1. DC2 pulsars analysis: a population point-of-view
Gamma-ray Large Area Space Telescope
DC2 pulsars analysis: a population point-of-view
Massimiliano Razzano
(INFN - Pisa)
LAT DC2 Closeout Meeting
(GSFC, May 31-June )
06/02/2006
M.Razzano - DC II Closeout Meeting

2. Outline Analysis overview; Scripting in Python: the pyPulsar tool;
Cuts and thresholds;
The EGRET pulsars;
Detecting pulsars in the LAT source catalog;
The flux distribution;
The brightest MSP;
Beyond the LAT source catalog;
Conclusions and DC2 follow-up

3. Yes, I was involved in the DC2 pulsars generation…but….
Analysis overview
I will present an analysis to detect pulsars with radio counterparts in the D4;
An automatic procedure for testing the periodicity of pulsars has been conducted to find high-confidence identification;
More detailed analysis on some pulsars (e.g. EGRET pulsars or MSP pulsars);
The basic characteristics of the detected pulsars have been derived;
A preliminary note:
Yes, I was involved in the DC2 pulsars generation…but….
We put a lot of pulsars in the DC2 sky…
DC2 pulsar truth is on my PC in some obscure directory with some obscure name…
It is almost impossible to me to remember names like PSR XYZM+IJKF12345…
Through hypnosis Julie and Seth forced me to forget everything…
Believe me…by doing this analysis, I completely ignored the DC2 truth..

4. Why scripting? Why Python?
Scripts are very useful to create some automatized analysis (see also D.Smith’s talk);
Scripts allow to perform the basic data reduction in automatic way (applying cuts, barycentering, etc…);
Scripts can be done using bash/tcsh shell or using scripting languages like Python
For these analysis a set of Python scripts have been developed
With Python more complex operations can be done easily (e.g. parsing);
A lot of Python extension are availabe (e.g. pyROOT, pyFits, sihyppo);
See for much informations about Python;

5. Python scripting: pyPulsar
pyPulsar is a set of Python classes and scripts for doing pulsar analysis
pyAnScripts
(scripts for various pulsar-related analysis)
pyGtTools
(classes for interfacing with LAT Pulsar Science Tools)
pyGtBary
pyGtPsearch
pyGtEphcomp
pyGtPhase
pyGtSelect
pyPulsar
(class for managing data and analysis of a pulsar)
Analysis results
LAT Pulsar Tools

6. Python scripting: pyPulsar (II)
A list of pulsar is given in input (extracted from D4, from coincidences with LAT Source catalog, etc…);
For every pulsar the ROI is selected and FT1 barycentered;
Periodicity test is applied (here only Chi2) and phase assigned;
The output file contain chance probability of the periodicity test;
Plots of detected pulsars and determination of population parameters (e.g. using ROOT macro);

7. About cuts and thresholds… What is the threshold for pulsar detection?
For each pulsar an specific selection in ROI, energy, etc. can be done;
For each pulsar list the analysis was performed for 2 energy ranges and with corresponding selection of the ROI according to the PSF;
Energy
Radius of ROI
E>100MeV 3°
E>1GeV 1°
What is the threshold for pulsar detection?
After some tests and some discussions I decided to use a more “conservative” detection limits, according to the EGRET pulsars papers
Energy
c2 chance probability
Reference
No detection
P > 2 x10-3
The same as below
Low-confidence detection
10-9 < P < 2 x10-3
Limit of B (Ramanamurthy et al. 1996)
High-confidence detection
P < 10-9
Limit of B (Ramanamurthy et al. 1995)

8. The EGRET pulsars
Vela
The first exercise is the analysis of the EGRET pulsars, testing periodicity and recovering lightcurve
Seeing that pulsars does pulse….
Geminga
Crab

