MINOS 1 Beam e ’s from antineutrinos David Jaffe and Pedro Ochoa September 27 th 2007 ...
-
date post
20-Dec-2015 -
Category
Documents
-
view
214 -
download
0
Transcript of MINOS 1 Beam e ’s from antineutrinos David Jaffe and Pedro Ochoa September 27 th 2007 ...
1
MINOS
Beam e’s from antineutrinos
David Jaffe and Pedro Ochoa
September 27th 2007
Preliminaries Data & MC Expected sensitivities Preliminary results Outlook
− Preliminary Results −
2
MINOS
3) Fit resulting distribution (top right) using shapes from the MC scaled by parameters parHE and parLE:
Preliminaries
x parLEx parHE
pHE-LE+C (simulated)
2) Apply a correction C for
from and parents
C=pHE-)LE
(+)pHE (+)LE
Goal is to measure the antineutrinos from + decay (brothers of beam e’s)
Antineutrinos from + are the most affected when changing the beam configuration. The technique for the measurement is:
1) Scale pHE and LE DATA to same POT and subtract
3
MINOS
Measurement can also be done with pME data (minos-doc 2706)
Systematic error from background uncertainty is practically negligible on LE
Errors in +)LE determination from horn & target systematics are in the order of ~5-10%
Systematic errors associated with hadron production uncertainties are yet to be determined. Some on this at the end.
Systematics were addressed in minos-docs 2909 & 3230. In particular:
Preliminaries
More details on the method in minos-doc 2783
Statistical error with 1.6x1019 POT of pHE data was expected to be ~15% (minos-doc 3230)
4
MINOS
Data & MC used:
DATA le010z185i runI: 2.46x1019 POT
DATA le010z185i runII: 2.21x1019 POT
DATA le250z200i runII: 1.41x1019 POT
MC le010z185i: 4.44x1019 POT
MC le250z200i: 1.19x1019 POT
Data & MC
cedar_phy
daikon-cedar
The le010z185i data used corresponds to the same data used in the latest CC analysis and is evenly distributed along that period (runI + beginning of runII) Thanks to Tricia for these pans !
POT values for DATA are after “good beam” cuts.
all available !
5
MINOS
Data & reweighted MC antineutrino spectra:
le010z185i
le250z200i
All MC parent parentL parent+ parentBackground
All MC parent parentL parent+ parentBackground
le010z185i
le250z200i
data/MC (no SKZP)data/MC
data/MC (no SKZP)data/MC
Note: SKZP “PiMinus_CedarDaikon”, run I configuration (more details in slides 16-17)
6
MINOS
Expected sensitivities
Before fitting the data tested the routine with fake data.
Used smoothed MC histograms (shown in grey) to construct scenario.
(+)pHE(+)LE
(-,K-)LE (-,K-)pHE
Background LE Background pHE
Fake data is produced by statistically fluctuating the histograms.
The fit is done “manually” (no Minuit)
7
MINOS
Wassup with the bias
Χ2best fit = 28.6
Scenario 1: “best possible”
LE DATA POT pHE DATA POT LE MC POT pHE MC POT
∞ 1.41x1019 ∞ ∞
One fake experiment
Best fit
Distribution of 1000 fake experiments:
Best possible stat. error
Accuracy of contour confirmed by distribution
of fake experiments
90% C.L.
68% C.L.
This is the best measurement we can do with the current amount of pHE data:
8
MINOS
Scenario 2: “now”
LE DATA POT pHE DATA POT LE MC POT pHE MC POT
4.67x1019 1.41x1019 4.4x1019 1.19x1019
Χ2best fit = 28.5
One fake experiment
Best fit
90% C.L.
68% C.L.
Distribution of 1000 fake experiments:
This is the kind of measurement we expect to do now:
9
MINOS
Prob(25.4,28) = 60.6%
χ2=16.9
Preliminary results
Χ2
Our results, with statistical uncertainties only:
pHE(data) – LE(data) + C(MC)
Best fit
Nominal case(parLE=parHE=1)
Best fit: parLE=1.525 ± 0.37parHE=0.522 ± 0.19
90% C.L.68% C.L.
