Space weathering effects and potential spectral alteration on Phobos and the Moon: Clues from the Fe content of olivine

Context. Olivine responds to space weathering in the fastest and most profound way, which results in signiﬁcant space weathering spectral alteration effects (SWSAEs) on airless silicate bodies. Although Mg-rich olivine (Fa 10 ) has been subjected to extensive studies, SWSAEs of Fe-rich (Fa# > 20) or Fa-dominant (Fa# (cid:62) 50) olivine are still poorly understood. Aims. We aim to systematically characterize the space weathering effects and the associated spectral alterations of Fe-rich olivine on the surface of Phobos and the Moon. Methods. We conducted nanosecond pulsed laser irradiation experiments on a set of synthetic Fe-rich olivine (Fa 29 , Fa 50 , Fa 71 , and Fa 100 ) with energy levels simulated for Phobos and the


Introduction
The surface of airless bodies is subjected to long-term space weathering.While various processes may simultaneously contribute to space weathering, for inner Solar System bodies, the dominant processes are cosmic irradiation, solar wind, and micrometeoroid bombardment (Hapke 2001;Pieters & Noble 2016).For a specific celestial body, the effects of space weathering are determined by a combination of factors, including its position in the Solar System (e.g., impact velocity and flux, radiation environment, and temperature), compositions, physical properties, and exposure history of the surface materials (Pieters & Noble 2016;Weber et al. 2020).
Previous studies of the Moon and S-type asteroids suggested that olivine responds to space weathering in the fastest and most profound way due to the formation of nanophase iron (np-Fe 0 ) particles (see e.g., Kohout et al. 2014;Sasaki et al. 2001).The formation of np-Fe 0 particles along with other nanophase opaque particles, amorphization, and the formation of vesicles are the primary cause of alteration of optical properties of the airless bodies, namely, space weathering spectral alteration effects (SWSAEs), including reduced spectral reflectance (darkening), increased spectral slope (reddening), and reduced particular absorption bands in the ultraviolet to visible and nearinfrared (NIR) reflectance spectra (see e.g., Chrbolková et al. 2022;Hapke 2001;Pieters & Noble 2016).In contrast, pyroxene, another primary rock-forming mineral on a silicate celestial body, is more resistant to the formation of np-Fe 0 , and therefore its reflectance spectra are less affected than those of olivine during space weathering processes (Pieters & Noble 2016;Sasaki et al. 2002;Yamada et al. 1999).
Since np-Fe 0 particles were discovered in the lunar samples returned by the Apollo mission, substantial efforts have been made to investigate space weathering effects on olivine, including the analysis of meteorites and returned samples, laboratory simulation experiments, spectral analyses, and modeling.These studies included a comparison of space weathering effects on olivine induced by solar wind implantation versus micrometeoroid and dust impacts (see e.g., Chrbolkova et al. 2021;Yamada et al. 1999); formation mechanisms of np-Fe 0 in olivine (see e.g., Keller & McKay 1997;Noble et al. 2007;Vandemoortele et al. 2007; Wang et al. 2012); atomic-level mechanisms for the favorable formation of np-Fe 0 in olivine rather than in pyroxene (Quadery et al. 2015;Shilobreeva et al. 2014); the influence of sizes and amounts of np-Fe 0 particles on reflectance spectra of olivine (see e.g., Kohout et al. 2014;Markley & Kletetschka 2016); and the infrared spectral inversion model for detecting and quantifying olivine (see e.g., Cloutis et al. 2015;Han et al. 2020;Pinet et al. 2022;Yang et al. 2017).However, previous studies almost exclusively focused on Mg-rich olivine (e.g., San Carlos olivine) and rarely involved Fe-rich olivine.Although a spectral library of the synthetic forsterite-fayalite solid solution set is available (Dyar et al. 2009;Pinet et al. 2022), it generally lacks data of the same set after being space-weathered.How Fe-rich olivine (Fa# > 20; Fa# represents Fe mol%), or even fayalite-dominant olivine (Fa # 50), responds to a specific space weathering process (e.g., micrometeoroid impacts, solar wind sputtering) in terms of the microscopic characteristics and optical properties is poorly understood.
Newly returned Chang'E-5 (CE-5) lunar samples (Li et al. 2021) and the planned exploration mission to the Martian moons, such as the Martian Moons exploration mission by the Japan Aerospace Exploration Agency (Kuramoto et al. 2022) make the need for this perspective more urgent.The CE-5 lunar regolith is young and composed of fayalitic olivine compared to the older and Mg-rich Apollo sample (Li et al. 2021).Different dominant space weathering processes and new np-Fe 0 formation mechanisms (i.e., thermal decomposition, Fe disproportionation) have been identified in olivine through these less mature regolith samples (Guo et al. 2022;Li et al. 2022;Mo et al. 2022).
Martian moons (Phobos and Deimos) generally lack apparent diagnostic absorption features and are considered the most enigmatic airless bodies (Pieters & Noble 2016).The originating materials of Deimos and the dominant red unit of Phobos are similar (Fraeman et al. 2012).However, whether Phobos and Deimos originated by capture (i.e., D-type asteroid) or formation in situ due to impact events around Mars is difficult to ascertain (Usui et al. 2020).More recent numerical simulations have proposed that Phobos and Deimos may have formed in situ from debris related to giant impacts (Bagheri et al. 2021;Canup & Salmon 2018;Citron et al. 2015;Hyodo et al. 2017).If that is the case, the original building blocks of Phobos and Deimos may involve Fe-rich olivine, as it is globally present in the Martian crust (Koeppen & Hamilton 2008).Subsequent space weathering of Fe-rich olivine might be the key attribute to the substantial dark signatures and the lack of mineralogical absorption signatures of these moons compared to other asteroids.
This study is one of a series of works systematically investigating space weathering effects on Fe-rich olivine.In this study, we conducted nanosecond pulsed laser irradiation experiments using a set of synthetic Fe-rich olivine (Fa 29 , Fa 50 , Fa 71 , and Fa 100 ).The pulsed laser irradiation experiments best simulate the thermal effects induced by micrometeoroid impacts (Hamann et al. 2016;Weber et al. 2020).Three energy levels (5 mJ × 1, 5 mJ × 5, and 5 mJ × 15) were examined to simulate the levels for Phobos, and two energy levels (15 mJ × 5 and 15 mJ × 10) were examined to simulate the levels for the Moon.This work provides essential microscopic characteristics, NIR and Raman spectral data, and quantitative inputs for spectral simulation of the space-weathered Fe-rich olivine.

Synthetic Fe-rich olivine and characterization
The fayalite (Fa 100 ) end-member was synthesized in the furnace using powders of Fe 2 O 3 and SiO 2 in a molar ratio of 1:1.002 at temperatures close to the melting point of fayalite under a controlled oxygen buffer (using a gas mixture of CO 2 and CO).The samples of Fa 71 , Fa 50 , and Fa 29 were then fabricated by mixing the synthetic fayalite with commercial forsterite.The mixtures were ground to a particle size of <25 µm and reacted at ∼85% of the solidus temperature in a one-atmosphere furnace for 20 h under a controlled oxygen buffer (Qi et al. 2021;Zhao et al. 2009).
The synthetic Fe-rich olivine samples were pure and crystalline, as confirmed by powder X-ray diffraction (Fig. Due to the excess of initial SiO 2 in synthesizing the Fa 100 , slight SiO 2 residuals were expected to be present in the synthetic olivine (Fig. A.2).As shown in previous studies, the SiO 2 residuals do not influence the physical, mechanical, or spectral characteristics of the synthetic olivine (see e.g., Qi et al. 2021;Zhao et al. 2009).However, caution was taken when interpreting the SiO 2 -relevant results on the microscopic scale, such as the TEM results.

