Dear colleagues,
after discussions of Rainer Haessner at this year's ENC, about possible
systematic errors in temperature calibration with CH3OH samples due
to the high Q of cryoprobes, I ran some tests here to figure out if
there is any significant effect.
Since our 900 cryoprobe was busy with important experiments, and our 600
cryoprobe is currently "on vacation / rehab" with the manufacturer, I
had to use a standard room temperature TXI at 600 MHz. Not exactly what
I would have thought of as a "high Q probe".
Anyway, what I tried was the following test program:
1. pure CH3OH with perfect tuning & matching
2. pure CH3OH with seriously detuned probe (increasing
the 90 deg. pulse length from ca. 12 us to 20 us)
3. pure CD3OD (the remaining 0,04% 1H are easily
sufficient for a nice 1H spectrum in one scan)
Deviation of calculated "real" temp. from display
correct T&M | detuned probe | CD3OD
Display rh Bruker| rh Bruker| rh Bruker "impure"
290,00 -2,61 -2,90 | -1,20 -1,55 | 0,32 -6,41
300,00 -2,74 -2,97 | -1,37 -1,53 | 0,09 -5,88
Evaluation was done with the Bruker "tempcalc" makro, which performs
automatic peak picking, calculation & display of the frequency
difference and subsequent calculation of the "real temp." ("Bruker" in
my table).
In this program the correlation "freq. diff./real temp." is only defined
up to 300K, so I'm showing only the numbers for 290 and 300 K here (I
measured up to 350 K). The program has a choice between "neat methanol:
Y/N", so for the pure CH3OH sample I used "yes", and for the CD3OD
sample I tried "no" (denoted as "Bruker 'impure'" in the table).
Alternatively, I used Rainer Haessner's online tool
(
http://www.nmr.de/html/tricks/tempcalib.htm), these are the numbers
denoted "rh" in the table.
From the data, I find three aspects remarkable:
a) although "tempcalc" and Rainer use different correlations from the
literature, the "real temps." I get differ only (?) by ca. 0.2-0.3 K.
(For CD3OD I therefore skipped the "Bruker / neat methanol" option.)
b) in spite of the standard probe, I got quite a significant difference
between tuned and detuned probe (ca. 1.5 K)! In our case, the required
correction of the displayed temp. changed from ca. 2.8 K (tuned) to only
ca. -1.4 K for the detuned case. This is a much larger effect than what
I had (probably naively) expected for a room temp. 600 MHz probe!
c) what is puzzling me: I expected switching to CD3OD to go in the same
direction as detuning the probe (you greately dilute 1H concentration,
thus avoiding the effects from the additional field generated by the
sample itself).
Rainer's program uses literature data for "pure CH3OH" (=samples
containing NO deuterium whatsoever; great fun to shim this sample ...).
So the calculated corrections for CD3OD of almost zero relative to the
display should be purely incidental.
However, with the "methanol - neat: NO!" option of Bruker's "tempcalc",
I get a correction of ca. -6 K!!! Not only is this really MUCH larger
than all the other corrections - it is also "in the wrong direction",
assuming that detuning the probe with CH3OH will give an improvement.
I talked to Wolfgang Bermel from Bruker Karlsruhe, and he confirmed that
this option is valid for a sample of 96% CD3OD / 4% CH3OH. I would
expect that using my sample of 99.96% CD3OD / 0.1% CH3OH would at least
come close to this ... obviously not!?
Sorry for this lengthy mail but now I am kind of confused. Of course I
could just ignore all this and continue using the "erroneous"
corrections we always used in the past ...
What are you using to calibrate your spectrometer temperature display?
- is anyone using a "96% CD3OD" sample? (I just digged into all our
standard / reference / calibration samples that were once delivered
together with our spectrometers - and found no such sample).
- have you ever tried the effect of detuning your probes?
If I'll get interesting new insights into this from the group (which
won't let me look too ignorant), I'll happily post a summary in this forum.
Servus from Munich,
gg
--
==========================================
PD Dr. Gerd Gemmecker
Dept. Chemie, OC II
TU Muenchen
Lichtenbergstr. 4
D-85747 Garching
Germany
Raum/Room 32 303
Tel. +49 (89) 289-13308
Fax +49 (89) 289-13210
e-mail: Gerd.Gemmecker_at_ch.tum.de
Internet: http://www.gemmecker.de
--
==========================================
PD Dr. Gerd Gemmecker
Dept. Chemie, OC II
TU Muenchen
Lichtenbergstr. 4
D-85747 Garching
Germany
Raum/Room 32 303
Tel. +49 (89) 289-13308
Fax +49 (89) 289-13210
e-mail: Gerd.Gemmecker_at_ch.tum.de
Internet: http://www.gemmecker.de
Received on Wed Jun 07 2006 - 11:00:19 MST