Department of Astronomy, University of Padova Department of Physics, University of Padova
Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )
description
Transcript of Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )
![Page 1: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/1.jpg)
1
Systematic uncertainties for the inelastic J/y PHP analysis
A. Bertolin (INFN-Padova)R. Brugnera (Padova Uni.)
18/5/2012
![Page 2: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/2.jpg)
2
s vs z for different pt ranges: diffractive background
• diffractive component quantified by comparing the z(rec.) distribution measured in data with an HERWIG (signal) + EPSOFT (background) MC mixture
• increase / decrease the EPSOFT fraction while keeping a reasonable agreement between data and MC mixture
• redo all calculations
only one bin, at low pt and high z, with cross section variations > ± 5 %
![Page 3: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/3.jpg)
3
s vs z for different pt ranges: hadronic energy resolution
• z = f (E-Pz(J/y),E-Pz(ZUFO))• using the true J/y kinematic
work out the true E-Pz• decrease or increase the
difference E-Pz(ZUFO) - E-Pz by 20 % event by event
• redo all calculations
variations < ± 5 %may be 20 % seems “large” but even with this “large” value the results are stable … 20 % is also the value we used in the previous papers (no jets, visible hadronic system is soft …)
![Page 4: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/4.jpg)
4
s vs z for different pt ranges: BMUI chambers efficiency
• efficiency in data computed from two tracks J/y events, known within some statistical uncertainties (due to the finite number of two tracks J/y events)
• data efficiency plugged into the MC at the analysis level (eaze)
• decrease or increase the efficiency for the barrel section, rear section unchanged
• redo all calculations
variations in the range ± 5 %, the size of the stat. uncertainties on the efficiencies
![Page 5: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/5.jpg)
5
s vs z for different pt ranges: RMUI chamber efficiency
• efficiency in data computed from two tracks J/y events, known within some statistical uncertainties (due to the finite number of two tracks J/y events)
• data efficiency plugged into the MC at the analysis level (eaze)
• decrease or increase the efficiency for the rear section, barrel section unchanged
• redo all calculations
variations in the range ± 5 %, the size of the stat. uncertainties on the efficiencies, for z > 0.75variations much smaller for z < 0.75
![Page 6: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/6.jpg)
6
s vs z for different pt ranges: l helicity parameter
• l related to the polar distribution of the m in the J/y rest frame
• l = 0: isotropic• l is weekly dependent on z
and pt• from the ZEUS measurements
(HERA I+II) we know that | l | < 0.5 “everywhere”
• l = ± 0.5 at the event level• redo all calculations
largest sys. error of the analysisunavoidable (even if you go to p p instead of PHP)
![Page 7: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/7.jpg)
7
s vs z for different pt ranges: n helicity parameter
• n related to the azimuthal distribution of the m in the J/y rest frame
• n = 0: isotropic• n is weekly dependent on z
and pt• from the ZEUS measurements
(HERA I+II) we know that | n | < 0.5 “everywhere”
• n = ± 0.5 at the event level• redo all calculations
largest sys. error of the analysisunavoidable (even if you go to p p instead of PHP)
![Page 8: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/8.jpg)
8
s vs z for different pt ranges: HERWIG MC pt spectrum
• the HERWIG MC J/y pt spectrum is reweighted to the data
• can make the MC spectrum harder or softer while keeping a reasonable agreement between data and MC
• additional weight given by exp(a pt2) at the event level
• redo all calculations
small effectas expected based on the experience with the past papers
![Page 9: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/9.jpg)
9
s vs z for different pt ranges: EPSOFT MC Mx spectrum
• the EPSOFT MC Mx spectrum can be fitted with the function 1/Mx
• E(FCAL) is the observable mostly sensitive to the Mx spectrum
• can make the MC spectrum harder or softer while keeping a reasonable agreement between data and MC
• additional weight given by 1/Mxa at the event level
• redo all calculations
small effectMx has a small impact on z(rec) i.e. E-Pz in FCAL is small
![Page 10: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/10.jpg)
10
s vs z for different pt ranges: EPSOFT MC W spectrum
• the EPSOFT MC Wgp spectrum is flat … unphysical …
• reweight to a linear dependence: observe good agreement between data and MC for 2 tracks events at high z (diffractive background rich region we cut out in the analysis)
