Skip to main content

Full text of "SCUBA Observations of NGC 1275"

See other formats


Mon. Not. R. Astron. Soc. 000, hHia (2000) Printed 1 February 2008 (MN M£X style file vl.4) 



SCUBA Observations of NGC 1275 



Judith A. Irwin, 1 J. M. Stil, 1 and T. J. Bridges 2 

IDept. of Physics, Queen's University, Kingston, Canada, K1L 3N6 

2 Anglo- Australian Observatory, PO Box 296, Epping, NSW 2121, Australia 



Accepted xxx. Received xxx; 



o 
o 

(N 



> 

O 
CO 

o 
r-> 

o 

o 

a,: 
6 

CO 

03 

"^ 1 INTRODUCTION 

• i—i . 

S^ , Until recently, it was not thought that centrally-located clus- 

J_j ' ter galaxies contained significant amounts of dust. However, 
recent observations have begun to reveal large amounts of 
dust in some central galaxies. The existence of dust is in- 
ferred from a diverse set of observations, including the ratios 
of emission lines in optical filaments (Hu 1992; Donahue & 
Voit 1993; Crawford et al. 1999), dust lanes in optical im- 
ages (e.g. McNamara, O'Connell & Sarazin 1996; Pinkney 
et al. 1996), near infrared (NIR) imaging and spectroscopy 
(e.g. Krabbe et al. 2000; Donahue et al. 2000; Jaffe, Bremer, 
& van der Werf 2000) , and X— ray absorption (e.g. White et 
al. 1991; Allen & Fabian 1997; Arnaud & Mushotzsky 1998; 
Allen 2000). 

Searches at longer wavelengths have now also revealed 
emission from dust directly, especially cooler dust which may 
be much more plentiful than the warmer components. The 
Infrared Astronomical Satellite (IRAS) detected several cen- 
tral cluster galaxies at 60 and 100 pirn; Cox, Bregman, & 
Schombert (1995) find that ~ 10% of central galaxies have 
far infrared (FIR) emission, with dust masses of ~ 10 7 M©. 
More recently, Lester et al. (1995) have detected 100 /im 
emission from NGC 1275 with the Kuiper Airborne Obser- 
vatory (KAO), and there have been Infrared Space Observa- 



ABSTRACT 

Deep SCUBA observations of NGC 1275 at 450/itm and 850/im along with the 
application of deconvolution algorithms have permitted us to separate the strong core 
emission in this galaxy from the fainter extended emission around it. The core has a 
steep spectral index and is likely due primarily to the AGN. The faint emission has a 
positive spectral index and is clearly due to extended dust in a patchy distribution out 
to a radius of ~ 20 kpc from the nucleus. These observations have now revealed that a 
large quantity of dust, ^6 x 10 7 M Q , 2 orders of magnitude larger than that inferred 
from previous optical absorption measurements, exists in this galaxy. We estimate the 
temperature of this dust to be ~ 20 K (using an emissivity index of (3 = 1.3) and the 
gas/dust ratio to be 360. These values are typical of spiral galaxies. The dust emission 
correlates spatially with the hot X— ray emitting gas which may be due to collisional 
heating of broadly distributed dust by electrons. Since the destruction timescale is 
short, the dust cannot be replenished by stellar mass loss and must be externally 
supplied, either via the infalling galaxy or the cooling flow itself. 



Key words: 

methods: data analysis - ISM: dust - galaxies: individual: NGC 1275 
flows - galaxies: clusters: individual: Perseus. 



galaxies: cooling 



tory (ISO) detections at 60—200 /jm of a few central galaxies 
(e.g. Hansen et al. 2000). Annis & Jewitt (1993) carried out 
an unsuccessful search for dust emission in 11 cooling flow 
clusters using the UKT-14 bolometer system on the James 
Clerk Maxwell Telescope (JCMT). More recently, however, 
Edge et al. (1999) have detected dust emission at 850 /jm 
in two cooling flow clusters using the more sensitive Submil- 
limetre Common-User Bolometer Array (SCUBA) camera. 
The origin of the dust in central cluster galaxies is a 
matter of some controversy. Central cluster galaxies live 
in complicated environments, with many at the centers 
of cluster cooling flows (e.g. Fabian 1994), and interac- 
tions/mergers may also be common. Dust in central cluster 
galaxies may thus originate in several ways: from the cooling 
flow itself or in cold, dense clouds condensed from the flow; 
from ongoing star formation, perhaps initiated by the cool- 
ing flow; or from the merger of a gas/dust rich galaxy. Direct 
detections of dust in central cluster galaxies, and good de- 
terminations of the dust properties (mass, temperature, and 
correlations with other components) will help enormously to 
decide between these alternatives. For example, the bulk of 
the dust in cooled gas clouds is expected to be quite cold 
with T < 15K (e.g. Fabian, Johnstone, & Daines 1994; John- 
stone, Fabian & Taylor 1998) [However, this conclusion is 
disputed by O'Dea, Baum & Gallimore (1994) and Voit & 



© 2000 RAS 



Judith A.Irwin et al. 



Donahue (1995), who claim minimum dust temperatures of 
~ 20 K.] 

The very existence of dust in central cluster galaxies 
is in itself surprising, since any dust should be destroyed 
quickly by sputtering from X— rays from the hot cluster gas, 
on timescales < 10 7 years (e.g. Draine & Salpeter 1979). 
Either the dust is shielded against sputtering (e.g. at the 
centers of cold, dense molecular clouds), or the dust is re- 
plenished on simila r tim escales. This issue is discussed in 



more detail in Sect. 4.2 



NGC 1275 is a complex and fascinating object. It has 
been studied in virtually every waveband, including the op- 
tical (e.g. N0rgaard-Nielsen et al. 1993; McNamara et al. 
1996), NIR (e.g. Donahue et al. 2000; Krabbe et al. 2000) 
and far infrared (FIR) (Gear et al. 1985; Lester et al. 1995). 
It has been classified as a Seyfert (Seyfert 1943) and has an 
active galactic nucleus (AGN) in the form of the highly vari- 
able radio source 3C 84 (e.g. Pedlar et al. 1990). Minkowski 
(1957) discovered two distinct filamentary systems around 
NGC 1275. The low velocity system (LVS) has the same 
velocity (~ 5300 km s" 1 ) as that of NGC 1275 itself, and 
has been studied in detail by several authors (e.g. Lynds 
1970; Hu et al. 1983; Caulet et al. 1992). The high velocity 
system (HVS) has a velocity ~ 3000 km s~ x higher than 
NGC 1275 and consists of giant HII regions to the North 
and North- West of the galaxy center; it has generally been 
concluded that the HVS is a late-type spiral galaxy. At 
least part of the HVS is in front of the nucleus, as deter- 
mined from HI absorption against the nuclear radio source 
(De Young, Roberts, & Saslaw 1973) and Lya absorption 
(Briggs, Snijders, & Boksenberg 1982). The HVS could be 
merely a chance projection onto NGC 1275, but several au- 
thors have argued that the HVS is in fact colliding with 
NGC 1275 (e.g. Hu et al. 1983; Unger et al. 1990; Caulet et 
al. 1992; N0rgaard-Nielsen et al. 1993). Recent observations 
of gas at velocities intermediate between the LVS and HVS 
also strongly argue for a physical connection between the 
two systems (Ferruit et al. 1998). 

It has been known for some time that the Perseus cluster 
hosts a large cooling flow (e.g. Fabian et al. 1981); analysis of 
recent Advanced Satellite for Cosmology and Astrophysics 
(ASCA) data show that the mass infall rate is ~ 300 M© 
yr _1 (Allen et al. 1999). Rontgen Satellite (ROSAT) and 
Chandra X— ray observations of NGC 1275 and the Perseus 
cluster (Bohringer et al. 1993, and Fabian et al. 2000, re- 
spectively) have also revealed two holes in the NGC 1275 
X— ray emission, thought to be caused by displacement of 
hot gas by the radio lobes. NGC 1275 is the only cooling 
flow central galaxy with detected CO emission, with an in- 
ferred molecular H2 mass of ~ 1O 1O M0; see Bridges & Irwin 
(1998) for a summary of the many previous CO studies. 

