Experimental Studies of the 13C NMR of Iodoalkynes in Lewis-Basic Solvents

Article abstract of DOI:10.1021/jo035584c

The ¹³C NMR spectra of two different iodoalkynes, 1-iodo-1-hexyne (1) and diiodoethyne (2), exhibit a strong solvent dependence. Comparisons of the data with several common empirical models, including Gutmann's Donor numbers, Reichardt's E^NT, and Taft and Kamlet's β and π*, demonstrate that this solvent effect arises from a specific acid-base interaction. Solvent basicity measures such as Donor numbers and β values correlate well with the α-carbon chemical shift of 1, but polarity measures such as E^NT and π* do not correlate. The similarity of the solvent effect for 1 and 2 suggests that carbon-carbon bond polarization may not play a role in the change in chemical shift, as previously hypothesized.

Full text of DOI:10.1021/jo035584c

Exp er im en ta l Stu d ies of th e 13C NMR of Iod oa lk yn es in
Lew is-Ba sic Solven ts
J effrey A. Webb, J aap E. Klijn, Philip Aru Hill, J ordan L. Bennett, and Nancy S. Goroff*
Department of Chemistry, State University of New York, Stony Brook, New York 11794-3400, and
Department of Chemistry and Biochemistry, Denison University, Granville, Ohio 43023
nancy.goroff@sunysb.edu
Received October 28, 2003
The 13C NMR spectra of two different iodoalkynes, 1-iodo-1-hexyne (1) and diiodoethyne (2), exhibit
a strong solvent dependence. Comparisons of the data with several common empirical models,
including Gutmann’s Donor numbers, Reichardt’s ENT, and Taft and Kamlet’s â and π*, demonstrate
that this solvent effect arises from a specific acid-base interaction. Solvent basicity measures such
as Donor numbers and â values correlate well with the R-carbon chemical shift of 1, but polarity
measures such as EN and π* do not correlate. The similarity of the solvent effect for 1 and 2
T
suggests that carbon-carbon bond polarization may not play a role in the change in chemical shift,
as previously hypothesized.
In tr od u ction
We recently reported an unusual solvent effect in the
13C NMR spectra of iodoalkynes.1,2 Depending on the
solvent, the chemical shift of the R-carbon can change
by as much as 15 ppm. Because of the unusual magnitude
and direction of this solvent effect, we have decided to
study it both computationally and experimentally. Ab
initio calculations, which have already been described,
demonstrated that, in the gas phase, specific intermo-
lecular interactions between the Lewis-acidic iodoalkyne
and basic solvent (Figure 1) can explain this change.2
Here we present experimental studies which probe the
role of bulk solvent in the chemical shift change. These
data describe the 13C NMR of two iodoalkynes, 1-iodo-1-
hexyne (1) and diiodoethyne (2), in a wide variety of
solvents.
Laurence and co-workers have carried out extensive
studies of iodoalkynes and their Lewis acidity.3-8 They
demonstrated that acid-base interactions can affect
iodoalkyne vibrational spectra, in particular the carbon-
iodine bond stretching frequency. Donation of electron
density from the base into the C-I antibonding orbital
weakens the C-I bond and lowers the frequency of
vibration. In the presence of Lewis bases in nonpolar
F IGURE 1. The interaction of a Lewis base (B) with an
iodoalkyne.
solvent, Laurence et al. observed separate C-I stretching
modes corresponding to the free and complexed iodo-
alkyne.4 They suggested that the difference in energy for
these two stretching modes corresponds to the strength
of the base.
Laurence and co-workers also examined the 13C NMR
spectra of different iodoalkynes. Iodine exerts a large
magnetic influence on neighboring nuclei through the
“heavy-atom effect” due to spin-orbit coupling of its
valence electrons.9-12 This relativistic effect lowers the
chemical shift of the adjacent carbon atom by about 60
ppm. Laurence found that the chemical shift of the
R-carbon in an iodoalkyne (C-I) is directly related to that
of the analogous terminal alkyne (C-H),8 with the
iodoalkyne chemical shift approximately 70 ppm lower
in frequency than the terminal alkyne value (i.e., around
0 ppm for an iodoalkyne in deuteriochloroform, CDCl3).
