Nature and timing of a significant reduction event on the L‐chondrite parent asteroid

About 17% of L6 chondrites (15/87) show significant reduction features in BSE images in thin section. Because some thin sections of these meteorites do not show reduction features, this percentage is a lower limit. Reduction features include: (1) 4–5‐μm‐thick BSE‐dark reduction rims on olivine and orthopyroxene grains and along fracture boundaries in these grains, (2) 4–12‐μm‐thick dark bands (probably poorly crystalline pyrrhotite) at the margins and along fractures in troilite grains, and (3) 2–5‐μm‐thick dark rinds of kamacite around some taenite grains. Only one of 70 L‐group chondrites (1.4%) of lower petrologic type exhibits minor reduction. The L6 chondrites showing major reduction have 40Ar/39Ar plateau ages ranging from 156 ± 1 Ma for Guangnan to 4543 ± 3 Ma for Thamaniyat Ajras. Reduction occurred after silicate, sulfide, and metal grains had attained their present sizes during parent‐body thermal metamorphism (and had been fractured by parent‐body collisions). The precise plateau age of Thamaniyat Ajras probably marks the timing of the L6 reduction event. It seems likely the reductant was a low‐viscosity fluid, plausibly CO, derived from oxidation of poorly graphitized and amorphous carbon within fine‐grained matrix. Water‐ice that had accreted to the L‐chondrite asteroid was heated and mobilized during metamorphism, causing oxidation. After peak metamorphism, ~75% of the water had been used up or lost; the remaining water facilitated continuing graphite oxidation so that, after this point, overall reduction effects exceeded those of oxidation. L chondrites of lower petrologic type were less affected by reduction due to their lower metamorphic temperatures.

Modification processes can operate sequentially.Some CM chondrites were aqueously altered and subsequently heated to temperatures of 200 to >750°C.These rocks underwent dehydration, loss of bulk volatile elements, and graphitization of organic matter (e.g., Lindgren et al., 2020;Nakamura, 2005;Quirico et al., 2018;Tonui et al., 2014).Because such heating occurred less than 2 billion years ago (Amsellem et al., 2020), the heating must have been caused by collisions.
As shown by Wasson et al. (1993), a few OCs experienced extensive whole-rock reduction; this process resulted in appreciably lower olivine Fa contents than in normal members of their respective compositional groups (Rubin, 1990): for example, H4 Cerro los Calvos: Fa 12.8 AE 0.6, n = 50 (cf.H4-6 chondrites: Fa 17.3-20.2);L3 Suwahib (Buwah): Fa 14.4 AE 1.6, n = 77 (cf.L4-6 chondrites: Fa 23.0-25.8).These reduced OCs were metamorphosed on their parent asteroids, either in the presence of a reducing agent or as boulders within a reduced host regolith.If reducing agents were indeed present on OC parent bodies, some OCs might also show more subtle evidence of reduction.

