
The Astrophysical Journal, 639:L103–L106, 2006 March 10
䉷 2006. The American Astronomical Society. All rights reserved. Printed in U.S.A.
A MODEL STUDY OF THE THERMAL EVOLUTION OF ASTROPHYSICAL ICES
M. J. Loeffler, B. D. Teolis, and R. A. Baragiola
Laboratory for Atomic and Surface Physics, University of Virginia, Thornton Hall, Charlottesville, VA 22904-4238;
[email protected], [email protected], [email protected]
Received 2005 October 20; accepted 2006 January 31; published 2006 February 20
ABSTRACT
We address the question of the evolution of ices that have been exposed to radiation from stellar sources and
cosmic rays. We studied in the laboratory the thermal evolution of a model ice sample: a mixture of water,
hydrogen peroxide, dioxygen, and ozone produced by irradiating solid H2O2 with 50 keV H⫹ at 17 K. The
changes in composition and release of volatiles during warming to 200 K were monitored by infrared spectroscopy,
mass spectrometry, and microbalance techniques. We find evidence for voids in the water component from the
infrared bands due to dangling H bonds. The absorption from these bands increases during heating and can be
observed at temperatures as high as ∼155 K. More O2 is stored in the radiolyzed film than can be retained by
codeposition of O2 and H2O. This O2 remains trapped until ∼155 K, where it desorbs in an outburst as water
ice crystallizes. Warming of the ice also drastically decreases the intrinsic absorbance of O2 by annealing defects
in the ice. We also observe loss of O3 in two stages during heating, which correlates with desorption and possibly
chemical reactions with radicals stored in the ice, triggered by the temperature increase.
Subject headings: comets: general — ISM: general — methods: laboratory —
planets and satellites: individual (Jupiter) — radiation mechanisms: general
Online material: color figures
techniques (Loeffler et al. 2006b): amorphous H2O (69.2%),
O2 (22%), H2O2 (!6.2%), and O3 (2.6%). No H2 was observed
(4140 cm⫺1) above the noise level. The values given for H2O
and H2O2 may include up to a few percent of OH, whose
infrared absorption bands are masked by the much stronger
water and hydrogen peroxide bands. Although pure interstellar
H2O2 aggregates are unlikely, the composition of the radiolyzed
ice used here is one of many possible for interstellar ice mantles
and a simple one for a “model” study because it contains just
two atoms: hydrogen and oxygen. We note that water is the
main molecule in most interstellar ice mantles. O2 is likely to
exist, although it has not been detected due to its low infrared
activity (Ehrenfreund et al. 1993). It has been proposed that
solid hydrogen peroxide (Tielens & Hagen 1982; Hasegawa &
Herbst 1993) and ozone (Ehrenfreund et al. 1993) may also
exist on grains in interstellar molecular clouds, but searches
for these molecules have been negative to date. In our study,
the radiolyzed ice mixture is warmed to 200 K while we simultaneously record its mass, its infrared reflectance spectra,
and the desorbed gases. This unusual combination of techniques
allows quantitative characterization during warming.
1. INTRODUCTION
The degree to which interstellar ices become integrated into
comets is still an unsolved problem (Irvine & Lunine 2004).
Icy grains can evaporate during inflow in the protoplanetary
nebula or by solar radiation if they migrate close to a protoSun, and the evaporated gas may later condense into comets
and be concurrently processed by the strong radiation environment (X-rays, stellar particles, UV light, cosmic rays). Alternatively, interstellar grains that have been irradiated in the
interstellar medium may be incorporated into comets with little
thermal processing in the outer regions of a protoplanetary disk.
In both cases, radiation processing will synthesize molecules
and store radicals in the ice to a degree that is not fully understood. Later, when a comet evolves in the solar system, it
warms and loses mass through sublimation of volatile gases
that reveal its composition. Warming may activate chemical
reactions involving frozen radicals, which alter the original ice
composition by forming or destroying molecular species. Related questions of the thermal evolution of radiation-processed
ice also arise when studying the surface properties of icy satellites immersed in the magnetospheres of giant planets.
To study the evolution of cometary ices in the laboratory,
one may start with samples of water ice containing different
gases, as in the seminal experiment by Bar-Nun & Laufer (2003
and references therein). The next step would involve adding
radicals/reactive species that form in the ice during condensation or by irradiation with energetic particles or photons.
Some studies along these lines were made on complex mixtures
of condensed gases by measuring the thermal evolution of either the composition of the radiolyzed ice or the desorbed gases
(Moore et al. 1983; Allamandola et al. 1988; Bahr et al. 2001).
Here we use a combination of measurement techniques to follow the thermal evolution of a processed ice mixture produced
by ion irradiation of pure amorphous H2O2 at 17 K (Loeffler
et al. 2006b).
The sample composition was determined by a combination
of thermal desorption, infrared spectroscopy, and microbalance
2. EXPERIMENTAL SETUP
The experiments were conducted in ultrahigh vacuum (base
pressure 10⫺10 torr), as described previously (Loeffler et al.
2006b). Our samples were produced by 50 keV proton irradiation
of amorphous H2O2 films deposited at 110 K on the gold mirror
electrode of a quartz-crystal microbalance (QCM; Sack & Baragiola 1993). The 910 nm films were thicker than the maximum
range (700 nm) of the 50 keV H⫹ (Ziegler 2003) to prevent the
projectiles from altering the substrate. The relatively high growth
temperature was chosen because our previous experiments
showed that adsorption of impurities such as H2, O2, and CO is
negligible above 100 K. The only impurity observed is CO2 at
∼0.002%, which may have formed in reactions of H2O2 with the
chamber walls. After depositing the hydrogen peroxide ice, it
was cooled to 17 K, where it was irradiated to a total fluence
L103
L104
LOEFFLER, TEOLIS, & BARAGIOLA
Vol. 639
Fig. 2.—Top: Mass loss due to sublimation; 1 mg cm⫺2 p 1.88 # 1016 O2 cm⫺2.
Bottom: (a) Mass loss rate vs. temperature during annealing at 0.2 K minute⫺1.
(b) Mass spectrometer reading at mass 32. The large rise in (b) beginning at
180 K is O2 from H2O2 decomposition (see text). [See the electronic edition of the
Journal for a color version of this figure.]
Fig. 1.—Infrared spectra during heating: (a) 95 K, (h) 178 K; the intermediate values are 121, 141, 142.5, 156.5, 160, and 163.5 K. The spectra are
displaced vertically for clarity. Top: Dangling bonds have been abbreviated
as DB, water as W, and H2O2 as HP. Middle and bottom: Sharp features at
higher temperatures (c–h) are crystalline H2O2. Bottom: X indicates CO2 impurity. [See the electronic edition of the Journal for a color version of this
figure.]
of 4.7 # 10 15 H⫹ cm⫺2 at 50 keV. Subsequently, the film was
warmed at 0.2 K minute⫺1, while we monitored its evolution
with a Thermo Nicolet Nexus 670 Fourier transform infrared
spectrometer at 2 cm⫺1 resolution, a Dycor quadrupole mass
spectrometer (MS) aimed at the sample surface, and the QCM.
The infrared spectra were divided by the reflectance spectrum
from the bare gold mirror substrate. The resulting ratios R(l)
were then converted to optical depth units, ⫺ln R(l) (Fig. 1),
and the band areas were calculated after subtraction of a baseline
that matched the continuum.
3. RESULTS
3.1. Mass Loss Using the Microbalance
and Mass Spectrometer
Thermal desorption spectra measured with the QCM and MS
are shown in Figure 2. The reading of the MS at 32 amu is
due not only to O2 desorbed from the ice but also to O2 formed
by fragmentation of the small amount of desorbed O3 and H2O2
in the MS and in collisions with the chamber walls, as seen in
previous sublimation experiments with H2O2 (Loeffler et al.
2006b) and O3 (Famá et al. 2002). Initially, there is a small
(∼3%) mass loss between ∼35 and 90 K that we attribute to
desorption of O2 and O3, on the basis of past measurements
with water films (Vidal et al. 1997; Baragiola & Bahr 1998;
Baragiola et al. 1999; Bahr et al. 2001). The remaining oxygen
stays in the film, which is mostly amorphous water ice, until
it leaves abruptly at ∼155 K (Fig. 2), a temperature at which
H2O2 desorption is measured to be negligible. This O2 outburst
is related to the crystallization of the water ice and has been
observed earlier with unirradiated water-oxygen mixtures (Vidal et al. 1997; Collings et al. 2004). The amount of oxygen
retained before the outburst is about twice the amount reported
for water-oxygen mixtures just below 130 K (Vidal et al. 1997;
Baragiola & Bahr 1998). A possible reason for this enhancement in oxygen retention is trapping at radiation defect sites
in irradiated water ice, which are not present in unirradiated
water-oxygen mixtures (Teolis et al. 2005). At temperatures
higher than that of the oxygen outburst, three distinct changes
in sublimation rate are observed on the QCM (Fig. 2, top).
These are due to desorption of H2O bound to other H2O molecules, of H2O bound in the H 2 O 2 7 2H 2 O dihydrate compound
(Loeffler & Baragiola 2005), and of pure H2O2, most of which
was not radiolyzed as it was below the ion penetration depth.
3.2. Infrared Spectroscopy
More detailed infrared spectra are shown in Figures 3 and 4.
The integrated absorbance of the O2 band at 1550 cm⫺1 decreased
during heating and became negligible at 140 K (Figs. 3 and 5),
but this was not accompanied by a commensurate release of gas
(Fig. 2). Thus, this decrease is not due to oxygen loss but rather
to a drop in the intrinsic absorbance of O2. We recall that the
infrared absorption of free O2 is dipole-forbidden because of
symmetry, and therefore the observed band arises from perturbing local fields in the solid, such as those near defects (Cairns
& Pimentel 1965). Thus, the decrease in the 1550 cm⫺1 band
can be interpreted to result from the annealing of defects. The
strong decrease of the 1037 cm⫺1 band of O3 between 90 and
120 K (Fig. 5, top), without mass loss, cannot be explained in
the same manner since this absorption is always allowed. We
return to this point below.
No. 2, 2006
THERMAL EVOLUTION OF ASTROPHYSICAL ICES
Fig. 3.—Infrared spectra of O2 and O3 during annealing at (a) 17 K, (f)
140 K. The intermediate temperatures are 42, 61, 92, and 120 K. The spectra
have been shifted vertically for clarity. [See the electronic edition of the Journal
for a color version of this figure.]
Also striking is the presence of three absorption features near
3700 cm⫺1 (Fig. 4, left). Similar features seen in amorphous
solid water and that shift with gas absorption have been assigned to dangling bonds at the internal surfaces of nanopores
(Buch & Devlin 1991). It is remarkable that the dangling bonds
appear in our irradiated starting material since in pure ice they
are destroyed by very low ion irradiation fluences (Raut et al.
2004; Palumbo 2005). The three features at 3706 (DB1), 3676
(DB2), and 3630 cm⫺1 (DB3) are shifted with respect to the
dangling bond absorptions in pure ice (3721 and 3696 cm⫺1).
These shifts are likely due to perturbations by attached adsorbates (Buch & Devlin 1991) such as O2 and O3. Based on the
observation of a 3635 cm⫺1 band in water ice with trapped
ozone (Chaabouni et al. 2000a), we assign DB3 to ozone attached to a dangling bond. The other two absorptions, DB1
and DB2, have not been observed in water-ozone mixtures but
appear in water-oxygen mixtures grown in our laboratory.
Above 60 K, the DB1 and DB2 absorptions increase, while
DB3 decreases. This evolution can be visualized in the plot of
the area of each dangling bond band versus temperature (Fig. 5,
bottom) and of the derivative of the optical depth (Fig. 4, right).
The DB1 and DB2 absorptions, related to O2, increase between
90 and 120 K, become most pronounced between 120 and
140 K, and disappear completely by 156.5 K (Fig. 5), the temperature of the oxygen outburst. The DB3 feature, related to
ozone, decreases steadily until it disappears by 90 K, as also
seen in ozone-water mixtures (Borget et al. 2001; Chaabouni et
al. 2000a).
L105
Fig. 4.—Evolution of the dangling bond bands (shifted vertically for clarity)
for both irradiation (left: normal spectra) and annealing (right: derivative spectra). Curves a and b were taken at irradiation fluences 0.25 and 4.7 # 1015
H⫹ cm⫺2, respectively. After irradiation, the film was warmed continuously at
0.2 K minute⫺1, and spectra were measured at (c) 60 K, (d) 90 K, (e) 104 K,
(f) 140 K, and (g) 152 K. [See the electronic edition of the Journal for a color
version of this figure.]
encounter a trap. In particular, molecules can aggregate in the
voids in increasing numbers (Johnson & Jesser 1997). This
growth decreases the surface-to-volume ratio of the oxygen
aggregates and therefore the perturbation acting on the O2 molecules, explaining the lowering of the infrared absorbance seen
in Figs. 3 and 5. The dangling bonds on the surface of the
voids will be attached to O2. Since water is not mobile below
4. DISCUSSION
To absorb infrared radiation, O2 in the ice must be perturbed
by interactions with defect sites. These sites can include vacancies, voids, and the surfaces of small clusters. Upon warming, individual O2 molecules become mobile (Vidal et al. 1997;
Baragiola & Bahr 1998) and diffuse out of the ice unless they
Fig. 5.—Top: Evolution of the integrals of the O2 and O3 infrared absorption
bands. Bottom: Evolution of the dangling bond features and the mass spectrometer reading. [See the electronic edition of the Journal for a color version
of this figure.]
L106
LOEFFLER, TEOLIS, & BARAGIOLA
∼115 K (Baragiola 2003), the pressure in the voids increases
as more oxygen molecules diffuse into them. We propose that
the resulting stress produces crazes (microcracks) that can then
be filled by O2. These crazes increase the internal surface area
and therefore the strength of the dangling bond absorptions, as
seen in the O2-related bands DB1 and DB2 (Fig. 5). At higher
temperatures, where self-diffusion of water becomes important,
the O2 aggregates can coalesce into large inclusions, thereby
decreasing the internal surface area. This is consistent with the
fall of DB1 and DB2 above 130 K.
The slow decrease of the O3 absorption band below 90 K is
possibly from surface desorption, as suggested by the corresponding mass loss peak at low temperatures (Fig. 2). The
strong decrease of the O3 absorption band at higher temperatures is not accompanied by mass loss, and it is unlikely due
to drastic changes in intrinsic absorbance, as occur in O2, since
the O3 molecule is infrared active. Also, the infrared absorption
of O3 does not change between 90 and 140 K in ozone-water
mixtures that have not been irradiated (Chaabouni et al. 2000b).
Therefore, we suggest that the decay of the ozone band above
90 K is caused by irradiation-produced radicals, such as OH,
which are known to become mobile in water ice at around these
temperatures (Plonka et al. 1984). The following reactions with
OH lead to O3 destruction (and O2 production):
OH ⫹ O 3 r HO 2 ⫹ O 2 ,
(1)
OH ⫹ HO 2 r H 2 O ⫹ O 2 .
(2)
This conversion of O3 into O2 is not detectable in the O2 absorption band because of the low O3/O2 ratio and furthermore
because, at these temperatures, the O2 product is inactive in
the infrared (Fig. 5).
Vol. 639
5. CONCLUSIONS
The use of complementary measurement techniques has allowed us to obtain a detailed and quantitative picture of the
thermal evolution of a model radiolyzed ice mixture. An interesting implication of our results concerns the trapping and
release of gas by comets and the surfaces of icy satellites and
rings. We found that ion irradiation strongly increases the maximum amount of gas that can be trapped in ice just before
crystallization (∼155 K) over the amount achieved in unirradiated ice mixtures (Bar-Nun & Laufer 2003; Vidal et al. 1997;
Baragiola & Bahr 1998 and references therein). This is explained by the generation of radiation-induced defects in the
ice that trap gas. The large amount (∼20%) of O2 that is trapped
at temperatures relevant to the Jupiter system has implications
for the observations of surface oxygen at its icy satellites (Calvin et al. 1996; Spencer & Calvin 2002) and can also occur
on other icy satellites in the outer solar system. On the other
hand, the low intrinsic infrared absorbance of O2 that we find
during thermal processing of the ice suggests a reason for the
lack of detection of O2 condensed on grains in interstellar molecular clouds. Finally, our results provide further evidence that
reactions involving stored radicals can alter the chemical composition of ices during warming, an idea that we advanced in
our report of laboratory simulations of H2O2 synthesis on Europa (Loeffler et al. 2006a). Further studies will test the importance of such processes in condensed gases containing other
astronomically relevant species, such as carbon and nitrogen.
The authors thank W. A. Jesser and R. E. Johnson for insightful comments and discussions. This research was supported by the NASA Cosmochemistry, NASA Planetary Atmospheres, and NSF Astronomy programs. M. J. L. thanks the
Virginia Space Grant Consortium for a fellowship.
REFERENCES
Allamandola, L. J., Sandford, S. A., & Valero, G. J. 1988, Icarus, 76, 225
Bahr, D. A., Famá, M., Vidal, R. A., & Baragiola, R. A. 2001, J. Geophys.
Res., 106, 33285
Baragiola, R. A. 2003, Planet. Space Sci., 51, 953
Baragiola, R. A., Atteberry, C. L., Bahr, D. A., & Jakas, M. M. 1999, Nucl.
Instrum. Methods Phys. Res. B, 157, 233
Baragiola, R. A., & Bahr, D. A. 1998, J. Geophys. Res., 103, 25865
Bar-Nun, A., & Laufer, D. 2003, Icarus, 161, 157
Borget, F., Chiavassa, T., Allouche, A., & Aycard, J. P. 2001, J. Phys. Chem.
B, 105, 449
Buch, V., & Devlin, J. P. 1991, J. Chem. Phys., 94, 4091
Cairns, B. R., & Pimentel, G. C. 1965, J. Chem. Phys., 43, 3432
Calvin, W. M., Johnson, R. E., & Spencer, J. R. 1996, Geophys. Res. Lett.,
23, 673
Chaabouni, H., Schriver-Mazzuoli, L., & Schriver, A. 2000a, Low Temp. Phys.,
26, 712
———. 2000b, J. Phys. Chem. A, 104, 6962
Collings, M. P., Anderson, M. A., Chen, R., Dever, J. W., Viti, S., Williams,
D. A., & McCoustra, M. R. S. 2004, MNRAS, 354, 1133
Ehrenfreund, P., Burgdorf, M., D’Hendecourt, L., & Greenberg, J. M. 1993,
Adv. Space Res., 13, 465
Famá, M., Bahr, D. A., Teolis, B. D., & Baragiola, R. A. 2002, Nucl. Instrum.
Methods Phys. Res. B, 193, 775
Hasegawa, T. I., & Herbst, E. 1993, MNRAS, 263, 589
Irvine, W. M., & Lunine, J. I. 2004, in Comets II, ed. M. C. Festou, H. U.
Keller, & H. A. Weaver (Tucson: Univ. Arizona Press), 25
Johnson, R. E., & Jesser, W. A. 1997, ApJ, 480, L79
Loeffler, M. J., & Baragiola, R. A. 2005, Geophys. Res. Lett., 32, 17202
Loeffler, M. J., Raut, U., Vidal, R. A., Baragiola, R. A., & Carlson, R. W.
2006a, Icarus, 180, 265
Loeffler, M. J., Teolis, B. D., & Baragiola, R. A. 2006b, J. Chem. Phys., in
press
Moore, M. H., Donn, B., Khanna, R., & A’Hearn, M. F. 1983, Icarus, 54, 388
Palumbo, M. E. 2005, J. Phys. Conf. Ser., 6, 211
Plonka, A., Szajdzinska-Pietek, E., & Kroh, J. 1984, Radiat. Phys. Chem., 23,
583
Raut, U., Loeffler, M. J., Vidal, R. A., & Baragiola, R. A. 2004, Lunar Planet.
Sci. Conf., Abstract 1922
Sack, N. J., & Baragiola, R. A. 1993, Phys. Rev. B, 48, 9973
Spencer, J. R., & Calvin, W. M. 2002, AJ, 124, 3400
Teolis, B. D., Vidal, R. A., Shi, J., & Baragiola, R. A. 2005, Phys. Rev. B,
72, 245422
Tielens, A. G. G. M., & Hagen, W. 1982, A&A, 114, 245
Vidal, R. A., Bahr, D., Baragiola, R. A., & Peters, M. 1997, Science, 276,
1839
Ziegler, J. F. 2003, Stopping and Range of Ions in Matter: SRIM-2003,
http://www.srim.org