Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.naic.edu/~tghosh/a2234/references/methani/arp220_hcn.ps Äàòà èçìåíåíèÿ: Mon Apr 2 21:32:35 2007 Äàòà èíäåêñèðîâàíèÿ: Tue Apr 12 02:59:50 2016 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: arp 220

arXiv:astro­ph/0612122
v1
5
Dec
2006
Astronomy & Astrophysics manuscript no. aaltohnc c
# ESO 2006
December 5, 2006
Overluminous HNC Line Emission in Arp 220, NGC 4418 and
Mrk 231 ­ Global IR Pumping or XDRs?
S. Aalto 1 , M. Spaans 2 , M. C. Wiedner 3 , S. H˜uttemeister 4
1 Onsala Rymdobservatorium, Chalmers Tekniska H˜ogskola, S ­ 439 92 Onsala, Sweden
2 Kapteyn Astronomical Institute, PO Box 800, 9700 AV Groningen, The Netherlands
3 I. Physikalisches Institut, Universitt zu K˜oln, Z˜ulpicher Str. 77, D ­ 50937 K˜oln, Germany
4 Astronomisches Institut der Universit˜at Bochum, Universit˜atsstraúe 150, D ­ 44780 Bochum, Germany
September 29, 2006 / November 29, 2006
ABSTRACT
Context. In recent studies of 3mm J=1--0 HNC emission from galaxies it is found that the emission is often bright which is unexpected
in warm, star forming clouds. We propose that the main cause for the luminous HNC line emission is the extreme radiative and
kinematical environment in starburst and active nuclei.
Aims. To determine the underlying excitational and chemical causes behind the luminous HNC emission in active galaxies and to
establish how HNC emission may serve to identify important properties of the nuclear source.
Methods. We present mm and submm JCMT, IRAM 30m and CSO observations of the J=3--2 line of HNC and its isomer HCN in
three luminous galaxies and J=4--3 HNC observations of one galaxy. The observations are discussed in terms of physical conditions
and excitation as well as in the context of X­ray influenced chemistry.
Results. The ultraluminous mergers Arp 220 and Mrk 231 and the luminous IR galaxy NGC 4418 show the HNC J 3--2 emission
being brighter than the HCN 3--2 emission by factors of 1.5 to 2. We furthermore report the detection of HNC J=4--3 in Mrk 231.
Overluminous HNC emission is unexpected in warm molecular gas in ultraluminous galaxies since I(HNC) >
# I(HCN) is usually taken
as a signature of cold (10 ­ 20 K) dark clouds. Since the molecular gas of the studied galaxies is warm (T k >
# 40 K), we present two
alternative explanations to the overluminous HNC: a) HNC excitation is a#ected by pumping of the rotational levels through the
mid­infrared continuum and b) XDRs (X­ray Dominated Regions) influence the abundances of HNC.
HNC may become pumped at 21.5 µm brightness temperatures of T B >
# 50 K, suggesting that HNC­pumping could be common in
warm, ultraluminous galaxies with compact IR­nuclei. This means that the HNC emission is no longer dominated by collisions and its
luminosity may not be used to deduce information on gas density. On the other hand, all three galaxies are either suspected of having
buried AGN ­ or the presence of AGN is clear (Mrk 231) ­ indicating that X­rays may a#ect the ISM chemistry.
Conclusions. We conclude that both the pumping and XDR alternatives imply molecular cloud ensembles distinctly di#erent from
those of typical starforming regions in the Galaxy, or the ISM of less extreme starburst galaxies. The HNC molecule shows the
potential of becoming an additional important tracer of extreme nuclear environments.
Key words. galaxies: evolution --- galaxies: individual: Arp 220, Mrk 231, NGC 4418 --- galaxies: starburst --- galaxies: active ---
radio lines: ISM --- ISM: molecules
1. Introduction
In order to understand the AGN and star formation activity in the
centres of luminous galaxies it is essential to study the prevailing
conditions of the dense (n(H 2 ) # 10 4 cm -3 ) molecular gas. The
polar molecule HCN (dipole moment 2.98 debye) is commonly
used as a tracer of this dense gas­phase. In particular in distant
luminous (L IR > 10 11 L # , LIRGs) and ultraluminous (L IR >
10 12 L # , ULIRGs) systems the HCN 1--0 line is the prototypical
tracer of dense gas content (e.g. Solomon et al. 1992, Helfer &
Blitz 1993, Curran, Aalto &Booth 2000, Gao & Solomon 2004).
