Alteration of olivine
It is well known that mineral replacements, such as serpentinization, take place primarily by dissolution–precipitation processes (Putnis 2002; Lafay et al. 2012). Etch pits confirmed by SEM (Fig. 2c–f) on the surface of product grains could be the preferential dissolution features at the first step of serpentinization. The dissolution process was crystallographically controlled (Velvel 2009; Malvoisin et al. 2012; Lafey et al. 2018). Thus, these etch pits became deeper and elongated in the [010] and [001] direction as alteration proceeds (Velvel 2009; Malvoisin et al. 2012). Isolated conical etch pits (Fig. 2c, d) would evolve to a network of polyhedral and pyramidal mounts and ultimately to a mammillated (sawtooth) topography (Fig. 2e, f; Malvoisin et al. 2012). It indicates that the alteration degree in grain B (Fig. 2e, f) is advanced as compared with grain A (Fig. 2c, d).
Olivine dissolution proceeds by the breaking of Mg–O bonds and an early rapid reversible exchange of Mg for protons on the olivine surface, which then liberates the SiO44− anions directly into solution in acidic to neutral solution (Luce et al. 1972; Oelker et al. 2018). In contrast, Si–O structures preferentially dissolved ahead of cation dissolution in alkaline solution (Oelker et al. 2018). These processes could result in depletion of Mg or Si on the olivine surface. Mg-poor and Si-rich webbed-like structure covering the surface of olivine after the experiments could be precipitation features of dissolved silicate (Lisabeth et al. 2017). Water obtained in an additional experiment through distillation (see Additional file for details) in this study was approximately at pH 8. Although in situ pH condition of generated water during heating was not clear in this study, carboxylic acids could contribute to weakly acidic water. Decomposition of amides and hexamethylenetetramine, on the contrary, could generate NH3 (Iwakami et al. 1968), which could result in alkaline water.
Amorphous proto-phyllosilicates precipitated onto the etch pits; thus, growth and crystallization of serpentine or talc proceeded in the precipitation area within or on the etch pits (Plümper et al. 2012). Talc could grow with higher silica activity and CO2 concentration (Moore and Rymer 2007; Oelker et al. 2018). Thus, Si-rich flake-like particles shown in Fig. 3 and tubular fibers shown in Fig. 4 were precipitated features of proto-phyllosilicates, and these features could be the same material for the Mg-poor and Si-rich precipitation features given in Fig. 2c, d. The formation of tubular structures (Fig. 4) also indicated the incipient formation of serpentine or talc on the surface of olivine grains. However, we could not confirm highly crystalline phyllosilicate by electron diffraction because the incomplete crystallization structures resulted in unclear lattice images and electron diffraction spots (Lafey et al. 2016). A schematic image of serpentinization based on obtained characteristics in this study is shown in Fig. 5.
The rate of serpentinization typically depends on temperature, pH, water/rock ratio, and initial grain size (Malvoisin et al. 2012). Reaction progress of serpentinization from olivine with 38–50 μm grain size reached nearly 40 % in 250 days at 300 °C, and the major serpentine was lizardite (Malvoisin et al. 2012). Duration of the reaction in this study was 10 days and was likely too short to reach complete serpentinization, although serpentinization rate was the highest at 300 °C compared with those at higher or lower temperature (Wegner and Ernst 1983; Malvoisin et al. 2012). Yada and Iishi (1974, 1977), however, confirmed the formation of serpentine in shorter time (30 min to 10 days) at 250–400 °C at pH 3–13. In that case, conical chrysotile was the major serpentine in fairly wide ranges of pH and temperature, and lizardite existed at higher temperatures and longer durations. Under the alteration conditions in this study, finding proto-serpentine was reasonable, and it could grow toward chrysotile in several months.
Role of organic matter
Serpentinization demands water, but water was not present in the starting material in this study. Water, which is required for serpentinization, could be generated through organic reactions, such as dehydration condensation of carboxylic acids and alcohols in the starting material at 300 °C (Hirakawa et al. in press). Nakano et al. (2020) showed the formation of water from the same organic mixture we used in this study (Additional file 1: Table S1). The amount of water formed depended on the content of hydroxy group (–OH) of the starting materials (Nakano et al. 2020). Decomposition of these organic compounds could form smaller molecules, such as CO, CO2, CH4, H2, C2H6, and C3H8 (Nakano et al. 2003). It is consistent with the generation of CO, CO2, CH4, and H2O through the destruction of insoluble organic matter in thermally metamorphosed chondrite parent bodies suggested by Alexander et al. (2010). In addition, our previous study showed that decarboxylation was promoted by olivine at 300 °C (Hirakawa et al. in press). Magnesite, which could be formed via aqueous carbonation of olivine (see details below), as detected by XRD, also indicates the formation of CO2 from the organic mixture.
