I sent in a query about 13C shifts of residual solvents (e.g., CD2HOD) in
deuterated solvents on Jan. 4, 2005. The short answer is that my I've
found out what works for CD3OD solvent:
CD3OD CD2HOD CDH2OD CH3OD
49.15 49.43 ??? 49.99
If I reference my residual (CD2HOD) solvent peak in the HSQC or HMBC
spectrum to 49.43 ppm, all my 13C shifts derived from cross-peaks line up
with the 13C shifts from the 1D 13C spectrum, which is referenced to
solvent CD3OD at 49.15 ppm. Note that the deuterium isotope effect on 13C
shift is exactly proportional to the number of deuteriums substituted in
the CH3 group.
Here is a summary of the responses I got:
Original Query to AMMRL: I often reference HSQC spectra using the residual
solvent peak. This leads to confusion because solvent chemical shifts in
tables (e.g., Cambridge Isotope Labs NMR Solvent Data Chart) give 13C shift
of the fully deuterated solvent and 1H shift of the residual (one H, the
rest D) solvent. For example:
13C shift of CD3OD = 49.15 (from the chart)
13C shift of CD2HOD = ?
13C shift of CH3OD = 0.56 ppm downfield of CH2HOD
Clearly it's not right to set the CD2HOD peak in the HSQC spectrum to
49.15. Can I assume that the isotope effect is linear in the number of
deuteriums? That would mean that CD3OD=49.15, CD2HOD=49.43, CDH2OD=49.71
and CH3OD=49.99. Of course, I could add TMS and measure all of these, or I
could acquire a 1D 13C spectrum and use a sample peak to reference the
HSQC, but I'm looking for a quicker and easier way. And, final question,
can we trust those charts put out by the isotope people? They differ from
chart to chart.
I will always state in a publication what I reference to, but it would be
nice if this business was more standardized. Does anyone know of a paper
where these numbers are given definitively?
---------------------------------------------------------------------
all modern spectrometers offer the possibility to use the lock signal
(the deuterium signal of the fully deuterated solvent) as a secondary
reference for all nuclei. Maybe you have this option.
Of course there are two sources of errors:
- the environmental dependence of the solvent shift
- the table used for the different gyromagnetic ratios of the nuclei
In our experience we are very satisfied with all solvents
with the exception of water und the nuclei table shipped with
Bruker spectrometers seems to be o.K.
-------------------------------------------------------------------
Here is a paper that has several solvent "impurities" and their chemical
shifts.
Gottlieb, Kotlyar, Nudelman J. Org. Chem. 1997 62, 7512-7515
[lots of useful information but none on 13C shifts of "partially
deuterated" solvents]
----------------------------------------------------------------------
I typically use the SR values from the 1H, and 13C files. Assuming that
you have a Bruker.
i.e. 1_sr_value from the 13 spectrum, and 2_sr_value from the 1H
spectrum.
-----------------------------------------------------------------------
Are you using a Varian instrument? If so why not use the setref command,
here are a few excerpts from the "Getting Started" handbook.
"setref macro assumes that the system is locked (and that the lock solvent
is defined in /vim/solvents). If you are working without lock and still want
to use serif, you must ensure that the field offset has been previously
adjusted so that the lock frequency is on resonance with a sample of similar
susceptibility. To ensure that the field offset is adjusted, do the
following procedure:
1. insert a sample with reiterated solvent
2 adjust z0 in acquit so the lock frequency is on resonance
3 switch off the lock
4 insert the nondeuterated sample
..."
" The accuracy of the setref macro is mostly limited by the accuracy of the
chemical shift of the lock resonance line, which may depend on the
concentration and the chemical properties (acidity/basicity) of the
components in the sample...."
" The macros setref1 and setref2 are used to reference f1 and f2 in
multidimensional spectra, respectively. ... "
---------------------------------------------------------------------------
I don't have a literature reference for you, but I do know that the isotope
shift is linear with the number of deuterium (or whatever). I have had a
couple
of samples over the years with multiple substituted species present (CH3,
CH2D, CHD2)
and they are nicely linear. I also had a nice high resolution 19F spectrum of
CFCl3 that showed the Cl35/Cl37 isotope shifts very clearly.
As far as reporting is concerned, pick a number, use it consistently, and
report it in the publication.
[this was by far the best advice]
------------------------------------------------------------------------------
I would use without any discussion the IUPAC-defined "XSI"-Scale.
http://www.iupac.org/publications/pac/2001/7311/7311x1795.html
(provided, you have a TMS-referenced 1H).
For details see the practical guide
http://www.chem.wisc.edu/~cic/nmr/Guides/Other/Xi_chem_shift_scale.pdf
-----------------------------------------------------------------------------
I like the article in JOC vol 62 no 21, 1997 p7512-7515 HE Gottlieb et al.
It has several advantages (1) it is a literature reference that you
can cite to customers, not some chart by company X (2) solvent
shifts are reported with error ranges, lending some credibility
to the notion that an actual scientific measurement was made and
(3) it lists many common contaminants & their 1H and 13C shifts in
each of the common deuterated solvents. For example, CD3OD is listed
at 49.00 +/- 0.01 ppm and methanol (CH3OD) in CD3OD is listed at 49.86 ppm
I would be very interested in what others are doing,
especially if there is a consensus on a more rigorous approach
(aside from always using a solvent with TMS in it.)
-----------------------------------------------------------------------------
Usually, we have a 1D-13C spectrum, and we reference a peak in the gHSQC or
gHMBC to a known peak referenced to the fully-deuterated solvent.
Of course, it is becoming increasingly common to work with sample quantities
that are too small to observe 13C directly, and on these samples gHSQC and
gHMBC are the only 13C frequencies that we have to work with. I think your
interpolation of the fully-protonated vs. fully-deuterated would be accurate
enough (well within the resolution limitations of the 2D linewidths).
I have run into the problem of doing these experiments in D2O, where there
is no reference peak. The biochemists deal with this by referencing back to
the absolute proton frequency of DSS (based on the exact shift of water at
the specific sample temperature). Using the ratio of the gammas (13C/1H)
they calculate the exact reference frequency for 13C and 15N. This blew my
mind for awhile because some of them (the bio-NMR users) didn't even fully
understand where the numbers came from (Actually the method is from a
classic paper authored by David Wishart, Eric Oldfield, Jane Dyson, John
Markley, Brian Sykes, et. al: J. Biomolecular NMR (1995), 6(2), 135-40.)
I suppose that the method outlined in this classic paper could be applied to
any solvent, as long as one knows exactly where zero ppm is in that solvent.
A very interesting question, Niel. I have never given it much thought
because most of my users run these experiments at low enough resolution that
the isotope shift really doesn't matter (and, as I already said, we usually
have a properly referenced 1D carbon spectrum to use).
I really don't believe that there is an accepted, standard solution to this
problem.
------------------------------------------------------------------------------
The 13C chemical shift of CD3OD is 47.84 ppm [R. E. Hoffman, J. Magn.
Reson., 163, 325 (2003)] relative to 1H of TMS in CD3OD ª 0.2514502 [R. K.
Harris, et al., Pure Appl. Chem., 73, 1795 (2001)]. It is reasonable to
assume that the isotope shift of CHD2OD is two thirds of CD3OD. However, it
is preferable to use the 1H frequency of TMS or calculate it from CHD2OD =
3.30 (1st ref. or more accurately 3.305 at 25°C) and multiply by 0.2514502
to yield the carbon reference.
------------------------------------------------------------------------------------------
Thanks for the mail, which gives me an idea to avoid these calibration
problems:
For HSQCs, I would, one time, for every common solvent, add a small amout
of TMS, and then calibrate the
residual solvent peak with respect to TMS.
This would give the right value for the residual (one H, rest D solvent),
e.g CD2HOD.
With this procedure, one could have a list of chemical shifts for all
single protonated solvent.
----------------------------------------------------------------------------------------
Thanks and sorry for taking so long to compile a summary,
Neil
Neil E. Jacobsen, Ph.D.
NMR Facility Manager
Department of Chemistry
119 Old Chemistry
1306 E. University
University of Arizona
Tucson, AZ 85721
520-621-8146
FAX 520-621-8407
Received on Fri Mar 11 2005 - 14:49:25 MST