Interestingly, the two iodoalkynes that did not match
Laurence’s C-H/C-I correlation were those with Lewis-
(1) Gao, K.; Goroff, N. S. J . Am. Chem. Soc. 2000, 122, 9320-9321.
(2) Rege, P. D.; Malkina, O. L.; Goroff, N. S. J . Am. Chem. Soc. 2002,
124, 370-371.
(3) Laurence, C.; Queignec-Cabanetos, M. J . Chem. Soc., Dalton
Trans. 1981, 2144-2146.
(4) Laurence, C.; Queignec-Cabanetos, M.; Dziembowska, T.;
Queignec, R.; Wojtkowiak, B. J . Am. Chem. Soc. 1981, 103, 2567-
2573.
(9) Cheremisin, A. A.; Schastnev, P. V. J . Magn. Reson. 1980, 40,
459-468.
(5) Laurence, C.; Queignec-Cabanetos, M.; Wojtkowiak, B. J . Chem.
Soc., Perkin Trans. 2 1982, 1605-1610.
(10) Kaupp, M.; Malkina, O. L.; Malkin, V. G.; Pyykko¨, P. Chem.
Eur. J . 1998, 4, 118-126.
(6) Laurence, C.; Queignec-Cabanetos, M. Bull. Soc. Chim. Belg.
1982, 91, 377-377.
(11) Kaupp, M.; Malkina, O. L.; Malkin, V. G. J . Comput. Chem.
1999, 20, 1304-1313.
(7) Queignec-Cabanetos, M.; Laurence, C. J . Chim. Phys. Chim. Biol.
1982, 79, 603-607.
(12) Malkina, O. L.; Schimmelpfennig, B.; Kaupp, M.; Hess, B. A.;
Chandra, P.; Wahlgren, U.; Malkin, V. G. Chem. Phys. Lett. 1998, 296,
93-104.
(8) Laurence, C.; Queignec-Cabanetos, M.; Wojtkowiak, B. Can. J .
Chem. Rev. Can. Chim. 1983, 61, 135-138.
10.1021/jo035584c CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/07/2004
660
J . Org. Chem. 2004, 69, 660-664

13C NMR of Iodoalkynes in Lewis-Basic Solvents
basic end groups, namely compounds 3 and 4, each of
which had a higher R-carbon chemical shift than pre-
dicted from the terminal-alkyne NMR.
Laurence et al. measured the NMR spectra of iodo-
alkynes only in CDCl3. However, we have found that in
DMSO or pyridine, the R-carbon chemical shifts of
iodoalkynes move 10-15 ppm higher in frequency.1 Here,
we describe studies of the iodoalkyne chemical shift both
in nonpolar solvents with dissolved base and in neat basic
solvents. In both cases, the presence of a Lewis base has
a measurable effect on the R-carbon chemical shift.
Exp er im en ta l Section
1-Iod o-1-h exyn e (1).13 Several synthetic routes have been
used to prepare 1. The following procedure is preferred because
it provides 1 in good yields and high purity. Under argon
atmosphere, 1-hexyne (1.64 g, 20 mmol), AgNO3 (0.84 g, 5
mmol), and N-iodosuccinimide (9.89 g, 44 mmol) were added
to 200 mL of acetone. The reaction mixture was stirred at room
temperature for approximately 3 h. Cold water (100 mL) was
added to quench the reaction. Hexanes (100 mL) was added,
and the resulting organic layer was washed with water and
brine (∼250 mL each), and finally with 0.05 M Na2S2O3, to
remove I2, until the organic layer had no color. The organic
layer was dried with MgSO4 and concentrated in vacuo to yield
a brown liquid (2.50 g, 60%). A slight pink color is an indication
of the presence of I2 in the product. 13C NMR (62 MHz, CDCl3)
δ -7.7, 13.5, 20.5, 21.8, 30.5, 94.8.
Diiod oeth yn e (2). A typical synthetic procedure (modified
from Dehn14) is as follows: Acetylene gas was bubbled into a
mechanically stirred solution of KI (25.54 g, 154 mmol) in 50
mL of water. NaOCl solution (10-13% available chlorine) was
added dropwise to the solution at such a rate that the yellow
color of OI- anion would quickly disappear. Addition of NaOCl
solution and bubbling of acetylene gas were stopped when no
color change was observed upon addition of the NaOCl
solution. The reaction was cooled in an ice bath. The precipi-
tate was filtered and washed with cold water. The product was
dissolved in 75 mL of petroleum ether and dried with calcium
chloride. Recrystallization of the product from the decanted
petroleum ether yielded 2 (9.497 g, 44.5% based on KI). 13C
NMR (75 MHz, CDCl3) δ 0.62, mp 78-79 °C (lit. mp 79 °C).
