2023 年 57 巻 5 号 p. e9-e16
Elemental and isotopic analyses of individual submicron-sized particles in chondrite matrix were made by an inductively coupled plasma time-of-flight mass spectrometer (ICP-TOF-MS) and a multiple collector ICP-MS equipped with high-time-resolution ion counters (HTR-MC-ICP-MS). The particles were collected from Allende CV3 chondrite through a laser ablation-in-liquid (LAL) technique. Firstly, the abundances for four major elements (Si, Al, Mg, and Fe) were determined on total 6086 particles, indicating that the Allende matrix is a mixture of submicron-sized particles made mainly of olivine, pyroxene, spinel, Fe–Ni sulfide, and Fe–Ni metal, consistent with the predicted major constituent minerals by a nebular condensation model. The major elemental compositions revealed that Fe–Ni particles are minor components (about 0.3% in number fraction) in the Allende matrix. Then, to estimate the origin of these metallic particles, abundances for Ni and two minor elements (Os and Pt) were measured. Total 10417 particles of Ni–Os–Pt bearing particles were also found in the chondrite matrix. Majority of the particles were enriched in Ni. Os and Pt were present as separated particles, and no particles with presence of both the Os and Pt were found. Finally, with the HTR-MC-ICP-MS technique, 195Pt/194Pt value was measured on total 1545 particles. The resulting 195Pt/194Pt values agree with the solar composition within analytical uncertainties. This lack in isotopic anomalies of the 195Pt/194Pt can be explained either by majority of the Pt nuggets being produced from uniform reservoir in the solar system or by Pt being isotopically homogenized prior to the formation of the solar nebula.
Chondrite matrix is an opaque mixture of submicron-sized grains ranging from 10 nm to 5 μm in size and fills in the interstices between chondrules, CAIs, other components (Scott et al., 1988; Scott and Krot, 2007). Matrix commonly accounts for about 1–60 vol% of all chondrites regardless of chemical groups, and minerals found in the matrix include silicates, oxides, sulfides, metallic Fe–Ni, phyllosilicates, and carbonates (e.g., Lee et al., 1996; Brearley and Jones, 1998). Mineralogical and cosmochemical studies on chondrite matrices mainly focused on coarser grains (e.g., ≥1 μm), suggesting that matrix minerals are complex mixture of presolar materials and nebular condensates that were mixed with fine chondrule fragments and experienced aqueous alteration and metamorphism (Scott and Krot, 2007).
Cosmochemical study on chondrite matrix has been retarded by the extremely fine-grained nature of the constituents, and the difficulty of distinguishing primary and secondary mineralogical features (Brearley, 1996). Nevertheless, the fact remains that the chondrite matrix is one of the major components of the chondrite, and thus, origin of the individual grains can provide further insights of the solar system evolution. In a recent study, elemental analysis of fine grains (smaller than 1 μm) in chondrite matrix was conducted through imaging analysis using a fast ion bombardment time-of-flight type secondary ion mass spectrometry (FIB-TOF-SIMS). Aluminum-rich fine particles could be found with a spatial resolution of higher than 100 nm (Morita et al., 2022). It should be noted that the most of the pioneering researches focused on major components of the matrix particles and there were few previous reports concerning elements being heavier than Fe.
It is widely recognized that cosmochemical information concerning the beginning of the early solar system formation can be derived from elemental analysis of submicron-sized particles for refractory heavy elements, such as platinum group elements (Macdougall and Goswami, 1981). Moreover, elemental and isotopic signatures derived from the refractory metal elements can provide information concerning nucleosynthetic processes of the heavy elements, including s- and r-processes (e.g., Yokoyama and Tsujimoto, 2021). In fact, the relative deviations of the averaged Pt isotopic ratios (195Pt/194Pt) produced through the s- and r-processes from the solar composition (Anders and Grevesse, 1989) are about 59% and 2%, respectively, estimated based on the data by Arlandini et al. (1999). More importantly, individual 195Pt/194Pt values, not averaged values, of the original materials released through the supernova explosions can vary >100% estimated based on the data by Panov and Lutostansky (2020) as reflecting the physical conditions of the supernova explosions (e.g., neutron flux, temperature, time duration, system entropy, or neutron capture cross section of the nuclides). Thus, the isotopic signatures for individual particles in meteorites are important to unveil the detailed nucleosynthetic processes of the heavy elements. Hence, a new technique capable of elemental and isotopic analysis for heavy elements in superfine particles with high sensitivity is highly desired.
