Simultaneous dual-frequency radio observations of S5 0716+714: A search for intraday variability with the Korean VLBI Network
Jee Won Lee, Bong Won Sohn, Do-Young Byun, Jeong Ae Lee, Sungsoo S Kim
AAstronomy & Astrophysics manuscript no. Arxiv_0716_jwlee c (cid:13)
ESO 2018November 13, 2018
Simultaneous dual-frequency radio observations of S5 0716+714:A search for intraday variability with the Korean VLBI Network
Jee Won Lee , , Bong Won Sohn , (cid:63) , Do-Young Byun , Jeong Ae Lee , , and Sungsoo S. Kim Korea Astronomy and Space Science Institute 776, Daedeok-daero, Yuseong-gu, Daejeon 34055, Republic of Korea Department of Astronomy and Space Science, Kyung Hee University, 1732, Deogyeong-daero, Giheung-gu, Yongin-si, Gyeonggi-do 17104, Republic of Korea Korea University of Science and Technology, 217, Gajeong-ro, Yuseong-gu, Daejeon 34113, Republic of KoreaAccepted
ABSTRACT
This study aims to search for the existence of intraday variability (IDV) of BL Lac object S5 0716 +
714 at high radio frequencies forwhich the interstellar scintillation e ff ect is not significant. Using the 21-meter radio telescope of the Korean VLBI Network (KVN),we present results of multi-epoch simultaneous dual-frequency radio observations. Single-dish observations of S5 0716 +
714 weresimultaneously conducted at 21.7 GHz (K-band) and 42.4 GHz (Q-band), with a high cadence of 30-60 minute intervals. We observedfour epochs between December 2009 and June 2010. Over the whole set of observation epochs, S5 0716 +
714 showed significantinter-month variations in flux density at both the K- and Q-bands, with modulation indices of approximately 19 % for the K-bandand approximately 36 % for the Q-band. In all epochs, no clear intraday variability was detected at either frequency. The sourceshows monotonic flux density increase in epochs 1 and 3 and monotonic flux density decrease in epochs 2 and 4. In the flux densityincreasing phases, the flux densities at the Q-band increase more rapidly. In the decreasing phase, no significant flux density di ff erenceis seen at the two frequencies. The situation could be di ff erent close to flux density peaks that we did not witness in our observations.We find an inverted spectrum with mean spectral indices, ¯ α ( S ν ∝ ν − α ), of -0.57 ± ± α of + ± + ± ff ects rather than by anyextrinsic scintillation e ff ect. Key words.
Galaxies: active - Galaxies: BL Lacertae objects: individual: S5 0716 +
714 - Galaxies: jets - Radio continuum: galaxies
1. Introduction
Flux density variability on various time scales in active galacticnuclei (AGNs) has been reported over a broad range of the elec-tromagnetic spectrum. Intraday variability (IDV) in AGNs wasobserved in radio wavelengths as well as in the optical band (e.g.,Heeschen et al. 1987; Wagner et al. 1990; Quirrenbach et al.1992; Wagner & Witzel 1995; Kraus et al. 2003). Heeschenet al. (1987) categorized IDVs into two types according to timescales of flux density variability, namely, sources with variabil-ity longer than two days as type I and shorter than two days astype II. IDV has been detected in compact radio sources with flatspectra, such as BL Lacertae (BL Lac) objects and optically vio-lent variable (OVV) quasars. IDV has been explained as resultingfrom either superluminal motion of relativistic shock in an inho-mogeneous jet (Marscher & Gear 1985; Qian et al. 1991), rapidchanges of the direction of shocks inside a jet (Gopal-Krishna &Wiita 1992), or the extrinsic interstellar scintillation (ISS) e ff ectdue to scattering in the interstellar plasma screen between thesource and the observer (Spangler et al. 1989; Dennett-Thorpe& de Bruyn 2002). The ISS e ff ect is usually dominant at longwavelengths because the scattering e ff ect depends on frequency,according to the relationship ν − . (Rickett et al. 1984). There-fore, the ISS e ff ect is not significant at high radio frequency.S5 0716 +
714 ( z = . ± .