In this paper, we present new SCUBA 450/850 /im data 
for NGC 1275, in an attempt to study the cold dust in this 
galaxy in more detail. We are especially interested in such 
dust properties as: its spatial distribution and correlation 
with other components, total mass, temperature and ulti- 
mately, its origin. In Sect. 0, we present the observations and 
data reductions, unprocessed and processed images, Sect. 
presents the results, including an analysis of the AGN and 
extended dust components, separately, and correlations with 
other wavebands, Sect. W discusses the origin of the dust 
and dust heating, and Sect. tA presents the conclusions. For 



NGC 1275, we take D = 70 Mpc (H = 75 km s _1 Mpc -1 , 
Bridges & Irwin 1998). 



2 OBSERVATIONS AND DATA REDUCTION 

2.1 Maps of NGC 1275 

Table hi lists the observing log of SCUBA jiggle maps ob- 
tained on July 15, October 31 and December 14 1998. The 
total observation time was 5.5 hours, half of which was spent 
on the source. Chopping was performed in Azimuth, with a 
chop thow of 120 arcseconds. 

The standard reduction scheme for SCUBA jiggle maps 
was followed, consisting of flatfielding, correction for atmo- 
spheric extinction, removal of noisy bolometers, sky noise 
removal and despiking. These steps are described in de- 
tail in Sandell (1997). Flatfielding corrects for differences 
in sensitivity of the bolometers. The atmospheric opacity at 
450 and 850 /im was determined from skydip observations. 
Noisy bolometers were identified from noise measurements 
performed in between observations, and visual inspection of 
the signal of all bolometers as a function of time during the 
observations. 

Variations of the sky brightness are not canceled out 
completely by chopping because the atmosphere in the di- 
rection of the chopped beam is significantly different from 
the atmosphere in the direction of the source. This is a par- 
ticular concern for observations of an extended source, where 
a large chop throw is required (Holland et al. 1999). Varia- 
tions of the sky on short time scales were further eliminated 
by subtracting the median sky level measured by off-source 
bolometers for each 18 second integration. Although the sky 
level is determined with higher accuracy if more bolometers 
are considered, the likelihood that faint extended emission 
of the source is also subtracted, increases as well. A compro- 
mise must be sought between decreasing the sky noise and 
the possibility of subtracting source emission. The sky level 
was determined from the outer two rings of bolometers at 
450 fim, and the single outer ring of bolometers at 850 /im. 
In this way a significant aperture around the centre of the 
field of view is excluded from the determination of the sky 
level. The radius of the aperture which is never covered by 
any of the "sky" -bolometers is 28" at 450 /im and 46" at 850 
(im. As the sky level at any time is taken to be the median 
of a ring of 54 (450 /im) or 18 (850 (im) bolometers sur- 
rounding the source, only a smooth symmetric component 
would be subtracted in this way. 

Occasional spikes in the output of each bolometer were 
then removed by excluding signals deviating more than ±3<r 
from the mean signal. 

The data were averaged with weight factors Wi — 
(ti/Ni) 2 where ti and JVj are the integration time and 
noise level of observation i, to produce the maps shown in 
Fig. fu These weight factors take into account both the dif- 
ferent integration times and atmospheric opacities of the 
observations, since the noise in each map is dominated 
by the sky. The relative weights of the three nights are 
July:October:December=0. 48:1:0. 64. 

Recently, a potential calibration problem resulting from 
erroneous values 

of the hot and cold load temperatures was identified (see 



© 2000 RAS, MNRAS 000, §-^| 



SCUBA Observations of NGC 1275 3 



http : //www . j ach . hawaii . edu/ JACdocs/ JCMT/SCD/SN/002/ 
tau_analysis.html). This problem could affect the sky 
opacity derived from skydip observations. The calibration 
was repeated with opacities calculated from the CSO opac- 
ity and the revised relations for the opacities at 450/mi and 
850/im. We found that this procedure had no effect on our 
calibration. 

Taking into account the repeatability of the flux deter- 
mination over a variety of integration times as well as an 
assessment of the accuracy of the calibration factor includ- 
ing uncertainties in the calibrator fluxes, we estimate that 
the 450/mr calibration error is ~ 25% and the 850/im error is 
~ 10%. Typical pointing corrections were 1 to 2 arcseconds 
over the course of an hour and residual pointing errors will 
be smaller than this. 

The corrected, calibrated maps are shown in Fig. hk 
(450/im) and Fig. @c (850/im). 

2.2 Beam maps 

Beam maps were constructed for each night from calibrator 
maps with the same chop throw that was applied to the ob- 
servations of NGC 1275. These calibrator maps were reduced 
in the same way as the observations of NGC 1275, with 
one exception: noisy bolometers were not excluded from the 
analysis. This was done to avoid the occurrence of unsam- 
pled areas in the planet maps. However no significant noise 
was introduced, because at 850/im the primary calibrators 
Mars and Uranus are brighter than NGC 1275 by factors 
of 300 and 60 respectively. At 450/im these factors are even 
larger because the brightness of the planets rises, whereas 
the spectrum of NGC 1275 shows a decline with frequency. 
Therefore, it is safe to conclude that noisy bolometers in the 
beam maps do not affect our results. 

Calibrator sources are Uranus, Mars, CRL 618, and 
HL Tau. We used planet maps made with the same chop 
throw as the NGC 1275 maps, to define the shape of the 
beam for each night. Only for the 450/im map in December, 
we used the secondary calibrator CRL 618, as no 450/im 
planet map with the appropriate chop throw was avail- 
able. The calibration factors adopted here are those deter- 
mined for the October observations: 623.69 Jy/(Volt beam) 
at 450/im and 251.54 Jy/(Volt beam) at 850 /im. Our largest 
calibrator was Mars which was smaller than 5" over the 
course of the observations. This broadens the beam by less 
than 0.3" at 850/im and by less than 0.6" at 450/im, or < 
1.8% and < 6.7% of the beam size, respectively (see caption 
to Figure 2 for the fitted beam size). Thus, the finite size of 
the calibrators will not significantly affect the beam, or its 
far sidelobes. 

The coordinate system for the beam maps produced in 
this way was azimuth-elevation because chopping was done 
in azimuth, and the beam tends to be slightly elongated in 
the direction of chopping. For each observation of NGC 1275, 
the appropriate beam map was shifted to the position of the 
peak in the final map of NGC 1275, and rotated with the 
field rotation angle at the time of mid-observation. The re- 
sulting beam maps were coadded with the sam e weight fac- 
tors as the data for NGC 1275 (see Sect. 2A) to construct 
a map of the effective beam. The beam maps are shown 
in Fig. |lp and d. The most distinctive elongated feature in 
position angle ~ 40 degrees visible in the 850/im maps of 



Num Date 



Time 
(Local) 



Exp. time, 
(hour) 



Airmass tsso 



130 


Jul 15 


05:25 


0.5 


1.276 


0.23 


131 


Jul 15 


05:53 


0.5 


1.202 


0.23 


134 


Jul 15 


06:29 


0.5 


1.135 


0.26 


135 


Jul 15 


06:58 


0.5 


1.102 


0.26 


144 


Jul 15 


08:08 


0.5 


1.079 


0.29 


115 


Jul 15 


08:36 


0.5 


1.091 


0.29 


118 


Jul 15 


09:12 


0.5 


1.128 


0.30 


12 


Oct 31 


02:11 


1 


1.160 


0.16 


57 


Dec 14 


00:23 


1 


1.326 


0.18 



Table 1. Observation 
of NGC 1275. 



for SCUBA 450/850 /im observations 



NGC 1275 and the final beam map, can be identified with 
the chop direction of the October 31 observation, which re- 
ceived a relatively high weight due to the favourable atmo- 
spheric conditions. 

The resulting beam maps (Fig. nib and d) give the best 
possible representation of the JCMT beam applicable to the 
coadded maps of NGC 1275. As only one calibrator obser- 
vation with the correct chop throw was made in a night, we 
cannot correct for variations of the error lobes during the 
night. The main source of such variations is the sensitivity 
of the structure of the telescope to temperature variations. 
Large changes in the structure of the error lobes have been 
reported in the early evening (until 9pm) and in late morn- 
ing (after 8am). However, only one observation, the 850/im 
beam map of December 14, was made outside the relatively 
stable window between 9pm and 8am. 



