Molecular Evaluation of Deep Eutectic Solvents for Wet Algae Lipid Extraction and Biofuel Production Authors: Poonpat Dumnoenchanvanit, Young Jin Lee, Michaela Adolphus, Mehmet H. Can, Marcela J. Silvera Tafur, Phattiya Matrakul, Jenny Wang, Qin M. Qi Published: 2026 ABSTRACT Microalgae represent promising feedstocks for biofuel production, yet cell-wall disruption remains a critical bottleneck for scaling up lipid extraction. This study evaluates four cholinium-based ionic liquids (ILs) and deep eutectic solvents (DESs) as green, energy-efficient pretreatment agents to facilitate the lipid release from Chlorella vulgaris under wet conditions. A geranic acid-based DES (CG 1-2 DES) was identified as the optimal agent, demonstrating direct wet biomass extraction with substantial process energy savings at comparable overall lipid yields. To minimize the energetic costs of biomass drying and reduce CO2 emissions, lipid extraction was evaluated under wet conditions by using a one-pot method for DES recycling. FTIR spectroscopy and mesoscopic molecular dynamics simulations suggested the underlying mechanism: CG 1-2 DES disrupts cell wall lipids while maintaining specific miscibility to phase-separate from the product lipids, with water acting as a key process parameter that weakens hydrogen bonding and lowers mixture viscosity. These insights highlight key operational differences from conventional dry extraction and the previous use of CG 1-2 DES in skin permeation applications. Ultimately, utilizing IL/DES systems for wet extraction offers a sustainable, cost-effective pathway for biofuel processing where integrated experimental and computational screening can further accelerate material optimization. Introduction Microalgae are considered one of the most promising feedstocks for biofuels as a renewable energy source and for CO2 capture. − Compared to previous generations of biomass sources such as corn and wood, these organisms have a much higher energy density per land: they can grow in saline or sewage water, or in areas unsuitable for livestock and agriculture; they can generate up to 50% fatty acids from atmospheric CO2 due to high photosynthetic efficiency. Despite upstream successes in increasing the yield by refining microalgae strains with genetic engineering and synthetic biology, , the extraction of lipids from microalgae in downstream processing accounts for 30–50% of production cost. The expensive cost of lipid extraction is associated with the microalgal cellular structure: the cell wall is made up of various robust components such as cellulose polysaccharides, proteins, and lipids, each contributing to the strength and resistance to lipid extraction. Therefore, the disruption of cell walls plays a central role in microalgal lipid extraction and is critical to solving the bottleneck problem of the biofuel mass production. It is highly desirable to develop methods that are energy efficient, fast, low-cost, recyclable, and safe to humans and the environment. , Conventional lipid extraction technologies include physical means to disrupt cell walls as a pretreatment at high energetic demands, such as mechanical bead milling, autoclaving, steam treatments, and high-pressure homogenization. , Recently, ionic liquids (ILs) and deep eutectic solvents (DESs) are being developed as green chemical alternatives: with highly tunable chemical, physical, and biological properties, they can be prepared through simple synthesis and are capable of dissolving a wide range of cellular components for an efficient penetration through the cell walls; , thanks to their thermal stability and low volatility, there exists great recycling potential to reduce the energy and chemical usage as well as minimize the environmental impact; cholinium-based IL/DES stand out as low toxicity candidates, − which can be explained by their underlying connection to choline, an essential micronutrient. Despite various attractive features, the use of cholinium-based IL/DES for microalgae processing still faces various limitations: (1) dried algae powders are typically used for pretreatment, which significantly increases the energy cost compared to directly treating algae in their aqueous growth environment (up to 70% total energy cost); (2) physical methods are still used to facilitate the cell-wall disruption, hindering the energy reduction. − Switching from dry conditions to wet conditions to reduce the energy cost is not a straightforward process. One design principle used by various studies is to facilitate lipid transport by increasing the number of hydrogen bonds that can form between the IL/DES and cellulose, the most prevalent component of the cell wall. , However, this principle is violated under wet conditions, as the presence of water significantly disrupts the intermolecular hydrogen bonding network in DES and ILs. Further, simulations comparing a cholinium-based DES’ interaction energies with cellotriose (model cell wall cellulose) and tetrapeptide (model cell wall protein) reveal that the latter is more favored as a mechanism of cell-wall disruption. Changes in hydrogen bonding can also influence the physical properties of IL/DES: melting point, density, and viscosity decrease while electrical conductivity increases as water is added. After pretreatment, phase separation with the addition of nonpolar solvents could be utilized to conveniently collect lipids and IL/DES separately for recycling. However, the presence of DES, water, lipids, and nonpolar solvents could also lead to the formation of mesoscopic structures such as globular micelles and lipid vesicles, creating challenges to their separations. Thus, there exists an urgent need to reconsider design principles and the underlying mechanism of action to guide the discovery of cholinium-based IL/DES under energy-efficient operating conditions. The research of cholinium-based IL/DES in other applications could offer important insights into