The FIR properties of the most isolated
galaxies
U. Lisenfeld1,2 , L. Verdes-Montenegro2, J. Sulentic3 , S. Leon4 , D. Espada2 ,
G. Bergond2 , E. Garcı́a2, J. Sabater2 , J.D. Santander -Vela2 , and S.
Verley2,5,6
1
2
3
4
5
6
Departamento de Fı́sica Teórica y del Cosmos, Facultad de Ciencias,
Universidad de Granada, Spain ute@ugr.es
Instituto de Astrofı́sica de Andalucı́a (IAA/CSIC), Apdo. 3004, 18080 Granada,
Spain
Department of Astronomy, University of Alabama, Tuscaloosa, USA
Instituto de Radioastronomı́a Milimétrica (IRAM), Avda. Divina Pastora 7,
local 20, 18012 Granada, Spain
LERMA – Observatoire de Paris, 61 avenue de l’Observatoire, 75014 Paris,
France
INAF-Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, 50125 Firenze, Italy
Summary. We describe the mid- (MIR) and far- (FIR) infrared properties of a
large sample of the most isolated galaxies in the local Universe. This sample is
intended as a “nurture-free” zero point against which more environmentally influenced samples can be compared. We reprocess IRAS MIR/FIR survey data using
the ADDSCAN/SCANPI utility for 1030 out of 1050 galaxies from the Catalogue of
Isolated Galaxies (CIG) as part of the AMIGA project. The distribution of log(LFIR )
is sharply peaked from 9.0–10.5 with very few (<2%) galaxies above 10.5. The optically normalised luminosity diagnostic R = log(LFIR /LB ) shows a distribution
sharply peaked between 0.0 and −1.0. These results were compared to the magnitude limited CfA sample that was selected without environmental discrimination.
This modestly (e.g. compared to cluster, binary galaxy and compact group samples)
environmentally affected sample shows significantly higher mean log(LFIR ) and R,
whereas the mean log(LB ) is the same. Our sample shows a strong LFIR vs. LB
1.41
correlation, with a slope steeper than one (LFIR ∝ LB
). Interacting galaxies were
found above this correlation, showing an enhancement in LFIR . The results indicate
that the FIR emission is a variable enhanced by interaction, and that our sample
probably shows the lowest possible mean value. This attests to the utility of our
sample for defining a nurture-free zero point.
1 The AMIGA project
Although it is widely accepted that galaxy interactions can stimulate secular
evolutionary effects in galaxies (e.g. enhanced star formation, morphological
2
Lisenfeld et al.
peculiarities including transitions to earlier type, active nuclei) there are still
many open questions. Studies aimed at quantifying the level of interaction
enhancement have even produced contradictory results; e.g. some studies of
interacting pairs find a clear star formation enhancement ([1];[2]) while others
find only a marginal increase ([3]). Much of this uncertainty reflects the lack
of a statistically useful baseline. What is the amplitude and dispersion in a
given galaxy property that can be ascribed to “nature”?
The AMIGA project (Analysis of the interstellar Medium of Isolated
GAlaxies) involves the identification and study of a statistically significant
sample of the most isolated galaxies in the local Universe. Our goal is to
quantify the properties of different phases of the interstellar medium in these
galaxies which are likely to be least affected by their external environment. We
adopted the Catalogue of Isolated Galaxies (CIG: [4]; [5]), including 1051 objects, as a base sample. All CIG galaxies are part of the Catalogue of Galaxies
and Clusters of Galaxies providing reasonably uniform apparent magnitude
measures with mpg < 15.7 and δ > −3 deg. Redshifts are now virtually complete for this sample with only one of the compiled objects recognised as a
Galactic source (CIG 781 ≡ Pal 15) reducing the working sample to n = 1050
objects. AMIGA is compiling data that will characterise all phases of the
ISM: blue magnitude, mid- and far-infrared, Hα, and radio continuum fluxes,
as well as the emission of the atomic gas (HI) and of carbon monoxide (CO),
as a tracer of the molecular gas. The data are being released and periodically
updated at http://www.iaa.es/AMIGA.html.
