indeed the
first example of a stable phenylogous enol. The enol has
been isolated, and its structure was proven by X-ray crystal-
lography. The equilibrium between the keto- and enol-
tautomers has been studied and quantified in solution and
proven to be fully dynamic. We also attempt to explain why
BDT is such a special case, allowing enolization to occur, at
what appears at first to be loss of aromatic stabilization.
(15) (a) Moore, H. W. Science 1977, 197, 527–532. (b) Moore, H. W.;
Czerniak, R. Med. Res. Rev. 1981, 1, 249–280.
(16) (a) Kleyer, D. L.; Koch, T. H. J. Am. Chem. Soc. 1984, 106, 2380–
2387. (b) Kleyer, D. L.; Koch, T. H. J. Am. Chem. Soc. 1983, 105, 2504–2505.
(c) Boldt, M.; Gaudiano, G.; Koch, T. H. J. Org. Chem. 1987, 52, 2146–2153.
(d) Gaudiano, G.; Frigerio, M.; Sangsurasak, C.; Pierfrancesco, B.; Koch, T.
H. J. Am. Chem. Soc. 1992, 114, 5546–5553. (e) Gaudiano, G.; Frigerio, M.;
Pierfrancesco, B.; Koch, T. H. J. Am. Chem. Soc. 1992, 114, 3107–3113. (f)
Schweitzer, B. A.; Egholm, M.; Koch, T. H. J. Am. Chem. Soc. 1992, 114,
242–248. (g) Gaudiano, G.; Koch, T. H. J. Am. Chem. Soc. 1990, 112, 9423–
9425. (h) Bird, D. M.; Boldt, M.; Koch, T. H. J. Am. Chem. Soc. 1989, 111,
1148–1150. (i) Kleyer, D. L.; Gaudiano, G.; Koch, T. H. J. Am. Chem. Soc.
1984, 106, 1105–1109.
Results and Discussion
BDT in the Solid State. BDT is highly pigment-like and
shows limited solubility in many solvents. However, suitable
single crystals were grown from methyl-isobutyl ketone to
allow X-ray structure analysis. Figure 4a shows the contents of
the asymmetric unit, in which two independent conformers
(17) Kleyer, D. L.; Koch, T. H. J. Am. Chem. Soc. 1983, 105, 5154–5155.
(18) (a) Ramakrishnan, K.; Fischer, J. J. Med. Chem. 1986, 29, 1215–
1221. (b) Ramakrishnan, K.; Fischer, J. J. Am. Chem. Soc. 1983, 105, 7187–
7188. (c) Gaudiano, G.; Resing, K.; Koch, T. H. J. Am. Chem. Soc. 1994, 116,
6537–6444. (d) Bird, D. M.; Gaudiano, G.; Koch, T. H. J. Am. Chem. Soc.
1991, 113, 308–315. (e) Gaudiano, G.; Frigerio, M.; Pierfrancesco, B.; Koch,
T. H. J. Am. Chem. Soc. 1990, 112, 6704–6709. (f) Gaudiano, G.; Egholm,
M.; Haddadin, M. J.; Koch, T. H. J. Org. Chem. 1989, 54, 5090–5093. (g)
Egholm, M.; Koch, T. H. J. Am. Chem. Soc. 1989, 111, 8291–8293. (h) Boldt,
M.; Gaudiano, G.; Haddadin, M. J.; Koch, T. H. J. Am. Chem. Soc. 1989,
111, 2283–2292.
(19) (a) Hughes, N.; Newton, D. F.; Milner, D. J.; Deboos, G. A. World Patent
WO 94/12501. (b) James, M.; Bradbury, R. World Patent WO 95/28447.
J. Org. Chem. Vol. 75, No. 3, 2010 691

Lawrence et al.
JOCArticle
SCHEME 1. Synthesis of Locked Enol-Ether 7, R-Keto-Ester 8,
and R-Keto-Acid 9
hydroxyl and adjacent lactone carbonyl or exist as dimers as is
found for tropolone-type structures,20 the molecules form
intermolecular H-bonded chains between the enol OH (O1)
of one conformer and the carbonyl oxygen (O5) of the
opposite lactone ring of the other conformer.
IR analysis of BDT in the solid state (KBr) also supported
the enol-tautomer 6b. Absorbances at 1782 and 1692 cm-1
were observed for the two lactones; the higher frequency of
one of the lactone stretches is explicable due to H-bonding to
the enolic OH, as seen in the crystal structure. Further
absorbances at 1623 and 1579 cm-1 were also evident and
appear to be characteristic of the double bonds of the
quinone dimethide.
BDT in Solution; Keto/Enol Standards. In parallel with the
solid-stateinvestigations, we looked in depth at the behavior of
BDT in solution by NMR and UV-vis spectroscopy. To aid
investigation, enol-locked analogue 7 was prepared by treating
6a in toluene with TMS-diazomethane (Scheme 1).21 All ana-
lytical data were consistent with the proposed structure 7.
The IR spectrum (KBr) was similar to that for BDT-enol
6b above. Absorption bands were observed at 1785 and
1753 cm-1 for the two lactones and at 1620 and 1583 cm-1
for the quinone dimethide. In this case, both lactone stretches
had a similar frequency since no intermolecular H-bonding
exists, unlike for BDT-enol 6b.
FIGURE 4. (a) Asymmetric unit showing two independent con-
formers, both enol structures 6b, differing in the twist angles
(13.3(2)° and 24.9(2)°) of the exo-7-phenyl ring. (b) Intermolecular
H-bonding observed between the two conformers displayed in green
and blue. (c) Atom labels and key bond lengths (A; in blue) for one
conformer (13.3(2)° twist).
˚
1
The H NMR spectrum (CD3CN, Figure 5a) showed no
were observed; Figure 4b shows the intermolecular hydrogen
bonding, and Figure 4c shows the important bond lengths in
one of the conformers. The conformers differed only by the
twist angle of the 7-phenyl ring, which was 13.3(2)° and
24.9(2)° in each case (as defined by the dihedral angle
between the least-squares planes of the phenyl rings and
the five-membered rings to which they are attached). In the
crystal structure, both conformers exist as the enol-tautomer
6b. The central ring is clearly quinonoid rather than benzenoid
with the bond lengths of the formal CdC double bonds
(C4-C5, C7-C8) in the central ring being shorter (1.347(4),
benzylic proton resonance and showed the methyl group
as a 3-proton singlet at δ 4.33 ppm, indicating that it was
O- rather than C-linked. The two central ring proton reso-
nances were seen at δ 6.97 and δ 6.93 ppm, and the two ortho-
protons of the 7-phenyl ring appeared at δ 7.75 ppm. The
relatively high-field shifts for the central ring protons
are comparable to the shifts seen for 1,4-benzoquinone
(δ 6.76 ppm).22 The upfield shifts for these central ring
protons and the relatively low-field shift for the 7-phenyl
ortho-protons proved to be extremely characteristic of the
enol-quinonoid form. 13C NMR showed the two lactone
resonances at δ 168.0 and δ 162.2 ppm, and no lower field
ketone resonance was observed. Instead, the derived
CdC-OMe enolic carbon was observed at δ 145.1 ppm,
displaying a strong correlation in an HMBC experiment to
˚
1.341(4) A) than the formal single bonds (1.426(4)-1.442(4)
˚
A). The H-atom is found on the enolic OH group (O1) and the
benzylic carbon (C10) is sp2 hybridized with the three bond
angles at these centers totaling 360.0°. Rather than adopt
an intramolecular five-membered H-bond between the enol
(21) Coe, D.; Drysdale, M.; Philps, O.; West, R.; Young, D. W. J. Chem.
Soc., Perkin Trans. 1 2002, 2459–2472.
