Rapid Aggregation and Dissolution of Organic Aerosols in Liquid Methane on Titan

Complex organic aerosols in the upper atmosphere of Saturn's moon Titan reach the troposphere and surface, where a methane (CH4)‐based hydrological cycle occur. Previous studies have assumed no interactions between organic aerosols and liquid CH4, although the dissolution of low‐molecular‐weight photochemical products in liquid CH4 has been considered. Here we report experimental results of soaking a laboratory analog (so‐called tholin) of Titan's organic aerosols in liquid CH4 at 93–98 K for several hours and then evaporating the liquid, simulating wet–dry cycling on Titan. After wet–dry cycling, residual tholin particles form aggregates through cementation. Solid evaporitic deposits formed by evaporation of interacted liquid contain nitrogen‐bearing aromatics, suggesting selective dissolution of aromatics. Our results suggest that organic aerosols or high‐molecular‐weight compounds adsorbed on them partly dissolve in liquid CH4 on Titan, even during short‐term wetting events, promoting the growth of aerosols to dune particles via aggregation and providing aromatics to evaporites.

10.1029/2023GL103015 2 of 9 of evaporites of organic materials deposited in lowlands at middle-high latitudes (e.g., Barnes et al., 2011;MacKenzie et al., 2014). In arid low-latitude regions, seasonal CH 4 rainfall temporarily wets the surface (e.g., Turtle et al., 2011). Although previous studies on chemical compositions of Titan's lakes have usually considered only thermodynamic equilibria with the ground-level atmosphere and simply assumed that complex organic aerosols are insoluble in liquid CH 4 on Titan (Cordier et al., 2009(Cordier et al., , 2010(Cordier et al., , 2012Glein & Shock, 2013;Tan et al., 2013), laboratory experiments have indicated that tholin is moderately soluble in both non-polar and polar solvents at room temperature (Carrasco et al., 2009;He & Smith, 2014a, 2014bMaillard et al., 2018;McKay, 1996). Such dissolution experiments were intended mainly to elucidate the chemical structure of Titan tholin (e.g., Carrasco et al., 2009;He & Smith, 2014a, 2014bMaillard et al., 2018), and there has been no experimental study of the dissolution of tholin in liquid CH 4 at temperatures comparable to those on Titan's surface (∼93 K) (Jennings et al., 2019).
Here we experimentally investigated interactions between tholin and liquid CH 4 at low temperatures comparable to Titan's surface. To simulate wet-dry cycling on Titan, a cryo-chamber was developed in which tholin particles were immersed in liquid CH 4 at 93-98 K. After immersion and subsequent evaporation, we analyzed undissolved (residual) tholin and solid deposits formed from dissolved components (evaporitic deposits). Based on the chemical analyses and microscopic observations for both residual tholin and evaporitic deposits, we discuss whether Titan tholin is soluble to liquid CH 4 and implications for interpreting the surface features, such as evaporites and dunes, on Titan.

Tholin Particle Preparation
The experimental system for producing Titan tholin particles was similar to a previous study  and comprised a quartz tube of 38 mm outside diameter inserted into a copper coil equipped with matching networks, a 13.56 MHz radiofrequency power supply (Nihon Koshuha) for plasma irradiation, and a downstream glass trap ( Figure S1 in Supporting Information S1). A pre-mixed CH 4 /N 2 gas mixture (10.1%:89.9% by volume; Suzuki Shokan, Japan) was introduced into the quartz tube at a flow rate of 3 standard cm 3 min −1 (sccm) using a mass-flow controller to maintain the pressure at ∼200 Pa, as measured using a transducer gauge (ConvecTorr gauge, Agilent). Al substrates ∼1 cm square were placed within a downstream glass trap to collect tholin particles ( Figure S1 in Supporting Information S1). Before the experiments, the Al substrates were baked at 673 K for several hours to remove organic contaminations. Cold plasma irradiation with the RF system at 70 W generated tholin in the gas phase; it was transported with the gas flow and deposited on the Al substrates.

