Widespread CO 2 and other non-ice compounds on the anti-jovian and trailing sides of Europa from Galileo/NIMS observations

GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L01202, doi:10.1029/2007gl031748, 2008 Widespread CO 2 and other non-ice compounds on the anti-jovian and trailing sides of Europa from Galileo/NIMS observations Gary B. Hansen 1 and Thomas B. McCord 2 Received 20 August 2007; revised 18 October 2007; accepted 27 November 2007; published 11 January 2008. [1] An observation at wavelengths 0.7 5.3 mm of the anti- Jovian and trailing sides of the moon Europa by the Near Infrared Mapping Spectrometer on the Galileo spacecraft has been reprocessed to reveal new details in the >3.0 mm infrared spectrum. The revealed features include strong absorption bands centered at 4.25 and 4.0 mm, attributed to CO 2 and SO 2, as well as weaker bands, such as near 3.5 mm attributed to H 2 O 2. The calculated band depth distribution of both the CO 2 and SO 2 show that the CO 2 is strongly associated with the (possibly endogenic) dark (typically hydrate) regions on the surface, while the SO 2 is not strongly correlated with the CO 2, but has a similar sparse distribution. The association of CO 2 with the hydrates may indicate a CO 2 -rich ocean that is a potential environment for autotrophic organisms that might thrive near the rock-ocean interface, similar to the earliest life on Earth. Citation: Hansen, G. B., and T. B. McCord (2008), Widespread CO 2 and other non-ice compounds on the anti-jovian and trailing sides of Europa from Galileo/NIMS observations, Geophys. Res. Lett., 35, L01202, doi:10.1029/2007gl031748. subsurface [McCord et al., 1998b, 1999a; Orlando et al., 2005]. [3] The NIMS observed the Jovian system in the infrared spectral range from 0.7 5.2 mm. Absorption bands found in the 4-mm region of the reflectance spectrum of the icy Galilean satellites have been attributed to bond vibrations of simple radicals and compounds such as CO 2,SO 2, SH, CN and CH [McCord et al., 1997, 1998a, Hibbitts et al., 2000, 2003]. These discoveries were made first on Ganymede and Callisto only, because the large density of radiation spikes in the early data from Europa made useful analysis difficult in this spectral region, although hints of these absoprtion bands were seen earlier in the observation discussed here [McCord et al., 1998a]. Towards the end of the prime mission, some distant, low spatial resolution (200 300 km per NIMS pixel) observations of the trailing hemisphere of Europa were made with spike densities comparable to typical Callisto and Ganymede observations. Analysis of these data clearly revealed CO 2,SO 2,[Smythe et al., 1998] and H 2 O 2 [Carlson et al., 1999b] compounds. 1. Introduction [2] Europa is the second outward from Jupiter of the four Galilean satellites. It is similar in size to the Earth s moon, and has a largely crater-free surface, implying a relatively young surface age [Lucchitta and Soderblom, 1982; Zahnle et al., 1998; 2003; Schenk et al., 2004]. The interior is differentiated, with a dense rocky/metallic core comprising all but the outer 80 170 km of radius, which exhibits a density of about 1 g/cm3, and assumed to be primarily ice and/or water [Anderson et al., 1998]. Although the surface is solid (ice), the magnetic signature of Europa during several passes of the Galileo orbiter strongly implied the presence of a conducting liquid subsurface ocean [Kivelson et al., 2000]. Additionally, the icy surface [Lucchitta and Soderblom, 1982] was shown by spectra from the Near Infrared Mapping Spectrometer (NIMS) on Galileo to be partly covered by heavily hydrated non-ice materials such as sulfate salts [McCord et al., 1999a, 2002] and/or sulfuric acid hydrate [Carlson et al., 1999a, 2005]. The hydrated material is concentrated in the regions of darker brownish coloring that occur in some chaos regions and along linea, and may be linked to possible endogenic materials from the 1 Department of Earth and Space Science, University of Washington, Seattle, Washington, USA. 2 Bear Fight Center, Winthrop, Washington, USA. Copyright 2008 by the American Geophysical Union. 