Condensed O2 on Europa and Callisto
John R. Spencer1 Wendy M. Calvin2
Lowell Observatory 1400 W. Mars Hill Rd. Flagstaff, AZ 86001 (928) 774-3358 Fax: (928) 774-6296 spencer@lowell.edu
Dept. Geological Sciences, MS172 University of Nevada, Reno Reno, NV 89557
Corresponding author: John Spencer Submitted to the Astronomical Journal, April 30 2002.

ABSTRACT

High signal-to-noise spectra of Europa and Callisto show a 0.3% deep 5770 absorption band, due to condensed O2, at the same wavelength as a stronger band previously identified on Ganymede. Excellent longitudinal coverage for Europa shows that unlike Ganymede, where the band is much stronger on the trailing side, Europa shows no significant longitudinal variation in the O2 band strength.

1. INTRODUCTION

In recent years, spectroscopy of the surfaces of the icy Galilean satellites has revealed the presence of several species other than water ice. O2 (Spencer et al. 1995, Calvin et al. 1996) and O3 (Noll et al. 1996) have been identified on the trailing side of Ganymede, as well as SO2 on the trailing side of Europa (Lane et al. 1981, Noll et al. 1995), and the leading side of Callisto (Noll et al. 1997). Galileo NIMS has seen CO2, and possible SH, C-N, and C-H features on Callisto and Ganymede (McCord et al. 1998), hydrated salts or sulfuric acid on Europa (McCord et al. 1999, Carlson et al. 1999a), and H 2O2 on Europa (Carlson et al. 1999b). The presence of condensed O2 on Ganymede is inferred from a pair of weak (<2% deep) absorption bands at 5770 and 6275 , which require the absorption of a photon by two adjacent O2 molecules and so are produced only in high-density condensed O2. The high vapor pressure of condensed O2 at Ganymede surface temperatures suggests that the O2 is trapped in bubbles or crystal defects in a water ice matrix (Calvin et al. 1996, Johnson and Jesser 1997, Johnson 1999): the traps may themselves be produced by charged-particle irradiation. HST observations show that Ganymedes O2 is concentrated at low latitudes (Calvin and Spencer 1997): the warmer temperatures at low latitudes may allow the coagulation and growth of radiation-triggered bubbles in the ice, providing sites for concentration of the O2 (Johnson and Jesser 1997). Laboratory measurements of H2O/O2 ice mixtures have reproduced the O2 absorption bands seen on Ganymede (Vidal et al 1998, Baragiola and Bahr 1998), but in experiments so far the O2 is gradually lost at temperatures above 70 K, leading to the more radical suggestion that the O2 is exposed on the surface in small frost patches with daytime temperatures below 70 K, perhaps due to extremely high albedo, or even in an atmospheric haze (Baragiola and Bahr 1998; Baragiola et al. 1999, Bahr et al. 2001). While it might be thought that Ganymedes intrinsic magnetic field (Kivelson et al. 1996) would protect the surface from bombardment by the low energy particles (ion speed < corotation speed) that could produce the observed strong trailing side concentration of O2 on Ganymede, sufficiently low-energy particles (< 20 keV for protons) may be able to penetrate the field on the trailing side due to E B drift associated with the corotational electric field of Jupiter's magnetosphere (Cooper et al. 2001). Such particles might also be expected to produce a similar abundance of O2 on Europas trailing side, however, but earlier studies (Spencer et al. 1995) did not show an O2 band on Europa.
2. OBSERVATIONS AND DATA REDUCTION
We obtained new CCD spectra of the Galilean satellites with the Ohio State University CCD spectrograph at the Lowell Observatory Perkins 72 telescope in June and November 1997, using similar observational and reduction techniques to those used previously (Spencer et al. 1995), with the difference that in the new observations we used an off-axis autoguider to track a nearby guide star, providing more consistent centering of the satellite in the spectrograph slit. Typical total integration time per satellite per night was about 300 seconds. Observations are logged in Table 1. Various gratings and slit widths were used, giving variable spectral resolution. We concentrated on the 5770 rather than the 6275 band, because the latter is weaker and broader and also overlaps a sharp telluric 6280 CO2 band, and is thus harder to study from the Earths surface. To remove Fraunhofer lines as precisely as possible, we ratioed the icy satellite spectra to Io spectra taken the same night, rather than to a solar-type standard star. We chose Io because its ice-free surface composition made it less likely to have surface O2 than the icy satellites. Ratioing to Io introduced a strong curvature in the spectra due to Ios very different continuum shape: this was removed by dividing by a cubic polynomial fit to the spectrum, excluding the region where sodium emission from Io was prominent in the ratios. We did not correct for wavelength-dependent atmospheric extinction, but the airmass difference between the Io and icy satellite spectra was generally less than 0.04 (Table 1). Ratios of the same satellite at different airmasses show negligible telluric absorptions in the region, and placed an upper limit of 0.05% on the strength of any atmospheric features due to atmospheric extinction resulting from a 0.04 airmass difference. We calibrated wavelengths by comparing Frauhofer line positions to a solar spectrum (AHearn et al. 1983) before ratioing: estimated wavelength uncertainty is 1.

