Liquid–liquid phase separation and self-assembly of hexanoic acid and ethyl hexanoate in ethanol–water systems: a model for aged colloidal Baijiu

Abstract

Liquid–liquid phase separation (LLPS) in the pre-Ouzo region governs the early-stage aggregation of amphiphiles, dictating the non-equilibrium evolution of soft materials. However, classical turbidimetry, which defines phase boundaries by macroscopic cloudiness, is blind to this metastable regime. Here, we introduce an integrated fluorescence–microscopy ternary phase mapping approach that directly probes the pre-Ouzo region in a canonical ethanol–water–amphiphile system. This method reveals critical aggregation concentrations that are 2–3 orders of magnitude below the binodal—a regime inaccessible to turbidimetry. Applying this diagram to naturally aged Baijiu (1–20 years), a dynamically evolving colloidal system, uncovers a pronounced aging-enhanced solubilization: the dissolution ratios of key amphiphiles (hexanoic acid and ethyl hexanoate) exceed 96%, with dissolved concentrations far surpassing static equilibrium predictions. Mechanistic investigations show that this phenomenon arises from the synergistic restructuring of ethanol–water hydrogen-bond networks and the expansion of hydrophobic microdomains. Our work not only provides a high-resolution tool for mapping non-equilibrium phase behavior but also establishes a direct link between slow microstructural evolution and the emergence of kinetically stabilized, supersaturated states in complex fluids.

---

1. Introduction

Ternary phase diagrams of solvent–amphiphile systems are central to mapping colloidal stability, self-assembly, and release kinetics. Their construction traditionally relies on turbidimetric titration, in which macroscopic cloudiness marks the phase-separation boundary.^1–3 However, this approach lacks the sensitivity to detect nanoscale or submicron liquid–liquid phase separation (LLPS) that may occur in macroscopically transparent systems.^4–6 Furthermore, in complex systems such as polymer blends, turbidimetry may fail to distinguish between different phase-separation pathways (e.g., polymer–solvent vs. polymer–polymer LLPS).⁷ Consequently, turbidity-based thresholds often overestimate thermodynamic solubility and fail to capture the non-equilibrium pathways that dictate phenomena such as gelation and flavor release.^8,9 To overcome this limitation, we herein integrate steady-state fluorescence spectroscopy with wide-field optical microscopy, a correlative approach that provides both molecular-scale sensitivity and direct microstructural visualization. This combined methodology enables the construction of a ternary phase diagram that probes the pre-Ouzo region—a metastable regime inaccessible to classical turbidimetry, which detects only macroscopic phase separation at the binodal.

This study is fundamentally rooted in soft matter physics, focusing on the ethanol–water–amphiphile system—a canonical model for exploring the interplay between liquid–liquid phase separation (LLPS) and molecular self-assembly. The phase behavior in such systems is governed by a delicate balance of interactions, highly sensitive to solvent composition, amphiphile structure, and environmental conditions. Ethanol, as a co-solvent, uniquely modulates the solvation environment and influences amphiphile aggregation via cononsolvency. For elastin-like polypeptides and other polymers, the addition of ethanol within a low concentration range tends to promote insolubility and aggregation, while high ethanol concentrations can restore solubility and suppress aggregation.¹⁰ The length of the alcohol alkyl chain and the structure of its hydration shell further dictate its efficacy in disrupting or promoting self-assembly.¹¹ Within this framework, amphiphiles can organize into a rich polymorphism of soft structures—including micelles, vesicles, fibers, and liquid crystalline phases—whose morphology is a direct outcome of molecular design and environmental cues.¹² Critically, LLPS and ordered self-assembly are not isolated events but often exist on a spectrum from disorder to order.¹³ Beyond fundamental science, the non-equilibrium phase behavior in ethanol–water–amphiphile systems underpins significant applied phenomena, such as the maturation of flavors in aged spirits.¹⁴

A high-resolution phase diagram is particularly valuable for studying complex aging systems such as traditional Baijiu, which itself represents a quintessential and dynamically evolving soft material. The aging of Baijiu is a prototypical non-equilibrium colloidal aging process, sharing mechanistic parallels with the aging of gels, emulsions, and micellar systems. It involves the slow evolution of self-assembled structures formed by amphiphilic flavor compounds, exhibiting characteristic soft matter phenomena such as Ostwald ripening (evidenced by changes in particle size, count, and zeta potential), structural relaxation, and enhanced viscoelasticity over time.^4,5 Specifically, LLPS has been identified as a key driver, initiating the formation of micrometer-sized droplets in young Baijiu that evolve into core–shell structures over decades of aging, directly influencing flavor stability and release.^14,15 This evolution mirrors aging dynamics observed in other soft materials, from particle aggregation in clay suspensions¹⁶ to stretched-exponential relaxation in metallic glasses.¹⁵ Thus, deciphering the microstructure dynamics in Baijiu not only addresses a specific technological question but also provides a unique temporal window into universal aging mechanisms in complex fluids. It is important to note that flavor compounds collectively constitute less than 2% of Baijiu's composition.¹⁷ Among these, ethyl hexanoate is recognized as one of the most abundant flavor compounds in strong-flavor Baijiu, with reported concentrations ranging from 1000 to 6600 mg L⁻¹, and is often identified as the predominant aroma compound.^18,19 Hexanoic acid, as the precursor of ethyl hexanoate, plays a central role in flavor formation; however, its absolute concentration is generally lower than that of its ester counterpart.²⁰

