Document Type

Article

Publication Date

2004

Abstract

We critique the hypothesisthat the first starswere very massive stars(VMSs; M > 140 M).We review the twomajor lines of evidencefor the existence ofVMSs: (1) that therelative metal abundances ofextremely metal-poor Galactic halostars show evidence ofVMS enrichment and (2)that the high electron-scatteringoptical depth (e) tothe cosmic microwave backgroundfound by the Wilkinson Microwave Anisotropy Probe(WMAP) requires VMSs forreionization in a concordanceCDM cosmology. The yieldpatterns of VMSs explodingas pair-instability supernovae areincompatible with the Fe-peakand r-process abundances inhalo stars. Models includingType II supernovae and/or"hypernovae" from zero-metallicity progenitorswith M = 840 Mcan better explain theobserved trends. We usethe nucleosynthesis results andstellar evolution models toconstruct an initial massfunction (IMF) for reionization.With a simple metaltransport model, we estimatethat halo enrichment curtailsmetal-free star formation after108 yr at z 20. Because thelifetime-integrated ionizing photon efficiencyof metal-free stars peaksat 120 M and declinesat higher mass, anIMF with an approximatelower bound at M 1020 M and noVMS can maximize theionizing photon budget andstill be consistent withthe nucleosynthetic evidence. AnIMF devoid of low-massstars is justified independentlyby models of theformation of primordial stars.Using a semianalytic modelfor H I and He IIreionization, we find thatsuch an IMF canreproduce e 0.100.14,consistent with the rangefrom WMAP, without extremeastrophysical assumptions, provided thatmetal-free star formation persists107108 yr after starformation begins. Because starsin the mass range50140 M are the mostefficient sources of ionizingphotons but are expectedto collapse to blackholes without releasing metals,this IMF effectively decouplesearly metal enrichment andearly ionization. Such anIMF may allow theunique properties of thezero-metallicity IMF to persistlonger than they wouldin the pure VMScase and to contributesignificantly to the globalionizing photon budget beforehalo self-enrichment and/or interhalometal transport truncates metal-freestar formation. We conclude,on the basis ofthese results, that VMSsare not necessary tomeet the existing constraintscommonly taken to motivatethem.

Comments

Copyright 2004 American Astronomical Society.

Published article available at: http://dx.doi.org/10.1086/422571

DOI

10.1086/422571

Share

COinS