[5]HELOL phosphite: A helically grooved sensor of remote chirality

Article abstract of DOI:10.1021/ja001904n

Even when the first center of chirality is far from a functional group, the stereochemistries of chiral molecules can be sensed if the functional group is attached to a phosphorus embedded within the groove of a nonracemic helicene. The sensor used was a phosphite ester of [5]HELOL, structure 6, and ³¹P NMR spectroscopy was the method of analysis. Although there are seven methylenes between the hydroxyl and the first center of chirality, the ³¹P NMRs could distinguish the enantiomers of 8-phenylnonanol with baseline resolution. However, to achieve this resolution, the solvent has to be chosen appropriately. The probe can also analyze the enantiomers of other alcohols, amines, phenols, and, when they were coupled to 2-aminophenol, carboxylic acids. The sensitivity to remote chiral centers of the moieties, analyzed is attributed to their confinement to the helical groove.

Full text of DOI:10.1021/ja001904n

J. Am. Chem. Soc. 2000, 122, 10027-10032
10027
[5]HELOL Phosphite: A Helically Grooved Sensor of Remote
Chirality
Daniel J. Weix, Spencer D. Dreher, and Thomas J. Katz*
Contribution from the Department of Chemistry, Columbia UniVersity, New York, New York 10027
ReceiVed May 30, 2000. ReVised Manuscript ReceiVed August 3, 2000
Abstract: Even when the first center of chirality is far from a functional group, the stereochemistries of chiral
molecules can be sensed if the functional group is attached to a phosphorus embedded within the groove of
a nonracemic helicene. The sensor used was a phosphite ester of [5]HELOL, structure 6, and 31P NMR
spectroscopy was the method of analysis. Although there are seven methylenes between the hydroxyl and the
first center of chirality, the 31P NMRs could distinguish the enantiomers of 8-phenylnonanol with baseline
resolution. However, to achieve this resolution, the solvent has to be chosen appropriately. The probe can also
analyze the enantiomers of other alcohols, amines, phenols, and, when they were coupled to 2-aminophenol,
carboxylic acids. The sensitivity to remote chiral centers of the moieties analyzed is attributed to their
confinement to the helical groove.
Introduction
structure, chiral centers of X, even those distant from the point
at which X is attached to the phosphorus, could experience
significant diastereomeric interactions. Since 5 (X ) Cl) should
form when 6 is combined with PCl3, since 5 gives the same
Although the chiral purities of enantiomeric mixtures can be
analyzed by a number of chromatographic1 or NMR spectro-
scopic methods,2 when the centers of chirality are remote from
reactive functional groups, the analyses fail. For example, the
(R)- and (S)-enantiomers of 1 for n up to 2 can be distinguished
because the 31P, 77Se, and 19F NMR chemical shifts of the esters
they form with 2a, 3, and 4 differ significantly.3,4 However,
there is no indication that similar analyses succeed when n >
2.5 The detection limits of HPLC analyses using chiral solid
phases are only slightly better. A support in which (aR)-1,1′-
bianthracene-2,2′-dicarboxylic acid is bound to silica gel
separates the enantiomers of 1 (n ) 3),6 but there is no indication
that it or any chiral support can when n > 3.
product whether nucleophiles displace the chlorine with retention
or inversion, and since nonracemic 6, which we called [5]-
HELOL, is easy to prepare in appreciable amounts,7 we tested
whether this reagent could be used to analyze the enantiomeric
purities of molecules whose centers of chirality are far from
their functional groups. The results reported here include the
discovery that when [5]HELOL chlorophosphite (5, X ) Cl) is
used as a chiral derivatizing agent (CDA), the range of alcohols
of structure 1 that can be analyzed is extended enormously:
from the previous limit, n ) 3, to n ) 7. The reagent can analyze
the enantiomeric purity of structure 7! It can distinguish the
enantiomers of a number of alcohols, amines, and, with the aid
We speculated that the sensitivity to remote chiral centers
could be increased if a 31P probe used for NMR analyses were
embedded in a large chiral groove, as in structure 5. In such a
(1) Recent reviews: (a) Okamoto, Y.; Yamamoto, C.; Yashima, E. Synlett
1998, 344. (b) Okamoto, Y.; Kaida, Y. J. Chromatogr. A 1994, 666, 403.
Books: (c) Chromatographic Chiral Separations; Zief, M., Crane, L. J.,
Eds.; Dekker: New York, 1988. (d) Recent AdVances in Chiral Separations;
Stevenson, D., Wilson, I. D., Eds.; Plenum: New York, 1990. (e) Lough,
W. J. Chiral Liquid Chromatography; Chapman and Hall: New York, 1989.
(f) Chiral Separations by Liquid Chromatography; Ahuja, S., Ed.; ACS
Symp. Ser. 471; American Chemical Society: Washington, DC, 1991.