Focusing on two key amphiphilic flavor compounds in strong-flavor Baijiu—hexanoic acid and ethyl hexanoate—we first establish a high-sensitivity phase diagram for ethanol–water–amphiphile systems. We then apply this diagram to naturally aged Baijiu samples to reveal how their dissolution behavior deviates from the predictions of static phase equilibrium, exhibiting a pronounced aging-enhanced solubilization beyond conventional saturation limits. Finally, using steady-state fluorescence and Nile red probing, we elucidate the underlying molecular mechanism: the continuous restructuring of ethanol–water hydrogen-bond networks coupled with the expansion of hydrophobic microdomains. Therefore, this study delivers insights on two fronts: it deciphers the key microstructural driver for flavor maturation in aged Baijiu, and by doing so, provides a perspective on the dynamic microstructure–property relationships that govern non-equilibrium states in soft matter systems such as colloidal dispersions and emulsions (Fig. 1).

Fig. 1 A schematic diagram showing the phase transition mechanism of hexanoic acid and ethyl hexanoate in strong-flavor Baijiu during cellaring.

2. Materials and methods

2.1 Chemicals and materials

Hexanoic acid (analytical grade, ≥99.0%) was purchased from Aladdin Biochemical Technology Co., Ltd (Shanghai, China). Ethyl hexanoate (analytical grade, ≥99.5%) and Nile red (≥95%) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Absolute ethanol (analytical grade, ≥99.7%), n-amyl acetate, and methanol (HPLC grade) were supplied by Sinopharm Chemical Reagent Co., Ltd (Beijing, China). Ultrapure water (resistivity ≥ 18.2 MΩ cm) was prepared using a Milli-Q® system (Merck Millipore). Strong-flavor Baijiu (SFB) samples, serving as model colloidal systems with varying aging histories, were obtained from Huangshantou Group Co., Ltd (Hubei, China). Samples were collected from pottery jars after 1, 4, 7, 10, 13, and 20 years of natural aging. The alcohol content of all SFB samples was 69% (v/v).

2.2 Experimental methods

2.2.1 Turbidimetric titration. Solubility determination via turbidimetric titration: phase separation was ascertained by monitoring the changes in macroscopic solution clarity. Precisely 4.0 g of the target solute was dissolved in 0.5 mL of ultrapure water under magnetic stirring to yield an initial homogeneous solution. Anhydrous ethanol was added dropwise into the system using an acid burette until visually discernible turbidity appeared, with the cumulative volume of ethanol added meticulously recorded. While maintaining continuous stirring, ultrapure water was incrementally added to the turbid system until clarity was restored, and the cumulative water volume added was recorded. This operational cycle was iterated twelve times, yielding twelve discrete data points.

2.2.2 Construction of high-sensitivity phase diagrams.
2.2.2.1 Optical microscopy. To detect submicron liquid–liquid phase separation, 15 μL aliquots of each sample were pipetted onto pre-cleaned glass slides, covered with standardized coverslips, and immediately observed under an optical microscope (Olympus BX53). Images were captured under consistent exposure conditions (100 ms). The acquired micrographs were processed using ImageJ (v1.53k, NIH), and quantitative data were analyzed using OriginPro (v2023b) via nonlinear regression.
2.2.2.2 Steady-state fluorescence spectroscopy. The molecular aggregation transitions of hexanoic acid and ethyl hexanoate were examined by steady-state fluorescence spectroscopy across a range of concentrations in 55%, 65%, and 75% (v/v) aqueous ethanol. Emission spectra (290–450 nm) were recorded at 25 °C on an F-7000 spectrophotometer (Hitachi). To ensure that the measured fluorescence signals primarily originate from the solute molecules (hexanoic acid and ethyl hexanoate), the following control experiments were performed: (i) the fluorescence spectrum of the pure ethanol–water mixture at each studied composition (55%, 65%, and 75% v/v ethanol) was recorded under identical conditions and subtracted as background; (ii) the concentration-dependent fluorescence of the solutes was measured at 275 nm excitation, where the solvent background was found to be negligible compared to the solute signal at concentrations above the critical threshold. All fluorescence measurements were conducted in sealed quartz cuvettes within 30 min of sample preparation at 25 °C (triplicate, RSD < 5%).

