Massive, deeply embedded stars are among the youngest stars that can be observed. These regions are essentially ``mini-star burst galaxies" and their study should add greatly to the understanding of the star-burst phenomenon in luminous galaxies. Recently a significant number of hyper-compact H II regions, i.e. H II regions with extremely high emission measures and extremely small spatial sizes (< a few 1000 AU) have been observed. Some of these hyper-compact H II regions are unresolved or only marginally resolved even in the larger configurations at higher VLA frequencies. These regions may include powerful stellar winds and could be dynamically or magnetically confined Stromgren spheres. Hyper-compact H II regions are quite weak radio sources, but increase in strength at higher frequencies. We would like to spatially resolve these H II regions, as well as to measure the velocities of the ionized gas.
The completion of the 45 GHz system, in conjunction with the new data transmission system and correlator for much higher bandwidth performance will dramatically improve the continuum sensitivity available and allow detailed study of such sources. In addition, a further increase in angular resolution provided by the A+ configuration will allow us to probe both the radio continuum and line emission of hyper-compact H II regions on astrophysically interesting scales. Since the line-to-continuum ratio for radio recombination lines (RRLs) increases with frequency, the observation of RRLs in the 45 GHz band will be one of the best ways of determining the kinematics of the ionized gas. For example, a hyper-compact H II region with a total flux density of 5 mJy at 15 GHz and with spectral index of 1 would have a recombination line detectable in the 45 GHz band at a signal-to-noise ratio of 10:1 with 4 km/s velocity resolution with the expanded VLA in one transit in most cases. Such a line is completely undetectable with the present system at any wavelength with reasonable observing times. With flexible steering of several IFs within the band, observations of a number of different recombination lines could be observed and averaged together to further improve the signal-to-noise ratio.
As hyper-compact H II regions evolve, they should be seen as
ultra-compact and compact H II regions on time scales of order 
to
years. These are complex sources, often displaying
cometary or jet-like extensions and hosting a variety of associated
molecular masers. Most interferometric observations of ultra-compact
H II regions have lacked the sensitivity to any associated
low surface brightness components and large spatial scales to image
any such associated structure well. Images of a number of regions
from multi-configuration VLA data show remarkably complex structures,
but even these observations are limited by u-v coverage on the
shortest and longest baselines. The increased surface brightness
sensitivity afforded by the proposed E configuration, and the
increased angular resolution of the A+ configuration, would aid
immensely in imaging the structure of these sources.
Recent VLA data have revealed a synchrotron emitting source at the
center of expansion of the H
O masers in the massive-star forming
region W3OH. The synchrotron emission appears elongated in the same
direction as the HO masers and thus is almost certainly involved
with their acceleration. Synchrotron emission implies a population of
ultra-relativistic electrons, and strong shocks may be needed to
accelerate these electrons. The nature of the central source powering
the system is unclear: one possibility is a late O-type star with a
fast (1000 km/s) stellar wind. The synchrotron source is weak (
mJy at 5 GHz) and is marginally resolved along its major axis (and
unresolved along its minor axis) at all wavelengths with the VLA in
the A-configuration. Longer baselines (the A+ configuration) are
needed to characterize this source, or any others that may resemble it
in other star-forming regions. Other instruments, such as MERLIN or
the VLBA will lack the combination of sensitivity and angular
resolution needed to study such objects.