Settings of laser irradiation simulation experiments
The synthetic olivine samples were ground and sieved into powder with a grain size <75 µm.They were dried at 80 • C for 8 h and then compressed into pellets (13 mm diameter and 2 mm thickness) to serve as the laser irradiation targets.
The choice of laser irradiation energies was based on the micrometeoroid flux and impact velocities at the Phobos and Moon locations and the size of the laser spots (Table 1).The equivalent space weathering timescale by laser irradiation in this study can be roughly estimated by the ratio of the calculated energy density (i.e., the total deposition energy by laser per square meter on the samples) to the deposited energy per square meter per year on the airless bodies (Zhang et al. 2022).The micrometeoroid impact flux was estimated to be the same for Phobos and the Moon at 10 −4 m −2 s −1 (Grün et al. 1991).For Phobos, the micrometeoroid impact velocity ranged from 8.62 Notes.The equivalent space weathering years by our simulation experiments can be roughly estimated by assuming that all kinetic energy from micrometeoroid bombardments on asteroids would convert into heat and that the micrometeoroid flux does not change over time (Zhang et al. 2022).
The equivalent years of space weathering was calculated using the formula C = B / A, where B denotes the energy density (i.e., total deposited energy by laser per square meter in our experiments) and A denotes the deposited energy per square meter per year on airless bodies.The formula to calculate B is B = TE/π r 2 , where TE represents the laser total energy in our experiment (unit in J) and r represents the radius of each laser spot (i.e., 2.5 × 10 −4 m) in our experiments.The formula to calculate A is A = 1/2 mv 2 FT, where m represents the mass of each dust particle (i.e., 10 −15 kg), v represents impact velocity (i.e., 15 km s −1 for the Moon and 8.62-12 km s −1 for Phobos), F represent impact flux (i.e., 10 −4 m −2 s −1 ), and T represents impact time (i.e., 365 × 24 × 3600 s).
to 12 km s −1 (Ballouz et al. 2019;Divine 1993;Miyamoto et al. 2021), so the total energy for the Phobos surface ranged from 2.23 to 178 mJ.For the lunar surface, the micrometeoroid impact velocity was estimated to be 15 km s −1 (Divine 1993); thus, the total energy ranged from 69.6 to 139 mJ.Due to the different micrometeoroid impact velocities on the surfaces of Phobos and the Moon and the corresponding kinetic energies, we assigned one level of single-shot energy to simulate Phobos (5 mJ) and another to simulate the Moon (15 mJ).Considering the size of the laser spots (0.5 mm in diameter), we ultimately set three energy levels for Phobos: 5 mJ × 1(P5mJ), 5 mJ × 5 (P25mJ), and 5 mJ × 15 (P75mJ), and we set two energy levels for the Moon: 15 mJ × 5 (M75mJ) and 15 mJ × 10 (M150mJ).The conditions P75mJ (5 mJ × 15) and M75mJ (15 mJ × 5) were set to compare the effects caused by the single-shot energy level when the total energy level remained the same.The equivalent space weathering timescale by laser irradiation in this study (listed in Table 1) can be roughly estimated by the ratio of the calculated energy density (i.e., the total deposition energy by laser per square meter on the samples) to the deposited energy per square meter per year on the airless bodies (Zhang et al. 2022).
The pulsed laser irradiation simulation experiments were conducted using the ultrahigh vacuum surface analysis system at the Institute of Geochemistry, Chinese Academy of Sciences.A Continuum Surelite I-10 laser was used with a wavelength of 532 nm, and the pulse duration was 6-7 ns, which is comparable to the timescale of micrometeoroid impact (Yamada et al. 1999).The olivine pellets were individually placed in a custom-made sample tank, fixed, and placed in the vacuum chamber.The vacuum condition during the irradiation experiment was 1.8 × 10 −4 to 2.4 × 10 −6 Pa.The laser light path was horizontal, and the target pellets were placed perpendicular to the laser light path.Each set of laser irradiations produced a clear bombardment pit of 0.5 mm in diameter.To enable subsequent sample analysis, the bombardment pits were arranged as a matrix on the pellet surface.The target pellets were stored in the vacuum desiccator before and after the experiments.

Sample analysis
Near-infrared reflectance spectra were collected in the Planetary Spectroscopy and Mineralogy Laboratory at the University of Hong Kong.A Nicolet Continuum Infrared microscope attached to the Thermal Nicolet iS50 spectrometer with an Indium Gallium Arsenide (InGaAs) detector was used to collect NIR reflectance spectra in situ on the pellet samples.Reflectance spectra were collected in a range from 1 µm to 2.5 µm with a spectral resolution of 8 cm −1 using a gold reflectance standard.To obtain high-quality spectra of each sample, a 150 µm aperture size and 128 scan integrations over 4 min were used.
The Raman spectra were collected before and after laser irradiation using a Renishaw inVia laser Raman spectrometer with a 532 nm laser and scanned from 100 cm −1 to 4000 cm −1 .The spectral analysis rate was 1.3 cm −1 , and the spectral repeatability was less than ±0.15 cm −1 .For each target pellet, ten pits were analyzed under a 50× microscope with a laser power of 1 mW.The integration time per spectrum was 30 s.
The surface of the laser-irradiated bombardment pits on each target pellet was sampled by a focused ion beam (FIB) for detailed characterization.The FIB sampling was conducted using FEI Scios Dual-Beam scanning electron microscopy (SEM) at the Institute of Geochemistry, Chinese Academy of Sciences.Before FIB sampling, the sample surface was plated with a platinum coating to prevent damage caused by the Ga + ion.For each pellet, the FIB sampling positions were consistent and confined at the center bottom of the bombardment pits to enable comparison across the different pellets.The FIB-sampled thin foils were characterized using a Talos F200S transmission electron microscope (TEM) at 200 kV at the Analysis and Testing Center of Guangdong University of Technology.The TEM analysis included TEM bright field imaging, dark-field imaging, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) imaging, and energy dispersive spectrometry (EDS) X-ray energy spectroscopy (detection limit 0.1-0.5 wt%).
Four samples (Fa 100 -P75mJ, Fa 100 -M150mJ, Fa 71 -P75mJ, and F 71 -M150mJ) were analyzed by electron microscopeelectron energy loss spectroscopy (EELS) for the distribution of Fe valence.These samples represent Fa 100 and Fa 71 , which experienced the highest total irradiation energies on the surface of Phobos and the Moon, respectively.The TEM-EELS analysis was conducted using a Gatan GIF Quantum ER system Model 965 parallel EELS spectrometer attached to a Hitachi HF5000 TEM at the Shanghai Institute of Ceramics, Chinese Academy of Sciences.The accelerating voltage was 200 kV, and the energy resolution at the zero-loss peak was 0.5-0.7 eV.The full width at half maximum (FWHM) spectra were acquired in DualEELS mode using a probe current of 100 pA.The acquisition time A115, page 3 of 18 A&A 672, A115 (2023) for each point analysis was 4.5 s.Reference standards of Fe 2+ and Fe 3+ for the TEM-EELS analysis were prepared using the synthetic fayalite and hematite extract, respectively, from an ordinary chondrite meteorite (GRV051874).

Characteristics of laser irradiation pits
Before laser irradiation, the olivine powders showed gradual variations from white (Fa 29 ) to dark green (Fa 100 ) with increasing Fe content (Fig. A.3).After irradiation, the laser-irradiated pits were visually darkened, and the color deepened with increasing irradiation energy (Fig. 1).The SEM images showed the melting and cementing of initial olivine particles (Fig. A.4) and produced a glassy surface from which the FIB foils were taken (Fig. A.5).