2.2.3 Quantitative analysis of dissolved and supersaturated phases in aged Baijiu.
2.2.3.1 Phase separation by dissolution and sample pretreatment. Aged Baijiu samples were subjected to centrifugation (9000 rpm, 4 °C, 5 h) to separate the supernatant. Anhydrous sodium sulfate was added to the supernatant under vigorous mixing. After 30 minutes of static dehydration, residual water was removed to prevent peak distortion in GC–MS analysis. The mixture was then filtered through a 0.22 µm organic membrane, and 1 mL of the filtrate was collected for subsequent analysis.
2.2.3.2 Quantitative analysis by GC–MS. Quantification was performed using an internal standard method with n-amyl acetate. The internal standard was added to the samples at a 1 : 10 ratio (standard-to-analyte). Hexanoic acid and ethyl hexanoate in aged Baijiu were quantified via internal standard calibration. The resulting data were used to construct solubility curves as a function of aging time within the high-sensitivity phase diagrams.

2.2.4 Microstructural dynamics characterization.
2.2.4.1 Steady-state fluorescence of ethanol–water clusters. To investigate the ethanol cluster fluorescence, emission spectra (280–450 nm) were recorded from a series of strong-flavor Baijiu samples (aged 1–20 years) using an F-7000 spectrophotometer (Hitachi) with an excitation wavelength of 240 nm.
2.2.4.2 Nile red fluorescence probing of hydrophobic microdomains. Fluorescence measurements were performed on an F-7000 spectrophotometer (Hitachi). Nile red was dissolved in chromatographic-grade ethanol to prepare a 0.03 mmol L⁻¹ stock solution. For analysis, an aliquot of this stock solution was added to each 5 mL Baijiu sample to achieve a final probe concentration of 60 nmol L⁻¹. After magnetic stirring for 10 min, the mixtures were equilibrated in the dark for 30 min. Emission spectra were then recorded using an excitation wavelength of 552 nm, monitoring changes in fluorescence intensity. All measurements were conducted in triplicate, and the average values were used for data analysis.

3. Results and discussion

3.1 Construction and validation of a high-sensitivity ternary phase diagram

Turbidimetric titration is a classical method for determining binodal curves in ternary systems, as demonstrated by Mohsen-Nia et al.,²¹ who used cloud point measurements to identify the phase transition from homogeneous to heterogeneous mixtures. However, as demonstrated by Iglicki et al.,²² the cloudiness curve does not necessarily coincide with the binodal curve. Compositions located between the binodal and the cloudiness curve can remain visually transparent while residing within the metastable Ouzo domain, where liquid–liquid phase separation (LLPS) has already initiated at the nanoscale. Thus, turbidimetry captures the onset of macroscopic turbidity, not the thermodynamic binodal itself. This distinction is critical: the binodal defines the equilibrium phase boundary, whereas the cloud point marks a kinetically determined transition that depends on nucleation and growth rates. Similar phenomena occur in freshly distilled Baijiu, which, despite optical clarity, contains abundant micrometer-scale spherical droplets,⁴ motivating the need for multi-scale characterization.

Given this limitation, we employed optical microscopy to directly detect sub-micron liquid-liquid phase separation (LLPS), following a quantitative image analysis approach similar to that used in droplet microfluidic studies.²³ The criteria for determining the phase separation state are as follows: if spherical droplets with a diameter of ≥0.8 µm are observed in ≥3 fields of view, it is determined that phase separation exists; conversely, if such droplets are not observed in all fields of view, it is determined as a homogeneous system. As shown in Fig. 2 for a representative 65% ethanol system, droplet formation and morphology varied systematically with the concentration of hexanoic acid or ethyl hexanoate, visually demonstrating concentration-dependent phase evolution. Quantitative image analysis confirmed that at concentrations below the defined threshold, no droplets were detectable, establishing a robust microscopic boundary for phase separation.

Fig. 2 Concentration-dependent sub-micron liquid–liquid phase separation in 65% ethanol–water systems. (a)–(d) Hexanoic acid at (a) 2 g L⁻¹, (b) 1 g L⁻¹, (c) 0.5 g L⁻¹, and (d) 0.1 g L⁻¹. (e)–(h) Ethyl hexanoate at (e) 5 g L⁻¹, (f) 3 g L⁻¹, (g) 1 g L⁻¹, and (h) 0.3 g L⁻¹.

