搜档网
当前位置:搜档网 › VLBI imaging of OH absorption The puzzle of the nuclear region of NGC 3079

VLBI imaging of OH absorption The puzzle of the nuclear region of NGC 3079

a r X i v :a s t r o -p h /0404347v 1 17 A p r 2004

Mon.Not.R.Astron.Soc.000,000–000(0000)Printed 2February 2008

(MN L A T E X style ?le v2.2)

VLBI imaging of OH absorption:The puzzle of the nuclear

region of NGC 3079

Yoshiaki Hagiwara,1?Hans-Rainer Kl¨o ckner,2,1Willem Baan 1

1

ASTRON,Westerbork Observatory,P.O.Box 2,Dwingeloo,7990AA,The Netherlands

2Kapteyn

Astronomical Institute,University of Groningen,Postbus 800,Groningen,The Netherlands

2003

ABSTRACT

Broad hydroxyl (OH)absorption-lines in the 1667MHz and 1665MHz transition to-wards the central region of NGC 3079have been observed at high resolution with the European VLBI Network (EVN).Velocity ?elds of two OH absorption components were resolved across the unresolved nuclear radio continuum of ~10parsecs.The velocity ?eld of the OH absorption close to the systemic velocity shows rotation in nearly the same sense as the edge-on galactic-scale molecular disk probed by CO(1–0)emission.The velocity ?eld of the blue-shifted OH absorption displays a gradient in al-most the opposite direction.The blue-shifted velocity ?eld represents a non-rotational component,which may trace an out?ow from the nucleus,or material driven and shocked by the kiloparsec-scale superbubble.This OH absorption component traces a structure that does not support a counter-rotating disk suggested on the basis of the neutral hydrogen absorption.

Key words:galaxies:active –galaxies:individual (NGC 3079):ISM –galaxies:Seyfert –radio lines:galaxies

1INTRODUCTION

NGC 3079is an edge-on Sc galaxy housing a LINER (Low Ionisation Narrow Emission Line Region)nucleus (Heckman 1980)with a long history of observations at various wave-lengths.The nucleus is also classi?ed as a type 2Seyfert nu-cleus (Ford et al.1986).The galaxy has a systemic velocity of 1116km s ?1(Irwin &Seaquist 1991),which gives the dis-tance of 16Mpc and hence 1milliarcsec (mas)corresponds to approximately 0.08pc,adopting H 0=75km s ?1Mpc ?1.Observations of optical emissions with the Hubble Space Telescope (HST)have shown a number of out?owing ?la-ments in a ’superbubble’,produced by a combination of stel-lar winds and supernova explosions in a site of massive star formation.X-ray observations with Chandra show a clear correspondence of H α-line ?laments with those in X-rays at the distance of ~65pc from the nucleus (Cecil et al.2002).Throughout several X-ray observations,a highly ob-scured active galactic nucleus (AGN)and iron line emission towards the nucleus were found (Cecil et al.2002).The cen-ter of the galaxy has been resolved by radio interferometry and reveals a double radio lobe extending 1.5kpc on either side of the major axis of the galaxy and a puzzling nuclear structure on parsec scales (Irwin &Seaquist 1988).Due to the complexity of the radio structure,the interpretation of

?

E-mail:hagiwara@astron.nl (YH)

the location of a true nucleus is not straightforward.

Broad (200–400km s ?1)and deep (τ~0.5)neutral hy-drogen H I absorption and weaker but equally broad and deep OH absorption are found towards those compact radio sources in the nucleus.These absorptions have been studied in detail at various angular resolutions by radio interferom-eters.Based on the Very Large Array (VLA)observations at 1arcsec resolution,Baan &Irwin (1995)considered that the two components of H I and OH absorptions are associ-ated with an obscuring torus in the nuclear region.Multi-Element Radio Linked Interferometer Network (MERLIN)observations at sub-arcsecond resolution resolved the rota-tion of the H I absorption (Pedlar et al.1996)in the same sense as the rotational trend traced by H I emission and CO(1–0)emission on galactic scales (Irwin &Seaquist 1991;Irwin &Sofue 1992).According to Very Long Baseline Inter-ferometry (VLBI)observations at an angular resolution of 10-15mas,three H I absorption components viewed against the resolved double-peaked radio source appear to be ev-idence for counter rotation relative to the rotation in the outer galaxy (Sawada-Satoh et al.2000).