Dissolution Experiments
Dissolution of Titan tholin was carried out using a cryochamber developed by the present study ( Figure S2 in Supporting Information S1). The cryo-chamber comprised a stainless-steel vessel with bellows (ADCAP Vacuum Technology) connected to a CH 4 gas cylinder (purity 99.999%; Suzuki Shokan) and a rotary pump through a gas sampling system including a liquid-nitrogen cold trap. The vessel had a two-stage bottom. The Al substrates (with tholin particles attached) were placed on a 100 μm Ti mesh over a small dent in the base of the chamber, and an Au plate was placed below the mesh to collect evaporitic deposits ( Figure S2 in Supporting Information S1). The interior of the cryo-chamber was observed through a viewport at the top. The chamber was cooled by immersion in a liquid-nitrogen bath, and the interior temperature was monitored using a Pt resistance temperature probe (Ptc100, Lakeshore) coupled to a temperature controller (Model 122C, CryoCon) to maintain the temperature at 93-98 K.
Dissolution experiments proceeded as follows. With tholin particles (∼5 mg) on the Al substrates, the cryo-chamber was evacuated to a pressure of ∼1 Pa ( Figure S2a in Supporting Information S1). After cooling to 110 K, CH 4 was introduced and the temperature lowered to ∼90 K as the bellows extended during gas influx ( Figure S2b in Supporting Information S1). The cryo-chamber was then immersed in liquid nitrogen, condensing liquid CH 4 ( Figure S2b in Supporting Information S1). The temperature inside the chamber was controlled to 93-98 K during the experiments with gas pressures of 15-40 kPa, consistent with the saturation vapor pressure of CH 4 at 93-98 K (see Figure S3 in Supporting Information S1). The amount of liquid CH 4 produced was estimated to be ∼10 mL, based on height of liquid surface in the chamber as seen from the viewport on the top of the chamber.

of 9
After dissolution for 6 hr ( Figure S2c in Supporting Information S1), the CH 4 solution was slowly (∼30 min) evaporated by evacuation of CH 4 in the gas phase of the chamber by adjustment of leak valves, avoiding boiling ( Figure S2d in Supporting Information S1). The temperature was maintained at 93-98 K. During evaporation, the amount of liquid CH 4 decreased, and dissolved matter in the liquid would be concentrated. The liquid that has decreased in the amount was gathered at the bottom of the interior of the chamber ( Figure S2e in Supporting Information S1). Finally, the droplets at the bottom disappeared ( Figure S2e in Supporting Information S1), leaving dissolved matter as evaporitic deposits on the Au plate in the small depression ( Figure S2f in Supporting Information S1). After complete evaporation, the liquid-nitrogen bath was removed and the interior temperature raised to room temperature (∼20°C) for the collection of reacted samples from the Al substrates and Au plate. During the increase in the interior temperature of the cryochamber, we closed a valve on the top of the cryochamber to avoid losing moderate volatiles degassing from the evaporitic deposits, if present. Then, we set a cold trap (77 K or 160 K) in the downstream of the cryochamber and opened the valve to collect moderate volatiles in the cryochamber. Volatiles trapped at the cold trap was collected with a gas-tight syringe and analyzed with gas chromatograph mass spectrometry (GC-MS) (Supporting Information S1 for the method).
The reproducibility of the experiments was confirmed by IR and XANES spectra of residual tholin aggregates and evaporitic deposits, based on multiple runs (see Figures S4 and S5 in Supporting Information S1). A control experiment with no tholin added to liquid CH 4 confirmed there was no formation of evaporitic deposits.