Fluorescence spectroscopy is sensitive to molecular packing and the local microenvironment, enabling detection of pre-nucleation aggregation events well below the turbidimetric detection limit.²⁴ To capture the intermolecular interactions that occur prior to the formation of optically resolvable droplets, this study utilizes the background fluorescence of ethanol–water systems excited at 275 nm to monitor changes in fluorescence intensity upon the addition of hexanoic acid/ethyl hexanoate. It is important to note that the ethanol–water solvent itself exhibits weak intrinsic fluorescence upon UV excitation, arising from hydrogen-bonded clusters formed between ethanol and water molecules.²⁵ Molecular dynamics simulations have shown that amphiphilic flavor compounds such as hexanoic acid and ethyl hexanoate can engage in hydrogen bonding and electrostatic interactions with the ethanol–water solvent matrix.²⁶ It is therefore plausible that the addition of these solutes is accompanied by local changes in the microenvironment, which could in turn influence the fluorescence response.

Using this approach, we determined the critical aggregation concentration (CAC) of hexanoic acid and ethyl hexanoate in ethanol–water mixtures (55%, 65%, and 75% v/v). The CAC was identified as the inflection point of the fluorescence intensity–concentration curve, where the slope change exceeded three times the baseline noise. As shown in Fig. 3a and b, a slight non-monotonic variation in fluorescence intensity was observed near the CAC. This arises because fluorescence intensity often shows non-monotonic changes during molecular self-assembly, as different aggregated states coexist and exhibit distinct quantum yields.^27–29 Similar behavior has been reported in the aggregation of amyloid proteins, where the coexistence of monomers and oligomers leads to non-linear fluorescence responses.³⁰ The CAC values obtained by fluorescence spectroscopy were consistently lower than those determined by optical microscopy under the same solvent compositions (Fig. 3b and c). Strikingly, these CAC values (0.03–0.50 g L⁻¹) lie 2–3 orders of magnitude below the binodal boundaries determined by classical turbidimetry (22.6–497.7 g L⁻¹), placing them within the metastable pre-Ouzo region (Fig. 3d, inset). For instance, in a 65% ethanol–water system, fluorescence spectroscopy yielded a CAC of 0.05 g L⁻¹ for hexanoic acid and 0.20 g L⁻¹ for ethyl hexanoate, while optical microscopy gave the corresponding CAC values of 0.12 g L⁻¹ and 0.30 g L⁻¹, respectively. In sharp contrast, turbidimetry determined the saturation thresholds of ethyl hexanoate in this system to be in the range of 22.6–497.7 g L⁻¹, and the saturation threshold of hexanoic acid was of the same order of magnitude, which was significantly higher than the CAC values obtained from the above two methods.

Fig. 3 Phase behavior and critical aggregation concentration (CAC) of flavor compounds in ethanol–water mixtures. Fluorescence intensity of (a) ethyl hexanoate and (b) hexanoic acid as a function of concentration in 55%, 65%, and 75% (v/v) ethanol–water mixtures. Dashed lines indicate the CAC values determined from inflection points. (c) CAC values of ethyl hexanoate (blue) and hexanoic acid (red) in 50–80% (v/v) ethanol–water mixtures, determined by optical microscopy. The shaded area represents the metastable region without macroscopic phase separation. (d) Ternary phase diagram of the ethyl hexanoate–hexanoic acid–ethanol–water system, showing phase boundaries (solid lines, turbidimetry) and CAC values. Inset: CAC values vs. ethanol concentration.

This systematic difference does not reflect a limitation of turbidimetry, but rather the distinct physical events each method probes. Turbidimetry defines the thermodynamic solubility limit (the binodal), while optical microscopy resolves submicron droplets within the metastable Ouzo region. Fluorescence spectroscopy, in contrast, responds to nanoscale density fluctuations and pre-nucleation aggregation occurring well before visible turbidity develops. The CACs obtained by fluorescence spectroscopy are consistently lower than the thresholds determined by microscopy (Fig. 3), suggesting that fluorescence captures aggregation events prior to or at the onset of the Ouzo region. The finding that fluorescence detects concentrations that are 2–3 orders of magnitude lower than those detected by turbidimetry reflects its higher sensitivity to sub-binodal events—a distinction physically grounded in spinodal decomposition and the Ouzo effect, where metastable supersaturation persists without macroscopic phase separation.^31,32 Thus, the methods provide complementary views of the phase separation pathway.

3.2 Pre-Ouzo supersaturation revealed by phase mapping

Gas chromatography–mass spectrometry (GC–MS) is often utilized for the identification and quantification of volatile organic compounds in Baijiu, leveraging its hyphenated capabilities in chromatographic separation and mass-selective detection to achieve high-resolution analysis of trace flavor constituents.^33,34 GC–MS was employed to quantify hexanoic acid and ethyl hexanoate in both dissolved and supersaturated phases across strong-flavor Baijiu with 1–20 years of aging. Fig. 4a and b shows effective supersaturated phase removal via centrifugation/filtration through comparative microscopy of a 10-year Baijiu sample, establishing a foundation for phase-resolved quantification. As shown in Fig. 4c, the concentration profiles of these compounds covaried throughout aging, resulting from the collective effects of the dynamic esterification equilibrium, ethanol-modulated phase partitioning, competitive side reactions, and flavor synergy mechanisms. Notably, the dissolution efficiencies of both hexanoic acid and ethyl hexanoate increased progressively with aging time (Fig. 4c), suggesting that the colloidal matrix undergoes continuous restructuring that favors solubilization over precipitation.