(2) (a) Rothchild, R. Enantiomer 2000, in press. (b) Hulst, R.; Kellogg,
R. M.; Feringa, B. L. Recl. TraV. Chim. Pays-Bas 1995, 114, 115. (c) Parker,
D. Chem. ReV. 1991, 91, 1441. (d) Silks, L. A. P.; Wu, R.; Dunlap, R. B.;
Odom, J. D. Phosphorus, Sulfur, Silicon 1998, 136-138, 209.
(3) (a) Alexakis, A.; Mutti, S.; Mangeney, P. J. Org. Chem. 1992, 57,
1224. (b) Wu, R.; Odom, J. D.; Dunlap, R. B.; Silks, L. A., III
Tetrahedron: Asymmetry 1995, 6, 833.
(4) Takeuchi, Y.; Itoh, N.; Satoh, T.; Koizumi, T.; Yamaguchi, K. J.
Org. Chem. 1993, 58, 1812. These authors showed that reagent 4 is better
than the related one of Mosher, which is barely effective when n ) 2.
(5) In fact, 2b, which is a good probe of enantiomeric excess in other
cases, fails for 1 (n ) 2). See: Alexakis, A.; Frutos, J. C.; Mutti, S.;
Mangeney, P. J. Org. Chem. 1994, 59, 3326.
(6) Oi, S.; Ono, H.; Tanaka, H.; Shijo, M.; Miyano, S. J. Chromatogr.
A 1994, 679, 35.
(7) Dreher, S. D.; Katz, T. J.; Lam, K.-C.; Rheingold, A. L. J. Org. Chem.
2000, 65, 815.
10.1021/ja001904n CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/23/2000

10028 J. Am. Chem. Soc., Vol. 122, No. 41, 2000
Weix et al.
Table 1. Differences between the 31P Chemical Shifts (∆δ, in
ppm) of the Diastereomers 5 (X ) OR*) that Are Formed by
Alcohols CH3CHR(CH2)nOH, R ) Ph and Et (Solvents Are Either
CDCl3 or CD3CN)
of a linker not previously used for this purpose, carboxylic acids,
substances for whose analysisssurprisingly, considering how
common they aresonly a few CDAs have previously been
designed.8 In addition, we report how altering the solvents can
significantly improve the sensitivity of the analyses.
R ) Ph
R ) Et
n
in CDCl3
in CD3CN
in CDCl3
0
1
2
3
4
5
6
7
8
0.74
1.37
0.30
0.21
0.32
0.27
0.018b
0.027b
a
Results
0.087
0.43
0.20
0.043
a
Following Alexakis,5 the reagent for the analysis, 5 (X )
Cl), was prepared as needed. The procedure, which is a variant
of ones used for 1,1′-binaphthalene-2,2′-diol9 and for the amine-
precursors of 2a and 2b,5 combines 6 and a small amount of
4-(dimethylamino)pyridine (DMAP) with a solution of PCl3 in
CH2Cl2 and Et3N. The material to be analyzed is then added,
followed by CDCl3. The solvent can also be replaced and
analyses then carried out when the samples are dissolved in,
for example, CD3CN. The transformations cleanly gave 5 (X
) OR* or NHR*), and when the reactant was (()-2-methylbu-
tanol, the product (5 with X ) OCH2CHEtMe) was isolated,
purified by chromatography, and characterized. Like related
phosphite esters,3a,10 the [5]HELOL phosphites were difficult
to purify, presumably because they react quickly with water
and oxygen in the air. However, we could isolate a sample in
excellent yield (87%) whose 1H NMR spectrum showed
resonances for aromatic, methoxy, and alkyl protons in the
required intensity ratio and whose mass spectrum showed the
mass of the parent peak correctly. Moreover, the strongest peak
in its IR spectrum, at 906 cm-1, is characteristic of phosphite
esters,11 as are the next most intense peaks, at 731, 789, 738,
0.15
0.070
a Unresolvable in either solvent. b Not baseline resolved.
Table 2. Differences between the 31P Chemical Shifts (∆δ) of the
Diastereomers 5 (X ) NHR*) Formed by Amines
amine
∆δ (ppm)a
1-phenylethylamine
1-(1-naphthyl)ethylamine
sec-butylamine
2-phenylpropylamine
1-aminoindan
2-(2-aminoethyl)-1-methyl-pyrrolidine
0.77
0.30
0.81
0.38
2.45
0.91
a The solvent was CDCl3.
Table 3. Differences between the 31P Chemical Shifts (∆δ) of the
Diastereomers 5 (X ) OR*) Formed by Some Alcohols and
Phenols (R*OH)
and 835 cm-1 11,12 There are peaks at 1164 and 1213 cm-1
,
.
R*OH
∆δ (ppm)a
where aryl phosphites absorb,11,13 but their intensity is only
medium. The 31P NMR spectrum, analyzed with proton decou-
pling, consists of two peaks, at δ 131.97 and 131.88 ppm relative
to external 85% H3PO4, chemical shifts expected for the
phosphite structure.14 The ratio of their intensities was 0.98:1.
