Status of Sirene

Status of SireneMaarten de Jong

What?

Sirene is yet another program that simulates the detector response to muons and showers

It uses a general purpose collections framework for PDF tables‒ allows for optimisation of accuracy and speed of

interpolations‒ facilitates I/O

JTools A Java inspired collections framework in C++ with multi-

dimensional interpolation functionality – different collection types

• each collection has map functionality– variable bin size O(log(N)) access time– equidistant bins O(1) access time

– allows for any number of dimensions and bins• binning in dimension n may depend on bin in dimension n-1

– allows for different interpolation algorithms• polynomial• spline

– allows for a mixture of different collection types (new)• use of type lists

JPhysics Formalism for light production, propagation and detection

‒ semi-analytical PDFs, based on single scattering approximation Custom classes for PDF/CDF tables

‒ treatment of weight functions to improve accuracy of interpolations (or equivalently reduce number of bins)

‒ may serve as input to Sirene Custom muon energy loss cross sections and shower

energy generation (Paul Kooijman)– Bremmstrahlung (threshold 10 MeV)– pair production (threshold100 MeV)– implementation photo-nuclear interactions pending

Procedure

1. Make PDF tables‒ at present, JPhysics table generators are used ‒ other applications can be interfaced readily

2. Convert PDFs to CDFs‒ allows to make various checks

3. Run event processor‒ read detector geometry‒ setup radiation tables (to speed up determination of cross

section and shower energy for pair production)‒ setup tables with maximal light yield (eliminate dependence on

orientation of PMT)‒ event loop

Standard Antares Monte Carlo production

“km3 + geasim”

Processing steps1. Read Event from MonteCarloEventWriter output file2. Remove existing hits, if any3. Propagate muon(s)

‒ simulate energy loss and EM-showers‒ generate hits

4. Process shower particles from primary vertex‒ simulate hadronic showers as an EM shower, but assign different weights to

the energies of different particle species‒ generate hits as one shower but use correct longitudinal profile of individual

showers5. Merge hits (to speed-up TriggerEfficiency)

‒ DTmax typically 15 ps

6. Write Event to MonteCarloEventWriter compatible output file

En [GeV]

Trigger effective volume

Volu

me

[m3 ]

km3 (V3.7)Sirene

± 10%

Ratio

Muon light yieldlig

ht y

ield

[n.p

.e]

direct scattered

km3 (V3.7)Sirene

En [GeV] En [GeV]

Muon PDF

Dt [ns] Dt [ns]

direct scattered

km3 (V3.7)Sirene

light

yie

ld [n

.p.e

]

EM-shower light yield

direct scattered

km3 (V3.7)Sirene

En [GeV] En [GeV]

light

yie

ld [n

.p.e

]

EM-shower PDF

Dt [ns] Dt [ns]

direct scattered

km3 (V3.7)Sirene

light

yie

ld [n

.p.e

]

Performance

En [GeV]

time

[ms]

¶ Events with ≥ 1 hits

Average time per event ¶

First tests KM3NeTMonte Carlo production

KM3NeT Detector

‒ 154 towers, each with 40 optical modules, each with 31 PMTs‒ number of PMTs 154 x 40 x 31 = 190,960‒ geometry ref154_3inch31pm180_204006.det

PMT‒ type 3 inch, standard Bialkali‒ light collector ring acceptance += 20%

Generation‒ particles muon neutrinos‒ Energy spectrum E-1.4

‒ Energy range [102,107] GeV Running time

– km3 (V4.3) ~135minutes / 35,000 events– Sirene ~12minutes / 35,000 events

muon light yield

direct scattered

km3 (V4.3)Sirene

En [GeV] En [GeV]

light

yie

ld [n

.p.e

]

Muon PDF

Dt [ns] Dt [ns]

direct scattered

km3 (V4.3)Sirene

light

yie

ld [n

.p.e

]

EM-shower light yield

direct + scattered

En [GeV]

km3 (V4.3)SireneIn km3 V4.3

direct light fromEM-showers is

tabulated together with scattered light from EM-showerslig

ht y

ield

[n.p

.e]

EM-shower PDF

Dt [ns]

direct + scattered

km3 (V4.3)Sirene

light

yie

ld [n

.p.e

] In km3 V4.3direct light fromEM-showers is

tabulated together with scattered light from EM-showers

Performance

En [GeV]

time

[ms]

¶ Events with ≥ 1 hits

Average time per event ¶

Summary & Outlook Sirene is a new program that simulates detector

response to muons and showers Antares

‒ trigger effective volume is similar‒ distributions of arrival times are slightly different¶

‒ Sirene is ready for use... KM3NeT

‒ distributions of arrival times in km3 V4.3 seem to be more reliable than V3.7 (thanks to Clancy James)

‒ Sirene is ready for tests...

¶ problem in Sirene with light from EM-showers fixed (shower energy)

Summary & Outlook (II)

Possible improvements– include photo-nuclear cross sections– include [multiple] scattering of muon

Possible enhancements‒ generation of neutrino vertex and simulation of

detector response, possibly including real-time trigger, could operate in a single step for the run-by-run Monte Carlo simulation