3. RESULTS

The ratio spectra generally contain artifacts larger than the noise at some wavelengths, due to imperfectly canceled Fraunhofer lines, or telluric features. However, the improved observations and analysis reveal a previously unseen weak 5770 O2 band in the Europa/Io and Callisto/Io ratios (Figure 1). Though the individual spectra are noisy, all spectra show a consistent drop in reflectance between 5805 and 5770 , as expected from O2 absorption. The shape of the absorption band on Europa is more apparent when all Europa/Io ratio spectra are averaged (Figure 2). The band seems to have identical shape to that on Ganymede, and is seen with similar strength on the leading and trailing sides. The single Callisto spectrum, also shown on Fig. 2, has unexplained features that may be artifacts centered at 5550. The Callisto O2 band is not much stronger than these unexplained features, but its perfect wavelength match to the Europa and Ganymede features provides good evidence that it is a real feature. The depth of such broad shallow bands, ratioed to Ios complexly curved continuum, is difficult to measure precisely. In Figure 2 the region of the O 2 absorption -3-
is excluded from the cubic fit so that the fit does not decrease the band depth, at the risk of introducing artifacts due to treating the region of interest differently from its surroundings. Band depth is 0.30% for the average Europa spectrum, 0.34% for the leading side, and 0.24% for the trailing side: the leading/trailing difference is probably not significant. Depth is 0.33% in the single Callisto spectrum, which covers Callistos trailing hemisphere. For comparison, the maximum depth of the O2 band on Ganymedes trailing side (Spencer et al. 1995, and Figure 2) is 1.8%, though as this depth was obtained from a Ganymede / Callisto ratio spectrum, the true band depth on Ganymede may be slightly higher if the Callisto spectrum also contained O2. The absorption minimum is at , consistent with the reported previously for Ganymede (Spencer et al. 1995).

4. DISCUSSION

The O2 band was not seen in previous reductions of Europa spectra (e.g, Figure 7 of Spencer et al. 1995), probably because all spectra were ratioed to Callisto rather than Io. Europa and Callisto (at the one longitude observed so far) apparently have very similar O2 absorption band strength, so the band disappears in Europa/Callisto ratios. The similar O2 bands on Europa and Callisto, despite their very different mean surface compositions, might lead to suspicion that the band is an artifact, perhaps due to a feature in Ios spectrum. However, we consider this unlikely. If the feature were on Io, it would have to be a reflectance excess with the same shape and wavelength dependence as the indubitably real Ganymede O2 absorption, but of opposite sign, and this seems highly unlikely. No instrumental artifact should appear only in Europa and Callisto spectra, but not concurrent Io spectra (Io does have a smaller mean distance from Jupiter, and thus more potential for artifacts due to Jupiter light contamination, but on several dates Europa was at a similar or smaller distance to Jupiter than was Io). We thus consider the only plausible explanation of the data, however surprising, to be that a weak O2 band is present at similar strength, 0.3%, at all longitudes on Europa and at least one longitude on Callisto. O2 abundance is difficult to constrain from these observations, as the intrinsic strength of the 5770 band is a very strong function of the density of the O2, which is unknown. On Ganymede we estimated a maximum absorption SDWKOHQJWKRIPLIWKHEDQGZDVDV VWURQJDVWKDWRIVROLG.22 as reported by Landau et al. (1962), (Spencer et al. 1995). In Fig. 2 we use new measurements of the strength of the 5770 band in 22 by Calvin et al. (2002) to match the observed band depths on Callisto, Ganymede, and Europa. The EDQGLVVHYHUDOWLPHVZHDNHULQSKDVH22WKDQLQWKH.SKDVH,ISKDVHEDQGVWUHQJWKV were appropriate for the Galilean satellite O2SDWKOHQJWKVRIURXJKO\PRQ&DOOLVWR PRQ(XURSDDQGPon Ganymedes trailing side would be implied, though it can EHVHHQIURP)LJWKDWSKDVH22 does not match the band shape precisely. Translation from path length to absolute abundance then requires knowledge of the typical visiblewavelength photon path length in the H2O matrix (if the O2 is dispersed in H2O ice), and is thus even more uncertain. The presence of condensed high-density O2 on Europa and Callisto constrains hypotheses for its formation on all the icy satellites, though we leave detailed exploration of these

