1. RHIC pC polarimeter what has been achieved and what needs to be done Osamu Jinnouchi RBRC 2/10/05 RSC meeting

2. 2/10/2005Osamu Jinnouchi RBRC 2 Outline ﻪ Brief description of system and data ﻪ Some measurement performances (Run-04) ﻪ Summary of the systematic errors ﻪ Calibration using Jet-target : obtained A N ﻪ Offline analysis results of Run-04 ﻪ Consistency check with online result ﻪ Corrections with newly obtained A N for past runs ﻪ The things can be done with Run-04 data

3. 2/10/2005Osamu Jinnouchi RBRC 3 Detector layout ﻪ Strips are parallel to the beam direction ﻪ 72 strips x 2rings ﻪ 45degree detectors (1,3,4,6) see reduced A N (by 1/sqrt(2)) Ultra thin Carbon ribbon Target (3.5  g/cm 2 ) 1 3 4 5 6 2 Si strip detectors (TOF, E C ) 15cm 10mm 2mm pitch 12 strips p + implants ~150 nm depth 72 strips in total Thin dead layer for low energy carbon spectroscopy

4. 2/10/2005Osamu Jinnouchi RBRC 4 DAQ and Data types ﻪ Offline data typically typically ﻩ 15M carbon selected events (27M overall events) (27M overall events) corresponds to 170MB size corresponds to 170MB size ﻩ DAQ time : ~10 seconds ﻩ Readout time: 3min Wave Form Digitizer (WFD) 20M events / 20sec - Pulse Height - Bunch ID - TOF - Integral (Q) Select carbons at on-board LUT Scaler data Scaler data Asymmetry calculation Asymmetry calculation Online results (to experiments) Online results (to experiments) Event by event data Stored in on-board memory Stored in on-board memory Used for offline detailed study Used for offline detailed study ~50ns ~100mV Bunch Next Bunch (213ns) Shaped Si Signal offline online

5. 2/10/2005Osamu Jinnouchi RBRC 5 Strip by strip asymmetry study Each strip can be considered as individual polarimeter (  vs.  asymmetry) The deviation more than statistical error turned out to be negligible (2 parameters) (Used data set is what is used for Jet calibration, i.e. Blue ring 100GeV) Great achievement in Run-04 Number of carbons in i-th strip for Up(Down) spin Deviations from Fit (projection) Fit with sine function (2 parameters)

6. 2/10/2005Osamu Jinnouchi RBRC 6 Strip by strip (azimuth angle study) RotatorONRotatorOFF RotatorON Best fit of all blue points -Fill by fill fluctuation is explained as a statistical behavior -To reach d  = 20mrad (the error size of red points), 32 CNI runs needed -Good sensitivity to spin angle measurement BLUE 100GeV With Jet

7. 2/10/2005Osamu Jinnouchi RBRC 7 Strip by strip at 24 GeV (in blue) -16 CNI runs of 24GeV data taken with Jet (including longer runs) - Angle: zero consistent with good statistics -  2 /ndf : not so bad

8. 2/10/2005Osamu Jinnouchi RBRC 8 Beam profile issue ?? ﻪ We claimed that we observed significant polarization profile ﻪ Is it a fine statement ? ﻪ Within a FWHM we cannot tell the difference ﻪ If we control it by mm precision, it would be fine Yellow target scanned vertically Yellow In different fill 1mm polarization beam intensity

9. 2/10/2005Osamu Jinnouchi RBRC 9 Beam profiles (in other ring) ﻪ Blue is also the same situation ﻪ Only one scan saw a very weird profiles in both intensity and polarization Blue horizontal scan Blue vertical scan

10. 2/10/2005Osamu Jinnouchi RBRC 10 Beam profile issue will be gone, anyway ﻪ In the new application tool for the RHIC polarimeter (available from Run-05), the beam center finding (fast target scan) will be done automatically and easily ﻪ As long as the target is sitting away from the beam edges, we can avoid the profile issue ﻪ Also with the new RHIC application, the plan is to make regular measurements with the target moving back and forth ﻪ Seth Nemesure (CAD) is working on this application ﻪ We need commissioning of this new tool at the beginning of pp run