1-phenylbut-3-en-1-ol
1-indanol
4,4′-dimethylbenzoin
4-sec-butylphenol
vitamin E (8)
â-citronellol (3,7-dimethyl-6-octenol)
2.29
3.21
8.05
0.59
0.21
0.32
As summarized in Tables1-4, the ratios of the diastereomers
formed by the other materials analyzed could be determined
similarly by 31P NMR spectroscopy, but because the phosphorus
derivatives react with air and because the analyses were not
improved by it, the samples were not purified before the spectra
were recorded. In every one of more than 30 cases, the
diastereomeric excesses (de’s) equaled within 3% the enantio-
meric excesses (ee’s) of the materials analyzed.15 Even 4,4′-
dimethylbenzoin, a molecule that is crowded about the alcohol
a The solvent was CDCl3.
Table 4. Differences between the 31P Chemical Shifts (∆δ, in
ppm) of the Diastereomers 9 that Are Formed by Carboxylic Acids
CH3CHPh(CH2)nCO2H and How They Change with Solvent
n
in CDCl3
in other solvents
1
2
3
4
5
6
0.96
1.83
0.42
0.19a
0.076a
a
0.26b,c
0.16d
(8) (a) Reference 2c, p 1453, and references therein. (b) Peng, J.; Barr,
M. E.; Ashburn, D. A.; Lebioda, L.; Garber, A. R.; Martinez, R. A.; Odom,
J. D.; Dunlap, R. B.; Silks, L. A., III J. Org. Chem. 1995, 60, 5540. (c)
Brown, E.; Chevalier, C.; Huet, F.; Le Grumelec, C.; Le´ze´, A.; Touet, J.
Tetrahedron: Asymmetry 1994, 5, 1191. (d) Koos, M.; Mosher, H. S.
Tetrahedron 1993, 49, 1541, and references therein. (e) Heumann, A.; Loutfi,
A.; Ortiz, B. Tetrahedron: Asymmetry 1995, 6, 1073.
0.021d,e
a Unresolvable. b In Et2O with a very small amount of CDCl3 to
provide a lock signal. c ∆δ ) 0.11 when the solvent was CD3CN. d In
CD3CN. e Not baseline resolved.
(9) (a) Greene, N.; Kee, T. P. Synth. Commun. 1993, 23, 1651. (b)
Bredikhin, A. A.; Bredikhina, Z. A.; Nigmatzyanov, F. F. Russ. Chem. Bull.
1998, 47, 411.
(10) (a) Reference 2b, p 124. (b) Said, M. A.; Swamy, K. C. K.; Veith,
M.; Huch, V. J. Chem. Soc., Perkin Trans. 1 1995, 2945.
(11) Corbridge, D. E. C. In Topics in Phosphorus Chemistry; Grayson,
M., Griffith, E. J. Eds.; Interscience: New York, 1969; Vol 6, pp 278-
281.
(12) Daimay, L.-V.; Colthup, N. B.; Fateley, W. G.; Grasselli, J. G. The
Handbook of Infrared and Raman Characteristic Frequencies of Organic
Molecules; Academic Press: New York, 1991; pp 270-273.
(13) Shagidullin, R. R.; Chernova, A. V.; Vinogradova, V. S.; Mukhame-
tov, F. S. Atlas of IR Spectra of Organophosphorus Compounds; Kluwer:
Boston, 1990.
(14) (a) Koenig, T.; Habicher, W. D.; Hahner, U.; Piontech, J.; Rueger,
C.; Schwetlick, K. J. Prakt. Chem. 1992, 334, 333. (b) Gorenstein, D. G.;
Shah, D. O. In Phosphorus-31 NMR, Principles and Applications; Goren-
stein, D. G., Ed.; Academic Press: New York, 1984; pp 558-559.
group, gave its diastereomers without the amount of either being
noticeably enriched. If DMAP was omitted from the procedure,
the transformations were not clean, and the ratios in which the
diastereomers of 5 (X ) OR* or NHR*) were formed sometimes
did not equal the ratios of the enantiomers derivatized. All the
spectra show, in addition to the peaks for 5, a 31P resonance at
-1 ppm, attributable to the diarylphosphonate.14a The intensity
of this peak was small if the CDCl3 had been filtered through
basic alumina.
1
The instrument used for the analyses was a 300 MHz H
NMR spectrometer. Although it depends on the line widths,
the instrument resolved the 31P resonances of diastereomers to
(15) Three percent is approximately the accuracy of the NMR analyses.

[5]HELOL Phosphite
J. Am. Chem. Soc., Vol. 122, No. 41, 2000 10029
Figure 3. 31P NMRs of 9 [R*CO ) CH3CHPh(CH2)5CO] in (a) CDCl3
and (b) CD3CN.