9. All-sky count map of the simulation of 55 days for DC2
E>200MeV
ROI with R=3°
Periodic at C.L. > 99.99%
E>100 MeV
ROI with R=1°
Periodic at C.L. > 99.85%
E>100MeV
Periodic
at C.L. > 99.99%
E>300MeV
ROI of R=1.5°
All-sky count map of the simulation of 55 days for DC2
06/02/2006
M.Razzano - DC II Closeout Meeting

10. Pulsars in the LAT source catalog?
As first step a search for coincidence between LAT sources and pulsars in D4
Radius dependent on energy (LAT_Cat_v2)
49 pulsars found

11. Pulsars in the Lat catalog: summary
The first step is the analysis for E>100MeV
Low confidence
High confidence
No detection 13
Low confid. 14
High confid. 22

12. Pulsars in the Lat catalog: populations
Youngest non-MSP have age≈103 y (Vela & Crab) and B≈5x1012 G
Low confidence
High confidence
1014G
103yr
109G
Oldest non-MSP have age≈108 y (Vela & Crab) and B≈2x1011 G
Low confidence
High confidence
Name RA
Dec
period
pdot
PSR_J1955p2149
298,98
21,82 0,
1,87E-20
PSR_J0904m5008
136,07
-50,14 0,
3,94E-20
PSR_J1659m5022
254,98
-50,37 0,
7,64E-21
PSR_J0717m1235
109,45
-12,59 0,
7,23E-21
PSR_J1735m5757
263,96
-57,96 0,
3,36E-20
PSR_J1745m0952
266,28
-9,876 0,
9,50E-20
4 MSP detected at high confidence and 1 at low confidence
(B<109gauss, age>109y)

13. PSR J0904-5008, the brightest MSP
Phase curve after periodicity test
Period=2,31 ms
In the LAT Catalog its flux is 1.7 x 10-7 ph/cm2/s
Fit with power-law + exponential cutoff
a=1.01±0.01
Ecut=37,1±1,3 GeV

14. Pulsars in the LAT catalog: DC fluxes
Low confidence
High confidence
An interesting plot is the LogN-LogS (see D.J.Thompson ’03)
The low coincidence show a change of slope.
Various possible origins (background, geometry,luminosity distributions)
Slope=-1.04±0.19
The high confidence pulsars follow a similar distribution

15. Beyond the LAT catalog: Summary(I)
The next step is the scan on the entire D4 database. First for E>100MeV
We expect to find more pulsars since there is the help of the periodicity signature
E>100MeV
Low conf
High conf
No detection 42
Low confid. 26
High confid. 29

16. Beyond the LAT catalog: Summary(I)
Then I repeat the scan for E>1GeV
The pulsar detected will be different because of many reasons (different ROI,different spectra of pulsars, etc…)
E>1GeV
Low conf
High conf
No detection 38
Low confid. 34
High confid. 25
The number of high+low is increased. In order to have a final sum we must make an OR of the results at 2 energies

17. Beyond the LAT catalog:population
E>100MeV
After merging E100MeV and E1GeV results
No detection 14
Low confid. 51
High confid. 32
E>1GeV
More MSP at E>1GeV
(Spectral shape)
Find the
differences….

18. Conclusions and DC2 follow-up
LAT pulsars will be a big amount, then an automatic system for pre-analysis should be very useful;
Python could be a good candidate, pyPulsar is under development: available in CVS (/users/razzano/pyPulsar);
The LAT sensitivity will permit study of different sub-population of pulsars: this was the case of DC2;
At least 4 MSP can be detected during DC2 in the LAT source catalog;
At E < 10-7 ph/cm2/s the logN-logS shows a change in slope: it will be the same with real data?
By scanning the D4 32 high-confidence pulsars have been found: why discrepancies with other analyses? Most probably differences in putting threshold and in cuts;
Detection is very sensitive to cuts and threshold: study of optimal cuts is strongly needed;
Explore other test statistics, in particular the H-test: try different strategies for detecting more pulsars;
Special thanks to Dave, Patrizia, Alice, Isabelle, Gottfried and the other people of PSR group for the very interesting discussions and suggestions !