Intermineral oxygen three-isotope systematics of silicate minerals in equilibrated ordinary chondrites

Abstract

High precision oxygen three-isotope ratios were measured for four mineral phases (olivine, low-Ca and high-Ca pyroxene, and plagioclase) in equilibrated ordinary chondrites (EOC) using a secondary ion mass spectrometer. Eleven EOCs were studied that cover all groups (H, L, LL) and petrologic types (4, 5, 6), including S1–S4 shock stages, as well as unbrecciated and brecciated meteorites. SIMS analyses of multiple minerals were made in close proximity (mostly <100 μm) from several areas in each meteorite thin section, to evaluate isotope exchange among minerals.

Oxygen isotope ratios in each mineral become more homogenized as petrologic type increases with the notable exception of brecciated samples. In type 4 chondrites, oxygen isotope ratios of olivine and low-Ca pyroxene are heterogeneous in both δ¹⁸O and Δ¹⁷O, showing similar systematics to those in type 3 chondrites. In types 5 and 6 chondrites, oxygen isotope ratios of the four mineral phases plot along mass dependent fractionation lines that are consistent with the bulk average Δ¹⁷O of each chondrite group. The δ¹⁸O of three minerals, low-Ca and high-Ca pyroxene and plagioclase, are consistent with equilibrium fractionation at temperatures of 700–1000°C. In most cases the δ¹⁸O values of olivine are higher than those expected from pyroxene and plagioclase, suggesting partial retention of premetamorphic values due to slower oxygen isotope diffusion in olivine than pyroxene during thermal metamorphism in ordinary chondrite parent bodies.

INTRODUCTION

Equilibrated ordinary chondrites (EOCs) represent 80% of all modern meteorite falls. They are called as such because of their equilibrated, homogenized, olivine and pyroxene major-element compositions, and because of their ubiquity. There are three chemical groups of EOCs, H, L, and LL, based on bulk metal content and total iron. EOCs are classified into petrologic types from type 4 to 6 that represent the degree of thermal metamorphism they have experienced in their parent asteroids (Van Schmus and Wood 1967). Type 3 ordinary chondrites are referred to as unequilibrated ordinary chondrites (UOCs) because of the heterogeneous major element compositions of olivine and pyroxene; they are considered precursors of EOCs. Among EOCs olivine and low calcium pyroxene MgO and FeO compositions are equilibrated in all petrologic types. However, while distributions of major element compositions among coexisting low-Ca pyroxene and high-Ca pyroxene are equilibrated in type 6 chondrites and some type 5 chondrites, they are not equilibrated in type 4 chondrites (e.g., McSween and Patchen 1989). EOCs are considered to have experienced metamorphic temperatures from a low of ~550 °C for type 4 chondrites up to peak temperatures of ~1000 °C for type 6 chondrites (e.g., Harrison and Grimm 2010). The heat source of thermal metamorphism was likely due to decay of the short-lived radionuclide ²⁶Al (half-life of 0.7 million years) that existed at the beginning of the solar system (Lee et al. 1977). Evidence for localized melting, brecciation, shock deformation of minerals, and veining are observed among EOCs, and these features are attributed to planetesimal impacts onto the parent body throughout the long history of the Solar System (Keil et al. 1997; Rubin 1995, 2004). L chondrites in particular often show evidence of later impact heating (e.g., Bogard et al. 1995), and an extensive L chondrite parent body breakup and subsequent meteor shower is considered to have occurred in the Ordovician ~470 million years ago (e.g. Heck et al. 2004; 2010; Korochantseva et al. 2007).

For each group of EOCs there are no significant differences in their bulk chemistry and isotope compositions that are related to petrologic type. As such, it is hypothesized that the parent asteroids of ordinary chondrites consist of layered structures with progressive thermal metamorphism (higher grade at depth), which is called an “Onion-shell structure” (e.g., Herndon and Herndon 1977; Minster and Allègre 1979; Miyamoto et al. 1981; Pellas and Storzer 1981; Trieloff et al. 2003; Harrison and Grimm 2010).

The onion-shell model suggests that ordinary chondrite parent bodies were ~100 km in radius, were internally heated by the decay of ²⁶Al, and that the interior of the body was heated to higher metamorphic temperatures and cooled more slowly than the surface. This model predicts that type 6 chondrites formed near the center of the parent body, with higher peak metamorphic temperatures and slower cooling rates. Type 4 chondrites were metamorphosed near the surface with lower peak temperatures and a faster cooling rate and UOCs made up the surface materials. The radiogenic heat generated from the interior of the asteroid could have produced a temperature gradient as a function of depth, due to the balance between thermal diffusion and radiative loss from the asteroid surface. The onion-shell model has been shown to work well, at least for the H chondrite parent body, to explain the observed correlation between phosphate Pb-Pb ages and petrologic types (Göpel et al. 1994); type 4 chondrite ages are as old as a few million years after CAIs, while type 6 chondrite ages are younger than the oldest type 4 chondrites by as much as ~60 Ma. Trieloff et al. (2003) further demonstrated that the cooling rates of EOCs, which are estimated from multiple chronometers having different closure temperatures, decrease with petrologic types, which is predicted from the onion-shell model. In contrast, it has long been discussed that metallographic cooling rates, inferred from Ni concentration zoning in taenite grains, do not show systematic differences among types 3–6 ordinary chondrites, which conflicts with the onion-shell model (Scott and Rajan 1981; Taylor et al. 1987). More recently, Scott et al. (2014) confirmed the lack of a correlation in metallographic cooling rates from a systematic study of 30 H chondrites with low shock stages. In particular, some H4 chondrites cooled slower than some H6 chondrites, and H3 chondrites did not cool faster than higher petrologic type H chondrites. A few H4 chondrites with very old Pb-Pb ages have extremely fast cooling rates of ≥5,000 °C/Myr at 500°C, which is 50 times faster than predicted cooing rates from ²⁶Al heating within undisturbed onion-shell structures. Furthermore, thermometry using assemblages of olivine-spinel (Kessel et al. 2007) and two-pyroxenes (Ganguly et al. 2013) suggests higher cooling rates of ~100°C/kyr at ~750°C for type 6 chondrites and 25–100°C/kyr above 700°C for H5/H6 chondrites, respectively. These data suggest that parent bodies of ordinary chondrites initially had onion-shell structures, but later experienced significant impact processes after they reached peak metamorphic temperatures, which either excavated some materials or resulted in the parent body break-up and re-assembly (Scott et al. 2014; Ganguly et al. 2016).

The ancient solar nebula is inferred to have experienced a large mass independent fractionation (MIF) of oxygen isotopes in ¹⁶O, according to analyses of materials from primitive meteorites and implanted solar wind from returned Genesis mission detectors. The ¹⁸O/¹⁶O and ¹⁷O/¹⁶O ratios of each of these materials, expressed as δ¹⁸O and δ¹⁷O respectively, have permil (‰) deviations from SMOW (Standard Mean Ocean Water) that plot along a slope ~1 line, with values ranging from −60‰ to +200‰ (e.g., Clayton 1993; Sakamoto et al. 2007; McKeegan et al. 2011). The MIF effects among bulk ordinary chondrites are very small (≤ 1‰), but they do plot along a slope ~1 line, referred to as the equilibrated chondrite line (ECL). The δ¹⁸O and δ¹⁷O values of EOCs generally increase in order of H, L, and LL chondrites, though there are significant overlaps between L and LL chondrites (Clayton et al. 1991). Chondrules in LL3 chondrites show systematically lower δ¹⁷O and δ¹⁸O than those of bulk LL chondrites but are similar to those of bulk H chondrites, indicating ¹⁶O-poor water (ice) in ordinary chondrite parent bodies contributed to their bulk oxygen isotope ratios (Kita et al. 2010).

Clayton (1993) suggested that oxygen isotope fractionations among olivine (Ol), low-Ca pyroxene (Lpx), and plagioclase (Pl) in EOCs provide reasonably concordant temperatures (600–1000°C) that generally increase with increasing petrologic type. In the study of Itokawa particles analyzed by secondary ion mass spectrometry (SIMS), Nakashima et al. (2013) reported oxygen three-isotopes of olivine, low-Ca pyroxene, high-Ca pyroxene (Hpx), and plagioclase, in Guareña (H6) and St. Séverin (LL6) chondrites that are not fully consistent with inter-mineral O-isotope equilibrium. In particular, while Δ¹⁸O_Pl-Hpx (=δ¹⁸O_Pl – δ¹⁸O_Hpx) values are consistent with peak metamorphic temperatures of EOCs (800–1000°C), isotope fractionations between low-Ca pyroxene and olivine are too small (<0.5‰) when compared to those expected from plagioclase and high-Ca pyroxene. The δ¹⁸O values of olivine are higher than those of high-Ca pyroxene in both samples, which is opposite from that expected during equilibrium fractionation. As a result, Nakashima et al. (2013) suggested that oxygen isotope diffusion in olivine and low-Ca pyroxene was too slow to reset Δ¹⁸O-derived equilibrium temperatures at sub-solidus conditions during parent body thermal metamorphism. Nakashima et al. (2013) proposed that neither high-Ca pyroxene nor low-Ca pyroxene was in equilibrium with olivine. They also hypothesize that mass dependent fractionation between high-Ca pyroxene and plagioclase represents recrystallization temperatures of these minerals or the closure temperature established by O-isotope diffusion during retrograde metamorphism. To build upon the results of Nakashima et al. (2013), we report the oxygen isotope systematics of olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase from 11 EOCs. The EOCs investigated cover all groups (H, L, LL) and metamorphic types (4, 5, 6); they include S1–S4 shock stages, as well as unbrecciated and brecciated meteorites. In contrast to previous studies, in which analysis locations of each mineral were randomly selected, we examine oxygen isotope ratios of these four minerals that are in contact or within 100 μm of each other. Such systematic analyses should test if oxygen isotope exchange and equilibration was achieved locally, even if the entire meteorite is not homogenized.

