Abstract– Maribo is a new Danish CM chondrite, which fell on January 17, 2009, at 19:08:28 CET. The fall was observed by many eye witnesses and recorded by a surveillance camera, an all sky camera, a few seismic stations, and by meteor radar observatories in Germany. A single fragment of Maribo with a dry weight of 25.8 g was found on March 4, 2009. The coarse-grained components in Maribo include chondrules, fine-grained olivine aggregates, large isolated lithic clasts, metals, and mineral fragments (often olivine), and rare Ca,Al-rich inclusions. The components are typically rimmed by fine-grained dust mantles. The matrix includes abundant dust rimmed fragments of tochilinite with a layered, fishbone-like texture, tochilinite–cronstedtite intergrowths, sulfides, metals, and carbonates often intergrown with tochilinite. The oxygen isotopic composition: (δ17O = −1.27‰; δ18O = 4.96‰; Δ17O = −3.85‰) plots at the edge of the CM field, close to the CCAM line. The very low Δ17O and the presence of unaltered components suggest that Maribo is among the least altered CM chondrites. The bulk chemistry of Maribo is typical of CM chondrites. Trapped noble gases are similar in abundance and isotopic composition to other CM chondrites, stepwise heating data indicating the presence of gas components hosted by presolar diamond and silicon carbide. The organics in Maribo include components also seen in Murchison as well as nitrogen-rich components unique to Maribo.
A bright fireball was seen heading west over the Baltic Sea at 20:08 CET, January 17, 2009 (Fig. 1). The fireball was caught on a surveillance video camera in southern Sweden (Fig. 2), an all sky camera in the Netherlands (Fig. 3), and observed by three German meteor radar observatories (Keuer et al. 2009). Three bright explosions are seen on the video and a glowing trail that persisted about 1 s. The event was also detected by the European Freball network. Due to cloud cover they did not see the trail, but several stations recorded the light curve of the fireball. Explosions were heard a few minutes after the fireball in southern Zealand and on the eastern part of the island of Lolland, where the meteorite fell. Five hundred and fifty eye witness reports were received by the Danish fireball network. Supersonic booms were registered by an infrasound station in southern Germany, two seismometers on the Danish island Sjælland, and nine seismometers in Germany.
As many eyewitnesses on the south coasts of Lolland and Falster reported seeing the fireball descend over the Baltic Sea to the South, it was originally concluded that any meteorite fragments would have fallen in the sea. Despite these observations, Thomas Grau proceeded to try and constrain the fireball trajectory, primarily through direct interviews with eye witnesses from Germany and Denmark. Using a combination of visual and in particular sound observations, he concluded that the fall area was probably near the town Maribo. After 6 days of searching in the suspected fall area, he found a single 30 g fragment of the meteorite on March 4, 2009 (Fig. 4). Following an initial study, the meteorite was classified as a CM2 (Weisberg et al. 2009).
Physical Characteristics of Maribo
Unfortunately, only one walnut-sized fragment of the meteorite was found. The meteorite was found in a cherry plantation where it had penetrated a few centimeters into a grass surface. The vegetation was crushed and had clearly not grown since the impact, consistent with the expectation that this was a fragment from the January 17 fireball. The meteorite appeared intact when found but fell into many pieces when it was touched with a magnet. Surface fragments display a thin (0.1 mm) vesicular fusion crust with numerous fractures. These fractures have probably facilitated the fragmentation of the meteorite when they became filled with water, which subsequently froze. The weather was wet with temperatures around 0 °C throughout most of the 6 weeks during which the meteorite was lying on the ground. The meteorite had therefore probably suffered some terrestrial alteration prior to recovery. Following the initial discovery, extensive efforts were made to find additional fragments. Unfortunately, these were unsuccessful. As required by Danish law, Thomas Grau handed in the original meteorite specimen to the Natural History Museum in Copenhagen. The meteorite was wet when found and had a mass of 30 ± 1 g. When the meteorite was received at the Natural History Museum, it was dry and the weight was 25.8 g. The weight of the largest fragment was 3 g. No further fragmentation was observed while the meteorite dried out.
The Maribo meteorite has been studied with a range of techniques at several different laboratories. Petrography was done in Münster and Copenhagen; oxygen isotopes at the Open University; noble gases in Mainz; cosmogenic radionuclides at the National Labs in Gran Sasso, Italy; bulk chemistry in Cologne (X-ray fluorescence [XRF]) and UCLA (instrumental neutron activation analysis [INAA]); and organics in Munich and at the Open University, UK.
Several thin sections of Maribo were studied by optical and electron microscopy. For optical microscopy in transmitted and reflected light, a ZEISS polarizing microscope (Axiophot) was used. Finer grained mineral assemblages within the meteorite were resolved using a scanning electron microscope (SEM) (JEOL JSM-6610 LV [Münster], Philips XL40 [Copenhagen]). Some mineral analyses were obtained with an energy dispersive system (EDS) attached to the JSM-6610 LV electron microscope at the Institut für Planetologie (ICEM [Interdisciplinary Center for Electron Microscopy and Microanalysis], Westfälische Wilhelms-Universität Münster, Germany). Using the SEM in Münster for quantitative analysis, samples and appropriate mineral standards were measured at an excitation voltage of 20 kV, with beam current stability controlled by a Faraday cup. For these EDS analyses, the INCA analytical program provided by Oxford Instruments was used. Most mineral analyses were obtained using a JEOL 8900 electron microprobe at the Institut für Mineralogie (ICEM) operating at 15 kV and a probe current of 15 nA. Natural and synthetic standards of well-known compositions were used as standards. For water-bearing phases (tochilinites, cronstedtite), oxygen was analyzed quantitatively. The matrix corrections were made by the Φρ(z) procedure of Armstrong (1991).