HNC, the isomer of HCN, traces gas of equally high density. In
dense, Galactic, molecular cloud cores it has been suggested to
trace gas temperature: Neutral­neutral chemical models predict
that the HCN/HNC abundance ratio increases with increasing
temperature. This is supported by the fact that the measured HCN
HNC
abundance ratio is especially high in the vicinity of the hot core
of Orion KL (e.g. Schilke et al. 1992, Hirota et al. 1998).
Send o#print requests to: S. Aalto
It is therefore surprising that the HNC/HCN J=1--0 in­
tensity ratios are found to be low in luminous galaxies (e.g.
H˜uttemeister et al. 1995, Aalto et al. 2002 (APHC02)) and that
the HNC/HCN line ratio appears to increase with galactic lumi­
nosity (APHC02).
As an explanation to the abnormally bright HNC emission,
APHC02 suggest that the processes are dominated by fast ion­
neutral chemistry in moderately dense PDR­like regions, instead
of the neutral­neutral chemistry likely governing the hot dense
cores of the Orion cloud. In this case, one would expect the HCN
and HNC abundances and excitation to be very similar to each
other ­ independently of gas temperature. A study of the higher
rotational transitions should show equal brightness for the two
species.
Another possible scenario suggested by APHC02 is that, in­
stead of being collisionally excited, HNC is being radiatively
excited. HNC may be pumped by 21.5 µm continuum radiation
through vibrational transitions in its degenerate bending mode.
This would likely result in significant di#erences in the HCN
and HNC excitation.

2 S. Aalto, M. Spaans, M. C. Wiedner, S. H˜uttemeister: Overluminous HNC in Arp 220, NGC 4418 and Mrk 231
Table 1. Observational parameters
Transition # [GHz] HPBW [ ## ] #mb
JCMT:
HCN 3--2 266 20 0.69
HNC 3--2 271 19 0.69
HNC 4--3 353 14 0.63
CSO:
HCN 3--2 267 30 0.69
CO 2--1 230 34 0.69
IRAM:
HCN 3--2 267 9.5 0.46
HNC 3--2 271 9.5 0.46
Alternatively, Meijerink & Spaans (2005) suggest that the
HNC abundance may become enhanced over that of HCN in
warm, dense XDRs (X­ray Dominated Regions). The deeply
penetrating X­rays induce an ionization structure that di#ers
from the one in a PDR and that favors asymmetries which exist
in the HNC and HCN chemical pathways. This o#ers a chemical
explanation for the abnormal HNC luminosity that is linked to
the accreting black hole.
In order to investigate the underlying cause behind the bright
HNC emission in luminous galaxies, we have searched for HCN
and HNC J=3--2 emission in a sample of LIRG and ULIRG
galaxies with the JCMT and IRAM 30m telescopes. This will
allow us to investigate the excitation of the HNC and HCN
molecules. In this paper, we present the first results of this study
on the HNC 3--2 emission of two ultralumious galaxies, Arp 220
and Mrk 231, and one luminous IR galaxy, NGC 4418. For
these galaxies we find the HNC 3--2 luminosity to be greater
than that of HCN J=3--2 and we discuss two possible scenarios:
mid­infrared (mid­IR) pumping and X­ray dominated chemistry
(XDRs). From now on, we omit the J and refer to a rotational
transition only as: 3--2 instead of: J=3--2.
2. Observations and results
We have used the JCMT telescope to measure the HNC 3--2
(271 GHz) (Figure 1 and Figure 2) lines towards the ultralumi­
nous galaxies Arp 220 and Mrk 231. Observations were made
in April 2005, and the system temperatures were typically 450
K. Pointing was checked regularly on SiO masers and the rms
was found to be 2 ## . Furthermore, the HNC 4--3 line (353 GHz)
was observed in Mrk 231 in February 2006. In addition, we ob­
served the HCN 3--2 (Figure 3) line in Mrk 231 with the CSO
telescope in 1997. With JCMT we also observed the 12 CO 2--1
line of Mrk 231 and used it to compare with the same line ob­
served with the CSO towards the same galaxy to compare the
intensity scales of the two telescopes. The 12 CO 2--1 line inten­
sities from the two telescopes were found to agree within 10%.
