0801.1494/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: \def\etals{{ et al. }\rm}
4: \def\ga{\mathrel{\raise0.35ex\hbox{$\scriptstyle >$}\kern-0.6em
5: \lower0.40ex\hbox{{$\scriptstyle \sim$}}}}
6: \def\la{\mathrel{\raise0.35ex\hbox{$\scriptstyle <$}\kern-0.6em
7: \lower0.40ex\hbox{{$\scriptstyle \sim$}}}}
8: \def\co{CO {\it J}=1-0 }
9: \def\cothree{CO {\it J}=3-2 }
10: 
11: \newcommand{\mum}{\,\mu{\rm m}}
12: 
13: \shorttitle{Molecular gas in HE\,0450-2958}
14: \shortauthors{Papadopoulos, Feain, Wagg, \& Wilner}
15: \begin{document}
16: 
17: \title{A new twist to an old story: HE\,0450-2958, and the
18: ULIRG$\rightarrow $(optically bright QSO) transition hypothesis}
19: 
20: \author{Padeli \ P.\ Papadopoulos}
21: \affil{Argelander-Institut f\"ur Astronomie,  Auf dem H\"ugel 71,  D-53121 Bonn, Germany}
22: \email{padeli@astro.uni-bonn.de}
23: 
24: \author{Ilana J. Feain}
25: \affil{CSIRO Australia Telescope National Facility, P.O. Box 76, Epping, NSW 1710, Australia}
26: \email{Ilana.Feain@CSIRO.AU}
27: 
28: 
29: \author{Jeff Wagg}
30: \affil{NRAO, PO Box 0, Socorro, New Mexico, 87801,USA}
31: \email{jwagg@nrao.edu}
32: 
33: \and
34: 
35: \author{David J. Wilner}
36: \affil{Harvard-Smithsonian Center for Astrophysics, Cambridge, MA, 02138, USA}
37: \email{dwilner@cfa.harvard.edu}
38: 
39: %\altaffiltext{1}{Institut f\"ur Astronomie, ETH Zurich, 8093 Z\"urich, Switzerland}
40: 
41: \begin{abstract}
42: 
43: We report on interferometric imaging  of the CO J=1--0 and J=3--2 line
44: emission from the  controversial QSO/galaxy pair HE\,0450--2958.  {\it
45: The  detected CO  J=1--0 line  emission is  found associated  with the
46: disturbed  companion galaxy not  the luminous  QSO,} and  implies $\rm
47: M_{gal}(H_2)\sim (1-2)\times 10^{10}\,  M_{\odot}$, which is $\ga 30\%
48: $ of  the dynamical  mass in its  CO-luminous region.  Fueled  by this
49: large gas  reservoir this galaxy is  the site of  an intense starburst
50: with $\rm  SFR\sim 370\, M_{\odot  }\, yr^{-1}$, placing it  firmly on
51: the  upper  gas-rich/star-forming   end  of  Ultra  Luminous  Infrared
52: Galaxies  (ULIRGs,  $\rm L_{IR}>10^{12}\,  L_{\odot  }$).  This  makes
53: HE\,0450--2958 the  first case of  extreme starburst and  powerful QSO
54: activity, intimately  linked (triggered  by a strong  interaction) but
55: not coincident.  The lack of CO emission towards the QSO itself renews
56: the controversy regarding its host  galaxy by making a gas-rich spiral
57: (the  typical  host  of  Narrow  Line  Seyfert~1  AGNs)  less  likely.
58: Finally,  given  that  HE\,0450--2958  and similar  IR-warm  QSOs  are
59: considered   typical   ULIRG$\rightarrow   $(optically   bright   QSO)
60: transition candidates, our results raise the possibility that some may
61: simply  be  {\it  gas-rich/gas-poor (e.g.   spiral/elliptical)  galaxy
62: interactions} which ``activate'' an optically bright unobscured QSO in
63: the gas-poor  galaxy, and a starburst  in the gas-rich  one.  We argue
64: that such  interactions may  have gone largely  unnoticed even  in the
65: local  Universe   because  the  combination  of   tools  necessary  to
66: disentagle the progenitors (high  resolution and S/N optical {\it and}
67: CO imaging) became available only~recently.
68: 
69: \end{abstract}
70: 
71: \keywords{galaxies: active  --- galaxies: ISM --- galaxies: starburst --- ISM: molecules
72: --- quasars: individual (HE\,0450-2958) --- galaxies: mergers}
73: 
74: 
75: 
76: \section{Introduction}
77: 
78: 
79:  HE\,0450-2958  is an optically  bright ($\rm  M_v=-25.8$) IR-selected
80:  QSO at a redshift of z=0.286 (Low et al. 1988), and one of the few in
81:  the  local Universe  whose position  in  the (60$\mu  $m, 25$\mu  $m)
82:  versus (100$\mu $m, 60$\mu $m) far-IR color-color diagram lies in the
83:  area between Ultra Luminous  Infrared Galaxies (ULIRGs) and optically
84:  bright QSOs defined  by $\rm -0.8 < \alpha (100, 60)  < 0.4$ and $\rm
85:  -2.0 <  \alpha (60, 25)  < - 0.8$  (Canalizo \& Stockton  2001).  Its
86:  complex  environment, marked  by tidal  interactions,  was identified
87:  early  with ground-based  (Hutchings,  \& Neff  1988)  and {\it  HST}
88:  imaging  (Boyce et  al.  1996).   These efforts  revealed  a strongly
89:  interacting system consisting of the QSO and a disturbed galaxy $\sim
90:  1.5''$ away, setting this QSO/galaxy  pair on par with typical ULIRGs
91:  in terms  of morphology, and  IR luminosity ($\rm  L_{IR}\sim 5\times
92:  10^{12}\, L_{\odot}$).
93: 
94: The  scarcity  of such  IR  {\it  and}  optically luminous  QSOs  made
95:  HE\,0450-2958   an  early   favorite  candidate   for   undergoing  a
96:  ULIRG$\rightarrow $(optically  bright QSO) transition  (Hutchings, \&
97:  Neff 1988; Canalizo \& Stockton 2001) in a scenario first outlined by
98:  Sanders  et   al.   (1988).   The  latter  was   proposed  after  FIR
99:  color-color diagrams have  identified a population of dust-enshrouded
100:  QSOs via  their AGN-heated  dust component (de  Grijp, Miley,  \& Lub
101:  1987), and  it was motivated  by the similar  bolometric luminosities
102:  and  space densities  of optically  powerful QSOs  and ULIRGs  in the
103:  local  Universe.   It involves  two  gas-rich  galaxies whose  strong
104:  dynamical  interaction induces  a  starburst and  fuels  an AGN  {\it
105:  within} heavily dust-obscured  environments.  Then, as star formation
106:  uses  and  disperses  their  large molecular  gas  reservoirs,  their
107:  initially cool  Spectral Energy  Distribution (SED) of  dust emission
108:  changes  towards one  dominated  by warm  AGN-heated  dust with  much
109:  warmer  IR   colors.   Eventually  an   unobscured  optically  bright
110:  QSO~emerges out of the original  ULIRG.  The large dust and molecular
111:  gas reservoirs found in {\it optically} selected QSOs where they fuel
112:  intense starbursts  (e.g.  Alloin  et al.  1992;  Haas et  al.  2000;
113:  Evans  et al.  2001)  certainly support  an AGN-starburst  link, with
114:  dynamical  interactions/mergers  as   the  likely  trigger  for  both
115:  activities,  suggested also  by  optical studies  (e.g.  Canalizo  \&
116:  Stockton   2001).   In  this   context  the   pivotal  role   of  the
117:  ULIRG$\rightarrow $QSO  transition objects can then  be understood as
118:  the ``markers'' of a special  evolutionary stage during which the QSO
119:  emerges  from a  still ongoing  dusty starburst,  and lasting  only a
120:  fraction of already short  molecular gas consumption timescales ($\rm
121:  \la 10^8\,yrs$).
122: 
123: 
124: HE\,0450-2958  acquired further singular  significance when  Magain et
125: al.   (2005), after  carefully subtracting  the AGN  emission  from an
126: ACS/HST  image, failed  to  find  the QSO  host  galaxy expected  from
127: well-established  and  tight   correlations  of  (host  galaxy)-quasar
128: properties (e.g.   McLure \& Dunlop  2002; Floyd et al.   2004).  This
129: potentially  pivotal result  has been  upheld by  independent analysis
130: (Kim  et  al.  2007),  and  follow-up  VLT spectroscopic  observations
131: weakened the  hypothesis for a significantly  dust-enshrouded QSO host
132: galaxy  (Letawe  et  al.    2007).   The  possibility  of  an  ejected
133: (black-hole)+(fuel)  system as  responsible for  a ``naked''  QSO, its
134: serious consequences and observable  signatures have been discussed in
135: several papers.  Summarizing their  results here, such a configuration
136: can be produced either by a Newtonian three-body ``kick'' exerted on a
137: lighter  black hole  by a  black hole  binary residing  in  a gas-poor
138: elliptical (Haehnelt, Davies, \& Rees  2005; Hoffman \& Loeb 2006), or
139: by the recoil of a coalesced  binary of {\it spinning} black holes due
140: to an  asymmetric gravitational wave emission (Haehnelt  et al.  2005;
141: Loeb  2007).  The  stage for  both  scenaria demands  a strong  galaxy
142: interaction,  and both have  important consequences  for gravitational
143: wave  detections by  future  detectors  such as  {\it  LISA}, and  the
144: hierarchical build-up  of supermassive  black holes in  galaxy centers
145: (Haehnelt et al.  2005 and references~therein).