METHODS

Sample Selection and Preparation

The 11 EOCs investigated here represent a diverse selection (Table 1), showing variable petrologic types (4, 5, or 6) and chemical groups (H, L, or LL). Most meteorites were selected to complement the studies of Clayton et al. (1991) and Clayton (1993), who reported bulk oxygen three-isotope ratios and/or δ¹⁸O values of mineral separates. Bjurböle (L4), Allegan (H5), Bruderheim (L6), and Estacado (H6), have mineral-separate data from Clayton (1993) and Forest Vale (H4), Allegan (H5), Bjurböle (L4), Ausson (L5), Soko Banja (LL4), St. Mesmin (LL6), and Estacado (H6) were measured by Clayton et al. (1991) for bulk oxygen isotope ratios. The Mifflin (L5) meteorite is a recent fall and is included because Kita et al. (2013) reported bulk oxygen isotope ratios. Lorton is also a recent fall (Corrigan et al. 2010) that was selected because it is an unshocked L6, which is relatively uncommon among L6 chondrites. Tuxtuac (LL5) was previously studied for U-Pb chronology of phosphate by Göpel et al. (1994), who also reported phosphate ages of Forest Vale (H4), Allegan (H5), and Ausson (L5). Shock stages of selected meteorites are low (S1–S2), except for Bruderheim (S4) and Ausson (S3). Mifflin and Bruderheim are fragmental breccias and St. Mesmin is a regolith breccia (Table 1).

Table 1.

List of EOCs studied (shock stages, brecciation).

Type	Meteorite	specimen#	Shock	Brecciation	O-isotope
H4	Forest Vale*		S2 [1]		[7]
H5	Allegan	USNM 215	S1 [1]		[7, 8]
H6	Estacado	ME 760 sp.4	S1 [2]		[7, 8]
L4	Bjurböle	ME 1428 sp.12	S1 [3]		[7, 8]
L5	Ausson	104PE1	S3 [3]		[7]
L5	Mifflin	UWGM898	S2 [4]	fragmental [4]	[4]
L6	Bruderheim	ME 2476 sp.9	S4 [3]	fragmental [3]	[8]
L6	Lorton	USNM7591	S1 [5]
LL4	Soko Banja	2759PE			[7]
LL5	Tuxtuac	USNM6197	S2 [1]
LL6	St. Mesmin	369PE1		regolith [6]	[7]

^*Source: Dr. Naoji Sugiura, the original section was made by the Geological Survey of Japan

References:

^[1]Rubin (1994),

^[2]Folco et al. (1996),

^[3]Stöffler et al. (1991),

^[4]Kita et al. (2013),

^[5]Corrigan et al. (2010),

^[6]Schultz and Signer (1977),

^[7]Clayton et al. (1991),

^[8]Clayton (1993).

Small fragments (2–6 mm) of individual meteorites were cut to expose and create a flat interior surface. The Mifflin (L5) sample is made from a small fragment (1.5 mm), which was prepared as a part of the initial identification of the Mifflin fall and contains both dark matrix and a light lithology (Fig. 4 of Kita et al. 2013). Each sample was mounted in epoxy with San Carlos olivine grains (~ 500 μm) that served as a running oxygen isotope standard (Kita et al. 2010). Mounts were then ground and polished to a flat smooth surface. A thin section of Forest Vale was remounted into epoxy resin with an olivine standard, because the original thin section showed significant surface topography (20–40 μm range), which exceeds the range for accurate oxygen isotope analyses (e.g. Kita et al. 2009). Samples were carbon coated for scanning electron microscopy (SEM), electron probe microanalysis (EPMA), and SIMS analysis. Optical and/or electron microscope images of the polished meteorite sections are shown in Supporting Information 1.

Electron Microscopy

A scanning electron microscope (SEM, Hitachi S3400) was used to select target areas for SIMS analyses, by obtaining backscattered electron (BSE) and secondary electron (SE) images. BSE images were used to identify olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase in the meteorite thin sections. Semi-quantitative energy-dispersive X-ray spectroscopy (EDS) was applied to distinguish high-Ca pyroxene from olivine. This was particularly important for H chondrites, as olivine and high-Ca pyroxene appear to have the same brightness in BSE images. Calcium X-ray maps were also obtained by EDS to determine the outlines of high-Ca pyroxene grains for SIMS targeting purposes.

A total of six to nine target areas were selected in each of the 11 EOC samples for SIMS analyses, according to the criteria that individual target minerals were (1) larger than 20 μm; (2) devoid of inclusions or cracks; and (3) that the four minerals were adjacent to each other (within 100 μm). In breccias and in type 5 chondrites, not all four minerals could be located within 100 μm of each other. In such cases, the four minerals were located within 200 μm of each other. For the type 4 chondrites, only olivine and low-Ca pyroxene were selected for SIMS analysis because plagioclase and high-Ca pyroxene are smaller than 15 μm, which is the spot size of SIMS analysis in this study. The selected targets of olivine and low-Ca pyroxene in the type 4 chondrites were all in chondrule structures, as grain sizes in the matrix are too small to accommodate a 15 μm SIMS spot analysis.

Major element compositions of individual SIMS target minerals were measured with a CAMECA SX-51 electron probe microanalyzer (EPMA), equipped with five wavelength dispersive X-ray spectrometers (WDS) at the University of Wisconsin-Madison. Analytical conditions are similar to those described in Kita et al. (2013). WDS quantitative chemical analyses were performed using a 15 kV accelerating voltage and with a 10–20 nA beam current that was focused to approximately 1 μm in diameter. Two analyses were taken for each mineral region targeted for SIMS. The endmember compositions, Fo (forsterite) in olivine, En (enstatite) and Wo (wollastonite) in pyroxene, and An (anorthite) in plagioclase, are calculated for each unknown analysis and used for SIMS instrumental bias corrections.

Oxygen Isotope Analysis

Oxygen isotope ratios of EOC minerals were analyzed with the CAMECA IMS-1280 SIMS at the WiscSIMS laboratory, University of Wisconsin-Madison (Kita et al. 2009). Analytical conditions and measurement procedures are similar to those described in Kita et al. (2010) and Nakashima et al. (2013). We used a 15–μm Cs⁺ primary beam at 2.5 nA, which yielded secondary ¹⁶O⁻ ion count-rates of ~2.5×10⁹ cps at a mass resolving power of ~5,000 (10% peak height). The three stable oxygen isotopes were detected simultaneously using three Faraday cup detectors. Samples were analyzed in groups of eight to fourteen unknowns, which were bracketed by eight analyses (4 analyses each before and after) of the San Carlos olivine standard in the same mount. Typical spot-to-spot reproducibilities of δ¹⁸O, δ¹⁷O, Δ¹⁷O (=δ¹⁷O – 0.52×δ¹⁸O) were 0.3‰, 0.5‰, and 0.5‰, respectively (2SD). The ¹⁶OH⁻ signal was examined after each analysis, in order to correct for the ¹⁶OH interference to ¹⁷O (Heck et al. 2010). All ¹⁶OH corrections were <0.05‰. We analyzed multiple olivine, pyroxene, and plagioclase standards in the same SIMS sessions to evaluate the instrumental biases, which are functions of mineral compositions (Fo in olivine, En and Wo in pyroxene, and An in plagioclase). The bias correction methods are same as those described in Nakashima et al. (2013) and Tenner et al. (2013). Instrumental bias corrected δ¹⁸O and δ¹⁷O values are reported in VSMOW-scale (Baertschi 1976).

All SIMS analyses were made within a 6 mm radius from the center of the mount. This minimized instrumental biases (<0.3‰ and <0.15‰ in δ¹⁸O and δ¹⁷O, respectively) related to the location of analyses on the sample holders (Kita et al. 2009). After SIMS data were collected, SEM images were obtained for each SIMS pit to determine if desired phases were analyzed, and to verify that the pits were free of large cracks, voids, or overlapping of other mineral phases. Analyses with undesirable features were rejected from the dataset.

RESULTS

Petrology and Mineral Chemistry

Representative SEM-BSE images of each EOC are shown in Fig. 1. In general, chondrules show well-defined outlines in type 4 chondrites, are readily delineated in type 5 chondrites, and are poorly defined in type 6 chondrites (Van Schmus and Wood 1967). Olivine and low calcium pyroxene are large compared to plagioclase and high-calcium pyroxene, and are often associated with each other because most of them were recrystallized from chondrule mesostases. Typical grain sizes of high-Ca pyroxene and plagioclase are smaller than 10 μm in type 4 chondrites (Fig. 1a). Type 5 chondrites contain secondary plagioclase larger than 50 micrometers (Fig. 1b). However, there are abundant plagioclase smaller than 5 micrometers (Fig. 1c). Type 6 chondrites also contain large plagioclase that sometimes exceeds 100 micrometers in size. Plagioclase smaller than 5 micrometers are deficient in type 6 chondrites. The identification of clean plagioclase and high-Ca pyroxene larger than 20 μm was much more difficult in the regolith breccia St. Mesmin (LL6) when compared to unbrecciated type 6 chondrites. The detailed BSE images of individual areas for EPMA and SIMS analyses are shown in Supporting Information documents 2–4.