microalgal pretreatment. In pharmaceutics, as skin permeation enhancers, their material performance is often evaluated under wet conditions due to the ubiquitous presence of water in a living environment; the potency of skin permeation enhancement is related to the effective extraction of lipids from the top layer of skin cells, which enhances the transport of drugs across the skin barrier. Despite similarities to microalgae treatment, there exist important differences to consider for repurposing ILs/DESs, which have not yet been performed to the best of our knowledge. First, besides potent lipid extraction, lipids and ILs/DESs also need to be easily separated into two phases for the purification of lipids for biofuel production. Second, while high viscosity and gelling could be beneficial for topical formulations to stay on the skin for a sustained release of drugs, it is ideal to have low viscosity for the processing of algae, especially in a continuous manner for large-scale production. Nevertheless, various experimental and computational methods for the molecular evaluation of material performance may also be valuable for algae pretreatment, such as the Hansen solubility parameter to explain skin-CPE interactions based on dispersive, polar, and hydrogen bonding interactions. Here, we address the key green design principles of IL/DES for wet microalgae pretreatment based on molecular understandings gained from experiments and simulations. Using cholinium-based ILs/DESs previously studied for skin permeation enhancement but not yet for microalgae lipid extraction (oleate and geranate based), we elucidate important similarities and differences between their uses in two different settings. Given that these candidates already possess superior safety and low toxicity properties based on their previous applications, we focus on evaluating their lipid extraction efficiency under wet conditions with moderate heating (60 °C) and treatment time (3 h) to reduce energy cost as well as the material processability and recyclability. The removal of excess water during IL/DES recycling combines two drying steps in conventional methods into one to reduce the drying cost and CO2 emission during synthesis. Further, we provide molecular understandings of the material performance using attenuated-total-reflectance Fourier transform infrared (ATR-FTIR) and coarse-grained molecular dynamics simulations. The use of the latter technique, to the best of our knowledge, is the first of its kind in the application of microalgae lipid extraction. In contrast to popular simulation methods calculating interactions between a pair of individual molecules, ,− our simulations utilize model parameters related to cohesive energies and Hansen solubility parameters and offer more richer knowledge about phase separations and mesoscopic structures relevant to the evaluation of lipid extraction efficiency. Overall, our studies demonstrate that a potent and safe skin permeation enhancer can be repurposed as a microalgae pretreatment agent evaluated by its permeation enhancement efficiency, processability, and recyclability. Compared to the number of IL and DES candidates that have been studied for microalgal pretreatment, there exists a vast design space underexplored and thus a great promise of finding candidates with even better performance following the same principles and working mechanisms. Materials and Methods IL/DES Synthesis Choline bicarbonate (∼80% in H2O), geranic acid, oleic acid, phosphate-buffered saline (PBS), and deuterated DMSO were obtained from Sigma-Aldrich (St. Louis, MO, USA). The synthesis of IL/DES follows the procedure previously reported. The stoichiometric ratio between two ions was controlled by the amounts of salts and acids added during synthesis. For a matching ion ratio, the product is classified as an IL, and for a ratio of 1–2, the product is classified as DES. One equivalent of choline bicarbonate and one (IL) or two (DES) equivalents of geranic acid or oleic acid were combined in a 250 mL round-bottom flask. Methanol was added (1.7 mL per 1 mol choline bicarbonate) to improve fluidity, and the mixture was stirred until no more CO2 was released. The solvents were removed by first stirring at 60 °C for 2 h, followed by evaporation in a vacuum oven for 48 h at 60 °C. To characterize the formulation, dried and neat IL/DES were placed into an NMR tube containing a coaxial insert filled with chloroform-d1, and the 1H NMR spectrum was collected using a 400.09 MHz two-channel Bruker Avance400 AvanceIII HD Nanobay spectrometer (Billerica, MA, USA). The ion ratio of the final product was confirmed by NMR (Supporting Information). The stock IL/DES homogeneity was verified with NMR of multiple samples at various time points. Tetrakis (hydroxymethyl) phosphonium chloride was purchased from Sigma. Preparation of Microalgae Stock Chlorella vulgaris and its culturing supplies were purchased from Algae Research Supply (Vista, CA) and cultured following the manufacturer’s recommended protocol under normal sunlight and CO2 conditions in freshwater. After algae grew to a sufficient density indicated by their color, they were allowed to settle to the bottom of the culturing bottle. Liquid collected from the bottom was centrifuged at 2500 rpm for 5 min and 8 times to concentrate the algae as a wet paste for use in lipid extraction. To determine the percentage of dry mass, wet paste was dried on aluminum foil for at least 48 h to measure its change in mass. Rheological Characterizations A Discovery Hybrid Rheometer HR-3 and a 60 mm 1.004° cone–plate (Peltier plate Steel model 986191) were used to measure the viscosity of samples under different shear rates (flow curves). Each sample had a minimum volume of 0.99 mL. It was loaded with either a pipette for viscous liquids or a spatula for gel-like samples. Bubbles were avoided whenever possible. The plate was spun at a shear rate of 1.0 rad/s before