I suppose
Thomas J's post is a cue to finally get around to restating the definition (corrections/opinions welcome):
For the purposes of this thread, a pea is anything you want to call a pea that you feel like posting here. Artifacts, quasars, stars, whatever.
However, an OIII "Pea" is a more restricted definition:
1. The OIII designation is based on the shape of the spectral chart, and therefore if the target doesn't have one, it
can only be called a candidate.
2. On the spectral chart, the tallest peak must be tagged OIII, and it should be reasonably narrow-based, not
a broad pyramid.
3. The baseline of the chart should be relatively flat other than the OIII peak (or peaks) and possible peaks at OII,
H-alpha, and H-beta. This flatness is relative to the scale of the chart; a tall OIII peak will make the baseline appear flatter.
Ultimately, OIII "Peas" are grouped together because of their similar spectral charts, not the circumstances that caused them, which might vary. OIII "Peas" seem to be found as the cores of spiral galaxies, cores involved in mergers, and Blue Compact Dwarf Galaxies (BCDG). So far, an OIII chart has not been found from an elliptical or lenticular galaxy, or an edge-on spiral; the disc may block the signature spectral peak. There is apparently not a hard dividing line between OIII galaxies and some other types of Active Galactic Nucleus (AGN), and the spectral charts may appear similar.
The current understanding is that the OIII peaks represent an emission of light at a very limited wavelength from doubly-ionized oxygen atoms in an interstellar nebula. These are oxygen atoms missing two electrons, heated by very thin plasma in the vacuum of space, and bombarded by radiation from the core and its densely-packed stars. This creates "forbidden" interactions in which electrons bounce off the ionized oxygen atoms instead of joining them, and was considered impossible in the past, thus the term "forbidden". The light that we see on SDSS represents the atoms shedding that energy, somewhat similar to the way fluorescent tubes glow.
At a redshift of z=0.1 or less, the outer structures of OIII galaxies may be visible. At a greater redshift, OIII "Peas" are generally approximately round because the less-energetic, less-bright parts have faded from view. The color we see on SDSS varies, so it's not one of the criteria, though it may be informative.
The catalog of all OIII galaxies identified from SDSS is
here.
Some sample quasar charts for comparison to OIII "Pea" charts are
here.
And a post about quasars that look like OIII "Peas",
here.
