Raman Spectroscopy as a Tool to Measure Silanol as Evidence of Water–Rock Interactions for Astrobiological Exploration

2.1. Definitions and terminology

To mitigate potential confusion over terminology, the nomenclature for phases and structures within this document is as follows. The term “α-quartz,” refers to the low-temperature crystalline form of silica, whereas the terms “opal-A,” “opal-CT,” and “opal-C” refer to X-ray amorphous-to-disordered microcrystalline varieties of hydrated silica. Opaline silica is a general term used when rock material displays an opaline appearance yet the types and relative amounts of opaline silica varieties present in the material are not known. “Silanol” is a distinct atomic structure that denotes nonbridged Si-OH bonds that are presumed to have formed on former growth surfaces of the sample materials studied.

2.2. Samples

Raman spectra were obtained for semiprecious stones from the University of Aberdeen Geology Collection (No. 1028) and the author’s own sample collection. Samples include macrocrystalline, microcrystalline, and noncrystalline silica minerals (Table 1, Supplementary Data^†). The samples include a range of silica varieties as semiprecious stones that include agate, chalcedony, flint, jasper, opal, quartz (micro- and macrocrystalline quartz) geyserite (precipitate from hydrothermal environments), and a piece of a stromatolite. All of these samples are known to have formed in the presence of water or they were previously determined to contain intracrystalline water. To keep data and observations easily accessible to workers from a variety of fields, sample classification followed a common reference manual for semiprecious stones (Hansen, 2022), with rocks being assigned mineral or gemstone names according to visual attributes, a common practice in classifying semiprecious stones. The samples include cut, polished, but also unaltered rocks, and it is important to note that, except for one sample, spectra were acquired without further modification or cleaning of the sample surfaces.

The one sample that had its surface altered was sample P4 (Supplementary Table. S1 in Supplementary Data^†), which was cut by a diamond saw, and its pieces were crushed in a disk mill to make a powder with particle sizes <100 µm to study the effect of grain or crystal size. Particles were further separated by a 38 µm sieve. Particles <38 µm were collected and mixed with distilled water to further separate particles of <10, 10–20, and 20–30 µm based on their settling rate calculated by Stokes’ law. Spectra were subsequently obtained for each size range of particle to assess the dependency of acquired spectra on particle size.

Raman spectra were acquired using a Renishaw inVia reflex Raman spectrometer at the University of Aberdeen. An Ar⁺ green diode laser with an excitation wavelength of 514.5 nm was focused through a 50× objective lens in a Lecia DMLM reflected light microscope onto the surface of the sample held in a microscope stage. Each spectrum was obtained by scanning the sample five times with an integration time of 30,000 ms/scan. The spectra were recorded at a laser power of 50% (approximately 15 mW reaching the sample surface) in the range of 100–1400 and 100–2000 cm⁻¹. This was determined to be the minimum laser power required for silanol detection, based on a sensitivity test (Fig. 2).

OPEN IN VIEWER

FIG. 2.

Raman spectra of flint (sample P11 in Table 1 acquired under varying acquisition conditions). (a) Spectrum obtained using a 10× objective lens (5 μm spot size), compared with (b) the spectrum from a 50× objective lens (1 μm spot size). (c) Spectrum recorded with a low laser power of 1% (approximately 0.3 mW reaching the sample surface), resulting in lower signal-to-noise ratio, while (d) spectrum at 50% (approximately 15 mW reaching the sample surface) laser power shows enhanced peak intensity.

The Raman spectra were processed using an R script (R Core Team, 2013), which included reading and normalizing the spectra, performing background subtraction, and identifying peaks within the Raman shift range of 360–600 cm⁻¹. The spectra were smoothed, missing values were interpolated, and the peak intensities were extracted for Raman bands in the range 499–503 cm⁻¹ for silanol and 450–480 cm⁻¹ for quartz, respectively.

The detection threshold for distinguishing signal from noise in each spectrum was determined using the signal-to-noise ratio, as outlined by Rosso and Bodnar, 1995 .

To ensure that the measured peak intensities reflected true signals rather than background statistical fluctuations, any I₅₀₃ values that fell within 95% confidence of the background noise intensity calculated by Eq. 1 were considered as having Sil_prop = 0 (Fig. 3).n net ≥ 2 2 n b ≈ 3 n b(1)

OPEN IN VIEWER

FIG. 3.