(20) Shimanouchi, H.; Sasada, Y. Acta Crystallogr., Sect. B 1973, 29, 81–
(22) Tamura, Y.; Yakura, T.; Tohma, H.; Kikuchi, K.; Kita, Y. Synthesis
1989, 126-127, compound 3c.
90.
692 J. Org. Chem. Vol. 75, No. 3, 2010

Lawrence et al.
JOCArticle
the OCH3 enol ether carbon, which itself resonated at δ
60.8 ppm. The two central quinonoid ring methine carbon
resonances were also highly distinctive, observed almost
coalesced at δ 97.97 and δ 98.00 ppm.
UV-vis spectra of 7 were measured in numerous solvents,
and the extinction coefficients were calculated. Compound 7
displayed modest solvatochromism, which was ascribed to
solvent polarizability,23 with extremes of λmax 432 nm
(MeCN) and λmax 446 nm (o-dichlorobenzene) for the sol-
vents studied. To normalize for varying absorbance peak
shape across the studied solvents, the extinction coefficients
were multiplied by the fwhm (full width at half maximum) to
give the approximate peak area (ε ꢀ fwhm).24 Full results
are detailed in Table 1. With the exception of THF and
2-propanol, the ε ꢀ fwhm values were relatively consistent
across all solvents studied, with an average of 2.96 ꢀ 106
(range, 2.81 ꢀ 106 - 3.10 ꢀ 106). The values provided for
THF and 2-propanol are averages over five measurements.
In THF compound 7 gave a cloudy solution, and it was
highly insoluble in 2-propanol, leading to difficulties in
accurate measurement and explaining the outlying results.
These two solvents were thus omitted in calculating the
average ε ꢀ fwhm value.
During UV-vis analysis (below), it was observed that
dissolution of BDT 6a in methanol led to ring opening to give
R-keto-methyl ester 8. A standard of 8 was prepared by
stirring for 24 h at room temperature in methanol. The
acquired 1H NMR spectrum (Figure 5b) was fully consistent
with 8 existing solely as the keto-benzenoid tautomer in all
deuterated solvents other than DMSO-d6, and as such served
as a useful keto-tautomer NMR standard. In CD3CN, the
benzylic proton resonated at δ 5.15 ppm, the two central ring
protons at δ 6.91 and δ 7.61 ppm, and in this case
the ortho-protons of the 7-phenyl ring appeared upfield at
δ 7.26 ppm. The distinctive differences in the shifts of the
central ring and 7-phenyl ring ortho-protons in the ring-
opened R-keto-ester 8 are notable compared with enol-
ether 7. IR data in the solid state showed absorbances at
1791 cm-1 for the β,γ-unsaturated lactone, 1741 cm-1 for the
ester, and 1624 cm-1 for the intramolecularly H-bonded aryl
ketone. A single absorbance was seen in the UV region at
λ
max 368 nm with εmax of 4,250 and fwhm of 62 nm.
BDT in Solution; NMR Studies. 1H NMR studies of BDT
FIGURE 5. 1H NMR (CD3CN) key resonances for (a) locked enol-
tautomer 7, (b) keto-tautomer standard 8, (c) BDT showing 1:2
[keto 6a]:[enol 6b] ratio from integration.
were conducted in CD3CO2D and in a range of other deuter-
ated solvents containing a trace of methane sulfonic acid
(MSA) to determine the keto/enol constitution. The inclusion
of MSA was required to remove any ionized-BDT 6c from the
equilibrium (see UV-vis studies below); omission of MSA
gave broad spectra with multiple species. All solutions were
heated to ca. 100 °C or boiling (depending on solvent boiling
point) to ensure equilibrium was achieved from the solid-state
constitution. The peaks were sharp and well resolved indicat-
ing that keto-enol exchange was slow on the 1H NMR time
scale at ambient temperature. In all cases both keto- and enol-
tautomers were observed, and the tautomer ratios were
calculated from integration of the spectra.
As an example, Figure 5c shows the 1H NMR spectrum for
BDT in CD3CN compared to locked enol-tautomer 7
(Figure 5a) and keto-standard 8 (Figure 5b). Both tautomers
are clearly visible with a [keto]:[enol] ratio of 1:2 by integration.
For the BDT keto-tautomer 6a, the diagnostic protons resonate
at δ 5.23 ppm (Ha), δ 7.17 ppm (Hb), δ 7.27 ppm (Ho), and
δ 7.56 ppm (Hc), which are consistent with the respective
shifts of the protons in the keto-standard 8 at δ 5.17 ppm
(Ha), δ 6.95 ppm (Hb), δ 7.23 ppm (Ho), and δ 7.58 ppm (Hc).
Likewise, the diagnostic protons for the enol-tautomer 6b
resonate at δ 6.75 ppm (Hb), δ 6.87 ppm (Hc) and δ 7.75 ppm
(Ho), again consistent with the respective shifts of the protons in
the enol-standard 7 at δ 6.93 ppm (Hb), δ 6.97 ppm (Hc) and
δ 7.75 ppm (Ho). Results of keto/enol ratios for BDT calculated
from 1H NMR spectra in other solvents studied are summarized
along with UV-vis data in the next section in Table 3.
ꢀ
(23) Catalan, J.; Hopf, H. Eur. J. Org. Chem. 2004, 4694–4702.
(24) We use the ε ꢀ fwhm area approximation rather than ε0, the exact
area derived extinction coefficient, since the true area could not be measured
due to the overlap of the keto (6a) absorbance (λmax ca. 295 nm) with the
hypsochromic tail of the enol (6b) absorbance and with solvent absorption in
some cases (e.g., acetone) despite the use of a blank balancing cell.
J. Org. Chem. Vol. 75, No. 3, 2010 693

Lawrence et al.
JOCArticle
TABLE 1. UV Data for Locked Enol-Tautomer 7 in a Variety of
Solvents
TABLE 2. Assignment of Key 13C Resonances of BDT-Keto 6a and
BDT-Enol 6b Tautomer Mixture in CD3CN by Comparison to Keto-Ester
8 and Enol-Ether 7 Standards and to Literature-Reported Shifts for
Coumarandione25
λmax
εmax
fwhma
(nm)
ε ꢀ fwhm
solvent
acetic acid
acetone
(nm) (M-1 cm-1
)
(ꢀ106)
434 37,250
433 39,250
432 38,500
444 39,250
444 39,250
440 37,000
435 42,000
79
78
78
79
78
77
71
2.94
3.06
3.00
3.10
3.06
2.85
2.98
acetonitrile
benzonitrile
chlorobenzene
chloroform
cyclohexane/toluene
(90/10)
carbon
δ
C ppm 6a
δ
C ppm 8a
δ
C ppm 6b
δ
C ppm 7a
a
b
c
d
e
f
g
h
i
52.0b
50.6
139.2
106.6
159.0
162.5
188.8
110.7
111.3
146.5
174.2
122.9
141.3
97.4b
152.0
162.0e
144.3
117-120d
98.8b
120.1
138.7
98.0
149.8
162.2
145.1
116.6
98.0
o-dichlorobenzene
dichloromethane
DMSO
1,4-dioxane
ethyl acetate
446 36,500
438 38,500
445 35,750
437 36,500
433 38,500
78
77
84
77
78
78
78
80
77
2.85
2.96
3.00
2.81
3.00
2.98
2.52 ( 0.46
2.66
2.85
135.7
112.5b
164.9e
152.7c
178.8
methyl isobutyl ketone 435 38,250
117-120d
107.7b
142.4
2-propanolb
THFc
toluene
435 32,250 ( 6000
436 33,250
443 37,000
155.5
169.8
153.5
168.0
j
175.0c
aFull width at half maximum. bSparingly soluble in this solvent;
measured εmax may be inaccurate. cCloudy solution; measured εmax may
be inaccurate.
aAssigned by HMQC and HMBC analysis. bAssigned by HMQC.
d
cWeak peak observable just above noise threshold. Peak obscured by
solvent CN resonance at 117-120 ppm. ePossible that these two peaks
are inversely assigned.
