arXiv:0911.2673v1 [astro-ph.CO] 13 Nov 2009
Photodissociation chemistry footprints in the Starburst galaxy
Sergio Mart´ ın
Harvard-Smithsonian Center for Astrophysics, 60 Garden St., 02138, Cambridge, MA, USA
J. Mart´ ın-Pintado
Centro de Astrobiolog´ ıa (CSIC-INTA), Ctra de Torrej´ on a Ajalvir, km 4, 28850 Torrej´ on
de Ardoz, Madrid, Spain
Physics and Astronomy Department, University College London, Gower Street, London,
WC1E 6BT, UK
– 2 –
UV radiation from massive stars is thought to be the dominant heating mech-
anism of the nuclear ISM in the late stages of evolution of starburst galaxies, cre-
ating large photodissociation regions (PDRs) and driving a very specific chem-
istry. We report the first detection of PDR molecular tracers, namely HOC+,
and CO+, and confirm the detection of the also PDR tracer HCO towards the
starburst galaxy NGC253, claimed to be mainly dominated by shock heating and
in an earlier stage of evolution than M82, the prototypical extragalactic PDR.
Our CO+detection suffers from significant blending to a group of transitions of
13CH3OH, tentatively detected for the first time in the extragalactic interstellar
medium. These species are efficiently formed in the highly UV irradiated outer
layers of molecular clouds, as observed in the late stage nuclear starburst in M82.
The molecular abundance ratios we derive for these molecules are very similar
to those found in M82. This strongly supports the idea that these molecules
are tracing the PDR component associated with the starburst in the nuclear re-
gion of NGC253. The presence of large abundances of PDR molecules in the
ISM of NGC253, which is dominated by shock chemistry, clearly illustrates the
potential of chemical complexity studies to establish the evolutionary state of
starbursts in galaxies. A comparison with the predictions of chemical models for
PDRs shows that the observed molecular ratios are tracing the outer layers of
UV illuminated clouds up to two magnitudes of visual extinction. We combine
the column densities of PDR tracers reported in this paper with those of easily
photodissociated species, such as HNCO, to derive the fraction of material in
the well shielded core relative to the UV pervaded envelopes. Chemical models,
which include grain formation and photodissociation of HNCO, support the sce-
nario of a photo-dominated chemistry as an explanation to the abundances of the
– 3 –
observed species. From this comparison we conclude that the molecular clouds in
NGC253 are more massive and with larger column densities than those in M82,
as expected from the evolutionary stage of the starbursts in both galaxies.
Subject headings: galaxies: abundances — galaxies: ISM — galaxies: starburst —
galaxies: individual(NGC 253)
– 4 –
Intense UV radiation from massive stars is one of the main mechanisms responsible
for the heating of the interestelar medium in the nuclear region of starburst galaxies. This
mechanism is particularly important in the latest stages of starburst (SB) galaxies where
the newly formed massive star clusters are responsible for creating large photodissociation
regions (PDRs). This is the case for the prototypical SB galaxy M82, where the large
observed abundances of molecular species such as HCO, HOC+, CO+, and H3O+are
claimed to be probes of the high ionization rates in large PDRs formed as a consequence
of its extended evolved nuclear starburst (Garc´ ıa-Burillo et al. 2002; Fuente et al. 2006;
van der Tak et al. 2008).
Observational evidences point to a significant enhancement in the abundance of HOC+
in regions with large ionization fractions. The abundance ratio [HCO+]/[HOC+]= 270 is
found in the prototypical Galactic PDRs of the Orion Bar (Apponi et al. 1999). Similar
or even lower abundance ratios are observed in the PDRs NGC7023 (50-120, Fuente et al.
2003), SgrB2(OH) and NGC2024 (360-900, Ziurys & Apponi 1995; Apponi & Ziurys 1997),
and the Horsehead (75-200 Goicoechea et al. 2009), as well as in diffuse clouds (70-120,
Liszt et al. 2004). This is in contrast with the much larger ratios of ≫ 1000 found in dense
molecular clouds well shielded from the UV radiation. However, these low HCO+/HOC+
ratios are not found in other galactic PDRs. Large values of this ratio of ? 2000 are found
in the PDRs M17-SW, S140, and NGC2023 (Apponi et al. 1999; Savage & Ziurys 2004).
The HCO molecule has also been observed to be a particularly good tracer of the PDR
interfaces. Low ratios of [HCO+]/[HCO]∼ 2.5 − 30 are found in prototypical Galactic
PDRs (Schenewerk et al. 1988; Schilke et al. 2001). The large HCO abundance (> 10−10)
altogether with the low ratio [HCO+]/[HCO]∼ 1 in the Horsehead PDR is claimed to be
a diagnostic for an ongoing FUV-dominated photochemistry (Gerin et al. 2009). CO+is
– 5 –
also claimed to be particularly prominent in the chemical modeling of PDRs and high
abundances of this molecule appear to be correlated to similar enhancements of HOC+
(Sternberg & Dalgarno 1995; Savage & Ziurys 2004). [CO+]/[HOC+] ratios in the range of
1-10 are observed in a number of PDRs (Savage & Ziurys 2004), but only of ? 0.1. towards
the Horsehead PDR (Goicoechea et al. 2009).
As mentioned above, this set of PDR probes has been extensively studied towards
M82. However, no such complete studies have been carried out towards other prototypical
galaxies, but for the detection of HCO and HOC+towards NGC1068 (Usero et al. 2004)
and H3O+in Arp220 (van der Tak et al. 2008). M82 and NGC253 are the brightest
prototypes of nearby SB galaxies, at a similar distance and showing very similar IR
luminosities and star formation rates of about ∼ 3M⊙yr−1(Ott et al. 2005; Minh et al.
