Liquid-Vapor Equilibrium & Azeotropes theory

Vapor-liquid phase equilibrium represents a state of thermodynamic coexistence between a liquid phase and its vapor phase. At equilibrium, three fundamental conditions must be simultaneously satisfied: (1) thermal equilibrium, requiring uniform temperature across both phases (T^L = T^V); (2) mechanical equilibrium, requiring uniform pressure across the interface (P^L = P^V = P); and (3) chemical equilibrium, requiring that the chemical potential of every chemical species i be identical in both phases: μ_i^L(T, P, x) = μ_i^V(T, P, y). Because chemical potential represents the escaping tendency of a substance from a phase, equality of chemical potentials ensures zero net mass transfer across the vapor-liquid interface.



In practical phase equilibrium calculations, chemical potential is related to fugacity f_i, where μ_i - μ_i° = RT ln(f_i / f_i°). The equilibrium condition therefore translates to equality of fugacities: f_i^L = f_i^V. In the vapor phase at low to moderate pressures (such as atmospheric pressure P = 101.3 kPa), vapor behaves as an ideal gas mixture, and the vapor fugacity is simply the partial pressure: f_i^V = y_i P. In the liquid phase, deviations from ideal solution behavior are captured by the activity coefficient γ_i: f_i^L = x_i γ_i f_i*, where f_i* is the fugacity of pure liquid i at system temperature and pressure, which equals the saturation vapor pressure P_i*(T) (neglecting the small Poynting pressure correction). Equating vapor and liquid fugacities yields the modified Raoult's law:



y_i P = x_i γ_i P_i*(T)



This fundamental relation governs all real distillation and boiling equilibria: it connects the directly measurable quantities (total pressure P, equilibrium temperature T, liquid composition x_i, and vapor composition y_i) to the microscopic intermolecular interactions quantified by the activity coefficient γ_i.

Deviations from Raoult's law reflect differences between intermolecular forces among like molecules (1-1 and 2-2) versus unlike molecules (1-2). In an ideal solution, the intermolecular interaction energy between unlike species equals the arithmetic or geometric mean of like-species interactions (w₁₂ = [w₁₁ + w₂₂]/2), resulting in zero enthalpy of mixing (ΔH_mix = 0) and activity coefficients identically equal to unity (γ_i = 1).



In the ethanol-water binary system, profound molecular asymmetry causes strong positive deviations from Raoult's law (γ₁ > 1 and γ₂ > 1). Water molecules (H₂O) form an extraordinarily tight, three-dimensional, highly cooperative hydrogen-bonded network with high cohesive energy density (internal pressure ~2200 MPa). Pure ethanol (C₂H₅OH) also forms hydrogen bonds, but its bulky non-polar ethyl groups (−CH₂CH₃) disrupt hydrogen bonding, giving it a much lower cohesive energy density (internal pressure ~290 MPa).



When ethanol is mixed with water, the non-polar ethyl tails cannot participate in hydrogen bonding. Water molecules are forced to reorganize around the alkyl groups into highly structured, clathrate-like solvation cages (hydrophobic hydration), incurring an unfavorable entropic penalty. Thermodynamically, water molecules prefer to associate with other water molecules rather than accommodate hydrocarbon tails, while ethanol molecules prefer to associate with ethanol. This net mutual repulsion creates an unfavorable, endothermic excess enthalpy of mixing (ΔH_mix > 0) and an excess Gibbs free energy that is positive (g^E > 0). Consequently, both ethanol and water molecules experience an enhanced escaping tendency from the liquid solution into the vapor phase compared to an ideal mixture: the vapor pressure of each component is higher than predicted by Raoult's law, yielding activity coefficients substantially greater than unity (for dilute ethanol in water, γ₁^∞ ≈ 4.0 to 4.2).

An azeotrope (from the Greek 'boiling without change') is a liquid mixture whose coexisting vapor has exactly the same chemical composition as the liquid phase: x_i = y_i for all components i. Consequently, when an azeotropic mixture boils, it vaporizes without any change in liquid composition or boiling temperature, behaving macroscopically like a single pure substance and preventing further separation by ordinary fractional distillation.



The thermodynamic origin of azeotropes is governed by the rigorous Gibbs-Konovalov theorems, derived from the Gibbs-Duhem equation applied simultaneously to both coexisting phases. Konovalov's first rule states that the vapor phase is always richer in the component whose addition increases the total vapor pressure (or lowers the boiling point). Konovalov's second rule establishes that an extremum in the total vapor pressure curve (dP/dx₁ = 0 at constant T) or an extremum in the boiling temperature curve (dT/dx₁ = 0 at constant P) occurs if and only if the vapor and liquid compositions are identical: x₁ = y₁.



