Evidence for Magnetically-Driven Accretion in the Distal Solar System

Abstract

Paleomagnetic measurements of meteorites indicate that magnetic fields existed in the inner solar nebula capable of driving accretion at rates similar to those observed for young stellar objects with protoplanetary disks. However, the field strength in the solar system beyond ∼7 astronomical units (AU) and its role in accretion remain poorly constrained. Returned samples from asteroid (162173) Ryugu offer the possibility of determining the nebular field intensity in this distal region. Here, we report paleomagnetic studies of three Ryugu particles which reveal that alteration occurred in the presence of a null or relatively weak (<15.8 μT) field within 3 million years (Ma) after solar system formation. This resolves previously contrasting reports that Ryugu's parent body experienced alteration in the presence of a strong (>80 μT) magnetic field and weak or null field (<3 μT). In addition, we re-examine previous paleomagnetic and Mn-Cr chronometry studies of three other distally-sourced meteorites, Tagish Lake, Tarda, and Wisconsin Range 91600, which measured paleointensities of <0.9, <1.7 and 5.1 ± 4.5 μT respectively. While it was previously unclear whether these records were acquired while the nebula was present, our re-analysis suggests that their records are sufficiently old (i.e., <3.5 Ma after solar system formation) to be nebular in origin. Collectively, these data demonstrate that the distal solar system nebular field, while faint, was likely still strong enough to drive accretion at rates like those observed in the inner solar system.

Key Points

• It is unknown if the nebular field in the distal (>∼7 astronomical units) solar system drove disk accretion

• The paleomagnetism of asteroid Ryugu samples suggests that the nebular field was <15.8 μT in this region

• A reanalysis of the ages and paleointensities of three other meteorites indicates a weak distal field capable of driving accretion

Plain Language Summary

Magnetic fields likely played a key role in transporting mass toward the Sun and angular momentum away from the Sun in the early solar system. Previous magnetic studies of meteorites that formed within ∼7 astronomical units (AU) of the Sun indicate that a magnetic field could transport mass at rates comparable to those observed in other planetary systems. However, whether this field could transfer mass in the distal (>7 AU) solar system is unknown. Samples returned from asteroid Ryugu can fill this gap. We find that the magnetic record of three particles returned from asteroid Ryugu is consistent with formation in a null or very weak field at <3 million years after solar system formation. Further, we re-examine previous paleomagnetic reports from three other distally-sourced meteorites and show that there was indeed a field present at this time, although it was weak (10 times less than Earth's field). Together, these three meteorites and the Ryugu samples indicate the presence of a faint field, but one that could still transfer mass at rates comparable to the inner solar system.

1 Introduction

Magnetic fields have long been theorized to play a key role in driving angular momentum transport and hence the global evolution of protoplanetary disks (PPDs) and the solar nebula (Weiss et al., 2021). Depending on the location (e.g., distance from the sun) and physical conditions (e.g., sense of disk rotation relative to the field) in the nebula, magnetic fields can generate turbulence and/or launch disk winds, two processes that have major implications for the accretion of chondrules and the later stages of planet formation (Weiss et al., 2021). Angular momentum transport can also be mediated by other mechanisms such as hydrodynamic and gravitational instabilities, but these are only thought to be significant if the magnetic field is absent or very weak (Kratter & Lodato, 2016; Lyra & Umurhan, 2019).

Insights about the nebular field in our own solar system during its PPD phase can be gained through analyses of the paleomagnetic records in meteorites. Natural remanent magnetization (NRM) has been observed in chondritic and achondritic meteorites and attributed to a nebular field (Borlina et al., 2021, 2022; Bryson et al., 2023; Cournede et al., 2015; Fu et al., 2014, 2020, 2021; Maurel & Gattacceca, 2024). These data indicate that during the first 3–4 million years (Ma) after the formation of calcium aluminum-rich inclusions (CAIs), the nebular field was 20–80 μT at heliocentric distances of ∼1–3 astronomical units (AU), as inferred from meteorites from the non-carbonaceous reservoir (Fu et al., 2014, 2023; Maurel & Gattacceca, 2024), and 40–200 μT between ∼3 and 7 AU, as inferred from meteorites from the carbonaceous reservoir (Borlina et al., 2021; Bryson et al., 2023; Fu et al., 2021, 2023). Furthermore, paleomagnetic studies of younger materials in chondrites and achondrites indicate that the nebular field dissipated after ∼3–4 Ma (Borlina et al., 2022; Bryson, Weiss, Biersteker, et al., 2020; Fu et al., 2020; Gattacceca et al., 2016; Wang et al., 2017; Weiss et al., 2008, 2017). The measured nebular field strengths are consistent with torques from magnetic fields dominating accretion, in which case they correspond to accretion rates of 10⁻⁹ - 10⁻⁷ M_☉ yr⁻¹ (solar masses per year). Such accretion rates are consistent with astronomical observations of stellar accretion rates for PPDs during the main disk (Class II) phase (Hartmann et al., 2016) when most chondrules and chondrites are thought to have formed (e.g., Blackburn et al., 2017). However, nearly all of these paleomagnetic studies were conducted on meteorites that, based partly on their inferred gas density environments, oxygen isotopic composition and oxygen fugacities, likely formed on parent bodies located at distances <∼7 AU (Fu et al., 2023; Scott & Krot, 2007) or perhaps even <∼4 AU (Desch et al., 2018). This leaves the intensity of the field in the more distal solar system (defined here as >∼7 AU) unconstrained.

A new opportunity to determine the strength of the nebular field in this remote region is provided by particles returned from Cb-type asteroid (162173) Ryugu in December 2020 by the Hayabusa2 mission (Watanabe et al., 2017). The spacecraft collected ∼5 g of material during two sampling events at two different locations in 2019: one on the surface of the asteroid and another at a location where the subsurface down to a depth of ∼3 m was exposed by a controlled impact (Arakawa et al., 2020; Yokoyama, Nagashima, et al., 2023). An initial paleomagnetic study of two Ryugu particles concluded that they recorded a >80–800 μT nebular field (Sato et al., 2022) [value reported here has been multiplied by a factor of two to account for the effects of asteroid rotation during NRM acquisition (Fu et al., 2014, 2023)]. However, a later study found no clear nebular field record and suggested that the samples in the original study were magnetically overprinted prior to NRM demagnetization (Maurel et al., 2024). In addition to paleomagnetic measurements of Ryugu samples, the Mobile Asteroid Scout (MASCOT) magnetometer (MasMag) acquired measurements during and after landing on the surface of Ryugu, finding the remanent field to be <1 nT at an altitude of 0.2 m (Hercik et al., 2020).

This study seeks to resolve the conflicting paleointensity estimates above and determine if the distal magnetic field was sufficiently strong to drive stellar accretion at astronomically observed rates. We determine the origin and nature of the NRM in Ryugu and obtain a robust bound on the nebular field intensity in the distal solar system by providing an analysis of the paleomagnetic record of three previously unstudied mm-sized Ryugu particles taken from the two landing sites: A0397, C0085, and C0006. We also re-examine the ages and paleointensities of other distally-sourced meteorites for which there was previous uncertainty as to whether their paleomagnetic records were acquired prior to or after nebula dissipation. This work was first presented at a meeting (Mansbach et al., 2023) simultaneously and independently of another Ryugu paleomagnetic study (Maurel et al., 2023).

2 Ryugu and Other Distal Meteorites

2.1 Ryugu as a Distal Solar System Magnetic Recorder

Ryugu particles contain 3.6–6.8 vol.% magnetite (Ito et al., 2022), much of which has a framboidal habit (Kimura et al., 2023; Nakamura et al., 2023; Sato et al., 2022). The framboidal magnetite grains have diameters of 300–1,100 nm and are in the single-vortex domain state (Nakamura et al., 2023), such that they should be capable of retaining paleomagnetic records over the lifetime of the solar system (Kimura et al., 2023). Much of this framboidal magnetite is thought to have formed at the same time or before carbonates based on their spatial association and similar oxygen isotopic compositions (Yamaguchi et al., 2023; Yokoyama, Nagashima, et al., 2023). Such carbonate-associated magnetite therefore likely formed during aqueous alteration of the Ryugu samples (Kimura et al., 2023) and should have recorded the ambient magnetic field paleointensity as a grain-growth chemical remanent magnetization (CRM) (Bryson et al., 2023). As Ryugu particles were not heated above ∼100°C after alteration (Yokoyama, Nagashima, et al., 2023) and experienced average peak pressures of <∼2 GPa (Tomioka et al., 2023), it is unlikely that the magnetite was remagnetized or completely demagnetized after initial formation. The duration of the aqueous alteration that affected Ryugu samples has been suggested to be on the order of ∼1 Ma (Nakamura et al., 2023; Yamaguchi et al., 2023), while similar aqueous alteration in Tagish Lake and CM chondrites is proposed to have occurred over periods lasting 0.01 to 1 Ma (Bryson, Weiss, Lima, et al., 2020; Cournede et al., 2015). As such, any paleomagnetic record should be considered as averaged over these timescales. Even so, simulations suggest that the nebular field configuration was likely quasi-steady within ∼20 AU over at least 10³ yr due to strong dissipation from non-ideal magnetohydrodynamic effects (e.g., Weiss et al., 2021).