146: 
147: 
148: In the  present work we report  on the observations of  CO J=1--0, the
149: prime molecular line  used to trace metal-rich H$_2$  gas in galaxies,
150: using the  Australian Telescope Compact  Array (ATCA), and  the J=3--2
151: line  with the Smithsonian  Submillimeter Array  (SMA).  The  paper is
152: organized  as  follows:  a)  we  present the  observations  and  their
153: analysis (section 2),  b) estimate the molecular gas  and dust mass of
154: the system and discuss the properties of the companion galaxy (Section
155: 3),  c)  present the  implications  regarding  the  AGN and  its  host
156: (section   4),  and   d)  discuss   possible  ramifications   for  the
157: ULIRG$\rightarrow  $QSO evolutionary  scheme (section  5).  Throughout
158: this     work      we     adopt     a      cosmology     with     $\rm
159: H_{\circ}=71\,km\,s^{-1}\,Mpc^{-1}$,  $\rm  \Omega  _M=0.27$ and  $\rm
160: \Omega  _{\Lambda}=0.73$,   for  which  the   luminosity  distance  of
161: HE\,0450-2958  at  z=0.286  is   $\rm  D_{L}=1460.4\,  Mpc$  and  $\rm
162: 1''\rightarrow 4.28\,kpc$.
163: 
164: 
165: 
166: \section{Observations, data reduction, and imaging}
167: 
168: \subsection{The ATCA observations}
169: 
170: We  used the  ATCA to  image  the CO  J=1--0 line  emission ($\rm  \nu
171:  _{rest}=115.2712\, GHz$) towards  the QSO/galaxy system HE\,0450-2958
172:  during  four periods  in  April,  August 2006  and  May, August  2007
173:  utilizing the  hybrid H\,214 and H\,168  configurations which combine
174:  antennas on  both the  east-west track and  north track  with maximum
175:  baselines   of  168\,m  and   214\,m  respectively.    These  special
176:  configurations  provide good  brightness sensitivity  while achieving
177:  good u-v  coverage within  short ($\sim $6\,hrs)  observing sessions.
178:  This  is  important  for  mm  observations at the  ATCA  site  where
179:  atmospheric conditions usually limit  useful observing (i.e. with low
180:  system  temperatures  and good  phase  stability)  to time  intervals
181:  significantly  shorter than  typical full  synthesis  sessions ($\sim
182:  $12\,hrs).   For the  mean  redshift $\rm  z  = 0.2864$  of the  pair
183:  (Canalizo \& Stockton 2001), this CO transition is redshifted to $\rm
184:  \nu (z)= 89.608\, GHz$, well  within the tuning range of ATCA's 3\,mm
185:  receivers (available in  5 of the 6 antennas).   The correlator setup
186:  consisted of two IF modules  of $\rm 64 \times 2\,MHz$ channels each,
187:  and frequency resolution of $\rm \Delta \nu _{res}\sim 2.2 \Delta \nu
188:  _{ch}=4.4\, MHz  $.  The effective velocity coverage  was $\rm \Delta
189:  V_{(IF1+IF2)}\sim   (-300\rightarrow  +270)\,  km\,   s^{-1}$,  after
190:  flagging bad edge channels  and channel overlap (=7 channels) between
191:  the two IF modules.  Pointing was checked hourly, and typical offsets
192:  were $\sim  5''$ (rms), while the  field of view of  ATCA antennas at
193:  this frequency is  32$''$ (HPBW).  The array phase  center was placed
194:  at the  AGN's position, marked by  its radio core at  8.6~GHz (the C1
195:  component in Feain et al.  2007).
196: 
197: 
198:   Dedicated wideband  mm continuum observations  were conducted during
199:  periods in May, August 2006, and  August 2007 using the two IFs tuned
200:  in series around a center frequency $\rm \sim 94.53\, GHz$ (well away
201:  from the CO line), and the same array configurations (two tracks with
202:  H\,214, three tracks with  H\,168).  For these observations the phase
203:  center was  placed $3''$ away  from the QSO  as to avoid  any (highly
204:  unlikely)  small correlator DC  offsets masking  as a  weak continuum
205:  source at the phase  center.  Typical system temperatures during line
206:  and  continuum  observations   were  $\rm  T_{sys}\sim  (200-350)\,K$
207:  (including  atmospheric  absorption),   estimated  with  hourly  vane
208:  calibration performed per~antenna.
209: 
210: 
211: Amplitude/phase  calibration was  obtained with  0454-234 observations
212: interleaved  with  HE\,0450-2958  in 1min/(3-5)mins  calibrator/source
213: intervals, and passband calibration was achieved by observing 1921-293
214: or 2223-052.  The absolute flux density scale was set by bootstrapping
215: that of 0454-234  from Uranus and/or Mars observations  in each track.
216: These yielded $\rm  S_{3\,mm}(0454-234)=(1.9\pm 0.28)\, Jy$, where the
217: $\sim 15\%$ error  represents the flux scale uncertainty  of the data,
218: obtained as the dispersion of $\rm S_{3\,mm}(0454-234)$ over the 2006,
219: 2007 observing periods  (and thus is an upper  limit since it contains
220: also any real source~variability).
221: 
222: 
223: \subsection{The SMA observations}
224: 
225: A  search  for  the  redshifted  CO J=3--2  line  emission  ($\rm  \nu
226:  _{rest}=345.796\,   GHz$)   was   conducted   using   the   8-element
227:  SMA\footnote{The Submillimeter  Array is a joint  project between the
228:  Smithsonian  Astrophysical   Observatory  and  the   Academia  Sinica
229:  Institute  of  Astronomy  and  Astrophysics  and  is  funded  by  the
230:  Smithsonian Institution and  the Academia Sinica.}  interferometer on
231:  9th of December 2006 with  a median atmospheric opacity at 225~GHz of
232:  $\tau   _{225}\sim   0.08$.   This   transition   is  redshifted   to
233:  269.102~GHz,  covered by  the  lower sideband  (LSB)  of the  345~GHz
234:  receivers placed at 268.892 GHz,  while the upper sideband was offset
235:  by 10 GHz from this  center frequency.  The bandwidth covered by each
236:  is 2  GHz while the spectral  resolution was 3.25~MHz for  a total of
237:  768 channels.   This was smoothed so  that a single CO  J=3-2 map was
238:  created,  averaged over  a  velocity range  of  570 $\rm  km\,s^{-1}$
239:  centered at the redshift of the CO J=1-0 line emission.
240: 
241:  
242: The  SMA extended  configuration was  used, yielding  a beam  of $\sim
243: 1.20'' \times  0.94''$, PA=16.9$^\circ$ at the  observed frequency and
244: the  u-v coverage  attained.   The mean  system  temperature was  $\rm
245: T_{sys}=490\,K$ (DSB), and the  total on-source integration time $\sim
246: 5.07$ hrs.   Calibration of  the antenna gain  variations was  done by
247: observing the calibrators 0522-364  and 0455-462, at regular 10~minute
248: intervals.  Saturn  and 3C454.3 were  used to calibrate  the bandpass,
249: while  Uranus was  observed  for flux  calibration,  with an  expected
250: uncertainty  of $\sim$20\%.   The MIR  package was  then used  for the
251: calibration of  the visibility~data.   A search for  275~GHz continuum
252: yielded  an  upper  limit   of  $\rm  S_{275\,GHz}\leq  8\,  mJy/beam$
253: ($3\,\sigma$),   over  a   $\rm  \sim   3.936\,GHz$   bandwidth  (both
254: sidebands), while the noise in the line map was $\rm \sigma _{rms}\sim
255: 6.6\,mJy/beam$.
256: 
257: 
258: 
259: \subsection{Data reduction and imaging}
260: 
261: We used  the {\it  in-situ} ATCA phase  monitor operating on  a 230\,m
262: east-west baseline  (Middelberg, Sault, \& Kesteven  2006) to identify
263: the periods when atmospheric phase noise at 90\,GHz over this baseline
264: (comparable  to the longest  present in  our datasets)  exceeded $\sim
265: 30^{\circ}$ (rms)  and rejected the  data taken during  those periods.
266: The remaining  visibilities were then  edited for any  remaining large
267: temporal  amplitude/phase jumps.   The final  $\rm  V(u,v)$ visibility
268: dataset has  a residual phase rms of  $\sigma _{\phi}\sim 20^{\circ}$,
269: in accordance with  a mean coherence factor $\rm  \rho _{coh} =\langle
270: V(u,v)\rangle/S =e^{-\sigma  ^2 _{\phi}(rad)/2} \sim  0.95$, estimated
271: from imaging  the visibilities  of the calibrator  with an  input flux
272: density  S (and  antenna  amplitude/phase solutions  smoothed to  $\rm
273: T_{avg}  \sim   2\times  T_{cycle}(QSO\leftrightarrow  calibrator)\sim
274: 12\,mins$ intervals).  The u-v range of the final dataset is $\rm \sim
275: (12\rightarrow 75)\, k{\lambda}$ for both line and continuum.
276: 
277: 
278: Imaging was done  using MIRIAD's task INVERT with  natural weight used
279:   for maximum point-source sensitivity in our line and continuum maps.