2.3 Processed Images 

The beam shape begins to show irregular, non-gaussian 
patterns below the 5% level and it is desirable to remove 
such features from the total intensity image in order to see 
whether real residual emission remains. We attempted to do 
this by subtracting the scaled beam from the image; how- 
ever, this assumes that the strong central source due to the 
AGN is strictly a point source and the result leaves some 
significant residual emission at the map center. We there- 
fore treated the image like a typical radio map and applied 
deconvolution methods to remove the "dirty beam" and re- 
place it with a "clean beam" with the same characteristics. 
The resulting maps retain the information from both the 
strong central source as well as any fainter extended emis- 
sion, while removing the low level beam irregularities. Re- 
sults using the CLEAN algorithm (Clark 1980), using the 
Astronomical Image Processing System (AIPS) routine, AP- 
CLN, are shown in Fig. |a (450/im) and Fig. |i (850/im) and 
results using the maximum entropy (ME) method (Cornwell 
& Evans 1985) using the AIPS routine, VTESS, are shown 
in Fig. Ep and Fig. Be. A comparison of these maps pro- 
vides an indication as to what emission is real. As the ME 
maps are well known to enhance faint, broad-scale structure 
but are less accurate than the clean maps in characterizing 
the emission quantitatively, we use the clean maps for all 
subsequent quantitative work. 

Residual maps, i.e. after subtraction of all clean com- 
ponents, are shown in Fig. He and f. For these images, we 



© 2000 RAS, MNRAS 000, 0-|l| 



4 Judith A.Irwin et al. 



NGC 1275 



I II rf '. ' 

3132 D w 



O 



41°3042 



o 
o 
o 

CM 

~3 






2952 



O 

*^ i ii 

< 3132 



O 
UJ 
Q 



41°30 42 



I I _i 



^— TRW 



'•-'-(j* 





■ ii ^ v / 

2952 O I 




3 h 19 m 49 8 



44 s 



3 h 19 m 49 s 



44 s 



RIGHT ASCENSION (J2000) 



Figure 1. Maps of NGC 1275 and the coadded beam maps at 450/xm (top row) and 850/tm (bottom row). Contours in the 450/im 
map of NGC 1275 are drawn at —10, 10, 30, 50, 70, and 90% of the peak intensity (0.78 Jy/beam). Grayscales start at 3% level on a 
logarithmic scale. Contours in the 850/xm map are drawn at —0.3, 0.3, 1, 2, 5, 10, 30, 50, 70, 90% of the peak intensity (1.44 Jy/beam). 
Grayscales start at 0.3% of the peak intensity. The contours and grayscales in the beam maps represent the same percentiles of the peak 
intensity as in the corresponding maps of NGC 1275. 



show both negative and positive contours so that the noise 
is well illustrated. Since all subtracted clean components 
are associated with the central bright core, the residual 
maps represent core-subtracted emission. At 450 /im, the 
core-subtracted map is dominated by noise which is about 
10 times higher than that of the 850/im map (see figure 
caption). At 850/im (Fig. Hf), the noise is much lower, so 
some additional errors associated with subtracting the cen- 
tral source also show up, the largest of which are within the 
central 30" region. Since positive residual emission could be 
real, we take the negative residual emission in this region to 
represent the errors from the subtraction process. Applying 
a 2cr cutoff to the core-subtracted and total intensity maps, 
we formed a ratio map to indicate the relative error in this 
region, finding an average error of 3.0% and a maximum 
error of 6.7%. 

Is the resulting emission due to continuum alone? 



Bridges & Irwin (1998) have detected CO(J=3-2) emission 
in NGC 1275 at points within ± 7" of the core. Their 
CO(J=3-2) beam of 14" is similar to our SCUBA 850/im 
beam of 16". The CO line occurs at a redshifted frequency of 
339.9 GHz and is therefore within the 30 GHz- wide SCUBA 
band which is centered at 347 GHz. Bridges & Irwin find 
an integrated intensity of 12 K km s" 1 which corresponds 
to 8.6 mjy beam -1 when diluted by the 30 GHz bandwidth 
(assuming constant response across the band). The value 
will be smaller when the bandpass response is considered 
(see Holland et al. 1999). This corresponds to only 0.6% of 
the SCUBA flux of 1.42 Jy within a 16" beam and is well 
within the noise. Thus, there is no contamination by the CO 
line at the core and we do not expect any contamination at 
locations away from the core, either, where no CO emission 
has been detected. 



© 2000 RAS, MNRAS 000, §4l| 



SCUBA Observations of NGC 1275 5 



Cleswi 






D 

a 



41 31 3D 
15 

aa 

3D 45 
3D 
15 

aa 



41 31 3D 



■-■ .S" I 










29 45 



Uaxrnun Enfropy 

4 






Core Subtracted 












t %*; 



» ,- 



fc •' ' * ; • I 



'■'-■v..- ^■•/. , 



D3 1952 5D 49 45 D3 1952 5D 49 45 D3 1952 5D 49 45 

RIGHT ASCENSION [J2DDD) 




Figure 2. Processed (clean, maximum entropy and core-subtracted) maps of NGC 1275. (a) Clean map at 450/mi. The image was cleaned 
to the first negative component and all cleaned components are associated with the strong central source. The greyscale ranges from 
to 0.7 Jy beam -1 and contours are at 75 (2<r), 100, 140, 225, 450, and 700 mjy beam -1 with a peak at 756 mjy beam -1 . The clean 
beam size is 9.1" X 8.8" @ position angle PA = 63.3°. (b) Maximum entropy map at 450 /im. Contours are at 28, 50, 100, 200, 400, 
and 650 mjy beam -1 and the greyscale is from to 0.35 Jy beam -1 . The beam is the same as in (a), (c) Residual map formed from 
map (a) but with the strong central source subtracted. The greyscale ranges from to 0.7 Jy beam -1 and contours are at -100, -75, 75 
and 100 mjy beam -1 , (d) Clean map at 850/im using the same procedure as at 450/xm. The greyscale ranges from to 0.3 Jy beam -1 
and the contours are at 7 (« 2a), 8.5, 10, 13, 18, 25, 50, 100, 200, 400, 800, and 1350 mjy beam -1 . The peak flux is 1.44 Jy beam -1 . 
The clean beam has parameters, 17.1" X 15.7" at PA=1.7°. (e) Maximum entropy map. The contours are the same as in (d) except for 
the lowest level which has been set to 6 mjy beam -1 . The greyscale ranges from to 0.32 Jy beam -1 . The beam is the same as in (d). 
(f) Residual map after subtracting the central source. The greyscale ranges from to 0.3 Jy beam -1 and contours are at -25, -18, -13, 
-10, -8.5, -7, 7, 8.5, 10, 13, 18, 25, and 50 mjy beam -1 . 



3 RESULTS 

The processed images (Fig. y) show that the 450 fim and 850 
/im emission is dominated by the bright AGN in NGC 1275. 
However, fainter extended emission is also observed and ap- 
pears to be real. Since the core is quite prominent in the 
processed maps, it has been possible to separate it from the 
extended emission and results from this process, along with 
total fluxes and the spectral index, are given in Table H. 

3.1 The AGN 

The bright AGN in NGC 1275 has shown a gradual decline 
of a factor 4 in brightness at 375 GHz (800 fim) between 
1988 and 1992 (Stevens et al. 1994), providing compelling 
evidence for the non-thermal nature of this emission. From 
our data, there is no indication for variability between the 
three epochs of the present SCUBA observations to the level 



of the calibration error, or about 6%. Moreover, our mea- 
sured 1998 total flux at 850 fj, of 1.6 Jy agrees closely with 
the 800/xm flux of ~ 1.6 Jy measured by Stevens et al. (from 
their Fig. 2) in 1994. Thus, the strong decline seen before 
1994 does not appear to have continued over the 4 years 
between 1994 and 1998. 

The core is 1.8 times brighter at 850^im than at 450/im, 
resulting in a spectral index (Table G|) of a — -0.88 ± 0.09 
(S„ oc v a ). This agrees with the value found by Stevens et 
al. (1994) of a = -0.92 ± 0.17 between 375 and 150 GHz 
over a 6 year monitoring period. Such a steep spectral in- 
dex can be produced from synchrotron emission. Thus we 
are likely observing the high frequency extension of the syn- 
chrotron emission seen at lower frequencies. It is not possible 
to extrapolate the low frequency radio flux to the sum-mm 
because of the variability of the source over the different 
epochs of observation. However, since the low frequency (i.e. 