Obtaining clear 13C NMR data for the BDT tautomeric
mixture was problematic because of its low solubility. In
solvents such as DMSO where solubility was good, BDT
existed solely as its enolate 6c (properties discussed below).
In solvents offering lower but sufficient solubility, such as
acetone, dioxane, or 2-propanol, the equilibrium was dis-
placed too far toward the enol to resolve the keto resonances
with certainty above the noise. In acetonitrile-d3 (containing
33% of the keto-tautomer), signal-to-noise was just suffi-
cient to observe all key carbon resonances for both tauto-
mers (Table 2) with the exception of carbon Cg in both
tautomers (as defined below), which were lost under the
solvent nitrile peak. Assignments were made aided with
HMQC analysis, by comparison to the enol-ether and
keto-ester standards 7 and 8, and by comparison to literature
reported shifts for close analogue coumarandione (10;
Figure 6).25 Due to low BDT solubility, meaningful HMBC
data could not be obtained to allow direct unambiguous
assignment of all carbons. Therefore, assignment of those
carbons without an HMQC correlation could only be in-
ferred by comparison. The majority of resonances for the
keto-tautomer 6a were comparable to the keto-standard 8.
Of note are the benzylic methine carbon Ca at δ 52.0 ppm,
lactone carbon Cj at δ 175.0 ppm, and the central benzene
ring tertiary carbons Cc and Ch at δ 112.5 and δ 107.7 ppm. In
this case, Cc was at lower field than Ch (determined from the
HMQC data), unlike in the keto-standard 8, in the enol-
tautomer 6b, and in the enol-ether 7. The shifts for the ketone
and lactone carbons Cf (δ 178.8 ppm) and Ce (δ 152.7 ppm) in
the keto-tautomer 6a were significantly upfield, cf. standard
8. However, these shifts were consistent with those deter-
mined for the respective carbons of coumarandione at δ
177.1 and δ 155.9 ppm.25 The further shielding of these
carbons in the ring-closed R-keto-lactone tautomer of cou-
marandione 10 and BDT-keto 6a compared to the ring-
opened R-keto-ester 8 is consistent with antiaromatic char-
acter in this ring, a key point that is discussed in detail below.
FIGURE 6. Selected reported 13C shifts for coumarandione 10.25a
No evidence existed in the 13C spectrum for the possible
alternative ketone-hydrate structure. The chemical shifts for
the enol-quinonoid tautomer 6b were very similar to those
observed for the locked enol-ether 7. Those of note are the two
tertiary central quinonoid ring carbons Cc and Ch at δ 97.4
and δ 98.8 ppm, the enol carbon Cf at δ 144.3 ppm, and the
two lactone carbons Ce and Cj at δ 162.0 and δ 169.8 ppm.
BDT in Solution; UV-vis Studies. The UV-vis spectrum
of BDT was measured in a range of solvents, and data are
summarized in Table 3. Example spectra are reproduced in
Figure 7. In DMSO and DMF, BDT gave a bright purple
solution with a single absorbance in the visible region at
554 and 552 nm, respectively. In acetic acid, CHCl3, and
Cl3CCO2Et, a single absorption was seen in the visible region
at 432 nm. In all other solvents both absorptions were
observed at 425-444 and at 530-550 nm, the proportions
of which shifted on standing in favor of the bathochromic
530-550 nm absorbance. Adding a trace amount of methane
sulfonic acid (MSA) removed the 530-550 nm absorption
from the spectrum for the solvents other than DMF and
DMSO and left a dominant absorption at 425-444 nm.
Knowing that locked enol-ether 7 had an average λmax of
438 nm in the solvents studied, it followed that the absorp-
tion band seen for BDT at 425-444 nm in the same solvents
was due to the electronically comparable enol-quinonoid
tautomer 6b. The absorption at ca. 550 nm, giving rise to the
purple coloration, was explained by BDT being fully or
partially ionized, i.e., as its extended enolate 6c, which bears
a stronger donor leading to a bathochromic shift. This form
of BDT is discussed in more detail later in this article. In
acidic media, such as in acetic acid or by adding a trace of
(25) (a) Winkler, T.; Ferrini, P. G.; Haas, G. Org. Magn. Reson. 1979, 12,
101–102. (b) Galasso, V.; Pellizer, G.; Pappalardo, G. C. Org. Magn. Reson.
1980, 13, 228.
694 J. Org. Chem. Vol. 75, No. 3, 2010

Lawrence et al.
JOCArticle
TABLE 3. UV-vis and 1H NMR [keto 6a]:[enol 6b] Ratio Data Measured in Various Solvents Containing a Trace of MSA
εmax
fwhmb
(nm)
εmax
ꢀ
% enol by
UV-visc
% enol by
NMRd
ΔG
(M-1 cm-1
)
fwhm (ꢀ106)
log Ke
(kJ mol-1
)
a
solvent
acetic acid f
acetone
acetonitrile
anisole
benzonitrile
chlorobenzene
chloroform f,h
o-dichlorobenzene
DMFi
dimethyl phthalate
DMSOi
1,4-dioxane
ethyl acetate
ethyl trichloroacetate f
methyl isobutyl ketone
2-propanol
tetrahydrofuran
toluene
HBA βOH
λ
max (nm)
0.45
0.49
0.39
0.21
0.37
0.07
0.10
0.03
0.73
0.41
0.73
0.46
0.50
0.25
0.48
0.84
0.59
0.10
433
432
428
441
444
436
431
438
552
440
554
435
430
433
433
439
436
437
31,250
32,250
23,000
15,500
28,500
11,500
(8,750)h
11,250
35,000
26,250
35,000
32,500
30,250
16,250
33,500
33,500
34,750
11,500
81
81
80
82
74
82
82
85
2.53
2.61
1.84
1.27
2.11
0.94
0.72
0.96
86
88
62
43
90
88
66
0.954
0.865
0.288
-5.44
-4.94
-1.63
þ0.71
-2.22
þ1.88
þ1.00
þ 1.21
-0.122
0.389
71
32
25g
40
-0.327g
-0.176
-0.213
(24)h
32
38
exists as enolate
72
exists as enolate
86
81
46
91
99
81
2.13
0.410
-2.34
78
79
84
80
87
85
83
2.54
2.39
1.37
2.68
2.91
2.95
0.96
88
0.865
0.630
-0.070
1.005
2.299
1.33
-0.327
-4.94
-3.56
þ0.42
-5.73
-13.1
-7.53
þ1.88
99.5
95.5
100
32
aKamlet-Taft solvatochromic scale of H-bond acceptor strength (βΟH).33 bFull width at half maximum. c% Enol calculated from (εmax6b ꢀ fwhm6b)/
2.96 ꢀ 106 (2.96 ꢀ 106 is the average εmax ꢀ fwhm for enol-ether 7 from Table 1). d% Enol calculated from comparison of averaged keto/enol integrals in
the 1H NMR spectrum where measured. log K based on NMR enol content where available, otherwise calculated from UV-vis estimates. MSA
omitted due to intrinsic solvent acidity. glog K based on UV due to broad NMR spectrum. hBDT highly insoluble in this solvent; measured εmax may be
inaccurate. iMSA omitted; BDT existed as enolate and could not be fully protonated in this solvent.
e
f
MSA to the solvents (excluding DMF and DMSO), BDT
existed fully protonated and the 550 nm absorption dis-
appeared,26 whereas addition of a trace of Et3N gave a full
shift to the 550 nm absorption.