Figure 1. 31P NMRs of 5 [X ) O(CH2)6CHPhCH3] in (a) ether with
a small amount of CDCl3 added to provide a frequency lock signal,
(b) CDCl3, and (c) CD3CN.
Figure 2. 31P NMRs of 5 [X ) O(CH2)7CHPhCH3] in (a) CDCl3 and
(b) CD3CN.
Scheme 1
Figure 4. 31P NMRs of 5 [X ) O(CH2)2CHMe(CH2)3CHdCMe2] in
CDCl3: (a) prepared from (R)-â-citronellol, (b) prepared from (S)-â-
citronellol, and (c) a mixture comprised of 19.4% of the (R)-isomer
and 80.6% of the (S)-isomer, that is, with an ee of 61.2%. The ee found
according to the spectra displayed is 60.5%.
the baseline when peaks were as little as 0.04 ppm apart. As
Table 1 shows, this means that the derivatives of 1 could be
analyzed in CDCl3 solution even when n ) 3, 4, or 5. Moreover,
as Table 1 and Figures1 and 2 show, when the solvent was
changed to CD3CN, the analysis could be extended to 1 even
when n ) 6 and 7.
The analyses were applied not only to alcohols CH3CHR-
(CH2)nOH (R ) Ph and C2H5, Table 1) but also to amines (Table
2), to various other alcohols, and to phenols such as 4-(2-butyl)-
phenol and vitamin E (8, Table 3). The application to phenols
in turn suggested a way to analyze carboxylic acids (Scheme
1): to couple them with a rigid linker, 2-aminophenol, and to
analyze the resulting 2-amidophenols, 9, like other phenols, after
they had been coupled to 5 (X ) Cl). Table 4 shows the
analyses. The carboxylic acids that were precursors to alcohols
1 could be analyzed in CDCl3 solution up to n ) 3 (5-
phenylhexanoic acid). However, as in the case of the alcohols,
the analyses could be extended to n ) 4 (6-phenylheptanoic
acid) and to n ) 5 (7-phenyloctanoic acid) when other solvents
were used: diethyl ether plus CDCl3 for the former and CD3-
CN for the latter. Figure 3 shows how the distinction between
the two diastereomers of 9 (R*CO ) 7-phenyloctanoyl)
improves when the solvent is changed from CDCl3 to CD3CN.
To test the ability of 5 to analyze the composition of mixtures
of enantiomers, samples of â-citronellol (3,7-dimethyl-6-octenol)
were analyzed. Figure 4 displays the 31P NMR spectra of the
phosphite esters of (M,M)-5 and each of the following: (S)-â-
citronellol, (R)-â-citronellol, and a mixture of the two whose

10030 J. Am. Chem. Soc., Vol. 122, No. 41, 2000
Weix et al.
it extends the ability to distinguish configurations of chiral
centers from the previous limit, n ) 3, to n ) 7. For the
methylalkanols Et(Me)CH(CH2)nOH, it extends the limit from
n ) 3 to n ) 4.20,21
It also analyzes carboxylic acids. Three points are significant
in this regard. The first is that reagent 5 is effective, for it is
one of very few reported for the analyses of chiral acids.8 Second
is that of the few previously studied reagents, only one has been
reported to distinguish chiral centers that are remote from the
carboxyl group.8b,22 Although the compounds tested before are
different from those tested here, Table 4 shows that reagent 5
distinguishes chiral centers that are even further from the
carboxyl than the furthest analyzed before. The situation is
similar when chromatographic methods are considered.23 Third
is that the linker used here, 2-aminophenol, or related linkers
have not previously been used as auxiliaries with other reagents
that analyze alcohols and amines. The linker is short and should
therefore keep the attached centers of chirality within or near
the groove of the [5]HELOL ligand.
The mechanism by which the phosphorus atom distinguishes
the chirality of the remote centers is not known, but the three-
point rule26 implies that the distinction, when for example
molecules 1 are bonded to 5, originates from the relationships
between both the remote phenyl and methyl groups and the
helical structure. A model in which O(CH2)8CH3 in its all-trans
extended conformation is bonded to 5 (at X) was constructed
by means of the Macromodel computer program. It shows that
at least the first three carbon atoms must lie within the
[5]HELOL groove. Substituents even on the fourth carbon must
interact with the [5]HELOL skeleton, and if the chain either
does not point straight out along [5]HELOL’s C2 axis or if it
folds back, substituents even further away will as well. If the
chain folds into the groove, the interactions within the HELOL
ligand can be extended considerably. Figure 5 suggests that in
the case of 1 when n ) 7, the benzene ring at the penultimate
carbon in the conformation of lowest energy lies within the
helicene cleft in both diastereomers, oriented with an edge
perpendicular to the terminal aromatic ring and near to that ring
and to the phosphorus. The pictured model appears reasonable
because benzene rings commonly adopt edge-to-face orienta-
tions.27,28,29 Figure 5 also implies that the terminal methyl group
of the alkoxy chain points away from the helicene structure not
Figure 5. Conformations of minimum energy of (M,M)-5 with X )
O(CH2)7CHPhMe when the side chain has (a) the (R)-configuration
and (b) the (S)-configuration. Hydrogens and methoxyl groups have
been deleted for clarity. Oxygen atoms are shown as speckled and
phosphorus as striped.