constraints to future papers. The presence of O2 at all longitudes on Europa in similar amounts, in contrast to Ganymede, suggests that it is not generated by low-energy plasma bombardment, which on Europa, due to the lack of a deflecting magnetic field, strongly favors the trailing hemisphere. The idea that O2 might be formed at all longitudes on Ganymede, but destroyed or buried by micrometeorite bombardment on the leading side (Calvin and Spencer 1997), is also challenging to reconcile with the lack of an obvious leading/trailing asymmetry on Europa, which will have an even greater leading/trailing bombardment asymmetry due to its greater orbital speed. The reduced abundance of O2 on Europa and Callisto compared to Ganymede also requires explanation: possible explanations for Europa might include scavenging of the oxygen by sulfur, which is probably more abundant on Europas surface than on Ganymedes (Carlson et al. 1999a), or surface erosion by charged particle bombardment. For Callisto, the lower surface ice abundance than on Ganymede is likely to be part of the explanation for the lower O2 abundance. The observations confirm the highly oxidizing nature of Europas surface inferred from the earlier detection of H2O2 and possible H2SO4. If these oxidants can be transported from the surface to the interior, they could conceivably provide an energy source for possible Europan organisms (Chyba 2000).

ACKNOWLEDGMENTS

This work was supported by NASA grant NAG5-4262. Thanks are due to Ray Bertram and Mark Wagner for assistance with the OSU spectrograph, and to John Cooper for insights into plasma/satellite interactions.

REFERENCES

AHearn, M. F., J. T. Ohlmacher, and D. G. Schleicher 1983. U. Maryland Tech. Rep. TR AP83-044. Bahr, D. A., M. Fam, R. A. Vidal, and R. A. Baragiola 2001. Radiolysis of water ice in the outer solar system: Sputtering and trapping of radiation products. J. Geophys. Res. 106, 33285-33290. Baragiola, R. and D. A. Bahr 1998. Laboratory studies of the optical properties and stability of oxygen on Ganymede. J. Geophys. Res. 103, 25865-25872. Baragiola, R. A., C. L. Atteberry, D. A. Bahr, and M. Peters 1999. Reply. J. Geophys. Res. 104, 14183-14188. Calvin, W. M., R. E. Johnson, and J. R. Spencer 1996. O2 on Ganymede: Spectral characteristics and plasma formation mechanisms. Geophys. Res. Lett. 23, 673. Calvin, W. M., and J.R. Spencer 1997. Latitudinal Distribution of O2 on Ganymede: Observations with the Hubble Space Telescope. Icarus, 130, 505. Calvin, W. M., Anicich, V. G., and R. H. Brown 2002. Visible and near-infrared transmission spectra of condensed oxygen: Temperature and phase effects. Icarus, submitted.

