1: \documentclass{aastex}
2: \shortauthors{Dinh-V-Trung et al.}
3: \shorttitle{Nature of ULIRG...}
4: \usepackage{epsfig,psfig}
5:
6: \begin{document}
7:
8: \title{Nature of widely separated ultraluminous infrared galaxies}
9:
10: \author{Dinh-V-Trung\altaffilmark{1}, K.Y. Lo, D.-C. Kim}
11: \affil{Institute of Astronomy and Astrophysics, Academia Sinica, P.O. Box 1-87,
12: Nankang, Taipei 11529, Taiwan, ROC.}
13: \email{trung@asiaa.sinica.edu.tw, kyl@asiaa.sinica.edu.tw, kim@asiaa.sinica.edu.tw}
14: \author{Yu Gao}
15: \affil{Infrared Processing and Analysis Center, MS 100-22, California Institute of Technology, Pasadena, CA 91125.}
16: \email{gao@ipac.caltech.edu}
17: \and
18: \author{R.A. Gruendl}
19: \affil{Department of Astronomy, University of Illinois at Urbana-Champaign, Urbana, IL 61801.}
20: \email{gruendl@astro.uiuc.edu}
21: \altaffiltext{1}{On leave from Institute of Physics, P.O. Box 429 BoHo 10000, Hanoi, VietNam}
22: \begin{abstract}
23: In the complete sample of ultraluminous infrared galaxies (ULIRGs) compiled by Kim (1995)
24: about 5\% consists of widely separated galaxies which are presumably in the early phase of interaction.
25: This fact is contrary to the conventional view that ULIRGs are in the
26: final stages of the merger of two gas-rich disk galaxies.
27: We have undertaken high resolution CO(J=1-0) observations for the
28: ultraluminous infrared galaxies that have nuclear separations larger than 20 kpc.
29: We have detected the CO emission in 5 out of 6 systems, but only in
30: one component of the ULIRG pairs. 4 of them have LINER spectral type and 1 galaxy has
31: Seyfert II spectral type. In K'-band images these components are also brighter
32: than the other components which have either HII-region spectra or no detectable emission lines.
33: Using the standard conversion factor, the molecular gas content is estimated to
34: be a few times 10 $^{10}$ M$_\odot$, similar to that of the other ultraluminous galaxies. The result indicates that
35: the galaxy containing the molecular gas is also the source of most, if not all, of the
36: huge far-infrared luminosity of the
37: system. The optical and K'-band imaging observations and optical spectra
38: suggest multiple merger scenario for 1 system.
39: If the remaining systems are in an early stage of a binary tidal interaction,
40: the commonly accepted interpretation of the ULIRG phenomenon as the final merger stage of two disk
41: galaxies may need to be re-examined.
42: \end{abstract}
43:
44: \keywords{galaxies: interactions --- galaxies: ISM --- galaxies: starburst --- ISM: molecules}
45:
46: \section{Introduction}
47: %
48: Ultraluminous infrared galaxies (ULIRGs) are known to emit the bulk of their energy in the infrared and
49: their infrared luminosities ($L_{IR} \ge 10^{12}L_{\odot}$) are equivalent to the bolometric luminosities
50: of optically selected quasars.
51: For the past decade, there have been extensive studies of the nature of these ULIRGs (Sanders \& Mirabel 1996).
52: Imaging studies at optical and near-infrared wavelengths show that most of these galaxies are
53: either interacting or merging (Sanders et al. 1988, Zou et al. 1991, Leech et al. 1994, Murphy et al. 1996,
54: Borne et al. 2000, Kim et al. 2001). The fraction of ULIRGs in interacting or merging ranges from at least 61\%
55: (25 objects out of 41, Zou et al. 1991) to possibly as high as 100\%
56: for all 10 nearby ULIRGs (Sanders et al. 1988). The rather small fraction of interacting systems
57: derived by Zou et al. (1991) resulted from the use of POSS images
58: in which faint tidal tails are almost impossible to detect. Recent imaging data by Kim et al. (2001) showed that
59: close to 100\% of ULIRGs in the complete 1 Jy sample have faint tidal tails, a signature of recent interaction and merger.
60:
61: One of the most basic issues of ULIRGs is the energy source powering the enormous amount of their infrared radiation.
62: Spectroscopic studies in the optical, near-infrared, and mid-infrared suggest that about 70\% and 30\%
63: of these galaxies are powered by starburst activities and active galactic nuclei (AGNs)
64: respectively (Veilleux et al. 1999a, Veilleux et al. 1999b, Genzel et al. 1998a).
65: The huge infrared luminosity of the ULIRGs suggests that they contain a large reservoir of molecular
66: gas to provide fuel for the star formation activity and/or a central AGN.
67: Extensive searches for CO emission have indeed shown that all ULIRGs possess a huge amount of
68: molecular gas ($\sim$ 10$^{10}$ M$_\odot$) and that there is a trend of increasing molecular gas content with
69: the far-infrared emission (Downes et al. 1993, Solomon et al. 1997).
70: Recent interferometric observations of nearby ULIRGs (Downes \& Solomon 1998)
71: and luminous infrared galaxies (Bryant \& Scoville 1999)
72: have determined that the molecular gas is concentrated in a
73: circumnuclear disk of typically 1 kpc in size.
74: From the interferometric observations of double-nucleus ULIRGs with projected nuclear separations of 3-5 kpc,
75: Evans et al. (1999, 2000) find that one of the interacting pair contains the majority of molecular gas in warm ULIRGs
76: whereas both components contain molecular gas in cool ULIRGs.
77: Based on observational evidence, Sanders et al. (1988) proposed that ULIRGs originate from the merger of two
78: gas-rich spiral galaxies. The large amount of molecular gas accumulated in the center of the galaxies due to
79: the gravitational interaction can sustain strong star formation activities as well as provide
80: fuel to the central AGN. As the merger proceeds, the molecular gas reservoir is eventually depleted
81: by the star formation (Gao \& Solomon 1999) and dispersed by the AGN activities and supernovae. Finally the AGN reveals itself as an
82: optical quasar. The estimated lifetime of the ULIRG phase is
83: about 10$^7$ to 10$^{8}$ yr. Genzel et al. (1998b) reviewed most recent observational evidence in favor of
84: this scenario.
