Broad Radio Recombination Lines from Hypercompact H II Regions

ABSTRACTThe H92 recombination line was observed toward six massive star formation regions (MSFRs), and theH76 line was observed toward one MSFR. All seven MSFRs were suspected of harboring hypercompact (HC)H ii regions. The goal was to detect broad-line sources and to investigate their properties. The sources wereselected according to their small sizes, high brightness temperatures, and rising continuum spectra (typicalspectral index þ 1, S / ) at centimeter wavelengths. Two of the HC Hii candidates, G25.5+0.2 and NGC7538 (IRS 1), were previously known to have extremely broad lines (line widths of 160 and 180 km s 1 ,respectively). Sixteen separate, compact, radio continuum components were detected, fourteen of which weredetected in either the H92 or H76 line. Eight sources have line widths (FWHMs) greater than 40 km s 1 ;typical ultracompact (UC) H ii region line widths are 25–30 km s 1 . These broad lines may be produced by acombination of thermal, turbulent, and electron impact broadening, and large-scale motions (rotation, expansion,jets, shocks, inflow, disk, etc.). On the basis of one line and a relatively low spatial resolution, we are unable todetermine the relative contributions from each mechanism. All the MSFRs in the current sample are composed oftwo or more continuum components. The large projected separations between the continuum components withina given MSFR indicate that they are unlikely to be gravitationally bound massive protostars. Possible origins ofthe observed intermediate-sloped power-law spectral energy distributions (SEDs) are discussed. It is suggestedthat hierarchal clumping in HC H ii regions may produce the observed power-law SEDs.Subject headings: H ii regions — radio lines: ISM — stars: formation1. INTRODUCTIONThe formation and earliest evolution of massive stars is oneof astrophysics’ least understood problems. The observationalrecord is far too incomplete to piece together a coherent theoryof massive star formation and subsequent evolution. Theconventional model assumes that ultracompact (UC) H iiregions represent the earliest manifestation of massive starsafter forming via rapid accretion of ambient gas onto a pro-tostellar hydrostatic core. Norberg & Maeder (2000) andBehrend & Maeder (2001) proposed a ‘‘growing accretionrate’’ scenario, in which the accretion rate increases as themass of the protostar increases. However, the hypothesis offormation via accretion for massive stars (M > 8 10 M )hasbeen questioned by Bonnell et al. (1998), who have proposedan alternative formation mechanism based on mergers oflower mass protostars in dense young clusters. The accretionhypothesis predicts infall of molecular gas to form an equa-torial, Keplerian, accretion disk accompanied by bipolar out-flows along the protostar’s spin axis. Massive bipolar outflowshave been observed toward numerous massive star formationregions (MSFRs; Shepherd & Churchwell 1996; Ridge 2000;Beuther et al. 2002, and references therein). However, thereexist only a few massive protostars with candidate accretiondisks (see Garay & Lizano 1999; Churchwell 2002; Shepherdet al. 2002, and references therein). The small number ofsources is not surprising, because the observations are difficultwith current telescopes, and the phase of rapid accretion islikelytobeshortlived(<10