Pure by NMR ?

Article abstract of DOI:10.1021/ol802211p

(Chemical Equation Presented) Integration of a ¹³C-¹H satellite peak of a given ¹²C-¹H parent resonance within a quantitative ¹H NMR spectrum and comparison to the minor component represents a simple protocol for the accurate determination of diastereoisomeric ratios of up to 1000:1 (i.e., 99.8% de).

Full text of DOI:10.1021/ol802211p

ORGANIC
LETTERS
2008
Vol. 10, No. 23
5433-5436
“Pure by NMR”?
Timothy D. W. Claridge, Stephen G. Davies,* Mario E. C. Polywka,
Paul M. Roberts, Angela J. Russell, Edward D. Savory, and Andrew D. Smith
Department of Chemistry, Chemistry Research Laboratory, UniVersity of Oxford,
Mansfield Road, Oxford OX1 3TA, U.K.
steVe.daVies@chem.ox.ac.uk
Received September 22, 2008
ABSTRACT
1
1
Integration of a 13C- H satellite peak of a given 12C- H parent resonance within a quantitative 1H NMR spectrum and comparison to the minor
component represents a simple protocol for the accurate determination of diastereoisomeric ratios of up to 1000:1 (i.e., 99.8% de).
The determination of ratios of components in a mixture
is an everyday task within organic chemistry, which is
usually tackled by the application of 1H NMR spectroscopy
of quantitative NMR techniques have been introduced in
order to improve the precision with which the minor
components in a mixture can be quantified,2 including doping
of samples with a reference standard.3 However, signals in
1H NMR spectra corresponding to protons directly bound to
1
or HPLC analysis. H NMR spectroscopy is arguably the
most widely used tool for this analysis, as examination of
the relative integration of resonances rapidly allows access
1
(2) Hollis, D. P. Anal. Chem. 1963, 35, 1682. Thakur, K. A. M.; Kean,
R. T.; Hall, E. S. Anal. Chem. 1997, 69, 4303. Forshed, J.; Erlandsson, B.;
Jacobsson, S. P. Anal. Chim. Acta 2005, 552, 160. For reviews, see: (a)
Holzgrabe, U.; Diehl, B. W. K.; Wawer, I. J. Pharm. Biomed. Anal. 1998,
17, 557. (b) Malz, F. NMR Spectroscopy in Pharmaceutical Analysis;
Holzgrabe, U., Wawer, I., Diehl, B., Eds.; Elsevier: New York, 2008; pp
43-62. (c) Holzgrabe, U. NMR Spectroscopy in Pharmaceutical Analysis;
Holzgrabe, U., Wawer, I., Diehl, B., Eds.; Elsevier: New York, 2008; pp
131-137. (d) Wells, R. J.; Cheung, J.; Hook, J. M. NMR Spectroscopy in
Pharmaceutical Analysis; Holzgrabe, U., Wawer, I., Diehl, B., Eds.;
Elsevier: New York, 2008; pp 291-315.
to product ratios. When applied to routine analysis of H
NMR spectra, the term “pure by NMR” is generally
understood to indicate a purity (diastereoisomeric or other-
wise) at the level of 95% or above, although the accuracy to
which such figures are reported varies widely.1 A number
(1) For examples, see: Bray, B. L.; Muchowski, J. M. Can. J. Chem.
1990, 68, 1305. MacKeith, R. A.; McCague, R.; Olivo, H. F.; Palmer, C. F.;
Roberts, S. M. J. Chem. Soc., Perkin Trans. 1 1993, 313. Wimmer, R.;
Mu¨ller, N. Monatsh. Chem. 1998, 129, 1161. Jain, R. P.; Williams, R. M.
Tetrahedron Lett. 2001, 42, 4437. Hutin, M.; Bernardinelli, G.; Nitschke,
J. R. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 17655.
(3) O’Neill, I. K.; Pringuer, M. A.; Prosser, H. J. J. Pharm. Pharmacol.
1975, 27, 222. Lacroix, P. M.; Dawson, B. A.; Sears, R. W.; Black, D. B.;
Cyr, T. D.; Ethier, J.-C. J. Pharm. Biomed. Anal. 1998, 18, 383. Malz, F.;
Jancke, H. J. Pharm. Biomed. Anal. 2005, 38, 813.
10.1021/ol802211p CCC: $40.75
Published on Web 11/01/2008
2008 American Chemical Society

1
Figure 1. 500 MHz H NMR spectrum of a 1:1 mixture of oxazolidinones 1 and 2, recorded in CDCl3 at ambient temperature [signals in
black correspond to resonances attributable to both diastereoisomers].
carbon have two satellite peaks (one on either side) of the
“parent” 12C-1H peak due to coupling to the 1.108% of
naturally occurring, NMR-active carbon isotope 13C (for
which I ) 1/2). This produces a ratio of 178.5:1 for each
13C-1H satellite peak versus the corresponding 12C-1H
parent resonance. Provided that the resonances of interest
are well resolved and the signal-to-noise ratio in the sample
is acceptable (>10:1 for the smallest signal),4 a satellite
resonance may be used as an intrinsic internal standard for
accurate quantification of minor components of a mixture.4,5
For this procedure to represent a general method for the
accurate determination of very high diastereoisomeric ratios
or purities, there are a number of specific experimental
parameters that must be considered. In this manuscript we
describe a simple experimental NMR procedure to facilitate
quantitative analysis of the 13C-1H satellite resonances of a
given 12C-1H parent resonance of the major component in
a 1H NMR spectrum, which allows diastereoisomeric ratios
of up to 1000:1 (i.e., 99.8% de) to be accurately assessed;
this procedure is equally applicable to analysis of product
ratios or purity.
proton acquisition. Nevertheless, for the accurate quantifica-
tion of an NMR spectrum it is necessary to avoid (dif-
ferential) saturation of resonances caused by too rapid signal
averaging and this requires the complete equilibration of
magnetization between each transient. The recovery times
are dictated by the longitudinal relaxation time constants (T1
values) of the NMR nuclei and delays of 5T1 are required to
re-establish essentially complete recovery (>99%) after full
excitation of the NMR response. In the context of this
methodology, it is important to note that a 13C nucleus acts
as an efficient relaxation source for a directly attached proton,
unlike the NMR inactive 12C nucleus, implying that recovery
rates for the 13C satellites will be significantly faster, and
hence T1 values shorter, than for the parent 12C resonance.
Thus, recycle delays must be dictated by the relaxation time
The methodology follows the well-understood principles
for achieving quantitative NMR analysis.6 Although 13C-
decoupling during proton observation has been recommended
as a means of simplifying proton spectra for quantification
purposes,6b,7 the use of the 13C-1H satellites as an internal
reference is advantageous as it requires the employment of
only the simpler, standard technique of single-pulse, direct
Table 1. Proton Longitudinal Relaxation Time Constants (T1
Values) for the C(5)H Resonances of 1 and 2, where the Proton
is Attached to Either 12C or 13C
compound
δ
12C-1H (ppm) T1 12C-1H (s) T1 13C-1H (s)
1
Figure 2
.
Partial quantitative 500 MHz H NMR spectrum for 1,
showing the region containing the C(5)H resonances. The asterisk
denotes low level spectrometer receiver artifacts.
1
2
5.607
4.929
3.10
3.49
1.84
1.97
5434
Org. Lett., Vol. 10, No. 23, 2008

1
Figure 3. Partial quantitative 500 MHz H NMR spectra for diastereoisomeric mixtures of 1 and 2, showing the region containing the
C(5)H resonances. The asterix denotes low level spectrometer receiver artifacts.
1
constant of the slowest relaxing parent 12C resonances; it is
this that would usually be measured when determining T1
values with, for example, the well-known inversion-recovery
method.8 Failure to avoid saturation effects will lead to a
relative enhancement of the 13C satellite resonance intensities
and will result in an overestimation of the diastereoisomeric
or product ratio. For many small organic molecules studied
in (nondegassed) low-viscosity solvents such as chloroform,
proton T1 values are below 5 s, dictating recycle delays of
up to 25 s. The use of conservative recycle delays should
avoid the need for direct T1 measurement on every sample
analyzed. While relaxation reagents can be used to reduce
these requirements, the addition of these to precious samples
is often undesirable, and may lead to unacceptable line-
broadening.
For H NMR spectra recorded under these experimental
conditions, it is apparent that when the height of the 13C-1H
satellite resonances associated with the major diastereoisomer
are greater than that of the 12C-1H resonance of the minor
diastereoisomer, the diastereoisomeric ratio is >180:1 (>98.9%
de), thus allowing an approximate, yet ready, quantitative
assessment of the diastereoisomeric ratio by direct inspection
1
of the H NMR spectrum.
As a model system to demonstrate the effectiveness and
ease with which this protocol may be applied, the known
oxazolidinones 1 and 2, derived from (1R,2S)-ephedrine and
(1S,2S)-pseudoephedrine, respectively, were prepared and
recrystallized twice from CH2Cl2/hexane.9 Stock solutions
of 1 (1.675 g in 25 mL of CDCl3) and 2 (685 mg in 10 mL
1
of CDCl3) were prepared. H NMR spectroscopic analysis
To fully resolve the 13C satellites from the parent
resonance, its line shape must be sufficiently narrow at the
baseline, a condition that is usually met on modern, well-
shimmed spectrometers. In this context gradient shimming
routines can be advantageous in the avoidance of low-level
broadening or asymmetrical humps that may arise from
inadequate shim optimization. It is also likely to be beneficial
to avoid sample rotation and eliminate the possible appear-
ance of undesirable “spinning sidebands” either side of the
parent resonances.6b Furthermore, it is advantageous to
consider measurements derived from the simplest and
narrowest available multiplet structures such that the satellite
resonances remain resolved: singlets are to be preferred.
Likewise, the use of only moderate sensitivity enhancement
window (apodization) functions when processing the FID is
recommended such that the base of the parent resonance is
not made so broad that it extends to the satellites. In situations
when the signal-to-noise ratio of the minor isomer resonance
is sufficient, such enhancement functions may be avoided
altogether. Spectra should also be baseline corrected, at least
in the region of the peaks of interest, prior to integration.
of a 1:1 mixture of diastereoisomers 1 and 2 (prepared from
the stock solutions) indicated excellent resonance dispersion
in CDCl3, with the resonances due to C(4)H (1H, dq), C(5)H
(1H, d), and C(4)Me (3H, d) being well resolved within each
diastereoisomer, and from each other (Figure 1).
The inversion-recovery sequence was used to determine
T1 values for the doublet resonances associated with C(5)H,
for both the parent 12C-1H resonances and their correspond-
1
ing 13C-1H satellites, within both 1 and 2. Fourteen H
aquisitions were collected, with inversion recovery delays
ranging from 0-20 s, and the individual recovery curves
(4) Maniara, G.; Rajamoorthu, K.; Rajan, S.; Stockton, G. W. Anal.
Chem. 1998, 70, 4921.
(5) Dewey, R. S.; Schoenewaldt, E. F.; Joshua, H.; Paleveda, W. J., Jr.;
Schwam, H.; Barkemeyer, H.; Arison, B. H.; Veber, D. F.; Denkewalter,
R. G.; Hirschmann, R. J. Am. Chem. Soc. 1968, 90, 3254. Boate, A. R.;
Eaton, D. R. Can. J. Chem. 1976, 54, 3895. Brown, S. L.; Davies, S. G.;
Foster, D. F.; Seeman, J. I.; Warner, P. Tetrahedron Lett. 1986, 27, 623.
Fux, P. Analyst 1990, 115, 179. Vedejs, E.; Wang, J.; Lin, S.; Klapars, A.
U. S. Patent 5,900,427, 1999. Al-Deen, T. S.; Hibbert, D. B.; Hook, J. M.;
Wells, R. J. Anal. Chim. Acta 2002, 474, 125. Wells, R. J.; Hook, J. M.;
Al-Deen, T. S.; Hibbert, D. B. J. Agric. Food Chem. 2002, 50, 3366.
Forshed, J.; Andersson, F. O.; Jacobsson, S. V. J. Pharm. Biomed. Anal.
2005, 29, 495. For a review see ref 2.
Org. Lett., Vol. 10, No. 23, 2008
5435

were fitted to extract T1 values.10 These data demonstrated
significant differences between the values for the parent and
satellite resonances, as well as differences in values between
the diastereoisomers 1 and 2 (Table 1). Based on the T1
values, recovery delays of 20 s were adopted for all
subsequent analysis of diastereoisomers 1 and 2 by quantita-
tereoisomeric ratios was noted for the 10:1, 100:1, and 1000:1
mixtures (Table 2). In the case of the 10,000:1 mixture, the
Table 2. Correlation between Calculated and Experimentally
Observed Diastereoisomeric Excess Values for Solutions of 1
and 2 in CDCl3
1
tive H NMR.
1
The diastereoisomeric purity of 1 was determined by H
ratio calculated
relative
integration ratio
s/n
observed
de
NMR to be >10,000:1 (>99.98% de), since 2 could not be
detected under conditions of high signal-to-noise ratio (s/n
>70:1) for the 13C-1H satellites of 1 (Figure 2).
1:2
de
peak
1000:1
99.80
1 (13C-1H)
2 (12C-1H)
1 (13C-1H)
2 (12C-1H)
1 (13C-1H)
2 (12C-1H)
1.0000
0.1603
1.0000
1.7696
1.0000
18.3561
162
11
99.82
98.04
81.35
In order to investigate the levels at which the minor
diastereoisomeric component could be accurately determined
using this technique, diastereoisomeric mixtures of 1 and 2
in 10:1, 100:1, 1000:1, and 10,000:1 ratios, respectively, were
prepared by serial dilution of the stock solution of 2.
100:1
10:1
98.02
81.82
48
82
44
270
1
Quantitative H NMR spectra of these diastereoisomeric
resonance attributable to C(5)H of 2 could not be observed
above baseline noise. These results therefore demonstrate
that this technique may be applied to the accurate determi-
nation of diastereoisomeric ratios of up to 1000:1 (i.e., 99.8%
de), signal-to-noise ratio permitting.
In conclusion, the technique described in this manuscript
offers a robust method for accurate measurement of very
high diastereoisomeric ratios (>95% de), which may be
applied routinely to the analysis of product ratios or purity.
The application of this protocol to the determination of the
levels of selectivity resulting from a very highly diastereo-
selective olefination protocol are reported in the following
manuscript.
mixtures were acquired using 90° pulse tip angles (7.5 µs)
with recovery delays of 20 s and acquisitions times of 3.2 s
and were processed without apodization functions (Figure
3).
When the spectra were integrated over identical regions,
good correlation between the calculated and observed dias-
(6) (a) Claridge, T. D. W. High-Resolution NMR Techniques in Organic
Chemistry; Elsevier Science Ltd.: Oxford, 1999; pp 114-116. (b) Pauli,
G. F.; Jaki, B. U.; Lankin, D. C. J. Nat. Prod. 2007, 70, 589.
(7) Li, S.; Wilson, W. K.; Schroepfer, G. J., Jr. J. Lipid Res. 1999, 40,
764. Soininen, P.; Haarala, J.; Vepsa¨la¨inen, J.; Niemitz, M.; Laatikainen,
R. Anal. Chim. Acta 2005, 542, 178. Pauli, G. F.; Jaki, B.; Lankin, D. World
Patent WO 2008/051857, 2008.
(8) Claridge, T. D. W. High-Resolution NMR Techniques in Organic
Chemistry; Elsevier Science Ltd.: Oxford, 1999; pp 25-30.
(9) Oxazolidinones 1 and 2 were prepared according to the procedure
of Cutugno, S.; Martelli, G.; Negro, L.; Savoia, D. Eur. J. Org. Chem.
2001, 517 Via treatment of either (1R,2S)-ephedrine hydrochloride or
(1S,2S)-pseudoephedrine hydrochloride with CDI.
Acknowledgment. The authors wish to thank New Col-
lege, Oxford for a Junior Research Fellowship (A.D.S.).
(10) Curves were fitted using the T1/T2 relaxation module of Bruker
TOPSPIN software.
OL802211P
5436
Org. Lett., Vol. 10, No. 23, 2008