Fig. 1.

BSE Images of representative EOCs. (a) Soko Banja (LL4), (b) Allegan (H5), (c) Lorton (L6), (d) St. Mesmin (LL6 breccia). The four phases labeled are olivine (Ol), low-Ca pyroxene (Lpx), high-Ca pyroxene (Hpx), and plagioclase (Pl).

EPMA analyses of individual mineral phases are shown in the Supporting Information 5. They are homogeneous within each meteorite and are within the range of each chondrite group (Brearley and Jones, 1998), though variabilities in major element compositions are observed among breccia samples. Olivine Fa (fayalite) and low-Ca pyroxene Fs (ferrosilite) compositions are slightly variable (Fa_30-32 and Fs_23-26) in regolith breccia St. Mesmin (LL6). The plagioclase albite compositions from the fragmental breccia Bruderheim (L6) range from 73 to 83, while those of St. Mesmin (LL6) show a smaller spread of 82 to 87. As previously reported (e.g., McSween 1989), major element compositions of low-Ca pyroxene and high-Ca pyroxene are in equilibrium for type 6 chondrites, according to chemical geothermometers of paired minerals. However, chemical compositions of two pyroxenes in type 5 chondrites are not always in equilibrium (Kessel et al. 2007).

Oxygen Isotope Ratios of EOC Minerals

Results of 238 individual oxygen isotope analyses are shown in Tables 2–4, sorted by area and mineral phase. A total of 11 analyses were rejected after inspection of SIMS spots via SEM revealed overlaps with other phases and unusual pit shapes. Thus, not all areas in types 5 and 6 chondrites have the full set of four-phase SIMS data.

Table 2.

Oxygen isotope ratios of olivine and low-Ca pyroxene in type 4 EOCs

Meteorite (type)	Area	δ18 OVSMOW				δ17 OVSMOW				Δ17O				Remark
		Ol ± unc.		Lpx ± unc.		Ol ± unc.		Lpx ± unc.		Ol ± unc.		Lpx ± unc.
Forest Vale (H4)	T04	6.14	0.15	5.58	0.15	3.96	0.32	3.87	0.32	0.77	0.29	0.96	0.29
	T05	4.39	0.15	4.16	0.15	3.70	0.32	3.36	0.32	1.42	0.29	1.20	0.29
	T08	4.11	0.15	4.09	0.15	4.06	0.32	3.90	0.32	1.92	0.29	1.78	0.29
	T09	4.28	0.15	4.59	0.15	4.06	0.32	3.47	0.32	0.94	0.29	1.08	0.29
	T10	5.25	0.15	4.59	0.15	4.06	0.32	3.42	0.32	0.42	0.29	1.04	0.29	[1]
	T11	6.00	0.15	4.98	0.15	4.06	0.32	2.98	0.32	0.54	0.29	0.39	0.29
	T12	3.99	0.15	3.13	0.15	4.06	0.32	3.28	0.32	1.53	0.29	1.65	0.29
	T13	5.07	0.15	5.22	0.15	4.06	0.32	3.33	0.32	0.69	0.29	0.62	0.29	[2]
	T14	2.16	0.15	2.31	0.15	4.06	0.32	2.36	0.32	0.84	0.29	1.16	0.29
	Average	4.60		4.29		4.06		3.33		1.01		1.10
	2SD	2.42		2.06		4.06		0.92		1.01		0.87
Bjurböle (L4)	T01	3.27	0.15	4.51	0.15	2.23	0.71	2.78	0.71	0.53	0.73	0.44	0.73
	T06	4.72	0.15	4.85	0.15	3.21	0.71	3.31	0.71	0.76	0.73	0.79	0.73
	T09	4.66	0.15	4.85	0.15	3.23	0.71	3.61	0.71	0.80	0.73	1.09	0.73
	T11	5.51	0.15	4.85	0.15	4.18	0.71	2.76	0.71	1.31	0.73	0.24	0.73
	T13	2.11	0.15	3.60	0.15	1.92	0.71	3.16	0.71	0.82	0.73	1.28	0.73
	T16	5.55	0.15	5.71	0.15	3.55	0.71	4.06	0.71	0.66	0.73	1.09	0.73
	T21	4.47	0.15	5.47	0.15	3.80	0.71	4.10	0.71	1.48	0.73	1.26	0.73
	T23	4.54	0.15	4.16	0.15	3.05	0.71	2.68	0.71	0.68	0.73	0.52	0.73
	Average	4.35		4.75		3.14		3.31		0.88		0.84
	2SD	2.30		1.35		1.51		1.14		0.67		0.80
Soko Banja (LL4)	T03	4.12	0.14	5.26	0.14	2.15	0.30	3.37	0.30	0.01	0.26	0.64	0.26
	T04	−3.05	0.14	5.45	0.14	−4.62	0.30	3.48	0.30	−3.03	0.26	0.64	0.26
	T10	4.33	0.14	5.82	0.14	3.83	0.30	4.79	0.30	1.58	0.26	1.76	0.26
	T13	−10.44	0.14	4.89	0.14	−13.54	0.30	2.86	0.30	−8.11	0.26	0.32	0.26
	T15	2.28	0.14	5.66	0.14	1.75	0.30	4.30	0.30	0.57	0.26	1.36	0.26
	T18	4.98	0.14	5.49	0.14	4.22	0.30	4.30	0.30	1.63	0.26	1.44	0.26
	Average	3.93		5.43		2.99		3.85		0.95		1.03
	2SD	2.32		0.65		2.44		1.45		1.59		1.14

^[1]Lpx may not be in the same chondrule as Ol.

^[2]Ol and Lpx are in different chondrules.

Table 4.

Oxygen isotope ratios of olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase in type 6 EOCs