While ferromagnetic pyrrhotite is also present in Ryugu particles, we only focus on the paleomagnetic record of magnetite. This is because domain state observations and rock magnetic properties indicate that framboidal magnetite is the main remanence carrier (Kimura et al., 2023; Maurel et al., 2024; Sato et al., 2022). In particular, 94% of an anhysteretic remanent magnetization (ARM) (AC field 260 mT and bias field 100 μT) and of a 1 T isothermal remanent magnetization (IRM) are removed by alternating field (AF) demagnetization by the peak of the coercivity distribution of Ryugu pyrrhotite (175 mT) (Maurel et al., 2024; Sato et al., 2022) (see Text S10 and Figure S5 in Supporting Information S1).

Petrologic, mineralogic, noble gas abundances, and isotopic analyses on the returned samples indicate that Ryugu particles most closely resemble CI carbonaceous chondrites and are distinct from all other known carbonaceous chondrite groups (Hopp et al., 2022; Ito et al., 2022; Nakamura et al., 2023; Okazaki et al., 2023; Paquet et al., 2022; Spitzer et al., 2023; Yada et al., 2021; Yokoyama, Nagashima, et al., 2023). Although the location at which Ryugu's parent body experienced alteration is not well-constrained, a diversity of observations suggests that it accreted in the distal solar system. Most significant are Ryugu's high abundances of volatile elements (e.g., C) (Kawasaki et al., 2022; Yokoyama, Nagashima, et al., 2023), the rarity (15–20 ppm) and small sizes of its CAIs and chondrules (<30 μm diameter) (Nakashima et al., 2023), and similarities between the infrared reflectance spectra of its anhydrous olivine and pyroxene to those of comets and D-type asteroids (Brunetto et al., 2023). Additionally, it has been suggested that Ryugu's parent body formed at 13–25 AU based on the abundance ratio of Cb-type asteroids compared to all C-type asteroids (0.1–0.2) in the asteroid main belt and simulations of the scattering of planetesimals from the distal solar system into the main belt due to the growth and migration of the giant gas and ice planets (Hopp et al., 2022; Nesvorný et al., 2024; Raymond & Izidoro, 2017). Also, Ryugu's phyllosilicates have similar compositions to those of interplanetary dust particles, which are thought to originate from comets and therefore could indicate the exchange of material between asteroid and distal comet formation regions (Joswiak et al., 2023). The idea of potential material exchange and a distal origin for Ryugu's parent body is bolstered by the similar relative abundances of ¹⁶O-rich versus ¹⁶O-poor refractory inclusions in Ryugu particles compared to comet 81P/Wild2 (Kawasaki et al., 2022). Furthermore, the high ¹³C isotope values in Ryugu carbonates are similar to those of Tagish Lake, which also likely formed in the distal solar system (see Section 2.3). Ryugu samples also possess IDP-like δD and δ¹⁵N compositions (Hoppe et al., 2023; McCain et al., 2023). Additionally, the presence of fluid inclusions in pyrrhotite containing CO₂ and H₂O is consistent with the formation of Ryugu's parent body beyond the CO₂ and H₂O snow lines (Nakamura et al., 2023), with the former indicating formation at >4–10 AU (Desch et al., 2018; Fujiya et al., 2019). Lastly, accepting that the geochemical links between Ryugu and CIs indicate their parent bodies formed at a similar distance from the Sun, the depletion of refractory elements in CIs, which has been previously interpreted to correspond to formation at >15 AU, would indicate a similar formation distance for Ryugu's parent body (Desch et al., 2018).

Collectively the above evidence suggests Ryugu's parent body formed, and may have also been aqueously altered, at a distance of >∼7 AU and therefore distal to other carbonaceous chondrite groups other than CIs. We estimate the upper limit on the heliocentric distance to be 25 AU from scattering models (Hopp et al., 2022; Nesvorný et al., 2024; Raymond & Izidoro, 2017). However, we stress that the location of formation in this region remains unknown and that additional research is needed to confirm the distal origin of Ryugu's parent body. We note that an alternate view of the formation region of CIs and Ryugu comes from a recent study on Ni isotopes concluding that while Fe and Ni isotopes in these samples are distinct from other carbonaceous chondrites, the isotopic differences could be due to more efficient accretion of FeNi grains as a result of nebular gas photoevaporation in a similar location as carbonaceous chondrites but at a later point in time (Spitzer et al., 2024).

2.2 Timing of Ryugu Magnetite Formation

To establish whether the Ryugu particles can retain a record of the nebular field, we must determine whether their magnetite formed while the solar nebula was still present. Previous paleomagnetic analyses of meteorites in the non-carbonaceous reservoir suggest the nebular gas had dispersed by 3.71 ± 0.2 Ma (Wang et al., 2017; Weiss et al., 2017) at 1–3 AU, and by 2.7–5.1 Ma in the carbonaceous reservoir at ∼3–7 AU (Borlina et al., 2022). Given that chondrules are thought to have formed while the gas was present, the youngest chondrule age might be taken as an upper limit for the lifetime of the nebula (e.g., Desch et al., 2018); however, the absence of chondrules known with younger ages does not require prior gas dissipation since our meteoritical record could be biased and younger chondrules found in the future. We note that astronomical observations of solar-mass stars indicate that protoplanetary disks could exist up to or after 10 Ma (Ribas et al., 2015) and that some disks, called transitional disks, possess an inner cavity depleted in dust and gas while retaining material at further distances (Andrews et al., 2011). Therefore, the distal solar nebula might have existed after the inner nebular gas had dissipated. The lifetime of the nebula may also be longer than inferred from paleomagnetic measurements based on formation models of Kuiper Belt Objects such as Arrokoth (Bierson & Nimmo, 2019; McKinnon et al., 2020) and IDP-like parent bodies (Neveu & Vernazza, 2019), which suggest that nebular gas was present at ∼5–6 Ma after CAI-formation (Neveu & Vernazza, 2019). However, since uncertainties on the age of disk dissipation in the carbonaceous reservoir overlap with the non-carbonaceous reservoir from paleomagnetic analyses, we consider as an upper bound that alteration within 3.5 Ma (taken as the oldest age within a 2 standard deviation uncertainty from Wang et al. (2017)) would suggest that magnetite in Ryugu particles formed while the field was active. This upper limit is also consistent the nebula still being present based on the non-paleomagnetic analyses above.

Mn-Cr dating has been used to data carbonates in Ryugu, which are thought to be cogenetic with magnetite based on textural observations and oxygen isotope data (Yamaguchi et al., 2023). Therefore, we can use Mn-Cr ages to infer the timing of NRM acquisition. The first Mn-Cr dating study of Ryugu reported a nominal age of 5.2 [+0.8, −0.7; 2σ] Ma after CAI-formation via secondary ion mass spectroscopy (SIMS) (Yokoyama, Nagashima, et al., 2023). Taking into account additional systematic uncertainties [choice of standard and the poorly constrained relative sensitivity factor (RSF)] increases the age range to 3.1–6.8 Ma (Yokoyama, Nagashima, et al., 2023), which spans the time that the nebular field is thought to have decayed from paleomagnetic analyses (Borlina et al., 2022; Wang et al., 2017; Weiss et al., 2017). A later study, which included more accurate calibrations of the fractionation effects associated with laboratory ionization, reported SIMS ages of 1.03 ± 0.79 (2σ) and 0.45 ± 0.39 (2σ) Ma for two samples (McCain et al., 2023). In particular, McCain et al. (2023) used synthetic matrix-matched standards appropriate for the Fe content of the dolomite in Ryugu (McCain et al., 2020, 2023).