The HNC and HCN 3--2 lines of NGC 4418 were observed in
May and July 2006 with the IRAM 30 telescope. Pointing rms
ranged between 1.5 and 2 and system temperatures ranged be­
tween 400 and 600 K. Beam sizes and e#ciencies are shown in
Table 1. Molecular line ratios are presented in Table 2. Note that
line intensities were corrected for beam size when compared to
each other.
Table 2. Line ratios
Galaxy HNC
HCN 3--2 HNC 3-2
1-0 HCN 3-2
1-0
Arp 220 1.9 ± 0.3 a 1.8 ± 0.3 b 0.9 c
Mrk 231 1.5 ± 0.2 0.7 ± 0.2 b 0.3 ± 0.1 d
NGC 4418 2.3 ± 0.3 0.8 ± 0.2 e 0.3 ± 0.1
a) The HNC/HCN 3--2 line ratio is consistent with the one found by
Cernicharo et al (2006) b) HNC 1--0 from APCH02 c) HCN 3--2 from
Wiedner et al. (2004). d) HCN 1--0 data from Curran et al. (2000). e)
HNC 1--0 data from Monje et al. (in preparation). HCN 1--0 from Kohno
et al. (2004). The HNC 4--3/3--2 line ratio for Mrk 231 is 0.5 but the lack
of baseline for the HNC 4--3 data renders this value somewhat uncertain.
Stated errors are 2# rms errors.
Table 3. Gaussian HNC and HCN line fits
Transistion Arp 220 NGC 4418 Mrk 231
# I(HNC 3 - 2) [K km s -1 ] 18.6 ± 0.7 a 10.8 ± 1 b 2.9 ± 0.2 a
V c [km s -1 ] 5330 2120 12095
#V [km s -1 ] 390 150 313
T peak [mK] 44 63 9.0
# I(HNC 4 - 3) [K km s -1 ] -- -- 2.7 ± 0.2 a
V c [km s -1 ] -- -- 12144
#V [km s -1 ] -- -- 278
T peak [mK] -- -- 9.0
# I(HCN 3 - 2) [K km s -1 ] -- 5.2 ± 0.6 b 0.7 ± 0.1 c
V c [km s -1 ] -- 2120 12095
#V [km s -1 ] -- 240 178
T peak [mK] -- 20 4.2
a) JCMT b) IRAM 30m c) CSO
Integrated intensities and line intensities are in main beam brightness
scale. Note that the data are taken with di#erent beam sizes due to the
di#erent telescopes. This has been corrected for in the line ratio table.
2.1. Arp 220
The HNC 3--2 line is about a factor of two brighter than the HCN
3--2 line (see Table 2). For Arp 220, we find that the HNC ex­
citation is superthermal with a 3--2/1--0 ratio of 1.8. Wiedner et
al. (2004) find the HCN 3--2/1--0 ratio to be thermal with ratios
close to 0.9 (see Table 2). The HCN and HNC line widths are
similar although there is a di#erence in line shape.
2.2. Mrk 231
In Mrk 231, the HNC 3--2 line is a factor of 1.5 brighter than the
HCN 3--2 line (Table 2). In contrast to Arp 220 this is mainly due
to the larger line width of the HNC line, while the intensities for
the two lines (corrected for the di#erence in beam size) is similar
(Table 3). The HNC 3--2/1--0 ratio is about 0.7 while the HCN
3--2/1--0 ratio is 0.3 (Table 2). The HNC/HCN intensity ratio is
close to unity in the 1--0 transition (APHC02). The HNC 4--3 line
is clearly detected and the HNC 4--3/3--2 ratio is 0.5, but the lack
of available baseline renders the estimated integrated intensity
somewhat uncertain.

S. Aalto, M. Spaans, M. C. Wiedner, S. H˜uttemeister: Overluminous HNC in Arp 220, NGC 4418 and Mrk 231 3
Arp 220 HNC 3-2
Fig. 1. JCMT HNC 3--2 spectrum of Arp 220. The scale is in
T #
A and should be multiplied with 1/0.69=1.45 for the scale to be
transferred into main beam brightness (see Table 1). The velocity
resolution has been smoothed to 30 km s -1 . The Gaussian fit is
presented in Table 3.