Fig. 4 Phase-resolved quantification and solubility evolution of hexanoic acid and ethyl hexanoate in strong-flavor Baijiu during aging. (a) and (b) Comparative microscopic validation of supersaturated phase removal via centrifugation/filtration in a 10-year Baijiu sample. (c) Temporal evolution of dissolved concentrations (columns) and dissolution efficiencies (lines) for hexanoic acid (red) and ethyl hexanoate (blue) across 1–20 years’ aging; data are presented as mean (n = 3). (d) Ethanol-modulated phase diagram mapping dissolved concentrations against theoretical solubility thresholds (shaded region), reflecting aging-induced solubilization enhancement beyond the static binodal (n = 3).

Ethanol-modulated phase diagrams specific to Baijiu systems were constructed (Fig. 4d); dissolved-phase concentrations of hexanoic acid and ethyl hexanoate exhibited stage-dependent evolution: years 1–10: dissolution efficiencies increased from 58.45% to 96.86% (hexanoic acid) and from 54.23% to 96.91% (ethyl hexanoate), with absolute concentrations increasing from 0.12 to 0.79 g L⁻¹ (acid) and from 0.56 to 1.63 g L⁻¹ (ester); years 10–20: concentrations surged synergistically to 1.91 g L⁻¹ (acid) and 4.34 g L⁻¹ (ester), achieving dissolution efficiencies >96% for both compounds. When these dissolved concentrations were plotted on the newly established integrated fluorescence–microscopy phase diagram (Fig. 4d), a striking trend emerged: with prolonged aging, the data points shifted systematically from the predicted two-phase region toward the single-phase zone, yet their absolute values substantially exceeded the static saturation thresholds derived from the binary ethanol–water system. Jiang et al.⁴ demonstrated that amphiphilic colloidal particles in strong-flavor Baijiu undergo Ostwald ripening during aging, with the mean particle size increasing from 1.86 µm to 2.96 µm over 16 years and the particle number decreasing by 39.5%—providing direct evidence that the solvent microstructure evolves in situ during cellaring. Concurrently, Shang et al.³⁵ showed that ethanol clusters in soy sauce flavor Baijiu transform from hydrophilic configurations to dense hydrophobic lamellae over three decades, accompanied by a 13% reduction in free ethanol concentration; crucially, this structural evolution directly modulates the solubility of long-chain fatty acid ethyl esters, leading to a non-monotonic dissolution pattern. More recently, Jiang et al.¹⁴ quantified a 13.3% decrease in free ethanol over 10 years of strong-flavor Baijiu aging and identified LLPS as the core mechanism driving structural maturation. Collectively, these studies demonstrate that aged Baijiu is not simply a “model system plus time”, but a dynamically evolving colloidal system whose solvent microstructure reorganizes over decades of cellaring.

The colloidal nature of aged Baijiu and its stability have been systematically reviewed by Jia et al.,³⁶ who emphasized that long-chain fatty acid ethyl esters (LCFAEEs) are prone to form stable colloidal clusters during aging, thereby avoiding hydrolysis and contributing to the overall stability of the Baijiu colloidal system. This perspective aligns with our observation that dissolution efficiencies exceed 96% for both hexanoic acid and ethyl hexanoate after 20 years of aging, indicating that the system has evolved into a kinetically stabilized colloidal state rather than a thermodynamically equilibrium one. It is important to emphasize that the phase diagram in Fig. 4d was constructed using freshly prepared ethanol–water mixtures. When applied to aged Baijiu, this diagram serves as a static reference framework that delineates the onset of the (pre-)Ouzo region expected for a solvent with the same bulk composition (69% v/v ethanol) but without the microstructural evolution that occurs during decades of cellaring. Therefore, the progressive shift of the data points from the two-phase region toward the single-phase zone does not indicate a failure of the phase diagram to represent the aged system; rather, it reflects the dissolved concentrations measured in aged Baijiu relative to this static reference. This deviation places the aging trajectory of Baijiu between the binodal curve (determined by turbidimetry) and the (pre-)Ouzo onset boundary (defined by fluorescence and microscopy). The observed increase in dissolved concentrations with aging time aligns with the characteristic behavior of the Ouzo effect, wherein metastable, kinetically arrested nanodroplets form without crossing the binodal.²² In our system, the progressive solubilization suggests that long-term solvent restructuring gradually shifts the system deeper into the Ouzo region, stabilizing concentrations below the binodal but above the static reference boundary. It should be noted that while hexanoic acid and ethyl hexanoate are key flavor components in strong-flavor Baijiu,³⁶ the roles of other trace constituents during aging remain unknown. The potential synergistic effects of these accumulated surface-active species warrant systematic investigation in future work.