Schematic example of criteria for signal-to-noise ratio (S/N). Raman spectra of samples showing the intensity peaks at 464 and 503 cm⁻¹. The top spectrum illustrates a sample with high S/N where SiI_prop = I₅₀₃/I₄₆₄, calculated from the intensity ratio of the peaks at 503 cm⁻¹ and 464 cm⁻¹. The bottom spectrum shows a sample with low S/N where SiI_prop is treated as zero. The green horizontal bars represent the 95% confidence interval of the background noise.

Here, n_net is the net intensity of the peak (I₄₆₄ or I₅₀₃), and n _b is the average intensity of the 10 lowest points within the 360–600 cm⁻¹ range of each spectrum.

3.1. Raman spectra of silica minerals

Figure 4 presents examples of Raman spectra for opal, microcrystalline quartz, and macrocrystalline quartz. These Raman spectra illustrate end member spectra for a sample characterized as macrocrystalline quartz (Amethyst—sample 4471 in Table 1, Fig. 4a) where the main Raman band is at 464 cm⁻¹, with a smaller band at a Raman shift of 207 cm⁻¹. Cryptocrystalline or microcrystalline quartz (a sample of chalcedony—sample 3044 in Table 1, Fig. 4b) has Raman spectral features of quartz but also with a clear Raman band at 503 cm⁻¹.

OPEN IN VIEWER

FIG. 4.

Raman spectra for amethyst (sample code 4471) a macrocrystalline quartz, chalcedony (sample code 3044) a microcrystalline quartz, and opal predominantly amorphous silica but with some quartz.

Raman spectra for opal have broad poorly resolved bands from 200 to 450 cm⁻¹ and ∼700 to 900 cm⁻¹ (opal, Fig. 4c). Here, poorly resolved refers to the identification of maxima for individual peaks within the spectra, the intensity of these bands with respect to signal-to-noise ratio and intensity of features above background fluorescence.

Differences in the proportion of silanol are graphically represented in Figure 5 in the form of a proxy comparing the maximum intensity of the Raman band diagnostic for silanol at 503 cm⁻¹ (I₅₀₃) to that of quartz at 464 cm⁻¹ (I₄₆₄), here termed Sil_prop:Sil prop = I 5 0 3 / ( I 464 + I 5 0 3 )(2)

OPEN IN VIEWER

FIG. 5.

Evaluation of the silanol content in samples from Table 1 using the Sil_prop parameter. Data are grouped as opals, microcrystalline quartz, and macrocrystalline quartz.

The use of a peak at 503 cm⁻¹ as a measure of silanol content is based on work by Schmidt et al. (2012), which demonstrated a positive correlation between silanol concentrations measured by relative intensity of a Raman band at 503 cm⁻¹ (I_{503 cm−1}/I_{464 cm−1}) and also a silanol-derived adsorption feature at 4547 cm⁻¹ in NIR spectra. Although superficially the Raman bands of moganite and silanol may appear to overlap in the 500–503 cm⁻¹ range (Schmidt et al., 2012), the peaks observed in this study are predominantly attributed to silanol (see Supplementary Data^† for fuller explanation). In brief, the Raman spectra of moganite and silanol on α-quartz can be distinguished by comparing peak centers and FWHM of the bands at 501 and 503 cm⁻¹ (Supplementary Fig. S1, Supplementary Data^†). Moganite has a broader peak at 501 cm⁻¹ and does not have a band at 503 cm⁻¹. Applying this methodology, the peaks around 503 cm⁻¹ in this study are concluded to be predominantly composed of silanol (Supplementary Fig. S2, Supplementary Data^†). Consequently, the Raman spectroscopy revealed that 52 of the 71 quartz samples exhibited peaks around 503 cm⁻¹ (Table 1) and therefore contained silanol.

Based on the Sil_prop parameter (Eq. 2), the samples of macrocrystalline quartz have less silanol than samples of opal and microcrystalline quartz (significant for p < 0.001). Samples of opal and microcrystalline quartz have somewhat overlapping values of Sil_prop, (significant p < 0.09). The microcrystalline samples that have higher values of Sil_prop are samples of chalcedony and the samples of flint. There is some suggestion of a complex bimodal distribution with some samples having Sil_prop values close to 0 but certain microcrystalline samples having values closer to 0.1.