Phyllosilicates and magnesium carbonate (magnesite) are thermodynamically stable solids in the MgO–SiO2–H2O–CO2 closed system at 300 °C and under 1–100 bar of partial pressure of CO2 (Oelkers et al. 2018). Magnesium olivine (forsterite) favors reactions that form serpentine and magnesite in accordance with
$$\begin{gathered} {\text{2Mg}}_{{\text{2}}} {\text{SiO}}_{{\text{4}}} + {\text{ CO}}_{{\text{2}}} + {\text{ 2H}}_{{\text{2}}} {\text{O}} \to {\text{Mg}}_{{\text{3}}} {\text{Si}}_{{\text{2}}} {\text{O}}_{{\text{5}}} \left( {{\text{OH}}} \right)_{{\text{4}}} + {\text{ MgCO}}_{{\text{3}}} \hfill \\ \left( {{\text{Forsterite}}} \right)\quad \quad \quad \quad \quad \quad \quad \quad \left( {{\text{Serpentine}}} \right)\quad \left( {{\text{Magnesite}}} \right) \hfill \\ \end{gathered}$$
or talc and magnesite in accord with
$$\begin{gathered} {\text{4Mg}}_{{\text{2}}} {\text{SiO}}_{{\text{4}}} + {\text{ 5CO}}_{{\text{2}}} + {\text{ H}}_{{\text{2}}} {\text{O}} \to {\text{Mg}}_{{\text{3}}} {\text{Si}}_{{\text{4}}} {\text{O}}_{{{\text{1}}0}} \left( {{\text{OH}}} \right)_{{\text{2}}} + {\text{ 5MgCO}}_{{\text{3}}} \hfill \\ \left( {{\text{Forsterite}}} \right)\quad \quad \quad \quad \quad \quad \quad \quad \left( {{\text{Talc}}} \right)\quad \quad \quad \quad \left( {{\text{Magnesite}}} \right) \hfill \\ \end{gathered}$$
under higher CO2 concentration in water (Oelkers et al. 2018). Organic-induced serpentinization is possible with simultaneous water formation from dehydration of organic compounds. It indicates the possibility of in situ formation of phyllosilicates inside the H2O snow line because refractory organic compounds, in contrast to H2O, could survive during accretion of planetesimals in the area, as discussed below.
Water/rock ratio of ordinary chondrites was estimated as ~0.1–0.2 (Doyle et al. 2015). Water/rock ratios of carbonaceous chondrites were estimated by Marrocchi et al. (2018), the ratio of CM chondrite was 0.4, CR was 0.1–0.4, CV was 0.1–0.2, and CO was 0.01–0.10. Glavin et al. (2018) also summarized the water/rock ratio of carbonaceous chondrites based on Brearley (2006) and Krot et al. (2006), and the ratio of CI was 1.1–1.2, CM was 0.3–0.6, and CR was 0.4–1.1. Thus, water/rock ratio in this experiment (0.13, see Additional file for details) was consistent with CV, CO, and ordinary chondrites.
Possible aqueous alteration in the “Dry” parent bodies
Recent astronomical observations revealed the existence of organic molecules with hydroxy groups, such as CH3OH (methanol) (Walsh et al. 2016) and HCOOH (formic acid) (Favre et al. 2018), in protoplanetary disk, as well as CH3CH2OH (ethanol), CH3COOH (acetic acid), and HOCH2CHO (glycolaldehyde) were suggested by model calculation (Walsh et al. 2016). CH3CH2OH, CH3COOH, and HOCH2CHO were detected from the low-mass protostar IRAS 16293–2422 (Bisschop et al. 2008; Jerry Shiao et al. 2010; Jørgensen et al. 2012). They could be delivered inside the snow line on dust particles. Water and highly volatile organic compounds evaporated inside the snow line, but the other organic molecules could remain on the grain surfaces. At least, organic compounds, such as polyaromatic hydrocarbons, aliphatic hydrocarbons, alcohols, carboxylic acids, and macromolecular organics detected in cometary dusts, could survive inside the sublimation area of water ice (Herbst and van Dishoeck 2009; Walsh et al. 2014). In that case, organic matter could accumulate on planetesimals without water or with little water.
Ordinary chondrites originate from S-type asteroids located in the inner region of the asteroid belt inside the snow line. However, some ordinary chondrites show aqueous alteration features. For example, the Semarkona meteorite (LL3.0) has phyllosilicates (Alexander et al. 1989; Piani et al. 2015). The major phyllosilicate phase in Semarkona is smectite, and carbonate is calcium carbonate (Hutchison et al. 1987; Alexander et al. 1989), although proto-phyllosilicates and magnesium carbonate formed in the present study. These differences could be due to the glassy mesostases in Semarkona (Hutchison et al. 1987). The glassy mesostases are highly susceptible to hydration compared with crystalline olivine (Dobrica and Brearley 2020). The total amount of water generated from organic matter could be much less than that present in aqueously altered carbonaceous chondrite parent bodies. Thus, in the case of Semarkona, the small amount of water generated from organic matter could contribute to dissolution of glassy mesostases and some metal ions from the other minerals to form smectite and calcium carbonate (Hutchison et al. 1987), preceding the dissolution of olivine. In addition, the Semarkona parent body could have heterogeneous textures with variable porosity (Dobrica and Brearley 2020), and thus, water could move through the grain boundaries, while the alteration condition in our experiment was static with only one mineral (olivine) in a closed system.
Water abundance in the bulk Semarkona was estimated as 0.2–0.7 wt.% (Alexander et al. 2010). Assuming that primordial organic matter accumulated in the Semarkona parent body was similar to the organic mixture used in this study, at least 8–28 wt.% of organic matter has to initially present in the parent body necessary for producing the amount of water in Semarkona, since the water produced from the organic mixture was 2.5 wt.% (see Additional file for details). Organic carbon in the bulk Semarkona meteorite as insoluble organic matter is 0.36 wt.% (Alexander et al. 2007), thus, if the water was delivered from organics, some of organic matter should have been lost through subsequent thermal processes.
The results in the present study indicate that the hydration of anhydrous silicate was possible in their parent bodies with organic matter at least locally, even if water ice did not accumulate on them. However, there could be some differences in conditions between parent bodies and the experiment—such as mineral compositions and static/flow of water, and thus, further studies are needed to evaluate the details of such alteration processes. Some thermally metamorphosed chondrites classified in CV and CO chondrites with alteration features could have captured water near, or in some case inside, the snow line (Marrocchi et al. 2018); thus, the role of organic matter to generate water to produce hydrated silicates would be minor. The water generated from organic matter could significantly contribute to alteration of anhydrous minerals, only in case of the accretion inside the snowline.
Our results further implied that organic matter was one of the possible sources of deuterium-rich water in ordinary chondrites. For example, water in CM chondrites are homogeneously D poor (Alexander et al. 2012). The water could have originated from isotopic re-equilibration in the inner disk between gaseous H2O and the D-depleted solar H2 (Jacquet and Robert 2013). In contrast to these, the D/H ratios of water in ordinary chondrites are highly heterogeneous (Alexander et al. 2017). Some of them show significantly high D/H ratios, similar to some comets (Piani et al. 2015). The existence of D-rich water in ordinary chondrites was explained by an oxidation reaction of Fe metal (3Fe + 4H2O = Fe3O4 + 4H2; Alexander et al. 2010, 2017). Loss of isotopically light H2 gas would have enriched the residual water in deuterium (Alexander et al. 2017). Piani et al. (2015), however, pointed out that the process would have required a large amount of initial water and a significant amount of oxidation reactions to explain the high D/H values. They suggested that the D-rich water could originate from isotopically heterogeneous ice inherited from the interstellar ice (Piani et al. 2015). Alternatively, Remusat et al. (2016) suggested that the D-rich, recalcitrant organic matter in some ordinary chondrites could be interstellar origin. In addition to these hypotheses, Nakano et al. (2020) suggested that precometary organic matter could be another water source in ordinary chondrite parent bodies. High deuterium concentration in organic matter is well known in the interstellar medium (ISM) and the outer region of the Solar System (Owen et al. 1986; Mauersberger et al. 1988; Cecilia 2002; Parise et al. 2004). These organic molecules could be enriched in deuterium by ion molecular reaction (Millar 2003) in the gas phase and/or by grain surface reactions in molecular clouds (Parise et al. 2004; Watanabe and Kouchi 2008; Taquet et al. 2012). Thus, some organic matter could survive the inner region of the solar nebula (Walsh et al. 2014). Although deuterium tends to concentrate on methyl or ethyl group more than on hydroxy groups in alcohols in molecular clouds (Nagaoka et al. 2005; Oba et al. 2016), deuterated molecules, such as CH3OD, were confirmed in ISM (Mauersberger et al. 1988). Thus, both D-poor and D-rich water could be generated from hydroxy groups in organic matter, and such water could then contribute to both D-rich and D-poor hydrated minerals, as well as explain the heterogeneous D/H distribution in ordinary chondrites. Further studies are needed to understand the D/H fractionation between organic matter and phyllosilicates. In addition, organic matter in ordinary chondrites is generally 15N-poor as compared with 15N-rich organic matter in the outer Solar System. However, nitrogen abundances in ordinary chondrites are small, and 15N-rich organic matter may be preferentially removed during thermal metamorphism (Alexander et al. 1998).