The keto-benzenoid tautomer 6a was observed between
292 and 300 nm for solutions of BDT in solvents transpa-
rent to relatively high energies and which supported high
keto contents. Treatment with solvents that favored enol-
tautomer formation (e.g., 2-propanol) eliminated the ca.
300 nm band assigned to keto-BDT 6a, and the intensity of
enol absorption at ca. 430 nm correspondingly increased
(Figure 7).27
After removing the enolate 6c by addition of MSA, the
observed solvent dependency of the extinction coefficients
for the ca. 430 nm absorption was directly related to the
proportion of enol-quinonoid in that solvent at equilibrium.
The absorbance for 6b in acidified media showed a linear
relationship when plotted versus concentration, demonstrat-
ing that the keto:enol equilibrium was not perturbed by
concentration over the range studied.28 Knowing that the
average value of ε ꢀ fwhm for enol-ether 7 was 2.96 ꢀ 106 in
FIGURE 7. Selected UV-vis spectra for BDT (6a-c) and enol-
ether 7 in various media. Spectra for BDT in CHCl3 with and
without 2-propanol were measured at identical concentration.
the solvents studied, it was possible to use this measurement
to estimate the [keto]:[enol] ratio for BDT at equilibrium
by comparison, e.g., (εmax6b ꢀ fwhm6b)/2.96 ꢀ 106.29 This
assumed that the enol 6b had an extinction coefficient equal
to that for the enol-ether 7, which is not necessarily always
the case.30 However, in this instance, this assumption gave
figures consistent with the ratios observed by NMR in the
same solvents and thus served as a useful and simple method
to estimate the enol-quinonoid 6b content.31 The calculated
[keto]:[enol] ratios are given in Table 3 with comparison to
the observed NMR ratios where measured. In all cases, the
solutions were prepared using solvents with <50 ppm water
content and by heating to ca. 100 °C or boiling to ensure
equilibrium was reached. Solutions in relatively water-im-
miscible solvents, such as chlorobenzene, showed little
change in εmax on standing overnight. In other water-
miscible solvents, a slow decline in εmax over time was observed
(ca. 10-20% overnight), which was attributed to the slow
hydrolysis of the R-keto-lactone to the ring-opened R-keto-
acid 9. This was confirmed experimentally by leaving a 10%
water/dioxane solution of BDT containing a trace of MSA to
(26) Presumably the commercially obtained chloroform and ethyl chloro-
acetate used for the UV-vis studies contained sufficient residual acid to
protonate dissolved enolate without the requirement of additional MSA.
(27) The keto (6a) UV assignment is discussed further in Supporting
Information.
(28) One example in acetone is supplied in Supporting Information.
(29) Further notes on the selection of ε ꢀ fwhm for enol quantification by
UV-vis are available in Supporting Information.
(30) Rhoads, S. J.; Pryde, C. J. Org. Chem. 1965, 30, 3212–3214.
(31) Accuracy of measurements by UV-vis was probably less than by
NMR. When UV-vis spectra were run in triplicate, around 5% variation
was found in the calculated enol content. In addition, slight perturbations
from the averaged value for ε ꢀ fwhm of 2.96 ꢀ 106 for the enol-ether 7 in
each specific solvent may have existed.
J. Org. Chem. Vol. 75, No. 3, 2010 695

Lawrence et al.
JOCArticle
stand overnight to accelerate the process. After this time, no
BDT remained, and TLC showed that the resultant material
coeluted with an authentic sample of 9, prepared by treat-
ment of BDT 6a with dilute aqueous sodium hydroxide
(Scheme 1).
BDT in Solution; Studies on Equilibrium Dynamics. Having
a convenient method for estimating the BDT [keto 6a]:[enol6b]
ratio in solution, we investigated one example of the
keto / enol equilibrium dynamics, approaching from both
low and high initial keto content. Starting from a weakly acidified
o-dichlorobenzene solution (30% enol) and adding 1% EtOAc
(final concentration, 2.9 ꢀ 10-5 mol L-1), the increasing enol
content was monitored from the change in the visible absor-
bance at 438 nm over time at 25 °C. The final enol content was
54% (K= 1.16) at equilibrium, which was reached after 165 min
under these conditions. A plot of -ln([keto]t - [keto]eqm) versus
Solvent Parameters of Abboud and Notario,34 augmented by
a few other β values from standard compilations.35 In fact it
was found that the Gibbs energy difference ΔG = -2.303RT
log10 K gave a satisfactory linear correlation with solvent
HBA strength (or H-bond basicity) as shown in Figure 8.
Inclusion of any of the other Kamlet-Taft parameters in a
multilinear treatment was statistically insignificant. Thus,
stronger solvent bases interact more favorably with, and
stabilize, the relatively acidic enolic OH group of 6b resulting
in a displacement of the equilibrium toward the enol, i.e.,
6d (Figure 8). Extrapolation of the relationship shown in
Figure 8 to the gas-phase H-bond basicity of β = 0 predicts a
ΔG of about 5.9 kJ mol-1, suggesting that in the absence of
HBA solvent stabilization the keto-tautomer 6a is favored to
the extent [keto]:[enol] = ca. 90:10. This value will be related
to calculated relative energies of keto- and enol-tautomers in
the following section. Data for solvents in which BDT fully
ionizes to give 6c (DMSO, DMF, etc.) could not be included
since it was impossible to acidify the solutions sufficiently to
obtain the fully protonated enolic form. The four weakest
HBA solvents in the series (chloroform, toluene, chloroben-
zene, and o-dichlorobenzene) were not included in the
determination of the least-squares trend line shown in
Figure 8. In all of these the BDT tautomer ΔG differences
are approximately the same, and the observed ratios fall
somewhat below this line by up to 5 kJ mol-1, indicating the
enol-tautomer 6b is more stabilized in these solvents than
expected. In the absence of external H-bond stabilization of
the enol by solvent, the enol could adopt a conformation
where the enolic OH group now becomes syn to the adjacent
lactone carbonyl and in effect is stabilized by an intramole-
cular five-membered ring “H-bond” (as depicted in confor-
mer 6e)36 or favorable dipole-dipole interaction. An alter-
native source of stabilization could be intermolecular solute
OH to solute lactone carbonyl H-bonding, as observed in the
solid phase (Figure 4b). The highest observed stabilization is
about -13 kJ mol-1, for 2-propanol acting as a HBA base,
which is within the range typical for a normal H-bond
energy.37 The H-bond donor solvents (H-bond acids)
2-propanol and acetic acid do not deviate from the trend
shown in Figure 8. Any interaction they have with BDT
based on their H-bond donor ability is clearly having no
observable effect on the keto-enol equilibrium.
time gave a straight line showing gradient = k1 þ k-1
=
3.2 ꢀ 10-4 s-1, with r2 = 0.999.32 From this, k1 and k-1 for the
keto /enol process were calculated to be k1 =1.7ꢀ 10-4 s-1 and
k-1 = 1.5 ꢀ 10-4 s-1. Starting from a weakly acidified,
concentrated ethyl acetate solution (87% enol) and diluting to
an identical final solvent mixture of 1% EtOAc/99% o-dichloro-
benzene (final concentration, 2.3 ꢀ 10-5 mol L-1), the final enol
content was 51% (K = 1.06) at equilibrium, reached after
115 min. A plot of -ln([enol]t - [enol]eqm) versus time again
gave a straight line with gradient = k1 þ k-1 = 4.1 ꢀ 10-4 s-1
,
with r2 = 0.981. In this case, k1 and k-1 for the keto / enol
process were calculated to be k1 = 2.1 ꢀ 10-4 s-1 and k-1
=
2.0 ꢀ 10-4 s-1. Results for the final equilibrium constant and
forward and reverse rate constants were comparable when
approaching the equilibrium from either high or low keto
concentration. Clearly these kinetic data are specific to this
studied solvent system and the absolute values would be different
for others, but nonetheless the observations demonstrate that the
keto / enol equilibrium is fully dynamic, obeying first-order
equilibrium kinetics and show that the enol is thermodynamically
stable, since its concentration increases on addition of ethyl
acetate to the o-dichlorobenzene solution, which favors this
tautomeric form.
