A Recollimation Shock in a Stationary Jet Feature with Limb-brightening in the Gamma-ray Emitting Narrow-line Seyfert 1 Galaxy 1H 0323+342
Akihiro Doi, Kazuhiro Hada, Motoki Kino, Kiyoaki Wajima, Satomi Nakahara
aa r X i v : . [ a s t r o - ph . GA ] A p r Received 2018 March 16; revised 2018 March 30; accepted 2018 March 30
Preprint typeset using L A TEX style emulateapj v. 12/16/11
A RECOLLIMATION SHOCK IN A STATIONARY JET FEATURE WITH LIMB-BRIGHTENING IN THEGAMMA-RAY EMITTING NARROW-LINE SEYFERT 1 GALAXY 1H 0323+342
Akihiro Doi , Kazuhiro Hada , Motoki Kino , Kiyoaki Wajima , and Satomi Nakahara Received 2018 March 16; revised 2018 March 30; accepted 2018 March 30
ABSTRACTWe report the discovery of a local convergence of a jet cross section in the quasi-stationary jetfeature in the γ -ray-emitting narrow-line Seyfert 1 galaxy (NLS1) 1H 0323+342. The convergencesite is located at ∼ γ -ray-emitting sites in this NLS1, in analogy with the HST-1 complexin the M87 jet. Monitoring observations have revealed that superluminal components passed throughthe convergence site and the peak intensity of the quasi-stationary feature, which showed apparentcoincidences with the timing of observed γ -ray activities. Subject headings: galaxies: active — galaxies: Seyfert — galaxies: jets — radio continuum: galaxies— galaxies: individual (1H 0323+342) — gamma rays: galaxies INTRODUCTION
The detection of GeV γ -ray emission from narrow-line Seyfert 1 (NLS1) galaxies by the Large Area Tele-scope on board the Fermi Gamma-ray Space Telescope satellite (
Fermi
LAT; Abdo et al. 2009) raises issues ofthe γ -ray production processes and acceleration mecha-nisms of relativistic jets in this subclass of active galac-tic nuclei (AGNs). Radio observations have providedstrong evidence of a pole-on viewed relativistic jet byfinding very high brightness and rapid variability on theflat-spectrum core associated with the one-side jet (e.g,Doi et al. 2006, 2011; D’Ammando et al. 2013), whichare apparently quite similar to that of blazars (e.g.,Foschini et al. 2015).1H 0323+342 is the nearest (redshift of 0 . γ -ray-emittingradio-loud NLS1s. One of the detected γ -ray flaresshowed a flux-doubling time of 3 hr, which was the fastest γ -ray variability ever observed from NLS1s (Paliya et al.2014, 2015). Prominent radio flares with a timescaleof ∼
30 days have been evident only at high frequen-cies, while the source activity is very moderate at lowerfrequencies ( .
10 GHz; Angelakis et al. 2015). The γ -ray/radio correlation for 1H 0323+342 has not been [email protected] The Institute of Space and Astronautical Science, JapanAerospace Exploration Agency, 3-1-1 Yoshinodai, Chuou-ku,Sagamihara, Kanagawa 252-5210, Japan Department of Space and Astronautical Science, The Grad-uate University for Advanced Studies, 3-1-1 Yoshinodai, Chuou-ku, Sagamihara, Kanagawa 252-5210, Japan Mizusawa VLBI Observatory, National Astronomical Obser-vatory of Japan, Osawa, Mitaka, Tokyo 181-8588, Japan Department of Astronomical Science, The Graduate Univer-sity for Advanced Studies (SOKENDAI), 2-21-1 Osawa, Mitaka,Tokyo 181-8588, Japan Kogakuin University, Academic Support Center, 2665-1Nakano, Hachioji, Tokyo 192-0015, Japan National Astronomical Observatory of Japan, 2-21-1 Osawa,Mitaka, Tokyo 181-8588, Japan Korea Astronomy and Space Science Institute (KASI), 776Daedeokdae-ro, Yuseong-gu, Daejeon 34055, Republic of Korea investigated in detail. One-sided morphology and su-perluminal components were identified at parsec scalesby monitoring observations using very-long-baseline in-terferometry (VLBI; Fuhrmann et al. 2016; Lister et al.2016). On the other hand, two-sided radio morphology isseen at kiloparsec scales (Ant´on et al. 2008). These char-acteristics can be understood analogously to the unifiedscheme of radio-loud AGNs that considers radio galaxiesas non-beamed parent populations of blazars (Doi et al.2012).The presence of a quasi-stationary feature located at ∼ ∼