ee was 61.2 ( 0.1% [the (S)-isomer predominating].16 The
analyses show the (S)-â-citronellol to be >98% enantiomerically
pure, the (R)-â-citronellol to be >98% enantiomerically pure,
and the ee of the mixture to be 60.3 ( 2%. These data also
show that the enantiomeric purity of the [5]HELOL reagent was
g98%.
To analyze why 5 is effective in distinguishing the stereo-
chemistry of remote chiral centers, calculations were made to
find the conformations of lowest energy for the two diastere-
omers of 5 in which X was O(CH2)7CHPhMe. The Macromodel
V.6.0 program17 with AMBER* force field,18 operating on a
Silicon Graphics-O2 computer, was used for the conformational
searches. Initially, a structure of local minimum energy was
calculated for 5 (X ) O(CH2)7CHPhMe) in the selected solvent.
The structure was then subjected to a Metropolis Monte Carlo
conformational search using the standard parameters in Mac-
romodel. No parts of the molecule were constrained. The
conformations of lowest energy found for the two diastereomers
are displayed in Figure 5.
The results of these calculations showed that in the conforma-
tion of globally minimum energy, the (R) alkoxy chain lies
within the cleft of the helicene whether the solvent is water or
chloroform. When water was the selected solvent, all the other
conformations within 2 kcal/mol of the global minimum (there
were two) were also folded like this. However, when the solvent
was chloroform, 40% of the other conformations within 2 kcal/
mol of the global minimum (there were five) had the alkoxy
chain extended, with the phenyl and methyl groups outside the
cleft.19
(20) The reagent used for Et(Me)CH(CH2)3OH was R-cyano-R-fluo-
rophenylacetic acid.4
(21) Reagents 2a (in this experiment, NMe2 took the place of the pictured
Cl) and 2b are reported not to distinguish the enantiomers of 4-methylhex-
anol.3a,5 They do distinguish those of 2-methylbutanol, as does 33b and a
related reagent (Wu, R.; Odom, J. D.; Dunlap, R. B.; Silks, L. A., III
Tetrahedron: Asymmetry 1999, 10, 1465).
Discussion
(22) (a) Silks, L. A., III; Peng, J.; Odom, J. D.; Dunlap, R. B. J. Chem.
Soc., Perkin Trans. 1 1991, 2495. (b) Salvatore, B. A.; Smith, A. B., III
Tetrahedron Lett. 1994, 35, 1329.
The data in Tables 1-4 imply that [5]HELOL chlorophos-
phite (5, X ) Cl), presumably because it binds molecules within
a helical groove, is by far the most sensitive probe there is for
remote chirality. When combined with alcohols of structure 1,
(23) Chromatography on a chiral column was used to resolve lipoic acid.24
A binaphthyl column was used to separate the 3,5-dinitroanilides of the
4-phenylvaleric acids.25
(24) Fadnavis, N. W.; Koteshwar, K. Tetrahedron: Asymmetry 1997, 8,
337.
(25) Oi, S.; Ono, H.; Tanaka, H.; Matsuzaka, Y.; Miyano, S. J.
Chromatogr. A 1994, 659, 75.
(26) (a) Pirkle, W. H.; Pochapsky, T. C. Chem. ReV. 1989, 89, 347. (b)
Pirkle, W. H.; House, D. W. J. Org. Chem. 1979, 44, 1957. (c) Taylor, D.
R.; Maher, K. J. Chromatogr. Sci. 1992, 30, 67.
(16) Measured by the weights of the (R) and (S) materials that were
combined. The analyses of the (R) and (S) materials show that any slight
deviation from their enantiopurity cannot significantly change the calculated
ee.
(17) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440.
(27) Hunter, C. A. Chem. Soc. ReV. 1994, 23, 101.
(18) (a) AMBER: Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I.
R.; Merz, K. M., Jr.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell,
J. W.; Kollman, P. A. J. Am. Chem. Soc. 1995, 117, 5179. (b) AMBER* is
a variation used in the Macromodel program, ref 17.
(19) When the solvent was water, there were only two conformers within
4 kcal/mol of the global minimum, but when CHCl3, there were five within
2 kcal/mol of the global minimum and eight total within 3 kcal/mol.