280:   CO J=1--0 emission was detected independently in both the H\,214 and
281:   H\,168 array  datasets, with peak flux densities  differing by $\sim
282:   15\%$, consistent  with the calibration uncertainties.   An image of
283:   the   CO   J=1--0   emission   was   produced   by   multi-frequency
284:   synthesis\footnote{Fourier    transform   of   the    $\rm   V(u,v)$
285:   visibilities {\it per channel}, then averaging the resulting channel
286:   maps}  over  the  whole  band, and  combining  both  configurations.
287:   Deconvolution  using the  Clark algorithm  was then  performed.  The
288:   resulting CO map and mm continuum image of the same region are shown
289:   in Figure 1, and the  spectrum corresponding to the peak CO emission
290:   is shown in Figure 2.
291: 
292: \subsubsection{Source characteristics}
293: 
294: The velocity-averaged CO J=1--0  image yields the highest S/N possible
295: from our data ($\rm S_{peak}/\sigma_{rms}\sim $15, Figure 1), and thus
296: allows optimum determination of the  position of the peak emission and
297: the  source size.   A gaussian  source model  yields an  excellent fit
298: (i.e. after subtraction of the  model, the noise in the residual image
299: over the source area is $\rm \sigma _{rms} =0.45\,mJy/beam$, as in the
300: rest   of  the   map  area)   where  the   peak  CO   brightness  $\rm
301: S_{peak}=(6.7\pm  0.45)\,  mJy/beam$ is  located  at $(\Delta  \alpha,
302: \Delta \delta)  = (1.15''\pm  0.12'', -1.05''\pm 0.10'')$  relative to
303: the AGN's radio core at the phase center.  The small positional errors
304: are those expected  from the high S/N ($\rm  \delta \theta _{rms} \sim
305: 1/2\, \langle  \Theta _{beam} \rangle  (S/N)^{-1}\sim 0.09''$ relative
306: to the  phase center).  The  angular proximity of  the amplitude/phase
307: calibrator  to HE\,0450--2958 ($\rm  |\Delta \vec  k|\sim 6.5^{\circ}$
308: away) allows also for  excellent {\it absolute} astrometry by limiting
309: $\rm  \delta \theta  _{bas} =  (\delta  \vec B\cdot  \Delta \vec  k)/B
310: \approx (\delta \phi  _{bas}/2\pi )\langle \Theta _{beam}\rangle$ (the
311: uncertainty  due  to  the  phase  error $\rm  \delta  \phi  _{bas}\sim
312: (2\pi/\lambda) (\delta \vec B\cdot  \Delta \vec k)$) to $\sim 0.086''$
313: (for typical calibration errors of the baseline length of $\rm |\delta
314: \vec  B|\sim 1\,mm$).  {\it  Thus we  are confident  that the  peak CO
315: emission is not located at  the QSO position but $\sim 1.56''\pm0.17''
316: (\sim 6.7\,kpc)$ away.}
317: 
318: The total  flux density  of $\rm S  ^{(tot)} _{\nu }(CO)=  (10\pm 2)\,
319:  mJy$ emerges for  a source for which the  gaussian fit gives: $\theta
320:  _{maj}=3.95''  \pm 0.30'' $  and $\theta  _{min}=2.80'' \pm  0.22'' $
321:  (PA=$57^{\circ}\pm  9^{\circ} $).   Deconvolving  the restoring  beam
322:  ($3.21''\times 2.14''$, PA=$72^{\circ}$)  yields an intrinsic size of
323:  $2.5''\times  1.5''$  (PA=$35^{\circ}$),  larger  than  any  expected
324:  ``seeing''   disk.   The   latter,   deduced  from   images  of   the
325:  amplitude/phase calibrator, is  $\sim 0.30''-0.45''$ (consistent with
326:  the  expected   $\rm  \Delta  \theta   _{seeing}\sim  [\sigma  _{\phi
327:  }(rad)/2\pi]  \sqrt{8\,  ln2}\,  \langle \Theta  _{beam}\rangle  \sim
328:  0.34'' $,  for a  residual $\rm \sigma  _{\phi }\sim  20^{\circ}$ and
329:  $\rm  \langle \Theta  _{beam}\rangle  =\sqrt{\Theta _1\Theta  _2}\sim
330:  2.62''$).
331: 
332: No continuum emission is detected at 94.5\,GHz in either the starburst
333:  or the  QSO position ($\rm  S_{94\,GHz}<1.4\,mJy$ (3$\sigma$), Figure
334:  1).  All  relevant data, and the  characteristics of the  CO and dust
335:  emission in HE\,0450--2958 are summarized in Table 1.
336: 
337: 
338: \section{The molecular gas in HE\,0450--2958}
339: 
340: The CO J=1--0  emission overlaid with the 8.4\,GHz  radio continuum is
341: shown in  Figure~2 where  a good correspondence  with the  region C\,2
342: associated with the companion galaxy  (Feain et al.  2007) is evident.
343: Regriding the observed CO emission and its deconvolved model, and then
344: overlaying  them  onto the  HST/ACS  optical  frame of  HE\,0450--2958
345: further underlines  its association  with the companion  galaxy rather
346: than the QSO (Figure 4).
347: 
348: 
349: 
350:   The  corresponding  H$_2$  mass   is  given  by  $\rm  M(H_2)=X_{CO}
351: L_{CO(1-0)}$, where  $\rm L_{CO(1-0)}=  \int _{\Delta V}  \int_{A_s} T
352: _{b}\,dA\,dV  $   is  the  velocity/area-integrated   line  brightness
353: temperature at the  source reference frame, and $\rm  X_{CO}$ (in $\rm
354: M_{\odot  }  (K\,km\,s^{-1}\,   pc^2)^{-1}$  units)  is  the  CO-H$_2$
355: luminosity-mass conversion  factor.  Using standard  derivations (e.g.
356: Solomon et al.  1997) we have
357: 
358: \begin{equation}
359: \rm L_{CO(1-0)} =  3.25\times 10^7\,(1+z)^{-1} 
360: \left(\frac{\nu _{co,rest}}{GHz}\right)^{-2} \left(\frac{D_L}{Mpc}\right)^2
361:  \left(\frac{\int S_{CO(1-0)} dV}{Jy\,km\,s^{-1}}\right)\, K\, km\,s^{-1}\,pc^2.
362: \end{equation}
363: 
364: 
365: \noindent
366: For  the   observed  $\rm   \int  S_{CO(1-0)}\,dV$  (Table   1),  $\rm
367: D_L=1460.4\,Mpc$  and  $\rm \nu  _{co,rest}=115.271\,  GHz$ (the  line
368: rest-frame  frequency),  the  latter  gives  $\rm  L_{CO(1-0)}=(2.3\pm
369: 0.46)\times 10^{10}\, K\,  km\, s^{-1}\, pc^2$.  This is  close to the
370: maximum CO luminosity observed for starbursts, beyond which it remains
371: constant  across redshift  (Frayer  et al.   1999;  Evans, Surace,  \&
372: Mazzarella 2000),  a possible indication of  a self-regulating process
373: occuring in extreme (maximal?)  starbursts.  Using $\rm X_{CO}\sim 1\,
374: M_{\odot}  (K\,km\,s^{-1}\,pc^2)^{-1}$   (deduced  for  local  ULIRGs,
375: Downes  \& Solomon  1998), we  obtain $\rm  M_{gal}(H_2)\sim 2.3\times
376: 10^{10}\, M_{\odot}$,  {\it which places the companion  galaxy in this
377: galaxy/QSO pair  firmly in the  ULIRG category,} at its  most gas-rich
378: upper  end,  and  similar   to  starbursts  found  at  high  redshifts
379: (e.g. Greve et al.~2005).
380: 
381: 
382: A lower limit  in the H$_2$ gas mass can be  obtained by assuming that
383: the $ ^{12}$CO J=1--0 emission is optically thin. Using the derivation
384: in  Bryant \&  Scoville (1996)  and assuming  LTE, the  new conversion
385: factor becomes
386: 
387: \begin{equation}
388: \rm \frac{M(H_2)}{L_{CO(1-0)}} \sim 0.08 \left(\frac{[CO/H_2]}{10^{-4}}\right)^{-1}
389: \left[\frac{g_1}{Z}e^{-T_{\circ}/T_k} \left(\frac{J(T_k)-J(T_{bg})}{J(T_k)}\right)\right]^{-1} 
390: \frac{M_{\odot }}{K\,km\,s^{-1}\,pc^2},
391: \end{equation}
392: 
393: 
394: \noindent
395: where     $\rm    T_{\circ}     =     E_1/k_B\sim    5.5\,K$,     $\rm
396: J(T)=T_{\circ}\left(e^{T_{\circ}/T}-1\right)^{-1}$,                $\rm
397: T_{bg}=(1+z)T_{cmb}\sim  3.5\,K$ (CMB  temperature  at z=0.286),  $\rm
398: g_1=3$ (degeneracy factor of  n=1 level), $\rm Z\sim 2(T_k/T_{\circ})$
399: (partition  function),  and  $\rm  [CO/H_2]\sim  10^{-4}$  (for  solar
400: metallicity).   For  typical  star  forming  gas  where  $\rm  T_k\sim
401: 40-60\,K$,   the  last   equation  yields   $\rm   \langle  X^{(thin)}
402: _{CO}\rangle        _{T_k}        \sim       0.55\,        M_{\odot}\,
403: (K\,km\,s^{-1}\,pc^2)^{-1}$,  and  thus  $\rm M_{gal}(H_2)  _{min}\sim
404: 1.25\times 10^{10}\,M_{\odot }$.