There are at least two ways to interpret the different ages above. First, alteration may have occurred over a 3–5 Ma timescale, or, second, the inconsistency in the ages is a result of different analytical techniques and RSF values used to derive the ages. These two explanations could potentially be distinguished using new Mn-Cr measurements that do not rely on knowledge of the RSF and/or by determining whether SIMS measurements that are corrected with the same RSF yield consistent ages. The first approach was taken by Tanaka et al. (2024), who acquired new Mn-Cr ages using thermal ionization and inductively coupled plasma mass spectrometry (TIMS/ICP-MS). These yielded an intermediate alteration age of 4.12 [+0.62, −0.55; 2σ] Ma, with all but one sample lying on an isochron (Tanaka et al., 2024). However, the analysis is limited by several aspects. First, the concentration of radiogenic Cr measured in the bulk samples with TIMS is orders of magnitude less than measured by SIMS from carbonates. Second, if the sample not on the isochron was included in the age calculation, it would increase the age estimate; on the other hand, including this point is problematic since an isochron passing through it would be older than age of the solar system. Third, this study was conducted on multiple bulk samples and therefore does not directly place a limit on the formation age of any particular grain, such that it requires the demonstration of isotopic equilibration of the samples. We note as well that an alteration age of 5.2 [+1.8, −1.4; 2σ] Ma was reported by Yokoyama, Wadhwa, et al. (2023) using ICP-MS but that this age suffers some of the same limitations. Therefore, we consider that these studies, while important, do not conclusively distinguish between the two interpretations of the SIMS ages.

Here we assess the second hypothesis above by using the new RSF measurement in McCain et al. (2023) to correct the nominal age from Yokoyama, Nagashima, et al. (2023) and compare the two results. Referencing the same standard (absolute U-Pb age and initial ⁵³Mn/⁵⁵Mn ratio for the D’Orbigny angrite) as in McCain et al. (2023), we find the corrected age for the sample in Yokoyama, Nagashima, et al. (2023) to be 2.19 ± 0.74 (2σ) Ma (see Text S7 and Table S3 in Supporting Information S1). This corrected age is now in agreement with the two ages reported by McCain et al. (2023) above, favoring a difference in RSF values as the source of the previous discrepancy. In summary, all three of the RSF-corrected ages (1.03 ± 0.79, 0.45 ± 0.39, and 2.19 ± 0.74 Ma) are within the lifetime of the nebula. This would indicate that Ryugu magnetite is old enough to possess a record of the nebular field.

For the rest of this manuscript, we take this as our nominal interpretation. Even so, we recognize that future Mn-Cr analyses are required to conclusively determine the formation age of the magnetite. For example, more work is needed to calibrate the relationship between dolomite Fe content and RSF in order to reduce the uncertainties on the latter (McCain et al., 2020). This is exemplified by a recent study that determined a matrix-matched RSF via crystallization of dolomite with Fe, Mn, and Cr doping (Sugawara et al., 2024) rather than Cr implantation in a natural dolomite (McCain et al., 2023), and determined the alteration age of Ryugu carbonates as 4.5 [+1.0, −1.2; 2σ] Ma after CAI formation (Sugawara et al., 2024). Additionally, recent work to align the ages of achondrites dated with various short-lived radioisotope systems suggests that the initial abundances of ⁵³Mn/⁵⁵Mn and other radioisotope ratios would be homogenous in the solar nebula if the ages of CAIs are ∼1 Ma older than most Pb-Pb ages due to later resetting from transient heating (Desch et al., 2023). In this case, the ages reported above would be shifted ∼1 Ma younger such that Ryugu's parent body may then have experienced alteration after nebular dissipation. However, since the homogeneity of short-lived radioisotope abundances in the solar system is still debated (e.g., Connelly et al., 2023), we take the age of Ryugu magnetite formation to be within 3.5 Ma of CAI-formation but consider alternate interpretations as well.

2.3 Additional Distal Meteorites

Despite the unique opportunity offered by Ryugu samples, they are likely not only our only known record from the distal solar system. The ungrouped carbonaceous chondrite Tagish Lake has also been suggested to have a distal origin based on its similarity in infrared reflectance to D-type asteroids (Hiroi et al., 2001) and high abundance of interstellar grains (Brown et al., 2000). Additionally, the high measured ¹³C/¹²C ratio in carbonates in Tagish Lake and its inferred abundance ratio of CO₂ to H₂O are distinct from CM chondrites and similar to those of comets (formation at ∼16 to ∼30 AU; Levison et al., 2009). This is likely indicative of ¹³C-rich ice (Fujiya et al., 2019), which has been interpreted as suggesting formation beyond the orbits of Neptune and Uranus (Hoppe et al., 2023). Other lines of evidence suggesting Tagish Lake has a distal origin include enrichments in ¹⁵N (Nakamura-Messenger et al., 2006), potential accretion past the ammonia line (∼90 K) to form the observed magnetite morphologies (Sridhar et al., 2021), and thermal evolution models based on the magnetic properties of framboidal magnetite after annealing (Kimura et al., 2021). A previous paleomagnetic study on bulk samples of Tagish Lake concluded that the meteorite experienced alteration in the presence of a <0.9 ± 0.3 (2σ) μT field (Bryson, Weiss, Lima, et al., 2020; Maurel & Gattacceca, 2023) (see Text S6 in Supporting Information S1).

Another ungrouped carbonaceous chondrite, Wisconsin Range (WIS) 91600 is also thought to have formed in the distal solar system based in part on the similarity of its visible-near infrared spectra to those of T and D type asteroids and Tagish Lake (Hiroi et al., 2005). Additionally, its oxygen isotopic composition is similar to that of CI chondrites (Choe et al., 2010). However, WIS 91600’s bulk composition (e.g., refractory and volatile elements) differs from that of CI chondrites (Choe et al., 2010). WIS 91600’s mineralogy and extent of aqueous alteration also have an affinity with Tagish Lake (Bryson, Weiss, Biersteker, et al., 2020), although their differing oxygen isotopic compositions suggests they are not from the same parent body (Yamanobe et al., 2018). Lastly, the framboidal morphology of its magnetite, which is similar to magnetite in Tagish Lake and Ryugu, could also indicate formation past the ammonia ice line (Sridhar et al., 2021). A paleomagnetic analysis of the NRM in WIS 91600 found evidence for formation in a 5.1 ± 4.5 (95% confidence interval) μT field after heating via impacts (Bryson, Weiss, Biersteker, et al., 2020) (see Text S6 in Supporting Information S1).

A third ungrouped carbonaceous chondrite, Tarda, was an observed fall in August 2020, and like WIS 91600 and Tagish Lake, may have formed and experienced alteration in the distal solar system. Evidence for a distal source region is provided by its similar bulk oxygen isotopic composition and mineralogy (Chennaoui Aoudjehane et al., 2021), C¹³ and N¹⁵ isotopic composition (Marrocchi et al., 2021), IR spectra (Bates et al., 2024), and noble gas signatures (Avice et al., 2022) to those of Tagish Lake. Tarda has also been tentatively linked to D-types asteroids, again like Tagish Lake (Marrocchi et al., 2021). A recent paleomagnetic analysis of Tarda determined that the meteorite experienced alteration in presence of a weak (<1.7 μT) or null field (Bates et al., 2024) (see Text S6 in Supporting Information S1).

Finally, CI chondrites are proposed to have a distal origin for many of the same reasons as for Ryugu as described in Section 2.1 (e.g., depletion in refractory elements) in addition to enrichments in volatiles and hydrogen and nitrogen isotopes [see Discussion in Hopp et al. (2022)]. However, Tagish Lake has a different Ti and Cr isotope composition than CI chondrites and Ryugu. Therefore, Tagish Lake, Tarda, and WIS 91600 did not likely form in the exact same region as CI chondrites and Ryugu, but we suggest that it was still in the distal part of the nebula, between 7 and 30 AU. Where these two groups of samples formed within that region and relative to each other is unknown. Preliminary paleomagnetic analyses on the Orgueil and Ivuna CI chondrites found an upper paleointensity limit of 14 μT in the CI formation region (Tikoo et al., 2019) (see Text S6 in Supporting Information S1).

While there have been paleomagnetic studies of Tagish Lake, WIS 91600, Tarda, and CI chondrites, initial Mn-Cr dating on these samples suggested alteration and magnetite formation around the time of or after nebular dissipation. Tagish Lake and CI chondrite dolomites have reported ages of 3.8 [+1.1, −1.3; 2σ] Ma and 4.3 [+0.7, −0.8; 2σ] Ma respectively, where the CI chondrite age is the average of the reported ages for Orgueil and Ivuna (Fujiya et al., 2013). Therefore, their weak paleointensities might be due to prior or ongoing dissipation of the nebula rather than formation in the distal solar system during the nebula's main lifetime. Additionally, no magnetization ages have been reported for WIS 91600 and Tarda, although they were proposed to be similar to that of Tagish Lake given the meteorites' similar reflectance spectra, oxygen isotopes, and extent of aqueous alteration (Bryson, Weiss, Biersteker, et al., 2020).