lowering to a trim gap of 58.0 μm. The excess sample protruding from the sides of the plate was removed with a Kimwipe. The truncation geometry gap was then lowered and maintained at 29.0 μm. Before each experiment, the sample was preconditioned at 1.0 rad/s for 90.0 s. At least 10 min was used for temperature equilibration. A solvent trap was used to prevent evaporation. To obtain the fitted correlation of viscosity vs temperature, a Vogel–Fulcher–Tammann (VFT) fit was used: η = η∞ exp­(B/(T–T 0)) with T 0 being the glass transition temperature. To fit viscosity vs water content, double exponential fitting was used: η = a1e–b1x+ a2e–b2x . Errors were calculated from the covariance matrix in weighted least squares regression and the standard deviation of experimental data. ATR-FTIR Spectroscopy ATR-FTIR spectra were measured using a Thermo Fisher Nicolet iS50 Fourier transform infrared spectrometer (Waltham, MA, USA). Each measurement was for a 400–4000 cm–1 wavenumber range with a spatial resolution of 0.4822 cm–1 and averaged over 16 scans. The sample was loaded with a spatula (solids) or a pipet (liquid) onto a diamond ATR with a KBr beamsplitter. The baseline was corrected for each wavenumber region of interest. Lipid Extraction The method was adapted from published methods and summarized in Figure . , Wet algae were transferred into a 5 mL Eppendorf tube using a spatula. IL/DES were then added at the desired ratio. The Eppendorf tube was sealed with Parafilm, vortexed for at least 30 s, and incubated at 60 °C under continuous rotation for a controlled time duration. Following incubation, 500 μL of methanol and 250 μL of hexane were added per gram of wet algae to the mixture. The sample was vortexed for 30 s and centrifuged at 4500 rpm for 5 min to form a hexane phase, a methanol/water phase, and a solid residue phase. The hexane phase contained lipids to be purified as the final product; the methanol phase contained IL/DES to be purified for recycling. The methanol/water phase was washed with hexane three times. The combined hexane phase was then washed with water twice and dried using sodium sulfate to remove residual moisture for 24 h in an aluminum foil tin. Finally, the dried lipid extracts were weighed, and the lipid yield (Y) was determined gravimetrically relative to the dry algae mass using the following equation 1 Y=mlipidextractedmdryalgae=mlipidextractedmwetalgae×%drymass Schematic of lipid extraction from wet algae stock. The lipid yield using IL/DES ( Y IL / DES ) was compared to that from a modified Bligh and Dyer ( Y Bligh&Dyer ) method as a benchmark 2 relativelipidextractionefficiency(%)=YIL/DESYBligh&Dyer×100 To recycle the DES and reduce the overall energetic cost and CO2 emissions, the combined methanol/water phase was dried in a vacuum oven for 48 h at 60 °C and directly used for lipid extraction without further adjustment. To check the stoichiometric ratio and the residual amount of solvents, NMR was performed on the recycled sample using procedures identical to those for freshly synthesized samples. The solid residue phase was dried for at least 48 h before performing FTIR analysis. All lipid extraction experiments were performed in triplicate (n = 3). Data are reported as mean ± standard deviation. Simulations Dissipative particle dynamics (DPD) was employed to simulate the soft-core interactions between coarse-grained beads, as implemented in the open-source LAMMPS molecular dynamics simulator. Following the coarse-graining scheme illustrated in Figure S5, the average degree of coarse-graining was set to 4 for the ratio between heavy atoms and soft-repulsive beads, and the system number density ρ was fixed at 3. In brief, the equation of motion for each bead is governed by the following forces 3 midvidt=∑j≠iFijC+∑j≠iFijD+∑j≠iFijE+∑j≠iFijR This choice of the average degree of coarse-graining N m sets the self-repulsion parameter a ij in the conservative force F ij C between beads i and j, evaluated as follows 4 aii=κ−1Nm−12αρkBT,aij=(aii+3.27χij)kBT The dimensionless compressibility κ–1 and the equation of state coefficient α, with values of 15.9835 and 0.101, respectively, were used to match the isothermal compressibility of water. 5 FijC=−aii(1−rijrc)nij The Flory–Huggins χ-parameters for each interacting bead pair in F ij C map the degree of mixing determined by thermodynamics when k B T is the characteristic scale for thermal energy of the system. In this study, Monte Carlo-based fragment molecular orbital (FMO) calculations were performed to extract χ-parameters for all interacting pairs, and the results are summarized in Table S1. The χ-parameters between like pairs not shown in the table are set to 0. 6 FijD=−γ(1−rijrc)2(nij·vij)nij 7 FijR=ζ(1−rijrc)θijnij For the friction coefficient γ in the dissipative force F ij D, a value of 4.5 was adopted, while the fluctuation–dissipation theorem yielded a noise parameter ζ of 3.0 for Brownian motion. Here, v ij is the relative velocity between beads i and j, and θ ij is a Gaussian normal deviate that is independent at every time step. 8 FijE=qiqjΓ4π1−e−2βrij(1+2βrij+2β2rij2)rij3nij To incorporate the electrostatic interactions from charged ILs, electrostatic interactions F ij E were also included between charged pairs of ions, where a Slater-type charge smearing was applied with a real-space Ewald cutoff of 3.0σ and a smearing parameter β = 0.929σ–1. The relative permittivity was set to 78.5 for water and 2.85 for neat CG, while for mixtures of choline geranate and water, a weighted-average value was assigned according to their volume fractions. For molecules containing multiple beads interconnected as represented in Figure S5, a special type of bonding scheme with the force F ij S = −k s(r ij – r e) n ij was imparted only between those two interconnected beads. Specifically, interbead bonds were represented using a harmonic potential with a spring constant k s of 100 k B T/σ2, where σ denotes the