2007). However, both galaxies show very different chemical composition. The chemistry and
to a large extend the heating in the central region of NGC253 is believed to be dominated
by large scale low velocity shocks (Mart´ ın et al. 2006b). The similar chemical composition
found in the nuclear region of NGC253 to that in Galactic star forming molecular complexes
points to an earlier evolutionary stage of the starburst in this galaxy than that in M82
(Mart´ ın et al. 2003, 2005, 2006b).
Furthermore, our recent observations of the PDR component as traced by the easily
photodissociated HNCO molecule towards a sample of galaxies (Mart´ ın et al. 2008)
showed the non-detection of HNCO in M82, at a very low abundance limit. This low
HNCO abundance supports the scenario that the PDR chemistry dominates the molecular
composition of the ISM in this galaxy. However, from the HNCO measured abundance in
NGC253, it would be placed in an intermediate stage of evolution where photodissociation
should be starting to play a significant role in driving a UV-dominated chemistry which has
not been yet identified towards this galaxy. The presence of a significant PDR component
– 6 –
in NGC253 claimed from the HNCO abundance is also inferred from the similar intensity
of the atomic fine structure line intensities from PDR tracers like CII and OI (Carral et al.
1994; Lord et al. 1996) observed in both M82 and NGC253.
In this paper we present the first detection of PDR molecular tracers HOC+and CO+,
and confirm the detection HCO (tentatively detected by Sage & Ziurys 1995) in the central
region of NGC253 which allows the evaluation of the influence of the photodissociation
radiation in the nuclear ISM of this SB galaxy. The results presented here support the
scenario of the presence of a significant PDR component and clearly show the potential of
molecular complexity in estimating the contribution of the different heating mechanisms of
the ISM in the nuclei of galaxies.
2. Observations and Results
The observations presented in this paper were carried out at the IRAM 30m and
JCMT telescopes on Pico Veleta, Spain, and Mauna Kea, USA, respectively.
2.1. IRAM 30m
The IRAM 30m observations were performed in symmetrical wobbler switched mode
with a frequency of 0.5Hz and a beam throw of 4′in azimuth. The 516×1MHz filter banks
were used as spectrometers.
We have observed the transitions of C18O J = 1 − 0 (109.782GHz), HCO+J = 1 − 0
(89.188GHz), HOC+J = 1 − 0 (89.487GHz) and 3 − 2 (268.451GHz), and the HCO
10,1−00,0(86.670GHz). Beam sizes at these frequencies were 22′′, 28′′, and 9′′. The nominal
position for the observation was αJ2000= 00h47m33.s3,δJ2000= −25◦17′23′′for HCO+and
– 7 –
HOC+, matching up the position used for the 2mm line survey (Mart´ ın et al. 2006b). The
data on C18O and HCO were centered at αJ2000= 00h47m33.s5,δJ2000= −25◦17′27′′. The
two positions are separated by < 5′′which, considering the 3mm beam sizes, should have a
negligible influence in the relative intensities. Double Gaussian profiles have been fitted to
all observed transitions and the corresponding derived fitting parameters are summarized
in Table 1.
Fig. 1 shows the simultaneously observed J = 1 − 0 features of HCO+and HOC+
compared to the C18O J = 1 − 0 line profile. Although at this frequency the SIS receivers
image band rejection is larger than 20db, we observed the HOC+line tuned to two
different velocities (250 and 500 kms−1) in order to confirm that the observed profile was
not line emission coming from the upper side band. Fig. 1 shows the average of both
observations. The HCO+J = 1 − 0 was detected at the edge of the band covered in the
500 kms−1tuning of HOC+. Although pointing accuracy was of the order of 3′′, the
different shapes observed between the HCO+and HOC+is attributed to a small change in
the pointing position during the two observations. This effect accounts for an uncertainty
in the integrated intensity of < 10%. Even though the HCO+feature was not completely
covered by the backend, the integrated intensity we derive is in agreement within 5% with
that of theobserved by Nguyen et al. (1992) and we observe the line shape to be consistent
with that of C18O. As indicated in Table 1, only a coarse upper limit to the detection of
HOC+J = 2 − 1 was obtained.
The HCO J = 1 − 0 emission was observed in the same window as SiO 2 − 1 and
H13CO+1 − 0 and appears slightly blended to the latter. With a significantly improved
signal-to-noise ratio, we confirm the previous tentative detection of this HCO transition
reported by Sage & Ziurys (1995) with the NRAO 12m telescope. Moreover, using the
main beam brightness temperature from Sage & Ziurys (1995) of ∼ 1mK with at 72′′beam
– 8 –
and our observed ∼ 4mK with a 28′′beam we can make an estimate of the emitting source
extent of > 20′′. The double Gaussian profiles fitted to each species were constrained
to have similar linewidths. The resulting fitted line positions agree within the errors to
those expected from the rest frequencies of each line. Fig. 2 shows the results of the fit
superimposed on the observations as well as the position of the hyperfine structure lines
of HCO. Only the brightest of the group (F = 2 − 1) has been taken into account for the
fit. Assuming optically thin emission, the F = 1 − 0 and F = 1 − 1 transitions (at 86.708
and 86.777GHz) are expected to show an intensity half of the main transition but they are
completely blended to the H13CO+emission. The F = 0 − 1 transition at 86.805GHz is
expected to be even fainter by a factor of 5, well below our detection limit. Fig. 2 shows
in dotted line a synthetic spectrum of HCO assuming one velocity component centered at
255kms−1with a linewidth of 192kms−1(as derived if only one component is fitted to the
spectrum from the other lines) and a peak intensity of the HCO F = 2 − 1 line of 3.6mK.
This shows that the fainter HCO hyperfine transitions may account for up to a 10−20% of
the H13CO+integrated intensity.
JCMT observations were performed in beam switched mode with a frequency of 1Hz
and beam throw of 2′in azimuth. The ACSIS digital autocorrelator spectrometer was used
with a bandwidth of 1600MHz providing a resolution of ∼ 1MHz.