The strong and highly intensity-variable water maser emis-sion is known to exist in NGC 3079(Henkel et al.1984;Haschick &Baan 1985).Earlier attempts for resolving the maser emission have been made since the late 1980s us-ing VLBI (Haschick et al.1990).At present,it is under-stood that the H 2O maser components are distributed

2

Table1.Properties of NGC3079.Coordinates in this table were phase centre values used for the data correlation at JIVE.Veloc-ity with respect to LSR for radio de?nition is adopted from Irwin &Seaquist(1991).Optical classi?cation was made by Heckman (1980)and Ford et al.(1986).Radio?ux densities were measured by the VLBA(Sawada-Satoh et al.2000).Infra-red(IR)lumi-nosity is from Baan(1989).X-ray luminosity is from Cecil et al. (2002).We assume that H0=75km s?1Mpc?1.

within an elongated(disk-like)structure within a few par-secs from the nucleus(Trotter et al.1998).Most of the maser components detected with VLBI lie in a velocity range V LSR(LSR refers Local Standard of Rest.)=956–1190km s?1and are not associated with any of the jet components(Trotter et al.1998;Sawada-Satoh et al.2000). They are distributed nearly north-to-south,and are aligned roughly with the molecular disk traced by CO(1–0)(Position Angle(P.A.)=15?)with a radius of750pc(Irwin&Sofue 1992).Several new weak components were recently detected, which extend the velocity range to about V LSR=1350 km s?1(Hagiwara et al.2002a).Given the velocity distri-bution of all the maser features observed to date,the maser spectrum shows a rough symmetry w.r.t the systemic veloc-ity of the galaxy(V LSR?1116km s?1)(Hagiwara et al. 2002b).It has been argued that there is a rotating parsec-scale molecular gas disk or torus obscuring an AGN in NGC3079,but no concrete evidence has been presented in the literature like the one in NGC4258(Herrnstein et al. 1998).A recent study by Middelberg et al.(2003)presents evidence for?ve distinct radio continuum sources in the nu-cleus of NGC3079,complicating the interpretation of its puzzling structure.Besides the three components A,B,and C that are detected at higher frequencies and are seen at22 GHz together with the H2O masers,there are two other com-ponents E and F that are visible only at lower frequencies, following the convention used in Irwin&Seaquist(1988). These additional components have the same orientation and are equally spaced as the22GHz components A and B,but they are shifted about25mas?2pc to the east;they form the18-cm radio signature that is observed in H I and OH absorption experiments.

In this paper,we present VLBI observations of OH ab-sorption at the nuclear region of NGC3079.With observa-tions using European VLBI Network(EVN)at the highest spectral-line sensitivity of any existing VLBI facility,we aim to review the kinematics of the circumnuclear region in the galaxy.2OBSER V ATIONS

The central region of NGC3079was observed at18-cm on

2000November15with the EVN,which consisted of eight telescopes:Cambridge,E?elsberg,Jodrell Bank(Lovell), Medicina,Noto,Onsala,Torun,and the Westerbork

phased-array.We observed the1667and1665MHz main line transitions of the2 3/2,J=3/2ground-state of OH in absorption toward the central radio continuum source in

NGC3079.

NGC3079was observed in a phase-referencing mode over a period of 5.7hrs,interspersed with observations of a phase calibrator source J0957+5522,which is located about1degree from the target source.The total time of one observing cycle was13minutes with10minutes for NGC3079and3minutes for J0957+5522.The total time spent on the target was about3hours.The data were recorded in both left and right circular polarizations using a single intermediate frequency(IF)band with8 MHz bandwidth.The IF baseband was subdivided into128 spectral points,yielding a frequency resolution of62.5kHz, or11.2km s?1in velocity at the source distance.The IF velocity coverage is1440.5km s?1.In order to cover both the main line transitions,we centered the IF on V LSR= 1350km s?1and the Doppler velocity center was referenced to the1667MHz line.Hereafter,all velocities are in the radio convention and with respect to LSR.The system temperature and antenna sensitivity for each EVN antenna over the observing frequency range from30K?100K dur-ing the observing run.DA193was measured in the middle of the observations for calibrating the absolute amplitude gain and the bandpass correction.The data-recording rate was128Mbit per second with a MkIV terminal.

The correlation of the data was performed at the EVN MkIV correlator at the Joint Institute for VLBI in Eu-rope(JIVE).Data analysis was made using the NRAO AIPS package.After the delay and delay rate calibra-tion using J0957+5522,the bandpass corrections were ap-plied.We discarded the data from six baselines,Noto(NT)-Medicina(Mc),Mc-Torun(Tr),Tr-NT,Cambridge(Cm)-Tr, Cm-NT,and Cm-Mc baseline,due to inadequate data qual-ity.The continuum visibility data set was generated from the spectral-line visibility data set by averaging absorption-free channels using the AIPS task of UVLIN.The contin-uum visibility data set was then used for self-calibration to improve the image sensitivity.After iterations of the self-calibration in both phase and amplitude,the solutions were transferred to the spectral-line visibility data.The contin-uum and spectral-line visibilities were imaged employing IMAGR.

