One of the most common methods to measure chemical abundances in the
Universe is the study of the emission of the ionized gas that composes
a number of nebulae related to both star formation (HII regions)
and death (planetary nebulae, supernova remnants, stellar outflows).
However, it is well known that in all these nebulae the computed
abundance values depend on the kind of emission lines
considered. Specifically, optical recombination lines (ORLs) provide
chemical abundance values that are systematically larger than those
obtained using collisionally excited lines (CELs). The abundance
discrepancy factor between ORLs and CELs is usually between 1.5 and 3,
but in planetary nebulae it has a significant tail extending to much
larger values. This is generally known as the "abundance discrepancy
problem". It has been around for more than seventy years, and is one
of the major unresolved problems in nebular astrophysics.
Spectroscopic observations with the William Herschel Telescope of
three planetary nebulae have shed new light on the
problem. Astronomers from the Instituto de Astrofísica de Canarias
have shown that the largest abundance discrepancies (as high as 300 in
certain positions in the nebula, see Figure 1) are reached in planetary
nebulae that have a close binary central star. The spectroscopic
analysis supports the interpretation that two different gas phases
coexist in these nebulae: hot gas at 10,000 K with standard chemical
abundances metallicity where the CELs can be efficiently excited, and
a much cooler (~1000 K) plasma with a highly enhanced content of heavy
elements (which is the cause of the cooling) where only ORLs form.
This dual nature of the stellar ejecta is not predicted by mass loss
theories. How much each gas component contributes to the total mass,
and how they are distributed and mixed, is poorly known.