DVCSmeetingV4

E08-025 Deuterium results
Hall A DVCS Collaboration Meeting - Friday 20 December 2013
Camille Desnault – Ph-D Student

2
Calorimeter calibration :
→ results using the π° method
→ comparison with Malek's results
Run quality :
→ HRS and Calorimeter problems during the data taking
(= discarding some runs)
Deuterium analysis :
→ Contamination subtraction
→ Including the fermi motion for the LH2 target's proton
→ LD2 – LH2 targets subtraction
→ Comparison with Malek's results (in applying the same cuts)
To get Deuterium results … (my work so far)

3
Calorimeter calibration using the π° method
Elastic calibration (ep → e'p') :
→ 3 Elastics calibrations (October 26th, November 17th, December 14th)
→The polarity of HRS is reversed to detect the proton, the Elastic calibration is not possible
during the data taking (= dedicated runs)
π° calibration (ep → e'p'π° → e'p'γγ) :
→ π° Calibration is possible during the data taking (= same experimental setup as the DVCS runs)
→ π° Calibration allows to calibrate the calorimeter for each day of the experiment (= Monitoring)
Calibration coefficients
Minimization of χ2
:
Theoretical energy Signal amplitude
Theoretical energy calculation from :
→ Electron energy
→ π° position
→ Assuming exclusive event (Mx2 cut)
We perform several iterations of calibration to stabilize the results

4
Calorimeter calibration using the π° method
If the calibration works, we expect to see an improvement on :
→ Mx2 (closer to the theoretical value : Mp2
= 0.88 GeV2
)
→ Minv (closer to the theoretical value : Mπ° = 0.135 MeV)
Mx2
= 0.88 GeV2
Minv = 0.135 MeV
Mx2
Minv
Blue dots :
Before calibration
Red dots :
After calibration
Red dotted lines :
Elastic calibrations

5
Calorimeter calibration (Comparison with Malek's results)
Difference in the cut applied for the comparison :
→ My cut :
A fixed 2-Dimensions cut in Mx2 vs Minv :
0.5 < Mx2 + 17.5 * Minv - 2.31 < 1.2
→ Malek's cut :
A variable 2-Dimensions cut in Mx2 vs Minv as a function of σMx2
and σMinv
at each iteration
Difference on the groups of runs used for the calibration
Minv = 0.135 MeV
Mx2
Minv
Mx2 = 0.88 GeV2
Conclusion :
→ the two calibrations
improve the Mx2 and
the Minv
→ the two calibrations
are close to each
other

6
Run quality (= discarding the problematic runs)
HRS problem (Low number of hits in one of the
PMT of the Cerenkov detector for one run)
Acquisition system problem (Dead Time problem
for one run)
Normal arrival times of signals in one block of the
Calorimeter for one run
Abnormal arrival times of signals in one block of the
Calorimeter for one run
Conclusion : 10% from the totality of the runs affected

7
Contamination subtraction : accidentals 1 cluster contribution
≈ 24 % of
accidentals
contribution
Mx2
= 1.15 GeV2
Accidentals are :
→ Photons not related to the trigger electron detected in the [-3, 3] ns clustering window
(= not coming from the vertex)
→ Uniform contribution in the time on the 128 ns of the acquisition window
To remove the accidentals contribution, we shift in time the clustering window :
from [-3, 3] ns to [-11, 5] ns (and [5, 11] ns)

8
Raw data
1-cluster events
(DVCS + π°)
2-clusters events
(π° data)
N0 : 0 cluster
N1 : 1 cluster
N2 : 2 clusters
Raw data = DVCS + Accidentals + π°
π°(ep→ e'p'π° → e'p'γγ) :
Contamination
2-clusters events
1-cluster events
Contamination when only 1 of the
two photons from the π° decay is
detected by the calorimeter
1-cluster events
(DVCS + π°)
π° random
decays
N1 : 1 cluster
Projection of the
photons on the
calorimeter surface
Contamination subtraction : π° contamination

9
Example of π° contamination subtraction
≈ 31 % of π°
contamination
In the Blue curve :
we have the real π° but also accidentals π°
We have to remove the accidentals π° contribution to the real π° to subtract
only the real π° from the raw data

10
Accidentals π° (3 types) :
→ A) 2 photons related to a π°, so the both in coincidence with themselves but not in coincidence
with the trigger electron
→ B) 2 photons not related to a π°, with one of them in coincidence with the trigger electron
→ C) 2 photons not related to a π°, and none of them in coincidence with themselves or with the
trigger electron
Arrival time of second
cluster (in ns)
Arrival time of first
cluster (in ns)
Camsonne A.
-5 ns
-5 ns
3 ns
5 ns
-3 ns
A
B
C
To remove the accidentals contribution, we select the clustering windows to :
A) [-11, -5] ns and [-11, -5] ns
B) [-3, -3] ns and [5, 11] ns
C) [-11, -5] ns and [5, 11] ns
Contamination subtraction : accidentals π° contribution
To subtract the
total accidentals
π° contribution :
A + B - C

11
→ LD2 target : 20% to 30% accidentals π° contribution
→ LH2 target : 13 % accidentals π° contribution
Example of the accidentals π° contribution with the Minv
Accidentals 2 clusters
Type A
[-11, -5] ns and [-11, -5] ns
Type B
[-3, 3] ns and [5, 11] nsAccidentals 2 clusters
Type C
[-11, -5] ns and [5, 11] ns
Accidentals π° contribution is not negligible, so it's necessary
to subtract this contribution to the total 2-clusters events.

12
We can see the difference on the number of events
when we apply the accidentals π° subtraction
15.4%
difference on
the total
number of
events
23% difference on
the number of π°
Example of the accidentals π° contribution with the Mx2

13
Global results after contamination subtraction for each target
DVCS = Raw data - Accidentals 1 cluster - (π° - Accidentals 2 clusters)
≈ 47 % of accidentals contribution + π°
contamination
≈ 49 % of accidentals contribution + π°
contamination
Mx2
= 1.15 GeV2
Mx2
= 1.15 GeV2

14
Fermi motion added to the LH2 target
→ Proton at rest in the LH2 target but not in the LD2 target
→ necessity to add the fermi motion to the LH2 target's proton for the target subtraction
→ The fermi motion is a smearing on the proton momentum and the proton mass to
take into account the initial motion of the proton in the nucleus
Fermi momentum (in GeV)
Distribution of fermi momentum

15
Global results after LD2-LH2 targets subtraction
Mx2
= 1.15 GeV2
Conclusion :
→Normalization by the charge of each run was performed to subtract the targets
→Fermi motion was included to the LH2 target data
We notice a shift of the Mx2 peak between the LD2 target and the LH2 target :
→ due to the calorimeter calibration, fermi motion, π° subtraction method … ?

16
Comparison of 2 parallel analysis for the contamination subtraction
(same cuts applied)
0.8% 2.7%
4.8% 2.1%
LD2 Target : Malek results (blue) / My results (red)

17
(without fermi motion)
0.2% 2%
13.7%
7.4%
LH2 Target (without fermi motion) : Malek results (blue) / My results (red)
Too big difference (7.4%) due to
the π° subtraction method :
Improvement in progress

18
(without fermi motion)
0.2% 2%
13.7%
7.4%
LH2 Target (without fermi motion) : Malek results (blue) / My results (red)

19
(with fermi motion)
0.3% 2.5%
17.5%
9.3%
LH2 Target (with fermi motion) : Malek results (blue) / My results (red)

20
Conclusion :
→We notice the same shift of the Mx2 peak between the LD2 target and the
LH2 target for Malek results
Comparison of the LD2-LH2 targets subtraction

21
Comparison of the 2 analysis for the contamination subtraction to
improve
Investigation of the relative calibration of the targets (= shift of the Mx2
peak between LD2 and LH2)
Analysis of the kinematic kin2Low
Studying the impact of the cuts variations on the Mx2
To get Deuterium results … (the next tasks)

23
Contamination subtraction to the DVCS (ep(n) → e'p'(n')γ)
Exclusive (ep → e'p'° → e'p')
Inclusive (ep→ e'(X)π° → e'(X)γγ)
Associated-DVCS (ep → e'p'(X)γ) :
→Example : ep → e'p'π°γ , ep → e'p'π+
π-
γ ...
→First channel inclusive π° (ep → e'p'π°γ) with a missing mass square :
Mx2
= (Mp
+ Mπ°
)2
= 1.15 GeV2
We apply a cut
on the Mx2
(Mx2
<1.15GeV2
)
to discard the
inclusive π°
events from the
raw data.
Mx2
= 1.15 GeV2

24
cluster (in ns)
cluster (in ns)
-3 ns
-3 ns
3 ns
3 ns
2 different
[-3, 3] ns
Coincidence
windows
Accidentals :
→ DVCS photons in the [-3, 3] ns coincidence window
→ Photons not related to the trigger electron are detected in the [-3, 3] ns coincidence window
(= not coming from the vertex)
→ Uniform contamination in the time on the 128 ns of the acquisition window
Camsonne A.

25
Accidentals 1 cluster :
→ 1 photon detected in the coincidence window →[-11, -5] ns or [5, 11] ns
Accidentals 2 cluters (3 types) :
→ A) 2 photons related to a π°, so the both in coincidence →[-11, -5] ns and [-11, -5] ns
→ B) 2 photons not related to a π°, with one of them in coincidence →[-3, -3] ns and [5, 11] ns
→ C) 2 photons not related to a π°, and none of them in coincidence →[-11, -5] ns and [5, 11] ns
We shift in time the 6 ns acquisition window to take only accidentals events
cluster (in ns)
cluster (in ns)
Camsonne A.
-5 ns
-5 ns
3 ns
5 ns
-3 ns
A
B
C
5 ns

26
≈20% to 30%
Check of the accidentals 2 clusters subtraction with the Minv
Type A
[-11, -5] ns and [-11, -5] ns
Type B
[-3, 3] ns and [5, 11] ns
Type C
[-11, -5] ns and [5, 11] ns
Contamination on the Minv according to the energy
thresholds for the clustering and the photons
≈13%

27
Cross check of the LD2-LH2 targets subtraction
12.2%
LD2 – LH2 : M. Ben Ali results (blue) / My cross check results (red)
the LH2 target's results to
improve :

DVCSmeetingV4

More Related Content

What's hot (18)

Similar to DVCSmeetingV4 (20)

DVCSmeetingV4