To achieve this, two major analytical challenges must be considered. One is the elemental and isotopic analyses, which have to be made on fairly large number of particles in chondrite matrix. This is especially important to understand the abundances of minor components. Second is the elemental and isotopic analyses on individual superfine particles, not averaged values obtained through bulk analysis. In this study, we have developed a new analytical technique to derive elemental and isotopic signatures from individual superfine particles in chondrite matrix using an inductively coupled plasma time-of-flight mass spectrometry (ICP-TOF-MS) and a multiple collector ICP-MS system equipped with high-time-resolution ion counters (HTR-MC-ICP-MS). To evaluate the analytical capabilities of the present ICP-MS techniques, elemental and isotopic analyses were conducted from matrix components in the Allende CV3 chondrite. Allende CV3 chondrite would be the best chondrite for evaluation of analytical capabilities of the system because the CV3 chondrite matrices have been well studied for coarse particles and can be dominated by coarse iron-rich olivine (Fa30–60) (Scott et al., 1988; Brearley and Jones, 1998). The superfine particles in Allende chondrite matrix were collected through a laser ablation-in-liquid (LAL) technique (Okabayashi et al., 2011).
It should be noted that the mineralogy and cosmochemistry of the CV3 chondrite matrices prior to secondary alteration including aqueous alterations are not well studied. To overcome this, we have tried to derive the elemental and isotopic signatures from individual matrix particles being smaller than 1 μm in size using the newly developed analytical system setup. The elemental analysis of individual particles was conducted by ICP-TOF-MS, and 195Pt/194Pt analysis on individual particles was conducted by HTR-MC-ICP-MS technique installed at Univ. Tokyo (e.g., Hirata et al., 2019; Yamashita et al., 2020; Hirata et al., 2021; Kurihara et al., 2022).
In this study, elemental and isotopic signatures of individual submicron-sized particles in matrix of Allende CV3 were measured. The thick cut piece of the Allende meteorite (20 × 30 mm in size) was embedded in resin (CaldoFix-2: Struers ApS, Ballerup, Denmark) and then polished successively by a series of abrasive sheets (Lapping Film Sheet #600, #1200, #2000, #4000, #8000: 3M, Maplewood, USA). For elemental analysis using the ICP-TOF-MS system, the sample was re-polished by diamond paste 1 μm in size (Musashino Denshi, Inc., Musashino-city, Japan) to remove potential contaminants (i.e., Al2O3) from the lapping film sheets. Submicron-sized particles were collected from the sample through the LAL technique. Sampling was carried out with optimized laser conditions to reduce laser-induced breakage or following agglomerations. The resulting sample suspensions were diluted to about 10–100 times using deionized water to reduce a probability of coincidence and overlapped particle detection which causes erroneous measurements for elemental analysis from individual particles (e.g., Yamashita et al., 2020; Hirata et al., 2020; Yamashita et al., 2021). Based on the resulting particle number concentrations and present time resolution of the system (2 ms for elemental analysis and 30 μs for isotopic analysis), contribution of the coincidence and overlapped particle detection can be smaller than 1% (Shaw and Donard, 2016).