characteristic length scale. The equilibrium bond length r e was set to 0.86σ, which is shorter than the cutoff radius r c = σ = 1. This is because the average distance between beads in a system with ρ = 3 is about 0.86σ and is derived from the two times of Wigner–Seitz radius (3/4πρ)1/3 = 0.43σ. At this spacing, each molecule does not exhibit intramolecular residual stress at equilibrium. Therefore, the intramolecular force acts primarily to hold certain beads together by a spring force, while an interbead conservative force still exists within the molecule. Finally, the governing equations were integrated using the conventional velocity-Verlet algorithm with a time step of 0.01. Results Optimization of Lipid Extraction Performance under Wet Conditions Using C. vulgaris, a commonly studied freshwater alga as an example, , we assessed various IL/DES systems for the pretreatment of microalgae. We synthesized two classes of cholinium-based ILs and DESs using oleic acid and geranic acid (Table ). The stoichiometric ratio of hydrogen bond donors and acceptors distinguishes IL (1:1) and DES (1:2). A 1:2 ratio of DES has been reported to demonstrate optimal lipid extraction efficiency and reduced viscosity for pharmaceutical applications. The four IL/DES formulations were specifically selected to evaluate the impact of the stoichiometric ratio of hydrogen bond donors/acceptors and the anion type on wet microalgae extraction performance. This focused matrix provides a mechanistic foundation to evaluate lipid extraction efficiency while establishing proof-of-concept feasibility for wet-processing workflows. A tetrakis-based IL, commercially available and previously shown to efficiently extract lipids from C. vulgaris, was also investigated with the adapted methodology. Instead of a complete drying process to obtain powders, the lipid extraction protocol (Figure ) operates under wet conditions by concentrating algae through sedimentation and centrifugation to effectively remove the majority of water at low energy cost. The wet algae are estimated to contain 10.5 ± 0.4% dry mass (Figure S6). The incubation with various pretreatment candidates uses a mild temperature (60 °C) without mechanical disruptions, which is preferred for large-scale production at low energy cost. The one-pot method using hexane and methanol to form two phases simultaneously can increase the lipid extraction efficiency under wet conditions compared to multistep methods to form various phases sequentially. Solid residue may be processed for other products that we did not consider in this study. 1List of IL and DES Investigated in This Study List of IL and DES Investigated in This Study % Lipid extraction is summarized in Figure for various ILs/DESs and processing conditions. We only achieved a quarter of the reported efficiency for tetrakis IL, which can be explained by the difference in operating conditions (lower temperature and shorter incubation duration). Further, previous reports used dried algae powder, which increases the energy cost and may also disrupt the cell wall. Therefore, the efficiency of lipid extraction may not be comparable. Next, we used four different IL/DESs at varying mass ratios in comparison to the wet algae mass and incubation periods to perform lipid extraction. Among the tested conditions, we found the optimal conditions to be a 1:0.8 wet algae:DES ratio and a 3 h incubation for DES cholinium geranate 1:2 (CG 1:2). In particular, this DES significantly outperforms other materials and generated 50% lipid extraction efficiency. Although we did not achieve a single-pass lipid extraction efficiency exceeding 90% as reported previously, our processing conditions did not use other physical means or dried biomass. Combining our processing conditions with recycling, a net energy saving of ∼100 MJ/kg at a comparable overall yield is estimated (Section ). This maximum efficiency is reached rapidly within 3 h, with no statistically significant increases as we further increased the incubation time. Thus, a 3 h incubation is beneficial for increasing the processing throughput at scale. Lowering the amount of DES reduced the lipid extraction efficiency, while increasing the amount of DES also reduced the lipid extraction efficiency. Such a non-monotonic trend agrees with a previous study using a DES consisting of choline chloride and formic acid under wet conditions assisted by hydrothermal temperatures above 100 °C, although no reasons were provided. We hypothesize that this trend could be related to viscosity limits in mass transfer to be investigated in the following section. % Lipid extraction investigated in this study for various IL and DES at different mass ratios of wet algae to IL/DES and incubation time. Recycling Potential and Sustainability Evaluation for Lipid Extraction Uses Despite the well-known low volatility of IL and DES, very few studies have evaluated their recyclability potential, especially for cholinium-based ones. The wet algae lipid extraction procedure listed above enabled us to collect CG 1-2, our best-performing candidate, in the polar phase (methanol/water) and separate it from the product lipids. Upon phase separation, the polar phase was evaporated to recycle the IL. From a sustainability perspective, the synthesis of fresh IL/DES produces CO2 and requires a drying step for water removal. Conventional lipid extraction methods also require drying algae feedstock. We save the energetic cost compared to existing methods by combining two drying steps into one: the excess water we kept throughout wet lipid extraction acts similarly to that added during synthesis to reduce the viscosity. Based on estimates using previously reported data, the energy cost saving is 100 MJ/kg dry mass to reach a comparable lipid yield using such wet conditions (Supporting Information). In addition, recycling IL/DES enables us to reduce CO2 emissions compared to repeated synthesis of fresh