3RESULTS

Fig.1(left)shows the EVN spectrum of OH absorption integrated over the nuclear radio continuum source(Fig.4) with natural weighting with a spectral resolution of11.2 km s?1.The systemic velocity of the galaxy V LSR=1116 km s?1and the peak velocities of HI absorption and water maser emission are denoted.The rms noise level is0.7mJy beam?1per channel.The OH main-line transitions of1667

3

MHz and1665MHz are clearly detected against an unre-solved continuum source at the centre of NGC3079.Both transitions have two distinct components,one of which is blue-shifted(OH1)and the other is nearly centred on the systemic velocity(OH2).This is consistent with the spectra at lower resolutions(Baan&Irwin1995).By comparison with the single-dish spectrum,the OH absorption lines ob-served in these observations have lost their broad wings.The integrated intensities of the1667MHz OH absorption esti-mated from the single-dish and the EVN spectrum in Fig. 1(left)are about8.3Jy km s?1(Baan&Irwin1995)and 0.74Jy km s?1,respectively.Only8.9percent of the ab-sorption intensity was recovered in the EVN observations. This accounts for the narrower line-widths of the OH ab-sorption spectra obtained with the EVN(Table2).The OH emissions in the wing of the absorption at V LSR=910and around V LSR=1260km s?1suggested by Baan&Irwin (1995)coincide with weak emission features seen in our EVN spectrum,although both of them need to be con?rmed in further observations.Two Gaussian components have been ?tted to the absorption pro?les of each of the transition(Fig. 1right);the results are listed in Table2.The resultant opacity and column density are listed in Table3together with those of the previous VLA and VLBI observations ob-tained at di?erent angular scales.The OH absorption veloc-ity center(Gaussian-?tted)is V LSR(OH1)=1011.9±0.9 km s?1and V LSR(OH2)=1113.5±2.0km s?1,while the centre velocities derived from the VLA-A data are V LSR= 1013km s?1and V LSR=1114km s?1(Baan&Irwin1995). These values are consistent within the spectral resolution of one channel.The HI absorptions are peaked at V LSR= 1010km s?1,1120km s?1,and1265km s?1(Baan&Irwin 1995),where the red-shifted third component has no OH counterpart.The ratios of the double peaks in both transi-tions are quite consistent,which may imply that both main-line features are associated with the same continuum source. The line ratio1667to1665MHz in the Gaussian-?tted pro-?les is1.5,which is lower than the Local Thermodynamic Equilibrium(LTE)value of1.8.

Velocity-integrated maps of the1667MHz OH absorption integrated over11.2km s?1intervals and the CLEAN map of a18-cm radio continuum source are shown in Fig. 2(All are uniformly weighted.).Fifteen velocity channel maps cover the velocity range of the absorption from990.4 km s?1to1147.7km s?1.

The continuum emission was not resolved in this experiment. The peak?ux density of the continuum map produced with uniform weighting is8.8mJy beam?1and the rms noise level is0.25mJy beam?1.The peak?ux density and the noise level of the continuum map produced with natural weighting are11.8mJy beam?1and0.085mJy beam?1,respectively. The integrated?ux density of the continuum is14.3±0.7 mJy,while the21-cm integrated?ux density obtained by single-dish measurements is760±31mJy(Condon1983). Hence,more than95percent of the continuum?ux is miss-ing in this EVN experiment.Because the limited angular resolution of our EVN observations compared with the past VLBI observations at higher frequencies,the nuclear con-tinuum remains unresolved without showing any structure. The EVN synthesized beam is larger than the angular sep-aration between the various nuclear components(Fig.7), which limits the discussion on the possible associations of the OH absorbing gas with individual sources in later sec-tions.

A position-velocity(PV)diagram of the1667MHz absorp-tion is presented in Fig.3.A cut was made through the continuum emission at P.A.=15?(rotated clockwise by15 degrees),which aligns a cut along a major axis of the ro-tation traced by CO(1–0)(Sofue&Irwin1992;Koda et al. 2002)and which was also used from the PV diagrams in Baan&Irwin(1995).The gradients in the1665MHz line show the same trend as those seen in the1667MHz line. Note that there are signi?cant position o?sets of~7-8mas in declination between the two peaks of OH1(blue-shifted) and OH2(systemic),which could not be measured in the PV diagrams of Baan&Irwin(1995)at1arcsec resolution with the VLA-A.The velocity contours of OH1and OH2 are separate,as compared with the PV diagrams in Fig.7 in Baan&Irwin(1995).

Fig.4shows a naturally weighted18-cm continuum map superposed on the integrated1667MHz OH intensity(0th moment)map.The absorption is apparently concentrated towards the centre with a weak outward extension.This is similar to the east-west elongation of the H I and OH absorb-ing gas seen at1arcsec resolution and caused by the orien-tation and separation(~25mas or~2pc)of the L-band components E and F(Irwin&Seaquist1988;Baan&Irwin 1995;Middelberg et al.2003).