NMR Exp er im en ts. Solvents were used as purchased
without further purification. For dilute experiments, samples
were prepared by dissolving a 1:1 mixture of 1 and the base
of interest (for example, 0.1 mmol of each) in hexanes, and
diluting the sample to 1 mL in a volumetric flask. The
concentrations of 1 and base ranged from 0.010 to 0.50 M.
NMR spectra were obtained at 25 °C, using a Varian Inova
600 operating at 150.8 MHz for 13C. For neat solvent experi-
ments, samples were prepared by dissolving 1 (0.062 g, 0.30
mmol) or 2 (0.083 g, 0.30 mmol) in the solvent of interest and
diluting the sample to 1 mL in a volumetric flask. NMR spectra
of 1 in neat solvents were obtained at 25 °C, using a Varian
Inova 600 operating at 150.8 MHz for 13C. NMR spectra of 2
in neat solvents were obtained at 25 °C, using a Bruker AC-
F IGURE 2. Dilute solution studies of 1 complexed with
several bases in hexanes. Concentration of base and 1 kept
approximately equal.
dine can have a significant effect on the chemical shift
of the R-carbon of an iodoalkyne. To explore whether this
solvent effect comes from bulk or molecular properties,
we measured the chemical shift of 1 in the presence of
dilute base. For these experiments, hexanes was used as
a nonpolar, noninteracting solvent. We chose compound
1 as the substrate for these studies because it is easily
prepared, because the R-carbon chemical shift appears
far from other peaks in the NMR spectrum (-8.1 ppm
at 0.1 M in hexanes), and because 1 lacks other functional
groups that might complicate the interactions with
solvent. In each case, a 1:1 mixture of 1 and the
appropriate base was dissolved in hexanes, and the
R-carbon chemical shift was measured.
The dilute solution experiments (Figure 2) provide
qualitative evidence for the importance of specific mo-
lecular interactions and Lewis basicity to the observed
solvent effect, consistent with computational results. As
shown in Figure 2, even without added base, the chemical
shift changes slightly with increased concentration (-7.9
ppm at 0.4 M in hexanes), perhaps because of self-
aggregation of the iodoalkyne. But adding a Lewis base
to the solution affects the chemical shifts even more.
Pyridine and quinoline, the two strongest bases, cause
the biggest change in R-carbon chemical shift; for ex-
ample, the R-carbon of 1 produces a shift of -5.7 ppm in
a 0.4 M solution of pyridine.
Nonetheless, in each case, the effect of dilute amounts
of base is much smaller than that of the corresponding
bulk solvent. For instance, in pure pyridine, the chemical
shift of 1 is +3.85 ppm. The relatively small changes in
chemical shift in these dilute solutions make it impossible
to calculate a reliable association constant for 1 with any
of the bases examined. These experiments also raise the
question of whether other bulk properties, in particular
solvent polarity, might be significant contributors to the
observed solvent effect.
300, interfaced to a Techmagnetics Tecmag MacSpect
3
controller, operating at 75 MHz. For all experiments an
external lock signal was produced by using a sealed glass
capillary containing either cyclohexane-d12 (for experiments
with 1) or acetone-d6 (for experiments with 2).
Resu lts a n d Discu ssion
Iod oa lk yn es in Nea t Ba sic Solven ts. To examine
the role of bulk solvent properties in the observed NMR
shifts, we carried out studies in neat solvents of varying
basicity and polarity, including halogenated hydrocar-
bons, aromatic solvents, amines, and ethers. The ob-
served R-carbon chemical shift for 1, δ(1), spans a range
Iod oa lk yn es in th e P r esen ce of Dilu te Ba se. Initial
experiments indicated that basic solvents such as pyri-
(13) Nelson, D. J .; Blue, C. D.; Brown, H. C. J . Am. Chem. Soc. 1982,
104, 4913-4917.
(14) Dehn, W. H. J . Am. Chem. Soc. 1911, 33, 1598.
