Preliminary Study of Birk's Constant
Introduction
We have seen that there are features in shower response/shape that are
not well modelled.
In particular there are discrepancies in the "fracvars" type variables
of order 20-30%.
Changes in:
- hadronization model (carrot/daikon)
- reconstruction
(birch/cedar)
- hadron codes (gcalor/gheisha)
have not been
sufficient to completely describe the differences.
Example: GCALOR vs GHEISHA
- Nue Preselection cuts applied
- fiducial cuts
- tracklength < 25
- trklikeplanes < 18
(n.b. plots are mis-labled in this figure - cedar => gheisha)
What other effects can we explore?
Non-linearities in strip response could also introduce this kind of
effect.
The two I can think of are PMT/electronics non-linearities and
Birk's suppression in the scintillator.
Birk's Suppression
I'm studying Birk's suppression by generating MC samples with
+/-30% changes in the standard carrot Birk's constant.
Standard MC: 0.13 cm/GeV
(c.f. From Trish/Mike CalDet Studies: 0.1 m/GeV)
For this study I'm using values 0.1, 0.13, 0.16 m/GeV
So far I've only managed to process a handful of files due to some
problem with lsf/dcache? that's causing my jobs to be killed
prematurely.
POTS used for the study:
- Data : 1.274e18
POT
- Birk 0.13 : 5.427e17 POT
- Birk 0.10 : 3.174e17 POT
- Birk 0.16 : 3.789e17 POT
Some preliminary distributions follow.
I have only applied nue preselection cuts as listed above.
Shower PH:
- General shape in data is described by MC at 10% level.
- Changing Birk's constant only makes small changes - data/mc ratio
remains ~same.
Fraction of energy in 3 largest
consecutive planes:
- Changes due to Birk's constant become more obvious
- Clearly a higher value can shift the peak
- Large discrepancy remains
Event length:
- Event length distribution is partially responsible for
discrepancy in fracvars_3plane variable above
- Small changes to shape can be observed with Birk's constant
Fraction of Charge in Largest 4 strips:
ANN PID:
Summary
- My preliminary conclusion is that Birk's suppression cannot
account for
the effect, although it's value has non-negligable effects on the
distributions.
- Worth asking Nathaniel why he chose 0.13 over the 0.1 value used
at CalDet.
- Next would like to look at the effect of Intranuclear
rescattering on
these distributions.