No-D NMR (No-Deuterium Proton NMR) spectroscopy: A simple yet powerful method for analyzing reaction and reagent solutions

Article abstract of DOI:10.1021/ol049979+

(Equation presented) The title technique is a convenient and powerful method for directly monitoring or assaying any reaction mixture or reagent solution. Examples of some common processes (Fischer esterification, lithiation, butyllithium/THF compatibility, olefin metathesis, and a quantification assay), each interrogated in its native solvent, are presented. The spectral data are easy to acquire, and the information content makes a compelling case for routine use of No-D NMR spectroscopy.

Full text of DOI:10.1021/ol049979+

ORGANIC
LETTERS
2004
Vol. 6, No. 6
953-956
No-D NMR (No-Deuterium Proton NMR)
Spectroscopy: A Simple Yet Powerful
Method for Analyzing Reaction and
Reagent Solutions
Thomas R. Hoye,* Brian M. Eklov, Troy D. Ryba, Mikhail Voloshin, and
Letitia J. Yao
Department of Chemistry, UniVersity of Minnesota, 207 Pleasant St., SE,
Minneapolis, Minnesota 55455
hoye@chem.umn.edu
Received January 1, 2004
ABSTRACT
The title technique is a convenient and powerful method for directly monitoring or assaying any reaction mixture or reagent solution. Examples
of some common processes (Fischer esterification, lithiation, butyllithium/THF compatibility, olefin metathesis, and a quantification assay),
each interrogated in its native solvent, are presented. The spectral data are easy to acquire, and the information content makes a compelling
case for routine use of No-D NMR spectroscopy.
Most introductory discussions of proton NMR spectroscopy
include the fact that 1H NMR spectra are typically recorded
in solutions prepared with deuterated solvents. The rationale
is, of course, that this enables observation of the protons of
interest without the interference from solute protons. Al-
though there are certain and obvious advantages to following
this practice, the use of solvents lacking protons (CDCl3,
D2O, CCl4, etc.) is by no means a necessity. In fact, by using
the strategy described here of collecting NMR spectra in
nondeuterated solvents (No-D NMR), one can routinely
examine any (and all) reaction or reagent solution(s).1 Scale,
economics, and premeditation of the studies are not issues.
We show representative examples of 1H NMR spectra of
reaction mixtures that contain no deuterium atoms (other than
those of natural abundance). Consequently, these No-D NMR
spectra were recorded in unlocked mode. These examples
represent only a small sampling of the informative data that
can be readily obtained for any type of reaction or reagent.
We present them, in part, to show that the paradigm2 of
routinely recording proton NMR spectra in deuterium-
enriched solvents is unwarranted.
The concentration of most neat organic solvents is ca. 10
M. The concentration of the reactants in most reaction
solutions is in the range of 0.1-1 M (and of solutions of
most commercial reagents ca. 0.5-2.5 M). Thus, the ratio
of solvent to solute/analyte molecules is usually between
100:1 and 10:1 in most solutions of interest. It is a
(1) We do not mean to imply that this is a new concept. To be sure,
samples of biological macromolecules are routinely recorded in a 9:1 ratio
of H2O/D2O and nuclei other than protons (e.g., 31P, 19F, 15N, 11B, etc.) are
also often examined in nondeuterated solvents.2 Moreover, although more
specialized applications in various areas (e.g., LC NMR, food science,
mechanistic studies, MRI, oil and petroleum industries, and quality control)
can be found, use by preparative organic chemists is not at all routine.
(2) Anecdotal comments such as the following underscore the engrained
nature of the practice: “I have run countless 31P (or 19F) spectra directly of
aliquots from reaction mixtures in nondeuterated solvents, but never once
did I think to put the sample tube into a proton probe and collect a 1H
NMR data set!”
10.1021/ol049979+ CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/17/2004

straightforward matter for most NMR spectrometer hardware
to handle dynamic ranges of proton intensities of greater than
4 orders of magnitude.3
It is our intent here to show how the data are collected
and the variety of information types that can be obtained.
Although the examples presented are all of known processes,
each is representative of a class of reaction or reagent that
chemists might want or need to interrogate.
The first example (Figure 1) is the simple Fischer
esterification4 of acetic acid (1) with ethanol (2).5 A 1:4 molar
Figure 1. Fischer esterification of acetic acid in ethanol (1:4 molar
ratio).
ratio of AcOH/EtOH was treated with 1 mol % of concen-
trated H2SO4. A 600-µL aliquot was removed to a sample
tube and thereafter used to monitor reaction progress toward
an equilibrium mixture by No-D 1H NMR spectroscopy. The
excellent resolution and quantitative features of the spectra
are noteworthy. This is a good experiment for first-time users.
The second example (Figure 2) is the lithiation of benzene
(4)6a by n-BuLi and TMEDA6b in hexanes at room temper-
ature. Spectrum (a) shows the No-D NMR spectrum of
n-BuLi in hexanes (2.0 M) containing 1.0 equiv of benzene.
Notice that one can decipher the individual components in
the hexanes [methylcyclopentane (i), n-hexane (ii), iso-
hexane (iii), 2,3-dimethylbutane (iv), and cyclohexane (v)
Figure 2. Metalation of benzene with n-BuLi in hexanes with
TMEDA [spectra (a)-(d) are offset for clarity].
in the bottom expansion]. This sample had already stood for
1 week in the NMR tube sealed only with a standard plastic
cap. No conversion to phenyllithium (5) was observed before
addition of TMEDA. Spectrum (b) was recorded 10 min after
the addition of 1.0 equiv of TMEDA; some phenyllithium6c
(∼5%) is already apparent (top expansion). Spectra (c) and
(d) (24 and 96 h) show ∼90% and essentially full consump-
tion of n-BuLi. A full equivalent of TMEDA was required;
use of a substoichiometric amount of the TMEDA promoter
(data not shown) stopped at a much lower conversion to PhLi
(product inhibition?).6d
All of the spectra shown here were straightforward to
record. Our experience suggests that the art of learning to
reliably shim a sample7 has analogy with learning to ride a
bicycle. Typically, users new to the technique find that after
a dozen or so samples they are able to shim and collect
quality data sets with no greater investment of effort or time
than is required for an analogous deuterated sample.
(3) Solvent suppression routines were not used for any of the spectra
shown here. However, particularly for more dilute samples, they can lead
to better spectral quality. Another contributing factor to routinely achieving
the quality of the spectra presented here is that most modern NMR
spectrometers suffer very little from magnet drift and, therefore, give narrow
lines when run in unlocked mode.
(4) Fischer, E.; Speier, A. Ber. Dtsch. Chem. Ges. 1895, 28, 3252-3257.
(5) All spectra shown here were recorded on a Varian INOVA 500
instrument, but the technique is applicable to any spectrometer. Factors
influencing instrument choice for No-D experiments are no different than
those for conventional locked mode acquisition.
(6) (a) Langer, A. W., Jr. Trans. N.Y. Acad. Sci. 1965, 27, 741-747. (b)
Eberhardt, G. G.; Butte, W. A. J. Org. Chem. 1964, 29, 2928-2932. (c)
Ladd, J. A. Spectrochim. Acta 1966, 22, 1157-1163. (d) Rausch, M. D.;
Ciappenelli, D. J. J. Organomet. Chem. 1967, 10, 127-136.
954
Org. Lett., Vol. 6, No. 6, 2004

The well-known decomposition of tetrahydrofuran (6) by
n-BuLi at room temperature is the third example (Figure 3).8
diagnostic resonance at -0.95 ppm for the RCH2Li protons,
the broad resonances at 3.5 and -0.9 ppm are attributed to
n-BuOLi or n-BuO2Li and its mixed aggregate with n-BuLi,9
respectively. Spectrum (b) reports that 5 min after addition
of THF (6, 4.5 equiv), the onset of its decomposition into
ethylene (7) and the lithio-acetaldehyde enolate (8) is
observable. The RCH2Li resonance is shifted upfield by 0.2
ppm upon addition of THF in response to dynamic THF/n-
BuLi complexation.10 Over time [spectra (c) and (d) at 2
and 48 h] resonances for each of the decomposition products
grew. Even at partial conversion [∼40% in (c)] the large
adjacent THF resonance does not obscure the (O-Z)-vinyl
proton (Hb) of 8. After 48 h all n-BuLi was consumed. Even
though it was not the intent of this ambient temperature
experiment, we subsequently judged the half-life for decom-
position to be ∼2 h, quite consistent with the known value.8d
This well-resolved spectrum of LiOCHdCH2 (8) maintained
its character for at least a week for a sample (again) protected
only with a standard plastic cap.
To further demonstrate the sensitivity and generality of
monitoring reaction progress in nondeuterated solvents, the
fourth example includes spectra of a ring-closing metathesis
reaction (Figure 4). A solution of allyl acrylate (9, 0.5 M in
CH2Cl2) was exposed to 10 mol % Grubbs first generation
ruthenium precatalyst (G1). At 1 min [spectrum (a)],
conversion of a portion of G1 (δ 20.0 ppm) into the Ru-
loaded substrate (13) is evident (δ 18.9 ppm, t, J ) 4.4 Hz).11
Spectrum (b), taken at 4.5 min, shows that G1 is nearly
consumed. The extent of formation of styrene (10) [δ 6.7
(dd, J ) 16.9 and 10.9 Hz) and 5.7 (d, J ) 16.8 Hz) ppm]
correlates with consumption of G1.
Figure 3. Generation (and stability) of ethylene and acetaldehyde
enolate by THF decomposition with n-BuLi.
Formation of γ-butenolide (11) [δ 7.6 (dt, J ) 5.8 and
1.5 Hz) and 6.1 (d, J ) 5.8 Hz) ppm] and E- and Z-dimer
(12)12 is evident and continues as long as resonances due to
any resting carbene are observable. At 64 min [spectrum (c)],
the methylidene carbene 14 (δ 18.9 ppm, s) is the only
identifiable ruthenium-containing species. After 3 h no
carbene resonances remained observable (not shown). At 48
h no further conversion (compared to 3 h) had occurred, so
an additional 10 mol % of G1 was added. One minute later
[spectrum (d)] it was obvious that reaction had been
reinitiated. A similar profile of conversion and alkylidene
Spectrum (a) is of a sample of n-BuLi in hexanes that had
stood in a capped NMR tube for 24 h. In addition to the
(7) (a) Gilman, H.; Gaj, B. J. J. Org. Chem. 1957, 22, 1165-1168. (b)
Honeycutt, S. C. J. Organomet. Chem. 1971, 29, 1-5. (c) Jung, M. E.;
Lyster, M. A. Tetrahedron Lett. 1977, 43, 3791-3794. (d) Stanetty, P.;
Mihovilovic, M. D. J. Org. Chem. 1997, 62, 1514-1515 (t1/2 ) 1.78 h at
20 °C).
(8) There are several methods that can be used to shim the proton probe
to a No-D sample and tube. These include (a) the initial use of a reference
tube of the same volume of deuterated solvent, which is locked and shimmed
as usual and then replaced with the No-D sample; (b) recalling/reentering
a set of previously used shim settings known to be optimal for the solvent
in use; (c) the use of a capillary insert containing a deuterated sample of
the same solvent, which can be locked and shimmed; (d) gradient shimming;
(e) shimming the No-D sample using the FID (see “shimming using the
FID” below); and (f) shimming the No-D sample using the spectrum (see
“shimming using the spectrum” below). Method (c) is a reliable way to
start, and methods (c) and (d) are compatible with autosamplers. We use
methods (e) or (f) for nearly all samples (including all spectra shown here).
Although the following instructions refer to software (VNMR 6.1) on a
Varian, Inc. instrument,5 they are sufficiently generic to be helpful to users
of other manufacturers’ equipment. Initial Setup. Use a typical volume
(e.g., ca. 600-700 µL) of a solution of solute in nondeuterated solvent
with a concentration of, say, 0.1-1 M (ca. 100:1 to 10:1 solvent/solute).
Set the spectrometer parameters as you would for a deuterated sample [e.g.,
nucleus, solvent (if “known” to the spectrometer, otherwise choose
anything), initial shim parameters, spectral width, etc.]. Turn off the “Lock”
but spin the sample. Acquire a single scan spectrum. Phase this initial “un-
easier observation of small changes in the FID level). Adjust the shims,
allowing the FID level to stabilize before making each additional adjustment,
until a maximum FID level (numerical) has been achieved. (f) Shimming
using the Spectrum. Shim by monitoring the increase in reporter peak
intensity (numerical and/or graphical) and peak symmetry. Don’t be
discouraged by initial poor-looking peak shape. In general, note that the
response to a change in shim settings will occur more slowly here than it
will when shimming on a deuterated sample in locked mode. The process
of shim optimization is otherwise quite analogous for samples in nondeu-
terated vs deuterated solvents. Once the shims are optimized, exit “acqi”
and take another one pulse spectrum. If the overall spectrum quality is
acceptable, proceed with recording the spectrum. Shimming the second
through the nth sample at the same sitting will usually be easier.
(9) (a) McGarrity, J. F.; Ogle, C. A. J. Am. Chem. Soc. 1985, 107, 1805-
1810. (b) McGarrity, J. F.; Ogle, C. A.; Brich, Z.; Loosli, H.-R. J. Am.
Chem. Soc. 1985, 107, 1810-1815.
1
shimmed” No-D H NMR spectrum. Select, expand, and note a “reporter
resonance” of known peak shape (a solvent peak is usually a good choice,
but any peak of known multiplicity will suffice). Run the FID/Spectrum
macro gf and enter the interactive acquisition display process acqi (to
allow observation of the real-time FID or spectrum). (e) Shimming using
the FID. After performing the initial setup, select the FID button and then
increase the gain until the FID level is between 500 and 1000 (to allow
(10) Keresztes, I.; Williard, P. G. J. Am. Chem. Soc. 2000, 122, 10228-
10229 and references therein.
(11) Schwab, P.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1996,
118, 8, 100-110.
(12) Sueltemeyer, J.; Doetz, K. H.; Hupfer, H.; Nieger, M. J. Organomet.
Chem. 2000, 606, 26-36.
Org. Lett., Vol. 6, No. 6, 2004
955

against an internal standard (p-NaOPhCO2Na, 15).13 The
inset table shows the reported and observed values for a set
of such solutions. Integration of resonances in the No-D
spectra of the standards of known concentration (entries 1-4)
demonstrated the quantitative reliability of the measurements
(R2 ) 1.000). Assay of commercially available solutions
(entries 5-8) suggested systematic overreporting of ethanol
titers.
Figure 5. Quantitative assay of aqueous ethanol solutions.
Acknowledgment. These studies were supported by
grants awarded by the DHHS (GM-65597 and CA-76497).
We thank Drs. Dmitry O. Koltun and Tricia J. Vos for early
contributions to the developments reported here.
Supporting Information Available: Enlarged, uncom-
pressed (high resolution) versions of the figures in both
black-on-white (printable) and color (displayable) formats.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Figure 4. RCM of allyl acrylate to produce γ-butenolide. The
region from δ 18.5-20.5 is increased vertically (5×) for easier
visualization.
depletion [spectra (d)-(f)] emerged. A considerable amount
of information and insight is obtained from this simple
experiment.
Finally, a number of more specific applications are under
study. Among these are various “titration” methods for
quantifying the amount of a solute of interest (e.g., organo-
lithium, Grignard, and metal hydride species). To whet
palates here, we show (Figure 5) an assay of ethanol content
in various aqueous solutions as determined by integration
OL049979+
(13) A known quantity of p-hydroxybenzoic acid (typically 50-70 mg)
was added to a solution (1.00 mL) of aqueous ethanol, and 40% aqueous
NaOH (∼100 µL) was added to deprotonate and solubilize the standard
acid. No-D 1H NMR data were collected (4-8 transients) with an acquisition
time and a pulse delay of 20 s each to ensure complete relaxation and
quantitative proton integration. Pulse widths (transmitter power) were
reduced to remove baseline spectral artifacts and increase integration
accuracy. Using these protocols the Ho vs Hm resonances integrated reliably
to 1.00 ( 0.02.
956
Org. Lett., Vol. 6, No. 6, 2004