Meteorite (type)	Area	δ18OVSMOW								δ17OVSMOW								Δ17O								Remarks
		Ol ± unc.		Lpx ± unc.		Hpx ± unc.		Pl ± unc.		Ol ± unc.		Lpx ± unc.		Hpx ± unc.		Pl ± unc.		Ol ± unc.		Lpx ± unc.		Hpx ± unc.		Pl ± unc.
Estacado (H6)	T01	4.06	0.18	4.32	0.18	4.22	0.18	5.54	0.18	3.19	0.35	3.45	0.35	3.15	0.35	3.97	0.35	1.08	0.35	1.20	0.35	0.95	0.35	1.09	0.35
	T03	3.88	0.18	4.08	0.18	3.71	0.18	5.65	0.18	3.28	0.35	3.41	0.35	3.12	0.35	4.36	0.35	1.26	0.35	1.28	0.35	1.20	0.35	1.43	0.35
	T04	3.86	0.16	4.53	0.16	3.85	0.16	5.70	0.16	2.38	0.45	3.12	0.45	3.12	0.45	3.57	0.45	0.38	0.40	0.77	0.40	1.12	0.40	0.61	0.40
	T09	3.78	0.16	4.28	0.16	3.92	0.16	5.77	0.16	2.74	0.45	3.32	0.45	3.02	0.45	3.55	0.45	0.77	0.40	1.09	0.40	0.98	0.40	0.55	0.40
	T10	3.98	0.16	4.27	0.16	3.80	0.16	5.58	0.16	2.67	0.45	3.07	0.45	2.37	0.45	3.74	0.45	0.60	0.40	0.85	0.40	0.40	0.40	0.84	0.40
	T11	3.86	0.17	4.38	0.17	3.98	0.17	5.72	0.17	2.41	0.44	2.92	0.44	2.67	0.44	3.78	0.44	0.40	0.42	0.64	0.42	0.60	0.42	0.81	0.42
	T15	4.01	0.17	4.25	0.17	4.00	0.17	5.46	0.17	3.08	0.44	2.94	0.44	3.32	0.44	4.19	0.44	0.99	0.42	0.73	0.42	1.24	0.42	1.35	0.42
	T18	3.79	0.17	4.42	0.17	4.02	0.17	5.78	0.17	2.71	0.44	2.65	0.44	2.53	0.44	3.79	0.44	0.74	0.42	0.35	0.42	0.44	0.42	0.79	0.42
	Average	3.90		4.32		3.94		5.65		2.81		3.11		2.91		3.87		0.78		0.86		0.87		0.93
	2SD	0.20		0.27		0.31		0.23		0.68		0.55		0.68		0.58		0.63		0.63		0.68		0.65
Lorton (L6)	T02	4.52	0.27	4.86	0.27	4.54	0.27	6.45	0.27	3.90	0.83	3.44	0.83	3.37	0.83	3.96	0.83	1.55	0.73	0.91	0.73	1.01	0.73	0.61	0.73
	T04	4.60	0.27	5.08	0.27	4.84	0.27	6.28	0.27	3.04	0.83	3.47	0.83	3.88	0.83	4.46	0.83	0.65	0.73	0.83	0.73	1.36	0.73	1.19	0.73
	T08	4.53	0.26	4.83	0.26	4.37	0.26	6.25	0.26	2.97	0.63	3.41	0.63	3.38	0.63	4.13	0.63	0.62	0.54	0.90	0.54	1.11	0.54	0.87	0.54
	T34	4.21	0.26	4.88	0.26	4.47	0.26	6.15	0.26	3.09	0.63	3.57	0.63	3.39	0.63	4.10	0.63	0.90	0.54	1.03	0.54	1.06	0.54	0.90	0.54
	T31	4.45	0.26	4.84	0.26	4.52	0.26	6.36	0.26	2.87	0.63	3.77	0.63	2.70	0.63	4.18	0.63	0.56	0.54	1.26	0.54	0.35	0.54	0.87	0.54
	T27	4.54	0.13	4.84	0.13	4.49	0.13	6.30	0.13	3.39	0.53	3.68	0.53	3.82	0.53	4.71	0.53	1.03	0.54	1.16	0.54	1.48	0.54	1.43	0.54
	T38-39	4.10	0.13	5.07	0.13	4.57	0.13	6.14	0.13	3.71	0.53	4.19	0.53	3.23	0.53	4.74	0.53	1.58	0.54	1.16	0.54	0.85	0.54	1.55	0.54
	T09-10	4.31	0.13	4.88	0.13	4.41	0.13	6.24	0.13	3.61	0.53	3.60	0.53	3.46	0.53	4.49	0.53	1.37	0.54	1.06	0.54	1.17	0.54	1.24	0.54
	Average	4.41		4.91		4.52		6.27		3.32		3.64		3.40		4.35		1.03		1.04		1.05		1.08
	2SD	0.36		0.21		0.29		0.20		0.77		0.51		0.73		0.59		0.84		0.31		0.69		0.64
Bruderheim (L6)	T15			4.96	0.16	4.62	0.16	4.75	0.16			3.72	0.46	3.60	0.46	3.55	0.46			1.14	0.45	1.15	0.45	1.09	0.45	[1]
	T14	4.40	0.16	4.80	0.16	4.07	0.16	5.45	0.16	3.79	0.46	3.72	0.46	3.25	0.46	3.98	0.46	1.13	0.45	1.22	0.45	1.13	0.45	1.15	0.45
	T10	4.56	0.16	4.70	0.16	4.49	0.16	6.09	0.16	4.03	0.46	3.50	0.46	3.76	0.46	4.64	0.46	1.65	0.45	1.06	0.45	1.43	0.45	1.48	0.45
	T17	3.07	0.14	4.87	0.14	4.33	0.14	5.89	0.14	2.50	0.51	3.42	0.51	3.53	0.51	4.18	0.51	0.90	0.50	0.89	0.50	1.28	0.50	1.12	0.50
	T03	4.32	0.14	5.22	0.14			5.97	0.14	3.31	0.51	4.15	0.51			4.21	0.51	1.06	0.50	1.44	0.50			1.11	0.50	[2]
	T08	4.67	0.14	4.11	0.14	4.46	0.14	4.93	0.14	3.90	0.51	3.20	0.51	3.89	0.51	3.21	0.51	1.47	0.50	1.06	0.50	1.57	0.50	0.65	0.50
	T18	4.27	0.16	4.90	0.16	4.19	0.16	6.22	0.16	3.65	0.47	3.05	0.47	2.90	0.47	4.23	0.47	1.44	0.45	0.50	0.45	0.72	0.45	1.00	0.45
	T04	4.42	0.16							3.46	0.47															[3]
	Average	4.24		4.80		4.39		5.51		3.52		3.54		3.61		3.96		1.28		1.05		1.31		1.10
	2SD	1.07		0.69		0.42		1.13		1.03		0.73		0.49		1.02		0.57		0.58		0.37		0.53
St. Mesmin (LL6)	T01	4.91	0.21			5.00	0.21	6.39	0.21	3.94	0.41			3.87	0.41	4.05	0.41	1.38	0.41			1.27	0.41	0.72	0.41	[4]
	T02	3.97	0.21	5.45	0.21	5.19	0.21	5.88	0.21	3.12	0.41	3.91	0.41	3.66	0.41	4.80	0.41	1.06	0.41	1.08	0.41	0.96	0.41	1.74	0.41
	T03	4.91	0.21	5.96	0.21	5.60	0.21	6.21	0.21	3.94	0.41	4.94	0.41	4.27	0.41	4.54	0.41	1.39	0.41	1.85	0.41	1.36	0.41	1.31	0.41
	T04	5.09	0.21	5.81	0.21	5.41	0.21	6.53	0.21	3.93	0.41	4.21	0.41	4.02	0.41	4.82	0.41	1.28	0.41	1.18	0.41	1.20	0.41	1.42	0.41
	T05	4.61	0.31	5.79	0.31	5.35	0.31	6.42	0.31	3.67	0.29	4.54	0.29	4.05	0.29	5.08	0.29	1.27	0.37	1.53	0.37	1.27	0.37	1.74	0.37	[5]
	T06	4.74	0.31	4.65	0.31	5.37	0.31	5.76	0.31	3.86	0.29	3.65	0.29	4.01	0.29	4.32	0.29	1.40	0.37	1.23	0.37	1.22	0.37	1.33	0.37
	T011	4.72	0.31	5.41	0.31	5.13	0.31	6.46	0.31	3.49	0.29	3.86	0.29	4.32	0.29	4.31	0.29	1.03	0.37	1.04	0.37	1.65	0.37	0.95	0.37
	Average	4.71		5.51		5.29		6.23		3.71		4.19		4.03		4.56		1.26		1.32		1.28		1.32
	2SD	0.72		0.95		0.40		0.61		0.62		0.97		0.45		0.72		0.31		0.62		0.41		0.76

^[1]Ol analyses overlapped Hpx.

^[2]Hpx analyses hit crack.

^[3]Bad SIMS pits except for Ol.

^[4]Bad SIMS pit for Lpx

^[5]Ol-Hpx are 200–300μm from Lpx-Pl.

Type 4 Chondrites

Oxygen isotope analyses of olivine and low-Ca pyroxene from type 4 chondrites are shown in Table 2 and Fig. 2. One analysis per each mineral was obtained among several different regions within each meteorite sample. For a given olivine and low-Ca pyroxene pair, their corresponding analyses were within 100 μm of each other, and in most cases were from the same chondrules. These data distribute widely above and parallel to the terrestrial fractionation line, similar to those observed in LL3 chondrite chondrules (Kita et al. 2010). Two olivine analyses from Soko Banja (LL4) plot significantly below the terrestrial fractionation line (Fig. 2d) with Δ¹⁷O values down to −8.1± 0.3‰; low-Ca pyroxene in the same chondrules do not show ¹⁶O-enrichments. Therefore, these olivine grains are likely ‘relict” olivine, representing precursor materials that remained unmelted during the final chondrule formation. Such ¹⁶O-rich relict olivine has been reported in chondrules from LL3 chondrites, but are not frequently observed (Kita et al. 2010).

Fig. 2.

Oxygen 3-isotope ratios in olivine and low-Ca pyroxene in type 4 chondrites. (a) Forest Vale (H4), (b) Bjurböle (L4), (c-d) Soko Banja (LL4). TFL: terrestrial fractionation line. ECL: equilibrated ordinary chondrite line (Clayton et al. 1991). Bulk H, L, LL chondrite mass-dependent fractionation lines are also shown as δ¹⁷O = 0.52× δ¹⁸O + Δ¹⁷O_Group, where Δ¹⁷O_Group values are 0.73‰, 1.07‰, and 1.26‰ for bulk H, L, and LL chondrites, respectively.

Type 5-6 Chondrites

Figs. 3 and 4 show individual spot data from types 5 and 6 chondrites. Oxygen isotope data are more scattered in type 5 chondrites (Fig. 3) than unbrecciated type 6 chondrites (Fig. 4), but are less scattered than type 4 chondrites (Fig. 2). All individual meteorite analyses plot along respective bulk H, L, or LL chondrite bulk mass-dependent fractionation lines (slope ~0.52), regardless of their variability in δ¹⁸O. The δ¹⁸O values are systematically higher in plagioclase than in olivine and pyroxene, which show significant overlap in Figs. 3–4. Among type 5 chondrites, Allegan (H5; Fig. 3a) shows the largest variation in oxygen isotope ratios, while Mifflin (L5; Fig. 3c) shows variation that is systematically less than others. For each mineral phase, δ¹⁸O and δ¹⁷O values from two unshocked and unbrecciated type 6 chondrites, Estacado (H6) and Lorton (L6), are homogeneous within analytical uncertainties; however, there are δ¹⁸O variabilities of up to 2.5‰ among mineral phases (Fig. 4a–b). The δ¹⁸O values of olivine and high-Ca pyroxene are indistinguishable and only slightly lower than those of low-Ca pyroxene (≤0.5‰). These data are comparable to those reported for Guareña (H6) and St. Séverin (LL6) by Nakashima et al. (2013).

Fig. 3.

Oxygen 3-isotope ratios in olivine (Ol), low-Ca pyroxene (Lpx), high-Ca pyroxene (Hpx), and plagioclase (Pl) in type 5 chondrites. (a) Allegan (H5), (b) Ausson (L5), (c) Mifflin (L5), (d) Tuxtuac (LL5). TFL: terrestrial fractionation line. ECL: equilibrated ordinary chondrite line (Clayton et al. 1991). Bulk H, L, LL chondrite mass-dependent fractionation lines are also shown as δ¹⁷O = 0.52× δ¹⁸O + Δ¹⁷O_Group, where Δ¹⁷O_Group values are 0.73‰, 1.07‰, and 1.26‰ for bulk H, L, LL chondrites, respectively.