© 2000 RAS, MNRAS 000. 



m 



6 Judith A.Irwin et al. 



Table 2. NGC 1275 Measured Parameters 



Parameter 




450/im 




850/im 


Total Flux (Jyfl 




1.2 ± 0.3 




1.60 ± 0.16 


Core Flux (Jy)|j 




0.81 ± 0.03 




1.42 ± 0.03 


Core Position 










RA (h, m, s) 




03 19 48.26 ± 0.02 


03 19 48.29 ± 0.03 


DEC (°, ', ") 




41 30 45.1 ± 0.3 


41 30 43.5 ± 0.4 


Core Size c 










Major Axis (arcsec) 


4.7 ± 0.5 




-% 


Minor Axis (arcsec) 


3.2 ± 0.5 




C 


Position Angle 


(degrees) 


90 ± 10 




_c 


otTotQ 




-0.45 ± 0.28 




OtCore" 




-0.88 ± 0.09 





a The error in the flux reflects the calibration error noted in Sect, to . Errors associated with varying the 

box size to include or exlude low S/N peaks at larger radii are smaller than this. 

6 The core parameters were derived from gaussian fits to the core component alone, averaging results from 

fits to the total intensity maps as well as the map of the core alone which was reconstructured from the 

clean components. The errors include differences between the maps and, in the case of the flux, differences 

between values measured from integration of the core map, and values determined by the gaussian fit to the 

core. Note that the calibration error and the pointing error (< 1 arcsecond) are not included in these quoted 

errors. 

c A dash means unresolved. 

d S^ <x v a 



between A90 cm and A199 cm) spectral index within the cen- 
tral 1 arcminute region is dominated by a flat spectral index 
core (see Pedlar et al. 1990), the total synchrotron spectral 
index must turn over between the low frequency and sub- 
mm regimes. Given the steepness of the sub-mm spectrum, 
there is no need to postulate the presence of a dust compo- 
nent in the core at any significant level. A similar conclusion 
is reached by Leeuw, Samson & Robson (2000) for the el- 
liptical galaxy, NGC 4374, in which sub-mm emission from 
the AGN core is detected. Note that this does not, by itself, 
rule out the possibility of some dust in the core. Indeed, the 
presence of CO emission at the position of the core (Bridges 
& Irwin 1998, Reuter et al. 1993) suggests that dust is also 
present there. However, the core is unlikely to be dominated 
by dust emission or else the core spectral index would be sig- 
nificantly flatter or inverted. The sub-mm core is unresolved 
at 850/im but has a (deconvolved) size of ~ 4" at 450/im. 

If we use the core spectral index and extrapolate the 
core flux to 100/im, we find a 100/im core flux of 0.22 Jy. 
This is within the error bar of the total 100/im flux of 7.9 ± 
1 Jy measured by Lester et al.(1995) and suggests that the 
nonthermal core is a negligible contributor to the total flux 
at 100/im. Thus, the 100/im total flux (7.9 Jy) can be inter- 
preted as being entirely due to dust emission. This agrees 
with Lester et al.'s conclusion that less than 20% of the to- 
tal 100/im emission can be due to the nonthermal core based 
on the known variability of the core and the constancy of 
the 100/im flux. Thus, the total emission changes from be- 
ing non-thermal core-dominated to dust-dominated between 
450/im and 100/im. 



3.2 Extended Emission 

The extended sub-mm emission is much fainter than that 
associated with the core, but is seen at both frequencies 
and appears to be real. The clean and maximum entropy 
maps also show good agreement although there are some 
differences at low contour levels in the 850/im maps. Single 
epoch maps were also cleaned to confirm that the extended 
emission is present at all epochs to a level that is consistent 
with the shorter integration time per epoch and higher at- 
mospheric opacity in the July 15 observations. The extended 
emission constitutes ~ 33% of the total flux at 450/im and 
~ 11% of the total flux at 850/im. 

The 450/im core-subtracted map (Fig. Be) has a much 
higher noise level than the 850/im map, but does show sev- 
eral features which remain upon smoothing to lower resolu- 
tion and which also have counterparts in the 850/im map. 
The principle one is an elongated feature to the west of 
the core extending northwards. This will be referred to as 
the NW extension. Secondly, there are some residual peaks 
about 45" to the south-east of the core which again have a 
counterpart in the 850/im map. We will refer to this region 
as the SE peak. Another peak about 1' to the south-west 
could possibly be real, but this is not certain given its offset 
from a nearby 850/im peak. 

Extended emission is most readily seen in the lower 
noise 850/im maps (Fig. bid, e, and f). These show further 
extensions to both the east and west and "disconnected" 
features elsewhere which appear to be real, given their pres- 
ence in each frame. A fairly strong feature near the core is 
also seen to extend ~ 30" to the south-west. This latter fea- 
ture also appears to be real, though may contain a relatively 
larger error, given its proximity to the core. 

The spectral index of the total emission (Table 0) is 
flatter than that of the dominant core alone, suggesting that 



© 2000 RAS, MNRAS 000, §-^| 



SCUBA Observations of NGC 1275 7 



the spectral index of the faint extended emission may have 
an inverted (positive) spectral index. Indeed, we could mea- 
sure this spectral index in two positions, the NW extension 
and the SE peak, by first smoothing the 450/mi map to the 
850/im resolution. For the NW extension, flux measurements 
were made within a 6" diameter region centered at RA = 
03 h 19™ 45? 79, DEC = 41° 30' 57 "3, with the resulting 
spectral index found to be a = +2.6. For the SE peak, the 
measurement was made within a 6" diameter region centered 
at RA = 03 h 19 m 49? 17, DEC = 41° 30' 00 '.'3, and found 
to be a — +2.8. Thus, where the spectral index could be 
measured, the values are strongly positive. These are clearly 
dust components. 



3.2.1 Dust Temperature 

Dust is expected to be optically thin at sub-mm wave- 
lengths, and the temperature can, in principle, be deter- 
mined via S v = Q,B v {T)Q{v) (e.g. Stevens & Gear 2000), 
where S v is the flux density at frequency, u, f2 is the solid 
angle subtended by the dust, B V (T) is the Planck func- 
tion applicable to a dust temperature, T, and Q(y) is the 
frequency-dependent emissivity of the grains. In the opti- 
cally thin limit, Q(y) — ( — )' 3 , where (3 is the index of the 
emissivity law and vq is the frequency at which the dust 
becomes optically thin. If we restrict the dust temperature 
determination for NGC 1275 to regions for which fluxes are 
available at both frequencies (here, taking the NW exten- 
sion), and assume that the covering factor is the same at 
the two frequencies, we can estimate a temperature, adopt- 
ing a value for the unknown emissivity index, /3. Taking /3 = 
1.3 as was found for M 82 (Hughes, Gear, & Robson 1994; 
see also Leeuw et al. 2000), we find a dust temperature of 
19 K. For j3 = 2 (e.g. Domingue et al. 1999), we find T = 
11 K and if j3 = 1 (Lester et al. 1995), T = 29 K. Thus, for 
reasonable values of /3, the dust temperature is constrained 
from these observations to be between 10 and 30 K. 

While most other flux measurements of NGC 1275 have 



included the non-thermal core flux, we have noted (Sect. 3.1) 



that the core is a negligible contributor to the total 100/im 
flux. Thus, the 100/im point should represent dust emission 
and can also be used to constrain the dust temperature. If 
we include the 100/im flux (7.9 Jy, Lester et al. 1995) in this 
analysis, however, we find that the observed flux ratios can- 
not be reproduced within error bars using single values of 
Q, (3 and T. This suggests that (at least) a two component 
model is required. For (5 = 1.3, the 100/im/450/im ratio im- 
plies a temperature of 38 K in comparison to 19 K found 
from the 450^m/850^m value. Differences in (5 or Q are also 
possible. However, we note that there is some evidence for 
a two-temperature model from a line ratio analysis of the 
molecular gas as well (Bridges & Irwin 1998). We suspect 
that dust that is near the non-thermal core may be warmer. 



