For the past several decades, laboratory astrophysicists, notably Mayo Greenberg in Leiden, were conducting experiments to obtain infrared spectra of likely interstellar constituents at low temperatures. We noted their work but largely lacked the ability to surmount the atmosphere to obtain the necessary unobstructed spectra that could have been matched to laboratory data.
With ISO, all this has changed. At this meeting we saw beautifully clear spectra
of interstellar
ices of water vapor H 
O, methane CH 
, formaldehyde H CO, carbon
monoxide
CO, and carbon dioxide CO , both in the prevalent form 
CO and in
its minor
isotopic form 
CO . Spectral shapes showed mixtures of these species
in both
polar and apolar ices -signifying an embedding in a matrix rich, respectively,
in H O or
CO . Pascale Ehrenfreund also pointed out an important missing ingredient.
We have not
observed significant levels of admixed solid O in these ices. This is
puzzling, as I will
note below.
Ices are useful indicators of the temperatures of the star-forming regions in
which they are
observed. CO and O ices sublime at an astrophysically significant rate at
roughly 20 -
25K. This is appreciably lower than the sublimation temperatures of CO or
H O,
that lie, respectively, in the range of 55 - 80K and 100 - 150K. Chris Wright
showed us spectra
of several regions containing young stellar objects. In some, CO appeared fully
frozen; in
others partially frozen with some addition of gas; and finally, in still others,
in fully gaseous
form.
Molecular ices found on interstellar grains are likely to be deposits formed when stray atoms are frozen out on a grain and then combine to form molecules. Low-temperature laboratory samples of molecules ion-irradiated with 30 to 60 keV argon ions show spectra that significantly differ from spectra of the same substances before irradiation. Giovanni Strazzulla suggested that processing material this way in the laboratory may not only better simulate interstellar cosmic ray irradiation of grain material, but also dissociate deposited materials and allow them to recombine in ways that more accurately reflect the chemical processing that takes place on interstellar grains.
To date, laboratory astrophysicists have largely worked with materials more or
less deposited
in bulk on cryogenically cooled surfaces. Yet, we know that the diffuse
interstellar features
originate in grains that might be aggregates of no more than 
atoms.
In recent
years, however, studies of the properties of atoms such as sodium in clusters
numbering about
a hundred atoms, have revealed new physical characteristics that differ from
bulk properties of
the same atoms. Surface effects become important, bond angles change. We may,
therefore,
expect that the spectra of small aggregates of interstellar atoms and molecules
might
substantially differ from bulk spectra obtained in the laboratory. Though
technically difficult,
it should be fruitful to investigate laboratory spectra of small aggregates of
atoms, molecules,
and perhaps admixtures of ions, that are likely to be important interstellar
constituents.
While such considerations are important, there is no doubt that even the bulk
properties of
different minerals catalogued to date already yield striking results. Louis
d'Hendecourt showed
us a laboratory spectrum of the magnesium rich olivine mineral forsterite from
10 to 40
m, and its excellent fit to the dust spectrum of comet Hale-Bopp, again
illustrating the
usefulness of a broad compendium of spectra of widely differing molecular and
mineral
species. As Jacques Crovisier then added, the superposed 2.7 m H O water
vapor
bands yield an ortho-para ratio of 
for this comet,
suggesting equilibration
at a temperature of 25K, corresponding to a distance of order 100 AU from the
sun.