chemicals. We recovered 75.6 ± 13.3% DES by mass. We compared the NMR of recycled DES (Figure S7) and the freshly synthesized ones (Figure S4). Despite showing similarities in major peaks, we noticed a drop in the stoichiometric ratio by 17 ± 4 %, indicating a loss of geranic acid. This result is expected because geranic acid is more volatile compared to the CG 1-1 IL. We expect the presence of other microalgae components in our recycled IL, which do not precipitate or dissolve in the hexane phase. Dissolution of proteins is unlikely due to FTIR analysis discussed earlier. However, dissolution of chlorophyll, a major pigment of microalgae, is observed as we noticed a green color in recycled DES. , According to previous studies, extracted chlorophyll occupies 1% dry mass of C. vulgaris. , Thus, we expect a 0.2% chlorophyll accumulation in the recycled IL, which is negligible for at least five passes. Despite these noticeable but trace amounts of changes, we repeated the experiment using recycled DES under the optimal processing condition identified from Figure (1:0.8 wet algae-to-DES ratio and 3 h incubation). The recycled DES was used in an unadjusted state, allowing for an evaluation of the system’s raw operational robustness. A comparison of the % lipid extraction indicates no significant reductions (Figure ). Our results on recyclability potential are consistent with another study on phosphonium-based ILs that despite the presence of impurities in the recycled sample, lipid extraction efficiency is not significantly impacted, thanks to their superior thermal stability and low volatility. Further, although an equivalent loading ratio of 12.5% is slightly lower than 20% previously reported, the recyclability justifies the use of more CG 1-2 DES in each pass. While this single-cycle recovery serves as a proof-of-concept demonstrating DES stability without immediate chemical degradation, evaluating long-term performance across multicycle operations (5–10 cycles) remains an important focus for future industrial commercialization. A comparison of % lipid extraction between fresh DES and recycled DES at matching conditions of a 1:0.8 algae-to-DES ratio and a 3 h incubation. Stoichiometric ratios of the DES are also shown. Ease of Processability in a Continuous Setting From a scale-up perspective, lowering the solution viscosity of the IL/DES mixture becomes advantageous for energy savings during 3 h of incubation under continuous stirring. Therefore, to examine how the viscosity of the IL/DES changes across a wide range of shear conditions relevant to continuous processing, we first conducted rheological experiments to obtain flow curves and characterize the dynamic behavior of the IL/DES system. Many IL/DESs are known to exhibit pronounced non-Newtonian behavior in flow curves due to liquid-phase ordering that gives rise to their mesoscale structures. This behavior is also observed in our COL 1-1 and COL 1-2 systems, which display characteristic shear thinning responses. When the formation of mesoscale structures is present, high shear rates are required to reduce the viscosity, which, in turn, is associated with an increase in the energy cost of processing. In contrast, the CG 1-1 and CG 1-2 systems behave as Newtonian fluids, showing constant viscosity across all shear rates under the flow curve (Figure a). The viscosity of the CG systems appears higher than that of cholinium hydroxide and acetate ILs, comparable to that of benzoate, but lower than that of citrate. Rheological measurements. (a) Flow curves of CG 1-2 at various water concentrations at 25 °C. (b) Viscosity vs temperature for CG 1-2 and CG 1-1 and the fitted lines using the Vogel–Fulcher–Tammann (VFT) equation. (c) A comparison of viscosity vs water concentration between experimental data and weighted averaging between the viscosities of pure water and CG 1-2. In addition to room temperature measurements similar to previous studies, we also measured its viscosity under varying temperatures to be relevant to the mildly heated processing conditions during lipid extraction. The CG 1-2 at increasing temperature remains a Newtonian fluid with decreasing viscosity, which can be described by a Vogel–Fulcher–Tammann fitting (Figure b). The viscosity at our processing temperature of 60 °C is 18.5% of the value at room temperature, thus providing the benefit of facilitating the rate of mass transfer through mixing for enhanced extraction efficiency. On the other hand, CG 1-2 exhibits a lower viscosity with a more rapid thermal decay than CG 1-1, which is consistent with CG 1-2’s higher lipid extraction efficiency in Figure . This behavior can be attributed to the dense network of ionic bonds characteristic of pure ionic liquids. The addition of geranic acid in the 1:2 DES formulation disrupts these strong cation–anion interactions by introducing hydrogen bonds between the unprotonated geranic acid, the hydroxy groups, and the quaternary ammonium cations. This interpretation is consistent with findings by Zakrewski et al., where a 1:2 choline-to-carboxylic acid ratio was shown to yield the lowest glass transition temperature via differential scanning calorimetry, directly correlating with enhanced fluidity. Given the importance of water in our application, we also investigated the role of water on CG 1-2 viscosity (Figure a,c). Overall, the mixture still displays Newtonian behavior. As expected, the mixture viscosity decreases with water content ( x ), which explains the use of aqueous IL/DES in previous studies to improve its processability. Further, our measurements can be fitted using a double exponential η = 0.426e–0.383x + 0.312e–0.042x in contrast to a single rate of decay. At low water concentrations, the viscosity decreases rapidly with the addition of trace amounts of water, which has been observed for other ILs. The slower decay dominates at higher water content and is consistent with the weighted average between