(28) (a) Mu¨ller-Dethlefs, K.; Hobza, P. Chem. ReV. 2000, 100, 143 and
references therein. (b) Boyd, D. R.; Evans, T. A.; Jennings, W. B.; Malone,
J. F.; O’Sullivan, W.; Smith, A. J. Chem. Soc., Chem. Commun. 1996, 2269
and references therein.
(29) The reasons are unclear (see: Kim, E.-I.; Paliwal, S.; Wilcox, C.
S. J. Am. Chem. Soc. 1998, 120, 11192, references therein, and refs 27 and
28a), and molecular mechanics calculations do not account for them.27

[5]HELOL Phosphite
J. Am. Chem. Soc., Vol. 122, No. 41, 2000 10031
just in one of the diastereomers, but in both, and that the chain
folds to accommodate this requirement. The consequence is a
change in the orientation of the chain that, in turn, can alter the
angles at the phosphorus nucleus and therefore the phosphorus
chemical shift. Most importantly, Figure 5 shows that the alkoxy
group of 1 with n ) 7 does fit into 5’s cleft.
guish the enantiomers of 3-phenylbutanol, it cannot distinguish
the enantiomers of 4-phenylpentanol and can only partially
distinguish those of 3-methylpentanol. (The resolution is not to
the baseline.33) In contrast, 5 (X ) Cl) distinguishes the
enantiomers of all of them. The difference is likely a conse-
quence of the greater depth of [5]HELOL’s cleft.
In line with these thoughts, the effect that the change in
solvent from CDCl3 to CD3CN exerts on the 31P chemical shifts
of the diastereomers 5 [X ) O(CH2)6-7CHMePh] might be a
consequence of the latter, more polar solvent favoring more
compact conformations for these moleculessthat is, those in
which the alkyl chains lie within the cleft.30 The calculations
described above addressing this hypothesis indicate that when
the solvent is chloroform, while the lowest energy conformations
are similar to the ones in Figure 5, there are a few just slightly
higher in energy in which the chains are extended. When the
solvent is water, no such extended conformations are found to
have energies near that of the global minimum. However, when
an attempt was made to extend the analyses of alcohols HO-
(CH2)nCHMeEt to n ) 5 by changing the solvent from CDCl3
to CD3CN, only a single resonance was observed in the latter
solvent, and only a slightly split resonance in the former. While
it is unfortunate that the enantiomers of 6-methyloctanol could
not be analyzed, it would have been remarkable if the small
differences among an H, Me, and Et so remote from their point
of attachment to a reagent could be distinguished.
We have been unable to find data to compare the effects that
changes in solvents have on the 31P NMR chemical shifts of
other phosphite esters. However, in analyzing enantiomers by
derivatizing them with molecules similar to 3, Silks et al. studied
the effects solvents had on the 77Se chemical shifts of the
resulting selenones. They found that when the solvent’s polarity
increases, the shielding of the selenium nuclei increases and
the difference between the chemical shifts of the diastereomers
decreases. The data for the 31P resonances of the alcohol
derivatives studied here are exactly opposite. In the more polar
solvent (acetonitrile), the shielding of the 31P nuclei decreases,
and the difference between the chemical shifts of the diaster-
eomers increases (Figures 1 and 2). In the case of the derivatives
of the carboxylic acids (Table 4), the switch from CDCl3 to
CD3CN decreases ∆δ when n ) 4, but increases it when n )
5. In addition, the line widths became narrower in both cases
(the peak width at half-height when n ) 4 was 4.0 Hz; when n
) 5, it was 4.8 Hz). This last effect might be a consequence of
the viscosities decreasing.31 However, in the case when n ) 4,
contrary to these expectations, the lines were broader when the
solvent was Et2O, which is less viscous than CD3CN and
CDCl3.31 Also in this case ∆δ was larger, even though the
polarity was lower.
Conclusions
The chlorophosphite of [5]HELOL is by far the most sensitive
probe of remote chirality. Its preparation is easy, the analyses
are fast, and the reagent can be applied to a variety of
materials: alcohols, amines, and carboxylic acids. For these last,
2-aminophenol proved to be an effective linker that allowed
chiral centers remote from the acid function to be analyzed.
The experiments show that solvent effects can be used to extend
the range of the analyses.
Experimental Section
THF was distilled from sodium-benzophenone ketyl, CH2Cl2 and
Et3N from CaH2. CDCl3 was stored over 4 Å molecular sieves, and
just before it was used, it was filtered through basic alumina. PCl3 was
used as a 0.115 M solution in CH2Cl2. All materials were purchased
from Aldrich except for the basic alumina (Acros) and, unless noted,
were used without further purification. 1H NMR spectra were recorded
at either 300 or 400 MHz (TMS standard) and 13C NMR spectra at 75
MHz (CDCl3 standard). The usual procedure for determining 31P NMR
spectra was to record them at 121.5 MHz with continuous proton
decoupling. After 30° pulses (2.5 µs), the FIDs were acquired for 0.675
s. The sweep width was 200 ppm, centered at 80 ppm, and referenced
to zero for the resonance of 85% w/w H3PO4 in a sealed capillary tube.