405: 
406: \subsection{Constraints on the molecular gas excitation} 
407: 
408: 
409:  The CO  J=3--2 observations  yield a limit  of $\rm  S_{CO(3-2)}\la 3
410: {N_b}^{1/2}   \delta  S_{CO(3-2)}\Delta   V  $,   where   $\rm  \delta
411: S_{CO(3-2)}\sim 6.6\,  mJy/beam $ is the  noise of the  CO J=3--2 map,
412: $\rm N_{b}\sim 3.3$ is the number of SMA beams corresponding to the CO
413: J=1--0 emitting  region and $\rm \Delta  V\sim 570\,km\,s^{-1}$.  This
414: gives   $\rm  S_{CO(3-2)}\la   20\,  Jy\,   km\,  s^{-1}$,   and  $\rm
415: L_{CO(3-2)}\la  9\times  10^{9}\,   K\,  km\,  s^{-1}\,  pc^2$,  which
416: corresponds to  a (3--2)/(1--0)  brightness temperature ratio  of $\rm
417: R_{32}\la 0.39$.  This  is rather low for molecular  gas in starbursts
418: where  $\rm \langle  R_{32}\rangle \sim  0.65$ (e.g.   Devereux  et a.
419: 1994), though  within the range found  in such galaxies  (e.g.  Yao et
420: al.  2003).   We examined an extensive  grid of models  with our Large
421: Velocity  Gradient (LVG)  code  (based on  work  by Richardson  1985),
422: restricted by the upper limit  on $\rm R_{32}$, and the condition $\rm
423: T_{\rm  kin}\geq   T_{dust}\ga  50\,K$  (since  the   gas  cools  less
424: efficiently than  the dust, and  turbulent/photoelectric heating heats
425: the  gas more).   Most  conditions reproducing  this  ratio have  $\rm
426: \langle  n(H_2)\rangle   \sim  (10^2-10^3)\,cm^{-3}$  (volume-averaged
427: H$_2$ gas density),  rather low for star forming  H$_2$ gas.  This may
428: in turn signify  that not all of the  CO(1-0)-emitting gas is involved
429: in star formation (and thus  some of it is not CO(3-2)-bright).  Given
430: the minimum  $\rm R_{32}(min)\sim  0.22$ found in  ULIRGs (Yao  et al.
431: 2003),  deeper SMA  observations of  CO J=3--2  at a  lower resolution
432: (better matching the CO J=1--0 source size) should detect this line in
433: HE\,0450--2958 and shed light on its star-forming molecular gas phase.
434: 
435: 
436: \subsection{The companion galaxy: a typical ULIRG}
437: 
438: The  mm/IR dust  continuum SED  of  HE\,0450--2958 from  IRAS and  our
439: mm/sub-mm data (Figure  5) is typical for warm  ULIRGs and IR-selected
440: QSOs, where the  presence of warm dust was the  prime reason they were
441: considered  as $\rm  ULIRG\rightarrow QSO$  transition objects  in the
442: evolutionary  scenario linking  these two  classes.  A  cool/warm dust
443: emission  SED fit  yields $\rm  T^{(cool)} _{dust}=45-55\,K$  and $\rm
444: T^{(warm)}  _{dust}\sim  175-194\,  K$  (for  emissivities  of  $\beta
445: =1-2$).   Adopting $\beta  =1.5 $  as  our working  value yields  $\rm
446: T^{(cool)} _{dust}=48\,K$, $\rm M^{(cool)} _{dust}\sim 10^8\, M_{\odot
447: }$, and $\rm L_{FIR}=2.1\times 10^{12}\, L_{\odot}$ (the luminosity of
448: the  cool dust  component).   A starburst  in  the gas-rich  companion
449: galaxy  is  a  natural  source  of this  large  far-IR  luminosity  of
450: HE\,0450-2958, with a star formation  rate of $\rm SFR \sim 1.76\times
451: 10^{-10}(L_{IR}/L_{\odot})\,  M_{\odot}\,  yr^{-1}\sim 370\,  M_{\odot
452: }\,  yr^{-1}$,  and  a  star  formation  efficiency  of  $\rm  SFE\sim
453: L_{FIR}/M_{gal}(H_2)\sim  (90-165)\,  (L_{\odot}/M_{\odot})$,  typical
454: values  for ULIRGs (e.g.   Solomon et  al.  1997).   Moreover assuming
455: most of  the cool dust mass  in this system residing  in the companion
456: galaxy,  we obtain  $\rm  M_{gal}(H_2)/M^{(cool)}_{dust}\sim 125-230$.
457: These values are  well within the range found  in IR-luminous galaxies
458: (e.g.  Sanders Scoville, \& Soifer  1991, for the $\rm X_{CO}$ adopted
459: here), adding confidence to the estimated $\rm M_{gal}(H_2)$ as a good
460: measure of the bulk molecular gas mass present in HE\,0450--2958.
461: 
462: 
463: The  results above  are  certainly  in good  accord  with evidence  of
464: significant star formation and  reddening in the companion galaxy, and
465: no such  activity or dust obscuration  evident in a  putative QSO host
466: galaxy (Magain  et al.  2005;  Letawe, Magain, \& Courbin  2007). They
467: contradict the recent view of Kim et al.  2007 regarding the nature of
468: the companion  galaxy, but also  set HE\,0450-2958 apart  from typical
469: ULIRGs with double nuclei.  The  latter have molecular gas mass ratios
470: of $\sim 1:1-2:1$ (Sanders \& Ishida 2004; Evans, Surace, \& Mazzarela
471: 2000),  and their  AGNs (when  present)  reside in  the most  gas-rich
472: member  of the  interacting/merger  pair (e.g.   Evans  et al.   1999,
473: 2002).  Subtraction of  the CO emission model leaves  no residual such
474: emission   in  the   location  of   the  QSO   with   $\rm  S^{(peak)}
475: _{gal}/S_{QSO-host}\ga  5$ ($3\, \sigma$),  which corresponds  to $\rm
476: M_{gal}(H_2)/M_{QSO-host}(H_2) \ga  5:1$ (for a  common velocity range
477: and $\rm X_{CO}$~value).  A different (higher) $\rm X_{CO}$ factor for
478: the  molecular gas  in the  QSO host  could reduce  the aforementioned
479: asymmetry in  H$_2$ gas mass  distribution, while some CO  emission at
480: the QSO's location at velocities outside our velocity coverage of $\rm
481: \sim 570\,km\,s^{-1}$  could be missed.  However  the relatively small
482: velocity range traced by the  ULIRG, the QSO, and the highly excitated
483: ionized  gas  around  it  (Merritt  et  al.   2006)  make  the  latter
484: possibility rather small.  Sensitive wideband observations of e.g.  CO
485: J=3--2  with good  brightness sensitivity  would nevertheless  be very
486: valuable in the search for any QSO-related molecular gas.
487: 
488: 
489: \subsection{The dynamical  mass of the companion galaxy}
490: 
491: The  line  profile towards  the  peak CO  emission  shows  signs of  a
492: rotating disk (Figure~3), though the  lower S/N per channel across the
493: band  makes this far  from certain.   The CO  source model  shows good
494: overall agreement with  the galaxy's optical image (Figure  4), and at
495: $\rm  z=0.2864$  where  $\rm  2.5''\times  1.5''\rightarrow  10.7\,kpc
496: \times  6.4\, kpc$, its  dimensions are  comparable to  the H$_2$-rich
497: parts  of typical  spiral disks  (Regan et  al.  2001;  Helfer  et al.
498: 2003), including the Milky Way, but with $\sim 10-20$ times more H$_2$
499: gas mass.
500: 
501: For  an  underlying  disk   geometry  the  source  angular  dimensions
502: correspond   to   an   inclination   angle  $\rm   cos   (i)=   \theta
503: _{minor}/\theta   _{major}=0.6\Rightarrow   i\sim  53^{\circ}$.    The
504: enclosed dynamical mass then would~be
505: 
506: \begin{equation}
507: \rm M_{dyn}\sim \frac{R}{G} \left(\frac{V^{(obs)} _{rot}}{sin(i)}\right)^2 \sim 2.32\times 10^5
508: \left(\frac{R}{kpc}\right)\left(\frac{V^{(obs)} _{rot}/sin(i)}{km\,s^{-1}}\right)^2\, M_{\odot}.
509: \end{equation}
510: 
511: \noindent
512: For   $\rm   R\sim    R_{major}/2\sim   5.35\,kpc$,   $\rm   V^{(obs)}
513: _{rot}=(100-150)\, km\, s^{-1}$ (see Figure 2), it is $\rm M_{dyn}\sim
514: (1.9-4.4)\times 10^{10}\, M_{\odot}$.  Thus the galaxy's molecular gas
515: mass  amounts to  $\ga 30\%$  of the  dynamical mass  enclosed  in its
516: CO-emitting region, making this ULIRG a very gas-rich system.
517: 
518: 
519: Inadequate resolution and/or low S/N makes it notoriously difficult to
520: discern a  gaseous disk from  other possible configurations,  and this
521: has  led  to ambiguities  regarding  the  presence  of large  gas-rich
522: spirals, especially in the distant  Universe (e.g. Genzel et al. 2003;
523: Tacconi et  al.  2006).  In our  case the resolution  is inadequate to
524: discern   velocity/position  patterns  within   the  source   and  the
525: underlying dynamical configuration  could certainly be different (e.g.