the pure viscosities of water and DES. Our findings provide important mechanistic insights into various intermolecular forces governing the material properties and support two molecular mechanisms for the viscosity decay: a change in hydrogen bonding at low water concentrations as well as a simple dilution effect. Water has been observed to disrupt IL/DES networks through intermolecular hydrogen bonds and lead to changes in melting point and viscosity. Similar to observations above for neat IL/DES, the viscosity decay with water concentration appears to be steeper for the CG 1-2 than for CG 1-1 (Table S2). The strong ionic bond of the IL could generate intermolecular forces that resist the influence of water more than just the hydrogen bonding of the DES. From the interest of lipid extraction, even a trace amount of water present in wet algae decreases the mixture viscosity and enables the slurry of algae and DES to be mixed more efficiently compared to dry algae. The transition between the hydrogen-bonding-dominated regime and the dilution-dominated regime depends on the IL/DES studied. Here, for CG 1-2, the wet algae processing conditions are expected to be in the dilution-dominated regime. However, for a phosphonium-based IL, the wet processing conditions were reported to be dominated by hydrogen bonding. The presence of two regimes is consistent with a trade-off between the dilution effect to reduce lipid extraction efficiency and the viscosity reduction benefit to enhance the lipid extraction efficiency. After pretreatment, water changes the polarity of IL/DES and can be used to tune the separation performance during lipid extraction. Thus, water serves as an important design parameter in tuning IL/DES performance and explains the non-monotonic trend of lipid extraction efficiency vs IL concentration in Figure . Spectroscopic Analysis for Various Components To further understand the role of CG 1-2 in facilitating microalgal lipid extraction, we performed ATR-FTIR on samples obtained at various stages of the lipid extraction process and identified various wavenumber regions of interest (Figures a and S8). First, we compared the dried algae sample and the dried solid residue after lipid extraction for the alkyl vibration region (2800–3000 cm–1). As seen in Figure a, we observed a reduction in peak areas, indicating a reduction of lipid content in the algae samples due to the use of ILs, providing further proof of the success of the lipid extraction process. This observation is consistent with those findings reported for skin permeation enhancers with similar lipid extraction mechanisms and microalgae lipid extraction assisted by DES and hydrothermal treatment. On the other hand, we did not observe a significant shift of the broad peak between 3000 and 3600 cm–1 associated with the O–H functional groups. This peak shift change was related to a reconstruction of the hydrogen bonding network in previous studies. Further, the peak intensity in the region between 1500 and 1700 cm–1 (characteristic bonds for the protein chain) did not decrease as in previous studies using other DES. However, this finding is not surprising because CG 1-2 was previously found to stabilize the proteins to reduce skin irritation. Finally, we did not observe a reduction of the peak around 1050 cm–1 due to the stretching of polysaccharides, suggesting no significant depolymerization of carbohydrates. Analysis of the region between 1139 and 1190 cm–1 assigned to C–O–C asymmetric stretching vibration at the β-glucosidic linkage of the cellulose chain suggests no significant peak shifts or peak changes (Figure S9). Thus, it is unlikely that water hydrates cellulose to weaken DES-cellulose interactions. Overall, our examination of the solid samples implies that the DES solubilizes more lipids compared to cellulose or proteins, as found in other IL/DES systems for microalgae lipid extraction. FTIR spectra for various liquid and solid samples obtained from the lipid extraction process using CG 1-2. (a) A comparison of dry algae stock and dried solid residue after lipid extraction, with the close-up of two key wavenumber regions shown below. (b) A comparison of various liquid samples for the 1500–1600 cm–1 region. (c) A comparison of various liquid samples for the 1180–1280 cm–1 region. Next, we examined liquid samples, including pure DES, methanol, hexane, and mixtures of hexane and methanol phases after the 3 h lipid extraction process. It is worth noting that some characteristic CO stretching peaks (e.g., 1650–1750 cm–1) cannot be used to distinguish product lipids and DES (Figure S8). We identified a pronounced peak in the neat DES sample at 1550 cm–1 that is not present in pure hexane and methanol (Figure S8). This peak is related to the asymmetric stretching of the carboxylate group in the geranate ion and geranic acid. Examining the corresponding wavenumber for the hexane and methanol phases after 3 h extraction, we found a weak presence of such a peak in the methanol phase, where we expect DESs to be present (Figure b). We did not find the presence of such a peak in the hexane phase, indicating that DES does not exist at a detectable concentration. Thus, we prove the recycling potential of DES from the methanol phase and the purity of lipid products obtained from the hexane phase. In addition, we noted the emergence of a new peak near 1250 cm–1 in the hexane phase after 3 h extraction but absent from the methanol phase (Figure c). This peak is related to the −C–N–C– functional group seen in cholinium ions, amino acids, or aminolipids. Although CG 1-2 also presents a strong peak in this region, our earlier findings about the peak near 1550 cm–1 can help us to rule out the possibility of CG 1-2 being extracted into the hexane phase. Thus, we attribute this peak to the aminolipid product extracted into the hexane phase. Overall, the FTIR analysis provides further validation of the effectiveness and