Oxygen isotope analysis was carried out using an infrared laser fluorination system (Miller et al. 1999). Violent ejection of powders from sample wells can occur when dealing with volatile-rich material such as carbonaceous chondrites. To avoid this problem, all analyses were obtained on whole-rock samples (0.5–2 mg) that were heated progressively in the presence of BrF5 for periods lasting up to 1 h. After fluorination the O2 released was purified by passing it through two cryogenic nitrogen traps and over a bed of heated KBr. O2 was analyzed using a Micromass Prism III dual inlet mass spectrometer. Analytical precision (1σ) based on replicate analyses of international (NBS-28 quartz, UWG-2 garnet) and internal standards is approximately ±0.04‰ for δ17Ο; ±0.08‰ for δ18O; ±0.024‰ for Δ17O (Miller et al. 1999). Carbonaceous chondrites are a heterogeneous mixture of high and low temperature components, consequently precision is less than that obtained on international standards. The quoted precision (1σ) for the analysis of Maribo is based on replicate analyses. Oxygen isotope results are reported in standard δ notation, where δ18O has been calculated as: δ18O = [(18O/16Osample/18O/16Oref) − 1] × 1000 and similarly for δ17O using 17O/16O ratio. For ease of comparison with previous oxygen isotope studies of carbonaceous chondrites (Clayton and Mayeda 1999), Δ17O has been calculated as: Δ17O = δ17O − 0.52δ18O.
Two fragments of approximately 40 mg each were analyzed for content and isotopic composition of noble gases. Analyses were performed on the Mainz noble gas + nitrogen system (Mohapatra et al. 2009). Maribo was analyzed within a measurement series that focused on noble gases, however, and the iridium crucible essential in nitrogen analysis had been replaced by a standard Mo crucible within a tantalum tube oven system (e.g., Schelhaas et al. 1990).
After loading, the samples were heated overnight at 100 °C to remove lightly adsorbed gases. Extraction was in several temperature steps (three for Maribo 1, seven for Maribo 2; see the Results section). Re-extraction at 1900 °C (Maribo 1) and 1800 °C (Maribo 2) produced gases close to the blank level only. Evolved gases were cleaned by exposure to Ti and Al-Zr getters. He/Ne, Ar, Kr, and Xe fractions were separated using activated charcoal and admitted separately to the MAP 215-50 mass spectrometer for measurement using standard analysis procedures (Schelhaas et al. 1990; Schwenzer et al. 2007). Data reported in the tables have been corrected for extraction blank, interferences from HD+/H3+ for 3He, H218O and doubly charged Ar and CO2 for the Ne isotopes, and (negligible in this case) HCl as well hydrocarbon interferences at Ar. Sensitivity and mass discrimination corrections are based on regular analyses of a standard mix of noble gases with (except for He with 3He/4He ∼ 1) atmospheric isotopic composition. Uncertainties in the corrections have been propagated in the reported data.
Neon was partitioned into cosmogenic and trapped components. For the calculation, a value of 0.83 ± 0.03 was assumed for cosmogenic 20Ne/22Ne (e.g., Schultz et al. 1991; Welten et al. 2011). As for trapped neon, while HL-Ne from presolar diamond is dominant, Maribo contains additional trapped Ne components with different isotopic composition (see the Results section); therefore, the composition of trapped neon released at each individual temperature step was estimated from the position in the three-isotope plot of 20Ne/22Ne versus 21Ne/22Ne. Maribo contains more than one trapped component, as most clearly seen in the 1200 °C step of M2, where there is a clear signature of Ne-E. To determine the 20Ne/22Ne trapped ratio, a mixing line drawn from typical cosmogenic Ne (20Ne/22Ne = 0.83; 21Ne/22Ne = 0.90) through the measured data point, was extrapolated to 21Ne/22Ne = 0.03 (the typical value for this ratio in trapped Ne). The uncertainty of the estimate has been propagated in the uncertainty of the reported abundances of trapped 20Ne and cosmogenic 21Ne, but does not significantly add to the error in the totals. No reliable estimate of the shielding parameter for cosmogenic 22Ne/21Ne was possible; however, because of the preponderance of the trapped components. For the same reason, use of 3He (often assumed to be purely cosmogenic) for determination of cosmic-ray exposure (CRE) ages turned out to be problematic (see the Results and Discussion sections). For Ar, Kr, and Xe, cosmogenic contributions were negligible.
X-ray fluorescence analysis was performed at the University of Cologne laboratories using the procedures described in Wolf and Palme (2001). A small piece of Maribo was powdered to a grain size of 10 μm and homogenized in a small agate ball mill. The recovered mass of 114 mg was treated with nitrohydrochloric acid for 2 h at 130 °C to oxidize any metals present. After vaporization of the acid, 3600 mg Li2B4O7 was added as flux. The glass bead was produced by melting at about 1200–1300 °C in a platinum crucible at oxidizing conditions. For the analysis, a Philips PW 2400 sequential wavelength X-ray spectrometer was used. At least 20 geological standard rock samples, prepared in the same way, were used for each element. As discussed in Wolf and Palme (2001) the precision for all elements analyzed is below 1%, accuracies estimated from a comparison with well-determined Allende samples are better than 2%, except for Cr and Ni (better than 3%).
Instrumental Neutron Activation Analysis
A 282 mg sample of Maribo was analyzed for 23 elements by INAA following the general procedures of Huber et al. (2006). Short- and long-lived nuclides were measured during four successive counting periods. Samples were irradiated at the TRIGA Mark I reactor of the University of California, Irvine with a neutron flux of 1.8 × 1012 neutrons cm−2 s−1. Standards included the Allende meteorite (Kallemeyn et al. 1989), the USGS international reference materials granite GSP-1 and BHVO-1, and the Filomena fragment of the North Chile IIAB iron meteorite (Wasson et al. 2007). The uncertainties of the concentrations of different elements can be judged from histograms plotted in fig. 1 of Kallemeyn et al. (1989). For a single analysis, 90% confidence limits are generally approximately 4–7% relative; elements determined with less precision are Sb, Yb, Lu, and Os.
Four small specimens of the CM chondrite Maribo, named DK-CM-A (2.74 g), -B (3.12 g), -C (3.12 g), and -D (0.97 g), have been measured and analyzed by means of nondestructive gamma ray spectrometry. The measurements were performed using high purity Germanium detectors, in ultra low background configuration (25 cm of lead and an inner liner of 5 cm copper, Rn suppression by flushing the closed shielding with ultra pure nitrogen, inside an underground laboratory with 1400 m rock overburden). All four specimens were measured together on two separate occasions firstly 79 days after the fall and then again 113 days after the fall. The counting efficiency was determined with a thoroughly tested Monte-Carlo code based on the CERN library GEANT4 (Agostinelli et al. 2003). The density of the specimens was taken to be that of the CM chondrite average reported in Britt and Consolmagno (2003).
Py-GC×GC-ToFMS (Pyrolysis 2-D Gas Chromatography Time of Flight Mass Spectrometry)
Approximately 2 mg of Maribo (and 2 mg of Murchison for comparison) was crushed using an agate pestle and mortar within a Class 100 clean room at the Open University. There was insufficient sample to risk a solvent wash but the samples were subjected to thermal desorption at 300 °C to remove any superficial contaminants and labile organic species, then exposed to flash pyrolysis at 610 °C. Exact temperatures were recorded for each step using filament resistance (<600 °C) and a photodiode (>600 °C). Sample heating was undertaken using a Pyrola 2000 (Pyrolab, Sweden) for 2 s in a total flow of He at 1 mL min−1. The interface was held at 200 °C and coupled to the injector of a Pegasus 4D (GC×GC) ToFMS (LECO Corporation) maintained at 250 °C. The Pegasus 4D system comprises an Agilent 6890 Gas Chromatograph, a LECO GC×GC module and a Pegasus IV time-of-flight mass spectrometer (ToFMS).
GC×GC separation utilized a primary (first dimension) nonpolar column (SGE BPX5, 30 m × 250 μm × 0.25 μm) and a secondary (second dimension) polar column (SGE BPX50, 1.5 m × 100 μm × 0.1 μm). The secondary column was installed in its own oven inside the main GC oven. The carrier gas was helium (99.9999%) supplied at a column flow rate of 1 mL min−1. Samples were analyzed under the following GC oven program: The primary GC oven was held at 35 °C for 1 min and then raised to 300 °C at a rate of 5° min−1 and held for 5 min. The secondary oven was offset +15 °C from the primary oven temperature. A liquid N cooled modulator was used with a temperature offset of +30 °C above the primary oven temperature for modulation on the second column. Modulation frequency was 4 s. Acquisition rate was 133 spectra s−1 for a mass range of 33–400 amu. Data processing was performed on LECO ChromaTOF software.
ESI(−)-ICR/FTMS (Negative Ion Mode ElectroSpray Ionization Ion Cyclotron Resonance Fourier Transform Mass Spectrometry)
Small fragments of a total weight of about 15–20 mg were first washed by stirring for a few seconds with methanol (LC/MS CHROMASOLV grade Methanol from FLUKA) prior to crushing with an agate pillar in 1 mL methanol poured into the corresponding agate mortar. This procedure was shown to limit the number of signals resulting from terrestrial and human contamination, for example, fatty acids stemming accidentally through sample handling. The mixture (suspension) was transferred into a vial and submitted to an ultrasonic cleaning for not more than 10 min, then centrifuged. The supernatant liquid was readily removed with a microsyringe, ready for flow injection into the electrospray ionization-source (ESI-source).
The samples were introduced into the microelectrospray source at a flow rate of 120 μL h−1 with a nebulizer gas pressure of 20 psi (138 kPa) and a drying gas pressure of 15 psi (103 kPa) at 200 °C (Agilent sprayer). Spectra were first externally calibrated on clusters of arginine (1 mg L−1 in methanol) and internal calibration was systematically achieved using fatty acids reaching accuracy values lower than 0.05 ppm in routine day-to-day measurements. Similar to traditional mass spectrometers, the FTMS determines the mass to charge ratio (m/z) of the ions analyzed. The spectra were acquired with a time domain of 4 million data points in the mass range of 100–2000 m/z for a peak resolution >500,000 at m/z 400, reaching 1,000,000 at m/z 250. A total of 5000 scans were accumulated for each sample. The Fourier transform ion cyclotron resonance spectra were exported to peak lists at a signal to noise ratio (S/N) of two as commonly used for ESI(−)-ICR/FTMS. The high mass accuracy allows elemental formulas to be calculated for each signal in batch mode by a custom-written software tool (Hertkorn et al. 2007). In conjunction with an automated theoretical isotope pattern comparison, the generated formulas were validated by setting sensible chemical constraints (N rule, O/C ratio ≤ 1, H/C ratio ≤ 2n + 2, element counts: C ≤ 100, H ≤ 200, O ≤ 80, N ≤ 3, S ≤ 3) and mass accuracy window (set here at ±0.2 ppm). Final formulas were generated and classified into groups containing CHO, CHON, CHOS, or CHONS illustrated in O/C versus H/C van Krevelen diagrams (Hertkorn et al. 2008; Schmitt-Kopplin et al. 2010b). Based on these groups of m/z ratios, specific mass spectra could be reconstructed.
In thin section, the Maribo meteorite displays a well-preserved fusion crust (Fig. 5a), and consists of various coarse-grained components embedded in an opaque, fine-grained matrix (Figs. 5–7). Coarser-grained components include chondrules, fine-grained olivine aggregates, large isolated lithic and mineral fragments (often olivine), rare Ca,Al-rich inclusions (CAIs), cronstedtite-rich rounded objects, and compact tochilinite-rich objects. Cronstedtite and tochilinite are minerals that form during aqueous alteration. We are using the term “compact tochilinite-rich objects” here to describe larger objects that are almost entirely composed of tochilinite. These objects appear to have survived in the original shape and form in contrast to the tochilinite fragments described below. The coarser-grained components are typically rimmed by fine-grained dust mantles (Figs. 5–7), similar to those described by Metzler et al. (1992) and Metzler and Bischoff (1996). Metals occur (as a minor constituent) in chondrules and matrix. The matrix includes abundant dust rimmed fragments of tochilinite with a layered, fishbone-like texture; tochilinite–cronstedtite intergrowths and sulfides, metals, and carbonates often intergrown with tochilinite.