However, these reported ages were not corrected for the difference in RSFs associated with calcite versus dolomite lithologies and the Fe content in dolomite as was found to be important in accurately dating Ryugu particles (see Section 2.2). We obtain a corrected age of <2.34 Ma after CAI-formation for Tagish Lake using the standards in McCain et al. (2023) and an updated RSF from McCain et al. (2020) (see Text S8 and Table S4 in Supporting Information S1). Therefore, aqueous alteration likely occurred while the nebula was still active. This age is also likely applicable to WIS 91600 and Tarda as stated in Bryson, Weiss, Biersteker, et al. (2020) and Bates et al. (2024) respectively. The RSF-corrected ages for Orgueil and Ivuna are 1.40 ± 0.73 Ma and 2.64 ± 1.70 Ma after CAI-formation respectively (see Text S8 and Table S4 in Supporting Information S1). As with Tagish Lake, Tarda, and WIS 91600, Orgueil's alteration age is within our upper limit on the lifetime of the nebula, though the uncertainty on the age for Ivuna includes 3.5 Ma.

3 Materials and Methods

3.1 Samples

Three Ryugu particles were provided to our group by the Japan Aerospace Exploration Agency (JAXA) for this analysis: A0397, C0085, and C0006. Prior to our analyses, all three samples had not been analyzed by other investigators with the exception of catalog imaging (all samples) and MicrOmega and multiband spectroscopy (only C0085 and C0006) (Yumoto et al., 2022). A0397 is from sample chamber A which was collected when the Hayabusa2 spacecraft touched down on the undisturbed surface of Ryugu. The particle is 1.62 mm in length and has a mass of 1.2 mg (Yumoto et al., 2022). C0006 is from sample chamber C, which was collected from subsurface materials exposed by the prior impact of an artificial projectile. The particle is 4.51 mm in length and has a mass of 16.3 mg (Yumoto et al., 2022). Particle C0085 is also from sample chamber C and has a length of 1.98 mm and a mass of 1.2 mg (Yumoto et al., 2022). C0085 split during analysis into two mutually-oriented daughter samples: C0085a and C0085b. Particle A0397 was allocated for a full paleomagnetic study while particles C0085 and C0006 were provided for only a limited time.

3.2 NRM Demagnetization

Particles A0397 and the two C0085 specimens were placed in AF-demagnetized, acid-washed Ge 124 and Eagle XG, respectively, quartz wells for measurement following the procedure in Borlina et al. (2020). The wells were packed with non-magnetic silica nanopowder to prevent the samples from moving during measurements. Particles A0397 and C0085a,b were measured in the Massachusetts Institute of Technology (MIT) Paleomagnetism Laboratory using a 2G Enterprises Superconducting Rock Magnetometer (2G SRM) 755 [2σ noise floor of 0.99 × 10⁻¹² Am²; Figure S5 in Wang et al. (2017)] equipped with an automatic sample handling and coil system and housed in a magnetically shielded room (<200 nT). The 2G SRM was also used for AF demagnetization of the NRM of the three particles in steps of 0.5–1 mT from 0 to 145 mT. To correct for any gyroremanent magnetization acquired during AF demagnetization, we averaged the moments measured after AF application along each of the three orthogonal axes following the Zijderveld-Dunlop method (Stephenson, 1993). Between AF levels of 145 and 400 mT, the NRM was measured in steps of 15–20 mT, averaging the moments after AF applications along each axis as before.

The NRM of particle C0006 was measured at the Marine Core Research Institute, Kochi University. It was placed in a non-magnetic plastic box used in the Integrated Ocean Drilling Program (Natsuhara-Gaiken, Japan) with a moment below the noise floor of the 2G SRM 760–4.4 cm U channel system. The sample's NRM was AF-demagnetized with an automatic sample handing and coil system in steps of 5–10 mT from 0 to 80 mT. All measurements were performed in a magnetically shielded room (<200 nT).

Directions of NRM components and their maximum angle of deviation (MAD) values, which provide a measure of angular uncertainty in the components (Khokhlov & Hulot, 2015), were determined by principal component analysis (Kirschvink, 1980). Components were considered origin-trending if their deviation angles from the origin (DANG) was less their MAD values (Tauxe & Staudigel, 2004), which would be consistent with, but does not require, that they are primary records dating back to magnetite formation.

3.3 Paleointensities

After AF demagnetization of NRM, we estimated paleointensities using the ARM and IRM family of methods. ARMs were applied to each of the particles: A0397 and C0085a,b (260 mT AC field and 100 μT bias field), and C0006 (80 mT AC field and 100 μT bias field). The ARM in each sample was then AF-demagnetized using the same process as for the NRM. The ARM demagnetization was compared to the NRM demagnetization to determine the paleointensity. ARM paleointensities were calculated by comparing the ratio of the NRM and ARM lost over each component (∆NRM and ∆ARM, respectively). Following Stephenson and Collinson (1974) and Weiss and Tikoo (2014), for a sample whose NRM is a CRM, we have

$$${B}_{p,\text{ARM}}=\frac{{\increment}\text{NRM}}{{\increment}\text{ARM}}\cdot \frac{{B}_{\text{lab}}}{{f}^{\prime }}$$$

where B_p,ARM is the paleointensity, B_lab is the ARM bias field (100 μT), and f′ is an empirically determined ratio of CRM to ARM. For the paleointensities for the high coercivity (HC) range (see Section 4), we adopt f′ = 1.17 (arithmetic mean of f′ values) following a recent study of grain-growth of magnetite in sedimentary rocks and synthetic samples containing grain-growth CRMs (Maurel & Gattacceca, 2023). The geometric mean of the f′ distribution determined by Maurel and Gattacceca (2023) is 0.99 which, if adopted, would increase all reported paleointensities by a factor of 1.18. We do not include any paleointensity correction for pyrrhotite given its low contribution to the total magnetization and unstable domain state (Section 2.1). Additionally, pyrrhotite has an efficiency of TRM acquisition relative to saturation IRMs (sIRM) that is ∼5 times lower than that of magnetite (Kletetschka & Wieczorek, 2017), suggesting that if pyrrhotite were a significant NRM carrier, our paleointensity constraint would be even lower than that reported here.

IRM paleointensities were estimated for A0397 and C0006 via application and subsequent AF demagnetization of a 1 T sIRM. The paleointensity is calculated in a similar fashion as above (Gattacceca & Rochette, 2004),

$$${B}_{p,\text{IRM}}=\frac{{\increment}\text{NRM}}{{\increment}\text{IRM}}\cdot a$$$

where ∆IRM is the sIRM lost over each component and a is an empirically determined quantity relating CRM to IRM. For HC paleointensities, we adopt a = 4,000 (arithmetic mean of a values) (Maurel & Gattacceca, 2023). The geometric mean of the a distribution determined by Maurel and Gattacceca (2023) is 3,300 and, if adopted, would decrease all reported paleointensities by a factor of 0.82. The application of an sIRM is also useful for diagnosing the presence of magnetic overprints as the TRM or CRM intensity produced in a tens of μT ancient field is of order 1 to several percent of the sIRM (McClelland, 1996; Yu et al., 2007).

We also calculated the paleointensities of the Ryugu particles in this study via the residual ARM (AREMc) method, analogous residual IRM (IREMc) method, and the ratio of total NRM to total sIRM. These three methods differ from the above ARM and IRM paleointensities in that they are direct comparisons of NRM, IRM, and ARM values and therefore minimize the effect of spurious noise at high AF levels. Further explanations can be found in Text S2 in Supporting Information S1.

All HC range paleointensities reported here using the above two equations have been multiplied by a factor of 2 to account for the average decrease in paleointensity associated with a rotating sample in an external field (Fu et al., 2014, 2023). For any low coercivity (LC) and medium coercivity (MC) components, we adopt f′ = 3.33 (Weiss & Tikoo, 2014) and a = 2,100 (Maurel & Gattacceca, 2023) as the NRM/ARM and NRM/IRM values, respectively, since we interpret those components as viscous overprints (see Section 5.1). No correction factor of 2 is included for the LC and MC paleointensities as well since we interpret them to be recent overprints.

We note that Ryugu particles experienced peak pressures <2 GPa (Tomioka et al., 2023). Assuming this occurred after magnetite formation, this could have partially removed the original NRM. Previous studies of pressure or shock demagnetization of magnetite-bearing samples have found a decrease of up to 80% of an initial sIRM at 1.24 GPa and 27% of an initial TRM at 0.3 GPa (Bezaeva et al., 2010; Gattacceca et al., 2010). Therefore, our reported paleointensities could be underestimated by a factor of 5, but we do not include this correction in our final paleointensities as the reported pressure is an upper limit.

