On five- vs six-membered diacetal formation from threitol and the intermediacy of unusually stable protonated species

Article abstract of DOI:10.1021/jo9908952

The long known, but hitherto poorly understood, thermodynamically controlled diacetalation of rac-threitol with alkylaldehydes provided bicyclic, cis-tetraoxadecalin (TOD) ('66') and bi(dioxolanyl) (BDO) ('55') products, shown to be formed in acid-concentration and temperature-dependent ratio. The configurational and conformational isomeric diacetals obtained in four such reactions of substituted aldehydes (RCHO, R = CH₃, CH₂Cl, CH₂Br, CO₂CH₃) with rac-threitol were isolated and characterized. A variable acid- concentration analysis of the equilibrium mixture of products in one such case (R = CH₂Br) was performed and provided equilibrium constants and, hence, free-energy differences among these products and their relatively stable protonated intermediates. The latter were rationalized by the unusually high proton-affinity calculated for the cis-TOD ('66') form.

Full text of DOI:10.1021/jo9908952

1636
J . Org. Chem. 2000, 65, 1636-1642
On F ive- vs Six-m em ber ed Dia ceta l F or m a tion fr om Th r eitol a n d
th e In ter m ed ia cy of Un u su a lly Sta ble P r oton a ted Sp ecies1
Mikhail Grabarnik, N. Gabriel Lemcoff, Ravit Madar, Sarah Abramson, Sarah Weinman, and
Benzion Fuchs*
School of Chemistry (Raymond and Beverly Sackler Faculty of Exact Sciences), Tel-Aviv University,
Ramat-Aviv, 69978 Tel-Aviv, Israel
Received J une 2, 1999
The long known, but hitherto poorly understood, thermodynamically controlled diacetalation of
rac-threitol with alkylaldehydes provided bicyclic, cis-tetraoxadecalin (TOD) (“66”) and bi(dioxolanyl)
(BDO) (“55”) products, shown to be formed in acid-concentration and temperature-dependent ratio.
The configurational and conformational isomeric diacetals obtained in four such reactions of
substituted aldehydes (RCHO, R ) CH3, CH2Cl, CH2Br, CO2CH3) with rac-threitol were isolated
and characterized. A variable acid-concentration analysis of the equilibrium mixture of products
in one such case (R ) CH2Br) was performed and provided equilibrium constants and, hence, free-
energy differences among these products and their relatively stable protonated intermediates. The
latter were rationalized by the unusually high proton-affinity calculated for the cis-TOD (“66”)
form.
Sch em e 1. Ster eosp ecific Rea ction P r od u cts of
Er yth r itol a n d r a c-Th r eitol w ith F or m a ld eh yd e
In tr od u ction
The condensation of a 1,2,3,4-tetrahydroxybutane with
formaldehyde under acid catalysis can take place in
1,2;3,4-, 1,3;2,4-, or 1,4;2,3-fashion to give bicyclic diac-
etals “55”, “66”, or “57”, respectively, which are formed
in a stereospecific manner (Scheme 1). The most signifi-
cant and ubiquitous ones are the “66” type compounds,
namely, the trans- and cis-1,3,5,7-tetraoxadecalin (TOD)6
system (Scheme 2), formed from erythritol or threitol,
respectively.2-6 The conformationally stable form of the
trans isomer is a configurationally fixed double-chair,
while the cis isomer can exist in two possible diastere-
oisomeric chair-chair forms, Oinside (Oin ) and Ooutside
(Oou t). These can interconvert by conformational ring
inversion (Oin h Oou t) (Scheme 2), but bias can be
* To whom correspondence should be addressed. Fax: (+972-3) 640
9293; e-mail: bfuchs@post.tau.ac.il.
Sch em e 2. Dia ster eom er ic 1,3,5,7-tetr a oxa d eca lin
(TOD) System
(1) (a) New Supramolecular Host Systems. 12. Part 11: Star, A.;
Goldberg, I.; Lemcoff, N. G.; Fuchs, B. Eur. J . Org. Chem. 1999, 2033.
(b) We use consistently the 1,3,5,7-tetraoxadecalin nomenclature. Other
possible names are cis- or trans-2,4,7,9-tetraoxabicyclo[4.4.0]decane,
1,3:2,4-di-O-methylenethreitol, or (cf. CA) (4aR)-(4ar,8ac)-tetrahydro-
[1,3]dioxino[5,4-d]-1,3-dioxin. Also, due to a minor but basic omission
of the CIP rules, one can assign unequivocally configurations to chiral
cis-decalin systems, only by 9,10-helicity designation, e.g., molecule
3c is (2R,6R,9R;9,10-M)-2,6- bis(bromomethyl)-cis-1,3,5,7-tetraoxa-
decalin.
(2) (a) Lemieux, R. U.; Howard, J . Can. J . Chem. 1963, 41, 393. (b)
Nouguier, R.; Gras, J .-L.; Mchich, M. Tetrahedron 1988, 44, 2943. (c)
Gras, J .-L.; Poncet, A. Synth. Commun. 1992, 22, 405.
(3) (a) J ensen, R. B.; Buchardt, O.; J ørgensen, S. E.; Nielsen, J . U.
R.; Schroll, G.; Altona, C. Acta Chem. Scand. B 1975 29B, 373. (b)
Nørskov, L.; J ensen, R. B.; Schroll, G. Acta Chem. Scand. B 1983 37B,
133 and previous parts in this series.
introduced by substitution in the 2,6 positions. Thus, two
2,6-diequatorially substituted Oin and Oou t may undergo
conformational interconversion only by chemical (acid-
catalyzed) isomerization. It was shown that Oin is
energetically preferred in the parent molecule and in
simply substituted derivatives,2,6 but purposeful substitu-
tion in the 4,8 or 9,10 positions may alter this order of
stability.3-5
(4) (a) Burden, I. J .; Stoddart, J . F.; J Chem. Soc., Perkin Trans. 1
1975, 666. (b) Idem. Ibid. 1975, 675.
These systems have been studied most in the carbo-
hydrate field,4,6e,7 where they occur most often. A number
of significant contributions have also been made toward
the understanding of the stereochemical and conforma-
tional features of these, mainly “66” diacetals, substituted
in the 4(8) or 9(10) positions.3-6 However, while the
erythritol diacetal formation and isomerism had been
well investigated,2,4,7 in particular by Burden and Stod-
dart,4 there was no detailed documentation on the
isomerism of threitol diacetals. Generally, the reaction
(5) Santos, A. G.; Hoffman, R. W. Tetrahedron: Asymmetry 1995,
6, 2767.
(6) (a) Senderowitz, H.; Linden, A.; Golender, L.; Abramson, S.;
Fuchs, B. Tetrahedron 1994, 50, 9691. (b) Senderowitz, H.; Golender,
L.; Fuchs, B. Tetrahedron 1994, 50, 9707. (c) Abramson, S.; Ashkenazi,
E.; Goldberg, I.; Greenwald, M.; J atzke, H.; Vardi, M.; Weinman, S.;
Fuchs, B. J . Chem. Soc., Chem. Commun. 1994, 1611. (d) Frische, K.;
Greenwald, M.; Ashkenazi, E.; Lemcoff, N. G.; Abramson, S.; Golender,
L.; Fuchs, B. Tetrahedron Lett. 1995, 36, 9193. (e) J atzke, H.; Frische,
K.; Greenwald, M.; Golender, L.; Fuchs, B. Tetrahedron 1997, 53, 4821.
(f) Grabarnik, M.; Goldberg, I.; Fuchs, B. J . Chem. Soc., Perkin Trans.
1
1997, 3123. (g) Linden, A.; Beckhaus, H.-D.; Verevkin, S. P.;
Ru¨chardt, Chr.; Ganguly, B.; Fuchs, B. J . Org. Chem. 1998, 63, 8205.
10.1021/jo9908952 CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/01/2000

Five- vs Six-Membered Diacetal Formation
J . Org. Chem., Vol. 65, No. 6, 2000 1637
Sch em e 3. Aceta la tion Rea ction P r od u cts of
D-Th r eitol w ith Su bstitu ted Aceta ld eh yd es
H2(6) vs H4ax (8ax) and H9 (H10) supports the described
structure in these molecules.
Contrary to 1, the H NMR spectra of the new 2(eq),
1
6(ax)-disubstituted-TOD compounds (2) exhibit two sets
of signals. The first set, consisting of H2, H4ax, H4eq
and H10, is similar (in both chemical shifts and coupling
constants) to that observed in the 1H NMR spectra of the
symmetrical diequatorial system (1) in this work, as well
as in previous ones (vide supra). The second set, consist-
ing of H6, H8ax, H8eq, and H9, is different from the first
one and unprecedented in previously reported TOD
systems. Assignment of chemical shift was unequivocal
on the basis of NOE experiments and analysis of the
chemical shifts and coupling constants: irradiation of H2
enhanced the H4ax and H9 signals, whereas irradiation
of H6 enhanced H8â and that of H12 enhanced H8R. The
angular H9 proton is shifted downfield (as compared with
H9(10) in 1 and H(10) in 2) and coupled with H8R, J 9,8R
growing up to 5.9 Hz (as compared to 1.2 Hz in 1). H8R
was first concluded to be H8ax, and H8â is H8eq, but
the overall behavior was subsequently taken to stem from
a fluctuational behavior in the C6-O7-C8 moiety within
that ring.
The coupling constants in the second set, e.g., J 8R,9 5.8
Hz and J 8â,9 1.7 Hz, reflect the fact that the dihedral
angle H9-C9-C8-H8R becomes smaller (∼30°) and
H9-C9-C8-H8â wider (∼100°) as compared with the
diequatorial system (1). These changes in coupling
constants fit a twist-boat conformation of the second ring
in 2, which, following the above-described NOE results,
exists within a chair/chair-chair/twist-boat conforma-
tional equilibrium 2 h 7 (Scheme 4).
These structures are supported by 13C NMR spectra,
the details of which are presented in the Table 2.
Assignment of 13C signals was carried out using DEPT
for all compounds and C-H correlation for 2b.
Our conclusions about the structures of and the con-
formational equilibrium between 2 and 7 are well sus-
tained by molecular mechanics calculations of 1-3a (X
) H), carried out using the MM3 force field,8 which had
been parametrized for the anomeric effect in O-C-O
systems and which we have reparametrized (MM3-GE)6b
for the gauche effect in O-C-C-O containing systems.
This modified MM3(92)8c force field has been successfully
used in the meantime on several similar systems.5,6c-g
As evident, all our isomers contain both these types of
dioxa (O-C-O and O-C-C-O) units and have to be
treated accordingly. Relative stabilities of the lowest
conformations of 1-3a are presented in Table 3.
These results lend strength to the conclusion that, at
the reaction temperature (80° C) used, the unsymmetrical
diastereomer 2a , inferred by the NMR spectrum in
solution, is actually an equilibrium mixture of two stable
conformers of similar energy, 2a and 7a (Scheme 4). The
calculated data are in good agreement with the energy
differences calculated for the naked cis-TOD system,9 in
which the chair/twist-boat form is about 5 kcal/mol higher
with substituted aldehydes had been commonly taken to
be selective, in yielding only 2,6-diequatorially substi-
tuted “66” acetals in their cis-Oin forms, based on the
behavior of aromatic aldehydes.
We have been dealing with such systems in connection
with our probes of new types of host systems based, inter
alia, on cis-1,3,5,7-TOD “core” units.6 The problem we
have encountered in the course of these studies was the
lack of information concerning the preparative details of
this, apparently simple, acetalation reaction in particular
with aliphatic aldehydes and the considerable confusion
surrounding the conditions for obtaining exclusively or
selectively any of the bicyclic isomeric products. In this
paper we deal with this problem, analyzing the forma-
tion, structure, conformation, and thermodynamic pa-
rameters of the isomeric 2,6- disubstituted diacetals of
threitol.
Resu lts a n d Discu ssion
Condensation of rac-threitol with substituted acetal-
dehydes (XCH2CHO, X ) H, Cl, Br) (Scheme 3), provided
the 2,6-disubstituted-cis-TOD products, with the 2(eq),
6(eq)-derivative (1) as the main ones, accompanied by
minute amounts of the new 2(eq),6(ax)-derivatives (2)
(Scheme 3). In addition to those, there were variable
amounts (around 30%) of five-membered ring products
(4-6). Careful workup made possible separation and
isolation of all these products and their reliable charac-
terization by NMR spectroscopy.
The NMR spectral data of the TOD (“66”) products are
presented in Tables 1 and 2. The chemical shifts and
coupling constants in the CH-CH2 group of the six-
membered ring in 1 (Table 1) are well defined (only
J 4(8)eq,10(9) is mostly undinstinct and seen as signal broad-
ening) and in excellent agreement with literature data
and our previous investigations.3-6 NOE experiments on
(8) (a) MM3 is available from QCPE (latest public version); the
official distributors are Technical Utilization Corporation, Inc., 235
Glen Village Court, Powell, OH 43065, and Tripos Associates, 1699 S.
Hanley Road, St. Louis, MO 63144. (b) Allinger, N. L.; Yuh, Y. H.; Lii,
J .-H.; J . Am. Chem. Soc. 1989, 111, 8551 and subsequent articles, in
particular that on alcohols and ethers: (c) Allinger, N. L.; Rahman,
M.; Lii, J .-H. J . Am. Chem. Soc. 1990, 112, 8293, subsequently updated
in MM3(92) for the anomeric effect by P. Aped. (d) MM3-GE6b is a
reparametrized version for the gauche effect in O-C-C-O-containing
systems.
(7) (a) Barker, S. A.; Bourne, E. J . Adv. Carbohydr. Chem. 1952, 7,
138. (b) Mills, A. Adv. Carbohydr. Chem. 1955, 10, 1. (c) Stoddart, J .
F. Stereochemistry of Carbohydrates; Wiley: New York, 1971; p 210.
(d) Foster, A. B. In The Carbohydrates; Pigman, W., Horton, D., Eds.;
Academic Press: New York, 1972; Vol. 1A, Ch. 11. (e) Cf. ref 6e for a
short but relevant review on carbohydrate TOD derivatives.

1638 J . Org. Chem., Vol. 65, No. 6, 2000
Grabarnik et al.
Ta ble 1. 1H NMR Sp ectr a l Da ta (CDCl3/TMS, 25 °C) of 2(eq),6(eq)-bis(XCH2)-cis-TOD (X ) H, Cl, Br ) (1a -c) a n d
2(eq),6(a x)-bis(XCH2)-cis-TOD (X ) H, Cl, Br ) (2a -c) (δ, p p m ; J ,a Hz)
1a
1b
1c
2a
2b
2c
H2
4.78 (q)
4.10 (d)
3.89 (d2)
3.61 (m)
1.41 (d)
4.80 (d2)
4.23 (d)
3.94 (d2)
3.70 (m)
3.60 (d)
4.83 (d2)
4.23 (d)
3.93 (d2)
3.67 (m)
3.44 (d)
4.74 (q)
4.09 (d2)
3.85 (d2)
3.92 (d3)
1.41 (d)
5.39 (q)
4.17 (d2)
3.85 (d2)
3.72 (m)
1.38 (d)
4.79 (d2)
4.79 (d2)
H4eq
H4ax
H9
4.21 (d2)
4.22 (d2)
3.90 (d2)
3.89 (d2)
4.12 (d3)
4.11 (d3)
H11
H6
H8ax
H8eq
H10
H12
3.58 (d)
3.42 (d)
5.30 (d2)
5.30 (d2)
4.26 (d2)
4.26 (d2)
3.93 (d2)
3.92 (dd)
3.79 (d3)
3.78 (d3)
3.63 (d2)(a)
3.56 (d2)(b)
J 2,11a ) 4.8
J 2,11b ) 4.8
J 4eq,4ax ) 12.9
J 4eq,10 ) 1.2
J 4ax,10 ) 1.7
J 9,8ax ) 5.8
J 9,8eq ) 1.7
J 9,10 ) 1.7
J 6,12a ) 5.2
J 6,12b ) 4.1
J 8ax,8eq ) 12.3
J 12a,12b ) 11.4
3.47 (d2)(a)
3.44 (d2)(b)
J 2,11a ) 4.8
J 2,11b ) 4.8
J 4eq,4ax ) 12.7
J 2,11 ) 5.1
J 2,11 ) 5.1
J 2,11a ) 5.0
J 2,11b ) 5.0
J 4eq,4ax ) 12.6
J
2,11 ) 5.1
J 4eq,4ax ) 12.5
J 4ax,10 ) 1.3
J 4eq,10 < 1
J 4eq,4ax ) 12.5
J 4ax,10 ) 1.3
J 4eq,10 < 1
J 4eq 4ax ) 13.0
,
J 4eq,10 ) 1.2
J 4ax,10 ) 1.0
J 9,8ax ) 4.9
J 9,8eq ) 1.5
J 9,10 ∼ 1.5
J 4ax
,
10 ) 1.2
J 4eq
,
,
10 ) 1.3
10 ) 1.6
J 4ax
J 9,8ax ) 5.9
J 9,8eq ) 1.7
J 9,10 ) 1.7
J 6,12 ) 5.3
J 8ax,8eq ) 12.2
J 6,12a ) 5.0
J 6,12b ) 4.0
J 8ax,8eq ) 12.3
J 12a,12b ) 11.2
a
Multiplicity: d ) doublet (d2 ) dd, d3 ) ddd), t ) triplet, q ) quadruplet, m ) multiplet.
Ta ble 2. 13C NMR Sp ectr a l Da ta (CDCl3/TMS, 25 °C) of
2,6-Disu bstitu ted -TOD (1a -c, 2a -c) (δ, p p m )
isolated from the reactions of threitol with the mentioned
substituted aldehydes, they could be well characterized
by their NMR spectra, the data of which are presented
in Tables 4 and 5. It merits recalling at this point that a
related issue, 1,3-dioxolane vs 1,3-dioxane formation in
acetalation of glycerol, has been receiving continuing
attention.10
1a
1b
1c
2a
2b
2c
C2
98.9
69.4
69.6
21.0
100.1
69.4
69.4
43.8
100.1
69.5
69.5
30.9
98.4
69.7
70.9
19.0
94.0
65.8
62.2
20.9
99.6
69.5
71.3
43.6
95.3
67.0
63.3
43.9
99.2
69.6
71.3
31.0
94.7
67.0
63.3
31.5
C4
C9
C11
C6
C8
The 1H NMR spectra (Table 4) of each and all “cis-
,trans” compounds (5) are practically superimposed spec-
tra of the symmetrical “cis,cis” (6) and “trans,trans” (4)
diastereomers, except the H4 and H4′ signals, which are
close but changed places. This was established by ir-
radiation of the low field “ddd”-signal. The configura-
tional assignments within the 4,4′-BDO systems were
supported by NOE experiments on compound 5c, namely,
irradiation of the low-field H2 signal, enhanced those of
H5cis and H4′, and that of the high-field H2′ signal,
enhanced H5′trans and H4′. These findings are in very
good agreement with the available10d,e detailed NMR data
of cis- and trans-2-methyl-4-XCH2-1,3-dioxolanes.
Notably, in all cases the isomers of 4, 5, and 6 occur in
statistic ratio, i.e., 1:2:1 and are not substituent depend-
ent, which means that no significant free energy differ-
ences of those isomers should be anticipated. This was
nicely confirmed by MM3-GE9 calculations, the results
of which are presented in Table 6.
Our interest in the intimate behavior of substituted
TOD systems and the peculiar phenomena we encoun-
tered in the first stages of this investigation compelled
us to probe the reaction mechanism and optimization
conditions of the described double acetalation process. We
chose the reaction of rac-threitol with bromoacetaldehyde
as a model for this purpose, 2,6-bis(bromomethyl)-TOD
(1c) having been designed as a useful starting material
in our past6f and present investigations.
C10
C12
Sch em e 4. Tw o Dou ble-Ch a ir (2) a n d Ch a ir /
Tw ist-Boa t (7) F or m s of
2S,6R-Disu bstitu ted -cis-TOD
Ta ble 3. Ca lcu la ted (MM3-GE) Rela tive F r ee En er gy
Differ en ces of 2,6- Dim eth yl-TOD (1-3a ) in Th eir Low est
Min im a Con for m a tion s (k ca l/m ol)
conformationa
∆G353 rel
2(eq),6(eq)-dimethyl-cis-TOD (Oinside) (1a )
2(eq),6-dimethyl-1,3-c-5,7-tb-cis-TOD (Oinside) (7a )
2(eq),6(ax)-dimethyl-cis-TOD (2a ) (Oinside)
2(eq),6(eq)-dimethyl-cis-TOD (Ooutside)
2(eq),6-dimethyl-1,3-c-(Ooutside)-5,7-tb-cis-TOD
2(ax),6(ax)-dimethyl-cis-TOD (Oinside) (3a )
2(ax),6(eq)-dimethyl-cis-TOD (Ooutside)
2(ax),6(ax)-dimethyl-cis-TOD (Ooutside)
0.0
4.4
5.1
5.4
7.7
11.1
13.2
22.1
a
c ) chair; tb ) twist-boat.
than the chair/chair, with a ca. 10 kcal/mol barrier
between them. Moreover, there is good agreement of the
calculated free energy difference between the diastere-
omers 1a and 2a /7a with the experimental ratio of these
diastereomeric products (ca. 1000:1). Interestingly, the
high-energy 2(ax),6(ax)-cis-TOD-Oinside diastereomer 3a
could conformationally invert into the 2(eq),6(eq)-cis-
TOD-Ooutside one, which is of similar free energy as
conformers 2a and 7a , but was not found among the
reaction products. We will address this issue later.
Turning to the diastereomeric 2,2′-disubstituted-4,4′-
bi(1,3-dioxolanyl) (BDO) (4-6) products (Scheme 3)
The yields of crude product in this reaction were, as a
rule, nearly quantitative, and of the products (1c, 2c, 4c,
5c, 6c), 1c could be crystallized from methanol in pure
form and the rest was resolved when needed, by column
(10) (a) Hibbert, H.; Carter, N. M. J . Am. Chem. Soc. 1928, 50, 3120.
(b) Hill, H. S.; Hill, H. C.; Hibbert, H. J . Am. Chem. Soc. 1928, 50,
2242. (c) Aksnes, G.; Albriktsen, P.; J uvvik, P. Acta Chem. Scand. 1965,
19, 920. (d) Borremans, F.; Anteunis, Anteunis-De Ketelaere, F. M.
Org. Magn. Res. 1973, 5, 299. (e) Gras, J .-L.; Poncet, A. Bull. Soc. Chim.
Fr. 1991, 128, 566.
(9) Golender, L.; Senderowitz, H.; Fuchs, B. J . Mol. Struct.
THEOCHEM 1996, 370, 221.

Five- vs Six-Membered Diacetal Formation
J . Org. Chem., Vol. 65, No. 6, 2000 1639
Ta ble 4. 1H NMR Sp ectr a l Da ta (CDCl3/TMS, 25 °C) of 2,2′-bis(XCH2)-4,4′-BDO (X ) H, Cl, Br ) (4a -c, 5a -c, 6a -c) (δ, p p m ;
J ,a Hz)b
4a
4b
4c
5a
5b
5c
6a
6b
6c
H2
H4
H5t
H5c
H6
5.16 (q)
4.10 (m)
4.10 (m)
3.74 (d2)
1.36 (d)
5.30 (d2)
4.27 (m)
4.21 (d2)
3.92 (d2)
3.54, 3.56
(d2,d2)
5.28 (d2)
4.26 (m)
4.22 (d2)
3.92 (d2)
3.44, 3.38
(d2,d2)
5.18 (q)
4.17 (d3)
4.11 (d2)
3.67 (d2)
1.36 (d)
5.33 (d2)
4.33 (d3)
4.21 (d2)
3.90 (d2)
3.54, 3.56
(d2,d2)
5.32 (d2)
4.35 (d3)
4.22 (d2)
3.91 (d3)
3.38, 3.42
(2d2)
5.03 (q)
4.19 (m)
3.90 (m)
3.90 (m)
1.40 (d)
5.17 (d2)
4.32 (m)
4.02 (d2)
3.97 (bd2)
3.59, 3.61
(d2,d2)
5.15 (d2)
4.34 (m)
4.03 (d2)
3.96 (d2)
3.43, 3.46
(d2,d2)
H2′
H4′
H5’t
H5′c
H6′
5.04 (q)
4.12 (d3)
3.90 (d2)
3.79 (d2)
1.40 (d)
5.20 (d2)
4.25 (d3)
4.05 (d2)
3.95 (d2)
5.19 (d2)
4.23 (d3)
4.04 (d2)
3.96 (d2)
3.47, 3.44
(2d2)
J 2,6a ) 3.9
J 2,6b ) 3.9
J 4,5t ) 6.8
J 4,5c ) 6.6
J 4,4′ ) 4.2
J 5c,5t ) 8.2
J 6a,6b ) 11.1
J 2′,6′a ) 4.0
J 2′,6′b ) 4.0
J 4′,5’t ) 7.0
J 4′,5′c ) 6.7
J 5’t,5′c ) 8.3
J 6′a,6′b ) 11.2
3.59 (d2
)
)
3.56 (d2a
b
J
J 2,6 ) 4.8
J 2,6a ) 3.7
J 2,6b ) 3.7
J 5t,4 ) 6.3
J 5c,5t ) 7.7
J 5c,4 ) 6.0
J 6a,6b ) 11.8
J 2,6a ) 3.7
J 2,6b ) 3.7
J 5t,4 ) 7.7
J 5c,5t ) 7.7
J 5c,4 ) 6.2
J 6a,6b ) 11.2
J 2,6 ) 4.8
J 4,5t ) 6.8
J 4,5c ) 7.0
J 4,4′ ) 6.0
J 5c,5t ) 8.1
J 2′,6′ ) 4.8
J 4′,5’t ) 7.4
J 4′,5′c ) 5.4
J 5’t,5′c ) 8.2
J 2,6a ) 3.8
J 2,6b ) 3.8
J 4,5t ) 6.7
J 4,5c ) 6.5
J 4,4′ ) 4.3
J 2,6 ) 4.8
J 2,6a ) 3.8
J 2,6b ) 3.8
J 5c,4 ) 5.6
J 5c,5t ) 8.5
J 6a,6b ) 11.8
J 2,6a ) 3.8
J 2,6b ) 3.8
J 5c,4 ) 5.0
J 5t,4 ) 6.9
J 5c,5t ) 8.5
J 6a,6b ) 11.2
J 5c,5t ) 8.3
J 6a,6b ) 11.8
J 2′,6′a ) 3.9
J 2′,6′b ) 3.9
J 4′,5’t ) 7.0
J 4′,5′c ) 6.5
J 5’t,5′c ) 8.3
J 6′a,6′b ) 11.8
a
b
Multiplicity: d ) doublet (d2 ) dd, d3 ) ddd), t ) triplet, q ) quadruplet, m ) multiplet, b ) broad. cis and trans designations
(including subscripts) refer to the central 4,4′ bond.
Ta ble 5. 13C NMR Sp ectr a l Da ta (CDCl3/TMS, 25 °C) of
Sch em e 5. Th er m od yn a m ica lly Con tr olled Acid
2,2′-bis(XCH2)-4,4′-BDO (X ) H, Cl, Br ) (4a -c, 5a -c, 6a -c)
P r om oted Rea ction Mod es of Th r eitol w ith
2-Br om oa ceta ld eh yd e
(δ, p p m )
4a
4b
4c
5a
5b
5c
6a
6b
6c
C2 102.2 103.0 102.6 102.2 103.3 102.7 102.2 102.6 102.4
C4
C5
C6
C2′
C4′
C5′
C6′
75.8 76.2 76.3 76.1 76.5 76.6 76.3 76.1 76.3
66.6 67.0 67.1 66.4 67.0 66.9 66.0 66.4 66.5
19.8 44.5 32.4 19.8 44.5 32.4 19.5 43.8 31.7
102.2 103.1 102.6
76.3 75.9 75.8
66.1 66.4 66.4
19.5 43.9 31.7
Ta ble 6. Ca lcu la ted (MM3-GE) Rela tive F r ee En er gy
Differ en ces of 2,2′-Dim eth yl-4,4′-bi(d ioxola n yl) (4-6a ) in
Th eir Low est Min im a Con for m a tion s (k ca l/m ol)
diastereomer
conformation
∆G353rel
trans,trans-4a
H4-C4-C4′-H4′ anti
C5-C4-C4′-C5′ anti
O3-C4-C4′-O3′ anti
H4-C4-C4′-H4′ anti
C5-C4-C4′-C5′ anti
O3-C4-C4′-O3′ anti
H4-C4-C4′-H4′ anti
C5-C4-C4′-C5′ anti
O3-C4-C4′-O3′ anti
0.8
0.7
1.4
0.0
0.4
0.7
0.2
1.1
0.9
cis,trans-5a
cis,cis-6a
the entire mixture of diastereomers as 55 alongside the
TOD product (1c), henceforth 66, to probe the acid (A)-
promoted equilibrium.
Ta ble 7. Yield s of 1c in Va r iou s Rea ction Con d ition s
(Sch em e 3)
Thus, the following expressions could be set up, defin-
ing the equilibrium constants for the three equilibria (eqs
1, 2, and 3) and, since we are measuring de facto overall
concentrations [66]+[66A] and [55]+[55A] (eqs 4 and
5), the observed equilibrium constants (Kobs) can be
expressed as in eq 6, where [A] is a function of substrate
concentration ([S]) and starting concentration of PTSA
[A]o and can be expressed by eq 7. Hence, we performed
a number of reactions, varying the concentrations of
substrate (1c or 4c-6c) and PTSA. The reactions were
monitored by GC and/or NMR, and the results are
presented in Table 8. A plot of the resulting Kobs vs the
initial acid concentrations ([A]o) was made (Figure 1),
and the three equilibrium constants (K1, K2, K3) were
optimized in a nonlinear regression procedure, based on
eqs 6 and 7 and the experimental data.
no.
temp, °C
[PTSA]/[threitol]
time, h
yield, %
1
2
3
4
5
6
80a
157
7
16
16
2
2
5
12
70
60
65
25
60
55
80a
80a
116
19
9
80a
80a
110b
9
a
b
In refluxing benzene; In refluxing toluene.
chromatography. The yields of 1c in some such experi-
ments are presented in Table 7, showing that they
depend on the reaction temperature, on the concentration
of PTSA, and on the reaction time. These results were
taken as an indication that 4c-6c are produced under
“kinetic control” (Table 7, no. 1), after which an equili-
bration process between 4c-6c and 1c sets in, depending
on temperature and concentration of acid (PTSA). We
proceeded probing the mechanism of the reaction de-
picted in Scheme 5. As shown above, isomers 4c, 5c, and
6c were calculated to have nearly the same free energies,
and were obtained in statistical ratio. This justified using
[55A]
K1 )
(1)
[55][A]

1640 J . Org. Chem., Vol. 65, No. 6, 2000
Grabarnik et al.
Ta ble 8. Kobs in th e Equ ilibr ia betw een
2(eq),6(eq)-bis(br om om eth yl)-cis-Tod (1c) a n d All Isom er s
of 2,2′-bis(br om om eth yl)-4,4′-BDO (4-6c) (Sch em e 5)
start.
no. mater. [S] × 103 [A]0 × 103 [S]/[A]0 × 102 time, h Kobs
1
1ca
19.1
18.4
19.4
22.3
10.7
6.2
8.4
18.6
34.6
5.1
228
100
56
160
50
30
100
50
50
50
30
2
1.63
1.82
2.14
1.72
2.03
2.25
1.56
2.17
2.48
3.00
3.06
3.09
2
1ca
3
1ca
4
1cb
437
31
5
1cb
35.0
44.0
9.4
6
1cb
14
111
67
7
4-6ca
4-6ca
4-6ca
4-6ca
4-6ca
4-6ca
10.4
25.0
1.7
8
37.4
9
150
1.13
10
11
12
1.7
350
530
0.49
0.32
0.13
1
1.7
1
1
1.7
1340
a
b
By GC. By NMR.
F igu r e 2. The free energy diagram of the equilibrium between
2(eq),6(eq)-bis(bromomethyl)-cis-tetraoxadecalin (1c) and 2,2′-
bis(bromomethyl)-4,4′-bi(dioxolanyl) (4c-6c) in boiling ben-
zene in the presence of PTSA.
diagram for the set of three equilibria is presented in
Figure 2, in which the species involved appear in the
stability order (kcal/mol): 66A (0.0), 55A (0.8), 66 (2.5),
and 55 (2.7).
Making the reasonable assumption that the rearrange-
ment step itself (66A h 55A) has a lower rate constant,
i.e., a higher barrier (without going into the nature of
this process) associated with it, than the flanking two
and that the Curtin-Hammett principle does not apply,
we may conclude that the equilibrium between 66 () 1c)
and 55 () 4c-6c) depends on the free energy difference
between 66A and 55A. Increasing the PTSA concentra-
tion caused not only reduction of the time to reach
equilibrium (i.e., increase of the reaction rate), but also
growth of the 66A mole fraction and, consequently, Kobs
(Table 8). On the other hand, raising the temperature
caused a decrease in the yield of 1c (no. 6 in Table 7).
This relative increase in the population of 55A can be
easily rationalized in terms of its enhanced entropy on
two accounts: its (rotationally) flexible molecular frame
and its being a mixture of at least three (4c, 5c, and 6c)
diastereomers.
Notably, the equilibrium between 1c and 4c-6c de-
pends on the kind of acid used: all other things being
equal, Kobs for PTSA, Dowex, and TFA were about 2.8,
0.72, and 0.71, respectively. Thus, the association energy
depends not only on the “acidity”, but also on the
structure of the proton carrier (e.g., large steric interfer-
ence by Dowex); be that as it may, the highest association
energy was achieved by cis-TOD-(Oinside) with PTSA.
Theoretical support to this mechanistic picture can be
found in the results of a computational ab initio study
we have recently reported, in which, among the three
possible TOD diastereomers (Scheme 6),6g cis-TOD-Oin-
side exhibits the highest proton affinity, way outside the
range of O-C-O anomeric systems, which are known to
be relatively weak bases,11a but similar to the stronger
gauche O-C-C-O bases.11b This is due, in large mea-
sure, to the strong intramolecular hydrogen bonds within
the cavity of cis-TOD-Oinside (Scheme 6). The fact that
the 2(eq),6(eq)-disubstituted-cis-TOD-Ooutside species
(i.e., the ring-inverted form of 2(ax),6(ax)-cis-TOD-Oin-
side) is practically absent in the reaction product, despite
its calculated energy being similar to that of 2 and 7
F igu r e 1. Experimental and calculated dependence of Kobs
on initial concentration of PTSA, in the equilibrium between
2(eq),6(eq)-bis(bromomethyl)-cis-tetraoxadecalin (1c) and 2,2′-
bis(bromomethyl)-4,4′-bi(dioxolanyl) (4c-6c) in boiling ben-
zene.
[66A]
K2 )
(2)
(3)
[55A]
[66A]
K3 )
[66][A]
[S] ) [55] + [66] + [55A] ) [66A]
(4)
(5)
[A]° ) [A] + [55A] + [66A]
1
K3
K1K2*
+ [A]
(
)
[66] + [66A]
[55] + [55A]
Kobs
)
)
(6)
1 + K1[A]
K1K2 + K3
[A] ) -0.5
+ [S] - [A]° +
(
)
K1K3(1 + K2)
2
K1K2 + K3
K1K2 + K3
0.25
+ [S] - [A]°
+
*[A]°
(
)
K1K3(1 + K2)
K1K3(1 + K2)
x
(7)
In the optimization process, different starting param-
eters were used in the interval from 0 to 106, and the
calculated results (K1 ) 15 ( 4; K2 ) 3.2 ( 0.1; K3 ) 32
( 9 l/mol) correspond to a unique global minimum. The
suggested model appears, therefore, to be correct, and
the equilibrium constants can be safely used for the
calculation of thermodynamic parameters. A free energy
(11) (a) Ganguly, B.; Fuchs, B. J . Org. Chem. 1997, 62, 8892. (b)
Ganguly, B.; Fuchs, B. J . Org. Chem., in press.

Five- vs Six-Membered Diacetal Formation
J . Org. Chem., Vol. 65, No. 6, 2000 1641
Sch em e 6. Rela tive En er gies a n d P r oton
Affin ities (in k ca l/m ol) of tr a n s- a n d
cis-TOD-Oin sid e a n d -Oou tsid e, Ta k en fr om Ref 6g
(ca lcu la ted u sin g Ga u ssia n 94 a t th e MP 2/6-31+G*
level)
axial conformation (which explains why no ring inverted
tb species are seen in this case) is due to its intrinsical
low conformational energy (1.2 kcal/mol)12 (its periodicity
number is 6) and to a possible anomeric effect.6e This is
now being looked into.
Con clu sion
We have unraveled the thermodynamic features of the
diacetalation reaction of threitol with alkylaldehydes and
the details of the dependence of the ratios of the bicyclic
cis-tetraoxadecalin (TOD) (“66”) and bi(dioxolanyl) (BDO)
(“55”) products, on acid-concentration and temperature.
The isomeric diacetals obtained in four such reactions of
substituted aldehydes (RCHO, R ) CH3, CH2Cl, CH2Br,
CO2CH3) with rac-threitol were isolated and character-
ized configurationally and conformationally. A variable
acid-concentration analysis of the equilibrium mixture
of products in one such case (R ) CH2Br) was performed
and provided equilibrium constants and, hence, free-
energy differences among these products and their
unprecedented stable protonated intermediates. The
latter were rationalized by the unusually high proton-
affinity calculated for the cis-TOD (“66”) form (which can
be extrapolated to further threo-tetraol diacetals). At-
tempts to ferret out (mass spectrometrically) these and
similar stable protonated species are in course.
Sch em e 7. Aceta la tion Rea ction P r od u cts of
D-Th r eitol w ith Meth yl Glyoxyla te
Exp er im en ta l Section
Gen er a l. Melting points were recorded on a capillary
melting point apparatus and are uncorrected. The NMR
spectra were obtained on 200, 360, and 500 MHz NMR
spectrometers. All 1H chemical shifts are in ppm and δ-values
are relative to TMS as internal standard. 13C chemical shifts
are reported in ppm relative to CDCl3 (center of triplet δ 77.0)
Mass spectra (DEI-MS, DCI-MS, and FAB) were recorded on
an Autospec 250 mass-spectrometer. Commercial solvents and
reagents were used without purification. Chromatographic
separations were performed on a silica gel 60 column, eluting
with gradient petroleum ether:ethyl acetate (9:1 to 1:1).
Nonlinear regression optimizations were performed using
Origin 4.0. While we deal with racemic mixtures throughout,
the designation is made throughout on D-threitol products.1b
Elemental analyses were not performed, since we deal with
isomeric mixtures. Mass spectral (EI) analyses were performed
in each case, and the molecular ions were observed only for
the methyl-substituted TOD and BDO products; otherwise the
fragment M+ - CH2X was the first and most abundant one.
The 1H and 13C NMR spectra were the diagnostic tool (see text)
and are given in detail in Tables 1, 2, 4, 5, and 9.
Aceta la tion of r a c-Th r eitol w ith Aceta ld eh yd e. To a
mixture of rac-threitol (1.2 g, 10 mmol) and Dowex-H+ (2.6 g)
in dichloromethane (200 mL) was added 1.5 mL (27 mmol) of
acetaldehyde, and the mixture was refluxed with separation
of water in a Dean-Stark adapter. The organic solution was
filtered and evaporated. The crude product (1.6 g, yield 92%)
was separated by column chromatography. The first three
fractions eluted were the three 2,2′-dimethyl-4,4′-bi(1,3-diox-
olanyl) isomers in the order: 6a , 4a , and 5a . The fourth
fraction consisted of a minute quantity of the 2(eq),6(ax) isomer
(2S,6R,9R;9,10-M)-dimethyl-1,3,5,7-cis-tetraoxadecalin (2a ),
and the last fraction was the 2(eq),6(eq) isomer (2S,6S,9R;9,-
10-M)-dimethyl-1,3,5,7-cis-tetraoxadecalin (1a ) (1.5 g, 87%).
Aceta la tion of r a c-Th r eitol w ith 2-Ch lor oa ceta ld e-
h yd e. To a solution of PTSA (0.33 g, 1.7 mmol) in water (20
mL) was added chloroacetaldehyde dimethylacetal (5.7 mL,
50 mmol), and the reaction mixture was stirred at 60-65° C
(Table 3), is evidently due its much lower protonation
energy. It should be mentioned that a similar calculation
of the BDO system has not yet been performed, due to
the CPU exigency of the job.
It is at this point that we see fit to present the findings
that actually triggered the above-reported investigation
and which stemmed from our keen interest in TOD
systems with functionalized substituents, as cores for
novel macromolecular systems. These are the results of
the condensation of rac-threitol with methyl glyoxylate
(MeO2CCHO) (Scheme 7), viz., all three isomers of bis-
(methoxycarbonyl)-1,3,5,7-cis-TOD, which were individu-
ally isolated, viz., 2(eq),6(eq) (1d ), 2(eq),6(ax) (2d ) and
(the only existing case of) 2(ax),6(ax) (3d ) in 20%, 15%,
and 2% yield, respectively, along with the 2,2′-bis-
(methoxycarbonyl)-4,4′-BDO isomers: 4d , 5d , and 6d ,
which were obtained in 25% yield and 1:2:1 ratio (as
evaluated by NMR of the mixture) and could not be
separated. The detailed NMR data (strengthened by NOE
and DEPT studies) are given in Table 9 and are il-
luminating, due to the presence of the hitherto unavail-
able 2(ax),6(ax) form (3d ). Both the 1H and 13C NMR
spectra of the 2(eq),6(ax)-isomer (2d ), exhibit two sets of
signals, as before (vide supra). The first set in each
spectrum is similar to that observed in the spectrum of
the symmetrical diequatorial system (1d ), while the
second set resembles that of 2(ax),6(ax)-bis(methoxycar-
bonyl)-1,3,5,7-cis-TOD (3d ). The fact that axially 2(6)-
substituted species (2d , 3d ) are relatively well populated
in this case is clearly due to the methoxycarbonyl
substituent. The latter’s relative propensity to assume
(12) Eliel, E. L.; Willen, S. Stereochemistry of Carbon Compounds;
Wiley: New York, 1994; p 696.

1642 J . Org. Chem., Vol. 65, No. 6, 2000
Grabarnik et al.
Ta ble 9. 1H a n d 13C NMR Sp ectr a l Da ta (CDCl3/TMS, 25 °C) of 2(eq),6(eq)- Bis(m eth oxyca r bon yl) cis-Tod (1d ),
2(eq),6(a x)-bis(m eth oxyca r bon yl)-cis-Tod (2d ), a n d 2(a x),6(a x)-Bis(m eth oxyca r bon yl)-cis-Tod (3d ) (δ, p p m ; J ,a Hz)
1H NMR
H2
H4eq
H4ax
H9
OCH3
H6
H8ax
H8eq
H10
OCH3
1d
2d
3d
5.11 (s)
4.41 (bd)
3.93 (d2)
3.81 (bs)
3.85 (s)
J 4eq,4ax ) 12.8; J 4ax,10 ) 1.3
5.11 (s)
5.50 (s)
4.34 (d2)
3.96 (d2)
3.83 (s)
3.85 (s)
5.48 (bs)
4.22 (d2)
4.12 (d2)
4.15 (bs)
3.82 (s)
J 4eq,4ax ) 12.9, J 4eq,10 ) 1.0, J 4ax,10 ) 1.9, J 9,8ax ) 2.1, J 9,8eq ) 1.4, J 8ax,8eq ) 13.0
4.16 (d2)
4.07 (d2)
4.18 (bs)
3.83 (s)
J 4eq,4ax ) 13.1; J 4ax,10 ) 1.4
13C NMR
C2
C4
C9
OCH3
CdO
C6
C8
C10
OCH3
52.4
CdO
1d
2d
3d
95.9
95.9
92.4
69.5
69.8
64.7
69.9
70.2
64.1
52.9
53.0
52.5
165.3
165.5
168.1
92.3
64.5
63.7
168.1
a
Multiplicity: d ) doublet (d2 ) dd, d3 ) ddd), t ) triplet, q ) quadruplet, m ) multiplet, b ) broad.
for 5 h, until the solution became clear, indicating full acetal
hydrolysis (by NMR). This aqueous solution was continuously
extracted with dichloromethane for 48 h, and the chloroac-
etaldehyde solution was continuously transferred to a mixture
of rac-threitol (3 g, 24 mmol) and Dowex-H+ (5 g) in boiling
dichloromethane (250 mL). The resulting solution was filtered
and evaporated. The crude product (5.3 g, yield 91%) was
crystallized from ethyl acetate, to give the 2(eq),6(eq) (2S,
6S,9R;9,10-M)-bis(chloromethyl)-1,3,5,7-cis-tetraoxadecalin (1b)
(3.4 g, 58%). Evaporation of the mother liquor gave an oil,
which was separated by column chromatography. The first
three products were the 2,2′-bis(chloromethyl)-4,4′-bi(1,3-
dioxolanyl) isomers, in the order: 4b, 6b, and 5b. The fourth
fraction consisted of a minute quantity of the 2(eq),6(ax)
(2S,6R,9R;9,10-M)-bis(chloromethyl)-1,3,5,7-cis-tetraoxadeca-
lin (2b) and the last product of additional 1b.
Aceta la tion of r a c-Th r eitol w ith 2-Br om oa ceta ld e-
h yd e. To a solution of PTSA (4.85 g, 25 mmol) in water (25
mL) was added bromoacetaldehyde dimethylacetal (5 mL, 41
mmol), and the reaction mixture was stirred at 60-65° C for
4 h, until the solution became clear, indicating full acetal
hydrolysis (by NMR). rac-Threitol (2 g, 15.9 mmol) was added,
along with 150 mL of benzene. The mixture was refluxed
overnight under Ar, using a Dean-Stark adapter. After
cooling, the organic solution was washed with 10% aqueous
sodium carbonate (20 mL), dried with calcium chloride, and
evaporated. The crude product (5.1 g, 97%) was crystallized
from methanol, to give 1c (2.1 g, 40%). Evaporation of the
mother liquor provided an oil, which was separated by column
chromatography. The first three fractions were the 2,2′-bis-
(bromomethyl)-4,4′-bi(1,3-dioxolanyl) isomers, in the order: 4c,
6c, and 5c. The fourth fraction consisted of a minute quantity
of 2(eq),6(ax)-bis(chloromethyl)-1,3,5,7-cis-tetraoxadecalin (2c)
and the last fraction was additional 1c.
washed with water and aq K2CO3, dried on MgSO4, filtered,
and evaporated. The residue was separated by column chro-
matography, to give first three fractions (the NMR spectra are
assembled in Table 9):
(2S,6S,9R;9,10-M)-Bis(methoxycarbonyl)-1,3,5,7-cis-TOD
(2(eq),6(eq)) (1d ) (175 mg, 20%). mp: 136.5-137.5 °C (from
EtAc). TLC: Rf ) 0.28 (EtAc, silica gel). EI-MS (70 eV): m/z
(%) 261 ([M - H]+, ∼1), 203 (100); 115 (19). FAB-MS (70 eV):
m/z (%) 263 ([MH]+, 100), 203 (91). Anal. Calcd for C10H14O8
(262.22): C, 45.81; H, 5.38. Found: C, 45.98; H, 5.49.
(2S,6R,9R;9,10-M)-Bis(methoxycarbonyl)-1,3,5,7-cis-TOD (2d)
(2(eq),6(ax)) (130 mg, 15%). TLC: Rf ) 0.56 (EtAc, silica gel).
EI-MS (70 eV): m/z (%) 203 (100), 115 (47). FAB-MS (70 eV):
m/z (%) 263 ([MH]+, 66), 203 (100).
(2R,6R,9R;9,10-M)-Bis(methoxycarbonyl)-1,3,5,7-cis-TOD (3d)
(2(ax),6(ax)) (2 mg, 2%). TLC: Rf ) 0.73 (E.A, silica gel). EI-
MS (70 eV): m/z (%) 203 (100), 115 (40).
The following fraction (220 mg, 25%) was a mixture of the
2,2′-bis(methoxycarbonyl)-4,4′-bi(dioxolanyl) isomers: 4d , 5d ,
and 6d , in 1:2:1 ratio (from the NMR spectrum of the mixture).
Equ ilibr iu m Stu d ies. A solution of PTSA in benzene was
refluxed for 2 h, using a Dean-Stark adapter under Ar, after
which, was added a mixture of substrate (1c or 4c-6c) with
biphenyl (as an inert GC internal standard) in mole ratio 1:1.
Samples were taken through an immersed capillary tube using
positive Ar pressure, washed with 10% aqueous sodium
carbonate, and examined by GC or, after evaporation of
benzene, by 1H NMR in CDCl3. Every experiment was finished
when the ratio between 1c and 4c-6c remained unchanged.
Chromatographic analysis were run on a Varian 3400 instru-
ment with TCD on Megabor DB5 column (15 m × 0.53 mm,
He carrier gas flow 30 mL/min, column temperature from 60
to 200° C with gradient 15 deg/min). The retention times of
biphenyl, 4c-6c and 1c were 4.7, 8.1, and 8.4 min, respec-
tively.
Aceta la tion of r a c-Th r eitol w ith Meth yl Glyoxyla te.
To a solution of PTSA (0.25 g, 1.3 mmol) and dimethoxyacetic
acid methyl ester (methyl glyoxylate dimethyl acetal) (1.24 g,
9.3 mmol) in water (5 mL) was added rac-threitol (0.4 g, 3.28
mmol) followed by 40 mL of benzene. The mixture was refluxed
overnight under Ar, using a Dean-Stark adapter. The benzene
overlayer was decanted, and the residual viscous oil was
dissolved in methanol (40 mL). The resulting solution was
concentrated and taken up in chloroform. This solution was
Ack n ow led gm en t. This work was supported in part
by grants from the Israel Science Foundation and the
Ministry of Science and by an Intel scholarship (to
N.G.L.). Ms. Diana Glazer provided able undergraduate
student research assistance.
J O9908952