Chondrules and CAIs
The abundance of chondrules is low, certainly below 10 vol%. Chondrule diameters are generally <0.4 mm with a few up to 0.9 mm. Some chondrules have been partially altered to tochilinite and cronstedtite (Fig. 5e). Although the chondrules contain cronstedtite and tochilinite, they also contain metals and sulfides (Fig. 5). Several olivine grains having a shape like chondrules (Fig. 5b) have been observed as well as zoned olivine fragments. All are typically rimmed by fine-grained dust rims (Fig. 5b). Most olivines in the chondrules and other coarse-grained components (chondrule fragments) have Fa < 3 mole%, but some Fe-rich olivines have been found (Fa up to 82 mole%).
Several CAIs were detected in Maribo. They are all fine-grained. Two typical inclusions are shown in Fig. 6. The most abundant phase within these inclusions is spinel (MgAl2O4). Several inclusions with blue hibonite (CaAl12O19) were observed. Mar-20b (Fig. 6a) is a 125 × 65 μm hibonite–spinel inclusion dominated by elongated to subhedral hibonites with variable compositions: Their TiO2-concentration varies from 3 to 8 wt% (Table 1). The CAIs are rimmed by a fine-grained, Fe-rich (probably water-bearing) phase and occasionally Ca-pyroxene (Fig. 6b). Perovskite (CaTiO3) only occurs as a minor constituent of the CAIs (Fig. 6). So far, melilite and fassaite have not been found.
Table 1. Chemical composition of phases in Ca,Al-rich inclusions Mar-20b and Mar-39 (Fig. 6).
Note: All data in wt%; n.d. = not detected. The analyzed phases are hibonite (Hib), spinel (Sp), and Ca-pyroxene (Cpx).
The porous aggregates are irregular in shape and measure a couple of hundred micrometers across (Fig. 5c). They contain zoned grains of olivine and pyroxene, tochilinite, cronstedtite, pyrrhotite, and pentlandite.
Tochilinites have been found in two varieties: (1) as compact tochilinite objects (Figs. 7a and 7c) and (2) as fishbone-like aggregates (Fig. 7b). Both varieties have similar O- and S-concentrations. However, the data also clearly show that both types of tochilinite have distinct compositions. The compact tochilinites are richer in Ni and have lower Fe-concentrations compared to the fishbone-like tochilinites. We also found significant differences in the P and Cr compositions of the two phases. While the compact tochilinites have mean P- and Cr-concentrations of 0.76 and 1.29 wt%, respectively, these elements are not or only barely detectable within the fishbone-like tochilinites (Table 2). Mar-21 (Fig. 7c) is a 60 μm diameter round, mineralogically zoned, tochilinite-rich object. It consists of a core of compact tochilinite and a Fe-,Ni-,S-rich phase (in the fragmented area) and a cronstedtite-like rim. Between these two components a Fe-rich layer (perhaps magnetite) is present. The chemical compositions of its constituents are given in Table 2.
Table 2. Chemical composition of tochilinites, cronstedtites, and phases of objects Mar-21 and Mar-39.
Note: All data in wt%; oxygen has been quantitatively analyzed. n.d. = not detected. Except for the cronstedtites in the CAI Mar-39 all analyses are from Mar-21. Phases in Mar-21: A = Compact tochilinite; B = Cronstedtite-like phase; C = Fe-rich layer; D = Fe-,Ni-,S-rich phase (see Fig. 7c).
Cronstedtites also occur in different parageneses. They occur (1) within or at the boundaries of larger objects (like chondrules; Fig. 5e), (2) as a major constituent intergrown with tochilinite within small matrix objects, (3) within chondrule-like large objects with sizes up to several hundred micrometers (Fig. 7d), and (4) intergrown with CAIs (Fig. 6b). The average and range of compositions of cronstedtite are given in Table 2.
Using Stöffler et al.’s (1991) shock classification scheme, Maribo is very weakly shocked (S2) as indicated by only minor undulatory extinction and the lack of planar deformation features in olivine.
Oxygen isotope results for Maribo are given in Table 3 and plotted in Fig. 8. Also shown in Fig. 8 are previous analyses of CM2 chondrites (Clayton and Mayeda 1999) as well as analyses for other carbonaceous chondrite groups (Clayton and Mayeda 1999; Greenwood and Franchi 2004). Our analysis of Maribo plots close to the carbonaceous chondrite anhydrous mineral (CCAM) line of Clayton and Mayeda (1999) at the 16O-rich end of the CM2 field. The proximity of Maribo to the CCAM line and its position far below the terrestrial fractionation line (TFL) may indicate that it has experienced less parent body alteration than CM2 samples that are displaced toward the TFL, such as the heavily altered CM2 fall Sayama (Fig. 8) (Takaoka et al. 2001). Compared to the recently described Paris CM chondrite, Maribo is displaced slightly away from the CCAM line and also further from the TFL line. Paris is of special interest as it shows only minor evidence of aqueous alteration (Bourot-Denise et al. 2010; Zanda et al. 2010).
Table 3. Oxygen isotopic composition of Maribo.
The abundances of trapped gases in Maribo (Tables 4 and 5) are typical of CM chondrites, rather similar to Murray and Murchison. The isotopic compositions show that—as is the case for other CMs—the HF/HCl resistant Q-phase (Lewis et al. 1975; Reynolds et al. 1978) hosts most of Ar, Kr, and Xe, whereas diamond is the major host phase for Ne. The noble gas data are interpreted in terms of the several previously identified noble gas phases. These phases are the HL-phase (a component enriched in Heavy and Light Xe isotopes with interstellar diamonds as the host phase), P3 phase (an isotopically normal Xe-component), E-phase (a component composed of almost pure 22Ne with silicon carbide [SiC] as the host phase), the Q-phase (a characteristic component with unknown host phase), and the G-phase (a Xe-component enriched in isotopes produced solely by S-processes).
Table 4. Helium and neon abundances and isotopic compositions of Maribo samples for each extraction temperature.
Extr. temp. (°C)
Note: Concentrations are in units 10−8 cm3 STP g−1. Errors in the last digits are given in parentheses. Also listed are the abundances of trapped (tr) 20Ne and cosmogenic (cos) 21Ne.
Maribo 141.33 mg
Maribo 240.96 mg
Table 5. Argon, krypton, and xenon abundances of Maribo samples for each extraction temperature.