3.3 Mechanism of aging-enhanced solubilization

3.3.1 Restructuring of ethanol–water hydrogen-bond networks. To explore whether ethanol–water hydrogen-bond networks evolve during aging, we monitored the intrinsic fluorescence of ethanol–water clusters under 240 nm excitation. The steady-state fluorescence spectra in Fig. 5(a) demonstrate that under 240 nm excitation, the fluorescence intensities at 308 nm, 330 nm, and 373 nm—corresponding to (H₂O)(EtOH)_m, (H₂O)_n(EtOH)_m, and (H₂O)_n(EtOH) ethanol–water clusters, respectively—are progressively enhanced with prolonged aging of strong-flavor Baijiu (m ≫ n). Notably, the (H₂O)_n(EtOH)_m cluster, which exhibits relatively higher stability, shows a pronounced increase in fluorescence intensity at 330 nm. This trend is consistent with a time-dependent reorganization of the hydrogen-bond network. Studies have shown that enhanced molecular assembly and structural ordering are often accompanied by the stabilization and regularization of hydrogen-bonding interactions, which in turn lead to an increase in fluorescence intensity without a significant shift in the fluorescence emission peak position.³⁷ Raman spectroscopy has directly confirmed the progressive strengthening of hydrogen bonds during Baijiu aging.⁵ Yang et al.³⁸ revealed that ethanol–water clusters exist in two primary configurations—symmetric tetrahedral (water-dominated) and chain-like (ethanol-dominated)—with their relative proportions shifting continuously with ethanol concentration, and confirmed that typical Baijiu esters do not alter these fundamental cluster structures. Using DFT and MD simulations, Huang et al.²⁶ (2022) demonstrated that the ethanol–water system containing acids exhibits significantly stronger electrostatic interactions and a higher number of hydrogen bonds compared to the system containing esters. Shang et al.³⁵ suggest that in soy sauce flavor Baijiu, ethanol clusters undergo LLPS-driven transformation from hydrophilic configurations to dense hydrophobic lamellae over three decades, accompanied by a 13% reduction in free ethanol concentration. Such structural evolution directly modulates the solubility of flavor compounds, establishing a causal link between hydrogen-bond network restructuring and solubilization behavior.

Fig. 5 (a) Steady-state fluorescence spectra of Baijiu samples with different aging years under 240 nm excitation, showing cluster fluorescence signals at 308 nm, 330 nm, and 373 nm. (b) Fluorescence spectra of aged Baijiu samples from 1 to 20 years, determined using Nile red as a hydrophobic probe.

3.3.2 Concomitant expansion of hydrophobic microdomains. To further elucidate the cooperative solubilization mechanism, we next examined the evolution of hydrophobic microdomains, which are expected to provide favorable niches for amphiphile sequestration. Previous research studies have suggested^39–41 that Nile red, when embedded in the hydrophobic microdomains of amphiphilic nanogels (ANGs), exhibits fluorescence signals directly correlated with the local hydrophobicity: enhanced hydrophobicity leads to more pronounced fluorescence variations. These changes enable quantitative characterization of the polarity, structural properties, and dynamic behaviors within hydrophobic microdomains. Accordingly, Nile red was utilized as a fluorescent probe in this study to stain strong-flavor Baijiu samples with varying aging durations, facilitating the characterization of hydrophobic domain features through fluorescence monitoring. Using 552 nm as the optimal excitation wavelength, the results (Fig. 5b) revealed a progressive increase in fluorescence intensity at the 640 nm emission peak with extended aging periods. Given that Nile red is well established as a polarity-sensitive probe whose quantum yield increases in less polar environments,⁴² this trend suggests an aging-induced expansion or enhanced hydrophobicity of the microdomains that sequester the probe. The properties of Nile red reveal that hydrophobic microdomains in strong-flavor Baijiu undergo dynamic evolution during aging, manifested by increased quantity, expanded spatial distribution, or enhanced hydrophobicity.

3.3.3 A synergistic solubilization model. Based on the above observations, we propose a synergistic solubilization model. The progressive restructuring of ethanol–water clusters creates a dynamic polar scaffold with reduced free energy penalty for amphiphile insertion. Simultaneously, the expansion of hydrophobic microdomains provides increased capacity and a thermodynamically favored niche for the alkyl chains of the flavor molecules. These two processes likely reinforce each other in a positive feedback loop, driving the system toward a kinetically stabilized, supersaturated state. This framework provides a testable hypothesis for future studies combining spectroscopic, scattering, and computational approaches.