Sample Analyses
Fourier-transform infrared spectroscopy (FTIR; Frontier, PerkinElmer) was used for the analysis of the original tholin particles with a diffuse-reflection method and mixing with KBr powder. To eliminate the effects of atmospheric H 2 O and CO 2 during analysis, the laboratory was purged using nitrogen (>99.99995% purity). FTIR absorption spectra of evaporitic deposits and residual tholin aggregates were obtained using a Bruker (INVENIO) system equipped with an IR microprobe (Hyperion) (micro-FTIR). Unlike the bulk FTIR analyses for the original tholin particles, the micro-FTIR analyses were performed in air due to equipment limitations. The mercury-cadmium-telluride (MCT) detector was cooled using liquid nitrogen. Reflection-mode absorption spectra were obtained in a range of 7,000-650 cm −1 with 100 scans per sample at a resolution of 4 cm −1 , and with up to 50 areas of each sample being analyzed with a 50 × 50 μm aperture. The background spectrum of the Au plate was acquired before sample measurements. XANES spectra were obtained using a Photon Factory Beamline 19B system at the High-Energy Accelerator Research Organization, Tsukuba, Japan. Spectra in electron-yield mode around the C (280-300 eV) and N (390-410 eV) K edges were collected using an energy step of 0.1 eV and dwell time of 100 or 200 ms per step (see Figure S6 in Supporting Information S1 for N K-edge results). The XANES spectra were calibrated against the E0 absorption edge and normalized using Athena software v. 0.9.26. The diameter of the XAFS beam was ∼300 μm, covering evaporitic deposits of 10-200 μm size. To identify the location of evaporitic deposits and tholin particles, an optical microscope associated with XANES was used.
Samples collected from that Al substrate and Au plate were examined using a field-emission scanning electron microscope (FE-SEM; Regulus 8230, Hitachi High Technology) and FE electron probe microanalyzer (FE-EPMA; JXA-8530F, JEOL). Osmium vapor and carbon coating was applied to each sample for FE-SEM and FE-EPMA analyses, respectively. Since osmium vapor and carbon coating was done after Raman, IR, and XANES analyses, the coating does not affect the results of these spectroscopy.

Scanning Electron Microscope (SEM) Analyses
Microscopic observations revealed that the original tholin particles were coagulated after wet-dry cycling (Figures 2b and 2c). Individual particles of aggregates appear to be connected with necks (Figure 2c), suggesting cementation of tholin particles. In contrast, evaporitic deposits of typical size 10-200 μm were flat and ameboid in shape with rough surfaces (Figure 2d) and were distinct from residual tholin aggregates and original tholin particles. Evaporitic deposits consisted of sheets of plate-shaped materials (Figure 2e). Pits of diameter ∼0.5 μm were often observed in the plate-shaped materials (Figure 2f) where liquid CH 4 may have been trapped in evaporitic deposits and degassed upon evaporation. Alternatively, pit structures may be hopper crystals resulting from preferential depositions at edges (e.g., Penha et al., 2021). No evaporitic deposits were found in a control experiment without tholin particles, indicating that the deposits found in the experiments with tholin were produced by the dissolution of tholin particles. Despite dissolution, the morphologies of individual tholin particles appear unchanged (Figure 2), suggesting a rigid structure.

Infrared (IR) Analyses
IR analyses indicate that evaporitic deposits have strong absorption at 1,500-1,400 cm −1 (Figure 2). Alkanes exhibit absorptions at 1,500-1,400 cm −1 , but the lack of C-H absorption at ∼2,900 cm −1 in evaporitic deposits implies that alkanes were not the cause of their 1,500-1,400 cm −1 absorption; rather, the absorption is more likely due to semi-circle stretching of aromatics such as nitrogen-bearing hetero-aromatic rings (Larkin, 2017). Cold-plasma tholin contains nitrogen-bearing hetero-aromatics such as pyrroles (McGuigan et al., 2006), which would be selectively fractionated in liquid CH 4 and retained in evaporitic deposits. The presence of nitrogen-bearing aromatics in evaporitic deposits would be supported by IR absorption in the 1,600-1,570 cm −1 range, where absorption peaks are attributed to C=C and C=N bonds in aromatics (Imanaka et al., 2004); on the other hand, residual tholin aggregates and original tholin exhibit a peak at ∼1,630 cm −1 , which is attributed to conjugated C=C and C=N bonds (Figure 2; Imanaka et al., 2004). In the wavenumber range of 2,200-2,100 cm −1 , residual tholin aggregates contain a combination of multiple types of nitriles (Imanaka et al., 2004) such as conjugated nitriles (2,180 cm −1 ), aliphatic isocyanides and carbodiimides (2,150 cm −1 ), and aryl isocyanides (2,135 cm −1 ) (Figure 2). Nitriles in the structure of evaporitic deposits would be mainly carbodiimide and isocyanide, and conjugated nitriles and aryl isocyanide would be rare in evaporitic deposits as suggested by a single sharp absorption at 2,150 cm −1 (Figure 2 and Figure S7 in Supporting Information S1). The relatively small absorption at 2,150 cm −1 to those at 1,600 cm −1 and 1,500-1,400 cm −1 for evaporitic deposits suggests it contains a lower amount of nitriles than the original tholin particles.