526: two orbiting compact gas  reservoirs).  Nevertheless for the intrinsic
527: size and  velocity width of  our source most  dynamical configurations
528: would  produce similar  or  even smaller  $\rm  M_{dyn}$ values  (e.g.
529: Bryant \& Scoville  1996; Genzel et al.  2003), and  thus the ULIRG in
530: HE\,0450-2958 would remain a very gas-rich~object in terms of its $\rm
531: M_{gal}(H_2)/M_{dyn}$ ratio.
532: 
533: 
534: 
535: 
536: \section{Implications for the QSO's host galaxy and its AGN}
537: 
538:    A  suggestion for  a Narrow  Line Seyfert~1  (hereafter  NLSy1) AGN
539: (which have  lower mass black holes),  made by Merrit  et al.  (2006),
540: could reconcile the lack of a massive spheroid host galaxy reported by
541: Magain et  al.  (2005).  However this  rests on the weak  premise of a
542: good analogy between luminous QSOs  and the much less luminous Seyfert
543: galaxies, unsupported by the fact that NLSy1 AGNs are typically hosted
544: by gas-rich spirals with luminous starbursts, while optically luminous
545: QSOs with  $\rm M_v<-24$ reside mostly in  gas-poor massive elliptical
546: hosts (Floyd  et al.  2004).  For example  in I Zw 1  (the closest QSO
547: and a prototypical NLSy1), a  spiral disk hosts a massive cirumnuclear
548: starburst at a radius of  $\rm \sim 1.9\, kpc$ (Schinnerer, Eckart, \&
549: Tacconi 1998;  Staguhn et al.  2004)  fueled by a  large molecular gas
550: reservoir.  {\it None of that is evident in the vicinity of the AGN in
551: the  HE\,0450-2958  system}  whose  bulk  of  the  molecular  gas  and
552: starburst  activity  are  found   in  the  companion  galaxy  instead.
553: Indicatively, at  z$\sim 0.286$ our CO 1--0  observations could detect
554: the molecular gas  mass of the $\sim 10$ times  less far-IR luminous I
555: Zw  1 at  a S/N$\sim  5$.  Finally,  well-established  {\it empirical}
556: AGN-(host  galaxy)   correlations  also  predict   a  hitherto  absent
557: prominent  elliptical (Magain  et al.~2005),  and thus  this important
558: issue remains~open.
559: 
560: 
561: \subsection{The AGN-related molecular gas: could it be ejected?}
562: 
563: The  two-component dust continuum  SED fit  (section 3.2)  yields $\rm
564: T^{(warm)}  _{dust}(AGN)=184\,K$,   $\rm  M^{(warm)}  _{dust}(AGN)\sim
565: 5\times  10^4\, M_{\odot  }$ and  $\rm  L_{mid-IR}=2.6\times 10^{12}\,
566: L_{\odot}$ (the  luminosity of the corresponding  dust component), but
567: unlike the  typical scenario this  AGN-heated dust reservoir  now lies
568: {\it outside} the  ULIRG where the bulk of the  dust and molecular gas
569: mass  resides.  Using  the same  gas/dust  ratio as  in the  companion
570: galaxy  gives $\rm  M_{AGN}(H_2)\sim (0.6-1.15)\times  10^7\, M_{\odot
571: }$.  This  amounts to the mass of  a few GMCs, and  such molecular gas
572: quantities are  found within  $\rm \la 100\,pc$  of AGNs in  the local
573: Universe  (e.g. in  the Sy2  galaxy NGC  1068, Planesas,  Scoville, \&
574: Myers 1991; Schinnerer, Eckart, \& Tacconi~1999).
575: 
576: The  disturbed nature  of  the HE\,0450-2958  system  and the  extreme
577: starburst in  the gas-rich companion  galaxy $\sim 6.7$ kpc  away from
578: the QSO argues in favor of a strong dynamical interaction.  Could then
579: a  black hole,  along  with  the aforementioned  gas  mass, have  been
580: ejected out  of its  host galaxy  during such an  event?  In  order to
581: answer this  and check the  consistency of such proposals  against the
582: gas mass associated with this particular AGN, an estimate of its black
583: hole mass is first needed. A {\it lower} limit can be derived from the
584: fact  that the  dust within  an  AGN-bound gas  reservoir will  partly
585: obscure it (giving rise to the distinct warm IR colors), and thus
586: 
587: \begin{equation}
588: \rm L^{(AGN)} _{IR}= \frac{4\pi c G m_H\, f_c \epsilon _{Edd} M_{BH}}{\sigma _T}
589: = 3.3\times 10^{12}\, f_c\, \epsilon _{Edd} \left(\frac{M_{BH}}{10^8\,
590:  M_{\odot}}\right)\, L_{\odot}, 
591: \end{equation}
592: 
593: \noindent
594: where $\rm  f_c=L^{(AGN)} _{IR}/L_{AGN}\leq 1$ is the  fraction of the
595: AGN's intrinsic luminosity intercepted  by the dust and re-radiated at
596: IR wavelengths,  $\rm \epsilon _{Edd}=L_{AGN}/L_{Edd}$  ($\rm L_{Edd}$
597: is the Eddington  luminosity limit), $\rm \sigma _{T}$  is the Thomson
598: cross  section, and  $\rm M_{BH}$  the  black hole  mass.  Since  $\rm
599: L^{(AGN)}  _{IR}=2.6\times 10^{12}\,  L_{\odot }$,  the  last relation
600: yields,
601: 
602: \begin{equation}
603: \rm M_{BH}\sim 0.8 (f_c \epsilon _{Edd})^{-1}\, 10^8\,M_{\odot}.
604: \end{equation}
605: 
606: \noindent
607: For $\rm \epsilon _{Edd} \sim 0.8$ (typical for $\rm M_{V}(QSO)<-25 $,
608: Floyd  et  al.  2004),  and  $\rm  f_c=(1+R_{opt,IR})^{-1}\sim 0.71  $
609: (where  $\rm R_{opt,IR}=(L_{opt}/L_{IR})_{AGN}\sim  [\nu  _B f^{(AGN)}
610: _{\nu  }(B)]/[\nu_{60\mu  m}  f^{(AGN)}  _{\nu }(60\mu  m)]\sim  0.4$,
611: Canalizo  \&   Stockton  2001),   it  is  $\rm   M_{BH}\sim  1.4\times
612: 10^8\,M_{\odot }$. This is consistent with the low values advocated by
613: Merritt et al.  (2006), but given  that ours is only a lower limit, it
614: does not  settle the  issue of the  black hole  mass (and thus  of the
615: expected host galaxy) raised by Magain et al.~(2005).
616: 
617: 
618: If such  a black hole  is the recoiling  remnant of a  coalesced black
619: hole binary then, following Loeb 2007,  it can carry a gas disk with a
620: mass of
621: 
622: 
623: \begin{equation}
624: \rm   M_{disk}\sim    3\times   10^5   \alpha^{-0.8}    \eta   ^{-0.6}
625: \left(\frac{M_{BH}}{10^7\,M_{\odot}}\right)^{2.2}
626: \left(\frac{V_{ej}}{10^3\, km\, s^{-1}}\right)^{-2.8}\, M_{\odot},
627: \end{equation}
628: 
629: \noindent
630: where $\alpha \sim  0.1$ is the disk viscosity  parameter, $\rm \eta =
631: (\epsilon  _{BH}/0.1)(L_{AGN}/L_{Edd})\sim 0.8$,  and $\rm  V_{ej}$ is
632: the ejection  velocity of  the recoiled merged  BH product.   For high
633: ejection velocities  of $\rm V_{ej}\ga  1000\,km\,s^{-1}$ (e.g Hoffman
634: \& Loeb 2006) Equation 6 yields the maximum $\rm M_{disk}$ that can be
635: carried out  by the recoiling  black hole of $\rm  M_{disk}\la 2\times
636: 10^6\left[M_{BH}/(10^7\,M_{\odot})\right]^{2.2}\,  M_{\odot}$.   Given
637: that  $\rm   M_{BH}\ga  10^8\,M_{\odot  }$,   the  latter  comfortably
638: encompasses   the   molecular  gas   associated   with   the  AGN   in
639: HE\,0450-2958.
640: 
641: 
642: \subsubsection{How long could the AGN remain luminous?}
643: 
644: Apart  from partly  obscuring the  AGN in  HE\,0450-2958 via  its dust
645: content,  the molecular  gas mass  in  its vicinity  can in  principle
646: ``fuel'' it for
647: 
648: \begin{equation}
649: \rm T_{fuel}=\frac{M_{AGN}(H_2)}{L _{AGN}/(c^2\,\epsilon _{BH})}\sim
650: 1.4\times 10^8\, \epsilon _{BH}\, f_c \, \left(\frac{L^{(AGN)} _{IR}}{10^{12}\,L_{\odot}}\right)^{-1}
651: \left[\frac{M_{AGN}(H_2)}{10^7\,M_{\odot}}\right]\, yrs,
652: \end{equation}
653: 
654: \noindent
655: where $\rm \epsilon _{BH}$ is the black hole fuelling efficiency.  For
656: a  typical  $\rm\epsilon  _{BH}\sim  0.1  $  and  the  estimated  $\rm
657: M_{AGN}(H_2)$ the  latter yields $\rm  T_{fuel}\sim (2-4)\times 10^6\,
658: yrs$. The shortness of this timescale, and the difficulty of fueling a
659: black  hole  once  it  is  outside  its  host  galaxy,  may  point  to
660: fundamental  limitations  in  the  observability  of  such  events,  a
661: plausible reason for their hitherto scarcity.