The relaxation delay was 0.8 s. Zero-filling by a factor of at least 2
was applied before the Fourier transformation. No window function
was applied. The results were not noticeably different when inverse-
gated or no proton decoupling was applied or the relaxation delay was
extended. For the experiments in Figure 1, the acquisition time was
increased slightly, to 0.8 s.
Preparation of (M, M)-(-)-[5]HELOL. (M,M)-(-)-[5]HELOL (6)
was prepared following the published procedure.34 The enantiomeric
purity of its [5]helicene precursor was high. (M,M)-(-)-[5]HELOL (6)
was shown to be g98% enantiopure as follows. The de’s, according
to 31P NMR analyses (Figures 4a and 4b) of the derivatives 5 (X )
O(CH)2CHMe(CH2)2CHdCMe2) formed by combining 5 (X ) Cl,
prepared from 6 and PCl3 as described below) and (S)-(-)-â-citronellol
and (R)-(+)-â-citronellol, are >98%.
5 (X ) (S)-(-)-â-citronelloxy) 31P NMR: 129.2 ppm (S:R > 107:
1). 5 (X ) (R)-(+)-â-citronelloxy) 31P NMR: 128.8 ppm (S:R > 103:
1).
Preparation of Reagent 5 (X ) Cl) and Its Conversion into 5 (X
) OR* or NHR*). PCl3 (0.025 mmol as its solution in CH2Cl2),
followed immediately by Et3N (21 µL, 0.151 mmol), was added by
syringe to 20 mg (0.028 mmol) of 6 and a very small amount of DMAP,
contained along with a stir bar in a 1 dram vial that had been evacuated
and filled with N2. The resulting brownish yellow solution was stirred
for 10 min at 25 °C, and then the alcohol or amine (as a solution in
CH2Cl2 if it was a solid) was syringed in. The mixture was stirred for
2 h to allow the alcohol or amine time to react completely. The resulting
greenish solution, diluted with CDCl3, was pipetted into an NMR tube,
and the 31P spectrum was recorded.
If the peaks were insufficiently resolved, the solvent was changed.
Since decomposition seemed to occur if the solvent was evaporated at
this point, the solution was chromatographed on a short plug of silica
gel, which removes very polar impurities (Et3N‚HCl). The eluting
solvents were varying mixtures of CH2Cl2-hexanes or EtOAc-hexanes
Finally we note that Kolodiazhyni and co-workers have
reported a chiral derivatizing agent related to 5, dimenthyl
chlorophosphite, and showed how it can be used to analyze
protected amino acids.32 However, in a direct comparison with
5, we found that while dimenthyl chlorophosphite can distin-
(30) For changes in the conformations of aromatic molecules that occur
when the solvent is changed from chloroform to acetonitrile, see: Prince,
R. B.; Saven, J. G.; Wolynes, P. G.; Moore, J. S. J. Am. Chem. Soc. 1999,
121, 3114 and references therein.
(31) The viscosities of ether (0.224 P at 25 °C) and acetonitrile (0.369)
are both lower than that of chloroform (0.537) (CRC Handbook of Chemistry
and Physics, 73rd ed.; Lide, D. R., Ed.; CRC: Ann Arbor, MI, 1992).
(32) Kolodiazhnyi, O. I.; Demchuk, O. M.; Gerschkovich, A. A.
Tetrahedron: Asymmetry 1999, 10, 1729. They also report that the chemical
shift difference is even larger in the spectra of the corresponding oxide or
sulfide (it is unclear which). Usually chemical shift differences are smaller
in the derivatives of P(V) than in the corresponding ones of P(III). See:
refs 2b and 3a.
(33) The spectra are displayed in the Supporting Information.
(34) While ref 7 states correctly on p 820 that 82 g of KOH was used to
prepare the dehydro-derivative of 6, it states incorrectly that this amount is
55 mmol. It is 1.46 mol. A salt-ice bath kept the temperature of the reaction
mixtures at 0 ( 2 °C.

10032 J. Am. Chem. Soc., Vol. 122, No. 41, 2000
Weix et al.
that were chosen on the basis of TLC analyses. The column can be
very short because impurities such as 6 (yellow, as is 5) and 6’s
oxidation product (dark blue) need not be separated. The solvent could
then be evaporated and replaced. If necessary, the evaporation of solvent
and its replacement could then be repeated.
1.53 (s, 0.5 H), 1.41-1.33 (m, 1H), 1.28-1.26 (m, 1H), 1.19-1.12
(m, 0.5H), 0.45-0.15 (series of m, 5H), 0.09-0.00 (m and d under
the TMS resonance, 2H), -0.15 (d, 6.5 Hz, 1H) ppm. 31P NMR (122
MHz, external 85% w/w H3PO4 standard): 132.0 and 131.9 ppm. IR
(CCl4): 1213, 1164, 906, 835, 789, 738, 731 cm-1. HRMS (FAB, M
+ 1): m/z calcd for C53H43O7P 822.2746, found 822.2750.