Extr. temp. (°C)
Note: Concentrations are in units 10−8 cm3 STP g−1. Errors in the last digits are given in parentheses. Also listed is the isotopic composition of Ar and the Xe isotopic ratios 129Xe/132Xe, 130Xe/132Xe, and 136Xe/132Xe.
Maribo 141.33 mg
Maribo 240.96 mg
Argon, Krypton, and Xenon
Ar, Kr, and Xe are dominated by Q-gases (Lewis et al. 1975; Ott 2002). Figure 9 shows that Xe in Maribo (Table 5) plots near the Q composition. Only the lowest temperature points are slightly shifted toward air, showing a little contamination. At some intermediate temperature steps (800 and 1000 °C), there is a shift to the right, indicating the presence of HL-Xe from diamond. The 1200 °C release, on the other hand, shows a very small shift toward the upper left, consistent with a small contribution from s-process Xe carried by presolar silicon carbide (Lewis et al. 1994).
Helium and Neon
While there is evidence for solar neon released at the lowest temperatures, Ne overall is dominated by the HL component in presolar diamond (Lewis et al. 1987; Ott 2002) as shown in the Ne three-isotope plot of Fig. 10. The presence of presolar SiC is also evident, indicated by the dip toward Neon-E (Black and Pepin 1969; Black 1972) in 20Ne/22Ne (Tang and Anders 1988; Ott 2002) at 1800 °C for Maribo 1 and at 1200 °C in the high-resolution analysis of Maribo 2. The ratio of 4He to trapped 20Ne of approximately 235 (Table 5) is lower than the “primitive” value for the HL component of approximately 600 (Huss and Lewis 1994; Ott 2002)— and also lower than that for the P3 component in diamond—and is more akin to the “processed” composition of Huss and Lewis (1994). This leaves little room for any radiogenic 4He, which cannot be used for chronological purposes, therefore. Similarly, as a significant fraction (most likely more than half) of 3He must be of trapped origin, 3He cannot be usefully applied in the determination of a CRE age.
The isotopic excursions in Xe (Fig. 9) and Ne (Fig. 10) allow an estimate of the abundances of presolar diamond and SiC. Using the approach of Huss and Lewis (1995) and Huss et al. (2003) with a “constant” Ne-E concentration of SiC, from both the 1800 °C data point for Maribo 1 and the 1200 °C for Maribo 2, an abundance of approximately 10 ppm is inferred. This is slightly lower than the approximately 13 ppm estimated for Murray and Murchison (Huss et al. 2003). A similar estimate for presolar diamond, assuming a Xe-HL abundance typical of diamonds from CM2 meteorites (Huss et al. 2003), gives approximately 500 ppm. This is again slightly lower than the abundances estimated for Murray (614 ppm) and Murchison (740 ppm) (Huss et al. 2003). The abundances of diamond and SiC determined for Maribo are lower limits, since more may be hiding in the isotopically more “normal” steps.
With a significant contribution to 4He from trapped He, it is not possible to calculate a U/Th-He age. Nominal K-Ar ages, based on the 40Ar abundances in Table 5 (probably containing a significant air contamination) and a K abundance of 372 ppm (Table 6), are 1.14 and 1.37 Ga. The air contamination component is only significant in the first release temperature (Fig. 9). Avoiding the potentially contaminated 400 °C release for M2 decreases the age from 1.37 to 0.88 Ga.
Table 6. Bulk composition of Maribo based on instrumental neutron activation analysis.
Not only are cosmogenic contributions in Ar, Kr, and Xe swamped by trapped gas, but—as noted above—also 3He must contain a substantial trapped contribution. Therefore, we have calculated CRE ages following Eugster (1988) based on 21Ne only. However, these also suffer from the high abundance of trapped Ne. The problem is that a reliable determination of the shielding parameter (22Ne/21Ne)cos is not possible. Assuming “average shielding” (shielding parameter = 1.11) (Nishiizumi et al. 1980) leads to CRE ages of 0.69 Ma (Maribo 1) and 0.63 Ma (Maribo 2). An alternative approach can be based on the inference from the radionuclides that the pre-atmospheric radius must have been between 10 and 20 cm. With CM chemical composition (Tables 6 and 7) (Wasson and Kallemeyn 1988), 21Ne production rates for meteoroids in this size range (assuming a pre-atmospheric depth of at least 2 cm) lie between approximately 0.13 and approximately 0.21 × 10−8 cm³ STP per gram and Ma, according to the calculations of Leya and Masarik (2009). With the abundances of cosmogenic 21Ne as listed in Table 4, this translates into a possible age range of 0.8–1.4 Ma.
Table 7. Bulk composition of Maribo based on X-ray fluorescence analysis.
In Table 8 the measured activity concentrations for the detected short- and medium-lived cosmogenic radionuclides (7Be, 22Na, 26Al, 44Ti, 46Sc, 51Cr, 54Mn, 56Co, 57Co, 58Co, and 60Co) are given. The given count rates are already calculated back to the activities at the date of fall following the simple decay law, and taking into account the time that passed between the fall of the meteorite and its measurement. The fact that 51Cr (half-life 27.703 days) is seen in the first measurement period, whereas it is below the detection threshold in the second, clearly supports that the date of fall is coincident with the observed fireball. To derive an approximate size of the meteorite body, the data of 60Co, 54Mn, and 22Na were used. Normalizing the 60Co data to a concentration of its main target Co of 700 ppm, the resulting specific activity was compared to the calculations of Eberhardt et al. (1963) and Spergel et al. (1986). This comparison gave a possible range in the radius of (20–25) cm. The 22Na data were normalized to the main targets Si and Mg. Then, data were compared to the calculations of Bhandari et al. (1993) for H chondrites normalizing the Si and Mg abundances of the CM chondrite to the average ones of an H chondrite. The resulting possible range in the radius is 10–20 cm. Finally, the data of 54Mn have been normalized to the concentration of its main target Fe. Then, data were eventually compared to the calculations of Kohman and Bender (1967), giving a range for the radius of 13–20 cm. Combining the results of all three radionuclides, one gets a possible conservative size interval for a spherical meteoroid from 10 to 20 cm radius, corresponding to about 9–74 kg of mass of the parent body. If we assume the production rates in C chondrites for 26Al to be that of Leya and Masarik (2009), the expected saturation values for the above determined size range are 30 and 37 dpm kg−1, respectively, for the radii 10 and 20 cm. Thus, taking the measured value for 26Al of 33 ± 3 dpm kg−1, the exposure age is × 106 a. Within the rather large uncertainties, the value for the exposure age found through 26Al is in agreement with the range for the exposure age determined through Neon. Nevertheless, as pointed out in Wieler (2002) and the references reported therein, it cannot be excluded that the 26Al levels might be higher due to solar cosmic-ray (SCR) production and therefore show a higher exposure age with respect to Neon, for which no firm evidence of SCR contributions in chondrites could be found yet.