Fig. 4.

Oxygen 3-isotope ratios in silicate minerals in type 6 chondrites. (a) Estacado (H6), (b) Lorton (L6), (c) Bruderheim (L6), (d) St. Mesmin (LL6). TFL: terrestrial fractionation line. ECL: equilibrated ordinary chondrite line (Clayton et al. 1991). Bulk H, L, LL chondrite mass-dependent fractionation lines are also shown as δ¹⁷O = 0.52× δ¹⁸O + Δ¹⁷O_Group, where Δ¹⁷O_Group values are 0.73‰, 1.07‰, and 1.26‰ for bulk H, L, LL chondrites, respectively. The four phases included are olivine (Ol), low-Ca pyroxene (Lpx), high-Ca pyroxene (Hpx), and plagioclase (Pl).

For unbrecciated EOCs, there is a general trend that oxygen isotope ratios of individual mineral phases become more homogeneous with increasing petrologic type, including those of olivine and pyroxene in type 4 chondrites. However, data are more scattered in brecciated type 6 chondrites (Fig. 4c–d), showing variability that is somewhat similar to type 5 chondrite data (Fig. 3).

Comparison to mineral separate data

Clayton (1993) reported olivine, pyroxene, and plagioclase mineral separate oxygen isotope data (δ¹⁸O) from several EOCs, including Bjurböle (L4), Allegan (H5), Bruderheim (L6) and Estacado (H6) that are studied here. Fig. 5 compares the δ¹⁸O values of SIMS data from this study with those of mineral separates by Clayton (1993). We note that pyroxene separates in Clayton (1993) are dominated by low-Ca pyroxene mixed with minor high-Ca pyroxene. In contrast, we individually analyzed these minerals by SIMS. Most homogeneous Estacado (H6) data are consistent with those of mineral separate data, similar to those reported for St. Severin (LL6) by Nakashima et al. (2013). Although SIMS data are heterogeneous in three other EOCs, Bjurböle (L4), Allegan (H5), and Bruderheim (L6), their respective mineral separate data and SIMS data show a similar tendency of increasing δ¹⁸O in the order of olivine, (low-Ca) pyroxene, and plagioclase.

Fig 5.

The δ¹⁸O values of minerals from the four EOCs in this study compared to those of mineral separates from Clayton (1993). Clayton (1993) data are shown as grey shading and include plagioclase (Pl), pyroxene that is dominated by low-Ca pyroxene (Lpx), and olivine (Ol). Other data are SIMS analyses including high-Ca pyroxene (Hpx) from this study, showing individual spot data with 2SD error bars. Only olivine and low-Ca pyroxene SIMS data are shown for Bjurböle (L4).

DISCUSSION

Application of Oxygen Isotope Thermometry to EOC Minerals

In this paper, fractionations in δ¹⁸O among four minerals (olivine, low-Ca pyroxene, high-Ca pyroxene, plagioclase) for type 5 and 6 chondrites, and among two minerals (olivine and low-Ca pyroxene) for type 4 chondrites, are examined. If the four minerals were in oxygen isotope equilibrium, their δ¹⁸O values would differ according to equilibrium fractionation factors at a given metamorphic temperature, while their Δ¹⁷O values would be identical. Thus, the temperature of metamorphism for each meteorite can be determined. Throughout the discussion, equilibrium fractionation factors of oxygen isotopes from Clayton and Kieffer (1991) for olivine, high-Ca pyroxene, and plagioclase (assuming An₁₁Ab₈₉), and from Rosenbaum et al. (1994) for low-Ca pyroxene-olivine are used. The solid solution effect in plagioclase is minimal due to the small variability in plagioclase compositions (An₆ to An₁₅; Table A4), corresponding to changes in the fractionation factor by ~0.05‰ at 800°C.

In order to achieve oxygen isotope equilibrium among minerals, oxygen isotopes should be efficiently exchanged between adjacent minerals through recrystallization or volume diffusion. In asteroidal parent bodies, diffusive oxygen isotope exchange in minerals would have stopped at the closure temperature for each mineral, which depends on various factors, but mainly on the self-diffusion rate of individual minerals, grain sizes, and cooling rates (Dodson 1973). Nakashima et al. (2013) estimated closure temperatures of olivine, pyroxene, and plagioclase by using the equation given by Dodson (1973):

$$T_c=\frac{E/R}{ln\left[\frac{-AR{T_c}^2(D_0/a^2)}{E(dT/dt)}\right]}$$

In the above equation, T_c is the closure temperature (K), D₀ is the pre-exponential factor from the Arrhenius relation for diffusion (m²/s), E is activation energy (kJ/mol), R is the gas constant (J/mol/K), a is the radius of a sphere (m), A is the diffusion anisotropy parameter that is assigned to 55 for a sphere, and (dT/dt) is the cooling rate (°C/s). Nakashima et al. (2013) estimated T_c for olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase by applying experimentally determined self-diffusion rates of oxygen in San Carlos olivine, diopside, and albite (Gérard and Jaoul 1989; Ryerson and McKeegan 1994; Mathews et al. 1994), assuming average chondrite grain sizes (100 μm for olivine and low-Ca pyroxene and 10 μm for high-Ca pyroxene and plagioclase), and cooling rates (1, 10, 100, 1000°C/Myr) that are appropriate for ordinary chondrite parent bodies. Table 5 lists estimated closure temperatures, which are updated from values employed by Nakashima et al. (2013) because we found unit conversion errors on D₀ for olivine and albite. In addition, the dependence of the diffusion coefficient on oxygen fugacity with an exponent m, D₀ = D (fO₂ )^m exp(−E/RT) (Gérard and Jaoul 1989; Ryerson et al. 1989), is considered in the calculation of closure temperatures of olivine. We assume the oxygen fugacity of EOCs to be 2.5 log units below iron-wüstite (IW) buffer, which is within the range of IW–2 to IW–3 reported in literature (McSween and Labotka 1993; Kessel et al. 2004; Schrader et al. 2016), and replace D₀ in the equation by D₀×(fO₂)^m. Diffusion coefficient parameters for olivine are applied from Ryerson et al. (1989) that include experiments on the IW buffer. The estimated closure temperatures are 860–1090°C for olivine, 860–1030°C low-Ca pyroxene, 760–910°C for high-Ca pyroxene (10 μm), and 160–310°C for albite. These values are the same for pyroxene, but slightly higher for olivine and a significantly lower for albite compared to those in Nakashima et al. (2013). There are several aspects of uncertainties in the estimated closure temperatures given in Table 5. For example, the estimated closure temperatures would change by 50–100°C if diffusion coefficients varied by 1–2 log units, which should be considered because diffusion coefficient parameters are applied from experimental data that are often extrapolated to lower temperatures. For olivine, diffusion coefficients determined by the experiments of Gérard and Jaoul (1989) and Ryerson et al. (1989) are very similar over the range of temperatures (1200–1500°C) and oxygen fugacity (10^–6 to 10^–1 Pa) explored by both studies; however, Gérard and Jaoul (1989) obtained a higher activation energy of 318 KJ/mole and a stronger oxygen fugacity dependency (m =0.34) compared to those by Ryerson et al. (1989). Closure temperature estimates for olivine would increase by 50–90°C compared to the values listed in Table 5 by applying diffusion coefficient parameters from Gérard and Jaoul (1989). However, closure temperatures of olivine do not change more than ~20°C for the change of oxygen fugacity by 1 log unit. For pyroxene, oxygen fugacity dependency was not observed for diopside (Pacaud et al. 1999), though the closure temperature estimate for low-Ca pyroxene may have unseen errors as diffusion coefficient data are not available for orthopyroxene and experimental data for diopside are used.

Table 5.

Closure temperatures for diffusive exchange of oxygen isotopes for olivine, pyroxene, and albite at various cooling rates.

	D0 (m2/s)	E (kJ/mol)	mb	grain size (μm)	Log fO2 (Pa)c	Tc (°C) at cooling rates (°C/Myr)
						1	10	100	1000
Olivinea	2.6×10–10	266	0.21	100	–11 to –15	871	938	1013	1098
Low Ca-pyroxenea	4.3×10–4	457		100		852	905	964	1028
High Ca-pyroxenea	4.3×10–4	457		10		758	803	852	905
Albitea	2.0×10–16	90		10		163	204	253	312

^aDiffusion coefficient parameters (D₀, E, m) are from Ryerson et al. (1989), Ryerson and McKeegan (1994), and Matthews et al. (1994).

^bExponent on the oxygen fugacity dependency to the oxygen diffusion rate of olivine.

^cCorresponding to 2.5 log unit lower than IW buffer at the closure temperatures of olivine.

The closure temperatures estimated for olivine and low-Ca pyroxene are comparable to or higher than the peak metamorphic temperatures in EOCs (~900°C; Slater-Reynolds and McSween 2005), but those for high-Ca pyroxene and plagioclase are lower than the peak metamorphic temperatures. As suggested in Nakashima et al. (2013), oxygen isotope fractionation between high-Ca pyroxene and plagioclase is likely to reflect the closure temperature of high-Ca pyroxene, while olivine and low-Ca pyroxene would not have exchanged oxygen isotopes efficiently during parent body metamorphism.

In addition to volume diffusion, oxygen isotope exchange would have been facilitated by grain boundary diffusion among minerals not adjacent to each other, which is orders of magnitude faster than volume diffusion, especially in the presence of fluid (Eiler et al. 1993, Valley 2001). A small amount of aqueous fluid may have existed in the early stage of ordinary chondrite parent body metamorphism, as indicated from the presence of phyllosilicates and magnetites in the least metamorphosed UOCs (Alexander et al. 1989; Choi et al. 1998; Grossman et al. 2000). However, ordinary chondrites are generally considered to have been anhydrous during thermal metamorphism, meaning grain boundary diffusion would have been limited. If grain boundary diffusion was a limiting factor, there would likely be oxygen isotope heterogeneity across a sample, even if touching mineral grains were in equilibrium.

Limited Oxygen Isotope Exchange in Type 4

In type 4 chondrites, olivine and pyroxene δ¹⁸O values distribute similarly to those in chondrules found in LL3.0-3.1 chondrites (Kita et al. 2010), indicating that oxygen isotope exchange was limited during thermal metamorphism of type 4 ordinary chondrites. Among H4 and L4 data, the δ¹⁸O values of olivine and low-Ca pyroxene from the same chondrules show nearly a 1:1 correlation (Fig. 6) similar to those in LL3.0-3.1 chondrites (Kita et al. 2010), suggesting that oxygen isotope ratios of the original chondrules were not disturbed in these minerals. Furthermore, in the Soko Banja LL4 chondrite, significantly ¹⁶O-rich olivine grains are likely derived from relict olivine in chondrules, of which distinct isotope signatures were not erased during thermal metamorphism. Thus, even though the olivine and pyroxene compositions (Fa and Fs in Fig. 7) are homogeneous and represent H, L, and LL averages (Brearley and Jones, 1998), oxygen isotope exchange hardly occurred in type 4 olivine and low-Ca pyroxene.

Fig. 6.

Plot of δ¹⁸O in olivine (Ol) versus δ¹⁸O in low-Ca pyroxene (Lpx) for mineral pairs within 100 μm areas in type 4 chondrites. Error bars represent external reproducibility of 0.3‰.

Fig. 7.

Mineral compositions of olivine (Fa) and low-Ca pyroxene (Fs) of individual chondrites in this study. Data are calculated from electron microprobe data in Supplemental Information 3. The open symbols show compositions of co-existing olivine and low-Ca pyroxene in the three type 4 chondrites studied. Filled squares and circles show mean values in types 5 and 6 chondrites studied. The ranges of H, L, and LL chondrite data are shown as grey open ovals (data from Brearley and Jones 1998).

In general, type 4 chondrites are distinguished from type 5 chondrites by petrologic evidence for lower-degree thermal metamorphism, including small sizes of secondary plagioclase (<2 μm) and well-defined chondrules (Van Schmus and Wood 1967). However, as discussed in Harrison and Grimm (2010), results from mineral thermometry do not show clear distinctions between the two types. Unlike type 6 chondrites, type 4 chondrites and some type 5 chondrites did not reach chemical cation equilibrium when comparing two pyroxene compositions, which implies types 4 and 5 chondrite peak metamorphic temperatures were below ~850°C (Slater-Reynolds and McSween 2005). Kessel et al. (2007) argued that peak metamorphic temperatures of type 4–6 chondrites should have been higher than ~730°C, based on olivine-spinel thermometry. In contrast, and as shown in Figs. 2 and 3, type 4 chondrites show oxygen isotope systematics of olivine and pyroxene that are clearly different from those in type 5 chondrites. At this temperature range (730–850°C), oxygen isotope ratios in coarse (≥100μm) olivine and pyroxene would not be homogenized due to slow diffusion, but olivine and pyroxene compositions (Fa and Fs) would have been homogenized because Fe-Mg inter-diffusion coefficients in these minerals are orders of magnitude higher than oxygen isotope diffusion coefficients (e.g., Chakraborty 1997; Ganguly and Tazzoli 1994; Cherniak and Dimanov 2010).

Fractionation between High Ca-Pyroxene and Plagioclase in Type 5–6

In types 5 and 6 chondrites, oxygen isotope ratios of olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase plot along the mass dependent fractionation line of their respective bulk meteorite values, showing that oxygen isotopes were redistributed in the parent bodies. In Fig. 8, δ¹⁸O values of coexisting high-Ca pyroxene and plagioclase in types 5–6 chondrites are compared, and plot along a linear trend if they are in equilibrium at specific temperatures. In type 5 chondrites, multiple data per meteorite distribute along slope ~1 lines (Fig. 8a), suggesting high-Ca pyroxene and plagioclase exchanged locally, even though these minerals exhibit oxygen isotope heterogeneity on the scale of a meteorite section.

Fig. 8.

Plot of δ¹⁸O _Hpx versus δ¹⁸O _Pl for mineral pairs within 100 μm areas in types 5 and 6 chondrites. (a) type 5, (b) unshocked and non-brecciated type 6, (c) shocked and brecciated type 6. Equilibrium fractionation lines labeled with temperatures (T_eq) are calculated from Clayton and Kieffer (1991). The Inset in (a): as an example, the uncertainty in calibration of oxygen isotope fractionation at 800 °C is calculated (±0.16‰), according to experiments by Chiba et al. (1989).

It is important to consider that while grain boundary diffusion in anhydrous systems may be slow, it is not completely halted (Eiler et al. 2003; Valley 2001). With this in mind, type 5 chondrite data (Fig. 8a) show correlated high-Ca pyroxene and plagioclase δ¹⁸O values, indicating that minerals are heterogeneous on a mm-scale, but that adjacent minerals are concordant with equilibrium fractionation. As such, these data are consistent with a dry environment during thermal metamorphism, which severely limited grain-boundary diffusion.

Apparent oxygen isotope temperatures, estimated from high-Ca pyroxene - plagioclase pairs, vary significantly: 900–1200°C for Tuxtuac (LL5), 550–900°C for Ausson (L5), and 400–800°C for Allegan (H5). Unbrecciated type 6 chondrite data plot in a narrow range (Fig. 8b), corresponding to T_eq = 700–900°C. These temperatures correspond to the closure temperature of oxygen isotope diffusion in high Ca-pyroxene, increasing with cooling rate after peak metamorphic temperatures (Nakashima et al. 2013). A higher estimated closure temperature in Tuxtuac (LL5) indicates faster cooling than those in other type 5 chondrites, which is consistent with nm-scale K-feldspar exsolution in albite observed by transmission electron microscopy (Jones and Brearley, 2011). Orthopyroxene in Tuxtuac (LL5) often contains relict monoclinic structure (Noguchi et al. 2013), which is uncommon in type 5 chondrites (Supporting Information 6). Such twinning in Tuxtuac also indicates that the meteorite cooled from high temperature faster than the other type 5 chondrites studied and phase transition from monoclinic to orthorhombic was incomplete.

Regarding brecciated type 6 chondrites, the data show significant scatter, mostly from δ¹⁸O values of plagioclase, and towards smaller fractionations in Δ¹⁸O_Pl-Hpx, corresponding to higher temperatures (Fig. 8c). Bruderheim is a fragmental breccia and is moderately shocked (S4). In turn, Bruderheim plagioclase shows a greater variation in chemistry and oxygen isotope ratios, suggesting impact melting. The brecciated St. Mesmin (LL6) meteorite also shows less oxygen isotope homogenization than the other type 6 chondrites (Fig. 4). Mifflin (L5) is weakly shocked (S2), but contains μm-scale shock-melted veins (Kita et al. 2013). Data from Mifflin also show a similar trend towards higher T_eq (Fig. 8a). If it is assumed plagioclase δ¹⁸O values were homogeneous before impact, similar to those from unshocked type 6 chondrites (Fig. 8b), brecciation and/or impact processes may have preferentially modified plagioclase isotope ratios, due to faster oxygen isotope diffusion rates in plagioclase at elevated temperatures during impact-generated heat.

Oxygen Isotope Systematics among Mineral Phases

Homogeneous oxygen isotope ratios among individual mineral phases in unbrecciated type 6 chondrites suggest diffusional isotope exchange during thermal metamorphism, which resulted in equilibrium fractionation. To test the degree of oxygen isotope equilibrium among olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase in types 5 and 6 chondrites, plots of isotope fractionation between mineral pairs are compared with those of equilibrium fractionation at various temperatures in Figs. 9 and 10. Figure 9 compares the consistency of oxygen isotope ratios among three minerals, low-Ca pyroxene, high-Ca pyroxene, and plagioclase, by plotting the fractionation of Pl-Hpx (Δ¹⁸O_Pl-Hpx) and Pl-Lpx (Δ¹⁸O_Pl-Lpx) in the same analysis areas. Unbrecciated type 6 chondrite data in Fig. 9 plot along concordant lines with corresponding temperatures of 600–800°C, while Tuxtuac (LL5) data, except for one point, plot along the line at higher temperatures of 900–1000°C. These data suggest that unbrecciated pyroxene and plagioclase exchanged at temperatures of 600–1000°C. For other type 5 and brecciated type 6 chondrite data, many plot off the concordant line, indicating that low-Ca pyroxene might not exchange its oxygen isotopes as efficiently as high-Ca pyroxene and plagioclase, or that isotope systematics were disturbed by the impact generated heating.

Fig. 9.