4. Conclusion

In this study, we established an integrated ternary phase mapping method. This methodological advance enabled the detection of critical saturation concentrations for hexanoic acid and ethyl hexanoate at thresholds 2–3 orders of magnitude below the binodal determined by classical turbidimetry, thereby mapping the pre-Ouzo region. Applying this refined phase diagram to naturally aged strong-flavor Baijiu (1–20 years) revealed a pronounced aging-enhanced solubilization phenomenon. The dissolution efficiencies of both flavor compounds progressively increased to >96%, with their dissolved concentrations substantially exceeding the static saturation limits predicted by equilibrium phase models. This marked deviation underscores the dynamic, non-equilibrium nature of aging in ethanol–water–amphiphile colloidal systems.

Mechanistic investigations revealed that this aging-enhanced solubilization is synergistically driven by two interrelated microstructural evolutions: (i) the continuous restructuring of ethanol–water hydrogen-bond networks toward more stable cluster states and (ii) the concomitant expansion of hydrophobic microdomains within the colloidal matrix. The reorganized hydrogen-bond network provides a dynamic polar scaffold that reduces the free energy barrier for amphiphile incorporation, while the expanded hydrophobic domains offer increased capacity for sequestering the alkyl chains of the flavor molecules. This cooperative process stabilizes a non-equilibrium colloidal state, effectively pushing the apparent solubility beyond the limits of static phase equilibrium. We acknowledge that long-term stability characterization of the model ternary system (e.g., DLS monitoring of droplet size evolution and ζ-potential measurements) was beyond the scope of this methodology-oriented study, which focused on detecting the onset of the pre-Ouzo region. Future work should integrate time-resolved DLS, ζ-potential measurements, and in situ microscopy to systematically investigate the effects of ethanol concentration and amphiphile content on the long-term colloidal stability of ethanol–water–amphiphile systems. We also acknowledge that the complexity of the Baijiu matrix may introduce co-surfactant effects; future work should systematically disentangle solvent restructuring from the contributions of other surface-active species.

Author contributions

Chaoyu Zhao: writing – original draft, investigation, formal analysis, and data curation. Xinyue Jiang: writing – review and editing, supervision, and methodology. Yuqun Xie: writing – review and editing, supervision, project administration, methodology, and funding acquisition.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.

Acknowledgements

This work received no external funding.

