Assignment of the Absolute Configuration of α-Chiral Carboxylic Acids by H NMR Spectroscopy

Article abstract of DOI:10.1021/jo9916838

The prediction of the absolute configuration of α-chiral carboxylic acids from the ¹H NMR spectra of their esters with (R)- and (S)-ethyl 2-hydroxy-2-(9-anthryl) acetate [(R)- and (S)-9-AHA, 5] is discussed. Low-temperature NMR experiments, MM, semiempirical, and aromatic shielding effect calculations allowed the identification of the main conformers and showed that, in all esters studied, conformer ap is the most stable. A simple model for the assignment of the absolute configuration from NMR data is presented, and its reliability is corroborated with acids 6-31 of known absolute configuration. In addition to 5, other auxiliary reagents with open (32-38) and cyclic (39-42) structures have also been studied. trans-(+)- and (-)-2-phenyl-1-cyclohexanol (41) was found to be particularly efficient and produced Δδ^RS values similar to those of 5.

Full text of DOI:10.1021/jo9916838

2658
J . Org. Chem. 2000, 65, 2658-2666
Assign m en t of th e Absolu te Con figu r a tion of r-Ch ir a l Ca r boxylic
1
Acid s by H NMR Sp ectr oscop y
Mar´ıa J . Ferreiro, Shamil K. Latypov,† Emilio Quin˜oa´, and Ricardo Riguera*
Departamento de Quı´mica Orga´nica, Facultad de Quı´mica, Universidad de Santiago de Compostela,
15706 Santiago de Compostela, Spain
Received October 26, 1999
1
The prediction of the absolute configuration of R-chiral carboxylic acids from the H NMR spectra
of their esters with (R)- and (S)-ethyl 2-hydroxy-2-(9-anthryl) acetate [(R)- and (S)-9-AHA, 5] is
discussed. Low-temperature NMR experiments, MM, semiempirical, and aromatic shielding effect
calculations allowed the identification of the main conformers and showed that, in all esters studied,
conformer ap is the most stable. A simple model for the assignment of the absolute configuration
from NMR data is presented, and its reliability is corroborated with acids 6-31 of known absolute
configuration. In addition to 5, other auxiliary reagents with open (32-38) and cyclic (39-42)
structures have also been studied. trans-(+)- and (-)-2-phenyl-1-cyclohexanol (41) was found to be
particularly efficient and produced ∆δRS values similar to those of 5.
In tr od u ction
can be obtained from the difference in the chemical shifts4
of L1/L2 measured as ∆δRS
.
Although the determination of absolute configuration
by NMR spectroscopy is widely used in the cases of
secondary alcohols and primary R-substituted amines,1
the application of this technique to other functional
groups has been scarcely investigated or is not fully
reliable.
The method consists of the derivatization of the
substrate with the (R)- and the (S)-enantiomer of a chiral
auxiliary reagent usually containing an aryl ring (i.e.,
arylmethoxyacetic acids, AMAAs) that directs its aniso-
tropic cone selectively toward one of the substituents L1/
L2 of the asymmetric center of the substrate (Figure 1).
The chemical shifts observed for the L1/L2 groups of the
amine or alcohol under investigation reflect their spatial
relationship with respect to the aryl ring, and therefore,
the absolute configuration at the chiral center of the
substrate can be correlated to that of the auxiliary
reagent by means of the NMR spectra.
This phenomenon depends on the existence of a
preferred conformation in the ester or amide derivatives
(the same for the R and the S derivative), where the aryl
ring is oriented to produce shielding selectively on just
one of the substituents (L1/L2) of the amine/alcohol part.
The conformational characteristics and the role of the
aryl ring (ring current and orientation) are well under-
stood in the case of the AMAA2 derivatives of secondary
alcohols2a-h and primary R-substituted amines.2i,k The
esters of AMAAs with secondary alcohols are dominated
by two main conformers3 (ap and sp); in each one the aryl
group selectively shields one of the substituents (L1 or
L2) of the alcohol part. The average spectra can be related
to the structure of conformer sp, which is the predomi-
nant one, and the assignment of the R/ S stereochemistry
We have recently demonstrated the application of this
methodology to other substrates. In the case of â-chiral
primary alcohols,5 extensive energy calculations and
conformational analysis of their 9-anthrylmethoxyacetic
acid (9-AMA) esters were necessary to find the NMR-
significant conformer and the model that correlates their
absolute configuration with their 1H NMR spectra.
Nevertheless, a note of caution was stressed regarding
the uncertainty of the configuration deduced by applica-
tion of this method to substrates where the functional
group serving as a handle to the auxiliary reagent is more
than one bond away from the asymmetric center.
Carboxylic acids are also interesting substrates be-
cause they are frequently found in nature as optically
active compounds and there are almost no general ap-
proaches for the determination of their stereochemistry.
A few reports that correlate their absolute configuration
with the NMR spectra of certain derivatives have been
published in the past few years.6 PGDA (1) and PGME
(2) (a) Trost, B. M.; O’Krongly, D.; Belletire, J . L. J . Am. Chem. Soc.
1980, 102, 7595-7596. (b) Seco, J . M.; Latypov, Sh. K.; Quin˜oa´, E.;
Riguera, R. Tetrahedron Lett. 1994, 35, 2921-2924. (c) Latypov, Sh.
K.; Seco, J . M.; Quin˜oa´, E.; Riguera, R. J . Org. Chem. 1995, 60, 504-
515. (d) Seco, J . M.; Latypov, Sh. K.; Quin˜oa´, E.; Riguera, R.
Tetrahedron: Asymmetry 1995, 6, 107-110. (e) Latypov, Sh. K.; Seco,
J . M.; Quin˜oa´, E.; Riguera, R. J . Org. Chem. 1996, 61, 8569-8577. (f)
Kouda, K.; Kusumi, T.; Ping, X.; Kan, Y.; Hashimoto, T.; Asakawa, Y.