Table 4. Mid­IR and dust properties of Arp 220, NGC 4418 and
Mrk 231
Galaxy L FIR A V T K (gas) T D (dust)
[L # ] [mag] [K] [K]
Arp 220 1.1 â 10 12a 1000 b >
# 40 b >
# 85 c,d
NGC 4418 8 â 10 10e >
# 50 f ­ 85 g
Mrk 231 2.3 â 10 12 ­ > 34 h 128 i
a) Aalto et al. (1991). b) Sakamoto et al. (1999). c) Soifer et al. (1999)
find that the two nuclei of Arp 220 are probably optically thick at 24.5
µm. d) Gonz’ales­Alfonso et al. (2004) have modelled the far­infrared
spectrum of Arp 220 as a two­component model consisting of a nuclear
region of T=106 K, e#ective size of 0. ## 4 and optically thick in the far­
infrared ­ surrounded by an extended region of size 2 ## . e) Roche et al.
(1986). f) Spoon et al. (2001). g) Evans et al. (2003) found that the mid­
IR emitting source is compact and is consistent with a 70 pc source of
brightness temperature (dust temperature) 85 K. h) Bryant & Scoville
(1999). i) Soifer et al. (2000) find a compact source of radius less than
100 pc at 12.5 µm and with brightness temperature T B >
#
141 K. If the IR
emitting source has the same size at 25µm, the corresponding brightness
temperature is 128 K.
2.3. NGC 4418
We find that the HNC 3--2 emission is brighter than the HCN
emission by a factor of 2.3 (Table 2) which makes NGC 4418
the galaxy with the largest HNC/HCN 3--2 ratio measured so
far. The di#erence in integrated intensity is due to the HNC line
having a significantly higher intensity while its line width is nar­
rower than that of HCN (Table 3). Furthermore, the excitation
between HCN and HNC is distinctly di#erent: while HCN is
subthermally excited with 3--2/1--0 line ratios of 0.3, we find that
the corresponding ratio for HNC is closer to unity.
3. The origin of overluminous HNC
3.1. HNC in Galactic molecular clouds
In ion­neutral chemical processes HCNH + will recombine to
produce either HCN or HNC with (roughly) 50% probability. In
Mrk 231
Mrk 231
HNC 3-2
HCN 3-2
Fig. 2. Upper panel: JCMT HNC 3--2 spectrum of Mrk 231. The
scale is in T #
A and should be multiplied with 1/0.7=1.43 for the
scale to be transferred into main beam brightness (see Table 1).
The velocity resolution has been smoothed to 20 km s -1 . Lower
panel: CSO HCN 3--2 spectrum of Mrk 231. The scale is in T #
A
and should be multiplied with 1/0.69 for the scale to be trans­
ferred into main beam brightness (see Table 1). The velocity res­
olution has been smoothed to 20 km s -1 . The velocity scales are
di#erent due to two di#erent observing approaches. The spectra,
however, have been scaled to the same resolution and bandwidth
so that the di#erent line widths are apparent. The total available
bandwidth for the HNC spectrum is not shown here to enable
comparison with the HCN spectrum. The Gaussian fits are pre­
sented in Table 3.
addition, the reaction H 2 NC + + e # HNC + H, is suggested to
produce more HNC (e.g., Hirota et al. 1998). In the Milky Way,
these reactions occur mainly in dark clouds for kinetic tempera­
tures below 24 K. In clouds with higher temperatures, the above
reactions are replaced by neutral­neutral reactions which tend
to selectively destroy HNC resulting in X[HNC]< X[HCN]. In
these neutral­neutral reactions HNC reacts with hydrogen and
oxygen to form HCN, NH and CO. Thus, in Galactic, warm star
forming regions, HNC is underabundant compared to HCN (e.g.
Schilke et al. 1992).

4 S. Aalto, M. Spaans, M. C. Wiedner, S. H˜uttemeister: Overluminous HNC in Arp 220, NGC 4418 and Mrk 231
Mrk 231 HNC 4-3
Fig. 3. JCMT HNC 4--3 spectrum of Mrk 231. The scale is in
T #
A and should be multiplied with 1/0.63=1.6 for the scale to be
transferred into main beam brightness (see Table 1). The velocity
resolution has been smoothed to 20 km s -1 . The Gaussian fit is
presented in Table 3.
