Mid- infrared [Nell] line emission from the nucleus of NGC 253
Torsten Bokcr
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, U.S.A.
Alfred Krabbe
Max-Planck- Institut fur extraterrestrische Physik, Garching, Germany
DLR, Institute of Space Sensor Technology, Rudower Chaussee 5, Berlin, Germany
John W.V. Storey
University of New South Wales, School of Physics, Sydney Australia
Received Jan. 14, 1998; accepted March 6, 1998
- 2 -
ABSTRACT
We report on mid-infrared (MIR) continuum and line emission mapping of the nucleus of
NGC 253. The data, with a resolution of 1".4, reveal a double-peaked arc-like [Nell] emission
region. Comparison with published data shows that the [Nell] arc is centered on the nucleus
of the galaxy. The brightest [Nell] source coincides with the infrared continuum peak. The
interpretation of these results is complicated by the edge-on orientation of NGC 253, but a
self-consistent explanation is starformation triggered by dynamical resonances in a barred
potential.
Subject headings: infrared: ISM: lines and bands — galaxies: starburst — galaxies: individual
(NGC 253)
- 3 -
Introduction
The prototypical starburst galaxy NGC 253 is a highly inclined (i=78°, Pence 1981) disk galaxy
at a distance of about 3 Mpc (Fully 1988). The view towards its inner region is heavily obscured, and
the exact location of its nucleus has been the matter of some debate. The astrometry of the various
infrared sources in the central 10" has been discussed in detail by Kalas & Wynn- Williams (1994, hereafter
KW94) in the near-infrared and Keto et al. 1993 in the mid-infrared. From their wl" resolution H, K,
L, and M observations, KW94 conclude that the brightest source in all the maps (their peak 1) lies at
R.A.(1950)=00 /l 45 m 5 s .62, Decl.(1950)=-25°33'40".2. This position is in very good agreement with that
of the emission peak at MIR wavelengths as determined by Keto et al. (1993) at 10 /xm from astrometric
methods and Piha et al. (1992) from low-level contour fitting to published maps with lower resolution. It is
safe to assume that the location of peak 1 is independent of wavelength up to 20 /im, enabling us to adopt
the above coordinates for the continuum peak in our maps.
Peak 1 is not connected to any of the strong radio point sources found at 2 cm by Turner & Ho (1985,
hereafter TH85) and 6 cm by Ulvestad fc Antonucci (1991)| and Antonucci & Ulvcstad (1988). The radio
sources are highly aligned along the major axis of NGC 253 at P. A. 51°. All of them are most likely either
radio supernovae or HII regions, based on their spectral index. The only exception is the source TH 2 in
TH85, a powerful flat spectrum radio source that they proposed to be the nucleus of NGC 253. TH85 argue
that because of its high radio luminosity (10 4 L©) and its unique location at the center of a synchroton
disk this source probably is a compact synchroton source, similar to those observed in active galactic nuclei
(AGNs). It lies close to a secondary MIR peak (peak 2) « 2". 2 northeast of peak 1 (Piha et al. 1992, Keto
et al. 1993). However, as we will show, the radio nucleus is most likely not identical with any characteristic
infrared source. This view was strongly supported from 0.5" resolution NIR color maps by Sams et al.
(1994). Their maps were referenced to other observations by cross-correlation of low level contours and
show that the radio nucleus TH 2 coincides with an extinction maximum of Ay > 24. In addition, KW94
show that the exact position of peak 2 is wavelength dependent in the sense that for J, H, K observations it
lies w 1" further east than in M-band.
The physical nature behind the various infrared sources is uncertain. For example, according to Forbes
et al. (1993) , the emission of the Br7-line at 2.17 /im is strongest at peak 1. On the other hand, KW94 show
that peak 1 has a low PAH-feature/continuum emission ratio. Since PAH emission is usually indicative of
ongoing star formation, KW94 argue against peak 1 being an intense starburst which would be the natural
-4-
explanation for the Br7 emission. To investigate this matter further, we have observed NGC 253 at MIR
wavelengths, both in continuum and line emission. The forbidden 2 P3/2 — 2 P1/2 transition of singly ionized
Neon at 12.81 /im traces the photoionization regions of hot stars and thus carries similar information than
e.g. Br7, but at a much lower extinction: A2. 17^,1 » 2.5 • Ai2.8pm ( Lutz et al. 1996 , Gcnzel et al. 1998 ).