Tetrahedron Lett. 1996, 37, 4541-4544. (g) Seco, J . M.; Latypov, Sh.
K.; Quin˜oa´, E.; Riguera, R. Tetrahedron 1997, 53, 8541-8564. (h)
Latypov, Sh. K.; Seco, J . M.; Quin˜oa´, E.; Riguera, R. J . Am. Chem.
Soc. 1998, 120, 877-882. (i) Trost, B. M.; Bunt, R. C.; Pulley, Sh. R.
J . Org. Chem. 1994, 59, 4202-4205. (j) Latypov, Sh. K.; Seco, J . M.;
Quin˜oa´, E.; Riguera, R. J . Org. Chem. 1995, 60, 1538-1545. (k) Seco,
J . M.; Latypov, Sh. K.; Quin˜oa´, E.; Riguera, R. J . Org. Chem. 1997,
62, 7569-7574.
(3) The sp and ap conformers are defined by the synperiplanar and
antiperiplanar arrangements of the CR-OMe and CdO groups,
respectively. See ref 2c
† Present address: The Institute of Organic and Physical Chemistry
of the Russian Academy of Sciences, Kazan, 4200.1383 Tatarstan,
Russian Federation.
(1) For a review on the use of NMR for assignment of absolute
configuration and ee measurements, see: Uray, G. Houben-Weyl:
Methods in Organic Chemistry; Helmchen, G. R., Hoffmann, W.,
Mulzer, J ., Schaumann, E., Eds.; Thieme: Stuttgart, New York, 1996;
Vol. 1, p 253.
(4) ∆δRS represents the difference between the chemical shift of
the same group of the substrate in the (R)- and (S)-derivative (∆δRS
δR - δS).
)
(5) (a) Ferreiro, M. J .; Latypov, Sh. K.; Quin˜oa´, E.; Riguera, R.
Tetrahedron: Asymmetry 1996, 7, 2195-2198. (b) Latypov, Sh. K.;
Ferreiro, M. J .; Quin˜oa´, E.; Riguera, R. J . Am. Chem. Soc. 1998, 120,
4741-4751.
10.1021/jo9916838 CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/07/2000

Absolute Configuration of R-Chiral Carboxylic Acids
J . Org. Chem., Vol. 65, No. 9, 2000 2659
F igu r e 1.
(2) were proposed by Kusumi et al.,6a,d mandelic ester
(MA, 3) by Tyrrell et al.,6b and 1,1′-binaphthalene-8,8′-
diol [BNDO (4)] was used by Fukushi et al.6c
In this paper we present our results on the use of
alcohol 5 [ethyl 2-hydroxy-2-(9-anthryl) acetate, 9-AHA]
as a reagent for the determination of the absolute
configuration of R-chiral carboxylic acids by comparison
of the NMR spectra of their esters with (R) and (S)-5.
MM8 and AM1/PM39 methods were used to calculate
the geometry and relative stability of the conformers;
DNMR10 was employed to prove experimentally the
theoretical findings; and finally, aromatic shielding effect
calculations11 on each conformer and on the equilibrium
mixture were carried out and the results were compared
with the experimental NMR shielding. A preliminary
account of this study has already been published.12 In
the last part of this article, the usefulness of aryl alcohols
other than 5 is also described, with particular emphasis
on the cyclohexane derivatives 41-45 and the role of the
aryl ring in the effectiveness of these reagents.
Nevertheless, the ∆δ values reported are very small,
and the range of compounds of known absolute configu-
ration used to test the stereochemical predictions is, in
most cases, very limited. Moreover, no real understand-
ing of the factors that determine the experimental results
has been advanced; the conformations used to explain
the NMR data obtained with 1-4 are simply empirical,
and in fact, they could be substituted by other stereo-
chemical dispositions that explain the results equally
well.
Resu lts a n d Discu ssion
Con for m a tion a l Com p osition a n d P r efer en ce.
The main issue in relation to the potential use of chiral
aryl alcohols such as 9-AHA (5) as reagents for the
assignment of absolute configuration of carboxylic acids
is to demonstrate the existence of an NMR-significant
conformational preference that is the same in all of the
esters and independent of the nature of the carboxylic
acid.
To this end, we carried out MM and semiempirical
calculations on a series of more than 20 R-substituted
carboxylic acids (Figure 2). On the basis of previous
studies,2c it is known that the alcohol part of an ester
adopts a preferred conformation where H-CR′-O-CdO
are in a plane and the CR′H is in an sp disposition in
relation to the CdO bond (Figure 3a). Therefore, we
focused our attention on the rotation around the CR-CO
bond (Figure 3b), the most important process in these
compounds.
An important decision that has to be taken into account
in the development of new reagents concerns the selection
of the bond employed to link the auxiliary reagent to the
substrate (carboxylic acid). In the reports discussed
above, amide (PGDA and PGME)6a,d and ester (MA and
BNDO)6b,c bonds have been used. However, we reasoned
that linking the carboxylic acid to the auxiliary reagent
through an ester bond should be preferred to an amide
because the equilibrium around the CR-CO bond in
amides has been shown to be highly dependent on the
structure of the carboxylic acid (i.e., the nature of the
aryl ring in the case of AMAA amides),2j,k significantly
limiting the generality of the results. Moreover, the
COCdO skeletal fragment shows a clear preference for
the Z conformation, while the conformational distribution
of CNHCdO is much more complex, with Z/E ratios
ranging from 90:10 to 50:50.7 The mobility around the
CR-CO bond and the absence of any strong conforma-
tional preference about the NH-CO bond of amides
strongly suggest that less generality could be expected
in the use of those reagents6a,d,e apart from the com-
pounds tested and those with very closely related struc-
tures. The higher reliability of esters led us to concentrate
our efforts on aryl alcohols rather than arylamines as
chiral derivatizing reagents (CDA) for R-chiral carboxylic
acids.