3.2. HNC in luminous galaxies
In contrast to the findings for Galactic, warm starforming re­
gions, APHC02 found significant HNC 1--0 luminosities in the
centres of warm starburst galaxies suggesting that high abun­
dances of HNC were produced in warm environments. APHC02
suggested that this is due to ion­neutral reactions persisting de­
spite high temperatures in the gas (40­150 K). The presence
of PDRs (Photon Dominated Regions) could provide an envi­
ronment where X[HNC]=X[HCN] and the luminosities of the
two species would be similar. In this paper, however, we re­
port an overluminosity of HNC compared to HCN by factors
of 1.5 to 2.3. The PDR chemistry alone does not o#er a scenario
where I[HNC] >
# I[HCN] and in the following two subsections we
present two mechanisms that can reproduce the observed over­
luminosity of HNC: mid­IR pumping and XDR chemistry.
3.3. IR pumping of HNC
Both HCN and HNC have degenerate bending modes in the IR.
The molecule absorbs IR­photons to the bending mode (its first
vibrational state) and then it decays back to the ground state
via its P branch (#=1--0, #J=+1) or R­branch (#=1--0, #J=­1)
(see Figure 5). In this way, a vibrational excitation may produce
a change in the rotational state in the ground level and can be
treated (e#ectively) as a collisional excitation in the statistical
equations. Thus, IR pumping excites the molecule to the higher
rotational level by a selection rule #J=2.
For HNC, the bending mode occurs at #=21.5 µm (464.2
cm -1 ) with an energy level h#/k=669 K and an A­coe#cient
of A IR =5.2 s -1 . For HCN the mode occurs at #=14 µm (713.5
cm -1 ), energy level h#/k=1027 K and A IR =1.7 s -1 . It is there­
fore significantly easier to pump HNC, than HCN. The pumping
of HNC may start to become e#ective when the IR background
reaches an optically thick brightness temperature of T B # 50 K
and the gas densities are below critical.
The competition between collisions and radiative excitation:
The IR pumping will compete with collisions for the excitation
of HNC ­ when the IR field becomes more intense, higher
HNC 3-2
HCN 3-2
NGC 4418
NGC 4418
Fig. 4. IRAM HNC (upper panel) and HCN (lower panel) 3--2
spectrum of NGC 4418. The scale is in T #
A and should be multi­
plied with 1/0.46=2.17 for the scale to be transferred into main
beam brightness (see Table 1.). The velocity resolution has been
smoothed to 30 km s -1 . The Gaussian fit is presented in Table 3.
densities are required to successfully compete with the radia­
tive excitation. A simplified analysis includes comparing the
rate of collisional excitation with the rate of mid­IR photon
absorption. A wavelength of 21.5 µm corresponds to a photon
energy of E/k = 669 K and the IR pumping rate is roughly
P IR # A IR /(e 669/T B - 1) where the T B is the mid­IR brightness
temperature (optically thick dust temperature) and A IR is the
Einstein coe#cient for the mid­IR bending transition. For HNC,
the collisional rate is roughly 2 â 10 -10 n(H 2 ) s -1 where n is in
cm -3 . A IR for HNC is 5.2 s -1 , so when T B = 55 K, the pumping
rate is 2.7 â 10 -5 may dominate at gas densities less than 10 5
cm -3 . When T B is 85 K, then the IR pumping rate is almost 100
times faster than at 55 K. Radiative excitation through mid­IR
pumping may then dominate over collisional excitation up to
a density greater than n = 10 6 cm -3 . These rough numbers
illustrate how sensitive the competition between pumping and
collisions is to the background IR brightness.
For HCN, the pump rate is two orders of magnitude slower at
85 K, compared to HNC. This is because the A IR coe#cient is
lower and the energy level for the bending mode is higher (see
previous section).
Pumping in a two­phase molecular medium
In a situation where the pumping completely dominates the exci­
tation of HNC the excitation temperatures of the rotational levels
approach that of the brightness temperature of the background

S. Aalto, M. Spaans, M. C. Wiedner, S. H˜uttemeister: Overluminous HNC in Arp 220, NGC 4418 and Mrk 231 5
E
#
k
0
2
1
13
4
0
#
#=3.3 mm
2 = 1
669 K
1
J=
J=
µ
21.5
R
P
m
Fig. 5. A schematic picture of the pumping of the HNC rota­
tional levels via the mid IR bending transitions. The figure shows
how the rotational J=2 level may become populated through the
#J=2 selection rule through mid­IR pumping. The principle is
the same for higher J levels.
IR field. However, in reality the scenario is unlikely to be this
simple. The molecular ISM will, for example, consist of a range
of densities ­ also in the nucleus of the galaxy.
Consider a simple molecular ISM where the dense (n >
#
10 4 cm -3 )
cores in the galaxy nuclei are surrounded by lower density (n =
10 2 -10 3 cm -3 ), unbound gas (a ``raisin roll scenario ­ e.g. Aalto
et al. 1995, Downes and Solomon 1998, H˜uttemeister and Aalto
2001, Aalto 2005). In this scenario, the di#use gas contributes
a significant fraction of the lower­J 12 CO emission and has a
high volume filling factor. In the case of an intense mid­IR field,
all of the HNC molecules exposed to it will be dominated by
the radiative excitation. In a more dilute field, with T B = 50 K,
for example, the dense cores may be una#ected while the HNC
molecules in the more di#use gas may be experiencing radiative
excitation. Thus, the global HNC luminosity may be significantly
a#ected by pumping, even if the mid­IR field is moderate. In this
scenario one would expect a radial gradient where the e#ect of
the radiative pumping a#ects a larger fraction of the gas closer
to the nucleus of the galaxy.
The presence of such a di#use, lower density (n = 10 2 - 10 3
cm -3 ) molecular medium is suggested to be present in Arp 220:
A large 12 CO/ 13 CO 1--0 line intensity ratio (>30 Aalto et al.
1991) implies a low to moderate optical depth of the 12 CO line.
Together with subthermally excited low­J CO lines, this sug­
gests the presence of a large filling factor of di#use molecular
gas (Aalto et al. 1995). The concept of di#use gas in ultralumi­
nous galaxies ­ among them Arp 220 ­ was extensively modelled
and discussed by Downes and Solomon (1998).
For Mrk 231, a similar di#use gas phase may be suggested by the
large CO/ 13 CO 2--1 line intensity ratio (Glenn & Hunter 2001)
although further observations and radiative transfer modeling is
necessary. For NGC 4418, the presence of di#use molecular gas
is unexplored and multi­wavelength CO and 13 CO modeling is
required to address this issue.
3.4. XDR chemistry
The X­ray irradiation of molecular gas leads to a so­called
X­ray dominated region (e.g., Maloney et al. 1996, Lepp &
Dalgarno 1996) similar to PDRs associated with bright UV
sources (Tielens & Hollenbach, 1985). The more energetic (1­
100 keV) X­ray photons penetrate large columns (10 22 - 10 24
cm -2 ) of gas and lead to a di#erent ion­molecule chemistry.
Species like C, C + and CO co­exist (unlike the stratified PDR)
and the abundances of H + and He + are much larger. Molecules
like H 2 O and OH can be formed more e#ciently because molec­
ular gas resides at higher temperatures, a consequence of the
fact that ionization heating (rather than photo­electric heat­
ing in PDRs) dominates and is more e#cient (about 70%).
Models of XDRs by Meijerink & Spaans (2005) indicate that
the HNC/HCN column density ratio is elevated, and reaches a
value of # 2, for gas densities around 10 5 cm -3 . This while
PDRs and quiescent cloud regions exhibit ratios of unity or less.
Hence, ULIRGS that contain an AGN and possess high gas den­
sities are likely candidates for overluminous HNC emission (see
Meijerink, Spaans & Israel 2007 for details).
In order to compare with observations Figure 6 presents a
grid of one dimensional single­sided PDR/XDR slab models in
density and FUV/X­ray irradiation for HNC/HCN 3--2 line ra­
tios (Meijerink, Spaans and Israel 2007). The model clouds have
a size of 1 pc, but we have verified that our choice of cloud size
does not influence the computed line intensity ratios. We have
assumed that an individual cloud in a galaxy has a velocity line
width of 5 km/s. The impinging X­ray flux follows a -0.9 pow­
erlaw in energy between 1 and 100 keV, typical of a Seyfert nu­
cleus. Recall here that the total flux in the Habing field (between
6 and 13.6 eV) is 1.6 â 10 -3 erg s -1 cm -2 . Hence, compared to a
PDR, we consider enhancement factors of the ambient radiation
field of 10 3 -10 5 , typical of gas exposed to an AGN (Maloney et
al. 1996, Meijerink & Spaans 2005). HCN and HNC are formed
in equal amounts through dissociative recombination of HCNH +
(branching ratio 50­50). In XDRs, the degree of ionization is
much larger than in PDRs. Hence, reactions with, e.g., He + , H + ,
C + and many others, that lead to the formation and destruction of
HCN and HNC in an asymmetric way (with HCNH + and H 2 NC +
as intermediates), are much more important. HNC is particularly
favored over HCN for moderate, 1 - 10 erg s -1 cm -2 , values of
the impinging flux, at an ambient density of n # 10 5 cm -3 , where
the abundance of ions exceeds that in a PDR by about an order
of magnitude (Meijerink, Spaans & Israel 2006, their Figure 1).
Note that the di#erent abundance gradients that occur for HCN
and HNC (Meijerink & Spaans 2005, their Figure 10) at least
provides a necessary condition for the HCN and HNC line inten­
sity profiles to di#er, i.e., to sample di#erent kinematic regions
in the nucleus of an active galaxy.
In PDRs, the ionization degree is modest and set by the am­
bient cosmic ray ionization rate. Consequently, the HNC/HCN
emission line ratios level o# to unity at large, > 10 22 cm -2 ,
depths into the cloud. Note that a strongly elevated cosmic ray
ionization rate is expected in systems that exhibit a large star
formation rate, like the 250 M # yr -1 for Arp 220, because the
supernova rate will scale up proportionally for a Salpeter IMF.
The models of Meijerink et al. (2006) show that even a cosmic
ray rate of 5 â 10 -15 s -1 (roughly 200 M # yr -1 ) does not allow
cosmic ray boosted PDRs to behave as genuine XDRs in terms
of the HNC/HCN ratio. This is basically because the very high
energies of cosmic rays render their radiative interactions (cross
sections scale as 1/E 3 ) with atoms and molecules rather weak.
A top­heavy (non­Salpeter) IMF would lead to a higher ion­
ization rate for a fiducial star formation rate of 200 M # yr -1 .
Still, even for ionization rates as large as 3 â 10 -14 s -1 , which
would accommodate many extreme IMFs, cosmic­ray enhanced
PDRs would still not mimic XDRs in terms of HNC, HCN and
HCO + . In particular, HNC/HCN would still be limited to unity

6 S. Aalto, M. Spaans, M. C. Wiedner, S. H˜uttemeister: Overluminous HNC in Arp 220, NGC 4418 and Mrk 231
Fig. 6. Model line ratio HNC/HCN 3--2: PDRs (left panel) and XDRs (right panel). Note that PDR models cannot reproduce ratios
larger than unity, whereas XDRs can. All models are for a 1 pc cloud.
for the very high cosmic ray ionization rate mentioned above.
The reason is that X­rays couple much better to the gas than
cosmic rays, i.e. their energies are lower, while large columns
of gas can still be penetrated by them (unlike by UV photons),
leading to an active ion­molecule chemistry that favours HNC
and HCO + .
Finally, we note that the maximum HNC/HCN 3--2 line ratio
of 1.6 in our model clouds is smaller than the abundance ratio
of about 2 found by Meijerink & Spaans (2005). This is a con­
sequence of optical depth e#ects in the lines, i.e., one typically
reaches a # = 1 surface before the full HNC/HCN abundance
gradient is sampled. A larger line width than 5 km/s for individ­
ual model clouds would alleviate this.
4. Discussion
With existing data it will be di#cult to distinguish between the
two proposed scenarios of IR­pumping and XDRs. The reason
for this is that the densities required for HNC abundance en­
hancements in an XDR are high, n > 10 5 cm -3 (see previous
section), resulting in a high excitation of the molecule. This is
also the case for the IR­pumping, since the radiative excitation
helps populating higher J­levels.
4.1. A pumping scenario: Observational confirmation
The mid­IR and dust properties of the three observed galaxies
are summarized in Table 4 and the requirements for pumping of
HNC, discussed in 3.3, appear fulfilled in the nuclei of the galax­
ies. Below is a list of useful observational tests for IR pumping:
-- Line ratios and excitation If the HNC/HCN line intensity
ratio exceeds the theoretical abundance limit of 2 from an
XDR then it is necessary to invoke radiative pumping to ex­
plain the observed line intensity ratio. This is for instance
the case for NGC 4418 (see section 4.3.2.) Furthermore, a
significant di#erence in excitation between HCN and HNC
would be a strong indication of radiative excitation of HNC,
for instance if the higher transition HNC lines a