85:
86: Numerical simulations of interacting galaxies have shown that violent star formation activity can be triggered in
87: the merger of two massive disk galaxies (Barnes \& Hernquist 1996, Mihos \& Hernquist 1996, Springel 2000). Furthermore, star formation
88: activities are found to be dependent mainly on the galaxy structure and the orbital geometry plays only a
89: modest role (Mihos \& Hernquist 1994, 1996).
90: They suggest that galaxies with dense central bulges show strong star formation activities at the final stages
91: of merging, while bulgeless galaxies show weaker star formation activity at the early stage of interacting
92: and do not show any subsequent strong star formation activities. Thus, according to the numerical simulations,
93: the majority of ULIRGs would be in the final stages of merger.
94:
95: Recently, Kim et al. (2001) have completed an optical and near-infrared imaging survey of the
96: IRAS 1 Jy sample of ULIRGs (Kim \& Sanders 1998) and find that all of the ULIRGs show signs of tidal interaction
97: and therefore are in a merging or interacting process (Kim et al. 2001).
98: More interestingly, we also find that 6 out of 118 ULIRGs have projected nuclear separation larger than 20 kpc.
99: The large nuclear separation observed in these ULIRGs seems to contradict the conventional view of the origin of ULIRGs
100: and the numerical simulation results outlined above. To become an ULIRG in the early phase of interaction,
101: the galaxies must have accumulated a large amount of molecular gas and it is still unclear how the star formation activities
102: are triggered in these cases.\\
103: %
104: The position uncertainty associated with the poor resolution of IRAS in the far-infrared did not allow
105: us to determine the distribution of far-infrared emission as well as the molecular gas in these galaxies. Without that
106: information, it is difficult to clarify the nature of these unusual ULIRGs.
107: In this paper we present high angular resolution CO(J=1-0) observations and
108: new K'-band imaging to study the widely separated ULIRGs. CO emission and the
109: molecular gas are known to be associated with the source of far-infrared emission.
110: Thus observations of CO emission will allow us to determine the location and the distribution
111: of far-infrared luminosity of these ULIRGs, in addition to the distribution of molecular gas.
112: Also making use of the optical imaging data (Kim et al. 2001) and our new K'-band
113: images, we discuss the nature of these widely separated ULIRGs.
114: We will use $q_0=0.5$ and $H_0=75$ km s$^{-1}$ Mpc$^{-1}$ throughout this paper.
115: %
116: \section{Observations}
117: We used the BIMA (Berkeley-Illinois-Maryland Association) array (Welch et al. 1996) to search for the
118: CO(J=1-0) emission line from the selected ULIRGs.
119: Each galaxy in our sample was observed in various runs during 1998 and 1999 (Table. 1) in
120: the B-configuration with baselines ranging from 5 k$\lambda$ to
121: 80 k$\lambda$, and the C-configuration with baselines ranging from 3 k$\lambda$ to 30 k$\lambda$.
122: The digital correlator was configured to cover a velocity range of 1600 km s$^{-1}$ centered at the
123: velocity of the ULIRGs calculated from their redshift.
124: Phase calibration was achieved by observing nearby quasars every 20 to 30 minutes. We also observed planets to determine
125: the absolute flux scale.
126:
127: We used the MIRIAD data reduction package (Sault et al. 1994) to calibrate and process our data. After
128: phase calibration and flux calibration the visibilities were Fourier-transformed to make the images.
129: These images were then cleaned using the
130: standard task CLEAN of the MIRIAD package. The final maps have angular resolutions of about 4 arcsec for
131: B-configuration and 8 arcsec for C-configuration, respectively.
132:
133: The K'-band (2.1 $\mu m$) images of 4 galaxies IRAS F11180+1623, IRAS
134: F14394+5332, IRAS F17028+5817, IRAS F23327+2913 were obtained with the QUIRC 1024 x 1024 near-infrared camera
135: on July 15, 2000, at the UH 88" telescope on Mauna Kea.
136: The plate scale of the QUIRC is 0.1886$^{\prime\prime}$/pixel with f/10 and the seeing is estimated about
137: 0.5$^{\prime\prime}$ to 0.7$^{\prime\prime}$.
138: The sky frames were constructed by median filtering of 9 dithered
139: (20$^{\prime\prime}$ rectilinear motion) object frames.
140: Dome flats were obtained by turning a lamp in the dome on and off
141: at the beginning and end of each night.
142: The dark subtracted and normalized flats were used for flat-fielding the sky-subtracted images.
143: Flat-fielded images were then coadded after shifting in fractional
144: pixel units with respect to the object center.
145: The final coadded images were calibrated from observations
146: of infrared standard stars (Elias et. al. 1982).
147: K'-band images of two other galaxies (IRAS F10485-1447, IRAS F12359-0725) together with R-band images of all
148: the galaxies in our sample were taken from the work of Kim et al. (2001).
149: The optical spectra of the nothern object of IRAS F23327$+$2913 were taken in October 1995 at the UH 88" telescope with
150: spectral resolution of 8 $\AA$ in the spectral range of $\lambda\lambda$ 4500-8500$\AA$. The spectra
151: were processed with the standard data reduction procedures.
152: %
153: \section{Results}
154: We detected CO(J=1-0) emission from 4 galaxies IRAS F11180$+$1623, IRAS
155: F14394$+$5332, IRAS F17028$+$5817, IRAS F23327$+$2913 at $\geq$ 3$\sigma$ in the channel maps.
156: In IRAS F12359-0725 we found possible CO(J=1-0) emission of 3$\sigma$ with channel width of 200 km s$^{-1}$ which is
157: coincident within the synthesized beam with one of the galaxy in the system.
158: Fig 1. shows the integrated intensity maps for the galaxies observed in our
159: program together with their R-band and K'-band images. The CO emission line profiles of the detected nuclei
160: are shown in Fig. 2. Some of the properties of these ULIRGs are provided in Table. 2.
161: The CO emitting regions are not resolved by our synthesized
162: beams. In the case of IRAS F14394$+$5332, which is the only ULIRG mapped in B-array, we can infer an upper limit to the size
163: of the CO emitting region to be $\sim$ 1" or $\sim$ 2 kpc in linear scale.
