Hi,
This was my question:
We would like to be able to accurately integrate 2H data. So far we run
our experiments without lock and obtain quite good qualitative spectra.
However, the baseline is far from being perfect and occasionally we can
also observe some strong glitches. Temperature is stabilized and the
sample does not spin during acquisition. What else can be done in order
to improve stability of the instrument without lock?
I've never done experiments using 19F or any other nuclei than 2H for
lock signal. Maybe this would be the only solution. Can anybody tell me
what do I need to run this kind of experiment (probe, other hardware,
software)? Please send me any suggestions: general and specific for our
instruments.
We use an ARX-400 (with BBO probe - Z3918/0041 or BBI - Z3750/0144) and
XWINNMR 1.3. Our new ECX-300 is going to arrive soon. Maybe we can still
order some modifications that will allow us to lock on 19F (or something
else) and observe 2H.
Well, I haven't learned much about 19F lock but I was able to solve my
problem. Generally, everybody thinks that problems with our 2H spectra
will not be helped by 19F lock, which seems to require quite complicated
and expensive modification of the hardware. 19F lock may be useful for
long experiments (several hours) - our data acquisition usually takes
less than 10 minutes.
Many people suggested disconnecting the lock cable at the probe in order
to remove those glitches from the 2H spectra (turning off the lock and
setting lock power to -60 on the BSMS is not enough!) and setting
LOCKNUC = OFF (I didn't know such a parameter existed and I am not sure
what it does). IT WORKED! We obtained a very clean spectrum.
We have also greatly improved quality of the baseline by optimizing the
DE delay.
Thank you Ashok, Bill, Charles, Chris , Donna, Jane, Jose, Markus,
Matthias, Phil, Rainer, Rodger, and Wei!
Piotr
Piotr Dobrowolski, Ph.D.
NMR/EPR Manager
Department of Chemistry and Biochemistry
Utah State University
0300 Old Main Hill, Logan UT 84322-0300
telephone: (435) 797-1673
fax: (435) 797-3390
Here are all of the responses:
I don't think a 19F lock would help. Are you sure your lock channel is
completely off and detuned? This may be the source of glitches.
Baseline problems can be helped by linear prediction or baseline
correction algorithms if your software has them.
I don't really think that not locking is the source of problems with 2H
integration. If you are convinced that it is and want to lock on 19F,
Bruker used to sell a 19F lock as an option. I presume they still do,
but I don't know for sure. We have it on an ARX500. The user who
needed it has been gone for years and the probe we used is not in good
shape, but it used to work very well.
How long do your experiments take? Unless you are doing very long
experiments (several hours) I would have thought you'd be fine
without locking, unless you have a drifting magnet. The 2H peaks are
probably quite broad anyway.
Are you using the X-channel of your probes or the lock channel? Do
you have a lock switch? If you are using the X-channel you might see
some interference if you don't disconnect the lock channel. If you
are using the X-channel preamplifier I assume you're removing the 2H
filter....
I guess you did turn off the lock power (-60). If not, this could be the
reason for some of the glitches. Although theoretically it should no
longer interfere since the lock is turned out, it does.
Please file this under "my two bits"
Although it couldn't hurt and would be useful for long acquisition
experiments, I don't think 19F lock will help much with baseline or
glitches. The baseline may be due to filter or probe ringing, so using
RIDE or processing with linear prediction may help. As for glitches,
check for bleed through from our lock transceiver (may disconnect it
during acquisition?).
Confirm in [eda] -> LOCNUC is OFF and that the SWEEP (in [bsmsdisp] or
in the BSMS Keyboard) is OFF.
Also be sure that the 2H filter is not installed in the broadband
channel of the preamp (HPPR).
Have you tried removing the lock cable from the probe when you run your
experiments? Maybe you have some sort of interference creeping in to
give you the glitches.
I have seen spikes in 2H spectra without lock. If we disconnect the two
2H lock cables from console to preamp the spikes disappear. We also
remove the 2H filter between the preamp and the probe.
One of my colleagues recommends the following (in case you haven't
already) before resorting to a 19F lock.
He should use a 2H pass filter, that would take care of the
clitches. The base line problem could be addressed by using a
longer DE or base line correction. If he has to run 2H for a longer
periode of time, then he may activate the drift compensation.
Concerning the baseline, increasing the value of DE might be helpful
(200 Microseconds as a rough guess).
I use to run a lot of 2H spectra for a research group that used
deuterium as an incorporation label in metabolic pathway studies. These
were often 5-10% enrichments requiring over night runs. These were done
with the instrument unlock but with the field drift adjustment correctly
set to follow the field drift. You do not need the 19F lock. We had one
with the original AM console and it was not worth the price. The 19F
lock needs a lock transmitter receiver, a 19F lock preamp and a probe
with 19F lock channel. The price is over 50K. I am assuming that your
samples are fully enriched at one or more locations and reasonable
amounts, so that your run times are shorter than mine. The one problem
you will have with deuterium is that the FID dies so quickly (T2
relaxation due to quadruple relaxation) that if the early data points in
the FID are incorrect, you will have baseline distortions. You can
correct for this condition by zeroing out the early data points and use
linear prediction to back calculate what they should be.
We at JEOL saw your posting on AMMRL about running 2H spectra on your
existing ARX-400.
I can mention a couple of things that we discovered while learning the
subtleties of 2H gradient shimming on the lock channel. If the LOCK
channel is still radiating RF, gradient shimming will fail.
We actively turn off the lock oscillator so there is no spurious
deuterium signals radiating while observing 2H.
We run 2H deuterium shimming all the time and if there were glitches it
would be really sad - I would hate it.
I am not exactly sure if you are pulsing on the LOCK channel or the Low
frequency coil. Either way you would see the same effect if the lock
oscillator was not turned off.
2. A few things we do when pulsing on the low frequency coil
automatically
a. Turn off the 2H lock oscillator so there is no spurious radiation.
b. Detune the lock coil, so there is no coupling between the lock and
LF coil. c. Bypass the LOCK STOP filter so as to allow 2H RF to pass
through.
We have customers in the field who run 2H spectra (lock coil and LF
channel) and have not heard of any problems similar to what you state.
Additionally we have the smarts built in, so if you run nuclei that are
close to 2H (O17, Li6) we automatically bypass the 2H stop filter built
into the spectrometer so they can be observed automatically. I don't
know if there is interest in nuclei close to 2H, but I thought I would
mention it.
At present there is no 19F lock available on the ECX300. But I don't
see this as a problem for your application assuming that the 300 does
not have a drift problem.
Typically 300 magnets have a proton drift less than 0.1[Hz]/Hr.
On the ECX300 you should be able to do the following to observe 2H -
that is just the tip of the iceberg.
1. Pulse on the lock channel.
2. Pulse on the LF channel.
Have you also considered changing your integration technique? I mean,
simple integration is more impaired by the problems you describe than
for example deconvolution techniques especially when the latter is
also able to simultaneously fit a baseline and even allows you to
include external knowledge like our software is offering.
Received on Tue Oct 25 2005 - 12:17:04 MST