A multiwavelength study of the IRAS Deep Survey galaxy sample III. Spectral classification and dynamical properties
aa r X i v : . [ a s t r o - ph . C O ] N ov Astronomy & Astrophysics manuscript no. bettoni˙ref˙rev10 c (cid:13)
ESO 2018December 13, 2018
A multiwavelength study of the IRAS Deep Survey galaxy sample
III. Spectral classification and dynamical properties
D. Bettoni , P. Mazzei , and A. della Valle , INAF - Osservatorio Astronomico di Padova, Vicolo dell’Osservatorio, 5 -35122 Padova-ITALYe-mail: [email protected], [email protected] INAF - Osservatorio Astronomico di Bologna, Via Ranzani, 1 - 40127 Bologna-ITALYe-mail: [email protected]
ABSTRACT
Context.
The infrared deep sample (IDS), in the north ecliptical polar region (NEPR), is the first complete, far–IRselected sample, on which numerous studies of galaxy evolution are based. Such a sample allows direct investigation ofthe evolution of dusty galaxies up to a redshift of about 0.3, where the global star formation rate is known to evolvevery fast. As discussed in previous papers, we performed optical and IR (ISOCAM, 15 µ m,) follow-up of its galaxies andexploited our IR observations to correct the 60 µ m fluxes for confusional effects and observational biases. In them wefound indications of a significant incompleteness of IDS sample below S(60) ≃
80 ˙mJy. We constructed 15 µ m and far-IR(60 µ m) luminosity functions of a complete sample of 56 ISO/IRAS sources. Aims.
Here we present and analyze the spectral classification of several galaxies in the IDS sample together with rotationcurves which allow estimating the lower mass limits of a subsample of objects.
Methods.
We measured fluxes and intensity ratios of the emission lines in the visible region of the spectrum ( λ − A ) for 75 galaxy members. Moreover, for some of them (55%), the spectra obtained with the Keck II telescopehave sufficient wavelength and spatial resolution to derive their rotation curve. Results.
These galaxies turn out to be disk like systems, with a high fraction ( ∼ Conclusions.
The rest-frame FIR luminosity distribution of these galaxies spans the same range as that of the FIRselected complete sample, i.e. three orders of magnitude, with the same mean value, log( L F IR )=10.2. This emphasizesthat such galaxies represent FIR properties of the whole sample well. Moreover, their optical properties are typical ofthe sample itself since 62% of these belong to the 60 µ m selected complete sample. Key words.
Galaxies: kinematics and dynamics – Galaxies: fundamental parameters
1. Introduction
One of the most useful tools for studying the physical con-ditions in galactic nuclei is the analysis of the interstel-lar medium. The warm ionized gas ( ∼ K), for exam-ple, is present in the nuclei of nearby galaxies of all mor-phological types. Detected through optical emission lines,this component of the interstellar medium has served asa diagnostic of the physical conditions in galactic nuclei(Ho et al. , 1993), including their excitation source andchemical abundances, and it allows spectral classificationto define the star-forming LINERS and AGN regions. Avariety of optical emission-line ratio diagnostics have beenpresented and employed to determine metallicities, abun-dances and star formation rates (SFR). In particular, us-ing [NII] λ α - [OIII] λ β and [SII] λ α - [OIII] λ β ratios diagnostics brings out the sep-arations among the three classes (Baldwin et al. , 1981;Veilleux & Osterbrock , 1987). The separation of extra-galactic objects according to their primary excitation mech-anisms, together with their photometric and kinematical Send offprint requests to : D. Bettoni properties, leads to a more complete view of their struc-ture and evolution. The spatially extended resolved emis-sion lines provide powerful tracers of the kinematics of agalaxy. These kinematical data are very important becausethey allow us to derive dynamical masses and to comparethem with those obtained from models. In this view, thestudy of the dynamics of dusty galaxies in the very localuniverse gives more insight into their evolution.Far-IR galaxies are dusty objects where a mixture ofongoing AGN activity and star formation play their roles.They are mostly interacting and merging gas-rich spirals(Le Floc’h et al., 2005) and may represent a link in an evo-lutionary sequence from spiral galaxies to ellipticals, viamergers. With this connection to early galaxy formationand evolution, the local population can be thought of asa laboratory for studying processes in detail such as SFtriggering and starburst vs. AGN interplay. Recent stud-ies suggest many AGNs in IR luminous and massive galax-ies (Treister et al , 2010), and Melbourne et al. (2008) haveshown that most of dusty galaxies at z ∼ scans of the north ecliptic polar region and representingmore than 20 hours of integration time, is one of the best-suited samples. Mazzei et al. (2001) exploited ISOCAMobservations (range 12-18 µ m) of 94 IRAS Deep Survey(IDS) fields (Aussel et al. , 2000), centered on the nomi-nal positions of IDS sources, to correct the 60 µ m fluxesfor confusion effects and observational biases. They foundindications of a significant incompleteness of the IDS sam-ple below S(60) ≃
80 mJy. In della Valle et al. (2006), thefirst paper of this series (Paper I in the following), we pre-sented spectroscopic and optical observations of candidateidentifications of our ISOCAM sources, and the redshiftdistribution of the 60 µ m complete subsample defined byMazzei et al. (2001) comprising 56 sources. In the secondpaper, Mazzei et al. (2007) (Paper II in the following), wederived the 60 µ m luminosity function (LF) and the poorlyknown 15 µ m LF with the bivariate method of such a sam-ple.The optical data, together with IR and far-IR (FIR)ones, are very useful for an integral view of the propertiesof galaxies in our sample. In particular we will derive thespectral energy distribution (SED) extended over severalorders of magnitude in wavelength to analyze the evolutionof these dusty galaxies over several Gyr in look-back time,i.e., over an interval of time in which the average SFR inthe Universe is known to evolve strongly (in prep).In this paper we give new spectroscopic observations of75 galaxies and classification of 42 galaxies, thus extendingour previous far-IR (FIR) data to the optical, an essentialstep, in getting a complete view of these galaxy properties.Moreover, dynamical parameters are derived for 41 galax-ies in the sample. The plan of the paper is the following.In Sections 2 we summarize our spectroscopy runs, whichare fully described in Paper I, and present the more im-portant spectral corrections. In Section 3 we analyze thespectral properties of our sample inside a fixed physical ra-dius of 3 Kpc, to derive their general classification. Section4 focuses on the analysis of dynamical parameters and therotation curves available for 41 and 31 of them, respec-tively. This analysis allowed us to derive mass lower lim-its for 39 galaxies in the sample. Finally Section 5 givesour conclusions. Here we adopt Λ=0.7, Ω =0.3 and H =70km s − Mpc − .
2. Spectroscopic observations and data analysis
A program of optical imaging and spectroscopy was under-taken to observe our sample. We acquired B and R imagesand low-resolution spectra of all the objects (106) iden-tified in Aussel et al. (2000). Our targets were observedin different runs in the years 2000 – 2003 with telescopesand instruments whose set up is given in Table 1 (see alsoTable C.1 in Paper I). Sixty-five IRAS fields were observedand spectra of 81 ISOCAM sources taken. In general theslit was oriented along the apparent major axis of the ob-ject. In those fields where a double component is visible,as discovered in our previous works (Aussel et al. (2000),and Paper I), the slit was oriented to take the spectra ofboth galaxies simultaneously. We discovered that ten of ourtargets (3-04A, 3-10A, 3-44A, 3-49A, 3-57A, 3-65A, 3-78C,3-83A, 3-85A, 3-89A) were physical pairs of galaxies withthe same redshift estimate, and four objects (3-17A, 3-20B,3-72A, 3-86A) were stars, so the total number of galaxiesobserved is 87, and out of them 85 have spectra with a good S/N. In this sample three galaxies (3-13A, 3-57A2,3-68A) have an absorption line spectra, and six more ob-jects show only a very weak (S/N ∼ α emission line.For these galaxies we can measure the redshift, but theline is too faint to measure the flux. Finally one object, 3-40B, is a very peculiar case: its Keck spectrum, althoughvery noisy, shows only one emission line that we identifiedas the Ly- α at z = 2 .
954 (see Fig.1). For this ISOCAMsource, we only have upper limits both at 60 µ m and at15 µ m (Mazzei et al. , 2001). Thus the final sample that weanalyze is composed of 75 galaxies.During each observing night some spectrophotometricand radial velocity standard stars were observed with thesame configuration grism/slit of the target objects. All thedata were reduced using standard IRAF reduction pack-ages and a detailed description of the reduction is given inPaper I. Spectra were classified according to the presence orabsence of various emission lines. Observed emission linesincluded [O II] 3727, H β , [O III] 4959-5007, H α , [N II] 6548-6583, and [S II] 6717-6731. The measured redshifts are re-ported in Paper I.In addition to the measure of the lines intensity, thehigh resolution of the ESI spectra allowed us to measurethe rotation curves for almost all the galaxies observed withKeck II (31 galaxies). The optical spectra are contaminated and often dominatedby the absorption lines of the stellar component, which af-fect the strengths of most emission lines of interest. Themagnitude of this effect depends on the equivalent widths ofthe emission and absorption lines, and it is generally large inthe nucleus of galaxies. For accurate measurements of theirfluxes, we must remove the starlight contribution from ourspectra. Generally, there are two strategies for removingstarlight and thus obtaining a continuum subtracted, pureemission line spectrum: i) an off-nuclear spectrum (with-out emission lines) is subtracted from the spectrum of thenucleus, or ii) a template spectrum, free of emission lines,is properly scaled to and subtracted from the spectrum ofinterest.We used the second method, i.e., the technique ofFilippenko & Halpern (1984) and Ho et al. (1993). Forthis purpose, we selected a sample of 57 pure absorptionline galaxies from the data of the third data release of theSloan Digital Sky Survey (SDSS; Abazajian et al. (2005)),whose spectra have very similar characteristics to ours. Inparticular, to account for the underlying stellar continuumdue to the bulge component, template galaxies with differ-ent properties such as internal reddening and line strengthLick indices (Mg1, Mg2, Mgb, NaD), have been selected.The line strength indices are in the range typical of an oldstellar population, and they include all the values found byAnnibali et al. (2007). The internal absorption ranges fromzero to A v ∼
10 mag as derived from the H α /H β ratio inour sample.We proceeded with the following steps: i) correct ourown spectra and template spectra for Galactic reddening IRAF is distributed by the National Optical AstronomyObservatory, which is operated by the Association ofUniversities for Research in Astronomy (AURA) under coop-erative agreement with the National Science Foundation.2ettoni et al.: Multi–wavelength study of the IDS/ISOCAM sample.III
Table 1.
Instrument set-up for every night of observations (see also della Valle et al. 2006)
Telescope Run ( ′′ )/pix Grism Slit λ Resgrating ( ′′ ) range(˚A) km/secEkar+AFOSC 28/6/00 0.473 − − − Fig. 1.
Region of the spectrum of 3-40B source with the Ly α line visible.using the extinction values of Schlegel et al. (1998), ii) de-redshift the spectra to zero, iii) change the resolution of thetemplates spectra to the resolution of the NEPR ones, iv)subtract the template spectrum from the object spectrum.For each NEPR spectrum, we applied this procedure usingall the templates. The best subtraction was chosen usingthe χ minimization. This method works successfully withthe low-resolution spectra of TNG and Ekar telescopes, butit did not work properly with the spectra taken at Keck II.This is due to the large difference between the resolution ofthe Keck II spectra and that of the template spectra. Forthis reason we did not subtract the starlight continuumto the Keck II spectra. This implies that fluxes of suchemission lines cloud be underestimated.In conclusion, for the Ekar and TNG spectra we used thefluxes corrected by the template subtraction. For spectrataken with the Keck II telescope, we used the raw fluxes,corrected only for Galactic extinction. For the Ekar andTNG spectra, we continued our analysis quantifying theerror due to the choice of different templates. With thisaim for each NEPR spectrum, we compared the flux ofH α emission line measured using the different template-subtracted spectra: we found differences from 2% up to 10%in the case of high S/N and low S/N respectively. Figure2 shows two examples of template-subtraction for high andlow S/N spectra. In our observations the slit of the spectrograph was orientedalong the apparent major axis of our targets or with a po-sition angle that allows having two galaxies along the slit.For this reason we have to take differential atmospheric re-fraction (DAR) effects on our measured fluxes into account.To minimize these effects, our observations were made asclose as possible to the zenith. In general our targets wereobserved with air mass ∼ ≥ α )/F(H β ), 2.86. The internal ab-sorption ranges from A v ∼ A v ∼
14 mag as de-rived from the H α / Hβ ratio in our sample. Its averagevalue is A v = 6 . ± . Fig. 2.
Top panel: 3-38B, bottom panel: 3-77A. In each panel the bottom plot represents the best fitting template modelused to match the stellar component. The middle spectrum is the observed one and the difference at the top. The spectraare scaled by an arbitrary constant.jects. This is not surprising since our sample is composedof dusty galaxies whose far-IR properties were analyzed inMazzei et al. (2007). In Table 3 we give the A v values de-rived, when available.This correction, as expected, does not affect the spectralclassification.We measured all the main parameters of the emissionlines, i.e. fluxes with their errors (see Tables 3, 4), fittingthe emission lines with Gaussian functions. The error esti-mates of the line fitting are computed by error propagationassuming independent pixel sigmas and no errors in thebackground. In general in our spectra, the emission lines arealso extended along the spatial axis. We decided to performa measure adding together all the lines in a physical regionof 3 Kpc derived using the luminosity distance computed,for each galaxy, according to our cosmological model. Thisis because, for our more distant objects ( ∼
30% of our sam-ple) the 3 Kpc aperture coincides with the whole opticalextension of the galaxy. Thus, with this choice, all the sam-ple is consistently compared.In this inner region, the emission lines can be straight-forwardly fitted by a single Gaussian function for almost all our sample. However there are some peculiar cases, asexplained below: – The emission lines of the spectra of 3-27A and 3-78C2galaxies showed a double peak. The peaks’ separationis 5.4 ˚A and 3 ˚A respectively, clearly indicating twokinematical components, as discussed in Sect. 4.1. Forthe lines of these two objects, two Gaussian functionswere fitted, and the sum of their area was taken as totalflux. – The spectra of the galaxies 3-44A1, 3-70A, and 3-96Aexhibit a broad component in the H α emission line typ-ical of Seyfert I galaxies, which was fitted by adding asecond Gaussian function to the fit. The FWHM andthe flux of their broad line components are reported inTable 2. Figure 3 shows their spectra and our fits in theH α region.Note that [NII] λ α fluxes of these three . . . . . l ( Å ) F l u x ( - e r g c m - Å - s - ) . . . . l ( Å ) F l u x ( - e r g c m - Å - s - ) . . . . . . l ( Å ) F l u x ( - e r g c m - Å - s - ) Fig. 3.
The H α region for the three AGN (upper left 3-44A1, upper right 3-70A, and bottom 3-96A). The dashed linesshow our fits of the broad and narrow components.objects, since the H α line and [NII] doublet are completelyresolved in the Keck II spectra.When possible (47 spectra), the accuracy of the fit ofthe H α group was checked by evaluating the strength ra-tio [NII] λ / [NII] λ ∼
3. Thevalues we obtained are in the range 2 . ÷ .
3. Spectral properties
Figure 4 shows the redshift distribution of all the 85 galax-ies described in Paper I. The NEPR supercluster (NEPSC)found by Ashby et al. (1996) (see also Burg et al. , 1992)dominates the z –distribution between 0.08 and 0.09.Galaxy members of other expected clusters appear between0.05 and 0.06, 0.07, and between 0.11 and 0.12. Sixty-onepercent of our targets are galaxies with z < .
1, 25% with0.1 ≤ z ≤ z > .
2. The mean red-shift of the galaxies for which we can measure the rotation curves (see Section 4) is h z rot i = 0 . ∼
50% of thesehave redshift z ≤ . The optical spectral classification as Seyfert (Sy), LINER(L), and star-forming (SF) galaxies, is based on diagnos-tic diagrams of Veilleux & Osterbrock (1987), a revision ofthe pioneer work by Baldwin et al. (1981). They exploitedfour emission-line ratios of the most prominent bright emis-sion lines: [OIII] λ / H β , [NII] λ / H α , [OI] λ / H α and [SII] λλ , / H α . These line ratios take full advan- Fig. 4.
Redshift distribution of our ISOCAM sources, in 65 IDS/ISOCAM fields. The insert in the lower right cornercorresponds the source 3-40B (see text).
Table 2.
Broad line data for 3-44A1, 3-70A, and 3-96A
Galaxy FWHM Fluxname km s − − erg s − cm − ˚ A − ± ± ± tage of the physical distinctions between the various typesof objects and minimize the effects of reddening correc-tion and calibration errors. Diagnostic diagrams have beenused ever since as a standard to identify narrow–line activegalaxy nuclei (AGN, or Sy). Kewley et al. (2001) (see alsoGroves & Kewley (2008) and references therein) build upa detailed starburst model with large ranges of metallicityand ionization parameter, finding upper limits to separatestarburst from AGN on such diagrams. Star-forming (SF)galaxies fall onto the lower left-hand region of these plots,narrow-line Sy are located in the upper right, and Ls inthe lower right-hand zone. Thus, to separate the differenttypes of galaxies (Sy, L, SF galaxies) we used the theoreticalboundaries of Kewley et al. (2001).Figure 5 shows the position of our galaxies on thesediagrams, including galaxies in the 60 µ m complete sample(S60 ≥
80 mJy) of Mazzei et al. (2001). The data referto the fluxes measured in the inner physical 3 Kpc regionof the galaxies. Three galaxies, 3-44A1, 3-70A, and 3-96A,show the H α with a broad line component (Fig. 3), andwe classified them as Sy 1. In Fig. 5 they are residing inthe AGN region and will be discussed further below. Table2 presents the FWHM and fluxes of their broad line H α emission, and Tables 3 and 4 the fluxes of their narrow linecomponents used in the spectral classification. In our sample there are 25 objects (33%) with all thefour available diagnostic ratios (see Fig. 5). However, itis possible to address a spectral classification for other 17galaxies using only two diagnostic diagrams, so we classifya total of 42 galaxies (56% of the observed ones).After analyzing the diagnostic diagrams (Fig. 5) fol-lowing Kewley et al. (2006) and Groves & Kewley (2008),all the galaxies below the pure SF line defined byKauffmann et al. (2003) are SF. We found that 16 out of25 galaxies with all four ratios available, turn out to be SFgalaxies.To classify all the remaining galaxies we used the theo-retical one sigma boundaries between the regions occupiedby LINERS (L) and AGN in these diagnostic plots, as inKewley et al. (2006). We note that two galaxies are cer-tainly inside the region that defines the Ls, and seven moregalaxies are in the region defined by the one sigma bound-aries as Ls. The emission lines in the spectra of these sevengalaxies do not have a broad line component. This pointfurther favors their classification as Ls. Following all thesecriteria, the three AGN galaxies discussed above appear inthe upper right-hand region of our diagnostic plots as ex-pected.A tentative spectroscopic classification of 3-84A wasmade by Ashby et al. (1992) obtaining an ambiguous re-sult: they found that this galaxy has [NII] λ λ –
30 SF (71% and 40% of the classified and observedgalaxies respectively), – – F IR from 42.5 to 122.5 µ m, where L FIR = 4 πD ( F IR ) (1)
F IR = 1 . × − (2 . f + f ) W/m , (2)and f and f are in Jy (Helou et al. , 1988), of the35 ISO sources corresponding to the 42 galaxies in Table5. In particular, in Table 5, 30 SF galaxies correspondto 26 ISO/IRAS sources, and three of these, 3-78C,3-79C, and 3-92A, have no FIR fluxes (Mazzei et al. ,2001), so there are 23 ISO/IRAS sources classified asSF types in the right-hand panel of Fig. 6. In thisfigure, as in the following ones, K-corrections were de-rived from evolutionary population synthesis models tak-ing dust effects into account (Mazzei et al. , 1995), andluminosities were in units of solar bolometric luminosity,L=3.83 × erg/s. Moreover, upper limits to flux densi-ties were accounted for by exploiting the Kaplan-Meier(KM) estimator (Kaplan & Meier , 1958), as in Paper II.Calculations were carried out using the ASURV v 1.2 pack-age (Isobe & Feigelson , 1990), which implements methodspresented in Feigelson & Nelson (1985) and in Isobe et al.(1986). The KM estimator is a nonparametric, maximum-likelihood-type estimator of the true distribution function(i.e., with all quantities properly measured, and no upperlimits). The survivor function, giving the estimated propor-tion of objects with upper limits falling in each bin, doesnot produce, in general, integer numbers, but is normalizedto the total number. This is why noninteger numbers ofobjects appear in the histograms of our figures.The average value of the LFIR distribution, which ac-counts for 18 upper limits, is log(LFIR)=10.3 (Fig. 6, leftpanel), very close to the average value of the FIR selectedcomplete sample of Mazzei et al. (2007). Liner galaxiesshow the same average LFIR whereas, by accounting onlyfor SF types, this slows down slightly, 9.9 (Fig. 6, rightpanel).
4. The rotation curves
The spectra obtained with the Keck II telescope have bothwavelength and spatial resolution high enough to allow usto measure the galaxy rotation curves (RC). Only threegalaxies do not show emission lines and for this reason wehave such data for 31 out of the thirty-four galaxies ob-served with this instrument (41% of the whole sample ofobserved galaxies). To measure the RCs, a Gaussian func-tion was fitted on the H α emission lines with S/N ≥
2. Insome cases more than one component was detected, eitherbecause of the presence of two decoupled kinematic com-ponents as in 3-26B and 3-27A or because of interactingregions of two galaxies (3-78C2).Since the main goal of our spectroscopic observationswas to measure the redshifts of the NEPR galaxies, the slitposition of the spectrograph was not always placed alongthe galaxy major axis: for example, some objects were ob-served simultaneously in the same exposure (see Section2). The slit orientations therefore, have heterogeneous mis-alignment respect to the real major axes of the galaxies: for nine out of them the misalignment is less than ten de-grees, and the slit was oriented along an oblique axis forthe remaining 22 galaxies. This misalignment angle, δ , hasbeen measured for all the Keck II spectra, and it is re-ported in Table 6. The redshift distribution of 31 galaxiesfor which we measured the RCs is shown in the left-handpanel of Fig.7. The mean redshift of the distribution is h z rot i = 0 . z = 0 . z = 0 . z ≤ .
1. Two galaxies (3-44A2 and3-53A1) have z > .
3. The observed RCs have been mod-eled with the normalized arctan rotation curve function: v ( r ) = v + 2 π v ∗ arctan ( R ) (3)where R = ( r − r ) /r t , and v is the systemic velocity of thenucleus, r the spatial center of the galaxy, v ∗ the asymp-totic velocity, and r t the transition between the rising andflat parts of the rotation curve (Courteau , 1997). In Figs. 9and 10, the best fit (Eq. 3, continuous line) is shown.To enlarge the sample of kinematical data in Table 6, weincluded the results from some low-resolution spectra ob-tained with the TNG telescope. These spectra do not allowthe measure of the RCs; nevertheless, since the spectra ofthe ten galaxies observed with this telescope (3-23A, 3-31A,3-38A, 3-47A, 3-59A, 3-61A, 3-81A, 3-81B, 3-88A1, 3-93A)are spatially resolved, we succeeded in measuring at leastthe ∆ v , i.e., the velocity difference between the core and theouter regions. The redshift distribution of these galaxies isshown in the left-hand panel of Fig. 7. The mean redshiftof the distribution is h z rot i = 0 . z = 0 . z = 0 . . ′′ with the exception of 3-59A (1 ′′ ), 3-88A1 (1 ′′ ), and3-93A (2 . ′′ ). Then, we extracted spectra of two regionsin the outer part of the galaxies. For each of these spectra,the position of the blending group of H α and [NII] emis-sion lines was measured. For four galaxies (3-23A, 3-31A,3-59A, 3-61A), these three emission lines were deblended,while in the remaining ones only the positions of H α and[NII] were measured. Since [NII] is much weaker than H α ,we considered its contribution negligible. Finally, the veloc-ity difference, ∆ v , between the velocity of the nucleus andin the outer region was calculated. Most of the thirty-one RCs in Figs. 9 and 10 are very reg-ular, which is typical of disk galaxies; however, some showpeculiar motions, very disturbed rotation patterns, and infew cases, kinematic decoupled regions that were excludedfrom the fit of RC. In the following we summarize the puz-zling features. : this ISO/IRAS source is composed of twogalaxies, A1 and A2, at both the same redshift(della Valle et al. , 2006). Their disturbed RCs show theseare interacting/merging galaxies. : its RC extends beyond 8 Kpc, but only the re-gion of rigid rotation is visible. No imaging is available toconfirm this point. : also in this case, the RC extends more than 9 Kpc,showing only the rigid rotation zone; there is no imagingavailable for this source either. : this is a edge-on spiral galaxy. Its RC showsa double component in the northern region, towards thecompanion galaxy 3-26A. In the GALEX archival images,a bridge connecting the two galaxies appears in the sameregion where the RC curve shows the double component;the plateau of its RC is hinted at because the slit of thespectrograph, placed along the major axis, is smaller thanthe apparent dimension of such a galaxy. : two inner velocity components separated bymore than 200 km/s arise in the RC of this galaxy. Theycorrespond to the opposite sides of a fast rotating ring de-coupled from the disk motion. : the west side of its RC, with the photometriccenter 2.2arcsec far from the dynamical center, shows per-turbed behavior. This pattern may be a signature of anoutburst or of some type of interaction; however, no imag-ing is available to explain the peculiarity of such an RC. : this galaxy corresponds to PGC3086419(HYPERLEDA catalog), but also in this case no imagingdeep enough is available to solve the puzzle of its RC. : even if this galaxy corresponds to PGC2699424(HYPERLEDA) and to 2MASX J17562643+6723549, theresolution of available images cannot explain the very per-turbed RC of this galaxy that could be the signature of afast rotating inner bar. : this is the highest redshift (0.37) galaxy be-longing to the 60 µ m complete (S >
80 mJy) sample(Mazzei et al. , 2001), and its RC only refers to the slowlyrotating inner region. : this ISO/IRAS source is composed by a pair ofgalaxies, A1 and A2 (della Valle et al. , 2006); the north-west distortion of RC of A1 galaxy cloud be due to theinteraction with its companion galaxy, A2; however, ourimaging do not confirm tidal distortions. : there is an anomaly in the western region ofits RC, which can be explained, looking at the DSS image(the only available so far), by an outburst visible in thenorth-west side of this galaxy. : the anomaly arising in the north-east region ofits RC may be due to a companion galaxy. No imaging withenough resolution is available to elucidate this point. : in the south-est region of its RC a per-turbation arises that, considering the 2MASS image(2MASXJ18042694+6720481), may be due to a faint ob-ject near the galaxy. : there is a double component in the RC of thisgalaxy that may be due to interaction; however, its com-panion galaxy, 3-78C1, shows no signs of tidal interaction.We derived the spectral classification of 28 out of the31 galaxies with RCs. For three objects, namely 3-40, 3-45,and 3-76, no [OIII] λ We defined v o pt as the velocity measured at the maximumspatial extent in arcsec ( r o pt ) of either the approaching orthe receding side of the galaxy spectra. No rotation curveshows the flat region of the curve clearly. Velocities are corrected for projection on the sky, cos-mological stretch, and misalignment angle. Therefore, thecorrected v c is there v c = v (1 + z ) sin i cos δ (4)where the inclination along the line of sight, i , was esti-mated using the observed axis ratio derived from the CCDimages when available (see Paper I), and the POSS imagesfor the others; an intrinsic flattening, q , of 0.20 for all thegalaxies is assumed. We usecos i = q − q − q (5)where the adopted i for each galaxy is given in Table 6.The misalignment angle is a critical parameter forour measurements, because several spectra were observedfar from the major axis. Following the simulation ofGiovannelli et al. (1997), errors due to this correction arenegligible for position angle offsets less than 15 ◦ . For sev-enteen galaxies this condition is verified. For two, 3-65A2,and 3-70A, this correction is instead not applicable, be-cause they were observed very close to the minor axis. Forother two galaxies, 3-10A2, and 3-81A the mass estimatesis affected by this correction (see Table 6).Using the derived velocity v c , we estimated the massinside the last observed point of thirty-nine galaxies, i.e., alower limit to the mass of our galaxies, with the relation byvan der Bosch (2002): M vir = 2 . · M ⊙ (cid:18) r d Kpc (cid:19) (cid:18) v c
100 kms − (cid:19) (6)where r d is the virial radius. We used the radius R opt in Kpcas defined above as r d . These estimates are summarized inTable 6.The right-hand panel of Figure 7 shows the mass dis-tribution of galaxies in Table 6 and Fig. 8. The left-handpanel shows the same, accounting for KM estimator as-suming lower mass limits when the RCs are less extendedthan 8 Kpc, and for all the ∆v estimates. The mass dis-tribution spans more than three orders of magnitudes. Foreleven galaxies the masses are derived in the nuclear region,i.e., ( R opt < · M ⊙ , and4.93 · M ⊙ including all the galaxies. We point out thatthe high velocity, v opt , measured for the 3-61A galaxy (seeTable 8) may be due to dynamical perturbations inducedby interaction with a dwarf compact companion ≃ · M ⊙ , the same as derived from galaxies with RCmeasured.We note that the 67% of these galaxies belongs to the60 µ m selected, complete sample defined by Mazzei et al.(2001). The right-hand panel of Fig. 8 shows the rest-frameFIR luminosity distribution of 32 ISO/IRAS sources corre-sponding to galaxies in Table 6. Four galaxies i.e., 3-10A2,3-78C2, 3-83A2, and 3-83A2 are multiple optical coun-terparts of the same ISO/IRAS source (della Valle et al. , µ mBright Galaxy Sample (Sanders et al. , 2003) and for a nor-mal spiral galaxy, like the Milky Way (Mazzei et al. , 1992).The ultraluminous infrared galaxy 3-53A emits the maxi-mum FIR luminosity of the sample, nearly 100 times higherthan the mean value.
5. Conclusions
The spectral classification of 42 galaxies, with the emis-sion line ratio diagnostic diagrams, shows that the NEPRsample is predominantly composed of SF, starburst galax-ies (71%), while the fraction of LINERs (21%) and AGNs(7%) is smaller. Three new Sy 1 galaxies were identified,3-44A1, 3-70A, and 3-96A. The rest-frame FIR luminositydistribution of galaxies with spectral classification spansthe same range as the FIR-selected complete sample an-alyzed by Mazzei et al. (2007), i.e. three orders of mag-nitude, with the same mean value, log( L F IR )=10.2. Thisemphasizes that such galaxies represent FIR properties ofthe whole sample well. Moreover, their optical propertiesare typical of the sample itself since 62% of these belong tothe 60 µ m selected complete sample of galaxies defined byMazzei et al. (2001) (see della Valle et al. (2006)).Using the rotation curves and spatially resolved, low-resolution spectra, we are able to derive dynamical param-eters of 41 galaxies and mass estimates, inside the last pointviewed, of 39 galaxies in the sample. We point out that the67% of them belong to the 60 µ m complete sample citedabove. Moreover, also in this case, their rest-frame FIR lu-minosity distribution extends over the same range and hasthe same mean value as expected for the complete sample.The mass distribution extends over three orders of magni-tude with a mean value of h M i = 3 . · M ⊙ , slightlymore than the Milky Way, where Wong et al. (2004) find amass of 1.3 · M ⊙ within 12 Kpc of the Galactic center.Paper I concluded that two or more galaxies with veryclose redshifts may contribute to the ISOCAM/IRAS fluxin at least seven cases (3-04A, 3-10A, 3-57A, 3-65A, 3-78C, 3-83A, 3-89A). Dynamical perturbations of the rota-tion curves discussed here prove that 3-10A and 3-65A areinteracting/merger systems. Interactions also involve the 3-78C2 galaxy, and disturbed patterns appear in the rotationcurve of 3-83A1; however, 3-89A1 and 3-89A2 galaxies showunperturbed rotation curves.We find several systems with previously unsuspected de-coupled velocity components, 3-26B, 3-27A, 3-37A, and 3-40A. Moreover peculiar motions arise in the rotation curvesof 3-26A, 3-30A, 3-66A, 3-69A, 3-70AA, 3-71A, and 3-76A.Thus, 48% of the rotation curves have disturbed morpholo-gies and most part of these are SF galaxies. This empha-sizes the role of interactions in triggering starbursts and, inparticular, FIR emission in our sample of dusty galaxies. Acknowledgements.
We thank the anonymous referee whose com-ments greatly improved the paper. Some of the data presented hereinwere obtained at the W.M. Keck Observatory, which is operated asa scientific partnership among the California Institute of Technology,the University of California, and the National Aeronautics and Space Administration. The Observatory was made possible by the generousfinancial support of the W.M. Keck Foundation.
References
Abazajian, K. et al. 2005, AJ, 129, 1755Annibali, F., Bressan, A., Rampazzo, R., Zeilinger, W. W., & Danese,L. 2007, A&A, 463, 455Aussel, H., Coia, D., Mazzei, P. et al. 2000, A&AS, 141, 257Ashby, M., Houck, J., Hacking, P., 1992, AJ, 104, 980Ashby, M., Hacking, P., Houck, J., et al. 1996, ApJ, 456, 428Baldwin, J. A., Phillips, M. M., & Terlevich, R., 1981, PASP, 93, 5Burg, R., Giacconi, R., Huchra, J. et al. 1992, A&A 259, L9Courteau, S. 1997, AJ, 114, 2402della Valle A., Mazzei, P., Bettoni, D. et al. 2006, A&A 454, 453Feigelson, E., Nelson, P. 1985, ApJ, 293, 192Filippenko, A.V. 1982, PASP, 94, 715Filippenko, A. V & Halpern, J.P., 1984, ApJ, 285, 458Hacking, P., Houck, I.R. 1987, ApJS, 63, 311Hacking, P., et al. 1989, ApJ, 339, 12Giovannelli, R., Haynes, M. P., Herter, T., Vogt, N. P., da Costa, L.N., Freudling, W., Salzer, J. J. and Wegner, G. 1997, AJ, 113, 53.Groves, B., Kewley, L. 2008, APS Conf.S. 390, 293Helou, G., Khan, I., Malek, L., & Boehmer, L. 1988, ApJS, 68, 151.Ho, L. C., Filippenko, A. V. & Sargent, W. L. W., 1993, ApJ, 417, 63Isobe, T., Feigelson, E., Nelson, P. 1986, ApJ, 306, 490Isobe, T. Feigelson, E. 1990, BAAS, 22, 917Kauffmann, G., et al. 2003, MNRAS, 346, 1055Kaplan, E., Meier, P. 1958, J. Am. Stat. Ass. 53, p. 457Kennicutt, R.C.jr. 1998, ARA&A, 36 189Kewley, L. J., Heisler, C. A., Dopita, M. A. & Lumsden, S. 2001,ApJS 132, 37Kewley, L. J.,Groves, Kauffmann, G., Heckman, T. 2006, MNRAS372, 961Le Floc’h, E., Papovich, C., Dole, H., et al. 2005, ApJ, 632, 169Mazzei, P., de Zotti, G., Xu, C. 1992, A&A, 256, 45Mazzei, P., Curir, A., Bonoli, C. 1995, AJ, 422, 81Mazzei P., Aussel, H., Xu, C. et al. 2001, New Astr, 6, 265Mazzei, P., della Valle, A. & Bettoni. D., 2007 A&A, 462, 21Melbourne, J., Ammons, M., Wright, S. A., et al. 2008, AJ, 135, 1207Rosa-Gonz´alez, D., Terlevich, E., & Terlevich, R. 2002, MNRAS, 332,283Sanders, D.B., Mazzarella, J.K, Kim, D., et al. 2003, AJ, 126, 1607Schlegel, D. J., Finkbeiner, D. P., and Davis, M., 1998, Apj, 500, 525Treister, E., Urry, C. M., Schawinski, Ke., Cardamone, C. N., Sanders,D. B., ApJ 722, L238van der Bosch, 2002, MNRAS, 332, 456Veilleux, S. and Osterbrock, D. E., 1987, ApJS, 63, 295Wong, O. I., Drinkwater, M. J., Jones, J. B., Gregg, M. D., & Freeman,K. C., 2004, Dark Matter in Galaxies, 220, 213
Fig. 5.
Emission line diagnostic diagrams. Galaxies belonging to the 60 µ m complete sample of Mazzei et al. (2001)are represented with filled triangles; light (cyan) curves in all the three panels show the extreme starburst definitionof Kewley et al. (2001); bold curve (magenta) shows the pure SF limit of Kauffmann et al. (2003); and red lines theLINER/AGN divisions from Kewley et al. (2001) and Kewley et al. (2006). Upper panels: log [OIII] λ β versuslog [NII] λ α and log [OIII] λ β versus log [SII] λ α . Lower panel: log [OIII] λ β versuslog [OI] λ α . Fig. 6.
Left : The rest-frame FIR luminosity distribution of 35 ISO/IRAS sources included in this study. Far-IR data(Mazzei et al. , 2001) account for KM estimator (see text).
Right:
As in the left panel accounting for spectral classificationhere derived (Table 5).
Fig. 7.
Left: Redshift distribution of 41 galaxies in Table 6: gray histogram is for 31 galaxies with rotation curves, dashed histogramshows the remaining galaxies; the bin size is ∆ z = 0 .
05. Right: Mass distribution of 39 galaxies in the same Table (see text); thebin size is 0.2. 11ettoni et al.: Multi–wavelength study of the IDS/ISOCAM sample.III
Fig. 8.
Left: The same as in Fig. 7 but acconting for KM estimator. Right: The rest-frame FIR luminosity distribution for 32galaxies with mass estimates exploiting KM estimator to take upper limits to FIR fluxes into account (Mazzei et al. , 2001).12ettoni et al.: Multi–wavelength study of the IDS/ISOCAM sample.III
Fig. 9.
The observed rotation curves for the galaxies in our sample (at a mean z ∼ . Fig. 10.
The observed rotation curves for galaxies in our sample (at a mean z ∼ . e tt o n i e t a l.: M u l t i – w a v e l e n g t h s t ud y o f t h e I D S / I S O C A M s a m p l e . III
Table 3.
Fluxes of the principal emission lines: red spectral region [SII] [SII] [NII] H α [NII] [OI] [OIII] [OIII] A v Object 6731 6717 6583 6562 6548 6300 5007 4959 mag3-04A1 0.810 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± e tt o n i e t a l.: M u l t i – w a v e l e n g t h s t ud y o f t h e I D S / I S O C A M s a m p l e . III
Table 3. continued [SII] [SII] [NII] H α [NII] [OI] [OIII] [OIII] A v Object 6731 6717 6583 6562 6548 6300 5007 4959 mag3-65A1 2.543 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − erg s − cm − ˚ A − , are not corrected for internal absorption. For H α we report here only the narrow component. The fluxes of the[SII] doublet are given separately, if available, otherwise the total flux of the doublet is given in the middle of the columns of the single lines. ettoni et al.: Multi–wavelength study of the IDS/ISOCAM sample.III Table 4.
Fluxes of the principal emission lines: blue spectral region H β H γ [OII] [OII]Object 4861 4340 3729 37263-04A1 0.953 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Table 4. continued H β H γ [OII] [OII]Object 4861 4340 3729 37263-81B — — — —3-83A1 0.065 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − erg s − cm − ˚ A − . The fluxes of the [OII] doublet are given separately, if available,otherwise the total flux of the doublet is given in the middle of the columns of the single lines. Table 5.
Spectral classification
Name Class. Name Class. Name Class.3-10A1 SF 3-42A L 3-78C1 SF3-10A2 SF 3-44A1 Sy 3-78C2 SF3-12A SF 3-44A2 SF 3-79C SF3-16A SF 3-49A1 L 3-83A1 SF3-20A L 3-53A1 L 3-83A2 SF3-21A SF 3-54A1 SF 3-84A1 L3-24A SF 3-55A SF 3-88A1 SF3-26A L 3-57A1 L 3-89A1 SF3-26B SF 3-65A1 SF 3-90A SF3-27A SF 3-65A2 SF 3-91A SF3-29A L 3-66A SF 3-92A SF3-30A SF 3-69A SF 3-92C L3-36A SF 3-70A Sy 3-93A SF3-37A SF 3-71A SF 3-96A Sy
Table 6.
Kinematical parameters and masses.
Name z i δ r o pt R o pt v o pt v c Mass r t [ ◦ ] [ ◦ ] [ ′′ ] [Kpc] [km/s] [km/s] [log M ⊙ ] [ ′′ ]3-10A1 0.0875 48 22 5.99 9.79 100 133 11.65 0.63-10A2 0.0867 38 74 2.99 4.50 50 271 11.92 0.43-12A 0.0775 24 14 5.97 8.74 25 58 10.89 —3-16A 0.1176 51 6 3.31 7.33 117 140 11.57 1.83-20A 0.0736 71 15 6.78 9.43 90 92 11.31 —3-23A ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ Columns: (1) galaxy name, (2) redshift, (3) galaxy inclination, (4) misalignment angle, (5) and (6) maximum radius of therotation curve in arcsec and Kpc, respectively, (7) rotation velocity measured at R opt , (8) mass derived by equation (1), (9)circular velocity, i.e., rotation velocity accounting for inclination and misalignment corrections (10) transition radius between therising and flat parts of the rotation curve; ∗∗