164:
165: To estimate the molecular gas content from the total flux of CO(J=1-0) line we adopt the same conversion
166: factor as for the Milky Way galaxy (Bryant \& Scoville 1999):
167: \begin{equation}
168: {\rm M}_{\rm g} = 1.20\times 10^4 {\rm F}_{\rm CO} {\rm D}_{\rm L}^2 (1 + z)^{-1}
169: \end{equation}
170: where M$_{\rm g}$ is the molecular gas mass in M$_\odot$,
171: F$_{\rm CO}$ is the integrated flux in Jy km s$^{-1}$, D$_{\rm L}$ is the luminosity distance in Mpc,
172: and $z$ is the redshift. The molecular gas mass
173: derived for the galaxies in our sample is similar to that found in other ultraluminous galaxies (Solomon et al. 1997) and in
174: some high redshift galaxies (Combes et al. 1999) using the same Galactic conversion factor as ours. The 1$\sigma$ upper
175: limit of the CO flux is rms*$\Delta V_{\rm FWHM}$/$\sqrt{\Delta V_{\rm FWHM}/\Delta V_{\rm res}}$ where rms is the noise level in the channel maps,
176: $\Delta V_{\rm FWHM}$ is the linewidth determined from the gaussian fitting to the line profiles and
177: $\Delta V_{\rm res}$ is the channel width. The 1$\sigma$ upper limit to
178: the molecular gas mass of the companion is then derived following the above equation.
179: For the ULIRGs with detected CO emission, the companion may possess a molecular gas content of
180: up to 10\% to 20\% the molecular gas content of the main component (Table. 2).
181: The upper limit is in the range of several 10$^9$ M$_\odot$ which implies the companions are not as gas rich as the main
182: component but may possess a molecular gas content comparable to that of normal spiral galaxies.
183:
184: Gaussian fitting to the line
185: profiles gives the FWHM of 200 to 400 km s$^{-1}$ except for the case of IRAS F11180$+$1623. The weak CO(J=1-0) emission from
186: IRAS F11180$+$1623 did not allow us to determine accurately its linewidth. Thus, the value of 600 km s$^{-1}$ should be regarded as
187: upper limit to the true linewidth of IRAS F11180$+$1623.
188:
189: Although we did not succeed to detect CO(J=1-0) emission from IRAS F10485$-$1447, the 1$\sigma$ upper limit to
190: the molecular gas mass of $\sim$ 4$\times$10$^9$ M$_\odot$ is still quite high. In addition, this galaxy has been
191: observed only with the B-array which provides long baselines data with high phase scattering due to the atmospheric fluctuation.
192: It would be very useful to search again for CO emission with improved sensitivity.\\
193: Using optical spectra, Veilleux et al. (1999a) determined the spectral type of all ULIRGs in the 1 Jy sample.
194: In all cases, the galaxy detected in CO(J=1-0) line has LINER or Seyfert II spectral type while the companion has no
195: emission line except for IRAS F17028$+$5817E galaxy, which has spectral type HII, typical of starburst galaxies (Table 3.).\\
196: The magnitudes of each component of the ULIRGs are derived from our K'-band images and presented in Table. 3. In most cases the
197: galaxies have compact light distribution in K'-band (Fig. 1).
198: However, K'-band image of IRAS F14394$+$5332 shows the extended emission with
199: distortion in the E component, suggestive of a recent merger.
200: \section{Detailed Description of Individual System}
201: %
202: {\bf IRAS F10485$-$1447} -
203: This system is close to two bright foreground stars which are visible on the upper left of both R-band and K'-band
204: images (Fig. 1). On the R-band image, the E component shows short tidal tails in two opposite directions
205: and the W component has
206: a long curved tidal tail of about 10 kpc. The W component also has a distortion toward SW direction.
207: The R-band luminosity of the W component is only slightly brighter than that of the E component,
208: whereas in the K'-band the W component is more than one magnitude brighter
209: than the E component. The optical emission lines were detected from the W component with
210: LINER spectral type (Veilleux et al. 1999a). The presence of absorption lines in the spectrum of the E galaxy
211: (S. Veilleux, private communication) indicates that both galaxies are at the same redshift and are interacting.
212: Though we did not detect any CO emission in this system, the general trend of finding molecular gas in the
213: active component of the widely separated ULIRGs suggests that most of the gas could reside in
214: the W component.\\
215: %
216: {\bf IRAS F11180$+$1623} -
217: The optical morphology of the E component (Fig. 1) shows two tidal tails
218: in two opposite directions whereas the W component does not show any strong disturbances.
219: The spectral type of the E component is LINER (Veilleux et al. 1999a) and no detectable optical
220: emission lines were observed in the W component.
221: The CO emission was detected only on the E component.
222: The E component may itself be a merger because it has two opposite tidal tails.
223: One of the tidal tails on the E component tends to be curved and directed to the W component.
224: There is also a small distortion on the east side of the W component, suggesting that the
225: two galaxies are interacting.
226: The lack of CO emission as well as the non-detection of optical emission lines in the W component
227: imply that this galaxy contains only a small amount of gas and is in the early stage of tidal
228: interaction with the E component. Because there is no measured redshift for the W component,
229: we classify this pair as a tentative merger system.
230: There are 3 additional galaxies around the system, and if
231: they are at the same redshift as the two galaxies considered,
232: then this whole system is likely an example of the Hickson compact groups (Hickson 1982).\\
233: %
234: {\bf IRAS F12359$-$0725} -
235: On the R-band image (Fig. 1), the S component shows a curved tidal tail whereas the N component has no visible tail
236: except a small depression on the western part of the galaxy. The two galaxies are found to be at the same redshift and
237: are likely in the early phase of interaction.
238: A foreground star is present to the west of the system on both R-band and K'-band
239: images.
240: The R-band magnitudes of both components are similar,
241: whereas K'-band magnitude of the N component is 1.1 magnitude
242: brighter than that of the S component.
243: The N galaxy with LINER spectral type is more active than the S component which
244: has HII spectral type. We found possible CO(J=1-0) emission coincident with the
245: N component within the synthesized beam of 10"$\times$6". As the N component is the most active in the
246: system, the detection follows the same trend for other widely separated ULIRGs in our sample.\\
247: %
248: {\bf IRAS F14394$+$5332} -
249: This system has large nuclear separation
250: and shows long tidal tail ($\sim$ 35 kpc, Fig. 1).
