RUI: Dynamic UC HII regions in Sgr B2: Flickering and Ionized Flows

I. Introduction & Motivation

Massive stars impact their environments in dramatic and fundamental ways: forming deep in molecular cloud cores, depositing energy through molecular outflows, ionizing their surroundings, and injecting vast amounts of energy and material into their environments when they explode as supernovae (Zinnecker & Yorke 2007; Peters et al. 2011). Because of their short formation times (~105 years) and their location in high density regions, the youngest massive objects can be more difficult to observe and to understand than their low mass counterparts. The physical processes involved in the formation of massive stars are actively debated in the literature (Zinnecker & Yorke 2007; McKee & Ostriker 2007; Mac Low & Klessen 2004). Two current models seek to explain the formation process of high mass stars: either they form from the collision of multiple lower mass stars in a stellar cluster (e.g. Bonnell et al. 1998), or they form like low mass stars, by the infall of material through an accretion disk (McKee & Tan 2003). Such disks have been observed around high mass protostars (e.g. Davies et al. 2010), and their presence can solve some outstanding problems in massive star formation. Recent modeling of the accretion process has also given insight into a number of other outstanding issues in massive star formation.

The most massive stars in the Galaxy emit large amounts of ultraviolet radiation that ionize their environments, creating HII regions. The first HII regions studied (“classical” HII regions) were relatively large (D~100 pc). The advent of high-resolution radio interferometers revealed that the ionized gas surrounding many young massive stars is highly confined, and early studies (reviewed in Churchwell 2002) identified what were called ultracompact (UC) HII regions (D~0.1 pc) and hypercompact (HC) HII regions (D~0.01 pc). The earliest models to explain UC and HC HII regions assumed that they were similar to classical HII regions, only smaller. That is, that they were steadily expanding into their environments and that the ionizing star was fully formed (Galván-Madrid et al. 2011). These assumptions led to an apparent difficulty, dubbed the “lifetime problem” (Wood & Churchwell 1989; Kurtz et al. 1994). The problem was that these regions lasted longer than they should if they were simply expanding into their local environments. A number of recent observations have led to a revision of these assumptions. These include observations that (1) hot molecular cores are rotating and infalling (e.g. Keto 1990), (2) resolved small-scale ionized gas and shows accretion dynamics (e.g. Galván-Madrid et al 2009), (3) some UC and HC HII regions have rising spectral indices (e.g. De Pree et al. 2004) and (4) a sample of UC and HC HII regions have measured flux variations on timescales of years (e.g. Galván-Madrid et al. 2008). These observations strongly suggest that the morphology and characteristics of UC and HII regions may be related to the accretion processes that form massive stars.

II. Predictions and Detections of Flickering in Ionized Flows

It is only in the past few years that three-dimensional radiation-hydrodynamic numerical simulations of the formation of HII regions in accretion flows have become possible. These studies show that the dense, rotating, accretion flows required to form massive stars quickly become gravitationally unstable. Recent high-resolution simulations (Peters et al. 2010a) show that when accretion continues in the presence of ionizing radiation, the UC HII region can be gravitationally trapped, and fluctuate over time between trapped and extended states as infalling massive filaments of material interact with the radiation field of the young massive star. As these “flickering” UC HII regions expand and contract, they take on the shapes defined by the morphological classifications of Wood & Churchwell (1989), Kurtz et al (1994) and De Pree et al. (2005). Magnetic fields do not change the morphology of UC HII regions significantly. The resulting HII region flickers between HC and UC sizes throughout the main accretion phase, rather than monotonically expanding. Peters et al. (2010c) show that this behavior also solves the UC HII lifetime problem (Wood & Churchwell 1989), since accretion continues for a period ten times longer than the free expansion timescale for an HII region. The model predicts that UC HII regions can experience scale length and flux variations of 5% per year (Figure 1).