662: 
663: 
664: 
665: \subsection{The origin of the radio continuum}
666: 
667: 
668: The  system's radio  continuum contains  low-luminosity  jets emerging
669: from  a  radio-quiet  core  located  at the  QSO  position  (Feain  et
670: al. 2007).  An important question arising from that study is how could
671: the  location  of  HE\,0450-2958  near the  FIR/radio  correlation  be
672: explained if the  starburst occurs in the companion  galaxy while most
673: of the radio  emission is of AGN origin.  Clues  for a possible answer
674: may be offered by the case of 1821+643, another radio-quiet but IR and
675: optically luminous  QSO ($\rm  M_V\sim -28$, $\rm  L_{FIR}\sim 9\times
676: 10^{12}\, L_{\odot }$)  residing in a giant elliptical.   This was the
677: first radio-quiet QSO discovered  to have classic FR~I radio structure
678: with low  brightness diffuse jets  (Papadopoulos et al.  1995),  and a
679: high  brightness temperature  compact  core (Blundell  et al.   1996),
680: indicative of a black hole--based central engine.  Despite the obvious
681: AGN contribution in its radio  continuum it did not deviate decisively
682: from  the  far-IR/radio correlation  until  the  discovery  of a  much
683: larger, 300  kpc-sized, low-brightness FR I  radio structure (Blundell
684: \& Rawlings  2001).  A similar  scenario may apply  for HE\,0450-2958,
685: and the discovery  of such a jet morphology in  this system could also
686: offer independent clues regarding the black hole.
687: 
688: 
689: Interestingly,  FR  I-type  radio  structures  may  originate  from  a
690: precessing  jet  axis due  to  binary black  holes  that  have yet  to
691: coalesce  (Blundell \&  Rawlings  2001).  Such  binaries, residing  in
692: gas-poor  ellipticals (which  allows for  their long  survival against
693: gas-dynamical  friction),  are   a  fundamental  prerequisite  in  the
694: scenario of  ejected AGNs as the  outcome of a  three-body kick during
695: gas-rich/gas-poor galaxy interactions (Hoffman \& Loeb 2006).
696: 
697: 
698: \section{HE\,0450-2958: not an ULIRG$\rightarrow $QSO transition object, implications }
699: 
700: The  hereby  discovery  of  HE\,0450-2958,  considered  an  archetypal
701:  ULIRG$\rightarrow $QSO  transition object, as  a strongly interacting
702:  pair of a  gas/dust-poor galaxy (marked by a  residing unobscured and
703:  optically powerful  QSO) and a gas-rich extreme  starburst raises the
704:  prospect of {\it such interactions being misclassified in the context
705:  of the popular ULIRG$\rightarrow $QSO evolution scenario.}  More such
706:  cases  would then  signify  an important  variation  in the  standard
707:  evolutionary  picture  in which  starburst  and  QSO activity,  while
708:  triggered by a common  cause (a strong galaxy interaction/merger), do
709:  not  necessarily  involve two  gas-rich  progenitors with  comparable
710:  amounts   of  molecular   gas   (Sanders  \&   Ishida  2004).    Such
711:  gas-poor/gas-rich  galaxy  interactions  would  not  conform  to  the
712:  standard ULIRG$\rightarrow  $QSO emergence scenario,  though they can
713:  still ignite a starburst in  the gas-rich progenitor (Di Mateo et al.
714:  2007).  They may also be the observational answer to the expectations
715:  of  current  galaxy-formation theories  that  predict  the so  called
716:  ``wet-dry'' mergers  between gas-rich and gas-poor  progenitors to be
717:  the most common in the Universe (e.g.  Springel et al.  2005).
718:  
719: 
720: The  demanding combination of  observational tools  it took  to reveal
721: these intriguing  aspects of HE\,0450-2958  may come to  exemplify why
722: such interactions have been missed until now.  Moreover, it could also
723: resolve the conflicting views regarding  the host galaxies of the more
724: powerful optical  QSOs ($\rm M_v<-24$) in the  local Universe.  Unlike
725: the low luminosity AGNs such as Seyferts where a consistent picture of
726: AGNs  in  centers  of  gas-rich vigorously  star-forming  spirals  has
727: emerged  (e.g.   Schinnerer  et  al.   1998; Staghun  et  al.   2004),
728: detailed {\it HST}  studies by Dunlop et al.  2003  (see also Floyd et
729: al.   2004) conclude  that the  more powerful  AGNs reside  in typical
730: massive  ellipticals  ($\rm   R^{1/4}$  light  profiles,  old  stellar
731: populations).   On the  other hand  Scoville  et al.   2003 (see  also
732: Bertram et  al.  2007)  found large amounts  of molecular gas  in such
733: systems and concluded otherwise.
734: 
735: It is now clear that  the latter conclusions are premature. Indeed all
736: current  studies  of   the  hosts  of  luminous  AGN   lack  the  {\it
737: combination}  of:  a)  high  resolution  optical  imaging  along  with
738: specialized  deconvolution techniques  (allowing sensitive  probing of
739: the  host   galaxy  properties  against   the  bright  QSO),   and  b)
740: interferometric  CO  maps with  good  resolution,  high  S/N, and  the
741: excellent astrometry  that helped identify  the type of merger  in the
742: HE\,0450-2958 system.  Indicative of those limitations is PDS~456, the
743: most luminous QSO in the local Universe (also a ULIRG/QSO composite in
744: terms of its SED), where despite sensitive high resolution optical and
745: CO imaging observations, the QSO  host galaxy type and the location of
746: the  molecular gas  with respect  to its  AGN remain  unclear  (Yun et
747: al. 2004).
748: 
749: 
750: The short gas consumption timescales by a starburst, which are shorter
751: still for optically powerful  AGNs residing in gas-poor hosts, suggest
752: that these  twin ``beacons'' of  gas-rich/gas-poor galaxy interactions
753: may always be  in close proximity.  The advent  of sensitive mm/sub-mm
754: arrays,  along  with the  already  available  high resolution  optical
755: imaging capabilities now allows their study, and the ULIRG$\rightarrow
756: $QSO transition objects are excellent~candidates.
757: 
758: 
759: \section{Conclusions}
760: 
761: We performed  CO line emission  imaging observations of  the enigmatic
762: QSO/galaxy pair HE\,0450-2958  using the ATCA (CO J=1--0)  and the SMA
763: (CO J=3--2), detecting  the J=1--0 line and placing  an upper limit on
764: the J=3--2 line, our conclusions are as follows
765: 
766: 
767: \noindent
768: 1. The CO J=1--0 line emission is associated with the companion galaxy
769:  not  the QSO, it  corresponds to  $\rm M_{gal}(H_2)  \sim (1-2)\times
770:  10^{10}\,  M_{\odot}$ and  amounts to  $\ga  30 \%$  of the  enclosed
771:  dynamical mass  within $\sim 8$\,kpc. This large  gas reservoir fuels
772:  extreme  star formation at  a rate  of $\rm  \sim 370\,  M_{\odot }\,
773:  yr^{-1}$,  efficiencies  of  $\rm L_{IR}/M_{gal}(H_2)\sim  (90-165)\,
774:  M_{\odot }/L_{\odot }$, and places  the companion galaxy on the upper
775:  gas-rich and star  forming end of ULIRGs.  This  property, along with
776:  the  disturbed   nature  of  the  system,   solidifies  the  strongly
777:  interacting status of this controversial galaxy/QSO pair.
778: 
779: 
780: \noindent
781: 2. No  molecular gas  is detected  towards  the QSO  itself with  $\rm
782:   M_{QSO-host}(H_2)/M_{gal}(H_2)\leq 1/5$,  consistent with the little
783:   dust extinction or star  formation activity deduced for its putative
784:   host galaxy.   This weakens the case  for a host  galaxy typical for
785:   Narrow Line  Seyfert 1  AGNs since these  are gas-rich  spirals with
786:   luminous circumnuclear starbursts.   HE\,0450-2958 is thus the first
787:   known  case of  a gas-rich  extreme  starburst and  a powerful  QSO,
788:   intimately linked (via a strong interaction), but not~coincident.
789:  
790: 
791: \noindent
792: 3. Most of the radio continuum in this system is due to low-brightness
793:    AGN-driven jets rather than  star formation activity.  Deeper radio
794:    continuum observations  are needed  to reveal their  morphology and
795:    extent, which in turn can yield more clues regarding its AGN.
796: 
797: 
798: \noindent
799: 4. The early  prominence of HE\,0450-2958  as an IR-selected  QSO that
800:   exemplifies   the   ULIRG$\rightarrow   $(optically  luminous   QSO)
801:   transition  in  a  popular  evolutionary scenario,  and  its  hereby
802:   revealed nature  as a  strong interaction between  a gas-rich  and a
803:   gas-poor, possibly elliptical, galaxy  opens up the possibility that
804:   some  of these  so-called  transition objects  are such  interacting
805:   pairs  instead.  Such   gas-poor/gas-rich  galaxy  interactions  can
806:   activate a  starburst in the  gas-rich progenitor, and  an optically
807:   luminous AGN  in the gas-poor  one.  Their composite  dust continuum
808:   would then  be typical of IR-warm  ULIRG$\rightarrow $QSO transition
809:   objects.  Sensitive  optical and CO  imaging at high  resolution can
810:   now  reveal  and  study   such  dynamical  interactions,  and  these
811:   transition objects are excellent candidates.