251: The R-band luminosity of the E component is slightly higher than that of the W component.
252: On K'-band, the E component is about 1.1 magnitudes brighter than the W component.
253: Furthermore, the E component shows strong optical emission lines and has a Seyfert II
254: spectral type. In contrast, the W component shows no detectable emission lines.
255: We detected CO(J=1-0) emission from the E component only.
256:
257: The E component has Seyfert II spectra and bright near-infrared nucleus. Radio continuum observations at 20 cm with the VLA -
258: B array (4 arcsec resolution) detected a fairly strong continuum source of 38 mJy (Crawford et al. 1996) at
259: the same position of this galaxy.
260: This result is expected as Seyfert galaxies are known to be strong continuum sources. An obscured AGN at the center of the
261: E component is likely the source of most of the continuum emission.
262:
263: Our new K'-band image of IRAS F14394$+$5332 (Fig. 1) under good seeing condition (0.5$^{\prime\prime}$) reveals
264: that the E component has two nuclei in the center with nuclear separation of 1.29$^{\prime\prime}$.
265: Thus, K'-band morphology suggests that this galaxy is already a merging system.\\
266: %
267: {\bf IRAS F17028$+$5817} -
268: On the R-band image (Fig. 1), the W component is larger than the E component, but it has a more diffuse light distribution.
269: Faint emission resembling tidal tails can be seen in the W component, in the R-band images of
270: Kim et al. (2001) and Murphy et al. (1996).
271: On K'-band image, the W component is much brighter than the E
272: component and shows a compact light distribution.
273: The CO emission was detected in this W component. The W component also has a very red nuclear color
274: as observed in the typical merger ULIRGs (Kim et al. 2001).
275: Strong optical emission lines detected from both components indicate that they are at the same redshift.
276: The spectral types of the W \& E components are LINER and HII region, respectively.\\
277: %
278: {\bf IRAS F23327$+$2913} -
279: On the R-band image (Fig. 1), the S component resembles a normal spiral with very thick bar structure,
280: whereas the N component looks much more disturbed with a big plume on the northern direction and
281: has small protrusion on the eastern direction.
282: Both components have comparable R-band magnitudes.
283: In K'-band the S component shows a compact distribution and is brighter than the N component.
284: Interestingly, the optical emission lines were detected only from the S component and
285: not from the much disturbed N component, except for the strong stellar absorption line of
286: Mg Ib $\lambda$5175{\AA} (EW (Mg Ib) = 5.5{\AA}, Fig. 3).
287: The spectral type of the S component is LINER and the CO emission line was detected only from this component.
288: Radio continuum emission at 20 cm and 6 cm has been detected in this component (Crawford et al. 1996).
289: The continuum emission at 6 cm was resolved by the 4 arcsec beam of the VLA.\\
290: The large equivalent width of the Mg Ib absorption line which originates from the intermediate stellar population
291: ($\ge 10^9$ yr, Bica, Alloin, \& Schmidt 1990), and the lack of Balmer absorption lines
292: (originating from stellar population of age $10^8-10^9$ yr) both suggests that
293: star formation activity in this galaxy has already ceased some time ago.
294: It is unclear whether the big northern plume on the N component is a remnant of past tidal interaction
295: or caused by recent tidal interaction with the S component. The latter is more likely,
296: since if the plume was caused by the past tidal interaction more than 10$^9$ years ago then,
297: it should already have been relaxed at this time and not be observable.\\
298: %
299: \section{Discussion}
300: \subsection{Molecular Gas Mass}
301: The most interesting result of our observations is the detection of CO emission from only
302: one component in the pair of the selected ULIRGs. The molecular gas content in this
303: component is quite high (several times 10$^{10}$ M$_\odot$) if we adopt the Milky Way conversion
304: factor. However, the value of the conversion factor has been suggested to be lower for ULIRGs (Solomon et al. 1997).
305: By modelling the high resolution observations of CO lines from nearby ULIRGs, Downes \& Solomon (1998) derived
306: molecular gas masses which are a factor of 5 lower.
307: %
308:
309: From the IRAS fluxes at 60 $\mu m$ and 100 $\mu m$ we have estimated the dust temperatures in the
310: ULIRGs detected with CO emission. The dust temperatures are roughly the same in these galaxies,
311: $\sim$ 40 K (Table. 4). If we
312: assume that the gas is in thermal equilibrium with the dust, namely T$_{\rm gas}$ = T$_{\rm dust}$,
313: we can estimate the radius of the CO emitting region from the observed CO flux F$_{\rm CO}$ (Solomon et al. 1997):
314: \begin{eqnarray}
315: L^{'}_{\rm CO} & = & 3.25\times 10^7 \: F_{\rm CO}\nu_{\rm obs}^{-2} D^{2}_L (1 +z)^{-3} \\
316: R_{\rm CO} & = & (L^{'}_{\rm CO}/\pi T_{\rm gas} f_{\rm V} \Delta V_{\rm FWHM})^{0.5}
317: \end{eqnarray}
318: where f$_{\rm V}$ is the velocity filling factor, $\Delta$V$_{\rm FWHM}$ is the
319: linewidth and $\nu_{\rm obs}$ is the observed frequency in GHz, D$_L$ is the luminosity distance in Mpc.
320: For simplicity, we assume that
321: the velocity filling factor f$_{\rm V}$ is equal to unity. We also assume the thermalization and optical thickness
322: of CO(J=1-0) line, with the
323: CO(J=1-0) brightness temperature equal to the gas temperature. The dynamical (virial) mass (Table. 4)
324: contained inside the CO emitting region is R$_{\rm CO}\Delta V_{\rm FWHM}^2$/G, where G is the gravitational constant.