The calculations revealed a consistent conformational
composition over the whole series, with only two main
(8) Maple, J . R.; Dinur, U.; Hagler, A. T. Proc. Natl. Acad. Sci. U.S.A.
1988, 85, 5350-5354.
(9) (a) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J . J .
P. J . Am. Chem. Soc. 1985, 107, 3902-3909. (b) Havel, T. F. Prog.
Biophys. Mol. Biol. 1991, 56, 43-78. (c) Stewart, J . J . P. J . Comput.
Chem. 1989, 10, 209-220. (d) Komorinski, A.; McIver, W., J r. J . Am.
Chem. Soc. 1973, 95, 4512-4517.
(10) (a) Drakenberg, T.; Forsen, S. J . Phys. Chem. 1972, 76, 3582-
3586. (b) Grindley, T. B. Tetrahedron Lett. 1982, 23, 1757-1760.
(11) (a) Waugh, J . S.; Fessenden, R. W. J . Am. Chem. Soc. 1957,
79, 846-849. (b) J ohnson, C. E.; Bovey, F. A. J . Chem. Phys. 1958, 29,
1012-1014. (c) J onathan, N.; Gordon, S.; Dailey, B. P. J . Chem. Phys.
1962, 36, 2443-2448. (d) Dailey, B. P. J . Chem. Phys. 1964, 41, 2304-
2310. (e) Musher, J . I. J . Chem. Phys. 1965, 42, 4081-4083. (f) Haigh,
C. W.; Mallion, R. B. Ring current theories in NMR. In Progress in
NMR spectroscopy; Pergamon Press: Elmsford, NY, 1980; pp 303-
344.
(6) (a) Nagai, Y.; Kusumi, T. Tetrahedron Lett. 1995, 36, 1853-1856.
(b) Tyrrell, E.; Tsang, M. W. H.; Skinner, G. A.; Fawcett, J . Tetrahedron
1996, 52, 9841-9852. (c) Fukushi, Y.; Shigematsu, K.; Mizutani, J .;
Tahara, S. Tetrahedron Lett. 1996, 37, 4737-4740. (d) Kusumi, T.;
Yabuuchi, T.; Ooi, T. Chirality 1997, 9, 550-555. (e) Hoye, T. R.;
Koltun, D. O. J . Am. Chem. Soc. 1998, 120, 4638-4643.
(7) (a) Stewart, W. E.; Siddall, T. H. Chem. Rev. 1970, 70, 517-
551. (b) J acobus. J .; J ones, T. B. J . Am. Chem. Soc., 1970, 92, 4583-
4585.
(12) Ferreiro, M. J .; Latypov, Sh. K.; Quin˜oa´, E.; Riguera, R.
Tetrahedron: Asymmetry 1997, 8, 1015-1018.

2660 J . Org. Chem., Vol. 65, No. 9, 2000
Ferreiro et al.
F igu r e 2. R-Chiral carboxylic acids studied by theoretical calculations.
Ta ble 1. Con for m a tion a l P r efer en ce of r-Ch ir a l
Ca r boxylic Acid s 6-29 (RCO2H), Th eir Meth yl Ester s
(RCO2Me), a n d N-Meth yl Am id es (RCONHMe)a
RCO2H
RCO2Me
RCONHMe
MM PM3
acid MM AM1
PM3
MM
0.53
PM3
0.69
6
7
8
0.64
1.13
1.14
0.85
1.41
1.31
0.42
1.42
0.91
0.08
0.83
1.35
0.75
0.86
1.28
1.22
1.12
0.97
1.23
0.94
0.93
0.37
1.69
2.11
0.37
0.45
0.42
0.31
0.57
0.63
0.56
0.54
0.00
0.00
0.10
0.25
0.36
0.22
0.76
0.75
1.48
0.08
0.32
0.00
0.46
0.37
0.76
0.44
0.54
-0.91
0.04
0.63
0.71
0.79
0.52
0.31
0.75
0.30
0.49 -0.09
0.40 -0.40
0.34 -1.98
0.99 -0.91b
0.99
0.73
1.23
1.05
1.24
1.25
1.29
0.24
0.12
0.80
0.94
1.02
0.95
0.94
0.95
0.31
0.14
0.33
1.09
0.88
0.15
0.73
0.68
0.87
0.02
0.63
0.70
1.17
0.63
1.04
9
0.39
0.42
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
0.77 -0.41
0.48 -0.13
0.79 -0.01
0.90 -0.43
0.88 -1.59
-0.12 -1.40
0.37 -2.51
0.16 -0.10
0.29
0.94
0.73
0.91
0.65
0.62
0.70
0.00
0.30
0.00
0.99
0.56
0.86
1.13
0.72
0.56
1.01
0.91
0.84
0.91
1.01
0.73
0.62
0.14
1.37
1.83
0.62
0.21
0.26
0.98
0.09
2.37
1.96
0.15
-0.05 -0.41
0.24
0.34
0.30
1.73
0.07 -2.60
0.09
0.98
0.13
2.23
c
0.13
0.42
1.78
0.16
1.09
a
Energies (kcal/mol) are relative to the ap conformer, as defined
in Figure 3b. A positive value means that sp is higher in energy,
b
whereas a negative one corresponds to lower energy states cf.
AM1 data -0.52. c sp nonstable.