We have used the receiver A3 to observe the CO+transition at 236.062GHz. At
this frequency, the beam size of the telescope is 21′′and the main beam efficiency 0.69.
The observations were carried towards the nominal position αJ2000= 00h47m33.s1,δJ2000=
−25◦17′18′′(radio continuum position, Douglas et al. 1996). As seen in the HC3N
J = 25 − 24 profile, most of the emission is observed from one of the velocity components
– 9 –
at this position, which is due to the JCMT observed position being ∼ 6′′and 10′′away
from those observed with the IRAM 30m, respectively. This position is half beam away
from the positions observed with the IRAM30m, so the abundance ratios derived from this
observation might be affected by a larger uncertainty of up to a factor of 2. However, this
effect might be attenuated by the emission being extended over scales of > 20′′.
The CO+5/2 − 3/2F = 2 − 1 and 3/2 − 1/2F = 2 − 1 transitions are clearly detected
above the noise level (∼ 1.5mk in 30kms−1channels). However, we observe its profile
significantly blended to that of the group of transitions of13CH3OHJ = 5 − 4. This
overlap was not a problem in the case of the CO+detection towards M82 due to the
significantly lower abundance of CH3OH towards this galaxy (Mart´ ın et al. 2006a). The
CO+3/2 − 3/2F = 2 − 1 component was not detected due to its low relative intensity.
We have fitted the observed profile with a single Gaussian component CO+synthetic
spectra with radial velocity and linewidth fixed from those derived from HC3N. The relative
intensities of the CO+components were also fixed to those expected from optically thin
emission under under local thermodynamic equilibrium conditions. The derived line profiles
parameters are presented in Table 1. Additionally, we simultaneously fitted a synthetic
13CH3OH spectra, using the 1813CH3OH transitions in the observed frequency range, using
the same constraints as for the CO+line fit. The fit reproduces most of the observed
features and shows that the emission from13CH3OH may explain the observed non Gaussian
CO+profiles. Only the parameters derived for the three most intense components in the
group are given in Table 1. This is the first time that the13C isotopologue of methanol is
detected towards an extragalactic source. Regarding the accuracy of the fitted parameters
presented in Table 1, the integrated line intensities derived for CO+and13CH3OH are
likely underestimated by ∼ 20% due to the baseline determination. In the next Section, we
discuss the detection of13CH3OH in the context of the derived abundances with respect to
those of the main methanol isotopologue.
– 10 –
3. Molecular abundances and ratios
We have estimated the fractional abundances of the newly observed species in NGC253
assuming optically thin emission, LTE conditions, and similar spatial distribution for all
species. Under these assumptions, we have calculated the column densities of H13CO+,
HOC+, HCO, and CO+for an excitation temperature Tex= 15 ± 5K and an estimated
source extent for each velocity component of 10′′. The Tex= 15 ± 5K is assumed based
on the average rotational temperatures derived from most of the species detected towards
NGC253 (Mart´ ın et al. 2006b). Indeed the non detection of HOC+3 − 2 implies low
excitation temperatures of Tex∼ 10K. Both the excitation temperature and the emission
extent have an important impact in the absolute derived column densities by up to a factor
of 2, however, the fractional abundances and abundance ratios are mostly independent of
these assumptions. We assume that the emission extent is similar for all observed species.
Table 2 presents the column densities and fractional abundance ratios with respect to H2for
all the species. The total H2column density has been derived from the C18O column density
for each velocity component assuming an isotopic ratio of16O/18O = 150 (Harrison et al.
1999) and a CO/H2=10−4. The relative abundances derived for all molecules are presented
in Table 2. Additionally, abundance ratio of H13CO+with respect all species is also
presented in Table 2. The errors in the derived column densities take into account the
statistical error of integrated intensities and the uncertainty in the assumed excitation
temperature. These errors are subsequently propagated to the abundance ratios. The
HCO+abundance has been derived from that of H13CO+to avoid the opacity effects
affecting the main isotopologue. Indeed, if the12C/13C ratio of ∼ 40 derived for NGC253
and NGC4945 (Henkel et al. 1993, 1994; Mart´ ın et al. 2005) applies to all these galaxies,
an average opacity of τHCO+J=1−0? 1 is derived from the HCO+/H13CO+J = 1 − 0 line
ratio (this work and Usero et al. 2004). As a consequence, ratios may be underestimated by
a factor of ∼ 2 if derived from the main isotopologue. The CO+column density presents
– 11 –
a significant uncertainty due to its blending with the newly detected isotopologue of
methanol,13CH3OH. However, the CO+abundance should not be affected by more than a
factor of 2.
We have derived a column density for13CH3OH of ∼ 1.6(1.0) × 1014cm−2. If we
compare this column density with that of the main isotopologue from (Mart´ ın et al.
2006b) it results in a CH3OH/13CH3OH ratio of 12 ± 7, significantly lower than the ratio
12C/13C=40 (Henkel et al. 1993). Both the difference of 6′′in the observed positions and
a different filling factor of CH3OH might account for part of this difference. However,
the integrated intensities measured for the methanol group of transitions at 145.1GHz
by Mart´ ın et al. (2006b) and H¨ uttemeister et al. (1997) at positions differing by > 13′′
show a variation of < 6% so the difference in positions is not likely to contribute to
this difference. Thus, opacity is likely the dominant effect as observed in the Galactic
center (GC, Requena-Torres et al. 2006). From the12C/13C ratio in NGC253 we derive
a fractional abundance of methanol of ∼10−7, close to the abundances observed in the
Galactic center (Requena-Torres et al. 2006). Methanol is, after CO and NO, the most
abundant molecule in the nucleus of NGC253 (Mart´ ın et al. 2006b).
