Minutes from proton polarization meeting held June 11 2008 in 892-1-D20 These minutes are available also from: http://wwwcompass.cern.ch/compass/detector/target/Minutes/2008/ a) Jaakko: NH3 was loaded on March 10 2007. First TE April 4 and May 2 - 4. The first physics runs from beginning of June in transverse mode. Dilution cryostat blockage problem July 20 - 28. August 10 longitudinal physics data taking started. November 11 Drell-Yan test. TE-calibration started November 13. Target material unloaded November 20. First empty cell measurement started November 21. No material visible inside the cells and the cells were immediately cooled back to LN2 temperature without warming up. Unload empty cells November 26. The cells were warmed up to 30 - 40 C to remove any ice and residual ammonia. The caps were put on the filling holes. Second empty cell measurement started November 27 and finished December 2. List of NMR files by Takuma http://wwwcompass.cern.ch/compass/detector/target/NMR/NMRfiles07/2007target.htm and preliminary offline polarization from Kaori http://wwwcompass.cern.ch/elog/target_polar/41 The run logbook MySQL data base can be used to read run number, start time, end time and run type. b) Kaori: the Q-curve was taken manually only about every 2 hours during TE-calibration. The coil #10 was broken after second reload of the empty target cells. Also coil #9 tuning changed. Possible recovery by line shape analysis using correlation coefficient. Attempt to improve the background proton measurement by eliminating very noisy signals and averaging with a reversed signal. No improvement in TE signal area calculation. Quality check by using one Q-curve for all the signal files and doing this for all Q-curves. The histograms similar to the normal method with time ordered Q-curve and signal file analysis. Curie plots of integrated NMR signal area vs. 1/T for all coils. Both the TE-signals from ammonia and background protons fit well into Curie law. The first and second empty cell measurements give different Curie constants for the background protons. Coil #9 was re-tuned to see the influence of the Q-curve to the Curie constant. The error seems to be very small. Compare two cases. Use average of first and second empty cell measurement to correct for the background protons. Then use only second empty cell measurement. The Curie constants are -12 - -1 % lower in the latter case expect for coil coil #3 with +2 % difference. The maximum polarizations high with this method up to +-95 % compared to the SMC +-90 %. In transverse data taking try to scale the polarizations with coil #6 that was used to monitor the polarization continuously in the 0.63 T dipole field. Fit between the points before and after transverse data taking gives better results. Error estimate. TE-calibration 2 - 5 %, enhanced signal baseline removal < 0.1 %, field polarity 0.2 %, LF gain drift 0.1 % and microwaves < 0.1 %. Circuit non linearity and field shift need to be checked. Nitrogen-14 polarization was checked by SMC collaboration in NIMA 419 (1998) 60. Gerhard: possible cross polarization of nitrogen with protons. Jaakko: this happens at 56 mT when the proton NMR and nitrogen-14 quadrupole frequencies are close to each other. c) Yuri: TE-calibration for coils #1 - #3 is well reproducible. Coil #4 has discrepancy. Downstream coils #5 - #10 need to be checked using cross correlation with #1 - #3. Linear fit of the NMR peak shift vs. polarization gives a coefficient about 7 % lower than for the SMC signals. Fit a gaussian to the asymmetric polarized proton NMR line peak to extract the asymmetric part of the signal. Possible explanation for the asymmetric part indirect exchange through the nitrogen-14. Asymmetry depends on the gradient of the static magnetic field and mutual influence between polarized cells. Maximum correlation of the NMR signals during polarization build up to a reference signal allows to determine the difference between polarizations of these two signals. Needs careful checking of the correlation function. The cross correlation can improve the error from two background measurements 2 - 15 % to 1 - 3 %. Good agreement with the correction factors calculated from the polarized SMC signals. d) Takuma: ammonia TE-calibration good fit to the Curie law. Difference in Curie constants between first and second empty cell measurement 0 - 10 % for upstream coils #1 - #5 while the difference is much larger 18 - 30 % for downstream coils #6 - #9. This can be explained if some ammonia material was left in the