Long
baseline Neutrino experiments
Index:
Related pages:
Write-ups:
FNAL REPORT (May 23, 2007)
Current work
Organized by date:
January 2008
January 2008
January 24th 2008:
Matching the FNAL plot about liquid Argon detector, using
SK Monte Carlo.
My assumptions:
- Assuming that 300kt of water is equivalent to 100kt LAr
- There is no NC background in LAr
- I keep a signal efficiency of 80%
- Project X assumes 60 GeV protons and 120 X 10^20 POT
I did not have flux for 60 GeV protons so I used the flux I had
for 120GeV protons, but I kept 120x10^20 POT
(NB compare to the work presented in my report of May 23, 2007
This corresponds to increasing the number of POT
from 30x10^20 to 120x10^20
and removing the NC background, and having a signal
efficiency of 80% instead of 40%,
nothing else changed.)
Spectrum for sin^2(2th_13)=0.01 and delta=0 normal mass hierarchy, for 120x10^20 POT
Neutrino
Antineutrino
Contour plots:
mass hierarchy (eps)
CP violation (eps)
May 2007
May 13th 2007:
Matching the FNAL plot:
Input flux: he flux file is http://www.phy.bnl.gov/~bishai/nwg/fluxes
and the files is
wble120e250i002dr380dz1300km_12oa_flux.txt
Normalized for 300Kton, 1MW beam, 30E+20 POT and 1300km
Key issue: I had to place the likelihood such that I keep 40% of the signal in each energy bin.
(Before I was always placing the cut at 0, go see there)
So here is my plot and here is their
(their plot comes from
http://minos.phy.bnl.gov/%7Emdier/bnl/nwg/longbaseline/120GeV_0.5deg/WCC/30+30/spectra/spectra_WCC_120GeV_12km_30p_s0.04_nu_no_app.png
And I think this is a very nice agreements!!!
November 2006
Nov 13th 2006:
More plots:
delta: pink
= -135 red = 0 blue = 45
green is nue
background
Normal Hierarchy:
Here is a plot using sin^2(2th_13)=0.01 and using spectrum technic for
the background:
eps or gif
Here is a plot using sin^2(2th_13)=0.05 and using spectrum technic for
the background:
eps or gif
Inverted Hierarchy:
Here is a
plot using sin^2(2th_13)=0.01
and using spectrum technic for
the background:
eps or gif
Here is a plot using sin^2(2th_13)=0.05 and using spectrum technic for
the background:
eps or gif
Nov 10th 2006:
Producing plots:
Here are plots trying to reproduce
Mary's plot for nue appearance:
(sin2(2th13)=0.05;
delta: pink
= -135 red = 0 blue = 45 ;
green is nue
background:
- using spectrum technics (gif)
(eps)
Integrate number for BG are in [ ] in table below
- using MC technics (gif)
(eps)
Intgerated number for BG are in ( ) in table below
| (for 0.5-10GeV) |
Fanny |
Mary sin^2(2th_13)=0.05 |
| delta=0 tot sig+bg (sig only) |
1375
(1124) [854] |
1460
[950] |
| delta=45 tot sig+bg (sig only) |
1108
(857) [587] |
1212
[702] |
| delta= - 135 tot sig+bg (sig only) |
1885
(1634) [1364] |
1887
[1377] |
|
|
|
| all background (MC method) |
251.8 |
NAN |
| background beam nue |
46.95 |
NAN |
|
|
|
| all background [spectrum method] |
521.1 |
510 |
| background beam nue |
317.4 |
NAN |
Nue background (my method vs.
spectrum method)
| Efficiency (nue) |
0.-0.35 |
0.35-0.85 |
0.85-1.5 |
1.5-2 |
2-3 |
3-4 |
4-5 |
5-higher |
| precuts (true) |
0.94 |
0.80 |
0.61 |
0.46 |
0.36 |
0.31 |
0.25 |
0.20 |
| precuts (rec) |
0.38 |
0.51 |
0.42 |
0.37 |
0.41 |
0.45 |
0.49 |
|
|
|
|
|
|
|
|
|
|
| Likeli (rec) |
0.87 |
0.80 |
0.79 |
0.73 |
0.73 |
0.70 |
0.78 |
0.78 |
|
|
|
|
|
|
|
|
|
| all (rec) (QE only) |
0.78 |
0.77 |
0.76 |
0.65 |
0.64 |
0.58 |
0.62 |
0.64 |
| all(rec) (allCC) |
0.33 |
0.41 |
0.33 |
0.27 |
0.30 |
0.31 |
0.34 |
0.41 |
| all (true+rec) (all CC)
??? |
0.82 |
0.64 |
0.47 |
0.33 |
0.26 |
0.21 |
0.19 |
0.15 |
I need to figure out
exaclty what we did for the signal.