Fig.5displays the OH velocity?eld of the two components (1st moment).The velocity?eld of OH1varies from south-west to northwest in P.A.=60?,while that of OH2is seen approximately from north to south in P.A.=145?.The ve-locity gradients of OH1and OH2are~10km s?1pc?1 and~13km s?1pc?1,respectively.The kpc-scale CO(1–0)velocity gradient is found to be also in the north-to-south direction and has a value of0.85km s?1pc?1and a P.A.=15?(Sofue&Irwin1992).The compact nuclear CO(1–0)core at PA=4?,tilted by about10?w.r.t the kpc-scale CO(1–0)disk,within central125pc shows a rigid rotation with a velocity of~300km s?1(Sofue et al. 2001;Koda et al.2002).This yields the velocity gradient of 2.4km s?1pc?1.Just as Baan&Irwin(1995)also Sawada-Satoh et al.(2000)have interpreted the equivalent H I components to originate against the two continuum com-ponents A and B,which in reality were components E and F (Middelberg et al.2003).The velocity di?erences estimated from the CO velocity gradients between the two contin-uum components F(A)and E(B),separated by25mas or2pc,are only1.7km s?1or4.8km s?1.Such values are not consistent with the observed velocity di?erence of 101.6km s?1between OH1and OH2,which would suggest that the two detected OH absorptions arise from neither the kpc-scale CO disk nor the nuclear CO core.MERLIN ob-servations at~400mas resolution resolved the rotational motion of the H I absorbing gas(Pedlar et al.1996),which matches the sense of the rotation traced by the H I emission (Irwin&Seaquist1991).The directions of the H I velocity ?elds in the three H I VLBI absorption components have not been presented in Sawada-Satoh et al.(2000),which pre-vents a comparison with the OH velocity?elds in Fig.5. Comparisons of the intensity distribution of OH1and OH2 for each transition are presented in Fig.6.The intensity maps were produced from the uniformly weighted spectral-line cubes.We?nd that there is a di?erence in the spatial

4

distribution in both1667MHz and1665MHz absorption, which is even more distinct in the1665MHz line.The dis-tribution of the blue-shifted OH1is downward w.r.t that of the systemic OH2by approximately8mas,or0.6pc in dec-lination.This trend is also true in the PV diagram of Fig. 3,where the di?erence is clearly seen for both main-lines. 4DISCUSSION

4.1The puzzle of the nuclear radio continuum NGC3079is known to exhibit a complex radio structure in the nuclear region.Recent VLBI studies reveal a mul-tiple milli-arcsec-scale nuclear

source structure with?ve

frequency dependent components(Middelberg et al.2003).

The continuum components A,B and C are only present

in maps at22GHz but are totally obscured at L-band.

The identi?cation of a true nucleus in the galaxy has been

a subject to debate.If the H2O maser structure is indeed

con?rmed to be a disk with a rotation axis misaligned

by15–20?with respect to the kpc-scale CO(1–0)disk

at P.A.=30?,then the nucleus would lie close to the

A-C-B line(Fig.7).Trotter et al.(1998)proposed that

the dynamical center lies between components A and B,

more speci?cally at the intersection of the jet axis and a

north-south distribution of the H2O masers.The north-

south distribution of the H2O maser was more convincingly

con?rmed in Kondratko(2003)by measuring the positions

of the new red-shifted H2O maser features reported in

Hagiwara et al.(2002a).An H2O disk centered close to C

and having a systemic velocity of1120km s?1would be

consistent with the H2O components at1123km s?1at

C,the systemic velocity of the H I absorption of1116

km s?1,and the systemic OH component(OH2)at1114

km s?1.The models based on the velocity signature of

the OH and HI absorption(Baan&Irwin1995)and of

the H I VLBI signature(Sawada-Satoh et al.2000)refer to

components E and F and do not re?ect the location of the

nucleus.Various monitoring programs have been executed

to detect a velocity drift of the systemic components as

seen in NGC4258,but they have failed to?nd any drifting

component to con?rm the disk signature(e.g.Baan&

Haschick1996;Hagiwara et al.2002a).

Considering the available evidence,the true nucleus is

hidden at L-band and lies close to C on the line connecting

components C and B.The nature of the L-band components

E and

F is not yet certain.These components could be

circum-nuclear starburst regions but for that E has too

high a brightness temperature and could be a supernova

remnant.

4.2The nature of the OH absorption

An important result of our EVN observations is the inter-

pretation of the two di?erent velocity?elds traced by the

two OH absorption components on scales of10mas towards

the unresolved radio continuum nucleus.The spectrum in

Fig.1shows that most of the OH absorption has been re-

solved out in our data set,as compared with single-dish and

m

J

y

Vsys

H2O

1190

1123

1020

956

100411301250

HI

LSR velocity (km/s)

800 1000 1200 1400 1600 1800 2000

1

?1

?2

?3

?4

?5

?6

?7

?8

?9

Figure1.left:Hydroxyl absorption spectrum of the1667MHz

and1665MHz transitions,obtained with the EVN in November

2000.An rms noise per channel is0.7mJy beam?1.The velocity

resolution is11.2km s?1.Velocities in the spectra are scaled

in the radio LSR convention.The adopted systemic velocity of

NGC3079is V LSR=1116km s?1.The peak velocities of H I

absorption and H2O maser are denoted by arrows.The results

of2-D Gaussian-?tting of the spectrum are listed in Table1.

right:Gaussian models?tted to the OH spectrum of the1667

and1665MHz transition.The velocity is referenced to V LSR=

1013km s?1,the Gaussian-?tted peak velocity of the1667MHz

OH1in Table2.