In a related paper, Galván-Madrid et al. (2011) estimate that ~10% of observed UC and HC HII regions should have significant, detectable flux variations (of 10%) on timescales of ~10 years. Indeed, such fluctuations have been convincingly seen in a few sources with multi-epoch VLA observations (e.g. Cep A, Hughes 1988; NGC~7358~IRS1, Franco-Hernandez & Rodriguez 2004; G24.78+0.08, Galván-Madrid et al. 2008). Figure 2 shows one such fluctuation detected by Galván-Madrid et al. (2008). The primary scientific goals of this proposal are to search for size and flux variations in a large sample of UC and HC HII regions, and to examine the properties of the radio recombination line (RRL) emission arising from those flickering regions.

III. The Candidate Source: Sgr B2

The Sgr B2 star forming region, located near the Galactic center is one of the most luminous in the Galaxy, and is associated with a 106 MO giant molecular cloud (GMC). It is an ideal source to search for flickering of UC and HC HII regions, because of the large number of sources with a variety of morphologies located in a single field of view. The region is highly extincted at optical and infrared wavelengths, but has been extensively studied at radio wavelengths (Qiu et al. 2011, De Pree et al. 1998, Gaume et al. 1995). Gaume et al. (1995) published the first high-resolution (qbeam = 0.25”, ~2000 AU) radio images of the Sgr B2 Main, South and North star-forming regions. These original 1.3 cm Very Large Array (VLA) continuum images were followed by H66a (1.3 cm) radio recombination line (RRL) observations at the same resolution, lower resolution (qbeam = 2.5”) H52 a (7 mm) RRL observations (De Pree et al. 1996) and high resolution (qbeam = 0.65”, ~600 AU) 7 mm continuum observations (De Pree et al. 1998). These final 7 mm continuum observations revealed complex morphologies for a number of the sources first imaged at 1.3 cm. Our previous H52a line observations (De Pree et al. 1996) had insufficient spatial resolution and sensitivity to determine RRL parameters for the individual 7 mm continuum sources discussed in De Pree et al. (1998). Higher resolution H52a (7 mm) observations are presented in De Pree et al. (2011). The Sgr B2 Main massive star forming region is shown in Figure 3.

Galactic UC HII regions typically have high frequency recombination line widths of less than 25 km/s (Osterbrock 1989). For example, the thermal width, assuming a constant temperature inside an HII region of 8000 K, is 19.1 km/s (Keto et al. 2008). Jaffe & Martin-Pintado (1999) found in a survey of Galactic UC HII regions that a substantial fraction of the surveyed sources (~30%) had both radio recombination lines that were significantly broader than the typical thermal profiles DVFWHM > 50 km/s), and rising spectral indices (a > 0.4, where Su = ua). They designated such objects broad recombination line objects, or BRLOs. De Pree et al. (2004) found a similar fraction of BRLOs in their 7 mm recombination line and continuum study of W49A. The exact physical process that accounts for the presence of BRLOs is unclear, though the combination of kinematically broadened lines and rising spectral indices is consistent with ionized outflow, perhaps from a circumstellar disk, several examples of which have been detected, e.g. K3-50A (De Pree et al. 1994).

IV. Observations of Sgr B2 with the VLA and EVLA

In 1989, the VLA was used to image the source-rich Sgr B2 massive star forming region in the DnC, CnB and BnA configurations at 1.3 cm in the continuum and the H66á radio recombination line (Gaume et al. 1995; De Pree et al. 1995, 1996). These data were combined to produce high resolution 0.25” (2000~AU) images of the Sgr B2 Main and Sgr B2 North regions, which together contain ~50 individual UC and HC HII regions.

My graduate work (1992-1996) focused on the origin and evolution of ultracompact (UC) HII regions. As a graduate student and soon thereafter as an early career professor, I made some of the first high resolution 7 mm observations of UC HII regions (e.g. Carral et al. 1997, De Pree et al. 1996, De Pree et al. 2004), and worked to identify and catalogue the 100+ sources in the Sgr B2 and W49A galactic star forming regions (De Pree et al. 2005). Even when the 7 mm system was installed on all 27 VLA antennas (by 2000), high frequency work with the VLA was limited by the bandwidth and spectral resolution of the old VLA correlator, and the advent of the EVLA (with its new correlator) has opened up new vistas in observing high frequency radio recombination lines.

We have been awarded 20 hours of EVLA time to re-image this large sample of UC HII regions in Sgr B2 at 1.3 cm with the three hybrid arrays (BnA, CnB and DnC) in the continuum and H66á and H68á lines. At this wavelength, the Sgr B2 region contains 49 detected regions, 25 of which are hypercompact (HC), with physical diameters <5000 AU. The image from these combined data will have a beam of 0.25”. This work presents one of the first attempts to carry out time domain astronomy in a massive star-forming region, and detect the “flickering” of UC HII regions over a 22-year time baseline.

Specifically, these new 1.3 cm EVLA observations of the Sgr B2 will allow us to:

1.  Determine the frequency and magnitude of UC HII flux and size fluctuations over a 22 year time baseline (1989 to 2011) in one of the most source-rich massive star forming regions in the Milky Way.

2.  Constrain the theoretical models described in Peters et al. (2010a, 2010b, 2010c, 2011) and Galván-Madrid et al (2011).

3.  Observe recombination lines with the improved spectral resolution and bandwidth of the new EVLA correlator, and characterize line profiles and velocity gradients, and

4.  Examine the dynamics of sources with especially broad or multiply peaked line profiles discussed in De Pree et al. (2011), and compare the RRL properties of flickering sources with the predictions of Peters et al (2011b). We will systematically look for kinematic signatures of H II region flickering. These signatures could be shocks of ionized gas, line broadening or asymmetric line profiles.

We will also make continuum and RRL observations at 7 mm in only the BnA configuration to obtain morphological information in the continuum and RRLs at a short wavelength recombination line at the highest available angular resolution of 0.06” (650 AU). These observations will allow us to:

1.  Provide recombination line data at additional wavelengths (H52 and H53á) to diagnose the role of pressure broadening in the most compact sources as described in Keto et al. (2008) and De Pree et al. (2011), and

2.  Better resolve the dynamics of the ionized gas – which can result from a combination of rotation of the evaporating accretion flow and outflow – to test the predictions of Peters et al. 2011b

Details: Continuum Observations of Sgr B2

We will use the EVLA to observe Sgr B2 in the 1.3 cm continuum in the DnC, CnB and BnA configurations between January-September 2012. We expect a continuum rms noise of 10.3 microJy/beam (20.46 GHz) and 14.1 microJy/beam (22.36 GHz) and a continuum rms noise of 22-24 microJy/beam (45.45 GHz) and 17-19 microJy/beam (42.95 GHz). While noise characteristics are better away from the H2O line near H66á, the observations are planned in order to match frequencies exactly to the 1989 VLA observations. We will observe the 7 mm continuum in the BnA configuration only. At 1.3 cm, we were awarded 4 hour tracks in each configuration, resulting in 3 hours on source, and one hour for bandpass and phase calibration. For the 7 mm observations, we were awarded 8 hours total in BnA, to be split between Sgr B2 Main and Sgr B2 North, with 4 hours per pointing. The size of the primary beam at Q band will require separate pointings for Sgr B2 Main and Sgr B2 North.

To search for the predicted brightness fluctuations in the sources in Sgr B2 Main and Sgr B2 North, we will make 1.3 cm images using the DnC, CnB, and BnA observations. Since the original 1.3 cm multi-configuration observations were made in 1989, the proposed observations (to be made over a 9 month cycle from January to September 2012) will provide us with a 22 year time baseline over which to search for source "flickering". Based on the predictions of Peters et al. (2010c) and Galván-Madrid et al. (2011), at least two HII regions are expected to have variations larger than 50%. Other sources are likely to show smaller changes, so that ~5-7 of the ~50 sources should have flux variations above 10%.

We will compare the EVLA data with the archival 1989 VLA data to detect any differences in the flux densities and sizes of the sources. The re-expansion of other (previously undetected) sources, which is predicted to occur on ~100 year timescales, could lead to the detection of new HC HII regions. One advantage of looking for ``flickering'' in the Sgr B2 region (as opposed to an isolated source) is that we do not have to depend on the absolute flux calibrations of two observations separated by 22-years. We should be able to easily detect any changes in the relative brightness between sources imaged in the same field, since all of the sources in each field are imaged simultaneously. With the weakest detected sources having source brightnesses of 3-5 mJy/beam, and continuum rms noise measurements of approximately 10 microJy/beam, 10% fluctuations will be easily measurable.