Carlson, R. W., R. E. Johnson, and M. S. Anderson 1999a. Sulfuric acid on Europa and the radiolytic sulfur cycle. Science 286, 97-99. Carlson, R. W., M. S. Anderson, R. E. Johnson, W. D. Smythe, A. R. Hendrix, C. A. Barth, L. A. Soderblom, G. B. Hansen, T. B. McCord, J. B. Dalton, R. N. Clark, J. H. Shirley, A. C. Ocampo, and D. L. Matson 1999b. Hydrogen peroxide on the surface of Europa. Science 283, 2062-2064. Chyba, C. F. 2000. Energy for microbial life on Europa. Nature 403 381-382. Cooper, J. F., R. E. Johnson, B. H. Mauk, H. B. Garrett, and N. Gehrels 2001. Energetic Ion and Electron Irradiation of the Icy Galilean Satellites. Icarus 149, 133-159. Johnson, R. E., and W. A. Jesser, 1997. O2/O3 microatmospheres in the surface of Ganymede. Astrophys. J. Letters, 480, L79. Johnson, R. E. 1999. Comment on Laboratory studies of the optical properties and stability of oxygen on Ganymede by Raul A. Baragiola and David A. Bahr. J. Geophys. Res. 24, 2631-2634. Kivelson, M. G., K. K. Khurana, C. T. Russell, R. J. Walker, J. Warnecke, F. V. Coroniti, C. Polanskey, D. J. Southwood, and G. Schubert 1996. Discovery of Ganymedes magnetic field by the Galileo spacecraft. Nature 384, 537-541. Landau, A., E. J. Allin, and H. J. Welsh 1962. The absorption spectrum of solid oxygen in the wavelength region from 12,000 to 3300. Spectrochim. Acta 18, 1-19. Lane, A. L., R. M. Nelson, and D. L. Matson 1981. Evidence for sulphur implantation in Europa's UV absorption band. Nature 292, 38. McCord, T. B. and 12 colleagues 1998. Non-water-ice constituents in the surface material of the icy Galilean satellites from the Galileo near-infrared mapping spectrometer investigation. J. Geophys. Res. 103, 8603-8626. McCord, T. B. and 11 colleagues 1999. Hydrated salt minerals on Europa's surface from the Galileo near-infrared mapping spectrometer (NIMS) investigation. J. Geophys. Res. 104, 11827-11852. Noll, K. S., H. A. Weaver, and A. M. Gonnella 1995. The albedo spectrum of Europa from 2200 to 3300. J. Geophys. Res. 100, 119057-19060. Noll, K. S., R. E. Johnson, A. L. Lane, D. Domingue, and H. A. Weaver 1996. Detection of ozone on Ganymede. Science 273, 341. Noll, K. S., R. E. Johnson, M. A. McGrath, and J. J. Caldwell 1997. Detection of SO2 on Callisto with the Hubble Space Telescope. Geophys. Res. Lett. 24, 1139. Spencer, J. R., W. M. Calvin, and M. J. Person 1995. Charge-coupled device spectra of the Galilean satellites: Molecular oxygen on Ganymede. J. Geophys. Res.,100, 19049-19056. Vidal, R. A., D. Bahr, R. A. Baragiola, and M. Peters 1997. Oxygen on Ganymede: Laboratory studies. Science 276, 1839-1842.
TABLE 1: LOG OF OBSERVATIONS
Date 97/06/17 97/06/18 97/06/19 97/06/20 97/06/21 97/06/22 97/06/23 97/11/02 Europa Io Callisto Mean Mean Mean Airmass Airmass Airmass 1.62 1.73 1.56 1.56 1.68 1.56 1.68 1.66 1.61 1.63 1.57 1.56 1.64 1.57 1.64 1.70 1.56 Spectra l Resn., () Europa CML Callisto CML 258
Notes: Spectral resolution is defined as the full-width-half-maximum of an unresolved spectral line. CML = central meridian longitude.

FIGURES

270 Callisto Orbital Longitude

0.3% band depth 90

0 5500
5900 Wavelength, Angstroms
Figure 1 Spectra of Europa (single, unlabeled lines) and Callisto (double, labeled, line), divided by contemporaneous Io spectra to remove Fraunhofer lines, arranged according to the central longitude of Europa or Callisto at the time of the observation. The band depth scale is also shown. Spectra are divided by a cubic fit to correct for the large difference in continuum shape between Io and Ganymede (Spencer et al. 1995): the fit is shown by the horizontal lines. Sodium emission from Io, which appears as a negative feature at 5893 , has been cropped out of the spectra. A weak 5770 absorption band due to O2 is seen in all spectra as a drop in the relative reflectance between 5805 and 5771 (vertical lines).
C. Trailing E. Trailing E. Leading E. Average G. Trailing 0.995

Reflectance / Continuum

0.985 5600 Solid O2 (Landau)

0.980 0.975 5500

Figure 2 Averages of all Europa/Io ratios from Figure 1, and separate averages of Europas leading and trailing hemispheres (labeled E.), showing the weak O2 band. Our single Callisto/Io ratio, from Fig. 1 (labeled C.) and a Ganymede trailing hemisphere spectrum from 1994/04/05, ratioed to Callisto (labeled G.), from Spencer HWDOLVDOVRVKRZQIRUFRPSDULVRQDVDUHWUDQVPLVVLRQVSHFWUDIRUSKDVH22 ZLWKSDWKOHQJWKVRIDQGPIURP&DOYLQHWDOPDWFKHGWR&DOOLVWR Europa and Ganymede respectively. Unlike Fig. 1, the spectra are normalized to a cubic fit which excludes the region of the O2 band, to allow more accurate measurement of the band depth. Finally we show laboratory spectra of all three phases of solid O2, with arbitrarily vertical scaling, from Landau et al. (1962). All spectra are offset vertically for clarity. Vertical lines show the wavelengths of minimum reflectance (5771 ) and return to continuum (5805 ) of the Ganymede O2 feature, for comparison with Europa and Callisto.