120 pc ) from the 1H 0323+342 nucleus wassuggested by Wajima et al. (2014). Such a structure isreminiscent of the M87 ( z = 0 . γ -ray sites in M87(Cheung et al. 2007). The location of this γ -ray activityis much farther from the central engine than previouslythought in relativistic jet sources. This phenomenon canbe explained as a consequence of inverse-Compton up-scattering off ambient starlight photons (Stawarz et al.2006). Asada & Nakamura (2012) discovered that thecross section of the jet shows locally smaller at the HST-1 in the jet-width profile. They discussed the origin ofthe HST-1 as a consequence of a recollimation shock(Stawarz et al. 2006) due to overcollimation by a pres-sure imbalance between the jet and ambient medium(e.g., G´omez et al. 1997; Mizuno et al. 2015) around thesphere of the Bondi radius. The plausible γ -ray-emittingsites on a standing shock located far from the central en- If we adopt a black hole mass of M BH ∼ × M ⊙ (Landt et al. 2017, and references therein) and an inclination an-gle of ∼ ◦ (Abdo et al. 2009) for 1H 0323+342, 1 mas correspondsto 1.2 pc in projection or 4 . × R S in deprojection, where R S is the Schwarzschild radius. We assume a ΛCDM cosmology with H = 70 . − Mpc − , Ω M = 0 .
27, and Ω Λ = 0 . Doi, et al. Relative Right Ascension (mas)
Relative Declination (mas) (a) Relative Right Ascension (mas)
Relative Declination (mas) (b)
Fig. 1.—
Stacked image and result of model fitting for1H 0323+342 jet. (a) The total intensity contour map that wasmade by stacking images over all eight epochs, provided by theMOJAVE project. (b) Overlay plots for the positions and decon-volved sizes of components in the sky with colors depending onepochs. Black dashed line represents the jet axis at
P A = 125 . ◦ gine have also been proposed for distant blazars (BL Lac,Marscher et al. 2008; PKS 1510 − γ -ray productions by direct imag-ing using VLBI angular resolutions. In this Letter, wepresent the observation result that the quasi-stationaryfeature in the 1H 0323+342 has a converging/divergingstructure with limb-brightening, which is likely the resultof a recollimation/reflected shock. Furthermore, fromVLBI monitoring data, we show an indication of a pos-sible γ -ray site on the quasi-stationary feature, whichmuch farther from the 1H 0323+342 central engine. DATA AND DATA ANALYSES -4 -3 -2 -1 I n t en s i t y ( Jy / bea m ) Distance from core along jet axis (mas)
Distance from core along jet axis (mas) C o m ponen t s i z e ( m a s ) (b)(a) F l u x den s i t y ( m Jy ) C o m ponen t s i z e ( m a s ) Distance from core along jet axis (mas)
C4 flux density C4 component size (c)
Fig. 2.—
Spatial-domain plots for 1H 0323+342 jet components.(a) Deconvolved sizes of components with distance from the core.The vertical solid line and gray shaded belt represent the averageand standard deviation of the location of convergence site in thejet, respectively. (b) Slice profiles of intensity along the jet axis.Color variations denoted in the panel are shared with the panelFigure 1 (b). The vertical dashed line and red shaded belt repre-sent the average and standard deviation of the location of intensitypeak on the quasi-stationary feature “S” in the jet, respectively.(c) Evolution of component C4 regarding flux density and size inradius.