0094-8276/08/2007GL031748 2. Observation Details and Calibration [4] The NIMS builds up spectral images by recording a spectrum over 20 mirror positions (Nyquist sampled) and up to 408 wavelengths (Nyquist sampled for all 408). These wavelengths are sensed by 17 discrete detectors, each of which covers a small region of the spectrum. The third dimension of the spectral image is filled out by scanning the instrument field-of-view slowly perpendicular to the mirror motion [Carlson et al., 1992]. Each such scan is called a swath, and many NIMS observations are built up of two or more swaths, generally acquired in a raster mode. [5] The NIMS observations of the icy satellites are now being reexamined and recalibrated using new techniques [Hansen and McCord, 2004; Hansen et al., 2006a, 2006b]. This involves determining the best dark values and radiometric calibrations for each observation. The Europa observations awaited an improved technical process (made possible by the recent startup of funding by the NASA Outer Planets Research Program) to determine the best values for the spectra longer than 3 mm, where radiation spikes outnumber the good data by the ratio of 2:1, or more. The first Europa observation to be processed was from the E6 orbit, which was known to have fewer spikes overall than most other Europa-observing orbits. The first observation in this orbit was named TERINC (Terra Incognito), and consisted of two NIMS swaths (downloaded) covering the central latitudes of Europa between about 140 and 260 degrees western longitude with a phase angle of 36 degrees (see Figure 1). The southern swath was taken in a higher gain state and exhibits some saturation from 1.1 1.45 mm near L01202 1of5

the sub-solar point. This observation contained half (204) of the available wavelengths and one detector (2.40 2.67 mm) was not working by this time (giving a total of 192 wavelengths). [6] The observation was dark-corrected and radiometrically calibrated using the best dark and calibration values and wavelength list (the wavelengths measured varied with time as the grating motion was affected by radiation) [McCord et al., 1999b]. The wavelengths up to 2.4 mm were despiked using our usual procedure (also see Hibbitts et al. [2000] for information on wavelengths and despiking). The standard despiking uses a 3 3 3 NIMS pixel analysis that throws out outliers until a good fit to a 3-D hyper surface can be made; this hyper surface is used to predict the actual value at the center if it is a spike. The parameters to the procedure are a percentage of large outliers to eliminate on the first pass, then three passes to identify outliers with varying criteria. The short wavelength process yielded about 20% spikes. [7] For the wavelengths longer than about 2.75 mm the same despiking was used with much tighter first-pass parameters. This despiking was repeated three times to remove most of the visible spikes. Then a few hundred remaining small spikes were manually removed up to 4.4 mm. Beyond this wavelength the spectra exhibit just a few digital numbers (DN) of signal, and contain no information other than a general shape indicated by spatial averages of 25 100 pixels. The wavelengths beyond 4.4 mm were further processed by averaging and smoothing the spectrum over 4 4 NIMS pixels, making a block average. For each spectrum in the block, values lying outside the block average ±1.5 DN are marked as spikes and replaced with the block average. The final frequency of spikes in this region was 70 80%. Because of this aggressive approach, there is significant averaging over adjacent pixels or wavelengths that must be taken into account when comparing to observations of the other satellites. [8] Figure 1 is an image of the first spectral channel of the despiked observation (0.701 mm), which has been converted to