F igu r e 3. (a) Preferred conformation of the ester of ethyl
2-hydroxy-2-(9-anthryl) acetate (5) with an R-chiral carboxylic
acid. (b) Main orientations in the acid fragment. (c) Low-energy
orientation of the aromatic ring.
main conformers, sp and ap, was estimated by semiem-
pirical methods (AM1 and PM3, Table 1) and showed a
preference of approximately 1 kcal/mol for the ap over
the sp conformer for all of the acids investigated. This
means that, in the most stable conformer, CR-H and
CR′-H are antiperiplanar. When L1 or L2 are polar groups
(i.e.,alkoxy in 15-17), some deviation from planarity is
observed and the angle becomes smaller than 180° as a
result of the tendency of the alkoxy group to become
orientations of the CR-H bond relative to the CdO bond
(Figure 3b): the ap conformation where the CR-H bond
is antiperiplanar in relation to the CdO bond and the
sp conformation where these two bonds are in a syn-
periplanar disposition. The energy gap between these two

Absolute Configuration of R-Chiral Carboxylic Acids
J . Org. Chem., Vol. 65, No. 9, 2000 2661
F igu r e 4. Conformational equilibrium in the esters of (R)- and (S)-9-AHA (5) with an R-chiral carboxylic acid.
closer to the CdO bond. Fortunately, this effect is small
and does not represent a significant modification of the
general disposition of L1/L2 in relation to the aryl group.
MM and PM3 calculations were also carried out on the
methyl esters of acids 6-29, and the same geometry and
similar energy differences as in the free acids were found
(Table 1). Furthermore, calculations on the esters of acid
(S)-6 with alcohols (R)- and (S)-5 corroborated this
conformational composition with a 0.5-0.9 kcal/mol
preference for ap over sp (Figure 4).13
When the same calculations (MM and PM3, Table 1)
were carried out on the methyl amides of acids 6-29,
the absence of a common conformational preference was
found. In fact, rotation around the CR-CO bond was
found to be preferentially sp in 12 out of the 24 amides
investigated and preferentially ap in the other 12 cases.
This dependence of the conformational composition on
the nature of the acid confirms our previous observations2j,k
that alcohols should be preferred to amines as auxiliary
reagents for carboxylic acids.
F igu r e 5. Conformational equilibrium in the esters of (R)-
and (S)-9-AHA (5) with (S)-2-methylbutyric acid (6). Energy
calculated by AM1.
Because we used an ester derivative, the preferred
conformation has the OR group synperiplanar (sp) to the
CdO bond and the aryl ring is coplanar to the CR′-H
bond14 (Figure 3c), a situation that permits an effective
transmission of the shielding effect to the substituents
(L1/L2) of the acid.
NMR Resu lts. Once the existence of an unequivocal
conformational preference identical for all esters of R-
carboxylic acids with chiral secondary aryl alcohols and
the favorable orientation of the aryl ring with regards
to substituents L1/L2 was theoretically well established,
the next step was to find the necessary experimental
evidence.
All of these data allow us to assume that the confor-
mational composition of any ester of 5 with an R-chiral
carboxylic acid is as represented in Figure 4. Thus, the
L1 substituent in the (R)-9-AHA ester should be shielded
in the ap conformation, while the L2 substituent will
remain unaffected (Figure 4). In contrast, in the sp
conformation it is L2 that is shielded, while L1 remains
unaffected. In the case of the (S)-9-AHA ester, it is L2
that is shielded in the ap conformation and L1 in the sp
conformation. Because the population of the ap conformer
is higher than that of the sp conformer, L1 will be more
shielded in the (R)-9-AHA ester than in the (S)-derivative
(∆δRS < 0), and conversely, L2 will be more shielded in
the (S)-9-AHA ester than in the (R)-ester (∆δRS > 0). The
signs of the ∆δRS values will allow the assignment of the
To that end, we selected for our study the aryl alcohol
(R)- and (S)-9-AHA (5) and tested its ability to separate
1
the signals of (S)-2-methylbutyric acid (6). The H NMR
spectra of the resulting diastereomeric (R)- and (S)-9-
AHA esters showed clear differences in the chemical
shifts of the protons located around the asymmetric
center of the acid (L1/L2). The ethyl group in the (R)-ester
resonates at higher field than in the (S)-9-AHA ester,
whereas the signal for the methyl group moves in the
opposite sense and is observed at higher field in the (S)-
9-AHA ester in comparison to the (R)-9-AHA ester.
Therefore, the ethyl group corresponds to L1, and the
methyl group corresponds to L2 (Figure 5). This means
that if we compare the chemical shift of these groups in
both esters, the difference, ∆δRS, is positive (+0.10 ppm)
for the methyl protons and negative (-0.16, -0.10, and
-0.09 ppm) for the ethyl group in accordance with the S
absolute configuration of the acid 6. These results are
fully coherent with the conformational composition and
1
spatial position of L1/L2 by comparison of the H NMR
spectra of both diastereoisomers.
(13) (a) For the (R)-9-AHA ester of 6: DE ) 0.93 (AM1), 0.93 (PM3).
(b) For the (S)-9-AHA ester of 6: DE ) 0.48 (AM1), 0.39 (PM3).
(14) To maintain consistency with previous works, CR′ refers to the
carbon bearing the hydroxy group in the alcohol moiety, although it is
C2 in the systematic name. In those works, the sp conformer was
defined by the synperiplanar disposition of the CROMe and CdO
groups.

2662 J . Org. Chem., Vol. 65, No. 9, 2000
Ferreiro et al.
