I like to thank you all for your reply to my question(s) on DNMR. I find
the information I have recieved very very valuable and I will remember both
my experience and all your reply for ever toward better understanding of
conformational features of "organometallic complexes" in the future.
Since, many of us would be interested in knowing the response I am fortunate
to have recieved for my post, I am very delighted to post them accordingly
to Bruker-Users-Mail. They are as follows:
Thank You Very Much !!
Mahesh....
=========================================================
I think you describe a case were Keq may indeed play a role, and needs
to be considered in the line shape analysis.
What we have done - perhaps you can do the same. Obtain spectra for
several temperatures below coalescence (170K) or near your lowest
temperature region (where all resonances are clean sat 150-170)
allowing one to determine Keq. From that teperature dependence you can
calculate the populations at any given teperature near coalescence,
such that the k_exchange is the only variable in the simulation. We
als do similar things for changes in chemical shift and J coupling to
predict values in the coalescence region.
If they are still giving you grief (or you have difficulty reaching
the lower temperatures) you could try a higher field instrument -
expanding the lower temperature region where k_exchange is still slow
as not to effect the line shape.
Good Luck
Todd Alam
tmalam@sandia.gov
=============================================================
Both you and the referee are right to some extent. The equilibrium
constant will depend on temperature, but in your case, the dependence is
small. We faced the same problem (J. Phys. Chem. vol 99, 17338-17343
(1995)).
The equilibrium constant is given by
K = exp(-delta G/RT) = exp(- delta H/RT)exp(delta S/R),
so if delta G is non-zero (which it will be for unequal populations), K
will be temperature dependent. A common assumption is that delta S is
zero. Then you can calculate delta H at low temperature, and extrapolate
K. If you don't assume delta S = 0, then you need to know the equilibrium
constant as a function of a wide range of temeperatures, and try to
separate delta H and delta S. I'm not aware of anyone who has done a
really good job of this - most of us assume delta S is zero.
In your case, the population is close to equal, so delta G is small and
the temperature dependence of K is relatively small. Let me know if you
have questions.
Alex D. Bain email: bain@mcmaster.ca
Department of Chemistry phone: (905) 525-9140 ext 24524
McMaster University fax: (905) 522-2509
1280 Main St. West
http://www.chemistry.mcmaster.ca/faculty/bain
Hamilton, Ontario CANADA
L8S 4M1
=============================================================
Maheswaren,
I don't quite understand the system but I can tell you this: If there
are several,
more than one, species present, you need to know ther relative
concentrations as well
as the NMR parameters for each species. The ideal way to do this is to
obtain spectra
over a range of low temperatures within which the spectra are not perturbed by
exchange effects. Then if the ratios change draw a Hoffmann plot (lnK vs
1/T) and
extrapolate it into the temperature range wherein signal averaging takes place.
You just read off the equilibrium constants at the higher temperatures.
One hopes that the intrinsic NMR parameters for the different species
do not
change with temperature.
If your ratio does not chage with temperature that should be evident
also aty the lower temperatures. Further the place where the shifts average
should be the same
as the weighted average from the low temperature spectrum. You can check that.
Sometimes it isn't possible to obtain NMR data over a suitably wide low
temperaturerange due to solubility problems. You could still see where the
system averages. If you
explain all this in the paper it should be accepted.
If you like, send me your data, or the write up, but with notes and
diagrams
to explain what you see and what you did. I'll be happy to look at it.
Gideon Fraenkel
Professor of Chemistry
Ohio State University
Voice 614 292 4210
Fax 614 292 1685
=============================================================
Hello,
Speaking from an organometallic kineticist's point of view (I did my
graduate work
for
Bergman and we did quite a bit of this kind of stuff) . . .
Maheswaran Hariharasarma schrieb:
> Hello
>
> The stopped exchange region (170 K) of the dynamic 31P{1H} NMR spectrum of a
> cis-Mo(CO)4{P(OR)3}2 (where R group is chiral) in dichloromethane-d2
> consists of one doublet of a doublets (AX pattern) and a singlet. We have
> assigned these phosphorus resonances to two unequal populations (3:2) of
> enantiomeric pairs of two diastereomeric conformations of the complexes that
> are undergoing chemical exchange.
>
> Thus, both diastereomers undergo chemical exchange via ring_inversion at the
> phosphorus and thus, at room temperature only one sharp 31P NMR resonance is
> observed.
>
> A(four line AX pattern)<=> B(singlet)
>
> We have modeled this dynamic behaviour(over a 40 degree range from 170K )by
> making use of the the populations(3:2) obtained at the stopped exchange
> region while treating other parameters as variables for optimization by
> iterative
> procedure. (chemical shift did not vary more than 0.15 ppm over a 40 degree
> range) i.e., we have kept the populations obtained at 170K constant for
> the computation. However, when we treated populations as independent
> parameter, it seemed they were too ill defined for iteration and were
forced to
> work with low-temperature values.
>
> We have lately submitted this paper for publication and got an interesting
> opinion from a referee who finds amusement with the methodology we have
> used. He claims, "knowledge of both 'forward rate constant'(kf) and the
> 'equilibrium constant'(Keq) for the species at equilibrium is needed for the
> computation of dynamic NMR line-shapes.
This unfortunately I believe is correct. One needs to be able to rely on your
equilibrium constants that are used in the calculation of each of the line
shapes
at each of the temperatures.
> Since, these two parameters can
> only be independently obtained at low-temperatures (wherein separate
> resonances are seen, ca. 170K); any line-shape computation using
> low-temperature values of population is useless.
Since almost all equilibria have some difference in delta H, almost all
equlibria
do indeed change their populations as you change the temperature, hence, the low
temperature data is not so good to use at other temperatures (one of the
drawbacks
of trying to get rate data out of coalescing signals).
> It is my understanding, that one assumption that is common to all of the
> line-shapes theories is that the parameters which describe a spectrum in the
> slow exchange limit can be used to calculate spectra at any rate. (I
> understand, this assumption is difficult to avoid for intermediate and fast
> rates, specifically for chemical shifts, which we have treated as
> independent parameter in the computation)
>
Well, as I said above, I don't think that assumption is a good one. There
are some
things that you can do. Basically, you need to get data for the equilibrium
constant at the temperatures where you don't know it. This can be done in
several
ways . . . Get more data at different temperatures in the slow exchange
region so
you can plot 1/T vs. deltaG. If you can get enough temperatures in the slow
exchange region you get reasonable numbers for deltaH (slope) and deltaS
(intercept) and then you use these to extrapolate to the temperatures that
you need
(Not fantastic usually but better than just using the slow exchange equilibria).
This method is also usually hard to implement as you need lots of room in
the slow
exchange limit to drop the temperature -- it sounds like you do not. An
alternate
method is to switch to another nucleus with greater difference between
signals in
absolute frequency, thus in effect raising the temperature of the slow exchange
signals and giving you an equilibrium value at a new, warmer temperature.
Carbon
of the R group on the phosphite might work, although I have my doubts. Finally,
change spectroscopies to something that has a faster time scale and freezes
out the
two diastereomers. Try for instance IR -- once again, it may not work for your
system, but maybe . . . .
> Did I do any major mistake in line-shape calculation for paper to be rejected
> for publication ?
>
If you cannot do the methods above, you need to say that you cannot do the
methods
above in your paper due to limitations in your chemical system. The
question then
will just rely on whether you are able to further advance knowledge in the area
that is in question with your study.
T. Andrew Mobley
e-mail: mobleyt@rz.uni-leipzig.de
phone: don't got one yet
address: Ritterstrasse 12
Appt. 406
04109 Leipzig, Germany
=============================================================
ahesh,
In principle the referee is correct because the relative populations at
each temperature will also determine the lineshape. In practice you can do
two things:
1, assume that DS is almost zero in the temperature range you are working
in and so you assume that the DG doesn't change much and so the population
hardly varies over your temperature range. (You might even be able to work
out how large DS would need to be to give seriously different lineshapes by
changing K and calculating lineshapes).
2, plot the chemical shifts for as large a temperature range as possible
below the coalescence and then above the coalescence. Below coalescence
you will get the temperature variation of the chemical shifts which you can
extrapolate into the region of interest. (You should do this as a matter
of course anyway.) Above coalescence the chemical shift changes will be
governed by the intrinsic variation of the shift of each site with
temperature and any population changes due to temperature variation of K.
If your data show that this temperature variation of K is very small then
your assumption under 1 above is OK.
Oh, and there is another thing you could do: send it to a diffferent journal!!
Frank
Frank Riddell, School of Chemistry, University of St Andrews,
St Andrews, Fife, KY16 9ST, Scotland, UK.
Tel: + 44 (0) 1334 - 463815 Fax: + 44 (0) 1334 - 463808
E-mail: fgr@st-andrews.ac.uk
Research group page:
WWW: http://www.st-and.ac.uk/~www_sc/personal/fgr
Personal Photos page:
WWW: http://www.st-and.ac.uk/~www_sc/personal/fgr/Photos/
=============================================================
"Earth may have enough to cater to our needs, it surely will not put up with
our greed"
--Mahatma Gandhi