11. 2/10/2005Osamu Jinnouchi RBRC 11 List of systematic errors ItemsRun02Run03Run04 Difference between e90 vs. e45 (run by run) Statistically fluctuate with small deviation from zero (reasons were not understood) N/A but expected to be very small from the azimuth fit to the strip by strip asymmetries Radial component (run by run) Cross asymmetries (run by run) Bunch by bunch azimuthal fit (removing bad bunch) DONEN/A Bunch dependent systematic error (Bunch shuffling) Almost statistical behavior N/A Event cut dependence (mass 2vs.3 sigma cuts) N/A Run03 See the significance 0.005-0.015 in Polarization Run04 see the absolute difference relative 3% (dP=0.012 for P=0.40) (1) dP/P from Jet N/AN/A0.085 (2) Energy correction ambiguities Estimated ±5ug/cm 2 Same ±12ug/cm 2

12. 2/10/2005Osamu Jinnouchi RBRC 12 Energy correction estimation ﻪ Fits with non-relativistic kinematics ﻪ With the factor 1181.6, the fit is forced to the Carbon mass (11.17GeV/c 2 ) 20  g/cm 2 60  g/cm 2 100  g/cm 2 3  Mass cut Time of Flight (ns) Energy (keV)

13. 2/10/2005Osamu Jinnouchi RBRC 13 Determination of dead layer from fit Fit 400-1000keV Fit 200-550keV flattop injection 200-550 keV 200-700 keV Luminocity (or DAQ rate) correction

14. 2/10/2005Osamu Jinnouchi RBRC 14 What makes difference in dead layer btw injection/flattop ﻪ At flattop the carbon locus is closer to the timing threshold ﻪ At high luminosity, the carbon locus is distorted from the actual shape Real center Faked center (hand drawing) Fit range Looks OK High luminosity 100GeV 80ug/cm 2 Low luminosity 100GeV 64ug/cm 2 High luminosity 24GeV 52ug/cm 2

15. 2/10/2005Osamu Jinnouchi RBRC 15 Energy correction for Run-04 Silicons ﻪ Silicons are extracted from the same wafer for each ring ﻪ Fit within 300-1000keV: use the average value from injection measurements ﻪ Blue 57ug/cm 2, Yellow 53ug/cm 2 ﻪ The uncertainty is assigned as ± 12ug/cm 2 ﻩ From the fluctuation between strip by strip ﻩ This is also consistent with the systematical difference between injection/flattop ﻪ The correction may includes multiple effects from silicon, electronics, etc. si-1si-2si-3si-4si-5si-6 Estimated Si dead layer width (Fit Range 300-1000keV) si-1si-2si-3si-4si-5si-6 BLUEYELLOW Due to horizontal target

16. 2/10/2005Osamu Jinnouchi RBRC 16 Carbon cut dependence ﻪ Other systematic error is the cut dependences on the carbon mass BLUE 100GeV through the run BLUE 100GeV 1.030+/-0.0008  2 /ndf = 277/124 BLUE 24GeV 1.033+/-0.0007 = 136/31 Yellow 100GeV 1.033+/-0.0014 = 110/98 Yellow 24GeV 1.034+/-0.0010 = 67/24 So we expect 3% relative change due to the cut  Time in days (since 4/1) Errors are estimated with extended version of D.Kowell ’ s formula i.g. (when one data set is subset to the other)

17. 2/10/2005Osamu Jinnouchi RBRC 17 A N (t) reported at Spin-04 ﻪ pC A N (t) was reported in spin-04 Conference ﻪ Two main systematic errors (  P/P from jet 8.5%, Energy correction) were combined linearly (magenta band) ﻪ Since two components are independent, they should be summed in quadratic (light blue band) Revised systematic band

18. 2/10/2005Osamu Jinnouchi RBRC 18 Comparison of A N jet and A N E950 ﻪ Jet calibrated A N is smaller than the Larry’s fit to E950 ﻪ The ratio dramatically changes in momentum transfer –t ﻪ Since we do not have a new A N at 24GeV, we will use these for the offline values ﻩ 24GeV – use E950 A N (t) ﻩ 100GeV – use new 100GeV A N (t) Ratio is for each point

19. 2/10/2005Osamu Jinnouchi RBRC 19 Offline results of run-04 (compared with online values) BLUE 100GeV 3 sigma mass cut Use new A N (t) 1.247+/-0.011.079+/-0.005 Red points: offline Blue points: online Time in days (from 4/1) When online energy correction changed from 50ug/cm 2 to 66ug/cm 2 at May 1 st Offline/online Average P after May 1 st Only the CNI runs taken with event mode data are selected

20. 2/10/2005Osamu Jinnouchi RBRC 20 Run-04 blue 24GeV 0.94+/-0.004 3sigma mass cut Use E950 A N (t) Red points: offline Blue points: online Offline/online Average P after May 1 st

21. 2/10/2005Osamu Jinnouchi RBRC 21 Run-04 Yellow 100GeV 1.112+/-0.006 1.211+/-0.02 3sigma mass cut Use new A N (t) Red points: offline Blue points: online Offline/online Average P after May 1 st When online energy correction changed from 49ug/cm 2 to 60ug/cm 2 at May 1 st

22. 2/10/2005Osamu Jinnouchi RBRC 22 Run-04 Yellow 24GeV 1.05+/-0.01 0.98+/-0.008 3sigma mass cut Use E950 A N (t) Red points: offline Blue points: online Offline/online Average P after May 1 st

23. 2/10/2005Osamu Jinnouchi RBRC 23 Short history since RHIC retreat-04 in offline analysis P(jet-offline, selected runs) = 0.386+/-0.033 P(jet-offline, selected runs) = 0.386+/-0.033 P(pC-online, selected runs) = 0.359+/- small err P(pC-online, selected runs) = 0.359+/- small err factor = 0.386/0.359 factor = 0.386/0.359 = +1.075 = +1.075 P(jet-offline, all runs) = 0.392+/-0.026 P(jet-offline, all runs) = 0.392+/-0.026 P(pC-online, all runs) = 0.377 +/- small err P(pC-online, all runs) = 0.377 +/- small err factor = 0.392/0.377 factor = 0.392/0.377 = +1.040 = +1.040 [comment] [comment] from D.Kowell's formula, 1sigma of the two from D.Kowell's formula, 1sigma of the two data sets would be (one is subset of the other), data sets would be (one is subset of the other), sqrt(0.033**2-0.026**2) = 0.020 sqrt(0.033**2-0.026**2) = 0.020 then the second factor (for example) could have error then the second factor (for example) could have error (0.392+/-0.02)/0.377 = 1.040+/-0.053 (0.392+/-0.02)/0.377 = 1.040+/-0.053 Consistency check between all and selected runs at online P(jet-online) = 0.369 +/- 0.02 P(jet-online) = 0.369 +/- 0.02 P(pC -online) = 0.381 +/- small err P(pC -online) = 0.381 +/- small err factor = 0.369/0.381 factor = 0.369/0.381 = +0.969 = +0.969

24. 2/10/2005Osamu Jinnouchi RBRC 24 In the past runs, how the A N was determined ?? ﻪ In Run02 ﻩ Fixed A N values were used for all the runs ﻯ BLUE -t = 0.007-0.030 A N =0.0127+/-0.0040 ﻯ YELL -t= 0.006- 0.027 A N =0.0133+/-0.0041 ﻩ These values were obtained from J.Tojo, where he used the asymmetry information of both E950 and RHIC CNI to estimate A N value at RHIC-t range (lower -t) ﻩ Since the RHIC asymmetry saw the rather large asymmetry in the lower –t region, this analyzing power is estimated larger than simply estimated from E950

25. 2/10/2005Osamu Jinnouchi RBRC 25 In the past runs, how the A N was determined ?? –2 ﻪ In Run03 Offline ﻩ Used effective A N based on the Larry’s function fit to E950 data points: weighting by the number of event is the idea ﻩ -t range was chosen to Blue –t = 0.0134 – 0.0201 (600-900keV) Blue –t = 0.0134 – 0.0201 (600-900keV) Yellow –t = 0.0112 – 0.0201 (500-900keV) Yellow –t = 0.0112 – 0.0201 (500-900keV) This rather high –t values are due to large energy correction ﻪ In Run04 online ﻩ The same as in run-03, but in wider –t range Both rings –t = 0.00716 – 0.0224 (320-1000keV) Both rings –t = 0.00716 – 0.0224 (320-1000keV)

