New NMR Temperature Calibration Protocol and Correction of old calibration sets

• Based on Findeisen et al. (Magn. Reson. Chem. 2007; 45: 175–178 findeisen2006.pdf), we determined that the temperature calibration curves generated using neat methanol, in use prior to June 2007, have resulted in a systematic error in sample temperatures of about +2-3 deg, i.e. actual sample temperatures were about 2-3 degrees warmer than indicated from the calibration. You may wish to record and report this specifically for the data sets prior to June 2007.
• We have since generated new calibration curves, starting in June 2007, for each spectrometer, using 99.5% methanol-d4, as suggested by Findeisen et al. We have also confirmed using a thermocouple that the new calibration curves do provide accurate temperatures.
• Both sets of calibration curves will remain in the red log books at each spectrometer. This will allow you to correct your temperatures appropriately. We recommend that you use the new calibrations. The old calibration curves will be removed sometime in the next six months. We suggest you record which calibration curve you use.
• As an example, for 800US2#1, if your desired sample temp was 298K, using the old calibration curve you would have set the controller target temp in edte as 303.87K, but your temperature would actually have been c. 300 deg K. The new calibration curve shows that for a controller target temp in edte of 303.87K, your actual sample temp would have been 299.87K.
• For 800US2#1, using the new calibration curve to achieve an actual sample temp of 298K, you would now need to set the controller target temp in edte as 301.92. However, if you wish to match the older sample temp of 299.87K, in the example above, you would just set the controller target temp in edte as 303.87.
• The equations obtained from temperature calibration using a thermocouple are reported in the red log book at each spectrometer. In each equation Y=sample temp, X=EDTE temp set point. (NOTE: for 1.7 mm probe in 800us2#2, the calibration curve was generated using 99.5% methanol-d4)

-- MichaelGoger - 20 Jun 2007 * findeisen2006.pdf: findeisen2006.pdf