ANALYTICAL PROCEDURES
Polished thin sections of 157 L chondrites (10 L3, 22 L4, 38 L5, 87 L6) were examined at UCLA by backscattered electron (BSE) imaging with the JEOL JXA-8200 electron microprobe using an acceleration voltage of 15 keV and a working distance of $11 mm.For most chondrites, only one thin section was examined.Mineral compositions were determined with the JEOL microprobe, using a focused beam and natural and synthetic standards (chromite for Cr, forsterite for Si and Mg, albite for Na, Mn-rich garnet for Mn, orthoclase for K, grossular for Al and Ca, magnetite for oxidized Fe, sphene for Ti, millerite for S, synthetic 99.99% Fe for metallic Fe, and the Fe-Ni-Co alloy NBS 1156 for Ni and Co).The acceleration voltage was 15 keV and the sample current was 15 nA.We employed 20-s counting times per element and ZAF corrections.The Co concentration in metallic Fe-Ni was corrected for the overlap of the Fe K β X-ray peak on the Co K α peak.The detection limit was 0.03 wt%.
Reported ages were calculated using the 40 K decay constants of Steiger and Jager (1977).The irradiation parameters "J" used in the 40 Ar/ 39 Ar age calculations are derived from the Fish Canyon sanidine (FC), the Fire Clay tonstein sanidine (FCT), and hornblende from the Lone Grove Pluton (Hb3gr).The ages of FC and FCT have been determined by the U-Pb zircon method; The FC tuff has a "young" age mode of 28.206 AE 0.022 Ma (Wotzlaw et al., 2013).This age is in good (0.2‰) agreement with the commonly used, astronomically calibrated value (28.201Ma) of Kuiper et al. (2008).The FCT has a U-Pb age of 314.629AE 0.039 Ma and 314.6 AE 0.9 Ma (Machlus et al., 2020 andLyons et al., 2006, respectively).As described in Setera et al. (2020), Hb3gr is calibrated to a U-Pb titanite closure age of 1082.8AE 0.3 Ma, derived from the mean of titanite ages that were within uncertainty of the youngest titanite age in Blackburn et al. (2017).Comparison of the measured 40 Ar*/ 39 Ar ratio of the standard to the calculated 40 Ar*/ 39 Ar ratio based on the U-Pb ages provides an intercalibration between the U-Pb calibrated standards and the measured samples is AE5‰.
Cosmic-ray exposure (CRE) ages were calculated for the six samples using the relationship below, where the CRE age (t 38 ) equals the total production rate for 38 Ar c .The total 38 Ar cos is determined by applying the lever rule to the total measured 36 Ar and 38 Ar, using a value of 5.35 for terrestrial atmospheric and 0.65 for the cosmogenic 36 Ar/ 38 Ar ratio.See Turrin et al. (2023) for a more detailed description of the methods.
Samples were loaded into separate wells (to preserve their identities) that had been drilled into aluminum disks along with the mineral standards discussed above.The samples were irradiated for 141 h at the Oregon State TRIGA reactor in the Cadmium-Lined In-Core Irradiation Tube (CLICIT) facility.A more complete discussion is given in Appendix 1. 40 Ar* is the standard notation used to denote radiogenic 40 Ar from the decay of 40 K. Loss of 40 Ar* generally occurs sometime during a sample's history.

39
Ar K is the standard notation used to denote 39 Ar produced by neutron irradiation of 39 K, which serves as a proxy for the 40 K content; it is used to calculate the apparent age of a sample.When used in a discussion of a plateau age, it refers to the amount of 39 Ar K released in a step-heating experiment that defines a plateau.A plateau age is defined as the age representing the diagram line wherein at least 50% of the released 39 Ar K have ages that are analytically indistinguishable at the alpha-95% confidence level (unless otherwise noted).

Normal L Chondrites
We examined 157 L chondrites (10 L3, 22 L4, 38 L5, 87 L6) by BSE imaging (Table 1).Most showed no reduction features.In typical L6 chondrites, individual mineral grains do not appear zoned in BSE images; they display uniform gray shadings from edge to edge, interrupted only by black fractures (Figure 1).(Nevertheless, many fractures show edge effects-thin [micrometer-or submicrometer-thick] white bands flanking one or both sides of the fracture.These bands are image artifacts, produced because grain edges have more surface area than flat regions, allowing greater numbers of weak secondary electrons to escape.)Different mineral phases exhibit distinct shades of gray, depending on the average atomic number of the phase-for example, metallic Fe-Ni is bright, olivine and low-Ca pyroxene are medium gray, plagioclase is dark gray.Metallic Fe-Ni grains show differences in brightness between intergrown kamacite and taenite (taenite is brighter because its Ni content is higher).However, individual metal phases are typically uniform in brightness from grain center to grain edge, indicative of little compositional zoning.
In L3 chondrites, some olivine and low-Ca pyroxene phenocrysts in type II chondrules exhibit normal compositional zoning; grain edges are richer in FeO (and appear brighter in BSE) than grain centers (e.g., fig. 1 of Jones, 1990).A few olivine grains show reverse zoning (grain centers are brighter than edges in BSE; e.g., fig.4c of Jones, 1990).Some olivine grains exhibit oscillatory zoning.
Not all thin sections of a single L6 chondrite show the same level of reduction.For example, of the three examined sections of Osceola, one exhibited major reduction, one exhibited minor reduction, and one exhibited no apparent reduction.Because reduction features are heterogeneously distributed and only a single thin section was examined for most of the surveyed L chondrites, the proportion of L6 chondrites exhibiting reduction (15/87) must be considered a lower limit.
The heterogeneous nature of reduced features in these meteorites is probably not due to brecciation (wherein some clasts are reduced and others are not).Although some of the reduced samples contain shock veins, for example, Viñales (Baziotis et al., 2023;Yin & Dai, 2021) and Cactus Springs (this study), our examination of thin sections of all 15 L chondrites exhibiting reduction reveals no obvious signs of brecciation.This is in accord with previous studies (e.g., Wang & Rubin, 1987).