working principles of lipid extraction using DES, followed by hexane and methanol. Simulations To examine the thermodynamic role of CG 1-2 in microalgal lipid extraction in greater detail, we employed coarse-grained numerical simulations using dissipative particle dynamics (DPD). Compared to all-atom molecular dynamics simulations, DPD is much more efficient for investigating larger time and length scales relevant to lipid extraction, while still preserving coarse-grained chemical fidelity through the use of Flory–Huggins χ-parameters derived from the cohesive energies of polar/nonpolar and ionic/nonionic fragments (Table S1). In addition, its model parameters are inherently related to density functional theory (DFT) calculations and Hansen solubility parameters while offering richer details about mesoscopic structures influencing phase separation beyond miscibility predictions generated by COSMO models. ,− , Figure a shows the equilibrium mesoscopic structure of neat CG 1-2 obtained from the simulation, revealing a homogeneous phase without any noticeable mesoscopic structures indicative of ion aggregations or ordering. These observations are consistent with the Newtonian behavior measured experimentally in Figure a. In contrast, the COL 1-1 and 1-2 show noticeable stacked layers of alternating hydrocarbon chains and ionized head groups (Figure S10), which correlate with the gel-like properties seen experimentally. Such behavior can be attributed to the longer and less saturated carbon chains in oleic acid and the oleate ion, despite similar hydrogen-bonding interactions. Next, Figure b shows CG 1-2 mixed with varying volume fractions of water. CG 1-2 remains well dispersed across all water fractions, consistent with the characteristics of a Newtonian fluid described earlier. Simulation snapshots of (a) neat CAGE 1-2 and (b) mixtures of CAGE 1-2 with water at volume fractions of 0.3, 0.6, and 0.9 (from left to right). (c) Initial configuration of a DOPC lipid bilayer in which the upper leaflet is in contact with a 0.5 volume fraction CG 1-2/water mixture and the lower leaflet with pure water (left), and the configuration after 200τ (right). (d) Phase-separated structures obtained after 200τ of simulation for a system with a 0.4:0.4:0.1:0.1 volume ratio of hexane:water:CG 1-2:triacylglycerol (left), together with the corresponding profiles of reduced number density as a function of depth normal to the solvent interface (right). The color-coded bead representations for the coarse-grained chemical groups are shown in the bottom panel. After examining various equilibrium structures formed by CG 1-2 mixtures, we next investigated the potential of CG 1-2 to disrupt the cell wall and subsequently penetrate into it. Given the findings from FTIR studies suggesting that CG 1-2 primarily interacts with lipids in contrast to proteins and cellulose, we focus on modeling the lipid component of the cell wall using a self-assembled bilayer of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), shown in Figure c. When applied to the upper region of the bilayer, CG 1-2 interacts with DOPC due to the amphiphilic properties they share. Both cholinium and the carboxylate group of geranate have negative χ-parameters with the hydrophilic head groups of the bilayer, which promotes mixing, and the alkyl chain of geranate stabilizes the hydrophobic tails of DOPC that contain double bonds. These combined effects enable the complete infiltration of CG 1-2 into the cell wall and the subsequent cell membrane, thereby facilitating the subsequent extraction of lipids. As a final example, we investigate the last stage of spontaneous phase separation by modeling the thermodynamically driven demixing of a well-dispersed mixture of triacylglycerol (TAG), CG 1-2, water, and hexane. TAG is known as a major type of storage lipid located inside the microalgae that can be converted to biofuels. The simulation results shown in Figure c indicate that hexane and water rapidly form two distinct phases to minimize the Gibbs free energy of mixing, as reflected in the χ-parameter. TAG selectively dissolves into the hexane phase while CG 1-2 predominantly dissolves in water. Although no individual lipid molecules were detected in the bulk water phase, weak miscibility was observed at the interface, as evidenced by the reduced number density distribution shown in Figure d. We represent this phenomenon in more detail using the reduced number density profiles (Figure b). The reduced number density is defined as the volume fraction of a species within a finite slab located at depth r, normalized by the total species’ volume of that species, which is 0.4 for water and hexane and 0.1 for TAG and CG 1-2. Importantly, the TAG concentration at the water/hexane interface was approximately 1.45 times higher than in the bulk hexane phase. At the molecular level, the three alkyl chains of TAG drive its partitioning into the hexane phase while its three hydroxyl functional groups impart weak amphiphilicity and lead to accumulation at the interface. Likewise, CG 1-2 was also enriched at the water/hexane interface. Due to its higher amphiphilicity, its enrichment at the interphase was even more pronounced. Further, a small percentage of CG 1-2 dissolved into the hexane phase, which could potentially influence its recycling efficiency after extraction. It is worth noticing that despite the weak amphiphilicity of CG 1-2, we do not expect it to act as a surfactant, as demonstrated by our simulations. The presence of surfactants has also been shown to improve lipid extraction yields by disturbing the phospholipid bilayers, which can replace the use of a polar phase (methanol and water in our case). Such mechanisms may create issues for the recycling of IL/DESs, which were not considered in previous studies. Discussion Leveraging the broad design space of ILs/DESs holds high promise for developing sustainable agents to enhance energy-efficient