Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~lt/preprints/nics_ngc1068/Da262.ps.gz Äàòà èçìåíåíèÿ: Tue Sep 11 16:56:53 2007 Äàòà èíäåêñèðîâàíèÿ: Sat Dec 22 08:12:06 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: supernova remnant

Astronomy & Astrophysics manuscript no.
(will be inserted by hand later)
NICS--TNG infrared spectroscopy of NGC1068: the first
extragalactic measurement of [P II] and a new tool to constrain
the origin of [Fe II] line emission in galaxies.
E. Oliva 1;2 , A. Marconi 1 , R. Maiolino 1 , L. Testi 1 , F. Mannucci 3 , F. Ghinassi 2 , J. Licandro 2 , L. Origlia 4 , C.
Baffa 1 , A. Checcucci 1 , G. Comoretto 1 , V. Gavryussev 3 , S. Gennari 1 , E. Giani 1 , L.K. Hunt 3 , F. Lisi 1 , D.
Lorenzetti 5 , G. Marcucci 6 , L. Miglietta 1 , M. Sozzi 3 , P. Stefanini 1 , F. Vitali 5
1 Osservatorio di Arcetri, Largo E. Fermi 5, I­50125 Firenze, Italy
2 Centro Galileo Galilei & Telescopio Nazionale Galileo, P.O. Box 565 E­38700 S. Cruz de La Palma, Spain
3 CAISMI­CNR, Largo E. Fermi 5, I­50125 Firenze, Italy
4 Osservatorio Astronomico di Bologna, Via Ranzani 1, I­40127 Bologna, Italy
5 Osservatorio Astronomico di Roma, Via Frascati 33, I­00044 Rome Italy
6 Universit`a degli studi di Firenze, dipartimento di Astronomia, Largo E. Fermi 5, I­50125 Firenze, Italy
Received 26 January 2001 / Accepted 7 February 2001
Abstract. We report 0.9­1.4 ¯m spectroscopic observations of NGC1068 collected during the commissioning phase
of the near infrared camera spectrometer (NICS) of the Telescopio Nazionale Galileo (TNG). These yielded the
first extragalactic measurement of [P II] (1.188 ¯m) line emission. In the central 0.75''x2'' the [Fe II]/[P II] line­
intensity ratio is close to unity, similar to that measured in the Orion bar and a factor of ! ¸ 20 smaller than in
supernova remnants. This indicates that most of iron is locked into grains and, therefore, argues against shock
excitation being the primary origin of [Fe II] line emission in the central regions of NGC1068.
We propose the [Fe II]/[P II] ratio as a simple and effective tool to study and perhaps resolve the long debated
questions related to the origin of [Fe II] line emission and, more generally, to constrain the role of shock excitation
in active galaxies.
Key words. Line: formation; Line: identification; Galaxies: active ; Galaxies: individual: NGC1068 ; Galaxies:
Seyfert ; Infrared: galaxies
1. Introduction
Since the first infrared spectroscopic observations of galax­
ies and supernova remnants in the 80's, the emission
lines of [Fe II] have become a popular and debated issue
(Moorwood & Oliva 1988, Forbes & Ward 1993, Simpson
et al. 1996, Veilleux et al. 1997, Alonso--Herrero et al. 1997,
Mouri et al. 2000).
From the observational point of view, [Fe II] is weak in
HII regions and planetary nebulae while extremely strong
in shock--excited filaments of supernova remnants. Since
relatively bright [Fe II] emission is commonly found in the
IR spectra of normal and active galaxies, many authors
have considered the possibility of using this line as shock
tracer and, even, to count the number of supernova rem­
nants (e.g. Colina 1993, Vanzi & Rieke 1997, Engelbracht
et al. 1998).
Send offprint requests to: E. Oliva, e­mail oliva@tng.iac.es
From the theoretical point of view, a low density region
with normal abundances can become a strong source of
[Fe II] only if the following conditions are satisfied.
i) Most of iron must be in the gas phase, i.e. dust grains
must have been destroyed.
ii) The gas electron temperature must be large enough
to collisionally excite the upper levels of the lines, in
practice T e
? ¸ 5000 K.
iii) Most of iron must be singly ionized, i.e. Fe + /Fe must
be close to unity.
Given the low ionization potential of Fe + and the high ef­
ficiency of the Fe ++ + H o charge--exchange recombination
reactions, the latter two conditions are equivalent to say­
ing that bright [Fe II] lines can only be formed in regions
where hydrogen is partly ionized (e.g. Oliva et al. 1989).
The most efficient mechanisms for creating extended
regions of hot, partially ionized gas are shocks and pho­
toionization by soft X--rays. In both cases the volume emis­
sion measure of the partially ionized region could easily

2 E. Oliva, A. Marconi, R. Maiolino et al.: [PII] in NGC1068 and the role of shocks in AGNs
Fig. 1. NICS--TNG spectrum of the Seyfert galaxy NGC1068 compared with spectra of template objects from the literature
(see text Sect. 2). The break at '1.1 ¯m corresponds to the region of bad atmospheric transmission. The flux scales and zero
levels have been adjusted to facilitate the direct comparison between the spectra.
exceed that of the fully ionized gas. The only important
difference between the two mechanisms is that photoion­
ization is unable to destroy the toughest iron--based grains
which are otherwise easily sputtered by shock fronts.
In practice, therefore, a purely photoionized region
emitting [Fe II] can be easily recognized by measuring the
abundance of Fe + relative to any non­refractory species
which forms in the same partially ionized region. Finding
a very low iron abundance would unequivocally imply that
the gas has not been significantly processed by shocks.
This could be in principle obtained by comparing the
infrared [Fe II] and optical [O I] lines but, in practice, the
large difference in critical densities, reddening and the
problems of comparing data taken with different instru­
ments makes it impossible to obtain a clear--cut conclu­
sion (see e.g. Mouri et al. 1993, Alonso--Herrero et al.
1997, Larkin et al. 1998). A much more reliable deter­
mination of the iron relative abundance should be derived
from lines close in wavelengths and with similar critical
densities such as the [P II] and [Fe II] lines which are dis­
cussed in this Letter.
2. Observations and results
The data were collected at the Telescopio Nazionale
Galileo (TNG) in November 2000 during the commis­
sioning phase of NICS, the near infrared camera and
spectrometer expressly designed and built for this tele­
scope. This instrument is a FOSC--type cryogenic focal
reducer equipped with two interchangeable cameras feed­
ing a Rockwell Hawaii 1024 2 array. The camera used for
the spectroscopic observations has a projected scale of
0.25''/pixel (Oliva & Gennari 1995, Baffa et al. 2000).
The spectroscopic modes are achieved by means of a se­
ries of glass--resin grisms which can be inserted in the 22
mm collimated beam (Vitali et al. 1997). The spectrum of
NGC1068 was collected through a slit of 0.75'' (=3 pix­
els) width and using the IJ grism which yields a 0.89­
1.46 ¯m spectrum with a dispersion of 5.7 š A/pix. The slit
was oriented N--S (i.e. at PA=0 ffi ) and centered on the 1
¯m continuum peak. The acquisition consisted of a series
of four 5--minute exposures with the object set at differ­
ent positions along the slit followed by halogen flats. The
atmospheric spectral response and the instrumental effi­
ciency were determined using spectra of the O6.5V star
HD42088 whose intrinsic spectrum was approximated by
F – =1.7\Delta10 \Gamma9 (–/1.25) \Gamma3:7 erg cm \Gamma2 s \Gamma1 ¯m \Gamma1 .
The spectrum of the central 0.75''x2'' region is dis­
played in Fig. 1 where we also show, for comparison, spec­
tra of the Orion Bar (Walmsley et al. 2000) and unpub­
lished spectra of supernova remnants collected in 1992 us­
ing IRSPEC at the ESO­NTT telescope. The relative line
fluxes are summarized in Table 1. Evident is the differ­
ence between the very large [Fe II]/[P II] ratio measured
in SNR's and the much smaller values found in the Orion
Bar and in NGC1068.

E. Oliva, A. Marconi, R. Maiolino et al.: [PII] in NGC1068 and the role of shocks in AGNs 3
Table 1. Line fluxes in NGC1068 and template objects
Line (¯m) Relative fluxes (1)
NGC1068 Orion RCW103 LMC­N63A LMC­N49
[SIII] 0.9529 1200 29000 -- -- --
[CI] 0.985 72 20 -- -- --
[SVIII] 0.9913 55 !10 -- -- --
Paffi 1.005 61: 1400 -- -- --
HeII 1.012 120 !10 -- -- --
[SII] 1.033 170: 400 -- -- --
[NI] 1.040 60: 14 -- -- --
[FeXIII] 1.075 50: !10 -- -- --
HeI 1.083 1000 6600 31 -- --
Pafl 1.094 110: 2300 !15 -- --
[PII] 1.188 67 50 !8 !6 3
HeI 1.197 !20 48 !8 !6 !6
[SIX] 1.252 + HeI 1.253 72 a 66 b !8 !6 !6
[FeII] 1.257 100 100 100 100 100
HeI 1.279 50: 430 -- !6 !6
Pafi 1.282 230 4300 11 20 15
[FeII] 1.321 33 34 30 -- --
[FeII] intensity (2) ú20 c 4.4 12 3.0 4.4
(1) Normalized to I([FeII] 1.257)=100. Fluxes for Orion refer to position A of Walmsley et al. (2000), values for the supernova
remnants are from Oliva et al. (1990) and from the IRSPEC spectra displayed in Fig. 1. The error on line fluxes in
NGC1068 are tyipically \Sigma10% except for the entries marked with a ``:'' which are uncertain due to blending.
(2) Intensity of [FeII] 1.257 in units of 10 \Gamma4 erg cm \Gamma2 s \Gamma1 sr \Gamma1
a Contribution from HeI 1.253 should be '10, based on the observed intensities of the other HeI lines.
b Contribution from [SIX] is negligible.
c The absolute flux calibration is uncertain.
In principle, the emission feature peaking at 1.188 ¯m
could be contaminated by [Ni II] 1.191 ¯m. However, this
line was measured in the Crab nebula at a level of only
! ¸ 15% of [Fe II] (Rudy et al. 1994) and was not detected in
the supernova remnants listed in Table 1. Moreover, [Ni II]
has a critical density of only '500 cm \Gamma3 (Nusbbaumer
& Storey 1982), two orders of magnitude lower than the
critical density of [Fe II] and much lower than the electron
density pertinent to the region of NGC1068 under consid­
eration here. For these reasons we believe that [Ni II] does
not affect significantly the measurement of [P II].
It is also interesting to note that our data confirm the
identification of [SIX] by Marconi et al. (1996) and in­
clude the first detection of [Fe XIII] in an extragalactic
object. The latter identification is also supported by mea­
surements of the higher ionization green line of [Fe XIV]
(Kraemer & Crenshaw 2000).
3. Discussion
3.1. [Fe II], [P II] and the Fe/P abundance ratio
The near--IR lines of [P II] and [Fe II] have several interest­
ing similarities. They lie nearby in wavelength, have simi­
lar excitation temperatures and critical densities and their
parent ions have similar ionization potentials and radiative
recombination coefficients. Using the collision strengths of
Krueger & Czyzak (1970), Zhang & Pradhan (1995) and
the transition probabilities of Mendoza & Zeippen (1982),
Nussbaumer & Storey (1988) yields
n(Fe + )
n(P + ) ' 2 \Delta
I([Fe II] 1:257¯m)
I([P II] 1:188¯m) (1)
which is accurate within a factor of 2 for all the temper­
atures and densities of interest. The only important dif­
ference between the two species is that charge exchange
recombinations between P ++ and neutral hydrogen are
'2 orders of magnitude less efficient than the Fe ++ + H o
reactions (Kingdon & Ferland 1996). This implies
n(Fe + )
n(P + )
? ¸
n(Fe)
n(P)
(2)
which combined with Eq. (1) yields
n(Fe)
n(P)
! ¸ 2 \Delta
I([Fe II] 1:257¯m)
I([P II] 1:188¯m) (3)
i.e. the Fe/P abundance ratio is quite well constrained by
the [Fe II]/[P II] ratio and, if anything, it is overestimated.
Finally, it is interesting to note that, for a solar Fe/P'100
abundance ratio, one expects [Fe II]/[P II]'50 i.e. a ratio
similar to that measured in supernova remnants.

4 E. Oliva, A. Marconi, R. Maiolino et al.: [PII] in NGC1068 and the role of shocks in AGNs
3.2. Fe/P abundance ratio and dust destruction
Iron is a well known refractory species whose gas phase
abundance in the ISM is often found to be down by many
orders of magnitudes relative its cosmic value. The only
regions where iron is not found to be significantly depleted
are those associated with fast ( ? ¸ 100 km/s) shocks which
can effectively destroy the grains by sputtering. In normal
photoionized regions the depletion of iron ranges between
the factor of '0.1 measured in the Orion Bar (Baldwin
et al. 1996) to significantly lower values found in plane­
tary nebulae (e.g. Oliva et al. 1996, Perinotto et al. 1999).
These differences probably reflect the fact that a variable
(though small) fraction of iron is locked into relatively soft
grains which can be easily destroyed without the need of
fast shocks.
Phosphorus is a non­refractory species whose mea­
sured depletion in ionized gas is close to unity. Therefore,
the Fe/P relative abundance should give a direct estimate
of the iron depletion or, equivalently, of the presence of
robust dust in a given region. This is indeed confirmed by
the data presented here (Fig. 1 and Table 1) which show
variations by more than one order of magnitude between
the large, quasi--solar Fe/P ratio found in SNR's and the
much smaller values found in the other objects.
3.3. [Fe II]/[P II] and the origin of [Fe II] in galaxies
As already discussed in the introduction, determining the
origin of [Fe II] line emission is of crucial importance for
any program aiming at using [Fe II] for tracing shock fronts
and/or constraining the supernovae rate in galaxies. The
results obtained here indicate that the [Fe II]/[P II] ratio
could provide a clear--cut answer to this problem. The line
ratio is large ( ? ¸ 20) in regions excited by fast shocks while
low ( ! ¸ 2) in normal photoionized region and in NGC1068.
This indicate that shocks are not the dominant source of
[Fe II] in the central '200 pc of NGC1068 where, most
likely, the lines are produced by photoionization from the
active nucleus, as already indicated by detailed photoion­
ization modeling (e.g. Kraemer & Crenshaw 2000). The
copious flux of soft X--rays from the AGN creates a very
extended partially ionized region which is responsible for
the strong emission of low ionization species such as [S II],
[O I], [P II] and [Fe II]. However, the latter is relatively
weak because most of the iron is locked into dust grains.
The relative intensities of the low and higher ionization
lines depend on a complex combination of the ionization
parameter, density and of the spectral shape of the ioniz­
ing radiation. Nevertheless, the ratio between ``close rel­
atives'' such as [Fe II]/[P II] are little influenced by this
and, as discussed in Sect. 3.1, almost solely depend on
the Fe/P relative abundance which, in turn, is a direct
measurement of iron depletion (Sect. 3.2).
4. Conclusions
Given the above results and considerations, we propose
using the ratio between [Fe II] (1.257 ¯m) and [P II] (1.188
¯m) as a general tool for constraining the origin of [Fe II]
line emission in galaxies. The diagnostic works as follows
-- In objects with low [Fe II]/[P II] ratios shocks do not
play an important role in the lines excitation.
-- Large values of [Fe II]/[P II] ( ? ¸ 20) indicate that the
emitting gas has recently passed through a fast shock
which sputtered and destroyed most of the dust grains.
It is therefore likely that the lines are also produced
by shock excitation
Acknowledgements. This paper is based on observations made
with the Italian Telescopio Nazionale Galileo (TNG) operated
on the island of La Palma by the Centro Galileo Galilei of the
CNAA (Consorzio Nazionale per l'Astronomia e l'Astrofisica)
at the Spanish Observatorio del Roque de los Muchachos of
the Instituto de Astrofisica de Canarias. We are grateful to all
the technical staff and telescope operators for their assistance
during the commissioning phase of NICS.
References
Alonso--Herrero A., Rieke M., Rieke G., Ruiz M., 1997, ApJ
482, 747
Baffa C., Gennari S., Lisi F. et al., 2000, in The Scientific
dedication of the Telescopio Nazionale Galileo, conference
held in Santa Cruz del La Palma on November 3­5 2000
Baldwin J., Crotts A., Dufour R. et al., 1996, ApJ 468, L115
Colina L., 1993, ApJ 411, 565
Engelbracht C., Rieke M., Rieke G., Kelly D., Acthermann J.,
1998, ApJ 505, 639
Forbes D., Ward M., 1993, ApJ 416, 150
Kraemer S., Crenshaw D., 2000, ApJ 532, 256
Krueger T., Czyzak S., 1970, Proc Roy Soc London A 318, 531
Kingdon J., Ferland G., 1996, ApJS 106, 205
Larkin J., Armus L., Knop R., Soifer B., Matthews K., 1998,
ApJS 114, 59
Marconi A., van der Werf P., Moorwood A., Oliva E., 1996,
A&A 315, 335
Mendoza C., Zeippen C., 1982, MNRAS 199, 1025
Moorwood A., Oliva E., 1988, A&A 203, 278
Mouri H., Kawara K., Taniguchi Y., 1993, ApJ 406, 52
Mouri H., Kawara K., Taniguchi Y., 2000, ApJ 528, 186
Nussbaumer H., Storey P., 1988, A&A 193, 327
Nussbaumer H., Storey P., 1982, A&A 110, 295
Oliva E., Moorwood A., Danziger J., 1989, A&A 214, 307
Oliva E., Moorwood A., Danziger J., 1990, A&A 240, 453
Oliva E., Gennari S., 1995, A&AS 114, 179
Oliva E., Pasquali A., Reconditi M., 1996, A&A 305, L21
Perinotto M., Bencini C., Pasquali A., Manchado A.,
Rodriguez Espinosa J., Stanga R., 1999, A&A 347, 967
Rudy R., Rossano G., Puetter R., 1994, ApJ 426, 646
Simpson C., Forbes D., Baker A., Ward M., 1996, MNRAS 283,
777
Vanzi L., Rieke G., 1997, ApJ 479, 694
Veilleux S., Goodrich R., Hill G., 1997, ApJ 477, 631
Vitali F., Lorenzetti D., Ferruzzi D., Oliva E., 1997, TNG tech­
nical report n. 66
Walmsley M., Natta A., Oliva E., Testi L., 2000, A&A 364,
301
Zhang H., Pradhan A., 1995, A&A 293, 953