The 31P NMR chemical shifts (in ppm) of the products were as
follows. Table 1, n ) 0, R ) Ph: 143.18, 142.44; Table 1, n ) 1, R
) Ph: 128.51, 127.14; Table 1, n ) 2, R ) Ph: 131.51, 131.20; Table
1, n ) 3, R ) Ph: 128.01, 127.80; Table 1, n ) 4, R ) Ph: 128.76,
128.45; Table 1, n ) 5, R ) Ph: 129.40, 129.13; Table 1, n ) 6, R )
Ph, CDCl3: 129.16, 129.14, CD3CN: 135.40, 135.25; Table 1, n ) 7,
R ) Ph, CDCl3: 129.18, 129.16, CD3CN: 135.24, 135.17; Table 1, n
) 8, R ) Ph, CDCl3: 128.75, 128.74, CD3CN: 135.36; Table 1, n )
1, R ) Et: 132.19, 132.11; Table 1, n ) 2, R ) Et: 131.69, 131.25;
Table 1, n ) 3, R ) Et: 128.98, 128.79; Table 1, n ) 4, R ) Et:
129.04, 129.00; Table 1, n ) 5, R ) Et: 129.20; Table 2, entry 1:
150.44, 149.68; Table 2, entry 2: 150.26, 149.96; Table 2, entry 3:
149.21, 148.41; Table 2, entry 4: 143.20, 142.82; Table 2, entry 5:
148.48, 146.03; Table 2, entry 6: 145.04, 144.13; Table 3, entry 1:
146.07, 143.78; Table 3, entry 2: 140.36, 137.15; Table 3, entry 3:
149.80, 141.75; Table 3, entry 4: 137.08, 136.67; Table 3, entry 5:
137.53, 137.32; Table 3, entry 6: 129.15, 128.83; Table 4, n ) 1, R )
Ph: 134.06, 133.10; Table 4, n ) 2, R ) Ph: 132.33, 130.50; Table
4, n ) 3, R ) Ph: 132.87, 132.45; Table 4, n ) 4, R ) Ph, CDCl3:
132.63, 132.45, diethyl ether (with a small amount of CDCl3 to provide
a lock signal): 134.23, 133.97, CD3CN: 131.54, 131.43; Table 4, n )
5, R ) Ph, CDCl3: 133.00, 132.93, CD3CN: 133.32, 133.16; Table 4,
n ) 6, R ) Ph, CDCl3: 133.21, CD3CN: 133.08, 133.06.
Condensation of Acids with 2-Aminophenol. The acid (1 mol)
was stirred under N2 for 30 min with oxalyl chloride (11 mmol). The
excess oxalyl chloride was removed in vacuo, and the residue, dissolved
in CH2Cl2 (3.3 mL), was added by cannula to a solution, cooled to 0
°C, of 2-aminophenol (1.5 mmol) and pyridine (37 mmol) in CH2Cl2
(8.6 mL). The mixture was stirred at 0 °C for 1 h and at 25 °C for 2
h. After 1 M HCl had been added, the product was extracted with
EtOAc. The EtOAc solution was washed with brine and dried (Na2-
SO4). Removal of the solvent left a brown viscous oil or wax, which
after flash chromatography, eluting with 50-60% EtOAc-50-40%
1
hexanes, yielded a yellow-orange oil whose H NMR and 13C NMR
showed only the expected resonances.
Acknowledgment. We thank the NSF for supporting the
research (CHE98-02316) and D.J.W. (under its REU program).
We are grateful also for the support D.J.W. received under
Columbia University’s Rabi Scholar’s Program. We thank Stella
Wu for purifying some of the compounds and Dr. John Decatur
for helping us to analyze the NMR spectra.
Supporting Information Available: Preparations of alco-
hols and acids that could not be purchased, 31P NMR spectra
of all alcohols, amines, and acids analyzed and of the dimenthyl
chlorophosphite derivatives discussed in the text (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
Preparation of 5 (X ) OCH2CHEtMe). 5 was prepared as
described above except that all quantities were doubled. The product
was chromatographed (flash chromatography, eluting with 1:1 CH2-
Cl2-hexanes), yielding 27.0 mg (87% yield) of a bright yellow powder.
1H NMR (300 MHz, CDCl3): 8.30-8.26 (m, 2H), 7.97-7.78 (series
of m, 4H), 7.70-7.51 (series of m, 8H), 7.26 (d, 6.4 Hz, 2H), 7.19-
7.07 (m, 2H), 6.91-6.86 (m, 2H), 4.26 (d, 5.0 Hz, 6H), 4.17 (s, 6H),
JA001904N