Table 8. Detected cosmogenic radionuclides in the Maribo specimens. The table gives the specific activity at the time of fall, already taking into account the time that passed between the fall of the meteorite and its measurement.
Activity concentration (dpm kg−1)
Measured 79 days after fall
Measured 113 days after fall
Note: Reported uncertainties in the last digits (in parentheses) are combined standard uncertainties. The table gives the specific activity at the time of fall. All decay data of the radionuclides is taken from Nucleide (2011).
The activities of the short-lived radioisotopes, with half-life less than the orbital period, are representing, instead, the production integrated over the last segment of the orbit. The fall of the Maribo CM chondrite occurred during the solar minimum that is preceding the next-to-come solar cycle 24 as indicated by the publicly available neutron monitor data (Bartol 2011). As in the case of the Puerto Lápice eucrite (Llorca et al. 2009) the galactic cosmic ray flux was therefore high in the 6 months before the fall. The activity for the very short-lived radionuclides 46Sc and 54Mn is expected to be maximal following the sunspot number as reported in Llorca et al. (2009) and the references cited therein. This is confirmed by our measured values for 46Sc and 54Mn, normalized to their primary targets Fe and Ni.
The values measured for 7Be are in the range of the values that have been measured for ordinary chondrites (see Evans et al. 1982). The concentrations of the natural radionuclides 232Th and 238U as well as for K in the meteorite specimens are listed in Table 9. They are well in accordance with the average concentrations given in Wasson and Kallemeyn (1988), for CM chondrites. The value measured for potassium is somewhat smaller than what was determined by INAA (see Table 6). Nevertheless, both values agree within uncertainties.
Table 9. Naturally occuring nuclides Th, U, and K.
Concentrations (ng g−1)
Note: The reported uncertainties in the last digits (in parentheses) are standard uncertainties.
275 (24) × 103
Instrumental Neutron Activation Analysis
The INAA results are shown in Table 6 and Fig. 11 together with the composition of CM chondrites earlier analyzed at UCLA by Kallemeyn and Wasson (1981). With the exception of Na, the differences between Maribo and mean CM concentrations are the same within errors.
Table 7 shows the bulk chemistry of Maribo and Murchison based on XRF. Maribo has the typical Si/Mg ratio of carbonaceous chondrites. Element ratios that are variable in different types of carbonaceous chondrites are all within the CM-range (Ca/Mg, Fe/Mg, Cr/Mg, and Mn/Mg). The bulk chemistry of Murchison, the largest CM chondrite, is also very similar to the bulk chemistry of Maribo.
The total ion chromatogram of the newly recovered meteorite strongly resembles that of Murchison (Fig. 12). The trace is dominated by one-ring aromatic species, including benzene and up to C7-alkyl benzenes. There are traces of up to C13n-alkyl benzenes; however, their indigeneity is questionable. Naphthalene is also dominant, and up to C4-naphthalenes are identifiable within the pyrolysates. Molecular weight range extends as far as the four-ring species fluoranthene and pyrene. Volatile species separated using this technique include thiophene, acetone, low molecular weight ketones, and short-chain hydrocarbons. Both Murchison and Maribo show evidence within their pyrolysates of S, N, and O heterocyclic compounds including benzothiophenes and dibenzothiophenes, benzonitriles and methylbenzonitriles, benzaldehydes and dibenzofurans. As in Murchison, there is evidence in the pyrolysate of Maribo of branched and straight chain alkanes and alkenes, with the alkanes at least extending to C20.
Overall, the ESI(−)-ICR/FTMS spectra strongly resembles those of Murchison described earlier in Schmitt-Kopplin et al. (2010a), expressing a high organic chemical diversity of ten thousands of signals with more than 50 signals in each nominal mass (Figs. 13a–d). This is confirmed with very similar van Krevelen diagram representation of the CHNOS converted mass values (Figs. 13f and 13h). However, Maribo showed a distribution in m/z with higher abundances in the lower mass values compared to Murchison (Figs. 13a and 13c) and the higher signal intensities in masses of lower mass defect are representative of more aromatic type structures (Figs. 13b and 13d). Also, in the same experimental analysis conditions Murchison and Maribo only showed 8700 common signals; Murchison had 16,700 unique and Maribo 15,600 unique signals. Nitrogen containing compounds (both CHNO and CHNOS) of lower oxygen content and H/C ratio around one typical of heterocycles are more abundant in Maribo (see van Krevelens Figs. 13f versus 13h) than in Murchison and any carbonaceous chondrite analyzed with ICR-FT/MS so far (data not shown).
The oxygen isotopic composition, petrology, exposure age, trapped gas inventory, and bulk chemistry of Maribo are all very similar to other CM2 carbonaceous chondrites. The most unusual properties of Maribo are the abundant fishbone-like aggregates described in the petrology section and the unusual low metamorphic grade.
Nebula Versus Parent Body Alteration
CM chondrites have, as a group, been exposed to diverse levels of aqueous alteration (Brearley  and references therein). Some of the variation may be attributed to brecciation, but in situ variation is also observed (Metzler et al. 1992). Many previous authors favor a model where the aqueous alteration observed in CM chondrites took place in a single parent body (e.g., Kerridge and Bunch 1979; Rubin et al. 2007) but the heterogeneity has prompted others to argue for more complex models where at least parts of the alteration took place in the nebula and/or in a previous precursor parent body (e.g., Metzler et al. 1992).
Maribo hosts a selection of components which may potentially be used to infer the setting of the aqueous alteration that CM chondrites were exposed to. Many materials in Maribo are surrounded by compact, fine-grained rims. We find the rims around chondrules (Fig. 5e), unaltered metal grains (Fig. 5f), isolated olivine grains (Fig. 5b), porous aggregates containing variable abundances of water-bearing minerals (Figs. 5c and 5d), compact tochilinite objects (Figs. 7a and 7c), and around the fishbone-like aggregates (Fig. 7b). Tochilinite objects surrounded by rims have also been observed in other CM chondrites (Metzler et al. 1992). Whatever the nature of the process that formed the rims, it seems to be unrelated to the object they surround since the rims show no signs of zoning and since different types of objects such as metal and tochilinite have very similar rims.
Metzler et al. (1992) argued that these fine-grained zones are accretionary rims. If correct, this implies that the alteration products were exposed to a dusty environment in the nebula some time after they formed. Collisions of dust mantled chondrules were studied numerically by Ormel et al. (2008), who found that rims could facilitate sticking of chondrules and result in compression of the rims; their model yielded porosities as low as 67%, and they suggested that some nebular processes might yield still lower porosities. Trigo-Rodriguez et al. (2006) argued that the rims are too compact to have formed by single-grain accretion in the nebula (which is expected to produce porosities in the 50–90% range; Blum and Wurm 2008) and suggested that the low porosity is produced by multiple impact compaction events on the parent body, with greater degrees of compaction near incompressible materials such as chondrules. In their model, most of the compaction occurred before aqueous alteration. Rims around tochilinite objects such as the one shown in Fig. 7a are not easily explained with this model. It would require that the precursor of the tochilinite, presumably kamacite, was there when the compaction occurred, and that penetration of water through the permeable rim was responsible for the tochilinite formation.
Thus, the rims observed in Maribo and other CM chondrites may either have formed during asteroid compaction or as accretionary rims in the nebula. As asteroid compaction must have played a role, it is also possible that the observed features are a result of a combination of the two processes. Asteroid compaction is consistent with in situ aqueous alteration, whereas accretionary rims require that the altered products were exposed to a dusty environment in the nebula after they formed.
If the aqueous alteration took place in the parent body, should we then expect the aqueous alteration to be homogenous? Variable abundances of ice in the accreting material and variable distribution of porous channels that would allow the water to flow through the rock could conceivably result in an inhomogeneous distribution of alteration features. Local dissolution of various phases could also result in spatial variation of the water composition which would also, likely, result in inhomogenous alteration features (Brearley 2006). However, the observed variations are very large on a small spatial scale. We do find metal grains with little or no signs of alteration in contact with or within a few micrometers from heavily altered chondrules and cronstedtite/tochilinite objects (Figs. 5e and 5f). As there are no signs of brecciation and because we find it unlikely that the water abundance was variable on a micrometer scale, we infer that all of the objects must have been exposed to the same degree of parent body alteration. The existence of unaltered metal grains therefore implies that there was very little in situ aqueous alteration on the Maribo parent body. This is in contrast to the traditional view that CM chondrites experienced significant aqueous alteration on the parent body. If correct, it implies that the tochilinite and cronstedtite aggregates predate accretion of the Maribo parent body.
Objects very similar in structure to the fishbone-like aggregates have also been found in Acfer 094 (Sakamoto et al. 2007). In Acfer 094, the aggregates were found associated with tochilinite assemblages very similar in composition and structure to the compact tochilinites found in Maribo (Figs. 7a and 7c). The compact tochilinite structures in Acfer 094 are enriched in heavy oxygen isotopes (δ17,18OSMOW = 180 per mil) and Sakamoto et al. (2007) inferred that these inclusions must have formed in the nebula.
Earlier, Bischoff (1998) discussed the possibility that aqueous alteration occurred on precursor planetesimals, which produced water-bearing phases and were destroyed. He suggested that the products (fine-grained dust of water-bearing phases and compact objects of water-bearing phases) accreted in second-generation parent bodies. Such a process where aqueously altered products from an early generation parent body are mixed with unaltered components in the nebula, before accreting to form the Maribo parent body, may account for the observations in Maribo. The proposed first-generation parent body would have had a high volatile abundance and a higher 26Al abundance. Such a volatile-rich body, undergoing rapid heating, could be destroyed in a runaway process or the altered phases could have been ejected through gas vents (Wilson et al. 1999).
Metamorphic Grade of Maribo
The diverse degree of alteration features observed within Maribo makes it inherently difficult to assign a bulk metamorphic grade to the meteorite. Rubin et al. (2007) suggested that the oxygen isotopic composition and chemical composition of PCPs may be used to determine the metamorphic grade of CM chondrites. PCPs are clumps of serpentine-tochilinite intergrowths (also called “poorly characterized phases” or PCP).
Rubin et al. (2007) argue that the degree of alteration is negatively correlated with Δ17O—as the anhydrous components in Murchison have a Δ17O of −5.24, whereas the hydrated materials have a Δ17O of −1.87. The bulk Δ17O of the 11 CMs in Rubin et al. (2007) ranges from −3.07 (Murray) to −2.00 (Nogoya). In comparison, Maribo has a Δ17O of −3.85 which suggests that Maribo is less altered than any of the CMs studied by Rubin et al. (2007). Of the 34 CM chondrites analyzed by Clayton and Mayeda (1999), only two had Δ17O lower than Maribo. Clayton and Mayeda (1999) found a Δ17O of −4.16 for Asuka-881334 and −4.14 for Yamato-86720. The recently described CM chondrite Paris was classified as a type 3.0 ± 0.1 (Zanda et al. 2010). Paris has a 17ΔO of −3.11 suggesting that Maribo is equally unaltered. It is, however, important to note that Paris and Maribo contain altered phases and therefore may not qualify as type 3.0. It should also be noted that Rubin et al. (2007) only found a weak inverse correlation between Δ17O and petrologic type and that they only included meteorites from petrologic type 2.0 to 2.5.
Rubin et al. (2007) also found a correlation between the metamorphic grade and S/SiO2 and FeO/SiO2 in PCPs. This technique cannot be applied to Maribo because the typical PCP-objects found in other CM chondrites are rare in Maribo. In Maribo, tochilinites and cronstedtites are much coarser-grained than we have seen in any other CM chondrite. Because there was no need to analyze fine-grained tochilinite–cronstedtite intergrowths, individual phases were analyzed. The compositions of these phases are given in Table 2. Typically, the fishbone-like tochilinites occur in close relationship to cronstedtite. We could consider such coarse-grained tochilinite–cronstedtite aggregates as “PCP-analogs.” If we suggest a specific tochilinite to cronstedtite ratio, we can calculate the “FeO”/SiO2 and S/SiO2 for such aggregates by using the data from Table 2 (“FeO”/SiO2- and S/SiO2-values for all other tochilinite–cronstedtite ratios can be calculated from this table). If we consider an aggregate having 50% tochilinite and 50% cronstedtite, we would obtain “FeO”/SiO2- and S/SiO2-ratios of 4.56 and 0.69, respectively (Table 10). These values are much higher than the values published by Rubin et al. (2007) for the least altered CM chondrite, but similar to the ratios found in the new CM chondrite Paris (Bourot-Denise et al. 2010; Zanda et al. 2010). Even if we consider an aggregate having only 20% tochilinite and 80% cronstedtite, we obtain results similar to those for the least altered CM chondrites of Rubin et al. (2007) (Table 10).
Table 10. Estimated compositions of tochilinite–cronstedtite.
The distribution of solvent soluble organics species in Maribo also suggests that it experienced less alteration than Murchison. Chemical degradation would mainly lower aromaticity by breaking double bonds and lead to compounds having a more aliphatic character. In addition, it was shown with Murchison that aqueous alteration and reactivity with the mineral phase increased the amount of sulfur-bearing compounds as CHOS and CHNOS lowering consequently the amount of CHO- and CHNO-type of compounds and shifting the mass distribution toward higher mass compounds. Because Maribo involves more rather small and nitrogen-rich aromatic molecules than Murchison it can be hypothesized that Maribo may contain a more primitive type of organic composition that did not yet undergo such aqueous alterations as did Murchison. Another way to explain the unusually high nitrogen concentration in Maribo is to consider that additional nitrogens generally increase the hydrolysis rate of aromatic heterocycles (Reusch 2010). This could explain a more extended terrestrial alteration by water (and a consequent nitrogen depletion) of Murchison (old fall) than of Maribo (fresh fall), assuming both meteorites had initially a similar N content in their parent body. Fuchs et al. (1973) reported that several of the Murchison samples lost and even gained weight while being studied in different laboratories where the atmospheric conditions changed (presumably humidity). They suggested that the porous samples “breathed” and thus exchanged air with the surroundings.
The combined evidence from oxygen isotopes, chemical composition of tochilinite and cronstedtite, and organic chemistry suggest that Maribo is one of the most primitive, least altered CM chondrites.
Sodium Depletion in CM Chondrites
Maribo, Murray, and Yamato-74662 are depleted in Na relative to average CMs (Fig. 11). We suspect that this may be due to exposure to water on the Earth before the meteorites were recovered. Maribo was exposed to rain and snow for 6 weeks before it was recovered. A similar effect may be responsible for the low Na concentration found in the fall Murray, which was also exposed to rain for weeks prior to recovery. The search for Murray was initiated 32 days after it fell and intermittent rain hindered the search (Horan 1953). It is also entirely possible that Yamato Mountains-74662 was exposed to liquid water in Antarctica.
Among the other seven CMs in Fig. 11 that do not show depletion of Na, only Allan Hills 77306 is known to be a find. Cochabamba is officially a find, but it was found in the meteorite collection in Vienna labeled Cochabamba, Chile, without any information about the recovery (Kurat and Kracher 1975). Therefore, it is possible that Cochabamba is a fall, which would be consistent with its normal Na concentration.
Another possibility is that the depletion of Na is a nebular feature reflecting inhomogeneous distribution within the CM parent body as suggested by Spettel et al. (1978). Samples from the Murchison meteorite shower show large variations in their Na contents (Fuchs et al. 1973). Fuchs et al. (1973) found that Na2O varied from 0.19% to 0.71% (corresponding to 1.3–4.9 mg g−1 Na) among five different interior samples from five different fragments of Murchison. The variation among different Murchison fragments is therefore almost the same as the variation seen among the CMs (0.7–4.2 mg g−1). Although we have no information on how and when the different samples of Murchison were collected, the dry Australian climate and the fact that all samples were taken from the interior makes it highly unlikely that rainwater played a role. This supports the idea that the Na variation is not due to exposure to water on Earth but reflects a true variation.
Also, the soluble organic fraction in Maribo did not show features of pronounced aqueous alteration suggesting that terrestrial water alteration was minimized.
The Na variation among CM chondrites is therefore likely a nebular feature although exposure to rainwater may have been a contributing factor.
Conclusions and Summary
Maribo is a new CM fall. The meteorite was found as a result of dedicated search following a fireball observation 6 weeks earlier. The abundance of short-lived cosmogenic radionuclides shows that the fall of the meteorite coincides with the observation of the fireball. We therefore conclude that the meteorite did fall January 17, 2009 when the fireball was observed. While Maribo mainly contains the same components as other CM chondrites and the oxygen isotopes fall within the CM field it also has several unusual properties. Abundant tochilinite- and cronstedtite-rich objects with a fishbone-like texture and surrounded by fine-grained rims are omnipresent in the matrix of the meteorite. These objects are often found in close proximity to metal particles that show no signs of aqueous alteration. The oxygen isotopic composition of Maribo has an unusual low Δ17O suggesting that Maribo is among the least aqueously altered CM chondrites. This is supported by the analysis of the solvent extractable fraction that showed rather small-sized aromatic and nitrogen-rich nondegraded organic compounds. Based on the highly inhomogenous distribution of aqueous alteration features and the presence of fine-grained rims on altered phases, we conclude that the aqueous alteration did not take place on the Maribo parent body but rather on an earlier parent body which was subsequently destroyed.
Acknowledgments–– This project was supported by funds from the Danish National Research Foundation. We also thank Ulla Heitmann (Münster) for thin section preparation. Detailed and constructive reviews from Mike Zolensky and Eric Palmer are gratefully acknowledged.