For the elemental analysis of major elements (Si–Al–(Fe+Mg)) in the matrix submicron-sized particles, ICP-TOF-MS system (Vitesse, Nu Instruments, Wrexham, UK) was employed. To enhance the elemental sensitivity of the ICP-TOF-MS system, a desolvating nebulizer sample introduction system (Apex Omega, Elemental Scientific, Omaha, USA) was used (Hadioui et al., 2019; Yamashita et al., 2021). The measurement of the sample was conducted for 39 minutes with a time slice of 200 μs. Peak identifications and signal integrations for individual pulsed signals emanating from individual sample particles (particle events) were conducted using an in-house software (NP Shooter: Kurihara et al., 2022). Mass bias factors were corrected based on the relative sensitivity factors (RSFs) for Mg, Al, Si, and Fe defined by signal intensity data obtained by standard solutions and laser ablation on glass reference materials (GSE-2G and NIST SRM 610). Because of the limited instrumental sensitivity and signal input-output linearity of the present system, elemental analysis (Si–Al–(Fe+Mg)) was made on particles being greater than apparent size of 300 nm. Hence, the apparent size is defined as calculated size of particles based on signal intensities of 28Si. This suggests that the actual size should be greater than the apparent size if particles contained elements other than Si. For size calibration, commercially available SiO2 nanoparticles (500 nm in diameter from nanoComposix, San Diego, USA) were used as size reference materials.
As for the analysis of elemental and isotopic signatures, Ni–Os–Pt bearing matrix particles were measured. The Os and Pt abundances in small particles reflect different physicochemical conditions depending on formation processes such as nebula condensation or secondary alteration processes (Palme et al., 1982; Campbell et al., 2003). To do this, ICP-TOF-MS (icpTOF R: TOFWERK, Thun, Switzerland) was used. The sample was measured for 72 minutes with a time slice of 2 ms. The particle size was calibrated by 70 nm Pt nanoparticles (nanoComposix, San Diego, USA). Sample solution was introduced into the ICP through the desolvating nebulizer system.
After the elemental analysis of matrix particles, isotopic analysis of Pt on individual particles was conducted. For a precise analysis of Pt isotopic ratios in individual particles, HTR-MC-ICP-MS (Nu Plasma II: Nu Instruments, Wrexham, UK) was employed (Yamashita et al., 2020; Hirata et al., 2020). With our HTR-MC-ICP-MS system, at most six isotopes can be simultaneously detected. Hence, two isotopes (194Pt and 195Pt) were monitored considering a limited mass dispersion of the instrument and a limited configuration of ion detectors. To extend the signal input/output linearity of the ion detector, two Daly detectors equipped with the high-time resolution ion counters were used to measure ion currents of 194Pt and 195Pt. The in-house ion detection system using the Daly detectors enables us to monitor high-ion currents being 107 cps (Obayashi et al., 2017). The isotopic ratio was calibrated by a 1 ng g-1 Pt standard solution. The sample solution was introduced into ICP through a conventional solution nebulization technique. The signal intensities of 194Pt and 195Pt isotopes were acquired for 17 minutes with a dwell time of 30 μs. The raw signal intensity data were processed for identification of particle events using an in-house software (NanoQuant: Suzuki et al., 2019; Hirata et al., 2020). Details of the sampling using the LAL technique, instrumentations, and operational settings of the ICP-MS systems are described in supplemental information A.
Elemental analysis was carried out with the analytical throughput of about 150 particles per second, and 6086 particles were detected in total. The largest particle was smaller than 1 μm. The resulting abundance values for Si, Al, and (Fe+Mg) were listed in supplemental information C, and were plotted on a ternary diagram (Fig. 1). Particles being smaller than 300 nm were not plotted on the ternary diagram because of great uncertainties derived from counting statistics of the signals. The relative uncertainties associated to the signal intensities were more than 5% for particles smaller than 300 nm. For comparison, Si–Al–(Fe+Mg) abundance values for several minerals (e.g., olivine, pyroxene, Fe–Ni sulfide, Fe-bearing oxide, Fe–Ni metal, and spinel) were also given. The mineral classification was based on the resulting chemical compositions for individual particles. Since any crystallographic information is derived from the particles, presence of amorphous minerals cannot be ruled out. Nevertheless this, several pioneering researches revealed that most of the constituent minerals in the Allende matrix were crystals (e.g., Scott et al., 1988; Buseck and Hua, 1993; Keller et al., 1994; Brearley and Jones, 1998), and thus, the minerals were named as their crystalline form. The resulting number fractions for the individual mineral particles were given in Fig. 1. The majority of the particles is olivine or pyroxene (40%), followed by Fe–Ni sulfide including pentlandite (6.4%), spinel (5.1%), and Fe-bearing oxide (1.6%). These mineral compositions were consistent with previous studies of the Allende matrix (e.g., Scott et al., 1988; Buseck and Hua, 1993; Keller et al., 1994; Brearley and Jones, 1998; Neuland et al., 2021). It is widely recognized that the Allende matrix was suffered from secondary aqueous alterations (e.g., Bunch and Chang, 1980). The presence of Fe–Ni sulfide (6.4%) reflects the large contribution of the aqueous alterations of the Allende matrix. Figure 1 also represents that the Fe–Ni particles are minor components (about 0.3% in number fraction).