BDT in Solution; Keto-Enol Ratio and Empirical Solvent
Properties. Both the NMR and UV-vis data suggest that
higher enol-quinonoid content is found in H-bond acceptor
(HBA) or basic solvents and predominantly keto-benzenoid
content in less or non-HBA solvents. This relationship was
explored further by relating the equilibrium constant K
defined as K = [enol]/[keto] = [6b]/[6a] to empirical mea-
sures of solvent properties. The Kamlet-Taft solvatochro-
mic scales of solvent dipolarity/polarizability (π*), H-bond
donor strength (R), and H-bond acceptor strength (β) have
been widely exploited to investigate many aspects of solvent-
based phenomena.33 In the current context we chose the βOH
values recommended in the Critical Compilation of Scales of
Structural Dependence of BDT Enolization. We believe
that BDT is the first species for which enolization across a
benzene ring has been demonstrated. Given that the aro-
matic resonance stabilization of benzene is considered to be
around 135 kJ mol-1 38
the immediate reaction to this
,
observation is that the stabilization energy associated with
the central benzene ring of keto-BDT 6a is being lost on
enolization and must in some way be compensated for by the
energetics of the enol-tautomer 6b. Thus, the question of
interest is, where is this stabilization coming from to make
BDT such a unique species? Various structural features have
been considered and explored further with the aid of density
(32) Connors, K. A. In Chemical Kinetics, the Study of Reaction Rates in
Solution, Wiley-VCH: Weinheim, 1990.
(33) Abboud, J.-L. M.; Kamlet, M. J.; Taft, R. W. In Progress in Physical
Organic Chemistry; Taft, R. W., Ed.; Interscience-Wiley: New York, 1981;
Vol. 13, pp 485-630.
(34) Abboud, J.-L. M.; Notario, R. Pure Appl. Chem. 1999, 71, 645–718.
(35) (a) Marcus, Y. Chem. Soc. Rev. 1993, 22, 409–416. (b) Abraham, M.
H.; Buist, G. J.; Grellier, P. L.; McGill, R. A.; Prior, D. V.; Oliver, S.; Turner,
E.; Morris, J. J.; Taylor, P. J.; Nicolet, P.; Maria, P.-C.; Gal, J.-F.; Abboud,
J.-L. M.; Doherty, R. M.; Kamlet, M. J.; Shuely, W. J.; Taft, R. W. J. Phys.
Org. Chem. 1989, 2, 540–552.
(36) Bouchoux, G.; Hoppilliard, Y.; Houriet, R. Nouv. J. Chim. 1987, 11,
225–233.
(37) Reichardt, C. In Solvents and Solvent Effects in Organic Chemistry,
3rd ed.; Wiley-VCH: Weinheim, 2003; p 16.
€
(38) (a) Schleyer, P. v. R.; Puhlhofer, F. Org. Lett. 2002, 4, 2873–2876. (b)
Howard, S. T. Phys. Chem. Chem. Phys. 2003, 5, 3113–3119. (c) For a
discussion of the energetic aspects of cyclic π-electron delocalization, see:
~
Cyranski, M. K. Chem. Rev. 2005, 105, 3773–3811.
696 J. Org. Chem. Vol. 75, No. 3, 2010

Lawrence et al.
JOCArticle
found for enol 6b. First, in the higher energy minimum, the
exocyclic enolic OH is oriented away from the adjacent
carbonyl group (6b-anti). It has geometry close to that
determined in the solid phase, except that the torsion about
the phenyl-lactone bond is now 31° rather than the 13° and
25° found for the two independent molecules in the asym-
metric unit in the crystal (Figure 4a). This value is, however,
close to that found in an X-ray structure determination
for a benzodifuranone dye.40 The gas-phase difference in
energy between 6a and 6b-anti at the 6-311G(2d,p) level is
14.7 kJ mol-1 in favor of the keto-tautomer (6a). Relative
Gibbs energies calculated at 298 K, including zero point
energies, solvent plus cavitation, dispersion, and repulsion
terms, lowers the ΔG(298) values to 1.6, 6.6, and 5.3 kJ mol-1
for acetonitrile, chlorobenzene and THF, respectively, still in
favor of the keto-tautomer. Looking at the individual terms
that make up the difference between the gas phase and PCM
results in acetonitrile, we found that the main contribution is
the preferred stabilization of 6b-anti by the reaction field
solvent model, amounting to 13.1 kJ mol-1 preferential
lowering of 6b-anti compared to 6a. The contributions
arising from cavitation, dispersion, and repulsion are very
similar in 6a and 6b-anti and so largely cancel out, as do the
thermal corrections that convert ΔE to ΔG(298). Second, a
slightly lower enol energy minimum was found which corre-
sponds to a geometry in which the hydroxyl O-H bond is
oriented syn to the adjacent carbonyl (denoted 6b-syn). Re-
peating the comparison for 6b-syn (see Tables 4 and 5), we find
that the quinoidal enol form is now calculated to be the
preferred form in the gas and solvent phases. Furthermore,
the calculated preference for syn geometry in the gas phase
lends support to the conclusion, above, that in non-H-bond
donor solvents the enolic tautomer is energetically more
favored than the extrapolation from H-bonding solvents
would suggest, due to a stabilizing intramolecular H-bond.
Overall the calculated energetics of the tautomers of BDT do
not differ substantially from those deduced from the solvent
studies discussed above. The keto and enol forms are of similar
energy in the absence of stronger H-bond donor solvents.
The effect of delocalization of the conjugated extended
quinone electronic system into the pendant phenyl ring was
determined to be 15.9 kJ mol-1 (6-311G(2d,p)) by recalcula-
tion of enol 6b with enforced phenyl ring orthogonality.
There is no obvious phenyl interaction in the keto-tautomer
6a, where the energy-minimized geometry approaches a 90°
torsion about the phenyl-lactone bond, so this delocalization
must contribute to the overall enol stabilization.
FIGURE 8. Plot of Gibbs energy difference between enol 6b and
keto 6a tautomers of BDT against the solvent HBA parameter βOH
and two proposed solvent-dependent H-bonding systems, 6d and 6e.
functional calculations using the GAUSSIAN 03 suite of
programs.39 Gas phase geometries were optimized using a
range of basis sets and solvent effects were included within a
reaction field using the polarizable continuum model (PCM)
and a number of solvents. Solution-phase geometries were
also obtained using the PCM approach. The B3LYP ex-
change-correlation functional was used throughout.
We have discussed already that intermolecular H-bonding
is present with appropriate solvents and that the energetics
involved are sufficient to drive the equilibrium predomi-
nantly toward the enol-tautomer, but even in the absence of
H-bonding, the keto-tautomer is extrapolated from Figure 8
to be barely favored (ΔG ca. 5.9 kJ mol-1). The bulk of the
enol stabilization must, therefore, derive from elsewhere.
