Spectroscopic observations of the trailing face of Io at a resolution
of 1.02 cm-1 performed with the FTS on the 3.6-m Canada-France-Hawaii
telescope on Mauna Kea confirm the presence of two weak absorption fea
tures in the 2-mum region. The first feature occurs at 4704.9 +/- 0.2
cm-1 (2.12545 +/- 0.00010 mum); it is about 6.8% deep and 3 cm-1 wide
(FWHM). The second feature is at 5047.1 +/- 1 cm-1 (1.98135 +/- 0.004
mum) with a broad central core; its depth is about 8% and its width ap
proximately 7 cm-1. A laboratory investigation of spectra of solid SO2
with relatively thick samples, as well as of mixtures of SO2 with CO2
and H2S, indicates that the two features at 4705 cm-1 and 5047 cm-1 a
re best explained by the 3nu1 + nu3 and nu1 + 3nu3 modes of solid SO2
around 130 K, respectively. Previous work tentatively proposing CO2 cl
usters as an explanation for the first feature and condensed H2S for t
he second can no longer be supported. In addition, a careful a posteri
ori look at our Io spectrum shows the presence of a band at 3933 +/- 1
cm-1 (2.5426 +/- 0.0007 mum) (depth almost-equal-to 30%, FWHM almost-
equal-to 8 cm-1) due to the 3nu3 band of solid SO2, as predicted by ou
r laboratory experiments. The positions and widths of these bands indi
cate that a temperature gradient may exist between the surface of the
frost and several centimeters below and suggest that some SO2 may be m
ixed at the molecular level with some neutral component. Three differe
nt models are proposed to explain the apparent discrepancy between the
large variability with longitude of the strong 2457 cm-1 (4.07 mum) b
and and the nearly constant depth of the weak 4705 cm-1 (2.1254 mum) b
and. All three models lead to mean grain sizes from a few hundreds of
micrometers to about 1 mm. The first model invokes a large variation i
n grain sizes (factor 2-3) over Io's hemispheres. This model correctly
fits the observed depths of the 2- and 4.07-mum bands but not that of
the 3.78-mum (2645 cm-1) band. The second model assumes a considerabl
e variation in thermal flux with Io's longitude (at 2457 cm-1, up to 2
0% of the reflected continuum flux). This model can reconcile the band
strengths in the whole 2- to 4-mum range with about the same grain si
ze at all longitudes and with moderate variations in frost coverage, b
ut does not explain the albedo variability in the UV range. The third
model assumes an uniform layer of coarse grained SO2 frost (a few hund
red micrometers in size) with a variable longitudinal coverage of a th
in layer of very fine frost grains (less than a few millimeters of mic
rometer-sized grains). This model qualitatively explains the observed
IR and UV variations. Using the optical constants of solid SO2 measure
d in the laboratory, we predict the position and depth (greater-than-o
r-equal-to1%) of about 25 additional absorption features of SO2 frost,
in the 2800-5000 cm-1 (2.0-3.6 mum) range, that should be observed in
future high quality spectra of Io. (C) 1994 Academic Press, Inc.