Ta ble 2. 1H NMR Ch em ica l Sh ifts of (R)- a n d (S)-9-AHA
(5) Ester s of (S)-2-Meth ylbu tyr ic Acid (6) a t Differ en t
Tem p er a tu r es
Ta ble 4. Ca lcu la ted Ar om a tic Sh ield in g Effects for th e
Ester s of 5 a n d 41 w ith (S)-2-m eth ylbu tyr ic Acid (6)
config
conf
Me(5)
Me(4)
T, K
CRH
CH2-CH3
(R)-AHA
1.663/1.418
1.655/1.410
1.649/1.405
1.641/1.397
1.628/1.386
CH2-CH3
CH3(5)
CDA 5
R
S
sp
ap
sp
ap
0.89
-0.20
-0.17
0.95
-0.14
0.10
0.30
300
263
233
203
173
2.427
2.430
2.433
2.438
2.443
0.802
0.800
0.799
0.796
0.790
1.191
1.199
1.204
1.212
1.218
-0.04
CDA 41
R
S
sp
ap
sp
ap
0.40
1.01
0.92
0.04
0.64
0.09
0.08
0.36
(S)-AHA
1.728/1.532
1.732/1.566
300
193
2.430
2.446
0.950
0.980
1.091
1.065
Ta ble 3. 1H NMR Ch em ica l Sh ifts of (R)- a n d (S)-9-AHA
(5) Ester s of N-t-Boc-L-Va lin e16 (19) a t Differ en t
Tem p er a tu r es
tal values and confirm that the most important contri-
butions to the average NMR spectra come from the ap
conformer (Table 4).
T, K
CRH
CH(3)
CH3(4)
CH3(5)
Full experimental support for this interpretation of the
DNMR data and a demonstration that the conformational
characteristics of the (R)- and (S)-9-AHA esters of 6 and
19 are general for other substrates were obtained after
examination of many other carboxylic acids17 (6-31) of
known absolute configuration and with widely varied
structures.
The structures and ∆δRS values are shown in Figure 6
and serve to illustrate two important facts. First, the
distribution of the signs of ∆δRS is perfectly homogeneous
along the series18 and consistent with the absolute
configuration of these compounds and with the confor-
mational composition shown in Figure 4. Second, the data
also show that reagent 9-AHA (5) is much better than
other reagents (i.e.,1-3) because it produces ∆δRS values
that are much higher (see ref 6) and its influence is
exerted over much longer distances.
(R)-AHA
1.982
1.974
300
243
193
4.267
4.304
4.375
0.816
0.799
0.754
0.603
0.522
0.397
1.956
(S)-AHA
2.322
2.393
300
243
193
4.167
4.178
4.200
1.001
1.016
1.022
0.962
0.970
0.978
2.459
preference for the ap form derived from the calculations
(Figure 5).
Additional support for these findings was obtained
from low-temperature NMR experiments on the esters
of acid (S)-6 with (R)- and (S)-5 (Table 2), where an
increase in the shielding was observed in accordance with
the greater population of the preferred conformer ap.
Thus, in the NMR spectrum of the ester of (S)-6 with
(R)-5 the ethyl group resonances were shifted to higher
field (∆δT1,T2 ) 0.022CH2 and 0.006CH3 ppm)15 when the
probe temperature was decreased from 300 to 203 K,
whereas in the ester of (S)-6 with (S)-5 it was the methyl
Mod el for Assign m en t of th e Absolu te Con figu -
r a tion of Ca r boxylic Acid s. As we have shown, the
sign of the ∆δRS values of the esters of 9-AHA (5) is an
indicator of the spatial location of L1/L2 relative to the
aryl group and therefore allows the determination of the
R/S absolute configuration of the carboxylic acid. Experi-
mentally, this entails the preparation of the (R)- and (S)-
9-AHA esters of the acid of unknown configuration and
group signal that was shifted to higher field (∆δT1,T2
)
0.026 ppm) when the temperature was decreased to 193
K. These shifts indicate beyond doubt that the ethyl
group is located on the same side as the aryl ring in the
most stable ap conformer of the ester of (S)-6 with (R)-5,
whereas in the (S)-5 ester this side is occupied by the
methyl group. Signal broadening was not observed,
meaning that the rotation around the CR-CO bond
continues to be rapid, and this explains the small ∆δRS
values obtained.
Analogous behavior was observed in the esters of (R)-
and (S)-9-AHA with acid (S)-19 (Table 3) where larger
shielding of the isopropyl group of its ester with (R)-9-
AHA was observed at lower temperatures.
Moreover, the NMR signals due to the CRH proton of
(R)- and (S)-9-AHA esters of 6 and 19 are shifted to lower
field at low temperatures (see Tables 2 and 3). This effect
is perfectly consistent with the equilibrium represented
in Figure 4 and can be explained by the increase in the
population at lower temperatures of a conformer in which
the CRH is out of the shielding cone of the carbonyl group.
This condition is fulfilled only by the ap conformer, which
is the most stable.
1
comparison of their H NMR spectra.
This information should be interpreted in the light of
the equilibrium and structures shown in Figure 4.
Nevertheless, as the average spectrum is dominated by
conformer ap, the equilibrium can be simplified by
considering this form (Figure 7) as the only relevant one
and by using it to fix the spatial location of L1 and L2.
According to this model, the substituent that is more
shielded in the (R)-9-AHA ester than in the (S)-9-AHA
ester (∆δRS < 0) should be ascribed to L1 in the figure,
and conversely, the substituent that is more shielded
in the (S)-9-AHA ester than in the (R)-9-AHA ester
(∆δRS > 0) corresponds to L2. Identification of L1/L2 and
their placement around the asymmetric center can then
be directly obtained from the signs of ∆δRS as shown in
Figure 7.
(16) Some indication of coalescence in the 1H NMR spectra was
observed at ca. T ) 243 K. Only data for the major form are given in
the table. Assignment of the minor component is ambiguous because
of its very low population.
The calculated aromatic shielding effects for the ap and
sp conformers of the esters of (R)- and (S)-5 with (S)-2-
methylbutyric acid (6) compare well with the experimen-
(17) (R)-O-Phenyllactic acid was synthesized from (S)-ethyl lactate;
see: Tottie, L.; Baeckstrom, P.; Moberg, C.; Tegenfeldt, J .; Heumann,
A. J . Org. Chem. 1992, 57, 6579-6587.
(18) The 1H NMR data of the NHBoc group are very sensitive
to experimental conditions and should not be used to derive the
configuration (see ref 2h).