X-Ray Absorption Near-Edge Structure (XANES) Analysis
XANES analysis supports this view of fractionation of nitrogen-bearing aromatics from tholin ( Figure 3). XANES spectra of evaporitic deposits are distinct from those of original tholin particles and residual tholin aggregates ( Figure 3); the peak at 285.2 eV attributed to aromatic and/or olefin C=C bonds (Nuevo et al., 2022;Yabuta et al., 2014) is weak in evaporitic deposits (Figure 3). Considering the results of IR analysis (Figure 2), the strong XANES peaks at 286.5-288.0 eV for evaporitic deposits are likely attributable to nitrogen-bearing aromatics (Jimenez et al., 2000;Yabuta et al., 2014). This is because it has been shown previously (Jimenez et al., 2000) that additions of nitrogen to aromatic graphitic carbons result in a shift of XANES peaks from 285.2 eV (aromatic C=C) to 286-288 eV (aromatic C=N). The peak at 286.5 eV could also be attributed to aromatic ketones (De Gregorio et al., 2013), but an abundance of ketones in evaporitic deposits would have resulted in IR absorption at ∼1,700 cm −1 owing to C=O bonds, which was not observed (Figure 2).

Discussion
Our original Titan tholin was generated at ∼200 Pa from a N 2 -CH 4 gas mixture via cold plasma irradiation (see Section. 2.1). The IR spectrum of our original tholin (Figure 2) shows N-H, C-H, N=C=N, C≡N, C=C, and C=N bonds in the structure, whose characteristics are similar to those formed from N 2 -CH 4 gas mixtures at 26 and 67 Pa by Imanaka et al. (2004). Our results of XAFS analysis (Figure 3) for the original tholin also indicate that this contains aromatic and/or olefin C=C bonds, and nitrile and/or nitrogen-bearing aromatics. Based on comprehensive analyses of Titan tholin by Imanaka et al. (2004), including Raman spectroscopy and laser mass spectroscopy, the structure of Titan tholin formed by the previous study included clusters of nitrogen-bearing aromatics connected to polymer-like aliphatic/olefinic carbon and nitrogen chains (Imanaka et al., 2004).
Our IR and XANES results suggest that conjugated and olefinic hydrocarbon/nitrile chains of original tholin remain in residual tholin aggregates, whereas nitrogen-bearing aromatics tend to be fractionated with evaporitic deposits. A possible fractionation mechanism would involve clusters of nitrogen-bearing aromatics being detached from tholin particles through the dissolution of aliphatic carbon chains of the original structure, given the high predicted solubilities of aliphatic carbons in liquid CH 4 (e.g., Cordier et al., 2013;Cornet et al., 2015).
If low-molecular-weight molecules are the major components dissolved from tholin, these molecules would have been evaporated into gas phase upon evaporation of liquid CH 4 during the experiments. The lack of detection of gas species in the cold-trap during an increase in the interior temperature of the cryochamer after evacuation of liquid CH 4 ( Figure S8 in Supporting Information S1) supports the view that dissolved components are relatively high-molecular-weight molecules such as clusters of nitrogen-bearing aromatics, rather than volatile, low-molecular-weight hydrocarbons and nitriles. Alternatively, nitrogen-bearing aromatics in evaporitic deposits might have been derived from molecules adsorbed on tholin particles upon immersing into liquid CH 4 . Given the possibility that nitrogen-bearing aromatics may be precursor molecules of Titan tholin (e.g., He & Smith, 2014a, 2014b, these molecules could be adsorbed on tholin particles during the formation in our experiments and released into liquid CH 4 . In a previous study (McKay, 1996), tholin film formed by electrical discharge was immersed in liquid ethane (C 2 H 6 ) at 95-110 K for 15 min. Little dissolution (dissolution fraction <0.03%) was observed, with no residue apparent by visual inspection after evaporation of the C 2 H 6 . In our experiments, the dissolved tholin fraction may also be small, considering that residual tholin aggregates and original tholin particles display similar IR and XANES spectra after dissolution (Figures 2 and 3). Microscopic images indicate that the average area coverage by evaporitic deposits on the Au