5

990.4 KM/S

1001.6 KM/S

1012.9 KM/S

100500?50?100

1024.1 KM/S

1035.3 KM/S

1046.6 KM/S

1057.8 KM/S

1069.1 KM/S

1080.3 KM/S 1091.5 KM/S 1102.8 KM/S 1114.0 KM/S

1125.2 KM/S 1136.5 KM/S 1147.7 KM/S

Continuum

0?100

100100

?100

010050?50?100

0Rel R.A. (mas)

R e l D e c . (m a s )

Figure 2.Velocity channel maps of uniformly weighted continuum-subtracted spectral line cube where the OH absorption at 1667MHz is present.The contour levels are -8,-7,-6,-5,-4,-3,3×0.8mJy beam ?1(1σ).18-cm uniformly weighted continuum map is also shown with the contour levels at -5,5,10,15,20,25,30,35,40×0.25mJy beam ?1(1σ).The center (0,0)position is R.A.(J2000)=10h 01m 57s .805Dec.(J2000)=+55?40’47”.08.The synthesized beams are plotted at the top-left and the bottom-right corner.

V e l o c i t y (L S R , k m /s )

DECLINATION (J2000)

55 40 47.04

47.0547.0647.0747.0847.0947.1047.1147.12

1600

1500

1400

1300

1200

1100

1000

900

800OH1

OH2

OH1

OH2

1665 MHz

1667 MHz

Figure 3.Position-velocity (PV)diagram produced from the nat-urally weighted spectral-line visibility cube,sliced along P.A.=15degrees to align the major axis of the CO rotation (Sofue &Irwin 1992)and the cuts used in PV diagrams in (Baan &Irwin 1995).The contours are plotted at 10,20,30,40,50,60,70,80,90,100percent of the peak intensity of -6mJy beam ?1.The positive contour corresponds to +0.6mJy beam ?1.

-800-600-400-2000

M i l l i A R C S E C

MilliARC SEC

10050

-50-100

100

50

-50

-100

Figure 4.18-cm naturally weighted continuum map (contour)superposed with a 1667MHz OH intensity (0th moment)map (grey-scaled in Jy beam ?1m s ?1).The peak ?ux density of the continuum is 10.3mJy beam ?1.The contours are plot-ted at -5,5,10,20,30,40,50,60,70,80,90,100,110,120×0.085mJy beam ?1(1σ).The synthesized beam (70mas ×39mas,P.A.=–72?)is plotted on the lower left corner.

VLA spectra in Baan &Irwin (1995).This implies that we observe compact gas that extends no more than the beam size of 45mas,or ~4pc,which is almost equal to the size of the background continuum.The OH absorbing gas has been interpreted primarily to be a part of circum-nuclear gas traced by the CO(1–0)emission (Baan &Irwin 1995).A single black body ?t based on the measurement of far-infrared (FIR)?ux densities at 12–100μm in NGC 3079gives a temperature of 43K and the maximum extent of this large-scale FIR source (e.g.torus)of about 130pc.Our EVN data do not provide compelling evidence for such a large-scale torus.4.2.1

The absorbing gas structure

The nucleus of NGC 3079hosts a LINER or a type 2Seyfert nucleus.Therefore,the nucleus could be obscured by an edge-on dusty torus or intervening medium along the line of sight.Because of the systematic similarity of the H I and OH absorptions,Baan &Irwin (1995)suggested a possible association of the two H I and OH absorbing gas components with the double continuum source (Fig.7).With a 1.0arcsec (corresponding to 80pc)resolution of the VLA in A Con?guration,the data in Baan &Irwin (1995)were insu?cient to extract precise connections between these absorption components and the nuclear radio continuum structure.

Interpretation of the VLBI data in Sawada-Satoh et al.(2000)shows that the three H I components are resolved and that the column densities of each component against F

6

980

1000

1020

1040

1060

604020

-40-60-80

60

40

20

-40

-60

1010

1010

10501020

1040103010201010

1100

1120

1140

604020

-40-60-80

60

40

20

-40

-60

110011201110

1130

1140

s )

Figure 5.Mean velocity ?eld (1st moment)of the two 1667MHz OH absorption components.Contours are plotted every 10km s ?1in LSR velocities with the gray scales indicated on the top.left :blue-shifted component (OH1),right :systemic velocity component (OH2).

