Radio synchrotron spectra of star-forming galaxies
AAstronomy & Astrophysics manuscript no. gal˙spec˙astroph c (cid:13)
ESO 2017October 10, 2017
Radio synchrotron spectra of star-forming galaxies
U. Klein , U. Lisenfeld , , and S. Verley , AIfA, Universit¨at Bonn, Auf dem H¨ugel 71, 53121 Bonn, Germanye-mail: [uklein]@astro.uni-bonn.de Departamento de F´ısica Te´orica y del Cosmos, Universidad de Granada, Spaine-mail: [ute,simon]@ugr.es Instituto Universitario Carlos I de F´ısica Te´orica y Computacional, Facultad de Ciencias, 18071 Granada, SpainReceived month ??, ????; accepted month ??, ????
ABSTRACT
The radio continuum spectra of 14 star-forming galaxies are investigated by fitting nonthermal (synchrotron) and thermal (free-free)radiation laws. The underlying radio continuum measurements cover a frequency range of ∼
325 MHz to 24.5 GHz (32 GHz in caseof M 82). It turns out that most of these synchrotron spectra are not simple power-laws, but are best represented by a low-frequencyspectrum with a mean slope α nth = . ± .
20 ( S ν ∝ ν − α ), and by a break or an exponential decline in the frequency range of1 – 12 GHz. Simple power-laws or mildly curved synchrotron spectra lead to unrealistically low thermal flux densities, and / or tostrong deviations from the expected optically thin free-free spectra with slope α th = .
10 in the fits. The break or cuto ff energies arein the range of 1.5 - 7 GeV. We briefly discuss the possible origin of such a cuto ff or break. If the low-frequency spectra obtained herereflect the injection spectrum of cosmic-ray electrons, they comply with the mean spectral index of Galactic supernova remnants. Acomparison of the fitted thermal flux densities with the (foreground-corrected) H α fluxes yields the extinction, which increases withmetallicity. The fraction of thermal emission is higher than believed hitherto, especially at high frequencies, and is highest in thedwarf galaxies of our sample, which we interpret in terms of a lack of containment in these low-mass systems, or a time e ff ect causedby a very young starburst. Key words. galaxies: radio continuum, star formation, magnetic fields; radiation mechanisms: non-thermal, thermal
1. Introduction
Star-forming galaxies emit both thermal (free-free) and nonther-mal (synchrotron) radiation in the radio regime. It is known sincea few decades that the synchrotron component dominates at fre-quencies up to about 10 GHz (Klein & Emerson 1981; Gioiaet al. 1982) and that the overall slope of the observed spectra hasa mean value of about 0.70 to 0.75, with little dispersion (about0.10). And for decades, the perception has been that the spectraare the superposition of two power-laws, S tot ( ν ) = S th ( ν ) (cid:32) νν (cid:33) − . + S nth ( ν ) (cid:32) νν (cid:33) − α nth , (1)where α nth is the spectral index of the synchrotron radiationspectrum. The superposition of thermal and synchrotron radi-ation thus produces a flattening of the total radio spectrum to-wards high frequencies, with the asymptotic value of − .
1. This,however, can hardly be seen, owing to the onset of thermal radi-ation from dust, which becomes significant above about 40 GHz.In the past decade, numerous studies have been dedicatedto characterize the shape of the radio spectrum. The two ma-jor questions have been: What is the contribution of the ther-mal emission (thermal fraction, f th = S th / S tot )? What is thesynchrotron spectral index, α nth , and how does it change withfrequency? Past studies have given di ff erent answers, owing tothe di ff erences in the covered frequency range and the sampleselection. Gioia et al. (1982) found a mean spectral index of thetotal radio continuum emission of α tot ≈ .
7, based on a sampleof 56 galaxies, which they had observed at 408 MHz, 4.75 and10.7 GHz. They concluded that the spectral index of the syn-chrotron radiation is α syn ≈ .
8, with the synchrotron emission dominating below 10 GHz. They estimated the fraction of ther-mal radiation to be f th ,
10 GHz < α nth = . ± .
06, and a fraction of thermal emission of < ∼
40% at 10 GHz. Duric et al. (1988), however, claimed a largevariation of the thermal fraction in a sample of 41 spiral galaxies.Niklas et al. (1997) found a thermal fraction f th , = ± α nth changing from 0.45 to 0.69. These lat-ter two papers strongly indicate that the synchrotron spectra ofstar-forming galaxies cannot be simple power-laws, but probablydecline with increasing frequency in the cm-wavelength range.Recently, Tabatabaei et al. (2017) presented measurements of 61galaxies 1.4, 4.8, 8.5, and 10.5 GHz. Their analysis delivered atotal spectral index of α tot = . ± .
15, a mean synchrotronspectral index of α nth = . ± .
16, and a fraction of thermalradiation of f th , = ± < ∼ ff erent, and thefrequency coverage of the measurements culled from the liter- a r X i v : . [ a s t r o - ph . GA ] O c t . Klein et al.: Synchrotron spectra of galaxies ature too small. Third, the careful data selection described inSect.3 may not have been made in some of the works.The spectrum of the di ff use synchrotron emission is shapedby the processes that characterize the propagation of relativis-tic electrons, in particular the type of propagation (di ff usion orconvection), of energy losses (synchrotron, inverse-Compton,adiabatic losses or bremsstrahlung), and the confinement to thegalaxy. It is generally accepted that the acceleration takes placein the shocks of supernova remnants (SNRs), which have an av-erage spectral index of -0.5 Green (2014). From there, the rel-ativistic particles propagate into the interstellar medium (ISM)via di ff usion and / or convection, and lose energy. The dominantenergy losses at high frequencies are synchrotron and inverse-Compton losses, which steepen the synchrotron spectrum. Atlow frequencies thermal absorption would flatten the spectraand may even produce a turn-over at the lowest frequencies(see the papers and discussion by Israel & Mahoney 1990;Hummel 1991). However, for the frequency range consideredin the present paper, we are not concerned with this latter pro-cess, as it would require extreme emission measures in order tobecome relevant for the frequency range considered here.In this paper, we present an analysis of accurately deter-mined radio continuum spectra of 14 star-forming galaxies,comprising massive grand-design, closely interacting, and low-mass (dwarf) galaxies. The spectra cover a frequency range ∼
325 MHz – 24.5 GHz (32 GHz in case of M 82), with the24.5 GHz measurements performed with the E ff elsberg 100-m telescope in the mid 80’s, and mostly unpublished to date.All other flux densities were carefully selected from the litera-ture. For the first time, the data allow a reliable separation ofthe thermal and nonthermal components to be performed, hencean analysis of the resulting synchrotron spectra, and a firm ex-trapolation towards lower frequencies. This kind of work is alsoconsidered valuable towards establishing and interpreting low-frequency spectra of galaxies obtained with LOFAR observa-tions (see, e.g., Mulcahy et al. 2014). However, for this paper wedecided not to use measurements at frequencies below 325 MHz,since astrophysical processes that shape the spectra there wouldincrease the number of free parameters unnecessarily.In Sect. 2 we present the galaxy sample and their selectioncriteria. The data selection made for the present investigation isdescribed in Sect. 3. In Sect. 4 the fitting procedure of the spectrais described, and in Sect. 5 the results are presented. These arediscussed in Sect. 6, while Sect. 7 contains a summary and listsour conclusions.