812:   
813:   
814:  
815: \section{Acknowledgments}
816: 
817: Jeff Wagg is grateful for  support from the Max-Planck Society and the
818: Alexander von Humboldt Foundation. The anonymous referee is gratefully
819: acknowledged  for  numerous  suggestions  that greatly  clarified  the
820: original  document.  We  also thank  all the  people that  operate and
821: maintain  ATCA  and  in  particular  its  3\,mm  receivers  for  their
822: dedicated  support.   Padelis  Papadopoulos  thanks  Karl  Jesienowski
823: aboard  the  {\it  Undersea  Explorer} for  great  conversations,  and
824: Marcella  Carollo  for  bringing  this  controversial  object  to  his
825: attention.  Last  but not  least  he  would  like to  thank  Margarita
826: Zakalkas  for inspiration during  his time  in Zurich.   The Australia
827: Telescope is funded by the  Commonwealth of Australia for operation as
828: a National Facility managed by CSIRO.
829: 
830: 
831: \newpage
832: 
833: \begin{thebibliography}{}
834: \bibitem[]{} Aalto S., Booth R. S., Black J. M.,\& Johansson L. E. B. 1995, A\&A, 300, 369{}{}
835: \bibitem[]{} Alloin D., Barvainis R., Gordon M. A., \& Antonucci R. R. J. 1992, A\&A,
836:              265, 429{}{}
837: \bibitem[]{} Bahcall J. N., Kirhakos S., \& Schneider D. P. 1994, ApJ, 425, L11{}{}
838: \bibitem[]{} Barnes J. E., \& Hernquist L. 1992, ARA\&A, 30, 705{}{}
839: \bibitem[]{} Bertram T., Eckart A., Fischer S., Zuther J., Straubmeier C., 
840:              Wisotzki L., \& Krips M. 2007, A\&A, 470, 571{}{}
841: \bibitem[]{} Blitz, L. 1997, in {\it CO: Twenty-Five Years of Millimeter-Wave Spectroscopy},
842:              IAU Symposium No. 170, pg. 11{}{}
843: \bibitem[]{} Blundell K. M.,  Beasley A. J., Lacy M., \&  Garrington S. T. 1996, ApJ, 468, L91{}{}
844: \bibitem[]{} Blundell K. M., Rawlings S. 2001, ApJ, 562, L5{}{}
845: \bibitem[]{} Bonning E. W., Shields G. A., \& Salviander S. 2007, ApJ, 666, L13{}{}
846: \bibitem[]{} Boyce P. J., Disney M. J., Blades J. C., Boksenberg A., Crane P., Deharveng J. M.,
847:              Macchetto F. D., Mackay C. D., \& Sparks W. B. 1996, ApJ, 473, 760{}{}
848: \bibitem[]{} Bryant P. M., \& Scoville N. Z. 1996, ApJ, 457, 678{}{}
849: \bibitem[]{} Bryant P. M., \& Scoville N. Z. 1999, AJ, 117, 2632{}{}
850: \bibitem[]{} de Grijp M. H. K., Lub J., \& Miley G. K. 1987, A\&AS, 70, 95{}{} 
851: \bibitem[]{} Canalizo G., \& Stockton A. 2001, ApJ, 555, 719{}{}
852: \bibitem[]{} Devereux N., Taniguchi Y., Sanders D. B., Nakai N., Young J. S. 1994, AJ, 
853:              vol. 107, no. 6, p. 2006-2016{}{}
854: \bibitem[]{} Di Matteo P., Combes F., Melchior A.-L., \& Semelin B. 2007, A\&A, 468, 61{}{}
855: \bibitem[]{} Downes D., Solomon, P. M. 1998, ApJ, 507, 615{}{}
856: \bibitem[]{} Dunlop J. S., McLure R. J., Kukula M. J., Baum S. A., O'Dea C. P., \& 
857:              Hughes D. H. 2003, MNRAS, 340, 1095{}{}
858: \bibitem[]{} Evans A. S., Kim D. C., Mazzarella J. M., Scoville N. Z., \&
859:              Sanders D. B. 1999, ApJ, 521, L107{}{}
860: \bibitem[]{} Evans A. S., Surace J. A., \& Mazzarella J. M. 2000, ApJ, 529, L85{}{}
861: \bibitem[]{} Evans A. S., Frayer D. T., Surace J. A., \& Sanders D. B. 2001,
862:              AJ, 121, 3286{}{}
863: \bibitem[]{} Evans A. S., Mazzarella J. M., Surace J. A. \& Sanders D. B. 2002,
864:              ApJ, 580, 749{}{}
865: \bibitem[]{} Evans A. S., Solomon P. M., Tacconi L. J., Vavilkin T., \& Downes D.
866:              2006, AJ, 132, 2398{}{}
867: \bibitem[]{} Floyd D. J. E., Kukula M. J., Dunlop J. S., McLure R. J., Miller L.,
868:              Percival W. J., Baum S., \& O'Dea C. P. 2004, MNRAS, 355, 196{}{}
869: \bibitem[]{} Frayer D. T. et al. 1999, ApJ, 514, L13{}{}
870: \bibitem[]{} Genzel R., Baker A. J., Tacconi, L. J., Lutz D., Cox P., Guilloteau S., \&
871:              Omont A. 2003, ApJ, 584, 633{}{}
872: \bibitem[]{} Haas M., M\"uller S. A. H., Chini R., Meisenheimer K., Klaas U., Lemke D.,
873:              Kreysa E., \& Camenzind M. 2000, A\&A, 354, 453{}{}
874: \bibitem[]{} Haehnelt M. G., Davies M. B., \& Rees M. J. 2005, MNRAS, 366, L22{}{}
875: \bibitem[]{} Helfer T. T., et al. 2003, ApJS, 145, 259{}{}
876: \bibitem[]{} Hoffman L., \& Loeb A. 2006, ApJ, 638, L75{}{}
877: \bibitem[]{} Hutchings J. B., \& Neff S. G. 1988, AJ, 96, 1575{}{}
878: \bibitem[]{} Feain I. J., Papadopoulos P. P., Ekers R., \& Middelberg E. 2007,
879:              ApJ, 662, 872{}{} 
880: \bibitem[]{} Floyd D. J. E., Kukula M. J., Dunlop J. S., McLure R. J., 
881:              Miller L., Percival W. J., Baum S. A., O'Dea C. P. 2004, MNRAS, 355, 196{}{}
882: \bibitem[]{} Greve T. R., et al. 2005, MNRAS, 359, 1165{}{}
883: \bibitem[]{} Kim M., Ho L. C., Peng C. Y., Im M. 2007, ApJ, 658, 107{}{}
884: \bibitem[]{} Letawe G., Magain P., \& Courbin F. 2007, A\&A (in press), arXiv:0709.3743v1{}{}
885: \bibitem[]{} Loeb A. 2007, Physical Review Letters, 0031-9007/07/99(4)/041103(4){}{}
886: \bibitem[]{} Low F. J., Cutri R. M., Huchra J. P., \& Kleinman S. G. 1988, ApJ, 327, L41{}{}
887: \bibitem[]{} McLure R. J., \& Dunlop J. S. 2002, MNRAS, 331, 795{}{}
888: \bibitem[]{} Magain P., Letawe G., Courbin F., Jablonka P., Jahnke K., Meylan G.,
889:              \& Wisotzki L. 2005, Nature, Vol. 437, 318{}{}
890: \bibitem[]{} Merritt D., Storchi-Bergmann T., Robinson A., Batcheldor D., Axon D.,
891:              \& Cid Fernandes R. 2006, MNRAS, 367, 1746{}{}
892: \bibitem[]{} Middelberg E., Sault R. J., \&  Kesteven M. J. 2006, Publications of the
893:              Astronomical Society of Australia, Vol. 23, Issue 4, p. 147{}{}
894: \bibitem[]{} Moshir M. et al. 1990, Bulletin of the American Astronomical Society, Vol. 22, p.1325{}{}
895: \bibitem[]{} Papadopoulos P. P., Seaquist E. R., Wrobel J. M., \& Binette L. 1995, ApJ, 446, 150{}{}
896: \bibitem[]{} Planesas P., Scoville N. Z., \& Myers S. T. 1991, ApJ, 369, 364{}{}
897: \bibitem[]{} Regan W. M., et al. 2001, ApJ, 561, 218{}{}
898: \bibitem[]{} Richardson K. J. 1985, PhD Thesis, Queen Mary College, University of London{}{}
899: \bibitem[]{} Sanders D. B., et. al. 1988, ApJ, 328, L35{}{}
900: \bibitem[]{} Sanders D. B., Scoville N. Z., \& Soifer B. T. 1991, ApJ, 370, 158{}{}
901: \bibitem[]{} Sanders D. B., \& Ishida C. M. 2004, in {\it The Neutral ISM in Starburst
902:              Galaxies}, ASP Conference Series, Vol. 320, pg. 230{}{}
903: \bibitem[]{} Schinnerer E., Eckart A., \& Tacconi L. J. 1998, ApJ, 500, 147{}{}
904: \bibitem[]{} Schinnerer E., Eckart A., \& Tacconi L. J. 1999, ApJ, 524, L5{}{}
905: \bibitem[]{} Scoville N. Z., Frayer D. T., Schinnerer E., \& Christopher M. 2003, 
906:              ApJ, 585, L105{}{}
907: \bibitem[]{} Scoville N. Z. 2004, in {\it The Neutral ISM in Starburst Galaxies}, 
908:              ASP Conference Series, Vol. 320, pg. 253{}{}
909: \bibitem[]{} Solomon P. M., Downes D., Radford S. J. E., \& Barrett J. W. 1997, ApJ, 478, 144{}{}