(15) ∆δT1,T2 represents the difference between the δ value (ppm) at
the higher (T1) and lower (T2) temperature [∆δT1,T2 ) δ(T1) - δ(T2)].

Absolute Configuration of R-Chiral Carboxylic Acids
J . Org. Chem., Vol. 65, No. 9, 2000 2663
F igu r e 6. ∆δRS values in the esters of 9-AHA (5) with the acids shown.
F igu r e 7. Conformational model for the assignment of the
absolute configuration of R-chiral carboxylic acids by NMR
spectroscopy.
F igu r e 9.
roughly the same as if the more likely simplified model
shown in Figure 8b were considered.
Oth er Ar yl Alcoh ols a s Au xilia r y Rea gen ts. Ac-
cording to earlier observations in AMAA esters,2c,d the
conformational equilibrium around the CR-CO bond is
practically independent of the nature of the alcohol. This
means that it is reasonable to expect that other secondary
aryl alcohols would behave as 9-AHA and could therefore
potentially be used as auxiliary reagents for NMR
configuration assignment of carboxylic acids.
To test this idea we decided to explore the usefulness
of other alcohols with different substituents, skeletal
structures, and aryl rings as chiral reagents. To this end,
(S)-2-methylbutyric acid (6) was selected as the substrate,
and the ∆δRS values of its esters with alcohols 32-42
(Figure 9) were analyzed as before.
F igu r e 8. Simplified model for arylalkoxyacetic acid esters
of the acid 6 according to (a) ref 6b (MA) and (b) our data
(9-AHA).
In contrast to the equilibrium and structures described
in Figure 4 for the esters of 9-AHA (5), the results re-
ported using MA (3) as the reagent have been interpreted6b
assuming a completely different preferred conformation
(CR′-H not coplanar with CdO and L1/L2 cis to the CdO
group, see Figure 8a). This model is, however, inconsis-
tent with the well documented2c coplanarity of CR′-H
with CdO and with our calculations (Table 1) and NMR
data (Figure 6), and indeed, no specific experimental data
has been advanced to support it. Notwithstanding, the
application of this model has yielded correct R/S assign-
ments because those two errors cancel each other, and
the spatial location of L1/L2 relative to the aryl ring is
The results presented in Table 5 show that the ∆δRS
are, in most cases, smaller than those obtained with
9-AHA (5). Among the open chain 1-aryl alcohols exam-
ined, only two, 3719 and 38, produced shifts comparable
to those of 5, demonstrating once again2b,g the superiority

2664 J . Org. Chem., Vol. 65, No. 9, 2000
Ta ble 5. ∆δRS Va lu es for th e Ester s of Ar yl Alcoh ols 32-42 w ith (S)-2-Meth ylbu tyr ic Acid (6)
Ferreiro et al.
32a ,b
33a a
34a a
35a ,b
36a ,b
37a a
38a ,b
39a a
40a a
6c,d
42a a
6a ,b
Me(5)
Me(4)
-0.30
+0.06
-0.03
0.07
0.04
0.05
-0.03
+0.07
-0.03
+0.07
0.09
0.18
+0.10
-0.16
0.01
0.02
0.00
0.01
-0.10
+0.15
0.04
0.06
+0.10
-0.16
a
These compounds were used in racemic form, and ∆δ values were measured from the spectra of the mixture.
F igu r e 10. ∆δRS values for the esters of the carboxylic acids shown with trans-2-phenyl-1-cyclohexanol (41) (lightface) and 9-AHA
(5) (bold ).
of the anthryl ring in this kind of reagent. The cyclic
analogues 39 and 40 produced almost no separation of
signals, and a similar result was obtained with the
cyclohexane derivative 42.20 However, the commercially
available (1R,2S)- and (1S,2R)-2-phenyl-1-cyclohexanol
(41) gave ∆δRS values21 very close to those observed with
9-AHA (5), although this reagent bears a phenyl rather
than an anthryl ring.
Corroboration of the effectiveness of (1R,2S)- and
(1S,2R)-2-phenyl-1-cyclohexanol (41) as a reagent was
obtained after esterification with the R-chiral carboxylic
acids of known absolute configuration shown in Figure
10. The data shown can be summarized in the following
conclusions: (a) The signs of the shifts produced by 41
are regularly distributed in all of the compounds exam-
ined, so that ∆δRS values are positive for protons in L2
and negative for protons in L1. (b) The signs obtained are
opposite to those obtained with 5 (positive for L1 and
negative for L2). (c) The ∆δRS values obtained with 41
F igu r e 11. (a) trans-2-Aryl-1-cyclohexanols and (b) |∆δRS
|
are, in terms of absolute values, similar to those produced
by 5 for protons close to the chiral center. These facts
indicate that both 41 and 5 show basically the same
ability to separate the resonances of the enantiomeric
values for the esters of the acids 6, 12, and 19 with alcohols
41 (lightface), 43 (bold ), 44 (italic) and 45 (underlined).
protons of the substrates. Aromatic shielding effects have
been calculated for the ap and sp conformations of the
esters of (S)-2-methylbutyric acid (6) with reagents 41
and 5, and the results of these calculations are also
similar (Table 4).
To study the role of the aryl ring in 41 and to explore
the possibility of increasing the ∆δRS values by changing
the aromatic ring, the 1-naphthyl (43), 2-naphthyl (44),
and 9-anthryl (45) derivatives (Figure 11) were synthe-
(19) Racemic 1-(9-anthryl)-1-ethanol (37) was obtained by reduction
of 9-acetylanthracene with NaBH4 in MeOH.
(20) Transformation of the trans-alcohol (41) into the diastereomeric
cis-alcohol (42) was carried out by the Mitsunobu reaction; see:
Takahashi, M.; Ogasawara, K. Tetrahedron: Asymmetry 1995, 6,
1617-1620.
