Table Of Content

Conference Summary Brian W. O’Shea∗, Christopher F. McKee†, Alexander Heger∗∗ and Tom Abel‡ ∗TheoreticalAstrophysicsGroup,LosAlamosNationalLaboratory,LosAlamos,NM87545;[email protected] †DepartmentsofPhysicsandAstronomy,UniversityofCalifornia,Berkeley,CA94720; [email protected] ∗∗TheoreticalAstrophysicsGroup,LosAlamosNationalLaboratory,LosAlamos,NM87545;[email protected] 8 0 ‡KavliInstituteforParticleAstrophysicsandCosmology,StanfordUniversity,MenloPark,CA94025; 0 [email protected]) 2 n Abstract. Theunderstandingoftheformation,life,anddeathofPopulationIIIstars,aswellastheimpactthattheseobjects a hadonlatergenerationsofstructureformation,isoneoftheforemostissuesinmoderncosmologicalresearchandhasbeen J anactiveareaofresearchduringthepastseveralyears.Wesummarizetheresultspresentedat“FirstStarsIII,”aconference sponsored byLosAlamosNational Laboratory, theKavliInstitutefor ParticleAstrophysicsandCosmology, and theJoint 4 Institutefor NuclearAstrophysics. Thisconference, thethirdinaseries,tookplaceinJuly2007 attheLaFondaHotelin 1 SantaFe,NewMexico,U.S.A. ] Keywords: Cosmology;PopulationIIIstars;Conferencesummary h PACS: 98.80.-k p - o 1. INTRODUCTION and to then use systematic observation,experimentand r t theorytoconverttheseto"knownknowns." s In July 2007, more than 130 international researchers The study of the first stars is a new field. There are a [ met at the La Fonda Hotel in Santa Fe, New Mex- almost no “known knowns," except for the background ico, U.S.A. to discuss the formation, life, and death of cosmologyandan increasingset ofdata onabundances 1 zero-metallicity(PopulationIII)andverylowmetallicity in very metal poor stars. There are many “known un- v 4 stars,aswellastheimpactthattheseobjectshadonlater knowns:" What is the nature of dark matter? What is 2 generationsofstarsandonstructureformation.Thisfield the strength of the magnetic field? What are the abun- 1 has made significant theoretical and observational ad- dance and size distribution of dust? What is the rate of 2 vancessinceFirst StarsI,whichwasa MPA/ESO/MPE mixingofmetalsintoprimordialgas?Untilrecently,the . 1 Joint conferenceheld in 1999in Garchingb. München, effect of dark matter annihilation on the first stars was 0 Germany,andFirstStarsII,whichwasheldatThePenn- an“unknownunknown,"butthishasbecomea “known 8 sylvania State University in State College, Pennsylva- unknown"throughtheworkofFreeseandhercollabora- 0 nia,U.S.A.,in2003.Thoughmajoradvanceshavebeen tors. : v made,the understandingof PopulationIIIandverylow We conjecture that the number of “unknown i metallicitystarsisstillinitsinfancy,andmanyimportant unknowns"—i.e., the discovery potential—in a field X questionsremainunanswered. is proportional to the ratio of the number of “known r Donald Rumsfeld is known for many things, but his unknowns"to“knownknowns:" a contribution to the philosophy of science is less well recognized. In an oft-cited speech he delivered on Feb UU (cid:181) KU/KK. (1) 12,2002,hesaid: Ifso,thestudyoftheFirstStarsisanexcellentfieldfor There are known knowns. These are things youngpeople—asreflectedintheyouthoftheaudience we know that we know. There are known un- andtheorganizers(althoughnotinthesummarizer)! knowns.Thatistosay,therearethingsthatwe Inthe followingsections,we shallsummarizethe re- know we don’t know. But there are also un- sultspresentedatthisconference,andraisesomesignifi- known unknowns. There are things we don’t cantissuesthathaveyettobeexplored.Intheinterestsof knowwedon’tknow. conservingspace, we discuss only the results presented duringFirstStarsIII,anddirectinterestedreaderstoin- Heappliedthistointelligencedata,butonecandescribe dividualcontributionsformorein-depthinformationand ourtaskasscientistsistouncover“unknownunknowns" citationstorefereedpapers.Inaddition,wehavemadea bypurethoughtorbyserendipitousobservationorexper- concertedefforttoreferonlytocontributionsinthispro- iment, thereby converting them to “known unknowns;" ceedings,anddosobythelastnameofoneoftheauthors exceeds the “critical metallicity” discussed by Bromm, (typicallythefirst)ofeachcontribution. Glover, Omukai, Shull, and B. Smith – namely, the metallicity where the enhanced cooling properties of metal-enriched gas affects the star formation process, 2. A PROPOSEDNAMING possiblyresultinginadifferentIMF. CONVENTION FORPRIMORDIAL AND METAL-POOR STARS 3. FORMATIONANDIMF OFSTARS AT ZERO AND LOW METALLICITY AsignificantpointofconfusionintheliteratureonPop- ulationIIIandlowmetallicitystarsarisesfromtheabun- danceof (often contradictoryand/orconfusing)naming Theultimategoalofthetheoreticalandnumericalwork conventions. Given that essentially all researchers in- beingperformedbymanyinvestigatorsistogainafunda- volvedinthefieldwerepresentatFirstStarsIII,adiscus- mentalunderstandingoftheInitialMassFunction(IMF) sionofthissubjecttookplaceandthefollowingnaming of Population III (Pop III) stars. This is also true of re- conventionhasbeenproposed: search relating to star formationin the present-dayuni- PopulationIII: This is a blanketterm that describes verse–PopulationIIIstarformation,however,isconsid- allstarsofprimordialcomposition(i.e.,gaswhosecom- ered to be a more tractable problem, given the relative positionwasdeterminedduringBigBangNucleosynthe- simplicityofthephysicsinvolved. sisandasaresultiscomposedalmostentirelyofhydro- IntheabsenceofobservationsofPopulationIIIstars, genandhelium),regardlessofhow,when,orwherethey wemustrelyontheoryandsimulations.Asdiscussedby formed.Ithasbeenfoundthatthistermistoobroad,and Norman,boththeoryandsimulationspredictthatthefirst thustheneedfortheexistenceof“PopulationIII.1”and stars were massive, with an average mass that is much “PopulationIII.2”stars,asdescribedbelow. greater than stars in the Milky Way. This is consistent PopulationIII.1:Thesearethetrue“firstgeneration” withtheabsenceofobservedzero-metallicitystarsatthe stars of primordial composition, whose properties have presentday.SimulationsshowthatnotallPopulationIII been determined entirely by cosmological parameters stars form in exactly the same fashion, however,which andtheprocessofcosmologicalstructureformation,and suggestsabroadspreadofmasses.Inaddition,thereisno havenotbeensignificantlyaffectedbypreviousstarfor- evidenceforfragmentationinextremelyhigh-resolution mation.ExamplesofPopulationIII.1starformationare simulationsof PopulationIIIstar formation,evenwhen shownincontributionsbyNorman,Turk,andYoshida. theory suggests that there should be (though note that Population III.2: These are “second generation” Umemura and Suwa displayed preliminarycalculations stars,whichstillhaveprimordialcomposition.Theirfor- that indicate fragmentation in an asymmetric runaway mation,however,hasbeensignificantlyaffectedbypre- collapseofprimordialgas). vious generations of star formation through the injec- Turk and Yoshida presented simulations of Popula- tion of kinetic energy, photodissociatingor ionizing ra- tionIII.1starformationusingdifferentmethods.Yoshida diation, by cosmic rays, or by other as-yet-unsuggested used the Gadget-2 smoothed-particle hydrodynamics processes, which may change the mass range of these code, while Turk used the ENZO adaptive mesh refine- stars.ExamplesofPopulationIII.2starformationaredis- ment code. Both codes have been improved in the past cussedincontributionsbyAhn,Bromm,Bryan,Johnson, few yearswith the additionofextended-precisionarith- McGreer,Sato,Norman,UmemuraandWhalen. metic, particle splitting techniques (for SPH), updated PopulationII.5:Thisisthesuggestedtermforapos- chemistry,approximationsforcoolingthattake into ac- sibleclassofstarswithnon-zerometalcontentwherethe count the non-negligible optical depths at high densi- amountofmetalsinthegasthatthestarsareformedfrom ties, and modifications to the ideal gas law equation of wouldnotbesufficienttoaffectthecoolingpropertiesof state atextremelyhighdensity.Theseimprovementsal- the gas and thus the formation of the stars, but would low both codes to simulate large (∼ Mpc) volumes of play a non-negligible role in the star’s main-sequence the universe while at the same time following the col- evolution. An example of this class of object (as dis- lapseofgastoprotostellardensities,withacurrentmax- cussedbyMeynet)isarapidlyrotating,massivestarwith imumdensityofnH ≃1021 cm−3!Thesefundamentally ametallicityof≃10−6Z⊙,whichwouldhaveformedin different methods agree quite well to sub-parsec scales a manner identical to a star of primordial composition, – there is no evidence for fragmentationin the collaps- but would experience enhanced mass loss relative to a ing primordial cloud cores, and the inferred accretion Pop III star of otherwise identical propertiesdue to the rates of gas flowing onto the evolving protostellar core smallamountofmetalsinthestar. areextremelyhigh,peakingatM˙ ≃10−2−10−1M⊙/yr. Population II: These are stars whose metal content Atsmallerscales,somedifferencesareapparentintheir calculations. It is unclear, however, whether this is due Bromm (Greif et al.) also showed that the presence of tonumericalissuesorsimplyduetotheuseofdifferent acosmicraybackgroundcancausetheformationofsig- cosmologicalrealizations. nificant amounts of HD in high-density primordialgas, Gloverdiscussedissuesrelatingtouncertaintiesinthe againloweringgastemperatureto that of the CMB and 3-bodymolecularhydrogenformationreactionrates.At reducing the accretion rate. This is not to say that all thetemperaturesrelevanttoPopulationIIIstarformation Population III.2 star-forming regions are characterized (T <1000K),theoreticalcalculationsoftheseratesvary byreducedaccretionrates:Normanshowedresultsfrom by2−3ordersofmagnitude. simulationsofPopIIIstarsforminginthepresenceofa Theuns showed that Pop III star formation in warm molecular hydrogen photodissociating (Lyman-Werner) dark matter may occur in a very different manner than background,whichresultsinhigheroverallhalotemper- inaCDMuniverse,possiblyleadingtofilamentaryfrag- aturesandhigherinferredaccretionratesontotheproto- mentation and the formation of intermediate-mass pri- stellarcore. mordial stars. Freese presented results suggesting that The transition between Population III and Popula- if neutralinos comprise a significant fraction of dark tion II star formation was discussed at length during matter,theirannihilationcouldpossiblyoverwhelmany the conference. Bromm (Greif et al.) suggested the ex- cooling mechanism in high-density primordialgas, and istence of a “critical metallicity,” or Z , where metal crit may result in a “dark star” – a massive object powered line cooling dominates over molecular hydrogen cool- by dark matter annihilation instead of nuclear fusion. ing at low temperature (and hence presumably where ZhaoandXubothshowedpreliminaryresultsfromAMR the stellar IMF will transition from being top-heavy to simulationsincludingthe effects of magnetic fields and aSalpeter-typefunctionwithasignificantlylowermean their productionvia the Biermann battery in cosmolog- mass), at 10−4 < Zcrit/Z⊙ < 10−3. Glover and Shull ical simulations, which may have significant effects on bothsuggestthatZ maydependstronglyonenviron- crit the formation of Pop III stars at small mass and spatial ment. Omukai showed that dust is an extremely effec- scales. tive coolantcomparedto gas-phasemetals, and that the An importantissue with the presentedsimulationsof presence of a small amount of dust can radically lower PopulationIIIstar formationis thattheyarefundamen- thecriticalmetallicity.Thedegreewithwhichthistakes tallyCourant-limitedathighdensities,sothatthesimu- place dependsstronglyon the abundance,typeand size lationsinevitablygrindtoahaltasthecollapseproceeds. ofthedustgrains,andishighlyuncertain.B.Smithused Intheabsenceofa meansto avoidthis(e.g.,sinkparti- highly-resolvedadaptivemeshrefinementcalculationsto cles),understandingoftheshutoffofaccretionontopri- examinethefragmentationofmetal-enrichedgasatsmall mordialprotostarscanonlybeobtainedanalytically.To scalesandobservedthatthereisnotaclearrelationship thisend,Tanpresentedanalyticandsemi-analyticcalcu- betweenmetallicityandfragmentationscale–rather,the lationsoftheevolutionoftheinnerregionsofprimordial fragment scale is determined by the density of the gas halosthatfollowthegrowthandevolutionoftheproto- whenitreachesthetemperatureoftheCMB.Inthisway, star, formationof an accretion disk, and feedback from higher-metallicity gas may reach the CMB temperature theevolvingprotostarandmain-sequencePopulationIII at lower densities than lower-metallicity gas, resulting star. He showed that the effectiveness of feedback pro- in larger overall clumps. Jappsen performed SPH sim- cessesinhaltingaccretionontothe star(andthusdeter- ulations of metal-enriched gas collapsing in an ionized mining its final mass) depend on core rotation and on halo.She showedthat,fordensitiesn <104cm−3, the H the rate of accretion of gas onto the disk. Estimates of evolutionofdensityandtemperaturearenotchangedby accretionratesfromhigh-resolutioncosmologicalsimu- metallicity for Z <0.1Z⊙, because H2 is the dominant lations suggest stars with masses of 60−400M⊙, with coolant rather than metal fine structure lines. In addi- themain-sequencestellarmassinthefiducialcasebeing tion, she does not find evidence in her calculations for ≃160M⊙. the “critical metallicity” threshold proposed by Bromm Some aspectsof PopulationIII.2star formationwere (Greifetal.)Clark,however,usedhigh-densitySPHsim- also discussed. Bromm (Greif et al.), Bryan, Yoshida, ulations to show that fragmentation can easily occur in andMcGreerpresentedresultsshowingtheformationof the high-density,dust-dominatedregime at metallicities primordialstarsinregionsofpre-ionizedgas,usingtwo atorbelowZ=10−5Z⊙. differentmethods.All agree thatsignificantamountsof HD form in these regions, and result in a rapid cooling of gas down to the temperature of the CMB at the red- shiftofformation.Thislowertemperaturedirectlytrans- lates into a reduced accretion rate onto the protostellar core, suggestingthat the resulting PopulationIII.2stars will be less massive than their Pop III.1 counterparts. 4. SEARCHES FORPOPULATIONIII IIIstars.Kashlinskypresentedresultsfromobservations ANDVERY LOWMETALLICITY using the Spitzer Space Telescope, arguingthat source- subtracted IRAC imagescontain significantCIB fluctu- STARS ANDOBSERVED ationsthatareinexcessofthenear-infraredbackground ABUNDANCEPATTERNS expected from resolvedgalaxy populations.These fluc- tuations appear to come from clustered sources that do 4.1. The search forPopulationIII stars not correlate with Hubble ACS source catalog maps of thesamefield.HeclaimedthatthisimpliesthattheCIB Todate,nostarsofprimordialcompositionhavebeen fluctuationscome from dim populationsat high (z>8) directly detected, either in our galaxy or in the distant redshifts,withmuchlowermass-to-lightratiosthancur- universe.Thismaybebecausethesestarsareverymas- rent galaxies, and that have a projected numberdensity siveandthusshort-lived,orbecausesurveysarelooking of 5−10 sources/arcsec2 (within the confusionlimit of in thewrongplaces. Ifthese objectsstillexist, howcan present-day instruments, but resolvable by JWST), and they be detected? And, have we already indirectly de- claimed that this source populationis high-redshiftpri- tectedthetracesofPopulationIIIstars? mordialstars.Thompson,ontheotherhand,claimedthat If,asTansuggests,PopulationIII.1starshavemasses thepurportednear-infraredbackgroundexcessisdueto of on the order of 100−400M⊙, some of these stars impropermodelingandsubtractionofzodiacallight,and mayexplodeaspairinstabilitysupernovae(PISN),asde- thatthecosmicinfraredbackgroundfluctuationsdetected scribed by Woosley. These objects would have unusual byKashlinskyaremainlyduetolow-redshift(0.5<z< nucleosynthetic patterns, which have not yet been ob- 1.5)galaxiesandareinconsistentwithgalaxiesatz>10. servedinabundanceratiomeasurementsofgalactichalo Fernandezshowedthatmetal-freestarsare nottheonly stars.Thislackofevidencesetsstrongupperlimitsonon possible source of the near-infrared background excess thenumberofprimordialstarsinthismassrange. – her modelspredictthatstars with metalscan produce N.SmithpresentedobservationsofSN2006gy,aType thesameamountofdiffuseradiationinthe1−2micron IIn supernovain NGC1260(a S0/Sa galaxylocated ap- bandasprimordialstars,becausetheaverageintensityin proximately73Mpcaway).Thiswasthebrightestsuper- thiswavebandisdeterminedbytheefficiencyofnuclear novaknownatthe time of the conference,and the light burninginstars,whichisnotverysensitivetometallicity. curve and inferred expansion velocity are inconsistent with other Type II supernovae and cannot be explained bytheinteractionofastandardTypeIIsupernovawitha 4.2. The search forlowmetallicitystars circumstellar medium.The observedlight curveis consistent with results presented by Kasen showing theo- Metal-poor galactic halo stars are of great interest to retical predictions of pair production supernova model the community that studies structure formation in the lightcurvesandspectra,lendinghopethatPopulationIII earlyuniverseforseveralreasons.Extremelymetal-poor PISN exist.In principle,a pairinstabilitysupernovaeat stars are possibly fossil records of the heavy element z∼20wouldbebrightenoughatpeakluminositytobe abundances produced in a single Population III super- observable with JWST, though the time dilation of the nova.Theshapeofthelow-metallicitytailofthemetal- lightcurvecouldmakedetectionextremelydifficult.This licity distribution function (MDF) has the potential to would be much less of a problem if, as suggested by probe epochs of star formation in the early galaxy, and Schneider, a non-negligiblefraction of stars forming at thechangeoftheMDFwithdistancefromthecenterof z∼3maybeprimordial.Inaddition,ifasignificantfrac- the galaxy may provideuseful clues as to the assembly tionofPopulationIIIstarsareGRBprogenitors(assug- history of the Milky Way. In addition, the discovery of gested by Yoon), JWST followups of SWIFT-detected objects with enhancements of combinations of various gammarayburstsmayyielddirectobservationsofPopu- a -,s-,andr-processelementscanproberareeventsinthe lationIIIsupernovae.AnothermodelforSN2006gywas formation history of our galaxy. Ferrara and Salvadori, presented by Woosley, a pulsational pair instability su- however,arguethatalmostno true “second generation” pernovaofabout110M⊙.Inthisscenariothesupernova- stars(objectsthathavebeenenrichedbya singlePopu- like ejecta of subsequent pulses run into each other at lationIIIsupernova)shouldbeobservableatz=0. largedistanceandconverttheirkineticenergyatalmost The discovery of metal-poor galactic halo stars and 100%efficiencyintoradiation,makingforaverybright spectroscopic estimates of their elemental abundances display.Generally,themassrangeforpulsationalpairin- has become a major industry in the past several years. stabilityinnon-rotationprimordialcomposition(PopIII) TheSDSSSEGUEproject(asdiscussedbyBeers)isal- starsisabout100−140M⊙. readyunderway,andhasdiscoveredmorethan2400stars Examination of cosmic infrared background (CIB) with [Fe/H] <−2 and dozens with [Fe/H] <−3. Over fluctuationsmay yield indirectdetectionsof Population the course of the project lifetime, SEGUE is projected and nota main-sequencestar, and thatatomic diffusion to find roughly20,000stars with [Fe/H] <−2, and ap- mayhavesignificantlyalteredthesurfaceabundancesof proximately2000with[Fe/H]<−3.The0ZProject(as thestar(thoughnotenoughtofullyexplainthepuzzling summarizedbyCohen)describesthediscoveryofmore non-detectionoflithium).Frebelalsopresentedworkus- than1500candidateextremelymetalpoorstarsfromthe ing observationsof carbon and oxygen-enhancedmetal Hamburg/ESOSurvey,withsomestarshavingextremely poor stars to examine necessary conditionsfor forming peculiar abundance ratios. The ESO LAMOST Survey low-massstarsintheearlyuniversethroughcoolingvia andSouthernSkySurvey(describedbyChristlieb),will fine-structure lines, and showed evidence for a “transi- begin production in the near future, will cover signifi- tion metallicity” where the sum of the carbon and oxy- cantlylargerareasof thesky thanSEGUE (to a similar gen abundance is approximately 10−3.5 the sum of the magnitude)and will, as a result, find thousandsof stars solarabundancesfortheseelements,whichsupportsthe withmetallicitiesof[Fe/H]<−3inthenextfewyears. Brommideaofa“criticalmetallicity.” Giventhelargenumberofextremelymetal-poorstars Johnson argued that carbon-enhanced metal poor that have been observed, and the even larger number (CEMP)starsarecreatedbythepollutionofalow-mass of stars that should be found by surveys that are cur- star by a companion asymptotic giant branch (AGB) rently underway, a question arises – how much can we star. She suggested that the observed [C/N] ratios im- trusttheelementalabundancesmeasuredinthese stars? ply that a largenumberof primordialstars with masses This is a crucial issue when one is interested in com- of 2−3 M⊙ were created. This result was supported paringthe resultsfrom simulations(shownby Nomoto, by Schuler,who presented observationsof fluorine in a Woosley, and others) with observationally-determined CEMPhalostar.AGBstarsarebelievedtobeprodigious stellar abundances. Sneden discussed some aspects of producers of both carbon and fluorine, suggesting that thisimportantissue,anddemonstratedthatcorrectmea- AGBstarsmaybethesourceoftheobservedabundance surement of abundances depends strongly on accurate patternsinatleastsomeCEMPstars. Both oftheseob- determinationofthestar’seffectivesurfacetemperature, servational results are supplemented by Pols, who pre- and that abundances derived from a small number of sented simulations of binary stellar evolution showing strong lines may be unreliable. In addition, laboratory that the observed CEMP abundance ratios can be ex- data on transition probabilities of alpha elements is of- plained by binary pollution from AGB stars, and that ten contradictory or completely lacking, and that more thermohalinemixingisimportantinthemetal-poorcom- accuratemeasurementofthepropertiesofiron-peakele- panion star, and by Husti, who presents similar results. mentsisnecessaryinordertoaccuratelyinterpretabun- ThislattereffectisexaminedinmoredetailbyBisterzo, dance data. Venn suggests that stars with the lowest who presented numerical experiments detailing the ef- [Fe/H]abundances(thathave[Fe/H]<−5),whichseem fects of thermohaline mixing on nucleosynthetic yields to be chemically peculiar, may actually appear to be resulting from neutron-capture nucleosynthesis. These thatway dueto dust-gasseparation.If this is true, stars calculations show that the yields for CEMP stars can with−4<[Fe/H]<−3mayactuallysamplemetalen- bewell-matchedwhenthermohalinemixingisincluded. richmentfromsingle PopulationIII stars. She also sug- Lucatello showed results from the HERES survey suggests searching nearby dwarf galaxies for single pollu- gesting that there are different classes of CEMP stars, tioneventstarsinadditiontothehaloofourgalaxy. basedonthepresenceorabsenceofr-ands-processelements.Thesestarsmightbeenrichedbydifferenttypes ofstars,orviadifferentmechanisms,andthemonitoring 4.3. Observedabundance patterns in of stellar radial velocities may give some clues to their differentorigins.Rossi showedthatestimatesofcarbon low-metallicitystars abundances in CEMPs may be affected by calibration problems,andpresentedasetofrefinedestimatesofcar- Much can be inferred from the observed abundance bonabundancesforalargesampleofCEMPs. patterns of extremely metal-poor galactic halo stars, deMinkpresentedresultsfrombinaryevolutionatlow though the concerns discussed in the previous section metallicity. She showed that binaries at low metallicity mustbekeptinmindwheninterpretingtheseresults. experience mass transfer in a very different way than Frebel presented an update on the abundance of HE solar-metallicity stars do, which may have implications 1327-2326 using spectra from the VLT, confirming its for the metallicity dependence of the formation rate of extremelylow ironabundanceandcarbonenhancement [FeI/H]=−5.7or−5.5,dependingonassumptions,and various objects through binary evolution channels. She [FeII/H]<−5.4,and[C/Fe] =3.28usinga3Dmodel alsoshowedthatlow-metallicitybinariescanexperience LTE much higher rates of accretion before reaching a given atmospherecorrection).Thisworkwaselaboratedupon size comparedto solar-metallicity stars, suggestingthat by Korn, who showed that HE 1327-2326is a subgiant fewer low-metallicity binaries come into contact (and thusmerge)duringrapidmasstransfer. other star. This is an important step, however, towards Masseron presented results from the high-resolution beingabletocomparenucleosyntheticpredictionsforthe chemicalanalysisofalargesampleofCEMPstars,and s-andr-processusingthistechnique. showed that these stars naturally split into two groups: Lithium is an element of great cosmological signif- starsenrichedonlyin s-processelements,andthoseen- icance. There are two stable isotopes – 6Li and 7Li – riched by both r- and s-process elements. The former with 7Li being formed during the epoch of Big Bang group is well explained by AGB mass transfer, and Nucleosynthesis(BBN), butnot6Li. Thereis an appar- Masseron suggests that the latter type of star can be ent discrepancy between the measured amounts of 7Li explained by a primordial companion with a mass of and the predicted amounts based on recent determina- 8−10M⊙.Tumlinsonarguedthatthetwoobservedhy- tions of the baryon-to-photon ratio, with the observed per metal poor stars were most likely formed in mass- amount of 7Li being too low. Asplund pointed out that transferbinarieswithatop-heavyIMF,andthatthehigh there are actually two cosmologicallithium problems– frequency of CEMP stars at [Fe/H] < −2, and gradi- the observedabundanceof 7Li is inconsistentwith pre- entswithmetallicityandlocation,suggestthattheCMB dictionsfromstandardBBN,and6Liisapparentlyincon- setsthetypicalmassofearlyPopIIstarsindependentof sistentwithgalacticcosmicrayproduction.Hesuggests metallicity. Komiya also suggested that the IMF of ex- arangeofpossiblemechanismstocausetheseinconsis- tremelymetal-poorstarsistopheavy,andthatthisiscon- tencies, but cautions that they state inconsistencies are sistentwiththeobservedmetallicitydistributionfunction basedonextremelychallengingobservationsandanaly- inmetal-poorhalostars. sis, and should be taken with a grain of salt. Perez pre- Sobeckpresentedresultsfromastudyofcopperabun- sented observations of two metal-poor stars suggesting dancesin50metal-poorhalostarsusingtheVLTUVES isotopic abundance ratios of 6Li/7Li ≃ 0.04−0.05. In spectrograph. There appears to be a deficit of copper one case this is similar to results foundin the literature in these stars, implying the reduced extent of weak s- forthesamestar,andintheothercaseherderivedvalue processinmassivestarsatlow[Fe/H]orthedelayedpro- is significantly higher. Perez suggested that the derived duction from Type Ia supernovae. Boesgaard discussed abundanceratioisextremelysensitivetoparametersused berylliumabundancesinasampleof51metal-poorhalo intheanalysis.Sbordonealsofoundthissensitivity,and dwarf stars that have been examined at high resolution showed that effective temperature scales for extremely and signal-to-noise using the Keck HIRES and Subaru lowmetallicitystarsarestillpoorlycalibrated,leadingto HDSspectrographs.Thereisnoevidenceforaberyllium inaccurate determination of lithium isotopic abundance “plateau”–rather,[Be/Fe]remainsconstantwithchang- ratios. ing[Fe/H]–andalsonoevidenceforadifferenceinBe vs.[O/Fe]instarsthatarefoundinthehaloversusthose observedin thethick galacticdisk,contraryto previous 5. STELLAR EVOLUTION, lower-resolutionwork. EXPLOSIONSAND CowanpresentedresultsfromHSTandground-based NUCLEOSYNTHESIS ATZERO AND studiesofgalactichalostars,andarguedthatatthetime VERYLOW METALLICITIES these stars formed the galaxy was chemically unmixed andinhomogeneousinr-processelements,butnotina - 5.1. Stellar EvolutionandExplosions at elements, and suggests different environments for the synthesis of these elements. There is evidence for in- Very Low Metallicities creasing contribution of the s-process with metallicity (and thus galactic age). Krugler presented results from In his review talk, Woosley made several points. He autoMOOG analysis of over 6500 stars from SDSS-I claimsthattheknownabundancesinlowmetallicitystars and SEGUE that have estimated metallicities of [Fe/H] canbefitby“ordinary”supernovaeinthe10−100M⊙ < −2 and effective temperatures of 4500K < T < range, and that there is no need for hypernovaeor pair eff 7000K.Thistechniqueproducesestimatesof[Fe/H]and instabilitysupernovae.HearguesthatthefavoredPopIII [Ce/Fe] (or upper limits on these quantities). Roederer stellarmassesbasedonnucleosyntheticresultsare10− derived isotopic fractions of europium, samarium and 20M⊙,withexplosionenergiesofroughly1051ergs,and neodymium in two metal-poor giants with apparently littlemixing.Healsosuggestedthatmetal-deficientstars differentnucleosynthetichistories, extendingthe exam- will produce many more black holes than their metal- ination of the neutron-capture origin of multiple rare- rich counterparts, with masses as high as 40M⊙. Fi- earth elements in metal poor stars to the isotopic level nally, he argued that the pulsational pair instability can forthefirsttime.Theresultssuggestanr-processorigin give a wide range of light curves, from faintest to the for the rare earth elements in one of the stars, and can- brightestobservedsupernovae.Yoonpresentedsomere- not distinguish between an r- or s-process origin in the sults based on simulations of massive stars at very low metallicity,includingrotationandbinaryinteraction.He and suggested that these stars may be responsible for claims that the effects of rotation are particular impor- CEMP stars. Ludwig discussed hydrodynamicalmodel tant for the evolution of massive stars at low metallic- atmospheresofmetal-poorstars,andshowedthatmetal- ity, and that a large fractionof these stars may produce poorstellaratmospheresarepronetoexhibitingsubstan- GRBs or hypernovae with unique nucleosynthetic fea- tial deviations from radiative equilibrium. He also sug- tures. Additionally,he suggested that a significantfrac- gested that large abundancecorrectionsmay have to be tion of massive stars may be Wolf-Rayet stars at very madetotakeintoaccountassumptionsmadein1Datmo- lowmetallicity,andthatstarsofthistypeinclosebina- sphere calculations. In addition, the three-dimensional ries may also produce GRBs. Meynet also spoke about effects on the effective atmospheric temperature from the evolution of Pop III and very metal poor stars, and Balmer lines show a complex pattern, further compli- stated thatthe effectsof rotationonPop III stars issig- cating analysis. van Marle showed results from simula- nificant, but less extreme than in very metal poor stars. tions of continuum-drivenwinds from primordial stars, He also suggested that a tiny amount of metals (on the and argued that continuum driving can produce strong order of Z =10−8Z⊙) may make a big difference, and mass loss from stars, even without metals. Krticka pre- thatrotationisakeyparameterforverymetalpoorstars. sented the possibility that hot Population III stars may Ekstrom argued that under certain conditionsit may be experiencesignificantmasslossiftheiratmospheresare possibleforverymassiveprimordialstarstoavoidpair- enriched by CNO elements through metal line-driven instabilitysupernovaewiththehelpoftwoeffectsofro- winds, with a range of effects depending on the metal- tation: anisotropic winds and magnetic fields. This will licity ofthe atmosphere.Muijrespresentedresultsfrom happen even with the assumption of reasonable initial simulationsexploringthe effectof clumpingon predic- equatorialvelocities.Chiappinipresentedfurtherresults tions of mass-loss rate of early-type stars, and suggests regardingthe impactof stellar rotation on chemicalen- that the difference between theoretically expected and richmentatlow metallicity,showingthatfast stellar ro- empiricallyderivedmass-lossratesmaybeduetoinho- tation at low metallicities is the only thing that can ex- mogeneities.Shepredictsthatclumpingleadstoahigher plaintheobservedabundancesinmetal-poorhalostarsin mass-loss rate, with only modest clumping factors re- theabsenceofAGBbinarymasstransfer.Laupresented quired to match observed values of mass-loss rates in simulationresultsfromexplosionsofintermediate-mass, massive stars. Onifer showed attempts to calculate the zero-metallicity stars, and showed that the stellar en- mass-lossrateofaPopulationIIIWolf-Rayetstarusinga velopesof these stars will be enrichedby nitrogen,and modifiedversionoftheCAKapproximation.Heshowed thatthere is nos-processenrichment,owing to the lack that even a star with zero initial metallicity will experi- of a third dredge-up. Gil-Pons presented results from ence significant mass loss due to radiation pressure on simulations that examine the effects of overshooting in dredged-upnucleosyntheticproducts. theevolutionofintermediate-massstarsthatagreeswell Church presented multidimensional simulations of withLao’sresult.Brottdiscussedtheefficiencyofrota- primordial supernovae, which investigate the effects of tionalmixinginmassivestars,andfindsthatwhileinter- Rayleigh-Taylor-inducedmixingandasymmetriesinthe nal magnetic fields are necessary to understandangular explosion on the final composition of the escaped gas. momentumtransport(and thus rotationalbehavior),the She finds that for spherically-symmetric explosions, correspondingchemicalmixingmustbeneglectedinor- mixing has little effect on the shells interior to oxygen, derto reproduceobservations.She also showedthatfor and that some asymmetry is needed in the explosionin low metallicity stars, detailed initial abundances are of orderforelementsinteriortooxygentoescapefromthe primaryimportance,sincesolar-scaledabundancesmay star. Nozawa showed the evolution of dust grains that resultinsignificantcalibrationerrors.Tsurutapresented formed in Population III supernovae, and argued that results from simulations of the evolution of very mas- dust grain size and composition strongly affects their sive (500M⊙ and 1000M⊙) Population III stars, and evolution. He also demonstrated that small grains are finds that though these stars experience the pair insta- preferentially destroyed in primordial supernova rem- bility, they eventually undergo core collapse and create nants,with themaximumdestructionmassvaryingasa blackholesthat,inthemoremassivestar,maybeupto functionoftheambientmediumproperties. 500M⊙. Polspresentedtheoreticalresultspertainingtocarbon- enhanced metal-poor (CEMP) stars, arguing that these 5.2. Nucleosynthesis atZero and Very Low stars are in binary systems and have been polluted by Metallicities a former AGB companion, and also suggesting that it is important to study nucleosynthesis together with bi- Qian presenteda reviewofnucleosynthesisof metal- nary evolution. Suda discussed the stellar evolution of free and metal-poor stars. In his talk, he made several low- and intermediate-massextremelymetal poorstars, points;hesuggestedthattheoriginof6Liinlowmetal- hole accretion disks. She suggests that these accretion licity stars is still a puzzle; that the presence of nitro- disksmaybeimportantcontributorstothenuclearabun- gen in low metallicity stars indicates rotation; and that dancesintheoldeststars,particularlyforrarespeciesor the presence of lead in metal-poor binary members in- thosenotuniformlyobserved. dicates that the s-process occurs at low metallicity. He also arguedthatthe observedabundancesin metal-poor galactic halo stars suggests standard supernovae rather 6. EARLY FEEDBACK PROCESSES: than pair-instability supernovae, but that there are con- INFLUENCE ONSTRUCTURE tributions from both hypernovae and faint supernovae. FORMATION,REIONIZATION,AND Nomoto discussed nucleosynthesis in massive Popula- IGMABUNDANCEPATTERNS tion III stars, focusing on hypernovae and jet-induced explosions. He showed that explosions with large en- The first generations of stars are potentially prodigious ergydepositionrateswill create gammarayburstswith sourcesofmechanical,chemical,andradiativefeedback, hypernovae and that their yields can explain the abun- and may contribute significantly to the enrichment and dances of normal extremely metal-poor stars, and that reionization of the intergalactic medium. Reed showed explosionswithsmallenergydepositionrateswillbeob- that high-redshifthalosare stronglybiased, particularly servedasGRBs withouta brightsupernova,andcan be at small scales – this effect differs from low-redshift responsiblefortheformationofCEMPandhypermetal bias,andisa functionofmass.Theclusteringobserved poorstars.Kratz(Farouqietal.)presentedresultsexplor- willsignificantlyaffectthestatisticalpropertiesoffeed- ingnucleosynthesismodesinthehigh-entropy-windsce- backathighredshifts.Ciardiarguedthatfeedbackfrom nario of Type II supernovae, and showed that a super- Population III stars is generally not as efficient as one position of several entropy components can reproduce mightexpectfromenergeticsestimates,andthatobjects the overall solar system isotopic r-process residuals, as wellasthemorerecentobservationsofelementalabun- with masses of M >107−8M⊙ are not greatly affected by feedback. Wise presented AMR simulations model- dancesofmetal-poor,r-process-richhalostars.Pignatari ingtheformationofasignificantnumberofPopIIIstars presented results of simulations examining the weak s- inasinglevolume,andshowedthatdynamicalfeedback processatlowmetallicity,andshowedthatthes-process fromHII regionsand supernovaecan expelmostof the efficiencychangessignificantlyasafunctionofbothstel- gasinfirst-generationhalosandleadtolowbaryonfrac- larmassandmetallicity. tions in subsequent star-forming halos. In addition, he Woodward and Herwig (combined proceedings) dis- showed that metals are well-mixed within dwarf galax- cussed nucleosynthesis and mixing in the first genera- ies,andthatejectafromPopIIIsupernovaewillprovide tion of low- and intermediate-mass stars, with a focus onstudyingentrainmentatconvectiveboundaries.They amaximummetallicityof∼10−4Z⊙.Greifpresentedan SPH calculation of a Pop III pair-instability supernova pointoutthatonefeatureofstarsatzeroandlowmetal- expanding into a HII region formed by the progenitor licityisthatconvective-reactivemixingeventsarecom- star,andfoundsimilarresults–thesupernovacompletely mon,whichisnottrueinstarsofhighermetallicity.They disrupts the host halo and expands out to a significant show that 1D models have difficulty simulating these fractionofthesizeoftheHIIregion,ultimatelypolluting evolutionary phases and the mixing that comes from them,and show 2D and3D simulationsof thisprocess. ≃2.5×105M⊙ ofgaswithmetals.Nagakurapresented 1D, spherically-symmetric simulations of the evolution Cristallopresentedworkontheevolutionandnucleosyn- ofsupernovaremnantsintheearlyuniverse,payingpar- thesisoflow-massmetal-poorAGBmodelswithcarbon- ticularattentiontothethermalandchemicalevolutionof and nitrogen-enhanced opacities (to address observed theexpandingdenseshellofgas.Heshowedthatathigh carbon and nitrogen-rich metal poor stars), and shows redshift,regardlessofmetallicity,theminimumtemper- that the new opacities can cause significant changes in ature of dense gas in the supernova remnant is limited the chemical and physical evolution of the stars, and bytheCMBtemperature,suggestingthatfragmentation maycausenon-negligiblechangesintheamountofcar- of the shell depends more critically on the density of bon, nitrogen, and s-process elements created. Camp- the ambient medium and the supernova energy than on bell discussed the structural and nucleosynthetic evolu- the metallicity of the gas. Nozawa discussed dust evo- tionofmetal-poorandmetal-freelowandintermediate- lution in Pop III supernova remnants, and showed that massstarsfromZAMStotheendoftheAGBphase,and the transport of dust within the evolving SNR depends showedthatmanyofthesestarsexperienceviolentevo- stronglyonthe size andcompositionofthe dustgrains, lutionary episodes that are not seen at higher metallici- andthatsmalldustgrainsarepreferentiallydestroyedin ties.Theseepisodesmaybecoupledwithstrongmixing, theremnantevolution.Additionally,dustcanbestrongly causingsurfacepollution.Surmanpresentedsimulations segregatedfrom metal-rich gas, possibly explaining the results of nucleosynthesis in outflows from Kerr black abundancepatternsofiron,magnesiumandsiliconinthe reionization at z <6. Additionally, the model predicts lowest metallicity stars (assuming they were formed in thatmorethan80%ofionizingphotonsresponsiblefor theshellsofPopIIIsupernovaremnants). reionization were produced at z ≥ 7 from halos with Whalen presented2D radiation hydrodynamicalsim- M<109M⊙, implyingthat the bulk of the reionization ulationsofthe photoevaporationofminihalosbyneigh- sourcesathighzhavenotyetbeenobserved.Thismodel boringPopulationIIIstars,anddemonstratedthat,when isconsistentwiththehigh-redshiftQSOabsorptionline appropriatephysicsandcoordinategeometriesareused, measurements presented by Fan and with several other feedback from the I-fronts of neighboring halos will sets of observations, including the observed number of largelybepositiveorneutral,andthattheimpingingradi- Lyman-alpha emitters and damped Lyman-alpha sys- ationdrivesprimordialchemistrythatiskeytothehydro- tems. Trac and Shin presented results from large vol- dynamics of the halo, making multifrequency radiation ume, high-resolution dark matter-only simulations that transport necessary. Umemura and Sato confirmed this have been post-processed to model the effects of radia- result with 3D SPH simulations including multigroup tive feedback from Population III and metal-enriched radiation transport, and additionally suggested that the stars, and foundresults that are also in goodagreement shielding due to H created in the ionization front may with Fan’s observations of the Gunn-Peterson trough 2 allow the formation of multiple stars in a single halo. in high-redshift quasar absorption lines, and are also Incontrast,Ahnpresentedworkusinga1DLagrangian consistent with the most recent WMAP electron opti- radiation transport code that suggests that the result of caldepthmeasurements.Ioccopresentedresultsshowing animpingingI-frontwillberoughlyneutral.Bryanand thatPopIIIstarswillproduceadiffusehigh-energyneu- McGreershowedthatHDcoolingcanplayanimportant trinobackground–thestrengthofthisbackground,how- roleinlow-massnon-pre-ionizedhalos,reducingaccre- ever,willfallbelowdetectablethresholdsforallcurrent tionratesontotheprotostellarcloudcore.Johnson(Greif and planned neutrino telescopes even with highly opti- et al.) showed thatprimordialstars formingin relic HII mistic assumptions, ruling out the neutrino background regions would likely be protected from radiative feed- asausefuldiagnostictoolformetal-freestars.Thiswas backfromneighboringhalosduetothelargeamountsof confirmed by Suwa, who also estimated gravitational molecularhydrogenthatformevenatlowdensitiesdue wave emission from the collapse of isolated Population tothehighresidualelectronfraction. IIIstars,andsuggestedthatthegravitationalwaveback- Schneiderpresentedasetofcosmologicalsimulations groundfromtheseobjectscouldbedetectablebyfuture of5−10Mpcboxesthatincludeametallicity-dependent gravitationalwave interferometerssuch as BBO or DE- starformationandfeedbackalgorithmalgorithm.These CIGO. simulationsindicatethat,duetoinefficientmetalenrich- ment, Population III star formation may continue up to z ∼ 2.5 in pockets of primordial gas. The sites where 7. COMPACTOBJECTS ATHIGH Pop III stars form also change over time, moving from REDSHIFT thecentersofhalos(athighredshifts)totheoutskirts(at low redshifts). In addition, she suggested that only 1% 7.1. Gamma-RayBursts and Quasars ofmetal-enrichedgashasbeenpollutedpurelybyPopu- lationIIIsupernovae,makingnucleosyntheticsignatures Rockefellershowedthree-dimensionalsimulationsof hard to identify.Pieri performedsimulationsof compa- thecollapseofamassivePopulationIIIstar,demonstrat- rablevolumesthatalsotakeintoaccountphotoionization ingthatitispossibletocreatearapidly-accretingdiskof suppressionofhalocollapse.Hefindsthatradiativefeed- gas aroundthe central black hole, which can then eject back has a strong impact on mechanical feedback, and significantamountsofthestellarenvelope.Liusedthese reducesmetalenrichmentatessentiallyalloverdensities. simulationsasinitialconditionsfora three-dimensional Ricotti demonstrated that an X-ray backgroundfrom MHD calculation of the collapsing stellar core, show- massivePopulationIIIsupernovae,supernovaremnants, ing that the star’s magnetic energy can drive an outgo- and from accretion onto intermediate-mass black holes ing shock,which leadsto an explosion.In addition,the maybeanimportantsourceofionizationintheIGM,and magneticenergyisprimarilyin“bubbles”ofgasthatare thatthetopologyofthisreionizationwillbe“outside-in,” dominated by magnetic energy, and tend to evolve in a withvoidsbeingionizedfirst.Inaddition,redshiftedX- veryinhomogeneous(andcollimated)fashion.Thishas rays will contribute to HI and HeII reionization in the profoundimplicationsforPopulationIIIstarsasgamma low-density IGM. Ferrara presented a model for reion- rayburstprogenitors,andalsofornucleosynthesis. ization that includes effects from Pop III stars, Pop High-redshiftgammarayburstshaveproventobeex- II stars, and AGN. This model predicts that reioniza- tremelyusefultoolsforprobingtheISMandIGM.Chen tion started at z ∼ 20 by Population III stars and was discussed using spectroscopyof gamma-rayburstafter- 90%completebyz=7,andthatquasarsdominateonly glows to perform direct, detailed studies of the ISM in PopIIIstarswillnotaccreterapidlyenoughtobeSMBH thestar-formingregionsofdistantgalaxies,whichisim- progenitors(thisalsohasimplicationsfortheX-rayion- possiblewithquasars.AlargesampleofGRBsightlines, izationscenariodiscussedbyRicotti) coupledwith moderateresolutionafterglowspectra and Begelman suggested that it may be possible to form imagingofthehostgalaxies,willbekeytoolsforunder- extremelymassive“quasistars”bythedirectcollapseof standingthepropertiesoftheISMinhigh-redshiftgalax- gas in cosmological halos, without a stellar precursor. ies.PenpraseshowedmoreresultsrelatedtoGRBafter- Therequiredprecursorobjectcouldbegeneratedbythe glows,includinghighlydetailedestimatesoftheproper- formationofhaloswithT >104 K,orpossiblyinthe vir tiesofstar-formingregionsindampedLyman-alphasys- aftermath of a large halo merger. These objects would tems,anddemonstratedthe promiseofGRBs asprobes beradiation-dominated,andwouldfeaturetheformation of star-forming regionsin high-redshiftgalaxies. Yone- of a central black hole, which then becomes a source toku showed a strong correlation between the spectral ofenergythatwouldcreatea convectiveenvelope.This peak energy of prompt GRBs and the peak luminosity, object could cool via thermal neutrinos, allowing it to and used this to estimate the possible redshiftsfor hun- accreteattheEddingtonlimitoftheenvelope,ratherthan dredsof BATSE GRBs of unknownredshift.This anal- the blackholeitself. This“quasistar”wouldhavea low ysis predicts that a large number of massive stars were effectivesurfacetemperature(T ∼4000K),wouldhave eff formedintheearlyuniverse,assumingPopIIIstarsform alifetimeontheorderof106years,andcouldpotentially GRBs. beobservedbyJWST. Quasar absorption line studies are complementary Colgate argued that supermassive black holes form toolstotheexaminationofGRBafterglows,inthatthey naturally as a result of cosmological structure forma- are very useful probes of the IGM at high redshift. At tion – high-entropy gas creates ∼105M⊙ stars, which present,measuringtheGunn-PetersontroughinQSOab- collapse due to relativistic instabilities and become the sorption line spectra is one of the primary methods of central point of a massive accretion disk, which can constraining the end of the epoch of reionization. This thenchannelgasintotheblackholeatsuper-Eddington wasshownbyFan,whodemonstratedthattheevolution rates,usingacombinationofself-gravityinstabilitiesand of the ionization state of the IGM experienced a very Rossbyvorticesto efficientlytransportangularmomen- rapid change at z∼6, increasing by an order of mag- tumoutward. nitude in D z ≃ 0.5. This implies that z∼ 6 marks the endofthestageofinhomogeneous,overlappingbubbles of reionizingphotons.Thereis also evidencethatz∼6 8. SUMMARY quasars are quite old, suggesting a great deal of black hole buildup and chemical enrichment at high redshift. The properties of the first generations of stars, as well Inaddition,therearefartoofewquasarsatz≥6tocon- as their effects on later epochs of cosmological struc- tribute significantly to reionization,suggestingthat, un- tureformation,havelongbeenatopicofgreatinterestto less the high-redshift QSO luminosity function is very theastrophysicalcommunity.Progressinunderstanding steep, the dominant source of ionizing photons in the these objectshasbeensignificantand rapidlyaccelerat- high-redshiftuniverseisstellarpopulations. ing in the past few years, and many insights have been obtainedbytheoryandbylarge-scalenumericalcompu- tations.Progresshasalso beenmadeobservationally,in 7.2. Supermassive blackholeprogenitors awidevarietyofsettings,andmanysignificantquestions havebeenraised.Theimpressivearrayofnewfacilities Can Population III stars be supermassive black hole andsurveysthatarecurrentlyrunningorareplannedfor progenitors?Therearemanyexamplesofz∼6quasars, the next few years, together with increasingly powerful whicharebelievedtobeblackholeswithmassesM ≥ simulations, should yield crucial new insights into the BH 106M⊙ sittingatthecentersofmassivegalaxies.Trenti earliestepochsofstarformationintheuniverse. arguedthattheseareextremelyrareobjects(oneper ≃ 0.5Gpc3 at z = 6), and that rare, extremely high red- ACKNOWLEDGMENTS shift (z∼ 40−50) black hole seeds can be QSO progenitors. Alvarez, on the other hand, argued that high- redshiftblack holes form in halos where the progenitor This conference was generously supported by Los starhasexpelledthemajorityofthesurroundinggas,ini- Alamos National Laboratory through the Center for tiallysuppressingaccretion.Aftergascollectsinthehalo Space Science and Exploration, the Institute for Geo- again,therewillbestrongself-regulationofaccretionby physics and Planetary Physics, and the Theoretical radiativefeedback.Asaresult,blackholesformedfrom Division.Inaddition,supportwasprovidedbytheNew Mexico Consortium Institutes for Advanced Study, the