Previous AMIGA papers evaluated, refined and improved the sample in
different ways including: 1) revised positions ([7]), 2) sample redefinition, magnitude correction and full-sample analysis of the Optical Luminosity Function
(OLF) ([11]: Paper I) and 3) morphological revision and type-specific OLF
analysis ([9]: Paper II).
In the study presented here, we analysed basic mid- (MIR) and far-infrared
(FIR) properties using data from the IRAS survey. This article is a short summary of a journal paper (The AMIGA sample of isolated galaxies III. IRAS
data and infrared diagnostics, U. Lisenfeld et al., 2006, A&A, in press) which
contains a more detailed description of the methods and a more extensive
analysis of the data.
2 Reprocessing of IRAS data and data analysis
We reprocessed the IRAS data for 1030 galaxies covered by the IRAS survey
using the utility SCANPI/ADDSCAN provided by the Infrared Processing
and Analysis Center (IPAC). SCANPI/ADDSCAN is a one-dimensional tool
that coadds calibrated IRAS survey data making use of all scans that passed
over a specific position. It is 3-5 times more sensitive than the IRAS PSC
since it combines all survey data and is therefore more suitable for detection
of the total flux from slightly extended objects.
The FIR properties of the most isolated galaxies
3
For the data analysis we restricted the sample to the optically selected
subsample (with corrected Zwicky magnitudes between 11 and 15 mag) which
was found in Paper I to be 80% complete. We exclude 32 galaxies which were
found in paper II to possibly violate the isolation criterion, and two nearby
dwarf ellipticals for which only (very low) upper limits for LFIR were available.
We were left with a sample of 719 galaxies for which redshift data was available
for 701. Hereafter we refer to this sample as the AMIGA (FIR) sample.
For the statistical calculations we used survival analysis that takes upper
limits into account. We used the ASURV package which is described in detail
in [6].
3 FIR luminosity distribution
Fig. 1. The FIR luminosity distribution of the AMIGA sample. The full line shows
the distribution calculated with ASURV, the shaded area shows all galaxies detected
at both 60 and 100 µm, and the dashed line gives the non-detections.
Fig. 1 shows the distribution of FIR luminosity of the AMIGA sample.
The distribution peaks in the bin log(LFIR /L ) = 9.5–9.75 with the ASURV
estimated mean log(LFIR /L ) = 9.15. Practically all the galaxies have FIR
luminosities betwen log(LFIR /L ) = 7.5 and log(LFIR /L ) = 11.25. It is remarkable that the bulk of the FIR luminosities (98%) lies below log(LFIR /L )
= 10.5.
In order to judge whether the low FIR luminosities of our sample were
normal or exceptional, we compared the distribution of the FIR luminosity
4
Lisenfeld et al.
Fig. 2. Left: The percentage FIR luminosity distribution for the FIR detections in
the AMIGA sample restricted to (uncorrected) mzw ≤ 14.5 mag (shaded area)
and the corresponding for the CfA sample (dashed line). Right: The same for
R =log(LFIR /LB ).
of and R=log(LFIR/LB ) to that of the galaxy sample of the Center of Astrophysics (CfA), whose FIR properties, based on data of the IRAS FSC were
studied in [10] and [8]. The CfA sample consists of 2445 galaxies representing
a complete flux-limited sample (mzw ≤ 14.5) selected in Galactic coordinates,
but without any selection with respect to environment. In order to properly
compare the two data sets, we applied the same magnitude cutoff as in [10] to
our sample and made sure that the distances were derived in the same way. As
a test to find possible systematic differences we compared the distances, LFIR ,
LB and R for those galaxies that are in common in both samples (n = 98)
and found excellent agreement.
In Fig. 2 (left) we show a comparsion of our distribution of log(LFIR ) to
that of the CfA sample. Above log(LFIR /L ) = 10.2 a clear excess of CfA
galaxies in comparison to our sample is visible. The mean values differ by
3σ. We performed statistical two-sample tests in the package ASURV and
found that the two distributions were different with a probability of > 97%.
Therefore, there is strong statistical evidence that the AMIGA sample has a
lower LFIR than the CfA sample which is comparable but not selected with
respect to environment. This suggests that the FIR luminosity is a variable
driven by interaction.
Fig. 2 (right) shows the comparsion of R. We notice that the mean value
of R is higher for the CfA sample than for the AMIGA sample by 4σ. Again,
statistical tests confirm the significance of this difference. On the other hand,
the distribution of LB of both samples is very similar, confirming that the
difference is due to an enhancement of LFIR in the CfA sample.
The FIR properties of the most isolated galaxies
5
4 Correlation between LB and LFIR
Fig. 3. LFIR vs. LB for the optically selected sample. The full line indicates the
best-fit bisector slope derived with ASURV, the dotted line shows the result of the
regression adopting LB , and the dashed line adopting LFIR (dashed) as independent
variable.
Fig. 3 plots LFIR vs. LB of the AMIGA sample. The best-fit bisector
1.41±0.02
regression, calculated with ASURV, yielded LFIR ∝ LB
. We found no
noticeable relation between the slope of the regression and the Hubble type.
In Fig. 4 we show the correlation between LFIR and LB for different interacting samples. One sample consists of galaxies originally included in the
CIG, but rejected after a visual inspection using POSS2/SDSS as possibly
interacting (see Sulentic et al. 2006). These galaxies have been excluded from
the present effort to characterise the isolated sample but offer a useful internal
comparison sample. Regression analysis yields a steeper slope for this subsam1.52±0.12
ple (LFIR ∝ LB
), and we see that they clearly lie above the regression
line for isolated CIG galaxies. The same is the case for two interacting samples of Perea et al. (1997), most clearly their strongly interacing sample. This
shows that interaction enhances the FIR luminosity of galaxies.
Acknowledgements
We would like to thank M. Sauvage for making his data available to us. This
work has been partially supported by DGI Grant AYA 2005-07516-C02-01 and
6
Lisenfeld et al.
Fig. 4. LFIR vs. LB for various interacting subsamples. The filled squares and arrows
denote the galaxies from the CIG showing signs of interaction. The dashed line is
the regression fit to this subsample. The triangles indicate strongly interacting, and
the crosses weakly interacting galaxies from Perea et al. (1997). The full line is the
fit to the total AMIGA sample from Fig. 3.
the Junta de Andalucı́a. UL acknowledges support by the research project ESP
2004-06870-C02-02.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Bushouse, H. A., ApJ, 320, 49 (1987)
Bushouse, H. A., Lamb, S. S., & Werner, M. W., ApJ, 335, 74 (1988)
Bergvall, N., Laurikainen, E., & Aalto, S., A&A, 405, 31 (2003)
Karachentseva, V. E., Comm. Spec. Ap. Obs., USSR, 8, 1 (1973)
Karachentseva, V. E., Lebedev, V. S., & Shcherbanovskij, A. L., Catalogue of
Isolated Galaxies. Bull. Inf. CDS, 30, 125 (1986)
La Valley, M., Isoble, T., & Feigelson, E. D., in ASP Conf. Ser. 25, ADASS I.,
eds. D.M. Worrall, C. Biemesderfer, & J. Barnes (San Francisco: ASP 1992),
245
Leon, S., & Verdes-Montenegro, L., A&A, 411, 391 (2003)
Sauvage, M., & Thuan, T. X., ApJ, 429, 153 (1992)
Sulentic, J. W., Verdes-Montenegro, L., Bergond, G., et al., A&A, 449, 937
(2006) (Paper II)
Thuan, T. X., & Sauvage, M. 1992, A&AS, 92, 749 (1992)
Verdes-Montenegro, L., Sulentic, J., Lisenfeld, U., et al., A&A, 436, 443 (2005)
(Paper I)