4.Discussion: The PDR component in NGC253
4.1. NGC253 in context
Table 3 shows the HCO+/HOC+, HCO+/HCO, and HCO+/CO+abundance ratios
resulting from our measurements in NGC253 compared to those of the similar SB galaxy
M82, and the Seyfert 2 with nuclear SBs, NGC1068, and NGC4945, together with
prototypical galactic PDRs, where observations of these species have been reported. For
the shake of consistency, the abundance ratios of the other galaxies have been calculated
– 12 –
from the available line profiles obtained from the literature. As already explained before,
for this comparison we used H13CO+to derive the HCO+column densities in order to
avoid the opacity effects. We have assumed the isotopic ratio of12C/13C=40 (Henkel et al.
1993) to derive the column densities of the main isotopologue. For M82, the HCO+/HOC+
and HCO+/CO+abundance ratio have been derived from the observations by Fuente et al.
(2006) towards the Eastern molecular lobe. Our measured ratios are ∼ 59 and ∼ 0.8, which
are > 30% larger that the ratios derived by Fuente et al. (2006). This difference is due
to the significant missing flux of the H13CO+interferometric maps (Garc´ ıa-Burillo et al.
2002) they used for comparison, as well the higher12C/13C ratio they used to compare
with single dish HCO+observations. By comparing the convolved integrated intensity
from the H13CO+interferometric maps by (Garc´ ıa-Burillo et al. 2002) towards the western
lobe Fuente et al. (2006) and the single dish data of by Mauersberger & Henkel (1991)
towards a nearby position, we have estimated a ? 50% missing flux. Thus we used the
H13CO+1 − 0 data by Mauersberger & Henkel (1991) in our measured ratios. The ratios
towards NGC1068 were derived for the regions within the galaxy where lines intensities
were tabulated by Usero et al. (2004). The observations from Wang et al. (2004) were used
to derive the HCO+/HCO ratio towards NGC4945.
4.1.1. PDR abundance ratios in SB galaxies
We find that the three derived HCO+/HOC+, HCO+/HCO, and HCO+/CO+
abundance ratios in the two starbursts, NGC253 and M82, are equivalent within the
measurement errors. The ratios are also in reasonable good agreement to those found in
galactic sources with similar FUV fluxes (see Table 3). Such high abundances ratios of
HOC+, HCO, and CO+relative to HCO+have been claimed to be the evidence of M82
being mostly dominated by photodissociation. We notice that the average HCO+/HCO
– 13 –
ratio in NGC253 are even lower than that measured in M82. In the case of HCO, the
interferometric maps of M82 clearly resolve the spatial variations in this ratio across the
galaxy nuclear region. However, towards the region of peak HCO emission in the M82
maps, we find a ratio of H13CO+/HCO ∼ 0.12 ± 0.04, equivalent to the average observed
towards NGC253. Our data show that the ISM in the nuclear region in NGC253 must be
significantly pervaded by a strong UV radiation flux from the massive star clusters formed
in the starburst, as also suggested by the study of the abundances of the HNCO/CS ratio
(Mart´ ın et al. 2009). Moreover, these new observations would imply that photodissociation
plays a similar role in the ISM heating of both NGC253 and M82.
4.1.2.PDR abundance ratios in AGN galaxies
Both HCO+/HCO and HCO+/HOC+abundance ratios in NGC1068 are different by
a factor 2 − 3 from those of NGC253 and M82. Furthermore, HCO+/HCO is also found
to be up to a factor of ∼ 2 lower in the ring of star formation than towards the nuclear
region. Like in NGC253, these ratios are consistent with the decrease in the abundance of
molecules such as HNCO from the nuclear region to the starburst ring (Mart´ ın et al. 2009).
The tentative detection of HCO in the circunnuclear disk (CND) of NGC1068 (Usero et al.
2004), suggest a rough H13CO+to HCO line intensity ratio of ∼ 2 − 3, which turns into an
abundance ratio in the range of ∼ 0.09−0.13, closer to the values derived in SB dominated
galaxies. This value at the CND can be significantly biased by the emission from the star
forming ring covered at half-power by the 28′′beam. Therefore, it is clear that this ratio
does not significantly decrease towards the nuclear AGN in this galaxy with the typical
angular resolution of 20′′−30′′. On the other hand, the ratio HCO+/HOC+is only a factor
of 2 lower in the starbursts galaxies than towards the nuclear AGN in NGC1068. From
these observations, it is unclear whether photodissociation does play a major role in the
– 14 –
AGN dominated center of NGC1068. Unfortunately, no observations of HOC+are available
towards the SF ring in this galaxy.
Similar to NGC1068, the obscured Seyfert 2 nucleus in NGC4945 is surrounded by a
starburst ring more prominent than in NGC1068 (Genzel et al. 1998). The HCO+/HCO
ratio found in NGC4945 is even lower than that found in the other galaxies. However, the
fit to these lines was claimed to be very uncertain by Wang et al. (2004).
4.2. Comparison to PDR chemical models
We have compared our observed fractional abundances and abundance ratios with
those predicted by the UCL PDR model (Bell et al. 2006). The UCL PDR code is a time
and depth dependent one dimensional PDR model that simultaneously solve the chemistry,
thermal balance and radiative transfer within a cloud (see Bell et al. 2006, for more details).
We have adopted a hydrogen density of 105cm−3, a radiation field G0∼ 5000 in units
of Habing field, and a cosmic radiation rate of 10−16. The high density of 105cm−3is
derived from the multiline analysis of CS and HC3N (Bayet et al. 2008, 2009, ; Aladro et
al. In Prep). Our estimates of the cloud structure (see Sect. 4.3) depends on the averaged
radiation field which might be different in both galaxies. The averaged radiation fields in
both NGC253 and M82 have been inferred from the fine structure lines and they are of
2×104and 103, respectively, with large errors of a factor of 2 (Carral et al. 1994; Lord et al.