target holder after unloading. Compare Curie constants using average of first and second empty cell measurements and the case of using only the second empty cell measurement. Coils #1 - #4 less than 1.2 % difference while for coils #5 - #8 the difference is 1.3 - 2.6 %. For coil #9 6.3 % and coil #10 was broken in second empty cell measurement. Use coil #2 as a reference signal for line-shape correlation analysis. Make a plot of polarization difference versus correlation coefficient R. Problem does not have unique solution and the distribution of possible solutions needs to be considered. Using the line-shape analysis new Curie constants can be estimated. These are in general agreement with the values from TE-calibration, but the error is larger 1.1 - 5.9 %. Clear deviation from the TE-calibration is seen for coils #3, #6 and #8. Comparing the line-shapes to SMC ammonia target smaller correlation coefficients seen. Thus the line-shapes were a little different. Possible reasons: only 400 data points recorded, different NMR circuit or coils embedded into the material. Conclusion: line-shape analysis is useful for cross checking the polarization values and correcting the values for coil #9 and #10, where the TE-calibration data is incomplete. e) Jaakko: reusable code for NMR analysis using Root libraries developed http://wwwcompass.cern.ch/compass/detector/target/NMR/Ntoolsdoc/index.htm List of the analyzed results in http://wwwcompass.cern.ch/compass/detector/target/NMR/NMR07 The NMR coil and thermometer positions 2007 http://wwwcompass.cern.ch/elog/target_polar/15 May 2 - 4 helium-3 vapor pressure temperatures http://wwwcompass.cern.ch/elog/target_polar/19 November 13 - December 3 RuO thermometers http://wwwcompass.cern.ch/elog/target_polar/17 Comparison of the helium-3 thermometer to RuO thermometers in 2.5 T field http://wwwcompass.cern.ch/elog/target_polar/22 Gaussian fit the TE-signals. The center frequency unstable in the beginning of the November TE-calibration http://wwwcompass.cern.ch/elog/target_polar/28 The field map from the NMR center frequencies http://wwwcompass.cern.ch/elog/target_polar/31 gives a little lower field for the empty cells than for the ammonia material, which could be expected if the protons in empty cell measurement are on the target cells. Coil #4 anomalously high magnetic field in empty cell measurement. The field maps for first and second empty cell measurement are almost the same. The NMR line-widths for the two empty cell measurement are much wider than for the ammonia http://wwwcompass.cern.ch/elog/target_polar/32 Thus the protons are not from ammonia crystals. For upstream coils #1 - #4 the same line-width for the first and second empty cell measurements. For downstream coils #5 - #9 the line-width broader in the first empty cell measurement. Raw integrated areas are 2 - 14 % higher than the areas from the gaussian fits http://wwwcompass.cern.ch/elog/target_polar/34 The relaxation times 1 - 3 hours at temperature 1.0 - 1.5 K for both ammonia protons and empty cell protons http://wwwcompass.cern.ch/elog/target_polar/39 http://wwwcompass.cern.ch/elog/target_polar/40 Robust time cuts were used to determine the most likely thermal equilibrium time. Area times temperature histograms gave the Curie constants for the May and November TE-calibrations and for the two empty cell measurements http://wwwcompass.cern.ch/elog/target_polar/44 http://wwwcompass.cern.ch/elog/target_polar/45 Q-meter center frequency was stable during TE-calibrations with less than 50 ppm drift. Coil #9 had clearly smaller noise compared to other coils. http://wwwcompass.cern.ch/elog/target_polar/49 http://wwwcompass.cern.ch/elog/target_polar/48 Simulate full TE-calibrations with background noise and statistical uncertainty in temperature. Slow needs about 5 minutes to simulate 1 TE-calibration. For coil #6 the errors for the May and November TEs were 2.05 % and 1.67 %. The empty cell measurements had errors 5.21 % and 6.23 %. http://wwwcompass.cern.ch/elog/target_polar/53 The May and November TEs measure the same quantity, so the data can be combined for smaller statistical error. The two empty cell measurements do not agree for all the coils. If both empty cell measurements are used, systematic error needs to be added. General agreement: first empty cell measurement had extra protons that were removed in the second empty cell measurement. Preliminary results from dispersion signal correction after phase shifting. The phase error seems to be less than 1 - 2 degrees for all the coils.