325: We find that the dynamical
326: masses are a factor of 3 to 4 lower than the molecular gas masses. As a result, the H$_2$-to-CO luminosity
327: conversion factor for ULIRGs must be lower than the Galactic value. It was suggested by Downes et al. (1993)
328: and Solomon et al. (1997) that in ULIRGs the dynamical mass is dominated by the molecular gas mass. In this
329: case, the dynamical masses estimated above are the upper limit to the actual molecular gas mass.\\
330: %
331: \subsection{Far Infrared and Radio Continuum Emission}
332: %
333: The component containing most of the molecular gas is also the most active galaxy in the pair, having either
334: Seyfert II or LINER-like spectral type. The other component is less active having either a HII spectral type typical
335: of starburst galaxies or no detectable emission lines. Since there is a correlation
336: between CO emission and the 100 $\mu m$ flux from luminous and ultraluminous galaxies (Solomon et al. 1997),
337: the galaxy detected in CO is likely to be the source of
338: most of the far-infrared luminosity coming from the system and presumably the site of vigorous
339: star formation activity. Our result is consistent with the Evans et al. (1999, 2000) who found
340: most of the molecular gas in several warm double-nucleus ULIRGs to be associated with only the Seyfert nuclei.
341: All ULIRGs in our sample are also found to be associated with 20 cm continuum sources
342: (Becker et al. 1994, Condon et al. 1997). The continuum sources have flux densities ranging from
343: a few mJy to $\sim$ 40 mJy. For two galaxies IRAS F14394$+$5332 and IRAS F23327$+$2913, high angular continuum observations
344: have been reported (Becker et al. 1994, Crawford et al. 1996). The continuum sources in these two galaxies coincide with the position
345: of CO(J=1-0) emission. In particular, the spectral index of radio continuum emission derived from 20 cm and 6 cm
346: observations of IRAS F23327$+$2913 is -0.52 (Crawford et al. 1996), suggesting a non-thermal nature of the emission.
347: For other ULIRGs in our sample, due to the lack of observations at other frequencies,
348: we could not determine the nature of the continuum
349: emission. However, IRAS F14394$+$5332E, a Seyfert II galaxy, appears to have stronger radio
350: continuum flux relative to CO luminosity, compared to other ULIRGs in this sample. Significant fraction of
351: the radio continuum could arise from an AGN in this galaxy.
352: %
353: \subsection{What is the origin of the widely separated ULIRGs ?}
354: From the optical and K'-band images we found evidence that the active component of IRAS F14394$+$5332
355: possesses two nuclei. IRAS F14394$+$5332 and 3 other systems identified by Kim et al. (2001) in
356: the 1 Jy sample of ULIRGs by imaging and spectroscopic data can be classified as multipled mergers.
357: Recent survey with the HST images of
358: 113 ULIRGs by Borne et al. (2000) found that the fraction of multiple merger among ULIRGs might be as high as 20\%.
359: However, this percentage could be overestimated since Borne et al. (2000) did not identify real interacting members
360: with spectroscopic observations. Studies of the ULIRGs and luminous infrared galaxies (LIG), which are less
361: luminous in the infrared than ULIRGs, by Gao (1996, 1997) also found that ULIRGs and LIGs have
362: strong clustering. They suggested that some LIGs and ULIRGs were formed by the merger of groups of gas-rich galaxies.
363:
364: Other systems studied in our paper appear to be in the early stage of interaction. Thus, these systems do not fit into the
365: standard interpretation of the ULIRGs because the star-formation activity is expected to happen at the late stage of interaction.
366: Murphy et al. (1996) also found 7 ULIRGs in their sample with separation greater than 10 kpc. They suggest several
367: possible scenarios to explain these particular systems such as the presence of a third nucleus from previous encounter
368: or the ULIRG phenomenon can occur early in the interaction. If the second scenario applies to the ULIRGs considered here,
369: our detection of molecular gas in only one component of the system
370: indicates that the ULIRG phenomenon might be triggered during the early phase of interaction in a system
371: comprised of a gas-rich galaxy and a much less gas-rich galaxy.
372:
373: The IRAS F23327$+$2913 system is unique in the whole 1 Jy sample of the ULIRGs compiled by Kim \& Sanders (1998).
374: The presence of absorption features in the optical spectrum of the northern component suggests that the N component
375: did not experience star formation episode for the last $10^9$ yrs at least.
376: The molecular gas is detected in the Southern component which has a LINER spectrum. Thus, spectroscopic
377: evidence suggests that the S component is more active than the N companion.
378: More surprisingly, in the R-band image the S component appears as a spiral galaxy with a large bar-like structure at the center. Therefore, the morphology of
379: the S component is very different from other ULIRGs which have either elliptical appearance or strongly disturbed morphology (Sanders et al. 1988,
380: Borne et al. 2000, Kim et al. 2001). Numerical simulations of the merger involving two massive disk galaxies (Mihos \& Hernquist 1996, Springel 2000) showed
381: that the two galaxies are strongly disturbed during the interacting phase and the final product of the merger resembles an elliptical galaxy.
382: Therefore, IRAS F23327$+$2913 does not fit into the commonly accepted scenario of merger between two disk galaxies for ULIRGs.
383: Other mechanisms should be considered to explain the presence of the large reservoir of molecular gas as well as the huge far-infrared
384: luminosity of the S component. The extended 6 cm radio continuum emission reported by Crawford et al. (1996) indicates that
385: the star formation activity is not concentrated in the nuclear region as in other ULIRGs (Downes \& Solomon 1998, Bryant \&
386: Scoville 1999) but spread over the galaxy.
387: The existence of the thick bar in a system of two merging disk galaxies can
388: be observed in the early stage of the tidal interaction in the numerical simulations (eg. Noguchi 1988, Barnes 1992, Barnes \& Hernquist 1996).
389: If this is the case, then we are witnessing the earliest stage of interaction in ULIRGs.
390: In order to become an ULIRG at this early phase of the interaction, the S component
391: should have a large amount of gas initially or the timescale to collect gas into the nuclear region is very short.
392: In addition, the presence of a bar-like structure suggests that the molecular gas might be driven
393: to the center of the galaxy under the influence of a non-axisymmetric gravitational potential.
394: However, Ho et al. (1997) and Mulchaey \& Regan (1997) found that Seyfert and LINER galaxies do not show any significant
395: preference for bars. As a result, it's still unclear
396: whether a bar is effective enough to drive the large amount of molecular gas currently found in the S component.
397: Higher resolution observations of CO emission lines and radio continuum are needed to better understand the origin of this
398: galaxy.