plate was a few percent, whereas that of residual tholin aggregates on the Al substrate was 40%-50%. Considering the thin and flat appearance of evaporitic deposits (Figures 1e and 1f), the volume ratio of evaporitic deposits to residual tholin aggregates (i.e., the dissolved tholin fraction) in our experiments might be of the order of 1%, and the dissolution fraction might be orders of magnitude higher than that achieved (McKay, 1996) using liquid C 2 H 6 . This difference in dissolution fraction may be caused by differences in dissolution time, and/or solvent among experiments. In particular, the short dissolution time in the previous study (McKay, 1996) may have prevented effective dissolution. Alternatively, the difference in chemical structure of tholins may have caused the difference in solubilty. Our original tholin particles before the experiments have C/H = 0.7 and C/N = 2.1, which are average elemental compositions among Titan tholins formed at low pressures with cold plasma irradiations ( Figure S9: see Supporting Information S1 for the method); however, they have a distinct composition compared with tholin formed at 10 5 Pa with discharge by McKay (1996) ( Figure  S9 in Supporting Information S1). The higher C/H and C/N ratios of tholin formed by McKay (1996) suggests less aliphatic chains in the chemical structure than our cold plasma tholins, which may have caused the low solubility of tholins formed by McKay (1996).
The ease of dissolution of tholin depends on its chemical structure. Previous studies have shown that different types of tholin are formed under different conditions of, for example, energy source, temperature, pressure, and composition of starting gas mixtures (e.g., Cable et al., 2012;Carrasco et al., 2009;Imanaka et al., 2004;Li et al., 2022;McKay, 1996;Sciamma-O'Brien et al., 2014Trainer et al., 2013). If organic aerosols on Titan comprise mainly nitrile polymers, their dissolution in liquid CH 4 may be limited. However, aromatics are considered seeds of complex organic aerosols in the upper atmosphere of Titan (Gautier et al., 2017;Lavvas et al., 2011;Waite et al., 2007) and in the mesosphere, and stratosphere photolyzed aromatics would form large molecules (Trainer et al., 2013), which could be incorporated into aerosol structures. Furthermore, heterogeneous H and CH 3 additions to aerosols in the stratosphere may also generate polymer-like aliphatic components in aerosol structures (Peng et al., 2018; Sekine, Imanaka,  et al., 2008). If Titan aerosols contain clusters of aromatics connected by polymer-like carbon and nitrogen chains, perhaps similar to our cold-plasma tholin, their interaction with liquid CH 4 would provide solutions of aromatics. In addition, nitrogen-bearing aromatics in evaporitic deposits might have been derived from desorption from tholin particles in our experiments. Given that aromatic compounds are also present in Titan's atmosphere (e.g., Delitsky & McKay, 2010;Lopez-Puertas et al., 2013;Waite et al., 2007), desorption of aromatics from aerosol particles might happen when aerosols interact with liquid CH 4 on Titan.
With photochemically produced C2-C6 compounds only, evaporites on Titan would be enriched in acetylene (Cordier et al., 2013). The previous study (MacKenzie & Barnes, 2016) tried to distinguish the different compositions of evaporites on Titan using spectral features at the 5-μm window. Their results suggest the presence of differences between the individual instances of evaporites; however, the spectral features are insufficient to distinguish what compounds are responsible (MacKenzie & Barnes, 2016). Together with our experimental results, Titan's evaporites may be composed of mixtures of materials from different sources, including photochemically-produced C 2 -C 6 compounds and aromatic compounds dissolved from organic aerosols. The difference in degree of mixing might have cause the differences between the instances of evaporites on Titan.