1996). This would imply that the PDR envelope should be larger in the clouds of NGC253
than in M82. Since we do not aim to quantitatively model the particular abundances
measured in NGC253 but to investigate the physical conditions that would give rise to the
wealth of observed molecules in starburst, the value of G0= 5 × 103used is a geometric
mean value derived from the fine structure lines in both galaxies. Given that our estimates
of the Avfor the PDR are based on this geometric mean, the expected changes in Avfor
– 15 –
the two galaxies would be just a factor of 1.6.
We have ran two different models: Model A is a standard time dependent gas-phase
PDR model where the initial composition is atomic; while Model B is computed using a
coupled dense core-PDR model where the diffuse material, initially also purely atomic and
gaseous, collapses to reach a final density of 105cm−3. During the collapse the gas depletes
on the grains forming icy mantles which remain on the dust until irradiation from a UV
field is switched on, evaporation occurs and the typical PDR chemistry takes place. In
both models the temperature is calculated self-consistently at each depth and time step
by thermal balance. Fig. 4 shows the predicted abundances and abundance ratios as a
function of visual extinction (Av) for Model A (left panels) and B (right panels). HNCO
and CH3OH results are only shown for Model B in Fig. 4.
Observed abundances of HCO+and HOC+towards NGC253 (shown as horizontal
lines in the key of Fig. 4) are well reproduced by the models for very low extinction
of Av ∼ 1 − 2. The HCO abundance observed is a factor 2 − 6 above the maximum
predicted by the model B. It is important to take into account that while we assumed a
ratio CO/H2=10−4to calculate the fractional abundances, this ratio is not constant in the
models and hardly ever reaches this value. On the other hand, the abundance ratios of
the observed molecules, unaffected by the hydrogen determination uncertainty, agree well
with the model predictions. The model shows that the abundances of CO+, follows the
same pattern as HOC+. The correlation between these two molecules was also predicted
by previous theoretical studies (Sternberg & Dalgarno 1995; Savage & Ziurys 2004). Thus,
CO+measurement allows us to confirm the effect of photodissociation suggested by the
large abundance of HOC+.
The ions HOC+, CO+and HOC+are mostly formed at the edge of the cloud and
while the two models predict similar abundances for these ions, it is worth noting that
– 16 –
the observed abundances of other species such as HNCO, CH3OH, or HOCO+can only be
explained with Model B. However, HNCO and CH3OH only reach observable abundances
for Av∼ 5 magnitudes. This implies that, while photodissociation does play an important
role in the chemistry of NGC253, the molecular clouds affected by the UV radiation must
contain a dense core well shielded from the UV radiation and rich with gas phase icy mantle
molecules like HNCO and CH3OH.
4.3. The molecular clouds in SB galaxies
Mart´ ın et al. (2008) have used a comparison with the GC molecular clouds to propose
another PDR diagnostic based on the relative abundance of HNCO to CS. The large
variation of the HNCO/CS abundance ratio between UV radiated clouds to those well
shielded clouds only affected by shocks was interpreted as the fast photodissociation of the
fragile molecule HNCO, efficiently produced on the icy mantles and delivered into gas phase
by low velocity shocks. Similarly, in a sample of nearby galaxies Mart´ ın et al. (2009) found
changes of nearly two orders of magnitude from the shock dominated chemistry in M83 and
IC342 to UV dominated chemistry in M82. The extremely low HNCO abundance in M82
and the large abundances of HCO, HOC+and CO+support the idea that the HNCO/CS
ratio is a measure of the relative importance of the UV heating to shock heating and the
evolutionary state of the starburst in galaxies.
The detection of HCO, HOC+and CO+in NGC253 with similar column densities and
abundance to those in M82 suggest that the PDR component is similar in both galaxies
as suggested by the similar atomic fine structure line intensities in both galaxies. The
results of model B confirm the observational trends observed in HNCO and the other
PDR molecules in galaxies. For molecular clouds with Av∼ 1 − 2 (i.e column densities
of 1 − 2 × 1021cm−2) illuminated by a strong UV radiation field like in the galaxies in
– 17 –
our sample, HNCO and CH3OH are largely photodissociated and only HCO, HOC+and
CO+should be observed like in M82 (Mart´ ın et al. 2006a, 2009). Then, not very massive
molecular clouds and widely translucent to the UV radiation should dominate in M82.
On the other hand, for galaxies with massive molecular clouds (large visual extinction) or
low UV radiation fields, HNCO and CH3OH are well shielded and the abundance ratio
of HNCO/CS will reach its maximum value. Considering that M83 and IC342 represent
the stage of galaxies with an extremely low PDR component, the lower HNCO/CS ratios
measured for NGC253, NGC4945 and NGC1068 indicate that the PDR component must
be substantial, as observed in other PDR tracers.
For NGC253 we find that the PDR component is similar to that in M82, but the total
column density of dense gas is a factor of 2-3 larger in NGC253 than in M82 from the low
HNCO and CH3OH abundance in the latter. Considering the PDR column densities in both
galaxies are similar, this component should represent about 1/3−1/2 of the total molecular
column density in NGC253. This is roughly consistent with the decrease by a factor of
2 − 3 of the HNCO/CS ratio as compared with that of M83 or IC342 (Mart´ ın et al. 2009).
This suggests that the molecular clouds properties in M82 and NGC253 must be quite
different in terms of the total molecular column density, which implies that the sizes or
the densities are different, or a combination of both. Using the atomic fine structure and
CO emission lines, Carral et al. (1994) and Lord et al. (1996) have also proposed that the
clouds in M82 and NGC253 are quite different. The clouds in NGC253 are slightly smaller
than those in M82, but with masses a factor of 15 larger than for M82. The NGC253
average cloud column densities are therefore factor of 20 larger than in M82. Though the
total column densities of the M82 clouds inferred from the atomic fine structure lines are a
factor of 5 larger than those predicted from the PDRs tracers and the HNCO abundance
in this galaxy, similar constrains are derived both the molecular and the atomic tracers.
Therefore, the clouds in NGC253 are more massive than in M82.