2006, 110, 7985-7988 Published on Web 06/13/2006
Infrared Detection of HO2 and HO3 Radicals in Water Ice
Paul D. Cooper,*, Marla H. Moore, and Reggie L. Hudson
NASA/Goddard Space Flight Center, Astrochemistry Branch, Code 691, Greenbelt, Maryland 20771, and Department of Chemistry, Eckerd College, 4200 54th AVenue South, St. Petersburg, Florida 33711 ReceiVed: May 5, 2006; In Final Form: May 26, 2006
Infrared spectroscopy has been used to detect HO2 and HO3 radicals in H2O + O2 ice mixtures irradiated with 0.8 MeV protons. In these experiments, HO2 was formed by the addition of an H atom to O2 and HO3 was formed by a similar addition of H to O3. The band positions observed for HO2 and HO3 in H2O-ice are 1142 and 1259 cm-1, respectively, and these assignments were confirmed with 18O2. HO2 and HO3 were also observed in irradiated H2O + O3 ice mixtures, as well as in irradiated H2O2 ice. The astronomical relevance of these laboratory measurements is discussed.
Introduction Although the hydroperoxy (HO2) radical has long been known to be produced in irradiated water ice,1 its infrared (IR) spectral detection in H2O is challenging. Because HO2 is a strong hydrogen-bonding species,2 with O-H stretching absorption bands likely to occur in the same spectral region as, and be dominated by, strong broad water features, most IR studies of HO2 have been confined to its trapping in inert gas matrixes.3 Furthermore, the reactivity of HO2, coupled with the relative insensitivity of IR spectroscopy, make it difficult to establish an IR-detectable concentration of this radical in solid H2O-ice. Recent work by Ignatov et al.,4 has suggested the formation of HO2 in UV-irradiated H2O + O3 ices in the 3200-3600 cm-1 region, but their results are not conclusive. The problems that plague the condensed-phase study of HO2 also apply to HO3. Recent work has identified HO3 at ambient temperatures in the gas-phase using neutralization-reionization mass spectrometry5 and Fourier transform microwave spectroscopy.6 It was shown that at ambient temperatures HO3 has an appreciable gas-phase lifetime (>10-6 s) and is not merely a momentary complex.5 HO3 also has been detected in the infrared, but only when isolated in an Ar matrix.7 Direct detection in the more-reactive H2O matrix has not been reported but is desirable because HO3 has possible roles in atmospheric chemistry8 and has been implicated as a bactericidal oxidant by Wentworth et al.9 Previous work suggests that HO3, as well as HO2, can form strong hydrogen bonds to water10,11 and that such bonds might actually stabilize these radicals.11 Beyond the areas already mentioned, HO2 and HO3 are also of interest in planetary science, and it is on this area that we focus. Frozen water is the dominant ice on surfaces in the outer solar system (e.g., those on the Galilean and Saturnian icy satellites) as well as on interstellar grains. Such bodies range in temperature from 100 to 10 K, respectively, and exist in a variety of photon-, electron- and ion-radiation environments.
* Corresponding author. E-mail: paul.cooper@ssedmail.gsfc.nasa.gov. NASA/Goddard Space Flight Center. Eckerd College.
IR spectroscopy has played a major role in characterizing the composition of these ices and was the tool used to identify H2O2, a radiation product, on the surface of Europa.12 Small UV and visible signatures of two other radiation products, O2 and O3, also have been identified in spectra of the Galilean and Saturnian satellites.13-16 However, identifying radiation products and unraveling their chemistry remain important challenges for planetary chemists. Experimental Methods Ice samples typically were prepared by mixing the appropriate gas-phase components in a vacuum manifold, using standard manometric techniques, followed by deposition onto a polished, cold aluminum substrate. The substrate, situated in a highvacuum chamber (1 10-7 Torr), could be cooled to 9 K by an APD HC-4 closed-cycle helium refrigerator. The thickness of each ice sample ( 3 m) was determined by laser interferometry. The 0.8 MeV protons used for irradiations were produced from a Van de Graaff accelerator at the Goddard Radiation Facility, and infrared spectra of the ices were measured using a Nicolet Nexus 670 spectrometer at 4 cm-1 spectral resolution. Millipore-purified water was degassed by several freeze-pump-thaw cycles before use. Research-grade 16O (Matheson Tri-Gas) was used as received, as was 18O (Isotec; purity of >97%). Ozone was synthesized by a Teslacoil discharge in a glass bulb containing 100 Torr of O2 and trapped in liquid nitrogen before use. Urea-hydrogen peroxide (Aldrich) was heated to 313 K to produce H2O2 vapor. This was deposited directly onto the precooled substrate via an attachment to the outside of the cryostat, therefore avoiding any decomposition of H2O2 on the metal surfaces of the vacuum manifold. For additional experimental details, see ref 17. Results and Discussion Recent experiments in our laboratory on 0.8 MeV H+irradiated H2O + O2 ice mixtures (H2O:O2 ratios of 6:118) at 9 K show previously unassigned IR features at 1142 and 1259 cm-1 after a dose of 0.6 eV/16 atomic mass units (amu).19 These 2006 American Chemical Society