Χ2best fit = 25.4
pHE(data) – LE(data) + C(MC)
Prob(42.3,28) = 4.1%
10
MINOS
par LE par HE χ2best fit
Normal 1.52 0.52 25.4
No SKZP 1.85 0.42 22.8
No Bkgd substraction
1.52 0.77 27.5
SKZP
Consistent with expectation as described in minos-doc 2909
Preliminary results
Fit results in other conditions:
Difference with “No SKZP” case stems mainly from ~15% difference in low energy (< 10 GeV) region of C:
How much of this change is attributed to hadron production only by SKZP, and how much to other effects?
ratio
11
MINOS
Outlook
How much more can the result be improved? (without taking more data)
For an infinite amount of pHE MC, LE data
and LE MC
→max. goal
Contours calculated assuming same best fit value and 1.41x1019 POT of pHE data:
For an infinite amount of LE data and LE MC, with current
amount of pHE MC (1.3x1019 POT)
LE data =2x1020 POT LE MC =2.5x1020 POT
with current amount of pHE MC
LE data & MC POT of ~2x1020 POT is already “infinite” for our purposes.
With ~5 times more pHE MC can get close to the max. goal
LE data =2x1020 POT LE MC =2.5x1020 POT pHE MC = 7x1019 POT
12
MINOS
Summary
Our preliminary results confirm the SKZP prediction of to 1.4(statistics only)
A couple of things left to do:
Run with more data & MC. Need ~5 times more pHE MC. Assign a systematic error to our measurement
A much smaller systematic error could be obtained by doing the measurement with pME data, as shown in minos-doc 2706.
13
MINOS
Backup
14
MINOS
le010z185i
le250z200i
Data & raw MC antineutrino spectra:
All MC parent parentL parent+ parentBackground
All MC parent parentL parent+ parentBackground
le010z185i
le250z200i
data/MC
data/MC
15
MINOS
Applied SKZP to the MC:
All MC parent
L parent
parent
+ parent Background
raw MCSKZP MC le010z185i
ratio
ratio
ratio
ratio
ratio
ratio
16
MINOS
le250z200i bkgd (Reweighted for runI)
le250z200i bkgd(Reweighted for runII)
difference
Reweighting in runI or in runII modality does not change the antineutrinos at all.
What about the background (made mostly of CC ’s)?
Difference between the plots above is tiny !
Reweighted everything for runI.
→ What about the two running periods?
17
MINOS Scale component by best fit values and compare with data:
Preliminary results
le010z185i
le250z200i
All MC parent parentL parent+ parentBackground
All MC parent parentL parent+ parentBackground
le010z185i
le250z200i
data/MC (before)data/MC (after scaling)
data/MC (before)data/MC (after scaling)
18
MINOS
χ2=35.2
Preliminary results (no SKZP)
Real data
Best fit
Real data
Nominal case(parLE=parHE=1)
Best fit: parLE = 1.85 parHE = 0.44
19
MINOS
parLE = 1.495 ± 0.37
parHE = 0.502 ± 0.19
Χ2 = 1.0 contour
20
MINOS
Max goalWith 1.41x1019 POT of le250z200i data,
stat error in only le250z200i data-MC
With 1.41x1019 POT of le250z200i data, 2x1020 POT of le010z185i MC and 2.5x1020 POT of le010z185i Data
With 1.41x1019 POT of le250z200i data, 2x1020 POT of le010z185i MC and
2.5x1020 POT of le010z185i Data, and 4x1019 POT of le250z200i MC
Outlook
21
MINOS
With 1.41x1019 POT of le250z200i data, 2x1020 POT of le010z185i MC and
2.5x1020 POT of le010z185i Data, and 7x1019 POT of le250z200i MC
With 1.41x1019 POT of le250z200i data, 2x1020 POT of le010z185i MC and
2.5x1020 POT of le010z185i Data, and 1x1020 POT of le250z200i MC
With 1.41x1019 POT of le250z200i data, 2x1020 POT of le010z185i MC and
2.5x1020 POT of le010z185i Data, and 1.5x1020 POT of le250z200i MC