Sulfate minerals and organic compounds on Mars

Andrew Aubrey  Scripps Institution of Oceanography, University of at , La Jolla, California H. James Cleaves  92093-0212, USA John H. Chalmers  Alison M. Skelley   Department of Chemistry, University of California, Berkeley, California 94720, USA Richard A. Mathies  Frank J. Grunthaner Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, California 91109, USA Pascale Ehrenfreund Leiden Observatory, 2300 RA Leiden, Netherlands Jeffrey L. Bada* Scripps Institution of Oceanography, University of California at San Diego, La Jolla, California 92093-0212, USA

ABSTRACT as 3±5 Ma (Remeika and Lindsay, 1992). Strong evidence for evaporitic sulfate minerals such as gypsum and jarosite has recently Gypsum from the Haughton impact crater is been found on Mars. Although organic molecules are often codeposited with terrestrial assumed to date from the time of the impact evaporitic minerals, there have been no systematic investigations of organic components crater, 23 Ma (Parnell et al., 2004). The age in sulfate minerals. We report here the detection of organic material, including amino of the host rock of the Panoche Valley sam- acids and their amine degradation products, in ancient terrestrial sulfate minerals. Amino ples is 75±65 Ma (Presser and Ohlendorf, acids and amines appear to be preserved for geologically long periods in sulfate mineral 1987), but the age of the sulfate minerals is matrices. This suggests that sulfate minerals should be prime targets in the search for estimated as 40 Ma (middle Tertiary); this is organic compounds, including those of biological origin, on Mars. when the coastal ranges were raised in this area during the Sierra Nevada uplift, which Keywords: Mars, sulfates, evaporites, amino acids, gypsum, jarosite, anhydrite, kerogen. caused ocean water to withdraw and deposit evaporitic minerals in California's Central INTRODUCTION not conclusively disprove that there are organ- Valley. Strontium isotope analyses were con- The search for evidence of water and or- ic compounds present on the surface of Mars. ducted to verify the geologically deduced ages ganic compounds, including those of possible The only other opportunity to analyze samples of the Panoche Valley samples. The Panoche biological origin, is one of the major goals of from Mars has been provided by meteorites Valley gypsum 87Sr/86Sr ratio was 0.707745 the Mars exploration programs of both the Na- ejected from its surface and delivered to Earth. (Ϯ0.000005). Comparing this ratio to the tional Aeronautics and Space Administration However, contamination of these meteorites strontium isotope history of seawater (Hess et (NASA) and the European Space Agency by terrestrial organic material during their res- al., 1986) gives an age 40 Ma, consistent with (ESA). The NASA Mars Exploration Rovers idence times on Earth (102±104 yr) compro- the inferred geologic age. The Panoche Valley and the ESA OMEGA/Mars Express have mises their use in assessing whether organic anhydrite was estimated to be roughly the provided the best evidence to date that liquid compounds are present on Mars (Jull et al., same age because anhydrite forms by dehy- water was once present on Mars. Abundant 1998). dration of gypsum. The jarosite from Panoche sulfate minerals such as gypsum and jarosite Organic matter is often codeposited in ter- Valley was likely formed by aqueous alter- suggest that large acidic water basins were restrial evaporites, and similar deposition pro- ation of pyrite, and it is thus dif®cult to de- once present and that as they evaporated sul- cesses should occur on Mars if organic