In systems with strong positive deviations from Raoult's law (such as ethanol-water), the high activity coefficients (γ₁ > 1, γ₂ > 1) elevate the total vapor pressure above the ideal line. Because the pure component boiling points are relatively close (ethanol at 78.37 °C and water at 100.00 °C), the synergistic vapor pressure enhancement from both components produces a maximum in the total pressure curve P(x₁) at constant temperature. By the Gibbs-Konovalov theorem, a maximum in vapor pressure at constant temperature corresponds to a minimum in boiling temperature at constant pressure: a minimum-boiling azeotrope. At 101.325 kPa (1 atm), the ethanol-water azeotrope occurs at x₁ = y₁ ≈ 0.894 mole fraction (95.63% by mass or 96.5% by volume), boiling at T_b = 78.15 °C—which is 0.22 °C lower than pure ethanol (78.37 °C) and 21.85 °C lower than pure water (100.00 °C).

In solution thermodynamics, deviations from ideal behavior are described by the molar excess Gibbs free energy g^E = g - g^ideal. In 1895, Max Margules proposed expressing the excess Gibbs energy as a power series in mole fractions. For a symmetric binary regular solution, the excess Gibbs energy is represented by the 1-parameter Margules equation (also known as the Porter equation):



g^E / (R T) = A · x₁ · x₂



where A is a dimensionless empirical interaction parameter (interchange energy divided by RT: A = w / [k_B T]), and x₁ and x₂ are the mole fractions of components 1 and 2 (x₁ + x₂ = 1). The physical significance of A is direct: it quantifies the net energetic penalty of creating 1-2 contacts relative to 1-1 and 2-2 contacts. If A = 0, the solution is ideal (g^E = 0). If A > 0, the solution exhibits positive deviations from Raoult's law; if A > 2, the mixture becomes thermodynamically unstable and undergoes liquid-liquid phase separation (immiscibility).



To find the individual activity coefficients γ₁ and γ₂, we apply the partial molar excess Gibbs energy relation: R T ln(γ_i) = (∂[n_tot g^E] / ∂n_i)_{T, P, n_{j ≠ i}}. Carrying out this differentiation for component 1 yields:



ln(γ₁) = A · x₂² = A · (1 − x₁)²

ln(γ₂) = A · x₁² = A · (1 − x₂)²



This quadratic dependence has profound implications: (1) as x₁ → 1 (pure ethanol), x₂ → 0, so ln(γ₁) → 0 and γ₁ → 1, recovering Raoult's law for the solvent; (2) as x₁ → 0 (infinite dilution of ethanol in water), x₂ → 1, so ln(γ₁) reaches its maximum value: ln(γ₁^∞) = A, or γ₁^∞ = exp(A). Therefore, the Margules parameter A directly represents the logarithm of the limiting activity coefficient at infinite dilution. Plotting experimental ln(γ₁) against (1 − x₁)² transforms the non-linear VLE data into a straight line with slope A passing through the origin, providing a robust, unbiased method to extract thermodynamic parameters from laboratory observations.

In boiling point and vapor-liquid equilibrium metrology, a major source of systematic error is liquid superheating. When a liquid is heated in a conventional boiling flask, bubbles nucleate at the heated vessel wall, where local temperatures can exceed the true thermodynamic boiling point by 1.0 °C to 3.0 °C before detachment. If a thermometer or thermocouple is merely immersed in the boiling liquid, it records this elevated superheated wall temperature rather than the true equilibrium boiling point. Conversely, if the sensor is suspended in the vapor space above the boiling liquid, partial condensation on the sensor stem and heat loss to the cooler upper walls cause droplet subcooling, yielding a temperature reading that is systematically too low.



In 1919, Frederick Gardner Cottrell invented the thermo-lift boiling pump (Cottrell pump) to overcome this fundamental dilemma. The Cottrell apparatus consists of an inverted glass funnel placed over the boiling zone, terminating in two or three slender glass lift tubes that curl around the thermometer well. As boiling occurs, rising vapor bubbles become trapped beneath the funnel. As these bubbles expand and ascend through the narrow lift tubes, they act as an air-lift piston, pumping an intimate, turbulent mixture of boiling liquid and coexisting equilibrium vapor continuously over the thermometer bulb.