26. 2/10/2005Osamu Jinnouchi RBRC 26 How the correction factors are calculated ﻪ For Run-02 ﻩ Recalculate the effective A N (t) within the defined –t range, and simply replace it ﻪ For Run-03 ﻩ Need to take the ratio of effective A N (t) from E950 and Jet calibration ﻪ For Run-04 ﻩ Compare the online and offline (new A N is taken into account) results ﻪ All of these takes into account the N(t) (cross section shape) from the data ﻪ The result does not change by using elastic shape N(t) = N 0 Exp(-Bt)

27. 2/10/2005Osamu Jinnouchi RBRC 27 Correction factors If we simply take the ratio of A N (t) integrations 320keV 500keV 600keV RUN-03 BLUE RUN-03 YELLOW If we define the jet calibrated pC A N (t) with 66 ug/cm 2 energy correction (shift in –t ) Integration range is defined for each year and ring Run04

28. 2/10/2005Osamu Jinnouchi RBRC 28 Correction factors for the past runs Correction Factors ﻪ Run-02 Blue Pol goes up by 8.9% (+13.6%-11.3% of itself) Yellow Pol goes up by 3.8% (+13.4%-11.0% of itself) Yellow Pol goes up by 3.8% (+13.4%-11.0% of itself) ﻪ Run-03 Blue Pol goes up by 23% (+13.6%-12.7% of itself) Yellow Pol goes up by 21% (+13.0%-12.7% of itself) Yellow Pol goes up by 21% (+13.0%-12.7% of itself) ﻪ Run-04 Yellow Pol goes up by 11.2% (+12.3%-11.8% of itself) Online Blue Pol goes up by 7.9% (only Pjet error = ± 8.5%) Online Blue Pol goes up by 7.9% (only Pjet error = ± 8.5%) Note: only runs after 5/1 are considered Note: only runs after 5/1 are considered

29. 2/10/2005Osamu Jinnouchi RBRC 29 Consistency (relations) Ratio BLUE 1 R=1.079 dl=66ug/cm 2 R=1.247 dl=50ug/cm 2 Offline R=1.17 dl=57ug/cm 2  R=0.08  dl=7ug/cm 2  R=0.10  dl=9ug/cm 2 By changing A N (E950  Jet) Ratio YELLOW 1 R=1.112 dl=60ug/cm 2 R=1.211 dl=49ug/cm 2 Offline R=1.17 dl=53ug/cm 2  R=0.04  dl=4ug/cm 2  R=0.06  dl=7ug/cm 2 By changing A N  E=1  g/cm 2 ~  P/P=0.01

30. 2/10/2005Osamu Jinnouchi RBRC 30 The things need to be done with Run-04 data ﻪ 24GeV calibration ﻩ Not satisfying statistical accuracy, but worth to see any deviation from E950 curve ﻪ Spin turn by turn frequency measurement ﻩ By using the unique bunch crossing id, pursue the turn by turn asymmetry (spin precession) ﻩ Fourier frequency analysis is the way to go ﻩ (maybe) I can play with this data during the next week

31. 2/10/2005Osamu Jinnouchi RBRC 31 Summary ﻪ RHIC pC polarimeters are becoming very stable device and precise tool to measure the polarization ﻪ Good quality data samples enable to explore the interesting features of the beam properties ﻪ Still the energy correction is the major issue ﻩ Things will get better when the beam pickup is gone ﻩ Local non linearity in TDC will be improved by the new algorithm (I.Alekseev/D.Svirida) This will also help the better estimation and the better measurements

32. 2/10/2005Osamu Jinnouchi RBRC 32 BackUps…

33. 2/10/2005Osamu Jinnouchi RBRC 33 The time dependence during the fill?? BLUE 100GeV Eth=320keV 3Sigma MassCut The Ratio = Offline/Online not seen a strong correlation

34. 2/10/2005Osamu Jinnouchi RBRC 34 2sigma cut BLUE 100GeV Eth=320keV 2Sigma MassCut The Ratio = Offline/Online Sigma = 1.5  2.0 GeV during a long fill Cut 3sigma (6.7~15.7GeV)  (5.2~17.2GeV) 2sigma (8.2~14.2GeV)  (7.2~15.2GeV)

35. 2/10/2005Osamu Jinnouchi RBRC 35 The effect of the cross section -t Mc +/- 5GeV cut Red: pulse height Blue: integral Line: N exp(-68t) Use integral N counts for the weights Degradation in high –t is from - Timing cut - variation of mass width in Energy Differential Cross section Weighted by EXP(-68t) Weighted by Event spectrum