J . Org. Chem, Vol. 69, No. 3, 2004 661

Webb et al.
TABLE 1. Nea t Solven t NMR Da ta a n d Selected Solven t P a r a m eter s
e
g
solvent
δ(1)b
δ(2)c
DNd
EN
âf
π* f
Bsoft
δ(CHCl3)h
T
hexanes
-7.95
-7.96
-7.10
-6.89
-6.29
-6.10
-4.53
-4.40
-3.62
-3.33
-2.72
-1.56
0.27
2.78
3.55
3.75
3.85
5.09
7.10
7.17
-1.45
0.62
0.61
1.34
0.47
1.27
3.47
0.009
0.256
0.259
0.327
0.052
0.111
0.228
0.117
0.460
0.207
0.355
0.065
0.404
0.043
0.444
0.269
0.302
0.315
0.145
0.00
-0.08
CDCl3
CHCl3
a
0.00
0.58
1.0
0.20
dichloroethane
CCl4
benzene
ethyl acetate
diethyl ether
acetonitrile
tetrahydrofuran
acetone
dimethyl sulfide
dimethylformamide
triethylamine
dimethyl sulfoxide
quinoline
0.0
10.0
0.1
17.1
19.2
14.1
20.0
17.0
0.10
0.45
0.47
0.31
0.55
0.48
0.59
0.55
0.27
0.75
0.58
0.71
20.5
17.0
25.0
18.0
0.74
0.56
0.79
0.92
5.26
4.67
5.80
7.31
8.68
26.6
31.7
29.8
0.69
0.71
0.76
0.88
0.14
1.00
30.0
1.30
1.22
1.32
12.36
13.67
13.03
11.65
32.0
50.0
57.5
pyridine
HMPA
diethylamine
N-methylimidazole
33.1
38.8
0.64
0.87
1.56
2.06
16.40
a
Chloroform used as purchased, including 1% ethanol as stabilizing agent. b 13C chemical shift of the R-carbon of 1, relative to external
cyclohexane-d12 lock. c 13C chemical shift of the R-carbon of 2, relative to external acetone-d6 lock. Gutmann’s donor number, from ref
d
15. e Reichardt’s EN values, from ref 16. f Taft and Kamlet’s â and π* parameters, from refs 17-20. Laurence’s Bsoft, from ref 4. 1H
g
h
T
chemical shift of chloroform, as reported in ref 26.
of more than 15 ppm, from -7.9 ppm in hexanes to +7.2
ppm in N-methyl imidazole (Table 1). Interestingly, even
nonpolar solvents such as CCl4 exert a noticeable effect
on the chemical shift. Qualitatively, the more basic
solvents produced larger changes in δ(1).
To help understand these data, we turned to several
common empirical models of solvent behavior. These
models include Gutmann’s donor number system for
16
Lewis bases,15 Reichardt’s polarity measure, EN
,
and
T
Taft and Kamlet’s â and π* solvent model.17-20 In
addition, we compared our data with Laurence’s proposed
measure of soft basicity, Bsoft.4 Analysis with additional
empirical solvent models is described in the Supporting
Information.
For a given Lewis base, Gutmann has defined the
donor number (DN) as the negative enthalpy of formation
of a 1:1 complex of that base with the Lewis acid
antimony(V) chloride, SbCl5, in dilute solution, typically
in dichloroethane.15 The DN parameter reflects only the
molecular behavior of the solvent, not bulk properties.21,22
Nonetheless, a linear regression demonstrates that the
13C NMR data for 1 correlate very well with DN values
(Figure 3, R2 ) 0.873). This analysis yields eq 1, describ-
ing the chemical shift (in ppm) of the R-carbon, δ(1), as
a function of the donor number of the solvent used.
F IGURE 3. The R-carbon 13C NMR shift of 1 as a function of
Gutmann’s donor numbers.15
correlate significantly with the data in Table 1 (Figure
4, R2 ) 0.0495). The EN values are based on the
T
solvatochromic shift in the absorption maximum of a
betaine dye.16 This dye is not Lewis acidic, and its
absorption maximum is a good measure of nonspecific
solvent polarity effects.23,24
Taft and Kamlet have developed a solvatochromic
comparison method that includes parameters for both
basicity (â) and polarity (π*) effects.17-20 The â values
come from the solvatochromic shift in the absorption
spectrum of 4-nitroaniline. Unlike Gutmann’s donor
numbers, the â values therefore reflect bulk solvent
basicity, and in particular solvent hydrogen-bond accep-
δ(1) ) -7.99 + 0.32(DN)
(1)
In contrast, Reichardt’s EN parameters, which act as
T
an empirical measure of bulk solvent polarity, do not
(15) Gutmann, V. The Donor-Acceptor Approach to Molecular
Interactions; Plenum Press: New York, 1978.
