Home

Observations of the Active Galactic Nuleus Markarian 421

1. 1000 Entries 800 600 400 200 it Lon A 1 2 1 4 1 6 1 8 2 Angular Error degrees FIGURE 4 7 Plot showing the angular resolution of the Monte Carlo simulations For all energies the angle below which 63 of events occur is 0 29 TABLE 4 2 Angular Resolution for the Monte Carlo simulations over different energy ranges Energy Bin Angular Resolution 100 GeV 0 37 100 GeV 800 GeV 0 29 800 GeV 2 TeV 0 22 2 TeV 10 TeV 0 22 all energies 0 29 62 4 5 Comparison of Experimental and Monte Carlo Data les 8000 Entr 7000 6000 5000 4000 3000 2000 1000 0 20 40 60 80 100 120 140 160 180 200 Core Error m FIGURE 4 8 Plot showing the core resolution of the Monte Carlo simulations of all energies This data yielded a core resolution of 12 5 m Method 1 intersects the major axes of the second moment ellipses while Method 2 intersects the lines connecting the centers of gravity of the images with the source locations Method 1 works for all sources both point sources and extended sources Method 2 works only for point sources but achieves a better resolution Since Mrk 421 is a point source we use only Method 2 Shown in Figure 4 8 we obtain a core reso lution of 12 5 m over the full range of simulated energies The energy resolution is calculated using the true Etrue and reconstructed Erec energies of ea
2. D The points with error bars show the DCF computed for one of the 20 runs For this particular run the DCF peaks at a time offset of Tpcr 4 lhrs The negative sign shows that the 25 keV flux indeed leads the 3 keV flux The shaded band shows the range of DCF values obtained for the entire set of 20 runs One sees that the DCF values vary considerably especially at larger time offsets The fact that the DCF varies much more than suggested by the error bars originates from the correlation between different pairs of measurements and was discussed al ready by Edelson and Krolik 1988 Averaged over all 20 simulated runs the mean time offset at which the DCFs peak was found to be mer 4 8 hrs The minimum and maximum Tpcr values found in the 20 runs were 4 hrs and 6 hrs respectively These Tpcr values on the order of 5 hrs are substantially shorter than the corre sponding difference in cooling times of 16 hrs dh for all six models The numerical values are Figure 5 5 shows Tpcr versus At also given in Table 5 2 Most interesting is that the two times are clearly correlated but that Tpcr is always shorter than At The discrepancy varies between a fac 84 5 6 Measuring Time Lags with the Discrete Correlation Function DCF 1 24 0 24 48 72 time lag h 1 1 72 48 FIGURE 5 4 Sample plot of DCF for 3 keV vs 25 keV The time lags are shown in the observer s frame A binning of
3. Hinton J A Hofmann W Holleran M Horns D de Jager O C Johnston S Kh lifi B Kirk J G Komin N Konopelko A Latham I J Le Gallou R Lemi re A Lemoine Goumard M Leroy N Martineau Huynh O Lohse T Marcowith A Masterson C McComb T J L de Naurois M Nolan S J Noutsos A Orford K J Osborne J L Ouchrif M Panter M Pelletier G Pita S P hlhofer G Punch M Raubenheimer B C Raue M Raux J Rayner S M Redondo I Reimer A Reimer O Ripken J Rob L Rolland L Rowell G Sahakian V Saug L Schlenker S Schlickeiser R Schuster C Schwanke U Siewert M Skj raasen O Sol H Steenkamp R Stegmann C Tavernet J P Terrier R Th oret C G Tluczykont M Vasileiadis G Venter C Vincent P V lk H J and Wagner S J 2005b Astronomy and Astrophysics 442 1 Aharonian F Akhperjanian A G Aye K M Bazer Bachi A R Beilicke M Benbow W Berge D Berghaus P Bernl hr K Boisson C Bolz O Braun I Breitling F Brown A M Bussons Gordo J Chadwick P M Chounet L M Cornils R Costamante L Degrange B Djannati Ata A O C Drury L Dubus G Emmanoulopoulos D Espigat P Feinstein F Fleury P Fontaine G Fuchs Y Funk S Gallant Y A Giebels B Gillessen S Glicenstein J F Goret P Hadjichristidis C Hauser M Heinzelmann G
4. Jung L Kankanyan R Kestel M Konopelko A Kornmeyer H Kranich D Lopez M Lorenz E Lucarelli F Mang O Meyer H Mirzoyan R Moralejo A Ona Wilhelmi E Plyasheshnikov A de Los Reyes R Rhode W Ripken J Rowell G Sahakian V Samorski M Schilling M Siems M Sobzynska D Stamm W Vitale V Volk H J Wiedner C A and Wittek W 2003 Astroparticle Physics 20 267 Punch M Akerlof C W Cawley M F Chantell M Fegan D J Fennell S Gaidos J A Hagan J Hillas A M Jiang Y Kerrick A D Lamb R C Lawrence M A Lewis D A Meyer D L Mohanty G O Flaherty K S Reynolds P T Rovero A C Schubnell M S Sembroski G Weekes T C and Wilson C 1992 Nature 358 477 Rebillot P F Badran H M Blaylock G Bradbury S M Buckley J H Carter Lewis D A Celik O Chow Y C Cogan P Cui W Daniel M Duke C Falcone A Fegan S J Finley J P Fortson L F Gillanders G H Grube J Gutierrez K Gyuk G Hanna D Holder J Horan D Hughes S B Kenny G E Kertzman M Kieda D Kildea J Kosack K Krawczynski 152 Bibliography H Krennrich F Lang M J Le Bohec S Linton E Maier G Moriarty P Perkins J Pohl M Quinn J Ragan K Reynolds P T Rose H J Schroedter M Sembroski G H Steele G Swordy S P Valcarcel L Vassiliev V V W
5. VEventSeq Gets the VEventSeq for a given telescope It displays nwords for 4 telescopes This function is called automatically when the dialog is first opened Kill ALL VDAQ DACQ This button will kill any active vdaq or dacq processes for the telescope chosen at the top of the dialog box This is useful after doing bias curves to ensure everything is cleaned up and ready to take normal data again 133 C 2 Using the VAC Event Builder Telescope D 4 DACQ Run 11183 2 Reload Database Config EVTB Start Run Init VME Config EVTB End Run Get Singles Scalers Get Interrupt Status telescope trigger rate 627 Query VME Config 11 bad events E E 44 Force Query if active run events to disk 88 Node o Get EVTB Status Active EVTB Runs M T1 Mi M T2 T4 VEventSeq nwords 1 499 nwords 3 1 nwords 2 499 nwords 4 1 Kill ALL DACQ VADQ FIGURE C 6 Layout of the Event Builder subsystem window 134 C 2 Using the VAC Telescope Multiplicity v FIGURE C 7 Layout of the L2 sub system window Adjacency m Enable Expert Mode M Load Pattern Triggers cose e DACQ Contains a new set of buttons to control the VME DACQ systems through the Event Builder Reload Database Config Reload the configurations info from the database Init V ME should be called after this before a new run is started Init VME Config Initializes VME system and propogates
6. 1000 E 600 800H l P 400 7 600 i L 200 400 E pele ZOO E BEE EE Ee 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 theta degree theta degree FIGURE 4 10 Plots of 0 the square of the angular distance of the events from the source position for the Mrk 421 data On the left 6 for all data solid line and OFF data dotted line The line for OFF data was fit from the dark solid line assuming no excess for large 0 The right shows the residual excess in the 0 distribution Here a loose cut of MSCW lt 1 0 was used and extrapolated it to small values of 0 to represent all background events This fit is shown in Figure 4 10a along with the total 6 histogram The OFF line was subtracted from the full histogram leaving behind the ON excess histogram seen in Figure 4 10b The Q factor is then calculated for this residual histogram Figure 4 11 plots the Q factor as well as signal and background acceptances versus the cut in 0 This gives a maximum Q factor QMAX 2 87 with a signal acceptance of 5496 and a background rejection of 97 This equates to an optimal cut of 6 lt 0 025deg corre sponding to an angular resolution of the two telescope VERITAS system of 0 0 16 As evident in Section 4 5 our simulations appear to overestimate the angular errors as they yield a resolution of 0 29 In these early observations with a new system 66 4 6 First Stereo Result
7. 1992 3 1 2 Air Showers in the Atmosphere VHE rays gt 300 GeV require detector areas much larger than the 1 m of typical space borne telescopes in order to have any chance of detection This is not technically or financially feasible However the Earth s atmosphere is completely opaque to such high energy particles In order to detect y rays on the ground one must use an indirect technique 29 3 1 y ray Propagation Y FIGURE 3 1 When a y ray interacts with the Earth s atmosphere it pair produces and initiates a cascading shower of electrons and positrons y ray Induced Showers While impossible in free space due to energy and momentum conservation Lon gair 1992 y rays can pair produce in the Earth s atmosphere creating a cascading shower of electrons and positrons These in turn produce another high energy photon through bremsstrahlung and the process repeats Figured 3 1 schematically shows this cascading air shower The result is a tightly collimated beam of Cerenkov light that eventually hits the ground 30 3 1 y ray Propagation FIGURE 3 2 A particle traveling faster than the speed of light within a medium emits Cerenkov radiation at a specific angle given by Equation 3 3 Figure from http wikipedia org Cerenkov Radiation When a particle travels through a medium at a velocity v faster than the speed of light in that medium Cerenkov radiation is produced A shock front is creat
8. After disregarding all runs with rate spikes as well as those with irrepairable problems we were left with the final data set that was used to complete the following analysis A total of 32 runs over nine separate days corresponding to 14 3 hrs of observations were used Table 4 1 lists the data and laser calibration runs used TABLE 4 1 Run information that makes up the final data set from which this thesis is based Data runs are all of the source Mrk 421 during the April May 2006 dark run Associated laser calibration runs are also listed ON OFF pairs are denoted as ON OFF Tracking runs as Trk Wobble runs denoted as the offset distance and direction 0 3N implies the source is offset 0 3 to the North of the camera s center Date Run Number OFF Run Laser Run Run Type Elevation 2006 04 20 30336 laser 2006 04 20 30328 30329 30336 ON OFF 83 2006 04 20 30330 30332 30336 ON OFF 78 2006 04 20 30333 30336 Trk 67 2006 04 21 30358 30359 30336 ON OFF 82 2006 04 21 30365 30336 Trk 69 2006 04 23 30404 laser 2006 04 23 30394 30395 30404 ON OFF TT 2006 04 23 30396 30397 30404 ON OFF 83 2006 04 23 30398 30399 30404 ON OFF 75 2006 04 23 30400 30401 30404 ON OFF 64 2006 04 24 30424 30404 Trk 83 2006 04 26 30476 30477 30404 ON OFF 72 2006 04 26 30478 30479 30404 ON OFF 62 2006 04 26 30480 30404 0 3E 50 Continued on next page 57 4 5 Comparison of Experimental and Mo
9. Figure B 2 shows an example of data when the telescopes are not working as they should Notice the differences in the plots compared to those of Figure B 1 For example the time plot is not flat corroborating the note of a rate spike by the observers The Hillas parameter distributions are also decidedly skewed The two telescopes are also behaving very differently as the black and red histogram lines rarely overlap After a night like this the telescopes seriously need to be debugged The ROOT script also prints out diagnostic information useful in determining how well the telescopes are performing It will display a list of channels whose pedestals pedestal variances gains or time offsets vary more than 4 c from the mean It will list the number of events that have no maximum pixel as well as the number of pixels that never have the maximum value It will also list when an abnormally long gt 1sec time passes between consecutive events All these pieces of information are useful in determining the quality of the data taken the previous night 107 B 2 Analysis and Results events per time default bin size 45 5 relative gain D ntubes zoomed 50 pedestal variance int dc 5 10 15 20 25 30 35 40 45 ntubes ntubes 0 45 50 40 mean pedestal dc 5 10 15 20 25 30 35 Is E E 8 s E z s Is FIGURE B 1 Sample of daily data quality monitoring DDQM plots The
10. G H Grindlay J Hall T A Harris K Hillas A M Kaaret P Kertzman M Kieda D Krennrich F Lang M J LeBohec S Lessard R Lloyd Evans J Knapp J McKernan B McEnery J Moriarty P Muller D Ogden P Ong R Petry D Quinn J Reay N W Reynolds P T Rose J Salamon M Sembroski G Sidwell R Slane P Stanton N Swordy S P Vassiliev V V and Wakely S P 2002 Astroparticle Physics 17 221 154 Bibliography Weekes T C Cawley M F Fegan D J Gibbs K G Hillas A M Kowk P W Lamb R C Lewis D A Macomb D Porter N A Reynolds P T and Vacanti G 1989 ApJ 342 379 Weekes T C and Turver K E 1977 in R D Wills and B Battrick eds ESA SP 124 Recent Advances in Gamma Ray Astronomy pp 279 286 155
11. Lloyd Evans J Maier G Mendoza D Milovanovic A Moriarty P Nagai T N Ong R A Power Mooney B Quinn J Quinn M Ragan K Reynolds P T Rebillot P Rose H J Schroedter M Sembroski G H Swordy S P Syson A Valcarel L Vassiliev V V Wakely S P Walker G Weekes T C White R Zweerink J Mochejska B Smith B Aller M Aller H Ter sranta H Boltwood P Sadun A Stanek K Adams E Foster J Hartman J Lai K B ttcher M Reimer A and Jung I 2005 ApJ 630 130 Boettcher M Mause H and Schlickeiser R 1997 Astronomy and Astrophysics 324 395 Buckley J 1999 in International Cosmic Ray Conference pp 267 Buckley J H Akerlof C W Biller S Carter Lewis D A Catanese M Cawley M F Connaughton V Fegan D J Finley J P Gaidos J Hillas A M Kartje J F Koenigl A Krennrich F Lamb R C Lessard R Macomb D J Mattox J R McEnery J E Mohanty G Quinn J Rodgers A J Rose H J Schubnel M S Sembroski G Smith P S Weekes T C Wilson C and Zweerink J 1996 ApJL 472 L9 Burrows D N Hill J E Nousek J A Wells A Short A Turner M Citterio O Tagliaferri G and Chincarini G 2003 in G R Ricker and R K Vanderspek eds AIP Conf Proc 662 Gamma Ray Burst and Afterglow Astronomy 2001 A Workshop Celebrating the First Year of the HETE Mission
12. M Heinzelmann G Henri G Hermann G Hinton J A Hoffmann A Hofmann W Holleran M Hoppe S Horns D Jacholkowska A de Jager O C Kendziorra E Kerschhaggl M Kh lifi B Komin N Konopelko A Kosack K Lamanna G Latham I J Le Gallou R Lemi re A Lemoine Goumard M Lenain J P Lohse T Martin J M Martineau Huynh O Marcowith A Masterson C Maurin G McComb T J L Moulin E de Naurois M Nedbal D Nolan S J Noutsos A Orford K J Osborne J L Ouchrif M Panter M Pelletier G Pita S P hlhofer G Punch M Ranchon S Raubenheimer B C Raue M Rayner S M Reimer A Ripken J Rob L Rolland L Rosier Lees S Rowell G Sahakian V Santangelo A Saug L Schlenker S Schlickeiser R Schr der R Schwanke U Schwarzburg S Schwemmer S Shalchi A Sol H Spangler D Spanier F Steenkamp R Stegmann C Superina G Tam P H Tavernet J P Terrier R Tluczykont M van Eldik C Vasileiadis G Venter C Vialle J P Vincent P V lk H J Wagner S J and Ward M 2006 Science 314 1424 Aharonian F A Akhperjanian A G Andronache M Barrio J A Bernl hr K Bojahr H Calle I Contreras J L Cortina J Daum A Deckers T Denninghoff S Fonseca V Gonzalez J C G tting N Heinzelmann G Hem berger M Hermann G He M Heusler A Hofmann
13. Tavecchio 2005 and references therein The tight correlation between the X ray and TeV ray fluxes found for the blazars Mrk 421 and Mrk 501 Buckley et al 1996 Takahashi et al 1996 Krawczynski et al 2000 Sambruna et al 2000 is commonly taken as strong evidence for synchrotron Compton models in which a single population of electrons emits X rays as synchrotron emission and rays as inverse Compton emission off synchrotron target photons Since the X ray and y ray observations give complementary information about the same high energy electron population it becomes possible in principle to break model degeneracies A major goal of blazar observations is to determine the key parameters describing the jet plasma such as the bulk Lorentz factor T the jet Doppler factor the energy densities of the magnetic field the thermal particle component and the relativistic particle component X ray and TeV y ray observations reveal flux variability on time scales of minutes indicating that the emission originates very close 101 cm to 1 The jet Doppler factor is defined by Ve T 1 8 cos 0 with T the bulk Lorentz factor of the emitting plasma P its bulk velocity in units of the speed of light and 0 the angle between jet axis and the line of sight in the observer frame 72 5 2 Measurement of the Jet Magnetic Field the central engine Gaidos et al 1996 Catanese and Sambruna 2000 Krawczynski et al 200
14. W Hohl H Horns D Ibarra A Kankanyan R Kettler J K hler C Konopelko A Kornmeyer H Kestel M Kranich D Krawczynski H Lampeitl H Lindner A Lorenz E Magnussen N Mang O Meyer H Mirzoyan R Moralejo A Padilla L Panter M Petry D Plaga R Plyasheshnikov A Prahl J P hlhofer G Rauterberg G Renault C Rhode W R hring A Sahakian V Samorski M Schilling M Schr der F Stamm W V lk H J Wiebel Sooth B Wiedner C Willmer M and Wittek W 1999 Astronomy and Astrophysics 350 757 145 Bibliography Aharonian F A Hofmann W Konopelko A K and Volk H J 1997 Astropar ticle Physics 6 343 Albert J Aliu E Anderhub H Antoranz P Armada A Asensio M Baixeras C Barrio J A Bartelt M Bartko H Bastieri D Bavikadi S R Bednarek W Berger K Bigongiari C Biland A Bisesi E Bock R K Bordas P Bosch Ramon V Bretz T Britvitch L Camara M Carmona E Chilingar ian A Ciprini S Coarasa J A Commichau S Contreras J L Cortina J Curtef V Danielyan V Dazzi F De Angelis A de los Reyes R De Lotto B Domingo Santamar a E Dorner D Doro M Errando M Fagiolini M Ferenc D Fern ndez E Firpo R Flix J Fonseca M V Font L Fuchs M Galante N Garczarczyk M Gaug M Giller M Goebel F Hakobyan D Hayashida M Hengste
15. back online after a problem for example e Run Ys The run number used when a run specific command is selected If a run is selected in the main Run Information Table it becomes the default run number Multiple runs can be entered as long as they are separated by a space No other punctuation is allowed e Config Runs Configures specific runs within L3 The run number s listed above the button is used 130 C 2 Using the VAC Start Runs Start specific runs within L3 The run number s listed above the button is used End Runs Terminate specific runs within L3 The run number s listed above the button is used Pause Runs Pauses specific runs within L3 The run number s listed above the button is used Resume Runs Resumes specific runs within L3 The run number s listed above the button is used Abort Runs Aborts specific runs within L3 The run number s listed above the button is used Reset Does a soft reset of L3 Clear Error Clears L3 of error status Get Status Displays the status of L3 The status items are described below L3 Alive Identical to the indicator in the main window it tells if L3 is connected to arrayctl It is only updated when the user clicks Get Status L3 Status The box s title contains both a string describing the current L3 Status and a number referring to the L3 Status Bits The other items in the box are self explanatory run number run status config mask coincidence
16. cosmic ray shower propagates through the atmosphere 3 2 Detection Using an Imaging Atmospheric Cerenkov Telescope IACT 3 2 1 Atmospheric Technique While the Earth s atmosphere has negative effects on most astronomical observa tions with clouds and air currents distorting an astronomer s view of the heavens it is a very necessary component for ground based y ray observations In fact one is using the atmosphere as the detector medium to greatly increase the apparent collection area for such a telescope The main drawback of using the atmosphere is 33 3 2 Detection Using an Imaging Atmospheric Cerenkov Telescope IACT Nucleon Cascade e e e e EM Cascade EM Cascade EM Cascade EM Cascade FIGURE 3 4 A cosmic ray induced air shower is much more extended than that of a y ray induced shower Figure adapted from Jelley 1958 34 3 2 Detection Using an Imaging Atmospheric Cerenkov Telescope IACT its unpredictability Its transparency affects the amount and angular distribution of Cerenkov light seen on the ground Cloud layers can cause errors in the data includ ing a higher detection threshold and inaccurate energy reconstruction Hence data must be taken on clear nights to be completely effective 3 2 2 Imaging Cerenkov Radiation When trying to observe a y ray source in the night sky most of one s view is dominated by background noise and cosmic ray induced showers Cerenkov light is on
17. option should never be used it confuses L3 Observing Mode on off tracking survey parked drift engineering cal ibration other describes what specific type of run is being taken so the analysis programs handle the data properly Options appear in a pop up list For Observing runs the only options are on off tracking survey parked drift other For all other runs the only options are engineering calibration other Pointing Mode parallel convergent NA other describes where the tele scope is pointing for the given run Options appear in a pop up list For Observing runs the only options are parallel convergent other For all other runs the only options are NA other Source source name from the known sources list Options appear in a pop up list The source list can be edited by accessing the Database 119 C 2 Using the VAC subsystem Trigger Config normal external muon other defines the current con dition for the telescope to trigger Options appear in a pop up list 8 Trigger Multiplicity Number of telescopes needed for an array trigger 9 Trigger Coincidence Window width in ns to look for coincident events 10 11 12 13 14 15 16 Currently for multiplicity 1 this is 0 for multiplicity gt 1 it is 100 Comments any number of comments can be added to each run to de scribe situations not covered in any other field Comments may also be added later during th
18. problem occurs in the process it may be aborted and the later items will not be completed as requested C 2 5 Settings Menu Various settings used by the Event Builder are stored in the database These settings effect the CFDs and FADCs The Settings menu contains interfaces to alter both sets of settings change the current configuration and load save the settings to a file Note The CFD and FADC Settings are not yet complete and should not be used Doing so may adversely effect the system or the database records Put CFD Settings The Put CFD Settings window is shown in Figure C 10 This panel will run Liz s vdbput CFDSettings script with various options e Telescope Choose which telescope you are changing the settings for 1 4 141 C 2 Using the VAC e Threshold If this box is checked thresholds will be changed to the value listed Note en tering a POSITIVE number will yield a NEGATVE mV setting i e Threshold 100 set to 100 mV e Width If this box is checked widths will be changed to the value listed 4 25 ns e RFB If this box is checked RFB s will be changed to the value listed 0 127mV MHz e Put Settings Runs the script with the above options in a new terminal window It will also tell VDAQ to reload the settings from the DB When it has finished you may close the window and put settings for another telescope e Cancel Does nothing closes the window CFD Settings This secti
19. 0 3 HESS J1837 069 18 37 37 4 06 56 42 G25 5 0 0 AX J1838 0655 HESS J1303 631 13 03 00 4 63 11 55 HESS J1614 518 1614190 51 49 07 HESS J1702 420 1702446 42 04 22 HESS J1708 410 17 08 143 41 04 57 HESS J1745 303 1745 02 2 30 22 14 3EG J1744 3011 TeV J2032 4131 20 31 57 41 29 56 8 There are several classes of TeV y ray sources the first of which are supernova remnants SNRs The expanding shell of gas from a supernova explosion consists of stellar material altered by the explosion as well as parts of the interstellar medium swept up during expansion In shell type SNR like RX J1713 7 3946 the emission appears to come from an outer shell with no apparent central power source Plerions like the Crab Nebula are thought to be powered by a central pulsar The Crab Nebula is a particularly interesting TeV source for many reasons The supernova that created it exploded in 1054 AD and was observed by Chinese as tronomers It left behind a bright spot in the sky visible in daylight for weeks after The Crab was the first confirmed TeV y ray source discovered by Weekes et al 1989 It has since become known for its strong steady signal Today it is used as the standard candle by which all y ray observations are measured A view of the Crab with the Hubble Space Telescope can be seen in Figure 2 6 17 2 3 TeV y ray Sources FIGURE 2 6 The Crab Nebula as seen by the Hubble Space Tel
20. 15 minutes has been used This DCF peaks at a negative lag of 4 1 hrs corresponding to the 25 keV flux variability leading the 3 keV flux variability This is the plot from set D run 14 The bounding curves represent the range over which all 20 runs DCFs occurred The dotted vertical lines represent the range over which the maximum DCF occurs for the 20 runs The average observed lag of 4 8 hours is much shorter than the one computed from the synchrotron cooling times for this data set of 16 hours thick dashed line 85 5 6 Measuring Time Lags with the Discrete Correlation Function DCF TABLE 5 2 Time lags for each set Errors on DCF values are root mean square values calculated from averaging 20 runs per set The first two DCFs are for 3 keV vs 25 keV The final DCF is for 25 keV vs 1 TeV scaled by the appropriate factor Set Tpcr with IC h cp without IC h Tpoe 0 R c Atsynch h A 0 25 0 01 0 25 0 01 0 55 0 02 0 29 B 1 00 0 01 1 25 0 01 0 49 0 09 1 4 C 1 74 0 05 1 97 0 07 0 45 0 10 2 6 D 4 79 0 70 5 72 0 82 0 67 0 30 16 E 7 34 2 70 17 1 12 0 1 3 1 9 55 F 6 94 2 73 16 5 12 5 16 22 62 tor of 1 and a factor of 9 for the strongest A and weakest F simulated magnetic field values respectively For all but the smallest magnetic fields the spread of the Tpcr values derived from different runs is rather small Thus assuming an o
21. 1ES 23444 514 23 47 04 92 51 42 17 9 H 2356 309 23 59 07 8 30 37 38 Radio Galaxy M 87 12 30 49 42 12 23 28 0 Plerion Crab Nebula 05 34 31 97 22 0052 1 PSR B05324 21 Vela X 08 33 32 45 4342 PSR B0833 45 G313 34 0 1 14 18 04 60 5831 Rabbit R2 Kookaburra K3 Kookaburra 14 20 09 60 48 36 PSR J1420 6048 MSH 15 52 15 14 07 5909 27 PSR B1509 58 composite G18 0 0 7 1826030 13 45 44 PSR J1826 1334 Shell Type SNR RX J0852 0 4622 08 52 00 46 20 00 Vela Junior RX J1713 7 3946 17 13 00 39 4500 G347 3 0 5 G0 9 0 1 17 47 232 28 09 06 composite G12 82 0 02 18 13 36 6 17 50 35 Cas A 23 23 24 58 48 54 Pulsar LS 5039 18 26 15 14 50 53 6 LS I 61 303 02 40 31 67 61 13 45 6 also an X ray binary X ray Binary PSR B1259 63 13 02 47 65 63 50 08 7 Unidentified HESS J1616 508 16 16 23 6 505357 PSR J1617 5055 HESS J1632 478 16 3208 6 474924 IGR J16320 4751 HESS J1634 472 163457 2 471602 G337 2 0 1 IGR J16358 4726 Continued on next page 16 2 3 TeV y ray Sources TABLE 2 2 Continued Name RA 2000 Dec 2000 Notes HESS J1640 465 16 40 44 2 46 31 44 G338 3 0 0 3EG J1639 4702 HESS J1713 381 1713580 38 11 43 G348 7 0 3 HESS J1745 290 1745413 29 00 22 G359 95 0 04 SgrA East SgrA HESS J1804 216 1804 31 6 21 42 03 G8 7 0 1 PSR J1803 2137 HESS J1834 087 18 3446 5 08 45 52 G23 3
22. 2 Debugging Systems The following steps are taken to use the VAC before the telescopes have been brought into full operation or to debug systems at any time This process involves going to the individual subsystems menus and accessing them directly 1 Start the VAC program on the arrayctl computer gt VAC The following commandline options may also be used e f filename specifiy a different configuration file to use e n lt nameserver gt specify a different CORBA nameserver to use 114 C 2 Using the VAC e q host specify a different host to provide QuickLook data e t disable reading of temperatures of FADC boards e s disable automatic Update Status calls e d disable automatic checking for Free Disk Space remaining 2 Start up the necessary subsystems using the Start Subsystems option in the Observer menu Check which subsystems have established contact by looking in the status portion of the main window Take the appropriate steps to initialize systems with which there is no contact as outlined in the related wiki pages 3 Choose the appropriate item from the Subsystem menu see Sect C 2 4 This will bring up a spearate dialog from which you can access the debugging com mands for that system The options are as follows e L3 Ctrl4 L e Harvester Ctrl4 H e EventBuilder Ctrl E e L2 Ctrl4 X e L1 Ctrl4 Y display L1 rates currently unsupported e DB Ctrl D access items in
23. 6 3 The Future of y ray Astrophysics We hope that the observations will make it possible to identify unambiguously the nature of the accelerated particles protons or electrons positrons Once the emis sion mechanism is interpreted the observations will give information about the jet parameters magnetic field particle to magnetic field energy density etc and thus contribute to our understanding of the structure of AGN jets 98 Appendix A X ray Data Analysis of 1ES 19594 650 and Mrk 421 A 1 Multiwavelength Campaign Overview Observing blazars in TeV rays reveals only a small portion of the information that can be gained from them Blazars emit energy over a wide range of energies To truly understand how they work one must look at data from more than one energy band To this end multiwavelength campaigns have been mounted on several known blazars Through these many telescopes working at different energies observe the same source at the same time to reveal a wealth of information not available to a lone observer Looking at the SED for blazars it is obvious where one should probe besides y rays for information The first peak of the SED occurs in X rays Monitoring this 99 A 2 RXTE Data highly active energy band will then yield more insight into the complex processes inside these interesting sources This data analysis was a part of two separate multiwavelength campaigns The first looked at the TeV blaza
24. Daniel M D Vali M de la Calle Perez I Duke C Falcone A Fegan D J Fegan S J Finley J P Fortson L F Gaidos J A Gammell S Gibbs K Gillanders G H Grube J Gutierrez K Hall J Hall T A Hanna D Hillas A M Holder J Horan D Jarvis A Jordan M Kenny G E Kertzman M Kieda D Kildea J Knapp J Krawczynski H Krennrich F Lang M J Le Bohec S Linton E Lloyd Evans J Milovanovic A McEnery J Moriarty P Muller D Nagai 150 Bibliography T Nolan S Ong R A Pallassini R Petry D Power Mooney B Quinn J Quinn M Ragan K Rebillot P Reynolds P T Rose H J Schroedter M Sembroski G H Swordy S P Syson A Vassiliev V V Wakely S P Walker G Weekes T C and Zweerink J 2004 ApJL 608 L97 Krawczynski H 2004 New Astronomy Review A8 367 Krawczynski H 2005 ArXiv Astrophysics e prints astro ph 0508621 Krawczynski H 2006 ArXiv Astrophysics e prints astro ph 0610641 Krawczynski H Coppi P S and Aharonian F 2002 MNRAS 336 721 Krawczynski H Coppi P S Maccarone T and Aharonian F A 2000 Astron omy and Astrophysics 353 97 Krawczynski H Hughes S B Horan D Aharonian F Aller M F Aller H Boltwood P Buckley J Coppi P Fossati G G tting N Holder J Horns D Kurtanidze O M Marscher A P Nikolashvili M Remillard R A S
25. Djannati Atai A Dumora D Espigat P Fabre B Fleury P Fontaine G George R Ghesquiere C Gilly J Goret P Gouiffes C Gouillaud J C Gre gory C Grenier I A Iacoucci L Kalt L Le Bohec S Malet I Meynadier C Mols J P Mora de Freitas P Morano R Morinaud G Munz F Palatka M Palfrey T A Pare E Pons Y Punch M Quebert J Ragan K Renault C Rivoal M Rob L Schovanek P Smith D Tabary A Tavernet J P Toussenel F and Vrana J 1998 ArXiv Astrophysics e prints astro ph 9804046 Bednarek W 1997 MNRAS 285 69 Bednarek W and Protheroe R J 1999 MNRAS 310 577 146 Bibliography Blandford R D and Rees M J 1978 in A M Wolfe ed Pittsburgh Conference on BL Lac Objects Pittsburgh Pa April 24 26 1978 Proceedings A79 30026 11 90 Pittsburgh Pa University of Pittsburgh 1978 p 328 341 Discussion p 341 347 NATO supported research pp 328 341 Blazejowski M Blaylock G Bond I H Bradbury S M Buckley J H Carter Lewis D A Celik O Cogan P Cui W Daniel M Duke C Falcone A Fegan D J Fegan S J Finley J P Fortson L Gammell S Gibbs K Gillanders G G Grube J Gutierrez K Hall J Hanna D Holder J Horan D Humensky B Kenny G Kertzman M Kieda D Kildea J Knapp J Kosack K Krawczynski H Krennrich F Lang M LeBohec S Linton E
26. H Cui W Jernigan J G Morgan E H Remillard R Shirey R E and Smith D A 1996 ApJL 469 L33 Levinson A 2005 in KITP Program Physics of Astrophysical Outflows and Accre tion Disks 151 Bibliography Longair M S 1992 High Energy Astrophysics Volume 1 Particles photons and their detection Cambridge University Press Cambridge Lorenz E and Martinez M 2005 Astronomy and Geophysics A6 21 Maraschi L Ghisellini G and Celotti A 1992 ApJL 397 L5 Mastichiadis A and Kirk J G 1997 Astronomy and Astrophysics 320 19 Mimica P Aloy M A M ller E and Brinkmann W 2004 Astronomy and Astrophysics 418 947 Morselli A 2003 Chinese Journal of Astronomy and Astrophysics Supplement 3 523 Perkins J S 2006 Ph D thesis Washington University in St Louis Pian E Vacanti G Tagliaferri G Ghisellini G Maraschi L Treves A Urry M Fiore F Giommi P Palazzi E Chiappetti L and Sambruna R M 1998 ApJL 492 L17 Piner B G and Edwards P G 2004 ApJ 600 115 Piner B G and Edwards P G 2005 ApJ 622 168 P hlhofer G Bolz O G tting N Heusler A Horns D Kohnle A Lampeitl H Panter M Tluczykont M Aharonian F Akhperjanian A Beilicke M Bernl hr K B rst H Bojahr H Coarasa T Contreras J L Cortina J Denninghoff S Fonseca M V Girma M Heinzelmann G Hermann G Hof mann W
27. Henri G Hermann G Hinton J A Hofmann W Holleran M Horns D de Jager O C Kh lifi B Komin N Konopelko A Latham I J Le Gallou R Lemi re A Lemoine M Leroy N Lohse T Marcowith A Masterson C McComb T J L de Naurois M Nolan S J Noutsos A Orford K J Osborne J L Ouchrif M Panter M Pelletier G Pita S P hlhofer G Punch M Raubenheimer B C Raue M Raux J Rayner S M Redondo I Reimer A Reimer O Ripken J Rob L Rolland L Rowell G Sahakian V Saug L Schlenker S Schlickeiser R Schuster C Schwanke U Siewert M Sol H Steenkamp R Stegmann C Tavernet J P Terrier R Th oret C G Tluczykont M Vasileiadis G Venter C Vincent P V lk H J and Wagner S J 2005c Astronomy and Astrophysics 437 95 Aharonian F Akhperjanian A G Bazer Bachi A R Beilicke M Benbow W Berge D Bernl hr K Boisson C Bolz O Borrel V Braun I Breitling F Brown A M Chadwick P M Chounet L M Cornils R Costamante L Degrange B Dickinson H J Djannati Atai A O C Drury L Dubus G Emmanoulopoulos D Espigat P Feinstein F Fontaine G Fuchs Y Funk S Gallant Y A Giebels B Gillessen S Glicenstein J F Goret P Hadjichristidis C Hauser M Heinzelmann G Henri G Hermann G Hinton J A Hofmann W Holleran M Horns D
28. Nebula the Whipple 10 m telescope gives I 2 50 1 Rebillot et al 2006 The results on F are comparable to previous results from earlier experiments where Mrk 421 was in a high state e g the Whipple 10 m Krennrich et al 2002 HEGRA Aharonian et al 1999 and H E S S Aharonian et al 2005c 70 Chapter 5 SSC Modeling of Blazar Emission 5 1 Rationale The previous chapter describes the analysis of VERITAS observations of the blazar Mrk 421 Extracting astrophysical results will require one to observe the source with good observational coverage over the entire electromagnetic spectrum Such intensive multiwavelength observation campaigns are planned for the years 2007 and 2008 In this chapter we present a theoretical study relevant for such multiwavelength studies We explore whether X ray measurements alone can be used to constrain the magnetic field in jets If this were possible the combined X ray and y ray measurements could be used to break additional model degeneracies 71 5 2 Measurement of the Jet Magnetic Field 5 2 Measurement of the Jet Magnetic Field During the last decade time resolved multiwavelength observations of blazars have emerged as a powerful tool to study AGN jets The most detailed data so far comes from joint observations of satellite borne X ray telescopes ASCA BeppoSAX RXTE XMM and ground based TeV y ray telescopes CAT HEGRA Whipple see Krawczynski 2004 Krawczynski 2005
29. OO Num Trigger 5 A A Num Tubes 0 CC Let XSX9C KICK X90 NumDead42 CXXX 00009009 XO KC XC XC XC X 60 0900 EEL 995 900 os Primary 0 y Energy TeV 0 00 X 0 00 Y 0 00 Xcos 0 000 Ze 0 00 Ycos 0 000 Az 0 00 FIGURE 4 2 Examples of actual sky shower images by one VERITAS camera a y ray top left a hadron top right a muon bottom left and a noise event bottom right Pixels in grey are marked as broken in the camera and are not being read out the result of poor connections and cabling among other things Fixing these pixels is one of the many tasks yet to be done to bring VERITAS into full operational mode 48 4 1 Event Reconstruction Shower axis Cherenkov light image Ka i CT4 9 degree Telescope focal plane FIGURE 4 3 Reconstructing the arrival direction of stereo showers The shower impact point and the resulting images in each camera are shown on the left The images from each camera are then superimposed right Solid ellipses portray a shower vertically incident to the telescopes dotted ellipses correspond to a shower inclined by an angle 09 Figure from Aharonian et al 1997 49 4 1 Event Reconstruction and size values for the image in the ith telescope and r is that telescope s distance from the shower axis MSCW is more useful than simply looking at width distri butions for multiple telescopes because it takes into
30. Orphan Flares e a TeV Flux Crab 4 2l k x geri ik ko B x d diu de de m IM b 10 keV Flux keV cm s 0 002 e ga gue e poa INN TARTA c 3 25 keV Photon Index 2 5 o o o M ath TU ug oq So gas each T NOS Pg T5 Ex hes Li II oe e yapa jeje al dS o a al anne a 15 2 m irls pie ME prse qe E Mapriinides 0 41 14 5 GER Flux Jy 0 2 RER 4 kA e pde og A pd ue O en TE ap e sp pp EEN 20 25 30 35 40 45 Date MJD 52400 FIGURE A 1 Light curve of 1ES 19594 650 from 2002 multiwavelength campaign a Whipple stars and HEGREA circles integral TeV y ray fluxes in Crab units above 600 GeV and 2 TeV respectively The Whipple data are binned in 20 minute bins the HEGRA data are in diurnal bins b RXTE X ray flux at 10 keV c RXTE 3 25 keV X ray photon index d Absolute R magnitudes e The 14 5 GHz flux density Figure from Krawczynski et al 2004 103 A 3 Orphan Flares Blazejowski et al 2005 and early results indicated this one zone model could sat isfactorily describe a wealth of data However as Krawczynski et al 2002 showed this method fails even for the blazar Mrk 501 Clearly a one zone SSC model is too simple to account for the full complexity of a blazar s inner workings Further inves tigations into the phenomenon of orphan flares are necessary to better understand these processes 104 Appendix B Daily VERITAS D
31. Run Insane Events 1456 Events to Harvester 320 r r dueto Tel 1 0 Events to Disk 640 due to L3 0 Bad Events 80 Run File Size MB i Em i Messages esc CDE command arrayCtlToGui mesg Terminating run 11183 L2 Rate command arrayCtIToGui mesg Run 11183 successfully terminated I think VDAQ Deadtime Run has ended move the telescope if necessary New run created with ID 11184 L3 Triggers Recorded 0 command arrayCtlToGui mesg Preparing run 11184 L3 command arrayCtlToGui mesg Expect a delay of 5 to 30 seconds now Run Info for Telescope ALL Status command arrayCtiToGui mesg Run 11184 prepared command arrayCtlToGui mesg Starting run 11184 command arrayCtlToGui mesg Run 11184 started successfully I think L3 Rate Error reading free disk space info Error reading free disk space info T4 TEL EVTB TE F Temps F Active Run M Run in Progress 11184 2007 01 05 22 24 20 Tel 1 Events Recorded Array Deadtime L2 L3 Rates Run Management completed 11182 CRAB A 28 00 observing tracking paralle completed 11183 CRAB 28 00 observing tracking paralle approx seconds since sample mp T2L2 Tate T4 L2 s D F Log10 Rate Rescale Rate Clear Plot A Define Run Add Comment End Run Start Run FIGURE C 1 Layout of the main VAC window 116 C 2 Using the VAC Currently the connection is checked only once at program startup If there is a problem make sure the harvester is running and restart VAC T1 T4
32. TEL EVTB QI L2 VDAQ Temps Displays the status of telescope specific systems telectl eventbuilder Charge Injection L2 VDAQ and FADC Temperature readings for Telescopes 1 4 re spectively Information is updated automatically A checked box means the CORBA connection is present unchecked means there is a problem While the current indicator system allows the user to click and change the status of the systems it will be overwritten with the correct status automatically the next time the system does its automatic checking Currently QI L2 and VDAQ do not have properly functioning CORBA connections They will not be checked even when they are running properly Regarding FADC Temperatures if the crate temperatures get too high usually over 55 C a warning message is printed Try reading the temperatures again because the fluctuation is so high the state may pass A further warning is printed when a board exceeds 60 C In this case they should be immediately shut down until the temperature decreases If there is a problem in this process or if more information is needed the Observer Read FADC Temperatures menu option may be used to display all temperature values Temperatures are read every few minutes An OK statement will appear if none of the boards show excessively high temperatures If the Temps indicators are greyed out automatic temperature readings have been disabled T1 T4 Active Run For convenience this box displays w
33. Times lt lt 4 ko RR 87 5 6 Actual vs Calculated B Fields 88 5 7 DCF over Different Energy Ranges lll 89 5 8 DCF for Flares of Different Durations e 90 6 1 Observatory Sensitivity e ooh Rem Rmo REIR LIP CAM SOS 96 A 1 Orphan Flare of 1ES IOpn0tob0 103 B 1 Representative Good DDQM Plots 108 B 2 Representative Poor DDQM Plots 109 B 3 Thedt PA i rues I vnu p ra e es AS ee 111 C 1 Main VAC Window at e a M RET hh Ne seo eut 116 C 2 Define Run Window toro Vots oe HE Seok oh RS okt wh eie Ih el 118 C 3 Run Info Window es mura aa Ss dd bos bee vAS 128 Cd L3 Subsystem Window dee demo ERE VPE sias e 130 C5 Harvester Subsystem Window e gy ve y vu TES 132 C 6 Event Builder Subsystem Window 134 C 7 L2 Subsystem Window s d iron er COS fe oO E ee 135 C 8 Database Subsystem Window 137 C 9 Custom Night Window si rs e det Hey eas 140 C 10 Put CFD Settings Window uiuos e a de 28 141 vill List of Tables dolo ra Nomenclature s s o xc o Era ge ratus Ve Goes uario 5 2 25 TEV y ray o o ui amp open dr q monde dug Ner ru Uer ro at ede et 16 4 1 Final Mrk 421 Data Set aede A MP e eg ud 57 4 2 Monte Carlo Angular Resolution a 62 9 1 Parameters for Simulated Data Files 78 5 2 Calculated and Observed Time Lags 86 5 3 Observed and Calculated B fields ln
34. W Lin Y C Mattox J R Mayer Hasselwander H A Michel son P F von Montigny C Nolan P L Pinkau K Rothermel H Schneid 148 Bibliography E J Sommer M Sreekumar P and Thompson D J 1993 Astronomy and Astrophysics Supplement Series 97 13 Gaidos J A Akerlof C W Biller S D Boyle P J Breslin A C Buckley J H Carter Lewis D A Catanese M Cawley M F Fegan D J Finley J P Hillas A M Krennrich F Lamb R C Lessard R McEnery J Mohanty G Moriarty P Quinn J Rodgers A Rose H J Samuelson F Schubnell M S Sembroski G Srinivasan R Weekes T C Wilson C L and Zweerink J 1996 Nature 383 319 Gehrels N Chipman E and Kniffen D A 1993 Astronomy and Astrophysics Supplement Series 97 5 Ghisellini G 1999 Astroparticle Physics 11 11 Ghisellini G Tavecchio F and Chiaberge M 2005 Astronomy and Astrophysics 432 401 Gingrich D M Boone L M Bramel D Carson J Covault C E Fortin P Hanna D S Hinton J A Jarvis A Kildea J Lindner T Mueller C Mukherjee R Ong R A Ragan K Scalzo R A Theoret C G Williams D A and Zweerink J A 2005 ArXiv Astrophysics e prints astro ph 0506613 Ginzburg V L and Syrovatskii S I 1969 Ann Rev Astron Astrophys 7 375 Gould R J and Schr der G P 1967 Phys Rev 155 1408 Hartman R C Bertsch D
35. d used Ue tt op eel Ge nzen le DURS mue S cerle ii viii ix Contents 3 y ray Detection and VERITAS 28 Ab y ray Propapubloto i o eeu but AT ba ELA Ae 28 3 1 1 Propagation Through Space 23 xd aia 29 3 1 2 Air Showers in the Atmosphere 29 y ray Induced Showers a uoto id RE eh dire S 30 erenkov Radiation ad som debated ebat idet 31 Cosmic Ray Induced Showers 004 33 3 2 Detection Using an Imaging Atmospheric Cerenkov Telescope IACT 33 3 2 1 Atmospheric Technique a sal n deu edm id d 33 3 2 2 Imaging Cerenkov Radiation els 35 3 3 VERITAS Very Energetic Radiation Imaging Telescope Array System 37 Ao Telescope Array as a tia Yon ee A rax xq 37 39 2 AMERO uie ur al ated e Re ho e ica 39 pond QDEISBOE e er bote decens Aw Sp de di ore e Sel re 42 3 3 4 Data Acquisition ae sre sede eR Sd SERIO SOS 43 4 Mrk 421 Data Analysis 44 4 1 Event Reconstruction esca EE e ot AE e 44 4 1 1 Hillas Parameterization 694 4 4 4 8 E ie Pa 45 4 1 2 Stereo Reconstruction 47 4 1 3 EE EE EE EE oy 50 4 2 Data Cuts and Significances e ee 51 4 3 Spectral Reconstruction o oao sta ate xS 52 4 4 Mrk 421 Data from April May 2006 a 53 4 4 1 Observation Modes s 3 e 9X da 53 ON OEE Pairs Vna e EE RE PO M REPE RUP ES 54 Tracking FUIS A en 6 B AR 54 Wobble Runs ef aid daa ESAS ee iS 56 AAD malla urne ve dtu pus du ergo ED Yay oe E aloe fg 56 4 5 Comparison
36. entered into the database Save New Info The observer is allowed to change some other aspects of a run s information in the database in case an error was made in defining that run Weather run type pointing mode trigger config source RA DEC offsets wobble offset and wobble angle can all be changed Clicking this button commits those changes to the database If changes are made to any run it is updated in the Run Information Table in the main window e Add Comment Allows observers to add comments to any run in the Run Info Table without having to first open the Run Info window C 2 3 Test Runs Menu For debugging and when things just won t work correctly these shortcuts can be used to start a run without having to define all the necessary info When chosen a run is automatically defined prepared and immediately started It is tagged as a 5 minute chargelnjection engineering run parked with C weather and a default observer Which telescopes are involved depends on which option is chosen Options appear as separate menu items to make starting a test run as effortless as possible e Make Test Run ALL Configures and starts a test run on ALL telescopes in the current system As a shortcut use Ctrl M e Make Test Run T1 Configures and starts a test run on Telescope 1 As a shortcut use Ctrl 1 e Make Test Run T2 Configures and starts a test run on Telescope 2 As a shortcut use Ctr1 2 e Make Test Run T3 Configures a
37. of Experimental and Monte Carlo Data 58 4 6 First Stereo Results from VERITAS 64 e Me esc xc Due ask Den AA ews eke eh wh seule howe Heke 65 4 6 2 Cutting on Mean Scaled Width 67 4 6 3 Mrk 421 Light Curve oe Se ee edo a aah op dee used 68 4 6 4 Energy Spectrum da cala de le a eR s 70 5 SSC Modeling of Blazar Emission 71 ak Rationale s dene O ear Ae Mt Saige ae Modos c te Mrd oe ot oe E d ier e 71 5 2 Measurement of the Jet Magnetic Field 72 5 3 Synchrotron SelfCompton Simulations 74 5 4 Generating Data Sets id E io ER 76 5 5 Analysis Procedure 522514 a a RE ES 81 5 6 Measuring Time Lags with the Discrete Correlation Function DCF 84 Contents 5 7 Comparing of DCF Time Lags to Expected Results 91 Discussion 94 6 1 Summary of Thesis Results 44 4 qoe 4 xx 94 ox ier ee ES 94 6 2 V ERLEASPertormisneece c3 S eos Ea a xi ve RS 95 6 3 The Future of y ray Astrophysics 97 X ray Data Analysis of 1ES 19594 650 and Mrk 421 99 A 1 Multiwavelength Campaign Overview 99 2 2 RXTE Datis dae tcn oic Su ende p v ene we argues 100 A Orphan Flar s dole a xoxo Ux S XR B ote RA 102 Daily VERITAS Data Quality Monitoring 105 B 1 Motivation and Procedure 0 0 000000 105 B 2 Analysis and Results i opa wb ORS ee BY ee RS er 106 B 3 The dt Bump 05 65 078 succo ite P Sh ope p ers 110 VAC VERITAS Array Control GUI 112 Cu Star
38. parameters were altered and different scenarios tried in order to test the accuracy of this method Understanding the behavior of properties such as this will allow the breaking of model degeneracies and give insight into the physical processes involved in particle acceleration of blazars The text of this thesis is organized as follows An introduction to y ray astro physics is covered in Chapter 2 focusing on blazars and their observation A descrip tion of how 4 rays are detected on Earth and the VERITAS telescopes in particular are covered in Chapter 3 Following are details on the analysis and results from the first VERITAS stereo data in Chapter 4 A complete description of the SSC simu lations used to test the DCF as a tool for measuring the magnetic field of sources is given in Chapter 5 concluding with a discussion of all results as well as the future of y ray astrophysics and VERITAS in Chapter 6 The Appendices describe various other studies I have performed during the course of this thesis A brief overview of the X ray analyses of Mrk 421 and 1ES 1959 650 are covered in Appendix A A summary of daily data quality monitoring DDQM for the VERITAS telescopes is given in Appendix B Finally a description and User s Manual for the graphical interface VAC is presented in Appendix C Chapter 2 Astrophysics of Blazars 2 1 Introduction to y ray Astrophysics At the high energy end of the electromagnetic spectrum lie what
39. prevent their being used to detect y rays at energies gt 300 GeV To probe higher energy y rays it is necessary to use ground based detectors 2 2 2 Ground based Instruments The Earth s atmosphere is opaque to y rays However it is still possible to de tect the results of their interactions with the Earth s atmosphere This process is described more in Chapter 3 Ground based techniques have proven highly effective in observing and discovering new sources of y rays Cerenkov telescopes in particular have discovered TeV emission from seven blazars five of which were not detected by EGRET Horan and Weekes 2004 Aharonian et al 2005a Imaging Atmospheric Cerenkov Telescopes Taking over where space based detectors leave off Imaging Atmospheric Cerenkov Telescopes IACTs operate in the 30 GeV 30 TeV range First proposed by Weekes and Turver 1977 this technique is based on detecting flashes of Cerenkov light resulting from interactions as the primary y ray passes through Earth s atmosphere It is discussed further in Section 3 2 Stand alone IACTs have been operating for years with optical reflectors ranging from 3 17 m in diameter Examples of early and present telescopes include CAT in France Barrau et al 1998 the Whipple 10 m in Arizona USA Cawley et al 1990 and MAGIC in the Canary Islands Lorenz and Martinez 2005 An example of this type of telescope can be seen in Figure 2 3 10 2 2 Instruments to
40. the database e QI charge injection currently unsupported C 2 Using the VAC C 2 1 Main Window The main window of VAC is shown in Figure C 1 System Status e Harvester QuickLook L3 DB PCS Displays the status of array specific systems the Harvester QuickLook L3 the Database and the Positioning System Information is updated automatically A checked box means the CORBA connection is present unchecked means there is a problem While the current indicator system allows the user to click and change the status of the systems it will be overwritten with the correct status automatically the next time the system does its automatic checking Following L3 is a string showing the currently reported subsystem status The positioner currently does not have a CORBA interface so 1t will never be checked Quicklook is an exception to the CORBA connection Though done on the har vester it involves a direct connection and not the normal CORBA connection 115 C 2 Using the VAC VAC VERITAS Array Control File Observer Test Runs Subsystems Settings Help Run Info for Current Active Run System Status 221739 Wl 22 45 39 M Harvester 7 M L3 Howzaboy iv DB Run Number 11184 Elapsed 00 06 41 Remaining 00 21 19 Ti MELO EVTB Je 3 Ts pa fe T Temps M Active Run Harvester Event Builder 1 T2 M TEL lv EVTB FP F Temps M Active Run E r Sane Events 7280 Telescope Trigger Rate 4560 T3 l TEL MT EVTB r r Temps F Active
41. will lose their energy more rapidly than low energy electrons The flux variability at higher energies should lead the variability at lower energies and a measurement of the 73 5 3 Synchrotron Self Compton Simulations time lag between the flux variability observed in different bands should constrain the magnetic field inside the jet plasma With this aim various authors have used the Discrete Correlation Function DCF of Edelson and Krolik 1988 to search for a time lag between fluxes measured in different X ray bands and to constrain the jet magnetic field Unfortunately the measurements have not produced a clear verification of the basic picture Based on a four day uninterrupted observation of the blazar Mrk 421 with the ASCA satellite Takahashi et al 2000 used the DCF formalism to determine a time lag between the 0 5 1 keV and 3 7 5 keV fluxes Rather than revealing a constant time lag indicative of synchrotron cooling of electrons in a certain magnetic field the analysis showed that the time lag changed constantly in its value and its sign Based on four XMM observations of 10 hrs duration Sembay et al 2002 performed a DCF analysis and derived an upper limit on the time lag between the soft 0 1 0 75 keV and hard 2 10 keV emission of 300 sec This upper limit translates into a lower limit on the magnetic field of 2 6 10 G uncomfortably high for synchrotron Compton models 5 3 Synchrotron
42. 0 E 0 00614 4 78 9 8 0 4 2056 200 F 0 00562 4 34 8 9 0 1 2346 234 blazars Mrk 421 and Mrk 501 Krawczynski et al 2000 2001 2002 We use 6 45 throughout this analysis which agrees well with Mrk 501 and Mrk 421 data and which assures that the internal electron densities are sufficiently low such that internal pair production processes are negligible Figure 5 2 shows sample SEDs for each parameter combination These specific combinations were chosen so that the resulting SEDs would resemble the known blazars Mrk 421 and Mrk 501 To keep the results generic we give results in rest frame luminosities per solid angle and we do not apply extragalactic extinction Figure 5 3 shows sample 3 keV 25 keV and 1 TeV light curves for one of the parameter combinations Close examination shows that the 25 keV fluxes indeed lead the 3 keV fluxes as expected in the case of dominant synchrotron cooling The 1 TeV flux lags the 25 keV flux owing to the fact that the TeV flux traces the evolution of the electron densities convoluted with the evolution of the synchrotron target photons Our code mimics the effect that it takes approximately a light crossing time until a change in the target photon density has reached all the electrons inside 78 5 4 Generating Data Sets 44 a E S L S P42r a gt of ES i l le 12 le 16 1e4 20 le 24 le 28 v Hz FIGURE 5 2 Representative SED plots of each of the six paramete
43. 1 A measurement of the the jet properties at the jet base would constrain the processes of matter accretion and jet formation and would establish a crucial link between violent processes near a black hole of 107 Mgy 10 Mo pc radius and the kpc scale radio and X ray jets While snapshots of the Spectral Energy Distribution SED constrain the parame ters of emission models there is no single source for which the model parameters have been determined unambiguously In the case of red blazars MeV GeV blazars the intensity of external radiation fields which supply the target photons for in verse Compton processes is not well constrained see e g Tavecchio et al 1998 for a summary of constraints that can be derived from the broadband SED In the case of blue blazars TeV blazars models are simpler as external radiation fields are thought to be weak However a highly uncertain amount of extragalactic extinc tion owing to pair production processes of TeV rays interacting with intergalactic infrared photons Gould and Schr der 1967 Stecker et al 1992 renders the inter pretation ambiguous Additional constraints can be derived from the temporal evolution of the broad band energy spectra that reflect the evolution of the electron energy spectra owing to various physical processes If synchrotron Compton models do indeed apply and if synchrotron cooling dominates the energy losses of electrons high energy electrons
44. 120 Entries 445436 Mean 0 005656 10 Mean 0 003768 RMS 0 005351 H RMS 0 003568 dT Entries 296386 10 0 0 005 0 01 0 015 0 02 0 025 0 03 0 035 0 04 0 045 0 05 0 0 005 0 01 0 015 0 02 0 025 0 03 0 035 0 04 0 045 0 05 sec sec FIGURE B 3 Plots of dt the time between consecutive events for a run with a low and b high rates The bump which always occurs at 0 04 sec corresponding to 25 Hz is more prominent in the high rate data It blends more into the normal exponential decay of dt times for the lower rate data is minimal especially for the normal data rate which is currently relatively low and for the most part masks the dt bump under the normal decay of times While in the end not determined to be a major problem the issue of the dt bump is an example of how daily data quality monitoring noticed characteristics of the data that would not otherwise have been detected This simply emphasizes the importance of this task in bringing a new complicated system such as VERITAS online 111 Appendix C VAC VERITAS Array Control GUI The VAC GUI was designed as a user friendly way to interact with the VERI TAS array control system Its current incarnation can support the full array of four telescopes It allows observers to interact with many subsystems control all aspects of run definition and management and displays status information and plots while the telescopes are taking data The program was written in
45. 3 6 The four telescope array is still being constructed and tested Telescope 1 began 37 3 8 VERITAS Very Energetic Radiation Imaging Telescope Array System FIGURE 3 6 The current configuration of the four telescopes at the base camp of the Whipple Observatory is as shown Distances and locations are not optimal due to construction restrictions around the existing structures 38 3 8 VERITAS Very Energetic Radiation Imaging Telescope Array System operating as a prototype in 2004 with half of its PMTs and one third of its mirrors and became fully operational in February 2005 Holder et al 2006 Telescope 2 saw first light in September 2005 The first two telescopes operated separately for several months The stereo trigger became active and the two telescopes operated together as one starting in March 2006 Construction on the other two telescopes has progressed rather quickly with Telescope 3 coming online in Fall 2006 and Telescope 4 in January 2007 The four telescopes are identical Each consists of a 12 m diameter support struc ture holding a segmented reflector made up of 350 hexagonal mirrors of total area 110m arranged in a Davies Cotton configuration Davies and Cotton 1957 These focus incoming light onto the PMT camera described in Section 3 3 2 Each tele scope also has an electronics shed located right next to it The sheds house the high voltage supplies and controls the digitizing electronics th
46. 87 Abstract This thesis describes two projects the first of which is the analysis of data from the VERITAS Very Energetic Radiation Imaging Telescope Array System experiment VERITAS an array of ground based y ray telescopes in southern Arizona USA has been taking data in hardware stereo mode since March 2006 The April May 2006 dark run provided a large set of data from two telescopes on the known blazar Markarian Mrk 421 An initial analysis of the 14 3 hours of stereo data produced a light curve and confirmed a detection on the 39 sigma level with a y ray rate of 2 91 0 07y min reduced from an inferred value of 8 83 0 21y min before analysis cuts The analysis shows the two telescope array s energy threshold to be 165 GeV before cuts and 220 GeV after cuts with an angular resolution of 0 16 These data were also used to extract an energy spectrum for Mrk 421 This initial analysis allows a test of the performance of the two telescope array and gives an idea of the data that will come from the full system The remaining two VERITAS telescopes will be brought online by January 2007 As a second project computer simulations were used to model Synchrotron Self Compton SSC emission from blazars that will be relevant for future multiwavelength campaigns The Discrete Correlation Function List of Tables DCF was used to calculate the source s magnetic field based on the time lag between emission in different X ray energy band
47. C using the QT development environment and consists of approximately 13 000 lines of code Below is the User s Manual for VAC v3 168 112 C 1 Starting VAC C 1 Starting VAC The graphical user interface for the array control system is located on the arrayct1 computer Through this program you can gain access to various subsystems for de bugging purposes as well as accomplish all tasks needed for a night of observing define and start runs etc This program is a work in progress and will be evolv ing in look and functionality over time This version has recently been expanded to handle the full four telescope VERITAS system The array control and telescope control programs arrayctl and telectl are started automatically at system startup they should not be started manually Nor should any version of these programs be started on any other computer This causes problems with the system s normal operation However multiple versions of VAC may be running at once Also versions of the text based simple ui may be running as well though it is no longer recommended that you even try using this program under normal circumstances Commands from each of these interfaces will be handled in turn by arrayctl and each version of the interface programs will update with the results of these actions This allows for simultaneous debugging from multiple locations Currently VAC requires some subsystems to be started independently Once the
48. CGRO we were left without a reliable method to detect and report GRBs Then in 2004 NASA launched Swift Burrows et al 2003 which contained the Burst Alert Telescope BAT with five times the sensitivity of BATSE The satellite is also made up of telescopes for monitoring these bursts in X rays UV and optical bands This multiwavelength ability helps Swift to detect GRB positions within a few arc seconds 2 2 Instruments to Detect y rays FIGURE 2 2 Set to launch in 2007 GLAST represents the fu ture of space based y ray astron omy GLAST Set to launch in 2007 the Gamma ray Large Area Space Telescope GLAST see Fig 2 2 is the successor of EGRET and will be able to detect sources from 20 MeV 300 GeV Ritz et al 2005 Its sensitivity will be almost 10 times that of EGRET and it will have about twice the field of view Like EGRET GLAST is a pair production telescope The primary ray interacts with the detector and creates an electron positron pair The two charged particles are then tracked through the detector volume The tracks point back towards the incident direction of the primary y ray GLAST will also have a basic ability to detect GRBs The experiment is described in more detail in Section 6 3 Though many technological advancements are being made limits to the physical size of these space borne detectors as well as the sources steep spectra at higher 2 2 Instruments to Detect y rays energies
49. CORBA connection with these systems are better established and the programs are able to handle remote starts or are daemonized and come up when their computers are powered on VAC will be able to take over more aspects of their functionality C 1 1 Normal Operation The following steps are taken to use the VAC to take data on a normal night when the entire telescope array is functioning properly This process involves defining individual runs for the telescope to take and initiating them Data taking is done automatically 1 Start the VAC program on the arrayctl computer gt VAC It has been noticed that the program can hang if there is no connection to the database Be sure to check this if things are not working properly 2 Start up the necessary subsystems using the Start Subsystems option in the Observer menu Check which subsystems have established contact by looking in the status portion of the main window Take the appropriate steps to initialize systems with which there is no contact as outlined in the related wiki pages 3 Initialize the night by choosing Start Night from the Observer menu This takes care of calling the initialization routines for all the connected subsystems 113 C 1 Starting VAC 4 Define a new run by clicking the Define Run button You will be prompted to enter all information pertaining to that run run type source etc When you are done click Define Prepare 5 Start the run by clicking
50. Detect y rays FIGURE 2 3 Located on Mt Hopkins in southern Arizona USA the Whipple 10 m telescope is an example of an Imaging Atmospheric Cerenkov Telescope 11 2 2 Instruments to Detect y rays The current trend in IACTs is to use an array of telescopes all looking at the same source and requiring multiple telescope coincidences for the array to trigger This method first pioneered by HEGRA P hlhofer et al 2003 has several advantages over single telescopes Arrays of IACTs provide a large effective area gt 100 m excellent suppression of cosmic ray initiated air showers and local muons lower energy threshold improved angular resolution and better flux sensitivity as well as better energy resolution compared to their single telescope counterparts Recently most new discoveries have come from the H E S S High Energy Stereo scopic System array in Namibia Africa Aharonian et al 2005d Shown in Fig ure 2 4 H E S S consists of four 12 m telescopes and has been able to detect an astounding number 30 of new sources since coming online in 2003 Other arrays of IACTSs are currently being built around the world VERITAS described in Section 3 3 is nearing completion in southern Arizona USA The MAGIC Collaboration is also building a second telescope at their current site to create the two telescope array MAGIC II Cerenkov Solar Array Telescopes Cerenkov light can also be collected by the large mirror
51. GrISU 59 4 5 Comparison of Experimental and Monte Carlo Data ES o a ISI UE IEEE el T T T T T T Normalized Entries E o to Normalized Entries Laud acra aaa A 07 08 09 1 E 8 10 theta degrees MSCW o o o by of eL o o a o o Normalized Entries e e ek Normalized Entries E eo ier arose prem o eo N Aaa o o E o pio o er BD o a o o eo N o co 09 1 5 5 6 width log size FIGURE 4 5 Several plots showing both Monte Carlo simulated events blue as well as real data events The top two plots show array level parameters 0 and MSCW with the array data plotted in black The bottom two plots show telescope level data width and size with T1 plotted in black and T2 in red modeling the trigger accurately A cut of size gt 1000 would eliminate this difference between the two curves We have also used these Monte Carlo simulations to esti mate the energy thresholds angular core and energy resolutions of the two telescope VERITAS experiment The energy threshold is calculated from the energy distributions of the Monte Carlo simulations Figure 4 6 shows these distributions both before and after the optimized cuts discussed in Section 4 6 Where these histograms peak is commonly referred to as the energy threshold It is measured as 165 GeV before cuts and 220 GeV after cuts 60 4 5 Comparison
52. J A Hofmann W Holleran M Horns D de Jager O C Kh lifi B Komin N Konopelko A Latham I J Le Gallou R Lemi re A Lemoine Goumard M Leroy N Lohse T Martineau Huynh O Marcowith A Mas terson C McComb T J L de Naurois M Nolan S J Noutsos A Orford K J Osborne J L Ouchrif M Panter M Pelletier G Pita S P hlhofer G Punch M Raubenheimer B C Raue M Raux J Rayner S M Redondo I Reimer A Reimer O Ripken J Rob L Rolland L Rowell G Sahakian V Saug L Schlenker S Schlickeiser R Schuster C Schwanke U Siewert M Sol H Steenkamp R Stegmann C Tavernet J P Terrier R Th oret C G Tluczykont M Vasileiadis G Venter C Vincent P V lk H J and Wagner S J 2005a Astronomy and Astrophysics 436 L17 Aharonian F Akhperjanian A G Aye K M Bazer Bachi A R Beilicke M Benbow W Berge D Berghaus P Bernl hr K Boisson C Bolz O Braun I Breitling F Brown A M Bussons Gordo J Chadwick P M Chounet L M Cornils R Costamante L Degrange B Djannati Ata A O C Drury L Dubus G Emmanoulopoulos D Espigat P Feinstein F Fleury P Fontaine 143 Bibliography G Fuchs Y Funk S Gallant Y A Giebels B Gillessen S Glicenstein J F Goret P Hadjichristidis C Hauser M Heinzelmann G Henri G Hermann G
53. Jacholkowska A de Jager O C Kh lifi B Komin N Konopelko A Latham I J Le Gallou R Lemi re A Lemoine Goumard M Leroy N Lohse T Martin J M Martineau Huynh O 144 Bibliography Marcowith A Masterson C McComb T J L de Naurois M Nolan S J Noutsos A Orford K J Osborne J L Ouchrif M Panter M Pelletier G Pita S P hlhofer G Punch M Raubenheimer B C Raue M Raux J Rayner S M Reimer A Reimer O Ripken J Rob L Rolland L Rowell G Sahakian V Saug L Schlenker S Schlickeiser R Schuster C Schwanke U Siewert M Sol H Spangler D Steenkamp R Stegmann C Tavernet J P Terrier R Th oret C G Tluczykont M Vasileiadis G Venter C Vincent P Volk H J and Wagner S J 2005d Astronomy and Astrophysics 441 465 Aharonian F Akhperjanian A G Bazer Bachi A R Beilicke M Benbow W Berge D Bernl hr K Boisson C Bolz O Borrel V Braun I Brown A M B hler R B sching I Carrigan S Chadwick P M Chounet L M Coignet G Cornils R Costamante L Degrange B Dickinson H J Djannati Atai A Drury L O Dubus G Egberts K Emmanoulopoulos D Espigat P Feinstein F Ferrero E Fiasson A Fontaine G Funk S Funk S F fling M Gallant Y A Giebels B Glicenstein J F Goret P Hadjichristidis C Hauser D Hauser
54. L Bloom S D Chen A W Deines Jones P Es posito J A Fichtel C E Friedlander D P Hunter S D McDonald L M Sreekumar P Thompson D J Jones B B Lin Y C Michelson P F Nolan P L Tompkins W F Kanbach G Mayer Hasselwander H A M cke A Pohl M Reimer O Kniffen D A Schneid E J von Montigny C Mukherjee R and Dingus B L 1999 ApJS 123 79 Henri G and Saug L 2006 ApJ 640 185 Hermann G 2006 in Ground based Gamma ray Astronomy Towards the Future Hoffman C M Sinnis C Fleury P and Punch M 1999 Reviews of Modern Physics T1 897 Holder J Atkins R W Badran H M Blaylock G Bradbury S M Buckley J H Byrum K L Carter Lewis D A Celik O Chow Y C K Cogan P Cui W Daniel M K de La Calle Perez I Dowdall C Dowkontt P Duke C Falcone A D Fegan S J Finley J P Fortin P Fortson L F Gibbs K Gillanders G Glidewell O J Grube J Gutierrez K J Gyuk G Hall J Hanna D Hays E Horan D Hughes S B Humensky T B Imran A 149 Bibliography Jung I Kaaret P Kenny G E Kieda D Kildea J Knapp J Krawczynski H Krennrich F Lang M J Lebohec S Linton E Little E K Maier G Manseri H Milovanovic A Moriarty P Mukherjee R Ogden P A Ong R A Petry D Perkins J S Pizlo F Pohl M Quinn J Ragan K Rey
55. Lorentz factors required by this model are more in line with the radio observations 2 4 3 Markarian 421 For this study we looked at the known TeV blazar Markarian Mrk 421 This source is a nearby z 0 031 high energy peaked BL Lac object It was the first extragalactic source detected in the TeV y ray band Punch et al 1992 It is also visible in the Northern Hemisphere in spring when the data were taken the Crab Nebula the standard candle for y ray sources is only visible there in the fall Mrk 421 is a very active source frequently prone to flaring sometimes to bright ness levels exceeding 10 times that of the Crab Nebula Figure 2 9 shows its light curve in X rays over several years Mrk 421 shows loose correlation between the X ray and TeV y ray bands and has been the subject of many previous multiwavelength campaigns e g Blazejowski et al 2005 Takahashi et al 2000 The history and volume of knowledge on Mrk 421 make it an appropriate candidate for testing the new VERITAS system 26 2 4 Blazars Year 96 97 98 99 00 01 02 03 80 60 40 20 Mrk 421 FIGURE 2 9 X ray lightcurve of Mrk 421 showing extreme flux variability flares continuing over several years The y axis shows the 2 12 keV X ray flux in mCrab units Figure from Krawczynski et al 2004 27 Chapter 3 y ray Detection and VERITAS 3 1 y ray Propagation Very High Energy VHE 4 rays 10 10 eV can come from both
56. Mrk 421 as a strong test beam of y rays For this study Monte Carlo simulations were done on vertically incident y rays with energies from 30 GeV to 50 TeV A total of 2 615 000 events with a Crab like differential spectral index of 2 5 were simulated The showers were simulated over an area of radius 350 m While y rays behave differently when the telescope is pointed at lower elevations than at higher ones Kosack et al 2004 the generally low zenith angles of the runs used in this study did not require simulating different zenith angles Simulations were generated with the Grinnell ISU GrISU package that uses the KASCADE air shower simulation code of Kertzman and Sembroski 1994 followed by the simulation of the Cerenkov light emitted by the air shower and the telescope detector s response to that light Figure 4 5 shows several plots of the various Hillas parameters comparing the Monte Carlo simulations with real data Overall it can be seen that the two agree very well showing these simulations accurately model the air showers and the detec tor response For both simulated and experimental data only showers within a radius of 200 m from the telescopes were used Whereas most parameter distributions agree very well over the entire range the size distributions disagree at smaller size values This fact can be linked to both the steep spectrum of Mrk 421 and the difficulty in 1 See http www physics utah edu gammaray
57. Self Compton Simulations We use the time dependent SSC code of Coppi 1992 to create artificial data sets The code assumes a spherical emission region of radius R that approaches the observer with a bulk Lorentz factor I The emission volume is filled with an isotropic 74 5 3 Synchrotron Self Compton Simulations electron population and a tangled magnetic field of mean strength B We create a light curve by changing the rate of freshly accelerated electrons No t We assume an electron acceleration spectrum following dN dy SE No t x gor X exp 7 Ymax 5 1 with a spectral index p 2 and a high energy cutoff Ymax that is constant for each single simulation The code self consistently evolves the coupled partial differential equations describing the energy spectra of the the non thermal electrons and photons taking into account synchrotron emission and self absorption as well as continuous and non continuous energy losses owing to inverse Compton processes in the Thomson and Klein Nishina regimes respectively The photon density evolves as On c 5 2 g nA RIFA a where q de and p de are the rate of photons being produced inside and outside the energy interval e e de while the last term represents photons escaping from the emission region The electron density evolves as One Ne e Ycon el de Pe gt Ot Q Oy ET E 5 Lo esc 5 3 where Q y t is the rate of freshly ac
58. Source Save New info Date Time 2007 01 05 22 14 56 Update pe FIGURE C 8 Layout of the Database subsystem window 137 C 2 Using the VAC New Source Click to add a new source entry to the database The values of the above fields are all used Save New Info If you are altering the description or position of an existing source clicking here will update its information in the database For this to work you must not change the New Source Name from its original value e Date Time Date amp Time Current date and time from the database when the dialog box was opened Update Click to update the date and time from the database at any point e Telescope Info Telescope Select the telescope number for which you would like information The telescope s hostname north and east offsets from the array center altitude and mirror radius are displayed Hostname Hostname of the above chosen telescope Offset North Offset East Offset Altitude Mirror Radius Position etc of the above telescope Save New Info Saves the current telescope number and hostname associating the two with each other in the database e Run Info Run Number Enter the run number for which you would like to access database infor mation Active Run Info Opens a new window displaying all the database information for the cur rent active run If more than one run is active information will be dis played one run at a time Clicking
59. TAS Very Energetic Radiation Imaging Telescope Array System is supported by grants from the U S Department of Energy the U S National Science Foundation the Smithsonian Institution by NSERC in Canada by Science Foundation Ireland and by PPARC in the U K It is being built through a collaboration of nine primary universities and the Smithsonian Astrophysical Observatory as well as several other contributing institutions http veritas sao arizona edu write the entire graphical interface to arrayctl the array control program written by Marty Olevitch arrayctl is responsible for coordinating and overseeing all data taking processes for the entire array The graphical interface called VAC VERITAS Array Control can oversee many aspects of daily observation It is used to manage data taking and also displays feedback and data plots from the various subsystems to ensure system and data integrity VAC has become the most important and most depended upon piece of software for VERITAS telescope operation Observations have to be complemented with theoretical modeling to reveal the physical processes that produce the observed emission I worked with a Synchrotron Self Compton SSC simulation code to explore time lags in the light curves measured in different observational bands Using the Discrete Correlation Function DCF one can determine this lag From that it is a common approach to calculate the magnetic field of the source Various
60. WASHINGTON UNIVERSITY Department of Physics Dissertation Examination Committee Henric Krawczynski Chair James Buckley Ramanath Cowsik Mark Franklin Martin Israel Aaron Stump OBSERVATIONS OF THE ACTIVE GALACIC NUCLEUS MARKARIAN 421 WITH THE FIRST TWO VERITAS CERENKOV TELESCOPES by Scott Brandon Hughes A dissertation presented to the Graduate School of Arts and Sciences of Washington University in partial fulfillment of the requirements for the degree of Doctor of Philosophy January 2007 Saint Louis Missouri Acknowledgements I would like to thank my advisor Dr Henric Krawczynski for all his help through out my graduate school career Dr Jim Buckley and Dr Martin Israel have also been integral in my education both in serving on my committee as well as being around to provide fresh insight into my work Thanks to the rest of my thesis defense committee as well for their comments and opinions I also thank Marty Olevitch for provid ing both expert computer programming knowledge and sarcastic spiteful retorts I thoroughly appreciate the help from Trevor Weekes and the entire VERITAS collab oration for not only creating a project that I have contributed so much to but also for their knowledge and assistance with problems too large for me to tackle alone I gratefully acknowledge the McDonnell Center for Space Sciences for a fellowship that allowed me to jump start my research immediately upon entering grad school A
61. acceptance divided by the square root of the background acceptance for a particular cut 4 7 It scales with the statistical significance of weak signals obtained with a certain cut For the cut optimization all runs are treated as tracking runs As this is the first data taken with the new stereo VERITAS system many aspects pointing point spread function of the optics trigger nonuniformity and biases etc are not well understood Despite this fact the two telescope data on Mrk 421 are very good 4 6 1 Cutting on 0 The first cut to be optimized was 0 the square of the angular distance of the reconstructed arrival directions of the primary y rays from the source direction This is the most important cut for stereo data and is very powerful at removing background events from the source signal Initially a histogram is made of the 0 values for all events passing a very weak cut of MSCW lt 1 0 This removes the most obvious background events to make the cut optimization cleaner The resulting histogram shows a prominent peak close to 0 0 with a tail that decreases linearly as a function of 0 Since there should be no excess at values 0 gt 0 4 deg one can assume all contributions there are from background events We then fit a line to the histogram for 0 4 deg lt 6 lt 1 0 deg 65 4 6 First Stereo Results from VERITAS 2 F 1200F ui 1400 ui 1000 1200 800
62. account the shower s location relative to each of the telescopes weighting the individual width distributions by the appropriate factor to calculate this normalized parameter Monte Carlo simulations are also useful to help reconstruct the energy of the y ray shower We can create a look up table of energies using the median and 9096 width values of the logarithm of the size parameters as a function of the primary shower s energy E To reconstruct the energy of a shower we simply invert the table The shower s reconstructed energy is determined by averaging the energies obtained from each telescope with a trigger 4 1 3 Analysis Tools As VERITAS is still fairly new there is yet no fully established method for data analysis The official analysis package VEGAS VERITAS Experiment Gamma ray Analysis Suite is still in development and far from fully optimized or simple to use Using VEGAS will soon become standard but for this work we have relied on other packages that have been more thoroughly tested and are known to produce reliable results for y ray data To analyze the data we have used the University of Leeds eventdisplay and mscw_energy packages These packages handle the main aspects of data reduction and analysis for a stereo system The programs output files can be imported into 50 4 2 Data Cuts and Significances ROOT for further analysis and plotting The eventdisplay package also contains ex tensive display tools to vie
63. adun A and Schr der M 2004 ApJ 601 151 Krawczynski H Sambruna R Kohnle A Coppi P S Aharonian F Akhper janian A Barrio J Bernl hr K B rst H Bojahr H Bolz O Contreras J Cortina J Denninghoff S Fonseca V Gonzalez J G tting N Heinzel mann G Hermann G Heusler A Hofmann W Horns D Ibarra A Jung I Kankanyan R Kestel M Kettler J Konopelko A Kornmeyer H Kranich D Lampeitl H Lorenz E Lucarelli F Magnussen N Mang O Meyer H Mirzoyan R Moralejo A Padilla L Panter M Plaga R Plyasheshnikov A P hlhofer G Rauterberg G R hring A Rhode W Rowell G Sahakian V Samorski M Schilling M Schr der F Siems M Stamm W Tluczykont M V lk H J Wiedner C A and Wittek W 2001 ApJ 559 187 Krennrich F Bond I H Bradbury S M Buckley J H Carter Lewis D A Cui W de la Calle Perez I Fegan D J Fegan S J Finley J P Gaidos J A Gibbs K Gillanders G H Hall T A Hillas A M Holder J Horan D Jordan M Kertzman M Kieda D Kildea J Knapp J Kosack K Lang M J LeBohec S Moriarty P M ller D Ong R A Pallassini R Petry D Quinn J Reay N W Reynolds P T Rose H J Sembroski G H Sidwell R Stanton N Swordy S P Vassiliev V V Wakely S P and Weekes T C 2002 ApJL 575 L9 Levine A M Bradt
64. akely S P Weekes T C Zweerink J VERITAS Collaboration Aller M Aller H Boltwood P Jung I Kranich D Nilsson K Pasanen M Sadun A and Sillanpaa A 2006 ApJ 641 740 Reynolds P T Akerlof C W Cawley M F Chantell M Fegan D J Hillas A M Lamb R C Lang M J Lawrence M A Lewis D A Macomb D Meyer D I Mohanty G O Flaherty K S Punch M Schubnell M S Vacanti G Weekes T C and Whitaker T 1993 ApJ 404 206 Ritz S Grindlay J Meegan C Michelson P F and GLAST Mission Team 2005 in Bulletin of the American Astronomical Society pp 1198 Rothschild R E Blanco P R Gruber D E Heindl W A MacDonald D R Marsden D C Pelling M R Wayne L R and Hink P L 1998 ApJ 496 538 Rybicki G B and Lightman A P 1979 Radiative Processes in Astrophysics John Wiley amp Sons New York Sambruna R M Aharonian F A Krawczynski H Akhperjanian A G Barrio J A Bernl hr K Bojahr H Calle I Contreras J L Cortina J Denninghoff S Fonseca V Gonzalez J C G tting N Heinzelmann G Hemberger M Hermann G Heusler A Hofmann W Horns D Ibarra A Kankanyan R Kestel M Kettler J K hler C Kohnle A Konopelko A Kornmeyer H Kranich D Lampeitl H Lindner A Lorenz E Magnussen N Mang O Meyer H Mirzoyan R Moralejo A Padilla L Panter M P
65. and oz being the standard deviation of the 3 keV and 25 keV fluxes respec tively If experimental measurement errors are non negligible Equation 5 5 has to be modified as described by Edelson and Krolik 1988 In the sample DCF shown below we plot error bars which represent the error on the mean DCF value 8l 5 5 Analysis Procedure 1 2 MEE s uu Ix DCF r SH 5 7 For each simulated data set we determine the time offset ep at which the DCF reaches its maximum value For each of the six parameter combinations we use all of the 20 simulated data sets to compute the arithmetic mean of the observed time lags as well as the RMS of the distribution Further below we will compare the DCF time offsets with the differences in syn chrotron cooling times Considering the effect of synchrotron cooling alone the latter are calculated in the following way Assuming a delta functional approach to the synchrotron emissivity electrons of Lorentz factor y emit synchrotron photons of energy 3 gt E y e Bo sina 5 8 with sino V3 for an isotropic pitch angle distribution Photon energies F are Doppler boosted to energies E 6 E 5 9 in the observer frame Based on Equations 5 8 and 5 9 the Lorentz factors y and ya responsible for the radiation observed at energies Ej and E can be determined In the jet reference frame the synchrotron cooling time of electrons of Lorentz factor y is given by Rybicki a
66. and time lag behavior A more realistic modeling of electron acceleration by the second order Fermi process 91 5 7 Comparing of DCF Time Lags to Expected Results would even result in a lag of opposite sign i e the soft X rays leading the hard X rays Kirk and Mastichiadis 1999 Second even phases of falling fluxes do not exhibit a consistent time lag behavior At the beginning of the decaying phase of a flare softening of the X ray energy spectrum produces a time lag behavior However once the decaying phase is long enough that the peak of the synchrotron SED shifts to energies below those sampled by the low energy X ray observations the high and low energy fluxes decrease with the same decay constant As the DCF is calculated from data of the full light curves the net effect is a time lag Tpcr shorter than At Application of the method to real astrophysical data may even be more problem atic as current models may underestimate the complexity of flares Each flare might be produced by a different emission region with a different magnetic field The emis sion plasma may expand or compress during individual flares which would result in a change of the magnetic field and in adiabatic cooling or heating of the non thermal particles Coppi and Aharonian 1999 High energy particles may escape diffusively from the emission region with an energy dependent electron escape time The angle between the bulk motion of the emit
67. ans two runs Rate spikes can overwhelm autoscaling making the plot useless until the spike is cleared Log10 Rate Axis Clicking this check box will change the scale on the y axis to from logarithmic This makes the L3 rate easier to see when the L2 rate is very high Rescale Rates This button temporarily rescales the y axis based on the most recent values added It corrects the autoscaling that includes the rate spike However the 124 C 2 Using the VAC plot is autoscaled again on its next update Soon there should be a more long lasting way to rescale the axis e Clear Plot This button zeros out the entire plot Though annoying it will at least remove rate spikes and make the plot easier to read C 2 2 Observer Menu e Start Subsystems Opens a separate window that can run various start up scripts to handle many of the subsystems You can choose a specific script from the pop up menu or click Skip to skip over a script you do not need to do Clicking Run Script Di executes the current script Items in the list such as Move to Desktop DACQ T1 require the observer to drag the scripts window to the specified desktop before continuing running the scripts The button changes to Changed Desktop in this case so as not to confuse the observer into thinking this action will be completed for them The scripts list should be executed completely in order The current tasks are as follows T2 power on fadc crat
68. are called y rays Photons in this range have the shortest wavelengths and energies from around 500 keV to through TeV and even higher The energy range covered by y rays is more than that of the rest of the electromagnetic spectrum combined Weekes 2003 In order to talk about y rays it is appropriate to divide the energy range into smaller sections of similar behavior and detection techniques Table 2 1 lists the common nomenclature and their corresponding energy ranges In order to produce y rays charged particles electrons positrons protons or other nuclei must be accelerated to extremely high energies The particles then emit radiation which travels through space as y rays There are several relevant emission mechanisms Synchrotron radiation is emitted when a relativistic charged particle 2 1 Introduction to y ray Astrophysics TABLE 2 1 The common divisions within the span of y ray energies including common descriptive names Common Name Energy Range Low LE 500 keV 10 MeV Medium ME 10 MeV 30 MeV High HE 30 MeV 30 GeV Very High VHE 30 GeV 30 TeV Ultrahigh UHE 30 TeV 30 PeV Extremely High EHE 30 PeV and up spirals around a magnetic field line To generate TeV photons by this mechanism either rather strong magnetic fields gt 1 G or extremely high energy electrons or protons are required Another process is Bremsstrahlung which occurs when an electron is decelerate
69. at convert the camera signal so it can be processed and the Level 2 trigger system Figure 3 7 shows one of the VERITAS telescopes and its electronics shed These four telescopes send output to a central control hub which combines the information from all telescopes into the final data stream This is also the location from which nightly observations are commanded 3 3 2 Camera Mounted on each telescope is a camera consisting of 499 individual photo mul tiplier tubes PMTs Each acts as a single pixel to image the air showers These 39 3 9 VERITAS Very Energetic Radiation Imaging Telescope Array System FIGURE 3 7 Pictured here is one of the four VERITAS telescopes in southern Ari zona The support arms extending off the 12 m optical structure hold the PMT camera When not in use the camera rests at an access platform directly above the electronics shed 40 3 9 VERITAS Very Energetic Radiation Imaging Telescope Array System FIGURE 3 8 Each VERITAS camera consists of 499 PMT pix els that cover a 3 5 diameter field of view PMTs have their peak sensitivity in the blue UV to maximize their sensitivity to Cerenkov light The entire camera housing is 1 8 m square large enough to allow for future expansion The camera has a 3 5 diameter field of view Figure 3 8 illustrates the PMT camera Light cones are used to maximize the photon collection efficiency by focusing light onto the PMTs that would otherwise fa
70. ata Quality Monitoring B 1 Motivation and Procedure In order to ensure the VERITAS telescopes are performing as expected during this intense and crucial period where many systems are unstable and constantly evolv ing some sort of sanity check must be instituted to make sure weeks of data are not completely lost because no one caught the fact that something wasn t working properly Starting January 2006 we began to look at each night s data for any inconsis tencies or blatant instances of data errors Not knowing in advance what to really look for the tools and plots used evolved over the second half of the season When problems arose we found ways to better detect issues automatically However some 105 B 2 Analysis and Results visual inspection of runs was still necessary to confidently determine data reliability Each morning a few runs were chosen to be analyzed These were selected to include the various observing modes of the previous night as well as those runs with known issues such as rate spikes to get the best overview of the general data quality The runs were first analyzed using the eventdisplay package Then ROOT was used to run a custom script that generated a set of plots as well as printed out errors and other useful information about the runs to help determine if there were major data issues B 2 Analysis and Results Initially the night s laser run is looked at by hand event by event using the
71. beck T H hne D Hose J Hsu C C Isar P G Jacon P Kalekin O Kosyra R Kranich D Laatiaoui M Laille A Lenisa T Liebing P Lindfors E Lombardi S Longo F L pez J L pez M Lorenz E Lu carelli F Majumdar P Maneva G Mannheim K Mansutti O Mariotti M Mart nez M Mase K Mazin D Merck C Meucci M Meyer M Miranda J M Mirzoyan R Mizobuchi S Moralejo A Nilsson K O a Wilhelmi E Ordu a R Otte N Oya I Paneque D Paoletti R Paredes J M Pasanen M Pascoli D Pauss F Pavel N Pegna R Persic M Peruzzo L Picci oli A Poller M Pooley G Prandini E Raymers A Rhode W Rib M Rico J Riegel B Rissi M Robert A Romero G E R gamer S Saggion A S nchez A Sartori P Scalzotto V Scapin V Schmitt R Schweizer T Shayduk M Shinozaki K Shore S N Sidro N Sillanp A Sobczynska D Stamerra A Stark L S Takalo L Temnikov P Tescaro D Teshima M Tonello N Torres A Torres D F Turini N Vankov H Vitale V Wagner R M Wibig T Wittek W Zanin R and Zapatero J 2006 Science 312 1771 Baixeras C and et al 2005 in International Cosmic Ray Conference pp 227 Barrau A Bazer Bachi R Beyer E Cabot H Cerutti M Chounet L M De biais G Degrange B Delchini H Denance J P Descotes G Dezalay J P
72. benefits of both Tracking and ON OFF pairs into a single method which minimizes both the observation time and the systematic uncertainties ON OFF Pairs Frequently used on the Whipple 10 m telescope ON OFF pairs are good for getting high quality data but are only accurate and effective on cloudless nights First a 28 minute run is taken with the source at the center of the camera Then a second OFF run is taken offset 30 minutes in Right Ascension and taken 30 minutes after the ON run This second run sweeps across the exact same portion of the sky as the ON run for more accurate background subtraction that is not affected by the systematics of the telescope itself Many data from this dark run were taken as ON OFF pairs but have been processed as tracking to simplify the analysis procedure Tracking Runs Tracking runs require less time and can be useful in less perfect weather These runs simply keep the source at the center of the camera at all times For background subtraction an OFF region is chosen farther from the center of the camera so it is not be affected by the source itself Figure 4 4a depicts visually the ON and OFF regions of the field of view for a Tracking run The ON region is a circle with given 54 4 4 Mrk 421 Data from April May 2006 Tracking Wobble 0 3 4 camera center Area ON Area OFF FIGURE 4 4 The different observation modes have different ways in which back gro
73. bserva tion campaign of 20 days the precision to which the maximum of the DCF can be determined will exclusively be limited by the accuracy of the experimental flux mea surements and the observational coverage Figure 5 6 compares the magnetic field values inferred from Tpcr and Equation 5 15 with those used to simulate the data sets As can be seen from the figure and the values listed in Table 5 3 the magnetic field values are overestimated by factors of between 1 5 and 6 for the strongest and weakest simulated magnetic field values respectively We have investigated if the inverse Compton cooling of the electrons produces the short DCF time offsets For this purpose we ran the SSC code again this time suppressing the inverse Compton processes The 7pcr values calculated for these simulations are also given in Table 5 2 For all sets of model parameters but those with the smallest B field the mean Tpcr values computed with and without inverse 86 5 6 Measuring Time Lags with the Discrete Correlation Function DCF 100 I MEE E A T TOT OTTTTT T LEE 10 Po ES TbCF h UA I 10 100 Ar h sync FIGURE 5 5 The computed time lag from the discrete correlation function Tpcr plotted against the expected time lag from Equation 5 13 Atsynch See Table 5 2 for specific values Error bars are RMS values from averaging over 20 runs The dotted line represents the two values being equal TABLE 5 3 Actua
74. celerated electrons Ycont is the decrease in an electron s Lorentz factor and q dy and p dy are the rates of particles being produced or scattered inside and outside the Lorentz factor interval y y dy Krawczynski et al 2002 The SSC code assumes that the synchrotron emission provides the dominant target photon field 75 5 4 Generating Data Sets Mimicking the duration of a typical multiwavelength observation campaign we generated artificial data sets of 30 days duration observer frame The first 10 days were discarded from subsequent analysis so that the target photon fields and low energy electron distributions could reach a steady level We created each artificial data set by a series of N triangle bursts of accelerated particles N No t M Aih t T 5 4 i 1 Here the function h t T represents triangle pulses of constant width centered on the times 7 randomly chosen between 0 and the duration of the flaring period the heights A of the triangle pulses were chosen to vary by a factor of 4 with a preference toward smaller values Choosing triangle flares of tgare 10 hrs duration and N 30 this method produced light curves that closely resemble observed ones A sample of the output can be seen in Figure 5 1 5 4 Generating Data Sets We use six different sets of model parameters with magnetic field values ranging from 0 005 G to 0 2 G In the following we refer to the six parameter configurations as
75. ch Monte Carlo event as R Erec dd Errue energy E true 4 6 The value below which 63 of events occur is the energy resolution For all events Note For this preliminary analysis we have not attempted to optimize the Energy Estimator function 63 4 6 First Stereo Results from VERITAS 1 5 mb o LEVE AA O A A A AAA Log Reconstructed Energy 2 L L i L L L L L L L L L L i L L L L L L L L L L L L L L L L L L 0 2 1 5 1 0 5 0 0 5 1 1 5 2 Log True Energy FIGURE 4 9 The logs of the true and reconstructed energies for the Monte Carlo data are plotted against each other The red line shows the one to one relationship of the two values being equal the energy resolution is 34 Figure 4 9 shows the log of the true and reconstructed energies plotted against each other also demonstrating the accuracy of the energy reconstruction 4 6 First Stereo Results from VERITAS All the runs listed in Table 4 1 were analyzed using the eventdisplay package with the standard default values These output files were then processed with mscw_energy also with the default values The runs were then evaluated by a custom ROOT script to determine y ray rates and significances One purpose of the custom ROOT script was to optimize cuts on 0 and MSCW 64 4 6 First Stereo Results from VERITAS for this data set The optimization aims at maximizing the Q factor defined as the y ray
76. configuration A through F The magnetic field values have been chosen such that the electrons emitting synchrotron radiation at 10 keV observer frame have synchrotron cooling times ranging from the flare duration tgare to much longer than the flaring time The six parameter combinations are listed in Table 5 1 They are similar to those inferred from detailed modeling of observation campaigns of the 76 5 4 Generating Data Sets ON A UA Go m Rate of injected electrons arbitrary units NO A pl d I 30 cz time days FIGURE 5 1 The figure shows a sample of an artificially generated electron acceler ation history No t The curve was generated by superimposing a random succession of 30 triangle pulses each 10 hrs wide The first ten days are a transient period used to get the system going data is only taken from the last 20 simulated days The electron acceleration history is shown in the stationary observer frame Data shown are from input file 14 TT 5 4 Generating Data Sets TABLE 5 1 Magnetic field values and calculated cooling times for each set as well as other important initial conditions Also the relation of each cooling time to the flaring time 10 h Set B G R 109cm A R c h ua teooi 10 keV h _ teooi titare A 0 20179 1 30 2f 0 04 10 9 1 B 0 07063 2 17 4 4 0 7 52 7 5 C 0 04708 2 61 9 3 0 8 96 8 10 D 0 01412 4 34 8 9 0 7 589 6
77. ctic Nucleus 2 8 Sample Spectral Energy Distribution 2 9 X ray Light Curve of Mrk 421 with Flares 3 1 Electromagnetic Cascading Air Shower les 3 2 erenkov Radiation s sep s mehr iR mir DM E DOES Ue Deb 3 3 Charged Particle Polarizing the Surrounding Medium 3 4 Cosmic Ray Air Showers i s ue eem xe dex Xem dE TERN DER ch 3 5 Cosmic Ray vs y ray Air Showers 3 6 Layout of the Four VERITAS Telescopes 3 7 A 12 m VERITAS Telescope 5 2 oo a a Y uox AA TAI EN NN 4 1 Hillas AA O dde IRIS 4 2 Images of Sky Showers da a wey AA e 4 3 Stereo Shower Reconstruction ZA OOBBSERGEIGUS dek fi AA A IS e CUR A A 4 5 Comparing Monte Carlo Simulations to Data 4 6 Energy Threshold of Monte Carlo Simulations 4 7 Angular Resolution of Monte Carlo Simulations 4 8 4 9 Core Resolution of Monte Carlo Simulations True vs Reconstructed Energy for the Monte Carlo Simulations 4 10 Plots of 0 4 11 Q factor for 0 4 12 Plot of MSCW 4 13 Q factor for MSCW 4 14 Mrk 421 Light Curve 4 15 Energy Spectrum of Mrk 421 vil List of Figures 5 1 Sample Input of Simulated Flares sls 77 5 2 SEDs of Simulated Data Sets coser a a 79 5 3 Light Curves of Simulated Data Sets s 80 5 4 DCF from Simulated Data WES Go ES 85 5 5 Actual vs Calculated Lag
78. d Information should appear as soon as a run is prepared Run Info for Telescope More than one run can currently be active in the system yet VAC can only display info about one of them at a time This pop up list allows the user to select which run has its information displayed Since each telescope can only have one active run at a time telescope number is used as a screening process for the active runs In this way also initial testing of operating two telescopes separately will not be overly confusing to the users Selecting a telescope number from the list means that only runs in which that telescope is participating will be displayed in the Current Active Run panel If ALL is selected info will be displayed for the first run in the arrayctl internal run list only and will change to the next one when that run has finished Open QL Displays Runs a script to open various QuickLook tools for the current run This includes ql_display and ql monitor It is not yet available L2 L3 Rate Plot This plot is a composite of individual telescope L2 rates and the L3 rate All rate information comes from L3 Each curve appears as a separate color Modifications will need to be made to handle multiple runs active at the same time Current issues are as follows 1 2 The x axis is only approximate time since the data was sampled The rate plot only updates when a run is active the x axis times become very inaccurate when the plot sp
79. d I an energy histogram is filled weighing the Monte Carlo events so as to mimic the model of Equation 4 4 Both parameters are then varied in the user defined parameter space to minimize the X difference between the data and Monte Carlo histograms The values of Ny and I with the best fit along with associated errors give the source s energy spectrum 4 4 Mrk 421 Data from April May 2006 Data taken during the dark run from April May 2006 were the first set of true stereo data from the VERITAS telescopes The data on the known blazar Mrk 421 contain a large number of y ray events and can be used for calibration purposes 4 4 1 Observation Modes There are a variety of modes in which VERITAS data can be taken ON OFF pairs were the standard method in the past however they require an extremely clear sky to be useful because passing clouds will affect the source and background regions differently This method can seem wasteful in that half of the observation time is used to examine blank fields in the sky Tracking runs are useful because they require less time than ON OFF pairs and can still be effective in less than perfect weather 93 4 4 Mrk 421 Data from April May 2006 due to clouds altering both ON and OFF regions in the same way The method is somewhat problematic as the sensitivity is not uniform over the entire field of view of the telescopes We are also trying out an attractive method Wobble mode which combines the
80. d by the electromagnetic field of a charged particle or particles The braking radiation produced is a result of the energy loss by the electron and can be quite substantial Perhaps the most important process is inverse Compton scattering When a low energy photon collides with an energetic electron it can be scattered to much higher energies All these processes are described in detail in Section 2 4 2 rays can be detected through their interaction with matter Different energy ranges lend themselves to different dominant interaction processes For the lowest energy y rays photoelectric absorption is the dominant process A y ray can eject an electron from a tightly bound atom which also emits an X ray as the resulting hole in the atom is filled by an electron from a higher orbit Both the ejected electron and X ray can be used to detect the original y ray Mid range y rays prefer Compton 2 2 Instruments to Detect y rays scattering Here a y ray collides with a loosely bound electron giving up some of its energy Multiple collisions can occur for the same incident y ray For higher energy photons the dominant process is pair production If the y ray photon has energy E gt 2moc 1 022 MeV 2 1 where mo is the rest mass of the electron it can convert to an electron positron pair in the presence of an atomic nucleus required for momentum conservation 2 2 Instruments to Detect y rays Being uncharged and therefore u
81. d of light in the medium there is a build up of polarized charge just behind the moving particle Figure adapted from Jelley 1958 the medium so no net polarization is observed However if the charged particle is traveling faster than the speed of light in the medium the polarization of nuclei is not symmetric see Fig 3 3b The moving particle s charge is not propagated to the atoms of the medium until after the particle has passed creating a build up of positive charge just behind the moving electron The transmittance of the electron s charge is sent out radially and becomes cohesive along a wavefront at the angle 0 from the direction it is traveling In air the Cerenkov light reaching the ground has its peak emission in the UV blue 32 3 2 Detection Using an Imaging Atmospheric Cerenkov Telescope IACT portion of the spectrum Telescopes that observe Cerenkov light are designed to have peak efficiency in this range Cosmic Ray Induced Showers Cosmic rays are also constantly bombarding the Earth and they too produce cascading showers in the atmosphere This hadronic shower is much different than showers produced by a y ray and includes both pions and muons Hadronic showers are spread out over much larger areas than electromagnetic showers owing to the momentum of the nucleons and quarks that give rise to large transverse velocities of the secondaries of hardronic interactions Figure 3 4 shows schematically how a
82. dditionally the staff at Washington University has been most helpful in this impor tant process Sarah Jordan Julia Hamilton Allison Verbeck Jan Foster Christine Monteith Paul Dowkontt Ira Jung and Richard Bose among others have all been there at one time or another to help with administrative or technical problems that crept up along the way 11 I cannot even begin to thank my fellow graduate students past and present for their assistance both intellectual and not so intellectual In particular to Jeremy Perkins Karl Kosack Paul Rebillot Lauren Scott Randy Wolfmeyer Mairin Hynes Allyson Gibson Brian Rauch Kuen Vicky Lee Trey Garson Kelly Lave Frank Gyngard and honorary grad student Ellen Wurm as well as many others I am not able to list A special thanks goes out to Jason Rosch who continually helped me to get out and temporarily forget about my problems when the going got tough Also to Erica Barnhill who has been there since college whenever I needed someone to talk to and is always able to cheer me up In addition to Morgann Reilly who unintentionally showed me that things could be much worse and who was always up for a drink and a movie to distract us both from our troubles at work Several others have also helped provide me with joy and amusement outside of school during my stay in St Louis Most importantly the ANTM crew Danette Wilson Rose Martelli Amber Specter and Allyson Gibson who does
83. e field of view VERITAS is starting to use them as well This method combines the effective ness of ON OFF pairs with the convenience and speed of a Tracking run Wobbling involves offsetting the source by a given amount in the field of view The offsets are in varying directions usually opposite in order to cancel out systematic inhomo geneities in the camera itself Figure 4 4b depicts visually the ON and OFF regions for a wobble run Both regions are the same size and shape offset on opposite sides of the camera with the ON region centered on the source There is space between the two regions again to prevent the source from influencing the background region This method is not as valuable for single telescopes because their detection rate is more strongly affected by stars in the field of view and faulty pixels 4 4 2 Final Data Set Not every data run that was taken is useable Various problems exist with the data as this new experiment is slowly brought completely online Issues range from imperfect weather to rate spikes due to nearby car headlights to issues with the 56 4 4 Mrk 421 Data from April May 2006 hardware and software Ground based 4 ray telescopes usually operate at around a 1096 duty cycle The solar and lunar cycles especially towards summer as the number of daylight hours increases as well as the weather and Arizona s monsoons severely limit the amount of time it is possible to operate the telescopes
84. e in each Good Time Interval GTI ranged from 160 s to 4 43 ks for 1ES 1959 650 and from 168 s to 9 01 ks for Mrk 421 Spectra and light curves were extracted with FTOOLS For Mrk 421 spectral analysis was restricted to the 4 15 keV energy range Analysis of earlier RXTE data showed corrupted behavior exceptionally high or low count rates of individual bins not compatible with the energy resolution of the instrument below 4 keV Above 15 keV the data of most pointings were dominated by background Background models were generated with the tool pcabackest based on the RXTE Guest Observatory Facility GOF calibration files for a bright source with more than 40 counts s PCU Comparison of the background models and the data at energies above 30 keV showed that the model underestimated the background by 10 We corrected for this shortcoming by scaling the background model with a correction factor of 1 1 Response matricies for the PCA data were created with the script pcarsp The spectral analysis was performed with the Sherpa package A Galactic neu was used for all observations of tral hydrogen column density of 1 027 x 10 cm ES 1959 650 while a value of 1 31 x 10 cm was used for Mrk 421 Since the analysis is restricted to the energy region above 3 keV the hydrogen column density has only a very minor influence on the estimated model parameters Single power 101 A 3 Orphan Flares law models result
85. e run or after its completion This is done through the main window Author author of the above comment Each comment has an associated author and must be added with the addition of each comment Observers the observers for the new run These must be entered as a comma separated list i e SBH MAO KPK etc for the database to handle them properly You are required to enter at least one observer before the run can be defined A pop up window will remind you of this fact if you try to declare the run without doing so Duration length of the desired run Separate minute and second fields allow for easier entry Wobble Defines whether this is a wobble run or not Sets appropriate offsetRA and offsetDEC in the run information based on the source s true RA and Dec as well as offsetAngle and offsetDistance Wobbles can be done in the default NSEW directions or any arbitrary direction by entering any angle next to the radio button The offset degrees should always be POSITIVE If the previously defined run was a wobble run the next time the Define Run panel is opened the wobble offset will automatically be set to the same angle but opposite direction Telescope Configuration Selects which telescopes will be a part of the new run This is a graphical representation of the Config Mask Type The number in the upper right is the mask value to be used and represents the scheme currently displayed graphically Checked boxes mean the t
86. eV produce synchrotron radiation which becomes a large portion of the y rays one observes In this model the X rays also produced through synchrotron radiation come from a completely different population of electrons those generated by the blazar s magnetic fields within the jets as with the other two models Particle Acceleration To produce any of the above mentioned emission the electrons positrons may be accelerated by shocks in the jet see review by Kirk and Duffy 1999 For example some Sokolov and Marscher 2005 Mimica et al 2004 suggest electrons are accel erated as they pass back and forth across the interface where two relativistic shock fronts collide Alternatively the particles may be accelerated by the central engine itself Levinson 2005 Katz 2006 Krawczynski 2006 Many authors Piner and Edwards 2005 Henri and Saug 2006 Tavecchio 2005 25 2 4 Blazars have investigated the T problem wherein previous simulations of SSC emissions require a bulk Lorentz factor 25 an order of magnitude higher than what is ob served through VLBA Very Long Baseline Array observations Ghisellini et al 2005 get around this need for high Lorentz factors by assuming there is a layer and spine structure to the AGN jets Here two concentric volumes move at dif ferent velocities and therefore boost the emission seen by a factor of I I spine Diayor 1 Bspineblayer The lower bulk
87. ead out then continues digitizing data until the next event One important feature of the FADC boards is their ability to switch between a high and low gain signal path If a pulse comes through the FADC that exceeds the dynamic range of the high gain path a high low gain switch is flipped and the signal from the low gain path is digitized This allows the system to easily handle a wide variety of pulse heights with an 8 bit digitizer 43 Chapter 4 Mrk 421 Data Analysis 4 1 Event Reconstruction The images captured by the VERITAS cameras result from thousands of brems strahlung and pair production interactions in the atmosphere To gain information about the incident y ray one must reconstruct an event and determine a number of parameters characterizing the air shower The important result of the event re construction is the location and orientation of the air shower axis which points back to the arrival direction of the primary y ray Furthermore the event reconstruction gives information about the nature of the primary particle photon or hadron Last but not least we want to know the energy of the primary y ray In the following we describe the methods used in this thesis 44 4 1 Event Reconstruction 4 1 1 Hillas Parameterization Cerenkov light reaching the telescopes cameras is captured by the individual pixel PMTs The resulting image can then be parameterized Calculations of the first second and third m
88. eased in recent years as well due mostly to the H E S S telescopes in the Southern Hemisphere A similar increase in new sources in the Northern Hemisphere should happen shortly when VERITAS comes fully online Figure 2 5 shows a map of the sky in galactic coordinates with all known TeV sources labeled These sources are also listed in Table 2 2 14 2 3 TeV y ray Sources The High Energy Gamma Ray Sky 2006 Galactic coordinates lt E e A SS a5 M po 8 5 Msi A Mg Us ei PG M 185 In soy Zen IES 01 295 TES 19 8 65 wt Te ue 9 op is C y Ms Ess A esa ies Sosa iis A 15 59 09 ea ax A AGN 4 Plerion Shell type SNR Binary system E Other or unid Background colours indicating northern southern sky FIGURE 2 5 Map of the sky in galactic coordinates showing TeV y ray sources Figure from Hermann 2006 15 2 3 TeV y ray Sources TABLE 2 2 Known TeV sources as of July 2006 Sources are divided by class Table from Cui 2006 Name RA 2000 Dec 2000 Notes Blazar 1ES 1101 232 11 03 37 57 23 29 30 2 Mrk 421 11 04 27 31 38 12 31 8 Mrk 180 11 36 26 41 70 09 27 3 1ES 1218 304 12 21 21 94 30 10 37 1 H 1426 428 14 28 32 6 42 40 29 PG 1553 113 15 55 43 04 11 11 24 4 Mrk 501 16 53 52 22 39 45 36 6 1ES 1959 650 19 59 59 85 65 08 54 7 PKS 2005 489 20 09 25 39 48 49 53 7 PKS 2155 304 21 58 52 07 30 13 32 1
89. ed and the particle radiates away energy This results in a cone of Cerenkov light which has a fixed angle 05 with respect to the direction of particle motion This angle is 0c cos cos EN f 3 3 where cm is the speed of light in the medium and n is the index of refraction of that found by medium As usual 8 v c Figure 3 2 depicts this scenario graphically To visualize how this radiation manifests itself consider first a charged particle like an electron moving slowly through a medium Asit moves the electron polarizes the nearby atoms pushing the negative charges away from it see Fig 3 3a The atoms relax back to their normal configuration after the electron has passed Because the speed of the electron is relatively slow this produces a symmetric disturbance in 3l 3 1 y ray Propagation A A 000 D O_O QOO D O_O d Dee CS O OS O OO Ween o9 9 0 oo O ccoQao do O ORO Oi o 00 00g Q9 9 06 2 ADE aO c oF D OO Gas SE OU D eO d qe OO o OUS QS Dodo O o9 doo o E Ob 9 Co 25 oS E So vg epe pr Sus O O O S OOO 000 4 DNS EE oy O DD O O O O O ene Crs OO On 6 O OP Oct O OP OH co O69 H 0500 O gO eeu de OOOO Ora EQ Qu s cu QUO 090 9000 Ze ROO ORO E QOO oo ole 00 2 a VO OO Op DO 257 9 o 0080 A aor O Y Y B B FIGURE 3 3 a When a charged particle travels slowly through a medium it polarizes the surrounding atoms b When the particle moves faster than the spee
90. ed in statistically acceptable fits for all data sets Data for Mrk 421 were complemented by data from the ASM Fluxes were derived by averaging the summed band intensities acquired during one day A 3 Orphan Flares Overall observations during both multiwavelenth campaigns showed an extremely high confidence level for X ray and TeV y ray flux correlation 97 for Mrk 421 However when they were not correlated the difference was extreme widely different TeV y ray fluxes for a single X ray flux and vice versa The most extreme example of this uncorrellated flux is the so called orphan flare Most often regarding X ray TeV y ray correlation this occurs when the intensity in one band increases without a noticeable counterpart in the other Mrk 421 saw an orphan X ray flare on January 13 2003 Rebillot et al 2006 while LES 1959 650 saw an orphan y ray flare on June 4 2002 Krawczynski et al 2004 Figure A 1 shows a light curve of 1ES 1959 650 in various energy bands clearly showing this orphan TeV flare The existance of these orphan flares goes against the results of previous one zone SSC models of blazar emission Many groups have modeled the X ray and TeV y ray emission from Mrk 421 data Inoue and Takahara 1996 Bednarek 1997 Bednarek and Protheroe 1999 Boettcher et al 1997 Mastichiadis and Kirk 1997 Tanihata et al 2001 Krawczynski et aL 2001 Konopelko et al 2003 Kino et al 2002 102 A 3
91. elescope is included Start Run Runs can be defined as either manual runs or auto runs Manual runs are completely handled by the observer auto runs are defined and started by arrayctl When an auto run option is selected from the list Auto Run Type is enabled and the Define Prepare option becomes Define Auto runs are defined and prepared starting 1 minute before the requested start time This allows the required initializations to complete and still begin the run on time 120 C 2 Using the VAC Manually A manual run behaves as normal It is defined and pre pared immediately and the user is left to start it when he she sees fit Automatically start run in Runs can be started automatically a certain amount of time in the future This is currently the best way to handle on off runs having the second run start 2 minutes after the first run finishes The minimum offset time is 1 minute to accommodate the necessary run preparations However setting the offset to 0 minutes seconds are then ignored will start the first auto run immediately after it is defined When defining multiple auto runs be advised the offsets are all relative to the time the runs are DEFINED and not relative to when the previous run finishes Be sure to factor in the time needed to prepare the run into this offset time It is recommended to have the offset times be at least 2 minutes longer than the previous run s offset time plus run durat
92. es T3 power on fadc crates T4_power_on fadc_crates Move to Desktop DACQ T1 start fadc vdaq 1 make evtbuilder data dir 1 start evtbuilder 1 Move to Desktop DACQ T2 start fadc vdaq 2 make evtbuilder data dir 2 start evtbuilder 2 Move to Desktop DACQ T3 start fadc vdaq 3 make evtbuilder data dir 3 start evtbuilder 3 Move to Desktop DACQ T4 start fadc vdaq 4 125 C 2 Using the VAC make evtbuilder data dir 4 start evtbuilder 4 Start Night Should be called at the start of each observing night Takes care of initial izations needed for many of the subsystems currently L3 Array Control and EventBuilder In the future it will start more of the subsystems automatically End Night Should be called at the end of each observing night Currently it does a clean halt of L3 and lets Array Control do any nightly clean up it needs to Read FADC Temperatures Temperature information is read automatically every few minutes If instead you would like the temperatures immediately selecting this item will do so and display all temperatures read This can also be accomplished by pressing Ctrl T Check Free Disk Space The Harvester L3 Archive and VDAQ machines are queried automatically for the total available disk space for data Currently results are simply listed in the main message window If you would like to check the di
93. escope The image is 6 5 arcmin across corresponding to 3 4 pc at a distance of 2 kpc 18 2 3 TeV y ray Sources Another known source class for TeV emission is X ray binaries XRBs Though mainly emitting X rays they also have been known to produce sporadic y ray emission from the gas accreting onto the compact star in the binary pair So far only two XRB sources of 4 rays have been detected PSR B1259 63 Aharonian et al 2005b and LS I 4 61303 also a pulsar Albert et al 2006 The largest group of sources from which y rays have been detected are Active Galactic Nuclei AGN These galaxies contain a very compact core emitting an ex tremely disproportionate amount of energy compared to the rest of the galaxy This central engine is thought to be a supermassive black hole surrounded by an accretion disk see Fig 2 7 The disk is surrounded by fast moving clouds of dust that in some cases obscure the central engine from view though these clouds can produce Doppler broadened emission lines Farther from the nucleus in the direction perpen dicular to the plane of the accretion disk narrow emission lines are produced through scattering of the slower less Doppler broadening and less dense clouds surrounding the galaxy In some AGN jets of highly relativistic particles are ejected out the poles of the spinning central nucleus These jets contain large magnetic fields capable of producing synchrotron radiation up to X ray waveleng
94. eventdisplay package If something is wrong with the laser run all data from the night are affected since they are analyzed against this laser run Laser events should light up the entire camera It is very easy to tell if something is definitely wrong by flipping through a few of these events The custom ROOT script was used to analyze the data every morning For each analyzed data run it produced a series of plots used to determine at a glance if something major was wrong with the data The various histograms displayed include pedestals and pedestal variances gains time offsets the many Hillas parameters see Sect 4 1 1 number of tubes per event tube with the maximum value a rate plot time between consecutive events dt and centroid distributions 106 B 2 Analysis and Results Figure B 1 shows an example of the data output The plot distributions are representative of how things look when most everything is working as it should Many histograms show tight groupings in the data while others show a smooth decaying curve or a flat horizontal line The max pixel distribution however is not ideal It shows that one pixel is firing far more often than any other This could simply be the result of a star in the field of view or it could be a more important problem with faulty high voltage on that pixel While not obvious at first after a few days of daily analysis one learns to quickly recognize these shapes as good signs
95. f the past is any indicator these new experiments will set off an explosion of new learning in the field Not just extensions to older projects many of these telescopes utilize new ideas and technologies GLAST Gamma ray Large Area Space Telescope a new space borne y ray tele scope will use solid state detectors in place of the spark chambers of the past re ducing the required space while increasing efficiency Scheduled to launch in 2007 GLAST will detect y rays from 20 MeV 300 GeV with an effective collection area of gt 8000 cm It will have a much higher angular resolution and an order of magnitude better sensitivity than EGRET Ritz et al 2005 On the ground new telescopes are being built in addition to VERITAS The MAGIC Collaboration is building a second 17 m IACT at their site in La Palma and MAGIC II should be completed in 2007 Baixeras and et al 2005 The H E S S Collaboration is building a new telescope at their site in Namibia This new IACT will have a 28 m diameter mirror Placed in the center of the existing H E S S array the new telescope will work with the existing array as H E S S II achieving even lower energy thresholds than before Vincent 2005 These new telescopes will allow us to simultaneously monitor blazars at MeV to TeV energies Observing blazars with radio IR optical X ray and y ray telescopes will enable us to test the models with unprecedented spectral coverage and sensitivity 97
96. g in the run as well as the current run number e Harvester Displays values for some aspects of Harvester activity Currently this includes the number of sane and insane events as well as why if it was the telescope or L3 that caused them to be insane the current size of the data file being written in MB the number of telescope events and L3 triggers recorded and the telescope and L3 event rates from QuickLook The final two are an average over the last 10 bins of their respective rate vs time histograms and should start updating as soon as a run is PREPARED The number in the name of the telescope specific fields tells which telescope the statistics are for e Event Builder Displays values for various aspects of Event Builder activity These include telescope trigger rate number of events to the harvester number of events to disk and number of bad events The number following Event Builder tells which telescope the statistics are for e Telescope Displays values from L2 acquired by L3 These include the L2 Rate and VDAQ Deadtime The Deadtime is displayed as a percent 100 means totally dead 123 C 2 Using the VAC The number following Telescope tells which telescope the statistics are for Information should appear as soon as a run is prepared L3 Displays values from L3 including a string describing its status the Array Deadtime and L3 Rate Deadtime is displayed as a percent 10096 means totally dea
97. galactic and extragalactic sources Unlike GeV and TeV cosmic rays which are isotropized by galactic magnetic fields and bombard the Earth from all directions VHE rays come from particular objects in the sky Within the galaxy these sources are mainly pulsars X ray binary stars or supernova remnants while extragalactic sources are usually blazars One can easily pinpoint where the y rays are coming from because they are uncharged and therefore their trajectories are unaltered by magnetic fields that exist throughout space 28 3 1 y ray Propagation 3 1 1 Propagation Through Space The strength of extragalactic y ray signals is reduced by interactions with the intergalactic infrared IR background in pair production processes WHE Yin gt et e 3 1 The absorption is strong for a wide range of VHE rays above 20 GeV due to the broad peak in energy of the absorption cross section The peak occurs when Eve Bn cos0 2 m c 0 52 MeV y 3 2 where Eypg and Er are the energies of the VHE y ray and IR photon respectively and 0 is the angle of the collision between the two particles The mass of the electron and speed of light in vacuum are represented by m and c For 1 TeV photons this peak occurs when colliding head on with 0 5 eV photons However absorption is strong across a wide range of energies due to the spectral features of the extragalactic background Gould and Schr der 1967 Stecker et al
98. he most successful space based y ray telescope to date has been the Comp ton Gamma Ray Observatory CGRO Gehrels et al 1993 shown in Figure 2 1 Launched in 1991 it remained in orbit for over nine years It was built to observe y rays over the energy range of 15 keV 30 GeV with several different instruments 2 2 Instruments to Detect y rays Each of the instruments onboard the CGRO was designed for a specific purpose The Burst and Transient Source Experiment BATSE was able to detect y ray bursts GRBs on microsecond time scales in the 20keV 1 9 MeV energy range The Comp ton Telescope COMPTEL provided the first sky survey in the 1 30 MeV band The Oriented Scintillation Spectroscopy Experiment OSSE performed spectral ob servations in the 0 05 10 MeV energy range The final instrument was EGRET the Energetic Gamma Ray Experiment Tele scope Operating at the highest energies of any of CGRO s components 10 MeV 30 GeV EGRET was able to detect over 250 new ray sources in its lifetime 66 of which were blazars Thompson et al 1995 Hartman et al 1999 The detector itself was massive The amount of material needed to stop the high energy photons had a mass of 1900 kg and was approximately the size of a compact car yet had an effective collection area of only 1600 cm Fichtel et al 1993 More recent experiments are able to accomplish more with a lighter detector Swift With the retirement of the
99. hether a run is active on a particular tele scope Status is updated automatically This status tells whether eventbuilder for each telescope is currently processing a run or even if the run has just been prepared and therefore a new run cannot yet be defined for that telescope If the box shows the indeterminate value status cannot be determined because the connection to that telescope s eventbuilder cannot be established Run in Progress When a run is active this indicator is checked and the active run number s listed following it Information is updated automatically though it doesn t always coincide with when arrayctl commands are sent it may take a couple seconds to read accurately Date and Time This information is taken from the database and is updated automatically It is in the form yyyy mm dd hh mm ss Messages window All information regarding what the system is up to exceptions caught etc 117 C 2 Using the VAC Run Define Run Start Run Manuall Pick Custom Run Number T D Y Run Number C Automatically start run in Weather E min sec Run Type observing C Automatically start run Observing Mode on X Today Pointing Mode parallel C Automatically start run later Source Mrka21 Ss Trigger Config normal e c C Trigger Multiplicity 2 i Trigger Coincidence 100 Comments Wobble None C North C South y las 4 C East C West o 3 degree
100. ht for the harvester Close Closes the Harvester dialog box and returns to the main window Event Builder Subsystem The Event Builder subsystem window is shown in Figure C 6 Telescope The telescope used when a run specific command is selected Numbering is 1 4 however some errors may still report with the old 0 3 numbering scheme Run The run number used when a run specific command is selected If a run is selected in the main Run Information Table it becomes the default run number EVTB Start Run Instructs the event builder to start a given run The run number and telescope number are taken from the input lines above EVTB End Run Terminates the current run on a given telescope The telescope number is taken from the input line above The status of the run is then updated This status includes telescope trigger rate number of bad events number of events to the harvester and number of events written to disk Get EV TB Status Gets the event builder status for a given telescope The telescope number is taken from the input line above The status includes telescope trigger rade number of bad events number of events to the harvester and number of events written to disk This function is called automatically when the dialog is first opened Active EVTB Runs Sees if a run is currently active on a given telescope Status for 4 telescopes is given This function is called automatically when the dialog is first opened
101. idth MSCW KE 4 1 trig A woo ri sizei summing over all Nu telescopes with a trigger where width and size are the width 47 4 1 Event Reconstruction Run 1986 Event 87 Type 0 GPS 2005 E 5 59 57 33772 Max channel 500 Num Samples 24 Num Trigger 36 Num Tubes 53 Num Dead 42 Xcos 0 000 Ze 0 00 Ycos 0 000 Az 0 00 GEO c_x 0 06 c_y 1 08 dist 1 08 length 0 3622 width 0 1738 a 4 57 size 8053 92 Run 1986 Event 429 Type 0 GPS 2005 333 5 59 59 68064 Max channel 500 Num Samples 24 Num Trigger 35 Num Tubes 53 Num Dead 42 1009 oc ER 8o X 0000 0 0000000 X EAS ONUS Sooo OO o9 eee oO OCCT 00000400000 CX 000000000 coe 6 e Gd 8 Primary 0 Energy TeV 0 00 X 0 00 Y 0 00 Xcos 0 000 Ze 0 00 Ycos 0 000 Az 0 00 GEO c_x 0 46 c_y 0 17 dist 0 49 length 0 7211 width 0 6233 a 17 56 size 3142 10 Run 1986 Event 129 Type 0 GPS 2005 333 5 59 57 67546 Max channel 500 vc OO SR Num Trigger 19 JOA Go OL C ABC OQ OC Num Dead 42 XXX Yo Y eoo AA R 6 Se A CTT E E EE OA E E Primary 0 3 Energy TeV 0 X 0 00 Y 0 00 Xcos 0 000 Ze 0 00 Ycos 0 000 Az 0 00 GEO c x 0 10 c_y 1 06 dist 1 06 length 0 4172 width 0 1433 0 4 80 size 1749 40 Run 1986 Event 140 Type 0 GPS 2005 333 5 59 57 73382 Max channel 500 Num Samples 24 COX SESH
102. imes is weaker x 4 7 than the energy dependence of the synchrotron cooling time x 1 7 To see if our chosen range of 3 keV 25 keV could be causing problems additional energy ranges were considered For one data set we calculated the DCF between 3 keV and each of 25 keV 50 keV 75 keV 100 keV and 1 TeV The results comparing the calculated time lag and that predicted by Equation 5 14 are shown in Figure 5 7 We next considered flares of varying lengths The data sets were rerun using flares 89 5 6 Measuring Time Lags with the Discrete Correlation Function DCF 5 AAA p tsm de ope bene a 0 5 10 15 20 25 30 35 40 45 flare duration h FIGURE 5 8 The DCF times were calculated from runs with flares of 5 h 10 h 20 h and 40 h durations There is an obvious correlation with flare duration and DCF lag times of 5 10 20 and 40 hr durations As Figure 5 8 shows there is a strong dependence on flare duration of the DCF lag times Another case we examined was that of a single flare Using these light curves we calculated the DCF as before We also did so using just the peak of the light curve generated by the single flare Then we hand calculated the time difference between the maxima of the curves for 3 keV and 25 keV We also tried fitting an exponential to the decaying light curve to find a cooling time None of the methods yielded similar values We also examined other possibilities but none yielded better
103. in fact deserve to be mentioned again for that much needed downward spiral into trashy TV Anna MacKay has also provided entertainment though in much more than a guilty plea sure capacity Finally I must thank all my friends out in Phoenix Most important of these is Mike Drabick who not only let me crash at his place when I needed to get away but without whom I would never have met any of the many other wonderful people I have come to know in my visits out there 111 Contents Acknowledgements List of Figures List of Tables Abstract Copyright 1 Summary of Thesis Work 2 Astrophysics of Blazars 2 1 2 2 2 3 2 4 Introduction to y ray Astrophysics Instruments to Detect rays vu kos aa Seres mes 2 2 1 Space based Instruments 2k lt 4 4 25h E Ee a es Compton Gamma Ray Observatory A Cs aree SON e ee be EON GLASE EE 2 2 2 Ground based Instruments 44 241 va eR ps Imaging Atmospheric Cerenkov Telescopes Cerenkov Solar Array Telescopes o o o Particle Air Shower Aas ENE SEAM SSOMECESS db E cheek sta Sekt Andee Sek e GAS Sens NE C CC AN EEN 2 4 1 Spectral Energy Distributions e ER exe A 2 4 2 Emission Models and Particle Acceleration Synchrotron Self Compton y dict y oki d om Ye nego External Compra jes 2426 ond freund S mes Hadronic Models 4 e ac ROS Tert RR Rd e ea Rer Particle Acceleration o 4 aca dee RE po a Ty semis 2 du Markarian 2T s
104. in the table within a few seconds The run status is also updated automatically Define Run Before a run can be started it must be defined Choosing this option prompts the user for all information necessary to create a run see Figure C 2 The required fields are as follows 1 Run Number usually automatically chosen by the database If the field is left to zero this will still happen If you wish to suggest your own run number for the new run you must first click on the Pick Custom Run Number button A dialog will ask you to confirm overriding the automatic selection process and the run number may now be changed You will be asked a second time to confirm your choice of selecting a custom run number when you change the current run number from 0 Any unused run number is a valid choice If you change your mind simply click Use Default Run Number and the database will select the run number for you Weather A to C rating the current conditions Options appear in a pop up list Addional info can be supplied in the comments below Run Type observing chargelnjection laser pedestal bias curve other test describing the type of run to be taken so the proper analysis can later be done on it Options appear in a pop up list The additional options for defining chargeInjection runs are not supported but the identifier should still be used to tag these runs and parameters chosen from the separate QI GUI Also the other
105. ion If you are defining and ON OFF pair you don t have to worry about these specifics Automatically start run Today You can request the run to start at a specific UTC time The format for this time is hhmmss and refers to the current day only You cannot currently use the VAC to define runs for the following or other future UTC day The time refers to the time on the arrayctl computer so the observer should make sure that it is accurate Automatically start run Later This option is useful for setting up ON OFF or other auto run pairs without giving them a specific start time The runs will be set up when you click Define but the run timers will not start until you Activate the runs in the main window Auto Run Type For ease of observing ON OFF pairs can be auto defined automatically An unrelated series of runs can also be auto matically defined This section determines which is the case and how the auto runs are set up x ON OFF amp OFF ON An ON OFF or OFF ON run pair can now be automatically defined at once The runs are offset in time by two minutes plus the duration chosen a 28 minute duration yields two runs offset by 30 minutes etc The sequence starts according to the time offset or absolute time defined above The second run will then be properly spaced in time to follow The off run is given the correct RA offset in the database The two runs are also grouped in the database as an on off pai
106. ion of the various Hillas parameters 46 4 1 Event Reconstruction values for the different Hillas parameters 4 1 2 Stereo Reconstruction After an event has been parameterized for each telescope the images are combined into one stereo event This stereo reconstruction involves using the intersection of the major axes of the ellipses of all telescopes to find the shower direction and axis see Fig 4 3 From this reconstruction we gain a new parameter 0 the square of the angular distance of the shower core from the center of the camera Cutting on this distance is the most reliable way to discard the maximum number of background events while keeping as much signal as possible An important parameter we derive from the analyzed data is the mean scaled width M SCW which also proves to be very useful in separating hadronic from y ray showers MSCW and its similarly derived cousin mean scaled length MSCL utilize the fact that hardronic showers appear significantly wider in an IACT cam era due to the transverse momentum associated with the nucleons and mesons as their interactions cascade through the atmosphere Using Monte Carlo simulations see Sect 4 5 we create a look up table of expected median width values wm and the 90 widths of the distributions two as functions of both size and r the distance of the telescope from the shower axis From these tables we calculate a normalized width 1 Mtrig
107. l B used in each set and those calculated from the DCF for each set using Eq 5 14 Errors here are propagated from the errors on the DCF times Set Birue G Bbcr G 0 20179 0 32 0 01 0 07063 0 127 0 001 0 04708 0 088 0 002 0 01412 0 045 0 005 0 00614 0 034 0 012 0 00562 0 035 0 014 300O00u gt 87 5 6 Measuring Time Lags with the Discrete Correlation Function DCF 10 1 aval Bock G 0 1 E ee O0l ee E peca p n el 0 01 0 1 G true FIGURE 5 6 The magnetic field values calculated from the DCF lags plotted against the actual values used in the simulations RMS values are propagated from the RMS values on the DCF time lags The dotted line represents the two values being equal See Table 5 3 for specific values 88 5 6 Measuring Time Lags with the Discrete Correlation Function DCF LOH VI O 1 a EDI Ar h sync FIGURE 5 7 The DCF times were calculated between 3 keV and left to right each of 25 keV 50 keV 75 keV 100 keV and 1 TeV The comparison to the expected lag time is shown The dotted line represents the two values being equal Compton cooling are almost identical We explain this finding with two factors First synchrotron emission is the dominant cooling mechanism for all considered parameter combinations Second the energy dependence of the inverse Compton cooling t
108. l aim at intensive mul tiwavelength observations In this thesis a theoretical study is described illuminating the possibility of using the X ray data of such campaigns to constrain the jet mag 94 6 2 VERITAS Performance netic field The study has shown that simple methods used by other authors do not perform as expected 6 2 VERITAS Performance Single telescope comparisons in Holder et al 2006 show that even T1 by itself shows significant advances over the Whipple 10 m telescope The stereo set up already greatly reduces the number of background events causing the telescope to trigger average trigger rates are down from 160 Hz for T1 data to 90 Hz for stereo data This rate decrease is also in spite of the fact that the telescope trigger levels have been lowered as well capitalizing on the increased background rejection to ensure events are not dominated by noise The agreement between the Monte Carlo simulations and data taken on Mrk 421 demonstrates that the telescopes are behaving as planned The simulations take into account how the electronics of the system are supposed to behave The distributions of the parameters examined in this study are very similar Unfortunately for this calibration Mrk 421 is a very active source with its flux varying constantly One needs to study a more steady source such as the Crab Nebula to further solidify the extent of the improvement of VERITAS over previous generations of IACTs A
109. laga R Plyasheshnikov A Prahl J P hlhofer G Rauterberg G R hring A Sa hakian V Samorski M Schilling M Schmele D Schr der F Stamm W Tluczykont M Volk H J Wiebel Sooth B Wiedner C Willmer M Wittek W Chou L Coppi P S Rothschild R and Urry C M 2000 ApJ 538 127 Schmitt J L 1968 Nature 218 663 Sembay S Edelson R Markowitz A Griffiths R G and Turner M J L 2002 ApJ 574 634 Sikora M and Madejski G 2001 in F A Aharonian and H J V lk eds American Institute of Physics Conference Series pp 275 4 Smith D A Brion E Britto R Bruel P Bussons Gordo J Dumora D Durand E Eschstruth P Espigat P Holder J Jacholkowska A Lavalle J Le Gallou R Lott B Manseri H M nz F Nuss E Piron F Rannot R C Reposeur T and Sako T 2006 Astronomy and Astrophysics 459 453 153 Bibliography Sokolov A and Marscher A P 2005 ApJ 629 52 Stecker F W de Jager O C and Salamon M H 1992 ApJL 390 L49 Takahashi T Kataoka J Madejski G Mattox J Urry C M Wagner S Aharonian F Catanese M Chiappetti L Coppi P Degrange B Fossati G Kubo H Krawezynski H Makino F Marshall H Maraschi L Piron F Remillard R Takahara F Tashiro M Terasranta H and Weekes T 2000 ApJL 542 L105 Takahashi T Tashiro M Madejski G Kubo H Ka
110. lectromagnetic spec trum from radio all the way through y rays It is characterized by two broad peaks one in the optical to X ray band and the other in MeV GeV rays see Sect 2 4 1 The continuum emission is strongly polarized with a high variability on short time 20 2 4 Blazars scales It is believed to be produced by non thermal processes synchrotron and in verse Compton see Sect 2 4 2 most likely coming from the blazar s jets Blandford and Rees 1978 While technically divided into two subclasses flat spectrum radio quasars FS RQs and BL Lacertae objects BL Lacs some observations suggest that the dis tinction may not be so clear cut Ghisellini 1999 Blazars detectable in TeV 4 rays are all BL Lacs which lack the strong emission lines that distinguish them from FS RQs They get their name from BL Lacertae the first object to be identified with these properties Schmitt 1968 They were usually discovered as extragalactic coun terparts to strong radio sources Distances of BL Lac sources are also very difficult to measure due to their lack of spectral emission lines and the dominance of the nuclear emission over the emission of the host galaxy BL Lac objects are also rare which is consistent with the overall small probability that a source of this type has jets within 10 of the line of sight 2 4 1 Spectral Energy Distributions The Spectral Energy Distribution SED is usually plotted as the p
111. ll between them They also shield the camera from stray background photons that do not originate from the source direction A final design has now been chosen but at the time the data for this thesis were taken Telescopes 1 and 2 had light cones of different designs This fact does not adversely affect this early stereo data The PMTs are operated with a gain of 2 x 10 electrons photoelectron necessary to detect single Cerenkov photons from the air showers The high voltage system to bias the PMTs is housed in the control shed next to each telescope 41 3 8 VERITAS Very Energetic Radiation Imaging Telescope Array System The pre amplifier attached to the base of each PMT is set up to allow constant monitoring of the PMT currents as well as to inject charge pulses into the system for calibration and testing purposes The current monitoring system allows for the auto or manual suppression of individual channels with high currents owing to faulty electronics or a star in the field of view 3 3 3 Trigger In order for an event to register and be fully processed it must pass a three level triggering system These correspond to triggering on the pixel telescope and array levels The first level L1 consists of fast constant fraction discriminators CFDs processing the analog PMT signals The CFDs contain a programmable 6 ns delay to compensate for different PMT transit times and cable lengths From the CFDs the L1 triggering signa
112. ls are sent to the Level 2 L2 trigger This is a topological system that uses a programmable look up table of patterns to determine if a given number of neighboring pixels have triggered Currently the coincidence condition requires three neighboring pixels to trigger within 10 ns Each telescope L2 trigger is then sent to the Level 3 L3 trigger For the two telescope data used for this thesis the L3 trigger was set to require both telescopes to record an event within a 100 ns coincidence window This relatively wide window ensured the array would trigger even if there were some problems with timing between the two telescopes If an event passes the L3 trigger it is tagged with the GPS time and the event is then read out from the data acquisition system 42 3 3 VERITAS Very Energetic Radiation Imaging Telescope Array System 3 3 4 Data Acquisition One of the technical strengths of VERITAS is the data acquisition system At its heart is the 500 MHz flash ADC FADC system Buckley 1999 The signal from each PMT is digitized in 2 ns bins and stored with a total lookback memory of 32 us The FADC system allows for continuous data taking while the triggering system determines the validity of events as well as the possibility for advanced timing studies of the individual showers The FADC boards record data until they receive a signal from the L3 trigger that an event has occurred The data aquisition pauses until the relevant data are r
113. lt 0 22 was used 2000 1000 057 SIE 15 MSCW IAE E 32 2 E ba z 0 8 SE FIGURE 4 13 Plot of the Q factor as I 35 a function of the applied MSCW cut ger EN solid line right axis Also plotted I SES are the signal acceptance blue dashed 0 4 I Zos line and background acceptance red dot dashed line E 30 6 o2 II EI g a o2 O yc n Toca Eno ca scm E 05 0 5 10 15 cut in MSCW gives a maximum Q factor QMAX 2 15 with a signal acceptance of 61 and a background rejection of 92 for the cut of MSCW lt 0 21 4 6 3 Mrk 421 Light Curve After the cuts on 0 and MSCW were established the data were processed again 1 using these cuts This yielded an average y ray rate of 2 91 0 07 y min after 68 4 6 First Stereo Results from VERITAS TeV flux arbitrary units S 9 9 e E eg o a i d 53846 53848 53880 53852 83854 83856 date MJD FIGURE 4 14 Light curve from the April May 2006 dark run for Mrk 421 Each data run is represented by one point on the graph Error bars are on the one sigma confidence level cuts Correcting for the cut efficiencies the true rate inferred from these data is 8 83 0 21y min The total significance was found to be 390 for 14 3 hours of data The flux measured for each of these runs was used to produce a preliminary light curve of Mrk 421 seen in Figure 4 14 Ab
114. ly visible as a very weak 50 photons m very short 5 ns pulse While the intensity of the shower requires highly sensitive electronics to record the pulse duration is what allows us to distinguish them from much of the night sky background Many methods have been utilized to better discriminate between the y ray signal and background The biggest advancement in this field came with the development of technology to image the individual air showers A camera with several individual PMT pixels can be used to determine both the size shape and intensity of the shower Due to the differences in the lateral spread of various air showers they can lead to quite different looking images when seen from the ground Figure 3 5 shows the difference between a y ray induced shower and a cosmic ray induced shower propa gating through the Earth s atmosphere y rays produce a very small tight round image on the camera while hadronic showers produce a larger broader shape This difference is caused by the transverse momentum of nucleons and mesons present in cosmic ray showers but not in y ray showers 35 3 2 Detection Using an Imaging Atmospheric Cerenkov Telescope IACT FIGURE 3 5 y ray induced air showers are very tight left while cosmic ray induced air showers are much broader Both simulated showers were initiated by a particle of energy 100 GeV Red lines represent electrons positrons and photons and green and blue lines represen
115. mae T Kataoka J Kii T Makino F Makishima K and Yamasaki N 1996 ApJL 470 L89 Tanihata C Urry C M Takahashi T Kataoka J Wagner S J Madejski G M Tashiro M and Kouda M 2001 ApJ 563 569 Tavecchio F 2005 in M Novello S Perez Bergliaffa and R Ruffini eds The Tenth Marcel Grossmann Meeting On recent developments in theoretical and ex perimental general relativity gravitation and relativistic field theories pp 512 529 Tavecchio F Maraschi L and Ghisellini G 1998 ApJ 509 608 Ter Haar D 1950 Reviews of Modern Physics 22 119 Thompson D J Bertsch D L Dingus B L Esposito J A Etienne A Fichtel C E Friedlander D P Hartman R C Hunter S D Kendig D J Mattox J R McDonald L M von Montigny C Mukherjee R Ramanamurthy P V Sreekumar P Fierro J M Lin Y C Michelson P F Nolan P L Shriver S K Willis T D Kanbach G Mayer Hasselwander H A Merck M Radecke H D Kniffen D A and Schneid E J 1995 ApJS 101 259 Vincent P 2005 in International Cosmic Ray Conference pp 163 4 Weekes T 2003 Very High Energy Gamma Ray Astronomy Institute of Physics Publishing Bristol Weekes T C Badran H Biller S D Bond I Bradbury S Buckley J Carter Lewis D Catanese M Criswell S Cui W Dowkontt P Duke C Fegan D J Finley J Fortson L Gaidos J Gillanders
116. me as the simulated blazar goes through flaring cycles Such shifts have been also observed for example in Mrk 501 Pian et al 1998 2 4 2 Emission Models and Particle Acceleration Several theories have been presented to account for the unique SED of blazars More observations in particular simultaneous observations at many different wave lengths are necessary to break the model degeneracies and prove the mechanism by which particles are being excited to such extremely high energies Synchrotron Self Compton The Synchrotron Self Compton SSC model was originally proposed by Ginzburg and Syrovatskii 1969 It has since been expanded for spherically homogeneous sources and evolved to incorporate relativistic jets It is the simplest explanation for 23 2 4 Blazars blazar emission in that the same population of relativistic electrons is responsible for both the X ray and y ray peaks of the SED The blazar s strong magnetic field accelerates the electrons in its jets which radiate synchrotron photons in the process creating the lower SED peak Inverse Compton processes then cause the upper peak as these radiated photons collide with the same relativistically accelerated electrons that created them in the first place The most basic version of this scenario is the one zone model where emission comes from a shock front moving along the jet Sikora and Madejski 2001 This emission zone has a homogeneous magnetic field and
117. ms in the visible tab will be executed when Do Night is chosen The items are grouped by subsystem e Start Night L3 Init Night calls L3 s own night initialization routine Eventbuilder Init VME Config initializes the system for each telescope listed If settings from the database have changed they are reloaded Arrayctl Start Night performs Array Control specific initialization items Clear Internal Run List removes old completed runs from the Run Information Table seen in VAC e End Night 139 C 2 Using the VAC Custom Night Start Night End Night L3 Iv Init Night Event Builder Iv Init VME Config T1 di Iv Init VME Config T3 M Init VME Config T4 FIGURE C 9 Layout of the Custom Array Control Night window Iv Start Night Iv Clear Internal Run List Database r Check All Uncheck All cuca 140 C 2 Using the VAC Put CFD Settings Telescope Thresholds mV FiGURE C 10 Layout of the Put CFD Settings window Width ns RFB mV MHz Put Settings L3 Quit L3 safely terminates the L3 system Arrayctl End Night performs Array Control specific nightly clean up items e Check All Uncheck All For easier handling of a large number of separate steps these buttons select or deselect all of the above options regardless of their current state e Do Night Sequentially carries out the selected items in the currently visible tab If a
118. naffected by magnetic fields permeating the uni verse y rays are very directional and arrive at Earth from distinct points in the sky The Earth s atmosphere is opaque to 7 rays so the only way to see them directly is from space However indirect techniques have been developed to observe y ray sources from the ground as well see Sect 2 2 2 2 2 1 Space based Instruments Early balloon experiments detecting cosmic rays suggested that moving beyond the Earth s atmosphere might be advantageous to finding even higher energy rays Despite its small collection area and poor angular resolution the Explorer XI satellite launched in 1965 was able to prove the existence of y rays originating outside the Earth s atmosphere Clark et al 1968 The practice of using balloons for flying spark chambers to detect y rays was effectively ended in 1972 with the launch of 6 2 2 Instruments to Detect y rays FIGURE 2 1 A collection of several instruments including EGRET center and BATSE eight detectors one on each corner of the spacecraft the Compton Gamma Ray Observatory was highly successful in observing a wide range of y ray phenomena Figure from http cossc gsfc nasa gov NASA s SAS 2 The European Space Agency s COS B was launched soon after in 1975 By helping map the y ray sky in detail both these telescopes established y ray astronomy as a new and exciting field worthy of further study Compton Gamma Ray Observatory T
119. nd Lightman 1979 as 4 B x teoolly ore ss d 5 10 with the Thomson cross section op 6 652 x 107 cm and B 87 the magnetic 82 5 5 Analysis Procedure field energy density Jet frame times transform into observer frame times according to 1 t t 5 11 5 5 11 From Equations 5 10 and 5 11 we can derive the difference in synchrotron cooling times of the electrons with Lorentz factors and y2 responsible for the radiation observed at energies Ej and E At tasa Tafele 5 12 synch 4 B He 1 jos 52 E 2 5 13 Inverting the equation one can compute the jet magnetic field in the jet frame from the time lag Ati R 1 3 i i 1 3 done E Se 5 14 or At en E E V E Ej c Ltd BI 7 x 8 45 x 0 045G 5 15 Various authors have used Tpcr the time offset that maximizes the DCF as an estimator of At and have used Equation 5 15 to estimate the jet magnetic field synch As we will show below this procedure can produce incorrect results 83 5 6 Measuring Time Lags with the Discrete Correlation Function DCF 5 6 Measuring Time Lags with the Discrete Cor relation Function DCF For each of the six different parameter combinations we simulate 20 independent artificial data sets to estimate the statistical accuracy to which we can determine Tpcr As an example of our procedure Figure 5 4 summarizes the DCF results for parameter combination
120. nd starts a test run on Telescope 3 As a shortcut use Ctr1 3 e Make Test Run T4 Configures and starts a test run on Telescope 4 As a shortcut use Ctr1 4 129 C 2 Using the VAC Control L3 Status Howzaboy 87114001 Init L3 Night End L3 Night Run Number 324 Next Run Run Status Henric We used up all our funds Runs 11184 Config Mask i Coincidence Window 2 8 L3 Rate 28 3 Multiplicity 1 Total Deadtime 1964 00 Start Runs Resume Runs Pedestal Rate 20 9 L3 Deadtime 843 33 Config Runs Pause Runs Stop Runs Abort Runs Telescope Status Telescope 2 3 4 clear Error L2 Rate 15 3 4 5 QI Rate 1 25 2 5 375 High Multiplicity Rate 1 16667 2 33333 Fa Get Status M L3 Alive New Physics Rate 15 3 4 5 VDAQ Deadtime i 112 50 225 00 337 50 Se L3 Output Rate 12 24 3 6 Delay Config Type OBSERVING_NORMAL y L3 L2 Output Rate 1 14286 228571 3 42857 Load Delay Config FiGURE C 4 Layout of the L3 subsystem window C 2 4 Subsystems Menu L3 Subsystem The L3 subsystem window is shown in Figure C 4 e Init L3 Night Initializes the night for L3 This is called as part of Start Night from the main window but is here for when it must be called separately when L3 is brought back online after a problem for example e End L3 Night Concludes the night for L3 This is called as part of End Night from the main window but is here for when it must be called separately when L3 is brought
121. nolds P T Roache E T Rose H J Schroedter M Sembroski G H Sleege G Steele D Swordy S P Syson A Toner J A Valcarcel L Vassiliev V V Wakely S P Weekes T C White R J Williams D A and Wagner R 2006 Astroparticle Physics 25 391 Holt S Neff S G and Urry C M 1992 Testing the AGN Paradigm Springer Verlag Horan D 2001 Ph D thesis University College Dublin Horan D and Weekes T C 2004 New Astronomy Review 48 527 Inoue S and Takahara F 1996 ApJ 463 555 Jahoda K Markwardt C B Radeva Y Rots A H Stark M J Swank J H Strohmayer T E and Zhang W 2006 ApJS 163 401 Jelley J V 1958 Cerenkov Radiation and its Applications Pergamon Press New York Katz J I 2006 ArXiv Astrophysics e prints astro ph 0603772 Kertzman M P and Sembroski G H 1994 Nuclear Instruments and Methods in Physics Research A 343 629 Kino M Takahara F and Kusunose M 2002 ApJ 564 97 Kirk J G and Duffy P 1999 Journal of Physics G Nuclear Physics 25 163 Kirk J G and Mastichiadis A 1999 Astroparticle Physics 11 45 Konopelko A Mastichiadis A Kirk J de Jager O C and Stecker F W 2003 ApJ 597 851 Kosack K 2005 Ph D thesis Washington University in St Louis Kosack K Badran H M Bond I H Boyle P J Bradbury S M Buckley J H Carter Lewis D A Celik O Connaughton V Cui W
122. nte Carlo Data TABLE 4 1 Continued Date Run Number OFF Run Laser Run Run Type Elevation 2006 04 27 30504 laser 2006 04 27 30492 30504 0 3E 82 2006 04 27 30493 30504 0 3W T 2006 04 27 30494 30504 0 3N 70 2006 04 27 30495 30504 0 35 64 2006 04 27 30496 30504 0 3N 58 2006 04 27 30497 30504 0 35 52 2006 04 29 30540 laser 2006 04 29 30533 30540 Trk 83 2006 04 29 30534 30540 Trk 81 2006 04 29 30536 30540 Trk 69 2006 04 29 30537 30540 Trk 63 2006 04 30 30564 laser 2006 04 30 30553 30564 Trk 78 2006 04 30 30554 30564 Trk 73 2006 04 30 30555 30564 Trk 68 2006 04 30 30556 30564 Trk 62 2006 04 30 30557 30564 Trk 56 2006 04 30 30559 30564 Trk 45 2006 04 30 30560 30564 Trk 41 2006 05 01 30575 laser 2006 05 01 30570 30575 Trk 65 2006 05 01 30572 30575 0 35 52 4 5 Comparison of Experimental and Monte Carlo Data In order for the energy spectrum or other results relying on Monte Carlo simu lations to be reliable one must first prove that the simulations accurately describe 58 4 5 Comparison of Experimental and Monte Carlo Data the system they are intended to model By comparing the distributions of the vari ous parameters in the simulated data set to those from experimental data one can test how well the Monte Carlo simulations are performing and how accurate results obtained with them will be Here we use
123. of Experimental and Monte Carlo Data o o LEI t B W pe wy E N 0 8 E E L o ES 06 0 4 0 2 0 l L L L dE T En L L L i L L L L L L L En NN L L J L 2 1 5 1 0 5 0 0 5 1 1 5 2 log Energy FIGURE 4 6 Plot showing the log energy distributions of the Monte Carlo simula tions for a Crab like spectrum of E both before black and after red cuts The energy bin with the most entries is the energy threshold 165 GeV before cuts and 220 GeV after cuts One can also plot how the angular resolution of the simulations vary with energy For a given energy bin the angular resolution is defined as the angular distance 0 below which occur 63 of the events This resolution is plotted in Figure 4 7 and Table 4 2 lists the values as well as the energy bins used For the entire energy range the angular resolution is 0 29 in simulations Similarly the core resolution can be plotted This measures the difference in position of the reconstructed shower core position from the actual position Mears as V Cose m MOTE Haase m M Gase 4 5 where MCZcore MCYcore is the true shower core and Zcore Ycore is the recon structed core eventdisplay can calculate the reconstructed core in two different ways 61 4 5 Comparison of Experimental and Monte Carlo Data
124. omatically updated 127 C 2 Using the VAC in the Current Active Run section Auto runs can be started early by choosing to Activate them at any point before they have automatically started Run Info for Current Active Run When a run is in progress its status is updated automatically and the results displayed in the Currnet Active Run section of the main window Through the telescope listing the observer can select which of multiple active runs they wish to view information for When no run is in progress all fields display 1 and no times are listed on the progress bar Information is sampled every few seconds and doesn t always sync with arrayctl commands Hence when a run is started stopped it may take a couple seconds for the Run Info to reflect this For an immediate update of status choose Update Status button under Observer menu The error message like updateRunTimes arrayctlException Statusmon exception No status for run Thrown in ac cpp line 1041 is nothing to worry about It just signifies that the VAC is trying to get the status of the run before enough events have been processed for the status to exist Eventually QuickLook will be incorporated into this display and even more run information will be available e Progress Bar The start and end times for the current run are displayed along with a progress bar indicating how much of the run has completed Listed below are the time elapsed and time remainin
125. oments of the image distribution are used to fit an ellipse to each event s image This process is described more fully in Reynolds et al 1993 The resulting parameters describing the size length and width placement dist ance and miss and orientation a of the ellipse are known as Hillas parameters Figure 4 1 shows these parameters in detail Additional parameters used include size the total number of digital counts in the image This is directly related to the shower s energy One great advantage to analyzing images this way is that air showers caused by different particles look dramatically different in the camera Since more than 9996 of the events recorded come from cosmic ray showers it is beneficial to be able to distinguish them from the more important y rays in the data y ray initiated showers create an ellipse with its major axis pointing towards the camera s center The large transverse momentum associated with the strong hadronic interactions cause cosmic ray showers to create a more rounded concentration of light Muons on the other hand produce a large ring resulting from the initial interaction occurring relatively close to the camera itself Any other odd shape is just background noise Figure 4 2 shows examples of these different types of events for one VERITAS camera These differences in shape allow for more efficient rejection of background events based on 45 4 1 Event Reconstruction FIGURE 4 1 Illustrat
126. on is currently unsupported FADC Settings This section is currently unsupported 142 Bibliography Aharonian F Akhperjanian A Beilicke M Bernl hr K B rst H G Bojahr H Bolz O Coarasa T Contreras J L Cortina J Denninghoff S Fonseca M V Girma M G tting N Heinzelmann G Hermann G Heusler A Hofmann W Horns D Jung I Kankanyan R Kestel M Kohnle A Konopelko A Kornmeyer H Kranich D Lampeitl H Lopez M Lorenz E Lucarelli F Mang O Meyer H Mirzoyan R Moralejo A Ona Wilhelmi E Panter M Plyasheshnikov A P hlhofer G de los Reyes R Rhode W Ripken J Rowell G Sahakian V Samorski M Schilling M Siems M Sobzynska D Stamm W Tluczykont M Vitale V V lk H J Wiedner C A and Wittek W 2003 Astronomy and Astrophysics 403 L1 Aharonian F Akhperjanian A G Aye K M Bazer Bachi A R Beilicke M Benbow W Berge D Berghaus P Bernl hr K Boisson C Bolz O Braun I Breitling F Brown A M Bussons Gordo J Chadwick P M Chounet L M Cornils R Costamante L Degrange B Djannati Atai A O C Drury L Dubus G Emmanoulopoulos D Espigat P Feinstein F Fleury P Fontaine G Fuchs Y Funk S Gallant Y A Giebels B Gillessen S Glicenstein J F Goret P Hadjichristidis C Hauser M Heinzelmann G Henri G Hermann G Hinton
127. ort times spiking to its peak at the deadtime for the system currently 500 us The curve should then exponentially decay out to longer values of dt This is because ignoring the deadtime detectable events occur randomly in time meaning a gaussian distribution of the time between events Any deviation from this distribution means there is some bias within the system causing events to be detected at more regular intervals The dt bump shown in Figure B 3 was noticed to occur at different sizes relative to the surrounding curve It became a much more prominent effect when runs with very high data rates were analyzed Here the dt curve decayed to zero at a value less than 0 04 s This left the bump standing alone see Fig B 3b Much speculation surrounded what could be causing this bump Physically it corresponds to many events occurring with the same amount of time between them This means we had a fairly consistent 25 Hz signal on top of our real data Eventually the problem was traced down to the EventBuilder When its event buffer fills it must be flushed sending all the events further up the data chain This transfer time acts as a secondary deadtime putting the telescope on hold until the transfer is complete Since the data buffers are always the same size it takes roughly the same amount of time for each transfer This periodic disruption in the data flow 110 B 3 The dt Bump delta T 1193 d20051129 delta T 2332 d20060
128. ower emitted at each frequency per logarithmic energy interval vF versus frequency v on a log log plot Also called a power spectrum it is an easy and compact way to view information about the frequency distribution of the emitted power across the entire electromagnetic spectrum in one plot As mentioned above the SED of blazars exhibits two broad peaks see Fig 2 8 21 2 4 Blazars 44 ET un A oN o 242 gt bf 40 1ie 12 lerle 1e420 ler24 Les v Hz FIGURE 2 8 Typical Spectral Energy Distribution for AGN There are two distinct peaks one in the optical to X ray band and the other in y rays Colored bars represent the ranges over which various detection methods are used X ray the upcoming GLAST satellite and IACTs 22 2 4 Blazars The first peak is usually in the optical to X ray band and almost universally at tributed to synchrotron emission The second peak occurs in the MeV to GeV band but there is still much speculation as to what causes this emission The most accepted mechanism for this second peak is inverse Compton emission Various theories are discussed further in Section 2 4 2 Since the flux from blazars varies considerably over time their SEDs are also changing When BL Lac sources get brighter the emission peaks shift to higher energies Using computer code to model blazar emission developed by Coppi 1992 we have seen the SED peaks evolve over ti
129. ported will be used to display L1 rates for each telescope Database Subsystem The Database subsystem window is shown in Figure C 8 e Source Info If the source you want to observe is not in the source list you can add it here Simply type in the new source s name RA Dec and Epoc and then click New Source to confirm the addition Note RA and Dec are in RADIANS To easily convert from hhmmss and ddmmss to radians you can use the program RADEC 2rad located on the arrayctl computer under home observer shugh es RADEC2rad Source Lists all sources currently in the database in a pop up list Selecting one brings up its corresponding information in the fields below New Source Name Re lists the source name chosen from the above list so the source info can be edited Or the user may type in a new name to add a new source to the database RA Dec Epoch Right Ascension Declination and Epoch of the source currently described Description A brief description of the source can be added to its database entry 136 C 2 Using the VAC Source Info Telescope Info Source Telescope 0 New Source Name 1ESO120 Hostname RA 0 362774 Offset North 0000 Offset East 0000 Dec 0 599462 Offset Altitude 0000 Epoch 2000 Mirror Radius 0000 Description Save New Info Run Info Run Number 11183 4 Active Run info Get Run Info Put DB Start Time Put DB Stop Time Put Data Start Time Put Data Stop Time New
130. pp 488 490 Catanese M and Sambruna R M 2000 ApJL 534 L39 Catanese M and Weekes T C 1999 Publ Astron Soc Pac 111 1193 Cawley M F Fegan D J Harris K Kwok P W Hillas A M Lamb R C Lang M J Lewis D A Macomb D Reynolds P T Schmid D J Vacanti G and Weekes T C 1990 Experimental Astronomy 1 173 Celik O 2007 in VERITAS Collaboration Meeting 147 Bibliography Chiappetti L Maraschi L Tavecchio F Celotti A Fossati G Ghisellini G Giommi P Pian E Tagliaferri G Treves A Urry C M and Zhang Y H 1999 ApJ 521 552 Clark G W Garmire G P and Kraushaar W L 1968 ApJL 153 L2034 Coppi P S 1992 MNRAS 258 657 Coppi P S and Aharonian F A 1999 ApJL 521 L33 Cui W 2006 ArXiv Astrophysics e prints astro ph 0608042 Danaher S Fegan D J Porter N A Weekes T C and Cole T 1982 Solar Energy 28 335 Daum A Hermann G Hess M Hofmann W Lampeitl H P hlhofer G Aha ronian F Akhperjanian A G Barrio J A Beglarian A S Bernl hr K Beteta J J G Bradbury S M Contreras J L Cortina J Deckers T Feigl E Fernandez J Fonseca V Frass A Funk B Gonzalez J C Heinzelmann G Hemberger M Heusler A Holl I Horns D Kankanyan R Kirstein O Kohler C Konopelko A Kranich D Krawczynski H Kornmayer H Lindner A Lorenz E Magnus
131. proceeds relativistically down the jet as electrons are constantly injected The resulting spectrum of relativis tic electrons can be described by a broken power law with a shoulder at the break frequency Vp External Compton Similar to SSC External Compton EC models involve the same group of rela tivistic electrons radiating at lower energies by synchrotron radiation and at higher energies by inverse Compton IC radiation However the difference lies in the fact that the dominant seed photons for the IC emission come from outside the jet and are not the same photons already being radiated through synchrotron processes If the Compton scattering of ambient photons dominates the SSC emission the energy density of the external radiation measured in the jet frame must exceed the energy density of the jet produced synchrotron radiation This requires the ambient photons 24 2 4 Blazars to be upscattered far from the source of synchrotron radiation gt 10 cm so the newly created y rays are not lost to absorption by thermally emitted photons through pair production Maraschi et al 1992 Hadronic Models This theory involves a population of protons with very high gt 10 eV energy being created near the core of the AGN that travel down the jets An intense proton flux near the jet base produces pions both neutral and charged which decay into y rays and electrons respectively These high energy electrons gt 10
132. r Unrelated Many auto runs can be defined simultaneously and have no relation to ON OFF pairs This pop up allows the user to review change aspects of the many runs before the block is 121 C 2 Using the VAC defined as a whole Selecting Add new adds another auto run to the list retaining by default all the run parameters currently displayed If the start set in offset times are used each new run added will be given the minimum value required for it to not overlap with the previous run It is not recommended that you decrease this value the 2 minute time between runs is necessary to guarantee they do not cause conflict 17 Define auto runs only Initializes the auto runs The timers to auto start the run s are initiated and the runs are defined as necessary If the run was not set to start immediately it is defined and prepared 1 minute before it is scheduled to start 18 Define Prepare manual runs only Defines the currently described run in the database and carries out the steps needed to prepare the run to be started Preparing takes care of various tasks in the subsystems to ensure the run will be ready to start when the Start Run button is clicked Once a run has been defined using the VAC the next time you try to define a run it will default to the values of the previous run This makes it easier to do multiple identical runs and so you don t have to reenter the observers each time you define a ne
133. r LES 1959 650 in Summer 2002 Krawczynski et al 2004 The second involved Mrk 421 in December 2002 and January 2003 Rebillot et al 2006 Both campaigns combined data from the radio optical X ray and y ray bands In particular analysis of the X ray data is discussed in Section A 2 Section A 3 covers the phenomenon of orphan flares discovered for both sources during their campaigns A 2 RXTE Data Launched in 1995 the Rossi X Ray Timing Explorer RXTE is a satellite designed to observe fast moving X rays passing near Earth It consists of three individual in struments the All Sky Monitor ASM Levine et al 1996 the Proportional Counter Array PCA Jahoda et al 2006 and the High Energy X Ray Timing Experiment HEXTE Rothschild et al 1998 While some data from the ASM were used the analysis concentrated on data from the PCA The 15 250 keV HEXTE data were not used due to their poor signal to noise ratio Due to their close proximity in time analysis of data for both multiwavelength campaigns were nearly identical The X ray analyses was based on the 3 25 keV data from the PCA Standard 2 mode PCA data gathered with the top layer of the operational proportional counter 100 A 2 RXTE Data units PCUs were analyzed The number of PCUs operational during a pointing varied between two and four After applying the standard screening criteria and removing by hand abnormal data spikes the net exposur
134. r combinations listed in Table 5 1 The SEDs are shown in the observer frame 79 5 4 Generating Data Sets T TTTITI L LLLI 1e 46 1e 44 gt L mM E as E 1e 42 E 3 keV J 8 25 keV M 1e 40 1 TeV ly 0 5 10 15 20 25 30 time days FIGURE 5 3 Sample light curves 3 keV 25 keV and 1 TeV from run 14 of the model parameter set D Close examination shows that the 25 keV fluxes lead the 3 keV fluxes and the 1 TeV fluxes The light curves are shown in the observer frame 80 5 5 Analysis Procedure the emission volume In the following we will discuss the time lag behavior more quantitatively For each of the six different magnetic field values we simulate 20 different artificial data sets We apply a DCF analysis to each of the 20 data sets allowing us to study the statistical distribution of the derived parameters 5 5 Analysis Procedure For each simulation we bin the 3 keV and 25 keV lightcurves into 15 min bins observer frame Based on the binned flux values f3 t and fos t we compute the DCF as a function of time lags 7 that are multiples of 1 hr DCF r uc Y DCF r 5 5 where the sum runs over all M 7 pairs of binned flux values f3 and fas that are separated in time by 7 Here DCF rT is an estimate of the DCF derived from a single pair of fluxes DCF r f3 ti fs fos ti 7 fas 5 6 03 025 with c3
135. re or through the Database Subsystem The user can also add comments and observers to a given run through this window This window can also be brought up by pressing Ctrl I Through the Run Information window the observer can alter aspects of a run in the database Add Comment As with the button in the main window this allows the observer to add 127 C 2 Using the VAC Run Info Run Number bus 4 Run Status manually ended Weather A ej Point Mode parallel sl Run Type observing sl Observing Mode wobble y Trigger Config normal y Source wka21 y Trigger Multiplicity br RA Offset fo Trigger Coincidence ho DEC Offset o 00523599 Config Mask Type RN Wobble Angle o Distance los Comments Author PO Observers Mao sp DB Start Time 2007 01 05 221719 Data Start Time 2007 01 05 22 17 39 DB End Time 2007 01 05 22 24 35 Data End Time 2007 01 05 22 24 33 Run Duration 28 00 Add Comment Add Observer Save New info FIGURE C 3 Layout of the Run Info window 128 C 2 Using the VAC any number of comments each with an associated author to the run This can be done any time before during or after completion of the run Add Observer Any number of observers may be added to each run This can be done any time before during or after completion of the run Defining a new run requires at least one observer to be listed You should add only one observer at a time here for it to be properly
136. resh again The current methods it tries are ending all active runs ending all defined runs ending all prepared runs and aborting those runs in L3 canceling all auto runs and manually entering the DB End Time for any remaining orphan active runs as well as any active runs arrayctl no longer knows about e Load Run List Loads the arrayctl internal run list and fills the Run Information Table below If the table contains any runs the list has already been loaded this function will do nothing Updates to the run list are done automatically there should be no need to choose this item e RE load Run List Empties and reloads the Run Information Table All runs listed are reloaded from the database to ensure their displayed parameters are accurate and up to date This command does NOT clear the arrayctl internal run list e Clear Old Runs Clears the arrayctl internal run list Used when there are too many completed runs in the table for it to be useful Should be called at the start of each night While the internal run list is immediately cleared it may take a second for the table to empty e Run Information Brings up a new window showing all information in the database for a specific run see Fig C 3 A run must be selected in the Run Information Table before you can do this the selected run is the one whose information is displayed Information for runs no longer in the Table can be accessed by changing the Run Number displayed he
137. results These other 90 5 7 Comparing of DCF Time Lags to Expected Results variables included using different bin sizes for the DCF changing the escape times of particles in the SSC code using different doppler factors and assuming that the peak emission comes from various fractions of the peak frequency ve Using just the peaks from the lightcurves to calculate the DCF instead of using the entire curve yielded larger DCF values but also a much larger spread in those values over the 20 runs 5 7 Comparing of DCF Time Lags to Expected Results Various authors have used the maximum of the soft hard X ray DCF to constrain the magnetic field Chiappetti et al 1999 Krawczynski et al 2000 Takahashi et al 2000 Sembay et al 2002 We have used a Synchrotron Self Compton code to check the validity of the standard equation used in this analysis In the framework of our simple model flares are produced by a variation of the rate of accelerated electrons alone the standard approach overestimates the magnetic field by factors of between 1 5 and 6 The DCF searches for a linear correlation between two lightcurves with a constant time offset However differences in cooling times do not produce a constant time offset between the fluxes observed in different bands This has two important effects First as we assume instantaneous electron acceleration up to the highest electron energies phases of rising fluxes do not exhibit any interb
138. s none Telescope Configuration M T1 T4 Author N A T3 M T2 Observers SBH MAO Run Duration 28 E min lo E sec Define 4 Prepare Cancel FIGURE C 2 Layout of the Define Run window is displayed in the message window Most information of major importance is printed to the console as well so it can be reached in case the GUI itself crashes On startup various system information is listed including the program version number CORBA nameserver being used and program ID info Messages ap pear in different colors to try and gain the observer s attention when necessary Management Run Information Table Displays information for all runs currently on the arrayctl internal run list This includes completed active prepared and pending runs as well as auto runs that are counting down to their start For each run a variety of information is displayed in the different columns More detailed information is available through Run Info under the Observer menu Ctrl I The columns cur rently displayed are run status pending prepared active completed auto run 118 C 2 Using the VAC countdown to start different than the run status from the database run num ber source weather duration run type observing mode pointing mode trig ger method config mask multiplicity coincidence and observers The run list in the table is constantly updated so a run defined using another client or simple ui will appear
139. s This method used by different authors in the literature was shown to overestimate the magnetic field by as much as a factor of six Understanding the behavior of properties such as this will allow the breaking of model degeneracies and give insight into the physical processes involved in particle acceleration of blazars xi This work is licensed under the Creative Commons Attribution NonCommercial ShareAlike2 5 License To view a copy of this license visit http creativecommons org licenses by nc sa 2 5 or send a letter to Creative Commons 543 Howard Street 5th Floor San Francisco California 94105 USA SOME RIGHTS RESERVED Chapter 1 Summary of Thesis Work For this thesis I looked at the first true stereo data taken with two of the planned four VERITAS y ray telescopes The April May 2006 dark run provided a large amount of data on the known blazar Markarian Mrk 421 These data were used to verify that the VERITAS system is performing as expected They were also used to determine some loose data cuts to use in the future Also VERITAS s first energy spectrum of this source was derived The results of this initial data analysis help us understand what type of data we may get once the entire telescope system is up and running It will also allow us to tweak our simulations so they more accurately model the telescopes behavior Operating the VERITAS experiment requires software I worked to design and 1 VERI
140. s from VERITAS Acceptance Q Factor 2 5 FIGURE 4 11 Plot of the Q factor as 2 a function of the applied 6 cut solid black line right axis Also plotted E are the signal acceptance blue dashed line and background acceptance red i dot dashed line 0 5 Ee E 05 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 d cut in theta degree much about the telescopes including quality of pointing and mirror alilgnment re mains uncertain Further study will allow us to understand these behaviors and improve future simulations 4 6 2 Cutting on Mean Scaled Width The data are now processed through the custom ROOT script again From here we can optimize the cut on MSCW Data that pass the 0 cut will be placed into two separate histograms ON and OFF based on the reconstructed origin of the primary ray For this study the ON region is a circle of radius 0 22 centered on the source The OFF region is an annulus of inner radius 0 4 The complete histograms of MSCW for both the ON and OFF regions are shown in Figure 4 12 The Q factor is then calculated on the residual histogram Figure 4 13 plots the Q factor as well as signal and background acceptances versus the M SCW cut This 67 4 6 First Stereo Results from VERITAS E 7000 of 6000 5000 FIGURE 4 12 Plots of M SCW for us the Mrk 421 data for ON data solid E line and OFF data dotted line 3000 Here a loose cut of 0
141. s its energy spectrum is well known Crab observations can be used to determine the sensitivity of the new system and to test the absolute flux calibration 95 6 2 VERITAS Performance a sigma 50 hours gt 10 events _ CELESTE STACEE MAGIC Integral flux photons cm s Cerenkov detectors in operation Past experiments i Future experiments Akio Morel 5404 Photon Energy GeV FIGURE 6 1 Sensitivity of major experiments in high energy astrophysics Figure from Morselli 2003 The success of the two telescope VERITAS system paves the way for the full array of four telescopes set to come online by January 2007 The addition of two more telescopes along with requiring three or even four telescopes triggering to record an event will increase the array s sensitivity by an additional factor of 2 Figure 6 1 shows how the planned full VERITAS sensitivity compares with other high energy experiments The number of sources detected in the Northern Hemisphere as a result should be similar to the overwhelming success of the H E S S array in the South ern Hemisphere This multitude of y ray sources to study should reveal new and interesting physics as well as expand our current knowledge of existing sources 96 6 3 The Future of y ray Astrophysics 6 3 The Future of y ray Astrophysics This is a very exciting time in y ray astrophysics Several new telescopes are poised to come online in 2007 I
142. s of solar detectors Orig inally proposed by Danaher et al 1982 several groups such as STACEE Gingrich et al 2005 and CELESTE Smith et al 2006 have since implemented the tech nique Cerenkov radiation is reflected off the large solar mirrors and focused onto photomultiplier tubes PMTs The shower direction is inferred from the arrival 12 2 2 Instruments to Detect y rays FIGURE 2 4 The four H E S S telescopes located in Namibia Africa are an example of an array of IACTs 13 2 3 TeV y ray Sources times of the light from each of the solar panels Due to their large collection area these telescopes have lower operating energies than single IACT s Particle Air Shower Arrays Though originally built to study the properties of cosmic rays Ter Haar 1950 particle air shower arrays can be used to study y rays in the TeV PeV energy regime The most successful such detector is MILAGRO located in New Mexico USA which operates at 1 TeV Dingus et al 2000 Inside a large pool of water are 723 PMTs used to detect residual particles from air showers However this type of detector achieves a limited separation between charged cosmic rays and y rays see e g Catanese and Weekes 1999 2 3 TeV y ray Sources Just as there is not one mechanism to detect the whole range of y rays there are also many types of sources from which these rays can originate The number of sources detected has sharply incr
143. se plots show typical distributions for most parameters Under ideal conditions the max pixel histogram far left second from the bottom should not contain any prominent spikes T1 is plotted in black T2 is in red 108 B 2 Analysis and Results 500 sec 14 16 18 vraies PEE ERA IS 0 0 005 0 01 0 015 0 02 0 025 0 03 0 035 0 04 0 045 0 05 02 04 06 08 1 12 index of max ft 5000 4000 3000 2000 1000 d sec 90 degrees 80 eds 550 22600 22650 22700 22750 22800 22850 22900 10 20 30 40 50 60 70 events per time default bin size o 5 4 5 relative gain D 1 15 2 25 3 35 05 ntubes zoomed Ss 3 e S E ei E ME 3 3 E H E P E SS lo B s E ba do E la 3 ze SS 45 50 i 35 40 mean pedestal dc M 30 EM L 25 0 0 0003 00 0 00 150 002 00250 00 00350 0040 00450 005 10 15 20 01 02 03 04 05 06 07 08 09 El 5 E amp i SS SS SS S ak 2 FIGURE B 2 Sample of daily data quality monitoring plots These plots show atypical distributions for many parameters T1 is plotted in black T2 is in red 109 B 3 The dt Bump B 3 The di Bump During the early course of daily data quality monitoring it was noticed that there was a slight bump in the plot of dt the time between consecutive events This curve should be zero at very sh
144. sen N Meyer H Mirzoyan R Moller H Moralejo A Padilla L Panter M Petry D Plaga R Prahl J Prosch C Rauterberg G Rhode W R hring A Sahakian V Samorski M Sanchez J A Schmele D Stamm W Ulrich M Volk H J Westerhoff S Wiebel Sooth B Wiedner C A Willmer M and Wirth H 1997 Astroparticle Physics 8 1 Davies J M and Cotton E S 1957 Solar Energy 1 16 Dingus B L Atkins R Benbow W Berley D Chen M L Coyne D G Dorfan D E Ellsworth R W Evans D Falcone A Fleysher L Fleysher R Gisler G Goodman J A Haines T J Hoffman C M Hugenberber Kelley L A Leonor I McConnell M McCullough J F McEnery J E Miller R S Mincer A L Morales M F Nemethy P Ryan J M Shen B Shoup A Sinnis C Smith A J Sullivan G W Tumer T Wang K Wascko M O Westerhoff S Williams D A and Yang T 2000 in M L McConnell and J M Ryan eds American Institute of Physics Conference Series pp 642 4 Edelson R A and Krolik J H 1988 ApJ 333 646 Fenimore E E Laros J G Klebesadel R W Stockdale R E and Kane S R 1982 in R E Lingenfelter H S Hudson and D M Worall eds AIP Conf Proc 77 Gamma Ray Transients and Related Astrophysical Phenomena pp 201 209 Fichtel C E Bertsch D L Hartman R C Hunter S D Kanbach G Kniffen D A Kwok P
145. sk status immediately you can select this item This can also be accomplished by pressing Ctr1 Ss Update Status Choosing this item manually calls the update of all status information This is usually handled automatically by a separate thread and shouldn t have to be called on its own The update includes all System Status information as well as info for the Current Active Run This can also be accomplished by pressing Ctrl U Define Run Opens the Define Run panel same as the Define Run Button Prepare Run Runs must be both defined and prepared before they can be started Preparing runs usually takes place automatically If there is a problem with this you may select this option to manually prepare a run A run must be selected in the Run Information Table in order to do this the selected run is the one that will be prepared Start Activate Run Allows the observer to Start a prepared run or Activate an auto run Same as the Start Activate Run button This can also be accomplished by pressing Ctrl1 A 126 C 2 Using the VAC e End Cancel Run Allows the observer to End an active run or Cancel an auto run Same as the End Cancel Run button This can also be accomplished by pressing Ctr1 K e Kill All Open Runs If you run into problems defining new runs and can t seem to figure out why select this option It goes through a variety of methods to try and clear ALL runs from ALL telescopes in the system allowing you to start f
146. solute flux calibration has yet to be done but by comparison to the Crab rate of 6y min derived by Celik 2007 we estimate a variable flux for Mrk 421 on the order of 1 2 that of the Crab This high variability can be seen between runs taken on the same night even though Mrk 421 was not in its most active state Such variability over short time scales was evident in numerous observations in the past with the Whipple 10 m telescope The improved sensitivity of VERITAS will ultimately provide much better measurements of this variability on all time scales 69 4 6 First Stereo Results from VERITAS 107 a DS Flux arbitrary normalization T T ARAN e SS 10 L L L L L L L 1 L L L i L L L L L L L L L 1 1 0 4 0 2 0 0 2 0 4 0 6 0 8 1 log Energy ST E FIGURE 4 15 Reconstructed energy spectrum of the blazar Mrk 421 from the first VERITAS stereo data 4 6 4 Energy Spectrum In order to get a more accurate energy spectrum only ON OFF pairs were used in this process Wobble and Tracking mode data were ignored Due to the high significance of these runs the nine pairs of data during this dark run were enough to generate a spectrum with small statistical errors Data were fit to the Monte Carlo energy histogram as described in Section 4 3 This resulted in the energy spectrum shown in Figure 4 15 Based on our preliminary analysis we get a spectral index of T 2 26 0 06 For the Crab
147. t muons and hadrons respectively Images courtesy of F Schmidt CORSIKA Shower Images http www ast leeds ac uk fs showerimages html 36 3 3 VERITAS Very Energetic Radiation Imaging Telescope Array System 3 3 VERITAS Very Energetic Radiation Imaging Telescope Array System VERITAS an array of four IACTs is currently being built at the base camp of the Whipple Observatory in southern Arizona It is designed as a successor to the previous Whipple 10 m telescope still in operation on Mt Hopkins Being a next generation telescope VERITAS improves over Whipple in sensitivity and background rejection Weekes et al 2002 3 3 1 Telescope Array VERITAS is not just one but several y ray telescopes that act together with a single trigger to greatly increase background rejection in the data Originally planned as a grouping of seven identical telescopes budget cuts required scaling back to just four Due to several factors outside of the control of the VERITAS Collaboration the telescopes are initially being built on a temporary site at the Whipple Observatory base camp This put many restrictions on the construction process so that the op timal telescope configuration could not be obtained Originally the four telescope arrangement would have had one telescope in the center and the other three equidis tant from each other in a ring around the central telescope The current configuration resembles a trapezoid see Fig
148. the Close button will bring up the info for the next run Get Run Info Opens a new window with all the database information for the run number entered above If the run does not exist most of the info will be left blank 138 C 2 Using the VAC Run info can be altered in the database in this fashion However it is not recommended you use this method for runs currently listed in the main Run Info Table They can be modified by choosing Run Info from the main window through the Observer menu Put DB Start Time Manually enter the current date time in the DB Start Time field for the run number listed above Put DB Stop Time Manually enter the current date time in the DB Stop Time field for the run number listed above This causes the database to end the given run Put Data Start Time Manually enter the current date time in the Data Start Time field for the run number listed above Put Data Stop Time Manually enter the current date time in the Data Stop Time field for the run number listed above Charge Injection QI Subsystem This subsystem is currently unsupported Custom Night For debugging purposes each part of Start End Night can be called individually or in custom groupings The window for handling this is shown in Figure C 9 Start Night and End Night appear as separate tabs each listing the functions specific to each procedure Regardless of which items are checked in the other tab only the ite
149. the Start Run button After being defined the run should ve automatically been selected in the Run Information Table If it was not you need to select the run in the table before starting the run The run will last for the duration specified when it was defined or until manually terminated by the user by clicking the End Run button If the speakers are properly connected to the computer a sound will also play to alert you to the ending of the run At this point a new run may be defined and started If you would like to take an ON OFF run pair VAC can easily handle that for you See Section C 2 1 for more info on taking pairs and other automatically started runs The modifications to the Define Run sequence are as follows 4 Click Define Run from the main window and fill out the run information as normal but do not define the run quite yet 5 Under Start Run click Automatically start run in Make sure ON OFF is selected under Auto Run Type This sets up the automatic runs 6 Click Define Two runs are automatically created The first will start as soon as it is defined and prepared which happens automatically The second run will begin two minutes after the completion of the first run 7 Kick back and relax while arrayctl automatically defines and starts the run Don t forget to move the telescope in between runs The program may safely be quit by choosing Exit from the File menu or by just closing the main window C 1
150. through any newly reloaded data base configurations Get Singles Scalars Gets the Singles Rates on the current telescope Results are displayed below Get Interrupt Status Gets the interrupt status on the current telescope Results are displayed below Query VME Config Also self descriptive If the Force Query box is checked the call will be made even if a run is active This will adversely effect the current run The query is made of the node listed e Close Closes the Event Builder dialog box and returns to the main window L2 Subsystem The L2 subsystem window is shown in Figure C 7 e Telescope Set which telescope s L2 you wish to deal wih e Multiplicity Choose 3 fold or 4 fold multiplicity 135 C 2 Using the VAC e Load Pattern Triggers Check this box if the pattern triggers need to be reloaded Leave unchecked if you simply are resetting L2 e Init L2 Runs the L2 script with the above options Also makes sure ecc host is running e Enable Expert Mode You should only use this if you really know what you are doing Clicking this button enables the Adjacency option It also allows for any combination 1 5 for Multiplicity and Adjacency as long as Multiplicity gt Adjacency For clarity Multiplicity defines how many pixels must have CFD triggers in order for a patch to trigger Adjacency defines how many pixels must be NEXT to each other for a patch to fire L1 Subsystem Currently not sup
151. ths Inverse Compton scatter ing from the jets relativistic electrons can also produce y rays All but one AGN detected in TeV 4 rays are part of the subclass known as blazars discussed further in Section 2 4 The one exception is the nearby radio galaxy M87 Aharonian et al 2003 2006 Its jet is believed to be at an angle of 30 to the line of sight 19 2 4 Blazars Narrow Line Region Broad Line Region Accretion a Disk A Obscuring Torus 2 4 Blazars FIGURE 2 7 Model for the struc ture of an AGN consisting of a dense central emitting region and including jets of relativistic par ticles The radius of the central black hole is 107 pc while the jets can extend from 107 pc to kpc or even Mpc from the black hole When the jets point towards the Earth the source is known as a blazar Figure from Holt et al 1992 Blazars are a subclass of AGN defined in particular by having their jets orientated along the line of sight towards the Earth see Fig 2 7 This fact makes blazars espe cially interesting in that one can literally see straight down the beam of relativistic particles Due to relativistic boosting blazars are the brightest extragalactic sources in y rays They are also characterized by their flux variability on time scales as short as minutes This variability is strongly correlated across many energy bands Continuum emission from blazars is visible over the entire e
152. ting VAG emirato Sem del metet ang xe Lohan moa fu Ye 113 C 1 1 Normal Operation ads e eS e 113 C 1 2 Debugging Systems eege ee Yee a e 114 62 Using th VAC usc do Rb x os b oaa en 115 C21 Mam d oho A AA a qwe 115 SNE nus cok ee Me E Re DOR RUEDA RE 115 Rur Ea eme me r taa y don ien te efr arde cn t v ied ed 118 Run Info for Current Active Run 123 E23 Rate Plot As dares dde eer em oe PS 124 G22 Observer Menu 240 8 s aate AAA LR 125 52 3 Test Runs Menu coiere EM ERA 129 C 2 4 Subsystems Menu ia sehen ERES RIEN 130 DE DS MSU Gh teg of creat Ee Begreep rtr ad nee 130 Harvester Subsystem 132 Event Builder Subsystem 4 2 0 xel EG e ia 133 L2 Subsystem s a ee unter e Se aces ee aN ao 135 Te STS ye n LL m 136 Database Subsystem 136 Charge Injection QI Subsystem 139 Custom ENEE 139 Chao in A tes He Ta iut E HORIS ee I 141 Put CFD Settings a D aug xs ar a fe Sc VE SU 141 CEL SoEUmgss aeu X ese dt de Sentech one E 28 fon T il eed 142 DEET sod enu S mea e A NR po wee at Ses 142 vi List of Figures 2 1 The Compton Gamma Ray Observatory OGRO 2 2 The Gamma ray Large Area Space Telescope GLAST 2 9 Ehe Whipple 10 m Telescopes wd eG xe X omnc ux KR a SE 2 4 The E Telescopes sce tr dia de da Tere Pee tg 2 9 Sky Map of y ray Sources aia dee y at por e amp ete a 2 6 The Crab Nebula vs 2 gout Ae LER XR B AUR RE 2 7 Schematic of an Active Gala
153. ting plasma and the observer may change with time If the magnetic field is not randomly oriented the mean angle between the magnetic field and the line of sight might also change with time Our two main conclusions from our study and these arguments are the following i the determination of the magnetic field from the DCF peak is more complicated than previously thought and ii the results of a more detailed analysis will depend on the assumptions underlying the calculations Reliable estimates of the plasma 92 5 7 Comparing of DCF Time Lags to Expected Results parameters can not be derived from a timing analysis alone but should take advantage of as many observational constraints as possible 93 Chapter 6 Discussion 6 1 Summary of Thesis Results In this thesis we have shown first results from the VERITAS two telescope system The experiment behaves as expected with no surprises in the performance of major hardware components The comparison of Mrk 421 y ray data with Monte Carlo simulations shows excellent agreement overall Simulation of the two telescope system achieves an energy threshold of 220 GeV and angular core and energy resolutions of 0 29 12 5 m and 34 respectively Performance of the three and four telescope system will improve over these two telescope values The analysis presented here gives Mrk 421 fluxes and energy spectra similar to those observed in earlier campaigns Future blazar studies wil
154. und is subtracted from OFF regions For tracking runs left the OFF region is an annulus with the same area as the source region spaced a bit outside of the source region For wobble runs right the OFF region is the same shape and size as the ON region but offset on the other side of the telescope s center radius at the center of the camera The OFF region is then an annulus with the same area as the ON region centered on the source with a buffer zone of space between the two so the source does not contaminate the background region This method is less reliable than doing ON OFF pairs for several reasons Because the OFF region is farther from the camera s center the number of events in this region may be less than what one would expect This could be due in part to these events not triggering both telescopes Also because the centroids would occur at larger distances from the camera center the images could be more spread out introducing a bias in the width distributions In contrast to ON OFF pairs passing clouds would 95 4 4 Mrk 421 Data from April May 2006 effect both the ON and OFF regions in the same way making this behavior a tradeoff between the two methods Wobble Runs Wobble runs have proved an effective observation mode for stereo Cerenkov tele scope systems such as HEGRA Daum et al 1997 and H E S S Aharonian et al 2005c as the detection sensitivity is rather uniform over the central part of th
155. und or OFF region see Sect 4 4 1 for more on these regions The number of events in these regions Non and Norr respectively can be used to calculate the ray rate as well as its associated error Non Norr m v Non Norr time time pd hp 4 2 A large positive excess of at least the 5o level means the source has been detected This statistical significance is calculated as follows AR POP E e ANON Norr 4 3 Ar VNon Norr 4 3 Spectral Reconstruction Spectral reconstruction of the Mrk 421 data was performed using the forward folding method of Henric Krawczynski described fully in Rebillot et al 2006 It has been incorporated into the program wufit part of Washington University s own data analysis package wuparam Energies are reconstructed for every event in both the ON and OFF regions as well as for all Monte Carlo simulated events as described above The energies are histogrammed and the OFF energy histogram is subtracted from the ON energies to yield the energy excess The energy spectrum can then be fitted to the excess histogram with help from the Monte Carlo simulated energy histogram using the forward folding approach see e g Fenimore et al 1982 52 4 4 Mrk 421 Data from April May 2006 The data are fit using a power law model dN _ 4g No E 1 TeV 4 4 where Nj is the flux normalization at 1 TeV and I is the photon index For each trial parameter set No an
156. w individual events on each camera as well as individual FADC traces and other timing information For each night of data a calibration must be done to correct for minor fluctuations in weather and telescope electronics This is done through the use of a laser run At the start of each night a short run is taken where a laser is flashed at the PMT camera The constant flux of these pulses is used to adjust the analysis program to the subtle differences in how each pixel detects the light intensity as well as timing After the calibration is done each run for the night may be analyzed and param eterized Hillas parameters are calculated for each telescope and the stereo data is reconstructed to find the shower axis see Sect 4 1 1 Afterwards the mean scaled width mean scaled length and energy are reconstructed 4 2 Data Cuts and Significances After the events are fully parameterized cuts are applied to separate actual y rays from background events Once VERITAS is brought fully online and more data have been acquired and analyzed a standard set of cuts will become available Since the standard cuts have yet to be established we present preliminary work to establish cuts in both MSCW and 6 discussed in Section 4 6 To determine the excess one separates the events passing all data cuts into two categories events from the signal or ON region and events from a comparison 51 4 3 Spectral Reconstruction backgro
157. w run However if the GUI crashes or quits the default run info is lost and all information must be reentered for the next run defined Also auto run information is not saved Only the most recent run is listed individually and auto start options must be reselected Add Comment Any number of comments may be added to each run with an author associated with each comment This can be done any time before during or after completion of the run A run must be selected in the Run Information Table before you can add a comment the selected run is the one to which the new comment is added End Cancel Run If circumstances require a run to be terminated before it has executed for its full duration it can be ended manually A run must be selected in the Run Information Table before you can do this the selected run is the one ended The user is asked for confirmation before the run actually ended Auto runs can also be cancelled by clicking this button If the auto run has already been defined you must also click End Run for that run as well Canceling the auto run only disables the auto starting of the run not the run itself if the run has already been defined Start Activate Run Tells arrayctl to start a given run A run must be selected in the Run Information Table before you can do this the selected run is the one started The run will then execute for the prescribed duration or until cancelled While active its status information will be aut
158. window multiplicity pedestal rate L3 rate Total deadtime and L3 deadtime Clicking the Next Run button displays the information for the a different run currently known to L3 The button is greyed when there are less than two runs present If no runs are present all values are 1 Telescope Status The status items displayed are as follows telescope ID L2 rate QI rate high multiplicity rate new physics rate VDAQ deadtime L3 output rate and L3 L2 output rate 131 C 2 Using the VAC Run 11183 Start Harvest Run End Harvest Run Roll Back Run RolBackRun FiGURE C 5 Layout of the Harvester subsystem window Start Harvest Night End Harvest Night Close Close Closes the L3 dialog box and returns to the main window Harvester Subsystem The Harvester subsystem window is shown in Figure C 5 Run The run number used when a run specific command is selected If a run is selected in the main Run Information Table it becomes the default run number Start Harvest Run Starts a run in the harvester The run number is taken from the input line above End Harvest Run Ends a run in the harvester The run number is taken from the input line above Roll Back Run Roll back a run in the harvester The run number is taken from the input line above Start Harvest Night Starts the night for the harvester initializing internal routines 132 C 2 Using the VAC End Harvest Night Ends the nig

Observations of the Active Galactic Nuleus Markarian 421

Contents

Download Pdf Manuals

Related Search

Related Contents