08; Nilsson et al. 2008) is aflat-spectrum BL Lac object. The source is known as a type (cid:63)
Correspondence to: [email protected]
II IDV source at 2.7 GHz (Heeschen et al. 1987). IDV with apeak-to-peak amplitude of ∼
20 % has also been reported even at32 GHz (Krichbaum et al. 2001; Kraus et al. 2003). A close cor-relation between the radio and the optical band with a short timelag in this source was observed (Quirrenbach et al. 1991; Wagneret al. 1996). In multi-frequency campaigns from centimeter tosub-millimeter wavelengths, S5 0716 +
714 did not exhibit typeII IDV, whereas a monotonic increase in the flux density simi-lar to interday variability was detected (Agudo et al. 2006; Os-torero et al. 2006; Fuhrmann et al. 2008). Fuhrmann et al. (2008)found that the flux density variation is correlated over all ob-served wavelengths from centimeter to millimeter and that thereis a trend of increasing time lag toward lower frequencies. Theyalso found that the flux density variations tend to be stronger athigher frequencies. All of these behaviors of the variability wereinterpreted as evidence of a source-intrinsic origin rather than aspart of the ISS e ff ect. Gupta et al. (2012) detected the intradayvariability with the timescale of less than 1.5 days at the low ra-dio frequencies (2.7, 4.8, and 10.5 GHz). They found more rapidvariability (approximately 0.5 days) at 10.5 GHz and that thevariability of 10.5 GHz leads the variability of the lower frequen-cies by approximately 1 day. They also found the modulation in-dices decrease with increasing frequency. These behaviors sug-gest that ISS seems to be dominant at 2.7 and 4.8 GHz, whereasthe intrinsic contribution to the source variability predominatesat 10.5 GHz. Rani et al. (2013) reported that the broadband fluxdensity and variability of S5 0716 +
714 peak around 43 GHz that
Article number, page 1 of 6 a r X i v : . [ a s t r o - ph . GA ] M a r able 1. Observation dates
Epoch
Date MJD1 12 Dec 2009 - 15 Dec 2009 55177.57 - 55180.922 5 Jan 2010 - 11 Jan 2010 55201.04 - 55207.533 28 Jan 2010 - 31 Jan 2010 55224.80 - 55227.604 14 Jun 2010 - 16 Jun 2010 55361.86 - 55363.71they are saturated above this frequency, and that they are dampedat lower frequency due to an opacity e ff ect. Therefore, observa-tions at 43 GHz are advantageous for the study of the character-istics of variability in the flux density.This study was planned to search for type II IDV at the radiofrequency and for evidence of intrinsic flux density variability inS5 0716 +
2. Observations and data reduction
The Korean VLBI Network (KVN), which is a very long base-line interferometry (VLBI) facility for mm-wavelength, allowssimultaneous multi-frequency observations (Kim et al. 2004).KVN consists of three 21-meter Cassegrain radio telescopes thatare located in Seoul (Yonsei University), Ulsan (University ofUlsan), and Jeju Island (Jeju International University) in the Re-public of Korea .Single-dish observations were performed with the KVNYonsei radio telescope at four epochs between December 2009and June 2010 . The observation dates are listed in Table 1. Inepoch 1, we observed from UT 04:00 to UT 16:00 in 3.5 con-secutive days. In epochs 2 - 4, the observations were conductedover 6.5 days, 3 days, and approximately 2 days, respectively.However, the observations were ceased due to bad weather con-ditions for approximately 1 day in epoch 3. The target and thecalibrator were observed every 30 minutes in epochs 1-3 withan on-source integration time of 40 sec at the K-band and 20sec at the Q-band. In epoch 4, we observed the sources in everyhour with an on-source integration time of 80 sec and 40 sec atthe K-band and the Q-band, respectively, due to the high systemtemperature in summer.Simultaneous dual-frequency observations were conductedat both 21.7 GHz (K-band) and 42.4 GHz (Q-band). The ob-served bandwidth was 512 MHz and the system temperatures, T sys , were from ∼
80 to ∼
230 K at the K-band and from ∼
170 to ∼