– 18 –
We can even make a very rough estimate of the average properties and structure of
the molecular clouds in NGC253 and M82 by combining the complementary information
obtained from the PDR tracers presented in this paper and the HNCO column densities
from (Mart´ ın et al. 2009). While HCO, HOC+and CO+mainly trace the PDR region
up to Av= 4 − 5 (i.e. H2column densities of 5 × 1021cm−2), the HNCO emission only
arises from the well shielded core (Av> 7) of the molecular clouds. The HCO, HOC+and
CO+column densities in NGC253 and M82 indicate similar averaged column densities
in the PDR envelopes of the molecular clouds in both galaxies. The big difference in the
molecular cloud structure in both galaxies is in the size (H2column density) of the well
shielded cores of the molecular clouds. In the case of M82, where HNCO has not been
detected, we can set an upper limit to the HNCO column density of 7 × 1012cm−2. This
translates to a upper limit to the core H2column density of < 3 × 1020cm−2for the HNCO
fractional abundance of 2 × 10−8derived for the well shielded clouds in the galactic center
(Mart´ ın et al. 2008). The averaged shielded cloud cores in M82 are smaller by more than
one order of magnitude than the PDR envelope. In the case of NGC253, the H2column
densities of the shielded cloud cores is 1022cm−2, a factor of 2 larger than PDR envelope.
Assuming a similar averaged density distribution in the molecular clouds in both galaxies,
the clouds in NGC253 would be a factor 2 − 3 larger than in M82.
4.4.The contribution X-ray induced chemistry
The PDR model presented by Fuente et al. (2006) failed to reproduce the large CO+
column density of a few 1013cm−3observed towards M82. This lead Spaans & Meijerink
(2007) to explore the possibility of an enhanced X-ray induced chemistry in this galaxy.
Spaans & Meijerink (2007) concluded that such high formation of CO+can only be
explained by X-ray irradiated molecular gas with densities of 103− 105cm−3. Although the
– 19 –
X-ray luminosity of NGC253 is a factor of 2 − 4 below that of M82, both galaxies have
a significant X-ray emission in the range of ∼ 1040ergs−1(Cappi et al. 1999). Similar to
M82, the NGC253 total CO+column density is (3.6 ± 1.1) × 1013cm−2. Moreover, both
show a similar HCO+/CO+ratio of ∼ 30 − 40.
The models presented in this paper are able to produce such column densities for visual
extinctions of Av∼ 3−5 for model A, and Av∼ 0.5−1 for model B. These models have been
calculated with a radiation field, a cosmic ray flux and a H2density smaller by a factor of 2,
40 and 4, respectively, with respect to those assumed in the models of Fuente et al. (2006).
Furthermore, we are able to reproduce the abundances and abundance ratios measured for
all the other observed species presented in this paper. van der Tak et al. (2008) showed
how the measured abundance of H3O+in M82 can be both produced by PDR with a high
cosmic-ray ionization or by an XDR. Indeed, an increase of cosmic ray ionization rate in
PDR models may be qualitatively used to simulate XDR-like environments. Our models,
however, do not use particularly high cosmic ray fluxes (a factor of 5 higher than standard).
Thus, though the X-ray irradiation is substantial in SB galaxies, the PDR models presented
in this paper can reproduce the molecular abundances observed towards the brightest
prototypes, M82 and NGC253. Moreover, no significant changes in the abundances of
HOC+and HCO are found towards the nuclear AGN of NGC1068 where X-ray radiation is
significantly more important than in SB nuclei, as shown in Sect. 4.1.2. Unfortunately, no
CO+observation has been reported towards this Seyfert 2 nucleus.
5. Conclusions: The pervading UV field in evolved starburst
The comparison of model predictions with the observations presented show that
the abundance of the species observed in this work towards NGC253, namely HCO+,
CO+, and HCO, are most efficiently formed in the outer region of the molecular clouds
– 20 –
where the gas is highly irradiated by the incident UV photons from massive stars. The
high molecular abundances derived for these species in NGC253 suggest that the PDR
component in this galaxy is similar to that found in M82, claimed to be the prototype
of extragalactic PDR. The abundance ratios found for this limited sample of galaxies are
of the same order as those observed towards galactic PDRs, which stress the importance
of photo-dominated chemistry in galaxy nuclei. Large amounts of molecular material are
affected by photodissociation not only in NGC253, but also towards the star forming
regions around the Seyfert 2 nuclei in NGC4945 and NGC1068. This is consistent with
the HNCO/CS ratio in these galaxies which suggest that a fraction of HNCO has been
photodissociated in PDRs. The combination of the observations of HCO, HOC+and CO+
with that of HNCO seems to confirm that their abundances reflect the evolutionary stage of
the starbursts in these galaxies. Although photodissociation is the most likely scenario for
the enhancement of the observed reactive ion in starburst environments, X-ray dominated
chemistry has been claimed to be responsible for the high abundances observed around
AGNs in circunnuclear disk of NGC1068 (HOC+Usero et al. 2004) and towards the ultra
luminous infrared galaxy Arp220 (H3O+van der Tak et al. 2008).
Therefore, M82 is still outstanding not only as a PDR dominated galaxy,
but by the underabundance of complex molecules such as CH3OH, HNCO or SiO
(Mauersberger & Henkel 1993; Mart´ ın et al. 2006a,b), evidence for the lack of large
amounts of dense molecular material which would potentially fuel its nuclear starburst
as compared to other starburst galaxies like NGC253 (Mart´ ın et al. 2009). Our data in
combination with the HNCO abundances (Mart´ ın et al. 2009) indicate that the molecular
clouds in M82 are different from those in NGC253. Although having a similar overall
PDR component, the clouds in NGC253 have to be more massive and have larger column
densities those in M82.
– 21 –
This work has been partially supported by the Spanish Ministerio de Ciencia e
Innovaci´ on under project ESP2007-65812-C02-01, and by the “Comunidad de Madrid”
Government under PRICIT project S-0505/ESP-0237 (ASTROCAM).