910: \bibitem[]{} Springel V., et al. 2005, Nature, 435, 629{}{}
911: \bibitem[]{} Staguhn J. G., Schinnerer E., Eckart A., \& Scharw\"achter J. 2004, ApJ,
912:              609, 85{}{}
913: \bibitem[]{} Tacconi L. J., Genzel R., Blietz M., Cameron M., Harris A. I., \& Madden S. 1994
914:              ApJ, 426, L77{}{}
915: \bibitem[]{} Tacconi L. J., Neri R., Chapman S. C., Genzel R., Smail I., Ivison R. J.,
916:              Bertoldi F., Blain A., Cox P., Greve T., \& Omont, A. 2006, ApJ, 640, 228{}{}
917: \bibitem[]{} Yao L., Seaquist E. R., Kuno N., \& Dunne L. 2003, ApJ, 588, 771{}{}
918: \bibitem[]{} Yun M. S., Reddy N. A., Scoville N. Z., Frayer D. T., Robson E. I., 
919:              \& Tilanus R. P. J. 2004, ApJ, 601, 723{}{}
920: \end{thebibliography}
921: 
922: \clearpage
923: 
924: 
925: 
926: \begin{figure}[h]
927: \vspace*{12cm}
928: \special{psfile=./f1a.ps hoffset=-125 voffset=370 hscale=53
929:      vscale=53 angle=-90}
930: \special{psfile=./f1b.ps hoffset=165 voffset=370 hscale=53
931:      vscale=53 angle=-90}
932: \caption{{\it Left:} Map of  CO J=1--0 emission (NA-weighted, CLEANed)
933:  averaged over $\rm 570\,km\,s^{-1}$  (see section 2.3), and contours:
934:  $\rm (-2,2,4,6,8,10,12,14)\times  \sigma_{rms}$. {\it Right:}  The mm
935:  continuum emission at  94.53\, GHz with contours of  $\rm (-3, -2, 2,
936:  3)\times  \sigma_{rms}$.   The noise  in  both  maps  is $\rm  \sigma
937:  _{rms}\sim 0.45\, mJy/beam$,  and the restoring beam is  shown at the
938:  bottom  left  of  the  CO  image  ($\rm  \Theta  _{beam}=3.21''\times
939:  2.14''$,  $\rm PA=72^{\circ}$).  The  phase center  is marked  by the
940:  cross, the  diamond marks  the AGN position,  and the star  marks the
941:  peak of the CO emission (see  Table 1).  For the mm continuum imaging
942:  the phase center was set $3''$  to the north of the AGN's position to
943:  avoid  any  DC correlator  offsets  ``masking''  as  a spurious  weak
944:  source.}
945: \end{figure}
946: 
947: \clearpage
948: 
949: 
950: \begin{figure}
951: \includegraphics[angle=0,scale=0.8]{f2.eps}
952: \caption{The  spectrum  corresponding  to  the  peak  CO  J=1--0  line
953: emission (Figure  1). The velocity range is centered on redshift of the
954:  QSO/ULIRG system shown on the upper left.}
955: \end{figure}
956: 
957: 
958: 
959: \clearpage
960: 
961: \begin{figure}
962: \includegraphics[angle=0,scale=0.8]{f3.eps}
963: \caption{ATCA CO(1-0) greyscale image with ATCA 8.4GHz radio continuum
964: contours overlaid. Contour  levels start at $200\mu$Jy~beam$^{-1}$ and
965: increase in a geometric series  with a common ratio of $\sqrt{2}$. The
966: CO(1-0) and the 8.4\,GHz beams are  shown in the bottom left corner of
967: the  image.  The cross  marks  the peak  emission  of  the C\,2  radio
968: continuum emitting region associated  with the companion galaxy (Feain
969: et al. 2007).}
970: \end{figure}
971: 
972: \clearpage
973: 
974: 
975: \begin{figure}[h]
976: \vspace*{12cm}
977: %\leavevmode
978: %\includegraphics[angle=-90,scale=0.9]{HST_COemission.ps}
979: \special{psfile=./f4a.ps hoffset=-110 voffset=370 hscale=50
980:      vscale=50 angle=-90}
981: \special{psfile=./f4b.ps hoffset=170 voffset=370 hscale=50
982:      vscale=50 angle=-90}
983: \caption{{\it Left:}  CO J=1--0  emission overlaid on  the deconvolved
984:   HST/ACS image of the HE\,0450-2958 system, with the QSO contribution
985:   subtracted (from Magain  et al.  2005).  Contours are:  $\rm (-2, 2,
986:   5,7,9,11,13)\times       \sigma       _{rms}$,       with       $\rm
987:   \sigma_{rms}=0.45\,mJy/beam$. The  beam is shown at  the bottom left
988:   (HPBW=$3.21''\times 2.14''$, $\rm  PA=72^{\circ}$). {\it Right:} The
989:   {\it deconvolved}  CO source model  (see section 2.3.1)  overlaid on
990:   the HST  image after a $\Delta \delta_{o}=0.6''$  shift aligning the
991:   optical and radio AGN positions.   The good correspondence of the CO
992:   emission with the companion  ULIRG is~evident.  In the optical image
993:   the QSO (marked by the single bright pixel near the phase center) is
994:   surrounded by an AGN-excited gas cloud.  The cross size at the phase
995:   center   $(0'',0'')$   denotes   the  (dominant)   HST   astrometric
996:   uncertainties ($\rm \sim 0.5''-1.0''$), and the star symbol denotes
997:   the peak CO J=1--0 emission.}
998: \end{figure}
999: 
1000: 
1001: \clearpage
1002: 
1003: \begin{figure}
1004: \includegraphics[angle=0,scale=0.9]{f5.eps}
1005: \caption{Two   component  fit   of   the  mm/IR   dust  continuum   of
1006: HE\,0450-2958 (see  section 3.2). The  upper limits at  mm wavelengths
1007: were obtained  from this work  (see Table 1),  and the IRAS  fluxes at
1008: $12\,\mu$m, $25\,\mu$m, $60\,\mu$m, and $100\,\mu$m are from Moshir et
1009: al. 1990.  The emissivity laws adopted and  the resulting temperatures
1010: from the  2-component dust  emission fit are  shown in the  upper left
1011: (see also 3.2).}
1012: \end{figure}
1013: 
1014: \clearpage
1015: 
1016: \begin{deluxetable}{lcc}
1017: \tablecolumns{3}
1018: \tablewidth{0pc}
1019: \tablecaption{Properties of the HE\,0450--2958 system}
1020: \tablehead{
1021: \colhead{Property} & \colhead{AGN} & \colhead{Companion galaxy} }
1022: \startdata
1023: $\rm \alpha (J2000.0)$ & 04$\rm ^h$ 52$\rm ^m$ 30$\rm ^s$.1 & $+1.15''\pm 0.12''$\,\tablenotemark{a}\\
1024: $\rm \delta (J2000.0)$ & --29$^{\circ}$ 53$'$ 35.0$''$ & $-1.05''\pm 0.10''$\,\tablenotemark{a}\\
1025: $z$\,\tablenotemark{b}    & 0.2863 & 0.2865 \\
1026: $\rm S_{peak}$(1--0) (mJy/beam)\tablenotemark{c}  &   $ <1.35  $  & $6.7\pm 0.45$\\
1027: $\rm S_{peak}$(3--2) (mJy/beam)\tablenotemark{c}  &   $ < 20   $  & $ < 20 $\\
1028: $\rm \int  S_{CO(1-0)} dV$ (Jy\,km\,s$^{-1}$)\tablenotemark{c} & $<0.77 $ & $5.70\pm1.15$\\
1029: $\rm \int  S_{CO(3-2)} dV$ (Jy\,km\,s$^{-1}$)\tablenotemark{c} & $<11$ & $<20 $\\
1030: Intrinsic CO(1-0) source size & \nodata & $2.5''\times 1.5''$\\
1031: $\rm M(H_2)$ $\rm (\times 10^{10}\,M_{\odot })$\,\tablenotemark{d} &  \nodata  &  $1.25-2.3$\\
1032: $\rm S_{94\,GHz}$ (mJy)\tablenotemark{c}  & $< 1.4$ & $ < 1.4$\\
1033: $\rm S_{275\,GHz}$ (mJy)\tablenotemark{c}  & $< 8$ & $ < 8$\\
1034: $\rm L_{IR}$  $\rm (\times 10^{12}\,L_{\odot})$\tablenotemark{e}    & 2.6  & 2.1 \\
1035: \enddata
1036: \tablenotetext{a}{Position of the peak CO J=1--0 emission with respect to the\\
1037: \hspace*{0.5cm} AGN (see 2.3.1)}
1038: \tablenotetext{b}{Canalizo \& Stockton 2001}
1039: \tablenotetext{c}{All limits are $3\sigma$, point source assumed for AGN-related limits}
1040: \tablenotetext{d}{See section 3}
1041: \tablenotetext{e}{from the dust continuum SED fit in Figure 6}
1042: 
1043: \end{deluxetable}
1044: 
1045: 
1046: 
1047: \end{document}
1048: 
1049: 
1050: 
1051: 
1052: 
1053: 
1054: 
1055: 
1056: 
1057: 
1058: 
1059: 
1060: 
1061: 
1062: 
1063: 
1064: 
1065: