Modelling Starbursts in HII Galaxies: What do we need to fit the observations?

0.34 MB·English

by  M. L. Martin-Manjon

Star-formingDwarfGalaxies: Ariadne’sThreadintheCosmicLabyrinth
Modelling Starbursts in HII Galaxies: What do we need to
fit the observations?
9
0
0
2 Mart´ın-Manjo´n, M.L. 1, Molla´,M. 2, D´ıaz, A.I.1 & Terlevich, R.3

n 1. UniversidadAutnomadeMadrid,Madrid(Spain),2. CIEMAT,Madrid(Spain),3.
a
INAOE,Puebla(Mexico)
J

2 E-mail:[email protected]
1

] Abstract.
O
We have computeda series of realistic and self-consistent modelsthat have been shown
C
to be able to reproduce the emitted spectra of HII galaxies in a star bursting scenario. Our
.
h models combine different codes of chemical evolution, evolutionary population synthesis
p and photoionization. The emitted spectrum of HII galaxies is reproduced by means of the
-
o photoionizationcodeCLOUDY[1],usingasionizingspectrumthespectralenergydistribution
r (SED) of the modelled HII galaxy, calculated using the new and updated stellar population
t
s modelsPopStar([9],inprep.).This,inturn,iscalculatedaccordingtoastarformationhistory
a
[ andametallicityevolutiongivenbyachemicalevolutionmodel.Eachmodelischaracterized

by three parameters which are going to determine the evolution of the modeled galaxy: an
2
v initialefficiencyofstarformation,thewayinwhichbursttakeplace,andthetimeofseparation
6 betweenthesebursts. Somemodelresultsemergingfromthecombinationofdifferentvalues
8
for these three parametersare shown here. Our technique reproducesobservedabundances,
1
diagnosticdiagramsandequivalentwidth-colourrelationsforlocalHIIgalaxies.
1
.
1
0
9 AMS classification scheme numbers: 98.52.Wz, 98.58.Hf, 98.62.Ai, 98.62.Bj, 98.62.Lv,
0
98.62.Qz
:
v
i
X
r
a
ModellingStarburstsinHII Galaxies 2
1. Introduction
HIIgalaxiesarecharacterizedbystrongandnarrowemissionlinesandbyalowmetalcontent,
but this does not necessarily mean that these galaxies be young systems. The current burst
of star formation (SF) dominates the SED even if previous stellar populations are present,
makingdifficulttoknowthestarformationhistory(SFH) ofthegalaxy. Wehavemadeagrid
ofattenuated star-burstingmodels,based on [5], usingsimultaneouslythewholeinformation
available for the galaxy sample: the ionized gas, which defines the present time state of the
galaxy, and spectrophotometric parameters, related to its SFH. The models are computed in
a self-consistent way, that is using the same assumptions regarding stellar evolution, model
stellaratmospheresand nucleosynthesis,and arealisticage-metallicityrelation.
2. The Star-Bursting Model
The model consists in a set of successive instantaneous bursts of star formation in a region
with a total mass of gas of 100·106 M , which take place along the whole evolution of the
⊙
galaxy in 13.2 Gyr. The chemical evolution code used is based on [8]. We obtain for every
0.7 Myr time step the abundances of 15 elements: H, He, C, O, N, Ne, Na, Al, Mg, Si, S,
Ca, Ar, Ni, Fe, the star formation rate (SFR) and the corresponding age-metallicity relation,
Z(t). With this SFH and Z(t) we assign a SED from the library Popstar [9, in prep.], with a
Ferrini IMF to each time step stellar population. When more than one burst takes place the
resulting SED is the sum of the SEDs of every stellar generation convolved with the SFH.
The final result is the total luminosity at each wavelength of the whole stellar population,
including the ionizing continuum of the last formed stellar generation. This resulting SED is
usedasionizingsourceforthephotoionizationcodeCLOUDY[1],whichgivestheemission
linesproduced bythemodelled nebula. Thegas is ionizedby themassivestars ofthecurrent
burst of SF, which is characterized by a radius R, calculated according to the mechanical
energy output of the massivestars winds and SNeI explosions, a gas density, n , the number
H
ofLymanionizingphotonsQ(H),obtaineddirectlyfromtheSEDsoftheionizingcontinuum,
and thechemical abundances obtainedfromthechemicalevolutioncode.
Each modelischaracterized by threeinputparameters:
• Theinitialefficiency(ǫ): Itistheamountofgasconsumedtoformstarsinthefirstburst
of star formation. We present here the models made with the percentages of 33% and
10% , that is, in these models, the first burst of star formation involves 33·106 M (high
⊙
efficiency model), and 10·106 M (low efficiency model), respectively.
⊙
• Attenuation(k): Theinitialefficiencyofstarformation(SF)isattenuatedintwodifferent
ways:
(i) Byafactorwhichchangeswiththenumberoftheburst,n,followingtheexpression:
1
Ψ = ( )·Ψ
n 0
n
which correspondstoasoftattenuation.
ModellingStarburstsinHII Galaxies 3
(ii) By a constantfactor, k(n−1), according totheexpression:
Ψ = Ψ ·k(n−1)
n 0
wherenisthenumberofthecurrentburstandktheattenuationfactor. Forthiswork
wehavetakenk=0.65, correspondingto astrong attenuation.
• Timebetweenbursts(∆t): Everybursttakesplaceinstantaneouslyanditisfollowedby
quiet periods, whose duration can change. For this work we have taken ∆t= 1.3 Gyr for
the inter-burst time, that is, one burst every 1.3 Gyr. For comparison purposes, we are
goingtoshowsomeresultsofmodelswith∆t=0.1 Gyrand ∆t=0.05 Gyr.
3. Results
The initial efficiency of the star formation principally leads the star formation rate and
the initial oxygen abundance. The SFR and oxygen abundances of ourmodels are between
thevaluesobservedinHIIgalaxies[3,2]. Infigure1,leftpanel,wecanseethatthefirstburst
is strong, while the subsequent ones are less intense due to the decrease of the available gas
and attenuation. The two efficiencies chosen, 33% and 10%, give the upper and lower limits
respectively for HII galaxy oxygen abundance range, as can be seen in figure 1, right panel.
Modelswiththesameinitialefficiency,butdifferent attenuationtype, arevery similar.
3 9
M/yr)sun2.25 Soft Attenuation εε==1303%% 8.5 HII galaxies upper limit
R ( 1.5
SF 1
log 0.5 8
H)
030 2 4 6 8 10 12 +log(O/7.5 HII galaxies lower limit
M/yr)sun2.25 Strong Attenuation εε==1303%% 12 7 SSSootrffott nAAgtt ttAeennttuueaanttuiiooatnni oεεn== 31ε=30%%33%
R ( 1.5 Strong Attenuation ε=10%
SF 1 6.5
log 0.5
0 6
0 2 4 6 8 10 12 0 2 4 6 8 10 12
t (Gyr) t (Gyr)
Figure1. Left: SFRofthemodelswithsoft(upperpart)andstrongattenuation(lowerpart).
Bothefficienciesareshowedineachpanel.Right:Evolutionofoxygenabundanceformodels
withdifferentinitialefficienciesandattenuationvalues.
The initial efficiency also leads the behaviour of the ionized gas. The emission lines are
producedbytheionizingphotonsofthemassivestarspresentinthecurrentburst. Infigures2
wecanseethedifferencesbetweenthemodelswithahighinitialefficiency(33%),rightpanel,
whichreproducehighexcitationandhighabundancegalaxies,withhigh[OIII]λ5007/H ,due
β
to its high efficiency of SF, and those with low efficiency ( 10%),left panel, which reproduce
less metallic galaxies, with high [OIII]λ5007/H and low [OIII]λ5007/H ratios. Differences
β β
duetoattenuationarenot soimportantsincetheSFR ofbothmodelsarevery similar.
The attenuation of the bursts (k) determines the contribution of the underlying
population: TheSFRforthesuccessiveburstsandtheoxygenabundanceevolutionaresetby
ModellingStarburstsinHII Galaxies 4
2 Hard Attenuation k=0.5, ε=10% 2 Hard Attenuation k=0.5, ε=33%
1 1
0 0
β) β)
H H
OIII]/-1 OIII]/-1
log([-2 log([-2
-3 Hoyos & Diaz 2006 -3
Izotov et al. 2006
-4 -4 Hoyos & Diaz 2006
Izotov et al. 2006
-5 -5
-1 -0,5 0 0,5 1 1,5 -1 -0,5 0 0,5 1 1,5
log([OII]/Hβ) log([OII]/Hβ)
Figure2.DiagnosticdiagramfortheStrongAttenuationModelwiththelowefficiencycaseat
theleftandthehighefficiencycaseattheright. Observationaldatafrom[2,4]. Thedifferent
colouredlinesrepresenteachSFburst,fromthefirstoneocurredatt=0Gyr(blackline)tothe
lastoneatt=13Gyr(violetline).
theattenuationtoo,keepingthemwithintherangeoftheobservations. Furthermore, ahigher
attenuation implies a larger contribution from the previous bursts to the total SED. Then, the
most important characteristics given by the adjustment of the attenuation are the colours of
the continuum and the evolution of the equivalent width of H . The evolution of EW(H ) vs
β β
a pseudo-colour of the continuum, similar to U-V, has been plotted in figure 3. In order to
reproduce the trend of HII galaxies, shifted to red colours at low values of EW(H .) due to
β
thepresenceofanonionizingpopulation,thecontributionoftheunderlyingpopulationtothe
totalcontinuummustbehigherthanthecontributionofthecurrentburstwhichdominatesthe
emission line spectrum. This trend can not be reproduced by SSPs or increasing metallicity
or age separately [5], and a strong attenuation is needed, as can be seen in the central panel,
toreproduce thewholerangeinEW(H )and colourssimultaneously.
β
The time between burst (∆t) is a secondary parameter which have an effect on the
model similar to that of the attenuation. The reduction of the time between burst offsets
the effect of increasing the attenuation: colours of the models with shorter inter-burst time
are similar to those with soft attenuation (strong bursts), and require an extra reddening to
reproduce the effects of the underlying non ionizing populations. However, the EW(H )
β
decreases from burst to burst while in the case of a soft attenuation EW(H ) maintains a
β
high value . In order to reproduce the correct behaviour of the HII galaxies, the inter-burst
timecan notbeless than100 Myrforthesemodels.
Besides this combination of values for the parameters, there are other possible
combinationswhichcouldreproduce theobservedfeatures ofHIIgalaxies[?,6, 7]
4. Conclusions
Wehavemademodelswhichconsistininstantaneousstarformationburstsspread along13.2
Gyr. In order to reproduce the observable characteristics of HII galaxies it is necessary to
ModellingStarburstsinHII Galaxies 5
3 3 3
Soft Attenuation Strong Attenuation Strong Attenuation
ε=33% ε=33% ∆τ=0.1
2.5 ε=10% 2.5 ε=10% 2.5 ∆τ=0.05
2 2 2
13.2Gyr
)β 1.5 1.5 1.5
H
(
W
E
g 1 1 1
o
l
0.5 0.5 0.5
0 0 0
decreasing ∆τ
<10 Myr
-0.5 -0.5 -0.5
-0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4
log(I /I ) log(I /I ) log(I /I )
3730 5010 3730 5010 3730 5010
Figure3. EW(H )vs. log(I /I )fordifferentattenuationparameters: asoftattenuation
β 3730 5010
model(leftpanel),astrongattenuationmodel(centralpanel),andastrongattenuationmodel
withdifferentinter-bursttimes(rightpanel).Bothinitialefficienciesarerepresentedinthetwo
firstpanelsandinthelastone,only33%modelsfordifferentinter-bursttimesarerepresented.
adjust three principal parameters. With the Initial efficiency we can vary the amount of gas
involved in star bursts, which is going to lead the SFR, the oxygen abundance, and the range
of metallicity covered by the emission lines produced by the ionized gas of the current burst
of SF. The attenuation of the burst sets the contribution of the underlying continuum from
thepreviousstellargenerationsbornbeforethestarsofthecurrentburstwhichdominatesthe
spectrum. We can also change the inter-burst time, obtaining a similar effect in colours to
the change in attenuation: decreasing this parameter, we can produce a larger contribution
fromtheunderlyingcontinuum,asanincreaseintheattenuationoftheburstcoulddo,butthe
effect in colours would be thesame as decreasing theattenuation (making theburst stronger)
thus making the spectrum bluer. At the same time, we have a larger contribution of the
underlying continuum, thus decreasing, even slightly, EW(H ). Our method, based in these
β
three parameter model, reproduce all observable characteristics of HII galaxies-abundances,
coloursand emissionlines-at thesametime.
Acknowledgments
This work has been partially supported by DGPTC grant AYA-2007-67965-C03-03 of the
Spanish Ministry of Science and Innovation. Also, partial support from the Comunidad de
ModellingStarburstsinHII Galaxies 6
Madridundergrant S-0505/ESP/000237(ASTROCAM) isacknowledged.
References
[1] FerlandG.J.,KoristaK.T.andVernerD.A.,et.al.1998,PASP,110,761
[2] HoyosC.andD´ıazA.I.,2006,MNRAS,365,454
[3] Hoyos,C.,Guzma´n,R.,Bershady,M.A.,Koo,D.C.andD´ıaz,A.I.,2004AJ,128,1541H
[4] IzotovY.I.,Stasin´skaG.,MeynetG.,etal.2006,A&A,448,955
[5] Mart´ın-Manjo´n,M.L.,Molla´,M.,D´ıaz,A.I.andTerlevich,R.,2008MNRAS.385.854M
[6] Mart´ın-Manjo´n,M.L.,Molla´,M.,D´ıaz,A.I.andTerlevich,R.,2008ASPC,396,153M
[7] Mart´ın-Manjo´n,M.L.,Molla´,M.,D´ıaz,A.I.andTerlevich,R.,2008arXiv0810.2162M
[8] Molla´,M.andD´ıaz,A.I.2005MNRAS.358.521M
[9] Molla´M.andGarc´ıa-VargasM.L.,2009,inpreparation.