(blue-shifted H I ),F+E (systemic H I ),and E (red-shifted H I )are almost the same,suggesting that E and F are uniformly obscured by the ISM foreground to the nuclear sources.Our EVN data failed to resolve E from F,so that we cannot identify the individual contributions for OH ab-sorption and the OH column densities for each component with those of H I .Baan &Irwin (1995)introduced a rotating disk model,in which the torus is con?ned in the roughly north-south orientation with an inclination about –20?w.r.t the CO(1–0)disk standing at P.A.=15?.In this discussion,

the jet axis connecting the nuclear radio sources of E and F makes an angle of about 45?with the con?ning torus.The sense of the rotation of the torus proposed in Baan &Irwin (1995)is consistent with that of the CO disk/core and the edge-on H 2O maser disk proposed in Trotter et al.(1998)and Kondratko (2003).Sawada-Satoh et al.(2000)proposed a counter-rotating disk at P.A.=30?in order to explain the spatially resolved H I absorption.However,the directions and position angles of the torus and the CO and H 2O maser disks are completely inconsistent with those of the torus model in Sawada-Satoh et al.(2000)and the evidence for this counter-rotation is weak.

On the other hand,a counter-rotating structure re-versed to the kiloparsec-scale CO(1–0)disk has been ob-served in OH absorption in a type 2Seyfert galaxy NGC 5793(Hagiwara et al.2000).Similarly,two nuclear disks with radii ~100pc embedded in the outer kpc-scale gas disk have been clearly resolved in CO(2–1)in the merg-ing nuclei of the ultra-luminous infrared galaxy Arp 220(IC 4553).The rotational sense between these two disks is reversed due to the counter rotation of the two nuclei them-selves (Sakamoto et al.1999).4.2.2

Interpretations of the double OH peaks

The OH absorption in the EVN spectrum shows double peaks with a separation of 101.6km s ?1in velocity (Table 2).Baan &Irwin (1995)considered the association of OH1and OH2with the radio twin peaks E and F (Fig.7)with an OH velocity gradient between the two components along the jet axis of approximately 50cos φkm s ?1pc ?1.The pro-jected separation of the twin radio peaks E and F is ~25mas or 2pc and the projected angle (φ)lies between the radio axis E–F and the plane of the OH gas motion.The value of the gradient would be 10times larger than that of the CO(1–0)kpc-scale disk and the CO(1–0)nuclear core rotating in the north-south direction,unless cos φis unrealistically small.Consequently,it is not plausible to correlate the OH velocity di?erence between E and F and the internal velocity gradients of the CO(1–0)disk/core,although the direction of the OH velocity gradients agrees with that of the larger-scale CO(1–0)disk.This velocity gradient would also be about 10times larger than those seen in the merging galaxy Arp 220(Mundell,Ferruit &Pedlar 2001)and the Seyfert galaxies Mrk 231and Mrk 273(Kl¨o ckner,Baan,&Garrett 2003;Kl¨o ckner &Baan 2003),where HI absorption or OH maser emission reveal a rotating molecular torus with an in-ner radius of several tens of parsecs from the central engine.Considering the available evidence,the double peaks must arise from two kinematically independent systems in the nu-clear region.The direction of the velocity gradient of the sys-temic OH2is in good agreement with that of the CO and H 2O maser disks.In addition,the velocity range of OH2nearly coincides with the systemic velocity of NGC 3079.We know that the H I absorption at the systemic veloc-ity is seen against the whole nuclear continuum at 21-cm (Sawada-Satoh et al.2000).This can be accounted for by clumpy gas in a kiloparsec-scale disk.Accordingly,the sys-temic OH2probes gaseous components in the CO disk,and particularly the inner disk on scales of 10-100pc.

On the other hand,the velocity ?eld of the blue-shifted

7

OH1is very di?erent from that of OH2and of the CO disk. Baan&Irwin(1995)argued that a foreground and an ex-panding shell driven by the nuclear superbubble could ex-plain the large blue-shift of the OH1centroid velocity,while a receding shell may account for the other weak red-shifted OH absorption component around V LSR=1260km s?1,as observed with VLA.In the starburst galaxy M82,distinct out?ow components of molecular gas were discovered that extend over500pc above the plane of the disk and lie along the minor axis of the galaxy(e.g.Nakai et al.1987).

It is uncertain whether or not the out?ows traced by OH1 are associated with the(possible)starburst-related com-ponents E and F themselves because the velocity gradi-ent of OH1does not align with this E–F axis(Fig.7).It should be noted that the velocity range of OH1(V LSR= 956–1050km s?1)overlaps with that of several blue-shifted H2O features peaking at velocities V LSR=1012,1018,1034, and1035km s?1(Trotter et al.1998).Although these fea-tures have a velocity signature similar to that of the OH1 absorption component,there is no evidence for a physical as-sociation.Some of the H2O maser components outside the edge-on masering disk of the Circinus galaxy are interpreted as molecular out?ow components,ejected from the edge-on structure(Greenhill et al.2003).

The two OH emission features identi?ed at V LSR=910 km s?1and1230km s?1in Baan&Irwin(1995)and de-tected weakly in our data may be accounted for by the blue-shifted1667MHz and1665MHz OH features that lie at ve-locities just below the OH1out?ow components.They may arise in the shocked molecular components foreground to the out?ow structure.In the above picture,an association with molecular out?ows can explain the blue-shifted OH1,the weak OH emission,and also possibly the H2O maser com-ponents.However,this last association cannot be quanti-?ed due to di?culties in comparing the distribution of weak H2O masers and the OH absorption on very di?erent angu-lar scales.