approximate albedo by multiplying by a photometric correction of m 0.25 m 0 0.75, where m is the cosine of the emission zenith angle and m 0 is the cosine of the incidence zenith angle. The resultant values vary from 0.56 to 0.81. This albedo value is a good proxy for identifying dark material and associated hydrate distribution [McCord et al., 1998b, 1999a; Carlson et al., 2005] (low implies rich in hydrate, high implies rich in ice). Figure 1. Approximate albedo image of the first wavelength of the E6 TERINC observation. The albedo ranges from 56% (dark, hydrated regions) to beyond 81% (bright, icy regions). The latitude-longitude grid shown here is used on all subsequent images, but it is labeled only here. The letters A, B, and C indicate the locations from where the spectra shown in Figure 2 are taken. 3. Results [9] Average spectra from the three locations marked A (36 pixels), B (121), and C (36) in Figure 1 are plotted in Figure 2. The shortwave part of the spectrum is shown in part (a) and the dark longwave portion is shown in Figure 2b with error bars. Spectrum A is from an almost pure hydrate region (heavy line) while spectra B and C are from more icerich areas thin solid and dashed lines). The long wave portion (Figure 2b) shows a smooth rise to an inflection near 4.0 mm (possibly SO 2 ) and a band at 4.25 mm(co 2 )for the hydrate spectrum A. The icy spectra have a Fresnel peak near 3.1 mm and a smooth peak at 3.6 and dip at 4.5 mm characteristic of fine-grained water frost. Some ice-rich Figure 2. Example spectra from E6 TERINC. These are small averages taken from the locations marked A, B, and C on Figure 1. The A (hydrate) spectrum is multiplied by 2.2 to help separate it from the other spectra. Shown are (a) short wave part of the spectra and (b) the long wave part on a very expanded vertical scale. The strongest absorptions visible here are indicated by labels specifying the attributed chemical species: the SO 2 bond near 4 mm, and the CO 2 bond near 4.25 mm. A weak band visible near 3.5 mm inthe ice-rich spectra has been attributed to the H 2 O 2 compound, a radiation product from water ice. 2of5

Figure 3. Map of 4.25 mm band depths from 0 to 40% (see legend). The large band-depths seem to be correlated with the dark hydrate surface regions. Weak band-depths are better measured on the right side because the reflected light signal is larger there (near the sub-solar point). spectra (B) exhibit bands centered near 4.0 and 4.25 mm, while others (C) do not. A band near 3.5 mm attributed to H 2 O 2 [Carlson et al., 1999b] can also be seen in the ice rich spectra. [10] Since the 4.25-mm band (CO 2 ) is contained within one NIMS detector and is not affected by detector overlap offsets (like e.g. the 4.0-mm band), it was simple to map its distribution using methods described by Hibbitts et al. [2000]. This uses a normalized band shape taken from the data, and finds least square fits of the data to the band [shape]. If the band depth (continuum minus band center level divided by the continuum) is less than the noise level, it is not recorded. The process is also run for 3 3 and 5 5 spatial averaging of the data to improve the fits for stronger bands and find weaker bands (since the assumed noise is smaller for the 9-point and 25 point averages). The CO 2 band-depth map (Figure 3) shows band depths exceeding 30% in some areas. Compared to the image in Figure 1, it appears that the strong bands are concentrated in the hydrate-rich areas. This is demonstrated clearly in Figure 4, showing the fraction of all pixels of a given albedo that have band depths greater than a given level. Here it can be seen that nearly all the darkest pixels have band-depths greater than 15%, and a quarter of them have band-depths greater then 25%. [11] To map the 4.0-mm band it was necessary to modify the data so that measurements from two detectors near 4.0 mm were as well overlapped as possible in each pixel. This band is very broad, and inspection of the