Morphological and photometric evolution of ultra-luminous infrared galaxies: Nature of faint SCUBA sources

o
o
\o
o
(^
a^

:^

Oh'

o

U

X

Morphological and photometric evolution of ultra-luminous
infrared galaxies: Nature of faint SCUBA sources

Kenji Bekki, Yasuhiro Shioya^, and Ichi Tanaka
Astronomical Institute, Tohoku University, Sendai, 980-8578, Japan

Received ; accepted

^Center for Interdisciplinary Research, Tohoku University, Sendai, 980-8578, Japan

-2-

ABSTRACT

We investigate when and how a dusty starburst galaxy merger can be heavily
obscured by dust and consequently becomes an ultra-luminous infrared galaxy
(ULIRG), based on numerical simulations of chemodynamical and photometric
evolution of dusty gas-rich major galaxy mergers. We found that a major galaxy
merger is more likely to become an ULIRG preferentially in the merger late
phase, when the two disks become very close and the very high-density dusty gas
can obscure heavily the central secondary massive starburst. We furthermore
show how the optical and near-infrared morphology of a simulated ULIRG at
intermediate (z=0.4) and high redshift (z=l — 2) can be observed by the Hubble
Space Telescope (HST) in order to present a plausible explanation for the
origin of some host galaxies of the faint SCUBA sources recently observed by
Small et al. (1998). The results of our numerical simulations imply that some
SCUBA sources with apparently faint and compact HST morphology can be
higher redshift dust-enshrouded starburst mergers with their outer low surface
brightness tidal features hardly detectable in the present optics of the HST.

Subject headings: galaxies: elliptical and lenticular, cD - galaxies: formation -
galaxies: ISM - galaxies: infrared - galaxies: interaction - galaxies: structure

-3-

1. Introduction

Recent observational studies with the Sub-milhmeter Common-User Bolometer Array
(SCUBA) (Holland et al. 1999) on the James Clerk Maxwell Telescope have revealed
possible candidates of heavily dust-enshrouded starburst galaxies at intermediate and
high redshift, which could be counterparts of low redshift ultra-luminous infrared galaxies
(Small, Ivison, & Blain 1997; Hughes et al. 1998; Small et al. 1998; Barger et al. 1999;
Lilly et al. 1999). Although optical morphology of these sub-millimeter galactic sources
should be treated with caution owing to the absence of high resolution sub-mm imaging
capability (Richards 1999), more than 50 % of those are suggested to show the indication
of galactic interaction and merging (Small et al. 1998). The huge amount of molecular
gas (~ 5x 10^° Mq) derived from CO emission and the inferred large star formation rate
(~ lO^M0yr~^) in a few sub-millimeter extragalactic sources (e.g. SMMJ14011-I-0252)
are suggested to be consistent with the scenario that the formation of some intermediate
and high redshift elliptical galaxies is associated with galaxy mergers with their starburst
activities heavily hidden by dust extinction (Frayer et al. 1998; Frayer et al. 1999).
Furthermore observational studies of an extremely red object ERO J 164502-4626.4 (HR 10)
with the redshift of 1.44 by the Hubble Space Telescope and the SCUBA have found that
this high redshift galaxy is also a dust-enshrouded starburst galaxy with clear indication of
galaxy merging/interaction (Graham & Dey 1996; Cimatti et al. 1998; Dey et al. 1998).

Although low redshift ultra-luminous infrared galaxies (ULIRG) are generally
considered to be ongoing galaxy mergers with the triggered prominent nuclear activities
(starburst or AGN) heavily obscured by dust (Sanders et al. 1988; Sanders & Mirabel
1996), the origin of the faint SCUBA sources is not so clearly understood. In this Letter, we
numerically investigate both morphological and photometric properties of dusty starburst
galaxy mergers in an explicitly self-consistent manner and thereby demonstrate when and

-4-

how a gas-rich merger between two spirals becomes an ULIRG. We particularly discuss
the origin of very faint and compact HST optical morphology of possible host galaxies of
some faint SCUBA sources recently revealed by Small et al. (1998) by demonstrating how
intermediate and high redshift dust-enshrouded starburst galaxies can be seen by the HST
WFPC2 and NICMOS (NIC2).

Model

The most remarkable difference in investigating photometric evolution of dusty galaxies
between the present model and previous one- zone models (e.g., Mazzei, Xu, & De Zotti
1992; Franceschini et al. 1994; Guiderdoni et al. 1998) is that we derive spectral energy
distribution (SED) of galaxies based on the result of numerical simulations that can follow
both dynamical and chemical evolution of galaxies. The numerical techniques for solving
galactic chemodynamical and photometric evolution are given in Bekki & Shioya (1999a,b),
and accordingly we describe only briefly the present numerical model of dusty starburst
galaxy mergers here. We construct models of galaxy mergers between gas-rich disk galaxies
with equal mass by using Fall-Efstathiou model (1980). The total mass and the size of a
progenitor disk are 6.0 x 10^'^ Mq and 17.5 kpc, respectively. The coUisional and dissipative
nature of the interstellar medium with the initial total mass equal to 3.0 x lO^'' M© is
modeled by the sticky particle method ( pchwarz 1981| ). Star formation is modeled by

converting the coUisional gas particles into collisionless new stellar particles according to the
Schmidt law (Schmidt 1959) with the exponent of 2.0. Chemical enrichment through star
formation during galaxy merging is assumed to proceed both locally and instantaneously
in the present study. The fraction of gas returned to interstellar medium in each stellar
particle and the chemical yield are 0.2 and 0.03, respectively. Initial metallicity Z^, for each
stellar and gaseous particle in a given galactic radius R (kpc) from the center of a disk is

-5-

given according to the observed relation Z^ = 0.06 x iq^"'^^'''^*^^'^'^-' of typical late-type disk
galaxies (e.g., Zaritsky, Kennicutt, & Huchra 1994). Mean ages of old stellar components
initially in a merger progenitor disk at the redshift z=0.4, 1.0, and 1.5 are assumed to be
7.14, 3.80, and 2.46 Gyr, respectively. The orbital plane of a galaxy merger is assumed
to be the same as x-y plane and the initial distance between the center of mass of merger
progenitor disks is 140 kpc. Two disks in the merger are assumed to encounter each other
parabolically with the pericentric distance of 17.5 kpc.

It is assumed in the present study that the SED of a model galaxy is a sum of the
SED of stellar particles. We first calculate dust extinction of star light for each stellar
particle and dust temperature for each gaseous particle, based on the three dimensional
spacial distribution of stellar and gaseous particles. We then sum each stellar particle's SED
corrected by dust extinction and the dust re-emission of each gaseous particle. The method
to derive the dust extinction and re-emission for each particle is described as follows.
Absorption of star light of a stellar particle is modeled according to the following reddening
formulation (Black 1987; Mazzei et al. 1992); E{B-V) = A^(H)/4.77x lO^^cm'^x (Zg/0.02),
where A^(H) and Zg are gaseous column density and gaseous metallicity, respectively. Using
the derived redding E{B — V) and extinction law by Cardelli et al. (1989), we adopt the
so-called screen model and calculate the absorption of the stellar particle. By assuming the
modified black body radiation with the emissivity (e) law e oc z/^, we determine the dust
temperature of a gas particle such that total energy flux of dust absorption is equal to that
of dust re-emission. We here do not include the albedo for dust grains, which means that
only stars and new stars heat dust. For calculating the SED of each stellar particle with a
given age and metallicity, we use the spectral library GISSEL96 which is the latest version
of Bruzual & Chariot (1993). Using the derived SED, we investigate how dust-enshrouded
starburst galaxy mergers at z=0.4, 1.0, and 1.5 can be seen in the HST. The method
to construct the synthesized HST images of galactic morphology in the present study is

-6-

basically the same as that described by Mihos (1995). In the followings, the cosinological
parameters Hq and qo are set to be 50 km s"^ Mpc~^ and 0.5 respectively.

EDITOR: PLACE FIGURE |l| HERE.
EDITOR: PLACE FIGURE | HERE.

EDITOR: PLACE FIGURE | HERE.

3. Result

Figure 1 and 2 describe the time evolution of star formation rate and the rest-frame
SED in a dusty gas-rich major galaxy merger at z=1.0 and the mass distribution of stellar
and gaseous components of the merger at the epoch of the maximum secondary starburst,
T = L3 Gyr, respectively. From now on, the time T represents the time that has elapsed
since the two disks begin to merge at each redshift. As is shown in Figure 1, star formation
rate in the present prograde- retrograde merger becomes maximum (~ 378M0/yr) at T = 1.3
Gyr, when two disks of the merger become very close to suffer from final violent relaxation.
These results are qualitatively consistent with those in Mihos & Hernquist (1996). The
SED at T = 1.3 Gyr in Figure 1 is rather similar to the observed SED of typical ULIRGs,
which implies that the present dusty starburst merger model can be observed as an ULIRG
in the late of galaxy merging. The SED of a dusty merger depends on encounter parameters
in such a way that a prograde-retrograde merger shows stronger infrared re-emission in the
merger late phase than prograde-prograde one (Bekki & Shioya 1999b).

-7-

Figure 2 shows that both new stars and gas are more centrally concentrated than
old stellar components, which means that star light from new stars formed by secondary
massive starburst are more heavily obscured by dusty interstellar medium than that of old
stellar components. These results in Figure 1 and 2 clearly demonstrate that starburst
population in the merger is very heavily obscured by dusty interstellar medium {Ay ~ 2.46
mag) and consequently the dust re-emission in far-infrared ranges becomes very strong at
the maximum starburst {Lj^ = 1.59 x lO^^L©). The reasons for this heavy dust extinction
are firstly that column density of dusty interstellar medium becomes extremely high owing
to the strong central accumulation of gas and secondly that the central gaseous metallicity,
which is a measure of total amount of dust, also becomes rather large (~ 0.04) because
of efficient and rapid chemical evolution in galaxy merging with secondary starburst. The
galaxy merger 0.05 Gyr before and after the maximum starburst does not show the SED
characteristics of ULIRGs, which suggests that the time-scale during which the merger
remnant can be observed as an ULIRG is very short (less than 0.1 Gyr). Owing to the
heavy dust extinction, the merger shows very red rest-frame color V — 1=0.97 for a starburst
galaxy.

Figure 3 shows how the optical and near-infrared morphology of this ULIRG merger
at T = 1.3 Gyr can be observed by the HST WFPC2 and NICMOS if it locates at the
redshift z=0.4, 1.0, and 1.5 (See also Figure 2 for the raw image of the merger). Outer
diffuse large arc-like structure composed mainly of old stars, which is discernibly seen in
the upper left part of Figure 2, can not be seen even in the merger at z=0.4. Tidal feature
indicative of major galaxy merging can be discernibly seen at z=0.4 and can not be seen at
all at z=1.0 and 1.5. The faint and compact morphology of the present merger remnant
at z=1.0 and 1.5 is similar to some possible faint optical counterparts of the first SCUBA
sources observed by Small et al. (1998). We furthermore stress that it is hard to observe
the tidal feature even in the HST image with the exposure time equal to 1.24 x 10^ sec

corresponding to that for the Hubble Deep Field (HDF). These results imply that it is
not so easy task to clarify the origin of high redshift ULIRGs (z > 1) solely by optical
and near-infrared HST morphological studies; If the formation of high redshift ULIRGs is
closely associated with major galaxy merging, even the deepest existing HST surveys have
some difficulties in proving that. The 850 /im sub-millimeter energy flux of the ULIRG
merger is found to be 10.85 mJy for z=1.0 and 2.36 mJy for z=1.5. Considering the flux
value of the merger at z=1.0 and 1.5 exceeding the observed SCUBA flux 2.1 — 7.0 mJy by
Hughes et al. (1998) and 3.3 — 4.6 by Barger et al. (1998) and faint compact morphology
of the merger, a high redshift dust-enshrouded starburst merger (z~1.0) can be observed as
a very compact SCUBA source. Furthermore the apparently compact remnant shows very
red R — K color (4.70 mag for z=1.0 and 4.86 for z=1.5) for its K band magnitude (21.26
mag for z=1.0 and and 22.42 for z=1.5), which implies an evolutionary link between dusty
mergers and the high redshift extremely red objects (EROs). The present model describes
the detectability of only one merger model with the initial disk mass Mj = 6.0 x 10^°
M© in the present optics of the HST. Accordingly we must stress that the detectability of
tidal feature of a high redshift ULIRG in the HST can depend on mass and luminosity of
a merger and furthermore that the future improved HST Advanced Camera for Surveys
(ACS) is expected to reveal the more detailed morphology of a high redshift ULIRG.

4. Discussion and conclusion

Our numerical results have demonstrated that a dust-enshrouded starburst merger
can be observed as a SCUBA source with faint and compact morphology at high redshift.
The present results furthermore provide clues to the understanding of the evolutionary
link between the extremely red objects 'ERO' (e.g., Elston, Rieke, & Rieke 1988) and the
present day elliptical galaxies. Recent observational studies have revealed that ERO J

-9-

164502+4626.4 (HR 10) is not a red elliptical galaxy passively evolving at high redshift
but a dust-enshrouded starburst galaxy at redshift z=1.44 (Cimatti et al. 1998; Dey et
al. 1998). Peculiar morphological properties in rest-frame far-red and near-UV range, the
large 850 /xm energy flux, and the inferred huge star formation rate in this HR 10 are
qualitatively consistent with the present numerical results of dusty starburst mergers, which
implies that some EROs are dusty starburst major mergers and thus evolve eventually
into the present-day elliptical galaxies. Then, what fraction of the present-day ellipticals
have experienced the past dusty ERO phase ? Considering both the very short time-scale
(less than 0.1 Gyr in the present study) during which dusty mergers can be identified as
EROs and the high surface density of EROs (0.14 arcmin"^ for EROs with R — K > 6 and
K < 19.75) derived by Beckwith et al. (1998), the discovery of the possible dusty starburst
galaxy merger HR 10 implies that the fraction is likely to be fairly large. Furthermore,
the increase of number fraction of apparently interacting/merging galaxies with redshift in
the HDF (Driver et al. 1998), the very small fraction (less than 8 %) of early-type E/SO
galaxies in the HDF (Marleau & Simard 1998), and the observed high surface density of
EROs (Beckwith et al. 1998; Dey et al. 1998) combine to motivate us to interpret that a
sizable fraction of the present-day elliptical galaxies were previously ongoing major mergers
with dusty starburst at intermediate and high redshift. However, spectroscopic studies
of EROs have not been so accumulated yet which can reveal unambiguously the redshift
of EROs thus determine whether EROs are really dusty starburst galaxies or such high
redshift red elliptical galaxies as have been recently discovered by the VLT (Benitez et al.
1998). Future extensive spectrophotometric studies which can reveal whether the extremely
red colors of EROs result from dust effects or from galaxy aging will determine the relative
importance of major galaxy merging in the formation of high redshift EROs.

We are grateful to the anonymous referee for valuable comments, which contribute to

-10-

improve the present paper. K.B. and Y.S. thank to the Japan Society for Promotion of
Science (JSPS) Research Fellowships for Young Scientist.

-11-

REFERENCES

Barger A. J., Cowie, L. L., Sanders, D. B., Fulton, E., Taniguchi, Y., Sato, Y., Kawara, K.,
Okuda, H. 1998, Nature, 394, 248

Barger A. J., Cowie, L. L., Small, I., Ivlson, R. J., Blaln, A. W., Knelb, J-P. 1999,
astrb-pli/9903142

Beckwlth, S. V. W., Thompson, D., Manuccl, F., & Djorgovskl, S. G. 1998, ApJ, 504, 107

Bekkl, K., & Shloya, Y. 1999a, ApJ, 513, 108

Bekkl, K., & Shloya, Y. 1999b, In preparation

Benitez, N., Broadhurst, T., Bouwens, R., Silk, J., & Rosatl, P. 1998, |astro-ph/98 12205

Blretta, J, 1996, In the HST WFPC2 Instrumental handbook Version 4.0

"■> " )

Black, J. H. 1987, In Spectroscopy of Astrophyslcal Plasmas, eds. A. Dalgarno & D. Layzer,
Cambridge Univ. Press, Cambridge.

Cardelll, J. A., Clayton, G. C, & Mathls, J. S. 1989, ApJ, 345, 245

Clmattl, A., Andreanl, P., Rottgerlng, H., & Tllanus, R. 1998, Nature, 392, 895

Dey, A., Graham, J. R., Ivlson, R. J., Small, I., Wright, G. S., & Liu, M. C. 1999,

astro-ph/9902044

Driver, S. P., Fernandez-Soto, A., Couch, W. J., Odewahn, S. C, Windhorst, R. A.,
Phillips, S., Lanzetta, K., & Yahll, A. 1998, ApJ, 496, L93

Elston, R., Rleke, G. H., & Rleke, M. J. 1988, ApJ, 331, L77

Fall, S. M., & Efstathlou, G. 1980, MNRAS, 193, 189

-12-

Franceschini, A., Mazzei, P., De Zotti, G., & Danese, L. 1994, ApJ, 427, 140

Frayer, D. T., Ivison, R. J., Scoville, N. Z., Yun, M., Evans, A. S., Small, I., Blaln, A. W.,
& Knelb, J.-P. 1998, ApJ, 506, L7

Frayer, D. T., Ivison, R. J., Scoville, N. Z., Yun, M., Evans, A. S., Small, I., Barger, A. W.,
Blaln, A. W., & Knelb, J.-P. 1999, ApJ, 514, L7

Graham, J. R., & Dey, A. 1996, ApJ, 471, 720

Gulderdonl, B., Hlvon, E., Bouchet, F. R., & Maffel, B. 1998, MNRAS, 295, 877

Holland, W. S., et al. 1999, |astro-pli/980gT^

Holtzman, J., Burrows, C., Casertano, S., Hester, J. J., Trauger, J., Watson, A., & Worthey,
G. 1995, PASP, 107, 1065

Hughes, D. et al.. Nature, 394, 241

Krlst, J., & Hook, R. 1997, Tiny Tim Users Manual Version 4.4 (Baltimore: STScI)

Lilly, S. J., Eales, S. A., Gear, W.K.P., Hammer, F., Le Fevre, O., Crampton, D., Bond, J.
R., Dunne, L. 1999, |astro-ph/990"T047

MacKenty, J. W. et al. 1997, In the HST NICMOS Instrumental Handbook version 2.0
(Baltimore, STScI)

Marleau, F. R., & Slmard, L. 1998, ApJ, 507, 585

Mazzei, P., Xu, C., & De Zottl, G. 1992, A&A, 256, 45

Mlhos, J. C. 1995, ApJ, 438, L75

Mlhos, J. C., & Hernqulst, L. 1996, ApJ, 464, 641

-13-

Richards, E. A., 1999, ApJ, 513, L9

Rigopoulou, D., Lawrence, A., & Rowan- Robinson, M. 1996, MNRAS, 278, 1049

Sanders, D. B., Soifer, B. T., Elias, J. H., Modore, B. F., Mattehews, K., Neugbauer, G.,
Scoville, N. Z. 1988a, ApJ, 325, 74

Sanders, D. B., & Mirabel, I. F. 1996, ARA&A, 34, 749

Schmidt, M. 1959, ApJ, 344, 685

Schwarz, M. R 1981, ApJ, 247, 77

Smail, L, Ivison, R. J., & Blain, A. W. 1997, ApJ, 490, L5

Smail, L, Ivison, R. J., Blain, A. W., & Kneib, J.-R 1998, ApJ, 507, L21

Zaritsky, D., Kennicutt, R. C, Huchra, J. P. 1994, ApJ, 420, 87

This manuscript was prepared with the AAS L^T[t;X macros v4.0.

-14-

Fig. 1. — Upper panel shows the time evolution of star formation rate of a dusty gas-rich
major galaxy merger at z=1.0. The star formation rate of the merger becomes maximum
(~ 378M0/yr) at the time T = 1.3 Gyr. Lower panel shows the rest-frame SED of the
galaxy merger at T = 1.25 (dotted line), T = 1.3 (thick solid), and T = 1.35 Gyr (dashed).
For comparison, the SED of the merger without dust extinction and re-emission is also given
by thin solid line. Note that the far-infrared flux of the merger rapidly increases between
T = 1.25 and T = 1.3 Gyr and decreases between T = 1.3 and T = 1.35 Gyr. Furthermore
we can clearly see the effects of dust extinction and re-emission on the SED shape in the
merger at T = 1.3 Gyr by comparing the thick solid line and the thin solid one. In this
figure, the observed SED of Arp 220 by Rigopoulou, Lawrence, & Rowan-Robinson (1996) is
also given by a thin solid line with open squares. The reason for our failure to reproduce the
observed SED around lO^A is essentially that we do not include the effects of small grains
in the present study.

Fig. 2. — Mass distribution of a galaxy merger at T = 1.3 Gyr corresponding to the epoch
of maximum secondary starburst of the merger for total components (upper left), old stellar
components initially located in two disks (upper right), gaseous ones (lower left), and new
stellar ones formed by secondary starburst (lower right). Each of the four frames measures
64.4 kpc on a side.

Fig. 3. — The synthesized image of a galaxy merger at the redshift z=0.4, 1.0, and 1.5
projected onto the x — y plane corresponding to the orbital plane of the merger for the HST
WFPC2 (upper four panels) and the HST NIC2 (lower four panels). Here the morphology
of the merger at the epoch of the maximum starburst (T = 1.3 Gyr) is described. The
redshift of the merger is indicated in the upper right-hand corner of each panel and a bar
with the length of 10 kpc is superimposed in the lower right-hand corner. Except for the
extreme right panel, the exposure time for each synthesized morphology is set to be 10*^

-15-

sec (2000sec x5) both for the HST WFPC2 and the HST NIC2. The exposure time of the
synthesized optical and near-infrared morphology of the merger in the extreme right panel
is 1.24 X 10^ sec (2000sec x62) corresponding to that of the Hubble Deep Field (HDF). The
filter adopted in the present study is F814W for the WFPC2 and F160W for the NIC2. Each
frame measures 9. "9 on a side and one pixel size is 0."1 for the WFPC2 and 0."076 for the
NIC2. The raw image is convolved with a WFPC2 point-spread function (PSF) generated
by Tiny Tim ver-4.4 (Krist & Hook 1997) and the STSDAS FCONVOLVE task to convolve
the PSF is used. A magnitude zero point is 21.60 mag for the WFPC2 (Holtzman et al.
1995) and 25.68 mag for the NIC2 (MacKenty et al. 1997). We assume that a sky surface
brightness at V band is 22.9 mag/arcsec^ corresponding to 0.054 e"/pixel/sec (Biretta 1996).
Readout noise, gain, dark noise are set to be 5.2 (35) e^ /pixel/readout, 7.0 (5.4) e^/DN,
and 0.005 (0.05) e^/sec, respectively, for the HST WFPC2 (the HST NIC2). We adopted
these values, based on the WFPC2 Instrumental hand book ver-4.0 (Biretta 1996) and HST
NICMOS handbook (MacKenty et al. 1997). We used the NKNOISE package in IRAF in
order to add noise to the images. We did not consider other noise sources such as cosmic
rays and image anomaly in the present study.