At JPL I have become involved in several different projects described below.
While everyone agrees that magnetic fields exist in the interstellar medium, there is far less agreement about the importance of said magnetic fields. If the fields are strong enough, they can regulate star formation by providing support against gravitational collapse. Therefore studying the magnetic fields is important for understanding star formation rates, the lengths of various pre-main-sequence stages, and global cloud stability.Low-Mass Cores and the Outflow Axis
In magnetically regulated models of star formation, the magnetic field axis should align with the outflow axis. The observational signature of this scenario is that the magnetic field should look pinched towards the center of the core. I am collaborating with Giles Novak at Northwestern University to look for this signature in a sample of isolated low-mass cores with well-defined bipolar outflows. We are using SHARP , the Sharc-II Polarimeter installed on the CSO, the Caltech Submillimeter Observatory.
Observing at 350 microns from the CSO is a challenge; only about 1 in 6 nights has usable weather on average. The survey program was initiated in 2006, and I joined in 2008. By the end of 2008, three cores had been successfully observed: L1527, IC348_SMM2, and B335. The first two show clear signatures of a pinch in the magnetic field, aligned with the outflow axis. B335 however shows a toroidal field in the core. This might be explained by the fact that the B335 outflow is many times older and wider than the outflows in L1527 and IC348_SMM2. The first paper from this study, on L1527, is in preparation by a member of the group (Davidson et al). A preliminary figure based on those data is shown below.
Our most recent observing run in September 2009 was more successful. We obtained data for two more cores, though more time is probably needed on the fainter source, L1448-IRS2. Data reduction is proceeding on these sources. As this project continues, I expect to obtain more results.The Magnetic Field in Taurus
I am also interested in the larger-scale magnetic field within clouds. Unlike above, where I am using the polarized emission from the submillimeter, it is also possible to use the polarized absorption of background starlight seen through the cloud. Such studies are typically done in the optical or near-infrared. Historically, the percent polarization has been found to decrease as extinction increases, perhaps due to collisional dealignment of dust grains at higher densities, or grain growth. However, these studies have typically been limited to regions of modest extinction.
I have obtained near-infrared polarization data using Mimir on the 1.8 meter Perkins telescope in Flagstaff, AZ. The region mapped is a little more than half a square degree in B213 and L1495, with another 0.5 sq. degrees in a nearby off field for comparison. The data has not been processed yet (the data pipeline is still being developed). As the extinction increases, I am interested in how the magnetic field changes:
- in direction
- polarization percentage
As a preliminary while the pipeline is being developed, I am investigating techniques for computing the field strength. The classical method is Chandrasekhar-Fermi, where the dispersion in polarization direction gives you an estimate of the plane-of-sky magnetic field strength. Stronger magnetic fields should be more resistant to dispersion due to turbulence.
This method is appealing since it is direct and makes no assumptions about the structure of the cloud. However, it does not account for large scale spatial variations in the magnetic field that may not be due to turbulence. Therefore, blindly applying C-F can lead to underestimates of the field strength. I am experimenting with implementing the Hildebrand et al. (2009) paper which describes a method to account for the spatial and turbulent components of the magnetic field.
The two gray regions are the ones I observed with Mimir. The six blue regions are my attempts to group the polarization vectors into "obvious" related regions. Preliminary results suggest the total cloud strength is about 40 microgauss, higher in region 4 where there is some IR polarization, and lower in regions 5 + 6, which are at low extinction.
High-Mass Star Formation
The theoretical framework for high mass star formation is much less clear than that for low mass star formation. Orion is the nearest massive star forming region, allowing for better (in terms of spatial resolution and sensitivity) study of individual cores compared to further away regions. Li et al. (2007) identified a number of high-mass cores from 350 micron observation. These are likely at the early stages of star formation because of their distance from the trapezium cluster, no known IRAS sources or outflows, and single dish line observations of NH3 and N2H+.Core structure
The Li et al. (2007) survey was made using the CSO and therefore has 9 arcsecond resolution. Furthermore, their mass estimates are derived from the 350 micron data and require an estimate of the dust temperature. Our goal is to model the density and temperature structure of these cores, however, the 350 micron wavelength is not as sensitive to temperature as shorter wavelengths (see the figure below). Therefore, we obtained Spitzer MIPS SED (50-100 micron) spectra for each core.
The MIPS SED data has all been taken and reduced. Five of the detected cores are sufficiently isolated from their neighbors to allow us to extract reliable spectra. We are currently modeling the cores with a code we wrote called COREFIT. This is adapted from an earlier code called DISKFIT that has been used to analyze Spitzer and ground-based data of circumstellar disks (Marsh et al. 2005). COREFIT simultaneously combines all the available continuum and spectral data for each core to parametrically estimates of the temperature and density profile. An example of a typical fit is shown below. Currently we are trying to understand the failure to accurately fit the short wavelength SED data. This may indicate the presence of hot embedded sources in these cores, or perhaps our parametric models need to be improved. Our final goal is to derive improved mass estimates for these cores to see if they are supercritical.
Using COREFIT, we can model our cores, but we would like to have some higher resolution data to compare our models against. We frequently find the short wavelength SED data do not fit our COREFIT models. One possibility is that there are embedded protostars even in these quiescent massive cores. To address this possibility, we have selected two of these cores that may represent the different physical conditions in massive cores. Based on the 350 micron data, ORI2_6 is 12 solar masses while ORI8NW_2 is 46 solar masses. We observed these two cores in N2H+ 1-0 and HCO+ 1-0 using CARMA, the Combined Array for Research in Millimeter-wave Astronomy. With CARMA we achieve 1-2 arcsecond resolution, significantly better than the 9 arcsecond resolution of the 350 micron data.
Of particular note, the N2H+ does NOT peak at the location of the continuum for either core, which is relatively rare. However, for ORI8NW_2 the HCO+ does peak at the position of the 3.2 mm continuum. One possibility is that there is an embedded source in this core, which has heated the CO into the gas phase, which will destroy the N2H+:
N2H+ + CO -> HCO+ + N2
There are significant negative contours in the ORI8NW_2 data, suggesting that we are missing a lot of the extended N2H+ emission. Furthermore, the N2H+ in ORI2_6 also does not peak at the same position as the continuum. It is possible that the same thing that is occurring in ORI8NW_2 is also occurring in ORI2_6. So far we have only obtained B-array data for HCO+ in ORI2_6. We need additional data before we can create an image of the HCO+ emission. We have proposed for additional CARMA time to address both of these shortcomings, and hope to have new results soon.