data showed that the wavelength center varied by ± one NIMS spectral channel. So the band measurement was repeated using three profiles, with the best fit being adopted. The map of best-fit band center is completely random, implying that there is no significance to this shift, and it probably arises from the averaging during the despiking process. A similar effect appeared on the 4.25-mm fits, but it was ignored as not significant. The 4.0-mm (SO 2 ) band-depth map (Figure 5) shows band-depths exceeding 30% and a different distribution from the 4.25-mm band. There are several locations where there are strong SO 2 bands and no CO 2 bands and vice versa. Some of the brightest ice-rich regions have a large 4.0-mm band depth and no 4.25-mm band (e.g. near the bottom at 160 and 240 W longitude). 4. Discussion and Conclusions [12] Absorption bands attributed to CO 2 and SO 2 have been found on the anti-jovian and trailing sides of Europa for the first time. Each of them occur with band depths up to 40% in places, implying concentrations comparable to those on Ganymede and Callisto [Hibbitts et al., 2000, 2003]. The detected CO 2 appears to be associated with the darker hydrate materials, as it has also been detected mainly in the non-ice materials on Callisto and Ganymede [Hibbitts et al., 2000, 2003]. Since the hydrate distribution in chaos and along linea may be associated with the interior of Europa [McCord et al., 1998b, 1999a], the source of the CO 2 may be internal, or possibly produced by some radiolytic processes on the surface. Callisto may also have an internal reservoir, since its strong surface CO 2 signatures are maintained while CO 2 is lost through a thin exosphere [Carlson, 1999]. Figure 4. The distribution of different 4.25-mm banddepths as a function of albedo. This demonstrates the correlation of CO 2 band-depth with darker, hydrate-rich regions. 3of5

Figure 5. Map of 4.0 mm band depths from 0 to 40% (see legend). The large band-depths are not as correlated with the dark hydrate surface regions as for the 4.25-mm band. For example, the distributions right of center and at the bottom are very different from those of the 4.25-mm band. Weak band-depths are better measured on the right side because the reflected light signal is larger there (near the sub-solar point). [13] Internal CO 2, if found in the putative ocean may indicate an environment amenable to simple autotrophic organisms that use CO 2 and H 2 to drive an anerobic metabolism [Shock, 1992]. Reaction steps between H2 and CO 2 probably evolved into the first autotrophic pathway at the origin of life on Earth [Russell and Martin, 2004; Martin and Russell, 2007; see also McCollom, 1999]. [14] Acknowledgments. Funding for this research was provided by the NASA Outer Planets Research Program. We thank Bob Pappalardo for many fruitful discussions. We thank Michael Russell for his contributions to the biological discussion. Comments from Robert Carslon and an anonymous reviewer were very helpful in improving the manuscript. References Anderson, J. D., G. Schubert, R. A. Jacobson, E. L. Lau, W. B. Moore, and W. L. Sjogren (1998), Europa s differentiated internal structure: Inferences from four Galileo encounters, Science, 281, 2019 2022, doi:10.1126/science.281.5385.2019. Carlson, R. W. (1999), A tenuous carbon dioxide atmosphere on Jupiter s moon Callisto, Science, 283, 820 821, doi:10.1126/science.283.5403.820. Carlson, R. W., P. R. Weissman, W. D. Smythe, and J. C. Mahoney (1992), Near-infrared mapping spectrometer experiment on Galileo, Space Sci. Rev., 60, 457 502. 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., et al. (1999b), Hydrogen peroxide on Europa, Science, 283, 2062 2064. Carlson, R. W., M. S. Anderson, R. Mehlman, and R. E. Johnson (2005), Distribution of hydrate on Europa: Further evidence for sulfuric acid hydrate, Icarus, 177, 461 471, doi:10.1016/j.icarus.2005.03.026. Hansen, G. B., T. B. McCord, C. A. Hibbitts, and L. W. Kamp (2006a), Progress on the recalibration and rectification of Galileo NIMS spectral images of the Galilean satellites, Eos Trans. AGU, 87(36), Jt. Assem. Suppl., Abstract P41B-05. Hansen, G. B., T. B. McCord, C. A. Hibbitts, and L. W. Kamp (2006b), Recalibration and rectification of additional Galileo NIMS spectral images of the Galilean satellites (abstract), Bull. Am. Astron. Soc., 38, 540. Hansen, G. B., and T. B. McCord (2004), Amorphous and crystalline ice on the Galilean satellites: A balance between thermal and radiolytic processes, J. Geophys. Res., 109, E01012, doi:10.1029/2003je002149. Hibbitts, C. A., T. B. McCord, and G. B. Hansen (2000), The distributions of CO 2 and SO 2 on the surface of Callisto, J. Geophys. Res., 105, 22,541 22,557. Hibbitts, C. A., R. T. Pappalardo, G. B. Hansen, and T. B. McCord (2003), Carbon dioxide on Ganymede, J. Geophys. Res., 108(E5), 5036, doi:10.1029/2002je001956. Kivelson, M. G., K. K. Khurana, C. T. Russell, M. Volwerk, R. J. Walker, and C. Zimmer (2000), Galileo magnetometer measurements: A stronger case for a subsurface ocean at Europa, Science, 289, 1340 1343, doi:10.1126/science.289.5483.1340. Lucchitta, B. K., and L. A. Soderblom (1982), The geology of Europa, in Satellites of Jupiter, edited by D. Morrison, pp. 521 555, Univ. Ariz. Press, Tucson. Martin, W., and M. J. Russell (2007), On the origin of biochemistry at an alkaline hydrothermal vent, Philos. Trans. R. Soc., Ser. B, 362, 1887 1925, doi:10.1098/rstb.2002.1183. McCollom, T. M. (1999), Methanogenesis as a potential source of chemical energy for primary biomass production by autotrophic organisms in hydrothermal systems on Europa, J. Geophys. Res., 104, 30,729 30,742. McCord, T. B., et al. (1997), Organics and other molecules in the surfaces of Callisto and Ganymede, Science, 278, 271 275. McCord, T. B., et al. (1998a), 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., et al. (1998b), Salts on Europa s surface detected by Galileo s near infrared mapping spectrometer, Science, 280, 1242 1246. McCord, T. B., et al. (1999a), Hydrated salt minerals on Europa s surface from the Galileo near-infrared mapping spectrometer investigation, J. Geophys. Res., 104, 11,827 11,852. McCord, T. B., G. B. Hansen, J. H. Shirley, and R. W. Carlson (1999b), Discussion of the 1.04-mm water ice absorption band in the Europa NIMS spectra, J. Geophys. Res., 104, 27,157 27,162. McCord, T. B., G. Teeter, G. B. Hansen, M. T. Sieger, and T. M. Orlando (2002), Brines exposed to Europa surface conditions, J. Geophys. Res., 107(E1), 5004, doi:10.1029/2000je001453. Orlando, T. M., T. B. McCord, and G. A. Grieves (2005), The chemical nature of Europa surface material and the relation to a subsurface ocean, Icarus, 177, 528 533, doi:10.1016/j.icarus.2005.05.009. Russell, M. J., and W. Martin (2004), The rocky roots of the acetyl-coa pathway, Trends Biochem. Sci., 29, 358 363. Schenk, P. M., C. R. Chapman, K. Zahnle, and J. M. Moore (2004), Ages and interiors: The cratering record of the Galilean satellites, in Jupiter: The Planet, Satellites and Magnetosphere, edited by F. Bagenal, T. Dowling, and W. McKinnon, pp. 427 456, Cambridge Univ. Press, Cambridge, U. K. Shock, E. L. (1992), Chemical environments of submarine hydrothermal systems, Origins Life Evol. Biosphere, 22, 67 107. 4of5

Smythe, W. D., R. W. Carlson, A. Ocampo, D. L. Matson, and T. B. McCord (1998), Galileo NIMS measurements of the absorption bands at 4.03 and 4.25 microns in distant observations of Europa, Bull. Am. Astron. Soc., 30, 1448. Zahnle, K., L. Dones, and H. F. Levison (1998), Cratering rates on the Galilean satellites, Icarus, 136, 202 222. Zahnle, K., P. Schenk, H. Levison, and L. Dones (2003), Cratering rates in the outer solar system, Icarus, 163, 263 289, doi:10.1016/s0019-1035(03)00048-4. G. B. Hansen, Department of Earth and Space Science, University of Washington, Seattle, WA 98195, USA. (ghansen@rad.ess.washington.edu) T. B. McCord, Bear Fight Center, Winthrop, WA 98862, USA. (mccordtb@aol.com) 5of5