References

1. P. Denk, A. E. Maangar, J. Lal, D. Kleber, T. Zemb and W. Kunz, J. Colloid Interface Sci., 2021, 590, 375–386.
2. Y. H. A. Leurs, W. van den Hout, A. Gardin, J. L. J. van Dongen, A. Rodriguez-Abetxuko, N. A. Erkamp, J. C. M. van Hest, F. Grisoni and L. Brunsveld, Nat. Commun., 2025, 16, 9598.
3. L. E. Meyer, M. B. Andersen and S. Kara, Angew. Chem., Int. Ed., 2022, 61, e202203823.
4. X. Jiang, D. Liu, S. Yang, X. Cheng and Y. Xie, Food Chem., 2024, 461, 140883.
5. X. Jiang, R. Liu and Y. Xie, Food Chem., 2024, 139420.
6. R. Liu, X. Jiang, Y. Zheng, G. Jiang, S. Yang, X. Cheng and Y. Xie, Int. J. Food Prop., 2024, 27, 951–959.
7. J. X. Liu, N. Bizmark, D. M. Scott, R. A. Register, M. P. Haataja, S. S. Datta, C. B. Arnold and R. D. Priestley, JACS Au, 2021, 1, 936–944.
8. S. M. Fenton, P. Padmanabhan, B. K. Ryu, T. T. D. Nguyen, R. N. Zia and M. E. Helgeson, Proc. Natl. Acad. Sci. U. S. A., 2023, 120, e2215922120.
9. B. Keshavarz, D. G. Rodrigues, J. Champenois, M. G. Frith, J. Ilavsky, M. Geri, T. Divoux, G. H. McKinley and A. Poulesquen, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e2022339118.
10. K. Dai, Y. Guo and B. D. Olsen, Biomacromolecules, 2026, 27, 1310–1316.
11. L. Meng, R. Wei, Z. Xie, D. Zhang, X. Wan, H. Han, W. Shi, Z. Zhu, X. Xiao, Y. Q. Gao and D. Zhao, Chem. Sci., 2025, 16, 15602–15610.
12. Y. Dong, Y. Yang, C. Lin and D. Liu, Acc. Chem. Res., 2022, 55, 1938–1948.
13. A. Jain, S. Kassem, R. S. Fisher, B. Wang, T. Li, T. Wang, Y. He, S. Elbaum-Garfinkle and R. V. Ulijn, J. Am. Chem. Soc., 2022, 144, 15002–15007.
14. X. Jiang, Y. Shang, R. Hajar, H. Yang, J. Peng, J. Li, W. Wang, Z. Zou and Y. Xie, Food Res. Int., 2026, 225, 118060.
15. Y. T. Sun, R. Zhao, D. W. Ding, Y. H. Liu, H. Y. Bai, M. Z. Li and W. H. Wang, Nat. Commun., 2023, 14, 540.
16. M. Shoaib, J. Mata, C. Didier, I. Kuzmenko, L. M. Voelker-Pop and E. Bobicki, J. Colloid Interface Sci., 2025, 693, 137536.
17. Y. Xu, J. Zhao, X. Liu, C. Zhang, Z. Zhao, X. Li and B. Sun, Food Chem., 2022, 369, 130920.
18. Y. Mu, J. Huang, R. Zhou, S. Zhang, H. Qin, H. Tang, Q. Pan and H. Tang, Food Chem., 2023, 424, 136429.
19. J. Hong, H. Huang, D. Zhao, J. Sun, M. Huang, X. Sun and B. Sun, Food Chem., 2023, 426, 136576.
20. L. Zhang, L. Pan, X. Chen, D. Xu, X. Wang, R. Zhou, I. Ali, B. Wang, P. Gao, M. Huang, H. Chen, X. Dai and R. Xue, Food Chem.: X, 2025, 27, 102349.
21. M. Mohsen-Nia, B. Jazi and H. Amiri, J. Chem. Thermodyn., 2009, 41(7), 859–863.
22. D. Iglicki, C. Goubault, M. N. Mahamoud, S. Chevance and F. Gauffre, J. Colloid Interface Sci., 2023, 633, 72–81.
23. A. Villois, U. C. Palmiero, P. Mathur, G. Perone, T. Schneider, L. Li, M. Salvalaglio, A. DeMello, S. Stavrakis and P. Arosio, Nat. Chem. Biol., 2022, 18, 2202606.
24. P. Y. Gu, F. Zhou, G. Xie, P. Y. Kim, Y. Chai, Q. Hu, S. Shi, Q. F. Xu, F. Liu, J. M. Lu and T. P. Russell, Angew. Chem., 2021, 133, 8776–8781.
25. B. Wu, Y. Liu, C. Han, X. Luo, J. Lu and X. Ni, Chin. Opt. Lett., 2009, 7, 159–161.
26. Z. Huang, Y. Zeng, Q. Sun, W. Zhang, S. Wang, C. Shen and B. Shi, Food Chem., 2022, 373, 131522.
27. J. Li, J. Zhang, J. Wang, D. Wang, Y. Yan, J. Huang and B. Z. Tang, ACS Nano, 2022, 16, 20559–20566.
28. A. Kaur, L. D. Adair, S. R. Ball, E. J. New and M. Sunde, Angew. Chem., 2022, 134, e202112832.
29. J. Xu, Y. Wang, M. Huang, X. Xu, Y. Zeng, X. Luo, S. Pei, K. Xu and W. Zhong, Talanta, 2023, 261, 124662.
30. M. A. Saraiva, Int. J. Biol. Macromol., 2021, 177, 392–400.
31. S. A. Vitale and J. L. Katz, Langmuir, 2003, 19, 4105–4110.
32. T. N. Zemb, M. Klossek, T. Lopian, J. Marcus, S. Schöettl, D. Horinek, S. F. Prevost, D. Touraud, O. Diat, S. Marčelja and W. Kunz, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 4260–4265.
33. Y. Wang, J. Zheng, D. Zhang, S. Pan, X. Xu and F. Yuan, J. Agric. Food Chem., 2025, 73, 17888–17900.
34. J. Niu, R. Liu, W. Li, Y. Lang, X. Li, W. Sun and B. Sun, Food Res. Int., 2025, 213, 116568.
35. Y. Shang, X. Jiang, Y. Zuo and Y. Xie, Food Chem., 2026, 505, 148095.
36. W. Jia, R. Ma, L. Hu and H. Mo, Food Chem.: X, 2023, 17, 100554.
37. X. Wang, Y. Yang, H. Yang and H. Dong, Angew. Chem., Int. Ed., 2024, 64, e202420567.
38. X. Yang, J. Zheng, X. Luo, H. Xiao, P. Li, X. Luo, Y. Tian, L. Jiang and D. Zhao, Matter, 2024, 7, 1724–1735.
39. L. Lauritsen, M. Szomek, M. Hornum, P. Reinholdt, J. Kongsted, P. Nielsen, J. R. Brewer and D. Wüstner, Sci. Rep., 2024, 14, 13748.
40. C. López-Iglesias, A. Markovina, N. Nirmalananthan-Budau, U. Resch-Genger and D. Klinger, Nanoscale, 2024, 16, 9525–9535.
41. W. Teo, M. L. Morgan and P. K. Stys, J. Neurochem., 2024, 169, e16203.
42. W. Teo, A. V. Caprariello, M. L. Morgan, A. Luchicchi, G. J. Schenk, J. T. Joseph, J. J. G. Geurts and P. K. Stys, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e2016897118.

---