Документ взят из кэша поисковой машины. Адрес оригинального документа : http://antares.sinp.msu.ru/docs/2010_Amsterdam_Kulikovskiy.pdf
Дата изменения: Mon Dec 16 17:03:27 2013
Дата индексирования: Thu Feb 27 20:27:41 2014
Кодировка:
Analysis of the ANTARES data for SN neutrino detection V. Kulikovskiy (MSU / INFN Genova)

SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

1/14


SN progenitor stars distribution


A half of the possible sources of SN type II explosion are at 10kpc. Sensitivity decreases as the square of the distance to SN



SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

2/14


Methods overview


Single rate 1. exclude OMs affected by bio bursts 2. Sum all hits of the selected OMs in the detector in 100ms 3. Compare with total detector rate in case of no SN (background is slow varying bioluminescence + 40K)



Coincidence rate 1. Obtain sum of true coincidences in the detector in 100ms 2. Compare with total coincidence rate in the detector in case of no SN

SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

3/14


Single rate: Bioluminescence cut


Collection of distributions of hit counts in 100ms for each OM during ~45min Usually these distributions consist of 2 parts ­ Poissonian (due to 40K and plankton bioluminescence) and long tail (due to bioluminescence bursts)



PMT

CP

MT

Fit with Poisson distribution. Free parameters: (mean value) PMT C (fit only until this hit count) P MT
PMT



=

PMT

If number of hits in 100ms in the OM Nhits< CPMT it will be used in the total detector hits H calculation.
SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010 4/14


Single rate: PDF of hits in the detector


with known CPMT and PMT, we know PDFs of hit counts in each PMT In each time slice the set of PMTs which passed bioluminescence cut is different So, the PDF of the detector is different for every time slice (100ms)

#1

#2

#3

... ...

#900

time slice 100ms #1

time slice 100ms #2

t
exp

Detector PDF could be approximated as a Gaussian with M =
SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010


N
pmts

m

and

S=




N
pmts



2 exp

5/14


Single rate: results
(H-M)/S distribution


1.7 larger than expected due to the slow change of bioluminescence activity



Significance as a function of distance simulations for 900 Oms in the conditions of run 39856(low bioluminesce) 40154(high bioluminescence)



SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

6/14


Double coincidences: how to obtain number of true coincidences
Time difference between hits in two OMs entries ~200 Cerenkov photons per decay

= 4 ns

e40

K



For every couple of OM in a storey: collect time difference between hits in the 2 OMs during 2000s Random rate gives constant pedestal K decay, which flashes both OMs, produces Gaussian peak with width ~4ns (compatible with distance between OMs and time resolution)
40







Concentration of 40K is very stable. Rate of true coincidences is 16±3Hz
7/14

SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010


Double coincidences: search method


Find rate of true coincidences for every couple of OMs, analyzing 2000s of data. apply quality cuts to exclude couples of bad OMs (problems with the high voltage or bad calibration). In one time slice (100ms) collect time difference distribution for all coincidences in the detector. Obtain total number of true coincidences in the detector in the time slice. Compare total coincidence rate in the detector in 100ms with sum of coincidence rates of every OM couple.
8/14







SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010


Double coincidences: error on coincidence measurements
Two components ­ a statistical f uctuation of 40K l coincidence rate + f uctuation of a random l background(bioluminescence).
Black ­ error of the fit Red ­ statistical fluctuation Of 16Hz signal only



The plot could be used for error estimation:


of a OM distribution fit (in dependence of time slices) Of a time slice fit (in dependence of number of OMs)
9/14

SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010


Double coincidences: results


For 450 couples (runs 24/10/2009-29/10/2009) uncertanty for rate was 2.6 sigma, which is in correspondence with the simulations


5 sensitivity in dependence of the background rate and a distance to SN for 900 PMT

SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

10/14


Method with triples


Simulations show the best ratio of signal to noise.



Very low statistics the 100ms time slice: the fit of the time difference distribution for triples is not possible. Only the total number of triple coincidences is considered in the time slice by time slice analysis.



SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

11/14


Triple coincidences: Results


For January 2009 uncertainty is in correspondence with statistical fluctation


5 sensitivity in dependence of the background rate an a distance to SN for 900 PMT

SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

12/14


Implementation of the trigger


The sensitivity of the 3 methods is comparable (methods with doubles and triples don't have a decreasing efficiency for very close SN). 5 sensitivity gives 90% probability to detect event and ~1 fake event in 17 minutes ­ fine for writing coincidences, but not raw data. 5 is achievable only for distances to SN ~4kpc, so if we want to be able to detect closer SN, fake event will be much higher (for 10kpc it's about 1 every 2 time slices). SNEWS could help to save a raw data for offline analysis. Delay is maximum 10 minutes (this time is due to the slowest experiment in the network and probably could be minimized as we don't need precise coordinates). Two proposals:








Use 2 sec buffer of raw data to store coincidences - sufficient to cover the 10' SNEWS latency time & can be used by the current analysis methods. Possibly a big 10' buffer to store all data for future improvements.
13/14



SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010


Summary


ANTARES works with high level of noise from the environment. Different SN detection methods were introduced. Uncertainties are understood. Efficiency of methods is comparable for 900 OMs ANTARES. SNEWS seems to be useful for






Raw data writing for offline analysis (10min buffer is needed) Coincidence data writing (500 times smaller buffer is enough)
14/14



SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010


Uncertainty verification


Double: difference between number of true coincidence from the fit in time slice and sum of the hits from ever OM rates (~450 OMs). Sigma 2.6 is in correspondence with simulations. Triple: difference between number of true coincidence from the fit in time slice and sum of the hits from ever OM rates (~85 storeys). Sigma is in respondence with statistic fluctuation.





SN detection, V.Kulikovskiy, ANTARES meeting in Amsterdam 22-09-2010

Backup slide 1