Here is what I think today (might change
later...)
flux* sigma_QE*eff_QE + flux* sigma_QE *
eff_QE * (NQE/QE)
and this is ok only if NQE/QE = [flux *
sigma_NQE * eff_NQE] /[ flux *
sigma_QE * eff_QE]
Need to be checked !
Nov 9th 2006:
Plots and integrated numbers:
Finally things seem to be working. Here
are some event spectrum plots and some integerated numbers.
(To be compared with Milind's and Mary's
plots and numbers)
Nue appearance
Anti-nue
appearance
Where the black line is signal +
background for delta=135°
and the red line is delta=0° and
the blue line is delta=45°
Green
is the nue
background, purple is the NC background
and blue is the CCnumu background
| (for 0.5-10GeV) |
Milind (from
paper, p.13) |
Fanny |
Mary sin^2(2th_13)=0.05 |
| delta=0 tot sig+bg (sig only) |
1293
(727) |
1113
(881) |
1460
(950) |
| delta=45 tot sig+bg (sig only) |
1114
(548) |
827
(575) |
1212
(702) |
| delta=135 tot sig+bg (sig only) |
1711 (1145) |
1626
(1374) |
1887
(1377) |
|
|
|
|
| all background |
566 |
251.8 |
510 |
| background beam nue |
272 |
46.95 |
NAN |
Nov 7th 2006:
Concern about the oscillation
probability
Here is
a plot of the nue appearance for:
black: sin2(2th13)=0.103 delta=0
red: sin2(2th13)=0.038 delta=0
green: sin2(2th13)=0.038 delta=135
Nov 6th 2006:
Update on number summary:
The
differences that I saw in the flux tables are only due to binning. With
the 75MeV bins it is impossible (using just paw) to get a perfect
number for the 1-3 GeV since there is no bin starting at 1 GeV.
The following plot overlay the two
binnings. And convince me that the effect cannot be due to that.
Nov 2nd 2006:
Numbers
summary:
Trying
to keep tracks of what my numbers are:
I
think my event spectrum before oscillation and detector simulation is
good because I can (nearly perfectly, up to 1 % ) reproduce the numbers
which are on Milind's webpage
(version 14):
|
Milind's QE numu |
Fanny's QE numu |
Errors |
| 0-15GeV |
11766.77 |
11630 |
off by 1% |
| 0.5-1 GeV |
1628 |
1400 |
low by 16% |
| 1-3 GeV |
6439 |
6682 |
high by 3% |
| 3-15 GeV |
3150 |
3379 |
high by 7% |
NB those are numbers where the
distance is taken to be 2450km and for 5 years of running.
Need to check where those differences come from, but probably not for
overall factors since they vary as a function of energy.
Now let's check the numbers after I apply smearing, and likelihood
(output of resolution80nonqe.F)
| (for 0-10GeV) delta=135 |
Milind (from paper, p.13) |
Fanny |
Errors |
| total signal+background |
1711 |
1392 |
low by 21% |
| signal only |
1145 |
1218 |
high by 10% |
| all background |
566 |
174.1 |
low by factor of 3 |
| background beam nue |
272 |
20.19 |
low by factor of 13 |
Here is a plot where similar to the
one that you get when clicking on (Fanny) in the table above but where
I added a histogram where I don't consider the detector effect
(semaring and likelihood) for the signal.
Work on ratio of CCQE and CCnonQE
NB: in order to include CCnonQE in our signal we compute the ratio of
CCQE to CCnonQE using superk MC and then we just apply it to our CCQE
sample.
I had never updated the part of the code
which deals with the ration of CCQE and CCnonQE in evspec.F (in /codes)
I had assumed that the fit which was done for 0-4GeV would still be ok
for 0-10GeV. Since now I need to look for small modifications, I redid
the fit.
The 1st thing to know it that I couldn't reproduce exactly Okumura-san
fit parameters:
data
a/-0.4601E-01, 0.3646E-01, 0.5687, -0.2435, 0.3106E-01/
but I got:
-0.98390E-01, 0.31835, 0.16382,
-0.70817E-01, 0.74570E-02
which give very similar results if plotted, as you can see there.
(red
is my fit, green is
Okumura-san's fit)
Then I extended the fit to 0-10GeV and
redid the event spectrum plot.
(The black line is using the new ratio,
the red line is using the old
ratio.)
This doesn't seem to explain why I have
so much more signal events.
Nov 1st 2006:
Trying to understand why my peak for
signal is around 2 GeV while the peak on p.13 of hep-ex/0608023 is
around 2.5.
Guess 1:
The energy smearing
that I apply to signal is asymmetric, so I compared my usual plot
versus a signal with:
Note in both plots above the
background is taken with reconstructed energy (and thus smeared).