250 K at the Q-band over the whole set of observation epochs.Flux density measurements were performed by cross-scanson the source in the azimuth and elevation directions. One cross-scan set consists of ten sub-scans in each direction. The standarddeviation of the pointing o ff sets over the whole set of observa-tion epochs were ∼ (cid:48)(cid:48) in the azimuth direction and ∼ (cid:48)(cid:48) in theelevation direction. Due to a deviation of the beam alignment,the pointing o ff set between the K-band and the Q-band was ∼ (cid:48)(cid:48) in the elevation direction. Pointing o ff sets larger than 50 % of theHPBW (half-power beam width) at the Q-band occurred often inthe elevation direction in epoch 1. This is because of thermal de-flection by the antenna mounting structure. Therefore, to avoidthis e ff ect, we adjusted the antenna pointing during monitoringfrom epoch 2 to epoch 4. The primary calibrator, 3C 286, andthe secondary calibrator, 0836 + http://kvn.kasi.re.kr/index.html/main.html the antenna temperature to the flux density and to calibrate thetime-dependent systematic variations, respectively. The final er-rors of the flux densities are 0.16 Jy at 22 GHz, 0.42 Jy at 43 GHzin epoch 1, and 0.08 Jy at 22 GHz and 0.2 Jy at 43 GHz in epochs2-4. S5 0716 +
714 and 0836 +
710 were observed alternately and3C 286 was observed with a cadence of one or two hours. Tomeasure the sky opacity at a zenith, τ , we performed sky-dipobservation every one or two hours. A summary of the observa-tions is provided in Table 2. The data reduction processes for the cross-scan measurementswere as follows: first, we removed the sub-scans that had largefluctuations in power level. Those large fluctuations seem to havebeen a ff ected by atmospheric and instrumental fluctuations. Weaveraged the sub-scans in the azimuth ( az ) and elevation ( el ) di-rections separately. Then, we measured the opacity-corrected an-tenna temperature T ∗ A , the pointing o ff set, and the HPBW usinga Gaussian profile fit to the average of the sub-scans. Here, T ∗ A is defined by T ∗ A = T A · exp τ secz , where T A is the antenna tem-perature, τ is the opacity at zenith, and z is the angle at zenith.Then, the corrected pointing o ff set antenna temperature is T ∗ A , corr , i = T ∗ A , i · exp (cid:16) · ln 2 · x j θ (cid:17) , i (cid:44) j (1)where, T ∗ A , i is the opacity corrected antenna temperature in the i -direction and x j indicates the pointing o ff set in the j -direction; θ is the HPBW (see Table 2) (Fuhrmann 2004; Park et al. 2013).The directions i and j are the azimuth and the elevation, respec-tively. We next determined the corrected antenna temperature T ∗ A , corr , as the arithmetic mean antenna temperature T ∗ A , corr , az and T ∗ A , corr , el . For pointing o ff sets larger than 50 % of the HPBW inthe elevation direction, only T ∗ A , corr , el was taken into account. Inorder to remove systematic time-dependent variations (i.e. rela-tive gain) of S5 0716 + T ∗ A , corr of 0836 + T ∗ A , corr , over a few days; we then computed therelative gain using ¯ T ∗ A , corr / T ∗ A , corr of 0836 + ≤ ≤ σ measurement error. Welinearly interpolated the relative gain of 0836 +
710 with timeand determined a relative gain value with time of 0716 + T ∗ A , corr ,g ain of S5 0716 +
714 was obtained by multiplying the rel-ative gain and T ∗ A , corr of S5 0716 + T ∗ A , corr ,g ain wasconverted to flux density using conversion factors obtained fromthe primary calibrator, 3C 286. Here, the flux density values of3C 286 were 2.64 Jy at the K-band and 1.5 Jy at the Q-band;these values were obtained from measurements of Mars (Sohn etal. in prep.). Errors in the calibrated flux density of S5 0716 +
3. Results
The light curves of S5 0716 +
714 (top) and 0836 +
710 (bottom),obtained at the K-band (red symbols) and the Q-band (blacksymbols) are shown in Fig. 1. The light curves include data overthe whole set of observation epochs. As can be seen in Fig. 1, S50716 +
714 exhibits significant flux density variation at both fre-quencies. During our observations, the source was brightest in2 able 2.
Descriptions of observations
Frequency 21.7 GHz (K-band) 42.4 GHz (Q-band)Bandwidth 512 MHz 512 MHzHalf-power beam width (HPBW) 130 (cid:48)(cid:48) (cid:48)(cid:48) System temperature * ( T sys ) 100, 80, 80, 230 K 200, 170, 170, 250 KObservation mode Cross-Scan in AZ-EL directionFlux density calibrator 3C 286Relative gain calibrator S5 0836 + Notes. ( * ) System temperatures are written for epoch 1, 2, 3, and 4 in sequence. F l ux* ( J y ) F l ux ( J y ) Fig. 1.
Light curves of S5 0716 +
714 (top panel) and of the secondarycalibrator, 0836 + σ measurement error. epoch 1, with mean flux densities of 4.5 ± ± ± ± ± ± ± ± σ measurement error, and thus weconsidered that those variations are insignificant. In our multi-epoch observations, we detected two di ff erent trends in the fluxdensity of S5 0716 + ff erent trends. The flux density of the source decreases at low (Epoch 1) F l ux* ( J y ) K-bandQ-band (Epoch 2) F l ux* ( J y ) F l ux ( J y ) F l ux ( J y ) F l ux ( J y ) F l ux ( J y ) (Epoch 3) F l ux* ( J y ) (Epoch 4) F l ux* ( J y ) F l ux ( J y ) F l ux ( J y ) F l ux ( J y ) F l ux ( J y ) Fig. 2.
