VERPS and analysis of CTA1

ANDREW LOOColumbia University

August 7, 2011

This paper discusses the creation of VERPS (VERI-TAS Remote Proposal System) and the onset of thesearch for off-pulse emission from SNR CTA1 withFermi data.

1 Introduction

Figure 1: The Earth’s atmospheric transparency toelectromagnetic radiation differs throughout the elec-tromagnetic spectrum. The transparency of Earth’satmosphere at radio wave frequencies allow for de-tection of radio waves at ground-base telescopes. Onthe other hand, γ rays are completely blocked by theatmosphere, preventing any direct ground-base meth-ods of detection [3].

1.1 Introduction to Very-High EnergyGamma Rays

γ-ray astrophysics seeks to study the observable uni-verse at energy ranges of the highest frequencies. γ-rays extend the electromagnetic spectrum by 15 or-ders of magnitude [4]. The very-high-energy regimeare those at energies in the TeV scale (1012 eV). Theradiation emitted at this frequency is so energeticthat the cannot be produced through thermal pro-cesses. Instead, they are produced by violent non-thermal processes through interactions of high-speedparticles.

At the range of 300GeV to 10TeV, rarely any γ-rays reach the Earth’s surface. The detection of γ-rays at these high energies through outer space tele-scopes is greatly limited by the detectors effectivearea and by the low flux resulting from the high en-ergy. These two challenges delayed the study of VHEγ-rays until the 80s citeEster. As a result, studyingγ-rays in this VHE regime calls for a different methodthen lower γ-rays at lower energies. At lower ener-gies, there is enough flux such that a telescope witha decently sized effective area in outer space can takein a significant amount of data. For higher energies,ground-based γ-ray telescopes either detect the prod-ucts of extensive air showers (EAS) that survive tothe ground or the Cherenkov light produced by theEAS. For energies greater than 1 TeV, enough showerparticles reach the ground and detection methods in-clude arrays of scintillators or by water Cherenkovdetectors.

1

1.2 Introduction to IACTs

For lower energies, the amount of particles is too fewand the Cherenkov radiation is detected using Imag-ing Atmospheric Cherenkov Telescopes (IACT) thatuse mirrors to reflect light into photomultiplier tubes(PMT). As VHE γ-rays hit the Earths’ surface, theyinteract with nuclei in the earth’s atmosphere pro-ducing an EAS of particles. The longitudinal propa-gation of the shower is dominated by pair production(γ → e+ + e−) and Bremsstrahlung (e± → γ + e±.At approximately 10 km above ground level, the EASreaches its maximum intensity. At this point, whenthese particles travel faster than the speed of lightin that medium, they radiate blue light known asCherenkov radiation. This blue light creates a lightcone and a circular light pool on the correspondingground. If an IACT is within this light pool it cantake an ”image” of this using PMT.

Figure 2: Sketch of an extensive air shower of a γ-ray(left) and of a cosmic ray nuclei (right) [3].

The biggest obstacle in IACT image analysis isthe comic-ray background. To differentiate betweenhadronic (cosmic rays) and electromagnetic (γ-ray)showers, Monte Carlo simulations are run to com-pare the shape of the measured light on the PMTsto distinguish between the two types. Cosmic rayshowers tend to be more circular, larger, and havea wider spread owing to the greater complexities inhadronic showers, whereas γ-rays are more ellipti-cal. This difference reflects the vastly more complexhadronic EAS in which mesons, such as pions andkaons, and nucleons are produced. Other informationextracted from the PMT photograph include the cor-

relation between the energy of the coincident photonand the intensity of the flux in the image, as well asthe image axis, which corresponds to the direction ofthe shower.

1.3 VERITAS

10 k

m

120 m

1°~5 ns

!-ray100 GeV

Figure 3: VERITAS is composed of four 12-meterIACTs. The extended air shower’s ’size’ peaks atabout 10 kilometers from Earth’s ground. If theIACTs are within the light pool produced by theCherenkov radiation of the EAS, a reconstruction ofthe shower’s direction can be made. This design al-lows for improvements in angular resolution, energyresolution, background rejection, and sensitivity.