3.2.2 Dust Mass and Gas-to-Dust Ratio 

The dust mass is determined via Md = S v D 2 /\kdu B V {T)\, 
where D is the distance to the galaxy and fc<j„ is the dust 
mass absorption coefficient. We will determine the dust mass 
using the 850^m flux. This is because it has typically been 
found (e.g. Alton et al. 2000) that over 90% of observed 



850^im emission arises from cold ( ;J 20 K) dust, (consistent 
with our best temperature estimate of 19 K) and this cold 
component constitutes the bulk of galactic dust. Moreover, 
the lower noise in the 850^im images makes this a better 
choice for such a calculation. The value of k dv is rather un- 
certain and has the same frequency dependence as Q{y). 
Again, adopting j3 — 1.3 and extrapolating the mass ab- 
sorption coefficient from the 100/im value of 2.5 m 2 kg -1 
(Stevens & Gear 2000) yields k dv = 0.15 m 2 kg' 1 . (This is 
a factor of 2 higher than the value used by Dunne et al. 
2000 for their determination of dust mass functions in the 
local universe.) Applying a dust temperature of 20 K and 
using S v = 0.18 Jy (after core subtraction, see Table 2), we 
find a dust mass of Md = 5.9 x 10 7 Mq. A temperature of 
40 K would lower this by a factor of 2.5, using (3 — 2 would 
increase it by a factor of 4.5, and using the noisier 450/im 
(rather than 850/im) data would lower it by a factor of 2.6. 
Thus our best estimate of the dust mass is 6 x 10 7 Mq to 
within a factor of a few. This is the first estimate of the 
dust mass in this galaxy from spatially resolved emission 
measurements which correct for the AGN. 

This result can be compared with the much lower dust 
mass of 7.5 x 10 5 Mq (adjusted to our distance) found for 
NGC 1275 by Goudfrooij et al. (1994) using optical ab- 
sorption measurements. The absorbing dust is confined to 
a much smaller patchy region to the north-west of the core 
(Fig. |3|f) and constitutes only 1.3% of the total dust mass 
as determined from our 850/im measurements. Discrepan- 
cies between the absorbing and emitting dust masses have 
been noted by Goudfrooij & de Jong (1995) who suggested 
that the differences could be due to additional dust which 
is distributed more broadly than in front of the stellar light 
distribution alone. Our dust mass and the extended dust 
distribution seen at 850^m clearly support this view and in- 
dicate that there is much more dust in NGC 1275 seen in 
emission than previously determined from absorption mea- 
surements. 

Goudfrooij & de Jong (1995) have determined dust 
masses for 56 Shapley-Ames elliptical galaxies from IRAS 
observations, finding dust masses in the range, 10 4 to 5 x 
10 6 M . Bregman et al. (1998) have also found IRAS dust 
masses in the same range for E and SO galaxies. NGC 1275, 
therefore, has a higher dust mass, by about an order of 
magnitude, than the other early-type galaxies studied by 
these authors. However, it is well known that IRAS obser- 
vations sample somewhat warmer dust than SCUBA and 
may thereby overestimate the dust temperature, underesti- 
mating the resulting mass (Dunne et al. 2000). Therefore 
the discrepancy between the high dust mass in NGC 1275 
and the lower dust masses in E and SO galaxies may not 
be as high as an order of magnitude. Also, these galaxies 
are typically not in a cooling flow environment. Fewer pre- 
vious measurements have been made of central cooling flow 
galaxies. Edge et al. (1999), however, find masses between 
3 to 5 x 10 7 M Q for A1835 and A2390 (adjusting to H = 
75 km s _1 Mpc -1 ), in closer agreement with our value for 
NGC 1275. 

Armed with these data, we can compute a global 
gas/dust ratio. For the molecular gas mass, we use 1.6 x 
10 10 M (Bridges & Irwin 1998) which uses the Galactic 
CO to H2 conversion factor, and for the atomic hydrogen 
mass, we use the value found for the low velocity system 



© 2000 RAS, MNRAS 000. 



m 



8 Judith A.Irwin et al. 



333 MHz over Sub-mm 



(Core-subtracted) 



Halpha over Sub-mm 

(Core-subtracted) 



CM 

3 

z 
o 

< 



O 

LU 
Q 



41 31 30 
15 
00 
30 45 
30 
15 
00 

29 45 
41 31 30 

15 

00 

30 45 
30 
15 
00 



29 45 
03 19 52 





~i 1 1 1 — » r 



•«?a*\.. 



LVS-HVS 




J I L 



~i 1 1 r~, r 




(d) 



J I L_ 




A V 




50 48 46 44 52 50 48 46 44 
RIGHT ASCENSION (J2000) 



Figure 3. Contours of emission at other wavebands over greyscale of the sub-mm emission. (a) 333 MHz radio continuum (Pedlar et al. 
1990) over the core-subtracted 450/xm emission, the latter smoothed to 10" X 10" resolution. The greyscale ranges from 27 to 90 mjy 
beam" 1 and contours are at 13.5, 20, 28, 50, 100, 200, 400, 800, 1600, 3000, 5000, and 8000 mjy beam" 1 . The 333 MHz beam is 5 '.'5 
X 4 '.'9 at position angle PA = -74°. (b) Ho + [Nil] emission (Goudfrooij et al. 1994) from the LVS over the same 450/xm greyscale 
as in (a). Contours are in arbitrary units, (c) As in (a) but the 850/xm (un-smoothed) core-subtracted emission is the greyscale which 
scales from 7 to 25mjy beam -1 , (d) Contours as in (b) with the greyscale as in (c). (e) Ho emission showing the low velocity system in 
contours over the high velocity system in greyscale (Caulet et al. 1992) at the same scale as Figs, (a) through (d). The axes are in units 
of 0.406". (f) Optical absorption (A^) from Goudfrooij et al. (1994) at the same scale as Figs, (a) through (e). 



by Jaffe (1990), i.e. 5 x 10 9 M assuming a spin tempera- 
ture of 100 K. We find M(HI + H 2 )/M d ~ 360 with these 
assumptions. This is higher than typical Galactic values of 
100 to 150 (Devereux & Young 1990) but is consistent with 
what has been found for radio quiet quasars, spiral galaxies, 
and ultraluminous infra-red galaxies (« 260 to 540; Hughes 
et al. 1993). 



3.3 Correlations with other Wavebands 

Figure pi shows overlays of the core-subtracted sub-mm maps 
(in greyscale, 450/im at top and 850pm at bottom) with 
emission at other wavebands. Note that by subtracting the 
core, we cannot compare emission features within the cen- 
tral « 20" diameter region. Overlays are with the 333 MHz 
radio continuum emission (Fig. Ha, c) which traces the non- 
thermal component, and with the Ha low velocity system 
(Fig. ftp, d) which presumably traces condensing cooling 
flow gas. For comparison, the high velocity system, associ- 
ated with the infalling galaxy (Sect, hi), is shown as greyscale 



with the LVS contours in Fig. Be and the optical absorption 
is shown in Fig. Hf. 

The dust emission shows some weak correlations with 
the 333 MHz map. For example, the sub-mm NW extension 
occurs along the outer ridge of the radio continuum emission. 
There is also weak sub-mm emission just where the radio 
continuum emission falls off about 1 arcminute SW of the 
nucleus. A ridge of weak radio continuum emission extending 
from RA = 3 h 19 m 52 s , DEC = 41° 30' 55" to RA = 3 h 19 m 
49 s , DEC = 41° 29' 50" is also seen in the sub-mm. 

There is no obvious correlation between the LVS Ha 
emission and the sub-mm maps other than a possible one 
between the NW extension and an Ha extension at 450/im 
(Fig. Hb) . The HVS Hq emission (Fig. He) occurs in a patchy 
distribution mainly to the NW of and near the nucleus. 
Again, there is no clear correlation between the sub-mm 
emission and high velocity Ha system other than some cor- 
respondence in the same NW extension region. The optical 
extinction map (see Fig. Qf) does show several knots which 
correspond to enhanced 450/im or 850/im emission, suggest- 
ing that this dust is in the foreground. N0rgaard-Nielsen et 
al. (1993) and McNamara et al. (1996) have argued that 



© 2000 RAS, MNRAS 000, |^4T| 



SCUBA Observations of NGC 1275 9 



some of the dust, as observed in optical extinction, corre- 
sponds to the intervening (HVS) galaxy. 

We have also investigated whether the CO emission 
(not shown) resembles the sub-mm, since normally dust 
and molecular gas follow similar distributions. CO has been 
mapped with similar resolution as the sub-mm by Bridges 
& Irwin (1998) and Reuter et al. (1993); however, the most 
wide-scale mapping has been out to a radius of only ~ 30" 
and little structure is visible in existing CO maps. Bridges 
& Irwin do find a bend in the CO contours to the south- 
west, similar to the bright 850/im emission in that direction. 
Much higher resolution CO mapping, which cannot be di- 
rectly compared to our sub-mm distribution due to the core 
subtraction and difference in resolution, shows an east-west 
structure similar to the LVS (see Inoue et al. 1996). 

