--
Dr. Christophe Farès
Head of the NMR Department/Leiter der NMR Abteilung
Max-Planck-Institut für Kohlenforschung
Kaiser-Wilhelm-Platz 1
45470 Mülheim an der Ruhr, Germany/Deutschland
Tel: +49-208-306-2130
E-mail: fares {at} kofo.mpg.de
Web: http://www.kofo.mpg.de/en/research/service-departements/nuclear-magnetic-resonance-spectroscopy
----------------------------------------------
Original emails:
Hi Chris,
Reliable variable temperature continues to pose challenges for us. We train students to setup and run their own VT experiments. Shimming is hit or miss and Bruker's Topshim routine does not always work at low T. No problem for experienced user who will take the time to shim by hand but many students simply stop and run as is. Every week we need to look after students who forget to dismantle the low T set up or turn off the VT unit. People always want to ramp up or down the temperature quicker than they should (we use 10 C every 5-10 min).
Our approach is very hard on the hardware. If I had more time in the day, I would run things differently.
Marc
-----------------
Hi Chris,
I see no problems on doing low temp experiments as and when required. Whether you need to ramp temperatures will depend on your hardware and control settings. Modern systems seem to control well without over or under shooting targets, your experience may vary.
Samples will always need to be reshimmed at different temperatures due to solvent expansion/contraction and viscosity changes. Any other issues will depend on your hardware and the carefulness of your users.
Regards,
Phil.
----------------------------
Chris,
This was a significant aspect of managing the Northwestern facility when I was there. Three of our instruments were used for VT experiments on an almost daily basis. Our Bruker Avance-III 600 spent a good 1/3 of its time at -40 °C (the lowest we could get with a two-stage FTS chiller), about two days a month near -96 °C with the liquid N2 apparatus, and probably five days a month between +100 and +150 °C. Our Agilent 400MR spent a good10-20% of its time at +120 °C - probe's upper limit (people were making polymers that melted at +110 °C).
Shimming: It's difficult to start with RT-appropriate shims when you're well away from RT. The +120 °C people often needed reminding that they should try shimming at an intermediate temperature, say +70 °C before going up all the way. Those who saved their own shim files once they got good lineshape at high (or low) lineshape were happier, but they had to be specifically educated about how to do save and retrieve shim files. The Bruker Topshim routine worked very well near RT, but outside of -20 °C to +45 °C, it often failed. The Agilent gradient shimming, while more basic, was more robust for use in higher and lower temps. Perhaps there were ways of improving both that I was unaware of.
I generally tried to educate the students that they shouldn't just drop or raise the sample temp by huge amounts over a short period of time. For instance, if they were going to -40 °C, I'd suggest turning the FTS to -90 °C, then dropping the sample temp to -10 °C, and once it got close, then drop it again to -40 °C. O course, we had no way of enforcing that behavior, and students usually just set the temp to their final temp anyway.
Our Varian AutoX probe never had a problem with the repeated temp changes, but our Bruker BBI probe seemed like it needed servicing every year. It was a different problem every time (shields out of alignment (probably due to excessive airflow), cracked insert, broken capacitor), and none was clearly directly related to the VT use. However, it was clear that the probe that endured the most VT use - especially cold use - was the one that had to go to the shop the most.
Two other points of note:
1) Our students had to be reminded often that they had to return the spectrometer to RT by the time they were done, and it actually had to be at RT then. We had a spell where some were signing up for an hour just to take a "quick 1D" at -40 °C. Some would return the sample temp to +25 °C just a couple minutes before someone else needed to use it, and that person inevitably got confused about why their shims kept changing and why the temp wasn't stable.
2) I found that most of the students were unaware that NMR sample temperatures should be calibrated if they want accurate readings. I encountered one unfortunate pair who had been performing detailed thermodynamic and kinetic measurements, assuming the sample was actually at the temp shown on the screen, and in reality they were always between 1 and 5 °C off. So demonstrating temperature calibration is an important part of their training.... and maintenance. Even after training, they need to be reminded to calibrate their temps or at least look at our calibration chart.
I hope this helps!
- Josh
------------------
Dear Chris,
I don't do as much VT as you do. I usually take the temperature down slowly to -80 C with calibration of approximate temperature along the way. It may take me around 45 minutes to get down to low temperature, calibrate and then put in my sample. I would allow 5 minutes at least to equilibrate my sample to reduce thermal gradients.
I go up in temperature much more quickly, within about 20 minutes.
Best, Martha
--------------------------
Hi Chris,
This a great set of questions, one I just never had the forethought to ask of AMMRL.
We have done a lot of VT over the years, and have never worried about such issues; it has not been uncommon that we would have VT going 5-10 times per week on a spectrometer. And some have dynamic experiments where the temp is being changed during the day (e.g., from -50 to -130C). So we see a lot of temp changes and ramps. The experiments have occurred on both Bruker and Varian systems.