• can make the MC spectrum harder or softer while keeping a reasonable agreement between data and MC
• additional weight given by Wa at the event level
• redo all calculations
small effectW has a small impact on z(rec) with the kinematic of diffractive events
![Page 11: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/11.jpg)
11
s vs z for different pt ranges: EPSOFT MC pt2 spectrum
• the EPSOFT MC pt2 spectrum was set to -1 and -0.5 at the generation level and the two samples added
• observe good agreement between data and MC for 2 tracks events at high z (diffractive background rich region we cut out in the analysis)
• can make the MC spectrum harder or softer while keeping a reasonable agreement between data and MC
• additional weight given by exp(a pt2) at the event level
• redo all calculations
small effectsizable only for z > 0.75 at low pt
![Page 12: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/12.jpg)
12
s vs z for different pt ranges: invariant mass fit
• invariant mass procedure is fitting the non resonant background away from the mass peak with a smooth function
• an invariant mass window is defined for the signal: [2.85,3.3]
• count the events in the window and subtract the integral of the non resonant background function over the signal window
• change the window by ± 50 MeV (both in data and MC)
• redo all calculations
at most a10 % effect in the low z bins, there the S/B ratio is decreasing with respect to the bins with z > 0.45
![Page 13: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/13.jpg)
13
s vs z for different pt ranges: H1 track multiplicity cut
• H1 analysis: ask for at least 5 vertex track, with pt > 125 MeV and | h | < 1.75 and DO NOT consider any diffractive background after this
• redo the analysis “à la H1”
two bins with 20 % variations, one at high z and one at low z
this is testing the diffractive background procedure but also how well the track multiplicity cut is corrected for via MC
![Page 14: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/14.jpg)
14
s vs pt2 for different z ranges: diffractive background
• same steps shown previously
• same steps done for DIS11
![Page 15: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/15.jpg)
15
s vs pt2 for different z ranges: hadronic energy resolution
![Page 16: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/16.jpg)
16
s vs pt2 for different z ranges: BMUI RMUI chamber efficiency
![Page 17: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/17.jpg)
17
s vs pt2 for different z ranges: l n helicity parameters
![Page 18: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/18.jpg)
18
s vs pt2 for different z ranges: HERWIG MC pt spectrum
![Page 19: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/19.jpg)
19
s vs pt2 for different z ranges: EPSOFT MC Mx and W spectrum
![Page 20: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/20.jpg)
20
s vs pt2 for different z ranges: EPSOFT MC pt2 spectrum
![Page 21: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/21.jpg)
21
s vs pt2 for different z ranges: invariant mass fit
![Page 22: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/22.jpg)
22
s vs pt2 for different z ranges: H1 track multiplicity cut
![Page 23: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/23.jpg)
23
2S to 1S cross sections ratios
stat. uncertainty of about 15 % due to the small number of 2S events, unavoidable …
sys. sources:1. diffractive background: cancel in the ratio2. hadronic energy resolution: expect cancellations in the ratio, see next
slide3. BMUI chambers efficiency: cancel in the ratio, same hardware for 1S and
2S4. RMUI chambers efficiency: cancel in the ratio, same hardware for 1S and
2S5. helicity parameter l: cancel in the ratio6. helicity parameter n: cancel in the ratio7. HERWIG MC pt spectrum: tiny for 1S, furthermore will cancel in the ratio8. EPSOFT MC Mx spectrum: tiny for the 1S, furthermore will cancel in the
ratio9. EPSOFT MC W spectrum: tiny for the 1S, furthermore will cancel in the
ratio10. EPSOFT MC pt2 spectrum: tiny for the 1S, furthermore will cancel in the
ratio11. invariant mass fit: see next to next slide
![Page 24: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/24.jpg)
24
2S to 1S cross sections ratio vs pt: hadronic energy resolution
red: statistical uncertaintyblack: sys. uncertainty on E-Pz
negligible due to cancellations …
![Page 25: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/25.jpg)
25
2S to 1S cross sections ratio vs pt: invariant mass fit
• red: statistical uncertainty• black: sys. uncertainty on the
1S fitting range (± 50 MeV for both data and MC)
insignificant due to large S/B in the phase space selected for the 2S to 1S ratio
• red: statistical uncertainty• black: sys. uncertainty on the
2S fitting range (± 50 MeV for both data and MC)
smaller than the stat.
![Page 26: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/26.jpg)
26
2S to 1S cross sections ratio vs W: hadronic energy resolution
red: statistical uncertaintyblack: sys. uncertainty on E-Pz
negligible due to cancellations …
![Page 27: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/27.jpg)
27
2S to 1S cross sections ratio vs W: invariant mass fit
• red: statistical uncertainty• black: sys. uncertainty on the 1S
fitting range (± 50 MeV for both data and MC)
insignificant due to large S/B in the phase space selected for the 2S to 1S ratio
• red: statistical uncertainty• black: sys. uncertainty on the
2S fitting range (± 50 MeV for both data and MC)
smaller than the stat.
![Page 28: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/28.jpg)
28
2S to 1S cross sections ratio vs z: hadronic energy resolution
red: statistical uncertaintyblack: sys. uncertainty on E-Pz
negligible due to cancellations …
![Page 29: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/29.jpg)
29
2S to 1S cross sections ratio vs z: invariant mass fit
• red: statistical uncertainty• black: sys. uncertainty on the
1S fitting range (± 50 MeV for both data and MC)
insignificant due to large S/B in the phase space selected for the 2S to 1S ratio
• red: statistical uncertainty• black: sys. uncertainty on the
2S fitting range (± 50 MeV for both data and MC)
smaller than the stat.
![Page 30: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/30.jpg)
30
p flow along and against the J/y direction
stat. uncertainty: most unfavorable case, small (< 5%) for small values of p flow and very large (> 20 %) for large values of p flow … the bin widths are already increasing as the p flow increases … can not optimize more …shape measurement: both data and MC predictions normalized to 1
sys. sources:1. diffractive background: evaluated2. hadronic energy resolution: evaluated3. BMUI chambers efficiency: cancel after normalizing to 14. RMUI chambers efficiency: cancel after normalizing to 15. helicity parameter l: evaluated6. helicity parameter n: evaluated7. HERWIG MC pt spectrum: evaluated8. EPSOFT MC Mx spectrum: evaluated9. EPSOFT MC W spectrum: evaluated10. EPSOFT MC pt2 spectrum: evaluated
all uncertainties
of the MC model
![Page 31: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/31.jpg)
31
p flow along and against the J/y direction: diffractive background
• red: statistical uncertainty• black: sys. uncertainty due to the diffractive
background
![Page 32: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/32.jpg)
32
p flow along and against the J/y direction: hadronic energy resolution
• red: statistical uncertainty• black: sys. uncertainty due to E-Pz(ZUFO)
resolution
![Page 33: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/33.jpg)
33
p flow along and against the J/y direction: l helicity parameter
• red: statistical uncertainty• black: sys. uncertainty due to the l helicity
parameter
![Page 34: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/34.jpg)
34
p flow along and against the J/y direction: n helicity parameter
• red: statistical uncertainty• black: sys. uncertainty due to the n helicity
parameter
![Page 35: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/35.jpg)
35
p flow along and against the J/y direction: HERWIG MC pt spectrum
• red: statistical uncertainty• black: sys. uncertainty due to the HERWIG MC pt
spectrum
![Page 36: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/36.jpg)
36
p flow along and against the J/y direction: EPSOFT MC Mx spectrum
• red: statistical uncertainty• black: sys. uncertainty due to the EPSOFT MC Mx
spectrum
![Page 37: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/37.jpg)
37
p flow along and against the J/y direction: EPSOFT MC W spectrum
• red: statistical uncertainty• black: sys. uncertainty due to the EPSOFT MC W
spectrum
![Page 38: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/38.jpg)
38
p flow along and against the J/y direction: EPSOFT MC pt2 spectrum
• red: statistical uncertainty• black: sys. uncertainty due to the EPSOFT MC pt2
spectrum
![Page 39: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/39.jpg)
39
sys. errors are not visible in some binsstat. are dominant
![Page 40: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/40.jpg)
40
errors are mostly sys. at low pt2 and mostly stat. at high pt2
![Page 41: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/41.jpg)
41errors are mostly sys. at high z and mostly stat. at low
z
![Page 42: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/42.jpg)
42
uncertainties of the MC model: boxes
![Page 43: Systematic uncertainties for the inelastic J/ y PHP analysis A. Bertolin (INFN- Padova )](https://reader036.fdocuments.in/reader036/viewer/2022062520/5681658f550346895dd85bc7/html5/thumbnails/43.jpg)
43
Conclusions
the systematic uncertainty evaluation for the inelastic J/y PHP analysis have been presented in detail