Oxygen isotope fractionation among low Ca pyroxene (Lpx), high-Ca pyroxene (Hpx), and plagioclase (Pl) in type 5 and 6 chondrites, compared to concordant equilibrium fractionation factors. Equilibrium fractionation lines labeled with temperatures (°C) are calculated from Clayton and Kieffer (1991) and Rosenbaum et al. (1994). Inset in (a): examples of uncertainties in calibration of oxygen isotope fractionation at 800 °C are calculated to be ±0.16‰ and ±0.48‰ for the x- and y-axis, respectively, based on experiments by Chiba et al. (1989) and Rosenbaum et al. (1994).

Fig. 10.

Oxygen isotope fractionation among olivine (Ol), low-Ca pyroxene (Lpx), high-Ca pyroxene (Hpx), and plagioclase (Pl) in type 5 and 6 chondrites, compared to concordant equilibrium fractionation factors. Equilibrium fractionation lines labeled with temperatures (°C) are calculated from Clayton and Kieffer (1991) and Rosenbaum et al. (1994). The Inset in (a): examples of uncertainty in calibration of oxygen isotope fractionation at 800 °C are calculated (±0.16‰ and ±0.37‰ for the x- and y-axis, respectively) based on experiments by Chiba et al. (1989) and Rosenbaum et al. (1994).

In order to evaluate the degree of equilibrium of olivine relative to pyroxene and plagioclase, plots of Pl-Hpx versus Lpx-Ol pairs were constructed (Fig. 10). They show that the fractionations of Lpx-Ol (Δ¹⁸O_Lpx-Ol) in most EOCs are not consistent with those of Pl-Hpx and that the data plot below the concordant line. Tuxtuac (LL5) is the lone exception, as most data plot on a concordant line at T ~1000°C despite not obtaining clean analyses of the four minerals in all areas. For the other meteorites, the Δ¹⁸O_Lpx-Ol values are less than 0.5‰, which is too small compared to equilibrium fractionation values below 1000°C (Clayton and Kieffer 1991; Rosenbaum et al. 1994). Even within target areas where all four minerals share grain boundaries, they do not show consistent equilibrium fractionation. Similar results were reported by Nakashima et al. (2013) for Guarenña (H6) and St. Séverin (LL6) chondrites, who suggested that Δ¹⁸O_Pl-Hpx represents recrystallization temperatures of these minerals, while olivine and low-Ca pyroxene were not in equilibrium due to slow diffusion rates of oxygen at metamorphic temperatures.

As discussed earlier, closure temperatures for oxygen isotope diffusional exchange may be highest in olivine among four minerals (Table 5). If EOCs were heated to higher temperatures for longer periods of time, then it may be expected that O-isotopes of olivine would become more homogenized. Indeed, the 2 SD value of δ¹⁸O in olivine grains decrease as a function of increasing petrologic type: ~2‰ for type 4 chondrites, 0.5–1‰ for type 5 chondrites, and ≥0.3‰ (the same as analytical uncertainties) for unbrecciated type 6 chondrites (Table 2–4). Regarding type 6 chondrites, it is possible that upon cooling after peak metamorphic temperatures, diffusional isotope exchange would have closed first in olivine at temperatures >800°C. In turn, the two-pyroxenes and plagioclase may have continued to exchange oxygen isotopes at lower temperatures, in which case high-Ca pyroxene could have acquired lower δ¹⁸O, with values becoming similar to those of olivine. In type 5 chondrites, the olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase did not collectively homogenize at peak metamorphic temperatures, although pyroxene and plagioclase might have locally exchanged at lower temperatures. In the case of Tuxtuac (LL5), δ¹⁸O values in olivine show smaller 2SD values (0.5‰) than those of other type 5 chondrites (0.7–1‰) and δ¹⁸O values in olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase are nearly concordant at T_eq ~1000°C. These data suggest that Tuxtuac experienced higher peak metamorphic temperatures and faster cooling rates than other type 5 chondrites, which is consistent with previous studies indicating Tuxtuac experienced fast cooling (Jones and Brearley 2011).

Thermal History of Ordinary Chondrite Parent Bodies

The inter-mineral oxygen isotope data in this study are consistent with a scenario in which the ordinary chondrite parent bodies initially started from an onion-shell structure that was later disturbed by impacts. The level of δ¹⁸O homogenization in olivine correlates well with petrologic types, from highly heterogeneous in type 4 chondrites to completely homogenized (within the SIMS analytical precision of ~0.3‰) in unshocked type 6 chondrites. These data imply that peak temperatures were the highest in type 6 chondrites and that temperatures were successively lower in type 5 and type 4 chondrites, which matches predictions from the onion-shell model. The temperatures estimated from the high-Ca pyroxene - plagioclase pairs in unshocked type 6 chondrites are mostly 700–900°C, while type 5 chondrites are estimated to have experienced a larger range of temperatures, from 600–1000°C. Systematic changes in the equilibrium temperatures of high-Ca pyroxene and plagioclase pairs with petrologic types are not observed. Oxygen isotope ratios in individual mineral grains of Tuxtuac (LL5) are not homogeneous, although co-existing olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase have fractionations consistent with high-temperature equilibrium (~1000°C). The observation of exsolution textures among feldspars in Tuxtuac suggests fast cooling, which would have halted O-isotopic exchange from high temperatures, as proposed by Jones and Brearley (2011). In contrast, Allegan (H5) has highly variable values of δ¹⁸O and δ¹⁷O (Fig. 3a) and larger inter-mineral fractionations indicating the lowest temperature of ~600°C for the high-Ca pyroxene - plagioclase pair. As such, Allegan may not have experienced high temperatures and may have cooled slowly. Thus, the results suggest a wide range of thermal histories experienced by type 5 chondrites.

CONCLUSIONS

Oxygen three-isotope ratios were measured in olivine, low- and high-Ca pyroxene, and plagioclase in 11 equilibrated ordinary chondrite spanning all groups (H, L, LL) and petrologic types (4–6), and including some that were either shocked (≥S3) or brecciated. Sets of olivine, low-Ca and high-Ca pyroxene, and plagioclase that are touching or spatially close to each other (≤100 μm) were selected from multiple sub-areas in each meteorite. Systematic changes in oxygen isotope fractionations among these mineral sets are observed when comparing type 4 to type 6 chondrites. In addition, shocked and brecciated type 6 chondrites have different inter-mineral oxygen isotope systematics relative to unshocked and unbrecciated type 6 chondrites.

1. Olivine and low-Ca pyroxene in type 4 ordinary chondrites are heterogeneous in oxygen isotope ratios within individual grains and show a range of δ¹⁸O values similar to chondrules in type 3 chondrites. The peak metamorphic temperatures of type 4 samples were not high enough to cause oxygen isotope equilibration in olivine and low-Ca pyroxene.

2. δ¹⁸O values of high-Ca pyroxene and plagioclase in type 5 ordinary chondrites are locally consistent with those of equilibrium fractionation on the 100 μm scale. The apparent temperatures of oxygen isotope equilibration between high-Ca pyroxene and plagioclase record closure of oxygen exchange by diffusion or recrystallization of these minerals, and range from 600°C to 1000°C.

3. Oxygen isotope ratios of individual mineral grains in type 6 chondrites are homogeneous within the precision of SIMS analyses (±0.3 ‰ 2SD). They are interpreted to have experienced isotope exchange at peak metamorphic temperatures that likely exceeded 800°C. If assuming a slower oxygen diffusion rate in olivine, relative to pyroxene and plagioclase, it is possible that diffusional oxygen isotope exchange in olivine closed at temperatures exceeding 800°C. At temperatures below 800°C low-Ca pyroxene, high-Ca pyroxene, and plagioclase would have continued to exchange oxygen isotopes. If true, this would explain why olivine is observed to be out of equilibrium with low-Ca pyroxene, high-Ca pyroxene and plagioclase in type 6 chondrites.

4. Brecciated and shock metamorphosed type 6 chondrites have heterogeneous oxygen isotope ratios similar to type 5 chondrites. Impact processes induced localized melting and promoted isotope exchange among minerals, as well as mixing of materials with slightly different isotope ratios.

These results are generally consistent with the onion-shell model as a mechanism for the genesis and evolution of parent bodies. However, the variety of oxygen isotope systematics observed among type 5 chondrites could also be explained by major impact processes during the cooling of the parent bodies.

Supplementary Material

Supporting information

Supporting Information 1: Polished sections of 11 EOCs analyzed for SIMS.

Supporting Information 2: BSE images of individual SIMS analyses areas for type 4 chondrites.

Supporting Information 3: BSE images of individual SIMS analyses areas for type 5 chondrites.

Supporting Information 4: BSE images of individual SIMS analyses areas for type 6 chondrites.

Supporting Information 5. Electron microprobe data of silicate minerals in 11 EOCs.

Supporting Information 6: Optical microscope photographs of orthopyroxene in type 5 chondrites.

Table 3.

Oxygen isotope ratios of olivine, low-Ca pyroxene, high-Ca pyroxene, and plagioclase in type 5 EOCs

Meteorite (type)	Area	δ18OVSMOW								δ17OVSMOW								Δ17O								Remarks
		Ol ± unc.		Lpx ± unc.		Hpx ± unc.		Pl ± unc.		Ol ± unc.		Lpx ± unc.		Hpx ± unc.		Pl ± unc.		Ol ± unc.		Lpx ± unc.		Hpx ± unc.		Pl ± unc.
Allegan (H5)	T11	4.33	0.27	4.21	0.27	3.86	0.27	5.41	0.27	3.48	0.46	3.19	0.46	2.91	0.46	3.99	0.46	1.22	0.44	1.00	0.44	0.91	0.44	1.18	0.44
	T20	4.22	0.27	4.13	0.27	3.61	0.27	5.94	0.27	3.16	0.46	3.07	0.46	2.42	0.46	3.92	0.46	0.97	0.44	0.93	0.44	0.54	0.44	0.83	0.44
	T07	4.51	0.30	4.30	0.30	3.09	0.30	5.24	0.30	2.51	0.46	3.42	0.46	2.56	0.46	3.82	0.46	0.16	0.50	1.18	0.50	0.95	0.50	1.09	0.50
	T09	4.99	0.30	4.45	0.30	3.88	0.30	5.62	0.30	3.34	0.46	3.36	0.46	2.88	0.46	3.99	0.46	0.75	0.50	1.04	0.50	0.86	0.50	1.07	0.50
	T16	3.77	0.30	4.22	0.30	2.71	0.30	4.86	0.30	2.79	0.46	2.40	0.46	2.55	0.46	3.23	0.46	0.83	0.50	0.21	0.50	1.14	0.50	0.70	0.50
	T14	3.53	0.25	4.11	0.25	3.79	0.25	5.65	0.25	2.25	0.45	3.23	0.45	2.76	0.45	3.63	0.45	0.42	0.51	1.10	0.51	0.79	0.51	0.69	0.51
	T04	4.66	0.25	5.20	0.25	4.56	0.25	6.71	0.25	3.10	0.45	3.45	0.45	3.64	0.45	5.44	0.45	0.68	0.51	0.74	0.51	1.26	0.51	1.96	0.51
	T02			4.42	0.25	3.44	0.25	6.89	0.25			3.69	0.45	2.99	0.45	4.87	0.45			1.39	0.51	1.21	0.51	1.28	0.51	[1]
	Average	4.29		4.38		3.62		5.79		2.95		3.23		2.84		4.11		0.72		0.95		0.96		1.10
	2SD	1.01		0.71		1.12		1.40		0.90		0.76		0.76		1.42		0.70		0.71		0.48		0.82
Ausson (L5)	T01	4.84	0.15	5.01	0.26	4.03	0.26	5.24	0.26	3.72	0.71	3.80	0.35	3.54	0.35	3.95	0.35	1.20	0.73	1.20	0.39	1.44	0.39	1.23	0.39
	T02	4.75	0.26	5.19	0.26	4.63	0.26	6.26	0.26	4.08	0.35	4.59	0.35	3.83	0.35	4.57	0.35	1.61	0.39	1.89	0.39	1.42	0.39	1.31	0.39
	T03	4.33	0.26	4.72	0.26	4.08	0.26	5.89	0.26	3.44	0.35	3.86	0.35	2.88	0.35	5.05	0.35	1.19	0.39	1.41	0.39	0.77	0.39	1.99	0.39
	T04	3.89	0.26	5.40	0.26	4.45	0.26	5.84	0.26	3.63	0.35	4.23	0.35	3.48	0.35	4.53	0.35	1.61	0.39	1.42	0.39	1.16	0.39	1.50	0.39
	T05	4.26	0.26	4.55	0.26	4.47	0.26	6.00	0.26	3.60	0.60	3.80	0.60	3.66	0.60	4.51	0.60	1.39	0.69	1.43	0.69	1.34	0.69	1.39	0.69
	T09	4.74	0.26	5.19	0.26	3.57	0.26	5.94	0.26	3.80	0.60	4.10	0.60	3.07	0.60	4.78	0.60	1.34	0.69	1.41	0.69	1.22	0.69	1.68	0.69	[2]
	T10	4.75	0.26	5.10	0.26	4.71	0.26	5.96	0.26	3.22	0.60	3.49	0.60	3.71	0.60	4.24	0.60	0.75	0.69	0.84	0.69	1.26	0.69	1.14	0.69
	T11	4.61	0.26	5.22	0.26	4.70	0.26	6.03	0.26	3.84	0.60	3.91	0.60	4.12	0.60	4.11	0.60	1.44	0.69	1.20	0.69	1.68	0.69	0.98	0.69	[3]
	Average	4.52		5.04		4.33		5.90		3.67		3.97		3.54		4.47		1.32		1.35		1.29		1.40
	2SD	0.66		0.56		0.81		0.59		0.53		0.66		0.80		0.71		0.56		0.59		0.53		0.64
Mifflin (L5)	T01	4.28	0.29	4.86	0.29			5.56	0.29	3.15	0.37	3.38	0.37			3.77	0.37	0.93	0.37	0.85	0.37			0.87	0.37	[4]
	T02	4.58	0.29	4.72	0.29	4.52	0.29	5.59	0.24	3.34	0.37	3.38	0.37	3.67	0.37	4.22	0.58	0.95	0.37	0.93	0.37	1.32	0.37	1.31	0.37
	T03	4.48	0.29	4.97	0.29	4.49	0.29	5.91	0.29	3.43	0.37	3.76	0.37	3.54	0.37	3.93	0.37	1.10	0.37	1.18	0.37	1.20	0.37	0.85	0.37
	T04			4.91	0.29	4.01	0.29	5.74	0.29			3.98	0.37	3.10	0.37	4.20	0.37			1.42	0.37	1.01	0.37	1.22	0.37	[5]
	T06	3.51	0.24	4.89	0.24	4.62	0.24	5.40	0.24	2.63	0.58	3.79	0.58	4.19	0.58	3.94	0.58	0.81	0.60	1.25	0.60	1.79	0.60	1.13	0.60	[6]
	T09	4.23	0.24	4.89	0.24	4.75	0.24	5.40	0.24	2.90	0.58	3.41	0.58	3.10	0.58	3.94	0.58	0.70	0.60	0.87	0.60	0.63	0.60	1.13	0.60
	T13	4.32	0.24	4.81	0.24	4.67	0.24	5.63	0.24	3.00	0.58	3.20	0.58	3.60	0.58	4.24	0.58	0.75	0.60	0.70	0.60	1.17	0.60	1.32	0.60
	Average	4.23		4.86		4.51		5.60		3.08		3.56		3.53		4.03		0.87		1.03		1.19		1.12
	2SD	0.76		0.16		0.52		0.37		0.59		0.57		0.81		0.37		0.30		0.52		0.76		0.38
Tuxtuac (LL5)	1-T01	4.55	0.18	5.36	0.18	5.07	0.18	6.19	0.18	3.85	0.44	4.22	0.44	3.82	0.44	4.81	0.44	1.49	0.41	1.43	0.41	1.19	0.41	1.59	0.41	[7]
	1-T04	5.00	0.18			5.23	0.18	6.27	0.18	3.91	0.44			3.99	0.44	4.77	0.44	1.31	0.41			1.27	0.41	1.51	0.41	[8]
	1-T05	4.70	0.18	5.50	0.18	4.12	0.18	5.01	0.18	3.71	0.44	3.77	0.44	3.48	0.44	3.85	0.44	1.26	0.41	0.91	0.41	1.34	0.41	1.25	0.41
	1-T06	5.01	0.18	5.77	0.18	4.64	0.18	5.98	0.18	3.46	0.44	4.41	0.44	3.77	0.44	4.24	0.44	0.85	0.41	1.41	0.41	1.36	0.41	1.13	0.41
	2-T01	5.40	0.21	5.90	0.21	4.80	0.21	6.81	0.21	3.27	1.00	3.96	1.00	4.22	1.00	5.43	1.00	0.46	0.93	0.90	0.93	1.73	0.93	1.88	0.93	[9]
	2-T03	5.03	0.21	5.57	0.21			5.50	0.21	3.88	1.00	4.02	1.00			4.38	1.00	1.27	0.93	1.12	0.93			1.52	0.93	[10]
	2-T04	4.84	0.21	5.75	0.21	5.38	0.21	6.63	0.21	3.68	1.00	4.36	1.00	4.28	1.00	4.58	1.00	1.17	0.93	1.37	0.93	1.48	0.93	1.13	0.93
	2-T05	4.83	0.21	5.66	0.21	5.52	0.21	6.44	0.21	4.35	1.00	4.62	1.00	3.83	1.00	4.72	1.00	1.84	0.93	1.68	0.93	0.96	0.93	1.37	0.93
	Average	4.92		5.64		4.96		6.10		3.76		4.19		3.91		4.60		1.20		1.26		1.33		1.42
	2SD	0.51		0.36		0.97		1.20		0.65		0.59		0.55		0.93		0.82		0.58		0.48		0.52

^[1]Adjacent to T04.

^[2]Pairs of Ol-Lpx and Pl-Hpx are ~200 μm apart.

^[3]Pairs of Ol-Hpx and Pl-Lpx are ~300μm apart.

^[4]Bad SIMS pit for Hpx. Pl is ~300μm from Ol-Lpx.

^[5]Bad SIMS pit for Ol.

^[6]Lpx is in T04 and >200μm from Hpx-Pl.

^[7]Ol-Lpx, Hpx, Pl are >200μm to each other.

^[8]Bad SIMS pit for Lpx.

^[9]Pl is ~300μm from Lpx-Hpx.

^[10]Hpx SIMS pit overlaps with Lpx.