mol- termine its actual age of the alteration veins, fate minerals were precipitated (Squyres et al., ecules were present in the early oceans but it is probably somewhat younger than 40 2004; Langevin et al., 2005; Gendrin et al., (Mancinelli et al., 2004). To our knowledge, Ma. The modern gypsum sample is from the 2005). Although it is unknown how long these there have been no systemic investigations of South Bay salt works in Chula Vista. The area bodies of water existed, they could potentially organic compounds in sulfate minerals on is rich in marshes and tidal ¯ats, and there is have provided an environment capable of sup- Earth. We report here the determination of the continuous evaporite formation during tidal porting life. organic carbon and nitrogen contents of sev- ¯uctuations. Because of the poor water quality The presence of organic compounds on eral sulfate minerals, as well as the abundance in this region of , the salts typ- Mars is uncertain. The Viking missions in of amino acids and their degradation products. ically include signi®cant amounts of organic 1976 detected no organic compounds above a material. threshold level of a few parts per billion in METHODS The surface of each sample was thoroughly near-surface Martian soils (Biemann et al., The samples investigated included gypsum rinsed with doubly distilled water (ddH2O) 1976). However, key biomolecules such as from the Anza-Borrego Desert, California (4 followed by 1M ddHCl, then again with amino acids would not have been detected by Ma), gypsum from the Haughton impact cra- ddH2O. The identity of each mineral was ver- the Viking Gas chromatograph and mass spec- ter, Canada (23 Ma), and gypsum, anhydrite, i®ed by X-ray diffraction (XRD) analyses us- trometer (GCMS) even if several million bac- and jarosite samples from Panoche Valley ing a Scintag XDS-2000 powder diffractom- terial cells per gram were present (Glavin et (California) (40 Ma). A modern gypsum sam- eter. Samples were analyzed for total organic al., 2001). In addition, oxidation reactions in- ple from a salt evaporation pond in Chula Vis- carbon and nitrogen using a Costech elemental volving organic compounds on the Martian ta (California) was also analyzed. Sample ages combustion C-N analyzer. Carbon and nitro- surface would likely produce nonvolatile were estimated based on the geology of their gen isotopic ratios were determined with a products such as mellitic acid salts that also respective localities. Thermo®nnigan Delta-XP Plus stable isotope would not have been detected by Viking (Ben- The evaporite formations from the Anza- ratio mass spectrometer. In order to remove ner et al., 2000). Thus, the Viking results did Borrego Desert have been studied extensively. carbonate from the samples, they were pre- The gypsum investigated here was collected treated with an excess of 3N ddHCl and dried *E-mail: [email protected]. from the Fish Creek area and has been dated down on a vacuum centrifuge at 45 ЊC for 1

᭧ 2006 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. Geology; May 2006; v. 34; no. 5; p. 357±360; doi: 10.1130/G22316.1; 1 ®gure; 3 tables. 357 TABLE 1. TOTAL ORGANIC CARBON (TOC), TOTAL ORGANIC NITROGEN (TON), AND rosite was showed a much higher value of ISOTOPIC ANALYSES 1.28 mg´C/g. The nitrogen trends were similar Location (Ma) TOC TON ␦13C ␦15NAAϩAmines AAϩAmines in that the gypsum and anhydrite samples

(mg/g) (mg/g) (½) (½) TOC TON ranged from 0.01 to 0.03 mg´N/g. The jarosite (%) (%) was signi®cantly more nitrogen rich with a South Bay Gypsum (0) 6.91 1.01 Ϫ17.3 ϩ11.0 0.041 0.117 content of 0.15 mg´N/g. Anza-Borrego Gypsum (4)* 0.29 0.02 Ϫ34.9 ϩ1.7 0.042 0.265 The organic C/N ratios in old gypsum and Haughton Crater Gypsum (23)² 0.77 0.03 Ϫ31.3 ϩ0.1 0.007 0.105 jarosite range from 9 to 30, indicating that the Panoche Gypsum (40)§ 0.12 0.01 Ϫ30.0 ϩ13.1 0.110 0.399 Panoche Anhydrite (40)§ 0.09 0.02 Ϫ29.0 ϩ4.2 0.359 0.718 major organic component present in the sul- Panoche Jarosite (ϳ40)# 1.28 0.15 Ϫ26.2 ϩ3.9 0.012 0.038 fate minerals is likely a humic acid- and/or Note: The last two columns represent the mass percent of the TOC and TON accounted for by amino acids kerogen-like material (Ertel and Hedges, and amines. The uncertainties are roughly Ϯ5% for TOC, Ϯ10% for TON and Ϯ 0.5±1.0½ for the isotopic values. 1983), a conclusion that is consistent with the *33Њ00ЈN, 116Њ10ЈW; (Remeika and Lindsay, 1992) measured carbon and nitrogen isotopic values. ²75Њ22ЈN, 89Њ41ЈW; (Cockell and Lee, 2002). §36Њ35ЈN, 120Њ42ЈW; (Presser and Ohlendorf, 1987). The exceptions are the Chula Vista gypsum #36Њ27ЈN, 120Њ39ЈW. with a C/N ratio of ϳ7 and the Panoche an- hydrite with a ratio of ϳ5. These numbers are more indicative of recent biological material. h before analyses for total organic carbon and Extracts were analyzed for amino acids and Even though amino acids and amines consti- total organic nitrogen. amines by reverse-phase high performance tute only a fraction of the total organic carbon Amino acids were isolated by vapor-phase liquid chromatography (RP-HPLC) using pre- and nitrogen present in the sulfate minerals acid hydrolysis (6 N HCl, 24 h, 100 ЊC) of column derivatization with o-phthalaldehyde/ (Table 1), they are readily detected and char- ground samples followed by desalting (Ame- N-acetyl L-cysteine using a Shimadzu RF-530 acterized (Fig. 1; Table 2). lung and Zhang, 2001). Amines were isolated ¯uorescence detector (Zhao and Bada, 1995) The detected levels of amino acids and their by microdiffusion from the powdered mineral and a Phenomenex Synergi Hydro-RP column enantiomers, as well as methylamine (MA) treated with 1N NaOH, into a 0.01N HCl so- (250 ϫ 4.6 mm). Quanti®cation of amino ac- and ethylamine (EA), the decarboxylation lution at 40 ЊC for 6 days (Conway, 1963). ids and amines included background level cor- products of glycine and alanine, on average rection using a serpentine procedural blank account for ϳ0.1% of the total organic carbon and a comparison of the peak areas with those and ϳ0.27% of the total organic nitrogen for of an amino acid standard. A D/L-norleucine the 6 samples (the Panoche gypsum showed internal standard was added to normalize ami- an unknown peak that eluted near valine and no acid recoveries from desalting and deriva- was not included in the average). tization. The recovery of the amines carried Adenine was detected in every sample ex- through the extraction procedure was found to cept the Anza-Borrego gypsum. The adenine be near 100% using spiked procedural levels detected (2±6 ppb) indicate that the E. samples. coli equivalents per gram of sample (ECE/g) To investigate the possible presence of are in the range of 106±107 cells/g. With the modern bacterial contamination in the various exception of the Anza-Borrego gypsum, minerals, we determined total adenine concen- which is the most pristine, the various samples trations. Both a liquid extraction involving all have low cell counts, indicating that some treatment of 1 g of sample with 2 mL of 95% of the organic matter is likely associated with formic acid solution for 24 h at 100 ЊC and a bacterial remains. sublimation extraction method at 500 ЊC for 5 min were performed. Adenine concentrations DISCUSSION were quanti®ed by HPLC with ultraviolet The organic matter detected in each mineral (UV) absorption detection (260 nm) and con- is likely ancient organic matter trapped within Figure 1. Combined reverse-phase high per- verted to bacterial cell densities (E. coli equiv- the matrix, along with material derived from formance liquid chromatography chromato- alents/g) as described in Glavin et al. (2004). the remnants of more recent sulfate-reducing grams of recovery-corrected amino acids (hydrolyzed/desalted, 5±25 min) and amines microbial communities. The presence of the (microdiffusion, 27±35 min) in sulfate min- RESULTS D-enantiomers (produced by racemization) of erals. Chromatograms represent elutions The XRD results veri®ed each mineral several amino acids in the gypsum samples with identical gradients. Chromatograms on identity. The gypsum samples that were ob- suggests that these compounds are mostly left and the standard have later retention tained from Anza-Borrego, Panoche Valley, original components of the depositional envi- times because they were separated with a and Haughton crater were selenite, pure gyp- different buffer than samples on right. ronment and not recent contaminants; how- Detection limits were ~1 ppb for amino sum in discrete layers. The Chula Vista gyp- ever, their presence could partially be due to D ؉ L)- sum sample was the least pure. bacterial cell wall material. The correlation) ؍ acids and ~0.5 ppb for amines. 1 glu- The organic carbon and nitrogen data are between the ratios of the amino acids glycine ؍ D ؉ L)-serine, 3) ؍ aspartic acid, 2 -D ؉ L)-alanine, tabulated in Table 1. The organic carbon con) ؍ glycine, 5 ؍ tamic acid, 4 ,and alanine and their degradation products ؍ ؍ ؍ 6 L-valine, 7 methylamine, 8 ethyla- tent range was 0.12±0.77 mg´C/g in the 3 gyp- residual ammonia from ex- MA and EA, respectively, also suggests that ؍ mine, asterisk resin impurity. Under- sum samples; the contemporary gypsum from the organic material is a component of the ؍ traction process, x lined numbers indicate resolution of L- and South Bay showed signi®cantly higher percent original evaporite. The ratio of degradation D-enantiomers for that respective amino organic carbon (6.9 mg´C/g) due to the sam- amines to the amino acids increases with the acid. Peaks labeled with question mark are ple's origin in a highly contaminated region. age of the sample. Amines are not typically currently unidenti®ed. Y-axis scales are list- ed on left for each separation. Right axis The Panoche anhydrite was consistent with detected in ancient terrestrial carbonate min- has been labeled accordingly if amine data the gypsum samples with an organic carbon erals (Glavin and Bada, 1998), presumably be- are at different attenuation. content of 0.09 mg´C/g and the Panoche ja- cause they are volatile and lost from the basic

358 GEOLOGY, May 2006 TABLE 2. AMINO ACID CONCENTRATIONS OF VARIOUS SULFATE MINERALS

Location (Ma) Asp Ser Glu Gly Ala Val MA EA Z* South Bay Gypsum (0) 77.7² 1495² 176 1591 3172² 731 37.2 173 0.04 Anza-Borrego Gypsum (4) 137² 8.3² 116 10.3 30.0² Trace 10.2 8.8 0.5 Haughton Crater Gypsum (23) 21.7 65.9 13.5 N.D. 6.7² Trace 17.7 18.7 5.4 Panoche Valley Gypsum (40) 5.6² 3.3² 234 N.D. 2.9² ?? 40.3 21.2 21.2 Panoche Valley Anhydrite (40) 55.5² 587 93.4² N.D. 58.9² 46.2 72.4 7.9 1.4 Panoche Valley Jarosite (ϳ40) 62.8 50.3² 39.0 28.1 120² 55.8 11.9 10.5 0.15 Note: All values are blank-corrected and reported in mass ppb. Uncertainties in the measurements are Ϯ10%. MA ϩ EA *z ϭ gly ϩ ala ²D-enantiomer detected. N.D.ÐNot detected above blank level. ??ÐValine is not possible to evaluate because of interference from an unknown component.

ϭ ϩ mineral matrices. However, they are apparent- where AAt glycine alanine concentration say, 1992) although average temperatures over ϩ ly retained in these minerals, perhaps as their at time t, AA0 is the original glycine alanine the past 5 m.y. have likely been somewhat nonvolatile sulfate salts. concentration in the sample, and kDC is the cooler, especially during Pleistocene ice ages. The amino acids should be racemic (D/L ϭ rate of decarboxylation of glycine and alanine. The modern Panoche Valley average temper- 1) in the ancient samples because they have Assuming amines are retained in gypsum and ature is ϳ17 ЊC, but over the past 40 m.y. ages in excess of several million years (Bada that the major sources of MA and EA are gly- depositional history of the region, the average et al., 1999), but this was not the case in all cine and alanine decarboxylation, respectively, exposure temperature was likely higher (Park of the samples. The Anza-Borrego gypsum is then: and Downing, 2001). the only sample in which the D/L alanine ratio The values in Table 3 can be used to esti- ϭ ϩ is close to unity, and this sample therefore ap- AA0 AAttAMINES , (2) mate the half-life to decarboxylation in gyp- pears to be the most pristine. Anza-Borrego sum at the temperatures characteristic of Mars ϩ using the Arrhenius equation: also has no detectable adenine, so we are con- where AMINESt is the MA EA concentra- ®dent that the amino acids are not from recent tion at time t. This also assumes that only trace t (T ) E ´⌬T contamination and are an ancient biosignature. levels of amines were present in the original ln 1/2 1 ϭ A , (4) The other samples may contain some levels of gypsum, which is what we found for the mod- []t1/2(T 2)R´T 1´T 2 organic matter of more recent origin, espe- ern gypsum from Chula Vista (Z ϭ 0.04). Sub- where t (T ) and t (T ) are the half-lives of cially the anhydrite and the jarosite samples. stituting equation 2 into equation 1 yields: 1/2 1 1/2 2 The ratio Z, the concentrations of MA ϩ decarboxylation at temperatures T1 and T2, re- spectively, E is the activation energy, and R EA divided by the concentrations of glycine ln(1 ϩ Z) ϭ k ´ t. (3) A ϩ DC is the universal gas constant (8.314 alanine, increases with age in the gypsum Ϫ1 Ϫ1 samples (Table 2). In both the Panoche Valley J´mol ´K ). Assuming a mean exposure tem- Equation 3 was used to estimate the rate of perature of ϳ0 ЊC for Haughton crater and ϳ20 jarosite and anhydrite samples, Z is less than decarboxylation (kDC) in gypsum from the ЊC for Anza Borrego and Panoche Valley, and expected for gypsum minerals from the same various sample localities and these values using the average of the aqueous Arrhenius ac- deposit, suggesting that the rate of decompo- were used to calculate the half-lives (t1/2) for tivation energies for glycine and alanine (ϳ160 sition of amino acids is slower in iron-rich or decarboxylation (Table 3). kJ/mol) determined in Li and Brill (2003), the dehydrated sulfates, that volatile amines are The calculated t1/2 values are in general half-lives of glycine and alanine decarboxyl- more easily lost from such minerals, or that consistent with the estimated average expo- ation in sulfate minerals at Martian tempera- the organic material is much more recent in sure temperature history of the gypsum sam- tures can be estimated (Table 3). origin in comparison to that in the gypsum ples since deposition. Although the present Temperatures on Mars over the past 4 b.y. samples. temperatures at Haughton crater are very cold are considered to be similar to those (Ͻ0 ЊC) Assuming that the change in the relative (average annual temperature of Ϫ16 ЊC), that prevail today (Shuster and Weiss, 2005). amounts of MA ϩ EA and glycine ϩ alanine when the crater and sulfate minerals were If the surface paleotemperature on Mars has in the various gypsum samples is entirely due formed there was an extended period of high- averaged 0 ЊC, then t for decarboxylation of to decarboxylation, then the data in Table 2 1/2 temperature hydrothermal activity (Parnell et glycine and alanine in gypsum is estimated to would be expected to obey the following ir- al., 2005). Even at 2±3 Ma, temperatures in be ϳ8 ϫ 108 yr based on the Anza-Borrego reversible ®rst-order kinetic relationship: this region were likely signi®cantly warmer results, ϳ1 ϫ 109 using the Panoche sample, than today (Brigham-Grette and Crater, 1992). and ϳ2 ϫ 106 for the Haughton crater sample. AA ln t ϭϪk ´t, (1) At the Anza-Borrego site, present average If the surface temperature on Mars has aver- ΂΃AA DC ϳ Њ Ϫ Њ 0 temperatures are 23 C (Remeika and Lind- aged 20 C, then t1/2 for decarboxylation of glycine and alanine is much longer, ranging from ϳ2 ϫ 1011 yr based on the Anza- TABLE 3. ESTIMATED RATES OF DECARBOXYLATION IN GYPSUM Borrego gypsum, to ϳ3 ϫ 1011 for Panoche ϳ ϫ 9 Location Average KDC t1/2 (years)* t1/2 on Mars t1/2 on Mars Valley, and 2 10 from the Haughton cra- Exposure T (years)Ϫ1 Ϫ20ЊC 0ЊC ter gypsum. The apparently shorter half-lives Њ ( C) (years) (years) predicted using the Haughton crater gypsum Anza-Borrego Gypsum 20 1.0 ϫ 10Ϫ7 6.8 ϫ 106 2.2 ϫ 1011 8.5 ϫ 108 may be explained by its exposure to either a Haughton Crater Gypsum 0 8.1 ϫ 10Ϫ8 8.6 ϫ 106 2.3 ϫ 109 8.6 ϫ 106 Panoche Gypsum 20 7.8 ϫ 10Ϫ8 8.9 ϫ 106 2.9 ϫ 1011 1.1 ϫ 109 warmer climate or hydrothermal conditions after deposition, or this may be the result of 0.693 *t ϭ more recent amino acid contamination dilut- 1/2 k DC ing the Z value. Because of the pristine nature

GEOLOGY, May 2006 359 of the Anza-Borrego gypsum, we consider the REFERENCES CITED San Diego, University of California, Scripps In- t values based on this sample to be the most Amelung, W., and Zhang, X., 2001, Determination of stitution of Oceanography, 172 p. 1/2 amino acid enantiomers in soil: Soil Biology & Langevin, Y., Poulet, F., Bibring, J.-P., and Gondet, B., reliable. These results imply that at modern Biochemistry, v. 33, p. 553±562, doi: 10.1016/ 2005, Sulfates in the North Polar region of Mars Martian surface temperatures, amino acids in S0038-0717(00)00195-4. detected by OMEGA/Mars Express: Science, gypsum should be preserved for periods in ex- Bada, J.L., 2001, State-of-the-art instruments for de- v. 307, p. 1584±1586, doi: 10.1126/science. tecting extraterrestrial life: National Academy of 1109091. cess of several billion years. Sciences Proceedings, v. 98, p. 797±800, doi: Li, J., and Brill, T.B., 2003, Spectroscopy of hydro- The estimated decarboxylation rates in sul- 10.1073/pnas.98.3.797. thermal reactions, part 26: Kinetics of decarbox- fate minerals on Mars are so slow that the lim- Bada, J.L., Wang, X.S., and Hamilton, H., 1999, Pres- ylation of aliphatic amino acids and comparison iting factor in the survival of amino acids is ervation of key biomolecules in the fossil record: with the rates of racemization: International Jour- Current knowledge and future challenges: Royal nal of Chemical Kinetics, v. 35, p. 602±610, doi: likely to be radiolysis in the upper 1±2 m of Society of London Philosophical Transactions, 10.1002/kin.10160. the regolith by galactic cosmic radiation ser. B, Biological Sciences, v. 354, p. 77±87, doi: Mancinelli, R.L., Fahlen, T.F., Landheim, R., and Klov- (Kminek, 2003) and UV-induced (Benner et 10.1098/rstb.1999.0361. stad, M.R., 2004, Brines and evaporites: Analogs Benner, S.A., Devine, K.G., Matveeva, L.N., and Pow- for Martian life: Advances in Space Research, al., 2000) or metal-catalyzed oxidation (Sum- ell, D.H., 2000, The missing organic molecules v. 33, p. 1244±1246, doi: 10.1016/j.asr. ner, 2004). The Martian iron-oxide rich soils on Mars: National Academy of Sciences Proceed- 2003.08.034. may provide a barrier against cosmic radia- ings, v. 97, p. 2425±2430, doi: 10.1073/ Park, L.E., and Downing, K.F., 2001, Paleoecology of tion, and organic material preservation may be pnas.040539497. an exceptionally preserved arthropod fauna from Biemann, K., OroÂ, J., Toulmin, P. III, Orgel, L.E., Nier, lake deposits of the Miocene Barstow Formation, increased at greater depths in the regolith. If A.O., Anderson, D.M., Simmonds, P.G., Flory, , U S.A: Palaios, v. 16, jarosite is present at these locations, then its D., Diaz, A.V., Rushneck, D.R., and Biller, J.A., p. 175±184. high iron content might assist preservation by 1976, Search for organic and volatile inorganic Parnell, J., Lee, P., Cockell, C.S., and Osinski, G.R., compounds in two surface samples from the 2004, Microbial colonization in impact- offering further shielding against radiolysis. Chryse Planitia region of Mars: Science, v. 194, generated hydrothermal sulphate deposits, While pure crystalline gypsum is transparent p. 72±76. Haughton impact structure, and implications for to visible and UV light (Parnell et al., 2004), Brigham-Grette, J., and Crater, L.D., 1992, Pliocene sulphates on Mars: International Journal of Astro- impure Antarctic gypsum crusts are essentially marine transgressions of northern Alaska: Cir- biology, v. 3, p. 247±256, doi: 10.1017/ cumarctic correlations and paleoclimatic interpre- S1473550404001995. opaque to radiation below 400 nm (Hughes tations: Arctic, v. 45, p. 74±89. Parnell, J., Osinski, G.R., Lee, P., Green, P.F., and Bar- and Lawley, 2003), and a few millimeters of Cockell, C.S., and Lee, P., 2002, The biology of impact on, M.J., 2005, Thermal alteration of organic mat- a similar mineral on Mars should be able to cratersÐA review: Biological Reviews, v. 77, ter in an impact crater and the duration of post- p. 279±310. impact heating: Geology, v. 33, p. 373±376, doi: shield gypsum from UV penetration. In the Conway, E.J., 1963, Microdiffusion analysis and 10.1130/G21204.1. absence of UV light, the contribution of volumetric error: New York, Chemical Pub. Co., Presser, T.S., and Ohlendorf, H.M., 1987, Biogeochem- metal-catalyzed oxidation should be minimal. p. 195±200. ical cycling of selenium in the San Joaquin Val- Therefore, amino acids and other organic Ertel, R.E., and Hedges, J.I., 1983, Bulk chemical and ley, California, USA: Environmental Manage- spectroscopic properties of marine and terrestrial ment, v. 11, p. 805±821, doi: 10.1007/ compounds should be extremely persistent in humic acids, melanoidins and catechol-based syn- BF01867247. sulfate minerals at the low temperatures on thetic polymers, in Christman, R.F., and Gjessing, Remeika, P., and Lindsay, L., 1992, Geology of Anza- Mars. E.T., eds., Aquatic and terrestrial humic materials: Borrego: Edge of creation: San Diego, California, Ann Arbor, Michigan, Ann Arbor Science Pub- Sunbelt Publications Inc. , 208 p. lishers, p. 143±163. Shuster, D.L., and Weiss, B.P., 2005, Martian surface CONCLUSIONS Gendrin, A., Mangold, N., Bibring, J.P., Langevin, Y., paleotemperatures from thermochonology of me- These results demonstrate that amino acids Gondet, B., Poulet, F., Bonello, G., Quantin, C., teorites: Science, v. 309, p. 594±597, doi: Mustard, J., Arvidson, R., and LeMouelic, S., 10.1126/science.1113077. and other organic compounds are well pre- 2005, Sulfates in Martian layered terrains: The Skelley, A.M., Scherer, J.R., Aubrey, A.D., Grover, served in terrestrial sulfate minerals. Based on OMEGA/Mars express view: Science, v. 307, W.H., Isvester, R.H.C., Ehrenfreund, P., Gruntha- these results, it is predicted that organic matter p. 1587±1591, doi: 10.1126/science.1109087. ner, F.G., Bada, J.L., and Mathies, R.A., 2005, Glavin, D.P., and Bada, J.L., 1998, Isolation of amino Development and evaluation of a microdevice for should be preserved over billions of years on acids from natural samples using sublimation: amino acid biomarker detection and analysis on Mars. Amino acids are excellent indicators for Analytical Chemistry, v. 70, p. 3119±3122, doi: Mars: National Academy of Sciences Proceed- the presence of other organic material because 10.1021/ac9803784. ings, v. 102, p. 1041±1046, doi: 10.1073/ they can be detected at very low levels using Glavin, D.P., Schubert, M., Botta, O., Kminek, G., and pnas.0406798102. Bada, J.L., 2001, Detecting pyrolysis products Squyres, S.W., Grotzinger, J.P., Arvidson, R.E., Bell, modern analytical techniques, including from bacteria on Mars: Earth and Planetary Sci- J.F. III, Calvin, W., Christensen, P.R., Clark, B.C., ones that potentially can be used to carry out ence Letters, v. 185, p. 1±5, doi: 10.1016/S0012- Crisp, J.A., Farrand, W.H., Herkenhoff, K.E., spacecraft-based in situ analyses (Skelley et 821X(00)00370-8. Johnson, J.R., KlingelhoÈfer, G., Knoll, A.H., Glavin, D.P., Cleaves, H.J., Schubert, M., Aubrey, A.D., McLennan, S.M., McSween, H.Y. Jr., Morris, al., 2005). Their structural diversity and chi- and Bada, J.L., 2004, New method for estimating R.V., Rice, J.W. Jr., Rieder, R., and Soderblom, rality may also provide a unique biological bacterial cell abundances in natural samples by L.A., 2004, In situ evidence for an aqueous en- signature, making amino acids excellent tar- use of sublimation: Applied and Environmental vironment at Meridiani Planum, Mars: Science, Microbiology, v. 70, p. 5923±5928, doi: 10.1128/ gets in the search for evidence of life on Mars v. 306, p. 1709±1714, doi: 10.1126/science. AEM.70.10.5923-5928.2004. 1104559. (Bada, 2001). Investigations of sulfate-rich Hess, J., Stott, L.D., Bender, M.L., and Schilling, J.G., Sumner, D.Y., 2004, Poor preservation potential of or- evaporite deposits, such as those seen at Mer- 1986, The Oligocene marine microfossil record: ganics in Meridiani Planum hematite- Age assessments using strontium isotopes: Paleo- idiani Planum, should be potential targets for bearing sedimentary rocks: Journal of Geophysi- ceanography, v. 4, p. 655±679. cal Research, v. 109, p. E12007, doi: 10.1029/ organic compounds on Mars. Hughes, K.A., and Lawley, B., 2003, A novel Antarctic 2004JE002321, doi: 10.1029/2004JE002321. microbial endolithic community within gypsum Zhao, M., and Bada, J.L., 1995, Determination of ␣- crusts: Environmental Microbiology, v. 5, ACKNOWLEDGMENTS dialkylamino acids and their enantiomers in geo- p. 555±565, doi: 10.1046/j.1462-;2920.2003. logical samples by high-performance liquid chro- National Aeronautics and Space Administration 00439.x. grant NAG5-12851 supported this work. We collected matography after derivatization with a chiral Jull, A.J.T., Courtney, C., Jeffrey, D.A., and Beck, J.W., adduct of o-phthaldialdehyde: Journal of Chro- the various samples, with the exception of the Haughton 1998, Isotopic evidence for a terrestrial source of matography A. v. 690, p. 55±63, doi: 10.1016/ crater sample, which was generously provided by John organic compounds found in Martian Meteorites 0021-9673(94)00927-2. Parnell and Pascal Lee. Bruce Deck determined the iso- Allan Hills 84001 and Elephant Moraine 79001: tope and total carbon and nitrogen data at Scripps In- Science, v. 279, p. 366±369, doi: 10.1126/ Manuscript received 6 October 2005 stitution of Oceanography. We thank Tabitha Hensley science.279.5349.366. Revised manuscript received 4 December 2005 for carrying out the preliminary strontium isotope anal- Kminek, G., 2003, The effect of ionising radiation on Manuscript accepted 14 December 2005 yses and Ron Amundson for sharing his extensive amino acids and bacterial spores in different geo- knowledge of California's Panoche Valley. and cosmochemical environments [Ph.D. thesis]: Printed in USA

360 GEOLOGY, May 2006