Metallicities in long GRB host galaxies at z<0.5 calculated by the detailed modelling of optical and infrared line ratios
aa r X i v : . [ a s t r o - ph . GA ] D ec Mon. Not. R. Astron. Soc. , 1– ?? (2009) Printed 16 July 2018 (MN L A TEX style file v2.2)
Metallicities in long GRB host galaxies at z < M. Contini
School of Physics and Astronomy, Tel Aviv University, Tel Aviv 69978, Israel
16 July 2018
ABSTRACT
We revisited the line spectra emitted from long GRB (LGRB) host galaxies at z β in LGRB 980425 host sat-isfactorily reproduce the HeII/H β and [ArIII]/H β line ratios. The modelling of theobserved [SIV]10.51 µ m /[SIII]18.71 µ m and [NeIII]10.6 µ m /[NeII]12.81 µ m line ratiosfrom LGRB 031203 host galaxy at z=0.105 shows that the mid-IR lines are emit-ted from geometrically thin shock dominated filaments which are not reached by thephotoionizing flux, while the optical lines are emitted from the radiation dominatedoutflowing clouds. Key words: radiation mechanisms: general — shock waves — ISM: abundances —galaxies: GRB — galaxies: high redshift
Long duration γ -ray bursts (LGRB) derive from the deathof very massive star (e.g. Paczynski 1998), they are flashesof cosmic high energy ( ∼ −
10 GeV) photons (Fish-man & Meegan 1995) and explode in star forming galaxies.LGRB and their afterglows are associated with broad linedSN Ic (e.g. Hjorth et al 2003, Stanek et al 2003). The analy-sis of LGRB host galaxy emission lines provides informationabout star-forming galaxies at high z (e.g. Kr¨uhler et al 2015,Blanchard et al 2015 and references therein). The fluxesof significant oxygen lines, a few nitrogen lines as well asH α and H β are available from the surveys at redshifts z β and H α . The [OIII]4363 line,which plays a dominant role in the modelling process, isweak and not always available. Some spectra contain also[NeIII]3869, [SII]6717, [SII]6731 and seldom He, Fe and Arlines (e.g. Hammer et al 2006). The interpretation of thespectra leads to more or less converging theories about the distribution on z of star formation rates, ages, star massesand temperatures. The most interesting parameters are re-lated to metallicities, such as the O/H relative abundances,followed by the N/H one. Metallicity is one of the main pa-rameters which affects the evolution of massive stars as wellas their explosive deaths (Vergani et al 2011, Piranomonteet al 2015, Sollerman et al 2005, Woosley 1993, etc). Niinoet al (2016) claim that the relation between metallicity andLGRB occurrence rate is not understood quantitatively be-cause, even when the redshifts of the host galaxies are wellknown, the host galaxy is ”not studied in detail”. Such acrucial issue, which yields many consequences in a large zrange deserves a more precise analysis.In a previous paper (Contini 2016) we presented newresults calculated by the detailed modelling of LGRB lineratios and we compared them with those obtained by differ-ent investigations, in particular for the element abundancesin GRB host galaxies at relatively high redshifts. The emis-sion line spectra emitted from LGRB host galaxies on alarge z range were collected from the surveys of Kr¨uhleret al, Savaglio et al (2009), Sollerman et al (2005), Castro-Tirado et al (2001), Graham & Fruchter (2013), Levesque et c (cid:13) M. Contini al (2010), Vergani et al (2011), Piranomonte et al (2015), theLGRB line and continuum spectra with Wolf-Rayet (WR)features from the Han et al (2010) survey. In particular, theGRB 980425 host spectrum (Sollerman et al 2005) was se-lected because the survey refers to relatively low redshiftgalaxies (z < < L ⊙ andrelatively low metallicity. The spectra include the He I 5876line which is also significant in SN Type 1c hosts relativelyto WR stars.Contini (2016, fig. 6 top diagram) shows the results ofO/H and N/H calculated in each of the LGRB hosts bythe detailed modelling of the spectra. O/H is close to solarin most objects. We will conventionally define ”solar” rel-ative abundances (O/H) ⊙ =6.6 - 6.7 × − and (N/H) ⊙ =9. × − (Allen 1976, Grevesse & Sauval 1998) that werefound suitable to local galaxy nebulae. Moreover, these val-ues are included between those of Anders & Grevesse (1989)(8.5 × − and 1.12 × − ,respectively) and Asplund et al(2009) (4.9 × − and 6.76 × − , respectively ). Regardingnitrogen, we note a large distribution of N/H abundanceratios from solar to lower than solar by a factor >
10 inGRB hosts. Subsolar N/H indicates external gas acquisitionthrough galaxy merging processes. Some observed LGRBhosts, however, show high metallicity. Kr¨uler et al (and ref-erences therein) point out that ”several metal rich GRBwere discovered”. Graham et al (2015) recently confirmedthat high metallicities can be found in LGRB. The near so-lar O/H relative abundances calculated from the observedline ratios by Contini (2016) in LGRB hosts exceed theO/H values generally obtained by other modelling methods.In particular at high z, lower than solar O/H in the hostgas (e.g. Kr¨uhler et al 2015) were predicted by modellingthe [OIII]/H β and [OII]/H β line ratios using the strong-linemethods (see Sect. 2). These results are generally obtainedby the author majority. This issue (see Contini 2014 and ref-erences therein) will be explained in the following section.The number of LGRB galaxy host spectra at relativelylow z investigated by Contini (2016) was insufficiently small.In this paper we try to fill this gap, although by a relativelysmall number of objects, by modelling in detail the LGRBhost spectra presented by Niino et al (2016) at z µ m , [OI]63 µ m in the far-IR(FIR) and [SIV]10.51 µ m , [SIII]18.71 µ m , [NeIII]15.56 µ m ,[NeII]12.81 µ m , etc. lines in the mid-IR, respectively. In thispaper we will investigate whether the optical lines and themid-IR and FIR lines from the GRB 0301203 host galaxyare emitted from clouds of gas in similar physical conditions.We use for the calculation of the line ratios the codeSUMA which simulates an emitting gaseous cloud within agalaxy hosting an AGN or a starburst, heated and ionized bythe coupled effect of the primary and secondary photoioniza-tion fluxes and shocks. Shocks throughout the host galaxiesare the product of star explosions and of cloud collisions bye.g. galaxy merging (Contini 2016 and references therein).The ”detailed modelling” method is briefly explained andcompared with the strong-line methods in Sect. 2. The anal-ysis of Niino et al observed line spectra is presented in Sect.3 and those of Christensen et al and Hammer et al relativelyto the LGRB 980425 host galaxy are investigated in Sect.4. The optical and IR emission lines observed from LGRB031203 host are discussed in Sect. 5. Concluding remarksfollow in Sect. 6. To calculate the O/H abundance ratio from the observa-tions, ”direct methods” (see e.g. Modjaz et al 2008) are gen-erally adopted. They date back to the early ’70ties whenactive galaxy line spectra started to appear (see Seaton1975, Pagel et al 1992). Later, they were updated. Deter-mination of the O/H relative abundance from the oxygenline ratios to H β by current strong-line methods focus on[OIII]/H β and [OII]/H β . The line fluxes depend mostlyon the fractional abundance of the ions and on the rel-ative abundance of the elements. In particular, for inter-mediate ionization- level lines such as [OII] and [OIII], thefractional abundance of the relative ions are high in the ra-diation dominated zone of the cloud, while strong shocksare necessary to obtain strong lines from high ionization-levels (see, e.g. Sect. 5) and recombined element lines. Bydetailed modelling, the line fluxes are integrated through-out the recombination region where a large zone of gas ischaracterized by T e < K (see Sect. 5). O + /O and O /Ofractional abundances peak at different temperatures. In thecool gas region both are relatively low. Therefore, to repro-duce high [OII]/H β and [OIII]/H β line ratios, relatively highO/H are invoked. By the strong-line models, a single tem-perature of 10 K is generally adopted corresponding to highO + /O and O /O fractional abundances, so relatively lowO/H are suitable to fit the observed line ratios. It was ex-plained by Contini (2014) that the metallicities in terms c (cid:13) , 1– ?? GRB host galaxies at z < Table 1.
Modelling [OII]3727+, [OIII]5007+, H α and [NII]6583 line ratios to H β from Niino et al (2016) LGRB host galaxy spectraz [OII]/ [OIII]/ H α / [NII]/ V s n D O/H N/H T ∗ U H β H β H β H β H β km s − cm − cm 10 − − K - c c ,
0. 125 4.15 2.2 3.1 < c c - 4.2 1.8 3. 0.3 - - - - - - - -mN9 - 4.1 1.8 2.96 0.46 160 250 0.36 5.8 0.2 6.6 0.008 0.016061021 < , in erg cm − s − (calculated at the nebula); Christensen et al (2008); the subscript c refers to the corrected line ratios; Th¨one etal (2008); Prochaska et al (2004); Sollerman et al (2005); Niino et al (2016); Levesque et al (2010a); Schulze et al (2014); Levesque et al (2010b); Niino et al (2016); Kr¨uler et al (2015); Kr¨uler et al (2015); Garnavich et al (2003); Graham &Fruchter (2013); Kr¨uhler et al (2015); of O/H and N/H obtained by the strong-line methods arelower limits. Moreover, less consistent results are obtainedcomparing the data with diagnostic diagrams calculated forgeneral cases. For spectra rich in number of lines from dif-ferent elements in various frequency ranges, the CLOUDY,SUMA and other codes were assembled. SUMA accounts forboth the photoionization and the shock because 1) it wasnoticed that lines from recombined elements and from veryhigh ionization levels could not be reproduced consistentlyusing photoionization alone, 2) the radio continuum showsthe synchrotron power-law created by the Fermi mechanismat the shock front in most of the spectral energy distribu-tion (SED), 3) low [OIII]5007+/[OIII] 4363 line ratios (suchas observed in LINERs) could not be reproduced withoutthe shocks and above all, because 4) in galaxies at high z,that derive mostly from mergers, shocks are created by col-lision, etc. By the detailed modelling method the simulatedemitting clouds in the galaxy are characterized by the pa-rameters of the shock and of the photoionization, by the ele-ment abundances , by the geometrical thickness due to frag-mentation, etc. We calculate the emission lines from the gasadopting a set of the most significant input parameters. The calculation process is briefly described in the following. Someof the parameters are suggested by the observations (e.g. theFWHM gives a hint to the shock velocity choice, the electrondensity is roughly deduced from the [SII] 6716/6731 doubletratio, etc)). Moreover, some published grids of model results(Contini & Viegas 2001a,b) indicate in a general way howto choose the first set of input parameters. The lines calcu-lated by the code are more than 200, because line emissioncontributes to the cooling process of the emitting gas in therecombination zone. The number of calculated lines is inde-pendent from the observations. Indeed, when only a few linesare observed the initial parameter set is less constrained. Sowe create a grid of models in order to reproduce the obser-vations and to avoid degeneracy. In the modelling process,we aim to reproduce the observed line ratios for each el-ement. Each line has a different strength which translatesinto the different precision by the fitting process. A mini-mum number of significant lines ([OIII] 5007+, [OII]3727+,[OIII]4363, [NII], H α , H β ) is necessary to constrain themodel. We deal with line ratios to avoid distance and mor-phological effects. A perfect fit of the observed line ratiosis not realistic because the observed data have errors, both c (cid:13) , 1–, 1–
0. 125 4.15 2.2 3.1 < c c - 4.2 1.8 3. 0.3 - - - - - - - -mN9 - 4.1 1.8 2.96 0.46 160 250 0.36 5.8 0.2 6.6 0.008 0.016061021 < , in erg cm − s − (calculated at the nebula); Christensen et al (2008); the subscript c refers to the corrected line ratios; Th¨one etal (2008); Prochaska et al (2004); Sollerman et al (2005); Niino et al (2016); Levesque et al (2010a); Schulze et al (2014); Levesque et al (2010b); Niino et al (2016); Kr¨uler et al (2015); Kr¨uler et al (2015); Garnavich et al (2003); Graham &Fruchter (2013); Kr¨uhler et al (2015); of O/H and N/H obtained by the strong-line methods arelower limits. Moreover, less consistent results are obtainedcomparing the data with diagnostic diagrams calculated forgeneral cases. For spectra rich in number of lines from dif-ferent elements in various frequency ranges, the CLOUDY,SUMA and other codes were assembled. SUMA accounts forboth the photoionization and the shock because 1) it wasnoticed that lines from recombined elements and from veryhigh ionization levels could not be reproduced consistentlyusing photoionization alone, 2) the radio continuum showsthe synchrotron power-law created by the Fermi mechanismat the shock front in most of the spectral energy distribu-tion (SED), 3) low [OIII]5007+/[OIII] 4363 line ratios (suchas observed in LINERs) could not be reproduced withoutthe shocks and above all, because 4) in galaxies at high z,that derive mostly from mergers, shocks are created by col-lision, etc. By the detailed modelling method the simulatedemitting clouds in the galaxy are characterized by the pa-rameters of the shock and of the photoionization, by the ele-ment abundances , by the geometrical thickness due to frag-mentation, etc. We calculate the emission lines from the gasadopting a set of the most significant input parameters. The calculation process is briefly described in the following. Someof the parameters are suggested by the observations (e.g. theFWHM gives a hint to the shock velocity choice, the electrondensity is roughly deduced from the [SII] 6716/6731 doubletratio, etc)). Moreover, some published grids of model results(Contini & Viegas 2001a,b) indicate in a general way howto choose the first set of input parameters. The lines calcu-lated by the code are more than 200, because line emissioncontributes to the cooling process of the emitting gas in therecombination zone. The number of calculated lines is inde-pendent from the observations. Indeed, when only a few linesare observed the initial parameter set is less constrained. Sowe create a grid of models in order to reproduce the obser-vations and to avoid degeneracy. In the modelling process,we aim to reproduce the observed line ratios for each el-ement. Each line has a different strength which translatesinto the different precision by the fitting process. A mini-mum number of significant lines ([OIII] 5007+, [OII]3727+,[OIII]4363, [NII], H α , H β ) is necessary to constrain themodel. We deal with line ratios to avoid distance and mor-phological effects. A perfect fit of the observed line ratiosis not realistic because the observed data have errors, both c (cid:13) , 1–, 1– ?? M. Contini random and systematic. The set of parameters which leadsto the best fit of the observed line ratios and continuumSED, is regarded as the result of modelling. The results areacceptable when the observed strongest line ratios are re-produced within 10%, and the weakest by ∼
50 %.
By the SUMA code (Contini 2015 and references therein)line and continuum emissions from the gas are calculatedconsistently with dust-reprocessed radiation in a plane-parallel geometry. The calculations start at the shock frontwhere the gas is compressed and thermalized adiabatically,reaching the maximum temperature in the immediate post-shock region ( T ( K ) ∼ . × ( V s /
100 km s − ) , whereV s is the shock velocity). T decreases downstream follow-ing the cooling rate. The input parameters such as V s ,the atomic preshock density n and the preshock magneticfield B (for all models B =10 − Gauss is adopted) definethe hydrodynamical field. They are used in the calculationsof the Rankine-Hugoniot equations at the shock front anddownstream. They are combined in the compression equa-tion which is resolved throughout each slab of gas in orderto obtain the density profile downstream. The input param-eters that represent the primary radiation from the hoststarburst (SB) are the effective temperature T ∗ and the ion-ization parameter U . A pure black-body radiation referringto T ∗ is a poor approximation for a starburst, even adoptinga dominant spectral type (see Rigby & Rieke 2004). How-ever, it is the most suitable because the line ratios that areused to indicate T ∗ also depend on metallicity, electron tem-perature, density, ionization parameter, the morphology ofthe ionized clouds, and, in particular, they depend on thehydrodynamical field. The primary radiation source is inde-pendent but it affects the surrounding gas. In contrast, thesecondary diffuse radiation is emitted from the slabs of gasheated by the radiation flux reaching the gas and collision-ally by the shock. In our model the gas region surroundingthe radiation source is not considered as a unique cloud,but as an ensemble of fragmented filaments. The geometri-cal thickness of these filaments is an input parameter of thecode ( D ) which is calculated consistently with the physicalconditions and element abundances of the emitting gas. Pri-mary and secondary radiations are calculated by radiationtransfer throughout the slabs downstream. The fractionalabundances of the ions are calculated resolving the ioniza-tion equations for each element in each ionization level. Thedust-to-gas ratio ( d/g ) and the abundances of He, C, N, O,Ne, Mg, Si, S, A, Fe, relative to H, are also accounted for.The uncertainty in the calculations is due to the atomic pa-rameters (within 10 %) which are often updated. Niino et al presented spectroscopy results for three LGRBhost galaxies at z m + l og ( O / H ) d −4 −3 −2 −1 0−4−3−2−10 U m U d Figure 1.
Top : comparison of 12+log(O/H) calculated by de-tailed modelling (m) with those calculated using indirect diagnos-tic calibrations (R23, N2) by Niino et al (d). Bottom : comparisonof U calculated by detailed modelling (m) with those calculated byNiino et al (d). open circles: model calculations; asterisks: modelcalculations referring to reddening corrected data l og ( O / H ) log(z) a) a) −2.5 −2 −1.5 −1 −0.5 0 0.5−3−2−10 l og ( N / H ) log(z) b) b) Figure 2.
Calculated log(O/H) (top) and log(N/H) (bottom)(in units of 10 − , see Table 1) as a function of z on an extendedz range : symbols as in Fig. 1. Moreover, green stars : LGRB(Kr¨uhler et al); yellow hexagrams : SGRB (de Ugarte Postigo etal); magenta plus : LGRB with WR stars (Han et al).c (cid:13) , 1– ?? GRB host galaxies at z < spectroscopy. GRB 130427A host emission lines were alreadyreported by Xu et al (2013) and Kr¨uhler et al (2015), butthe [NII] detection was marginal in both. The host was re-visited by Subaru/FOCAS. GRB 111225A host has been an-alyzed by archival data (Th¨one & de Ugarte Postigo 2014)but they were not reported. The host was revisited with Sub-aru/FOCAS spectroscopy. Recently, Perley et al (2016) sug-gested that GRB 020819B host at z=0.410 reported by Niinoet al happens to lie close to the line of sight of GRB020819Bat z=1.9621, but it is entirely unrelated to the GRB. So itwas removed from the modelling list.In Table 1 we compare the calculated with the observedline ratios from Niino et al table 5. The observed line spectrawere corrected for foreground Galaxy extinction. The theo-retical H α /H β line ratios at the emitting nebula should be ∼
3, considering that the emitting gas has a distribution ofdensities and temperatures in the recombination zone down-stream of the shock front, and behind the ionization frontcreated by the radiation source. For some galaxies in theNiino et al sample the observed H α /H β are > > α /H β values can be found in highdensity gas ( > cm − ) where some self-absorption oc-curs in the Balmer lines (Osterbrock 1974). This leads tothe strengthening of the H α line relatively to the otherlines of the Balmer series (see also Contini 2003). For theforbidden lines, in particular [OII], [SII] etc. the critical den-sities for collisional deexcitation are > × cm − and forthe [OIII] 5007+ lines ∼ × cm − , therefore high den-sity gas in the host galaxies could be revealed by relativelystrong permitted lines and abnormally high H α /H β ratios.Alternatively, considering that the line fluxes are affected bygas and dust through their path to Earth, the data shouldby further reddening corrected. In Table 1 first column theidentification number of the galaxies is given followed bythe redshift and by the line ratios presented in the differentsurveys, reported by Niino et al. In the rows next to thosereferring to observations, the best fitting calculated line ra-tios and the parameters adopted in the selected models aregiven in columns 3-6 and columns 7-13, respectively. In thelast column of Table 1 the H β line fluxes calculated at thenebula are reported. For galaxies 980425 , 060505, 120422,130427A and 020819B, we present both the data given byNiino et al and those reddening corrected (e.g. 980425 c ).Models mN1, mN2, mN6, mN8 and mN13 split into e.g.mN1a and mN1b, where the former refers to Niino et aldata and the latter to the corrected line ratios.The results of modelling give shock velocities between120 and 200 km s − , pre-shock densities between 80 and280 cm − , in agreement with previous results obtained fora large sample of LGRB hosts (Contini 2016). Except for060614, for all the galaxies the geometrical thickness of theemitting clouds is τ
1. There is a roughagreement for U calculated from modelling the reddeningcorrected spectra and U =q/c determined on the basis ofstrong-line methods (Kobulnicky & Kewley 2004) applied,however, to undereddened line ratios by Niino et al (Fig. 1,bottom diagram). The reddening corrected [OII]/[OIII] lineratios are higher than those presented by Niino et al, im-plying models with higher star temperatures, higher O/Hand lower U , because these parameters affect differently the [OII]/H β and [OIII]/H β line ratios. The O/H metallicitiescalculated by the detailed modelling and by the strong-linemethod are compared in Fig. 1 (top diagram). The uncer-tainty in the Niino et al metallicity is ∼ > × − ) and the Asplund et al. one(4.9 × − ), the O/H relative abundances calculated to fitthe Niino et al corrected line ratios are all lower than solar.This would imply that O/H ratios calculated by detailedmodelling for the LGRB host sample at z < × − could be seen as min-ima throughout the z range. The O/H ratios recover solarvalues towards local galaxies. Considering that lower thansolar metallicities derive from mixing with external matterduring galaxy merging, it seems that the merging process ishighly efficient at low z. Niino et al report the average spectrum of the LGRB 980425host galaxy presented by Christensen et al in table 2, lastrow. The uncertainties are very large in the line flux means,therefore in the following we refer to the single region spec-tra.
In Tables 2 and 3 we present the modelling of the spectra ob-served throughout the 980425 host and the selected modelswhich best reproduce the observed line ratios. The data wereobtained by the Very Large Telescope UT3 Melipal with theVIMOS integral field spectroscopy mode. In Table 2 the lineratios reported by Christensen et al are shown in columns 2-7 followed by the reddening corrected line ratios in columns9-14. In the next rows the calculated line ratios best fittingthe data (models mc1-mc26) and best fitting the correctedline ratios (models m1c-m26c) are reported. The models aredescribed in Table 3. Most of the observed H α /H β are > τ <
1, therefore we have corrected for redden-ing the line ratios. In Fig. 3 (top diagram) the observedH α /H β are given as function of the observed [OII]/H β . Thethree observed spectra corresponding to models mc6, mc11and mc17 show H α /H β = 12.7, 10.4 and 11.9, respectively.The correction factor is very high and leads to high corrected[OII]/H β , in particular for the latter, which corresponds toan already high uncorrected [OII]/H β . High [OII]/H β aregenerally found in collisionally dominated nebulae, suggest-ing that shocks are at work. The H α /H β ratios for the twoformer spectra (referring to mc6 and mc11) correspond torelatively low [OII]/H β indicating that the reddening correc-tion makes sense. The [OIII]/H β versus [OII]/H β correctedand not- corrected line ratios (Fig. 3 bottom diagram) show c (cid:13) , 1– ?? M. Contini β H α / H β β ) l og ( [ O III]/ H β ) Figure 3.
Christensen et al data. Top : observed H α /H β ver-sus observed [OII]/H β . Bottom : observed [OIII]/H β ver-sus [OII]H β (red pentagrams); corrected [OIII]/H β versus[OII]/H β (blue squares) that even if the trend corresponds to a radiative situationwith small U (Contini 2016, fig. 1), the data are scattered,confirming that another mechanism, e.g. the shocks cannotbe neglected.Table 3 shows that the H β flux calculated at the emit-ting nebula is relatively high in the WR star region, and thesame should be for H α ( ∼ β ) confirming a relativelyhigh SFR. Metallicities in terms of O/H and N/H relativeabundances are close to solar. In the extreme east, in thehost regions at 10.1” × × × − and 5. × − , respectively. Fig. 4 showsthat O/H values calculated by detailed modelling exceedthose obtained by Christensen et al. using the strong-linemethod by a factor >
2. Table 3 shows that sulphur is almostunderabundant everywhere ((S/H) ⊙ = 2. × − ). S is easilytrapped into dust grains and subtracted from the gaseousphase by factors < α /H β .We have chosen the symbols showing darker areas with in-creasing parameter in each diagram. They are described inTable 4. Fig. 5 shows the following features. T ∗ increasestowards the south-west zone and reaches T ∗ =8.4 10 K in m + l og ( O / H ) d Figure 4.
Comparison of log(O/H)+12 calculated by detailedmodelling (m) with those calculated using indirect diagnostic cal-ibrations by Christensen et al (d). open circles: model calcula-tions; asterisks: model calculations referring to reddening cor-rected data. the SN region, slightly lower than T ∗ > K which wascalculated by Contini (2016) modelling the Han et al (2010)survey. Han et al suggest that WR and O stars are present insome LGRB (e.g. 980703, 990712) and in other host galax-ies (Contini 2016, table 10). The ionization parameter hasan opposite trend. A diluted U in the SN region indicatesthat the radiation source is far from the emitting gas or thatthe photoionizing flux is prevented from reaching the gas byobstructing matter. The H α /H β observed line ratio is < >
300 cm − are revealed. The den-sities inside the gaseous clouds increase by a factor > × -4.0” in GRB 980425 is reproducedadopting an accretion model, in agreement with Michalowskiet al (2016) who claim that accretion is more adapted thanoutflow in starburst galaxies. Michalowski et al recently pub-lished [CII] 158 µ m and [OI] 63 µ m line fluxes in the FIR.They claim that [CII] emission exhibits a normal radial pro-file while [OI] is concentrated close to the WR zone. We havecalculated the [CII] and [OI] line fluxes consistently withthe optical lines using the same models (Table 3) in eachregion. Our results (Table 5, Fig. 5) show that [OI] are weakthroughout the host with higher values in the regions withina band roughly oriented from north-east to south-west. For[CII] the radial structure can be roughly recognized. In theWR and SN regions, the calculated values are within themean, suggesting that clouds different from those emittingthe optical lines contribute to the FIR emission lines. Ac-cording to the results presented for the LGRB 031203 hostgalaxy in Sect. 5., they could be represented by shock domi-nated ( U =0) filaments, but the data are not enough to con-strain the models. c (cid:13) , 1– ?? GRB host galaxies at z < Table 2.
Modelling [OII]3727+, [OIII]5007+, H α , [NII]6583, [SII]6717, 6731 line ratios to H β =1 from GRB 980425 host regions(Christensen et al 2008)[OII]/ [OIII]/ H α / [NII]/ [SII]/ [SII]/ [OII]/ [OIII]/ H α / [NII]/ [SII]/ [SII]/H β H β H β H β H β H β H β H β H β H β H β H β galaxy 5.39 3.87 4.02 0.44 0.63 0.15 7.1 3.74 3. 0.33 0.45 0.11mc1 5.41 3.97 2.97 0.44 0.58 0.66 m1c 7.2 3.8 3. 0.39 0.45 0.5WR region 4. 6.24 3.46 0.27 0.3 0.23 4.57 6.14 3. 0.23 0.25 0.2mc2 4. 6.24 2.94 0.3 0.27 0.3 m2c 4.5 6.18 3. 0.3 0.26 0.3SN region 3.45 3.22 5.95 0.67 1.22 0.91 6.6 3. 3. 0.34 0.56 0.4mc3 3.4 3.2 2.96 0.68 0.78 0.9 m3c 6.7 3.2 3. 0.36 0.54 0.6WR(-8.7) 3.1 5. 2.95 0.21 0.21 0.17 3. 5. 3. 0.21 0.21 0.17mc4 3. 5.2 2.95 0.2 0.27 3. m4c 3. 5.2 3. 0.22 0.27 0.32SN(-5.3) 2.86 3.36 6.86 0.82 1.53 1.12 6.24 3. 3. 0.36 0.59 0.42mc5 2.79 3.33 2.96 0.86 0.83 0.9 m5c 6.4 3.1 3. 0.24 0.53 0.56(-11.4) 1.08 1.03 12.7 0.146 2. 1.52 4.2 0.87 3. 0.034 0.38 0.28mc6 1.1 1.01 2.99 0.17 1.3 1.5 m6c 4.3 0.86 3. 0.1 0.35 0.38(-8.0) 1.8 1.09 6.28 1.08 1.65 1.29 3.6 1. 3. 0.52 0.71 0.54mc7 1.8 1.12 3. 0.9 1.64 1.7 m7c 3.86 1. 3. 0.6 0.7 0.7(-6.7) 1.5 1.26 6.98 0.86 1.22 1.04 3.3 1.1 3. 0.37 0.46 0.39mc8 1.7 1.25 3. 0.8 1.23 1.29 m8c 3.4 1.19 3. 0.4 0.48 0.49(-4.0) 1.84 1.5 4.7 0.59 0.73 0.53 2.81 1.42 3. 0.37 0.44 0.3mc9 1.9 1.5 2.98 0.6 0.6 0.62 m9c 2.97 1.4 2.98 0.36 0.45 0.46(-2.0) 3.2 1.72 5.53 0.78 0.94 0.77 5.7 1.6 3. 0.42 0.47 0.37mc10 3. 1.77 2.98 0.7 0.8 0.73 m10c 5.88 1.6 3. 0.4 0.47 0.42(-1.3,-6.1) 2.23 1.86 10.4 1.2 1.84 1.29 7.2 1.6 3. 0.35 0.44 0.3mc11 2.1 1.92 3. 0.9 1.42 1.22 m11c 7.4 1.6 3.1 0.46 0.42 0.35(-1.3,10.0) 1.47 0.84 5.42 0.76 1.06 0.86 2.56 0.78 3. 0.42 0.53 0.43mc12 1.5 0.84 3. 0.7 1.07 0.93 m12c 2.57 0.82 3. 0.43 0.58 0.5(-0.7) 2.89 2.16 4.11 0.53 0.43 0.3 3.89 2.08 3. 0.38 0.3 0.2mc13 2.96 2.02 3. 0.66 0.46 0.39 m13c 3.7 2.1 3. 0.4 0.3 0.26(-0.0) 1.66 3.13 5.7 0.92 1.23 0.92 3.0 2.9 3. 0.48 0.59 0.43mc14 1.67 3.1 3. 0.9 1.04 1.04 m14c 3.1 2.97 3. 0.6 0.59 0.58(2) 0.0 0.99 6.25 0.87 1.5 1.28 0.0 0.91 3. 0.42 0.65 0.54mc15 0.7 1.0 3. 0.7 0.65 1.18 m15c 0.8 0.96 2.97 0.48 0.46 0.85(2.7) 1.03 1.12 4.1 0.61 0.66 0.44 1.38 1.08 3. 0.44 0.46 0.30mc16 1.02 1.13 3.05 0.63 0.47 0.51 m16c 1.34 1.18 3. 0.4 0.36 0.4(2.7,-4.0) 6.9 1.84 11.9 0.44 0.62 0.4 25. 1.57 3. 0.44 0.62 0.41mc17 7.0 1.84 3.17 1.8 2.6 2.2 m17c 25.6 1.4 3.25 0.42 0.55 0.56(5.4) 0.93 1.2 4.3 0.63 0.77 0.56 1.3 1.15 3. 0.44 0.51 0.37mc18 0.91 1.25 3.05 0.5 0.74 0.79 m18c 1.45 1.1 3.1 0.46 0.57 0.57(6.1) 1.22 1.38 3.5 0.53 0.76 0.54 1.4 1.35 3. 0.45 0.64 0.45mc19 1.1 1.4 3.05 0.6 0.77 0.8 m19c 1.3 1.35 3. 0.48 0.52 0.47(8.7) 1.22 0.68 2.4 0.31 0.68 0.42 0.99 0.7 3. 0.39 0.88 0.54mc20 1.21 0.63 3. 0.4 0.7 0.6 m20c 1.04 0.7 3. 0.3 0.7 0.58(9.4) 1.63 2.87 2.47 0.27 0.4 0.3 1.36 2.93 3. 0.33 0.5 0.38mc21 1.75 2.9 2.99 0.4 0.44 0.44 m21c 1.4 2.85 3. 0.32 0.4 0.42(10.1) 0.09 2.45 4. 0.49 0.68 0.48 0.12 2.37 3. 0.37 0.49 0.34mc22 0.12 2.47 3. 0.1 0.02 0.03 m22c 0.23 2.2 3. 0.4 0.4 0.5(11.4) 1.78 2.98 2.22 0.37 0.33 0.25 1.3 3.08 3. 0.5 0.46 0.35mc23 1.8 2.97 2.99 0.38 0.37 0.37 m23c 1.37 2.95 3. 0.44 0.37 0.37(12.1) 0.00 1.3 4.78 0.67 0.9 0.77 0.0 1.23 3. 0.42 0.53 0.45mc24 0.7 1.2 2.98 0.7 0.66 1.22 m24c 0.8 1.2 2.98 0.48 0.4 0.7(14.8,3.3) 3.98 2.34 2.7 0.29 0.56 0.39 3.6 2.36 3. 0.47 0.66 0.61mc25 3.99 2.24 2.95 0.3 0.38 0.5 m25c 3.67 2.3 2.95 0.4 0.45 0.6(14.8,10.) 2.15 1.86 2.09 0.33 0.44 0.4 1.53 1.94 3. 0.47 0.66 0.61mc26 2.3 1.94 3. 0.5 0.45 0.39 m26c 1.5 2.1 3. 0.48 0.62 0.53 To check whether the models calculated to reproduce theline ratios observed by Christensen et al from the SN andWR regions are able to explain the data reported by otherobservers, we apply the detailed modelling method to thespectroscopic VLT/FORS2 observations presented by Ham- mer et al (2006, table 1) which contain a relatively largenumber of lines covering an extended range of frequen-cies and of elements. In Table 6 we compare model re-sults (mSN, mWR and mrg4) with the data. The line ra-tios were corrected on the basis of previous considerations.However, the spectrum from region ”4” is characterised by c (cid:13) , 1–, 1–
Modelling [OII]3727+, [OIII]5007+, H α , [NII]6583, [SII]6717, 6731 line ratios to H β =1 from GRB 980425 host regions(Christensen et al 2008)[OII]/ [OIII]/ H α / [NII]/ [SII]/ [SII]/ [OII]/ [OIII]/ H α / [NII]/ [SII]/ [SII]/H β H β H β H β H β H β H β H β H β H β H β H β galaxy 5.39 3.87 4.02 0.44 0.63 0.15 7.1 3.74 3. 0.33 0.45 0.11mc1 5.41 3.97 2.97 0.44 0.58 0.66 m1c 7.2 3.8 3. 0.39 0.45 0.5WR region 4. 6.24 3.46 0.27 0.3 0.23 4.57 6.14 3. 0.23 0.25 0.2mc2 4. 6.24 2.94 0.3 0.27 0.3 m2c 4.5 6.18 3. 0.3 0.26 0.3SN region 3.45 3.22 5.95 0.67 1.22 0.91 6.6 3. 3. 0.34 0.56 0.4mc3 3.4 3.2 2.96 0.68 0.78 0.9 m3c 6.7 3.2 3. 0.36 0.54 0.6WR(-8.7) 3.1 5. 2.95 0.21 0.21 0.17 3. 5. 3. 0.21 0.21 0.17mc4 3. 5.2 2.95 0.2 0.27 3. m4c 3. 5.2 3. 0.22 0.27 0.32SN(-5.3) 2.86 3.36 6.86 0.82 1.53 1.12 6.24 3. 3. 0.36 0.59 0.42mc5 2.79 3.33 2.96 0.86 0.83 0.9 m5c 6.4 3.1 3. 0.24 0.53 0.56(-11.4) 1.08 1.03 12.7 0.146 2. 1.52 4.2 0.87 3. 0.034 0.38 0.28mc6 1.1 1.01 2.99 0.17 1.3 1.5 m6c 4.3 0.86 3. 0.1 0.35 0.38(-8.0) 1.8 1.09 6.28 1.08 1.65 1.29 3.6 1. 3. 0.52 0.71 0.54mc7 1.8 1.12 3. 0.9 1.64 1.7 m7c 3.86 1. 3. 0.6 0.7 0.7(-6.7) 1.5 1.26 6.98 0.86 1.22 1.04 3.3 1.1 3. 0.37 0.46 0.39mc8 1.7 1.25 3. 0.8 1.23 1.29 m8c 3.4 1.19 3. 0.4 0.48 0.49(-4.0) 1.84 1.5 4.7 0.59 0.73 0.53 2.81 1.42 3. 0.37 0.44 0.3mc9 1.9 1.5 2.98 0.6 0.6 0.62 m9c 2.97 1.4 2.98 0.36 0.45 0.46(-2.0) 3.2 1.72 5.53 0.78 0.94 0.77 5.7 1.6 3. 0.42 0.47 0.37mc10 3. 1.77 2.98 0.7 0.8 0.73 m10c 5.88 1.6 3. 0.4 0.47 0.42(-1.3,-6.1) 2.23 1.86 10.4 1.2 1.84 1.29 7.2 1.6 3. 0.35 0.44 0.3mc11 2.1 1.92 3. 0.9 1.42 1.22 m11c 7.4 1.6 3.1 0.46 0.42 0.35(-1.3,10.0) 1.47 0.84 5.42 0.76 1.06 0.86 2.56 0.78 3. 0.42 0.53 0.43mc12 1.5 0.84 3. 0.7 1.07 0.93 m12c 2.57 0.82 3. 0.43 0.58 0.5(-0.7) 2.89 2.16 4.11 0.53 0.43 0.3 3.89 2.08 3. 0.38 0.3 0.2mc13 2.96 2.02 3. 0.66 0.46 0.39 m13c 3.7 2.1 3. 0.4 0.3 0.26(-0.0) 1.66 3.13 5.7 0.92 1.23 0.92 3.0 2.9 3. 0.48 0.59 0.43mc14 1.67 3.1 3. 0.9 1.04 1.04 m14c 3.1 2.97 3. 0.6 0.59 0.58(2) 0.0 0.99 6.25 0.87 1.5 1.28 0.0 0.91 3. 0.42 0.65 0.54mc15 0.7 1.0 3. 0.7 0.65 1.18 m15c 0.8 0.96 2.97 0.48 0.46 0.85(2.7) 1.03 1.12 4.1 0.61 0.66 0.44 1.38 1.08 3. 0.44 0.46 0.30mc16 1.02 1.13 3.05 0.63 0.47 0.51 m16c 1.34 1.18 3. 0.4 0.36 0.4(2.7,-4.0) 6.9 1.84 11.9 0.44 0.62 0.4 25. 1.57 3. 0.44 0.62 0.41mc17 7.0 1.84 3.17 1.8 2.6 2.2 m17c 25.6 1.4 3.25 0.42 0.55 0.56(5.4) 0.93 1.2 4.3 0.63 0.77 0.56 1.3 1.15 3. 0.44 0.51 0.37mc18 0.91 1.25 3.05 0.5 0.74 0.79 m18c 1.45 1.1 3.1 0.46 0.57 0.57(6.1) 1.22 1.38 3.5 0.53 0.76 0.54 1.4 1.35 3. 0.45 0.64 0.45mc19 1.1 1.4 3.05 0.6 0.77 0.8 m19c 1.3 1.35 3. 0.48 0.52 0.47(8.7) 1.22 0.68 2.4 0.31 0.68 0.42 0.99 0.7 3. 0.39 0.88 0.54mc20 1.21 0.63 3. 0.4 0.7 0.6 m20c 1.04 0.7 3. 0.3 0.7 0.58(9.4) 1.63 2.87 2.47 0.27 0.4 0.3 1.36 2.93 3. 0.33 0.5 0.38mc21 1.75 2.9 2.99 0.4 0.44 0.44 m21c 1.4 2.85 3. 0.32 0.4 0.42(10.1) 0.09 2.45 4. 0.49 0.68 0.48 0.12 2.37 3. 0.37 0.49 0.34mc22 0.12 2.47 3. 0.1 0.02 0.03 m22c 0.23 2.2 3. 0.4 0.4 0.5(11.4) 1.78 2.98 2.22 0.37 0.33 0.25 1.3 3.08 3. 0.5 0.46 0.35mc23 1.8 2.97 2.99 0.38 0.37 0.37 m23c 1.37 2.95 3. 0.44 0.37 0.37(12.1) 0.00 1.3 4.78 0.67 0.9 0.77 0.0 1.23 3. 0.42 0.53 0.45mc24 0.7 1.2 2.98 0.7 0.66 1.22 m24c 0.8 1.2 2.98 0.48 0.4 0.7(14.8,3.3) 3.98 2.34 2.7 0.29 0.56 0.39 3.6 2.36 3. 0.47 0.66 0.61mc25 3.99 2.24 2.95 0.3 0.38 0.5 m25c 3.67 2.3 2.95 0.4 0.45 0.6(14.8,10.) 2.15 1.86 2.09 0.33 0.44 0.4 1.53 1.94 3. 0.47 0.66 0.61mc26 2.3 1.94 3. 0.5 0.45 0.39 m26c 1.5 2.1 3. 0.48 0.62 0.53 To check whether the models calculated to reproduce theline ratios observed by Christensen et al from the SN andWR regions are able to explain the data reported by otherobservers, we apply the detailed modelling method to thespectroscopic VLT/FORS2 observations presented by Ham- mer et al (2006, table 1) which contain a relatively largenumber of lines covering an extended range of frequen-cies and of elements. In Table 6 we compare model re-sults (mSN, mWR and mrg4) with the data. The line ra-tios were corrected on the basis of previous considerations.However, the spectrum from region ”4” is characterised by c (cid:13) , 1–, 1– ?? M. Contini
Table 3.
Models calculated to reproduce the spectra by Christensen et al (2008) reported in Table 2V s n D O/H N/H S/H T ∗ U H β km s − cm − cm 10 − − − K - mc1 150 100 0.4 4.9 0.17 0.04 9. 0.0076 0.006m1c 150 100 0.4 6. 0.14 0.03 9.4 0.0056 0.0047mc2 150 100 0.4 5. 0.17 0.02 8.4 0.018 0.01m2c 150 100 0.4 5.5 0.17 0.02 8.4 0.016 0.01mc3 150 100 0.3 5.5 0.5 0.08 6.5 0.014 0.01m3c 150 100 0.3 6.3 0.15 0.04 8.3 0.0055 0.0046mc4 150 100 0.3 6.6 0.2 0.03 6.5 0.032 0.015m4c 150 100 0.3 6.6 0.2 0.03 6.5 0.032 0.015mc5 120 100 0.3 6.6 0.8 0.1 6. 0.018 0.01m5c 120 100 0.3 6.6 1. 0.04 8.4 0.0044 0.0037mc6 120 100 0.3 6. 0.3 0.3 3.4 0.15 0.027m6c 120 100 0.3 6. 0.06 0.04 5. 0.003 0.003mc7 130 80 0.3 6.6 1. 0.3 4.4 0.02 0.0084m7c 130 80 0.3 6.4 0.4 0.08 5. 0.005 0.0034mc8 130 80 0.3 6.4 1. 0.22 4.4 0.024 0.0095m8c 130 80 0.3 6.4 0.3 0.06 4.8 0.008 0.0048mc9 130 80 0.3 6.2 0.8 0.1 4.4 0.03 0.01m9c 130 80 0.3 6.4 0.3 0.06 4.7 0.012 0.0063mc10 120 70 0.2 6.6 0.6 0.1 5.2 0.01 0.0037m10c 120 60 0.2 6.6 0.2 0.04 5.9 0.003 0.0015mc11 120 50 1. 6.7 1. 0.2 5.5 0.015 0.0042m11c 120 50 1. 6.6 0.2 0.03 7.8 0.0014 8.3e-4mc12 120 50 2. 6.3 1. 0.2 4.4 0.015 0.0046m12c 120 50 2. 6.2 0.4 0.08 4.8 0.0055 0.0026mc13 120 50 2. 6.3 0.6 0.05 6.1 0.008 0.0032m13c 120 50 2. 6.4 0.3 0.03 6.4 0.006 0.0027mc14 120 80 2. 6.5 1.2 0.14 6.1 0.03 0.013m14c 120 80 2. 6.5 0.5 0.06 7. 0.01 0.0067mc15 130 320 0.1 6.0 1.4 0.3 3.0 0.8 0.22m15c 130 320 0.1 6.0 0.8 0.2 2.9 0.9 0.23mc16 120 100 5.5 6.0 1. 0.1 4.5 0.03 0.02m16c 120 100 5.5 6.0 0.7 0.07 4.6 0.025 0.019mc17 120 50 0.17 6.6 1. 0.26 5.8 8.e-4 4.e-4m17c
54 200 0.0035 6.3 0.15 0.14 3.2 0.003 1.6mc18 110 100 5.5 6. 1. 0.16 4.5 0.044 0.024m18c 100 120 6.5 6.4 0.6 0.1 5.5 0.01 0.013mc19 120 100 5.5 6. 1. 0.15 4.6 0.04 0.022m19c 120 100 5.5 6. 0.7 0.1 5.0 0.02 0.017mc20 120 50 5. 6.3 0.6 0.15 4.2 0.016 0.006m20c 120 50 5. 6.3 0.6 0.15 4. 0.03 0.008mc21 120 80 2. 6.5 0.6 0.06 5.6 0.033 0.0124m21c 120 80 2. 6. 0.5 0.06 5.2 0.059 0.016mc22 100 200 0.1 4.5 2. 0.3 4. 0.95 0.058m22c 100 160 2. 4. 2. 0.09 3.9 9.4 0.11mc23 120 80 2. 6.5 0.5 0.05 5.6 0.033 0.012m23c 120 80 4. 6. 0.6 0.05 5.8 0.042 0.015mc24 130 320 0.1 6. 1.4 0.3 3.1 0.83 0.22m24c 130 320 0.1 6. 0.8 0.16 3. 0.83 0.22mc25 150 140 0.18 5. 0.14 0.04 6.3 0.0094 0.01m25c 150 140 0.18 5. 0.2 0.05 6.1 0.011 0.01mc26 120 50 1. 6.6 0.6 0.06 5.3 0.015 0.0042m26c 120 50 1. 6.6 0.6 0.09 4.9 0.015 0.0043 in erg cm − s − (calculated at the nebula); calculated by an infalling model. a relatively low H β flux, leading to H α /H β =8.217. Theobserved [OII]/H β =11.65 skips to 30 by reddening cor-rection, which looks very high. We reproduce it by ∼ U (0.0004). The [OIII] 4363/H β is over-predicted by >
50% by model mrg4. This suggests that theabnormally high H α /H β is not only due to dust absorp- tion, but some self absorption across high density gas hasreduced the H β flux reported by Hammer et al in their ta-ble 1. [ArIII]7136/H β observed from region ”4” is well re-produced by model mrg4, while [ArIII]7136/H β data fromthe WR and SN regions are underpredicted by the modelsby a factor of ∼
3, adopting Ar/H=3.3 × − . c (cid:13) , 1– ?? GRB host galaxies at z < −15−10−5051015−15−10−5051015 arcsec a r cs e c Pre−shock density −15−10−5051015−15−10−5051015 arcsec a r cs e c Star temperature −15−10−5051015−15−10−5051015 arcsec a r cs e c Ionization parameter −15−10−5051015−15−10−5051015 arcsec a r cs e c (H α /H β ) obs −15−10−5051015−15−10−5051015 arcsec a r cs e c [CII] −15−10−5051015−15−10−5051015 arcsec a r cs e c [OI] Figure 5.
Christensen et al data. Distribution of the physical parameters throughout the galaxy GRB 980435 host. n (top left); T ∗ (topright) ; U (middle left) ; (H α /H β ) obs (middle right); [CII]158 µ m flux calculated at the nebula(bottom left); [OI]63 µ m flux calculatedat the nebula (bottom right); large blue circle : the SN place and large red square : the region where WR stars were detected. Symbolsrelative to the different parameters are explained in Table 4. Table 4.
Symbols in Fig. 5 diagramsleft-top right-top left-middle right-middle left-bottom right-bottomn T ∗ U (H α /H β ) [CII] line flux [OI] line flux cm − K - - 0.01 erg cm − s − − s − open circle 50 < < < > > > > observed at Earth; calculated at the nebulac (cid:13) , 1–, 1–
Symbols in Fig. 5 diagramsleft-top right-top left-middle right-middle left-bottom right-bottomn T ∗ U (H α /H β ) [CII] line flux [OI] line flux cm − K - - 0.01 erg cm − s − − s − open circle 50 < < < > > > > observed at Earth; calculated at the nebulac (cid:13) , 1–, 1– ?? M. Contini
Table 5.
Flux (in 0.01 erg cm − s − ) of [CII] and [OI] lines calculated at the nebula by the m4c-m26c models (Table 3)model [CII]158 µ m [OI]63 µ m model [CII]158 µ m [OI]63 µ mm4c 0.11 0.207 m16c 1.80 0.098m5c 0.29 0.430 m17c 0.002 0.000m6c 1.03 0.972 m18c 3.29 0.384m7c 0.21 0.015 m19c 1.34 0.099m8c 0.161 0.161 m20c 0.438 0.034m9c 0.120 0.009 m21c 0.360 0.033m10c 0.074 0.009 m22c 0.09 0.018m11c 0.588 0.588 m23c 0.72 0.07m12c 0.403 0.030 m24c 0.176 0.0220m13c 0.269 0.235 m25c 0.028 0.0074m14c 0.871 0.062 m26c 0.105 0.0075m15c 0.176 0.022 - - - Table 6.
Modelling the line spectra from Hammer et al (2006, table 2)SN SN (corr) mSN WR WR (corr) mWR reg 4 reg 4 (corr) mrg4[OII] 3727 4.44 6.23 6.7 1.24 2.08 3. 11.65 30.13 17.4[NeIII] 3869 0.61 0.81 0.68 0.32 0.49 0.49 0.239 0.53 0.66[OIII] 4363 < < β α s (km s − ) - - 120 - - 150 - - 140n (cm − ) - - 100 - - 100 - - 40 D (10 cm ) - - 0.3 - - 0.3 - - 0.02T ∗ (10 K) - - 8. - - 6.5 - - 4.6 U - - 0.004 - - 0.032 - - 4e-4He/H - - 0.1 - - 0.1 - - 0.1N/H (10 − ) - - 0.3 - - 0.2 - - 0.5O/H (10 − ) - - 6.6 - - 6.6 - - 7.0Ne/H (10 − ) - - 1. - 1. - - 0.7S/H (10 − ) - - 0.04 - 0.06 - - 0.15Ar/H(10 − ) - - 0.033 - 0.033 - - 0.033H β (erg cm − s − ) at the nebula - - 0.0037 - - 0.015 - - 1.e-4 The lines in the optical range from GRB 031203 host galaxywere observed and modelled in different ways in the lastyears (Niino et al and references therein, Contini 2016, etc).We have reproduced the line ratios by the radiation dom-inated (RD) mN3 model (Table 1), characterized by rela-tively low O/H and N/H and a relatively high T ∗ . Themodel accounts consistently also for the shock.Recently, Watson et al (2016) presented the first mid- IR spectrum of a GRB host (HG031203) by low andhigh resolution spectroscopy with Spitzer-IRS . Watson etal found that the IRS spectra show strong high ionizationfine structure emission lines such as [SIV] 10.51 µ m whichsuggest a hard radiation field in the galaxy and there-fore strong ongoing star formation and a very young stel-lar population. They claim that the absence of PAH sup-ports this idea as well as the hot dust peak temperature.We remind that high ionization lines are strong in pres-ence of the shocks which heat the gas to temperatures ∝ c (cid:13) , 1– ?? GRB host galaxies at z < (V s ) . Moreover, PAH grains are very small ( < µ m )and easily sputtered throughout the shock front. We referto the observed [SIV]10.51 µ m /[SIII]18.71 µ m (=1.63) and[NeIII]15.56 µ m /[NeII]12.81 µ m (=15.14) line ratios. By thecode SUMA line fluxes from far-UV to far-IR are calculatedconsistently for each model. The IR line ratios calculatedby model mN3 do not reproduce the observed line ratios.On the other hand, the shock dominated (SD) model cal-culated adopting the same shock input parameters as mN3but with U =0, approximates the data by an error of ∼ D =1.17 × cm.To better understand the results, the profiles of the elec-tron temperature and density within the emitting clouds andof the fractional abundances of the H, O, Ne and S ions indifferent ionization stages are shown in Fig. 6 for both theRD (top diagrams) and the SD (bottom diagram) models.The RD model corresponds to the case in which the gasmoves outwards from the starburst, therefore the photoion-izing radiation reaches the edge (on the right of the right topdiagram) of the cloud opposite to the shock front (on the leftof the left top diagram). The emitting cloud correspondingto the RD model is divided into two halves represented bythe left and right top diagrams. The left diagram shows theregion close to the shock front and the distance from theshock front on the X-axis scale is logarithmic. The right di-agram shows the conditions downstream far from the shockfront, close to the edge reached by the photoionization fluxwhich is opposite to the shock front. The distance from theilluminated edge is given by a reverse logarithmic X-axisscale. The two edges of the cloud are bridged by the sec-ondary radiation from both sides. In the SD case (bottomdiagram) the gas is collisionally heated by the shock and bysecondary radiation emitted by the slabs of gas heated bythe shock. The gas recombines in the downstream region.Fig. 6 shows that the optical lines calculated by the mN3model are emitted from RD clouds moving outwards, whilethe corresponding mid-IR ones come from cloud fragmentswhich are not reached by the primary radiation flux. We revisited the line spectra reported by Niino et al forgalaxies at z z < α /H β ratios (which reach values as high as5.5 in some of the observed spectra) cannot originate fromhigh density gas which is not predicted by the observed [SII]doublet and by strong forbidden lines in general. Therefore,the line ratios were reddening corrected. We have found inLGRB host galaxies at z U reveals that in the SNregion the emitting gas is far from the radiation source inagreement with Fynbo et al (2000) who claim by high spa-tial resolution imaging, that the GRB and the associatedSN did not occur in the regions where the WR stars and Ostars are located, corresponding to rich and compact clus-ters of SFR, but several hundreds parsec away. Moreover,our results show that O/H is slightly lower than solar, N/His lower than solar and S is depleted from the gaseous phasein nearly all the regions throughout the host. Comparisonof the pre-shock density map with the H α /H β one confirmsthat high H α /H β are due to dust reddening rather thanto self absorption by high density gas. Following Chevalier(1982) theory, two shocks are formed after the SN explo-sion, one propagating towards the SN and one proceedingoutward throughout the ISM. We suggest that the line ratiosobserved by Christensen et al in the SN region are emitteddownstream of the outward shock front, where the pre-shockdensity and the shock velocity are suitable to the ISM.Region 2.7” × -4.0” line spectrum (Table 2) in the GRB980425 host, which shows an abnormally high [OII]/[OIII]line ratio, could be reproduced only by an accretion model,i.e. the emitting gaseous cloud is infalling towards the ra-diation source. Accretion rather than outflow is supportedby Michalowski et al in starburst galaxies. Modelling re-sults of [CII] 158 and [OI] 63 FIR line fluxes observed byMichalowsky et al in the SN and WR regions are similar inaverage to those obtained in the other regions of the hostgalaxy.The models calculated for the SN and WR regions onthe basis of the Christensen et al data, were constrained onlyby a few oxygen, nitrogen and sulphur line ratios to H β . Wehave checked them by modelling the line ratios observed byHammer et al. We have found that the same models repro-duce satisfactorily also the HeII/H β and [ArIII]/H β line ra-tios. High temperature stars in the SN region are confirmed.The modelling of [SIV]10.51/[SIII]18.71 and[NeIII]10.6/[NeII]12.81 line ratios observed by Watsonet al from GRB 031203 host galaxy at z=0.105 indicatesthat the mid-IR lines are emitted from geometrically thinshock dominated clouds which are not reached by thestarburst photoionizing flux, while the optical lines areemitted from the radiation dominated outflowing clouds. ACKNOWLEDGEMENTS
I am very grateful to the referee for many critical suggestionswhich substantially improved the presentation of the paper. c (cid:13) , 1– ?? M. Contini l og ( T e , n e ) T e N e a) a) −3−2−10 l og ( X i / X ) III IIII I b) b) l og ( X i / X ) c) c) a) a) −3−2−10 III IIII I b) b) c) c) l og ( T e , n e ) T e N e a) a) −3−2−10 l og ( X i / X ) III II III b) b)
15 15.02 15.04 15.06 15.08 15.1−3−2−10 log distance from shock front [cm] l og ( X i / X ) c) c) Figure 6.
Top diagrams : the RD model mN3 for GRB 031203 (see text). Top panels : the electron temperature and the electron densitythroughout the emitting cloud. Middle panels : red lines : O /O (dot-dashed, III) and O + /O (dashed, II): black solid lines : H + /H (II)and H /H (I). Bottom panels : blue lines refer to Ne: Ne /Ne (dotted), Ne /Ne (solid), Ne /Ne (dash-dotted), Ne + /Ne (dashed),Ne /Ne (large circles). Black thick lines refer to S : S /S (dotted), S /S (solid), S /S (dash-dotted), S + /S (dashed), S /S (largecircles). Bottom diagram : the SD model for GRB 031203. Symbols as in the top diagrams. The vertical black dotted line shows the edgeof the emitting cloud REFERENCES
Allen, C.W. 1976 Astrophysical Quantities, London:Athlone (3rd edition)Anders, E., Grevesse, N. 1989, Geochimica et CosmochimicaActa, 53, 197Asplund, M., Grevesse, N., Sauval, A.J., Scott, P. 2009,ARAA, 47, 481Blanchard, P.K. et al 2015 arXiv:1509.07866Castro-Tirado, A.J. et al 2001, A&A, 370, 398Chevalier, R. A. 1982, ApJ, 259, L85Christensen, L., Vreeswijk, P.M., Sollerman, J. et al 2008,A&A, 490, 45Contini, M. 2016, MNRAS, 460, 3232Contini, M. 2015, MNRAS, 452, 3795Contini, M. 2014, A&A, 564, 19Contini, M. 2003, ApJ, 339, 125Contini, M. & Viegas, S.M. 2001a ApJS, 132, 211Contini, M. & Viegas, S.M. 2001b ApJS, 137, 75de Ugarte Postigo, A. et al 2014, A&A, 563, 62 Drake, S. A. , Ulrich, R. K. 1980 ApJS, 42, 351Fishman, G.J., Meegan, C.A. 1995 ARA&A, 33, 415Fynbo, J.P.U. et al 2000, ApJ, 542, L89Garnavich, P.M. et al 2003, ApJ, 582, 924Graham, J. F., Fruchter, A. S. 2013, ApJ, 774,119Graham, J.F. et al 2015 arXiv:1511.00667vGrevesse, N. , Sauval, A.J. 1998, Space Science Reviews,85,161Hammer, F. et al 2006, A&A, 454, 103Han, X. H., Hammer, F., Liang, Y. C., Flores, H., Rodrigues,M., Hou, J. L., Wei, J. Y. 2010, A&A, 514, 24Hjorth, J. et al. 2003, Nature, 423, 847Hook, I.M., Jorgensen, I., Allington-Smith, J.R., Davies,R.I., Metcalfe, N., Murowinski, R.G., Crampton, D.2004, PASP, 116,425Kobulnicky, H.A., Kewley, I.J. 2004, ApJ, 617, 240Kr¨uhler, T. et al 2015 A&A, 581, 125Levesque, E. M., Berger, E., Kewley, L. J., Bagley, M. M.2010a, AJ, 139, 694 c (cid:13) , 1– ?? GRB host galaxies at z < Levesque, E M., Kewley, L.J., Berger, E., Jabran Zahid, H.2010b,AJ, 140, 1557Michalowski, M.J. et al 2016 arXiv:1609.01742Modjaz, M. et al 2008, AJ, 135, 1136Niino, Y. et al 2016 Publ. Astron. Soc. Japan,arXiv:1606.01983Osterbrock, D. E. 1974 in Astrophysics of gaseous nebulae,San Francisco, W. H. Freeman and Co., 1974. 263 p.Paczynski, B. 1998 AIPC, 428, 783Pagel, B.E.J., Simonson, E.A., Terlevich, R.J., Edmunds,M.G. 1992, MNRAS, 255, 325Perley, D.A. et al 2016 arXiv:1609.04016Piranomonte, S. et al 2015, MNRAS, 452, 3293 2005, NewA,11, 103Prochaska, J. X. et al. 2004, ApJ, 611, 200Rigby, J.R., Rieke, G.H. 2004 ApJ, 606, 237Savaglio, S., Glazerbrook, K., Le Borgue, D. 2009, ApJ, 691,182Schulze, S. et al 2014, A&A, 566A, 102SSeaton, M.J. 1975, MNRAS, 170, 475Sollerman, J., ¨Ostlin, G., Fynbo, J. P. U., Hjorth, J.,Fruchter, A., Pedersen, K. 2005, NewA, 11, 103Stanek, K.Z. et al 2003, ApJ, 591, L17Th¨one, C.C. et al 2008, ApJ, 676, 1151Th¨one, C.C. & de Ugarte Postigo, A. 2014, GRB Coordi-nates Network 16079Vergani, S.D. et al 2011 A&A 535, A127Watson, D. et al 2014, arXiv:1010.1793Woosley, S.E. 1993, ApJ, 405, 273Xu, D. et al 2013, ApJ, 776, 98 c (cid:13) , 1–, 1–