In the next section we will describe how the data were obtained. We also discuss the special observing
techniques in the MIR and their implications on data reduction. We present the results of our observations
in Sec. || and interpret them in Sec. ^.
Observations and data reduction
Our observations were carried out with MANIAC, the Mid- And Near- Infrared Array Camera (Boker ei
al. 1997a| ) built at the Max-Planck-Institut fur extraterrestrische Physik in Garching, Germany. NGC 253
was observed during five nights at the ESO 2.2m telescope in LaSilla, Chile in October 1996. The average
seeing was below 1", therefore the data have a spatial resolution defined by the diffraction limit of about
I" A.
We used three narrow band filters « 0.015) centered on the [Nell] line (12.81 /im) and two
continuum positions shortwards (12.62 /im) and longwards (13.00 ^irn). This is essential for accurate
continuum subtraction, because strong Silicate absorption features and emission from polycyclic aromatic
hydrocarbon (PAH) make it difficult or even impossible to determine the continuum emission underlying
the [Nell] line from N-band photometry. The mean of the two continuum images should represent the
emission underlying the [Nell], its subtraction from the 12.81 /xm map then reveals the distribution of the
[Nell] emission. The total on-source integration time was about 9600 s on the [Nell] filter and 4800 s on
each of the two continuum filters.
Due to the high thermal background of the atmosphere at wavelengths around 10 /im, rapid readout
of the detector, chopping of the telescope secondary, as well as frequent beam switching is essential. We
typically used a chop amplitude of 15" so that the source remained on the MANIAC detector all the time.
The correct telescope movement for the beam switch was defined with high precision by alternate guiding
on each of the two (chopped) images of a nearby guide star.
The sky signal through the filters allowed single frame integration times of 80 ms before saturating
the detector. Frames taken immediately after the chop movement were discarded because of the settling
- 5 -
time of the secondary. After 1 min, the data (two sums of about 375 frames for each chop position) were
transferred to the MANIAC PC and the beam switch was performed. After subtraction of the two beams,
the residual background emission in our data is typically reduced by a factor of 10 5 , as can be determined
from empty regions in our images.
During the course of the night one gathers many such exposures, which have to be coadded. Because
of refraction by the atmosphere, the source slowly changes position on the detector as the zenith distance
changes. We correct for this drift by coadding a few exposures both at the beginning and the end of the
observations to increase the signal-to-noise ratio, and determining the centroids of the well-defined emission
peak. Their difference in position gives the drift over this timeline. Assuming it to be linear with time, we
can estimate the drift per exposure and correct each before coadding them.
The chop amplitude and direction, required for the correct overlay of the three source images on the
detector, was accurately determined by observing a point source in an identical way, and fitting Gaussians
to its three images. Finally, the resulting maps were calibrated for (slightly) different filter transmissions
and a mean image for each filter was generated from the data of the individual nights. Because the 12.81/mi
filter showed a ghost reflection during the October 1996 run, the filter maps had to be deconvolved with
their point spread functions (PSF) before subtraction. We used a CLEAN algorithm (Eogbom 1974)
with observations of reference stars as the PSF for each filter. Since the observations described here are
diffraction limited, the PSF is much more stable and the deconvolution more reliable than is normally the
case.
Flux calibration of our data was performed by assuming that the 12.8 /im flux density of the nearby
star (3 Ceti is 53 Jy. This number was extrapolated from the 12 /im IRAS flux and a blackbody temperature
of 4200 K. The flux density inside a 5". 5 aperture at 12.8 /im is 17 Jy which should be compared to the
value of 11.2 Jy given by Rieke & Low (1975) in the same aperture at 12.6 /mi. The total flux density in our
map is 35 Jy, about a factor of 2.5 higher than the result of Keto et al. (1993), measured at 12.5 /mi. On
the other hand, the [Nell] flux we derive from the central 7". 6 is 2 • 10~ 14 W/m 2 , roughly 60% of the result
by Roche et al. (1991). The data thus seem to agree to some extent, although it should be stressed that a
direct comparison between various data sets is problematic because of the differences in filter passbands
and bandwidths. This is especially true in the MIR since this regime is rich in spectral features.