We have made some mistakes, see some problems, and have instituted various issues to try to alleviate them.
1. One mistake made many years ago was having the lift gas being air; terrible idea with water condensation "rusting" the shim stack. That magnet was salvaged, but I was horrified (but not too surprised) when the rt stack was pulled. Fortunately, the arrangement was setup by my predecessors; I made sure that all gas going into the magnet gap was N2 every since then whenever we're running VT.
2. We use gas purged through the shim stack (always N2), and place heating tape about the top and bottom of the shim stack that is turned to warm (not very warm, but enough to feel by finger touch) to help lessen condensation at those spots.
3. We run temps up and down freely. Shimming can be an issue, but students are used to saving their shims at specific temps, and working from there. On our new Bruker systems, topshim often does fine in shimming things in, but that doesn't always work.
4. We do ramp the temps especially going up from being cold in 20C steps. Turn it up then give it 5-10 min at the new temp before going further. I'm no completely certain this is important, but we don't like to see the heater on max for extended (>30s) times.
5. If cold (e.g., -80C) for a while, once at room temp we wait>15min before disconnecting the LN2 tank to help lessen condensation in the gas flow areas.
6. It can take> 30 min for shims to settle when changing temps significantly. For the careful work,> 60 min may be needed for the metal structure to stabilize. We make it the VT user that has to schedule in the delay (i.e., they must have the probe back to ambient) 30 min prior to next user's time.
We not seen that anything really bothers the probes; i.e., they seem to hold up fine. In particular, we've had Bruker BB probes reworked about every 10 yrs to replace the slider sections. But that's about the same on VT probes as on our "routine" probes. The Varian probes held up fine to the cycling as well.
I'm sure I'm not thinking of everything, but that's what comes to mind now. Really looking forward to seeing what others chime in with.
Cheers,
Charlie
-----------------------
Dear Chris
We do low temperature experiments here on a regular basis, although not as often as you describe. In a nutshell, I would agree with you that shimming is no longer a big issue. We have Bruker instruments and gradshim or in particular Topshim to an excellent job in this regard. We are still concerned about the mechanical stress these large T-changes produce and ask our students to change the temperature in 10-15C increments followed by a 5-10 min equilibration period. This also allows the sample to cool down properly. For example, on our instruments with cryoprobes, we have restricted the T-changes to go in 2C/min steps. Yes, it will take some time, but I argue that the sample is ready to go when the temperature is reached and the probes are better protected. The temptation of a large, quick temperature change because the experiment has to be done before going home is rather big and we had instances, where the quartz inserts cracked, most likely by thermal stress. Since you are doing this so often, the manufacturer might give you some limits to what they consider safe. This might be faster then our procedure.
For long term experiments, I would be concerned about the bottom O-ring. If you have ice build-up or constant cooling over en extended period (days), the O-ring might be affected and the vacuum in the magnet break. I heard of such instances after much shorter times when the N2 dewar was not hooked up properly, allowing too much cold gas to get at the bottom O-ring.
Hope this helps,
Regards,
Markus
------------------------
Greetings, Chris,
We have 5 walk-up magnets with no VT use restrictions, except:
1) Every probe has its temperature range
2) Temperature changes be made gradually in steps of less than 20 degrees C.
3) Users return to probe to room temperature prior to departure.
We do train people carefully, and individually. You are welcome to use the attached Varian-oriented document.
I do think temperature changes are hard on the probes, but it seems like something we have learned to live with here. The most common problem is that shimming changes, but shimming is still achievable. If you have gradient shimming, even better. The glass insert sometimes breaks, and that is typically a repair of about $3000 and several weeks.
Most VT here happens in the evenings because the whole process is too slow for daytime use.
Dean
-----------------------
If temperature is changed too rapidly, probe glassware could break and data acquisition might start before the sample temperature is the same as the probe environment. The latter - except of course with the new software/gadgets that measure actual sample temperature. Maybe newer hardware is more robust, but I try to be careful in any case. Only selected researchers are allowed to run VT and they are trained. If they have a one time need, then I run it for them (for "free" :-))
I don't have the new stuff that directly measures sample temp.
I have my users do it through automation with a custom temperature changing routine which works out to around a degree every 8-20 seconds, with a delay waiting for temperature stabilization every 5 degrees.
Only experiments labelled VT change the temperature through a variable I make readily accessible; all other experiments (non-VT experiments) are set to reset the temperature to a standard value (24C) in case the user goofs up. At the endpoint, temperature stabilization is checked and a little extra time is given to stabilize. Eventually I would like to add something that checks lock signal which can indicate if temperature of sample is still changing, and also load temperature control parameters at various temperature points. Sample is then tune/match, gradient shims for Z, and aftertune touchup of Z-Z3, Z,Y, XZ, YZ done.