None of those effect seems to be big enough, but if I use a background
where I used the true energy then my peak shifts back as it can been
seen here.
In that plot:
black line is signal (no energy smearing)
+ background (using true energy)
red line is signal
(with assymetric smearing) + background (using reconstructed energy)
colored box are
background using true energy.
October 2006
Oct 24th 2006:
Made a new webpage which summarize every normalization and units
used in T2KK and other longbaseline studies.
September 2006
Sept 4th 2006:
Notice I made a mistake before in the nue
and nuebar plots, the link below (Sept 1st) are now corrected.
Made plot of sigma/E for nue and nuebar: here
There is a factor of ~2 discrepancy between this plots and the plot of
the combined paper ((hep-ex/050164)
I haven't found the reason yet, nor how to explain it by just change of
units.
October
24th:
The
reason is that for T2K cross section, the fraction of neutron (protons)
is included in the cross-section so that you just need to know the
number of nucleons in the target. (See
there)
But for FNAL and
BNL cross-section this is not the case.
Sept 1st 2006:
Plots of the numu flux for NUMI beam 120GeV protons (neutrinos/GeV/m^2/proton-on-target at 1 km)
Cross-section (nue and nue_bar) (unkown units!!!!
(corrected on Sept 4th, binning mistake))
the hbk files containing those
plots is: crsqe_flux.hbk
NB: originally those plots where binned in 75 MeV
bin and I changed it to 10 MeV bin. I did check the normalization
during that
change so I am
pretty sure that those plots are equivalent to the orignal ones.
In the flux plots there is a few
negative values at the very beginning and at the very end, those are
due to the interpolation
that I had to do to rebin. (since
they appear only at the very beginning and at the very end, I hope this
is ok, I might have to
fix that manually to make them positive in case in
creates bugs later).
The original files I used are: 28gev_on_numi.hbk
crsqe.txt
both coming from: http://nwg.phy.bnl.gov/~diwan/nwg/hst/
How to use the code
As of end of
October 2006
Starting from
the cross-section files and the flux files given by Diwan, here I
describe the process I go through to be able to use the T2KK
oscillation analysis code. Every piece of code is currently (061024)
stored in /net/sukatmd1/work23/fdufour/oscillation/t2kk-2006/
with making event rate histograms in evspectrum_10gev
For units and
normalization questions go see there
(p.186
of notebook 1)
Step 1 :
Make flux and
cross-section files so that they range from 0 to 10 GeV in 10 MeV bins
Input files:
Flux file
28gev_on_numi.hbk for FNAL
numuv14.hst for BNL (used those)
Cross-section file
crsqe.txt for both cases (this is just cross-section, no neutron fraction included)
Output files:
Flux and cross-section in same file: bnl14_numu_crsqe_flux.hbk
(or nue)
Dump any hbook file I have into some text file using hbktotxt.kumac
Then using crs_txt_hbk.F,
I changed what ever been I had to files ranging from 0 to 10 GeV with
10 MeV bins.
The output of the subroutines are hbook files.
NB
make sure that the normalization of what ever file you are dealing with
make sense! For example, whatever the binning is for the cross-section
file, the value of xsec at 1 GeV should not change, since the xsec is
binned only for computational reasons. Where it is exactly the opposite
for flux files.. (ok, I should stop writing obivous stuff.. but who
know I made that mistake once and I'd rather not do it again..
embarassing...)
Make sure that everything is write by checking that the xsec files
still peaks at the same values and that the integral of the flux files
for a given energy window is the same than for the original files.
Step 2:
Make event
spectrum histograms and figure out normalization factor
Input: bnl14_numu_crsqe_flux.hbk (or nue)
Output: evspec_bnl14_numu.hbk (or nue)
Using the program
evspec.kumac (or evsepc_061024.kumac) I mutliply the flux by the
cross-section to get and event spectrum.
In this program I also take care of the normalization factor. as
described below:
Flux is given in neutrino/GeV/m2/POT
at 1m from the target
Warning the
units areback tob per GeV but the binning is per 75MeV, so I need to
multiply my flux by 0.075
to have something correct
Cross-section are
given in 1E-38 cm2 in
file bnl14_numu_crsqe_flux.hbk (bnl14_nue_crsqe_flux.hbk)
NB
(In this case the fraction of neutrons in water is NOT included.. need
to be taken care off earlier in evspec.kumac so that neutrino and
antineutrinos can be handle properly separately)
Beam
power
1
Time
5*1.7 E+7 sec (=5
years!!)
Mass of
target 300 kton of water = 0.3E+12
gr of water = 1.81
E+35 nucleons
So 2500 Kton MW 107sec
corresponds to 1.12
E+22 POT
Distance
1290km
--> 6.0E-7
Normalization factor: 0.075 *
1E-42 * 6.0E-7 * 1.12E+22 * 3.01E+35 = 1.51 E+8
Step 3:
Make
background files
Input:
table_10gev/evspec_bnl14_numu_061024.hbk
(for the BNL flux file)
background_10gev/flux_sk_10gev.hbk
(for the SK flux file)
+ regular SK atmospheric MC.
NB: the SK flux file is obtained using background_10gev/flux_10gev.F
Output:
background_10gev/bg_bnl14.hbk
To compute the background, we
use the atmospheric superk MC and we reweight each events by the ratio
of the BNL flux to the SK flux.
That way we can apply our usual precuts to the MC and apply different
likelihood efficiencies for different source of background.
The main program used is mkbg.kumac and it requires to
manually dump the histograms in a file after having run it.
We need a normalization factor in order to reweight to two fluxes, here
are the details:
FNAL:
Flux is given in neutrino/GeV/m2/POT
at 1km from the target
Warning
the
units are per GeV but the binning is per 75MeV, but I would have to
multiply by the binning to be able to compare with the SK flux.
Cross-section 1E-38 cm2
Beam
power
1MW
Time
5*1.7E+7 sec
Mass of
target 300
kton
Distance
1290km
--> 6.0E-7
Binning
75
MeV bins (well
it is 10 MeV, but it was 75 and it still needed to be properly
normalized)
And the number of POT in 1
second for a 1 MW beam with 28GeV protons is 2.23E+14.
SK:
Flux cm-2
sec-1
MeV-1
Cross-section 1E-38 cm2
Mass 22.5
kton
Time 100
yr of MC = 3.15E+9 seconds
Binning
10
MeV bins
So
FNAL/SK = (1/1E+4) * (1/1E+3)
*(5*1.7E+7/3.15E+9) *(300/22.5)
*(2.23E+14) *(6.0E-7) *
75/10
m-2/cm-2
GeV-1/MeV-1
Time
Mass POT to seconds
Distance Binning
= 35.3
Step 4:
Use
oscillation analysis code to modify signal event spectrum. Add
oscillation probabilty, energy smearing, and likelihood effiency.
Input: all the files in table_10gev
Output:
quick_codes_10gev/bnl14+_70_75_061102.hbk
This require using the
#check# version of the resolution80nonqe.F
program.
Use resolution_check_fnal.sh
to start
running it. It is also in that file that you'll find a list of all the
input files that you need.
Make
sure that the phase and the value for sin2(2th13) are set correctly
(hard-coded in resolution80nonqe.F).
Step 5:
Make event spectrum including oscillation and background to be able to
compare with Diwan results
Input:
quick_codes_10gev/bnl14+_70_75_061102.hbk
(for signal event spectrum)
background_10gev/bg_bnl14.hbk (for
background event spectrum)
Output: Whatever I want !
Just
run the kumac evspectrum_fnal.kumac
Be aware of the final binning that you want.
All the normalization should be set to 1 if you want to simulate 5
years of running at 1290 km with 1.12 E+22 POT/year (2500Kton, MW E+7
sec)
Here
is a results using those factors
(eps)
Where the black line is signal +
background for delta=135°
and the red line is delta=0°
Green
is the nue
background, purple is the NC background
and blue is the CCnumu background
The integral of the 135° histogram is:1628
The integral of
the 0°
histogram is: 1185
The assumptions are:
sin2(2theta13)
= 0.04
2500KtonMW 10^7sec
Step 6:
Run oscillation analysis code
Directory /codes_10gev
There is several substep in here:
But for old readme file go there.
Step 6a: Compiling everything with the
right header files
The script compile.sh
should be used before trying to do anything.
setup.h.mine
takes care of: the time of running (for nu and nub)
the volume of detector
the power of the beam
the location of detector
the number of bins (kind of.. this is not fool-proof
some things are still
hard-coded)
osc_param.h
takes care of: the number of background
events
the number of systematic errors
Never change the file setup.h
manually but always make a newfile setup.h.newfile and change it compile.sh
Step 6b: Making oscillation table
Input:
/osc_calc_10gev/prob_fnal_1290-.dat
(oscillation table)
/table_10gev/evspec_bnl14_numu_061024.hbk
(BNL flux file)
/table_10gev/like-eff_10gev.hbk
(Likelihood efficiency table)
Smearing tables:
/table_10gev/nu_ccqe.hbk
/table_10gev/nu_cc.hbk
/table_10gev/nubar_ccqe.hbk
/table_10gev/nubar_cc.hbk
Output:
bnl14+.dat and bnl14-.dat (for example)
Then just run the kumac mktable.sh
Step 6c: Looping over tables to make
sensitivity tables
Talks
Fermilab Workshop (Sep 16-17, 2006)
From others:
Useful links:
