Weathering of ordinary chondrites from the Atacama Desert, Chile, by Mössbauer spectroscopy and synchrotron radiation X-ray diffraction

Authors


Corresponding author. E-mail: scorza@cbpf.br

Abstract

Some terrestrial areas have climatic and geomorphologic features that favor the preservation, and therefore, accumulation of meteorites. The Atacama Desert in Chile is among the most important of such areas, known as DCA. This desert is the driest on Earth, one of the most arid, uninhabitable localities with semiarid, arid, and hyper-arid conditions. The meteorites studied here were collected from within the DCA of San Juan and Pampa de Mejillones, located, respectively, in the Central Depression and the Coastal Range of the Atacama Desert. 57Fe Mössbauer spectroscopy was used for quantitative analysis of the degree of weathering of the meteorites, through the determination of the proportions of the various Fe-bearing phases and in particular the amount of oxidized iron in terrestrial alteration products. The abundance of ferric ions in weathered chondrites can be related to specific precursor compositions and to the level of terrestrial weathering. The aim of the study was the identification, quantification, and differentiation of the weathering products in the ordinary chondrites found in the San Juan and the Pampa de Mejillones areas of the Atacama Desert. The 57Fe Mössbauer spectroscopy study was complemented by synchrotron radiation X-ray diffraction and magnetic susceptibility measurements. The results allow a clear differentiation of the rate of weathering in meteorite samples collected from the San Juan versus the Pampa de Mejillones areas of the Atacama Desert.

Introduction

The study of meteorites is hindered by the fact that it relies largely on finding samples that may have been under the influence of the terrestrial environment, and thus altered by weathering. An interdisciplinary study is under way with meteorite samples from the Atacama Desert (AD), one of the oldest and driest deserts in the world (Dunai et al. 2005), in an attempt to understand the weathering processes acting on these primitive materials and the conditions of the accumulation surfaces that have preserved them. These samples may have undergone weathering processes that may result in large changes to their primary phases. Oxidation is certainly the most important reaction to consider and its products are usually crystalline and amorphous oxides and hydroxides associated to the iron-bearing primary minerals.

The degree of weathering in ordinary chondrites (OC) can be quantitatively measured using Mössbauer spectroscopy, which can be used to determine the proportions of the Fe-bearing phases, occurring mainly as Fe0 in the Fe-Ni metal, Fe2+ in the ferromagnesian silicates and troilite, and Fe3+ in terrestrial alteration products. In freshly fallen, equilibrated OC the amounts of Fe0 (kamacite and taenite) and Fe2+ (olivine, pyroxene, and troilite) are known within narrow limits. Thus, the abundance of oxidized iron in weathered chondrites can be related to specific starting compositions and to the level of terrestrial weathering.

The aim of this study was the identification and quantification of the weathering products of OC found in the San Juan (SJ) and Pampa de Mejillones (PM) areas of the AD. The 57Fe Mössbauer spectroscopy study was complemented by other techniques such as synchrotron radiation X-ray diffraction (SR-XRD) and magnetic susceptibility.

The Atacama desert

Some terrestrial areas, known as dense collection areas (DCA), have favorable climatic and geomorphologic features for the preservation and accumulation of meteorites. The AD is one of the most important of such places. Atacama is located between the western central Andes and the Pacific coast and extends from the southern border of Peru (18 °S) to Copiapó, Chile (30 °S). This desert is the driest in the world and it is one of the most arid, uninhabited, and old desert localities on Earth and comprises areas with semiarid, arid, and hyper-arid conditions. This hyper-aridity and associated very low erosion rates have persisted for over 25 Ma (Dunai et al. 2005). As a consequence, in addition to very low erosion rates, this extreme aridity has resulted in a number of unusual and unique features including the accumulation of a range of unusual salts.

Meteorites were collected within the DCA of SJ (25°26.4′ S; 69°52.2′ W) and PM (23°5′ S; 70°26′ W) corresponding to central depression (CD) and coastal range (CR) of the Chilean AD, respectively (Fig. 1). Hyperarid conditions occur in the CD area, where the precipitation is less than 10 mm yr−1 and measurable rainfall (1 mm or more) may be as infrequent as once every 5–20 yr. Nitrate and salar salts, typical of extreme aridity, are abundant in the CD of Atacama. These salts are concentrated in some places in this area and the deposits are formed in response to the extreme desiccation conditions of interior lakes, which originated from the groundwater flows from the Andes or regional aquifers. The arid conditions occur principally in all the CR where occasional winter rainfall and marine fog precipitation take place. The CR is directly affected by coastal fog and its influence varies markedly between different localities based on the topographic frame and the elevations (Muñoz et al. 2007).

Figure 1.

General map and section of the Chilean Andes showing the geomorphologic units present in the Atacama Desert. The SJ and PM areas are indicated by a star and circle, respectively. Adapted from Gattacceca et al. (2011).

Other hot-desert localities, which are also highly successful meteorite-find sites, have a more complex paleoclimatic history compared with AD. For example, the Roosevelt County, the Sahara, and the Nullarbor regions have experienced several humid/arid cycles over the last 40 ka (Bland et al. 1998). Likewise, the surface stabilities of the other principal hot-desert meteorite-find sites are relatively young compared with AD.

Experimental methods

The 57Fe Mössbauer spectroscopic measurements, in transmission geometry, were performed at room temperature (RT) and liquid helium temperature (4.2 K) using a 512 channel Halder spectrometer. The drive velocity was calibrated using a 57Co in Rh matrix source and an iron foil both at RT. All measurements were performed at high velocity (±12 mm s−1), with an average recording time of 12 h per sample, absorbers and source at same temperature. Mössbauer absorbers were prepared with 100 mg cm−2 of the bulk meteorite sample. Normos code (Brand 1995) was used for spectral analysis based on a least-squares fitting routine, assuming each spectrum to be a sum of Lorentzian lines grouped into quadrupole doublets and magnetic sextets. Magnetic relaxation effects in low-temperature spectra are fitted using the spheric relaxation model of the Normos program. The isomer shifts are given relative to α-Fe at RT.

The Mössbauer measurements were performed on 32 meteorites collected in the SJ and other localities (SJ and OL) area and 11 meteorites from Pampa de Mejillones (PM) in the AD. The two meteorites from other localities correspond geographically to the CD or to its transition to the high cordillera. The meteorites are weathered OC and include the chemical groups H (22 samples), L (18 samples), and LL (3 samples). The geographic coordinates for all studied meteorites are reported in the Meteoritical Bulletin 87 (Russell et al. 2003), 95 (Weisberg et al. 2009), and 97 (Weisberg et al. 2010).

X-ray diffraction was carried out at the Brazilian Synchrotron Light Laboratory (LNLS) in Campinas, Brazil. Synchrotron light was employed to enhance the intensity, to allow a choice of wavelength to prevent iron fluorescence, and to provide a very monochromatic X-ray. The D10A-XRD2 beamline was used with wavelength 1.127130 Å, a germanium detector, 1 mm primary slit and 4 mm secondary slit, and θ–2θ geometry. The data were recorded using 2θ range from 10° to 110° with the acquisition time determined by the total photon arriving at a calibration monitor. The measured diffraction pattern showed high intensity and well-defined peaks with a reasonable low background.

Rietveld refinement was done with the MAUD software (Lutterotti 2010) with a pseudo-Voigt function to describe the peaks and a Caglioti function to describe the full width at half maximum (FWHM), as a function of the angle. The compositions of all phases were kept constant during the least squares fitting; however, the iron content in the olivine and pyroxene were adjusted to the values found by microprobe. No preferred orientation correction was employed for any of the phases.

Magnetic susceptibility (χ in 10−9 m3 kg−1) at RT was measured on large samples following the procedure described by Rochette et al. (2003). In OC, χ is a proxy for the FeNi metal amount, it is less specific than Mössbauer spectroscopy but it has the advantage of avoiding nugget effects due to measurement of large and unprocessed volumes. Previous investigations have shown that the value of χ in falls has a narrow range for each group and decreases as a function of weathering grade. This decrease has been shown in finds from the Sahara (Rochette et al. 2003) and Atacama (Rochette et al. 2013).

57Fe Mössbauer Spectra of Weathered Meteorites

The Mössbauer spectra at room and low temperatures show an overlapping of paramagnetic and magnetic phases as seen in Fig. 2 for two selected samples from the SJ area, showing a low (SJ 013) and high (SJ 025) weathering degree. Depending on their magnetic properties, iron oxides appear in the RT-Mössbauer spectra as doublets associated with paramagnetic or superparamagnetic species (small particles of goethite, akaganeite, and lepidocrocite) or sextets associated with magnetically ordered species (maghemite, goethite, magnetite, and hematite). At low temperature, the paramagnetic/superparamagnetic doublets split into magnetic sextets depending on the ordering/blocking temperature.

Figure 2.

Mössbauer spectra of some selected meteorite samples showing low (SJ 013) and high (SJ 025) degree of weathering from the San Juan area: a) room temperature, b) low temperature (4.2 K).

The Mössbauer spectra exhibit the primary phases: olivine, pyroxene, troilite, and Fe-Ni (kamacite/taenite). They also exhibit the weathering products: goethite, maghemite and/or hematite, and akaganeite. These different phases observed in meteorites are identified by their hyperfine parameters at 300 K and 4.2 K. The complete identification of the oxidation products is only possible with measurements at 4.2 K. The low-temperature measurements also minimize the differences in recoil-free fraction of the Fe sites in the meteorite sample, and hence give more accurate results for the relative areas of the oxidation products. Therefore, in our analysis we use the results obtained at 4.2 K.

In Atacama meteorites, the total oxidation (%), defined as the sum of all phases containing Fe3+ product of weathering shows a wide range, as shown in Fig. 3 for some samples from SJ, OL, and PM areas. Selected hyperfine parameters derived from the spectra of Fig. 3 are given in Table 1a. Phase quantification was done using the relative areas taken from fitted Mössbauer spectra (Table 2). The relative areas of absorption in a Mössbauer spectrum give the relative proportions of Fe in the Fe-bearing phases in the sample. In this case, it is assumed that all the phases have the same recoil-free fraction.