5CONCLUSIONS AND SUMMARY

The broad OH absorption towards the nuclear continuum source in NGC3079was imaged using VLBI techniques.The EVN observations reveal two kinematically independent OH absorption components,where OH1is the blue-shifted ab-sorption,and OH2is the absorption at the systemic velocity of the galaxy.The understanding of the kinematics of these components is limited by the fact that the18-cm background continuum and the OH components have not been spatially resolved.

1.The OH2component shows a distinct velocity gradient in roughly north-to-south direction viewed against the un-resolved nuclear continuum at18-cm consisting of compo-nents E and F.This velocity gradient of the systemic OH2 is consistent with that of the clumpy gas component in the kpc-scale CO disk(P.A.=15?).

2.The blue-shifted component OH1appears to be asso-ciated with molecular out?ows and is possibly associated with the nuclear super-bubble due to circum-nuclear star-burst activity at components E and F and possibly at other locations.The velocity gradient of the OH1component is not yet understood butis almost reversed from that of the systemic component OH2.The weak OH emission features marginally detected in our EVN observation may also be associated with shocked gas in these nuclear out?ows.

3.The H2O masers in the galaxy have been attributed to

a compact disk structure around the true nucleus located west of the radio continuum component C.The OH1veloc-ity range is similar to that of the blue-shifted H2O maser components and there may be some relation.

The combined knowledge of our OH data,of previous VLBI data on the radio continuum,of H2O masers,and of the HI absorption on scales of100pc down to sub-parsec scales,provides a clear and consistent interpretation of the various spectral components seen at the nucleus.VLBI ob-servations at even higher angular resolution and high sensi-tivity could resolve the nature of the OH and H I components in relation to H2O maser components in the nuclear region of NGC3079.

ACKNOWLEDGMENTS

This research has made use of the NASA/IPAC Extragalac-tic Database(NED),which is operated by the Jet Propul-sion Laboratory,California Institute of Technology,under contract with the National Aeronautics and Space Admin-istration.The European VLBI Network is a joint facility of European,Chinese,South African and other radio astron-omy institutes funded by their national research councils. We thank the anonymous referee for useful comments. REFERENCES

Baan W.A.,1989,ApJ,338,804

Baan W.A.,Irwin J.A.,1995,ApJ,446,602

Baan W.A.,Haschick A.,1996,ApJ,473,269

Braatz J.A.,Wilson A.S.,Henkel C.,1996,ApJS,106,51 Cecil G.,Bland-Hawthorn J.,Veilleux S.,Filippenko A.V., 2001,ApJ,555,338

Cecil G.,Bland-Hawthorn J.,Veilleux S.,2002,ApJ,576, 745

Condon J.J.,1983,ApJS,53,459

Ford H.C.,Dahari O.,Jacoby G.H.,Crane P.C.,Ciardullo R.,1986,ApJ,311,L7

Greenhill L.J.,Booth R.S.,Ellingsen S.P.,et al.,2003,ApJ, 590,162

Hagiwara Y.,Diamond P.J.,Nakai N.,Kawabe,R.,2000, A&A,360,49

Hagiwara Y.,Henkel C.,Sherwood W.A.,Baan W.A., 2002a,A&A,387,29L

Hagiwara Y.,Henkel C.,Sherwood W.A.,2002b,in Mi-genes V.,Reid M.J.,eds,IAU Symp.206,Cosmic Masers: From Protostars to Black Holes,p.392

Haschick A.D.,Baan W.A.,1985,Nature,314,14 Haschick A.D.,Baan W.A.,Schneps M.H.,Reid M.J., Moran J.M.,G¨u sten R.,1990,ApJ,356,149

Heckman T.M.,1980,A&A,87,152

Henkel C.,G¨u sten R.,Downes D.,Thum C.,Wilson T.L., Biermann P.,1984,A&A,141,L1

Herrnstein J.R.,Greenhill L.J.,Moran J.M.,Diamond P.J., Inoue M.,Nakai N.,Miyoshi M.,1998,ApJ,497,L69 Irwin J.A.,Seaquist E.R.,1988,ApJ,335,658

8

Irwin J.A.,Seaquist E.R.,1991,ApJ,371,111

Irwin J.A.,Sofue Y.,1992,ApJ,396,L75

Kl¨o ckner H-R,Baan W.A.,Garrett M.A.,2003,Nature, 421,821

Kl¨o ckner H-R,Baan W.A.,2004,A&A,in press

Koda J.,Sofue Y.,Kohno K.,Nakanishi H.,Onodera S., Okumura S.K.,Irwin J.A.,2002,ApJ,573,105 Kondratko P.T.,2003,research exam paper,Harvard Uni-versity