microalgal lipid extraction. Our report on the novel use of CG 1-2 DES in microalgae lipid extraction brings new perspectives to several common notions based on conventional dry conditions. First, we demonstrate that the interaction between DES and the small fraction of lipids on the cell wall of C. vulgaris creates sufficient disruptions to facilitate the release of storage lipids. Unlike the popular approach of designing IL/DES for cellulose dissolution, our CG 1-2 does not have multiple electron-donating groups and smaller space-site resistance groups on the anion. Complementing the FTIR and simulation evidence presented earlier, a Hansen solubility parameter (HSP) analysis further substantiates this mechanism. Based on literature precedents, we expect the Hansen solubility parameters of CG 1-2 to resemble those of CG 1-1 (δD = 16.6, δP = 16.8, δH = 11.2) and geranic acid (δD = 16.2, δP = 4.2, δH = 9). The coexistence of ionic and neutral domains in ILs/DESs explains their wide range of solvency, which is reflected in the spacing between two domains on the HSP map. However, compared to TAG (δD = 16, δP = 4.8, δH = 3.8), cellulose (δD = 8.3, δP = 8, δH = 13.9), and proteins (δD = 15.85, δP = 12.95, δH = 15.17), there is no obvious advantage for CG 1-2 to have close miscibility with one cellular component to break the cell wall. Thus, our findings suggest a shift of research focus from cellulose dissolution to a more generalized breakdown of the transport barrier. Next, we demonstrate that hydrogen bonding is not the sole factor affecting the extraction performance under wet conditions. Oleate and geranate form similar degrees of hydrogen bonding but show distinct lipid extraction performance. We introduce rheological measurements that are rarely used for this application to elucidate the critical transition between hydrogen-bonding-dominated and dilution-dominated regimes. Although shear thinning behaviors in ILs are often associated with hydrogen bonding, we did not observe this behavior for the neat CG 1:2 system during viscosity measurements. Another important insight is that an ideal IL/DES should not only interact with lipids but also have distinct properties capable of separating into distinct phases when polar and nonpolar solvents are applied. From this perspective, amphiphilic ILs are less ideal, and the balance of various types of interactions must be carefully considered. We deploy coarse-grained molecular dynamics simulations that can complement experimental approaches to facilitate screening. Compared to other widely used simulation methods that focus on isolated pairwise molecular interactions, our multicomponent simulation framework provides deeper, mesoscopic insights into macroscopic phase-separation behaviors within complex mixtures. Further, we calculate the overall energetic cost with additional considerations of processability and recyclability for large-scale production. The mild heating conditions employed herein significantly reduce viscosity to facilitate mass transfer, thereby enhancing lipid extraction efficiency at a low energy penalty. ,, Despite high single-pass lipid yields optimized under dry conditions, the operational efficacy of conventional extraction agents fundamentally degrades in the presence of water. Conversely, the proposed CG 1-2 DES workflow extracts lipids directly from wet biomass (50% efficiency) while facilitating simultaneous solvent and product recoverya process loop rarely realized in traditional microalgae workflows. By consolidating energy-intensive thermal drying steps, the net process energy consumption is drastically minimized according to our energy cost estimate. Rheological profiles further confirm that continuous fluid transport can proceed with low pumping costs. While this study establishes a foundational framework, several avenues remain for optimizing the extraction protocol. First, evaluating variations among diverse microalgae species will be essential to ensure broad protocol adaptability. Beyond optimizing ionic structures, the stoichiometric ratio of hydrogen bond donors to acceptors represents a critical tuning parameter; refining this ratio between 1:1 and 1:2 could mitigate interfacial ion accumulation, thereby enhancing solvent recyclability. While this study focuses on evaluating gravimetric extraction efficiency and solvent interactions, comprehensive lipid compositional profilingincluding GC-MS analysis of fatty acid and FAME distributionsremains an important direction for future investigation to fully evaluate downstream biodiesel quality and maximize overall process economics. Finally, future iterations of this process should replace conventional partitioning solvents (hexane and methanol) in the current protocol with bio-derived green alternatives (e.g., ethyl acetate or limonene) to further maximize the net sustainability of the extraction loop. The simulation framework we established can be systematically expanded to represent more complex systems relevant to these future studies. Conclusions We examined the use of 4 cholinium-based IL/DESs to facilitate microalgal lipid extraction under energy-efficient wet conditions. CG 1-2 DES is evaluated as the optimal agent in terms of the lipid extraction efficiency, processability, and recyclability. FTIR and mesoscopic simulations were used to elucidate the working mechanism, which exhibits similarities and differences compared to the previous application of these ILs/DESs as skin permeation enhancers: CG 1-2 DES disturbs cell wall lipids while retaining distinct miscibility to phase-separate with the product lipids for its recycling. Under wet conditions, water significantly reduces hydrogen bonding and viscosity for the continuous processing. Our findings renew the design principles of microalgae lipid extraction under wet conditions, which have great promise for mass production at reduced energetic and economic costs. Supplementary Material