Previous experimental work has demonstrated that the ratio of the intensity of the silanol peak at 503 cm⁻¹ (I₅₀₃ to α-quartz at 464 cm⁻¹ I₄₆₄) can be changed by heating (Schmidt et al., 2012) and altering particle size (Tsukada et al., 2024). Based on this, it might have been expected that Sil_prop would be highest in the samples of opal (such as the gemstone-opal, jasper, and geyserite) that have experienced lesser geological heating, but this is not the case. Instead, samples of microcrystalline quartz have higher Sil_prop values. The simplest interpretation is that the samples of microcrystalline quartz have a much greater proportion of quartz surfaces that can host silanol; such samples have a higher quartz content and smaller crystal size. Additionally, it can also be seen that samples that are replacement or epigenetic mineral phases and that precipitated from water have among the highest values (triangles in Fig. 5). For the samples in Table 1, elevated Sil_prop values are seen where silica replaced different primary phases such as melanophlogite, carbonate, and wood, which suggests that the replacement process enhances Sil_prop values as much as the mineral or material being replaced. Thus, the presence of significant α-quartz and the prevalence of a free water phase can lead to high Sil_prop. Samples of macrocrystalline quartz have lower values of Sil_prop than microcrystalline quartz, which can be explained by water being eliminated as microcrystallites recrystallize into coarser crystals, which typically occurs at high temperatures and later stages of diagenesis and even metamorphism. This is consistent with experimental work (Schmidt et al., 2012: Tsukada et al., 2024).

Although a few samples may exhibit Raman spectra that match only the Raman spectra of end member mineral components such as α-quartz or opal-A, most samples have sufficient mineralogical heterogeneity that this is not the case. In fact, many of the samples of opal, after averaging of five measurements, yielded Raman spectra with strong characteristics of α-quartz. Thus many samples classified as gemstone-opal did not yield Raman spectra dominated by Raman bands characteristics of opaline minerals (often a fluorescent background dominated).

The discrepancy between the characteristics of a sample in hand specimen and its Raman spectra can be explained by both mineralogical heterogeneity and the strong Raman response of α-quartz compared with other mineral phases. The effect of mineralogical heterogeneity was evaluated by considering different particle sizes in cryptocrystalline or microcrystalline samples that contain both opal-CT and α-quartz (Fig. 6a–c). The smallest particle sizes of such samples (Fig. 6a, b) have Raman spectra with a broad band indicative of opal-CT at Raman shifts from 200 to 450 cm⁻¹, accompanied by subordinate peaks at 215, 252, 304, and 400 cm⁻¹. Conversely, peaks for α-quartz at 464 and 394 cm⁻¹ are most evident in the large particle sizes (Fig. 6c). Within the coarse particles (20–30 µm) there is additional heterogeneity depending on the spot position on the samples where measurements were made, with a difference between Raman measurements dominated by features of opaline minerals (Fig. 6d–g) and Raman spectra dominated by α-quartz (Fig. 6h). There is also a lesser difference in the relative intensity of subordinate peaks diagnostic for opal—such as the band ∼ 400 cm⁻¹ that corresponds to the most intense Raman band of tridymite or cristobalite; for example, compare the resolution of the peak at 400 cm⁻¹ in Figure 6g and e to the broadband from 200 to 450 cm⁻¹ in Figure 6f. Thus, the resultant Raman spectra of many samples of silica will heavily depend on the spot where the measurement is made and the relative proportion of silica minerals that exhibit the strongest Raman scattering.

OPEN IN VIEWER

FIG. 6.

(a) Raman spectra for different size fractions of a crushed sample (sample P4, Table 1). (b) Repeat measurements for the 20–30 µm size fraction.

The relative strength of signal generated is not quantitatively evaluated here but can be visually assessed by comparing the relative spikiness or noise in the spectra of Figure 6g and h. After rescaling the maximum intensity of the band for quartz at 464 cm⁻¹ causes the spikey small noise-peaks to be no longer discernible in Figure 6h.

Therefore, a laser’s micrometer-sized area of illumination may not illuminate sufficient opaline minerals to generate discernible Raman spectra features in many cases, even if opaline minerals comprise a high proportion of the sample. In contrast to the opaline minerals, even partial illumination of α-quartz-domains within a sample may induce sufficient Raman scattering for the detection of α-quartz and also quartz-hosted silanol. Consequently, in aggregated spectra (e.g., superposition of one or more original spectra) and averaged spectra (addition of spectra divided by the number of spectra), Raman scattering generated by quartz phases will typically be stronger and more prominent than scattering generated by opal phases.