The energy-minimized gas-phase geometries of keto-
benzenoid-BDT 6a and enol-quinonoid-BDT 6b were calcu-
lated and are shown in Figure 9. Two energy minima were
The theoretical calculations were used to explore the
aromaticity within the systems of interest. In particular the
nucleus independent chemical shift (NICS) index of aroma-
ticity was calculated for each of the rings comprising both
tautomers 6a and 6b. The NICS index reflects the nature of
the ring current induced by an external magnetic field, where
a negative NICS value implies a diatropic ring current, and
aromaticity for that ring.41 We report here NICS(0) values,
obtained at the centroid of each ring. Since the four rings in
6a and 6b are not coplanar, it is not straightforward to define
(39) Gaussian 03, Revision C.02, Frisch, M. J.; Trucks, G. W.; Schlegel, H.
B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, Jr., J. A.;
Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.;
Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda,
R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.;
Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.;
Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski,
V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.;
Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith,
T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill,
P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.
Gaussian, Inc.: Wallingford, CT, 2004.
˚
NICS(1) centers (i.e., 1 A above the molecular plane).
(40) Malone, J. F; Bullock, J. F. Personal communication.
(41) Chen, Z.; Wannere, C. S.; Corminboeuf, C.; Puchta, R.; Schleyer, P.
v. R. Chem. Rev. 2005, 105, 3842–3888.
J. Org. Chem. Vol. 75, No. 3, 2010 697

Lawrence et al.
JOCArticle
TABLE 4. Relative Energies (kJ mol-1) in the Gas Phase and PCM
Solvent Model with Different Basis Sets Using the B3LYP Functional
E(6b-anti) - E(6a)
E(6b-syn) - E(6a)
phase/basis set 6-31G(d,p) 6-311G(2d,p) 6-311Gþ(2d,p) 6-311G(2d,p)
gas
20.9
7.3
12.2
10.8
14.7
1.6
6.6
5.3
14.4
-4.9
-3.5
-2.7
-2.9
acetonitrile
chlorobenzene
THF
TABLE 5. Relative Free Energies at 298 K (kJ mol-1) Including Zero
Point Energies in the Gas Phase and PCM Solvent Model with Different
Basis Sets Using the B3LYP Functional
G(6b-anti) - G(6a)
G(6b-syn) - G(6a)
phase/basis set
6-31G(d,p)
6-311G(2d,p)
6-311G(2d,p)
gas
27.5
3.3
11.7
10.5
acetonitrile
chlorobenzene
THF
5.0
11.3
8.2
-6.4
-5.0
-5.4
TABLE 6. NICS Valuesa for the Four Rings within Keto-BDT 6a and
Enol-BDT 6b
FIGURE 9. Calculated energy-minimized [6-311G(2d,p)] gas-
phase geometries of keto-BDT 6a, enol-BDT 6b-anti, and enol-
BDT 6b-syn.
ring
Furthermore, we are concerned here with the change in the
NICS values rather than the absolute values themselves, and
so the NICS(0) values should be adequate.
A
B
C
D
keto-BDT (6a)
enol-BDT (6b)
þ 0.8
-1.3
-2.1
-7.5
-0.4
þ 7.1
þ 7.2
-2.1
-9.3
-8.3
-7.3
þ 1.0
difference on enolizationa
The results are shown in Table 6. As expected, the pendant
phenyl ring D (as defined in Table 6) retains its aromaticity
for both tautomers. The central six-membered ring B is
aromatic in the keto-tautomer (NICS(0) = -7.5) but loses
this aromaticity on enolization (NICS(0) = -0.4). This
again is expected on the basis of the conventional structural
representations for the two tautomers and lies at the heart of
our interest in the energetic basis for the ease of enolization of
BDT. The phenyl-substituted lactone ring A has a NICS(0)
value of þ0.8 in the keto-tautomer and -1.3 in the enol,
implying at best a weak or negligible tendency to aromatiza-
tion on enolization. The result for ring C in BDT is more
surprising. In the keto-tautomer its NICS(0) value of þ7.2
indicates an unexpected and significant anti aromaticity.
This result is by no means intuitive. Moreover, the NICS(0)
value of -2.1 in the enol implies aromatization or, more
significantly, the removal of the antiaromatic character in 6a.
The tempting qualitative inference is that the loss of aroma-
ticity and corresponding decreased stabilization energy as-
sociated with the central ring B is more or less compensated
by removal of the antiaromatic destabilization of ring C
as it enolizes. To quantify this effect, we would like to be
able to relate the NICS changes to aromatic stabilization
energies (ASE). While a reasonable correlation between
NICS and ASE has been shown for a set of more than 100
heteroaromatic ring systems, we note the warning that
aNegative values correspond to aromaticity; positive values to
antiaromaticity. bNegative differences correspond to stabilization on
enolization.41
“energetic... and magnetic descriptors of aromaticity... do
not speak with the same voice”42 and in the absence of stronger
justification, regrettably conclude that translation of the NICS
values into ASE for BDT is currently unwarranted.43 However,
we do feel that removal of the antiaromatic destabilization of
ring C is a surprising, feasible, and at least contributory
explanation for the relative ease of phenylogous enolization
for this molecule.
Literature precedent for antiaromaticity in R-dicarbonyl
heterocycle systems analogous to BDT appears to be limited
to a single study of the nitrogen analogue isatin 11.44
Thermochemical considerations allowed construction of
the isodesmic reaction connecting isatin with o-xylene as in
Figure 10. This reaction was estimated to be exothermic by
38 kJ mol-1, leading to the conclusion that isatin is appre-
ciably antiaromatic. Since the heterocycle in isatin is isoelec-
tronic with the keto-lactone ring C of keto-BDT, it seems fair
to conclude that the antiaromaticity of this ring already
(43) More recent studies continue to develop links between energetic and
magnetic criteria of aromaticity, as for example circuit resonance energies of
polycyclic aromatic hydrocarbons. Aihara, J.-i. J. Am. Chem. Soc. 2006, 128,
2873–2879.
(44) (a) Matos, M. A. R.; Miranda, M. S.; Morais, V. M. F.; Liebman, J.
F. Org. Biomol. Chem. 2003, 1, 2566–2571. (b) Matos, M. A. R.; Liebman, J.
F. Top. Heterocycl. Chem. 2009, 19, 1–26.
~
(42) Cyranski, M. K.; Krygowski, T. M.; Katritzky, A. R.; Schleyer, P. v.
R. J. Org. Chem. 2002, 67, 1333–1338.
698 J. Org. Chem. Vol. 75, No. 3, 2010

Lawrence et al.
JOCArticle
TABLE 7. Comparion of 13C NMR Shifts (DMSO-d6 þ Et3N) for
BDT Enolate 6c and Enol-Ether 7
FIGURE 10. Isodesmic reaction connecting isatin (11) with
o-xylene.44
carbon
δ
(ppm) enolate 6c
δ(ppm) enol-ether 7
δ
(ppm) difference
a
b
c
d
e
124.7
128.3
125.8
134.4
97.4
169.4
144.1
142.6
92.0
129.6
129.0
128.6
129.4
120.1
168.0
138.7
153.5
98.0
-4.9
-1.6
-2.8
þ5.0
-22.7
þ1.4
þ5.4
-10.9
-6.0
þ0.5
þ16.5
-9.4
þ14.8
þ1.2
FIGURE 11. Reported equilibrium constitution of 1,2-cyclopenta-
nedione 12 in aprotic and aqueous/alcoholic media.45,47
f
g
h
i
j
k
l
indicated by the NICS calculations should also be reflected
by an appreciable thermochemical aromatic destabilization
energy. Again, we demur from giving it a value.