Chemical compositions of submicron-sized particles in matrix of Allende CV3 chondrite. The resulting compositions of individual submicron-sized particles were plotted on Si–Al–(Fe+Mg) ternary diagram, and were used to estimate minerals such as olivine, pyroxene, Fe–Ni sulfide, Fe-bearing oxide, Fe–Ni metal, spinel, and unidentified minerals including feldspar and feldspathoid. For comparison, stoichiometric compositions of minerals in the CV chondrites reported by Buseck and Hua (1993) are shown as open circles.
Another important feature derived from Fig. 1 is that the number fraction for the “Mixture” exceeds 40%. This is also consistent with the previous report that the olivine particles contain small pentlandite (Fe–Ni sulfide) (Brearley, 1999) or spinel (Weinbruch et al., 1994). Moreover, some of the submicron-sized particles can be present as agglomerates of small minerals (Toriumi, 1989), and thus, the chemical composition of the particles can be identified as the mixture. Important point is that reliable elemental information can be obtained even from individual particles extracted by the LAL technique.
Ni–Os–Pt bearing particlesElemental fractionations among the refractory siderophile elements including Os and Pt, reflect the physicochemical conditions during the particle formation and subsequent alterations, such as condensation, oxidation/reduction, or metal segregations (Palme et al., 1982; Campbell et al., 2003). Moreover, the particles enriched in refractory elements can survive secondary alterations, suggesting that these particles can preserve inherent isotopic signatures of the heavy elements at the early solar system. A considerable number of metal particles has to be collected from the Allende matrix in order to obtain reliable elemental and isotopic signatures from minor components. To achieve this, LAL sampling was conducted from wider area in the matrix (560 μm × 470 μm) with a higher-repetition rate of the laser (10,000 Hz) (see supplemental information B). The resulting sample solution was subsidized to trace-element analysis using the ICP-TOF-MS.
With the present ICP-TOF-MS (icpTOF R), total 10417 particles enriched in Ni, Os, or Pt were found. With the icpTOF R system, because of both the high contribution of the background counts on m/z 57 and their oscillations mainly due to instability of the notch filter, reliable abundance values for Fe could not be obtained, and thus, the Fe abundance values for the Ni-bearing particles were not used for further discussion. The resulting Ni, Os, and Pt abundance values for individual particles were plotted on Ni–Os–Pt ternary diagram (Fig. 2). The resulting apparent sizes (in nm) for the individual particles were plotted on the vertical axis. Hence, the apparent size was defined under an assumption that the particles consist of three elements (i.e., Ni, Os, and Pt). This suggests that, likely with the size analysis of the major matrix minerals, actual sizes become greater when the particles contain other elements such as Fe or S.
Ni–Os–Pt ternary diagram for Ni, Os, or Pt-bearing particles collected from the Allende matrix. Apparent size of the particles was given in vertical axis (see text for details). The numbers in parentheses represent the number of the particles detected.
Analytical uncertainties in the Ni, Os, and Pt abundance measurements were basically controlled by the counting statistics (i.e., shot noise) of the signals. Since the typical sizes of Ni-bearing particles were greater than 50 nm, the uncertainties in the Ni abundance measurements were about 10% or smaller. In contrast, the uncertainties for Os and Pt were about 50%. Because of the large data uncertainties, no quantitative discussion was made for Os/Ni or Pt/Ni values of individual particles in this study. To minimize the contribution of the counting statistics for the analytes, the ICP-MS system with the higher ion transmission is highly desired.
Figure 2 shows that the majority of the particles are Ni-bearing particles (about 97% in number fraction), and the number fractions for Pt- or Os-bearing particles are less than 2%, demonstrating that the number fraction ratios for Os/Ni and Pt/Ni are about 10–2. With the combination of the Os/Ni and Pt/Ni values and the number fractions of Ni-bearing particles in the overall matrix minerals (6.4% as sulfide and 0.3% as Fe–Ni bearing particles, and 6.7% in total), the number fractions of the Os- and Pt-bearing particles are about 500 ppm (particles per one million particles). Now we can estimate the Os/Ni and Pt/Ni values for the Allende matrix based on the measured ion counts for Ni, Os, and Pt; 460,000 counts for Ni, 400 counts for Os, and 450 counts for Pt. The calculated Os/Ni and Pt/Ni values were about 10–3, demonstrating that these are about 10-times greater than those estimated from the solar abundances for Os and Pt (i.e., Os/Ni and Pt/Ni being 10–4 [Anders and Grevesse, 1989]). The measured greater elemental ratios could reflect high robustness of the Os- and Pt-bearing particles through aqueous alterations at the sampling areas. The contribution of the secondary alterations was also supported by the number fractions of the major minerals found in this study (Fig. 1).
Nickel was present as Fe–Ni metal or sulfide (6.7% of all particles), and thus, the particles with presence of both the Os and Pt (Os–Pt particles) could be produced through oxidation processes of the Fe and Ni. However, this fails to explain the lack in Os–Pt particles because it is natural to consider that both the Os and Pt would present as solid solution in Fe–Ni metals. One possible explanation for the absence of the Os–Pt particles is that these particles were formed as direct condensates from nebula gas. In fact, the condensation temperatures for Pt and Os are distinctively different (Sylvester et al., 1990), and therefore, Os and Pt would be separately condensed from the nebula gas. At a moment, we do not have any further evidence to support this model, so this must remain as a possibility. Nevertheless, other possibilities mainly require geochemical fractionation processes of the elemental ratio, such as generation of silicate or sulfide melts in deep regions of parent bodies (Barnes et al., 1985). However, these geochemical events should be ruled out due to the characterization of Allende as CV3. Despite the possible contribution of the secondary alterations, we believe that the new scientific insights for the chondrite matrix could be derived through the analysis of individual small particles (<1 μm).
195Pt/194Pt values for individual particlesIsotopic signatures can also provide piercing information concerning the solar system formation or origin of the heavy elements. In this study, the isotopic signature of Pt (i.e., 195Pt/194Pt values) was measured by the HTR-MC-ICP-MS technique. Likely with the sample solutions for Ni–Os–Pt particles, slightly high energy fluence and repetition rate were used at the LAL sampling (see supplemental information B), and the resulting sample solution was subsidized to the isotopic analysis of 195Pt/194Pt. To reduce a contribution of the counting loss due to a detector’s refresh time, ion signals higher than 107 cps (e.g., 300 counts per 30 μs dwell time) were not used for further calculations.