3. Results
- 6 -
3.1. Morphology of the continuum and [Nell] emission
The results of the above data reduction are presented in Fig. |l| (Plate 1). We show in panel A) the
"raw" 12.81 fim filter map before deconvolution with the PSF and in B) the deconvolved 12.81 /im map.
Panels C) and D) contain the two deconvolved continuum maps at 12.62 /im and 13.00 /im. Finally, panel
E) shows the resulting [Nell] line map, calculated as described in Sec. ^. The continuum maps show a
good overall agreement with other MIR maps at comparable resolution ( Pina et al. 1992 , Keto et al. 1993 ),
although there are differences in the detailed emission morphology. This emphasizes our point that a reliable
continuum subtraction is only possible with narrow band continuum filters neighboring the emission line.
Since both the strongest two continuum sources and the elongated emission structure along the major
axis of the NGC 253 disk at P.A. 51° ( [Pence 1981 ) are easily identified, we can safely adopt the astrometry
and terminology used by KW94 (Sec. [l]). For the remainder of this paper, we will identify our strongest
continuum source with peak 1 and the secondary continuum source with peak 2, keeping in mind that at
NIR wavelengths peak 2 is offset by « 1" to the east.
The most striking result of the [Nell] map is that the emission forms a double-peaked, arc-like structure.
The brightest source coincides with peak 1, adding to the evidence that it is indeed powered by a burst of
star formation. The opposite end of the arc seems to coincide with peak 2, but is somewhat elongated in
E-W direction. The [Nell] morphology is in excellent agreement with the Bi'7 map of Forbes et al. (1993)
It is important that the [Nell] emission from the central ~ 5" is aligned with the major axis of the
NGC 253 optical disk. In Fig. || we show again contour plots of the [Nell] and continuum emission
to demonstrate its relative orientation. The line of nodes at P.A. 51° is indicated by the dashed line.
The dotted line shows the orientation of the NIR bar detected first by |5covillc et al. (1985)| . The [Nell]
morphology and its relation to the NIR bar will be further discussed in Sec. |j.
3.2. The radio nucleus
Adopting the coordinates of peak 1 of KW94, we can relate our maps to the astrometry at longer
wavelengths. In Fig. ||, we have marked the location of the compact radio sources of TH85 on both the
[Nell] and the continuum emission. As can be seen, the major axis of the [Nell] emission coincides well
with the radio axis. The radio nucleus TH 2 - or extinction peak according to Sams et al. (1994) - does
fall on none of the [Nell] or continuum peaks. Rather, TH 2 lies inside the arc-like [Nell] emission region.
-7-
This supports the view of TH85 that TH 2 is most likely not an HII region, but a synchroton source in the
dynamical center of NGC 253. All other radio point sources are distributed rather randomly in the plane of
the [Nell] emission, supporting the idea of them being radio supernovae (TH85).
To summarize, the location of the nucleus of NGC 253 lies in the center of the [Nell] emission arc,
about 2". 2 northeast of peak 1 in a region with high extinction.
4. Interpretation and discussion
4.1. The bar in NGC 253
From 10" resolution K-band imaging, Bcoville et al. (1985) find a prominent NIR bar at P.A. 68°.