Table 1a. Hyperfine parameters for some selected meteorite samples from SJ, OL, and PM areas obtained by fitting the spectra at room temperature
MeteoritePrimary mineralogyWeathering products
OlivinePyroxeneFe-NiTroiliteParamagnetic Fe3+GeothiteMaghemite
ISQSISQSISQSBhfISQSBhfISQSISQSBhfISQSBhf
  1. Isomer shift (IS ± 0.05 mm s−1), quadrupole splitting (QS ± 0.05 mm s−1), hyperfine field (Bhf ± 0.5 T).

Collected in San Juan & OL Area
SJ 0311.142.931.152.090.04−0.0533.90.75−0.1731.20.380.78
SJ 0181.152.911.152.080.01−0.0732.60.77−0.1331.30.400.68 0.330.0950.0
SJ 0151.152.961.162.11−0.04−0.0933.50.79−0.1931.10.350.730.33−0.1632.00.350.1049.4
SJ 0061.152.951.172.110.00−0.0433.20.72−0.1830.90.340.680.35−0.1731.60.360.0148.3
SJ 0201.152.971.152.070.01−0.0633.80.750.1830.10.370.670.38−0.1631.70.330.0149.0
SJ 0051.163.001.152.130.00−0.0533.80.77−0.1431.10.380.660.35−0.1732.60.330.0149.4
Lutschaunig1.222.891.092.16−0.100.0233.90.72−0.2131.40.600.64
Collected in Pampa de Mejillones Area
PdM 0141.162.981.172.120.02−0.0633.60.76−0.1131.60.400.71
Pampa B1.102.811.082.17−0.070.0233.00.74−0.0630.90.410.480.33−0.2632.60.370.0149.2
PdM 0041.173.011.182.12−0.060.0633.50.75−0.1331.20.380.720.33−0.2631.50.340.0150.1
Pampa G1.173.011.182.12−0.200.0633.00.760.0931.90.370.720.37−0.1533.30.340.0149.9
PdM 0121.162.981.172.08−0.050.0633.30.720.0931.90.360.670.37−0.1532.30.320.0149.1
                   
Table 1b. Hyperfine parameters for some selected meteorite samples from SJ, OL, and PM areas obtained by fitting the spectra at 4.2 K
MeteoritePrimary mineralogyWeathering products
Olivine 1Olivine 2Pyroxene 1Pyroxene 2Fe-NiTroiliteAkaganeiteGeothiteMaghemite 1Maghemite 2
ISQSBhfωISQSBhfωISQSBhfωISQSBhfωISQSBhfISQSBhfISQSBhfISQSBhfISQSBhfISQSBhf
  1. Isomer shift (IS ± 0.07 mm s−1), quadrupole splitting (QS ± 0.07 mm s−1), hyperfine field (Bhf ± 0.8 T), relaxation rate (ω ± 0.2 MHz).

Collected in San Juan and OL Area
SJ 0310.852.694.87.61.132.3513.60.21.252.701.30.151.23−0.7424.50.400.12−0.0834.10.82−0.2032.80.41−0.1249.50.46−0.2550.7
SJ 0180.453.834.44.51.373.039.83.51.412.371.40.711.18−0.7925.30.340.22−0.0234.50.81−0.1332.60.42−0.2248.50.53−0.2350.20.39−0.0153.80.470.0052.6
SJ 0150.373.984.53.01.393.0910.33.01.362.281.10.911.04−0.7426.50.440.17−0.0334.90.90−0.1932.90.43−0.1748.70.51−0.2250.20.46−0.0153.60.410.0052.5
SJ 0060.483.574.43.01.263.0210.11.91.262.451.30.151.05−0.7924.80.440.12−0.0932.70.80−0.1732.90.44−0.2448.60.47−0.2250.50.40−0.0153.90.440.0052.5
SJ 0200.733.514.53.01.403.0310.01.91.392.481.30.151.23−0.7725.50.440.11−0.0934.50.88−0.2232.90.43−0.2348.20.49−0.2050.30.46−0.0153.70.480.0052.7
SJ 0050.553.403.93.01.423.139.91.91.432.431.30.151.22−0.8725.90.440.13−0.0934.50.87−0.1533.40.43−0.2149.20.49−0.2150.90.49−0.0153.90.450.0052.5
Lutschaunig0.333.794.73.01.233.0210.43.01.202.391.10.91.05−0.6925.10.40.01−0.1234.30.780.1633.00.44−0.2348.50.36−0.1950.9
Collected in Pampa de Mejillones Area
PdM 0140.493.734.53.51.233.1210.12.51.212.310.90.91.01−0.6925.60.40.09−0.1334.00.78−0.1432.90.44−0.2649.20.34−0.1550.5
Pampa B0.303.874.53.01.233.1210.31.91.252.281.30.21.15−0.8325.41.00.15−0.0934.30.78−0.1033.00.46−0.2048.60.39−0.2850.10.41−0.0153.50.400.0052.7
PdM 0040.433.794.23.01.253.069.91.91.272.281.10.21.12−0.8624.91.00.12−0.0934.60.82−0.0832.80.42−0.2148.40.37−0.2450.00.42−0.0153.50.460.0052.4
Pampa G0.553.764.73.01.283.089.91.91.282.341.30.21.19−0.8025.31.00.11−0.0734.70.82−0.1232.60.39−0.2248.80.37−0.2450.30.42−0.0153.40.440.0051.5
PdM 0120.833.153.43.01.253.1210.11.91.262.461.30.21.27−0.8024.71.00.11−0.0734.80.82−0.1233.20.40−0.1948.30.36−0.2050.30.47−0.0153.60.450.0052.1
                                   
Table 2. Relative area for primary phases and weathering products obtained from the fitting of the Mössbauer spectra of meteorites recovered in the SJ, OL, and PM areas: a) room temperature. b) low temperature. The shock stage, chemical groups, and log χ (in 10−9 m3 kg−1) are also indicated
a
MeteoriteShock StageTypePairing groupPrimary mineralogyWeathering productsTotal oxide (%)Paramagentic Fe3+
Olivine (%)Pyroxene (%)FeNi (%)Troilite (%)Goethite (%)Maghemite (%)Paramagnetic Fe3+ (%)IS (mm s−1)QS (mm s−1)
Collected in San Juan area
SJ 029S1H3(3.2/3.5) 20.112.414.85.116.96.424.247.50.370.84
SJ 010S2H3(3.8) 21.616.62.33.627.74.923.455.90.360.71
SJ 006S2H3(3.9) 26.016.45.18.716.54.822.643.80.350.74
SJ 027S2H3-5(3.2) 22.212.79.15.221.117.012.750.80.370.71
SJ 011S2H4 23.617.73.24.822.110.518.150.70.350.65
SJ 021S1H4 18.911.03.98.623.321.412.857.50.360.77
SJ 003S3H5 17.210.62.712.731.38.217.556.90.360.80
SJ 012S2H5 24.215.74.37.717.913.017.248.10.360.78
SJ 016S2H5 27.115.98.210.718.16.613.438.10.370.69
SJ 020S3H5 22.612.79.64.117.816.816.451.00.370.67
SJ 023S3H5332.017.212.615.3nd4.918.123.00.320.83
SJ 030S3H5321.413.521.53.013.16.321.240.60.360.73
SJ 025S3H5 16.714.83.94.231.29.519.860.50.360.71
SJ 028S3H5 23.418.224.515.3ndnd18.518.50.380.77
SJ 032S3H5/6 29.116.67.610.615.76.813.736.20.370.67
SJ 005S3H6 18.311.02.610.320.013.524.357.80.380.69
SJ 007S2H6 25.617.02.97.617.713.415.746.90.360.77
SJ 017S2H6 22.013.15.112.719.610.717.047.20.360.70
SJ 031S3L3(3.8/3.9) 43.218.78.120.3ndnd9.89.80.380.74
SJ 013S3L3(3.9) 42.018.512.512.9ndnd14.114.10.390.82
SJ 004S3L4 40.427.54.213.4ndnd14.614.60.370.67
SJ 018S3L5 44.816.65.314.1nd3.216.019.20.350.69
SJ 014S3L6839.319.02.93.514.04.117.235.30.380.76
SJ 015S3L6839.814.31.914.77.44.916.929.20.410.82
SJ 019S3L6 43.220.67.514.3ndnd14.414.40.370.71
SJ 022S4L6939.718.36.213.3nd5.616.922.60.350.65
SJ 024S3L6940.518.86.410.8nd6.417.223.50.360.77
SJ 026S1L6 39.919.23.13.016.8nd18.034.80.360.80
SJ 008S3LL6 56.721.91.01.0ndnd19.419.40.360.78
Collected in Pampa de Mejillones area
PdM 012S2H4 12.08.33.12.231.431.011.974.30.450.59
PdM 002S2H5 21.813.54.215.313.38.223.745.20.330.70
PdM 014S3L/LL4-6 45.523.24.013.60.0nd13.413.40.310.65
La Yesera 003S3L4 25.316.13.13.220.59.622.352.40.330.77
Pampa BS4L4/5 34.514.82.411.316.95.814.337.10.460.63
Pampa DS2L5 30.017.32.92.421.414.012.047.40.410.54
Pampa GS2L5 20.413.83.62.527.715.816.259.70.430.62
La Yesera 004S2L6 35.318.93.16.713.112.210.735.90.430.65
PdM 004S3L6 29.416.82.43.818.113.116.447.70.430.59
PdM 007S3L6 19.412.42.22.524.923.315.363.50.410.61
La Yesera 02S2LL5 38.026.83.611.7nd3.017.020.00.420.67
Collected in other localities
Estación ImilacS4H5 26.521.211.113.610.85.011.827.60.370.69
RencoretS3H6 27.914.46.05.910.414.021.445.90.370.67
Lutschaunig's stone L6 38.021.77.621.8ndnd10.910.90.320.83
b
MeteoriteShock stageTypePairing groupPrimary mineralogyWeathering productsTotal oxide (%)Log χ
Olivine (%)Pyroxene (%)FeNi (%)Troilite (%)Akaganéite (%)Goethite (%)Maghemite (%)
  1. PdM=Pampa de Mejillones; nd = not detected.

  2. IS = isomer shift; QS= quadrupole splitting; Pairing group proposed by Gattacceca et al. (2011).

Collected in San Juan area
SJ 029S1H3(3.2/3.5) 21.713.313.54.612.126.68.146.94.93
SJ 010S2H3(3.8) 21.715.82.43.714.435.16.956.44.64
SJ 006S2H3(3.9) 27.716.85.39.414.520.95.540.84.90
SJ 027S2H3-5(3.2) 21.811.96.66.211.125.117.253.44.98
SJ 011S2H4 23.814.63.65.720.721.79.952.35.04
SJ 021S1H4 20.213.54.58.35.729.118.853.64.94
SJ 003S3H5 18.911.02.210.28.940.28.557.65.18
SJ 012S2H5 25.717.34.37.87.725.012.244.95.06
SJ 016S2H5 24.216.46.512.110.621.98.440.95.16
SJ 020S3H5 22.513.99.64.18.323.218.550.04.98
SJ 023S3H5331.317.010.913.78.010.98.327.15.23
SJ 030S3H5321.113.321.34.710.621.77.339.65.24
SJ 025S3H5 18.714.21.84.710.239.410.960.55.19
SJ 028S3H5 21.919.823.414.86.513.7nd20.25.29
SJ 032S3H5/6 27.516.47.810.313.015.99.137.95.27
SJ 005S3H6 20.211.21.211.617.126.712.055.84.43
SJ 007S2H6 26.716.22.99.79.021.713.744.45.14
SJ 017S2H6 23.112.34.812.313.122.611.747.54.91
SJ 031S3L3(3.8/3.9) 43.720.08.420.45.32.2nd7.54.98
SJ 013S3L3(3.9) 41.018.412.015.76.56.3nd12.94.73
SJ 004S3L4 41.625.56.513.47.45.5nd13.04.78
SJ 018S3L5 43.616.74.814.410.35.54.720.44.82
SJ 014S3L6838.718.24.43.712.416.06.534.94.38
SJ 015S3L6839.115.23.015.86.715.54.726.94.53
SJ 019S3L6 41.220.37.315.44.511.3nd15.74.63
SJ 022S4L6938.718.46.211.76.810.87.425.14.43
SJ 024S3L6941.219.95.48.96.79.48.624.64.66
SJ 026S1L6 37.618.92.94.08.927.6nd36.64.44
SJ 008S3LL6 55.022.81.01.010.69.6nd20.23.18
Collected in Pampa de Mejillones area
PdM 012S2H4 12.88.72.03.012.332.628.773.64.77
PdM 002S2H5 20.613.12.316.811.123.412.747.24.88
PdM 014S3L/LL4-6 45.118.63.314.29.99.0nd18.94.55
La Yesera 003S3L4 27.215.91.24.39.531.810.151.44.21
Pampa BS4L4/5 36.416.52.011.17.620.36.134.14.18
Pampa DS2L5 30.415.91.94.010.522.814.647.94.52
Pampa GS2L5 21.613.22.03.28.336.815.060.14.30
La Yesera 004S2L6 37.220.01.33.56.319.212.538.14.39
PdM 004S3L6 29.215.31.72.86.729.714.550.94.34
PdM 007S3L6 20.412.31.53.27.634.220.762.54.60
La Yesera 02S2LL5 39.126.91.49.64.714.34.123.13.87
Collected in other localities
Estación ImilacS4H5 27.422.311.915.07.515.9nd23.45.32
RencoretS3H6 25.815.63.48.210.323.213.446.95.04
Lutschaunig′s stone L6 39.520.76.720.67.55.1nd12.64.62
Figure 3.

Mössbauer spectra of selected meteorite samples from a) SJ area and b) PM and OL areas obtained at room (left column) and low (right column) temperature showing different range in total oxidation.

Weathering of Primary Minerals

Determining differences in alteration rates between phases in a meteorite is a rather difficult analytical problem. Mössbauer spectroscopy is a useful tool in this regard. Observing how the spectral area of a primary phase varies with the total amount of oxidation allows the identification of phases that are most susceptible to weathering: an increase in a ferric component should be accompanied by a concomitant decrease in an absorption associated with a primary mineral component. Comparing the spectral area of opaque phases (Fe-Ni and troilite) and ferromagnesian silicates with that of total oxidation, a decrease in primary phases with an increase in oxidation is observed, suggesting that all Fe-containing minerals within the meteorite are affected by weathering to some degree (Figs. 4a and 4b). Interestingly, although ferromagnesian silicates appear to be weathered at a constant rate, there was no detectable difference in the rate of weathering between olivine and pyroxene (Fig. 4c). It can be seen that the spectral areas of opaque and ferromagnesian silicates against total oxidation exhibit the same behavior for SJ, OL, and PM suggesting that the weathering of these minerals occurs with very similar relative rates regardless of location.

Figure 4.

Spectral area of total a) ferromagnesian silicates and b) opaques against total oxidation derived from the 4.2 K Mössbauer spectra recorded from weathered meteorite samples of SJ, OL, and PM areas. The results show that the weathering process occurs almost at a constant rate for all samples. c) Spectral area of olivine (Ol) to olivine + pyroxene (Ol+Px) as a function of oxidation indicating that both olivine and pyroxene are equally susceptible to oxidation.

Oxidation Products

Atacama meteorites show a wide total oxidation range. In samples from SJ and OL area, the total ferric phase may range from 8% to 60%. Samples from PM show a range from 19% to 74%, with higher average oxidation, in agreement with the more humid climate (Table 2b, Fig. 6). There is a clear tendency, in a given site, to have larger oxidation in H than in L and LL. This is easily explained by the different initial ratio of Fe° over Fe2+, the former (larger in H) being more rapidly oxidized than the latter (Bland et al. 2002).

For all samples collected from the SJ, OL, and PM areas, the paramagnetic Fe3+ component contributes 10–24% to the total area (Table 2a). The paramagnetic Fe3+ phase orders magnetically at low temperature, splitting into akaganeite and goethite. In samples with low oxidation (<20%), the paramagnetic Fe3+ is dominated by akaganeite (Munayco et al. 2010).

According to Buchwald and Clarke (1989) the weathering mechanism that could explain this suite of oxidation products is the following: akaganeite precipitates at a reaction front, incorporating Cl ions from the environment into ion exchange sites. Over time, akaganeite ages and transforms to maghemite and goethite, making Cl available for further depassivating corrosive action. Thus, akaganeite may be identified as the first oxidation product formed in meteorites collected in the SJ, OL, and PM areas.

The Mössbauer spectra at low temperature are very complex due to the magnetic relaxation effects of the ferromagnesian silicates attributed to spin-spin relaxation of Fe2+ ions and the overlapping of ferric phases. These problems make the correct identification of maghemite and/or hematite difficult due to the close similarity in their hyperfine fields. To explore the presence of maghemite, magnetic separation was performed using a hand magnet. Extraction of the maghemite/metal particles from the bulk sample was accomplished by grinding the meteorite fragments into a powder using a mortar under acetone. Mössbauer spectra at room and low temperatures of maghemite/metal particles extracted from some selected samples, from both, SJ, OL, and PM areas, clearly show the presence of ferrimagnetic phases: maghemite, goethite, and akaganeite (Stevens et al. 1998) (Fig. 5). In all the samples studied, hematite was not detected. The observation of the presence of maghemite and the absence of hematite has also been seen in some of these meteorites by thermal demagnetization of the remanent magnetization (Uehara et al. 2012).

Figure 5.

Mössbauer spectra at room and low temperature of the Fe-oxide and metal particles extracted from the San Juan 020 meteorite, showing clearly the presence of maghemite.

In the SJ, OL, and PM areas studied, the same oxidation products are observed: akaganeite, goethite, and maghemite. The contribution of these phases to the total spectra is shown in Table 2b. As expected for an area with a lower level of oxidation, SJ and OL have a higher akaganeite percentage, since the transformation into other oxides is still lower than in regions with higher oxidation levels such as PM in which maghemite is detected in a higher percentage.

Differences in Weathering Rate between Sites

Samples collected from the PM area show a wide total oxidation range in contrast with samples from SJ and OL area. Fig. 6 shows total oxidation-frequency histograms for the two areas, after we have accounted for pairing of the meteorites. In PM, pairing is likely not an important issue due to the small number of samples and low concentration samples per km2 (approximately 0.3 meteorite km−2). Nevertheless, we checked possible pairing based on proximity (distance less than 3 km) and close petrographic, magnetic, and Mössbauer data. No obvious pairing emerged, although one may suspect possible pairing among the L5 chondrites Pampa D and G (within 1.9 km) and L6 PdM 004 and 007 (2.8 km), or between Pampa D and PdM 004 (within 1.8 km). Still, these distances are quite large and the match in all data is not perfect, so we decided not to consider pairing for the PM area. In contrast, in SJ, where there is a high concentration (approximately 14–19 meteorite km−2), this question becomes an important consideration. We used the pairing proposed by Gattacceca et al. (2011), that resulted in six meteorites being paired (see Table 1b). However, in our study, the frequency histograms of oxidation show no significant difference with or without pairing. The histograms exhibit two peaks near 15% and 50% oxidation for the SJ and OL area, and one peak at 50% for PM area. These two peaks in the SJ and OL area reflect the existence of two different populations that can be related to the different chemical groups (types H, L, and LL) of the samples. The peak at 15% corresponds preferentially to samples of type L/LL, while the maximum at 50% can be related to samples of type H. Therefore, for a given chemical group, oxidation is clearly larger for PM than for the SJ and OL areas. On the other hand, based on the much larger concentration observed at SJ, the average terrestrial age for SJ is expected to be much larger than in PM. Therefore, the difference in oxidation we observe is clearly linked to the much larger oxidation rate in PM, which itself is due to much more humid climatic conditions and possibly higher marine salt input. When compared with Bland et al.'s (1998, 2002) data, the range of oxidation at PM is comparable to U.S. or Australia hotdesert finds, while SJ and OL are more comparable to Sahara finds or even Antarctic meteorites.

Figure 6.

Frequency against oxidation percentage (in 11% bins) for all the Atacama Desert samples obtained in the present study at 4.2 K.

A plot of the quadrupole splitting (QS) versus isomer shift (IS) for paramagnetic Fe3+ observed at RT (Fig. 7) reveals that the samples from the SJ and OL areas show one cluster with average IS = 0.36 mm s−1 and QS = 0.73 mm s−1 consistent with iron site in akaganeite and goethite (Barrero et al. 2006). A very different result is obtained for samples from PM that exhibit two broad clusters, one with the same IS and QS as found for the SJ and OL area, and a second with average IS = 0.43 mm s−1 and QS = 0.61 mm s−1. These two clusters were also identified by Bland et al. (1998) and labeled A and B. This dichotomy may be due to a difference in crystallinity or size of superparamagnetic particles of goethite and akaganeite, reflecting different conditions in the process of weathering of the samples from SJ, OL, and PM areas. It is noteworthy that both clusters are present in all hot-desert meteorites studied by Bland et al. (1998), although the B cluster is less represented in the Sahara compared with Australia and the United States. On the other hand, Antarctic meteorites exhibit only the A cluster as observed in the SJ and OL samples.

Figure 7.

Quadrupole splitting (QS) versus isomer shift (IS) for paramagnetic Fe3+ exhibiting two different clusters labeled (A) and (B) for meteorites collected in the SJ, OL, and PM areas.

Synchrotron Radiation X-Ray Diffraction (SR-XRD)

Typical SR-XRD diffractograms for samples from the AD can be seen in Fig. 8 for six selected samples with different degrees of weathering, from SJ (SJ 015, SJ 025, SJ 031) and PM and OL (Lutschauning, Pampa B, PdM 012).

Figure 8.

SR-XRD diffractograms for some meteorites recovered in the SJ, PM, and OL areas.

SR-XRD allowed the discrimination and quantification of the iron and the noniron phases. Up to 11 different phases were considered in the Rietveld modeling. The results show the primary phases: olivine, orthopyroxene, diopside, albite, anorthite, Fe-Ni, troilite, and the weathering products: goethite, maghemite, akaganeite (Munayco et al. 2011), and an unknown phase, which could be associated, from its crystallographic parameters, with nchwaningite (a hydrated manganese silicate (Mn2SiO3(OH)2·H2O) a rare mineral discovered in 1995 (Nyfeler et al. 1995). This phase was detected in 1 PM sample and in two thirds of SJ samples. Additional studies, such as SEM-EDS and SR-XRD under special conditions, are necessary to confirm the presence of this mineral.

The scattering factors for the iron and magnesium sites of the orthopyroxene and olivine phases were calculated using mean compositions of these elements obtained from the electron microprobe analysis. Table 3 shows the weight percent (wt%) of the different phases obtained by refinement of the diffractograms.

Table 3. Weight percent of the primary phases and weathering products obtained from the Rietveld refinement method for meteorites recovered in the SJ, OL, and PM areas
MeteoriteOlivine (wt%)Pyroxene (wt%)Plaglioclase (wt%)Opaques (wt%)Total oxide (wt%)
OlivineOrthopyroxeneDiopsideNchwaningiteAnorthiteAlbiteFe-NiTroiliteGoethiteAkaganeiteMaghemite
  1. nd = not detected.

Collected in San Juan area
SJ 1340.135.77.40.73.76.31.23.41.6ndnd
SJ 1442.035.09.01.9nd5.8nd0.64.1nd1.6
SJ 1540.232.19.21.0nd5.4nd4.65.8nd1.6
SJ 1733.536.19.02.1nd5.90.83.86.2nd2.6
SJ 2048.736.65.20.72.4ndnd2.83.6ndnd
SJ 2130.941.38.51.0nd5.10.41.56.00.84.5
SJ 2229.028.011.7ndnd7.10.53.813.8nd6.1
SJ 2349.334.87.9nd2.3ndnd1.62.8nd1.3
SJ 2634.532.96.8nd1.96.7nd0.911.51.33.5
SJ 2747.831.86.8ndnd5.80.41.06.5ndnd
SJ 2932.841.67.01.11.47.03.72.62.8ndnd
SJ 3049.222.53.90.7nd6.91.50.711.31.12.3
SJ 3131.136.25.50.1nd7.56.91.98.4nd2.3
SJ 3251.621.913.51.33.11.21.44.51.4ndnd
SJ 3328.643.98.21.7nd6.10.73.25.2nd2.5
Collected in other localities
Lustchaunig's stone46.935.58.0nd0.9nd1.74.52.5ndnd
Collected in Pampa de Mejillones
Pampa B41.443.42.9nd4.2ndnd1.55.00.70.9
La Yesera 0347.927.03.00.5nd5.6nd1.612.50.51.4
Pampa G37.633.55.4ndnd6.00.60.413.3nd3.3
PdM 01231.432.35.6ndnd5.4nd0.612.21.610.9

In the SR-XRD diffractograms, it is difficult to detect akaganeite due to poor crystallinity or small particles size. However, it is clearly observed by Mössbauer spectroscopy at temperatures approximately 50 K, which shows magnetically split peaks, indicating superparamagnetic behavior or a higher degree of structural disorder.

The difficulty in the detection of akaganeite is a disadvantage for SR-DRX in relation to Mössbauer spectroscopy in the study of the weathering of meteorites. The comparison between Rietveld and Mössbauer phase quantification was done with the necessary conversion of the X-ray results of the Fe-containing phases from weight percentage to molar percentage of iron. This conversion was made using the iron percentage inside each crystalline phase, the number of molecules inside the unit cell, and total cell mass of each phase. Figures 9 and 10 show the comparison between Mössbauer and SR-XRD results for silicates (olivine and orthopyroxene), opaques (Fe-Ni and troilite), and oxidation products (akaganeite, goethite, and maghemite) expressed in % of the Fe-containing phases.

Figure 9.

Comparison between Fe(silicates)/Fe(total) (and Fe[opaques]/Fe[total]) ratio (%) determined by Mössbauer spectroscopy at 4.2 K and that calculated from SR-XRD for meteorites recovered in the SJ, OL, and PM area. The straight line indicates the 1:1 relation.

Figure 10.

Comparison between Fe(weathering products)/Fe(total) ratio (%) determined by Mössbauer spectroscopy at 4.2 K and SR-XRD for meteorites recovered in the SJ, OL, and PM area. The straight line indicates the 1:1 relation.

Silicate, opaque, and oxidation product identification shows good agreement between Mössbauer and SR-XRD. The quantification of silicates obtained by SR-XRD is always above that obtained by Mössbauer spectroscopy. This is perhaps due to the fact that in the conversion process of the XRD results, an average value obtained from microprobe measurements is used for Fe in olivine and orthopyroxene.

Total oxidation (Fig. 10) obtained from SR-XRD is below that estimated from Mössbauer. This is probably related to the difficulty of identifying the akaganeite phase in the SR-XRD data.

Magnetic Susceptibility

Total oxidation from Mössbauer spectroscopy shows a significant correlation with magnetic susceptibility, when separating H and L that have different initial ranges of ferromagnetic FeNi minerals (Fig. 11a). The magnetic susceptibility value observed for H and L falls are 4.87 ± 0.10 and 5.32 ± 0.10, respectively. By separating L and H, one can see that in the present database H chondrites show on average higher oxidation, but a slower decrease in magnetic susceptibility as a function of degree of oxidation. This important dispersion in the correlation may be due to the fact that both parameters were measured on different samples from the same meteorite, thus possibly showing heterogeneous weathering on the scale of the samples used for Mössbauer spectroscopy (approximately 150 mg). Another explanation could be that magnetic susceptibility is affected by the type of oxidation products, ferrimagnetic maghemite being able to contribute significantly to susceptibility, while the other paramagnetic and antiferromagnetic phases have a negligible contribution. To test this hypothesis, we plotted logχ as a function of total oxidation minus maghemite% (Fig. 11b). We computed the R2 values (correlation coefficient) for the combined H and L populations, and observed that the R2 values change from 0.388 to 0.503 for L and 0.378 to 0.336 for H populations, when maghemite was removed. The correlation was only marginally improved for L populations. When plotted as a function of maghemite%, or maghemite+metal%, magnetic susceptibility decreases. All this suggests that maghemite does not contribute significantly to magnetic susceptibility. Metal% and magnetic susceptibility are poorly correlated (Fig. 11c). This is eventually due to nugget effect on metal amount determination by Mössbauer.

Figure 11.

Magnetic susceptibility χ in 10−9 m3 kg−1 versus Mössbauer parameters, separating H and L; a) versus total oxidation%; b) versus total oxidation minus maghemite%; c) versus metal%.

Conclusions

The 57Fe Mössbauer spectroscopy results showed higher average oxidation in metal-rich (H) than in metal-poor (L and LL) chondrites and a clear differentiation of meteorite samples collected from the SJ, OL, and PM areas. The degree of weathering in PM area meteorites is higher than in those from the SJ and OL areas, for a given chemical group. These results show a broad total oxidation range between approximately 8 and 60% for meteorites collected in SJ and OL area and between approximately 19 and 74% for meteorites from PM area. The oxidation index was determined by the presence of Fe-oxides appearing as magnetically ordered Fe3+ (maghemite and goethite) and paramagnetic Fe3+ compounds, identified as akaganeite and detected by Mössbauer spectroscopy in all samples and by SR-XRD in some samples. The decrease in primary phases with increasing oxidation suggests that all iron-containing minerals within a meteorite were affected by weathering to some degree.

The presence of two different populations in the oxidation-frequency histogram can be related to different chemical group of the samples. In SJ and OL meteorites, the lower peak corresponds to meteorites from type L and LL and the highest peak to type H. In PM meteorites the L and LL peak is shifted to higher values with respect to SJ and OL samples, which suggests that the meteorite population in PM is more oxidized. There are not enough H meteorites in PM to compare with the H population from SJ and OL.

The synchrotron radiation X-ray diffraction allowed the quantification of all crystalline phases with and without iron. Thus, this characterization technique has an important advantage over Mössbauer spectroscopy, which can only identify the Fe-containing phases. On the other hand, poorly crystallized phases or very small particles cannot be detected by SR-XRD. Nonetheless, SR-XRD can be used as a complementary technique in the study of weathering of meteorites since it confirms and enhances Mössbauer spectroscopy data.

Mössbauer parameters of paramagnetic ferric iron from the coastal site PM are comparable to results from other hot deserts (Sahara, Australia, United States), whereas the range observed in the CD (SJ & OL) is comparable only to data from Antarctic meteorite. These results show that Atacama is unique among hot deserts, with less evolved iron oxyhydroxides.

Acknowledgments

R. B. Scorzelli would like to thank FAPERJ and CNPq for financial support, P. Munayco is grateful to FAPERJ, and J. Munayco is grateful to CLAF/CNPq for his fellowship.

Editorial Handling

Dr. A. J. Timothy Jull

Ancillary