Facilities: IRAM 30m,JCMT.
– 22 –
Apponi, A. J., & Ziurys, L. M. 1997, ApJ, 481, 800
Apponi, A. J., Pesch, T. C., & Ziurys, L. M. 1999, ApJ, 519, L89
Bayet, E., Lintott, C., Viti, S., Mart´ ın-Pintado, J., Mart´ ın, S., Williams, D. A., & Rawlings,
J. M. C. 2008, ApJ, 685, L35
Bayet, E., Aladro, R., Mart´ ın, S., Viti, S., Mart´ ın-Pintado, J. 2009, Submitted to ApJ
Bell, T. A., Roueff, E., Viti, S., & Williams, D. A. 2006, MNRAS, 371, 1865
Cappi, M., et al. 1999, A&A, 350, 777
Carral, P., Hollenbach, D. J., Lord, S. D., Colgan, S. W. J., Haas, M. R., Rubin, R. H., &
Erickson, E. F. 1994, ApJ, 423, 223
Douglas, J. N., Bash, F. N., Bozyan, F. A., Torrence, G. W., & Wolfe, C. 1996, AJ, 111,
Fuente, A., Rodr´ ıguez-Franco, A., Garc´ ıa-Burillo, S., Martın-Pintado, J., & Black, J. H.
2003, A&A, 406, 899
Fuente, A., Garc´ ıa-Burillo, S., Gerin, M., Rizzo, J. R., Usero, A., Teyssier, D., Roueff, E.,
& Le Bourlot, J. 2006, ApJ, 641, L105
Garc´ ıa-Burillo, S., Mart´ ın-Pintado, J., Fuente, A., & Neri, R. 2000, A&A, 355, 499
Garc´ ıa-Burillo, S., Mart´ ın-Pintado, J., Fuente, A., Usero, A., & Neri, R. 2002, ApJ, 575,
Genzel, R., et al. 1998, ApJ, 498, 579
Gerin, M., Goicoechea, J. R., Pety, J., & Hily-Blant, P. 2009, A&A, 494, 977
– 23 –
Goicoechea, J. R., Pety, J., Gerin, M., Hily-Blant, P., & Le Bourlot, J. 2009, A&A, 498, 771
Harrison, A., Henkel, C., & Russell, A. 1999, MNRAS, 303, 157
Henkel, C., Mauersberger, R., Wiklind, T., Huettemeister, S., Lemme, C., & Millar, T. J.
1993, A&A, 268, L17
Henkel, C., Whiteoak, J. B., & Mauersberger, R. 1994, A&A, 284, 17
H¨ uttemeister, S., Mauersberger, R., & Henkel, C. 1997, A&A, 326, 59
Liszt, H., Lucas, R., & Black, J. H. 2004, A&A, 428, 117
Lord, S. D., Hollenbach, D. J., Haas, M. R., Rubin, R. H., Colgan, S. W. J., & Erickson,
E. F. 1996, ApJ, 465, 703
Mart´ ın, S., Mauersberger, R., Mart´ ın-Pintado, J., Garc´ ıa-Burillo, S., & Henkel, C. 2003,
A&A, 411, L465
Mart´ ın, S., Mart´ ın-Pintado, J., Mauersberger, R., Henkel, C., & Garc´ ıa-Burillo, S. 2005,
ApJ, 620, 210
Mart´ ın, S., Mart´ ın-Pintado, J., & Mauersberger, R. 2006a, A&A, 450, L13
Mart´ ın, S., Mauersberger, R., Mart´ ın-Pintado, J., Henkel, C., & Garc´ ıa-Burillo, S. 2006b,
ApJS, 164, 450
Mart´ ın, S., Requena-Torres, M. A., Mart´ ın-Pintado, J., & Mauersberger, R. 2008, ApJ,
Mart´ ın, S., Mart´ ın-Pintado, J., & Mauersberger, R. 2009, ApJ, 694, 610
Mauersberger, R., & Henkel, C. 1991, A&A, 245, 457
Mauersberger, R., & Henkel, C. 1993, Reviews in Modern Astronomy, 6, 69
– 24 –
Minh, Y. C., Muller, S., Liu, S.-Y., & Yoon, T. S. 2007, ApJ, 661, L135
Nguyen, Q.-R., Jackson, J. M., Henkel, C., Truong, B., & Mauersberger, R. 1992, ApJ, 399,
Ott, J., Weiss, A., Henkel, C., & Walter, F. 2005, ApJ, 629, 767
Requena-Torres, M. A., Mart´ ın-Pintado, J., Rodr´ ıguez-Franco, A., Mart´ ın, S., Rodr´ ıguez-
Fern´ andez, N. J., & de Vicente, P. 2006, A&A, 455, 971
Savage, C., & Ziurys, L. M. 2004, ApJ, 616, 966
Sage, L. J., & Ziurys, L. M. 1995, ApJ, 447, 625
Schenewerk, M. S., Jewell, P. R., Snyder, L. E., Hollis, J. M., & Ziurys, L. M. 1988, ApJ,
Schilke, P., Pineau des Forˆ ets, G., Walmsley, C. M., & Mart´ ın-Pintado, J. 2001, A&A, 372,
Spaans, M., & Meijerink, R. 2007, ApJ, 664, L23
Sternberg, A., & Dalgarno, A. 1995, ApJS, 99, 565
St¨ orzer, H., Stutzki, J., & Sternberg, A. 1995, A&A, 296, L9
Usero, A., Garc´ ıa-Burillo, S., Fuente, A., Mart´ ın-Pintado, J., & Rodr´ ıguez-Fern´ andez, N. J.