(16) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry,
2nd ed.; VCH: New York, 1990.
(17) Kamlet, M. J .; Taft, R. W. J . Am. Chem. Soc. 1976, 98, 377-
383.
tor ability. The π* parameters, similar to Reichardt’s EN
T
values, were determined via solvatochromic shifts in the
absorption spectra of several nitroaromatics, including
(18) Taft, R. W.; Kamlet, M. J . Org. Magn. Reson. 1980, 14, 485-
493.
(19) Taft, R. W.; Abboud, J .-L. M.; Kamlet, M. J . J . Am. Chem. Soc.
1981, 103, 1080-1086.
(20) Taft, R. W.; Gramstad, T.; Kamlet, M. J . J . Org. Chem. 1982,
47, 4557-4563.
(21) Rais, J .; Okada, T. J . Phys. Chem. A 2000, 104, 7314-7323.
(22) Masuda, Y. J . Phys. Chem. A 2001, 105, 2989-2996.
(23) Grodkowski, J .; Neta, P. J . Phys. Chem. A 2002, 106, 11130-
11134.
(24) Imai, Y.; Chujo, Y. Macromolecules 2000, 33, 3059-3064.
662 J . Org. Chem., Vol. 69, No. 3, 2004

13C NMR of Iodoalkynes in Lewis-Basic Solvents
F IGURE 4. The R-carbon 13C NMR shift of 1 as a function of
F IGURE 6. The R-carbon 13C NMR shift of 1 as a function of
Laurence’s Bsoft values.4
Reichardt’s normalized EN polarity model.16
T
remaining deviations. The linear relation between δ(1)
and â is shown in eq 2.
δ(1) ) -7.72 + 12.50(â)
(2)
In addition to the models described above, we also
compared our data with Bsoft, a parameter developed by
Laurence and co-workers to quantify “soft” basicity.4 Like
iodoalkynes, Laurence et al. found that iodine cyanide
(ICN) gives two distinct IR peaks for the C-I stretch
when in the presence of dilute solutions of various Lewis
bases. They defined Bsoft as the difference in the C-I
stretching frequency of ICN, free vs complexed, and
offered it as a general measure of “soft” basicity. The
chemical similarity between Laurence’s system and the
one described here suggests that there should be a strong
correlation between Bsoft and our experimental 13C NMR
shift. Yet Bsoft reflects a snapshot on the IR time scale of
the vibrational properties of separate free and complexed
ICN, without regard to the equilibrium constant for
complexation. It should act as a measure of the enthalpy,
but not the entropy, of complexation. The δ(1) values, on
the other hand, represent a weighted average of all
iodohexyne molecules present in the solution, necessarily
incorporating both enthalpy and entropy effects. The two
data sets may therefore diverge if entropy plays a more
significant role in complexation for some solvents than
others. In fact, the NMR data correlate about as well with
the Bsoft values (Figure 6, R2 ) 0.857) as with DN or â,
suggesting that there are no major changes in the entropy
of interaction as the solvent changes. The lack of entropy
effects implies that steric interactions are not significant
for this system, understandable given the linear geometry
of the iodoalkyne.
F IGURE 5. The R-carbon 13C NMR shift of 1 as a function of
Taft and Kamlet’s solvatochromatic parameters: (a) â and (b)
π*.17-20
These comparisons with empirical models demonstrate
that the observed solvent effect results from specific
solute-solvent interactions. They also suggest a strong
similarity between the interactions of solvent with 1 and
with hydrogen-bond donating solutes. For example,
solvent effects on the NMR spectra of dissolved CHCl3
have been extensively studied, and its 13C NMR chemical
shift has been shown to depend on the basicity of the
solvent used.18,25-27 Two groups have correlated these
data with Taft and Kamlet’s â values, with one study
4-nitroanisole, and N,N-diethyl-3-nitroaniline. Linear
regression analysis shows that the NMR data for 1
correlate well with â alone (Figure 5a, R2 ) 0.838), but
not with π* (Figure 5b, R2 ) 0.316). In addition, linear
analysis with both parameters offers no significant
improvement over using â alone (R2 ) 0.845). This
analysis indicates that although â is not perfectly cor-
related with the NMR data, π* cannot explain the
suggesting that solvent polarity also plays a role.18,27
A
further confirmation of the similarity in the two systems
J . Org. Chem, Vol. 69, No. 3, 2004 663

Webb et al.
F IGURE 7. The R-carbon 13C NMR shift of 1 as a function of
F IGURE 8. The R-carbon 13C NMR shift of 2 as a function of
the R-carbon 13C NMR shift of 1.26
1H NMR shifts of CHCl3.22
1
comes from comparing the 13C NMR data for 1 to the H
NMR data for CHCl3 (Figure 7, R2 ) 0.885).
complexed to only one Lewis base. We are currently
addressing these two possibilities via computer modeling.
In addition, the direction of the solvent effect is the
same for both chloroform and compound 1. For example,
the 13C chemical shift of chloroform in HMPA is 4.2 ppm
higher than that in cyclohexane.26 Nonetheless, it is
noteworthy that the observed 13C chemical shift range
for 1 is significantly larger than that for chloroform.
Iodoalkynes have a unique sensitivity to solvent basicity
in their NMR spectra.
Con clu sion s
The experimental studies described here have demon-
strated the generality of the solvent effect previously
observed in the NMR signal of iodoalkynes. Correlations
with various empirical models of solvent basicity, includ-
ing Gutmann’s donor numbers and Taft and Kamlet’s â
values, indicate that acid-base interactions are respon-
sible for the observed change in chemical shift. Solvent
polarity does not appear to play a role, as demonstrated
Diiod oeth yn e. Calculations have suggested that car-
bon-carbon bond polarization is an important component
of the mechanism for this solvent effect.2 To examine the
role of bond polarization in the NMR of iodoalkyne, we
turned to studies of diiodoethyne (2). While iodohexyne
(1) is a monodentate iodoalkyne with no competing
functional groups, compound 2 contains two Lewis-acidic
iodine atoms, each bonded to one of the carbon atoms of
the lone triple bond. Because the triple bond cannot be
polarized in both directions at once, the symmetry of 2
might lead to a lessening of the NMR solvent effect,
compared to that of 1.
by the poor correlation to Reichardt’s EN and Taft and
Kamlet’s π* parameter.
T
Nonetheless, there is still more to learn about the
mechanism by which the solvent-solute interactions lead
to an increase in chemical shift. Calculations indicate
that polarization of the carbon-carbon triple bond plays
a role. But experiments on the symmetric substrate
diiodoethyne (2) suggest that the situation may be more
complicated than originally proposed.
Figure 8 shows a plot of the R-carbon chemical shift of
1 vs the carbon chemical shift of 2 in the same solvent.
As can be seen, the solvent effect is virtually the same
in 1 and 2 (R2 ) 0.977), but the NMR signal of 2 is
slightly more sensitive to solvent (slope ) 1.10). These
data suggest that either bond polarization is not a
significant component of the solvent effect or that, on the
NMR time scale, many of the diiodoethyne molecules are
Ack n ow led gm en t. We thank C. Grey, J . Osegovic,
J . Marecek, F. Raineri, and P. Rablen for helpful
discussions. We thank P. Rege and A. Nasir for initial
contributions to this research. We acknowledge the
National Science Foundation (CHE-9984937) and the
donors of the Petroleum Research Fund, administered
by the American Chemical Society, for support of this
research.
Su p p or tin g In for m a tion Ava ila ble: Details of each
linear regression analysis. This material is available free of
charge via the Internet at http://pubs.acs.org.
(25) Lichter, R. L.; Roberts, J . D. J . Phys. Chem. 1970, 74, 912-
916.
(26) Martin, M. L. Ann. Phys. 1962, 7, 35-55.
(27) Ueji, S.; Nakamura, M.; Kitadani, M.; Sera, A. Bull. Chem. Soc.
J pn. 1982, 55, 1665-1666.
J O035584C
664 J . Org. Chem., Vol. 69, No. 3, 2004