The radio emission from normal galaxies is dominated by synchrotron
emission from relativistic cosmic ray electrons (CRE) and thermal
free-free emission from ionized gas at 
K. The two
processes are distinguishable mainly by their spectra: synchrotron
emission is generally much brighter at low frequencies (
to 
), while thermal emission has a more
slowly-varying spectrum (
) at higher frequencies and may
become opaque at low frequencies. Synchrotron emission therefore
dominates up to about 20 GHz, while thermal emission is most important
between 
and 
GHz. Disentangling the two in any
individual galaxy requires sensitive imaging at the same resolution
over as wide a range of frequencies as possible. As the separation
depends on the behavior of the total flux density with
frequency, it is essential to retain sensitivity to large-scale
structures at all frequencies. This makes wide-field spectral
imaging a particularly demanding problem in terms of sensitivity and
u-v coverage. The enhanced VLA will be ideal for this work,
combining high sensitivity and scaled-array capability across a huge
frequency range with the ability to image a large range of spatial
scales.
Synchrotron emission from galaxies is most readily studied at
frequencies below a few GHz, since both the flux density and the size
of the primary beam increase as one moves to lower and lower
frequencies. As the loss rate varies as 
, regions with little
ongoing particle re-supply/acceleration have steeper spectra; the
range from 
to 
has been observed in nearby
galaxies. The sites and length scales of this spectral evolution,
both in and out of the disk, tell us about where CRE are accelerated,
and about the diffusion and convection processes that affect CRE
within the disk.
The lowest frequencies sample the lowest-energy CRE which move
farthest from their last acceleration point before their synchrotron
emission is quenched. Radio ``halos'' of low-energy particles (
to ) have been observed kiloparsecs above the
disks of a number of edge-on galaxies at 1.4 GHz. In the few cases
that have also been imaged at 327 MHz, even more extended structures
have been found. The full extents and energy reservoirs of such
cocoons of ancient particles around spiral galaxies are still unknown.
The current 8-antenna system at 74 MHz has neither the sensitivity nor
the spatial sampling to image (faint) galaxy halos properly; here we
make a special plea for outfitting all 27 VLA antennas with 74 MHz
receivers, to study such steep-spectrum sources. With an expected
sensitivity of 5-10 mJy in 12 hrs, one could image NGC 891 at 74 MHz
with 
(
kpc) resolution. Such an image would be as
sensitive as any current higher-frequency one, given only the
electrons we already know about from 1.4 GHz images; since we expect a
much older population as well, the 74 MHz images should show more
emission, even further from the disk. With the A+ configuration one
could achieve 
(
kpc) resolution, offering an
unprecedented view of both the disk-halo interface and of any
structure in the halo itself. By probing the oldest and most extended
cosmic ray populations, 74 MHz data offer the unique prospect of
measuring the total time over which cosmic rays have been produced, as
well as the extent of the galaxy's magnetic field. We also need the
higher resolution available around 1 GHz to study the the detailed
changes in spectra with height above the disk and so to place
direct constraints on models for cosmic ray diffusion/convection, as
well as theories about the origin of chimneys, bubbles, and other
phenomena connecting disk to halo. With the enhanced sensitivity, one
could match individual SNRs and H II regions to individual outflows
above the disk, and track the aging of the electrons as they escape
from the plane.
At 0.3-2.4 GHz, we can study the spectral aging of particles within and somewhat above the disk. The big gains of the enhanced VLA are (1) higher sensitivity, allowing observations of more normal galaxies at higher resolution; and (2) better frequency coverage, allowing more detailed imaging of subtler spectral changes.
The thermal component dominates the synchrotron emission from normal galaxies at 30-40 GHz. This regime is currently unexplored, because no imaging instrument has simultaneously the high sensitivity, high resolution, and large field-of-view to make this work practical. The VLA Development Plan addresses all these difficulties. The particular aspects of interest are
9mm in E configuration would be about 
, roughly like
that of the B configuration at 20cm);
9mm gives a 
field of view).
The diffuse ionized gas with emission measures of a few might be
detectable with long integrations and 10 GHz bandwidths: a
beam filled with the warm ionized medium as seen in the Milky Way
would give a flux density of 0.2 microJy/beam at 8mm.
(This would incidentally give the best possible limits on diffuse
ionized gas in ellipticals, since the deepest optical techniques
rely on having fairly narrow and well-known velocities.) More
reasonable integrations could trace faint star-forming complexes
or regions of high electron density for the first time at radio
wavelengths. Knowledge of the true thermal emission is also
critical to disentangling the thermal and non-thermal emission, as
discussed in the next section.
The primary difficulty in disentangling synchrotron emission with varying spectral index from thermal emission is the limited spectral range now available. The discussion above focussed on improvements at a few frequencies, but the real strength of the enhanced VLA comes when all those frequencies are combined. With so many measurements of the radio spectrum available as a function of position in a galaxy, one can finally hope to image the thermal and non-thermal emission separately, following the spectral aging of the synchrotron component with good spectral index sensitivity over a wide range of spatial scales. The combination of high- and low-resolution observations will allow both the identification of SNRs and H II regions, and their removal from the low-resolution images needed to trace more diffuse structures. This is essential to any detailed understanding of galaxies in general and the interstellar medium in particular. To take a single example: one of the great puzzles found by the Infrared Astronomical Satellite (IRAS) was the FIR-radio correlation, which shows an incredibly tight relation between FIR and radio emission over many orders of magnitude. The origin of this relationship, and the heat sources for the dust which emits the FIR radiation, remains controversial. While the dust must be heated by stars, whether those stars are mostly the very bright ones in H II regions, or the more diffuse population of A and B stars, is still debated. On the radio end, the synchrotron emission can presumably be traced eventually to SNRs, but the details of this are not at all clear. The Infrared Space Observatory (ISO) and such ground-based instruments as SCUBA on JCMT will provide high-resolution probes of the FIR emission; the enhanced VLA will make similarly-detailed images available at radio wavelengths, and more importantly, show whether the FIR-radio correlation is tighter when done with pure thermal/non- thermal emission (either or both might be expected...). Assuming we do understand this correlation, various authors (esp. Helou) have used the details of the relative FIR/radio distribution to constrain cosmic ray diffusion lengths; these models can only be tested and refined through high-resolution, multi-wavelength studies.