2. The Galaxies
The galaxy sample used for our analysis was determined by theinclusion of high-frequency radio continuum data (see Sect. 3.2),which are avaliable for a small number of galaxies mapped withthe E ff elsberg 100-m telescope years ago. This results in a mix-ture of rather di ff erent galaxy types, spanning the whole massrange of star-forming galaxies and incorporating closely inter-acting pairs and ongoing mergers. These galaxies have veryhigh star formation rates per unit surface, hence all of them arerather radio-bright. In what follows we shall briefly introducethe galaxies and their pertinent properties, subdividing them intoseveral categories. All subsequent tables of this paper follow thiscategorization using dashed horizontal lines. Dwarf galaxies : There are five low-mass galaxies in oursample. II Zw 40 and II Zw 70 are classical blue compact dwarfgalaxies (BCGDs). BCDGs are also called H ii galaxies becausethey are dominated by giant H ii regions occupying much of their total volumes. Their emission-line spectra indicate that they aremetal-poor, while they are gas-rich and are experiencing intensebursts of star formation. II Zw 40 was observed in the radiocontinuum by Klein et al. (1984b), Klein et al. (1991), and byDeeg et al. (1993). II Zw 70 is a small and distant BCDG so thatexisting radio continuum measurements by (Klein et al. 1984b;Skillman & Klein 1988; Deeg et al. 1993) only provided inte-grated flux densities. The high-frequency radiation of these twogalaxies is purely thermal.IC 10, NGC 1569, and NGC 4449 are nearby starburstingdwarf galaxies. They have properties that are rather similar tothose of BCDGs. They are gas-rich, too, but their metallicitiesare not so low.Because of its proximity to the Galactic plane, IC 10 is atricky case for radio continuum observations. In particular at lowfrequencies, at which interferometric observations are indispens-able, imaging su ff ers from contamination by spurious sidelobesfrom radio continuum structures in the Galactic plane. Anotherworry is thermal absorption through the plane at the lowest fre-quencies. Useful measurements for our purpose have been car-ried out Klein et al. (1983), Klein & Gr¨ave (1986), Chy˙zy et al.(2003), and by Chy˙zy et al. (2016).NGC 1569 is considered as a template of a low-mass galaxywith an evolved starburst. Israel & de Bruyn (1988) were the firstto notice that the synchrotron spectrum is not a simple power-law, but ‘has a high-frequency cuto ff at 8 ± α emission. Numerous radio con-tinuum studies at a multitude of frequencies, which aimed at anunderstanding of cosmic-ray propagation into the halo regime ofthis dwarf galaxy (Klein & Gr¨ave 1986; Lisenfeld et al. 2004;Kepley et al. 2010; Purkayastha 2014) prove its intense star for-mation. It is among the radio-brightest in our sample and pos-sesses a low-frequency radio halo.Another such nearby template of a starburst dwarf galaxy,NGC 4449, has been observed in the radio continuum by Klein& Gr¨ave (1986, 24.5 GHz), Chy˙zy et al. (2000, 4.9, 8.5 GHz),Chy˙zy et al. (2011, 2.7, 4.9 GHz), Srivastava et al. (2014, 150,325, 610 MHz), and by Purkayastha (2014, 350 MHz). Thisgalaxy, too, is characterized by a high star formation rate (SFR),which was most likely triggered by the close passage of an-other - yet lower-mass - galaxy. NGC 4449 also possesses a low-frequency radio halo. Interacting galaxies : Upon close inspection, almost allgalaxies are gravitationally interacting with another to some ex-tent. In the present sample, NGC 4490 /
85, NGC 5194 /
95 (M 51),and NGC 4631 (with nearby dwarf elliptical NGC 4627 and themore massive disk galaxy NGC 4652) have nearby and obvi-ously interacting companion galaxies. These interactions are thelikely cause of the intense ongoing star formation in these galax-ies, with the close companions stirring up the gas and givingrise to gas compression and subsequent star formation out of themolecular gas.Radio continuum studies of NGC 4490 /
85, a closely inter-acting pair of galaxies, were reported by Viallefond et al. (1980,1.4 GHz), Clemens et al. (1999, 1.49, 4.86, 8.44, 15.2 GHz),and by Klein et al. (1983, 24.5 GHz). Nikiel-Wroczy´nski et al.(2016) published a multi-frequency radio continuum study ofthis system using archival VLA data and new GMRT observa-tions at 610 MHz. The highest frequency at which NGC 4490 /
2. Klein et al.: Synchrotron spectra of galaxies dio continuum spectrum and studying cosmic-ray propagation(Klein et al. 1984a), and later on focused on investigating thestructure of the large-scale magnetic field in it (e.g. Neininger1992; Fletcher et al. 2011; Mao et al. 2015). The rich set of flux-density measurements available for this galaxy required a carefulselection, with only the most reliable data in terms of confusionand missing short spacings culled by us. Klein et al. (1984a) in-ferred a simple power-law for the synchrotron spectrum of M 51.After the discovery of an extended radio continuum haloat 610 and 1412 MHz around NGC 4631 by Ekers & Sancisi(1977) there has been a large number of observations at higherfrequencies (Hummel & Dettmar 1990a; Hummel et al. 1991;Golla 1999; Irwin et al. 2012). These mostly aimed at study-ing cosmic-ray transport out of the disk and at investigating themorphology of the magnetic field of this galaxy. The high star-formation rate is probably caused by the gravitational interactionwith the neighbouring galaxies NGC 4627 and NGC 4652, ren-dering this galaxy a rather radio-bright one.
Merging galaxies : In case of ongoing mergers, star forma-tion is extreme, owing to the collision of molecular clouds in themerging centres of the two galaxies.NGC 4038 /
39, the ‘Antenna Galaxy’, represents one of theclassical nearby ongoing mergers of two galaxies. Chy˙zy &Beck (2004a) and Basu et al. (2017) performed thorough ra-dio continuum studies, with the emphasis on the linear polar-ization and on the resulting morphology of the large-scale mag-netic field in NGC 4038 /
39. In separating the thermal from thenonthermal radio emission, they inferred a simple power-law forthe total synchrotron spectrum of this system.NGC 6052, which became known as a so-called ‘clumpy ir-regular galaxy’ under its label Mkn 297 (Heidmann 1979), is re-vealed as a close merger of two galaxies oriented perpendicu-larly by HST images (Holtzman et al. 1996). It has a high radiobrightness. High-resolution radio continuum observations of thissystem were reported by Deeg et al. (1993).
Starburst galaxies : NGC 2146 and NGC 3034 (M 82) areclassical nearby starburst galaxies, again with the intense star-forming activity resulting from gravitational interaction with anearby galaxy or from a past merger event. A comprehensiveradio continuum study of NGC 2146 was reported by Lisenfeldet al. (1996), who investigated the cosmic-ray transport in thisgalaxy. NGC 3034 (M82) has been the target of a large num-ber of radio continuum studies across a large frequency range.Adebahr et al. (2013) presented observations at λλ
3, 6, 22, and92 cm, using the VLA and the WSRT. Varenius et al. (2015)made dedicated measurements with LOFAR to image the radiocontinuum of M 82 at 118 and 154 MHz. However, such low-frequency measurements are not relevant for our present study,since at frequencies below about 1 GHz various e ff ects mayshape the synchrotron spectra in such a way as to produce devia-tions from a simple power-law (see Sect. 3.2). Klein et al. (1988)published observations of M 82 at 32 GHz, the highest frequencyat which thermal radiation from dust is still insignificant.NGC 3079 and NGC 3310 are starburst which can be consid-ered as transition phases towards AGN (active galactic nucleus)activity. NGC 3079 exhibits a focused wind emerging from itscentre and giving rise to a ‘figure-8’ bow-shock structure seenin the nonthermal radio continuum (Duric et al. 1983; Duric &Seaquist 1988). NGC 3310 has recently swallowed one of itsdwarf companion galaxies (see, e.g., Kregel & Sancisi 2001),this having led to the enormous starburst. High-resolution radiocontinuum observations were performed by Duric et al. (1986),with the aim to study the distribution of cosmic rays in thisgalaxy.
3. The data
The observations at 24.5 GHz had been performed in perfectweather during test measurements in February 1982. The K-band maser receiver system with a bandwidth 100 MHz had beeninstalled in the primary focus of the E ff elsberg 100-m telescope.The half power beam width (HPBW) was measured to be 38 (cid:48)(cid:48) ± (cid:48)(cid:48) . The frontend was equipped with two horns, giving an angularbeam separation of 114 (cid:48)(cid:48) ± (cid:48)(cid:48) on the sky. The aperture e ffi ciencyof the 100-m telescope was 19% at the time of the observations.The pointing and focus of the telescope were checked using thepoint source 3C 286, with small maps centered on this sourceproviding also the flux calibration. The flux density scale is thatof Baars et al. (1977). For each galaxy, six coverages were takenbetween 50 ◦ and 80 ◦ elevation (corresponding to system temper-atures of 79 K and 49 K, respectively, on the sky), using a scanseparation of 10 (cid:48)(cid:48) . These were stacked to yield final maps with anroot mean square (rms) noise of 2.9 mJy / beam area. The observ-ing procedure and reduction technique is essentially the same asdescribed by Emerson et al. (1979). The sample selection wasgoverned by the radio brightness and sizes of the target galax-ies, dictated by the rather high frequency and the comparativelysmall bandwidth of the maser receiver. Since those were the first(and only) measurements of this kind, it was attempted to in-clude star-forming galaxies with a variety of properties (low- andhigh-mass galaxies, interacting systems). In what follows we describe the compilation of the data usedbelow 24.5 GHz. All published flux densities used here werechecked for the calibration scale and, wherever necessary, werescaled to the common scale of Baars et al. (1977). At the highestfrequencies (24.5 and 32 GHz), the Baars scale is ∼
1% higherthan that properly extended to higher frequencies by Perley &Butler (2013). This small dii ff erence does not a ff ect our analysis.A careful selection was made to ensure that the most reliabledata would be used. We discarded interferometric measurementsthat might underestimate the flux density as a result of miss-ing short spacings (mostly relevant above about 1.4 GHz), butalso single-dish measurements that might overestimate the fluxdensity owing to source confusion (mostly relevant below about5 GHz). If several measurements are available at the same ora neighbouring frequency, we calculated the error-weighted av-erage at the corresponding mean frequency. The resulting datacompilation is presented in tabular form in App. A. For eachgalaxy, the tables give the mean frequency in GHz (Col. 1), theflux density (Col. 2) and its error (Col. 3) in mJy, and the refer-ence to the flux density measurement (Col. 4). In a few cases, wehave determined flux densities at 325 MHz and 1.4 GHz usingthe Westerbork Northern Sky Survey (WENSS) (Rengelink et al.1997) and the NRAO VLA Sky Survey (NVSS) (Condon et al.1998), respectively. These are galaxies with small angular sizesso that flux loss by interfreometric measurements is not an issue.Finally, some flux densities were recently obtained in the courseof other observing programmes (VLA, E ff elsberg).We decided to utilize only flux densities above 0.3 GHz,since at lower frequencies a number of astrophysical e ff ectssuch as thermal absorption may shape the synchrotron spectrain such a way as to produce deviations from a simple power-law.It would require more free parameters, rendering the fit resultsless reliable when taking these processes into account. The syn-
3. Klein et al.: Synchrotron spectra of galaxies chrotron spectra obtained here may serve though to extrapolatethem to lower frequencies and thus provide a firm ‘leverage’ forlow-frequency studies.
Apart from the radio data, we collected a set of ancillary datafrom the literature for the galaxies that allowed us to quantifytheir star formation rate (SFR), H α flux, stellar mass, metallicity,and optical sizes. The data, together with the distances, are listedin Table 1.As a measure of the stellar mass we used the infrared K S luminosity, L K , which we calculated from the total (extrapolated) K S flux, f K , as L K = ν f K ( ν ) 4 π D (where D is the distance and ν is the frequency of the K-band , 1.38 × Hz). The fluxes inthe K S (2.17 µ m) band were taken from the 2MASS ExtendedSource Catalog (Jarrett et al. 2000), the 2MASS Large GalaxyAtlas (Jarrett et al. 2003), and from the 2Mass Extended ObjectsFinal Release.We calculated the star formation rate (SFR) as a combinationof 24 µ m and H α fluxes, following Kennicutt et al. (2009):S FR = . × − L (H α ) + L (24 µ m)erg s − M (cid:12) yr − . (2)For the 24 µ m flux we used, whenever possible, the flux mea-sured with MIPS on the Spitzer satellite. Only in those casesfor which no
Spitzer data was available (IC 10 and NGC 4038),data from the IRAS satellite at 25 µ m were used, neglecting thesmall central wavelength di ff erence. By combining mid-infraredand H α fluxes we take into account both unobscured and dust-enshrouded star formation and circumvent the uncertain extinc-tion correction of the H α flux.
4. Spectral fitting
Following Eqn. 1 we fitted the radio data with the sum of thermaland nonthermal radio emission.In order to predict the synchrotron emission of a modelgalaxy, the first step is to calculate the relativistic electron parti-cle density, N ( E ), i.e. the number density of relativistic electronsat energy E within the energy interval dE . The synchrotron emis-sion can then be calculated by convolving this distribution withthe synchrotron emission of a single electron: S nth ( ν ) ∝ (cid:90) ∞ (cid:32) νν c (cid:33) . e − ν/ν c N ( γ ) d γ, (3)where γ = E / ( m e c ) is the Lorentz factor, m e the electron restmass and ν c = π e B ⊥ m e c γ (4)is the critical frequency, e is the electron charge and B ⊥ thestrength of the magnetic field component perpendicular to thedirection of the relativistic electron’s velocity. The critical fre-quency is close to the peak of the electron spectrum and rep-resents, within a factor of a few, the frequency where the rela-tivistic electrons emit most of their energy (see, e.g., Klein & Note that - unfortunately - there exits definitions for the K-band inboth, optical and radio astronomy, both of which have to be used in thepresent paper and must not be confused.
Fletcher 2015). The shape of the synchrotron spectrum emittedby a single electron is such that for frequencies above ν c theemission decreases exponentially, whereas for lower frequenciesthere is power-law, ν . .If the electron energy distribution is a power-law, N ( E ) ∝ E − g , we can calculate the synchrotron emission to agood accuracy by assuming that an electron with E emits the en-tire synchrotron radiation at the critical frequency (see Klein &Fletcher 2015), S nth ( ν ) = N [ E ( ν c )] dEdt (cid:12)(cid:12)(cid:12)(cid:12)(cid:12) syn · dEd ν ∝ B g + ⊥ ν − g − . (5)In the following we are going to explain the four di ff erentmodels that we considered for the synchrotron spectrum. Wechose these models with the goal to cover the range from aspectrum with constant slope to a maximally curved synchrotronspectrum within simplified yet realistic scenarios. Spectrum with a constant slope:
In the first case we adopta synchrotron spectrum with a constant slope over the entire fre-quency range, S nth ( ν ) = S nth , (cid:32) νν (cid:33) − α nth , (6)where we take ν = ff erent scenarios. Spectrum curved due to energy losses:
Here, we considera steady-state, closed-box model. Since we are only interested inthe total N ( E ), integrated over the entire galaxy, we do not needto take into account propagation of the relativistic particles, aslong as it is energy-independent. In this approximation, the cor-responding equation for the relativistic electron particle densityis: ∂∂ E (cid:20) b ( E ) N ( E ) (cid:21) = (cid:18) Em e c (cid:19) − g inj q S N ν S N . (7)This equation takes into account acceleration in super-nova remnants (SNRs) with a source spectrum as γ − g inj q S N ν S N (where ν S N is the supernova rate, q S N is the number of relativis-tic electrons produced per supernova and per unit energy inter-val, and g inj is the injection spectral index), and the radiative en-ergy losses of the relativistic electrons dE / dt = b ( E ). The injec-tion spectral index, g inj is predicted by shock acceleration theoryto be 2.1 (Drury et al. 1994; Berezhko & V¨olk 1997). Eqn. 7 canbe solved by integration and gives N ( E ) = (cid:18) Em e c (cid:19) − g inj + q S N ν S N m e c b ( E )( g inj − . (8)Thus, the spectral shape of N ( E ) and of the resulting syn-chrotron emission depend on the energy dependence of the en-ergy losses. The most relevant energy losses for CR electrons inthe GHz range are inverse-Compton and synchrotron losses, b ( E ) syn + iC = d E d t (cid:12)(cid:12)(cid:12)(cid:12)(cid:12) syn + iC = C syn + iC E , (9)with C syn + iC ∝ ( U rad + U B ), U rad being the energy density ofthe radiation field below the Klein-Nishina limit, and U B theenergy density of the magnetic field. We furthermore consider
4. Klein et al.: Synchrotron spectra of galaxies
Table 1. Ancillary data for the galaxy sample
Galaxy
D d + log(O / H) Ref. log( L K ) log( L µ m ) F (H α ) Ref. log(SFR) [Mpc] [kpc] [ L (cid:12) , k ] [erg s − ] [erg s − cm − ] [M (cid:12) yr − ]IC 10 0.9 1.6 8.30 1 8.79 40.61 − .
63 2 − . − .
81 1 0.02II Zw 70 23.1 4.8 7.86 3 8.94 41.83 − .
08 1 − . − .
05 2 − . − .
64 2 − . − .
86 2 0.03NGC 4631 6.3 26.4 8.75 6 10.35 42.69 − .
70 2 0.02NGC 5194 7.6 30.2 8.86 6 10.90 43.03 − .
79 2 0.26NGC 4038 21.3 33.4 8.74 8 11.12 43.65 − .
93 3 2.01NGC 6052 70.4 16.8 8.85 10 10.76 43.74 − .
66 3 1.04NGC 2146 22.4 34.5 8.68 5 11.21 44.11 − .
18 3 1.20NGC 3034 3.8 12.1 9.12 6 10.63 43.85 − .
11 2 0.93NGC 3079 19.1 45.3 8.89 7 10.99 43.22 − .
42 3 0.35NGC 3310 18.1 9.4 8.75 7 10.40 43.39 − .
95 3 0.64 (1)
Optical diameter in kpc, calculated from d in LEDA. (2) References for the oxygen abundance: (1) Magrini & Gonc¸alves (2009), (2) Thuan & Izotov (2005), (3) Kehrig et al. (2008), (4) Kobulnicky &Skillman (1997), (5) Engelbracht et al. (2008), (6) Moustakas et al. (2010), (7) Robertson et al. (2013), (8) Bastian et al. (2009), (9) Pilyugin &Thuan (2007), (10) Sage et al. (1993). (3)
Decimal logarithm of the luminosity in the K-band in units of the solar luminosity in the K S -band ( L K , (cid:12) = . × erg s − ). (4) Decimal logarithm of the luminosity at 24 µ m, derived from Spitzer
MIPS fluxes as L µ m = ν f µ m × π D . The fluxes were obtained fromEngelbracht et al. (2008, II Zw 40, NGC 1569, NGC 2146, NGC 3310, NGC 4449, NGC 3079), Dale et al. (2009, NGC 3034, NGC 4490,NGC 4631, NGC 5194), Brown et al. (2014, NGC 6052), or directly from the Spitzer archive (https: // irsa.ipac.caltech.edu) (II Zw 70). For twogalaxies (IC 10 and NGC 4038), no Spitzer
MIPS data were available and we used IRAS 25 µ m instead, neglecting the slight di ff erence in thecentral wavelength. (5) Decimal logarithm of the H α flux, corrected for N ii contribution and for Galactic foreground extinction. (6) References for the H α fluxes. (1) Gil de Paz et al. (2003) (2) Kennicutt et al. (2008), (3) Moustakas & Kennicutt (2006). (7) Decimal logarithm of the SFR, calculated with Eqn. 2. bremsstrahlung and adiabatic losses, which depend linearly on E and can thus become relevant at lower energies / frequencies. b ( E ) ad + brems = d E d t (cid:12)(cid:12)(cid:12)(cid:12)(cid:12) ad + brems = C ad + brems E . (10)Here C ad + brems is a constant that depends on the gas densityand advection velocity gradient. Taking both kinds of energylosses into account results in an energy distribution of the rel-ativistic electrons that exhibits a (very shallow) change of slopecentered at the break energy, E b = C ad + brems / C syn + iC , N ( E ) ∝ E − g inj E / E b + , (11)The corresponding curved synchrotron spectrum reads S nth ( ν ) = S nth , (cid:16) νν (cid:17) − α nth (cid:16) νν b (cid:17) . + , (12)where ν b = ν ( γ b ) is the break frequency at which the changeof slope takes place, and α nth = g inj − is the low-frequencysynchrotron spectral index. Spectrum with a break:
We can produce a much more pro-nounced break in an open-box model, in which the relativis-tic electrons can escape from a finite halo with vertical extent z halo (see Lisenfeld et al. 2004, for a detailed discussion). Thesharpest break occurs when the relativistic electrons propagateby convection because the scalelength (i.e. the distance overwhich relativistic electrons can propagate before losing their en-ergy) of di ff usion, being a stochastic process, has a much weakerdependence on energy and therefore produces a shallower break. We only take into account inverse-Compton and synchrotronlosses, according to Eqn. 9. The propagation equation for therelativistic electrons accelerated in SNRs in the galactic planeat z = V c in thez-direction perpendicular to the disk is in this case ∂ N ( E , z ) ∂ z V c − ∂∂ E (cid:20) C syn + iC E N ( E , z ) (cid:21) = δ ( z ) (cid:18) Em e c (cid:19) − g inj q SN ν SN , (13)with δ ( z ) being the one-dimensional δ -function. The solution ofthis equation is given by N ( E , z ) = q SN ν SN V c (cid:18) Em e c (cid:19) − g inj (cid:18) − zz max (cid:19) g inj − z < z max z ≥ z max (14)where z max = V c / E C syn + iC is the maximum distance that a rela-tivistic electron can travel before having lost its energy to below E . The total relativistic electron density in the galaxy is calcu-lated by integrating Eqn.14 from z = z = min( z max , z halo ). Wethen assume that the relativistic electrons emit all their energy at ν c and obtain for the synchrotron emission, S nth ( ν ) = S nth , (cid:16) νν (cid:17) − α nth − . (cid:20) − (cid:32) − (cid:113) νν b (cid:33) g inj − (cid:21) ν ≤ ν b S nth , (cid:16) νν (cid:17) − α nth − . ν > ν b (15)The resulting spectrum has a break at ν b = ν c ( E b ), with breakenergy E b = V c / ( z halo C syn + iC ). The spectral index changes from α nth + . ν > ν b to α nth for ν (cid:28) ν b , i.e. it changes by 0.5.The break produced is much more pronounced than the shallow
5. Klein et al.: Synchrotron spectra of galaxies change of spectral index due to di ff erent energy losses (this canbe clearly seen in the figures in App. B where the synchrotronspectra of the di ff erent models are shown and can be compared.) Spectrum with an exponential cuto ff : Finally, we con-sider the case that the distribution of relativistic electrons endsabruptly at a certain Lorentz factor γ max . Assuming continuouspitch-angles, thus following the model of Ja ff e & Perola (1973),an exponential cuto ff is produced in the synchrotron spectrumaccording to Eqn. 5 whenever there is an abrupt cuto ff in the en-ergy spectrum of the relativistic electrons (Ja ff e & Perola 1973;Kardashev 1962). A cuto ff in energy can occur in a single-injection scenario because the highest-energy electrons lose theirenergy much faster than lower-energy electrons so that their pop-ulation becomes completely depopulated after a characteristicenergy loss time. The location of the exponential cuto ff in theradiation spectrum depends on the relative time scales for accel-eration, energy losses, and escape of the particles (Schlickeiser1984). The synchrotron spectrum in this case is S nth ( ν ) = S nth , (cid:32) νν (cid:33) − α nth e − νν b (16)Table 2 summarizes the four models, together with the freeparameters determined in the spectral fitting. For each galaxy, we fit four models: constant, curved, break, andcuto ff . The shapes of the model radio spectra are presented in theprevious subsection and recapitulated in Table 2, along with thelist of free parameters. The only fixed parameter was the slope ofthe (optically thin) thermal emission, i.e. α th = .
1. We require S th to be positive for all four models. In addition, we provide thefollowing bounds for some parameters: S nth ≥ ≤ ν b ≤
50 GHz for the break model and ν b ≥ ff model.For each model, the best fit is obtained via the Trust RegionReflective (”trf”) algorithm for optimization, well suited to ef-ficiently explore the whole space of variables for a bound-constrained minimization problem (Branch et al. 1999). InApp. B, we provide more detailed information and we show thebest fits of all four models (constant, curved, break, and cuto ff )for an easier comparison, together with a table of their best-fitparameters.
5. Results
Fig. 1 shows the best fits out of the four tested cases (constant,curved, break, or cuto ff , solid red line). In these plots, we alsodisplay the free-free (thermal, dotted cyan line) and synchrotron(nonthermal, dashed green line) components of the models. Theobserved radio continuum flux densities minus the fitted non-thermal component is represented by cyan stars, while the ob-served minus the fitted thermal component is depicted by greencrosses. The optimal values of the parameters are listed in thelower left part of the figure, and their errors are given in termsof the standard deviation of 1000 generated models (for moredetailed information see App. B).In Table 3 we compile the fit results obtained from the best-fit model, with the total flux density at 1 GHz and its error listedin Col. 2, the fraction of thermal emission at 1 GHz and its errorin Col. 3, the nonthermal spectral index and its error in Col. 4,the break frequency in GHz and its error in Col. 5, the reduced χ ν in Col. 6, and the name of the best-fit model in Col. 7. In general, the best fit was selected as the lowest reducedchi-squared, χ ν . Apart from this, the resulting thermal emissionwas an asset of whether the fits are physically meaningful. Thus,we rejected fits in spite of an acceptable χ ν if the fitted thermalradio emission was unphysically low. In addition, an importanttest was to look at the resulting thermal flux densities obtained ateach frequency after subtracting the computed nonthermal fluxdensity from the total (observed). A fit makes only sense if theresulting thermal flux densities so obtained obey to the expectedoptically thin spectrum with slope − .
1. This is shown by theblue data points in each diagram, with the errors taken fromthose of the measured total flux densities.The most striking result of our fits is that most objects needstrongly curved (break or cuto ff ) synchrotron spectra in order tosatisfactorily fit the total radio spectrum. Apparent exceptionsare the three dwarf galaxies IC 10, II Zw 40, and II Zw 70.IC 10 and II Zw 70 can be well fitted by a constant synchrotronspectrum, while in case of II Zw 40 it is only at the low-est frequencies that nonthermal emission becomes evident. Wetherefore decided to fit a power-law only to its high-frequencydata, obtaining an optically thin thermal spectrum with slope α th = . ± .
02. This spectrum is then extrapolated to the lowerfrequencies at which we retrieve the nonthermal flux densities bysubtracting the extrapolation from the measured fluxes. In Fig. 1the dashed cyan line shows the fit to the three high-frequencydata, yielding a perfect optically thin thermal spectrum. At thetwo low frequencies we have subtracted the extrapolated thermalflux densities from those observed, such as to yield the nonther-mal spectrum.We have also produced fits to this spectrum in which α nth islimited to 1.0, 0.9, 0.8, 0.7, and 0.6. We note that α nth may takevalues as low as 0.8, and the total fit is still compatible with theerror bars of the observational data. The thermal spectrum is alittle lower in this case and, as expected, χ ν is worse. Values of0.7 and 0.6 make the total fit to go below the first data point. Wealso tried to fit with maximum values of 0.5 and 0.4, but the fitsdid not converge.Obviously, the BCDGs and IC 10 are dominated by the ther-mal radio emission ( f th ≥ . / curved fits are rather low. Compared to the H α datathey would require internal extinctions of 2 . m . m ff . For 5 galaxies, the χ ν values arevery similar in both cases (within a factor of 1.5), and for all10 galaxies both break and cuto ff give acceptable fits. In addi-tion, the broad radio range covered by our data in most of thesegalaxies renders both, the steepening of the spectrum startingat ∼ − GHz, due to the strong curvature of the synchrotronemission, and the subsequent flattening due to the dominanceof the thermal radio emission (above ∼
10 GHz) visible in themeasured spectra!
The most noticeable cases are NGC 2146,NGC 4449, NGC 4631, and NGC 5194, but also in NGC 1569,NGC 3034, and NGC 3310 this trend is visible. The fact that wedirectly observe the steepening and subsequent flattening in thedata lends support for the reality of the pronounced curvature ofthe synchrotron spectra that our fitting procedure yields.
6. Klein et al.: Synchrotron spectra of galaxies
Fig. 1. Radio continuum spectra of the sample galaxies. Except for II Zw 40 (see text for the procedure and plotted lines here), theplots show the following: measured flux densities (blue squares), best fit of the radio continuum spectrum (solid red line). Thebest parameters are listed in the lower left part of the figure along with the reduced chi-square ( χ ν ). The free-free (thermal) andsynchrotron (nonthermal) components of the models are depicted by the dotted cyan line and dashed green line, respectively. Thegreen crosses represent the observed nonthermal component (i.e., the modelled thermal component removed from the observationaldata). The cyan stars delineate the observed thermal component (i.e., the modelled nonthermal component subtracted from theobservational data). The vertical yellow dash-dotted line marks the break frequency ν b .
7. Klein et al.: Synchrotron spectra of galaxies
Table 2. Shapes of the radio (synchrotron and free-free) emission considered in the fitting process
Model name Radio spectrum free parametersconstant S th , (cid:16) νν (cid:17) − . + S nth , (cid:16) νν (cid:17) − α nth S th , , S nth , , α nth curved S th , (cid:16) νν (cid:17) − . + S nth , (cid:18) νν (cid:19) − α nth (cid:18) νν b (cid:19) . + S th , , S nth , , ν b , α nth break S th , (cid:16) νν (cid:17) − . + S nth , (cid:16) νν (cid:17) − α nth − . (cid:20) − (cid:32) − (cid:113) νν b (cid:33) g inj − (cid:21) for ν ≤ ν b S th , , S nth , , ν b , α nth S th , (cid:16) νν (cid:17) − . + S nth , (cid:16) νν (cid:17) − α nth − . for ν > ν b cuto ff S th , (cid:16) νν (cid:17) − . + S nth , (cid:16) νν (cid:17) − α nth e − νν b S th , , S nth , , ν b , α nth Notes.
We take ν = Table 3. Fit results
Galaxy S tot , f th , α nth ν b χ ν best fit[mJy] [GHz]II Zw 40 32 0 .
80 0 . − − -II Zw 70 5 . ± . . ± .
03 1 . ± . − .
33 constantIC 10 446 ±
40 0 . ± .
07 0 . ± . − .
18 constantNGC 1569 494 ±
12 0 . ± .
02 0 . ± .
02 12 . ± . .
12 cuto ff NGC 4449 342 ±
32 0 . ± .
03 0 . ± .
08 4 . ± . .
17 cuto ff NGC 4490 1051 ±
206 0 . ± .
04 0 . ± .
09 6 . ± . .
20 cuto ff NGC 4631 1637 ±
75 0 . ± .
01 0 . ± .
03 6 . ± . .
08 cuto ff NGC 5194 1788 ±
464 0 . ± .
03 0 . ± .
10 7 . ± . .
62 cuto ff NGC 4038 683 ±
160 0 . ± .
04 0 . ± .
09 8 . ± . .
01 breakNGC 6052 129 ±
10 0 . ± .
01 0 . ± .
04 2 . ± . .
32 breakNGC 2146 1359 ±
18 0 . ± .
01 0 . ± .
02 6 . ± . .
47 cuto ff NGC 3034 9043 ±
175 0 . ± .
01 0 . ± .
02 11 . ± . .
45 cuto ff NGC 3079 1111 ±
104 0 . ± .
02 0 . ± .
01 9 . ± . .
59 breakNGC 3310 470 ±
16 0 . ± .
01 0 . ± .
03 1 . ± . .
40 break8. Klein et al.: Synchrotron spectra of galaxies
6. Discussion α emission We have used the thermal flux densities resulting from our spec-tral decomposition to compare them with what is predicted fromthe observed H α fluxes corrected for N ii emission and Galacticextinction (Table 1). We used the relation derived by Lequeux(1980), converted from H β to H α : S th , H α = . · (cid:18) ν GHz (cid:19) − . (cid:18) T e K (cid:19) . (cid:34) F (H α )erg s − cm − (cid:35) mJy . (17)The extinction is then calculated via A (H α ) = − . (cid:32) S t h , H α S t h , f it (cid:33) . (18)Since extinction is caused by the dust in the galaxy planes,we have plotted in Fig. 2 the extinction resulting from Eqn. 18vs. the metallicity, for which we have collected the quantity12 + log(O / H) from the literature (Table 1). As is to be expected,there is a trend of increasing extinction with increasing metal-licity, with the highly inclined spiral galaxies having the largestextinctions. Dwarf galaxies are known to possess lower metal-licities and thus have a low dust content, hence they are foundin the lower left part of the diagram. The values for IC 10 andII Zw 40 are even negative, which is most likely due to an over-estimate of the Galactic extinction which is very high in bothcases (3 . m . m In Fig. 3 we show the fraction of thermal emission at 1 GHz re-sulting from our spectral fits, plotted vs. the K-band luminosityof the galaxies. Since this luminosity is a measure for the stellarmass, the diagram clearly indicates a decreasing relative amountof nonthermal radiation as we move to small stellar masses. Infact, the left half of this plot contains all the dwarf galaxies in oursample. This result corroborates the early conjecture of Kleinet al. (1991) that the lowest-mass galaxies are unable to retainthe recently produced cosmic rays. Alternatively, the high ther-mal fraction in the most extreme galaxies II Zw 40 and II Zw 70might be the result of a temporal e ff ect. Both galaxies are cur-rently experiencing a strong and very recent starburst, with anage of only 3 − ff yet. Hence, the observed syn-chrotron radiation may have been produced during a previousepoch of star formation. We also compare the thermal fractionwith the SFR (from Tab. 1 but found no correlation.It is important to note here that for some of the galaxies in oursample the amount of thermal emission resulting from our anal-ysis is higher than thought hitherto, which has consequences forany conjectures based upon the thermal radio continuum in star-forming galaxies. This is particularly true for higher frequencieswhere the cuto ff/ break lowers the synchrotron emission consid-erably. At 10 GHz, the mean thermal fraction for the galaxies is0.6 – considerably more than what was believed so far. In Fig. 4 a superposition of the low-frequency spectral indices ofthe synchrotron radiation of our galaxy sample with the spec-tral indices of Galactic supernova remnants (SNR) is shown,the latter data taken from the catalogue of Green (2014). Thetwo distributions are rather similar, albeit the statistics are vastlydi ff erent (203 vs. 14 objects). The resulting values of the meanand standard deviation are < α SNR > = . σ α SNR = .
33 and < α
Gal > = . σ α Gal = .
20 for the SNR and galaxies, respec-tively. Both, the mean and the rms are rather similar, possiblysuggesting that the synchrotron spectra that we measure at lowradio frequencies reflect the injection spectra of the SNR.An outlier is II Zw 70, which has a low-frequency spectralindex of α nth = . ± .
12. Such a steep spectrum is indicativeof the impact of synchrotron and inverse-Compton losses whichhave steepened the injection spectral index by + .
5. It is unclearwhy II Zw 70 is so di ff erent from the rest of the galaxies. TheSFR per L K , which can be taken as a measure of the capabilityof the galaxy to expell material into the halo, is similar to thatof the other starburst galaxies, such as M 82 or NGC 1569 (butmore than a factor of 10 lower than for the otherwise similarBCDG II Zw 40).Table 4. Radio sizes, magnetic fields, and cuto ff/ break energies Galaxy Radio Size B E Shape[kpc] [ µ G] [GeV]II Zw 40 0.8 29 - -II Zw 70 < <
40 - constantIC 10 1.3 14 - constantNGC 1569 2.5 16 6.8 cuto ff NGC 4449 5.0 14 4.6 cuto ff NGC 4490 10 20 4.6 cuto ff NGC 4631 22 13 5.6 cuto ff NGC 5194 15 18 5.0 cuto ff NGC 4038 18 27 - breakNGC 6052 4.0 71 - breakNGC 2146 10 40 3.1 cuto ff NGC 3034 1.7 66 3.3 cuto ff NGC 3079 13 32 - breakNGC 3310 4.4 39 - break (1)
The magnetic field is calculated with the minimum energy assump-tion.
For the majority of our galaxies, the fitting of their radio spectrashows the need of a break of the synchrotron spectra in the rangeof 1 −
12 GHz, corresponding to particle energies of 1 . − ∼ ff ).As outlined in Sect. 4.1, a sharp break can be explained in anopen box-model with convection. The break is produced becausehigh-energy electrons, emitting at frequencies above the break,
9. Klein et al.: Synchrotron spectra of galaxies
Fig. 1. (continued)
10. Klein et al.: Synchrotron spectra of galaxies
Fig. 1. (continued)Fig. 2. H α extinction vs. metallicity. Highly inclined ( >
75 de-grees) and edge-on galaxies are marked with red dotssu ff er substantial synchrotron or inverse-Compton losses beforethey arrive at the edge of the halos, whereas low-energy elec-trons do not. We can estimate the half-lifetime of the relativisticelectrons emitting the synchrotron radiation at the break freqe-uncy, which is determined by synchrotron and inverse-Comptonlosses (see, e.g., Klein & Fletcher 2015): t / = . · · (cid:32) B µ G (cid:33) + (cid:32) B eq µ G (cid:33) − · (cid:18) E GeV (cid:19) − yr . (19)Here, E is the energy of the relativistic particles, B is the totalmagnetic-field strength, and B eq is the field strength equivalent Fig. 3. Fraction of thermal emission at 1 GHz vs. K-band lumi-nosity, the errors resulting from our fits (see Table 3).to the energy density of the local radiation field. The latter canbe obtained via B eq = (cid:114) c · L / bol d . (20)Here, L bol is the bolometric luminosity of a region of size d . Kennicutt et al. (2007) have measured the quantity ν · L ν at24 µ m (which approximates the luminosity) for a number of H ii regions in M 51, with typical values of 10 erg s − , obtainedwith aperture sizes of 13 (cid:48)(cid:48) (hence, d =
500 pc). Plugging thisinto Eqn. (20), we obtain B eq ≈ µ G. The strength of the mag-netic field in our sample galaxies is between B = µ G and
11. Klein et al.: Synchrotron spectra of galaxies
Fig. 4. Histogram of the spectral indices of the low-frequencysynchrotron radiation of our sample galaxies (red) and ofGalactic supernova remnants (black). The vertical dashed linesindicate the mean (thick) and variance (thin) of each distribution. B = µ G (Table 4). This implies a range for the particle half-lifetimes of between 7 . · yr and 2 . · yr (Tev and PeVparticles enter the GeV regime on time scales much shorter thanany dynamical time scales under conditions considered above).Assuming a vertical halo size of about 250 pc, this implies con-vection speeds of between 10 and 300 km s − , which is reason-able. In order to test this interpretation, it would be useful tocarry out a similar analysis for quiescent galaxies with a lowstar-formation rate per area for which we would expect muchlower vertical propagation of the relativistic electrons.A cuto ff in the relativistic electron distribution is more dif-ficult to explain. Energy losses can produce a cuto ff only ina single-injection scenario where radiative ageing depopulatesthe high-energy part of the relativistic electron distribution withtime. This situation is unrealistic for the integrated radio emis-sion of an entire galaxy. The acceleration process of cosmic raysin SNRs is e ff ective up to TeV energies and the synchrotronspectra of young SNRs can be observed up to X-ray energies.At high energies, the short synchrotron energy loss time in thestrong magnetic fields ( > µ G) in SNR (V¨olk et al. 2005) isexpected to produce a steepening of the electron injection spec-trum. This can be seen from Eqn. 19, which for a particle energyof, say, 100 GeV, in a 100 µ G magnetic field yields a lifetimeof ∼ yr. This is comparable to the Sedov phase of a SNR( τ ∼ · yr) so that at energies above ∼
100 GeV synchrotroncooling should be relevant and produce a steeper injection spec-trum above ∼ · Hz. However, the breaks / cuto ff s that weinfer are at much lower particle energies and do therefore notprobe this process. Thus, in the framework of the simple mod-els considered here, it is unclear which process could produce acuto ff in the synchrotron spectrum. It is, however, obvious thatthis must have to do with the relative time scales of energy gain,losses and escape of the relativistic electrons, as discussed bySchlickeiser (1984).
7. Summary and conclusions
We have analyzed the radio continuum spectra of 14 star-forming galaxies by fitting nonthermal (synchrotron) and ther-mal (free-free) radiation laws to carefully selected measure-ments, covering a frequency range of ∼
300 MHz to 24.5 GHz(32 GHz in case of M 82). The 24.5-GHz measurements, mostlyunpublished to date, are crucial for a more reliable separation ofthe thermal and nonthermal components in this analysis.We find that the majority of the synchrotron spectra are notsimple power-laws as believed hitherto. The curved shape of thesynchrotron spectrum is clearly visible in many of the total radiospectra, which steepen over a range of several GHz and flattenat higher frequencies due to the thermal radio emission. Our fit-ting shows that the lowest values of the reduced χ ν result forsynchrotron spectra with a mean slope α nth = . ± .
20 in thelow-frequency regime, and a break or an exponential decline inthe frequency range of 1 −
12 GHz. There are only one galaxythat shows pure power-laws (the dwarf galaxy IC 10). In the caseof the BCDGs II Zw 40 and II Zw 70 only the slope (and not theshape) of the synchrotron spectrum can be worked out, since adeviation from a purely thermal spectrum is only evident at thelowest frequencies involved.For the bulk of the sample galaxies, simple power-laws ormildly curved synchrotron spectra lead to unrealistic low ther-mal flux densities (frequently S th = / or to strong devi-ations from the expected optically thin free-free spectra withslope α th = .
10 in the fits. Assuming energy equipartitionbetween relativistic particles and magnetic fields, the cuto ff and break frequencies translate into energies in the range of1 . − α fluxes yields the extinc-tion, which increases with metallicity. The fraction of thermalemission at 1 GHz is higher in some of our galaxies than believedhitherto, and the discrepancy increases towards higher frequen-cies where the mean thermal fraction is ∼ .
6. It is highest inthe dwarf galaxies of our sample, which we interpret either interms of a lack of containment of synchrotron emission in theselow-mass systems, or alternatively, in the case of II Zw 40 andII Zw 70, as a time e ff ect due to a very young starburst.The asymptotic low-frequency synchrotron spectra derivedhere provide a firm leverage for low-frequency studies, e.g. withLOFAR. In order to significantly incraese the number of ra-dio continuum spectra of galaxies allowing analyses as pre-sented here, expensive mapping with large single-dish telescopes(E ff elsberg, GBT, Sardinia Radio Telescope) in the frequencyrange of 25 – 40 GHz (so-called Ka-band in radio astronomy)are indispensible.An interpretation of the rapid declines in the synchrotronspectra (break or exponential cuto ff ) at GeV particle energiesis not obvious, since the relativistic particles gain energies upto the TeV range within the supernovae. In a simple model, aconvective wind carrying away the low-energy relativistic elec-trons could explain a sharp break. The break energy must dependon the relative time scales of energy gain (acceleration), losses(synchrotron and inverse-Compton) and escape of the relativisticelectrons. Acknowledgements. We wish to thank H. Lesch and R. Schlickeiser for help-ful discussions. UK acknowledges financial support by the German DeutscheForschungsgemeinschaft, DFG project FOR 1254, and is very grateful forthe kind hospitality at the Departamento de F´ısica Te´orica y del Cosmos,
12. Klein et al.: Synchrotron spectra of galaxies
Universidad de Granada. UL and SV acknowledge support by the researchprojects AYA2014-53506-P from the Spanish Ministerio de Econom´ıa yCompetitividad, from the European Regional Development Funds (FEDER)and the Junta de Andaluc´ıa (Spain) grants FQM108. This research has madeuse of the NASA / IPAC Extragalactic Database (NED), which is operatedby the Jet Propulsion Laboratory, California Institute of Technology, undercontract with the National Aeronautics and Space Administration. We alsoacknowledge the use of the HyperLeda database (http: // leda.univ-lyon1.fr).This research made use of astropy , a community-developed core python ( ) package for Astronomy (Astropy Collaborationet al. 2013); ipython (P´erez & Granger 2007); matplotlib (Hunter 2007); numpy (van der Walt et al. 2011); scipy (Jones et al. 2001–). Finally, we are very greatfulto the referee for her / his comments, which helped to improve the manuscript. References
Adebahr, B., Krause, M., Klein, U., et al. 2013, A&A, 555, A23Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558,A33Baars, J. W. M., Genzel, R., Pauliny-Toth, I. I. K., & Witzel, A. 1977, A&A, 61,99Balkowski, C., Chamaraux, P., & Weliachew, L. 1978, A&A, 69, 263Bastian, N., Trancho, G., Konstantopoulos, I. S., & Miller, B. W. 2009, ApJ, 701,607Basu, A., Mao, S. A., Kepley, A. A., et al. 2017, MNRAS, 464, 1003Beck, R., Klein, U., & Wielebinski, R. 1987, A&A, 186, 95Becker, R. H., White, R. L., & Edwards, A. L. 1991, ApJS, 75, 1Berezhko, E. G. & V¨olk, H. J. 1997, Astroparticle Physics, 7, 183Branch, M. A., Coleman, T. F., & Li, Y. 1999, SIAM J. Sci. Comput., 21, 1Braun, R., Oosterloo, T. A., Morganti, R., Klein, U., & Beck, R. 2007, A&A,461, 455Brown, M. J. I., Moustakas, J., Smith, J.-D. T., et al. 2014, ApJS, 212, 18Chy˙zy, K. T. & Beck, R. 2004a, A&A, 417, 541Chy˙zy, K. T. & Beck, R. 2004b, A&A, 417, 541Chy˙zy, K. T., Beck, R., Kohle, S., Klein, U., & Urbanik, M. 2000, A&A, 355,128Chy˙zy, K. T., Drzazga, R. T., Beck, R., et al. 2016, ApJ, 819, 39Chy˙zy, K. T., Knapik, J., Bomans, D. J., et al. 2003, A&A, 405, 513Chy˙zy, K. T., We˙zgowiec, M., Beck, R., & Bomans, D. J. 2011, A&A, 529, A94Clemens, M. S., Alexander, P., & Green, D. A. 1999, MNRAS, 307, 481Condon, J. J. 1987, ApJS, 65, 485Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998, AJ, 115, 1693Dale, D. A., Cohen, S. A., Johnson, L. C., et al. 2009, ApJ, 703, 517de Bruyn, A. G. 1977, A&A, 58, 221de Jong, M. L. 1967, ApJ, 150, 1Deeg, H.-J., Brinks, E., Duric, N., Klein, U., & Skillman, E. 1993, ApJ, 410, 626Drury, L. O., Aharonian, F. A., & Voelk, H. J. 1994, A&A, 287, 959Dumke, M., Krause, M., Wielebinski, R., & Klein, U. 1995, A&A, 302, 691Duric, N., Bourneuf, E., & Gregory, P. C. 1988, AJ, 96, 81Duric, N. & Seaquist, E. R. 1988, ApJ, 326, 574Duric, N., Seaquist, E. R., Crane, P. C., Bignell, R. C., & Davis, L. E. 1983, ApJ,273, L11Duric, N., Seaquist, E. R., Crane, P. C., & Davis, L. E. 1986, ApJ, 304, 82Ekers, R. D. & Sancisi, R. 1977, A&A, 54, 973Emerson, D. T., Klein, U., & Haslam, C. G. T. 1979, A&A, 76, 92Engelbracht, C. W., Rieke, G. H., Gordon, K. D., et al. 2008, ApJ, 678, 804Fletcher, A., Beck, R., Shukurov, A., Berkhuijsen, E. M., & Horellou, C. 2011,MNRAS, 412, 2396Gil de Paz, A., Madore, B. F., & Pevunova, O. 2003, ApJS, 147, 29Gioia, I. M. & Gregorini, L. 1980, A&AS, 41, 329Gioia, I. M., Gregorini, L., & Klein, U. 1982, A&A, 116, 164Golla, G. 1999, A&A, 345, 778Green, D. A. 2014, ArXiv e-printsGregory, P. C. & Condon, J. J. 1991, ApJS, 75, 1011Gri ffi th, M. R., Wright, A. E., Burke, B. F., & Ekers, R. D. 1994, ApJS, 90, 179Haynes, R. F., Huchtmeier, W. K. G., & Siegman, B. C. 1975, A compendium ofradio measurements of bright galaxiesHeidmann, J. 1979, in Extragalactic Astronomy - Meeting “Sol-Espace”, ed.C. Balkowski-Mauger, 205Heidmann, J., Klein, U., & Wielebinski, R. 1982, A&A, 105, 188Holtzman, J. A., Watson, A. M., Mould, J. R., et al. 1996, AJ, 112, 416Hummel, E. 1980, A&AS, 41, 151Hummel, E. 1991, A&A, 251, 442Hummel, E., Beck, R., & Dahlem, M. 1991, A&A, 248, 23Hummel, E. & Dettmar, R.-J. 1990a, A&A, 236, 33Hummel, E. & Dettmar, R.-J. 1990b, A&A, 236, 33 Hummel, E., Pedlar, A., Davies, R. D., & van der Hulst, J. M. 1985, A&AS, 60,293Hunter, J. D. 2007, Computing In Science & Engineering, 9, 90Irwin, J., Beck, R., Benjamin, R. A., et al. 2012, AJ, 144, 44Irwin, J. A. & Saikia, D. J. 2003, MNRAS, 346, 977Israel, F. P. & de Bruyn, A. G. 1988, A&A, 198, 109Israel, F. P. & Mahoney, M. J. 1990, ApJ, 352, 30Israel, F. P. & van der Hulst, J. M. 1983, AJ, 88, 1736Ja ff e, W. J. & Perola, G. C. 1973, A&A, 26, 423Ja ff e, W. J., Perola, G. C., & Tarenghi, M. 1978, ApJ, 224, 808Jarrett, T. H., Chester, T., Cutri, R., et al. 2000, AJ, 119, 2498Jarrett, T. H., Chester, T., Cutri, R., Schneider, S. E., & Huchra, J. P. 2003, AJ,125, 525Jones, E., Oliphant, T., Peterson, P., et al. 2001–, SciPy: Open source scientifictools for Python, [Online; accessed ¡today¿]Kardashev, N. S. 1962, Soviet Ast., 6, 317Kaz`es, I., Le Squeren, A. M., & Nguyen-Quang-Rieu. 1970, Astrophys. Lett., 6,193Kehrig, C., V´ılchez, J. M., S´anchez, S. F., et al. 2008, A&A, 477, 813Kellermann, K. I., Pauliny-Toth, I. I. K., & Williams, P. J. S. 1969, ApJ, 157, 1Kennicutt, Jr., R. C., Calzetti, D., Walter, F., et al. 2007, ApJ, 671, 333Kennicutt, Jr., R. C., Hao, C.-N., Calzetti, D., et al. 2009, ApJ, 703, 1672Kennicutt, Jr., R. C., Lee, J. C., Funes, Jos´e G., S. J., Sakai, S., & Akiyama, S.2008, ApJS, 178, 247Kepley, A. A., M¨uhle, S., Everett, J., et al. 2010, ApJ, 712, 536Klein, U. 1983, A&A, 121, 150Klein, U. 1988, PhD thesis, Habilitation Thesis, Univ. Bonn, (1988)Klein, U. & Emerson, D. T. 1981, A&A, 94, 29Klein, U. & Fletcher, A. 2015, Galactic and Intergalactic Magnetic FieldsKlein, U. & Gr¨ave, R. 1986, A&A, 161, 155Klein, U., Gr¨ave, R., & Wielebinski, R. 1983, A&A, 117, 332Klein, U., Hummel, E., Bomans, D. J., & Hopp, U. 1996, A&A, 313, 396Klein, U., Weiland, H., & Brinks, E. 1991, A&A, 246, 323Klein, U., Wielebinski, R., & Beck, R. 1984a, A&A, 135, 213Klein, U., Wielebinski, R., & Morsi, H. W. 1988, A&A, 190, 41Klein, U., Wielebinski, R., & Thuan, T. X. 1984b, A&A, 141, 241Kobulnicky, H. A. & Skillman, E. D. 1997, ApJ, 489, 636Kregel, M. & Sancisi, R. 2001, A&A, 376, 59Lequeux, J. 1971, A&A, 15, 30Lequeux, J. 1980, in Saas-Fee Advanced Course 10: Star Formation, ed.A. Maeder & L. Martinet, 9999Lisenfeld, U., Alexander, P., Pooley, G. G., & Wilding, T. 1996, MNRAS, 281,301Lisenfeld, U., Wilding, T. W., Pooley, G. G., & Alexander, P. 2004, MNRAS,349, 1335Maehara, H., Inoue, M., Takase, B., & Noguchi, T. 1985, PASJ, 37, 451Magrini, L. & Gonc¸alves, D. R. 2009, MNRAS, 398, 280Mao, S. A., Zweibel, E., Fletcher, A., Ott, J., & Tabatabaei, F. 2015, ApJ, 800,92Marvil, J., Owen, F., & Eilek, J. 2015, AJ, 149, 32McCutcheon, W. H. 1973, AJ, 78, 18Mora, S. C. & Krause, M. 2013, A&A, 560, A42Moustakas, J. & Kennicutt, Jr., R. C. 2006, ApJS, 164, 81Moustakas, J., Kennicutt, Jr., R. C., Tremonti, C. A., et al. 2010, ApJS, 190, 233Mulcahy, D. D., Horne ff er, A., Beck, R., et al. 2014, A&A, 568, A74Neininger, N. 1992, A&A, 263, 30Nikiel-Wroczy´nski, B., Jamrozy, M., Soida, M., Urbanik, M., & Knapik, J. 2016,MNRAS, 459, 683Niklas, S., Klein, U., Braine, J., & Wielebinski, R. 1995, A&AS, 114, 21Niklas, S., Klein, U., & Wielebinski, R. 1997, A&A, 322, 19P´erez, F. & Granger, B. E. 2007, Computing in Science and Engineering, 9, 21Perley, R. A. & Butler, B. J. 2013, ApJS, 204, 19Pfleiderer, J., Durst, C., & Gebler, K.-H. 1980, MNRAS, 192, 635Pilyugin, L. S. & Thuan, T. X. 2007, ApJ, 669, 299Purkayastha, A. 2014, PhD thesis, Thesis (Ph.D.) – University Bonn, 2014Rengelink, R. B., Tang, Y., de Bruyn, A. G., et al. 1997, A&AS, 124, 259Robertson, P., Shields, G. A., Dav´e, R., Blanc, G. A., & Wright, A. 2013, ApJ,773, 4Sage, L. J., Loose, H. H., & Salzer, J. J. 1993, A&A, 273, 6Schlickeiser, R. 1984, A&A, 136, 227Segalovitz, A. 1977, A&A, 54, 703Skillman, E. D. & Klein, U. 1988, A&A, 199, 61Slee, O. B. 1995, Australian Journal of Physics, 48, 143Sramek, R. 1975, AJ, 80, 771Srivastava, S., Kantharia, N. G., Basu, A., Srivastava, D. C., & Ananthakrishnan,S. 2014, MNRAS, 443, 860Sulentic, J. W. 1976, ApJS, 32, 171Tabatabaei, F. S., Schinnerer, E., Krause, M., et al. 2017, ApJ, 836, 185
13. Klein et al.: Synchrotron spectra of galaxies
Thuan, T. X. & Izotov, Y. I. 2005, ApJS, 161, 240Tovmassian, H. M. 1968, Australian Journal of Physics, 21, 193van der Hulst, J. M. 1979, A&A, 71, 131van der Kruit, P. C. & de Bruyn, A. G. 1976, A&A, 48, 373van der Walt, S., Colbert, S. C., & Varoquaux, G. 2011, Computing in ScienceEngineering, 13, 22Varenius, E., Conway, J. E., Mart´ı-Vidal, I., et al. 2015, A&A, 574, A114Viallefond, F., Allen, R. J., & de Boer, J. A. 1980, A&A, 82, 207V¨olk, H. J., Berezhko, E. G., & Ksenofontov, L. T. 2005, A&A, 433, 229Werner, W. 1988, A&A, 201, 1Wielebinski, R. & von Kap-Herr, A. 1977, A&A, 59, L17Williams, P. K. G. & Bower, G. C. 2010, ApJ, 710, 1462Wynn-Williams, C. G. & Becklin, E. E. 1986, ApJ, 308, 620
14. Klein et al.: Synchrotron spectra of galaxies
Appendix A: Flux densities
15. Klein et al.: Synchrotron spectra of galaxies
Table A.1. II Zw 40
Frequency Flux density Error Reference0.325 38 4 Deeg et al. (1993)1.459 30.1 0.4 Klein et al. (1991), Deeg et al. (1993)4.88 21.5 1.9 Ja ff e et al. (1978), Klein et al. (1984b), Klein et al. (1991)10.63 20.1 1.5 Skillman & Klein (1988), Klein et al. (1984b)24.500 18 4 Klein et al. (1984b) Table A.2. II Zw 70
Frequency Flux density Error Reference0.327 11.2 2.0 Skillman & Klein (1988)0.609 6.5 0.8 Skillman & Klein (1988)1.440 4.56 0.34 Wynn-Williams & Becklin (1986), Balkowski et al. (1978)4.795 3.04 0.13 Klein et al. (1984b), Wynn-Williams & Becklin (1986), Skillman & Klein (1988)10.700 2.73 0.20 Skillman & Klein (1988), Klein et al. (1984b)
Table A.3. IC 10
Frequency Flux density Error Reference0.327 352 20 WENSS, this work1.43 377 6 this work2.64 283 20 Chy˙zy et al. (2011) (recomputed)4.82 219 8 Klein et al. (1983), Klein & Gr¨ave (1986), Becker et al. (1991)10.575 162 8 Klein & Gr¨ave (1986), Chy˙zy et al. (2003)24.5 118 18 Klein & Gr¨ave (1986)
Table A.4. NGC 1569
Frequency Flux density Error Reference0.350 750 20 Purkayastha (2014); Ph.D. thesis, Univ. Bonn0.610 610 20 Israel & de Bruyn (1988)1.415 425 20 Hummel (1980), Israel & de Bruyn (1988)2.700 318 20 Pfleiderer et al. (1980), Sulentic (1976)4.800 233 20 Klein & Gr¨ave (1986), Gregory & Condon (1991)8.350 176 5 this work10.700 155 10 Klein & Gr¨ave (1986)15.36 116 13 Lisenfeld et al. (2004)24.500 96 8 Klein & Gr¨ave (1986)
Table A.5. NGC 2146
Frequency Flux density Error Reference0.327 2520 100 WENSS, this work1.43 1094 13 NVSS, Braun et al. (2007)2.695 720 70 Haynes et al. (1975)5.000 472 25 de Bruyn (1977)6.630 360 30 McCutcheon (1973)10.625 239 10 Niklas et al. (1995), Israel & van der Hulst (1983)24.500 167 10 this work16. Klein et al.: Synchrotron spectra of galaxies
Table A.6. NGC 3034
Frequency Flux density Error Reference0.327 13830 690 Adebahr et al. (2013)0.750 10700 500 Kellermann et al. (1969)1.388 7805 385 Hummel (1980), Adebahr et al. (2013)2.695 5700 300 Kellermann et al. (1969)5.000 3900 200 Kellermann et al. (1969)10.700 2250 60 Klein et al. (1988)14.700 1790 40 Klein et al. (1988)24.500 1190 30 Klein et al. (1988)32.000 1020 60 Klein et al. (1988)
Table A.7. NGC 3079
Frequency Flux density Error Reference0.327 2440 100 WENSS, this work0.615 1740 90 Irwin & Saikia (2003)1.365 800 10 Braun et al. (2007)2.640 526 30 this work4.750 377 20 Gioia et al. (1982)10.650 205 11 Gioia et al. (1982)24.500 123 8 this work
Table A.8. NGC 3310
Frequency Flux density Error Reference0.327 904 30 WENSS, this work0.610 650 50 van der Kruit & de Bruyn (1976)1.433 357 10 Condon (1987), Hummel et al. (1985)2.695 240 30 van der Kruit & de Bruyn (1976)4.750 146 8 Gioia et al. (1982)10.650 100 6 Gioia et al. (1982), Israel & de Bruyn (1988), Niklas et al. (1995)24.500 80 5 this work
Table A.9. NGC 4038 / Frequency Flux density Error Reference0.408 1250 60 Slee (1995)1.410 544 30 van der Hulst (1979), Hummel (1980), this work2.700 353 30 de Jong (1967), Tovmassian (1968), Kaz`es et al. (1970)4.850 246 17 Gri ffi th et al. (1994)10.550 146 7 Niklas et al. (1995), Chy˙zy & Beck (2004b) re-computed24.500 92 9 this work Table A.10. NGC 4449
Frequency Flux density Error Reference0.376 514 30 Purkayastha (2014)0.609 450 40 Klein et al. (1996)1.427 278 15 Condon (1987), NVSS, this work Klein & Emerson (1981)2.695 204 12 this work4.875 137 9 Klein & Gr¨ave (1986), Sramek (1975)10.650 87 5 Israel & van der Hulst (1983), Klein & Gr¨ave (1986), Klein et al. (1996)24.500 67 10 Klein & Gr¨ave (1986) 17. Klein et al.: Synchrotron spectra of galaxies
Table A.11. NGC 4490 / Frequency Flux density Error Reference0.327 2040 50 WENSS, this work0.408 2099 50 Gioia & Gregorini (1980)1.430 835 16 Lequeux (1971), Viallefond et al. (1980), Condon (1987)2.695 520 50 Kaz`es et al. (1970)4.81 331 13 Gioia et al. (1982), Nikiel-Wroczy´nski et al. (2016)6.630 262 40 McCutcheon (1973)10.700 185 27 Klein & Emerson (1981)24.500 106 10 Klein (1983)
Table A.12. NGC 4631
Frequency Flux density Error Reference0.327 2993 110 Hummel & Dettmar (1990b), WENSS, this work0.419 2900 71 Gioia & Gregorini (1980), Israel & van der Hulst (1983)0.610 2200 100 Ekers & Sancisi (1977), Werner (1988)0.835 2000 100 Israel & de Bruyn (1988)1.423 1282 20 Ekers & Sancisi (1977), Hummel & Dettmar (1990b), Braun et al. (2007)2.688 835 32 Wielebinski & von Kap-Herr (1977), Werner (1988)4.750 544 20 Wielebinski & von Kap-Herr (1977), Israel & van der Hulst (1983), Werner (1988)8.35 310 16 Mora & Krause (2013)10.625 260 10 Israel & van der Hulst (1983), Dumke et al. (1995)24.500 172 15 this work
Table A.13. NGC 5194 / Frequency Flux density Error Reference0.327 3812 150 WENSS, this work0.408 3640 700 Gioia & Gregorini (1980)0.610 2790 200 Segalovitz (1977)1.365 1420 10 Braun et al. (2007)2.695 780 50 Klein et al. (1984a)4.750 488 20 Gioia et al. (1982), Israel & van der Hulst (1983), Beck et al. (1987)8.35 306 26 Klein & Emerson (1981)10.700 239 11 Israel & van der Hulst (1983), Klein et al. (1984a)14.700 190 20 Klein et al. (1984a)22.800 142 15 Klein et al. (1984a)
Table A.14. NGC 6052
Frequency Flux density Error Reference0.325 244 20 Deeg et al. (1993)1.445 105 5 Deeg et al. (1993); NVSS, this work4.75 41.7 0.8 Klein et al. (1984b), Klein et al. (1991)10.7 22.8 2.2 Klein et al. (1991), Heidmann et al. (1982), Maehara et al. (1985)23.7 12.5 2.1 Heidmann et al. (1982), Klein et al. (1991)32.0 <
12 Klein et al. (1991)18. Klein et al.: Synchrotron spectra of galaxies , Online Material p 1
Appendix B: Fit plots
In this appendix, we show the best fits of all four models (con-stant, curved, break, and cuto ff ) for an easier comparison, alongwith the parameter spaces and we give the best-fit values forthe free parameters in Tab. B.1. The fitting method has been ex-plained in Sect. 4.2. In short, we have tested the four di ff erentmodels presented in Sect. 4.1 for each galaxy (which are rep-resented by the solid red line): (i) simple power-law (i.e. withconstant log-log slope), (ii) slightly curved law, (iii) power-lawwith a break, and (iv) power-law with an exponential decline.The only fixed parameter was the slope of the (optically thin)thermal emission, i.e. α th = .
1. The optimal values for the pa-rameters are listed in the lower left part of the figures, along withthe reduced chi-square, χ ν .Then, for a given model, we generate 1000 spectra similarto the observed one. At each frequency, the generated value isallowed to vary within the error of the observational data. Thisvariation follows a Gaussian distribution around the observeddata, and the observational error represents 3 σ of the Gaussiandistribution. The 1000 spectra generated randomly this way arerepresented by faint grey lines in the main plots of the App. B.Each generated spectrum is then fitted in the same way as theobservational data (see Sect. 4.2) and yields a value for eachfree parameter. For each model, the distribution of the possible1000 values taken by the parameters are shown in two parameterspace plots under the main plot, allowing an assessment of therobustness of the fits.In case of IC 10 (IIZw 40) , the least-squares minimizationfails for the break model (cuto ff model) so that this model cannotbe shown. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 2 c on s t a n t c u r v e db r ea k c u t o ff G a l S t o t f t h α n t h χ ν S t o t f t h α n t h ν b χ ν S t o t f t h α n t h ν b χ ν S t o t f t h α n t h ν b χ ν [ m J y ][ m J y ][ GH z ][ m J y ][ GH z ][ m J y ][ GH z ] II Z w . . . . - . . . . . . . II Z w . . . . . . . . . . . . . . . I C . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . NG C . . . . . . . . . . . . . . . T a b l e B . . O p ti m a l p a r a m e t e r s ob t a i n e dby t h e b e s t fi tt o eac h m od e l ( c on s t a n t , c u r v e d , b r ea k , a nd c u t o ff ) . . Klein et al.: Synchrotron spectra of galaxies , Online Material p 3
Apart from the reduced chi-squared, χ ν , an important testwhether the fits are physically meaningful is to look at the re-sulting thermal flux densities obtained at each frequency aftersubtracting the computed nonthermal flux density from the total(observed). A fit makes only sense if the resulting thermal fluxdensities so obtained obey to theexpected optically thin spectrum with slope − .
10. This isshown in each diagram by the cyan data points, with the errorstaken from those of the measured total flux densities. The result-ing slopes are listed in Table B.2.
Galaxy constant curved break cuto ff II Zw 40 − . − . − .
09 –II Zw 70 − . − . − . − . − . − .
10 – − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . − . . − . − . − . − .
38 0 . − . − . − . − . − . − . − . − . − . − . − . − . − . − . Table B.2. Slopes of the thermal flux densities. These linear fitsare depicted by solid magenta lines in the figures of the presentappendix.
Caption for the plots of this appendix:
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,,
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 10 . Klein et al.: Synchrotron spectra of galaxies ,,
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 10 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 11 . Klein et al.: Synchrotron spectra of galaxies ,,
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 10 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 11 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 12 . Klein et al.: Synchrotron spectra of galaxies ,,
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 10 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 11 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 12 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 13 . Klein et al.: Synchrotron spectra of galaxies ,,
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 10 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 11 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 12 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 13 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 14 . Klein et al.: Synchrotron spectra of galaxies ,,
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 10 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 11 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 12 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 13 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 14 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 15 . Klein et al.: Synchrotron spectra of galaxies ,,
In the main plots,the caption is as follows: the best fit of the radio continuumspectrum is represented by a solid red line. The best parame-ters are listed in the lower left part of the figure along with thereduced chi-square ( χ ν ). The faint grey lines represent the ran-domly 1000 spectra generated in the fitting process. The free-free (thermal) and synchrotron (nonthermal) components of themodels are depicted by the dotted cyan line and dashed greenline, respectively. The green crosses represent the observed non-thermal component (i.e., the modeled thermal component re-moved from the observational data). The cyan stars delineatethe observed thermal component (i.e., the modeled nonthermalcomponent subtracted from the observational data). The ther-mal component is drawn only if the fitted coe ffi cient, S th , ,is strictly positive. If present (curved, break, and cuto ff models),the vertical dash-dotted yellow line depicts the break frequency, ν b . The magenta line delineates a linear fit to the thermal fluxdensities. . Klein et al.: Synchrotron spectra of galaxies , Online Material p 4 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 5 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 6 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 7 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 8 . Klein et al.: Synchrotron spectra of galaxies , Online Material p 9 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 10 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 11 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 12 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 13 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 14 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 15 . Klein et al.: Synchrotron spectra of galaxies ,, Online Material p 16 . Klein et al.: Synchrotron spectra of galaxies ,,