3.4 Paleointensity Fidelity Tests

Paleointensity fidelity tests are a method for estimating the minimum recoverable field from a sample across the HC range given our paleointensity methodology and instrumentation (Tikoo et al., 2012). In this test, ARMs of varying bias fields are applied to a sample to create pseudo-NRMs. These are then AF demagnetized in a similar manner to NRM demagnetization. Using the ARM paleointensity method above, we calculate the retrieved paleointensity over the HC range from these pseudo-NRMs and compare it to the known applied laboratory field to determine the paleointensity recording fidelity of the sample. The metrics used to assess the fidelity of the sample are its agreement with the known laboratory field and the uncertainty in the recovered field value relative to the magnitude of the known field, known as the difference (D′) and error (E), respectively (Bryson et al., 2017; Tikoo et al., 2012). We have:

$$${D}^{\prime }=\frac{L-I}{L}$$$

$$$E=\frac{W}{L}$$$

where L is the applied (known) laboratory field, I is the recovered paleointensity, and W is the 95% confidence interval for the retrieved paleointensity (calculated via Student's t test) (Tikoo et al., 2012). As metrics for passing the paleointensity fidelity test for a certain known laboratory field, we use −0.5 < D′ < 1 and E < 0.5 (Bryson et al., 2017). A paleointensity fidelity test was not performed on C0006 due to the short time the sample was available for analysis.

3.5 Viscous Remanent Magnetization (VRM) Experiments

We conducted a VRM study on particle A0397 to constrain the source of the LC and MC components. The rates of VRM acquisition and decay for particle A0397 were assessed by demagnetizing the sample to an AF level of 145 mT and subsequently leaving it outside of the MIT magnetically shielded room in Earth's magnetic field (∼45 μT). After 19 days, it was then brought into the shielded room (field <200 nT) and continuously measured to determine the amount of VRM gained by the sample over that time period and the decay rate of the VRM.

4 Results

Our AF demagnetization of the NRM of A0397 to 400 mT (Figure 1a) revealed a non-origin-trending LC component that unblocked between 0 and 10.5 mT and a weak and non-origin-trending possible MC component that unblocked between 11 and 23.5 mT. We label the AF range above this final NRM component to 145 mT as the HC range and note that it contains no recoverable NRM component. AF demagnetization of the NRM of C0085b to 400 mT (Figure 1b) revealed two components: an LC component that unblocked between 0 and 15 mT and an origin-trending possible MC component that unblocked between 15.5 and 23.5 mT. Lastly, AF demagnetization of the NRM of C0006 to 80 mT (Figure 1c) revealed only an LC component that unblocked between 0 and 20 mT, but the coarse step size of the demagnetization (5–10 mT) compared to that of A0397 and C0085b (0.5–1 mT) at AF levels <25 mT may have precluded identification of an MC component. We confine our discussion of the results of AF demagnetization of the NRM of C0085a to the Supplementary Materials due to evidence for prior artificial remagnetization of this sample (see Text S4 and Figure S4 in Supporting Information S1).

Figure 1

Alternating field (AF) demagnetization of natural remanent magnetization (NRM) in Ryugu particles. Orthographic projection of endpoints of NRM vectors on the northeast (N–E) and up-east (Z–E) planes for (a) A0397, (b) C0085b, and (c) C0006. The low coercivity (LC) component is shown in red, the medium coercivity (MC) component in blue, and the high coercivity (HC) range in black. Arrows denote the directions of the components in both projections and selected AF steps are labeled.

The ARM LC paleointensities for A0397, C0085b, and C0006 were 22.1 ± 5.07 μT (all uncertainties here reported as 95% confidence intervals estimated from Student's t-test), 118 ± 27.5 μT, and 7.87 ± 12.2 μT, respectively (Figure 2). The ARM paleointensities of the possible MC components in A0397 and C0085b were 8.04 ± 5.54 μT and 3.60 ± 6.85 μT, respectively. The ARM paleointensities determined for the HC range are within error of zero: 4.98 ± 10.8 μT (A0397), 26.3 ± 30.5 μT (C0085b), and 8.76 ± 16.5 μT (C0006). The IRM paleointensities for the LC and MC components and HC range in A0397 (C0006) were 9.81 ± 1.61 μT (3.15 ± 3.94 μT), 6.48 ± 2.89 μT, and 12.4 ± 6.80 μT (4.57 ± 7.82 μT), respectively.

Figure 2

Paleointensities of Ryugu particles. Shown is anhysteretic remanent magnetization (ARM) lost (CRM-equivalent 85 μT bias field in 260 mT AC field) versus natural remanent magnetization (NRM) lost during AF demagnetization. (a) A0397. (b) C0085b. Over the high coercivity (HC) range, the ARM applied to A0397 continues to demagnetize while the NRM does not, such that the paleointensity is within error of zero. The same is true for C0085b, although the sample has a noisier ARM demagnetization. We label the paleointensity for the HC range derived from the slope of the line shown (see Section 3.3). We note that the paleointensities labeled here account for all corrections discussed in Section 3.3 as well.

The results of the paleointensity fidelity tests for A0397 and C0085b are shown in Figure 3. We find that the minimum recoverable paleofield is >171 μT for C0085b and >34 μT for A0397 (factor of 2 included). Other paleointensity methods retrieved values between 4.82 and 28.9 μT across all three samples for the HC range (Table 1). As the calculated paleointensities and the results of the paleointensity fidelity test are all justifiable techniques for determining an upper limit on the paleofield when no HC component is present (see Section 5.3 for further discussion regarding paleointensity upper limits), we take the lowest value amongst all the methods to be our upper limit. While the lowest value is 4.82 ± 7.82 μT from the IREMc method for C0006 (Table 1), the coarse demagnetization of that sample could lead to an over- or underestimation of the paleointensity. Therefore, we take the ARM paleointensity for A0397 plus its uncertainty, 15.8 μT, as our upper limit on the intensity of the distal field where Ryugu's parent body experienced alteration.

Figure 3

Paleointensity fidelity test results. (a) A0397 and (b) C0085b. The solid blue and red lines and axes denote D' and E, respectively. The blue and red dashed lines and arrows denote the regions in which the D' and E, metrics, respectively, are passed. All chemical remanent magnetization-equivalent bias fields have been multiplied by a factor of 2 to account for asteroid rotation.

Table 1. Paleointensities for Ryugu Particles

Sample	Component	Bp,ARM (μT)	Bp,IRM (μT)	Bp,AREMc (μT)	Bp,IREMc (μT)	Bp,NRM/IRM (μT)	Bfidelity (μT)
A0397	LC	22.1 ± 5.07	9.81 ± 1.61	–	–	17.4	–
	MC	8.04 ± 5.54	6.48 ± 2.89	5.19	8.64		–
	HC	4.98 ± 10.8	12.4 ± 6.80	28.9	19.1		<34
C0085b	LC	118 ± 27.5	–	–	–	–	–
	MC	3.60 ± 6.85	–	19.0	–		–
	HC	26.3 ± 30.5	–	18.7	–		<171
C0006	LC	7.87 ± 12.2	3.15 ± 3.94	–	–	6.33	–
	HC	8.76 ± 16.5	4.57 ± 7.82	12.6	4.82		–
C0005	LC	1.91 ± 2.49	0.68 ± 0.61	–	–	1.96	–
	HC	<2.74	–	8.68	2.19		<40
C0154a	–	–	–	–	–	12.0	–
C0154b	–	–	–	–	–	14.4	–
A0026	C1	–	147 ± 211	–	1,240	288	–
	HC	–	<125	–	423		–
C0002-4-f	C1	–	526 ± 311	–	–	168	–
	HC	–	13.5 ± 53.5	–	88.7		–

• Note. Data for particles A0397, C0085b, and C0006 are from this study, data for particles C0005 and C0154a,b taken from Maurel et al. (2024), and data for A0026 and C0002-4-f taken from Sato et al. (2022). The first column denotes the sample name. The second provides the component label (LC = low coercivity, MC = medium coercivity, HC = high coercivity). For the two samples from Sato et al. (2022), we use C1 as the identifier for the component labeled by that study as being a record of the nebular field. The third and fourth column provide the ARM (B_p,ARM) and IRM (B_p,IRM) paleointensities, respectively. The fifth and sixth columns provide paleointensities calculated via the AREMc (B_AREMc) and IREMc (B_IREMc) methods, respectively (see Text S2 in Supporting Information S1). The sixth column provides the paleointensity calculated as total NRM/sIRM. The final column list the results of the paleointensity fidelity experiments. For all paleointensity calculations of the HC and C1 components, f’ = 1.17 and a = 4,000 (Maurel et al., 2024). For all paleointensity calculations of the LC and MC components, f’ = 3.33 and a = 2,100 (Maurel et al., 2024). Highlighted in green is the upper limit that we prefer for the paleointensity of the nebular in the region where Ryugu's parent body experienced alteration.

The VRM acquisition and decay experiment (see Text S3 and Figure S1 in Supporting Information S1) on particle A0397 indicates an upper limit on the VRM acquired by the sample after 4.5 years (i.e., time since sampling on Ryugu) of 9.11 ± 0.05 × 10⁻¹⁰ Am² after correcting for VRM decay in our shielded room for 3 days before the start of NRM demagnetization. This value is 45% of the vector-summed LC and MC components for that particle and 68% of just the LC component. The total VRM gained is also 42% of the vector-summed LC and MC components in C0085b and 38% of just the LC component.