Mineralogy and Petrology of Representative L6 Chondrites Exhibiting Reduction
Osceola (total known weight, 1099 g) was observed to fall in Florida in 2016; the meteorite is unweathered (W0).The rock is highly recrystallized; plagioclase grains range up to 100 μm.Poorly defined, relict chondrules range in apparent diameter from 450 to 1100 μm; textural varieties

Note:
For most of these meteorites, only a single thin section was imaged by BSE.
Cactus Springs (total known weight, 176 g) was found in Nevada in 2008; it is only slightly weathered (W1).The meteorite is very recrystallized; plagioclase grains range from 30 to 200 μm.Relict chondrules range in apparent diameter from 450 to 3240 μm; textural  varieties include BO, PO, and POP.Cactus Springs exhibits weak mosaic extinction, characteristic of shock stage S4.Troilite grains are polycrystalline.One small (0.5 × 1.0 μm) grain of metallic Cu was found.Also present are small irregular troilite grains within metallic Fe-Ni, fractured chromite grains, chromite-plagioclase assemblages, rare chromite veinlets, and a 4-5-mm-long shock vein containing troilite surrounding fragmented silicate grains.Maskelynite is absent.
Olivine (Fa 24.2 AE 0.3) and orthopyroxene (Fs 20.8 AE 0.2 Wo 1.4 AE 0.2) in Cactus Springs are compositionally homogeneous and in the range of equilibrated L chondrites.Kamacite contains 0.83 wt% Co, well within the L-chondrite range.Also present are Ca pyroxene (Fs 8.1 AE 0.6 Wo 44.7 AE 0.8), plagioclase (Ab 63.1 AE 3.6 Or 12.9 AE 2.6), taenite and tetrataenite (Tables 3 and 4).As in Osceola, plagioclase in Cactus Springs is significantly depleted in Na 2 O and enriched in K 2 O relative to equilibrated L chondrites.

Mafic Silicates
In those L6-chondrite thin sections showing major reduction, there are 4-5-μm-thick reduction rims (dark bands in BSE images) on nearly every olivine and orthopyroxene grain and along the boundaries of most fractures within these grains (Figure 2).A typical large olivine grain in Osceola has a composition of Fa 23.6 AE 0.2 (n = 10) in the center and Fa 23.2 AE 0.1 (n = 2) in the reduction rim (Table 5).[Because the rims are so narrow, it is difficult to obtain quantitative electron microprobe data.The rim olivine Fa value is an upper limit due to the narrowness of the rim and the possibilities of beam overlap and secondary fluorescence of Fe from outside the rim.]The MnO/FeO ratios in the grain center and the reduction rim are sub-identical (0.021, 0.020).The focused electron beam was too wide to resolve differences in low-Ca pyroxene between grain centers (Fs 19.9 AE 0.2 Wo 1.5 AE 0.1, n = 10) and reduction rims (Fs 20.0 Wo 1.6, n = 1; Table 5); centers and rims have sub-identical MnO/FeO ratios (0.036, 0.035).
Unlike the reduction rims on ureilite olivine grains (e.g., fig.1c of Goodrich, 1992), there are no small blebs of low-Ni metallic Fe at the mafic silicate grain margins in Osceola or the other L6 chondrites exhibiting reduction features.The reduction features described here also differ from the lower Fa contents in the portions of olivine grains adjacent to zoned taenite + kamacite particles in equilibrated H chondrites; these features are a consequence of localized olivine-orthopyroxene-metal reactions during cooling from peak-metamorphic temperatures (Reisener et al., 2006).

Sulfide
Every grain margin and nearly every fracture in every troilite (FeS) grain in the Osceola thin section showing major reduction are lined by a 4-12-μmthick band that appears dark in BSE images (Figure 3).(Similar features are evident in Cactus Springs, Muroc Dry Lake, and NWA 1857.)The BSE-dark phase is likely to be poorly crystalline pyrrhotite (Fe 1-x S).The composition of troilite grain centers in Osceola (n = 10) is 62.7 AE 0.2 wt% Fe and 37.5 AE 0.2 wt% S (Fe/S = 1.67);BSE-dark rims (n = 8) contain 62.3 AE 0.2 wt% Fe and 37.7 AE 0.1 wt% S (Fe/S = 1.65;Table 5).(As with the silicates, the reported amount of Fe in the rim is an upper limit due to the narrowness of the rim and the possibility of secondary fluorescence of nearby Fe.) Sulfide grains appear unaffected in the nine L chondrites that exhibit only minor reduction in olivine and orthopyroxene.