Samples whose Raman spectra indicated silanol underwent additional examination to assess the presence of CM based on Raman shifts between 100 and 2000 cm⁻¹ to investigate the potential for a correlation between silanol and CM. In total, 12 out of 52 silanol-bearing samples were found to have at least one G-band (graphite-type lattice vibration) or D-band (graphite-type disordered lattice vibration), which demonstrated the coexistence of CM and silanol at the same point of focus in the sample. This shows the potential of Raman spectroscopy to detect both evidence of past habitability in the form of water and also CM as potential chemical evidence of past life.

An example of the co-occurrence of silanol and CM is shown for a sample of onyx (Fig. 7a), which has distinct black and white bands but is also cut by a quartz vein. While the sample is composed mostly of silica phases, there are important differences in the phases found in different parts of the sample. The most obvious difference is between macrocrystalline vein-quartz and the microcrystalline quartz comprising black bands, some of which show the existence of α-quartz and CM (Fig. 8). It is particularly notable that CM was only detected where silanol-bearing α-quartz was observed. The silanol bands attest to the presence of free water when the initially amorphous silica phase crystallized, whereas the disordered carbonaceous phase could be a product of past microbial activity (Götze et al., 2012) or hydrothermal activity (bitumen mobilized by a hydrothermal system). The Raman spectra of the macrocrystalline quartz vein lack bands diagnostic for both CM and silanol.

OPEN IN VIEWER

FIG. 7.

(a) Banded onyx (sample 3229, Table 1), labeled are white and black bands, and quartz vein. (b) Chalcedony pseudomorphed after cubic fluorite (sample 3044, Table 1), prismatic quartz on back of sample.

OPEN IN VIEWER

FIG. 8.

Raman spectra for different regions of onyx sample 3229, Table 1. (a)–(c) Raman spectra illustrating the presence of quartz and silanol. (d)–(f) Raman spectra illustrating the presence of carbonaceous organic matter.

Another instance of the codetection of quartz, silanol, and CM within Raman spectra is shown in Figure 9. This sample comprises silica phases pseudomorphed after fluorite (Fig. 7b) and primary prismatic quartz (Fig. 7c). Raman spectra obtained from the centimeter-sized cubic crystals (chalcedony pseudomorphed after fluorite) have peaks for α-quartz, silanol, and CM (Fig. 9b, d), but Raman spectra measured on the primary quartz phase have spectra with peaks for α-quartz (Fig. 9a, c). The fluorite pseudomorphs can be taken as an example of a replacive silica phase, in which the original melanophlogite has been replaced (Adrian-Iulian et al., 2023). Initially deposited as a silica gel, these phases subsequently undergo opalization, dissolution, precipitation, and recrystallization, and ultimately form microcrystalline quartz with associated structural defects that resulted from the incorporation of trapped water in the form of silanol (Adrian-Iulian et al., 2023). Additionally, CM may be incorporated as an exotic component during formation (Foucher et al., 2013). A parent fluid that comprises both water and organic matter can initially serve as a microbial shelter but also has the potential to form inclusions that preserve evidence of habitation (González-Ramírez et al., 2023), with the caution that carbon signatures in Raman spectra may not necessarily be a true biomarker (Edwards et al., 2014).

OPEN IN VIEWER

FIG. 9.

Raman spectra for different sides of sample 3044, Table 1. (a) and (b) Raman spectra illustrating the presence of quartz and silanol. (c) and (d) Raman spectra illustrating the presence of carbonaceous organic matter.

4.1. Silanol in the context of other methods of trace water detection

Residual water and water films (O’Niell et al., 2023) held by porous media are known to support endolithic microbial communities (Cockell et al., 2002). While porewater is considered to be physically bound (often by capillary forces) and thus better resists evaporation, it is not chemically isolated and can be removed by prolonged drying or destruction of pore structures. The size of pores can vary, but typically within coarse-grained sedimentary rocks pore structures (individual pore throats and pores) are of a scale less than 100 s of micrometers across, although this can be greater for fractures or where grains are plucked from lithic surfaces. Although by strict definition pore spaces filled by an authigenic mineral phase are lost (they are no longer pore spaces), the replacing mineral phase may encapsulate pore fluids and microbes (Parnell and Baron, 2004; Benison and Karmanocky, 2014). Additionally, pore-filling cements and biofilms may also possess stable isotope signatures that preserve evidence of dissimilatory metabolic processes (Chan et al., 2019). Thus, the speciation of water at the pore scale is important for astrobiological exploration so long as the samples analyzed have preserved pore structures.