5 Discussion

5.1 Comparison to Previous Ryugu Paleomagnetic Analyses

As stated in Section 1, two previous paleomagnetic studies have been conducted on Ryugu particles, with one reporting detection of NRM attributed to a nebular field with an intensity of >80–800 μT (Sato et al., 2022) and the other that a weak (less than a few μT) or null field environment (Maurel et al., 2024). The latter analysis was agnostic as to whether the near-zero field was the result of Ryugu's parent body having been altered after the nebular field dissipated or that a nebular field was present but weak in the region in which Ryugu's parent body experienced alteration.

With regard to whether our Ryugu particles contain a clear record of the nebular field, because (a) the VRM that could be gained prior to our analysis can account for ∼45% of the magnetization of the LC and MC components in A0397 and C0085b and (b) the paleointensities of these two components (22.1–118 μT for LC and 3.6–9.0 μT for MC) are comparable to the field strength of the reported spacecraft engine field (Sato et al., 2022) and/or Earth's field, the source of the magnetization in these components in our samples is likely a VRM acquired on the spacecraft after sampling and/or from Earth's field prior to our analysis. In any case, the weak LC and MC components are nearly all not origin-trending and therefore not primary. Therefore, the weak LC and MC components observed in the Ryugu particles analyzed in this study are overprints. Because the HC paleointensities for the particles in this study are within error of zero, this indicates that there are ferromagnetic grains with coercivities >24 mT capable of recording an ARM (Figure 4) but that did not unblock any NRM. Combined with a lack of a recoverable magnetization component, this suggests a weak or null field at the time of the alteration when the magnetite formed. In summary, we agree with the conclusion drawn by Maurel et al. (2024) that there is no clear nebular field record in Ryugu particles.

Figure 4

AF demagnetization of natural remanent magnetization (NRM), anhysteretic remanent magnetization (ARM), and isothermal remanent magnetization (IRM) in A0397. (a) Moment normalized to pre-AF demagnetization value versus AF level for NRM, IRM, and various ARMs. Fields are chemical remanent magnetization (CRM)-equivalents for ARM bias fields (using CRM/ARM = 1.17 (Maurel & Gattacceca, 2023)). (b) Moment (not normalized) versus AF level for the NRM and various ARM applications.

The demagnetization behavior and magnetic properties of our samples most closely resemble those of particles A0154-a,b and C0005 in Maurel et al. (2024) (Figure 5). The NRMs of particles A0154-a, A0154-b, and C0005 are 8.16 × 10⁻⁴, 1.11 × 10⁻³, and 2.89 × 10⁻⁴ Am² kg⁻¹, respectively. This is comparable to the NRM intensities of A0397 and C0006 in our study, which are 1.9 × 10⁻³ and 6.3 × 10⁻⁴ Am² kg⁻¹, respectively. Samples A0154-a and C0005 in Maurel et al. (2024) possessed a poorly-defined NRM component between 0 and 8 mT, after which no clear component was observed and the NRM behaved erratically, while our samples A0397 and C0085b exhibited two poorly-defined components up to 24 mT and then erratic behavior at higher AF levels. The potential difference between the groups of samples could be due to the longer time spent in a null field environment (11 months for A0154-a and 20 days for C0005 vs. 3 days for our samples) prior to NRM demagnetization.

Figure 5

Comparison to previous Ryugu paleomagnetic studies. (a) Specific natural remanent magnetization (NRM) versus AF level for samples in all three studies. (b) NRM (normalized to pre-AF demagnetization value) versus AF level for samples in all three studies. Labels are the same as in panel (a). (c) Specific NRM and specific anhysteretic remanent magnetization (ARM) versus AF level for A0397 (this study) and C0005 (Maurel et al., 2024). An ARM bias field of 85 μT was applied to A0397 and C0005, though the AC fields differed (260 and 100 mT, respectively). Fields are chemical remanent magnetization (CRM)-equivalents for ARM bias fields (using CRM/ARM = 1.17 (Maurel & Gattacceca, 2023)).

The paleointensity fidelity experiments on sample C0005 in Maurel et al. (2024) seem to indicate that the weakest recoverable field from that particle is 40 μT (Table 1), which is similar to sample A0397 in our study that failed the paleointensity fidelity test at 34 μT. However, Maurel et al. (2024) argued that the upper limit on the field is likely to be a few μT for two reasons: (a) the NRM demagnetization patterns of the Ryugu particles that they observed are similar to those for Tagish Lake, which experienced alteration in the presence of a <0.9 ± 0.3 μT field; and (b) since the NRM/sIRM values of Ryugu particles matches that of Tagish Lake. The samples in our study have NRM/sIRM of 7.91 × 10⁻⁴ and 2.17 × 10⁻³ for C0006 and A0397, respectively. This is similar to the NRM/sIRM of A0154-a,b and C0005, which span from 2.3 × 10⁻⁴ to 1.5 × 10⁻³ (Maurel et al., 2024). Tagish Lake has an NRM/sIRM of 4.4 × 10⁻⁴ (Maurel et al., 2024). We note that NRM/sIRM is one method of determining paleointensities, which provides an upper limit of 6.33, 17.4, 1.96, and 3.52 μT for the HC ranges in C0006, A0397, C0005, and Tagish Lake, respectively.

We agree with Maurel et al. (2024) that the high paleointensity reported by Sato et al. (2022) is likely not a true nebular field record but is instead due to magnetic contamination after arrival on Earth for the following reasons. First, the reported components by Sato et al. (2022) are not origin-trending [see Table S1 in Supporting Information S1, DANG > MAD in Figure 9 in Sato et al. (2022)] and therefore are unlikely to be a primary record. Second, Ryugu magnetite has microcoercivities up to ∼300 mT (Maurel et al., 2024; Sato et al., 2022), but the NRM of the samples studied by Sato et al. (2022) were unblocked only to <30 mT during tumbling AF demagnetization. Stephenson (1983) shows that for an assemblage of uniaxial grains with a distribution of microcoercivities up to the saturation remanence field, tumbling AF fully removes a saturation ARM when the peak AF level reaches 50% of the field required to produce the ARM. However, IRM acquisition shows that the maximum coercivity of magnetite in Ryugu is ∼300 mT (orange curve in Figure 1 of Maurel et al., 2024), consistent with the maximum theoretically achievable value [i.e., for magnetite whiskers dominated by shape-anisotropy (Dunlop & Özdemir, 1997)]. To summarize, the fact that the NRM is fully demagnetized by tumbling AF with a peak <30 mT field that is only 10% of the saturating field indicates that the high-microcoercivity (>100 mT) magnetite grain population is unmagnetized. As a caveat, we note that the analysis in Stephenson (1983) only considers single grains with uniaxial anisotropy, but it should generally apply to SV grains as well given that their unblocking should depend on the angle of the applied AF to the vortex axis.

Next, the magnitude of the NRM and its behavior during demagnetization for particles A0397, C0085b, and C0006 observed here differ greatly from those of particles A0026 and C0002-4-f reported in Sato et al. (2022) (Figures 5a and 5b). The ratios of NRM to sIRM for A0026 and C0002-4-f after AF demagnetization to 2 mT (Sato et al., 2022) are 13.8% and 9.5% respectively. By comparison, the NRM/sIRM for A0397 and C0006 are 0.21% and 0.08%, respectively. Additionally, the ratios of the NRM lost after AF demagnetization to 24 mT to the sIRM lost at 24 mT are 25% and 19% for A0026 and C0002-4-f, respectively (Sato et al., 2022), while the comparable values for A0397 and C0006 (this study) are 0.1% and 0.2%, respectively. The high NRM/sIRM values for the A0026 and C0002-4-f samples in Sato et al. (2022) indicate that they had been exposed to a strong field (see Section 3.3) and acquired a partial IRM that overprinted grains up to an AF level ∼30 mT. This is consistent with the fact that the components identified as nebular in origin are not origin-trending. We note that this is different from the overprints observed in the Ryugu particles in our study, which we interpret as a VRM.

In addition to these clues from the paleomagnetic data, there are aspects to the histories of the samples measured by Sato et al. (2022) prior to paleomagnetic analysis that could lead to contamination of their NRM records. First, prior to paleomagnetic measurements, the samples in Sato et al. (2022) were heated to 110°C in the presence of Earth's magnetic field (Sato et al., 2022), which could lead to the acquisition of a partial thermoremanent magnetization (pTRM). Second, the samples had also been previously analyzed with an electron microprobe (Sato et al., 2022), which uses an electromagnet that could lead to a weak IRM overprint. In their experiments, the apparent paleointensity estimates decrease with increasing AF level, which is characteristic of an IRM leaking into a high coercivity range with near-zero NRM [see Section 3.6 of Gattacceca and Rochette (2004)]. This is expected given that an IRM is single-axis in nature while tumbling AF activates all directions [see Figure 5 of Stephenson (1983)]. Generally, as the particles in Sato et al. (2022) were analyzed with other instruments prior to NRM demagnetization, it is possible they were exposed to an overprinting magnetic field while being handled or analyzed. By comparison, the samples in this study and in Maurel et al. (2024) were not analyzed with additional instruments prior to NRM demagnetization.

