Dark spots on Mercury show no signs of weathering during 30 Earth months

Mercury’s surface is unusually reduced and volatile-rich, with low Fe and Ti but elevated S, C, and Cl. Volatile-related landforms such as pyroclastic deposits and hollows are widespread, and a growing consensus holds that graphite is a major opaque phase potentially derived from an early magma ocean. However, the precise darkening agents, formation mechanisms, and lifetimes of low-reflectance dark spots associated with hollows remain unclear. Prior work suggested these spots might be short-lived volatile products and potentially distinct from hollows in origin and evolution. This study asks: what are the dominant darkening phases in dark spots, what is their sub-pixel texture, and do they undergo observable reflectance changes on decadal timescales? Using the complete MESSENGER imaging dataset, the authors aim to constrain composition, formation processes (e.g., outgassing), and temporal stability of Mercury’s dark spots, thereby informing models of the planet’s crustal building blocks and space weathering of graphite.

Previous studies identified Mercury’s low-reflectance material (LRM) and proposed graphite as a key darkening phase, while other candidates (chlorides, sulfides such as MgS and CaS) were also advanced to explain hollow-related absorptions near 600 nm. Earlier spectral sampling often mixed multiple hollow facies (halos, floors, host terrains), complicating interpretation of devolatilization signatures. Dark spots, young and diffuse low-reflectance materials beyond bright hollow halos, were hypothesized to arise from energetic outgassing and to be shorter-lived than hollows. Some preliminary spectral analyses proposed sulfides as a major component and darkening phase for dark spots, though their high reflectance and thermal instability at Mercury day temperatures challenge this view. The authors build on global MESSENGER color mosaics, LRM spectral definitions (including BD600 near 560 nm linked to graphite content), and prior phase-ratio approaches to assess sub-resolution texture and track temporal changes.

• Data: Full MESSENGER MDIS dataset (NAC single band at 750 nm; WAC 11 narrow bands from ~430–1020 nm); global 8-band color mosaic at 665 m/pixel with ~2% photometric residuals was used for quantitative spectral comparisons. Single-frame images were processed with USGS ISIS, using the Kaasalainen–Shkuratov photometric model.
• Inventory update: Using high-resolution NAC and multi-band WAC imagery from the entire mission, the authors applied the established definition of dark spots (isolated, dark, diffuse materials around hollows, not impact or volcanic origin) to produce an updated global catalog.
• LRM relationship and spectral metrics: Adopted the spectral definition of LRM and computed BD600 (band depth near 560 nm) from the global mosaics to evaluate relative graphite enrichment and the spatial association between dark spots and LRM.
• Reflectance spectra extraction: For seven relatively large dark spots (>5 km²), regions of interest were carefully selected using manually georeferenced high-resolution NAC over the global color mosaic to avoid shadows, margins, ejecta rays, bright halos, or pyroclastics. Spectra were compared to typical LRM, low-reflectance blue plains (LBP), and flat hollow floors.
• Sub-pixel roughness: Phase-ratio imagery R(g1)/R(g2) from co-registered NAC frames with similar pixel scales and incidence angles but sufficiently different phase angles (>20°) was used to infer relative sub-pixel roughness/particle size via shadow-hiding behavior. Dark spots, bright halos, hollow floors, and background regolith were compared.
• Temporal imaging: For three dark spots with suitable repeated NAC coverages (similar geometry and resolution) spanning 6, 18, and up to 30 Earth months, high-precision coregistration (ISIS coreg with manual control points) enabled before–after reflectance ratio analyses. Cross-mission visual comparisons with Mariner 10 images (1974) were also made to assess multi-decade changes, acknowledging lower resolution and calibration fidelity of Mariner 10.

• Global inventory: 59 dark spots identified. They are not preferentially located at longitudes 0° or 180°. About 83% occur within LRM; 61% develop in areas with positive BD600 values.
• Spectral properties: Dark spots have the lowest reflectance among surveyed mercurian materials and show shallow kinks near 560 and 750 nm. Their BD600 values are significantly larger than those of LRM and LBP, while hollow floors tend to have negative BD600.
• Composition inference: Using the empirical BD600–graphite relation for LRM, dark spots may have up to ~5 wt% higher graphite content than the global average, consistent with their very low reflectance and young ages.
• Sub-pixel texture: Phase-ratio analyses indicate dark spot materials are rougher (coarser grains and/or rougher surfaces) than bright halos and background regolith, but smoother than collapsed lag deposits on flat hollow floors. Thickness-proximal materials (closer to hollows) appear rougher than distal deposits.
• Temporal stability: No detectable intrinsic reflectance changes over 30 Earth months for three cases; observed variations (<~3%) match residual photometric calibration errors (≈1.8–2.3%) and track background changes. Visual comparisons with Mariner 10 images suggest no obvious changes over >40 Earth years (at ≤~730 m/pixel), implying stability over ~247 Mercury days.
• Implications for sulfides: Rapid thermal decomposition expected for MgS/CaS at Mercury day temperatures conflicts with the observed stability; combined with their higher intrinsic reflectance, these sulfides are unlikely to be major components of dark spots.
• Origin: Phase-ratio results support an energetic outgassing origin, potentially entraining graphite-rich material.

The findings address the core questions by showing that dark spots are an LRM subtype with exceptionally strong 560 nm absorption consistent with graphite enrichment, rather than sulfide dominance. Their rougher sub-pixel textures and spatial relation to hollows support deposition from energetic outgassing. The absence of observable reflectance evolution over 30 months (and likely decades) implies that space weathering-driven graphite alteration in dark spots proceeds too slowly to detect with MESSENGER, contrasting with expectations for thermally unstable sulfides. This stability constrains models for hollow growth and dark spot evolution: either hollow formation involves non-uniform rates (rapid initiation, steady sublimation, episodic wall collapse) and/or graphite reduction is less efficient in dark spot veneers than along hollow walls. The requirement of substantial subsurface gas to drive outgassing (>100 m/s) presents challenges given regolith porosity and the need for gas accumulation; nevertheless, an outgassing mechanism remains consistent with observed textures and distributions. The results provide a benchmark for laboratory simulations of graphite space weathering and for future orbital spectroscopy to refine compositional constraints.

This work updates the global catalog of Mercury’s dark spots (59 cases) and demonstrates that they are graphite-enriched LRM deposits likely emplaced by energetic outgassing. Spectrally, dark spots exhibit the deepest ~560 nm absorption and lowest reflectance among mercurian units, with sub-pixel textures rougher than halos/regolith but smoother than hollow-floor lag. Crucially, no intrinsic reflectance changes were detected over 30 Earth months and likely none over >40 years, arguing against thermally unstable sulfides (e.g., MgS, CaS) as major components and implying slow space weathering effects on graphite at MESSENGER sensitivities. These observations refine the understanding of volatile-related processes and provide a reference for modeling graphite’s spectral evolution. Future research should leverage BepiColombo (MERTIS, SIMBIO-SYS) for higher-fidelity spectroscopy and imaging to resolve composition, detect subtle temporal evolution, and test outgassing mechanisms and subsurface gas accumulation scenarios.

• Temporal analyses could be performed for only three dark spots with suitable repeat coverage; small sample size may limit generality.
• Detected reflectance variations are at or below the residual photometric calibration errors (≈1.8–2.3%), constraining sensitivity to subtle changes.
• Cross-mission comparisons (Mariner 10 vs. MESSENGER) are limited by Mariner 10’s lower resolution and less sophisticated calibration.
• Composition is inferred indirectly from spectral parameters (BD600) and empirical relations; definitive mineralogical identification remains uncertain, and the stability/optical behavior of chlorides on Mercury is not well constrained.
• Spatial resolution (global mosaics at 665 m/pixel) and small areal extent of many dark spots may mask localized or very thin-surface changes.