However, the NIR continuum of the central w 10" in all high resolution NIR maps (e.g. Forbes et al. 1993,
KW94, Sams et al. 1994) is oriented at P.A. 51°, roughly parallel to the major axis of the optical disk. This
change in orientation is also visible as an isophote twist in Scoville et al. (1985) and all MIR maps (Piha et
al. 1992j , |Keto et al. 19931 , Fig. |). What is the reason for this behavior? From kincmatical studies of the
CS(2-1) line at 98 GHz, Peng et al. (1996) have tried to answer this question. They derived a model for the
dynamical processes in the center of NGC 253. In short, they find five prominent CS emission spots. The
two innermost spots are aligned along the axis of the radio knots, i.e. the major axis of NGC 253. They
are separated by 3" and positioned symmetrically on either side of the radio nucleus. Two other knots
lie again symmetrically to the nucleus, but are aligned with the large scale NIR bar. Their observations are
well explained by a rotating gas torus around the dynamical center of NGC 253. The torus is the response
of the viscous molecular gas to the potential of the NIR stellar bar. Gas clouds in the inner w 10" move
along elliptical x-i orbits whose major axis is aligned perpendicular to the bar (Athanassoula 1992a, b). At
the inclination of NGC 253, these orbits appear to be oriented parallel to the major axis of the optical disk
as demonstrated in Fig. 5 of Peng et al. (1996) . The gas clouds outside of w 10" move on X\ orbits, their
major axis being aligned with the bar. The existence of xi orbits depends on the presence of at least one
Inner Lindblad Resonance (ILR). For the case of NGC 253, Arnaboldi et al. (1995) have shown that there
are in fact two ILRs, one at a radius of 25", the other close to the nucleus. It is at the inner ILR, where the
orientation of the dominant orbits changes from x^ to x\ (e.g. Felesco & Decher 1988). Because of orbit
crowding, massive star formation occurs predominantly at the apocenters of the Xi orbits (Athanassoula
1992b ), oriented symmetrically on either side of the nucleus. Therefore, Peng et al. (1996) interpret their
CS results as evidence for a rotating torus of dense gas around the radio nucleus TH 2. These results are in
- 8 -
agreement with observations of other tracers of high density gas like HCN (Paglionc et al. 1995)
4.2. The [Nell] arc - a starformation ring
Based on their model, Peng et al. (1996) predict that higher resolution line mapping should prove that
gas emission from the central 10" (145 pc) of NGC 253 is aligned with the radio knots rather than with
the optical disk. In that context, our results provide an independent confirmation of this scenario. Firstly,
the [Nell] emission is well aligned with the radio knots (Figs. |] (Plate 1) and||). Secondly, based on the
structure of the [Nell] emission, we also confirm the distribution of the star forming material in a ring
around the radio nucleus. The ring has a rotation velocity of ~ 60 km/s as seen in the CS data by Peng
et al. (1996), the western side moving away from the observer. Unfortunately, our observations are unable
to resolve the velocity structure of the [Nell], this has to await higher resolution line mapping, e.g. with a
cryogenic Fabry-Perot interferometer.
Our data only show weak emission from the southeastern half of the ring. On the other hand,
the [Fell] map of Forbes et al. (1993) does show emission from this region (their source B). Thus, we
favor the interpretation of the double-peaked morphology as a starburst ring with a diameter of « 4"
(60 pc). The total observed flux from a circular aperture with 1".4 (20 pc) diameter centered on peak 1
is 1.7 • 10~ 15 W/m 2 . After dereddening with Ay = 24 (Sams et al. 1994) in a mixed case scenario, this
corresponds to 7.7 ■ 10 5 Lq|^.
With the conversion factor
'[Nell]
64 (Genzel et al. 1998), the intrinsic Lyman-continuum luminosity
is 4.9 • 10 7 Lq. This value should be used with caution given its uncertainties, but it compares well to other
bright star forming regions like the core of 30 Doradus in the LMC (Brandl et al. 1996) or the brightest
SFR in the gaseous ring of IC 342 (Boker et al. 1997b). This supports the view that the [Nell] emission
probably stems from individual giant molecular clouds (GMCs) that actively form stars and are located in
a ring at the ILR. We will extend the comparison to a similar structure in IC 342 in the next section.
1 We have adopted Ai 2 . 8/i m = 0.04 ■ A v ( |Genzel et al. 1998[ )
- 9 -
4.3. NGC 253 and IC 342 - two of the same kind?
It is interesting to note the close similarities between NGC 253 and IC 342, another nearby late type
spiral. Boker et al. (1997b) describe the dynamical processes in the central 100 pc of IC 342 as derived
from NIR integral field spectroscopy. Table [l] compares the two galaxies. The dynamical processes and
the morphology of the molecular gas are almost identical. This was also noted by Paglione et al. (1995).