Middelberg E.,Kurichbaum T.P.,Roy A.L.,Witzel A., Zensus J.A.,2003,in Romney J.D.,Reid M.J.,eds,Proc. Future Directions in High Resolution Astronomy:A Cele-bration of the10th Anniversary of the VLB.Astron.Soc. Pac.,San Francisco,in press

Mundell C.G.,Ferruit P.,Pedlar A.,2001,ApJ,560,168 Nakai N.,Hayashi M.,Handa T.,Sofue Y.,Hasegawa T., Sasaki M.,1987,PASJ,39,685

Pedlar A.,Mundell C.G.,Gallimore J.F.,Baum S.A., O’Dea C.P.,1996,Vistas Astron.,40,91

Sakamoto K.,Scoville N.Z.,Yun M.S.,Crosas M.,Genzel R.,Tacconi L.J.,1999,ApJ,514,68

Sawada-Satoh S.,Inoue M.,Shibata K.M.,Kameno S.,Mi-

genes V.,Nakai N,Diamond P.J.,2000,PASJ,52,421 Sawada-Satoh S.,Inoue M.,Shibata K.M.,Kameno S., Nakai N.,Migenes V.,Diamond P.J.,2001,in Schilizzi R.T.,Vogel S.N.,Paresce F.,Elvis M.S.,eds,Proc.IAU Symp.205,Galaxies and their Constituents at the Highest Angular Resolutions,p.196

Sofue Y.,Irwin J.A.,1992,PASJ,44,353

Sofue Y.,Koda J.,Kohno K.,Okumura S.K.,Honma M., Kawamura A.,Irwin J.A.,2001,ApJ,547,L115

Trotter A.S.,Greenhill L.J.,Moran J.M.,Reid M.J.,Irwin J.A.,Lo K.-Y.1998,ApJ,495,740

This paper has been typeset from a T E X/L A T E X?le prepared by the author.

Rel R.A. (mas)

R

e

l

D

e

c

.

(

m

a

s

)

1667 MHz

OH1

OH2

40200?20?40?60 50

40

30

20

10

?10

?20

?30

?40

?50

Rel R.A. (mas)

R

e

l

D

e

c

.

(

m

a

s

)

OH1

OH2

1665 MHz

40200?20?40?60 50

40

30

20

10

?10

?20

?30

?40

?50

https://www.sodocs.net/doc/241789317.html,parison of the intensity distribution of the OH double peaks OH1and OH2,integrated over16velocity channels (corresponding to~180km s?1)for each peak.Contour levels

run from-2to-1.2mJy per beam in steps of0.2mJy per beam. The synthesized beam is included at the left corner of each plot. left:1667MHz,right:1665MHz.The dashed and solid lines show

OH1(blue-shifted absorption)and OH2(near systemic velocity), respectively.

9

Figure7.Schematic views of the nuclear region of NGC3079.The sizes of components are drawn roughly in scale.Oval indicates the outline of the unresolved continuum nucleus in Fig.2.Upward thin arrow denotes the direction of OH1velocity gradient against E and F,both of them are not resolved in our EVN observations.Downward thick arrow shows the sense of the OH2velocity gradient that is roughly consistent with the kpc-scale CO rotation,indicated by thin dashed arrows.A thin dashed line denotes the axis of the kiloparsec-scale super-bubble consisting of wide-angle out?ows ejected from the nuclear region(Cecil et al.2001).Inset displays the innermost nuclear region of NGC3079,where three radio sources A,B,and C are aligned along P.A.=55?.The weak component C may be shaded by the inclined torus or disk traced by the H2O masers(Trotter et al.1998).Components E and F are visible at lower frequencies at from1.7to5.0GHz(Kondratko2003;Middelberg et al.2003)and may result from circum-nuclear starburst.The component labels are adopted from the nomenclature de?ned in Irwin&Seaquist(1988)and Middelberg et al.(2003).Components E and A show similar spectral characteristics,that are di?erent from other components.The spectra of E and A are inverted at2.3and 5.0GHz,respectively(Middelberg et al.2003).

Table2.Gaussian-?tted parameters of the OH absorption and emission spectra.VLA data are from Baan&Irwin(1995).Vc refers a line-peak velocity.Sp and?V are peak?ux density and linewidth(FWHM),respectively.The EVN measurements of the OH emission are tentative.

OH ABSORPTION

OH1(1667MHz)101322841011.9±0.98.5±0.438.7±2.2 (1665MHz)135415-1361.6±1.3 6.1±0.437.2±3.1 OH2(1667MHz)111427921113.5±2.0 5.2±0.372.1±5.0 (1665MHz)146218-1450.7±2.9 3.6±0.367.8±7.3

OH EMISSION

(1667MHz)910 4.761(~900)(0.5)-

(1665MHz)1260 1.7-(~1235)(0.5)-

EVN Global-VLBI VLA(A Con?guration)

OH1(Blue-shift)OH2(Systemic)OH1OH2OH1OH2

相关主题