(21) For a given proton of the acid, ∆δRS represents the difference
between its chemical shift in the (1R,2S)-derivative and that in the
(1S,2R)-derivative.

Absolute Configuration of R-Chiral Carboxylic Acids
J . Org. Chem., Vol. 65, No. 9, 2000 2665
F igu r e 12. Frontal and side views of the calculated minimum energy conformers of the (a) trans-2-phenyl-1-cyclohexanol and
(b) trans-2-anthryl-1-cyclohexanol esters of 2-methylpropionic acid.
sized from the corresponding arylbromides and cyclohex-
ene oxide,22 and their ability to produce selective shifts
was tested with acids 6, 12, and 19, of known absolute
configuration. The results shown in Figure 11 indicate
that no significant improvement in ∆δRS values is as-
sociated with the presence of naphthyl or anthryl rings.
MM calculations showed why trans-2-aryl-1-cyclo-
hexanols are more qualified than their cis counterparts
as CDA reagents for acids. In both cases, the chair
conformers with an equatorial aryl group proved to be
F igu r e 13. Schematic representation of the direction of
maximun shielding in the esters of (a) 5 and (b) 41.
the most stable ones, but in the trans isomers only the
the direction of maximum shielding points to positions 4
aromatic ring is placed to cause maximum shielding
or 5 bonds away from the asymmetric center,2g whereas
effect on the acid substituents (equatorial-equatorial
in the esters of 41, the maximum shielding is focused on
relationship). Furthermore, the aryl rings were roughly
the immediate surroundings of the asymmetric carbon
coplanar with the C(2′)H bond, with the fragment CR′-
(Figure 13). Thus, if the phenyl group of 41 is replaced
HOCOCRH adopting the same conformation as in the
by a larger aromatic system, as in 43-45, the area of
shielding becomes very much larger. This should now
affect not only the substituent on the same side (L2) but
esters of 5 (CRH-CO and CR′H coplanar and ap as in
Figure 7). In the case of 43-45, the only modification
observed when compared with 41 was a small rotation
also to some extent the other substituent (L1), and so on
of the aryl plane, which departs by about 10° from
coplanarity with C(2′)H. Figure 12 shows frontal and side
average the ∆δRS values will not be as high as expected.
In the case of 5, the focus of the anisotropic ring is so far
views of the calculated minimum energy conformers of
away that even if its effect is greatly increased, it will
the trans-2-phenyl-1-cyclohexanol and trans-2-anthryl-
never affect more than one substituent in each conformer.
Figure 14 shows a graphic summary that illustrates
how the absolute configuration of an R-chiral carboxylic
acid can be deduced from the signs of the experimental
1-cyclohexanol esters of 2-methylpropionic acid where the
above structural features can be observed.
The fact that 43-45, with naphthyl and anthryl rings,
do not produce higher shifts than 41, which has a
∆δ values in their esters with any trans-2-aryl-1-cyclo-
practically identical conformation, seems to be in conflict
hexanol.
with data obtained for 9-AMA esters. Nevertheless,
inspection of the structures of esters of 5 indicates that
Con clu sion s
In summary, we have demonstrated that the esters of
R-substituted carboxylic acids with aryl secondary alco-
hols such as 9-AHA (5) are composed of two conformers
in equilibrium, with the CHR′-O-C(dO)-CHR system in
the same plane and the form defined by an anti relation-
ship between CHR and CdO groups being predominant.
(22) (a) Eis, M. J .; Wrobel, J . E.; Ganem, B. J . Am. Chem. Soc. 1984,
106, 3693-3694. (b) Mizuno, M.; Kanai, M.; Iida, A.; Tomioka, K.
Tetrahedron 1997, 53, 10699-10708. Compounds 43-45 were used
in their racemic form. The ∆δRS values of the esters 19e, 12f, 19f, 12g,
and 19g were measured after HPLC separation, whereas those of 6e,
12e, 6f, and 6g were directly measured from the spectra of the
corresponding mixture.

2666 J . Org. Chem., Vol. 65, No. 9, 2000
Ferreiro et al.
F igu r e 14. Conformational model for the determination of the absolute configuration of R-chiral carboxylic acids by NMR of
their trans-2-phenyl-1-cyclohexanol esters. ∆δRS defined as in ref 21.
recorded on 500, 300, or 250 MHz NMR spectrometers.
Chemical shifts (ppm) are internally referenced to the tetra-
methylsilane signal (0 ppm) in all cases. One- and two-
dimensional NMR spectra were measured with standard pulse
sequences. 2D Homo- (COSY) and heteronuclear (HMQC) shift
correlation experiments were carried out using pulsed field
gradient technique. Apodization with a shifted sine bell and
baseline correction was implemented to process 2D spectra.
1D 1H NMR spectra: size 32 K, pulse length 2.8 ms (30°),
16 acquisitions. 2D COSY spectra: sequence D1-90-t1-G1-90-
G2-AQ, relaxation delay D1 ) 1 s, 90° pulse 8.5 µs, gradient
ratio 1:1. 2D TOCSY spectra: relaxation delay D1 ) 2 s;
mixing time 41.3 ms; 90° pulse 8.5 µs; TPPI-mode, NS ) 64.
2D Proton-detected heteronuclear multiple quantum correla-
tion (HMQC) experiments: sequence D190(1H)-D2-90(13C)-t1/
2-G1-180(1H)-G2-t1/2-90(13C)-G3-D2-AQ (GARP(13C)), relax-
ation delay D1 ) 2s, D2 ) 3.45 ms, 90° pulse (1H) 8.5 µs; 90°
pulse (13C) 10.5 µs, gradient ratio 5:3:4. For DNMR spectros-
copy, the probe temperature was controlled by a standard unit
calibrated using a methanol reference; samples were allowed
to equilibrate for 15 min at each temperature before recording
spectra.