One important difference, however, is that IC 342 does not house a central synchroton source. In fact, in
the case of IC 342, there is no radio emission from the dynamical center ( Turner fc Ho 1985 ). Rather, it
is occupied by a cluster of young red supergiants that dominate the NIR continuum. For NGC 253, there
is no obvious NIR component at the location of the radio nucleus, but the edge-on orientation and patchy
extinction might complicate its detection. Both the diameter and rotation velocity of the molecular rings
are very similar. This adds to the increasing evidence that stellar bars and small star formation rings on
scales of 50-100 pc are a common feature in late type spirals. This points to an evolutionary connection
between spirals of different Hubble type, as suggested by Pfenniger et al. (1994).
5. Summary
We have reported on MIR line and continuum mapping of the central 200 pc of NGC 253. The main
results can be summarized as follows:
1) The continuum maps confirm that the nucleus of NGC 253 is not identified with any strong IR source.
2) The [Nell] map reveals two peaks separated by about 4" and connected by an arc-shaped emission region.
3) The nucleus lies inside the [Nell] arc.
4) We interpret the morphology of the [Nell] map as evidence for a bar-triggered starformation ring with a
diameter of 50 pc, similar to that seen in IC 342.
It is a pleasure to thank Thomas Lehmann for his invaluable help before, during, and after the
campaign. We are indebted to the ESO 2.2m-team for their continuous support, and to Rolf Chini and
Craig Smith who helped a great deal with the observations. We also would like to acknowledge Thomas
Ott for his expertise on deconvolution algorithms, and the anonymous referee whose comments led us
to re-examine the point spread function and thereby substantially improve this paper. JWVS thanks
the Alexander von Humboldt Stiftung for their support during the development of MANIAC, and the
Australian DIST for travel support.
- 11 -
REFERENCES
Antonucci, R. R. J. & Ulvestad, J. S. 1988, ApJ, 330, L97
Arnaboldi, M., Capaccioli, M., Cappellaro, E., Held, E. V., & Koribalski, B. 1995, AJ, 110, 199
Athanassoula, E. 1992a, MNRAS, 259, 328
Athanassoula, E. 1992b, MNRAS, 259, 345
Boker, T., Storey, J. W. V, Krabbe, A., & Lchmann, T. 1997a, PASP, 109, 827
Boker, T., Forster-Schreiber, N. M., & Genzel, R. 1997b, AJ, 114, 1883
Brandl, B., et al. 1996, ApJ, 466, 254 & Wild, W. 1990, ApJ, 348, 434
Forbes, D. A., Ward, M. J., Rotatiuc, V., Blietz, M., Genzel, R., Drapatz, S., van der Werf, P., & Krabbe,
A. 1993, ApJ, 406, Lll
Genzel, R. et al. 1998, accepted for publication in ApJ
Hogbom, J. A. 1974, A&AS, 15, 417
Kalas, P. & Wynn- Williams, C. G. 1994, ApJ, 434, 546
Keto, E., Ball, R., Arcns, J., Jernigan, G., Meixner, M., Skinner, C, & Graham, J. 1993, ApJ, 413, L23
Lutz, D., et al. 1996, A&A, 315, L269
Madore, B. F. & Freedman, W. L. 1992, PASP, 104, 362
McCall, M. L. 1989, AJ, 97, 1341
Newton, K. 1980, A&A, 191, 169
Paglione, T. A. D., Tomaka, T., & Jackson, J. M. 1995, ApJ, 454, L117
Pence, W. D. 1981, ApJ, 247, 473
Peng, R., Zhou, S., Whiteoak, J. B., Lo, K. Y., & Sutton, E. C. 1996, ApJ, 470, 821
Pfenniger, D., Combes, F. & Martinet, L. 1994, A&A, 285, 79
Pina, R. K., Jones, B., Puettcr, R. C, & Stein, W. A. 1992, ApJ, 401, L75
Roche, P. F., Aitken, D. K., Smith, C. H., & Ward, M. J. 1991, MNRAS, 248, 606
Rieke, G. H. & Low, F. J. 1975, ApJ, 197, 17
Sams, B. J. III., Genzel, R., Eckart, A., Tacconi-Garman, L. E., & Hofman, R. 1994, ApJ, 430, L33
Scovillc, N. Z., Soifer, B. T., Neugebauer, G., Young, J. S., Matthews, K., & Yerka, J. 1985, ApJ, 289, 129
Telesco, C. M. & Decher, R. 1988, ApJ, 334, 573
Tully, R. B., 1988, Nearby Galaxies Catalog (Cambridge: Cambridge Univ. Press)
Turner, J. L. & Ho, P. T. P. 1983, ApJ, 178, 623
Turner, J. L. & Ho, P. T. P. 1985, ApJ, 299, L77 [TH85]
Ulvestad, J. S. & Antonucci, R. R. J. 1991, AJ, 102, 875
This manuscript was prepared with the AAS IATgX macros v4.0.
