Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~gcresci/papers/subnaco.ps Äàòà èçìåíåíèÿ: Wed Feb 21 13:56:25 2007 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 10:01:26 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: active galaxy

Astronomy & Astrophysics manuscript no. Gd071 July 5, 2004
(DOI: will be inserted by hand later)
Nuclear star formation in the quasar PG1126-041 from adaptive
optics assisted spectroscopy #
G. Cresci 1 , R. Maiolino 2 , A. Marconi 2 , F. Mannucci 3 , and G. L. Granato 4
1 Dip. di Astronomia ­ Universit‘a di Firenze, Largo E. Fermi 5, I­50125 Firenze, Italy; e­mail: gcresci@arcetri.astro.it
2 INAF, Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, I­50125, Firenze, Italy
3 Istituto di Radioastronomia, sezione di Firenze, Largo E. Fermi 5, I­50125, Firenze, Italy
4 INAF, Osservatorio Astronomico di Padova, Vicolo dell'Osservatorio 5, I­35100 Padova, Italy
Received ; accepted
Abstract. We present adaptive optics assisted spectroscopy of three quasars obtained with NACO at VLT. The high angular
resolution achieved with the adaptive optics (# 0.08 ## ), joined to the diagnostic power of near­IR spectroscopy, allow us to
investigate the properties of the innermost 100 pc of these quasars. In the quasar with the best adaptive optics correction,
PG1126-041, we spatially resolve the Pa# emission within the nuclear 100 pc. The comparison with higher excitation lines
suggests that the narrow Pa# emission is due to nuclear star formation. The inferred intensity of the nuclear star formation
(13 M # yr -1 ) may account for most of the far­IR luminosity observed in this quasar.
Key words. Galaxies: active -- Galaxies: nuclei -- quasars: emission lines -- Galaxies: starburst -- Techniques: high angular
resolution -- Instrumentation: adaptive optics
1. Introduction
The study of the stellar populations and star formation in the
host galaxies of Active Galactic Nuclei (AGNs) is fundamen­
tal to understand the connection between black­hole growth
and galaxy formation. Several evidences for such a connec­
tion have been obtained in local low­luminosity AGNs, i.e.
Seyfert galaxies (e.g. Cid Fernandes et al. 2001, Heckman et
al. 1997, Maiolino et al. 1995, Oliva et al. 1999). At higher,
quasar­like luminosities the investigation of star formation and
stellar population is much more di#cult because the tracers of
star formation are heavily diluted by the strong active nucleus.
Nevertheless a few observational programs have been success­
ful in disentangling the host galaxy and star formation from
the quasar light. The separation of the two components has
been achieved either through a spectroscopic decomposition
or through high angular resolution imaging observations (e.g.
Canalizo & Stockton 2001, Courbin et al. 2002, Dunlop et al.
2003, Jahnke et al. 2003, Schade et al. 2000). By using such
observating strategies, evidence both for active star formation
and relatively quiescent hosts was found. However, all these
studies probed the stellar population on scales of a few kpc,
while the stellar activity within the central 100 pc, where any
interplay with the quasar activity is expected to occur, remains
poorly explored.
Send o#print requests to: G. Cresci
# Based on data obtained at the VLT through the ESO program
71.B­0453(A).
During recent years the performances of adaptive optics
systems have improved significantly, achieving very high angu­
lar resolutions ( <
# 0.1 ## ) even on relatively faint sources. Such
high angular resolution, coupled with spectroscopic capabili­
ties of some of these systems, allows us to e#ciently disen­
tangle components associated with stellar activity even in the
circumnuclear region of luminous quasars. In this letter we
present NACO observations of three quasars at z #0.06 selected
to have an excess of far­Infrared emission. In one of the quasars
we find evidence for star formation within the central 100 pc
with a rate of # 13 M # yr -1 , which may account for most of
its far­IR emission. We assume a ''standard'' cosmology, with
H 0 = 70 km s -1 Mpc -1
,# m = 0.3
and# # = 0.7.
2. Sample selection, observations and data
reduction
The targets were selected from the sample of Andreani et al.
(1999) who provide mid­IR and far­IR data for a large num­
ber of optically selected quasars. We selected quasars match­
ing the following constraints: 1) z#0.06, so that the CONICA
projected resolution on the source is better than 100 pc (in the
K band) and, at the same time, both Pa# and [SiVI]1.97µm
emission lines are shifted into the K band; 2) far­IR luminos­
ity L FIR > 10 44 erg s -1 , i.e. objects more IR luminous than
the average of quasars at the same redshift and, therefore, sus­
pected to harbor starburst activity; 3) radio quiet, to avoid syn­
chrotron contamination to the far­IR radiation; 4) visible mag­

2 G. Cresci et al.: Nuclear star formation in the quasar PG1126-041 from adaptive optics assisted spectroscopy
Name RA(J2000) DEC(J2000) z m V L FIR Date seeing FWHM SR
IRAS09149-6206 09 16 09.4 -62 19 29 0.057 14.1 7.9 Apr 20, 2003 1.0 ## 0.09 ## (100 pc) 5%
PG1126-041 11 29 16.6 -04 24 08 0.060 14.5 3.2 Jun 17, 2003 0.4 ## 0.07 ## (81 pc) 26%
PG2214+139 22 17 12.3 +14 14 21 0.066 14.5 1.5 Jun 18, 2003 0.8 ## 0.09 ## (114 pc) 11%
Table 1. List of the bserved quasars, along with their properties and log of the observations. L FIR is in units of 10 44 erg s -1 ; seeing is the median
DIMM seeing during the observation; FWHM is the Full Width Half Maximum of the PSF on the Adaptive Optics corrected images, in arcsec
and in parsec; SR is the Strehl Ratio measured on the images.
nitude V<14.5 mag, so that the Adaptive Optics correction can
achieve good performances with Strehl ratios of #0.2 by using
the quasar itself as a reference to close the loop. Four quasars
were selected and three of them, whose properties are listed
in Tab. 1, were observed in service mode with NACO. Both
images and spectra were obtained in the K band. Images were
obtained in the IB 2.18 filter (to avoid saturation) with the S27
camera, yielding a pixel scale of 0.027 ## /pix. Spectra were ob­
tained with a 0.086 ## slit at PA=0 # , with the Grism2 coupled
with the SK order sorting filter, yielding a spectral resolution
R = #/## = 1400, and with the S54 camera yielding a pixel
scale along the slit of 0.054 ## /pix. During each observing run
we obtained a di#raction limited nuclear spike of # 0.08 ## (cor­
responding to #100 pc projected on these quasars). In the spec­
troscopic mode the pixel size undersamples the PSF. The Strehl
Ratio ranges from 5% to 26% (see Tab. 1).
The spectra were reduced following the standard steps for
NIR spectroscopy. The spectra obtained at di#erent positions
along the slit were subtracted from each other to remove the
background, then flat fielded, aligned, co­added and calibrated.
Atmospheric features were corrected by dividing for the spec­
trum of a reference star. B3III and B5V stars were used for
PG1126 and IRAS09149 respectively, and in these cases the
spectra were then multiplied by a blackbody at 20000 K to re­
establish the correct slope of the continuum. For PG2214 we
used a G0V star, and the spectra were then multiplied by the so­
lar spectrum to remove the stellar features and continuum slope
(see Maiolino et al. 1996 for details). Images were instead re­
duced using the eclipse software and the CONICA pipeline.
3. Analysis and results
In this letter we focus on the analysis of the spectra, since they
provide the most interesting information. Of the three quasar
spectra we obtain spatially resolved features only for the case
with the best Adaptive Optics correction, i.e. PG1126-041
which was observed with a Strehl Ratio of 26%. The integrated
spectrum is shown in Fig. 1. In the following we will concen­
trate on the analysis of this spectrum.
Pa# is clearly broad and certainly dominated by the Broad
Line Region (BLR). It is more di#cult to indentify the pres­
ence of a narrow component associated either with a Narrow
Line Region (NLR) or with star formation. A method to inves­
tigate the presence of a narrow component, while exploiting
the high angular resolution delivered by the AO, is to study the
spatial variation of the Pa# profile along the slit. The BLR is
certainly unresolved at our resolution (R<1pc) and therefore
the profile of the Pa#, if totally attributed to the BLR, should
2 2.1 2.2 2.3 2.4
0
4
8
Pa a
Br g *
Br g
Br d [SiVI]
Fig. 1. Continuum subtracted spectra of PG1126-04. The Br#* intro­
duced by the division of the reference B star is marked.
not change spatially along the slit, and should simply scale in
intensity following the profile of the PSF. We have investigated
the variation of the Pa# profile along the slit by subtracting the
continuum of all spectra at di#erent locations along the slit and
then rescaled the spectra to match the Pa# flux on the central
spectrum. In Fig. 2 we show this comparison for the spectrum
0.054 ## north of the central spectrum for PG1126: the solid line
shows the (continuum subtracted) central spectrum, while the
dashed line shows the rescaled o#­nuclear spectrum. The lower
panel shows the di#erence between the two spectra which, al­
though noisy, shows the presence of a excess narrow compo­
nent (FWHM # 200 km/s) in the o#­nuclear spectrum. This
demonstrates the presence of a narrow component of Pa#, as­
sociated either with a NLR or with star forming regions.
Once proved the existence of a narrow component, sepa­
rated from the BLR, we can follow its spatial behavior along
with other spectral features. We have fitted the Pa# of PG1126
profile with two broad gaussians, which reproduce the broad
(BLR) component, and a narrow component. The resulting fit
for the central spectrum is shown in Fig. 3. In Tab. 2 we list
the best fit parameters for the components. For what concerns
the blend of [SiVI]1.96µm and Br#, each of the two lines was
fitted by using two gaussian components only (additional com­
ponents are not required given the lower S/N for these fainter
lines) 1 . [SiVI] in particular is well fitted by a broad and a nar­
row component. The presence of both narrow and broad com­
ponents of coronal lines (such as [SiVI]) is well known in
AGNs. A component with a width intermediate between BLR
and NLR (# 1000-2000km s -1 ) is commonly observed and as­
cribed to an ``Intermediate Emission Line Region'' (Giannuzzo
1 Note that the profile of the Pa# cannot be used to fit the Br# be­
cause di#erent hydrogen lines originating from the BLR have often
di#erent profiles, like in this case.

G. Cresci et al.: Nuclear star formation in the quasar PG1126-041 from adaptive optics assisted spectroscopy 3
Line Flux v-v sys FWHM
Pa# ­ Broad 1 953.3 ± 43.3 14 ± 37 3003 ± 134
Pa# ­ Broad 2 457.1 ± 41.7 37 ± 32 1312 ± 142
Pa# ­ Narrow 27.6 ± 7.0 71 ± 32 Unresolved
Br# ­ Broad 129.0 ± 10.4 -75 ± 191 2277 ± 258
Br# ­ Narrow 13.8 ± 6.3 -124 ± 123 697 ± 413
[SiVI] ­ Broad 89.7 ± 6.8 -203 ± 333 2797 ± 240
[SiVI] ­ Narrow 8.0 ± 1.7 42 ± 89 360 ± 185
Table 2. Best fit parameters for the components of the central spec­
trum of PG1126-041, for Pa#, Br# and [SiVI]. Flux is in units of
10 -16 erg s -1 cm -2 ; the shift the various component from the systemic
velocity (v-v sys ) and their FWHM are in km s -1 .
Narrow
a
Pa
Paa
Fig. 2. Comparison between the central spectrum and the spectrum
0.054 ## to the north, around the Pa# region. In the upper panel the solid
gray line shows the (continuum subtracted) central spectrum, while
the black dashed line shows the rescaled o#­nuclear spectrum. The
loer panel shows the di#erence between the two spectra.
et al. 1995). The narrow component of the coronal lines comes
instead from the ``classical'' NLR and extends for a few 100 pc
(Maiolino et al. 2000, Thompson et al. 2001). The broad com­
ponent requires also the profile to be asymmetric, and specif­
ically with a blue wing. Such a blue asymmetry is also com­
monly observed on high excitation lines from the nuclear re­
gion of several Seyferts (e.g. Marconi et al. 1996, Oliva et al.
1994).
The line profiles on the o#­nuclear spectra were fitted by
keeping frozen the parameters of the broad components (ex­
cept for a scaling factor) while the narrow components were left
free to vary. In Fig. 4 we show the intensity of the various lines
as a function of the o#set from the nucleus. The broad com­
ponent of the Pa# gives the profiles of the PSF along the slit
(and it is also consistent with the profile of the continuum, not
shown). The most interesting result, expected from the analysis
discussed above, is that the intensity of the narrow component
of Pa# does not follow the PSF profile and shows an excess of
Fig. 3. Fit of the Pa# profile for the central spectrum of PG112. The
fit is overplotted on the original spectrum, together with the single
gaussian components (see text).
emission to the North, tracing an emission line region extend­
ing for 0.05 ## --0.1 ## , or 60--120 pc, from the nucleus.
We investigate the nature of the narrow Pa# emitting region
by comparing its flux with the observed [SiVI] emission. The
narrow component of the latter traces the NLR as it is not pro­
duced by star forming regions, while the narrow Pa# may either
come from star formation or from the NLR. The di#erent trends
of the narrow Pa# and [SiVI] in Fig. 4 suggest that the two
lines have di#erent origins. We have further investigated this
issue by comparing the ratio between narrow Pa# and [SiVI]
in the narrow line region of Sy2 galaxies observed by other au­
thors (Sy1 cannot be used because the hydrogen lines include
the broad component and generally a decomposition of narrow
and broad components is not given). Pa# is not observable in
low redshift Sy2s, but it can be easily derived from Br# by us­
ing the case B recombination ratio (which holds for the NLR):
Br#/Pa#=0.082. The additional problem is however that in Sy2
galaxies Br# may well be contributed by star formation, since
there are several evidences for nuclear and circumnuclear star
formation, as discussed in Sect. 1. For this reason we decided
to focus on two specific templates: NGC 1068 and Circinus.
These are very nearby Sy2's, extremely well studied, for which
there is no evidence in the central region for significant con­
tribution to Br# by active star formation (star formation occurs
in rings at a radius of # 10 ## - 30 ## ). The Pa#/[SiVI] ratio in­
ferred for the NLR is 2.2 in Circinus (Maiolino et al. 1998) and
1.1 in NGC 1068 (Reunanen et al. 2003). In PG1126 the ra­
tio between narrow Pa#/[SiVI] is 3.4 on the nucleus and 6.4 in
the northern region at 60 pc from the nucleus. This comparison
indicates that there is a strong excess of narrow Pa# emission
with respect to that expected from a NLR. Such a Pa# excess
is probably due to active star formation in the nuclear and cir­
cumnuclear region, within the central #100 pc.
Assuming that all the narrow Pa# is associated with star
formation we can estimate the star formation rate in the central
region of this quasar. Unfortunately, our data do not provide a
bi­dimensional information of the Pa# distribution in the cen­
tral region. However, we can roughly estimate the circumnu­
clear integrated Pa# emission by assuming that the emission is
well represented by the average of the northern and southern

4 G. Cresci et al.: Nuclear star formation in the quasar PG1126-041 from adaptive optics assisted spectroscopy
Fig. 4. Intensity of the various lines as a function of the o#set from
the nucleus, normalized at the flux in the central spectrum. The peak
emission of the narrow Pa# is not aligned with the nucleus, but shows
an excess #60 pc to the north.
sides in our spectrum. The inferred total narrow Pa# flux esti­
mated in this way is F tot (Pa# Narrow ) = 2.15 10 -14 erg s -1 cm -2 .
From case B recombination (Pa#/H# = 0.107), we infer
a total H# luminosity due to star formation of L(H#) SF =
1.7 10 42 erg s -1 . By using the relation between H# luminos­
ity and star formation rate given in Kennicutt et al. (1998)
we derive a nuclear star formation rate of SFR # 13 M #
yr -1 . According to the relation between SFR and far­IR emis­
sion obtained by Kennicutt et al. (1998), such a star forma­
tion rate is expected to produce a far­IR luminosity of #
3 10 44 erg s -1 , which matches the luminosity observed from
this quasar (Tab. 1). Therefore, the nuclear star formation de­
tected in the nuclear few 100 pc may well account for most of
the far­IR luminosity emitted by this quasar and could explain
the excess of far­IR emission in this object (Sect. 2). A similar
result was found at higher redshift by Alexander et al. (2004),
who derived that star formation appears to dominate the bolo­
metric output of AGNs hosted in bright SCUBA galaxies.
Of course, the estimated nuclear star formation rate is de­
rived from a monodimensional spectrum and, in particular, the
inferred nuclear (R<250 pc) Pa# luminosity has required a
large aperture correction (a factor of #3). As consequence, a
confirmation of these findings is certainly required with inte­
gral field spectroscopy. Another source of uncertainty is the
possible star formation activity occurring in the host galaxy at
radii larger than 250 pc; indeed the limited signal­to­noise ratio
of our spectra in the outer regions can only provide a relatively
loose upper limit of 0.18 M # yr -1 kpc -2 on the star formation
rate per unit surface area. Deeper observations are required to
further constrain the star formation in the host galaxy.
In the other two quasars the analysis of the line profile is
more complex due to imperfect subtraction of some deep at­
mospheric absorption features and to a much lower Strehl ratio
than in PG1126. There are some hints of a resolved narrow
component of Pa#, but need to be confirmed with higher qual­
ity spectra and higher Strehl Ratios.
4. Conclusions
We have obtained K­band spectroscopic observations assisted
by adaptive optics of three quasars at z#0.06 which are lumi­
nous in the far­IR. The adaptive optics correction allows us the
reach a di#raction limited angular resolution (# 0.08 ## , cor­
responding to # 100 pc), with Strehl ratios ranging from 5%
to 26%. In the quasar observed with the highest Strehl ratio
(PG1126-041) we spatially resolve a narrow component of
Pa# on scales of #50--100 pc. By comparing the spatial dis­
tribution of the narrow Pa# with the distribution of the coronal
line [SiVI]1.96µm, we infer that most of the narrow Pa# is due
to star formation in the nuclear and circumnuclear region of the
quasar. We also derive that most of the far­IR emission of this
quasar is due to the nuclear star formation detected by us.
These observations clearly demonstrate that adaptive optics
assisted spectroscopy is a powerful tool to investigate the nu­
clear region even in bright quasars.
Acknowledgements. This work was partially supported by the Italian
Ministry of Research (MIUR). We are grateful to the ESO sta# on
Paranal for having performed these observations in service mode.
References
Alexander, D. M., Bauer, F. E., Chapman, S. C., et al. 2004, astro­
ph/0401129
Andreani, P., Franceschini, A., & Granato, G. 1999, MNRAS, 306,
161
Canalizo, G. & Stockton, A. 2001, ApJ, 555, 719
Cid Fernandes, R., Heckman, T., Schmitt, H., Delgado, R. M. G., &
Storchi­Bergmann, T. 2001, ApJ, 558, 81
Courbin, F., et al. 2002, A&A, 394, 863
Dunlop, J. S., et al. McLure, R. J., Kukula, M. J., Baum, S. A., O'Dea,
C. P., & Hughes, D. H. 2003, MNRAS, 340, 1095
Giannuzzo, E., Rieke, G. H., & Rieke, M. J. 1995, ApJ, 446, L5
Heckman, T. M., Gonzalez­Delgado, R., Leitherer, C., et al. 1997,
ApJ, 482, 114
Jahnke, K. & Wisotzki, L. 2003, MNRAS, 346, 304
Kennicutt, R. C. 1998, ARA&A, 36, 189
Maiolino, R., Ruiz, M., Rieke, G. H., & Keller, L. D. 1995, ApJ, 446,
561
Maiolino, R., Rieke, G. H., & Rieke, M. J. 1996, AJ, 111, 537
Maiolino, R., Krabbe, A., Thatte, N., & Genzel, R. 1998, ApJ, 493,
650
Maiolino, R. Alonso­Herrero, A., Anders, S., et al. 2000, ApJ, 531,
219
Marconi, A., van der Werf, P. P., Moorwood, A. F. M., & Oliva, E.
1996, A&A, 315, 335
Oliva, E., Salvati, M., Moorwood, A. F. M., & Marconi, A. 1994,
A&A, 288, 457
Oliva, E., Origlia, L., Maiolino, R., & Moorwood, A. F. M. 1999,
A&A, 350, 9
Reunanen, J., Kotilainen, J. K., & Prieto, M. A. 2003, MNRAS, 343,
192
Schade, D. J., Boyle, B. J., & Letawsky, M. 2000, MNRAS, 315, 498
Thompson, R. I., Chary, R., Corbin, M. R., & Epps, H. 2001, ApJ,
558, L97