Ar-Ar and Cosmic-Ray Exposure Ages of L6 Chondrites with Reduction Features
To determine if the reduction exhibited by the L6 chondrites was the result of a single event (possibly connected to the disruption of the L-chondrite parent asteroid $470 Ma ago ;Heymann, 1967;Korochantseva et al., 2007;Swindle et al., 2014;Turner, 1988) or a series of stochastic heating events, we determined the 40 Ar/ 39 Ar ages of six of the L6 chondrites showing major reduction features: Cactus Springs, Guangnan, Muroc Dry Lake, NWA 1857, Osceola, and Thamaniyat Ajras (Table 6).In heating experiments, reliable plateau ages can be assigned if three or more consecutive steps are indistinguishable from adjacent ones at the 95% confidence level and together account for ≥50% of the total 39 Ar K released.In some of the samples, we list "forced" plateaus that do not meet these criteria.The isochron results are summarized in Table 7.
NWA 1857, measured in duplicate, produced a disturbed release spectrum, and did not yield a plateau age as defined above.The first three low-temperature steps of run 23184 indicate an 40 Ar* loss event $1200 Ma ago, and the last five high-temperature steps ($17% of the total 39 Ar K released) of run 23185 produce a forced plateau age Note: For the sulfide analyses, Cr, Co, and Ni were below the detection limit of $0.03 wt%.The edge may consist of poorly crystalline pyrrhotite.
of 3066 AE 16 Ma.When the Ar-isotope data are cast on an isotope correlation diagram, the high-temperature steps for both runs produce an array of points that correspond to isochron ages of 3055 AE 13 and 3056 AE 47 Ma, respectively.The mean sum weighted deviates (MSWDs) are greater than two, suggesting that the scatter in the points that compose the two isochrons is greater than what would be expected given the analytical errors.Taken at face value, these results suggest that this stone experienced a thermal event at $3060 Ma (Figure 5a,b).Thamaniyat Ajras has a plateau age of 4543 AE 3 Ma, accounting for $85% of the 39 Ar K released, and a concordant integrated/total fusion age of 4540 AE 8 Ma.The Ar-isotope data also yield a concordant age of 4540 AE 9 Ma (Figure 5c).Our reported age for Thamaniyat Ajras is one of the oldest reported ages for L6 chondrites.Other "old" L6 ages include MIL 05029 (4517 AE 11 Ma; Weirich et al., 2010); Park, which experienced shock and subsequent annealing (Ruzicka & Hugo, 2018;4526 AE 5 Ma;Ruzicka, Clay, et al., 2015); and an L4-chondrite clast recovered from Almahata Sitta (AhS) AhS 100 (4535 AE 10 Ma; see 40 Ar/ 39 Ar age compilation of the L chondrites; tab.S6.1, Turrin et al., 2023).
Muroc Dry Lake, measured in duplicate, produced a disturbed release spectrum and did not yield plateau ages as defined above.The first two low-temperature steps suggest an 40 Ar* loss event 2600-2700 Ma.Unlike NWA 1857, the isotope correlation diagrams for Muroc Dry Lake are scattered and do not yield an isochron result (Figure 5d,e).
Osceola, measured in duplicate, produced disturbed saddle-shaped step-heating release spectra.In sample split 23193, 7 of 10 steps produced a plateau age of 468 AE 2 Ma, defined by $44% of the 39 Ar K released.When cast on an isotope correlation diagram, the plateau steps define a concordant isochron age of 485 AE 12 Ma.The integrated/total fusion age is 623 AE 2 Ma; this is older than the plateau age, reflecting the older hightemperature steps.Sample split 23198 produced a two-step plateau age of 515 AE 6 Ma, defined by $35% of the 39 Ar K released.On an isotope correlation diagram, the Arisotopic data do not define an age.Like sample split 23193, the integrated/total fusion age (702 AE 5 Ma) is older than the plateau age.
Whereas the plateau age of sample split 23193 is just short of the 50% 39 Ar K release criteria, the age is in good agreement with the stratigraphic age of L-chondrite fossil meteorites found in 480-Ma-old Ordovician limestones in Sweden (Heck et al., 2004;Schmitz et al., 1997Schmitz et al., , 2001 Cactus Springs produced a concordant data set consisting of a five-step plateau age of 4038 AE 9 Ma, characterized by 100% of the 39 Ar K released, and integrated/total fusion and isochron ages of 4037 AE 16 and 4047 AE 32 Ma, respectively.The only other L6 chondrite of similar age is Leedey (significantly shocked, but little annealed; Ruzicka & Hugo, 2018), with an "averaged 39 Ar/ 40 Ar age of 3950 Ma" and a "partially reset 39 Ar/ 40 Ar plateau age of 3800 AE 100 Ma" Bogard et al. (1987;Figure 5h).
Guangnan has disturbed, saddle-shaped release spectra and did not produce a plateau or isochron age.
Step-heating of the sample resulted in an integrated/total fusion age of 156 AE 1 Ma (Figure 5i).The three youngest ages in the saddle spectrum range from 64 to 90 Ma.As discussed in the Osceola results above, these young ages are probably the result of only partial resetting of the K-Ar clock.
The only other L6 chondrite we found to exhibit major reduction in at least one thin section is Vi ñales.This stone fell as a shower in Cuba in 2019; 50 kg were recovered.Although we did not analyze noble gases in Vi ñales, Smith et al. (2021) determined gas-retention ages (equivalent to K-Ar ages) in two samples of Vi ñales to be 533 AE 55 and 707 AE 72 Ma.These values are similar to the integrated ages of the two Osceola samples.
The CRE ages for our samples (Table 8) are representative of the population of reported L6-exposure ages (see Herzog & Caffee, 2014;Marti & Graf, 1992).Guangnan and Thamaniyat Ajras share similar ages of 6.5 AE 0.1 and 7.5 AE 1.1 Ma, respectively.Exposure ages of Muroc Dry Lake, Cactus Springs, and Osceola (19.2 AE 1.8, 23.2 AE 0.4, 24.8 AE 0.3 Ma, respectively) all lie within a broad cluster of reported CRE ages of $15 to 30 Ma. Lastly, NWA 1857 has a CRE age of 41.5 AE 0.5 Ma, which is concordant with a commonly reported value from other L6 chondrites (Marti & Graf, 1992).Our reported CRE ages do not match their associated 40 Ar/ 39 Ar ages, indicating that the launch events of the materials that ultimately became these meteorites did not cause sufficient heating to reset the K-  Ar system.This is consistent with the observation of Swindle et al. (2014) that there are no CRE ages that match their associated 40 Ar/ 39 Ar ages.

Relative Timing of Reduction
Because the reduction process affecting L chondrites involved L6 chondrites nearly exclusively, it is probable the reduction event was either syn-or post-metamorphic.
The silicate, sulfide, and metallic Fe-Ni grains had thus coarsened to their present sizes prior to reduction.If reduction had occurred before grain coarsening during metamorphism, there likely would have been unreduced (uniformly BSE-shaded) overgrowths surrounding reduced mafic silicate grain cores (e.g., fig. 2 of Rubin, 2006).Because reduction zones line the fractures in many olivine and low-Ca pyroxene grains, these grains must have grown to their present sizes and been fractured (presumably by collisions) prior to reduction.

Absolute Timing of Reduction
The variations in Ar-Ar ages reflect stochastic collisions affecting those L6 chondrites exhibiting major reduction.No single impact event (such as the disruption of the L asteroid $470 Ma ago) caused widespread reduction in the L-chondrite parent body.This is evident in the absence of reduction features in most L6 chondrites with an 40 Ar/ 39 Ar age of $470-500 Ma: Bruderheim, Chico, Louisville, NWA 091, Paranaiba, Wethersfield (1971) (tab. 1 of Swindle et al., 2014).Osceola, with a plateau age of 469.1 AE 2.2 Ma (Figure 5e), is an exception.
Thamaniyat Ajras has a precise (AE0.7‰)plateau age of 4543 AE 3 Ma (Figure 5f); it was not subsequently disturbed.The major reduction of this meteorite provides the timing of Ar closure during post-metamorphic cooling.
If reduction were caused by a single event, it would probably have happened around the time of Ar closure in Thamaniyat Ajras.If reduction had occurred sporadically after this time (e.g., due to stochastic collisions), it would have affected L3, L4, and L5 chondrites as well.The fact that only one of 70 L chondrites of petrologic type <6 shows evidence of reduction indicates that sporadic reduction events were uncommon during the history of the L-chondrite parent asteroid.

Nature of Reductant
In those L6 thin sections exhibiting major reduction, reduction is pervasive; it is evident along most every olivine, orthopyroxene, and troilite grain boundary and along nearly every fracture (including narrow fractures) within these grains.This suggests the reductant was a low-viscosity fluid, possibly CO.Plausible reduction reactions include: Free silica was searched for in the reduced L6 chondrites by EPMA but was not found.Thus, there is no evidence that the hypothetical excess silica liberated from the reduction of ferroan olivine and orthopyroxene by these reactions produced many grains of free silica in these meteorites.The excess silica could have combined with olivine to make small amounts of additional low-Ca pyroxene: Electron microprobe analyses also show no evidence of metallic Si (formed by reduction of SiO 2 ) in the metallic Fe-Ni grains (Table 4).The 2-5-μm-thick kamacite rinds around taenite grains in Osceola (Figure 4) appear to be the site where Fe 0 derived mainly from reduced Fe 2+ in olivine and orthopyroxene diffused into metallic Fe-Ni grains.
A potential source for CO is oxidation of C derived from the suite of C-rich aggregates and C-rich chondritic clasts in OCs (Scott et al., 1988).These objects are composed of fine-scale intergrowths of poorly graphitized C, amorphous C, metallic Fe-Ni, and minor chromite (Brearley, 1990); they range up to $1 mm in size and constitute up to several volume percent of some type 3 OC, including L3.5 ALH 77011 (McKinley et al., 1981).

Oxidation of Graphite
The extent of graphite oxidation (C + ½O 2 → CO) at various temperatures is likely to depend on oxygen fugacity.Under terrestrial-surface redox conditions, the proportion of graphite that would be oxidized at temperatures corresponding to OC petrologic types (Xiaowei et al., 2004) can be estimated: type 3 (0%-4%), type 4 (4%-11%), type 5 (11%-17%), and type 6 (17%-26%; Figure 6).Under the more reduced conditions pertaining during OC thermal metamorphism (e.g., Rubin et al., 1988), less graphite would have been oxidized.We suggest that, during progressive thermal metamorphism of OCs, there are two competing redox reactions: (1) mobilization of water (derived from ice or phyllosilicates within the matrix), causing oxidation of silicates, oxides, and metallic Fe-Ni and (2) concomitant oxidation of any graphite that may be present, resulting in release of CO and reduction in these same mineral phases.After peakmetamorphic temperatures were reached in L6 chondrites, $75% of the water had been used up or lost; the remaining water facilitated the continuation of graphite oxidation so that, after this point, overall reduction effects exceeded those of oxidation (Figure 7).
The L3, L4, and L5 chondrites were much less affected by reduction due to their lower metamorphic temperatures.Nevertheless, one L4 chondrite (NWA 8144) exhibits minor reduction.It is not clear why this particular meteorite (one out of 70 L3 to L5 chondrites) is an exception.We speculate that NWA 8144 experienced a brief impact-induced localized heating episode during a period when CO was present.Although NWA 8144 is listed as shock stage S1, it appears to have been shocked and subsequently annealed after it was buried beneath materials heated by the impact event (e.g., Rubin, 2004).Shock features evident in NWA 8144 include: (a) minor silicate darkening (caused by thin curvilinear trails containing small blebs of metallic Fe-Ni), (b) thin ($1-μmthick) chromite veinlets within olivine grains, and (c) small irregular troilite grains within metallic Fe-Ni.Annealing at subsolidus temperatures could have repaired the shock-induced damage of the olivine crystal lattices (e.g., Ashworth & Mallinson, 1985;Bauer, 1979), resulting in an apparent shock stage of S1.
In some thin sections of the L6 chondrites that show major reduction, there are 4-5-μm-thick BSE-dark reduction rims on nearly every olivine and orthopyroxene grain and along the boundaries of most fractures within these grains.For example, olivine in rims has slightly lower fayalite contents than grain centers: Fa 23.2 versus Fa 23.6.In addition, every grain margin and nearly every fracture in every troilite grain are lined by a 4-12-μm-thick dark band, probably poorly crystalline pyrrhotite.There are also 2-5-μm-thick dark rinds of kamacite around some taenite grains.
The six measured L6 chondrites with major reduction features have 40 Ar/ 39 Ar integrated/total fusion or plateau ages ranging from 156 AE 1 Ma for Guangnan to 4543 AE 3 Ma for Thamaniyat Ajras.Reduction occurred after the silicate, sulfide, and metallic Fe-Ni grains had attained their present sizes toward the end of the epoch of parent-body thermal metamorphism.The precise plateau age of Thamaniyat Ajras marks the timing of Ar closure during post-metamorphic cooling.Because this meteorite was not subsequently disturbed, it seems likely the L6 reduction event occurred around 4543 Ma ago.
The reductant was probably a low-viscosity fluid, possibly CO, derived from oxidation of carbon (poorly graphitized carbon and amorphous carbon) within submillimeter-size C-rich aggregates in the meteorite matrix.
Water that had accreted to the L-chondrite parent body (plausibly as ice) was heated and mobilized during metamorphism, causing oxidation of components.After peak-metamorphic temperatures were reached in L6 chondrites (750-950°C), about 75% of the water had been used up or lost; the remaining water facilitated the continuation of graphite oxidation so that, after this point, overall reduction effects exceeded those of oxidation.L chondrites of lower petrologic type were much less affected by reduction due to their lower metamorphic temperatures.
4538 AE 9 Ma with a MSWD of 1.2 and a trapped 40 Ar/ 36 Ar (trp.) of 60 AE 7. The three ages (plateau, integrated, and isochron ages) are concordant at the alpha-95 % confidence level.Arguably, the plateau age (4543 AE 3 Ma) represents a cooling age for the Thamaniyat Ajras meteorite.
Lab #: 23194 and Lab #: 23195: The L6 chondrite Muroc Dry Lake was measured in duplicate; both sample splits produce increasing age spectra that are slightly concave downward, typical of samples that experienced 40 Ar loss.The integrated ages of the two splits (23194 and 23195) are 3684 AE 2 and 3724 AE 4 Ma, respectively; they are similar but analytically different.As with NWA 1857, we can compare the Thamaniyat Ajras (TA) cooling age (4543 Ma) to the integrated ages of NWA 1857.This comparison (3700 Ma/4543 Ma) indicates an approximate 20% 40 Ar loss for the Muroc Dry Lake meteorite.The first low-temperature steps from the two samples suggest that the age of this 40 Ar loss event may have occurred about 2600-2700 Ma.Like the NWA 1857 sample, the degassing history of this sample of Muroc Dry Lake shows a complex distribution on isotope correlation diagrams.With a large MSWD and excessively high trapped 40 Ar/ 36 Ar, this sample did not yield useful age information.
Lab #: 23196 and Lab #: 23198: The L6 chondrite Osceola was measured in duplicate.Run 23,196 produces a saddle-shaped 39 Ar K release spectrum and an integrated total fusion age of 623 AE 2 Ma.Seven of 10 steps satisfy the "more than three contiguous steps that are analytically indistinguishable at the 95% confidence level" criterion.However, the steps constitute only 45% of the total 39 Ar K released; thus, we overrode the 50% criterion to obtain a plateau age of 468 AE 2 Ma.When plotted on an isotope correlation diagram, the plateau points define an age of 484 AE 12 Ma with a MSWD of 0.7 and a trapped 40 Ar/ 36 Ar (trapped) of À3 AE 1.0 (1 s).
Split 23198-02: An eight step-heating release spectrum did not produce a valid plateau.By relaxing the plateau criteria to only two points and to a greater than 30% 39 Ar K release, we obtained a two-point plateau with an apparent age of 516 AE 6 Ma.The isotope correlation diagram has a MSWD greater than 2 and does not indicate a meaningful age.Two runs were measured on split 23198: Run 23198-01 and 23198-01.Neither of these runs produces plateaus.When plotted on an isotope correlation diagram, the data do not produce any meaningful results, as indicated by the large MSWDs.Thus, there are no useful age results indicated for split 23198.Similar to the other L chondrites discussed here, the general shape of the three Osceola L6-chondrite stepheating experiments suggests a major resetting event at 469 AE 2 Ma.The age of the target is consistent with an apparent age of $1000 Ma.
Lab #: 23201: A single split of the L6 chondrite Cactus Springs was measured; it produced a five-step plateau age of 4037 AE 9 Ma comprised of 100% of the total 39 Ar (K) released with a concordant integrated age of 4036 AE 16 Ma.When the five plateau steps are plotted on an isotope correlation diagram, the data produce an isochron age of 4045 AE 56 Ma with a MSWD of 0.1 and a trapped 40 Ar/ 36 Ar (trapped) of À17 AE 93, indistinguishable from zero.The three ages (plateau, integrated, and isochron ages) are concordant at the alpha-95 % confidence level.Arguably, the plateau age (4037 AE 9 Ma) is from a completely reset part of the original L-chondrite parent body.
Lab #: 23203: A single split of the L6 chondrite Guangnan was measured.The disturbed saddle-shaped release spectrum did not produce a plateau.The maximum step with a measurable age of 39 Ar K is 1300 AE 6 Ma and the minimum step age is 64.9 AE 0.3 Ma.The integrated age is 157.8AE 0.6 Ma.When plotted on an isotope correlation diagram, the data points do not indicate any well-defined array and the large MSWD does not produce any meaningful age information.In summary, this sample did not produce any age information.

FIGURE 1 .
FIGURE 1.The Lanxi L6 chondrite fall showing no evidence of reduction.Olivine (medium-light gray) and orthopyroxene (opx; medium gray) grains exhibit no evident compositional zoning; grains appear homogeneous in shade from center to edge.Plagioclase (plag; dark gray) and small grains of metallic Fe-Ni (white) are also present.Image courtesy of Chi Ma (Caltech).Backscattered electron image.(Color figure can be viewed at wileyonlinelibrary.com)

FIGURE 2 .
FIGURE 2. The Osceola L6-chondrite fall exhibiting major reduction features.Nearly every grain margin and every internal fracture in olivine (light gray) and orthopyroxene (opx; medium gray) in this thin section have 4-5-μm-thick BSE-dark bands (arrows).These bands are $0.4 mol% lower in Fa and Fs than the lighter-shaded grain centers.The white regions around some fractures result from charging in the electron microprobe.Backscattered electron image.
). Subsequent work by Korochantseva et al. (2007) and Weirich et al. (2012) confirmed a 470-Ma age for collisional disruption of the L-chondrite parent body.As noted in Swindle et al. (2014), most of the young Lchondrite 40 Ar/ 39 Ar ages are centered about 500 Ma due to only partial resetting of the K-Ar clock (as evident in sample split 23198; Figure 5f,g).

FIGURE 6 .
FIGURE 6. Oxidation of graphite under terrestrial-surface redox conditions (after Xiaowei et al., 2004).Temperatures corresponding to those inferred for ordinary-chondrite petrologic types are shown.Less graphite would have been oxidized under the more reduced conditions pertaining during ordinary-chondrite thermal metamorphism.(Color figure can be viewed at wileyonlinelibrary.com)

TABLE 1 .
Examined L-group chondrites that exhibited no reduction features in examined thin sections.

TABLE 2 .
L-group chondrites exhibiting reduction features in at least one thin section.

TABLE 3 .
Mean silicate and oxide mineral compositions (wt%) in Osceola and Cactus Springs.

TABLE 4 .
Mean compositions (wt%) of metallic Fe-Ni minerals in Osceola and Cactus Springs.

TABLE 5 .
Mean compositions (wt%) of grain centers and edges in L6 Osceola.

TABLE 6 .
Summary table of 40 Age is equivalent to a total fusion measurement; n/n total = the number of consecutive steps out of the total number of heating steps that defines the plateau age.Preferred ages are in bold.

TABLE 7 .
Summary table of 40 Ar/ 39 Ar isochron results.: 40 Ar/ 36 Ar Trp = 40 Ar/ 36 Ar of the gas incorporated into the sample at the measured apparent age.Italicized items indicate there was no isochron.
Note(g) Osceola, second laboratory run.(h) Cactus Springs.(i) Guangnan.Also shown is the MSWD (Mean Sum Weighted Deviates) value; this is a statistical "goodness-of-fit" test.Higher MSWD values indicate poorer fits; MSWD values less than 2.5 are generally accepted as good fits.The MSWD value for Thamaniyat Ajras is 0.62.(Color figure can be viewed at wileyonlinelibrary.com)