Fluid inclusions form where a precipitating mineral phase encapsulates microscopic regions of the parent fluids (Benison and Karmanocky, 2014; Parnell and Baron, 2004). In addition to parent fluids, other components can be encapsulated, which include microorganisms (Grant et al., 1998). Inclusions may survive for millions of years, but they can be ruptured, change shape, and be destroyed if the host mineral recrystallizes. Despite such challenges, inclusions have been shown to preserve a paleoenvironmental record that can survive for billions of years (Dutkiewicz et al., 2006). Fluid inclusions can be visible with the naked eye (millimeters in size) though they are typically much smaller, with the size of inclusions ranging from 1 to 100 um in size.

Water is chemically bound within phyllosilicate and sulfate minerals (Basilevsky et al., 2003; Flahaut et al., 2015), and these mineral phases are recognized as being important records and evidence of the presence and habit of water in past environments.

Various silica minerals and polymorphs other than quartz contain chemically bound water, and this association is especially strong in the case of hydrothermal deposits. Such deposits are not widespread on Mars but are present, such as in the Gusev crater, where the Alpha Particle X-ray spectrometer detected a silica content exceeding 90% (Squyres et al., 2008). While such silica-rich sediments are likely primarily opaline minerals, reworked siliciclastic sediment that consists mainly of quartz, feldspar, and phyllosilicate minerals was identified in the Antoniadi crater (Bandfield et al., 2004; Bandfield, 2006).

From the data presented here, samples of either opal (opal as a material and not a pure mineral phase) or reworked detrital sediment enriched in quartz (quartz is the most mechanically resistant SiO₂ polymorph) might both be expected to produce Raman spectra with silanol. Thus, Raman spectra might be measured on quartz in sediments both proximal and distal to primary sources (e.g., an outcrop and sediment weathered from an outcrop, respectively), to look for silanol to obtain records and evidence of water in past environments. Relative to other observations of water within rocks and sediments, silanol is the smallest in scale and the most firmly intracrystalline molecule, able to survive mechanical attrition and chemical degradation since it exists as closed inclusions isolated from the outer atmosphere (Gaweł et al., 2020).

4.2. Silanol occurrence by lithology

If water is present, then silanol groups can form during crystal growth, deformation, cracking, and healing (Kronenberg, 1994). Furthermore, silicon that has dissociated from quartz crystals and is in an aqueous phase can form silicic acid, which may go on to form silanol as silica minerals precipitate (Dove and Rimstidt, 1994). Thus, silanol might be expected to be extremely prevalent on the surfaces of quartz when water is relatively abundant during mineral formation, and therefore Raman spectra measured on quartz might commonly be expected to possess bands generated by silanol. However, based on the data presented here, this is not the case; many samples were observed to have a baseline value of the Sil_prop parameter (less than from 0.015 to 0), which indicated that most quartz surfaces do not characteristically generate Raman spectra indicative of silanol.

The occurrence of silanol was observed to be highest in cases where the portion of microcrystalline quartz is highest relative to other silica minerals (Fig. 10). Samples of macrocrystalline quartz, and ancient stromatolite samples, such as the 3.4 Ga Strelley Pool Chert from the Pilbara Craton (sample ID P7), showed lower intensity of silanol peak than microcrystalline quartz. During heating experiments, silanol bands were shown to vanish from Raman spectra following a 6-hour heating process at 700°C (Schmidt et al., 2012). Thus, recrystallization of microcrystalline quartz to macrocrystalline quartz, or crystallization under anhydrous conditions and high pressure and temperatures as can occur in ancient cratons on Earth, likely eliminates crystal defects, removes hydroxyls, and results in very low silanol and Sil_prop values.

OPEN IN VIEWER

FIG. 10.

Schematic diagram illustrating silanol in the context of water speciation in different silica mineral phases, and the meaning of Sil_prop.

For the samples in this study, it was notable that replacement phases of microcrystalline quartz had the highest Sil_prop values (0.175–0.35, Fig. 10), which perhaps indicates that direct precipitation from an aqueous phase provides the greatest opportunity for the incorporation of water within quartz as silanol.

Even though both free molecular water and surface-bound water are generally known to be higher within opal-A and other microcrystalline opaline minerals (Graetsch, 1994; Herdianita et al., 2000), it was found in this study that many opals (both as gems and lithologies) had low Sil_prop values, or silanol was not detected. Since the Raman peak around 503 cm⁻¹ is α-quartz related and not hosted by amorphous silica phases, the very low values of Sil_prop reflect the lack of crystalline SiO₂ and crystalline quartz.