- 13 -
Fig. 1.— MIR maps of the central 15" (200 pc) of NGC 253. A) "raw" 12.81/xm filter map
before CLEANing. B) 12.81/im filter map after CLEANing. C) 12.62 /im continmim map after
CLEANing. D) 13.00 /im continuum map after CLEANing. E) [Nell] line map calculated as B -
(C+D)/2. The resolution in all maps is 1".4 The coordinate origin is the continuum peak (peak 1) at
R.A.(1950)=00 /l 45 m 5 s .62, Decl.(1950)=-25°33 / 40".2. The contour levels for the three continuum maps
are 4, 8, 12, 16, 20, 30, 40 ••• 100% of the peak continuum flux at 12.8,um (1.6 • 10~ 16 W/m 2 /a") 1 and
6, 12, 18, 24, 30, 40, 50 . . . 100% of the peak line flux (2.3 Jy/D") for the [Nell] map.
Fig. 2. — 12.8 fim continuum emission (top) and [Nell] line emission (bottom) from the
central 15" (200 pc) of NGC 253. The coordinate origin is the continuum peak (peak 1) at
R.A.(1950)=00' l 45 m 5 s .62, Decl.(1950)=-25°33'40".2 (see text for discussion). The resolution in both maps
is 1".4. Contour levels arc 4, 8, 12, 16, 20, 30, 40 • • • 100% of the peak flux (2.3 Jy/D") for the continuum and
6, 12, 18, 24, 30, 40, 50 . . . 100% of the peak flux (1.3 • 10~ 15 W/m 2 /D") for the [Nell] map. The location of
the non-thermal radio nucleus TH 2 (coincident with the NIR extinction peak of Sams et al. 1994) as well as
the thermal 2cm radio point sources of TH85 are marked by the indicated symbols for Nuc and SNR, resp.
The dashed line shows the orientation of orbits that form the starburst ring. Incidentally, this is identical
with the major axis of the optical disk of NGC 253 at P.A. 51°. The dotted line denotes the orientation of
the NIR bar at P.A. 68°.
-14-
Table 1. NGC 253 and IC 342 - a comparison
Property
IC 342
NGC 253
Ref.
Distance [Mpc]
1.8
3
1,2
Inclination
25°
78°
3.4
P.A.
39°
51°
3.4
Morph. type
Scd
Sc
5
Radio core
no
yes
6,7
P.A. of NIR bar
0°
68°
8,9
No. oflLR's
1
2
8,10
Ring diameter [pc]
70
GO
8,11
L Lyc from brightest SRF [10 7 L ]
1.7
4.9
8,12
Note. — References: 1) |McCall 1989j |Madore fc Freedman 1992
2) [Tully 1988] 3) |Newton 198C| 4) |Pence 1981] 5) NASA extragalactic
database (NED) 6) [Turner fc Ho 1983| 7) [Turner fc Ho 1985| 8)
Bdkcr et al. 1997bj 9) Scoville et al. 1985 10) Arnaboldi et al. 1995
11) Peng et al. 1996, |Paglione et al. 1995 , this work, 12) this work
A) 'raw' Continuum + line (12.81/um)
B) CLEANed Continuum + line (12.81/im)
-10
-5
Offset (arcsec)
1 2. 8/zrri — Continuum
-5 +5
Offset (arcsec)