The VERITAS (Very-High-Energy Gamma RayInstrumentation Telescope Array) is array of fourIACT to detect γ-rays in the range of 100GeV to30TeV. Other IACTs include HESS, CANGAROO,and the planned CTA. These IACT’s complementother telescopes, such as Fermi, in order to observevery-high-energy extra-galactic and galactic sources.Built in fall of 2007, VERITAS employs four 12-meterIACTs in Amado, Arizona pointed at the NorthernHemisphere. It has a field of view (FOV) of 3.5 andan angular resolution of r68 < 0.1. The collabora-tion includes 22 institutions in 5 countries, includingUS, Canada, UK, Ireland, and Germany. It also has

2

support from DOE, NSF, SAO (U.S.), STFC (U.K.),NSERC (Canada), SFI (Ireland). With four IACTs,VERITAS benefits from an increase angular resolu-tion, sensitivity, background rejection, energy resolu-tion, and EAS reconstruction. Currently, more than38 sources have been observed with VERITAS with8 source classes [4].

The telescopes were designed using the Davies-Cotton design for practical and financial purposes.There are 345 hexagonal mirror facets per telescope[4]. After the Cherenkov light is reflected from themirrors, they are collected by a camera which con-verts the light image to electrical pulses. The pho-tomultiplier tubes (PMTs) generate an analog signalin response to the light. The signal is then digitizedby the Flash Analog-to-Digital Converters (FADCs)to facilitate the analysis.

2 Search for off-pulse emissionfrom SNR CTA 1 with Fermi

2.1 Pulsars

Figure 4: 2D diagram of several proposed pulsarmodels. The light cylinder marks a cylindrical areacentered at the rotational axis at which the co-rotating velocity velocity equals the speed of light[5].

At the end of a massive stars life with a mass of8M ≤ M ≤ 25M, the star collapses under itsgravitational force. Eventually, the collapse is sta-bilized by the degeneration of neutrons. Becauseof a partial conservation in magnetic flux, after thestabilization of the collapse, the pulsar star (aroundR = 10 km) magnetic field increases by a factor of1010 or more amounting to fields around the orderof 1012 G. Similarly, a partial conservation in angu-lar momentum causes the neutron star to rotate atrapid speeds up to periods of milliseconds. Radia-tion pours out from the pulsar causing an ejectionof particles at two points at opposite ends. Whenthis magnetic field is misaligned with the rotationalaxis, which is called a oblique rotator, the radia-tion is emitted in a lighthouse effect. These observedrotating pulsars emitting radiation periodic pulses, isknown as a pulsar [3]. Many of the pulsars thus farfound show a gradual increase in P . This can be ameasured observable P . The pulsar rotation is slowedas kinetic energy is lost through the radiation of thedipole. Ostriker & Gunn estimate that the loss of ro-tational energy of the Crab pulsar with the measuredP and P could power all non-thermal radiation seenin the Crab Nebula.

Measurements of P and P allow for fairly good es-timations of pulsar age, magnetic field strength, andspin-down power. The spin-down power is given bythe loss of rotational kinetic energy.

E =d(1/2IΩ2)

dt= IΩΩ = 4π2IP /P 3 (1)

Where I is the moment of inertia of the pulsar, andΩ = 2π/P . For pulsar age, assuming that the P0 P the pulsar age is given by

τc ≡P

2P(2)

2.2 PWNe

The observable pulsed emission accounts for only asmall percentage of the total spin-down energy. Mostof the energy is thought to leave he pulsar’s magne-tosphere through magnetized winds of charged parti-cles. Eventually the magnetized wind interacts with

3

Figure 5: The infamous P − Pdiagram. Lines fromtop left to bottom right indicate lines of inferred con-stant surface magnetic field of pulsar and lines frombottom left to top right indicate constant E. Thebottom left population are those known as ”millisec-ond pulsars,” while those in the top center are theregular pulsars. Note how many of the pulsars inthe ”millisecond pulsar” region are binaries. It is be-lieved that the pulsar consuming its neighbor leads toa greater acceleration. Sources found in the shadowedregion at the top of the diagram are no longer consid-ered pulsars but magnetars. Pulsars travel from thetop left to the bottom right. When the pulsar reachesthe graveyard line, it is believed they no longer haveenough energy to produce electron pair production,which powers the pulsars spin, and they essentially’die’ out [5].

the interstellar material, emitting synchrotron radia-tion and inverse-Compton radiation. It is these emis-sions which are termed the the pulsar wind nebula(PWN).

2.3 Objective

The objective was to measure the Inverse Comptoncomponent of PWN’s detected by VERITAS. VHEinstruments are used to measure Inverse Comptonin the VHE regime, while HE instruments, such asFermi, was used to detect the corresponding InverseCompton in the HE regime. This allows for a morecomplete picture of the Inverse Compton function,as well as a clearer picture as to the function’s max-imum. However, because the pulsar component ofthe spectrum dominates the PWN spectrum in theHE regime, the off-pulse emissions were measured.

Figure 6: A view of pulsar CTA 1 by Fermi. Overlaidon top are the corresponding view from radio waves(black) and x-rays (green). Note the typical SNRshell shown by the radio wave.

2.4 CTA 1

The off-pulse emission of CTA 1 was recently ob-served by VERITAS making it the perfect target foranalysis. CTA 1 was also the first pulsar discov-ered through GeV γ-rays by the Large Array Tele-scope (LAT) of Fermi. CTA 1 is about 13 kyr and1.4±0.3 kpc. It is about 1 in size and the pulsar

4

spin-down is E = 4.5 × 1035 erg/s [1]. Although theshell-type supernova remnant (SNR) CTA 1 was dis-covered by the Owens Valley Radio Observatory inthe 1960s (CTA1FermiPulsarDiscovery.pdf), the as-sociated pulsar was not identified until 2008 by thenewly commissioned LAT [1]. The discovery helpedfortify the opinion that unidentified SNRs and star-forming regions are young pulsars (CTA1 Discovery).

2.5 Method

Using Fermi analysis data and tools, we plan to firstuse timing analysis to phase fold the data by barycen-tering photons. In essence, the Fermi tools com-pensate for the arrival time of the photons duringthis process. Next, the corresponding off-pulse emis-sion is identified in the light curves produced. Then,the significance of the off-pulse emission is calculatedabove the background. Using various methods, wewill attempt to identify whether this radiation is fromthe magnetosphere or the PWN. The measured Pand P in 2008 for CTA 1 was P = 316.86ms andP = 3.614× 10−13 [1].

Counts map were created to have a general under-standing of the data and what to be expected fromthe data. Specifically, since pulsars have a strongcutoff below 10 GeV, the presence of a strong sourceat high energies ( > 10 GeV or > 50 GeV) suggeststhe presence of a PWN. counts maps were created forthe 200 MeV to 1GeV range, 1 GeV to 10GeV range,and 10 GeV to 100 GeV range. The strong source atthe high energies is probably a result of the CTA 1pulsar.

Since this undertaking has already been done withCrab, we planned to replicate the results of Acker-mann et al. [2] , and then apply a similar analysis toCTA 1. The same ephemerides was used as the Crabpaper. First, LAT photon and spacecraft data wasdownloaded from fermi in the form of .fits files. Thephoton events were selected using gtselect and timeselection using gtmktime. Then the photon data wasphase-folded using gtpphase. To replicate the 2-phaselight curves present in the (papers) data was dumpedin an ASCII file using fdump. Next, a ROOT filewas written to import the data in the ASCII file to aROOT tree.

2.6 Results

(a) 200 MeV to 1 GeV

(b) 1 GeV to 10 GeV

(c) 10 GeV to 100 GeV

Figure 7: CTA 1 countmaps at different energyranges. The strong sourceat high energies suggeststhe PWN. Fermi toolswere used to make thesemaps.

The count maps recon-firmed the existence ofthe PWN of CTA 1 andindicated that we werelikely to see the pulsarwind nebula in our anal-ysis. During the repro-duction of the crab lightcurves using the Fermidata, our histogramsshowed a similarity to-wards that shown inthe paper. However,our counts were lowerthan the crab’s and theposition of our phaseseemed to be phase-shifted. In retrospect,this may have been amistake caused by tak-ing data using the SNRCTA 1 instead of the lo-cation of the pulsar ofCTA 1. When a lightcurve was produced forCTA 1, the data alsoseemed similar to theone produced in paper(). However, countswere significantly lowerand the correspondencebetween the two lightcurves was much lower.The significantly lowercounts are attributed tothe fact that half thedata could not be ob-tained from the Fermidatabase since the CTA1 paper included datafrom the commissioningperiod. The incongruitybetween CTA 1 lightcurves prompted a new retrieval of data and anal-

5

ysis. This new light curve vastly differed from theCTA 1 light curve in the paper and in pulsar lightcurves in general. This difference is attributed to thedifference between the old and new Fermi data andtools which were recently changed.

Figure 8: Crab light curves. Both the light curvefrom the paper (top) [2] and the reproduction (bot-tom) are similar. However, the reproduced lightcurves show a markedly decrease in counts and aslight phase shift.

Figure 9: CTA 1 light curves. The light curve fromthe paper (top) [1] is markedly different form that ofthe reproduction (bottom). There was clearly a greaterror in producing the second histogram.

3 VERPS

3.1 Introduction

Because VERITAS can only be used to study pre-viously known sources (i.e. it cannot be used tosearch for new sources, but it can be used to searchfor sources visible in other wavelengths but not yetfound in the VHE regime), a method for organizing

how much time is spent on a particular source mustbe put in place. Each year, all VERITAS memberssubmit proposals to the Time Allocation Committee(TAC), which decides how much time is to be allottedto any particular source.

3.2 Motivation

Cycle 4 (this past year) saw a large amount of pro-posals. There were ∼ 64 proposals, many with mul-tiple targets (i.e. sources) amounting to a total of> 2000 hours of observation. The members of TACbuilt an excel spreadsheet for the evaluation and dis-cussion of the targets. The tedious and laboriousmethod employed in submitting and organizing pro-posals prompted the suggestion for an automatedproposal system.

Figure 10: Excel spreedsheet of proposals for TACevaluation and discussion

3.3 Structure

The proposal system was designed with efficiency, or-ganization, and simplicity in mind. Not only wasthe ease of submitting proposals stressed, but alsothe ease of which submitted proposals can be ac-cessed and modified. The proposal system consistsof ”Cover”, ”Target Form”, and ”Science Justifica-tion” sections (See Figure 11).

6

3.4 VERPS Submission Process

3.4.1 Genera Process for the VERPS Sub-mission

The general process of submitting a proposal usingthe VERPS system can be summarized as:

1. Choose to either start a new proposal fromscratch/retrieve cookie information if availableor to revise an already submitted proposal.

2. Enter the proposal data into the RPS form.

3. Validate the proposal data for completeness. Forthis, use the ”Validate” button at the end of eachsection (Cover, Targets, and Scientific Justifica-tion) on the RPS form. Required missing infor-mation will be shown in red and required filledout information will be shown in green. If cook-ies are enabled in user’s computer, the enteredand validated information will be rememberedand stored for the next time.

4. Upload a pdf file for user’s proposal (a maximumof 3 pages).

5. When done, user’s should click on Make Cover tomake a cover sheet. This will enable the submitbutton.

6. Submit the proposal by clicking on the Submitbutton. User will get a proposal ID and anemail with a confirmation receipt. User’s pro-posal data will be forever stored in a VERITASdatabase at UCSC.

Figure 11: On the left is the webpage when it is firstopened. On the right is the actual proposal form.

3.4.2 Specific Help for the VERPS

Below contains a more detailed description of theVERPS submission process.

New Proposal/Revised submitted Proposal

• Clicking New Proposal will allow users to sub-mit a new proposal or retrieves data from cook-ies to continue working on the previous pro-posal. Clicking Revised submitted Proposal willprompt the user for the Proposal id of a submit-ted proposal as well as the corresponding PI lastname. If either are missing or incorrect, accesswill be denied.

Cover

• (See Figure ??) The information needed in thissection includes the proposal title, a two lineabstract and data on the investigators such asname, university and email. Please note thatproper email is important as the proposal systemwill email user with information on user’s sub-mitted proposal after submission. Other techni-cal details related to the proposed observationsare to be indicated here such as MW require-ments, name of the analyzers or if this is partof a PhD thesis. Clicking ”Validate” helps userssee what has been filled out already that is re-quired (green) and what still needs to be filledout (red), as well as storing all filled out informa-tion in this section as a cookie. Clicking ”Clear”will clear all the information in this section only.

Target Form

• (See Figure 12) The information needed in thissection include Source, right ascension, decli-nation, minimum elevation, requested exposure,minimum exposure, observation mode, numberof telescopes, moonlight, weather, when it is ob-servable from, and when it is observable till.RA and dec must be filled in using J2000 HHMM SS.S. Once Source, RA, and dec have beenfilled properly, the TeVCat Visibility link shouldbe enabled. Clicking ”Validate” helps users see

7

what has been filled out already that is required(green) and what still needs to be filled out (red),as well as storing all filled out information in thissection as a cookie. Clicking ”Clear” will clearall the information in this section only. Userscan click the ”Add” button to bring up as manytarget forms as they need.

Science Justification

• The information needed in this section is the sci-entific justification attached as a pdf file no morethan 3 pages long. Clicking ”Validate” helpsusers see what has been filled out already that isrequired (green) and what still needs to be filledout (red), as well as storing all filled out informa-tion in this section as a cookie. Clicking ”Clear”will clear all the information in this section only.

Once all required information has been filled, the”Make Cover” button on the bottom will be enabled.Users should click this button to view a preview oftheir cover sheet. At this point the Submit buttonshould be enabled. Users should click on the submitbutton to submit their proposal.

3.5 Technical details

The VERITAS Remote Proposal System (VERPS)was made using HTML, JavaScript, MySQL, andPHP. Ajax was used to seamlessly allow communica-tion between server-side and client-side. The jQueryJavaScript library was used to slightly reduce theamount of code at certain times. The PEAR pack-age DB was used to simplify communications betweenPHP and MySQL.

3.6 Beta testing and Final Comments

The proposal system was presented at the SummerCollaboration Meeting in Montreal in July. After ap-proval, plans for a beta testing were quickly put inplace. Throughout the week of July 25th, over 10members of VERITAS beta tested VERPs. The betatesting uncovered new errors and provided many sug-gestions on improving the proposal system. It wasrealized that the ”Target Form” did not work in old

versions of Firefox, such as version 3.0.7. The pdffile validation to prevent users from submitting morethan 3 pages did not work for some pdf files. A coversheet system and improved security were added upon.

Figure 12: ”Target” section

4 Conclusion: final thoughtsand future plans

After our failed attempt at producing a light curve forCTA 1, we will continue trying to replicate the CTA1 light curve and finish the CTA 1 analysis. Theupgrade in Fermi tools and data had come in a mostinopportune time. Careful steps must be taken toextract the methods used in previous Fermi analysis,and apply them to the new one. If this is successful,we plan to apply a similar analysis to other pulsars.

The final revisions of VERPS completion is under-way. THe final version of VERPS is planned to beput in place by Sunday, August 6th.

5 Acknowledgements

I would like to first thank Ester Aliu for her help andassistants on my projects even during late nights dur-ing this summer. I would also like to thank ReshmiMukherjee for letting me work with the VERITASgroup; I learned about the people and work involvedin this wonderful group. I would like to thank John

8

Parsons for organizing the REU program at Nevisand allowing me to participate in such a programand everyone who helped to make this program pos-sible. Finally, I would like to thank all the people whohelped in the beta testing for pointing out several bigbugs that I somehow missed during my coding andtesting.

References

[1] Abdo, A. A et al. The Fermi Gamma-Ray SpaceTelescope Discovers the Pulsar in the YoungGalactic Supernova Remnant CTA 1. Science322.5905 (2008): 1218-221.

[2] Ackermann M. et al. Fermi-LAT Search for PulsarWind Nebulae Around Gamma-Ray Pulsars. TheAstrophysical Journal 726:35 (2011).

[3] Ester Aliu Fuste. VHE γ-ray observations ofNorthern sky pulsar wind nebulae with theMAGIC Telescope. 2007.

[4] Luis Valcarcel. VERITAS, 1ES 1218 + 30.4 andthe Extragalactic Background Light. 2008.

[5] Victoria M. Kaspi and Mallory S. E. Roberts. Iso-lated Neutron Stars. 2004.

9