399: %
400: \section{Conclusion}
401: We have detected CO(J=1-0) emission line in 5 out of 6 widely separated ULIRGs selected from the complete 1 Jy sample. The molecular gas mass
402: present in these ULIRGs indicates that they are gas-rich. Our high angular resolution observations also reveal the concentration of
403: the molecular gas in only one galaxy of the pair.
404: Optical spectra suggest that galaxy is also the most active in the pair, having either LINER or Seyfert II spectral type.
405: Using optical and K'-band images of the
406: ULIRGs, we found evidence in support of the multiple merger scenario for IRAS F14394+5332. Other systems are likely
407: in the early phase of interaction.
408: In particular, optical image of IRAS F23327$+$2913 reveals that a large amount of molecular gas is
409: concentrated in a spiral-like galaxy with no evidence for strong tidal interaction.
410: Our observation results suggest that
411: the conventional view of ULIRG phenomenon as the final phase of the merger of two massive disk galaxies
412: probably does not apply to the widely separated ULIRGs. Other mechanisms are needed
413: to explain their origin. Future sensitive and high resolution observations will help to clarify the
414: nature of these unusual ULIRGs.
415: %
416: \acknowledgments
417:
418: We would like to express our thanks to an anonymous referee for critical
419: comments which are very helpful in improving our paper. We also thank Dr. S. Veilleux
420: for providing us the data prior to publication.
421:
422: \begin{references}
423: \reference{ba91} Barnes, J.E., \& Hernquist, L.E., 1991, ApJL, 370, L65
424: \reference{ba92} Barnes, J.E., 1992, ApJ, 393, 484
425: \reference{ba96} Barnes, J.E., \& Hernquist, L.E., 1996, ApJ, 471, 115
426: \reference{be} Becker, R. H., White, R. L., \& Helfand, D. J. 1994, ASP Conference Series, v. 61, eds. D. R.
427: Crabtree, R. J. Hanisch, J. Barnes, p. 165
428: \reference{bi} Bica, E., Alloin, D., \& Schmitt, H.R., 1994, A\&A 285, 868
429: \reference{bo} Borne, K.D., Bushouse, H., Colina, L., \& Lucas, R.A., 2000, ApJL, 529, L77
430: \reference{br} Bryant, P.M., \& Scoville, N.Z., 1999, AJ, 117, 2632
431: \reference{com} Combes, F., Maoli, R., \& Omont, A., 1999, A\&A 345, 369
432: \reference{con} Condon, J.J. et al., 1998, AJ, 115, 1693.
433: \reference{cr} Crawford, T., Marr, J., Partridge, B., \& Strauss, M., 1996, ApJ 460, 225
434: \reference{dow93} Downes, D., Solomon, P.M., \& Radford, S.J.E., 1993, ApJL, 414, L13
435: \reference{dow98} Downes, D., \& Solomon, P.M., 1998, ApJ 507, 615
436: \reference{el}Elias, J.H., Frogel, J.A., Matthews, K., \& Neugebauer, G., 1982, AJ 87, 1029
437: \reference{ev99} Evans, A.S., Kim, D.C., Mazzarella, J.M., Scoville, N.Z., \& Sanders, D.B., 1999, ApJL, 521, L107
438: \reference{ev00} Evans, A.S., Surace, J.A., Mazzarella, J.M., \& Sanders, D.B., 2000, in Science with the Atacama Large Millimeter Array,
439: ed. A. Wootten, in press.
440: \reference{gao96} Gao, Y., 1996, Ph.D thesis, Univ. of New York at Stony Brook
441: \reference{gao97} Gao, Y., 1997, PASP, 109, 1189
442: \reference{gao99} Gao, Y., \& Solomon, P.M., 1999, ApJL, 512, L99
443: \reference{gea} Genzel, R., Lutz, D., Sturm, E., Egami, E., Kunze, D., Moorwood, A.F.M., Rigopoulou, D.,
444: Spoon, H.W.W., Sternberg, A., Tacconi-Garman, L.E., Tacconi, L., \& Thatte, N., 1998a, ApJ 498, 579
445: \reference{geb} Genzel, R., Lutz, D., \& Tacconi, L., 1998b, Nature 395, 859
446: \reference{hi} Hickson, P., 1982, ApJ, 255, 382
447: \reference{ho} Ho, L.C., Filippenko, A.V., \& Sargent, W.L.W, 1997, ApJ, 487, 591
448: \reference{kim95} Kim, D-C., 1995, Ph.D. thesis, Univ. of Hawaii
449: \reference{kim98} Kim, D-C., \& Sanders, D.B. 1998, ApJS, 119, 41
450: \reference{kim01} Kim, D-C. et al., 2001, in preparation
451: \reference{le} Leech, K.J., Rowan-Robinson, M., Lawrence, A., \& Hughes, J.D., 1994, MNRAS 267, 253
452: \reference{lu} Lutz, D., Veilleux, S., \& Genzel, R., 1999, ApJL, 517, L13
453: \reference{mi94} Mihos, J.C., \& Hernquist, L. 1994, ApJL, 431, L9
454: \reference{mi96} Mihos, J.C., \& Hernquist, L. 1996, ApJ, 464, 641
455: \reference{mul} Mulchaey, J.S., \& Regan, M.W., 1997, ApJL, 482, L135
456: \reference{mu} Murphy, T.W., Armus, J.L, Soifer, B.T., Mazzarella, J.M., Shupe, D.L., Strauss, M.A., \&
457: Neugebauer, G., 1996, AJ, 111, 1025
458: \reference{no} Noguchi, M., 1988, A\&A 203, 259
459: \reference{san88} Sanders, D.B., Soifer, B.T., Elias, J.H., Neugebauer, G., \& Matthews, K., 1988, ApJL, 328, L35
460: \reference{san96} Sanders, D.B., \& Mirabel I.F., 1996, ARAA, 34, 749
461: \reference{sau95} Sault, R.J., Teuben, P.J., \& Wright, M.C.H., 1995, in ASP Conf. Proc. 77,
462: Astronomical Data Analysis Software and Systems IV, ed. R.A. Shaw, H.E. Payne, J.J.E. Hayes (San Francisco: ASP), 443
463: \reference{so} Solomon, P.M., Downes, D., Radford, S.J.E., \& Barret, J.W., 1997, ApJ, 478, 144
464: \reference{sp} Springel, V., 2000, MNRAS 312, 859
465: \reference{tan98} Taniguchi, Y., \& Shioya, Y., 1998, ApJL, 501, L167
466: \reference{veil99a} Veilleux, S., Kim, D-C., \& Sanders, D.B., 1999a, ApJ, 522, 113