Relative to spectra for opal, spectra for quartz-hosted silanol were much easier to acquire. This observation was particularly notable when analyzing powders. Thus, drilling and crushing during the collection of samples by rovers (Vago et al., 2017) may make it easier to measure silanol within opaline samples, especially considering the effect of a larger spot size or issues caused by diffuse illumination under sub-optimally focused lasers applied to a range of grain sizes (Foucher et al., 2013).

However, the detection of silanol in silica-based minerals, such as quartz and opal, through Raman spectroscopy is significantly influenced by acquisition parameters, particularly spot size and laser power. These parameters are critical when applying Raman techniques in planetary exploration, such as on the martian surface, where precise identification of mineralogical features is crucial. For example, spot size in Raman microscopy is determined by the magnification of the objective lens. Analysis of the same sample with the use of a 50× objective lens (producing a 1 µm spot size) reveals distinct silanol peaks, whereas a 10× objective lens returns a spectrum from the same point of focus with slightly different peak positions (Fig. 2a, b). Heating effects from the Raman laser also shift and reduce the peak positions for α-quartz at 464 cm⁻¹ and silanol at 503 cm⁻¹ (Chio et al., 2003). In addition, Sil_prop increases with heating since the reduction in intensity of the silanol peak at 503 cm⁻¹ is less than that of α-quartz at 464 cm⁻¹ (Tsukada et al., 2024). This demonstrates that although silanol can still be detected at lower spatial resolutions and magnification, the position of a peak and the resultant Sil_prop can be affected by laser-induced heating of a sample. This variability is particularly significant in planetary exploration, where a larger spot size may result in a lower Sil_prop compared with lab-based measurements.

In addition to spot size, the level of incident laser power significantly affects the detection of silanol. Figure 2c and d illustrate the Raman spectra obtained from flint particles at varying settings of laser power. At a lower laser power of 1%, the signal-to-noise ratio for the peak intensity at 503 cm⁻¹ is insufficient for reliable detection of the 503 cm⁻¹ peak (Fig. 2c). However, increasing the power to 50% enhances the signal-to-noise ratio, allowing silanol to be discernible above the background noise, although a potential risk of sample-heating exists (Fig. 2d). These variations highlight the influence of magnification and laser power on silanol detection and overall spectral quality. The findings indicate that a minimum laser power threshold of 50% (approximately 15 mW at the sample surface) is necessary for consistent detection of silanol on α-quartz, although it is important to balance this improved consistency against the risk of sample degradation or the alteration of spectra due to excessive heating.

When applying these findings to Mars exploration, taking the ExoMars rover equipped with the Pasteur payload as an example, the particle size distribution created by the onboard rock crusher should also be taken into account. The rock crusher on the rover generates particles that range from a few micrometers to a maximum size of 500 µm, with a median particle size of approximately 250 µm (Vago et al., 2017). Given these particle sizes, a Raman spot size of around 50 µm would cover multiple particles and lead to spectra that represent a mixture of minerals (Rull et al., 2017). This mixing effect complicates the detection of silanol and other fine structural features, as heterogeneously distributed phases might be averaged within the spot size as seen in the spectra of Figure 6.

Silanol detection in opal or quartz, especially in powder form, is relatively straightforward, as demonstrated by the ease of obtaining quartz spectra (Fig. 6h). However, this process also has its challenges. One concern is that mechanical actions like crushing and drilling, commonly used in sample collection by rovers, may artificially enhance silanol detection by increasing surface area and exposing fresh quartz surfaces. Additionally, this procedure may introduce difficulties related to diffuse illumination, uneven laser focusing, and spectral interferences from mixed mineral grains, particularly with larger spot sizes (Fig. 6). Another issue is the need for caution when interpreting Sil_prop, as both the peak at 464 cm⁻¹ produced by Si-O vibration and the silanol peak at 503 cm⁻¹ are affected by particle size. Specifically, the intensities of these peaks decrease as particle size diminishes (Tsukada et al., 2024). For instance, the intensity of the 464 cm⁻¹ peak, characteristic of α-quartz, drops significantly when particle size falls below 10 µm (Chio et al., 2003). As a result, the acquisition parameters may need to be optimized to gather consistent data under varying surface conditions on Mars.