10.1021/jp062765k CCC: $33.50
7986 J. Phys. Chem. A, Vol. 110, No. 26, 2006
Letters to hydrogen bond interactions with the ice lattice. Previous work has shown that the 3 band of HO2 shifts 19 cm-1 upon complexation of this radical with a single H2O molecule.2 The influence of many H2O molecules, such as an ice matrix, will enhance the shift so that the values in Table 1 are about as expected. Our spectral assignments are also consistent with the initial composition of the ice mixture and the known radiation chemistry of water. Ionizing radiation results in, among other things, decomposition of H2O molecules into H and OH radicals, so that HO2 can form by an H-addition reaction to O2 as follows:

H + O2 f HO2

Alternatively, HO2 might also form through the reaction
Figure 1. Infrared spectra of (a) 6:1 H2O + O2 and (b) 6:1 H2O + 18 O2 ice mixtures irradiated with 0.8 MeV H+ at 9 K to a dose of 0.6 eV/16 amu. The IR spectrum of unirradiated 6:1 H2O + O2 is shown in (c).

O + OH f HO2

TABLE 1: Vibrational Frequencies for HO2 and HO3 Absorption Bands in Water-Ice Compared with Those in an Argon Matrix (2, 6)
3(H16O2) 3(H18O2) isotopic shift 3(H16O2)/ (cm-1) (cm-1) 3(H18O2) (cm-1) in H2O-ice in Ar matrix shift (cm-1) 1142 1101.(H16O in H2O-ice in Ar matrix shift (cm-1) 36 (cm-1)
Here, OH is produced from the dissociation of water molecules and O is produced from the dissociation of O2. However, in H2O + 18O2 experiments reaction 2 would produce an H16O18O molecule and we see no evidence for this. In pure water, HO2 is usually considered to be produced from the abstraction of H from H2O2, formed previously by OHradical dimerization:
1078 1039.(H 3) (cm-1) 30
64 61.6 isotopic shift (cm-1) 39 33
1.059 1.0592 (H16O3)/ (H18O3) 1.032 1.028
OH + OH f H2O2 H2O2 + OH f HO2 + H2O

(3) (4)

absorptions are shown in Figure 1. H2O2 and O3 were also produced and observed as absorptions at 2850 and 1040 cm-1 respectively. The band intensities at 1142 and 1259 cm-1 decreased as the irradiated sample was warmed, and disappeared by 100 K. To help identify these absorptions, the experiment was repeated using H2O + 18O2, whereon the bands shifted to 1078 and 1220 cm-1, respectively. Table 1 summarizes the isotopic shifts for the 1142 and 1259 cm-1 features and compares them to data reported for HO2 and HO3. The 18O isotopic shifts of 39 and 64 cm-1 we observed in H2O-ice compare favorably with those for HO2 and HO3, respectively, in solid argon. Our tables final column shows that the ratios of band positions, 16O/18O, also agree closely with results for HO2 and HO3 in argon. The small discrepancy in the ratio of positions for HO3 in ice and Ar may be due to the different lattices slightly altering the relative contributions of bending and stretching components7 in the observed vibrational mode. In any case, the results summarized in Table 1 strongly suggest assignments of the 1142 and 1259 cm-1 IR bands to HO2 and HO3, respectively. Supporting these spectral assignments are the matrix shifts, H2O-ice to argon, of Table 1. These show a consistent value of 35 ( 6 cm-1 for all four isotopologues, a similarity suggesting that HO2 and HO3 hydrogen bond similarly in H2O-ice. Furthermore, the absolute values of the matrix shifts are reasonable. In the case of HO2, this radical bonds to a single water molecule in a cyclic way, similar to that of the water dimer, so that its H atom acts as a hydrogen bond donor (i.e., OOHOH2) and its terminal O atom acts as a hydrogen bond acceptor (i.e., HOOHOH).2 Consequently, the 3 mode of HO2 (OO stretch) is sensitive to shifts of vibrational frequency due

We do observe H2O2 formation in our ices by the growth of an absorption at 2850 cm-1, but the HO2 band at 1142 cm-1 appears at low doses when the H2O2 feature is still very weak. With the high abundance of O2 molecules in our H2O + O2 ices, an H-addition reaction to O2 seems more likely, particularly because H-addition reactions have been observed in other H+irradiated ices.20-22 Previous work in this laboratory20 found a substantial enhancement of the H2O2 yield in irradiated H2Oice when O2 was added, an enhancement presumably due to the sequential addition of two H atoms to O2 through an HO2 intermediate. The present observations indicate that HO2 is forming in this manner, supporting the earlier proposed mechanism for H2O2 production in H2O + O2 mixtures.20 An assignment of the 1259 cm-1 band to HO3 suggests that this radical might also form by an H-addition reaction, this time to O3 produced radiolytically from the O2 present in the ice mixture.

H + O3 f HO3

To test this possibility, we synthesized ozone and performed radiation experiments with H2O + O3 ices at 9 K. A 1259 cm-1 band was observed after ion irradiation, shifting to 1220 cm-1 when 18O3 was used. Figure 2 shows that in our H2O + O3 experiments the intensity of the 1259 cm-1 band decreased with increasing dose, a behavior consistent with an HO3 assignment. At higher doses, less HO3 was produced because much of the O3 precursor had been destroyed. As expected, HO2 also was observed (at 1142 cm-1, 1078 cm-1 when 18O3 was used) in irradiated H2O + O3 ice mixtures, probably due to H-addition to O2 made from irradiated O3. Consistent with the assignments was the observation that the HO2 band did not decrease in intensity like HO3. This clearly shows that the two bands are from two different species. The HO2 and HO3 bands also were seen after pure H2O2 ices (prepared from the thermal decomposition of the urea-H2O2 complex) were ion-irradiated at 9

Letters

J. Phys. Chem. A, Vol. 110, No. 26, We also note that the present work clears up a minor mystery in the literature. An earlier paper by Gerakines et al.23 listed band positions for HO2 in H2O-ice that are essentially identical to those for HO2 trapped in argon. However, the data in our Table 1 show that HO2 undergoes substantial hydrogen-bonding to H2O, with matrix shifts near 40 cm-1. Thus we believe that the HO2 reported earlier23 was actually HO2 measured in Ar layers that enveloped the ice sample. A similar detection, and an unexpected absence of a matrix effect, was reported for two OH bands,23 features now known to be from site effects induced by an Ar matrix on the OH stretch of an H2OHO complex.24 Conclusions In summary, we have assigned spectral bands at 1142 and 1259 cm-1 in irradiated H2O + O2 ices, at both 9 and 80 K, to HO2 and HO3, respectively. With a re-interpretation of previous work,22 this is the first such detection of either radical in H2Oice by IR spectroscopy. These species, which persist to 100 K, are also produced in irradiated H2O + O3 and H2O2 ices, and are formed from H-addition reactions to O2 and O3. Consequently, we believe that HO2 and HO3 can form on Galilean and Saturnian icy satellites, contributing to the inventory of oxidants. Along with O2, O3, and H2O2, the HO2 and HO3 radicals will have the ability to oxidize other molecules, such as organics, or even provide a source of chemical energy to sustain microbial life.25 Beyond our solar system, HO2 and HO3 probably also form by UV and cosmic-ray bombardment of ice grains in the interstellar medium and in proto-planetary disks. The HO2 and HO3 radicals have been overlooked by astrochemists and planetary scientists but appear to readily form in irradiated water ices containing O2. Thus these radicals should be considered in future theoretical models of ice radiation processes. For observational astronomers, our experimental results confirm theoretical predictions of the stability of HO3 in the presence of H2O molecules,11 and imply that both HO2 and HO3 radicals could exist in extraterrestrial environments. Both species should be sought in astronomical observations of icy bodies. Acknowledgment. This work was supported by NASA through the Planetary Atmospheres and Planetary Geology and Geophysics programs. P.D.C. held a NASA Postdoctoral Fellowship. Claude Smith, Eugene Gerashchenko and Steve Brown of the NASA/Goddard Radiation Facility are thanked for assistance with the proton irradiations. References and Notes

(1) Taub, I. A.; Eiben, K. J. Chem. Phys. 1968, 49, 2499-2513. (2) Nelander, B. J. Phys. Chem. A 1997, 101, 9092-9096. (3) Smith, D. W.; Andrews, L. J. Chem. Phys. 1974, 60, 81-85. (4) Ignatov, S. K.; Sennikov, P. G.; Jacobi, H.-W.; Razuvaev, A. G.; Schrems, O. Phys. Chem. Chem. Phys. 2003, 5, 496-505. (5) Cacace, F.; de Petris, G.; Pepi, F.; Troiani, A. Science 1999, 285, 81-82. (6) Suma, K.; Sumiyoshi, Y.; Endo, Y. Science 2005, 308, 1885-1886. (7) Nelander, B.; Engdahl, A.; Svensson, T. Chem. Phys. Lett. 2000, 332, 403-408. (8) Shroder, D. Angew. Chem., Int. Ed. 2002, 41, 573-574. (9) Wentworth, P., Jr.; Wentworth, A. D.; Zhu, X.; Wilson, I. A.; Janda, K. D.; Eschenmoser, A.; Lerner, R. A. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 1490-1493. (10) Aloisio, S.; Francisco, J. S. J. Phys. Chem. A 1998, 102, 18991902. (11) Aloisio, S.; Francisco, J. S. J. Am. Chem. Soc. 1999, 121, 85928596.
Figure 2. Normalized intensities of the HO3 absorption band produced from (a) 6:1 H2O + O2 and (b) 5:1 H2O + O3 ice mixtures as a function of radiation dose. For comparison, the inset shows normalized intensities vs dose for the 1040 cm-1 O3 absorption band in (c) 6:1 H2O + O2 and (d) 5:1 H2O + O3 ice mixtures.
Figure 3. Infrared spectra of 6:1 H2O + O2 ice mixtures irradiated with 0.8 MeV H+ at (a) 9 K and (b) 80 K to a dose of 0.6 eV/16 amu.
K. Here, O2 and O3 presumably were produced by the radiolytic destruction of H2O2, followed by H-addition to form HO2 and HO3. It is also possible that some HO2 was made through reaction 4. The 9 K low-temperature experiments so far described might not fully represent the radiolytic processes occurring on the surfaces of icy Galilean or Saturnian satellites, which have temperatures in the 65-130 K region. As already stated, the HO2 and HO3 we observed was retained when the irradiated H2O-ice was warmed from 9 to 100 K, but do these radicals form and persist at temperatures typical of the Jovian and Saturnian systems? We repeated our experiments with a 6:1 H2O + O2 mixture deposited at 9 K but subsequently warmed and irradiated at 80 K. Figure 3 compares spectra from our 9 and 80 K experiments and shows that HO2 and HO3 also are produced at the higher temperature, although their band intensities are smaller than at 9 K. At 80 K, hydrogen is lost from the sample more efficiently, as H2, than at 9 K because the H atom mobility is greater. In addition, above 30 K, some O2 is lost from the ice because of sublimation. Evidence for this was seen as an increase in the vacuum chamber pressure at 30 K, and in the disappearance of the 1550 cm-1 O2 absorption above this temperature. The smaller HO2 and HO3 intensities at 80 K also might be caused by destruction processes, such as radicalradical reactions with OH, or even other HO2 and HO3, which are efficiently trapped at 9 K.

7988 J. Phys. Chem. A, Vol. 110, No. 26, 2006
(12) Carlson, R. W.; Anderson, M. S.; Johnson, R. E.; Smythe, W. D.; Hendrix, A. R.; Barth, C. A.; Soderblom, L. A.; Hansen, G. B.; McCord, T. B.; Dalton, J. B.; Clark, R. N.; Shirley, J. H.; Ocampo, A. C.; Matson, D. L. Science 1999, 283, 2062-2064. (13) Spencer, J. R.; Calvin, W. M. Astron. J. 2002, 124, 3400-3403. (14) Spencer, J. R.; Calvin, W. M.; Person, M. J. J. Geophys. Res. 1995, 100, 19049-19056. (15) Noll, K. S.; Roush, T. L.; Cruikshank, D. P.; Johnson, R. E.; Pendleton, Y. J. Nature 1997, 388, 45-47. (16) Noll, K. S.; Johnson, R. E.; Lane, A. L.; Domingue, D. L.; Weaver, H. A. Science 1996, 273, 341-343. (17) Hudson, R. L.; Moore, M. H. Radiat. Phys. Chem. 1995, 45, 779789. (18) A relatively high O2 fraction has been used to enable the detection of the title radicals. However, although the 6:1 ratio exceeds the abundance
of O2 on any known satellite, there is spectroscopic evidence13 that O2 is present at high densities on a small scale, and under such circumstances HO2 and HO3 may form. (19) The radiation dose is standardized to units of eV/16 amu to facilitate meaningful comparison between ice mixtures of varying constituents. Note that 1 eV/16-amu-molecule is 600 Mrad or 6.00 MGy. (20) Moore, M. H.; Hudson, R. L. Icarus 2000, 145, 282-288. (21) Hudson, R. L.; Moore, M. H. Icarus 1999, 140, 451-461. (22) Moore, M. H.; Hudson, R. L. Icarus 1998, 135, 518-527. (23) Gerakines, P. A.; Schutte, W. A.; Ehrenfreund, P. Astron. Astrophys. 1996, 312, 289-???. (24) Cooper, P. D.; Kjaergaard, H. G.; Langford, V. S.; McKinley, A. J.; Quickenden, T. I.; Schofield, D. P. J. Am. Chem. Soc. 2003, 125, 60486049. (25) Chyba, C. F. Nature 2000, 403, 381-382.