467: \reference{veil99b} Veilleux, S., Sanders, D.B., \& Kim, D.-C., 1999b, ApJ, 522, 139
468: \reference{wel} Welch, W.J., et al. 1996, PASP, 108, 93
469: \reference{zou} Zou, Z., Xia, X., Deng, Z., \& Su, H., 1991, MNRAS, 252, 593
470: \end{references}
471:
472:
473: \begin{figure*}[ht]
474: \setlength{\unitlength}{1cm}
475: \begin{picture}(16.0,20.0)
476: \put(0,13){\resizebox{16cm}{!}{\includegraphics*{1048.ps}}}
477: \put(0,6.85){\resizebox{16cm}{!}{\includegraphics*{1118.ps}}}
478: \put(0,0){\resizebox{16cm}{!}{\includegraphics*{1235.ps}}}
479: \end{picture}
480: \caption{Left frame: the contour plots of the CO(J=1-0) emission in ULIRGs overlayed on the
481: R-band images. The first CO(J=1-0) contour is at 50\% of the peak. Central frame: contour plot
482: of the same R-band images at the same scale. Right frame: K'-band images.}
483: \end{figure*}
484:
485: \addtocounter{figure}{-1}
486:
487: \begin{figure*}[ht]
488: \setlength{\unitlength}{1cm}
489: \begin{picture}(16.0,20.0)
490: \put(0,13){\resizebox{16cm}{!}{\includegraphics*{1439.ps}}}
491: \put(0,6.85){\resizebox{16cm}{!}{\includegraphics*{1702.ps}}}
492: \put(0,0){\resizebox{16cm}{!}{\includegraphics*{2332.ps}}}
493: \end{picture}
494: \caption{Continued. Left frame: the contour plots of the CO(J=1-0) emission from ULIRGs in our sample overlayed on their
495: R-band images. Central frame: contour plot of the same R-band images at the same scale. Right frame: K'-band images.
496: The small image superposed on the right frame of IRAS F14394+5332 is the K'-band blown-up image of the E component.}
497: \end{figure*}
498:
499: \begin{figure*}[ht]
500: \setlength{\unitlength}{1cm}
501: \begin{picture}(16.0,12.0)
502: \put(0,5.){\resizebox{6cm}{!}{\includegraphics*{s1118.ps}}}
503: \put(6.5,5.){\resizebox{5.5cm}{!}{\includegraphics*{s1235.ps}}}
504: \put(12.5,5.){\resizebox{5.5cm}{!}{\includegraphics*{s1439.ps}}}
505: \put(0,0){\resizebox{6cm}{!}{\includegraphics*{s1702.ps}}}
506: \put(6.5,0){\resizebox{5.5cm}{!}{\includegraphics*{s2332.ps}}}
507: \end{picture}
508: \caption{The CO(1-0) line profiles from 5 ULIRGs. The velocity
509: is relative to the galaxy's velocity measured from optical spectra.}
510: \end{figure*}
511:
512: \begin{figure*}[ht]
513: \setlength{\unitlength}{1cm}
514: \begin{picture}(10.0,10.0)
515: \put(0,-2){\resizebox{16cm}{!}{\includegraphics*{ir2332n.ps}}}
516: \end{picture}
517: \caption{Optical spectra of IRAS F23327$+$2913 N taken from the University of Hawaii 2.2m
518: telescope on Mauna Kea Observatory. \label{fig3}}
519: \end{figure*}
520:
521: \begin{table}
522: \begin{center}
523: \footnotesize
524: \caption{{\footnotesize Summary of the observations.}}
525: \vspace{0.5cm}
526: \begin{tabular}{lcccl}
527: \tableline
528: Name & Redshift & Obs. Freq & Num. of tracks & remarks \\
529: & & (GHz) & & \\ \hline
530: IRAS F10485$-$1447 & 0.133 & 101.739 & 4 B-array tracks & no detection \\
531: & & & & \\ \hline
532: IRAS F11180+1623 & 0.166 & 98.860 & 4 B-array tracks & no detection \\
533: & & & 2 C-array tracks & CO(J=1-0) detection \\ \hline
534: IRAS F12359$-$0725 & 0.138 & 101.291 & 3 B-array tracks & no detection \\
535: & & & 1 C-array tracks & possible detection of CO(J=1-0)\\ \hline
536: IRAS F14394+5332 & 0.105 & 104.317 & 4 B-array tracks & CO(J=1-0) detection \\
537: & & & 3 C-array tracks & CO(J=1-0) detection \\ \hline
538: IRAS F17028+5817 & 0.106 & 104.222 & 4 B-array tracks & no detection \\
539: & & & 2 C-array tracks & CO(J=1-0) detection \\ \hline
540: IRAS F23327+2913 & 0.106 & 104.160 & 1 B-array track & no detection \\
541: & & & 3 C-array tracks & CO(J=1-0) detection \\
542: \tableline
543: \end{tabular}
544: \end{center}
545: \end{table}
546:
547: \begin{table}
548: \begin{center}
549: \footnotesize
550: \caption{{\footnotesize Observational properties.}}
551: \vspace{-0.25cm}
552: \begin{tabular}{cccccccc}
553: \tableline
554: Name &\multispan{2} Coordinates (J2000.0)$^a$ & D$_L$ & Beam & Total CO flux & M$_{\rm H_2}$ & $\Delta$M$_{\rm H_2}$$^b$ \\
555: & R. A. & Dec. & (Mpc) & (${''}\times{''}$) & (Jy km s$^{-1}$) & (10$^{10}$ M$_\odot$) & (10$^{10}$ M$_\odot$) \\
556: \tableline
557: IRAS F10485$-$1447 & 10$^{\rm h}$51$^{\rm m}$03$^{\rm s}$.42 & -15$^{\rm o}$03'19".8 & 548.6 & 5.5$\times$4.3 &$\le$ 1 & $\le$ 0.4 & 0.4 \\
558: IRAS F11180+1623 & 11$^{\rm h}$20$^{\rm m}$41$^{\rm s}$.52 & 16$^{\rm o}$06'56".6 & 689.5 & 10.2$\times$9.1 & 11 & 5.4 & 0.8 \\
559: IRAS F12359$-$0725 & 12$^{\rm h}$38$^{\rm m}$31$^{\rm s}$.67 & -07$^{\rm o}$42'32".1 & 569.8 & 10.5$\times$6.6 & 6$^c$ & 2 & 0.4 \\
560: IRAS F14394+5332 & 14$^{\rm h}$41$^{\rm m}$02$^{\rm s}$.86 & 53$^{\rm o}$20'07".9 & 430.5 & 4.6$\times$3.8 & 21 & 4.2 & 0.4 \\
561: IRAS F17028+5817 & 17$^{\rm h}$03$^{\rm m}$42$^{\rm s}$.75 & 58$^{\rm o}$13'44".5 & 433.7 & 6.6$\times$5.5 & 14 & 2.8 & 0.3 \\
562: IRAS F23327+2913 & 23$^{\rm h}$35$^{\rm m}$11$^{\rm s}$.87 & 29$^{\rm o}$30'05".8 & 438.9 & 7.2$\times$5.6 & 8 & 1.7 & 0.2 \\
563: \tableline
564: \end{tabular}
565: \end{center}
566: $^a$ From Kim et al. (2000).\\
567: $^b$ 1$\sigma$ upper limit of the molecular gas mass of the companion.\\
568: $^c$ Weak emission at 3$\sigma$ level in 200 km s$^{-1}$ channel maps.
569: \end{table}
570: %
571: %
572: \begin{table}
573: \begin{center}
574: \footnotesize
575: \caption{{\footnotesize Physical parameters of ULIRGs.}}
576: \vspace{0.5cm}
577: \begin{tabular}{llccccccc}
578: \tableline
579: Name & Spectral type$^a$ & z & Proj. Separation & F$_{\rm IR}$ & F$_{\rm 20cm}$ & \multispan{2} Total magnitude \\
580: & & & (kpc) & (10$^{12}$ L$_\odot$) & (mJy) & R-band & K'-band \\
581: \tableline
582: IRAS F10485-1447E & absorption lines & 0.134$^b$ & & & & 18.1 & 15.9 \\
583: IRAS F10485-1447W & LINER & 0.133 & \raisebox{2.0ex}[0cm]{20.0} & \raisebox{2.0ex}[0cm]{1.48} & \raisebox{2.0ex}[0cm]{4.4} & 17.7 & 14.7 \\
584: \hline
585: IRAS F11180+1623E & LINER & 0.166 & & & & 17.9 & 14.7 \\
586: IRAS F11180+1623W & no emission lines & - &\raisebox{2.0ex}[0cm]{21.8} & \raisebox{2.0ex}[0cm]{1.74} & \raisebox{2.0ex}[0cm]{4.7} & 19.4 & 16.1 \\
587: \hline
588: IRAS F12359-0725N & LINER & 0.138 & & & & 18.2 & 14.8 \\
589: IRAS F12359-0725S & HII & 0.138 &\raisebox{2.0ex}[0cm]{22.9} & \raisebox{2.0ex}[0cm]{1.29} & \raisebox{2.0ex}[0cm]{4.5} & 18.2 & 15.9 \\
590: \hline
591: IRAS F14394+5332E & Seyfert II & 0.105 & & & 38.0 & 16.3 & 13.1 \\
592: IRAS F14394+5332W & no emission lines & - &\raisebox{2.0ex}[0cm]{50.5} & \raisebox{2.0ex}[0cm]{1.10} & - & 16.9 & 14.2 \\ \hline
593: IRAS F17028+5817E & HII & 0.106 & & & & 17.4 & 14.8 \\
594: IRAS F17028+5817W & LINER & 0.106 &\raisebox{2.0ex}[0cm]{23.3} & \raisebox{2.0ex}[0cm]{1.26} & \raisebox{2.0ex}[0cm]{16.7} & 17.1 & 13.5 \\
595: \hline
596: IRAS F23327+2913N & absorption lines & 0.106 & & & - & 16.5 & 14.1 \\
597: IRAS F23327+2913S & LINER & 0.107 &\raisebox{2.0ex}[0cm]{22.7} & \raisebox{2.0ex}[0cm]{1.15} & 8.4 & 16.8 & 13.6 \\
598: \tableline
599: \end{tabular}
600: \end{center}
601: $^a$ From Veilleux, Kim, \& Sanders (1999).\\
602: $^b$ S. Veilleux, private communication
603: \end{table}
604: %
605: %
606: \pagebreak
607:
608: \begin{table}
609: \begin{center}
610: \footnotesize
611: \caption{{\footnotesize Dynamical masses derived from CO data.}}
612: \vspace{0.5cm}
613: \begin{tabular}{ccccccccc}
614: \tableline
615: Name & \multispan{2} $f_{\nu}(\lambda)^\dagger$ & T$_{\rm dust}$ & L$^{'}_{\rm CO} $ & $\Delta$V$_{\rm FWHM}^*$ & Radius & Dynamical mass \\ \cline{2-3}
616: & 60 $\mu m$ & 100 $\mu m$ & (K) & (10$^9$K km s$^{-1}$pc$^2$) & (km s$^{-1}$) & (kpc) & (10$^9$ M$_\odot$) \\
617: \tableline
618: IRAS F11180+1623 & 1.19 & 1.60 & 41.5 & 11.0 & 600 & 0.38 & 31.0 \\
619: IRAS F14394+5332 & 1.95 & 2.39 & 43.3 & 8.6 & 380 & 0.40 & 13.3 \\
620: IRAS F17028+5817 & 2.43 & 3.91 & 36.9 & 5.8 & 330 & 0.40 & 10.0 \\
621: IRAS F23327+2913 & 2.1 & 2.81 & 37.9 & 3.4 & 200 & 0.38 & 3.5 \\
622: \tableline
623: \end{tabular}
624: \end{center}
625: $^\dagger$ From Kim \& Sanders (1998)\\
626: $^*$ The FWHM is derived from gaussian fitting of the line profile
627: \end{table}
628:
629: \end{document}
630: