Transcription

Draft version August 29, 2020Typeset using LATEX twocolumn style in AASTeX63Optimizing the Next GBT Pulsar SurveyG. Y. Agazie,1, 2 R. S. Lynch,2 T. Cohen,3 and J. K. Swiggum41 Departmentof Physics and Astronomy, West Virginia University, Morgantown, WV 26506, USA2 Green Bank Observatory, Green Bank, WV 24494, USA3 New Mexico Institute of Mining and Technology, Socorro, NM4 Lafayette College, Easton, PA, United StatesAbstractLarge-area pulsar surveys are important tools for increasing the population of known pulsars, whichare used to study various extreme physical phenomena. As the current large-scale survey on theRobert C. Bryd Green Bank Telescope (GBT), the Green Bank North Celestial Cap (GBNCC) surveyapproaches completion, it is important to begin planning next-generation pulsar surveys that arecompetitive with current and upcoming pulsar surveys conducted on other instruments. We present theresults of simulated pulsar population and survey studies conducted for potential GBT pulsar surveysusing the 820 MHz, L-Band, S-Band, Ultra-wide band, and the Focal L-Band Array for the GBT(FLAG) receivers. We determined that the most effective survey would be conducted at frequenciesof 1.1 to 1.9 GHz using the GBT L-Band receiver at an dwell time of 180 s over a 1700 square degreeregion along the galactic plane. Such a survey would detect approximately 500 new pulsars, about 50of which would be MSPs. We also compare these results to a possible survey with the FLAG receiverover the same region with a 1260 s dwell time, which is projected to discover 550 pulsars.the most results for a reasonable amount of telescope1. INTRODUCTIONtime. In section 4, we show the results of our analyses,Pulsars are rapidly spinning neutron stars with powand in section 5, discuss how our results support ourerful magnetic fields and emit beams of radiation fromchosen next-generation survey.each magnetic pole (Lorimer 2008). In astronomy, pulsars have been used to study extreme states of matter, general relativity, and led to the discovery of thefirst extra-solar planets (Cromartie et al. 2020; Krameret al. 2006; Wolszczan & Frail 1992). Millisecond pulsars(MSPs) are used in pulsar timing arrays, where correlated fluctuations in pulse arrival times are studied withthe goal of detecting stochastic background gravitationalwaves (Arzoumanian et al. 2018). There are currentlyover 2,800 known pulsars, roughly 8% of which are MSPs(ATNF Pulsar Catalogue v1.63; Manchester et al. 2005).The current large-area survey being conducted on theGBT, the GBNCC survey, is nearing completion, whichmotivates making plans to determine where the nextGBT survey should be focused in order to maximize thenumber of new pulsars, in particular MSPs (McEwenet al. 2020). In this paper we present the results of simulated pulsar surveys to show that the next GBT pulsarsurvey should be conducted using the L-Band receiverwith an dwell time of 180s. In section 2, we present information on several current large area pulsar surveysbeing conducted on other telescopes. In section 3, weoutline our methods of simulating pulsar populationsand determining which type of survey would produce2. CURRENT SURVEYS2.1. Green Bank North Celestial Cap SurveyThe GBNCC pulsar survey covers the entire sky abovedeclination (δ) 40 at 350 MHz with a 100 MHz bandwidth, and is conducted on the GBT (Stovall et al. 2014;McEwen et al. 2020). To date, the survey has discovered 161 new pulsars, including 25 MSPs and 24 rotatingradio transients (RRATs) (McEwen et al. 2020). Datacollection began in 2011 and is expected to be completedwithin the next few years.2.2. The Arecibo 327 MHz Drift-Scan SurveyThe Arecibo 327 MHz Drift-Scan (AODRIFT) Surveyis a 327 MHz drift-scan survey conducted on the AreciboTelescope from δ 1 to 38 (Martinez et al. 2019). Datacollection began in 2011 and to date, the survey hasdiscovered 85 new pulsars, including 16 recycled pulsars,10 MSPs, and 16 RRATs (Martinez et al. 2019).2.3. The Commensal Radio Astronomy FAST SurveyThe Commensal Radio Astronomy FAST Survey(CRAFTS) is a drift-scan survey conducted on theFive-hundred-meter Aperture Spherical radio Telescope

2(FAST) using a frequency bandwidth of 270-1620 MHzat δ 14 to 60 (Qian et al. 2019). Data collectionbegan in 2017 (Qian et al. 2019).2.4. The Pulsar Arecibo L-Band Feed Array SurveyThe Pulsar Arecibo L-Band Feed Array (PALFA) survey is an L-Band survey centered at 1375.5 MHz conducted on the Arecibo telescope. The survey has a lowgalactic longitude ( ) region from 32 to 77 and a high region from 168 to 214 (Cordes et al. 2006). Datacollection began in 2004 and the survey has discovered196 new pulsars, including 31 MSPs and RRATs1 .2.5. The High Time Resolution Universe PulsarSurveyThe High Time Resolution Universe (HTRU) Pulsarsurvey is split into a northern (HTRU-N) and southern(HTRU-S) sky survey. HTRU-N. HTRU-N is conductedon the Effelsberg telescope at 1360 MHz with 240 MHzbandwidth, and HTRU-S is conducted on the Parkestelescope and centered on 1352 MHz with a 340 MHzbandwidth (Barr et al. 2013; Keith et al. 2010). Bothhave three survey regions, a low galactic latitude (b)region with b 3.5 and 80 30 , a midlatitude region with b 15 and 120 30 ,and a high-latitude region of b 15 (Keith et al.2010). To date, HTRU-S has discovered over 100 newpulsars, including 9 MSPs2 .3. METHODSGalactic pulsar populations were simulated using PsrPopPy, a python-based software package used to simulate pulsar populations and surveys (Bates et al. 2014).Models of the galactic pulsar population were generatedusing PsrPopPy populate function and normalized tomatch the results of the Parkes Multibeam Survey, asit is currently the largest and most successful large areapulsar survey to date (Stairs et al. 1999; Bates et al.2014). To account for random fluctuations within a particular simulated population we averaged all results over100 pulsar population models. For each realization ofthe galactic pulsar population we generated a canonicalpulsar population with a mean spin period of 500 msand an MSP population with a mean spin period of 30ms. This accounted for limitations within PsrPopPy ingenerating a diverse pulsar population.We looked at potential future surveys using the following GBT recievers: 820 MHz (0.68-0.92 GHz), L-Band(1.15-1.73 GHz), S-Band (1.73-2.6 GHz)3 , Ultra-wideband (UWB) (0.7-4 GHz)4 , and the Focal L-Band Array for the GBT (FLAG) (1.365-1.515 GHz) (Rajwadeet al. 2019). For each receiver we used dwell times of 60s, 120 s, 180 s, 300 s, 600 s, and 1800 s. Detailed survey parameters are listed in Table 1. We compared thefuture surveys at each dwell time to the following extisting large scale surveys: GBNCC, AODRIFT, CRAFTS,PALFA, and HTRU. We simulated each survey on themodel populations using the PsrPopPy dosurvey function. For each galactic population we assigned eachpulsar to the survey that detected it with the highestS/N and created sky maps for each survey of all thepulsars the survey detected with the highest S/N (seeFigure 2). This measure of survey sensitivity helped determine which type of future survey would be the mostsensitive and accounts for regions of the galaxy with themost pulsars. This allowed us to optimize each surveyfor regions that would yield the most pulsars. In Figure 1, we plotted the period vs DM of pulsars detectedin a single pulsar model, where a 180 s integration timewas used for each GBT future survey with colors indicating the survey which detected a particular pulsarwith the highest S/N.Our first step was to define the best survey regions.This was done by dividing the sky into bins and calculating which survey detected the most pulsars withthe highest S/N in each sky bin. We saw three distinctregions where the GBT L-Band receiver detected themost pulsars (see Figure 3) and calculated the total survey area from which we determined the total numberof observing hours needed for each dwell time. Surveyduration was calculated by dividing the survey area bythe angular size of the L-Band receiver and multiplyingby the dwell time per pointing if we assume no overlapbetween pointings. If a half width half max overlap isassumed, this increases the survey duration by a factorof 2.25, and quarter width half max separation causes a1.5 factor increase in survey duration.The next step was to predict the total number of newpulsars within our newly defined survey region. Wereran surveys over the newly defined survey regions andsubtracted any pulsars within the new survey regionthat were detected by any of the current surveys. Toaccount for pulsars that have already been discovered,we determined the number of known pulsars listed inthe ATNF pulsar catalog within the survey region thatwere not discovered by any of the current surveys listed31http://www.naic.edu/ ry.org/science/instruments-20202030

3Dwell Time (s)120180300600in Table 2, and subtracted this from the total pulsarcount.4. RESULTS4.1. Survey TypeIn Figure 2, we have sky maps of all the pulsars a particular survey detected with the highest S/N. In this figure we have used 180 s dwell times for the future survey,and indicated the number of pulsars in each subplot inthe subplot title. Maps for all the other dwell times usedto study the future surveys can be found in AppendixA. The GBT L-Band survey discovered 565 pulsars witha higher S/N than all the other current and future surveys. The majority of pulsars appear to be clustered inregions of the Galactic plane that the FAST and Arecibotelescopes cannot observe, so a survey focused on the region L-Band is sensitive to would not be competing withthe FAST or Arecibo telescopes to find pulsars. Thesesky maps do not take into account discoveries that havealready been made by the current surveys so the number of pulsars assigned to L-Band does not represent thetrue expected yield from an L-Band survey.Area (deg2 )1191171219542397Duration (hr)22474845921522611Table 3. Best survey region area and duration for the LBand receiver for four dwell times. We have assumed nooverlap between pointings.For a 180 s dwell time L-Band survey, we wouldneed approximately 4845 observing hours if we do notuse overlapping pointings. For a half-width half maxoverlap, the survey would need 10,902 hours and for aquarter-width half max overlap, 7268 hours. We dividedthe survey area into three sub-regions which are definedin Table 4.Regionlow mid high range 126 to 81 57 to 18 100 to 137 b range 9 to 6 20 to 19 5 to 9 δ range 45 to 12 45 to 12 60 Table 4. Survey regions defined for the GBT L-Band surveywith a 180 s dwell time.4.3. Survey Yield4.2. Survey RegionIn Figure 3, we see that there are three regions inthe galactic plane where the GBT L-Band receiver, labeled in pink, detects more pulsars with the highest S/Nthan any of the other current and future surveys. Forease of reference we have over plotted lines of constantδ of 45 , 12 , and 60 . 45 is the lowest δ thatthe GBT can observe and 12 is near the lowest declination observable by FAST (actual limit is 14 ). Thislimit appears to be dwell time dependent, since at a highdwell time of 1800s, the L-Band survey region extendsto cover most of the galactic plane (see Figure 13).Since higher dwell times had higher sensitivity, we created maps for each dwell time and calculated the totalsurvey area per dwell time (see Appendix B and Table 3). Due to the very small survey region at 60 s adwell time, we excluded it from this analysis. We alsoexcluded the 1800 s dwell time as a large-area surveywith such a long dwell time would take more telescopetime than a survey could reasonably be awarded. Weconsidered a reasonable amount of telescope time to beless than the GBNCC total observing time of 6000hours.In Table 5, we have listed the survey yield for an LBand survey at dwell times of 120 s, 180 s, 300 s, and600s. For a 180 s dwell time L-Band survey we would expect to discover about 500 new pulsars including roughly50 MSPs. In comparison, the GBNCC survey has discovered 161 new pulsars, 25 of which are MSPs.Dwell time (s)120180300600MSP334973126Part. 1220Table 5. The number of MSPs, partially recycled, canonical, and total pulsars detected by the L-Band receiver at fourdwell times. MSPs are categorized as pulsars with spin periods less than 15 ms, partially recycled are spin periods between 15 ms and 200 ms, and canonical pulsars are those withspin periods above 200 ms. These numbers have been averaged over 100 pulsar populations and corrected for pulsarsfound by current surveys and known pulsars in the ATNFcatalog.5. DISCUSSIONTaking into account survey yields and duration, a 180s L-Band survey would be an ideal candidate for theGBT large area survey.

4Figure 1. Spin period vs DM for all pulsars in a single galactic population detected by a current or future survey. The colorrepresents which survey detected the pulsar with the highest S/N.ReceiverGain (K/Jy)Sampling time (ms)Tsys (K)Central Frequency (MHz)Frequency Bandwidth (MHz)Channel Bandwidth (MHz)FWHM (0 )820 MHz2.00.08192468202000.024415L 1500.30323.81UWB2.00.010920235015000.1839S Band2.00.020242221658000.04885.8Table 1. Table of survey parameters used for possible future GBT surveys in PsrPopPy to detect simulated pulsars. Each wasrun with dwell times of 60 s, 120 s, 180 s, 300 s, 600 s, and 1800 s.SurveyIntegratation Time (s)Gain (K/Jy)Sampling Time (ms)Tsys (K)Center Frequency (MHz)Bandwidth (MHz)Channel Bandwidth (MHz)FWHM (0 )δ range ( ) range ( )b range ( 14–60-180–1800–90PALFA268, 1808.50.064251375.5322.60.366043.60–3832–77, 168–2140–5HTRU-N & HTRU-S1500, 180, 90 & 4300, 540, 2701.5 & 0.60.05941 & 0.06421 & 251360 & 1352240 & 3400.5859 & 0.399.6 & 14-20–90 & -90–10-180–180 & -80–30, -120–30, -180–1800–3.5, 0–15, 15–90 for both N & STable 2. Table of existing survey parameters used in PsrPopPy to detect simulated pulsars. For the purpose of determiningoverall survey yield, HTRU North and South results were combined.

5Figure 2. Skymaps of pulsars detected with the highest S/N by each current survey and each future survey using a 180s dwelltime for the future surveys. Color scales indicate the number of pulsars found in a particular sky bin, and the total number ofpulsars per map is indicated in each title. These maps have been averaged over 100 galactic populations.Figure 3. Skymap in which each bin color indicates the survey that detected the most pulsars averaged over 100 galacticpopulations. The blue, orange, and green lines indicate constant δ of 45 , 12 , and 60 respectively. For this map, allpossible future GBT surveys use a 180 s dwell time. Regions labelled as ’None’ detected less than one pulsar for any of thesurveys after being averaged.

65.1. Comparisons with GBT FLAG receiver5.2. Future StudiesThe GBT FLAG receiver covers a similar frequencyrange to the L-Band receiver, although with a muchsmaller bandwidth and higher system temperature. Asseen in our dwell time maps, the FLAG receiver suffersfrom significantly lower sensitivity, however it has a fieldof view (FOV) seven times larger than the L-Band receiver and thus can significantly drive down the numberof pointings needed to cover the same area as the LBand receiver. Also, the larger field of view means thatthere is relatively uniform sensitivity across the area of asingle beam. Seven L-Band beams across the same areaas a single FLAG beam would have reduced sensitivityat the edges of each beam.If we use a longer dwell time, this can help account forthe reduced sensitivity of FLAG. Looking at the 180 s LBand survey area, if we were to use seven times the dwelltime on a FLAG survey (1260 s) we would use the sameamount of overall observing hours and detect about 50more pulsars than at L-Band (Table 6). The longer dwelltime would mean that we were more sensitive to nullingpulsars and RRATs, but less sensitive to short-orbitalperiod binaries.Further studies into this project would include examining potential timing precision of the MSPs discoveredin a single pulsar population to determine how manyPTA candidates might be expected from an L-Band survey. We would also run population studies on the pulsars discovered by L-Band vs FLAG in order to studythe impact of the longer FLAG dwell time on the typesof pulsars discovered.Dwell Time (s)180 L-Band1260 FLAGDuration (hr)48454845Number of Pulsars498552Table 6. Survey duration and yield for L-Band at 180 s andFLAG at 1260 s over the survey area defined for L-Band at180 s.ACKNOWLEDGMENTSThis work is supported by the Green Bank Observatory which is a major facility funded by the NationalScience Foundation operated by Associated Universities,Inc.This project was conducted using the python packagePsrPopPy described in Bates et al. (2014) and the version used by the authors of this paper can be found Arzoumanian, Z., Baker, P. T., Brazier, A., et al. 2018,ApJ, 859, 47, doi: 10.3847/1538-4357/aabd3bBarr, E. D., Champion, D. J., Kramer, M., et al. 2013,MNRAS, 435, 2234, doi: 10.1093/mnras/stt1440Bates, S. D., Lorimer, D. R., Rane, A., & Swiggum, J.2014, MNRAS, 439, 2893, doi: 10.1093/mnras/stu157Cordes, J. M., Freire, P. C. C., Lorimer, D. R., et al. 2006,ApJ, 637, 446, doi: 10.1086/498335Cromartie, H. T., Fonseca, E., Ransom, S. M., et al. 2020,Nature Astronomy, 4, 72, doi: 10.1038/s41550-019-0880-2Keith, M. J., Jameson, A., van Straten, W., et al. 2010,MNRAS, 409, 619, doi: 10.1111/j.1365-2966.2010.17325.xKramer, M., Stairs, I. H., Manchester, R. N., et al. 2006,Science, 314, 97, doi: 10.1126/science.1132305Lorimer, D. R. 2008, Living Reviews in Relativity, 11, 8,doi: 10.12942/lrr-2008-8Manchester, R. N., Hobbs, G. B., Teoh, A., & Hobbs, M.2005, VizieR Online Data Catalog, VII/245Martinez, J. G., Gentile, P., Freire, P. C. C., et al. 2019,ApJ, 881, 166, doi: 10.3847/1538-4357/ab2877McEwen, A. E., Spiewak, R., Swiggum, J. K., et al. 2020,ApJ, 892, 76, doi: 10.3847/1538-4357/ab75e2Qian, L., Pan, Z., Li, D., et al. 2019, Science China Physics,Mechanics, and Astronomy, 62, 959508,doi: 10.1007/s11433-018-9354-yRajwade, K. M., Agarwal, D., Lorimer, D. R., et al. 2019,MNRAS, 489, 1709, doi: 10.1093/mnras/stz2207Stairs, I. H., Lyne, A. G., Camilo, F., et al. 1999, arXive-prints, astro. https://arxiv.org/abs/astro-ph/9903290Stovall, K., Lynch, R. S., Ransom, S. M., et al. 2014, TheAstrophysical Journal, 791, 67,doi: 10.1088/0004-637x/791/1/67Wolszczan, A., & Frail, D. A. 1992, Nature, 355, 145,doi: 10.1038/355145a0

7APPENDIXA. DWELL TIME MAPSFigure 4. Skymaps of pulsars detected with the highest S/N by each current survey and each future survey using a 60s dwelltime for the future surveys averaged over 100 galactic populations.Figure 5. Skymaps of pulsars detected with the highest S/N by each current survey and each future survey using a 120s dwelltime for the future surveys averaged over 100 galactic populations.

8Figure 6. Skymaps of pulsars detected with the highest S/N by each current survey and each future survey using a 300s dwelltime for the future surveys averaged over 100 galactic populations.Figure 7. Skymaps of pulsars detected with the highest S/N by each current survey and each future survey using a 600s dwelltime for the future surveys averaged over 100 galactic populations.

9Figure 8. Skymaps of pulsars detected with the highest S/N by each current survey and each future survey using a 600s dwelltime for the future surveys averaged over 100 galactic populations.B. SENSITIVITY MAPSFigure 9. Skymap in which each bin color indicates the survey that detected the most pulsars averaged over 100 galacticpopulations. For this map, all possible future GBT surveys use a 60 s dwell time.

10Figure 10. Skymap in which each bin color indicates the survey that detected the most pulsars averaged over 100 galacticpopulations. The blue, orange, and green lines indicate constant δ of 45 , 12 , and 60 respectively. For this map, allpossible future GBT surveys use a 120 s dwell time.

11Figure 11. Skymap in which each bin color indicates the survey that detected the most pulsars averaged over 100 galacticpopulations. The blue, orange, and green lines indicate constant δ of 45 , 12 , and 60 respectively. For this map, allpossible future GBT surveys use a 300 s dwell time.

12Figure 12. Skymap in which each bin color indicates the survey that detected the most pulsars averaged over 100 galacticpopulations. The blue, orange, and green lines indicate constant δ of 45 , 12 , and 60 respectively. For this map, allpossible future GBT surveys use a 600 s dwell time.

13Figure 13. Skymap in which each bin color indicates the survey that detected the most pulsars averaged over 100 galacticpopulations. For this map, all possible future GBT surveys use a 1800 s dwell time.

Draft version August 29, 2020 Typeset using LATEX twocolumn style in AASTeX63 Optimizing the Next GBT Pulsar Survey G. Y. Agazie,1,2 R. S. Lynch,2 T. Cohen,3 and J. K. Swiggum4 1Department of Physics and Astronomy, West Virginia University, Morgantown, WV 26506, USA 2Green Bank Observatory, Green