Light curves of S5 0716 +
714 (top and middle panels) and of thesecondary calibrator, 0836 + +
714 cor-respond to the flux densities without the calibration, with relative gainof 0836 + + σ measurementerror. To quantify the variability of S5 0716 +
714 at the K- and the Q-bands for each epoch, we followed statistical variability analysismethods such as modulation index m , variability amplitude Y ,and reduced χ -test, as defined by Heeschen et al. (1987) andKraus et al. (2003). To characterize the variability strength ofthe source, the modulation index m and the variability amplitude Y were calculated. The modulation index m is defined as m [%] = · σ S (cid:104) S (cid:105) , (2)where σ S and (cid:104) S (cid:105) denote the standard deviation of the flux den-sity and the mean flux density, respectively. To take into accountthe residual variability due to calibration error of the calibratorsource, the variability amplitude Y was used with Y [%] = (cid:113) m − m , (3)where m is a modulation index of 0836 + χ -test with a hypothesis of a constant model as χ = N − N (cid:88) i = ( S i − (cid:104) S (cid:105) ∆ S i ) , (4)where S i is the individual flux density at time i , (cid:104) S (cid:105) is the averageof the flux density, ∆ S i is the individual measurement error, and N is the number of data points. To reject a constant model inwhich the source has no variability, we chose a p -value cut-o ff of 0.001. This cut-o ff corresponds to 99.9 % significance levelfor variability.We applied the above statistical analysis methods to both S50716 +
714 and 0836 +
710 at the K-band and the Q-band, respec-tively. The results of the analysis of the modulation index m ,the variability amplitude Y , the reduced χ -test, and reduced χ value corresponding to a 99.9 % significance level of variabil-ity are listed in Table 3. We finnd significant inter-month fluxdensity variations and frequency dependence of the variabilityin S5 0716 + m values of approximately 19 % at theK-band and approximately 36 % at the Q-band over the wholeset of observation epochs. These values are larger than thoseof m of the secondary calibrator, 0836 + m ineach epoch is thought to be due to atmospheric e ff ect and resid-ual calibration errors. Although the values of m for 0836 + m ofS5 0716 +
714 in the ranges of 6.5 % to 10.3 % at the K-bandand 8.8 % to 12 % at the Q-band in each epoch. These varia-tions seem to indicate the monotonic increase or decrease in theflux density of S5 0716 + α Fig. 3.
Daily average spectral index of S5 0716 +
714 over the whole setof observation epochs. the source at the higher frequency. Fuhrmann et al. (2008) foundsimilar frequency dependence of the variability from cm to sub-mm data and explained it as a source-intrinsic origin. In epoch2, the highest Y values seem to be due to the longest duration ofthe observation period.To measure a monotonically increasing and decreasing ratein the flux density, we applied linear least-square fitting to thelight curves. Because the light curves in epoch 2 have two di ff er-ent slopes, we divided the light curves into two parts from MJD2455201.04 to JD 2455203.97 and from JD 2455203.98 to JD2455207.53. For epoch 3, we did not conduct the linear least-square fitting to the light curves because of the absence of datafor the one day between MJD 2455225.7 and 2455226.5 and thelarge scatter around MJD 5227.5 that seems to be a ff ected bythe scattering of the secondary calibrator, 0836 + ff erent at the two frequen-cies. This behavior of the variability could be di ff erent close toother flux density peaks that we did not witness in our observa-tions. We will discuss this question in our next paper, which is inpreparation. The variation of the spectral index with time, α (defined by S ν ∝ ν − α , where S ν is the flux density at the observing fre-quency, ν ), of S5 0716 +
714 is shown in Fig. 3. In this figure,the data points indicate the daily mean spectral indices ¯ α . Thespectral indices vary significantly over the whole set of the ob-servation epochs, with changes in sign from negative to positive.There are values of a ¯ α of -0.57 ± + ± ± + ± able 3. Statistical results for S5 0716 +
714 and 0836 + ν , totalnumber of measurements N, mean flux density < s > , standard deviation of flux density σ s , modulation index m, variability amplitude Y, reduced χ χ , and corresponding value of the 99.9 % significance level of variability χ . . Epoch Source ν N < s > σ s m Y χ χ . (band) (Jy) (Jy) (%) (%)1 0716 +
714 K 67 4.5 0.3 6.7 15.7 2.96 1.63Q 67 6.6 0.6 9.1 19.7 2.81 1.630836 +
710 K 74 2.4 0.1 4.2 – 1.73 1.59Q 74 3.2 0.2 6.3 – 1.65 1.592 0716 +
714 K 411 2.9 0.3 10.3 28.2 12.96 1.23Q 411 2.5 0.3 12.0 29.9 2.27 1.230836 +
710 K 411 2.4 0.1 4.2 – 1.83 1.23Q 411 3.0 0.2 6.7 – 2.45 1.233 0716 +
714 K 110 3.1 0.2 6.5 15.8 3.69 1.47Q 110 3.4 0.3 8.8 20.1 1.32 1.470836 +
710 K 107 2.6 0.1 3.8 – 1.69 1.48Q 107 3.5 0.2 5.7 – 2.16 1.484 0716 +
714 K 15 1.9 0.1 5.3 11.7 1.28 2.58Q 15 1.8 0.2 11.1 24.8 2.84 2.580836 +
710 K 18 2.8 0.1 3.6 – 2.74 2.40Q 18 2.7 0.2 7.4 – 1.98 2.40Whole 0716 +
714 K 603 3.1 0.6 19.4 57.1 27.80 1.19epochs Q 603 3.6 1.3 36.1 104.3 21.28 1.190836 +
710 K 610 2.6 0.1 3.8 – 3.83 1.19Q 610 3.1 0.3 9.7 – 4.32 1.19 α Epoch 1 (K)Epoch 1 (Q)Epoch 2 (K)Epoch 2 (Q)Epoch 3 (K)Epoch 3 (Q)Epoch 4 (K)Epoch 4 (Q)r=-0.84 (K)r=-0.95 (Q)
Fig. 4.
Flux densities versus daily average spectral indices of S50716 +
714 for all four epochs. Red and black symbols correspond to theK-band and the Q-band, respectively. Circle, upward triangle, down-ward triangle, and diamond symbols indicate epochs 1, 2, 3, and 4, re-spectively. a tight correlation between the flux densities and the spectralindices. We estimated the Pearson correlation coe ffi cient r be-tween the flux densities and the spectral indices as shown inFig. 4. The correlations are strong (i.e., | r | >
4. Conclusions
In this work, we search for the existence of type II IDV in the fluxdensity of BL Lac object S5 0716 +
714 at radio frequencies. We
Table 4.
Results of linear least-square fitting to the light curves at theK- and Q-bands.
K-band Q-bandEpoch Slope χ Slope χ (Jy / day) (Jy / day)1 0.18 ± ± * -0.25 ± ± + -0.11 ± ± ± ± Notes. ( * ) J.D. 2455201.04-2455203.97 ( + ) J.D. 2455203.98-2455207.53 perform multi-epoch simultaneous dual-frequency observationsat the K- and the Q-bands, using the KVN Yonsei radio tele-scope. In conclusion, the source shows significant inter-monthvariation in the flux density at the K- and the Q-bands with alarge modulation index over the whole set of observation epochs.Despite several intensive observations, no typical type II IDVwas found in either frequency in any of the epochs. The sourceexhibits a monotonic flux density increase or decrease in eachepoch, with increasing variability amplitudes at high frequency.In the flux density increasing phase, the flux density varies morerapidly at the Q-band whereas in the decreasing phase, there areno significantly di ff erent rates for the two frequencies. We ob-served variations of the spectral indices over the whole set ofobservation epochs, with changes in sign from negative to posi-tive. We suggest that this variability behavior could have an in-trinsic origin rather than resulting from the extrinsic scintillatione ff ect. In our observations, we did not find statistically meaning-ful IDV phenomena at 22 and 43 GHz. To understand the originof variability on inter-month time scales, continuous flux densitymonitoring of S5 0716 +
714 will be required.
Acknowledgements.
We would like to thank the anonymous referee for impor-tant comments and suggestions that have improved the manuscript. This work as supported by the KASI-Yonsei DRC program of the Korea Research Coun-cil of Fundamental Science and Technology (DRC-12-2-KASI) and the Interna-tional Research & Development Program of the National Research Foundationof Korea (NRF) funded by the Ministry of Science, ICT and Future Planning(MSIP) of Korea (Grant number: NRF-2012K1A3A7A03049606). References