2004, A&A, 419, 897
van der Tak, F. F. S., Aalto, S., & Meijerink, R. 2008, A&A, 477, L5
Wang, M., Henkel, C., Chin, Y.-N., Whiteoak, J. B., Hunt Cunningham, M., Mauersberger,
R., & Muders, D. 2004, A&A, 422, 883
Ziurys, L. M., & Apponi, A. J. 1995, ApJ, 455, L73
– 25 –
This manuscript was prepared with the AAS LATEX macros v5.2.
– 26 –
Fig. 1.— IRAM 30m observations of the emission of HCO+, C18O and HOC+J = 1 − 0
transitions with the fitted line profiles superimposed. C18O emission has been multiplied by
a factor of 4 for comparison with the HCO+profile. The spectral resolution are the original
1MHz (∼ 3 kms−1) and smoothed to 4MHz (∼ 13 kms−1), respectively.
– 27 –
Fig. 2.— HCO profile in the same window as SiO and H13CO+observed with the IRAM
30m.. On top it is shown the triple Gaussian profile fitted to the transitions. Dashed lines
indicate the position of each hyperfine structure transition of HCO, where only the brighter
one has been fitted. The theoretical synthetic spectrum of HCO is shown with a dotted line
(see Sect. 2 for details). Velocity resolution has been degraded to ∼31kms−1.
– 28 –
Fig. 3.— JCMT observations of the CO+emission blended with transitions of13CH3OHJ =
5 − 4 and observed in the same window as HC3NJ = 25 − 24. This is the first detection
of the13C isotopologue of methanol. The overall spectral fitting to all the lines is shown
in grey. The contribution of the CO+emission is shown with dashed line. The position of
the three brighter transitions of the J = 5 − 4 group of13CH3OH are shown indicated with
vertical dashed lines. See Sect. 2 for details on the fitting to the spectra. Velocity resolution
has been degraded to ∼30kms−1.
– 29 –
Fig. 4.— Theoretical predictions for the fractional abundances relative to H2(Upper panels)
and abundance ratios (Lower Panels) for the observed species as derived from the two dif-
ferent PDR models: a pure gas-phase model A (Left panels) and a coupled dense core-PDR
model B (Right panels). Details are given in Section 4.2. The vertical position of the key for
each molecule and ratio shown only in the plots for model A correspond to the actual derived
parameters from the observations. Additionally, the fractional abundances of CH3OH and
HNCO are shown for model B.
– 30 –
Table 1: Parameters derived from the observed line profiles.
C18O1 − 03.1 ± 0.6183 ± 13100 ± 9116.75
6.0 ± 0.8283 ± 7100 ± 9a
HCO+1 − 021.98 ± 1.1177.9 ± 0.4 118.1 ± 0.3174.9
35.79 ± 1.4289.0 ± 0.3118.1 ± 0.3a
HOC+1 − 00.8 ± 0.2170 ± 10100 ± 207.6
1.0 ± 0.2282 ± 9100 ± 209.2
HOC+3 − 2< 0.8b
0.41 ± 0.08183 ± 14102 ± 4a
0.43 ± 0.08297 ± 13102 ± 4a
H13CO+1 − 01.26 ± 0.09176 ± 6 102 ± 4a
1.36 ± 0.10285 ± 5102 ± 4a
SiO 2 − 11.52 ± 0.11 182 ± 5 102 ± 4a
1.59 ± 0.11 292 ± 5 102 ± 4a
HC3N J = 5 − 41.05 ± 0.15 191 ± 795 ± 1810.5
CO+5/2 − 3/2F = 2 − 10.30 ± 0.10 191c
CO+3/2 − 1/2F = 2 − 10.17 ± 0.10 191c
13CH3OH 50,5− 40.4
0.10 ± 0.07191c
13CH3OH 5−1,5− 4−1.4
0.16 ± 0.07191c
aLinewidths forced to have the same value in the Gaussian fit.
b3σ upper limit assuming a 200 kms−1linewidth.
cParameters forced to equal those derived from HC3N Gaussian fit.
– 31 –
Table 2: Derived fractional abundances and H13CO+ratios for each velocity component
1.6 ± 0.54.8 ± 1.01
1.7 ± 0.62.7 ± 0.41
0.8 ± 0.32.4 ± 0.72.0 ± 0.8
1.1 ± 0.41.7 ± 0.41.6 ± 0.4
HCO12.3 ± 5.837 ± 100.13 ± 0.04
12.9 ± 6.1 20 ± 40.14 ± 0.03
1.7 ± 0.85 ± 20.9 ± 0.4
aWith N(H2) = 3.3±1.2×1022cm−2and 6.3±2.1×1022cm−2for each velocity component, respectively, as
derived from C18O with16O/18O = 150 (Harrison et al. 1999) and a CO/H2= 10−4.
– 32 –
Table 3: Abundance ratios of HCO+vs HOC+and HCO.
NGC25380 ± 305.2 ± 1.838 ± 15
63 ± 175.4 ± 1.3...
M8260 ± 28a
9.6 ± 2.8b
32 ± 16a
NGC1068 128 ± 28c
3.2 ± 1.2d
NGC4945...2.4 ± 1.2e
GC prototypical PDRs
Horsehead75 − 200f
Orion Bar< 166 − 270h,i
< 83 − 140i,k
NGC702350 − 120i
3.5− > 62j,l
aDerived from single dish data (Mauersberger & Henkel 1991; Fuente et al. 2006). See Sect. 4.1 for details.
bAverage ratio from the interferometric maps by (Garc´ ıa-Burillo et al. 2002)
cAverage over the whole line profile in the CND position (Usero et al. 2004).
dAverage value over the three positions in the circunnuclear starburst ring with HCO detections (Usero et al.
eFrom Wang et al. (2004).
fGoicoechea et al. (2009)
gGerin et al. (2009)
hApponi et al. (1999)
iOrion Bar ionization front and PDR-peak in NGC7023 Fuente et al. (2003)
jSchilke et al. (2001)
kSavage & Ziurys (2004)
lSchenewerk et al. (1988)
mSt¨ orzer et al. (1995)