Ramping is done via custom software; the standard hardware also does some ramping. (AVIII instrument)
The users are instructed to always run a 24C experiment at the end of the experiment - makes sense for me but it also makes sense for them as it tells them if heating/cooling the sample has changed the sample or not.
I also have a H2O/D2O standard in the autochanger and the user is supposed to run a custom experiment that redoes all the shims and stores it as a standard set. Also tunes/matches.
George
-----------------
Hi Chris,
How have you been?
I like your traditional way of having a reserved day for VT.
While the systems are more robust now, I would still take time going to "extreme" temperatures such as -80C (or +80C). Going from say 0 - 50C is easy enough, but you cannot get around the fact the things expand and contract, and resistivity in the RT shim coils also vary with temperature.
So, you may be able to change the sample temperature a little faster than you used to, but you still need to leave it at that target temperature for a while until the system reaches a thermal steady state.
I would not "ramp" up and down. The only motivation I can think of to do so is to lessen the amount of time needed to reach a steady state. I think you're better off to just wait.
Also, if a user is planning to look at several samples at low-T, I would instruct them to maintain the low-T when changing samples. They may even pre-cool their samples in a Dewar outside the probe, just be careful with condensation when changing samples.
---------
In a related matter: we made a small recent modification to deal with too much ice buildup on the probe during extended low-T experiments. We had a dry nitrogen purge blowing where the VT gas entered the probe, but it was insufficient to completely eliminate the ice buildup. So we wrapped a bag loosely around the base of the probe. With the positive pressure of dry-N2 gas in the bag, there was no way for the water vapour to condense on the cold surfaces.
Bye for now,
Bob.
---------------
Hi Chris,
I'd concur with your view that VT is not a huge issue these days. We still take it carefully, but tend to mix it up with ambient operation in order to stop it being heated/cooled too often in a short space of time.
Our only real restrictions in terms of low temperature are solvent limits (obviously), sensible hard and soft maximums and minimum temperature (sensibly within the probe specs, so anything less than -100 on a probe that is rated to -150 requires us to know about it), and finally we ask that they ramp the temperature if they are going very low (so more than 50 degrees below ambient), but this is usually faster than in the past (typically equilibrate the temperature every 20 degrees below -20, so it takes about 15-20 minutes to get to -100).
It is the last point (temperature ramping), which is probably the most unclear point, because no-one knows what is sensible until the probe cracks! So we take it conservatively.....
On the other hand, we tend to heat probes fairly routinely without controls on the ramping rates - but then we rarely heat more than 60-70 degrees above ambient, so ramping isn't a big deal here anyway.
Craig
-----------------
Dear Chris,
The issue of shimming is "fixed" only provided you have provisions for room temperature air to be flowing around the outside of the shims. If you have Bruker spectrometers, only the latest console, i.e. AVIII, has this. However, Bruker has had a flange that can be retrofitted to their older systems. If you don't have this, I would suggest you get it for any spectrometer that you use for VT. Not only does this largely solve the shimming issue, but when used properly it will avoid the more serious problem of the potential to quench the magnet if the dewar O-rings get too cold or hot. With this flange you can go to the limits of the temperature specs of the probes for extended periods without problems. If you have Varian/Agilent systems, I don't know what provisions they have for this.
The question of ramping the temperature is not something that I have a strong opinion about, but we don't do it. I will be interested in seeing what replies you get to that.
Jane
-----------------
I think high temperature VT could be quite hard on older probes. The
first generation of plated susceptibility matched rf coils could be
irreversibly damaged by prolonged heating and become more or less
unshimmable; also, there can be some air oxidation of the coils at high
temperature. As far as I know, the coils used today are much less
susceptible to these problems. These did not apply to low temperature VT
however.
The biggest remaining problem with VT experiments is the presence of
temperature gradients inside the sample, driving convection cells. These
greatly reduce gradient shim performance and also resolution, particularly
for any signals exhibiting temperature dependence. Keeping the sample as
short as practical, and maintaining spinning during any experiments where
this is possible, are good for limiting convection. Dropping the
temperature gradually may also be good for that, although a very small
temperature gradient can still drive a convection cell if the solvent
viscosity is low. I am not sure if very rapid but otherwise controlled
cooling or heating is bad for the probe in terms of thermal shock, but I
have always changed the temperature in steps of 20-30 degrees and I've
never had an insert crack on me--so why change?
Received on Mon Feb 25 2013 - 00:55:47 MST
This archive was generated by hypermail 2.4.0 : Sun Jun 18 2023 - 15:21:11 MST