The resulting 195Pt/194Pt values for individual particles were listed in supplemental information C, and were plotted against the signal intensity of 194Pt (Fig. 3). Hence, relative deviation of 195Pt/194Pt from the solar value was plotted on vertical axis. The error bars represent uncertainties estimated by 2-times counting statistics on 194Pt and 195Pt signals. Total 1545 Pt-bearing particles were found from the sample solution prepared from the Allende matrix. As described in the earlier section, relative number fraction of Pt-bearing particles against overall particles in the Allende matrix can be estimated by the number fractions of Ni-bearing particles (Fig. 1) and Pt-bearing particles (Fig. 2). Based on the present analysis protocol, to derive isotopic signatures from about 1000 Pt-bearing particles, greater than one million matrix particles must be subsidized to analysis. The resulting 195Pt/194Pt values for all particles were plotted within two curves representing the 95% confidence levels, suggesting that the 195Pt/194Pt values for the particles agreed with the solar value within analytical uncertainties deduced by the counting statistics, and no clear isotopic anomaly could be found. The lack in the isotopic deviation in the 195Pt/194Pt values suggests that majority of the Pt-bearing particles in the solar system were formed either by oxidation processes of Fe–Ni nuggets or by reduction processes from uniform reservoirs within a meteorite parent body. Another possibility is that Pt-bearing particles were formed as condensates from isotopically homogenized gas with the solar composition. From the homogeneous nature of the Pt isotopes for Pt-bearing superfine particle, the question now becomes how to mix and homogenize the Pt isotopes of various origin. One possible explanation for the lack in the isotopic anomaly of the 195Pt/194Pt values is that all the Pt-bearing particles were produced from an isotopically homogenized reservoir within the solar system through condensation or segregation processes. Besides this, another possibility is that the heterogeneity of the Pt isotopes had already been moderated during star formation processes. Platinum of various isotopic signatures produced under different nucleosynthetic conditions was accumulated and homogenized through star formation processes. The homogenized components and newly synthesized components inside stars were then mixed and released into space through explosions of supernovae. Several recent studies, however, revealed that supernova explosions cannot be a main source of the heavy elements mainly due to a low neutron flux achieved in supernovae (e.g., Fischer et al., 2012; Wanajo, 2013), suggesting that the supernova events acted as a homogenizer for heavy elements derived from former-generation nucleosynthesis. We do not have further material evidence for this explanation, so this must be remained as a possibility. The plausible conclusion of this study is that the combination of isotopic signatures of various elements with different volatility would be effective to understand further insights of the isotopic signature of the heavy elements in submicron-sized particles.
195Pt/194Pt values for individual particles collected from the Allende matrix. Relative deviation from the solar value was plotted against signal intensity of 194Pt. Error bars represent 2-times counting statistics.
Elemental and isotopic analyses were conducted on individual submicron-sized particles collected from the Allende matrix using an ICP-TOF-MS and HTR-MC-ICP-MS. Several important features can be derived from the data.
1. Minerals and their number fractions deduced by major phase-forming elements (Mg, Al, Si, and Fe) for individual particles suggested that the minerals presented in the Allende matrix are consistent with the major constituting minerals reported by previous studies (e.g., Scott et al., 1988; Buseck and Hua, 1993; Keller et al., 1994; Brearley and Jones, 1998).
2. Total 10417 of Ni–Os–Pt bearing particles could be found. The majority of the particles were Ni-bearing, and the number fractions for Os- and Pt-bearing particles did not exceed 2%. No Os–Pt particles could be found in this study. The lack in the Os–Pt particles can reflect the elemental fractionation of Os and Pt through formation sequence of the particles. One possible explanation for the absence of Os–Pt particles is that the particles were produced through separated condensations of Os and Pt, reflecting the large difference in the condensation temperatures for Os and Pt (Sylvester et al., 1990).
3. The 195Pt/194Pt values for total 1545 particles did not vary from the solar value within analytical uncertainties. This suggests that Pt-bearing particles were produced from a uniform reservoir or gas of solar composition. The lack in the 195Pt/194Pt anomaly strongly requires the contribution of homogenization process at the meteorite parent body and/or solar nebula.
Based on the elemental and isotopic data presented here demonstrate clearly that the ICP-TOF-MS and HTR-MC-ICP-MS have a potential to become a sensitive and fast analytical tool for individual submicron-sized particles. The techniques have an immediate potential as reconnaissance techniques for geochemistry dealing with the particles such as cosmic dust, regolith, airborne particle, suspension or sediment, or biomineral in tissues through large-scale data science.
There are no conflicts to declare.
We are grateful to Dr. Hiroaki Takahashi (Nuclear Regulation Authority, Japan), Dr. Hisashi Asanuma (Kyoto University, Japan), and Tetsuya Tamaki (Fab instruments, Japan) for technical support. This work was financially supported, in part, by a Grant-in-Aid for Scientific Research (JP21H04511) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