98.5
98.0
166.3
107.2
159.9
163.5
149.8
116.6
145.1
162.2
As well as consideration of the antiaromatic destabilization
in R-keto-lactone ring C, small (less than seven-membered)
alicyclic 1,2-dicarbonyl systems, for example, 1,2-cyclopenta-
nedione45 12a (Figure 11) and dihydrofuran-2,3-dione,46 are
known to exist exclusively as their enol-tautomers, i.e., 12b, in
aprotic, non-H-bonding solvents. In such rings, the locked
s-cis conformation of the carbonyl groups enforces an un-
favorable dipole-dipole interaction, which destabilizes the
diketo-tautomer12a.3f Additionally, the enol-tautomercanbe
further stabilized by an intramolecular hydrogen bond,36
a phenomenon that may occur for BDT in weak HBA
solvents, as discussed previously. However, in aqueous or
alcoholic media, the enol content of 1,2-cyclopentanedione is
low, resulting from preferable formation of the ketone-
hydrate/hemiacetal 12c.47 Interestingly, in nonscrupulously
dried water-miscible solvents, BDT still equilibrates to the
enol-tautomer 6b at the expense of aromatic stabilization
instead of forming an alternative benzenoid-ketone-hydrate,
which would presumably also remove antiaromaticity. How-
ever, in BDT, the resultant enolic-OH is delocalized into an
extended π-system, whereas in 1,2-cyclopentanedione 12b
only resonance into a single double bond exists.
In summary, there are several factors that contribute to the
energetics of the enolization of BDT, not all of which can
yet be quantified. The main drivers are (a) destabilizing
antiaromaticity of the keto-lactone ring of 6a, removed in
enol-tautomer 6b; (b) further stabilizing π-electron delocali-
zation in the extended conjugated ring system of 6b; (c)
relaxation of the destabilizing syn-1,2-dicarbonyl dipole-
dipoleinteractionin the keto-lactone ring;and (d) the pendant
phenyl ring of BDT, which contributes stabilization to the
enol-tautomer 6b by π-overlap effects, judging by the detri-
mental effect of removing its coplanarity. In addition to these,
H-bond acceptor solvents give intermolecular stabilization of
the enol 6b through H-bonding to the enolic OH group, and
m
n
result in keto-tautomer 6a giving way to almost total enol 6b
content in a strong H-bond acceptor like 2-propanol.
BDT Enolate. In strongly H-bonding and high relative
permittivity solvents with β values >0.6 such as DMF,
DMA, DMSO etc., complete ionization of BDT occurred
resulting in liberation of the intensely purple-colored free
enolate anion 6c with an intense broad absorption at λmax
550-555 nm with a bathochromicshoulderand εmax of 35,000
(Figure 7). No absorption was seen for the enol-tautomer at
ca. 440 nm in these solvents, even on addition of MSA. The
ionization is presumably stabilized by a combination of
favorable proton solvation by the basic solvent and dielectric
stabilization of the ions due to the high relative permittivities
of these solvents. Tertiary amines like triethylamine and
pyridine also ionized BDT, either as neat solution or on their
addition to BDT in other solvents, presumably because of
their strong basicity and proton solvation. The addition of
triethylamine to a DMSO or DMF solution gave no change in
the observed λmax or εmax, indicating further that BDT already
existed fully ionized in these solvents.
The nature of the enolate was studied by NMR spectro-
scopy. BDT enolate 6c in deuterated DMSO (containing a
small amount of triethylamine to further sharpen the
spectrum) showed a single entity strongly resembling the
enol-tautomer. The 1H spectrum displayed the characteristic
high field central ring proton resonances at δ 6.80 and δ
6.82 ppm and low field resonances for the 7-phenyl ortho-
protons at δ 7.83 ppm. The 13C NMR spectrum of 6c was
fully assigned with the assistance of HMQC and HMBC
analysis. When compared to 13C data for the enol-ether 7
(Table 7), this gave a qualitative insight into where negative
charge was localized on deprotonation, from the difference
in chemical shifts. The comparison is made here to the enol-
ether 7 rather than directly to the enol 6b, since the 13C NMR
spectrum of the enol-ether was fully assigned by HMQC/
HMBC experiments, and that of the enol 6b was itself made
by comparison to the enol-ether 7. Of particular interest are
carbons Ce and Cm (as defined in Table 7). The downfield
(45) Cumper, C. W. N.; Leton, G. B.; Vogel, A. I. J. Chem. Soc. 1965,
2067–2072.
(46) (a) Wasserman, H. H.; Ives, J. L. J. Org. Chem. 1978, 43, 3238–3240.
(b) Schank, K.; Beck, H.; Pistorius, S. Helv. Chim. Acta 2004, 87, 2025–2049.
Blank, I.; Lin, J.; Fumeaux, R.; Welti, D. H.; Fay, L. B. J. Agric. Food Chem.
1996, 44, 1851–1856.
(47) Bakule, R.; Long, F. A. J. Am. Chem. Soc. 1963, 85, 2309–2312.
J. Org. Chem. Vol. 75, No. 3, 2010 699

Lawrence et al.
JOCArticle
chemical shift for Cm (δ 159.9 ppm) relative to the enol-ether
(δ 145.1 ppm) indicates a higher degree of ketone character at
this position in the enolate, tending toward the shift seen in
the fully keto tautomer 6a (δ 178.8 ppm). However, despite
the increasing ketone character at Cm, the shift of the lactone
carbonyl Cn is largely unaffected, suggesting that the C ring
in the enolate has none of the antiaromatic character seen in
the keto-tautomer. Carbon Ce is significantly more shielded
in the enolate (δ 97.4 ppm) compared to the enol-ether
(δ 120.1 ppm). Similarly, smaller increases in electron density
are identified at Ci and weakly in the exo-phenyl ring at
Ca and Cc, which are intuitive from consideration of stan-
dard resonance canonical forms. Carbons Cg and Ck in the
central ring also show large differences compared to the enol,
tending closer to the shifts seen for these carbons in the keto-
standard 8 (δ 138.9, δ 161.7 ppm, respectively), suggesting
greater aromatic character in the central ring on deprotona-
tion. Lactone carbon Cf shows little change in the enolate,
indicating little delocalization of charge into this position.
Comparison of data, therefore, indicates that the nega-
tive charge in the enolate is not fully localized on oxygen,
but that the enolate exists part-way between an enol-like
and keto-like structure, with the bulk of the negative
charge split between the keto/enol-oxygen and the benzylic
carbon Ce.
J = 7.0), 7.37 (2H, m), 7.82 (2H, dd, J = 1.2, J = 8.3); δC
(100 MHz, DMSO-d6 þ Et3N; exists as enolate 6c) 8.9, 45.8, 92.0,
97.4, 98.5, 107.2, 124.7, 125.8, 128.3, 134.4, 142.6, 144.1, 159.9,
163.5, 166.3, 169.4. m/z (CIþ) 284 (100%), 298 (25%, M þ
NH4þ). Anal. Calcd for C16H8O5: C, 68.58; H, 2.88. Found: C,
68.64; H, 2.80.
3-Methoxy-7-phenyl-7H-benzo[1,2-b;4,5-b0]difuran-2,6-dione
(7). 7-Phenyl-7H-benzo[1,2-b;4,5-b0]difuran-2,3,6-trione (6a)
(2.8 g, 10 mmol) was suspended in toluene (100 mL). At rt a
solution of trimethylsilyl-diazomethane (2 M, 6 mL, 12 mmol)
was added dropwise over ca. 1 min. Gas evolution was seen
immediately, and the suspended solid turned from a dark purple
to a yellow-brown color. The suspension was stirred overnight.