Figure shows the 850/im core-subtracted map super- 
imposed on a colour image of the ROSAT HRI X— ray emis- 
sion which shows a deficit of X— ray flux at the positions 
of the radio lobes. This morphology has been interpreted as 
a displacement of the thermal gas by the outflowing radio 
lobes (Bohringer et al. 1993, Churazov et al. 2000). Fig. § 
shows the best correlation we have found with the sub-mm 
and helps to explain the partial correlation found between 
dust and the radio continuum (see above) since there is a 
relationship between the radio continuum and X— ray emis- 
sion. The brightest X— ray emission occurs in a ridge which 
extends from the east of the nucleus and curves to the south. 
This is also roughly where a weak radio continuum ridge 
occurs (as pointed out above). The 850/im emission approx- 
imately follows this curved ridge (cf. the "filled in" emission 
in the ME map, Fig. He also) although the peaks don't ex- 
actly coincide. For example, dust emission curves around 
the X-ray peak at RA = 3 h 19 m 50 ?5, DEC = 41° 30' 
15". The bright sub-mm extension to the SW of the nu- 
cleus truncates abruptly just where the X— ray emission also 
drops about 30" from the nucleus; this is where the SW ra- 
dio lobe occurs. To the west of the nucleus, both X— ray 
and dust emission show an extension to ~ 50 arcseconds, 
both ending at roughly the same position. The NW exten- 
sion correlates well with an X— ray peak on the NW side of 
the galaxy with the sub-mm peak slightly displaced to the 
NW of the X— ray peak. Even some of the the very faint sub- 
mm emission which we at first considered to be noise bears 
some resemblance to the X— ray. For example, to the north 
of the nucleus, the northern X— ray arc has several 850/im 
peaks occurring along its northern edge about 45" from the 
nucleus. About 1 arcminute from the nucleus to the NNW, 
weaker X— ray emission falls off and again, weak peaks of 
850/im emission can be seen there. 

Recent observations by the Chandra satellite (Fabian 
et al. 2000) confirm the general characteristics of the X— ray 
map shown in Figure H. One improvement is in the clearer 
view of an X-ray absorption feature ~ 40" NW of the nu- 
cleus which is seen as a small arc just to the west of the 
brightest X-ray ridge on the NW side of the nucleus (see 
the low X-ray emission at RA ~ 3 h 19 m 45 s , DEC ~ 41° 
30' 47"). This absorption feature appears to "wrap around" 
on the west side of the brightest X-ray/sub-mm peak NW 
of the nucleus and so suggests that there may be some re- 
lation between the dust and the infalling galaxy. However, 
we do not see enhanced sub-mm emission where the X-ray 
absorption occurs. This suggests that the feature which is 



producing the X-ray absorption includes some component 
other than dust and/or that the characteristics of the ab- 
sorbing dust are such that we are not clearly detecting it at 
450/im or 850/im. 

In summary, we find that the sub-mm dust emission 
correlates spatially with the X— ray emission, except that 
there may be offsets between the X-ray peaks and sub-mm 
peaks (for example, the strongest X— ray peak does not have 
a counterpart in the sub-mm). 



4 DISCUSSION 

4.1 Source of Dust Heating 

From these SCUBA data, we have estimated a dust temper- 
ature to be ~ 20 K. What are the possible heating sources 
for this dust? The possibilities include heating by stellar 
photons, heating by hot electrons in X— ray emitting gas, 
direct heating by X— rays, and heating by the AGN. 

Since dust emission is observed to a distance of ~ 1 
arcminute = 20 kpc from the AGN in NGC 1275, we can rule 
out AGN heating as a possible source for this extended dust. 
Rigopoulou, Lawrence, & Rowan- Robinson (1996) have also 
looked at the submillimetre and X— ray data from a number 
of ultraluminous infra-red galaxies and found that most are 
well- fitted by a standard model of stellar heating, even when 
there is evidence for an AGN. Also, since the 450/im/850/im 
spectral index of the dust emission is the same for dust in 
the NW extension (measured ~ 6 kpc from the nucleus) as 
for dust at 15 kpc, there is no need to postulate different 
heating sources for the extended dust at these frequencies. 

Goudfrooij & de Jong (1995) have considered the other 
heating mechanisms for elliptical galaxies and conclude that 
direct heating by X— rays is not feasible. This leaves heat- 
ing by stellar photons and heating by hot electrons as the 
remaining viable mechanisms. McNamara et al. (1996) have 
found a faint blue continuum extending out to ~ 30 kpc 
(1.5 arcminutes) in NGC 1275 along the filamentary Ha sys- 
tem, indicating that stars do indeed exist to large distances. 
However, the close resemblance of the sub-mm continuum in 
NGC 1275 to the X— ray structure and the apparent absence 
of a large population of stars more massive than ~ 5 Mq 
(Sarazin & O'Connell 1983, Smith et al. 1992, McNamara et 
al. 1996) points to hot electrons as the source of dust heating 
in NGC 1275. Lester et al. (1995) have also shown quanti- 
tatively, that hot electrons in NGC 1275 are energetically 
capable of supplying the dust heating in this galaxy. The 
fact that the 850/im emission avoids the brightest X— ray 
peak suggests that dust may be more rapidly destroyed in 
this region. 



4.2 Origin of the Dust 

These observations have shown that extended dust exists in 
NGC 1275 to very large distances from the core (~ 20 kpc) 
and that the total dust mass is high, i.e. 6 x 10 7 Mq, two 
orders of magnitude larger than that inferred from optical 
absorption studies. As indicated in Sect, pi, three possibilities 
for the origin of the dust present themselves: from the cool- 
ing flow or in dense clouds associated with the flow, from 



© 2000 RAS, MNRAS 000, 0-|l| 



10 Judith A. Irwin et al. 



ongoing star formation, or from the merger of a gas/dust 
rich galaxy. Let us consider these possibilities. 

If dust is mixed with the cooling flow gas, then its life- 
time against destruction by sputtering is found from t^ = 
2 x 10 4 (cm _3 /n)(a/0.01/im), where n is the density of the 
hot cooling flow gas and a is the grain radius (Draine & 
Salpeter 1979). The grain size is uncertain, but a reasonable 
upper limit is 0.1 micron (cf. Lester et al. 1995), and the hot 
gas density near the center of NGC 1275 is 0.05 - 0.1 cm -3 
(e.g. Fabian et al. 2000). This gives t d = 2 - 4 x 10 6 years. 
Thus, dust will be destroyed on short timescales if it is mixed 
with the hot gas and there must be a source of dust replen- 
ishment. The alternative is that dust is somehow shielded 
from the X— ray environment, possibly in dense molecular 
clouds for which there is (as yet) no observational evidence. 

If stellar mass loss is continuously replenishing the dust, 
we would expect this dust to be ejected and dispersed into 
the X— ray emitting environment, suggesting that the short 
lifetime determined above should be relevant. Determining 
a dust mass loss rate from stars is not straightforward and 
involves a knowledge of the initial mass function, the mass 
loss rates from stars, and the gas/dust ratio in the ejected 
material. For NGC 1275, there currently appear to be no 
stars more massive than 5 Mq and, if more massive stars 
have been produced in the past, this past massive star for- 
mation must have ceased 70 Myr ago or more (Smith et al. 
1992). Since this is longer than the lifetime of dust in the 
X-ray emitting environment, we do not expect young, mas- 
sive O type stars to contribute to the currently observed 
dust in NGC 1275. The next strongest source of mass loss, 
and that which will dominate in NGC 1275, comes from the 
older, more evolved stellar population. The remaining main 
sequence stars of 5 Mq or less contribute negligibly to global 
mass loss rates (see Snow 1982). 