We retrieved all of the calibrated visibility data for1H 0323+342 observed at 15.4 GHz using the very longbaseline array (VLBA) from the Monitoring Of Jetsin AGNs with VLBA Experiments (MOJAVE) project(Lister & Homan 2005) online page . The data were ac-quired from 210 October 15 to 2013 July 08 (eight epochsin total) in a series of observations coded BL149 andBL178. We analyzed the visibilities using difmap soft-ware (Shepherd et al. 1994). Radio structures were es-tablished using circular Gaussian model components in modelfit procedure.Fuhrmann et al. (2016) reported their results obtainedin the same manner; most of the identified componentswere consistent with our results. We were not confi-dent of model fitting for the data on 2012 December 23(the seventh epoch) because of significantly poor data recollimation shock in the γ -ray NLS1 1H 0323+342 3 S epa r a t i on f r o m c o r e ( m a s ) MJD (day)
Fig. 3.—
Historical location of the quasi-stationary feature S.The data before and after modified Julian date (MJD) 55484 (2010October 15) come from Wajima et al. (2014) and the present study,respectively. The solid line represents fitted proper motion. quality compared to the other epochs. For the presentstudy, we adopt only features cataloged with robustcross-identifications across epochs in the table of the MO-JAVE paper (Homan et al. 2015) for the seventh epoch.The formal errors of the model fit parameters were esti-mated to be σ z = θ/ (2 × SNR) and σ r = θ/ SNR for theposition and deconvolved component size, respectively(Fomalont 1999). In the case of very small componentsand high flux density, this tends to underestimate theerror. We included an additional minimum error of θ/ difmap . These imagesare used to investigate the intensity profiles of the jet. RESULTS
A converging Jet Structure
Pc-scale jet structures were modeled using a series ofdiscrete components. Figure 1 (a) shows a stacked imagemade from images from all the eight epochs. Figure 1 (b)shows the jet components regarding the positions and de-convolved sizes, overlaid with the stacked image. There isan emission gap in the jet at a radial distance of z ∼ z ∼ z = 7 . ± .
17 mas from the core ( d = 0 . ± .
13 masin size), which is the average and standard deviation ofmeasurements for one or two local minimums at eachepoch. We call this location the “convergence site.” Thejet appears to expand again at the downstream of theconvergence site.The jet-width profile over a wide distance range for 1H0323+342 is separately reported by Hada et al. (2018),which also confirms a locally narrowed jet based on multi-frequency (1.4–43 GHz).
A Quasi-stationary Feature in the Jet
Figure 2(b) exhibits slice profiles of intensity along thejet axis of
P A = 125 . ◦
0, on the basis of images regener-ated with a common restored beam of 0.75 mas, whichis the geometric mean of the major and minor axes oforiginal beam sizes. We found that the intensity maxi-mum in the jet is placed at z = 6 . ± .
17 mas from thecore, which is the average and standard deviation of thepeak positions at the all epochs. The location of the con-vergence site is slightly shifted from that of the intensitypeak in the jet feature.It has been pointed out that this jet feature is sta-tionary for a corresponding component; Wajima et al.(2014) reported a proper motion of µ app = − . ± .
083 mas yr − during ∼
15 year. Figure 3 shows thehistorical locations, including the measurements in ouranalyses for an additional further eight epochs in MO-JAVE data. We derived a proper motion of µ app = 0 . ± .
01 mas yr − and a historical position of 7 . ± . ∼
18 year. We found some degree of wobble in dis-tance position during the MOJAVE period. Hence, wecall this feature “quasi-stationary feature S.”We recognize the existence of another potential sta-tionary component very close to the core, z ∼ . Time Evolution of jet components
Figure 4 (b) is the plot for the positions of compo-nents in distance from the core. The identifications of aseries of components and proper motions are similar tothose in previous studies using the same MOJAVE data(Fuhrmann et al. 2016; Lister et al. 2016). The compo-nent C4 has been evidently identified as a superlumi-nal component with the highest apparent speed in thissource. A model with a constant apparent speed fits themeasurement points quite well, with χ / ndf = 5 . / p = 0 .
53, where ndf is the numberof degrees of freedom.Figure 2 (c) shows the evolutions of C4 regarding itsflux density and angular size. This component initiallyexpanded as it became dark, then turned to flare up asit approached the convergence site along the intensityprofile of the quasi-stationary feature S. The other com-ponents also showed similar evolution (not shown). Thisbehavior indicates how the part of the jet is maintainedas a quasi-stationary feature.
Limb-brightening Structure
Figure 5 shows the CLEAN image on the epoch of 2011July 15, which was convolved with a circular restoringbeam with 0.56 mas that was equivalent to the minor-axis size of the original synthesized beam. We foundlimb-brightening intensity profiles at both the upstreamand downstream with respect to the intensity peak andthe convergence site in the quasi-stationary feature S. Weperformed a pixel-based analysis by slicing in transversedirections of the jet at z = 6 and z = 9 on the image; adouble-peaked slice profile is evident in the both cases.Similar structures were also apparent at upstream at fourepochs and/or downstream at six epochs among the eightepochs. DISCUSSION
Doi, et al. (b) F l u x den s i t y ( Jy ) MJD (day)
CoreQuasi-stationary S (x 10) S epa r a t i on f r o m c o r e ( m a s ) MJD (day)
F1 F2 F3 C1 C0 C5 C2 C3 C4 SP CS (a)
Fig. 4.—
Time-domain plots for 1H 0323+342. (a) Measurementsof separations from the core for jet components. The solid linesrepresent fitted proper motions for components (the yellow shadedbelt on C4 shows fitting error). The horizontal solid line and redshaded belt represent the average and standard deviation of thelocation of convergence site (CS) in the jet. The horizontal dashedline and gray shaded belt represent the average and standard de-viation of the location of quasi-stationary feature’s intensity peak(SP) in the jet. (b) Radio light curves of the core region and quasi-stationary features S in MOJAVE images. The periods shaded bylight blue and denoted as F1, F2, and F3 represent known γ -rayactivity periods, which were defined by Paliya et al. (2014). Thevertical doted-dashed line represents the time of a fast γ -ray flareat MJD = 56 , .
13 (Paliya et al. 2014, 2015).
Our analyses of the radio imaging data for 1H0323+342 have provided us with two key results. Thefirst key result is the discovery of a recollimation siteas the converging shape in the pc-scale jet, by mea-surements for the deconvolved sizes of jet blobs. Fur-thermore, the intensity peak of the quasi-stationary jetfeature spatially coincides with, or slightly upstreamof, the convergence site. The second key result is thefinding of limb-brightening structures at both the up-stream and downstream with respect to the intensitypeak and the convergence site in the quasi-stationary fea-ture. These critical sites are considerably distant fromthis NLS1 central engine. If we adopt a black hole massof M BH ∼ × M ⊙ (Landt et al. 2017, and referencestherein) and an inclination angle of ∼ ◦ (Abdo et al.2009), the projected distance z ∼ TABLE 1Estimated MJDs of Passing Through the Two CriticalSites.
Component Quasi-stationary S Convergence Site(MJD) (MJD)(1) (2) (3)C3 56144 ±
145 56417 ± ±
45 56561 ± ±
101 57112 ± Note . — Column. (1) ID of components; Column (2) estimatedMJD of passing through the location of the intensity peak of thequasi-stationary feature S; Column (3) estimated MJD of passingthrough the location of the convergence site. ∼
120 pc or ∼ × R S in deprojection. Analogy with M87
These characteristic jet structures in 1H 0323+342 arereminiscent of the M87 jets in pc scales and the HST-1complex at a deprojected distance of 120 pc. The in-ner radio jet of M87 showing a limb-brightened structure(e.g., Kovalev et al. 2007). Asada & Nakamura (2012)discovered a jet structure locally convergent at the lo-cation of HST-1, which is the transitional boundary ofa jet acceleration region with a parabolically expand-ing streamline and a deceleration region with a conicalexpansion (e.g., Nakamura & Asada 2013; Asada et al.2014). 1H 0323+342 also shows a parabolic shape inthe inner jet and a hyperbolic/conical shape at largerscales (Hada et al. 2018). The location of the upstreamend in the HST-1 complex in M87 has been stable towithin ∼ γ -ray emission(Stawarz et al. 2006; Cheung et al. 2007). Limb-brightened Structure with a RecollimationShock
We propose that the combination of the observed limb-brightened structure and the convergence site in thequasi-stationary feature in the 1H 0323+342 jet can beexplained in the framework of the recollimation shock.We consider that the limb-brightening at the upstreamof the convergence site is possibly a radiation layer ofshocked materials flowing between a contact discontinu-ity that separates from the external medium and the rec-ollimation shock front. The low-brightness inner layeris considered to be an unshocked spine flow. Such apicture has been demonstrated by Bromberg & Levinson(2009) and Bodo & Tavecchio (2018) to explain the ac-tivity in a small region located at a considerable dis-tance from the central engine in several blazars andM87. The shape of the post-shocked dissipation layerafter the recollimation shock is very similar to the limb-brightening at the upstream of the convergence site recollimation shock in the γ -ray NLS1 1H 0323+342 5 - I n t en s i t y ( m Jy bea m ) Transverse Distance (mas)
Convergence siteIntensity peak z = 6 masz = 9 mas
Fig. 5.—
Jet structure of 1H 0323+342. Contours represent the total intensity map of MOJAVE data on the epoch MJD = 55757.Sub-figures show the sliced intensity profiles in transverse direction at the locations z = 6 (red color) and z = 9 (blue color) from the core.Bold lines illustrate a putative sheath layer in the jet as a shocked plasma after a recollimation shock. in 1H 0323+342. Bromberg & Levinson (2009) andBodo & Tavecchio (2018) also revealed that significantpressure loss caused by radiative cooling can lead to thejet being focused on a tiny cross-sectional radius by ex-ternal pressure. This picture can explain well the factthat the intensity peak and convergence site are adja-cent to one another, as observed in 1H 0323+342. Thelimb-brightening at the downstream in the 1H 0323+342jet can be caused by emission from the combination ofa shocked outer layer with a relatively high Γ and innerlayer strongly decelerated through a reflected shock, asshown in Bodo & Tavecchio (2018). Possible Coincidences of γ -ray Activity andJet-passing Events through the Critical Sites We also discuss possible correlations between γ -ray ac-tivities and passing events of the superluminal compo-nent through the intensity peak of the quasi-stationaryfeature S and the convergence site. This is the possiblethird finding in the present study (Figure 4 (b)). Themost striking event is that the component C4 reached the intensity peak of S during the known GeV γ -rayactive phase F2 (MJD = 56470–56500; Paliya et al.2014). Furthermore, the subsequent γ -ray active phaseF3 (MJD = 56531–56535) coincides with the componentC4’s arrival at the convergence site. The estimated datesof passing through the two critical sites are listed in Ta-ble 1.Table 1 also presents corresponding dates for the com-ponents C3 and C5. The date of C3 crossing the conver-gence site was coincident with the previous γ -ray activephase F1 (MJD = 56262–56310). The proper motion ofC5 is expected to cross the two critical sites at MJD ∼ Fermi
LAT daily andweekly quick-look light curves (not shown) show recur-rence in γ -ray in the duration of MJD ∼ γ -ray emission and passage eventsof jet components. The recollimation site in the quasi-stationary feature is a possible candidate for observed https://fermi.gsfc.nasa.gov/ssc/data/access/lat/msl_lc/ Doi, et al. γ -ray emission in 1H 0323+342.X-ray spectral modeling suggested a dominance bynonthermal jetted (inverse-Compton) emission duringboth F2 and F3, whereas in the quiescent (and F1)phase there is a significant dominance by thermal corona(Paliya et al. 2014). Given the hour-scale light curve inF3 and its estimated γ -ray luminosity, a good fractionof the total kinetic energy of the jet was converted intoradiation (Paliya et al. 2014). They proposed that thedissipation sites for all of these γ -ray phases are located ∼ R S ( ∼ .
002 pc) from the nucleus, with region sizesof ∼ . ∼ . ∼ . γ -ray productionsat the quasi-stationary feature.The C4 showed a progressively increasing flux densityby a factor of three (Figure 2 (c)). This indicates how thequasi-stationary feature keeps stationary, i.e., as a resultof the aggregation of passing components that brightenas they approach the convergence site. However, theseflux densities were, at most, at the mJy level. On the other hand, radio flares seen in the F-GAMMA lightcurves based on single-dish monitoring were on the or-der of 1 Jy (Angelakis et al. 2015). We examined the γ -ray/radio correlation on spatially resolved light curvesusing the MOJAVE images (Figure 4 (b), a similar plotwas also presented by Fuhrmann et al. 2016). However,we found no clear evidence of γ -ray/radio correlation, ei-ther in the core region or the quasi-stationary feature.These inconclusive results may be due to the MOJAVE’ssparse sampling ( ≥
50 days) compared to the flare’s timescales ( .
30 days) in the total radio flux (Angelakis et al.2015). Searching for component emergence events forboth the core and the stationary feature helps to iden-tify the γ -ray-emitting site (Doi, A. et al. in preparation).The superluminal ejections from the stationary locationin the HST-1 complex were observed during the 2005 TeV γ -ray event in M87 (Cheung et al. 2007; Giroletti et al.2012).This research has made use of data from the MO-JAVE database that is maintained by the MOJAVE team(Lister et al. 2009).-ray event in M87 (Cheung et al. 2007; Giroletti et al.2012).This research has made use of data from the MO-JAVE database that is maintained by the MOJAVE team(Lister et al. 2009).