The solid was then filtered off, washed with further toluene, and
dried overnight in a fan oven at 40 °C (mass = 2.25 g). The
material was recrystallized from a large volume of boiling methyl
isobutyl ketone, toobtainanalytically pure material(1.78 g, 60%)
as fine orange-brown needles: UV-vis see Table 1; FTIR (KBr
cm-1) νmax 1785 (lactone CdO), 1753(lactone CdO), 1620, 1583,
1369; δH (400 MHz, DMSO-d6) 4.31 (3H, s), 7.13 (1H, s),
7.29 (1H, s), 7.45-7.55 (3H, m), 7.79 (2H, m); δC (100 MHz,
DMSO-d6) 60.8, 97.97, 98.00, 116.6, 120.1, 128.6, 129.0, 129.4,
129.6, 138.7, 145.1, 149.8, 153.5, 162.2, 168.0. m/z (APCIþ) 295
(100%, M þ Hþ). Anal. Calcd for C17H10O5: C, 69.39; H, 3.43.
Found: C, 69.45; H, 3.28.
Methyl 2-(5-Hydroxy-2-oxo-3-phenyl-2,3-dihydrobenzofuran-
6-yl)-2-oxoacetate (8). 7-Phenyl-7H-benzo[1,2-b;4,5-b0]difuran-
2,3,6-trione (6a) (1.5 g, 5.34 mmol) was suspended in methanol
(60 mL), heated to boiling to effect a fine suspension, and then
allowed to stir overnight at ambient temperature. Residual trace
suspended solid was filtered off, and the solvent was evaporated
under reduced pressure. A small amount (ca. 5 mL) of methanol
was added to redissolve the resulting purple oily solid, and the
solution was allowed to stand overnight. The required product
was filtered off, washed with a small portion of ice-cold metha-
nol (20 mL), pulled dry under vacuum, and dried in a fan oven at
40 °C overnight to give 8 (0.74 g, 45%) as a yellow solid: UV-vis
(CH2Cl2 þ trace CH3SO3H) λmax 368 nm, εmax 4,250; FTIR
(KBr cm-1) νmax 1791 (lactone), 1741 (ester), 1624 (H-bonded
ketone); FTIR (CH2Cl2 cm-1) νmax 1819, 1744, 1644 and 1583;
Conclusions
The first example of a stable phenylogous enol has been
isolated, and its structure was proven by X-ray crystallogra-
phy. The equilibrium between the keto- and enol-tautomers
was quantified in solution, and the position of equilibrium
showed a linear correlation to the Kamlet-Taft solvato-
chromic scale for solvent H-bond acceptor strength (βΟH).
We attribute the surprising apparent net stabilizing effect on
enolization despite the loss of the aromatic resonance stabi-
lization of the central benzene ring to several energetically
favorable outcomes including a loss of destabilizing antiar-
omaticity of the keto-lactone ring of 6a that is not present in
enol-tautomer 6b, increased π-electron delocalization in the
enol extended conjugated ring system, removal of the desta-
bilizing syn-1,2-dicarbonyl dipole-dipole interaction pre-
sent in the keto-lactone ring of 6a, and π-overlap of the
pendant phenyl ring in the enol-tautomer 6b. In solution,
H-bond acceptor solvents gave further intermolecular stabi-
lization of the enol 6b through H-bonding to the enolic OH
group, leading to measurable differences in the observed
equilibrium position. In strongly H-bonding and high rela-
tive permittivity solvents with β values >0.6 BDT existed
exclusively as the free enolate anion 6c, part-way between an
enol-like and keto-like structure.
δ
H (400 MHz, CDCl3) 4.03 (3H, s), 4.95 (1H, d, J = 0.7), 6.91
(1H, d, J = 1.0), 7.20 (2H, m), 7.33-7.44 (3H, m), 7.57 (1H, s),
11.35 (1H, s). δC (100 MHz, CDCl3) 50.1, 53.3, 111.8, 115.2,
115.8, 128.1, 128.6, 129.3, 133.5, 138.9, 146.1, 161.7, 162.1,
173.5, 189.1. m/z (CIþ) 136 (100%), 262 (80%), 330 (15%, M
þ NH4þ). Anal. Calcd for C17H12O6: C, 65.39; H, 3.87. Found:
C, 65.41; H, 3.77.
2-(5-Hydroxy-2-oxo-3-phenyl-2,3-dihydrobenzofuran-6-yl)-2-
oxoacetic Acid (9). 7-Phenyl-7H-benzo[1,2-b;4,5-b0]difuran-
2,3,6-trione (6a) (2.7 g, 9.64 mmol) was suspended in 0.1 N
NaOH(aq) (200 mL) and stirred at ambient temperature for
30 min, after which time the solid had dissolved. The solution
was screened through a glass fiber filter to remove residual traces
of insoluble materials and then adjusted to pH 2 by dropwise
addition of 2 N HCl(aq). Sodium chloride (30 g) was added and
stirred for 2 h to aid precipitation of the required product. The
product was filtered off, washed with a small volume of ice-cold
water, and then dried in a fan oven at 40 °C overnight to yield a
pale brown solid (2.1 g, 73%), which was >99% organically
pure by 1H NMR. To obtain analytically pure material, a
portion of the obtained solid was recrystallized from boiling
methylated spirit: UV-vis (acetone þ trace CH3SO3H) λmax
357 nm, εmax 4,050; FTIR (CH2Cl2 cm-1) νmax 1817 (lactone),
1779, 1758, 1649, 1621, 1580; δH (400 MHz, CDCl3) 4.95 (1H,
br. s), 6.96 (1H, d, J = 1.0), 7.22 (2H, m), 7.37-7.44 (3H, m),
8.25 (1H, s), 11.39 (1H, br. s). δC (100 MHz, CDCl3) 50.4, 112.5,
115.2, 115.9, 128.2, 128.8, 129.5, 133.4, 139.7, 146.3, 161.5,
162.5, 173.9, 187.4. m/z (ES-) 297 (100%, [M - Hþ]-). Anal.
Experimental Section
Exemplary 1H and 13C data are supplied here. The behavior
of all compounds was studied in numerous solvents. Further
details and spectral assignments are available in Supporting
Information.
7-Phenyl-7H-benzo[1,2-b;4,5-b0]difuran-2,3,6-trione (BDT; 6a).
Prepared according to ref 19. UV-vis see Table 3; FTIR (KBr
cm-1; exists as H-bonded enol 6b) νmax 1782 (lactone CdO), 1692
(H-bonded lactone CdO), 1623, 1579; δH (400 MHz, DMSO-d6;
exists as enolate 6c) 6.80 (1H, s), 6.82 (1H, s), 7.15 (1H, tt, J = 1.2,
700 J. Org. Chem. Vol. 75, No. 3, 2010

Lawrence et al.
JOCArticle
Supporting Information Available: General experimental
Calcd for C16H10O6 0.45 NaCl: C, 59.21; H, 3.11. Found: C,
59.21; H, 2.98.
3
1
information. Assigned H, 13C, HMQC, and HMBC spectra
for all compounds measured in a range of solvents. X-ray
crystallographic data for 6b in CIF format. Details on solu-
tion-phase equilibrium dynamics measurements for 6a/6b.
Further discussion on the UV-vis spectrum of keto-BDT 6a.
Confirmation of a linear relationship between A and concentra-
tion for enol-BDT in UV-vis measurements. Gas-phase
Cartesian coordinates for 6a/6b from energy minima calcula-
tions. This material is available free of charge via the Internet at
http://pubs.acs.org.
Acknowledgment. We are grateful to the following people
who are or were based at The University of Manchester,
School of Chemistry: M. Chem. students Khalid Ali, Simon
McCormack, and Collette Turner for their experimental
input; Drs. Ian Watt and Peter Quayle for invaluable
discussion; and staff in the NMR department, especially
Mr Stephen Kelly, for their assistance in obtaining data.
J. Org. Chem. Vol. 75, No. 3, 2010 701