Both Faber & Gallagher (1976) and Knapp, Gunn 
& Wynn- Williams (1992) have estimated, through in- 
dependent means, mass-loss rates for an evolved pop- 
ulation of stars. Knapp et al. find Al = 2.1 x 
10" 12 (L 2 .2^m/L e ) Moyr' 1 and Faber & Gallagher find 
M = 1.5 x 10" 11 (Lb/Lq) Mq yr' 1 , where L 2 .2 M m and 
Lb are, respectively, the 2.2/xm and blue luminosities. Us- 
ing (from the NASA Extragalactic Database) the total cor- 
rected blue magnitude of 12.15, we find M = 0.81 Mq yr~ l 
and using the 2.2/im flux density of 159 mjy, we find 
M — 0.62 Mq yr~ x . The former value should be an over- 
estimate since it includes any contribution from the AGN 
and the latter should be an underestimate since the 2.2/im 
flux density is measured only within a central 32" aper- 
ture. Taking a mass loss rate of 0.7Mq yr~ x , then, over 2 
- 4 x 10 6 years, stars can contribute, at most, 2.8 x 10 6 Mq 
of gas to t he IS M in NGC 1275. With a gas/dust ratio of 
360 (Sect. 3.2.2), the dust mass generated by stars would 
be 8 x 10 3 Mq , which is 4 orders of magnitude lower than 
observed. It is possible that 2 orders of magnitude could be 
gained by allowing dust to survive over a timescale one order 
of magnitude longer, since then there could be a contribu- 
tion from hot massive stars which typically have higher mass 
loss rates than the older stars. However, it is unlikely that 
4 orders of magnitude could be gained and a dust shielding 
mechanism would then also be required. Thus, we conclude 
that dust cannot be replenished by stellar mass loss. Lester 
et al. (1995) came to the same conclusion for NGC 1275 and 



Goudfrooij & de Jong (1995) have also found that most el- 
liptical galaxies in which X— ray emission has been detected 
contain more dust than can be accounted for by stellar mass 
loss. Moreover, if the observed dust originated from stellar 
mass loss, we would expect the dust distribution to bear 
some resemblance to the stellar light distribution, which is 
not observed. 

An obvious source of dust is the infalling galaxy. Goud- 
frooij et al. (1994), in fact, make a case for all dust in ellipti- 
cal galaxies being supplied by external systems, and others 
(e.g. Pinkney et al. 1996, Koekemoer et al. 1999) have also 
supported the merger/accretion origin for dust in ellipticals. 
McNamara et al. (1996) and N0rgaard-Nielsen et al. (1993) 
have suggested, for NGC 1275, that two dust systems may be 
present, one associated with NGC 1275 and one associated 
with the the high velocity system. Martel et al. (1999) also 
find dust absorption at the north-west location of the HVS. 
These results find some support from our observations since 
the HVS is observed to the NW of the core of NGC 1275, 
as is the NW sub-mm extension. Thus, the infalling galaxy 
appears to contain dust and, one could argue, is a source of 
dust. 

Several lines of argument support this view. There is ac- 
cumulating evidence that the infalling galaxy is currently in- 
teracting with NGC 1275 (Hu et al. 1983; Unger et al. 1990; 
Caulet et al. 1992; N0rgaard-Nielsen et al. 1993; Ferruit et 
al. 1998). The HVS gas occurs in a patchy distribution over 
~ 20 kpc, in projection (Fig. Br), arguing for a broad in- 
teraction region (though the region containing dust is larger 
still by about a factor of 2) . The interaction timescale (im- 
pulsive approximation) is roughly tj ~ d/v, where d is the 
size of the target galaxy and v is the relative velocity of the 
perturber. If we take the size to be the entire projected re- 
gion over which dust emission is detected (~ 38 kpc) and 
v = 3000 km s _1 , we find t; ~ 1 x 10 7 yrs. This is a fac- 
tor of at least 2 higher than the dust destruction timescale, 
but this may not be largely discordant, given the approx- 
imation made. Finally, our dust mass (6 x 10 r Mq) and 
gas-to-dust ratio (360) are quite typical of spiral and IRAS 
galaxies, and our dust temperature (20 K) is also typical of 
spiral galaxies though somewhat cooler than that observed 
in IRAS galaxies (Alton et al. 2000, Dunne et al. 2000), ar- 
guing for an infalling spiral galaxy as the origin of the dust. 
In this scenario, the dust distribution would resemble the 
X-ray gas distribution (Fig. |4|) because the dust is widely 
distri but ed and is being collisionally heated as discussed in 
Sect. 4.1. The implication is that there should be additional 
cold dust, as yet undetected, which is widely distributed in 
NGC 1275. A difficulty with this scenario is that externally 
supplied dust, initially moving at 3000 km s _1 with respect 
to NGC 1275, must "settle" into the rest frame of NGC 1275 
while the hot stars and associated Ha emitting gas do not. A 
theoretical development involving an infalling spiral galaxy 
which also interacts with cooling flow gas should be brought 
to bear on this question. 

A final interesting idea is that the dust forms directly 
in the cooling flow (Fabian et al. 1994, Hansen et al. 1995), 
which would explain the observed spatial correlation di- 
rectly. Fabian et al. suggest that dust could form within 
clouds, even in the absence of dust in the cooling flow origi- 
nally, and would be shielded against destruction, giving it a 
longer lifetime than the above-determined value. The tem- 



© 2000 RAS, MNRAS 000, §-^| 



SCUBA Observations of NGC 1275 11 




29 45 hf ft 



02 19 54 



RIGHT ASCENSION (J 2000) 

Figure 4. Core-subtracted 850/jm contours (as in Fig. 2f) over a colour image of the X— ray emission taken from the ROSAT HRI 
archive (observer, W. Voges, 1994). The colours range (lowest to highest intensity) from grey through red, yellow, green, blue, violet, 
and black. The low intensity "holes" north and south-west of the nucleus are filled with radio continuum emission (see Fig. 3c); for a 
similar X— ray image, see Bohringer et al. (1993). 



perature of the dust grains is expected to be is very low, 
with most of the dust at less than 6 K and some as high as 
11 K. However, as the clouds become disturbed in the inner 
part of the cooling flow, external heating should make the 
dust more visible (and destroy the dust on short timescales 
again). Since the LVS is known to be associated with the 
cooling flow as well, however, we would also expect the 
dust (unless it is rapidly destroyed) to correlate with the 
LVS in this scenario, for which there is no strong evidence. 
Thus, this model seems less attractive, although should not 
be ruled out pending further work on disentangling density 
from temperature effects in the cooling flow region. 



5 CONCLUSIONS 

We have made SCUBA observations of the cD galaxy, 
NGC 1275, which is at the center of the Perseus Cluster 
cooling flow. Emission has been detected at both 850/im 
and 450/im and, applying standard radio astronomical tech- 
niques such as CLEAN and Maximum Entropy to these 
sub-mm maps, we have removed structures related to the 
beam shape in order to reveal real low level emission. The 
resulting sub-mm distribution can be easily separated into a 
bright central core and fainter extended emission, allowing 
us to determine the spectral indices of the core and extended 
emission separately. 

The core has a steep spectral index (-0.88) suggest- 



© 2000 RAS, MNRAS 000, 0-|l| 



12 Judith A. Irwin et al. 



ing that it can be explained primarily in terms of a high- 
frequency extension of the non-thermal AGN emission. The 
extended emission has a positive spectral index (+2.7), and 
is clearly due to dust which extends as far as 20 kpc from 
the nucleus. We estimate the dust temperature to be 20 K 
(though 10 K to 30 K is possible, depending on the assumed 
value of the dust emissivity index) and rule out the AGN 
as the main heating source, given the presence of dust far 
from the nucleus. The dust mass is 6 x 10 7 Mq (within 
a factor of a few); this dust mass is higher by 2 orders of 
magnitude than that inferred from optical extinction. Thus 
there is much more dust in this galaxy than previously re- 
alized. The gas-to-dust ratio is found to be 360. Note that 
the latter value is derived using an assumption of a Galactic 
CO to H2 conversion factor as well as a 100 K spin temper- 
ature for neutral hydrogen. Nevertheless, the derived dust 
mass, gas-to-dust ratio, and dust temperature are all typical 
of values being found for spiral galaxies. 

A remarkable correlation between the sub-mm dust 
emission and the X— ray emission is observed in this galaxy, 
although there are some displacements in emission peaks. 
When the various sources of dust heating are considered, 
the strongest case can be made for collisional heating by 
hot X— ray emitting electrons. This would explain the spa- 
tial correlation between the two components. Since the grain 
destruction timescale is short (~ 4 x 10 6 yrs), continuous 
dust replenishment is required to account for the sub-mm 
emission. We have ruled out stars as the source of dust and 
suggest instead that the infalling galaxy is the most likely 
candidate, given that the observed dust parameters are typ- 
ical for spiral galaxies. Thus, we have a picture in which 
the infalling galaxy is interacting with NGC 1275 over a 
large region, supplying it with dust which is then heated 
most effectively by collisions with hot electrons in the cool- 
ing flow gas. If this scenario is correct, then more cold dust 
should be present in NGC 1275 than we have detected. The 
alternative is that dust forms directly in the cooling flow 
and is shielded from destruction. Further spatially resolved 
multi- wavelength observations should be able to distinguish 
between these scenarios by disentangling density from tem- 
perature effects. 



ACKNOWLEDGMENTS 

This research has made use of data obtained through the 
High Energy Astrophysics Science Archive Research Cen- 
ter Online Service, provided by the NASA/Goddard Space 
Flight Center. We are grateful to A. Pedlar and P. Goudfrooij 
for providing FITS images of their published data and also 
to Dr. A. Caulet for her assistance. 



REFERENCES 

Allen S.W., 2000, MNRAS, 315, 269 

Allen S. W., Fabian A. C, 1997, MNRAS, 286, 583 

Allen S.W., Fabian A.C., Johnstone R.M., Arnaud K.A., Nulsen 

P.E.J. , 1999, astroph/9910188 
Alton P. B., Xilouris E. M., Bianchi S., Davies J., Kylafis N. 2000, 

A&A, 356, 795 
Annis J., Jewitt D., 1993, MNRAS, 264, 593 
Arnaud K.A., Mushotzsky R.F. 1998, ApJ, 501, 119 



Bohringer H., Voges W., Fabian A. C, Edge A. C, Neumann D. 

M., 1993, MNRAS, 264, L25 
Bregman J. N., Snider B. A., Grego L., Cox C. V. 1998, ApJ, 

499, 670 
Bridges T. J., Irwin J. A., 1998, MNRAS, 300, 967 
Briggs S.A., Snijders M.A.J., Boksenberg A., 1982, Nature, 300, 

336 
Carlson M. N. et al. 1998, AJ, 115, 1778 
Caulet A., Woodgate B.E., Brown L.W., Gull T.R., Hintzen P., 

Lowenthal J.D., Oliversen R.J., Ziegler M.M., 1992, ApJ, 388, 

301 
Churazov E., Forman W., Jones C, Bohringer H., 2000, A&A, 

356, 788 
Clark B. G., 1980, A&A, 89, 377 
Cornwell T. J., Evans K. F., 1985, A&A, 143, 77 
Cox C.V., Bregman J.N., Schombert J.M., 1995, ApJS, 99, 405 
Crawford C.S., Allen S.W., Ebeling H., Edge A.C., Fabian A.C., 

1999, MNRAS, 306, 857 
Devereux N. A., Young J. S., 1990, ApJ, 359, 42 
De Young D.S., Roberts M.S., Saslaw W.C. 1973, ApJ, 185, 809 
Domingue D. L., Keel W. C, Ryder S. D., White III R. E., 1999, 

AJ, 118, 1542 
Donahue M., Mack J., Voit G.M., Sparks W., Elston R., Maloney 

P.R., 2000, astroph/0007062 
Draine B.T., Salpeter E., 1979, ApJ, 231, 77 
Dunne L., Eales S., Edmunds M., Ivison R., Alexander P., 

Clements D. L., 2000, MNRAS, 315, 115 
Edge A.C., Ivison R.J., Smail I., Blain A.W., Kneib J.-P., 1999, 

MNRAS, 306, 599 
Faber S. M., Gallagher J. S. 1976, ApJ, 204, 365 
Fabian A.C., 1994, ARA&A, 32, 227 
Fabian A.C., Johnstone R.M., Daines S.J., 1994, MNRAS, 271, 

737 
Fabian A.C., Hu E.M., Cowic L.L., Grindlay J. 1981, ApJ, 248, 

47 
Fabian A.C., Sanders J.S., Ettori S., Taylor G.B., Allen S.W., 

Crawford C.S., Iwasawa K., Johnstone R.M., Ogle P.M., 2000, 

MNRAS, 318, 65 
Ferruit P., Adam G., Binette L., Pecontal E., 1998, NewA, 2, 345 
Gear W.K., Robson E.I., Gee G., Nolt I.G., 1985, MNRAS, 217, 

281 
Goudfrooij P., de Jong T., 1995, A&A, 298, 784 
Goudfrooij P., Hansen L., j0rgensen H. E., N0rgaard-Nielsen H. 

U., 1994, A&AS, 105, 341 
Hansen L., j0rgensen H. E., N0rgaard-Nielsen H. U. 1995, A&A, 

297, 13 
Hansen L., j0rgensen H.E., N0rgaard-Nielsen H.U., Pedersen K., 

Goudrooij P., Linden- Vornle M.J.D., 2000, A&A, 362, 133 
Holland W. S., et al. 1999, MNRAS, 303, 659 
Hu E.M., 1992, ApJ, 391, 608 
Hu E.M., Cowie L.L., Kaaret P., Jenkins E.B., York D.G., Roesler 

F.L., 1983, ApJ, 275, 27 
Hughes D. H., Gear W. K., Robson E. I., 1994, MNRAS, 270, 641 
Hughes D. H., Robson E. I., Dunlop J. S., Gear W. K., 1993, 

MNRAS, 263, 607 
Inoue, M. Y., Kameno, S., Kawabe, R., Inoue, M., Hasegawa, T., 

& Tanaka, M. 1996, AJ, 111, 1852 
Jaffe W. 1990, A&A, 240, 254 

Jaffe W., Bremer M.N., van der Werf P.P., 2000, astroph/0009418 
Johnstone R.M., Fabian A.C., Taylor G.B. 1998, MNRAS, 298, 

854 
Koekemoer A.M., O'Dea C.P., Sarazin C.L., McNamara B.R., 

Donahue M., Voit G.M., Baum S.A., Gallimore J.F. 1999, 

ApJ, 525, 621 
Knapp G. R., Gunn J. E., Wynn- Williams C. G. 1992, ApJ, 399, 

76 
Krabbe A., Sams III B.J., Genzel R., Thatte N., Prada F., 2000, 

A&A, 354, 439 



© 2000 RAS, MNRAS 000, §-^| 



SCUBA Observations of NGC 1275 13 



Leeuw L. L., Sansom A. E., Robson E. I., 2000, MNRAS, 311, 

683 
Lester D.F., Zink E.C., Doppmann G.W., Gaffney N.I., Harvey 

P.M., Smith B.J., 1995, ApJ, 439, 185 
Lynds R. 1970, ApJ, 159, 151 
Martel A. R., et al., 1999, ApJS, 122, 81 
McNamara B. R., O'Connell R. W., Sarazin C. L. 1996, AJ, 112, 

91 
Minkowski R. 1957, in Proceedings from J^th IAU Symposium, 

edited by H.C. Van de Hulst, CUP, pg. 107 
N0rgaard-Nielsen H. U., Goudfrooij P., j0rgensen H. E., Hansen 

L. 1993, A&A, 279, 61 
O'Dea C.P., Baum S.A., Gallimore J.F. 1994, ApJ, 436, 669 
Pedlar A., Ghataure H.S., Davies R.D., Harrison B.A., Perley R., 

Crane P.C., Unger S.W., 1990, MNRAS, 246, 477 
Pinkney J., et al., 1996, ApJ, 468, L13 
Reuter H.-P., Pohl M., Lesch H., Sievers A. W. 1993, A&A, 277, 

21 
Sandell G. 1997, "The SCUBA mapping 

cookbook. A first step to proper map reduction" . available at 



http://www.jach.hawaii.edu/JACpublic/stardocs/scll.htx/scll.html 



Sarazin C. L., O'Connell R. W. 1983, ApJ, 268, 552 

Seyfert C.K. 1943, ApJ, 97, 28 

Smith E. P. et al. 1992, ApJ, 396, L49 

Snow T. P. Jr. 1982, ApJ, 253, L39 

Stevens J. A., Gear W. K. 2000, MNRAS, 312, L5 

Stevens J. A., Litchfield S. J., Robson E. I., Hughes D. H., Gear 

W. K., Terasranta H., Valtaoja E., Tornikoski M., 1994, ApJ, 

340, 150 
Unger S.W., Taylor K., Pedlar A., Ghataure H.S., Penston M.V., 

Robinson A. 1990, MNRAS, 242, 33 
Voit M., Donahue M. 1995, ApJ, 452, 164 
White D.A., Fabian A.C., Johnstone R.M., Mushotzsky R.F., Ar- 

naud K.A. 1991, MNRAS, 252, 72 

This paper has been produced using the Royal Astronomical 
Society/Blackwell Science ffl^K style file. 



© 2000 RAS, MNRAS 000, §-|l|