The shielding produced by the aryl ring is selectively
directed to the substituent of the acid located on the same
side of the plane, and this shielding allows a correlation
to be established between the absolute configuration of
the alcohol and that of the acid. In practice, the esters of
the acid with the both enantiomers of ethyl 2-hydroxy-
2-(9-anthryl) acetate (5) have to be prepared, their 1H
NMR spectra compared, and the sign of ∆δRS interpreted
according to the model shown in Figure 7.
In addition to 5, other aryl alcohols with open (32-
38) and cyclic (39-42) structures have also been studied
as auxiliary reagents, and (1R,2S)- and (1S,2R)-2-phenyl-
1-cyclohexanol (41) was found to be particularly efficient,
producing ∆δRS values similar to those of 5. Because of
the geometry of compound 41, replacement of its phenyl
ring by naphthyl or anthryl (43-45) does not produce a
significant improvement in its capability to separate the
signals of the substrate. Thus, both 5 and 41 are reagents
of similar effectiveness for the assignment of absolute
configuration of R-chiral carboxylic acids by 1H NMR
spectroscopy. However, 41 produces much more compli-
Gen er a l. (R)- and (S)-ethyl 2-hydroxy-2-(9-anthryl)acetate
(5) were obtained by asymmetric reduction of ethyl (9-anthryl)-
glyoxylate with (R)- and (S)-ALPINE BORANE, respectively,
(prepared in situ from 9-BBN and (+)- and (-)-pinene by the
usual procedure).24 It was methylated to measure the ee by
HPLC (Chiral column) and then hydrolyzed to compare the
stereochemistry with that of 9-AMA. The absolute stereochem-
istry was confirmed by CD. (R): yield ) 85%, ee (OMe) )
1
cated H NMR spectra as a result of the presence of the
cyclohexane signals, which may overlap significant pro-
tons of the substrate, and its effect on hydrogens placed
at long distances from the chiral center is weaker than
the effect generated by 5.
90.36, [R]D ) -75.9. (S): yield ) 80%, ee (OMe) ) 60.6, [R]D
+50.9.
)
Exp er im en ta l Section
Com p u ta tion a l Meth od s. Molecular mechanics (employ-
ing the PC91 force field8) and AM19 (PM3) calculations were
performed using the Insight II package on a Silicon Graphics
Iris (SGI) computer. Initial molecular geometries were origi-
nated from the Builder Module of Insight II; 3D coordinates
were then generated from the bond lengths, bond angles, and
dihedral angles using the DG-II package.23 The conformational
space of each compound was scanned by MM optimization of
the sterically allowed conformations around key single bonds.
The MM simulations were carried out in vacuo. Analysis of
conformational transitions, identification of the low-energy
conformers, and calculation of the energy barriers between
these conformers were all carried out by MM. The energies of
conformations were minimized in Cartesian coordinate space
by the block diagonal Newton-Raphson method; minima
corresponded to rms energy gradients less than 0.001 kcal/
mol Å. The ground-state energies of the geometries were then
calculated by AM1 (PM3) using the MOPAC 6.0 program. For
all compounds, full geometry optimization used the Broyden-
Fletcher-Goldfarb-Shanno (BFGS) method and the PRECISE
option.9
Preparation of diastereomeric esters from the corresponding
carboxylic acids was carried out with DCC and DMAP in
CH2Cl2.25 The reaction mixture was filtered and purified by
flash chromatography on silica gel. Final purification was
achieved by HPLC (µ-Porasil, 3 mm × 250 mm or Spherisorb
S5W 5 µm, hexanes-ethyl acetate).
Ack n ow led gm en t. This work was financially sup-
ported by grants from CICYT (PM98-0227) and from
the Xunta de Galicia (XUGA-20908B97 and XUGA:
PGIDT99PXI20906B). S.K.L. acknowledges the Spanish
Ministry for Education and Science for a postdoctoral
research grant. M.J .F. thanks the Xunta de Galicia for
a grant.
Su p p or tin g In for m a tion Ava ila ble: Experimental data
(HPLC, NMR, MS, etc.) relative to compounds 6a ,b, 7a ,b, 9a ,b,
11a ,b, 12a ,b, 13a ,b, 14a ,b, 15a ,b, 16a ,b, 30a ,b, 17a ,b, 18a ,b,
19a ,b, 21a ,b, 25a ,b, 31a ,b, 32a ,b, 33a , 34a , 35a ,b, 36a ,b, 37,
37a , 38a ,b, 39a , 40a , 6c,d , 42, 42a , 7c,d , 9c,d , 11c,d , 12c,
11d , 14c,d , 30c,d , 17c,d , 19c,d , 21c,d , 31c,d , 25c,d , 43, 6e,
12e, 19e, 44, 6f, 12f, 19f, 45, 6g, 12g, and 19g. This material
is available free of charge via the Internet at http://pubs.acs.org.
Sh ield in g Effect. Calculations were carried out on a SGI
computer using a program (writen on Fortran 77) based on
the semiclassical model of Bovey and J ohnson.11 No corrections
for local anisotropic contributions11d,e were implemented.
Calculations were performed with π-current loops separation
of 1.39 Å.11b,f
J O9916838
1
NMR Sp ectr oscop y. H and 13C NMR spectra of samples
(24) Brown, H. C.; Pai, G. G. J . Org. Chem. 1985, 50, 1384-1394.
(25) Trost, B. M.; Belletire, J . L.; Godleski, S.; McDougal, P. G.;
Balkovec, J . M.; Baldwin, J . J .; Christy, M.; Ponticello, G. S.; Varga,
S. L.; Springer, J . P. J . Org. Chem. 1986, 51, 2370-2374.
in 4:1 CS2/CD2Cl2 or CDCl3 (ca. 2-3 mg in 0.5 mL) were
(23) Cioslowski, J .; Kertesz, M. QCPE Bull. 1987, 7, 159.