Star Formation in Galaxies Along the Huble Sequence

Star Formation In Galaxies Along The HubbleSequence

Robert C. Kennicut, Jr.(Annu. Rev. Astron. Astrophys. 1998. 36:189-231)

Christian Herenz

Extragalactic Science Club 2011

March 7th, 2012

Citations from ADS

The Author

Robert C. Kennicut, Jr.

Foto Source: http://www.flickr.com/photos/swilliams2001/4029565601/

Overview

Review is basically 2 Parts:I Diagnostic methods used to measure star-formation rates

(SFRs) in galaxies.I Systematics of SFRs along the Hubble Sequence.

This talk focuses on the first part, which is not merely asummary, but

“... a self-consistent set of SFR calibrations is presented as anaid to workers in the field.”

(from the Abstract)

Outline of this Talk

I Historical Overview of SFR MeasurementsI SFR from Evolutionary SynthesisI SFR Diagnostics

1. Continuum Luminosity - Color Scaling Relation2. UV Continuum Luminosity3. H Recombination Lines4. Forbidden Lines5. FIR Continuum Emission

I Summary

Historical Overview of SFR Measurements

I Late 1960s: First quantitative SFRs from evolutionarysynthesis models (e.g. Beatrice Tinsley 1968 - GalaxyColors)

I 1970s-80s: Development of precise direct SFR calibrators.I emission-line fluxesI (near) UV continuumI IR continuum

Application to large galaxy samples. Interpretation in termsof evolutionary properties.

I 1990s - present: Detection of star-forming galaxies athigh-z. Trace evolution of SFR density with look-back time(e.g. Madau plot)

SFRs from Evolutionary Synthesis

I Individual stars in galaxies typically unresolved−→ SFRs derive from integrated light (Colors, UV, IR,recombination lines).

I Basis of all calibrators: Evolutionary Synthesis Models(ESM)

I Grid of stellar evolution tracks = T?(t) & Lbol.? (t) for various

M?.I Stellar atmosphere modells or spectral libraries: T?(t) &Lbol.? → broadband colors or spectra (=Templates(t))

I∑

weighed by IMF Templates(t) = Luminosity, Color (andspectrum) for single age population

I For different SFH: Use linear-combination of single agepops.

I 4 free parameters (at least): age, metalicity /metal-abundance, IMF, SFH (constant / e−τ ...)

Synthesis modells can be downloaded / generated online(e.g. GALEXV - Bruzal & Charlot)

SFR Diagnostic I. – Continuum Luminosity - ColorScaling

I Color dictated by ratio of early to late-type starsI Color⇒ Fraction of young (t < 109y) massive starsI Knowledge of amount of massive stars: IMF→ SFRI Scales via broad band L to total stellar mass

109y old pop., Salpeter IMF, e−τ SFH

SFRs via Continuum Luminosity - Color Scaling

Pros:I Easy applicable for homogeneous sample, when no

absolute accuracy is required.

Cons / Gotchas:I IMF dependent, age, metalicity & SFH (holds for all

calibrators)I Reddening (!)I Imprecise & prone to systematic errors

SFR Diagnostic II. - UV Luminosity

Direct tracer: UV (1250 A – 2500 A) photons produced only bymassive O – B Type Stars (no attenuation by Lyα forest & nocontribution by old stars).

Fλ of O0 (red), B8 (blue) and G5×5000 (green) type stars

For Salpeter IMF 0.1 . . . 100M�, continuous SFH tPop = 108 yrs:

SFR[M�yr−1] = 1.4× 10−28 Lν [erg s

−1Hz−1] (1)

Pros:I Directly tied to photospheric emission of young-stellar

population.I Can be used for high-redshift galaxies in the optical.

Cons / Gotchas:I Not accesible from the ground for local galaxies.I Sensitive to extinction, form of IMF (large extrapolation,

since measured M? > 5M�)I Inappropriate e.g. for young (t ∼ 107yrs) star-burst (lower

SFR/Lν ratio, i.e. less luminos for same SFR)

SFR Diagnostic III. - H Recombination Lines

Direct Tracer: Only M? > 10M� stars (i.e. t? < 2× 107yr)contribute signifcantly to F (λ < 912 A).

200Å 2000Å 3000Å 4000Å912ÅΛ

FΛ

For Salpeter IMF 0.1 . . . 100M�, continuous SFH tPop = 108 yrs:

SFR[M�yr−1] = 1.08× 10−53Q(H0)[s

−1] (2)

Q(H0)= photons with λ < 912A

From Eq. (2), using your “favorite recombination scenario”,line-strengths can be derived (e.g. using tables compiled inOsterbrock’s monograph).

For example for Case B recombination with Te = 10, 000 K:

SFR[M�yr−1] = 7.9× 10−42L(Hα) [erg s−1] (3)

= 8.2× 10−53L(Brγ) [erg s−1] (4)= . . .

⇒ Recombination Lines = “Ionizing Photon Counters” (undercertain assumptions).

Parenthesis: Case B RecombinationGas region optically thick in Lyman-Lines (generally all gas regionsthat contain enough gas to be observable - because of high Lynline-absorption cross section - because ∝ 〈ψi| − er|ψf〉- because of . . . ,)

SFRs via hydrogen recombination lines

Pros:I High sensitivity, Hα easily measureable with small

telescopes. SFR measurable in individual regions innearby galaxies.

I Directly coupled to most massive stars.Cons / Gotchas:

I Escape of Q(H0) photonsI Extinction for Hα (but e.g. not H53α in the radio)I IMF & reliability of ESMI Hα - at high-z only with JWST.

Forbidden Lines ([OII])

I z ∼ 0.5 Hα λ6563 shifts to IR→ interest in strong bluerlines⇒ [OII] λ3727 (doublet).

I Excitation dependent on abundance and ionization state ofgas, i.e. not directly coupled to ionizing flux.

I Empirical calibration to Hα Eq. 3 (using a set of ∼ 170galaxies) yields:

SFR[M�yr−1] = (1.4 ± 0.4)× 10−41L([OII]) [erg s−1] (5)

Pros: bluer, stronger Cons: Less precise

FIR Continuum

I Simplest Case: Radiation field dominated by young stars,dust opacity high everywhere (dusty circumnuclearstarburst)

I Dust: Absorbs essentially bolometric luminosity andre-emits it as thermal emission (i.e. calorimetric SFRmeasure).

I Real Situation more complex - e.g. τ � 1 approximationnot valid, dust needs to depleted by stars (i.e. oldgeneration contributes to dust heating) - etc.

I Models from literature calibrated to IMF used for otherrelations (±30%):

SFR[M�yr−1] = 4.5× 10−44LFIRo(8− 1000µm) [erg s−1]

(6)10 – 100 Myr old starburst

Summary

Several tracers (UV, Lines, FIR) for SFR exist, but quantitativecalibration is tricky and requires a set of assumptions. Providedthat for a galaxy or sample of galaxies the assumptions given inthis review are valid, the formulas

SFR[M�yr−1] = 1.4× 10−28 Lν [erg s

−1Hz−1]

SFR[M�yr−1] = 7.9× 10−42L(Hα) [erg s−1]

= 8.2× 10−53L(Brγ) [erg s−1]

= . . .

SFR[M�yr−1] = (1.4 ± 0.4)× 10−41L([OII]) [erg s−1]

SFR[M�yr−1] = 4.5× 10−44LFIRo(8− 1000µm) [erg s−1]

can be used.