5.2 Comparison to MasMag

Comparison of our results to that from MasMag aboard the MASCOT (Ho et al., 2016) lander enables the first direct comparison between data from spacecraft magnetometry with laboratory paleomagnetic analyses on the same materials from the same body. An initial analysis of MasMag data showed that a 3 m diameter boulder had a weak local magnetization (<3 × 10⁻⁶ Am² kg⁻¹ for uniform magnetization at the ∼3 m scale) (Hercik et al., 2020). In comparison, after the removal of the LC and MC components, the magnetization of the mm-sized samples C0006 and A0397 are ∼2 × 10⁻⁴ Am² kg⁻¹ and ∼2 × 10⁻³ Am² kg⁻¹, respectively. There are at least two potential scenarios that reconcile these data sets. First, we can estimate the maximum strength of the nebular field assuming the boulder has the same magnetic properties as sample A0397 and was homogenously magnetized with a CRM. In this case, given that the magnetization of A0397 after application of an 85 μT CRM-equivalent ARM bias field is 1.01 × 10⁻² Am² kg⁻¹, an upper limit of 0.03 μT can be placed on the nebular field. In the second scenario, the boulder is comprised of smaller magnetic domains with magnetizations similar to the NRMs of the Ryugu particles in this study. Treating the boulder as a collection of 1 mm diameter grains (the sizes of the particles in this study) with randomly oriented magnetic moments matching A0397, the NRM per unit mass of the boulder would be 1.44 × 10⁻⁸ Am² kg⁻¹ [see Equation 3 in Biersteker et al. (2019)], which is consistent with the upper limit set by MasMag. The second scenario is likely more meaningful as there are observations of boulders on Ryugu that possess inclusions on the mm to sub-mm-scale (Grott et al., 2019; Jaumann et al., 2019). In summary, we find that the MasMag results are consistent with either alteration in a weak nebular field or with the boulders being breccias formed after initial alteration.

Further comparison of the results from this study and Sato et al. (2022) to MasMag provide additional evidence that the particles analyzed by Sato et al. (2022) were magnetically contaminated. The large magnetizations of the samples in Sato et al. (2022) (on the order of 10⁻² Am² kg⁻¹), if reflective of the asteroid generally, indicate that MasMag should have been capable of detecting a 1 nT or larger field from particles >6 mm at the relevant distance of 0.2 m (see Equation 2 in Hercik et al. (2020) and Figure S7 in Supporting Information S1). Possible clasts similar to these sizes were observed as inclusions in boulders by MASCOT on the surface of Ryugu (Jaumann et al., 2019), and therefore should have resulted in a field detection by MasMag if they were magnetized at the levels proposed by Sato et al. (2022). By comparison, the particles in this study have magnetizations on the order of 10⁻³ to 10⁻⁴ Am² kg⁻¹, which would only lead to a field detection by MasMag for uniformly magnetized clasts >2–3 cm (see Equation 2 in Hercik et al. (2020) and Figure S7 in Supporting Information S1). The lack of inclusions of this size observed by MasMag is therefore consistent with our results, although what constitutes a clast in surface materials remains somewhat ambiguous (Otto et al., 2023).

5.3 Implications for the Distal Solar Nebula Field

Considering that Ryugu's parent body was aqueously altered within 2.93 Ma of CAI-formation and therefore the nebula was likely present at this time (e.g., Weiss et al., 2017), the strictest paleointensity upper limit of 15.8 μT for A0397 indicates that the nebular field in the distal solar system was weak. The apparent lack of a nebular field record in the Ryugu particles is in stark contrast to the magnetization from nebular fields found in bulk, aqueously altered samples of the Winchombe CM chondrite (Bryson et al., 2023) and possibly CV chondrites (Fu et al., 2021), in addition to previous analyses on CO (Borlina et al., 2021) and LL chondrules (Fu et al., 2014), which recorded fields of >20 μT at <∼7 AU (Figure 5). Thus, our results support the hypothesis that the Ryugu parent body, and affiliated ungrouped carbonaceous chondrites, formed in a region distinct from other carbonaceous and non-carbonaceous chondrites (i.e., Nesvorný et al., 2024). In particular, assuming a typical Class II PPD accretion rate of 10⁻⁸ M_☉ yr⁻¹ (Hartmann et al., 2016), the upper paleointensity limit suggests alteration at a distance of >19.6 AU (using Equation 3 of Weiss et al. (2021) assuming vertical Maxwell stresses via magnetized disk winds drive accretion and their values for the relevant constants in Figure 8), which is consistent with the suggested 13–25 AU formation distance independently inferred from the chemical and isotopic studies of Ryugu and scattering models (Hopp et al., 2022; Raymond & Izidoro, 2017). If the age of alteration were in the future revised to be >3.5 Ma after CAI-formation [e.g., due to new measurements (Yokoyama, Wadhwa, et al., 2023) and/or due to subsequent advancements in our understanding of RSF factors and Mn-Cr dating and/or updated ages of CAIs (Sugawara et al., 2024)], then our upper limit is consistent with, but would not require, that the nebula had dissipated when the magnetite formed.

We note here that our favored upper limit is lower than the results of the paleointensity fidelity tests, which are designed to determine the minimum recoverable field using the ARM paleointensity method. The samples likely fail the fidelity test at relatively high paleointensities due to the poor ability of the particles to acquire the pseudo-NRM. As discussed above, unlike the NRM, which is a CRM, the pseudo-NRM used in our fidelity tests is an ARM. This ARM has a large contribution of spurious remanence from AC waveform as evidenced by the direction of ARM gained relative to the bias field direction and comparison to Tagish Lake (Figure S6 in Supporting Information S1) (see Tikoo et al., 2012 as well). Indeed, the ARM paleointensity is reliant on only one ARM application and demagnetization, while the paleointensity fidelity test requires two ARM applications, and could therefore lead to the sample failing due to a noisy demagnetization. The inability of Ryugu particles to acquire weak bias field ARMs was also observed by Maurel et al. (2024) [see Figure 3 in Maurel et al. (2024)]. In addition, the fact that the mechanism of remanence acquisition for Ryugu particles (CRM) differs from the paleointensity fidelity test (ARM) leads to uncertainty whether the paleointensity fidelity test is an accurate metric of the ability of Ryugu particles to record a weak field. Ideally, the paleointensity fidelity test would be conducted with CRM acquisitions, but it is not possible to dissolve and re-nucleate the framboidal magnetite in a laboratory setting (see Weiss et al. (2017) for an example of a TRM paleointensity fidelity test instead of using ARMs). Therefore, while we favor 15.8 μT as our upper limit on the strength of the distal nebular field, we acknowledge that the upper limit might be as large as 34 μT.

The scale of brecciation and the timing of magnetite formation relative to brecciation of the Ryugu particles could influence the interpretation of the above results and/or our upper limits. We put forth three potential scenarios and explain their implications. In Case 1, magnetite forms after brecciation and final accretion. This scenario is unlikely as the magnetite formed prior to or contemporaneously with the carbonates, which itself crystallized prior to brecciation as evidenced by carbonate fragmentation (Nakamura et al., 2023; Yamaguchi et al., 2023). Additionally, there is evidence of shear in framboidal magnetite aggregates that may be due to post-crystallization shock (Tomioka et al., 2023). In Case 2, magnetite formed prior to brecciation and the brecciation is at the spatial scale of the samples or finer (<1 mm). Evidence for this is that Ryugu's matrix consists of mostly <1 μm phyllosilicates and embedded clasts with sizes of 50–500 μm (Nakamura et al., 2023; Yamaguchi et al., 2023). In this scenario, the magnetite in the samples could have been magnetized by the solar nebula field, but subsequent brecciation and final accretion could lead to randomly oriented clasts, leading us to underestimate the bulk sample paleointensity.

Lastly, in Case 3, magnetite formed prior to brecciation and the brecciation is larger than the spatial scale of the samples (>1 mm). Evidence for this scenario is provided by the observed sizes of inclusions in boulders on the surface of Ryugu as observed by the MASCOT lander, which ranged from the mm-scale to sub-mm scale (Grott et al., 2019; Jaumann et al., 2019). If this scenario is accurate and a nebular field was present during alteration, we would expect our mm-sized samples to be mostly uniformly magnetized. We note that differentiation between Cases 2 and 3 could be achieved by examining variations in the spatial scale of the magnetization in Ryugu particles using a combination of quantum diamond microscopy (∼1 μm resolution; Glenn et al., 2017) and superconducting quantum interference device microscopy (∼150 μm resolution; Weiss et al., 2007). Here, we interpret our results in the context that the samples are non-accretional aggregates and instead are uniform (Case 3).

The weak paleointensity in the distal solar system from Ryugu particles is consistent with theory predicting a decrease of the field with increasing heliocentric distance for a uniform accretion rate with radial distance (Bai & Goodman, 2009; Borlina et al., 2021). Therefore, there is no evidence for a structural gap in the distal region as suggested for the inner solar system based on the mismatch in the inferred accretion rates in different regions of the inner disk (Borlina et al., 2021) (Figure 6). Based on our proposed alteration distance of 7–25 AU for Ryugu, the 15.8 μT upper paleointensity limit corresponds to accretion rates of <7.7 × 10⁻⁹–1.8 × 10⁻⁷ M_☉ yr⁻¹ (see Equation 3 in Weiss et al. (2021); we use f = m = 10 in that equation here and below), which are broadly consistent with previous paleomagnetic studies discussed above indicating accretion rates of 10⁻⁷–10⁻⁹ M_☉ yr⁻¹. However, we stress that this is an upper limit.

Figure 6

Nebular field measurements and bounds in meteorites and comets. (a) Field intensity as a function of heliocentric distance. (b) Field intensity as a function of time after calcium aluminum-rich inclusion formation. Nebular field intensities from previous paleomagnetic studies on meteorites and one comet from the non-carbonaceous reservoir (NC, red), carbonaceous reservoir (CC, green), and distal solar system (blue). New results from Ryugu and meteorites with updated paleointensities/ages are indicated by bolded labels. Data for Winchombe (Bryson et al., 2023), CV chondrites (Fu et al., 2021), LL chondrules (Fu et al., 2014), Tagish Lake (Bryson, Weiss, Lima, et al., 2020; Maurel & Gattacceca, 2023), CO chondrules (Borlina et al., 2021), WIS 91600 (Bryson, Weiss, Biersteker, et al., 2020), Tarda (Bates et al., 2024), CM chondrites (Cournede et al., 2015), Erg Chech 002 (Maurel & Gattacceca, 2024), CO chondrites (Borlina et al., 2022), angrites (Wang et al., 2017), Northwest Africa (NWA) 7325 (Weiss et al., 2017), and spacecraft measurements of comet 67P/Churyumov-Gerasimenko (Biersteker et al., 2019, see note in Text S9 in Supporting Information S1) are plotted as well. Chondrule-derived paleointensities may be inflated by a factor of ∼10–30 depending on their formation mechanism (i.e., nebular shocks) (Mai et al., 2018). We also include previous Ryugu paleointensities (Maurel et al., 2024; Sato et al., 2022). The paleointensity for Ryugu from Maurel et al. (2024) is taken as the anhysteretic remanent magnetization paleointensity of the HC range for particle C0005 (see Table 1), and we acknowledge that the upper limit provided by paleointensity fidelity tests for this study and Maurel et al. (2024) could be as high as 34 μT. CI chondrites are not plotted as the reports are preliminary. All locations of meteorite formation and/or alteration are suggestions. Uncertainties plotted are either 2σ or 95% confidence intervals. Only measurements acquired during the expected lifetime of the nebular field as inferred from inner solar system meteorites are shown. The Sun, proto-Jupiter [assumed to be at 3 AU (Desch et al., 2018)], and the gap around Proto-Jupiter are not to scale. The gray lines represent models of the midplane nebular field around a 1 M_☉ star under the assumption that magnetic stresses drive accretion. The solid and dashed line represent the two cases in which the nebular magnetic field, B, and the sense of disk rotation axis, Ω, are aligned, with the solid line denoting the field due to vertical-toroidal (zφ) Maxwell stresses [see Equation 3 in Weiss et al. (2021)] and the dashed line denoting the field due to radial-toroidal (rφ) Maxwell stresses [see Equation 2 in Weiss et al. (2021)]. The dotted lines represent the field due to vertical-toroidal Maxwell stresses assuming the nebular field and disk rotation are anti-aligned. After Figure 8 and Equation 3 in Weiss et al. (2021) with the addition of the new constraints from Ryugu, Tagish Lake, and WIS 91600.

While the upper paleointensity limit provided by our study does not provide evidence for a disk substructure lying between Ryugu's alteration region and <∼7 AU, these samples alone are unable to conclusively determine if the nebular field in the distal solar system was capable of driving solar accretion. A stronger constraint on the nature of the distal nebular field can be achieved by combining our results with the paleointensities of Tagish Lake, Tarda, and WIS 91600. The much more stringent <0.9 ± 0.3 μT upper limit provided by Tagish Lake places a constraint on the distal nebular field in the region where Tagish Lake experienced alteration, which was likely different than the alteration region for Ryugu (see Section 2.3). While this upper limit is close to the ∼0.9 μT cutoff [assuming accretion at 10⁻⁹ M_☉ yr⁻¹ at 30 AU, see Equation 3 in Weiss et al. (2021)] at which the nebular field would be incapable of driving solar accretion, the detection of a 5.1 ± 4.5 μT distal nebular field in WIS 61900 indicates that the field was capable of driving accretion at rates 8.0 × 10⁻¹⁰ M_☉ yr⁻¹ to 3.0 × 10⁻⁸ M_☉ yr⁻¹ at heliocentric distances 7–30 AU [(see Equation 3 in Weiss et al. (2021)].

In summary, our study of Ryugu particles indicates a weak to null distal nebular field in the region where Ryugu's parent body experienced alteration. While Tagish Lake, Tarda, and WIS 91600 likely formed in a separate region in the distal solar system, their low upper limits and weak paleointensitiesd are consistent with the upper limit set by Ryugu. Therefore, the samples considered here collectively indicate that a distal nebular field with an intensity at least an order of magnitude weaker than the field in the inner solar system could have been present during the first several Ma of solar system history. Furthermore, the detection of a nebular field record from WIS 91600 suggests that the distal nebular field could drive accretion at comparable rates to those inferred from paleomagnetic analyses of meteorites that formed in the inner solar system.

6 Conclusion

Previous paleomagnetic studies have shown that the intensity of the solar nebular field at heliocentric distances <∼7 AU was sufficiently high for magnetic fields to drive accretion at rates similar to those astronomically observed around young stellar objects in the PPD phase. However, the strength of the field in the more distal region of our solar nebula remains poorly constrained. To address this gap, we analyzed the paleomagnetic record of samples returned from asteroid Ryugu, thought to have formed in the distal solar system (>∼7 AU). We find that none of the three particles analyzed possess clear nebular field records at <3 Ma after CAI-formation and place an upper limit on the field intensity of 15.8 μT. The lack of a clear nebular record is in agreement with the results of Maurel et al. (2024) and with the lack of a detection of a remanent field >1 nT by the MASCOT lander on the surface of Ryugu. However, a re-analysis of the ages in meteorites WIS 91600, Tarda, and Tagish Lake, which are also thought to have formed in distal solar system, indicates that they should have undergone aqueous alteration while the nebular field was still active as well. Correcting the paleointensities for recent CRM/ARM efficiency factors where applicable, WIS 91600 has a positive field detection of 5.1 ± 4.5 μT and Tagish Lake and Tarda have null field detections with upper limits of 0.9 ± 0.3 and 1.7 μT respectively. Taken together, Ryugu, Tagish Lake, Tarda, and WIS 91600 paint the picture of a faint nebular field in the distal solar system, but one that was still capable of driving accretion at rates similar to the inner solar system.

Acknowledgments

We thank JAXA for providing samples of Ryugu for this study. We thank Motoo Ito and Yuhji Yamamoto for their assistance in demagnetizing C0006. We additionally thank John Biersteker and Richard Teague for enlightening conversations about spacecraft and astronomical detections of magnetic fields. We also thank Kevin McKeegan for fruitful discussions about Mn-Cr dating and Cauê Borlina for conversations about paleointensities. We thank as well Steve Desch and two anonymous reviewers for thoughtful reviews that improved the quality of the manuscript. BPW, EM, EAL, JBR, KM, and SC thank the NASA Laboratory Analysis for Returned Samples program (80NSSC20K0238) for funding.

Conflict of Interest

The authors declare no conflicts of interest relevant to this study.

Open Research

Data Availability Statement

Data needed to evaluate the conclusions presented in this paper can be found in two locations: (a) Raw magnetization data files for A0397 and C0085 are available on the Magnetics Information Consortium (MagIC) data set at https://earthref.org/MagIC/19927 (Mansbach et al., 2024b); and (b) Demagnetization data for C0006 and SQUID maps (.mat) files are available via Harvard Dataverse at https://doi.org/10.7910/DVN/SPNKFF (Mansbach et al., 2024a).