Structures of laser alteration layers on the olivine rim
The laser irradiation produced multilayer structures at the rims of the olivine grains, and the produced structures varied in response to different olivine types and irradiation energy levels.Pure fayalite (Fa 100 ) produced two altered structures depending on the irradiation energy.Under P5mJ and P25mJ conditions, the olivine surface produced a three-layered structure with a thin vapor deposition layer (top), an amorphous layer (middle), and the olivine host (basal; Figs.2a,c).Under P75mJ to M150mJ conditions, the irradiated rim also consisted of three layers.The vapor deposition layer on top was absent, and a subsolidus layer appeared beneath the amorphous layer (Fig. 2d).The crystallographic orientation of the subsolidus layer was identical to that of the olivine host (Figs.2h and A.6a), suggesting that it was not formed by recrystallization of the amorphous melts during cooling but was instead a remnant of the precursor olivine host (Fazio et al. 2018;Weber et al. 2020).The np-Fe 0 particles were identified throughout the laser-altered layers, having a lattice spacing of 0.203 nm, consistent with the lattice plane (110) of α-Fe (Fig. 2b).
Both Fa 71 and Fa 50 consistently produced three-layered structures (amorphous plus subsolidus plus basal olivine host) under all irradiation conditions (Figs. 2e-j), similar to those observed in Fa 100 under high-energy conditions (P75mJ to M150mJ).Under irradiation conditions from P5mJ to M75mJ, Fa 29 produced a four-layered structure.In this structure, the vapor deposition layer was consistently present on top of the amorphous plus subsolidus plus basal olivine host structure (Fig. 2k).However, under the highest energy condition (M150mJ), a vapor deposition layer with np-Fe 0 particles included was not observed (Fig. 2l).
In addition, irradiation induced splash materials, vesicles, and amorphous silica were also observed in the laser-irradiated olivine.The splash materials were deposited on top of the amorphous layer containing no np-Fe 0 particles (Fig. A.6). Two types of splash materials were identified.One shared the same composition as the underlying amorphous layer (Figs.A.6a-d), and the other was relatively Si-poor (Figs.A.6e-h).In addition, a large number of vesicles with various sizes were identified as being widely distributed in the amorphous layer and close to the np-Fe 0 particles (Fig. 3).Amorphous silica was also observed (Fig. 3a), but many residual SiO 2 originated from synthesizing the Fe-rich olivine samples (Fig. A.1).
Similar to the vapor deposition layer observed in magnesian olivine in lunar or analog samples (see e.g., Sasaki et al. 2001), Fa 29 consistently produced a vapor deposition layer.The The dark spots lined up in a matrix on each pellet are the laser irradiation spots (0.5 mm in diameter), and the light-tone area is the initial state of the olivine.The initial color of the pellets became darker with increasing Fa# values.After irradiation, the color of the irradiation spots became substantially darker.In addition, the color tones became darker as the total energies increased.
vapor layer was absent under the highest irradiation energy condition (M150mJ), which is consistent with a previous report that used Fa 09 and even higher energy (i.e., 15 mJ × 20; 300 mJ) (Weber et al. 2020).In contrast, most Fa-dominant olivine samples (Fa 50 , Fa 71 , and Fa 100 ) produced mainly amorphous and subsolidus layers without a vapor deposition layer.An exception to this occurred in Fa 100 under low-energy conditions (P5mJ and P25mJ), where a very thin veneer (a few nanometers thick) was present.

Thickness of the laser alteration layer
The laser alteration layer refers to the sum of altered layers above the basal olivine host, including the vapor depositional, amorphous, and subsolidus layers.For each type of olivine, the thickness of the laser alteration layer positively correlated with the irradiation energy (Fig. 4).Upon increasing from P5mJ to M150mJ, the laser alteration layer of Fa 29 to Fa 100 increased from ∼200-1700 nm.The two conditions P75mJ (5mJ × 15) and M75mJ (15mJ × 5) further suggested that higher single-shot energy would consistently produce a thicker laser alteration layer across all the olivine (Fig. 4).
A115, page 4 of 18 The relative thickness of laser alteration layers in the different types of olivine changed with irradiation energy levels.Positive correlations between the layer thickness and Fa# values were found only at the P5mJ condition.Under the P25mJ condition, Fa 29 , Fa 50 , and Fa 71 produced alteration layers similar in thickness, and the overall pattern was in the order of Fa 100 Fa 29 ≈ Fa 50 ≈ Fa 71 .At higher energy levels (P75mJ to M150mJ), the relative thickness patterns were in the order of Fa 100 Fa 29 > Fa 71 > Fa 50 .Of all the types of Fe-rich olivine, Fa 100 formed alteration layers substantially thicker than those of the Fo-Fa solid-solution samples, except under the lowest energy condition (Fig. 4).For most energy levels in our experiments, alteration layers of Fa 100 reached the micron level.In contrast, the other Fo-Fa solid-solution samples produced alteration layers only hundreds of nanometers thick.
Notably, the thickness of the alteration layers in Fa 29 , Fa 50 , and Fa 71 did not correlate to their Fa# values, as presumed.In particular, Fa 50 constantly produced a thinner alteration layer than Fo-dominant Fa 29 and Fa-dominant Fa 71 .This discrepancy might be attributable to the degree of nonideal solid-solution present in Fa 50 , as Fa 29 and Fa 71 demonstrate fewer nonideal characteristics when either forsterite or fayalite is predominant in their structures.Although we did not identify an outstanding difference in the initial Fa 50 compared to previous studies of Fo-Fa solid solutions (Kuebler et al. 2006), we infer that the case of nonideal solid solutions might be most pronounced in Fa 50 , and such a difference can be amplified during nonequilibrium processes, such as laser-induced alteration.Potentially similar nonideal conditions were observed in a grossular-pyrope solid solution in which the most significant deviations appeared A115, page 5 of 18  when end-members were mixed at a ratio of one to one (Du et al. 2017).

Characteristics of np-Fe 0 particles
The np-Fe 0 particles showed distinct characteristics in different layers.In the vapor deposition layer, the np-Fe 0 particles were evenly distributed and concentrated in the upper part of the layer if the layer thickness was greater than the size of the np-Fe 0 particles (Fig. A.7).In the amorphous and subsolidus layers, the np-Fe 0 particles were uniformly distributed in the subsolidus layer and tended to merge and migrate into the middle and lower parts of the amorphous layer (Figs.2f and A .8).The np-Fe 0 particles were smaller in the vapor and subsolidus layers than in the amorphous layer.The formation of np-Fe 0 particles was also accompanied by the redistribution of local elements.The Fe concentrated in np-Fe 0 particles and migrated into the amorphous layer, leaving an Mg-enriched signature in the subsolidus layer (Figs.2i,j).
The sizes of the np-Fe 0 particles positively correlated with the Fa# values of olivine and the irradiation energy levels (Fig. 5).In Fa 29 , the median size of np-Fe 0 increased from 4.4 ± 1.1 nm under P5mJ conditions to 10.0 ± 3.5 nm under M150mJ conditions (Figs. 5a-e).Similarly, in Fa 50 and Fa 71 , the median size of the np-Fe 0 particles increased from 12.4 ± 2.2 nm to 17.2 ± 3.6 nm and 14.8 ± 2.9 nm to 17.2 ± 7.5 nm, respectively, with increasing irradiation energy (Figs.5f-j; 5k-o).The largest np-Fe 0 particles were produced by Fa 100 , and the median size increased from 25.0 ± 5.0 nm to 33.8 ± 6.8 nm with increasing energy (Figs.5p-t).Across all the energy levels examined, the np-Fe 0 particles were less than 30 nm in Fa 29 , and they ranged from 6 nm to 40 nm in Fa 50 , 8 nm to 47 nm in Fa 71 , and from 10 nm to 55 nm in Fa 100 .For each type of olivine, the np-Fe 0 size distribution pattern changed from a right-skewed distribution to a normal distribution, with the tendency shifting to a left-skewed distribution with increasing irradiation energy levels (Fig. 5).Such a distribution is consistent with the Ostwald ripening process that describes the nanoparticle growth process (Gommes 2019;Thompson et al. 2017).In addition, higher single-shot energy consistently produced larger np-Fe 0 particles when compared to the P75mJ (5 mJ × 15) and M75mJ (15 mJ × 5) A115, page 6 of 18 conditions (Fig. 5, images c, d, h, i, m, n, r, s).As the np-Fe 0 particles became larger, the shape of the nanoparticle changed from spherical to subangular (Fig. A.9).For comparison, the number of np-Fe 0 particles negatively correlated with the Fa# values, and the amounts of np-Fe 0 particles in different types of olivine were in the order of Fa 29 > Fa 50 ≈ Fa 71 > Fa 100 (Fig. 5).
The size-quantity distribution of the np-Fe 0 particles showed a tendency to increase in size but decrease in quantity with increasing irradiation energy levels and Fa# values.In our experiments, Fa 29 under the lowest irradiation energy would produce the most abundant but smallest np-Fe 0 particles, and Fa 100 under the highest irradiation energy would produce the least abundant A115, page 7 of 18 The TEM-EELS line scan shows that the Fe in olivine and the matrix existed as Fe 2+ , and no evidence of Fe 3+ was found.
but most extensive np-Fe 0 particles (Fig. 5).These trends are consistent with previous laser irradiation experiments in which np-Fe 0 particles produced by San Carlos olivine (Fa 09 ) are mostly a few nanometers in size, which is smaller than those produced in Fa 29 under similar irradiation energy conditions (Loeffler et al. 2016;Weber et al. 2020;Yang et al. 2017).

Formation mechanisms of np-Fe 0 particles in Fe-rich olivine
The proposed formation mechanisms of np-Fe 0 particles relevant to micrometeoroid impacts on olivine include vapor deposition of elemental iron (Sasaki et al. 2001), decomposition (Guo et al. 2022), and disproportionation (Li et al. 2022).In this study, with Fe-rich olivine under various energy conditions, we observed vapor-deposited np-Fe 0 particles originating in the outermost vapor deposition layer and decomposition-originated np-Fe 0 particles in the amorphous and subsolidus layers.The two types of np-Fe 0 particles can be visually distinguished.
The disproportionate origin of np-Fe 0 particles was not observed in our experiment.The TEM-EELS line scan analyzed four laser-irradiated samples, including Fa 71 -P75mJ, Fa 71 -M150mJ, Fa 100 -P75mJ, and Fa 100 -M150mJ, which represent Fa-dominant olivine being space weathered at locations for Phobos and the Moon, respectively.The TEM-EELS analysis also included a linear scan through the np-Fe 0 , amorphous matrix, and crystalline olivine host for each sample.When comparing the TEM-EELS measurement data with the standard reference, we found that Fe in the amorphous matrix and crystalline olivine host existed as Fe 2+ (Figs.6e-h).No Fe 3+ was found in the amorphous matrix surrounding the np-Fe 0 .
The reason for the absence of disproportionation could not be determined in the current study.An essential factor suggested for the Fe disproportionation observed in the CE-5-returned lunar regolith is impact-induced high pressure during micrometeorite A115, page 8 of 18 Xu,J.,et al.: A&A proofs, impacts (Li et al. 2022;Mo et al. 2022) , which is not simulated by our laser irradiation experiments.We inferred that the high pressure or continuous heating required for Fe disproportionation cannot be achieved in our experimental settings.It is unclear whether the temperature needed for Fe disproportionation in fayalite would be lower than that in forsterite.
The presence of vapor-deposited np-Fe 0 and in situdecomposed np-Fe 0 is primarily controlled by the olivine type and irradiation energy conditions, which ultimately relate to the instantaneous temperature of laser irradiation and the melting points of the target olivine.The np-Fe 0 particles formed via in situ decomposition were ubiquitously present in the amorphous and subsolidus layers.In particular, the subsolidus layer represents a transition state between the solid and liquid phases, suggesting that the thermal effects transferred from impact to this layer have diminished but still reach the conditions for np-Fe 0 production.The initially formed particles in the subsolidus layer are small (a few nanometers), and continuous heating allows large nanoparticles to grow at the expense of smaller nanoparticles.With increasing distance from the olivine host, np-Fe 0 particles grow gradually in size, suggesting slower cooling as an effect of thermal diffusivity (Weber et al. 2020).Under the same irradiation energy, the lower melting point of fayalite (1250 • C) compared to forsterite (1890 • C) would allow a longer time for np-Fe 0 to grow, and the higher Fe content in fayalitic olivine would provide sufficient Fe for the formation of np-Fe 0 particles.Therefore, with the increasing Fa# values of the initial olivine, the growing supply of Fe and the low melting point of olivine would facilitate the formation of large np-Fe 0 via olivine decomposition.It thus can be extrapolated that other Fe-rich minerals with low melting points may be able to produce large np-Fe 0 as the Fe-rich olivine did during the space weathering processes.Previous laser irradiation experiments have consistently demonstrated that targets such as ilmenite (melting point 1050 • C) and kamacite (melting point 800-900 • C) produce large np-Fe 0 , similar to the Fe-rich olivine observed in this study (Wu et al. 2017).
We found only a few cases of vapor-deposited np-Fe 0 in our experiments, such as in Fa 29 under almost all energy conditions, except for the highest energy level (M150mJ), and in Fa 100 under the lowest irradiation energy conditions (P5mJ and P25mJ).Vapor-deposited np-Fe 0 can be largely absent in Fadominant olivine regardless of the energy levels and absent in Fo-dominant olivine under high irradiation energy.The formation of vapor-deposited np-Fe 0 may be primarily controlled by the instantaneous temperature during irradiation and may also be related to the melting point of olivine and the vaporization temperature of Fe.For Fo-dominant olivine, when the melting point is relatively higher than that of Fa-dominant olivine, the np-Fe 0 particles can condense from the vaporized materials under most energy conditions, except for the highest energy condition.For the fayalite, which has the lowest melting point of all the Fo-Fa solid solutions, the laser irradiation should produce substantially low instantaneous T under P5mJ conditions versus those under P75mJ conditions.High instantaneous T might account for the absence of vapor-deposited np-Fe 0 particles that Fe can vaporize and escape into the vacuum from the surface in our experiments.The lack of vapor-deposited np-Fe 0 and a low vapor-melt production ratio in Fa 09 olivine under 15 mJ × 20 conditions have been inferred to potentially be due to the instantaneous T of ∼7000 K (Weber et al. 2020).
Integrating the microscopic features of the laser alteration layer and np-Fe 0 particles, we suggest that Fo-dominant (Fa# < 50) and Fa-dominant (Fa# 50) olivine may demonstrate distinct characteristics during thermal alteration induced by micrometeoroid impacts.Under the same impact flux and time, the Fa-dominant olivine would produce thicker altered rims (micron level) and grow larger np-Fe 0 particles (e.g., a substantial fraction of sizes >10 nm or even >40 nm) compared to the Fo-dominant olivine.

Near-infrared reflectance spectra of altered olivine
With increasing Fa# values of olivine, the NIR reflectance spectra of initial olivine decrease in albedo, and the 1000 nm absorption feature (R 1000 ) shifts to longer wavelengths with increasing Fa# values (Fig. A.10), consistent with previous observations (Dyar et al. 2009;Pinet et al. 2022).After laser irradiation, the NIR spectra of olivine samples (Fa 29 , Fa 50 , Fa 71 , and Fa 100 ) under different conditions were compared (see Fig. 7).We used the absolute reflectance at 1900 nm (R 1900 ) to characterize the darkening effects in the NIR band, the ratio of reflectance at 2400 nm to 1900 nm (R 2400 /R 1900 ) to characterize the reddening effect, and the band depth at 1000 nm (BD 1000 ) to describe the decrease in absorption (Table 2).The changes in spectral parameters relative to the initial olivine under different irradiation conditions are also listed in Table 2 to quantitatively compare the extent of spectral alteration effects in these different parameters.
For Fa 29 , Fa 50 and Fa 71 , the NIR spectra of altered olivine exhibited a progressive spectral change with increasing irradiation energy under conditions simulating Phobos and the Moon, respectively (Figs. 7a,b,c,e,f,g).The irradiated olivine samples demonstrate a decrease in absolute reflectance (darkening), an increase in R 2400 /R 1900 slope (reddening), and attenuation in BD 1000 (flattening).We note that Fa 50 under M75mJ and M150mJ conditions showed similar reddening levels and thus no further reddening at higher energies.Comparing Fa 29 to Fa 71 olivine under P75mJ versus M75mJ conditions, the absolute reflectance R 1900 and BD 1000 under M75mJ are always higher than those under P75mJ, but the slopes of R 2400 /R 1900 are slightly lower under M75mJ than under P75mJ (Table 2).This indicates that under the same total irradiation energy, for Fa 29 , Fa 50 , and Fa 71 , the spectra produced via high single-shot energy but with fewer shot times (15 mJ × 5) are relatively brighter but with a lower extent of reddening and feature attenuation compared to their counterparts under low single-shot energy but with more shot times (5 mJ × 15).Thus, we can conclude that both singleshot energy and total irradiation energy have essential effects on spectral modification, and caution needs to be taken when comparing spectral features across different single-shot irradiation conditions.In addition, the influence of Fa# values alone on the spectral parameters (i.e., darkening, reddening, feature attenuation) is relatively minor, and there is generally a lack of increasing or decreasing trends corresponding with increasing Fa# values (Fig. 8).
Pure fayalite (Fa 100 ) largely validates observations in Fa 29 to Fa 71 samples but demonstrates interesting exceptions (Figs.7d,h; Table 2).From P5mJ to P75mJ, the reflectance spectra of Fa 100 demonstrate progressive darkening and reddening and feature attenuation with increasing irradiation energy.While the band depth (BD 1000 ) continuously weakened under M75mJ and M150mJ conditions (reduced by 80.5% and 85.1%, respectively; Table 2), following the trend under Phobos conditions, the spectral parameters started to reverse from darkening and reddening to brightening and reddening (Figs.7d,h; Table 2).Under M75mJ, the darkening and reddening effects were significantly weakened compared to the P75mJ condition.Under M150mJ, the absolute reflectance of R 1900 was equivalent to the initial A115, page 9 of 18 olivine, and the R 2400 /R 1900 slope further decreased compared to that of the M75mJ condition, resulting in brightening and reddening under the most irradiated condition.
Almost all irradiated olivine samples demonstrated darkening and reddening spectral effects, of which darkening was the primary effect relative to reddening in the NIR band (Table 2).The only sample that demonstrated comparable darkening (10.5%) and reddening (8.1%) effects is Fa 29 -P5mJ.In this particular sample, the np-Fe 0 particles a median size of 4.4 nm ± 1.1 nm, without grains larger than 10 nm (Fig. 5a).
In contrast, all the other samples contain substantial fractions of np-Fe 0 in the range of 10-40 nm or even larger (40-60 nm).We find the observations of np-Fe 0 fractions in the samples and the corresponding spectral alteration are consistent with those inferred by previous studies in the visible and NIR band, namely, the smallest size fraction (np-Fe 0 < 10 nm) controls reddening, the largest size fraction (np-Fe 0 > 40 nm) mainly results in darkening without reddening, and the intermediate size fraction (10-40 nm) has unique control over darkening or reddening based on the increasing iron concentration (see e.g., Cloutis et al. 2015;Kohout et al. 2014;Loeffler et al. 2016;Noble et al. 2007).Furthermore, increasing the volume fraction of the np-Fe 0 can also contribute to the darkening and reddening of the spectra and the weakening of the absorption bands.We suggest that in our Fe-rich olivine samples, a proportional increase in intermediate and large size fractions and an increase in the volume fraction of np-Fe 0 are likely to present and substantially contribute to the darkening of the spectra over the mostly reddening effect found in magnesian olivine (see e.g., Yamada et al. 1999;Hapke 2001).In addition, saturation of the space weathering effects is suggested for np-Fe 0 volume fractions 1%, as the spectrum flattens with very low reflectance (Escobar-Cerezo et al. 2018).In comparison, our maximum feature attenuation reached 85.1% under the Fa 100 -M150mJ condition and did not reach the saturation states of space weathering effects.
The brightening spectral feature produced by spaceweathered olivine has not been reported previously in any simulation experiments via pulsed laser bombardment, ion implantation, or simply heating.In particular, the reversal from darkening and reddening (under P25mJ to P75mJ) to brightening and reddening (under M75mJ and M15mJ) in Fa 100 in our experiments suggests that the irradiation energy (total energy and/or singleshot energy) and the Fe content in the initial olivine (fayalitic or Fe-dominant olivine) should be the key elements accounting for such unique spectral change.Because our synthetic olivine samples are relatively pure and the np-Fe 0 particles and laser alteration layers are the two primary space-weathered features, we infer that the substantial increase in thickness of the laser-altered layers on the fayalitic olivine may be critical, as the np-Fe 0 particles are known to produce darkening and reddening primarily in the NIR band.In Fa 29 , Fa 50 , and Fa 71 , the thickness of altered layers is always less than 1 µm (maximum 916.3 nm).In comparison, the altered layers of Fa 100 are mainly thicker than 1 µm, and they increase from 340.3 nm (P5mJ) to 1044.2 nm (P25mJ), 1429.3 nm (M75mJ), and 1697.0 nm (M150mJ).Although previous studies concluded that amorphization of mineral surfaces (in a few hundred nanometers) would lead to an overall darkening of the reflection spectra in the UV-Vis-NIR range even without considering np-Fe 0 (Zhang et al. 2022), as demonstrated in our experiments, when the altered glassy and amorphous rims became thick enough (>1 µm or >1.4 µm) and approached the wavelength of the infrared spectrum, they would substantially contribute to the spectral effects in the NIR band.The thickening alteration layers are composed of amorphous and subsolidus layers, and the np-Fe 0 particles increasing in size and volume reside within these layers.We infer that the substantially thick altered rims might effectively compensate for the darkening effects caused by the increasing size and volume of the np-Fe 0 particles, leaving reddening as the primary spectral effect in the sample.The thick alteration layers may also interfere by reducing the extent of the reddening effect.Depending on the relative contribution of the alteration layer thickness and np-Fe 0 particles, one can anticipate the spectral alteration from darkening and reddening to brightening and reddening in the Fe-rich olivine samples.Nevertheless, whether the spectra turn dark or bright, A115, page 10 of 18 the band depth of absorption gradually weakens with increasing irradiation energy, contributing to flattening in the spectral features.
We note that the single-shot and total energy conditions selected to simulate those of Phobos and the Moon in our experiments are conservative and considered to simulate low to medium degrees of space weathering (as defined in Zhang et al. 2022), and higher energy conditions are expected to occur on both Phobos and the Moon.For instance, considering a higher impact velocity (e.g., 12 km s −1 ; Table 1) and longer irradiation history (e.g., 4 Ga), a single-shot energy at 7 mJ and total energy at 178 mJ can be adopted for Phobos, which is higher than the present M150mJ conditions.Therefore, further investigation is required to determine whether the brightening and reduction in reddening found in Fa 100 under M75mJ and M150mJ can be extrapolated for at least fayalitic olivine under medium or high degrees of space weathering or if such features occur in other Fe-dominant olivine (Fa# > 50).

Raman spectral characteristics of altered olivine
The Raman spectra of each olivine type before and after laser irradiation are compared in Fig. 9.The initial olivine samples showed typical double peaks between 800 and 900 cm −1 .The positions of the double peaks shifted toward lower wavenumbers with increasing Fa# values, that is, from Fa 29 to Fa 100 .The wavenumbers of doublet peak #1 (DB1) decreased from 821 to 815 cm −1 , and those of doublet peak #2 (DB2) decreased from 851 to 833 cm −1 .In addition, the intensity of DB2 gradually decreased and appeared as a shoulder beside the DB1 peak with increasing Fa# values, consistent with previous studies (Kuebler et al. 2006).After laser irradiation, the wavenumber, shape, and FWHM of the doublet peaks did not change substantially (Table A.3).The newly formed np-Fe 0 particles are pure metallic iron that have no Raman effects, and the potential reduction of the fluorescence-dominated background due to the formation of np-Fe 0 may facilitate a better interpretation of the individual mineral peaks by the Raman spectrometer (Weber et al. 2021(Weber et al. , 2022)).Amorphization of the olivine surface may cause local disorder and result in the broadening of Raman peaks, as suggested in previous studies (Rouquette et al. 2008;Santamaria-Perez et al. 2016); however, the amorphous layer showed no substantial influence in our experiments.The spectral intensities were significantly reduced by laser irradiation, and the extent of reduction showed no difference under different energy conditions (i.e., P75mJ vs. M150mJ; Fig. 9).Therefore, a lack of changes in the characteristic DB1 and DB2 after laser irradiation may allow for the identification and extraction of olivine Fo-Fa compositions, as was done for fresh olivine before space weathering (Kuebler et al. 2006).
In this experiment, we measured Raman spectra before and after laser irradiation only at room temperature.Previous studies suggested that below 150 K, forsterite may show negligible temperature-related shifts but that fayalite would show temperature-dependent changes in the lattice modes and modification of Raman spectra at low Stokes frequencies, potentially due to the high amount of iron in the mineral structures (Weber et al. 2014).Since Raman spectroscopy has been proposed for in situ investigations of airless celestial bodies, such as Phobos and the Moon (Usui et al. 2020), how to accurately constrain the Fo-Fa compositions under low temperature, particularly for space-weathered Fe-rich olivine, requires further study.

Implications
The simulation experiments in this study mainly account for the thermal effects of micrometeoroid impacts on olivine; thus, cautions should be taken when comparing our findings with space weathering phenomena in natural samples that have experienced multiple processes and have complex weathering histories.In addition, in natural space-weathered materials, solar wind irradiation can cause faster, significant spectral change at a timescale of 10 4 -10 6 yr, compared to 10 8 -10 9 yr for micrometeoroid bombardment (Noguchi et al. 2011;Sasaki et al. 2001).Nevertheless, our experiments suggest that when subjected to micrometeoroid impacts on the lunar surface, fayalitic olivine responds faster to space weathering than magnesian olivine.For Fo-dominant olivine (e.g., Fa 29 ), np-Fe 0 formed via vapor deposition and in situ decomposition are both present, primarily in the size of <10 nm.Larger np-Fe 0 particles (10-30 nm) may also form in the amorphous layer during prolonged space weathering.For Fa-dominant olivine (Fa# 50), np-Fe 0 particles mainly formed via olivine decomposition and produced large np-Fe 0 particles (generally >10 nm; substantial fraction >40 nm).The spectra in red and blue represent irradiation conditions of P75mJ and M150mJ, respectively.The initial olivine samples showed typical double peaks between 800 and 900 cm −1 , and the positions of the double peaks shifted toward lower wavenumbers with increasing Fa# values.From Fa 29 to Fa 100 , the wavenumbers of doublet peak #1 (DB1) decreased from 821 to 815 cm −1 , and those of doublet peak #2 (DB2) decreased from 851 to 833 cm −1 .The intensity of DB2 gradually decreased and appeared as a shoulder beside the DB1 peak with increasing Fa# value in initial olivine.After laser irradiation, the wavenumber, shape, and FWHM of the doublet peaks did not change substantially.The spectral intensities were greatly reduced by laser irradiation, and the extent of reduction showed no difference under different energy conditions.
Similar to what was observed with Apollo samples, Mg-rich olivine (e.g., Fa 10 ) mainly produced vapor-deposited np-Fe 0 (a few nanometers in size) during micrometeoroid impacts (Noble et al. 2005;Pieters & Noble 2016).Therefore, other processes would be required to produce larger np-Fe 0 particles (tens of nanometers in size), such as impacts and gardening (e.g., in forming agglutinitic glass), solar wind ion irradiation, and friction between particles.In contrast, in the Fa-dominant or fayalitic olivine in the lunar regolith, a thick altered rim and large np-Fe 0 particles 40-50 nm formed directly via olivine decomposition can be expected.Since fayalitic olivine is more susceptible to thermal alteration than Mg-rich olivine, processes other than impact-induced thermal effects may also produce large np-Fe 0 particles, such as heating induced via friction.This supports that the np-Fe 0 particles on fayalitic olivine (Fa# 85) in the CE-5 lunar samples are not generated by micrometeoroid impacts; otherwise, a much thicker alteration rim would be expected to be present.The micrometeoroid impact effects on younger CE-5 lunar regolith during exposure (less than 100 Ma) should be relatively low, and the grains did not experience extensive heating events after exposure to the lunar surface (Guo et al. 2022).
For Phobos, the striking blue-red contrast on the surface is due to compositional variations or spatial variations in its physical surface properties (Ballouz et al. 2019;Rosenblatt 2011).A 0.65 µm absorption peak is present in the reddening region A115, page 12 of 18 Xu,J.,et al.: A&A proofs, of Phobos but absent in the blue regions, indicating a potential feature yielded via space weathering or possibly the presence of highly desiccated Fe-phyllosilicate minerals native to Phobos (Fraeman et al. 2014).In addition, metal-OH absorption peaks at 2.8 µm are present, originating through solar wind interactions with the weathering layer of the surface of Phobos (Fraeman et al. 2014).Regarding the origin hypotheses of Phobos, substantial olivine is expected to be present if Phobos had been captured from either organic-poor and water-poor outer Solar System bodies or inner Solar System bodies.Fe-rich olivine is still expected to be present if Phobos had been co-accreted with Mars or if it originated directly from Mars via giant impacts (Cho et al. 2021).Our experiments suggest that if the olivine on Phobos is Fe-rich or even Fa-dominant, which would be similar to the olivine widely present on the Martian surface (Koeppen & Hamilton 2008), space weathering by micrometeoroid impacts under various energy levels could produce dark and red units and bright and red units.If Phobos has been subjected to prolonged space weathering, such as ∼4 Ga if Phobos originated via a giant impact that formed the Borealis basin (Citron et al. 2015), a larger fraction of >40 nm np-Fe 0 particles and thick alteration rims would be expected to be produced.While large np-Fe 0 particles (10-60 nm) alone may result in primary darkening with minor reddening effects on the Phobos surface, substantially thick alteration rims induced via medium to high degrees of space weathering may inversely brighten the weathered surfaces.If considering only albedo and spectral slope change in the NIR band, a fayalite-rich region that has been subjected to intensive space weathering might be mistakenly treated as a relatively less weathered region because it appears brighter and less red in the former case.
We note that the Phobos surface is not uniformly weathered, the solar wind irradiation flux is 15-100 times higher, and the meteoroid impact flux is four times higher at the front surface compared to those on the far side (Christou et al. 2021;Nénon et al. 2021).Therefore, intensive space weathering features, including the formation of large-sized np-Fe 0 particles and thick alteration rims, are expected on the front side, contributing to substantially altering the NIR spectra of the Phobos surface.In addition, compared to the Moon, the flux and speed of the micrometeoroid impacts and solar wind ions decline substantially at locations of Phobos and Deimos or further, making the compositions of olivine a more critical factor controlling the optical properties of surface materials than the external space weathering factors.
Finally, several macroscopic characteristics of the laserirradiated Fe-rich olivine require further study, including the mid-infrared and magnetic characteristics and the quantitative inversion of the optical properties of laser-irradiated olivine samples.Further studies are needed to examine how the spaceweathered Fe-rich olivine may further contribute to the spectral slope behavior of olivine-pyroxene mixtures (e.g., for Q-and Stype asteroids) on different timescales and how the solar wind and micrometeoroid bombardments and the superposition of the two effects would influence the spectra of Fe-rich olivine.

Conclusions
In this study, we conducted nanosecond pulsed laser irradiation experiments using a set of synthetic Fe-rich olivine samples (Fa 29 , Fa 50 , Fa 71 , and Fa 100 ).Three energy levels (5 mJ × 1, 5 mJ × 5, and 5 mJ × 15) were examined to simulate the energy level of Phobos, and two energy levels (15 mJ × 5 and 15 mJ × 10) were examined to simulate the energy level of the Moon.The microscopic features and Raman spectra of space-weathered Fe-rich olivine were characterized in detail.The findings are summarized as follows: 1.The structure of laser alteration layers is primarily controlled by the irradiation energy and Fe content of olivine.The thickness of the layers roughly increases with irradiation energy.The laser alteration layer of Fa 100 reached a micronlevel thickness substantially thicker than those of Fo-Fa solid solution olivine; 2. The np-Fe 0 particles formed via in situ decomposition are ubiquitously present in all the Fe-rich olivine, and the np-Fe 0 particles formed by vapor deposition are primarily present in Fa 29 and rarely found in Fa-dominant olivine (Fa# 50).In the results from the laser irradiation experiment, we did not observe the disproportionate origin of np-Fe 0 particles; 3. The size-quantity distribution of np-Fe 0 particles showed a tendency to increase in size and decrease in quantity with increasing irradiation energy levels and Fa# values.
As the Fa# values increase from Fa 10 to Fe-rich (Fa 29 ) or Fa-dominant (Fa# 50), intermediate (10-40 nm) and large (>40 nm) size fractions of np-Fe 0 proportionally increase, which may substantially contribute to darkening of the spectra over reddening, as found in magnesian olivine; 4. The NIR spectra of the space-weathered olivine turned dark and red in most cases, with primary darkening over minor reddening effects occurring due to the formation of largesized and increasing-volume fractions of np-Fe 0 particles.Brightening and reduction in reddening of the NIR band were also found in Fa 100 with increasing irradiation energies, potentially due to the spectral effects of micron-level thickness alteration layers superimposed on the large np-Fe 0 particles.In either case, the band depth of absorption gradually weakens with increasing irradiation energy, contributing to flattening of the spectral features; 5.The Raman spectra show a reduction in intensity but no peak shifts in the weathered olivine, allowing for identification and quantification of Fo-Fa compositions on an airless body; 6.The distance of Phobos and Deimos from the Sun makes the Fa# value of olivine a more critical factor in controlling the SWSAEs.If the Phobos is Fe-rich or even Fa-dominant, thick alteration rims and large np-Fe 0 particles (10-60 nm) formed via space weathering may cause darkening-reddening and potentially brightening-reddening spectral effects on the surface of Phobos and Deimos.
Our study demonstrates that the iron content of olivine is a key factor influencing the SWSAEs.Understanding how Ferich olivine or Fa-dominant olivine responds to a variety of space weathering processes can aid data interpretation when exploring airless bodies potentially bearing Fe-rich olivine, such as the Moon, Phobos and Deimos, and asteroids.As stated in the previous section, several macroscopic characteristics of the laser-irradiated Fe-rich olivine samples require further study, including the mid-infrared and magnetic characteristics and the quantitative inversion of the optical properties.Further studies are needed to examine how space-weathered Fe-rich olivine may further contribute to the spectral slope behavior of olivinepyroxene mixtures (e.g., for Q-and S-type asteroids) on different timescales and how solar wind and micrometeoroid bombardments and the superposition of the two effects influence the spectra of Fe-rich olivine.
A115, page 1 of 18 Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0),which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.This article is published in open access under the Subscribe to Open model.Subscribe to A&A to support open access publication.
A.1).The X-ray diffraction patterns were collected through a Panalytical Empyrean diffractometer at the State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences.All patterns were collected using Cu-Kα radiation (λ = 1.5419Å) at 45 kV and 40 mA in the range of 10 • -60 • 2θ, a step size of 0.0263 • , and a counting time of 36.465s per step.The chemical compositions of Ferich olivine samples were determined by an inductively coupled plasma emission spectrometer (ICP735-ES) after acid digestion.The synthetic olivine samples comprised FeO, MgO, and SiO 2 , with minor Al 2 O 3 , CaO, MnO, and Ni contents (Table A.1).

Fig. 1 .
Fig. 1.Target pellets of four types of olivine after laser irradiations.Each pellet is 13 mm in diameter and 2 mm in thickness.The four olivine types are Fa 29 (a-e), Fa 50 (f-j), Fa 71 (k-o), and Fa 100 (p-t).The five irradiation energies used were P5mJ (first row), P25mJ (second row), P75mJ (third row), M75mJ (fourth row), and M150mJ (fifth row).The dark spots lined up in a matrix on each pellet are the laser irradiation spots (0.5 mm in diameter), and the light-tone area is the initial state of the olivine.The initial color of the pellets became darker with increasing Fa# values.After irradiation, the color of the irradiation spots became substantially darker.In addition, the color tones became darker as the total energies increased.

Fig. 2 .
Fig. 2. TEM images of different layered structures.Image (a) Fa 100 -P5mJ with a two-layer structure: amorphous layer plus olivine host.Image (b)High-resolution TEM image of np-Fe 0 distributed in the middle and bottom of the amorphous layer in image (a).The interplanar spacing is d = 0.203 nm, which is consistent with a crystal lattice plane (110) of α-Fe.Image (c) Fa 100 -P25mJ with a three-layer structure: vapor deposition layer plus amorphous layer plus olivine host.Image (d) Fa 100 -M150mJ with a three-layer structure: amorphous layer plus subsolidus layer plus olivine host that differs from that of image (c).Image (e) Fa 71 -P5mJ, (f) Fa 71 -P75mJ, (g) Fa 71 -M75mJ, (h) Fa 50 -P75mJ, and (i) Fa 50 -M75mJ in which the three-layer structure represents most of the samples: amorphous layer plus subsolidus layer plus olivine host.Image (j) EDS mapping of (i) showing the elemental distribution of Mg and Fe.Image (k) represents the characteristics of Fa 29 in the energy range from P5mJ to M75mJ and with a four-layer structure: vapor deposition layer plus amorphous layer plus subsolidus layer plus olivine host.Image (l) shows Fa 29 -M150mJ with a three-layer structure: amorphous layer plus subsolidus layer plus olivine host.This figure shows the variation in the layer structure produced by olivine with different iron contents at different energy levels.
Fig. 3. Vesicles and amorphous silica by laser irradiated in olivine.Image (a) TEM image of Fa 71 -M150mJ containing np-Fe 0 , amorphous silica, and vesicle.Image (b) Fe, (c) Si, and (d) Mg as EDS mapping images in image (a).Image (e) Fa 29 -P75mJ, (f) Fa 50 -P5mJ, (g) Fa 71 -M150mJ, and (h) Fa 100 -M150mJ showing features containing np-Fe 0 and vesicles.Vesicles support the in situ decomposition reaction of olivine.The EDS mapping shows that the area framed by the white dashed curves in Fig.3ais amorphous silica, which may be residual SiO 2 that originates from synthesizing the Fe-rich olivine samples.

Fig. 4 .
Fig. 4. Comparison of the thickness of the laser alteration layer formed in different types of olivine and irradiation energies.Four types of olivine, Fa 29 (yellow), Fa 50 (in purple), and Fa 71 (in green), Fa 100 (in blue), and five total energy levels (P5mJ, P25mJ, P75mJ, M75mJ, and M150mJ) were compared.The bars represented median values of the laser alteration layer thicknesses.The thicknesses of the laser alteration layer of Fa 29 , Fa 50 , and Fa 71 are not significantly different at the same energy, and the thickness of the laser alteration layer produced by Fa 100 is significantly thicker than the other groups with increasing energy.At different energies, the thickness of the laser alteration layer was positively correlated with the Fe content.

Fig. 5 .
Fig. 5. Particle size distribution of np-Fe 0 in different types of olivine and under different irradiation energy levels.The particle size labeled in each subfigure was the median particle size.Images (a)-(e) are Fa 29 , (f)-(j) are Fa 50 , (k)-(o) are Fa 71 , and (p)-(u) are Fa 100 .The numbers of np-Fe 0 particles are negatively correlated with and the np-Fe 0 particle sizes are positively correlated with increasing Fe content and irradiation energy.
Fig. 6.TEM-EELS line scan analysis and the corresponding spectrum results of Fa 71 -P75mJ, Fa 71 -M150mJ, Fa 100 -P75mJ, and Fa 100 -M150mJ samples.Twenty points were sampled on Fa 71 -P75mJ and Fa 71 -M150mJ samples, and fifteen points were sampled for Fa 100 -P75mJ and Fa 100 -M150mJ samples.Analysis points of the samples are shown in the top row, and the corresponding Fe L 3,2 TEM-EELS data for samples and standard references are shown in the bottom row.Images (a) and (e) Fa 71 -P75mJ; (b) and (f) Fa 71 -M150mJ; (c) and (g) Fa 100 -P75mJ; (d) and (h) Fa 100 -M150mJ.The standard references show that Fe 2+ and Fe 3+ have distinct edge shapes and peak positions at 706.2 eV and 708 eV, respectively.The TEM-EELS line scan shows that the Fe in olivine and the matrix existed as Fe 2+ , and no evidence of Fe 3+ was found.

Fig. 7 .
Fig. 7. NIR reflectance spectra of the Fe-rich olivine before and after laser irradiation in this study.The left column presents the NIR spectra of four types of Fe-rich olivine, and the right column presents the spectra but normalized at 1500 nm for comparison.Images (a) and (e) Fa 29 ; (b) and (f) Fa 50 ; (c) and (g) Fa 71 ; and (d) and (h) Fa 100 .In each subfigure, the black solid line represents the spectra of the initial olivine; colored solid lines represent irradiated olivine simulating Phobos conditions, P5mJ (teal), P25mJ (red), and P75mJ (brown); colored dashed lines represent irradiated olivine simulating lunar conditions, M75mJ (blue) and M150mJ (purple).In general, most irradiated olivine samples demonstrate a decrease in reflectance (i.e., darkening) compared to the initial olivine and an increase in the R 2400 /R 1900 spectral slope (i.e., reddening).Exceptions occur in Fa 50 and Fa 100 under M75mJ and M150mJ conditions.Fa 50 shows no further reddening under the M150mJ condition compared to that under the M75mJ condition.Fa 100 shows an inversed pattern, with an increase in reflectance (i.e., brightening) and a lower extent of reddening under M75mJ and M150mJ conditions compared to those under P5mJ to P75mJ conditions.Attenuation of absorption occurred in all the irradiated samples.

Fig. 8 .
Fig. 8. Variations in NIR spectral parameters with increasing Fa# values of olivine under different irradiation energy conditions.Images (a) and (b) R 1900 ; (c) and (d) R 2400 /R 1900 ; (e) and (f) BD 1000 .The initial olivine is shown in black.Images (a), (c), and (e) show irradiations simulating Phobos conditions, P5mJ (teal line), P25mJ (red line), and P75mJ (brown line).Images (b), (d), and (f) show irradiations simulating Moon conditions, M75mJ (blue dashed line) and M150mJ (purple dashed line).Under Phobos conditions, the decrease in reflectance and increase in R 2400 /R 1900 spectral slope are approximately parallel to the initial olivine with increasing Fa# values.Variations in band depth occured in different Fa# but overall the band depth attenuated after irradiations.Under the Moon conditions, the changes in reflectance, R 2400 /R 1900 spectral slope change, and band depth generally followed the trends of initial olivine, with exceptions in Fa 100 -M75mJ and -M150mJ.

Fig. 9 .
Fig. 9. Raman spectra before and after laser irradiations of the Ferich olivine.Image (a) Fa 29 , (b) Fa 50 , (c) Fa 71 , and (d) Fa 100 .Spectra in black are the Raman spectra of initial olivine before laser irradiation.The spectra in red and blue represent irradiation conditions of P75mJ and M150mJ, respectively.The initial olivine samples showed typical double peaks between 800 and 900 cm −1 , and the positions of the double peaks shifted toward lower wavenumbers with increasing Fa# values.From Fa 29 to Fa 100 , the wavenumbers of doublet peak #1 (DB1) decreased from 821 to 815 cm −1 , and those of doublet peak #2 (DB2) decreased from 851 to 833 cm −1 .The intensity of DB2 gradually decreased and appeared as a shoulder beside the DB1 peak with increasing Fa# value in initial olivine.After laser irradiation, the wavenumber, shape, and FWHM of the doublet peaks did not change substantially.The spectral intensities were greatly reduced by laser irradiation, and the extent of reduction showed no difference under different energy conditions.

Fig
Fig. A.5. Secondary electron images showing the locations from which the FIB slices were extracted from the laser-irradiated pits on the olivine targets.The locations are shown in green boxes.Each column represents olivine with the same Fa# values.From left to right, Fa 29 , Fa 50 , Fa 71 , and Fa100 100 are shown.Each row represents the same irradiation energy levels.From top to bottom, P5mJ, P25mJ, P75mJ, M75mJ, and M150mJ energy levels are shown.

Fig
Fig. A.6.TEM analysis of the impact-induced splash materials deposited on the olivine surface showing two distinctive compositions.Image (a) shows bright-field TEM mapping of Fa 50 -P75mJ with the splash materials and np-Fe 0 particles.Images (b)-(d) show the EDS mapping of Fe, Mg, and Si of the area highlighted by the white box in image (a).The splash material shares the same composition as the subsurface materials.Image (e) shows a high-angle annular dark-field (HAADF) TEM image of Fa 100 -M75mJ with the splash and np-Fe 0 particles.Images (f)-(h) show the EDS mapping of Fe, O, and Si of the area in image (e).The splash material is Si-poor compared to the subsurface materials.

Fig
Fig. A.7. TEM images showing the np-Fe 0 particles within the vapor deposition layer.Image (a) Fa 100 -P25mJ experiment, (b) Fa 29 -P5mJ experiment, (c) Fa 29 -P25mJ experiment, (d) Fa 29 -P75mJ experiment, and (e) Fa 29 -M75mJ experiment.The np-Fe 0 particles formed via vapor deposition were distinct in distribution and grain size compared to those in the middle and bottom of the amorphous layer.The former were evenly small and present close to the upper boundary of the vapor deposition layer, indicating different formation pathways from those in the amorphous and subsolidus layers.

Fig
Fig. A.8. High-resolution TEM images showing the formation of np-Fe 0 particles in the subsolidus layer and migration into the amorphous layer.Image (a) Fa 29 -P5mJ, (b) Fa 50 -P75mJ, (c) Fa 71 -P5mJ, and (d) Fa 100 -M150mJ.The lattice fringes of olivine and newly formed np-Fe 0 particles can be observed.

Fig
Fig. A.10. NIR reflectance spectra of the initial olivine samples Fa 29 , Fa 50 , Fa 71 , and Fa 100 without laser irradiation.Black line: Fa 29 ; blue: Fa 50 ; red: Fa 71 ; brown: Fa 100 .With increasing Fa# values, the reflectance spectra of the initial olivine decreased in albedo, and the 1 µm absorption feature shifted to longer wavelengths.