Because the thermometer bulb is continuously washed by a thin, flowing film of liquid in direct dynamic contact with its own saturated vapor, true vapor-liquid phase equilibrium is established right on the sensor surface. Any superheat in the liquid is immediately dissipated by rapid flash vaporization into the accompanying vapor bubbles, while any cooling is counteracted by latent heat of condensation. The Cottrell pump thus isolates the temperature measurement completely from flask wall conduction, radiation, and droplet subcooling, ensuring boiling point measurement precision of ±0.02 °C.

In vapor-liquid equilibrium investigations, measuring the chemical compositions of coexisting liquid and vapor phases must be fast, accurate, and non-destructive, requiring only microliter sample volumes so as not to perturb the steady-state equilibrium of the still. While gas chromatography (GC) and Karl Fischer titration are accurate, they require complex calibration, carrier gases, and long elution times. Optical refractometry using an Abbe refractometer provides an ideal, instantaneous alternative.



The Abbe refractometer operates on the principle of total internal reflection and the critical angle of illumination. When monochromatic light (typically the sodium D line, λ = 589.3 nm) passes from a high-refractive-index flint glass prism (n_prism ≈ 1.74) into a thin film of liquid sample (n_sample ≈ 1.33 to 1.36), the critical angle of refraction θ_c is governed by Snell's law: sin(θ_c) = n_sample / n_prism. In the refractometer eyepiece, this critical angle produces a sharp boundary between a light field and a dark field. The operator aligns this boundary with crosshairs, reading the refractive index n_D directly on an internal scale to ±0.0001.



Pure water has a refractive index of n_D²⁰ = 1.3330, while pure ethanol has n_D²⁰ = 1.3614 at 20.00 °C. Because the refractive index of a binary mixture is an intensive monotonic function of composition governed by the Lorentz-Lorenz polarizability equation, measuring n_D allows instant determination of mole fraction. Because refractive index varies significantly with temperature (dn_D/dT ≈ −4.0 × 10⁻⁴ K⁻¹ for water and −4.5 × 10⁻⁴ K⁻¹ for ethanol), the Abbe prisms must be water-jacketed and strictly maintained at 20.00 ± 0.02 °C. An uncertainty of ±0.0001 in n_D corresponds to an uncertainty of only δx ≈ ±0.0035 in ethanol mole fraction, providing research-grade compositional accuracy.

Conventional fractional distillation operates on the principle of repeated countercurrent vaporization and condensation stages within a distillation column. In a column operating at total or partial reflux, the vapor phase ascending each stage is enriched in the more volatile component according to the relative volatility α₁₂ = (y₁ / x₁) / (y₂ / x₂). In a non-azeotropic system where relative volatility exceeds unity (α₁₂ > 1), repeating this enrichment over multiple theoretical plates produces pure light component at the top of the column and pure heavy component at the reboiler.



However, in the ethanol-water system, as ethanol concentration approaches the azeotropic composition (x₁ = 0.894 mole fraction, 95.63 wt%), the driving force for separation vanishes: y₁ approaches x₁, and the relative volatility α₁₂ approaches exactly 1.000. At the azeotropic point, the vapor leaving the boiling liquid has the identical composition as the liquid; vaporizing and condensing this mixture produces zero separation. The azeotropic composition forms an insurmountable thermodynamic distillation boundary: starting with any dilute fermentation broth (e.g., 10% ethanol), a fractional distillation column can enrich ethanol up to 95.63 wt%, but cannot advance beyond this point regardless of column height or reflux ratio.



To break the azeotrope and produce anhydrous (100%) ethanol for biofuels or chemical synthesis, chemical engineers employ specialized advanced separation technologies: (1) Azeotropic Distillation with an Entrainer: adding a third chemical component (such as cyclohexane or historically benzene) forms a ternary minimum-boiling azeotrope that boils at 64.9 °C, carrying water overhead while dry ethanol discharges from the bottom; (2) Pressure-Swing Distillation: because azeotropic composition is pressure-dependent (at pressures below 11.5 kPa or ~86 mmHg, the ethanol-water azeotrope disappears entirely), running two interconnected columns at different pressures allows crossing the azeotrope without chemicals; (3) Extractive Distillation: adding a high-boiling polar solvent (such as ethylene glycol or ionic liquids) selectively alters the relative volatility of ethanol over water; and (4) Molecular Sieve Adsorption: passing the 95.6% ethanol vapor through a bed of synthetic Type 3A zeolites (pore size 3 Å), which adsorb water molecules (kinetic diameter 2.8 Å) while completely excluding the larger ethanol molecules (kinetic diameter 4.4 Å), producing 99.9% anhydrous ethanol cleanly and efficiently.