A possible mechanism for the aggregation of tholin in our experiments is that dissolved components precipitate on the surfaces of tholin particles during the evaporation of solution and connect (evaporitic cementation). Alternatively, Ostwald ripening could be an aggregation mechanism whereby smaller particles dissolve and larger particles grow owing to the precipitation of dissolved materials, forming necks among particles (Voorhees, 1985). Particle size would then change during aggregation, while remaining relatively unchanged in evaporitic cementation. Our observations indicate small particle size changes during dissolution (Figure 1 and Figure S10 in Supporting Information S1), suggesting that evaporitic cementation is the most likely aggregation mechanism.
On Titan, the mechanism responsible for the growth of organic aerosols on the surface to form dune particles is unclear, especially if dune particles originate from aerosols (Barnes et al., 2015). Organic aerosol in the troposphere is estimated to be of ∼1 μm size with aggregates of hundreds of spherical monomers (Tomasko et al., 2008), whereas a dune particle needs to be of several hundred μm size for it to saltate in wind to form dunes (Burr et al., 2014;Comola et al., 2022). Our observed aggregation of tholin particles during wet-dry cycling suggests the possibility of growth through aggregation of settled aerosol particles on Titan. Such aggregation could proceed through even short-term interaction with liquid CH 4 (6 hr in our experiments). Aerosols can be condensation nuclei for raindrops in the troposphere (Curtis et al., 2008;Yu et al., 2020), and dissolution of aerosol particles would proceed not only on the surface during seasonal rain but also in raindrops in clouds in the troposphere. Based on a modeled vertical profile of aerosol number density in Titan's troposphere (Tomasko & West, 2009), the number of aerosols at altitude below 30 km would be 2-4 particles/cm 3 for ∼1 μm-sized aerosol. Assuming the diameter of raindrop on Titan as ∼1 cm (Lorenz, 1995) and the altitude of CH 4 cloud formation as ∼15-20 km (Rannou et al., 2006), the number of aerosols that can be swept by one falling raindrop might be the order of 10 6 particles until reaching to the surface. If all swept particles are incorporated into the raindrop and coagulate upon drying similar to our experiments, then the size of one aggregate formed by drying of one raindrop would be around 10 2 μm. This size of particle is saltatable by wind to form dunes on Titan (Burr et al., 2014). Thereby, one cycle of falling and evaporation of raindrop could account for the growth of aerosols to dune particles.
The NASA Dragonfly mission will investigate the morphology and composition of dune particles in interdune fields Lorenz et al., 2021). The spacecraft will then move to multiple locations with different geological units, possibly including evaporites Lorenz et al., 2021). Given the fact that our tholin particles grow to aggregates after a wet-dry cycling (Figure 1), we predict that dune particles, at least partly, comprise large aggregates of organic aerosols on Titan. Our results of aromatic-rich evaporitic deposits (Figure 2) also suggest that evaporites on Titan may contain aromatics fractionated from organic aerosols into liquid CH 4 . Our results suggest that a geochemical process similar to or more effective than chemical weathering of rocks by water on Earth may be underway with liquid hydrocarbons on Titan, affecting its surface evolution (e.g., Malaska et al., 2020).

Conclusions
Based on the experimental results on dissolution of Titan tholin to liquid methane, we suggest a new view of chemical evolution of Titan's organic aerosols on the surface. Namely, our results suggest that aromatic compounds in organic aerosols would be rapidly fractionated in liquid methane at low temperatures (∼95 K). Remaining individual organic particles would coagulate to form large aggregates through cementation upon wetting and subsequent drying of liquid methane. This alteration of organic aerosols by wet-dry cycling may play an important role in the formation of evaporites and would be essential to explain the observed low-latitudal dunes with large particles on Titan. Our results provide an idea on interactions between organic aerosols and liquid methane for interpreting the surface evolution of Titan.