Introduction & Context

Multi-Stage Liquid-Liquid Extraction (LLE) is a fundamental mass transfer operation in process engineering used to separate components based on their relative solubilities in two immiscible liquid phases. Understanding the distinction between washing and extraction processes is essential when designing counter‑current extraction columns, such as packed or trayed towers, where a solute is transferred from a carrier liquid (feed) to an extracting solvent.

This methodology is typically employed in the chemical, pharmaceutical, and food processing industries—such as the removal of impurities from essential oils or the recovery of organic acids from aqueous streams. By determining the number of theoretical stages required to meet a specific purity target, engineers can size equipment, estimate capital expenditure, and optimize solvent-to-feed ratios to balance operational costs against separation performance.

Methodology & Formulas

The design relies on the Kremser equation, which assumes a dilute system where the distribution coefficient and phase flow rates remain constant throughout the column; for a detailed guide on how to perform a distribution coefficient calculation, see the dedicated page. The following algebraic steps define the calculation logic:

1. Extraction Factor Calculation: The extraction factor, E, represents the ratio of the solute's capacity in the solvent to its capacity in the feed.

\[ E = \frac{S \cdot K}{B} \]

2. Theoretical Stage Computation: The number of theoretical stages, N_theo, is derived using the logarithmic Kremser relationship, accounting for the feed concentration x_F, target raffinate concentration x_R, and solvent inlet concentration y_S. This form of the equation is valid for E ≠ 1; for the special case where E = 1, the simplified form N_theo = (x_F − y_S/K)/(x_R − y_S/K) − 1 must be used, and it highlights the importance of addressing incomplete solute recovery in the design.

\[ N_{\text{theo}} = \frac{\ln\left[ \left( \frac{x_{F} - \frac{y_{S}}{K}}{x_{R} - \frac{y_{S}}{K}} \right) \cdot \left( 1 - \frac{1}{E} \right) + \frac{1}{E} \right]}{\ln(E)} \]

3. Actual Stage Estimation: To account for non-ideal mass transfer and hydrodynamic limitations, the theoretical stages are adjusted by the tray efficiency, η.

\[ N_{\text{act}} = \frac{N_{\text{theo}}}{\eta} \]

Operational Validity and Empirical Ranges

Parameter	Constraint/Regime	Engineering Significance
Solute Concentration	x_F ≤ 0.05	Ensures dilute assumption and constant physical properties.
Distribution Coefficient	0.5 < K < 5.0	Validates the applicability of the linear equilibrium model.
Extraction Factor	E > 0.6, E ≠ 1.0	Lower values indicate insufficient driving force for separation; E = 1 requires the asymptotic form of the Kremser equation.
Tray Efficiency	0.2 ≤ η ≤ 0.5	Typical range for sieve trays in liquid-liquid systems.

To determine the optimal number of stages, process engineers typically utilize the following methodology:

Construct an equilibrium distribution curve based on the partition coefficient of the solute.
Apply the Hunter-Nash graphical method or the Kremser equation for dilute systems to calculate theoretical stages.
Adjust for stage efficiency, as real-world mass transfer limitations often require more stages than the theoretical ideal.
Perform a sensitivity analysis on the solvent-to-feed ratio to balance capital expenditure against operating costs.

Effective phase separation is critical for process efficiency and is influenced by several physical properties:

Density difference between the extract and raffinate phases, which drives gravitational settling.
Interfacial tension, which dictates the size of dispersed droplets and the potential for stable emulsion formation.
Viscosity of the continuous phase, which impacts the terminal velocity of droplets according to Stokes Law.
Presence of surfactants or impurities that may stabilize the interface and hinder coalescence.

Solvent selection is a multi-criteria decision process that should prioritize the following:

High selectivity for the target solute to minimize the co-extraction of impurities.
Significant distribution coefficient to reduce the required solvent-to-feed ratio and total stage count.
Ease of solvent recovery, typically via distillation, which requires a favorable relative volatility between the solvent and the solute.
Physical compatibility, including low toxicity, low flammability, and minimal solubility in the raffinate phase to prevent product loss.

Worked Example: Deterpenation of Citrus Oil by Multistage Liquid-Liquid Extraction

A citrus oil stream containing 4 wt% terpenes (solute) is to be purified by countercurrent extraction with pure ethanol. The ethanol is essentially immiscible with the oil and has a distribution coefficient K=2.0 for terpenes. The feed flow rate of the carrier phase (oxygenated aromatics) is 1000 kg/h, and a solvent rate of 2000 kg/h is chosen to achieve an extraction factor well above unity. The target raffinate must contain no more than 0.5 wt% terpenes. Tray efficiency is 0.4. Determine the number of theoretical and actual stages required.

Knowns

Carrier (feed) flow rate: B = 1000.0 kg/h
Solute mass fraction in feed: x_F = 0.040
Target solute mass fraction in raffinate: x_R = 0.005
Solvent flow rate: S = 2000.0 kg/h
Distribution coefficient: K = 2.0
Solute mass fraction in fresh solvent: y_S = 0.0
Tray efficiency: η = 0.4

Step-by-Step Calculation

Extraction factor \[ E = \frac{S \cdot K}{B} = \frac{2000.0 \cdot 2.0}{1000.0} = 4.000 \] (The value E = 4.0 satisfies the validity constraint E > 0.6 and is well above the practical guideline of E > 1.2 for efficient separation.)
Intermediate terms for the Kremser equation \[ \text{term}_a = x_F - \frac{y_S}{K} = 0.040 - \frac{0.0}{2.0} = 0.040 \] \[ \text{term}_b = x_R - \frac{y_S}{K} = 0.005 - \frac{0.0}{2.0} = 0.005 \] \[ \text{term}_c = \left(\frac{\text{term}_a}{\text{term}_b}\right) \cdot \left(1 - \frac{1}{E}\right) + \frac{1}{E} = \left(\frac{0.040}{0.005}\right) \cdot \left(1 - \frac{1}{4.0}\right) + \frac{1}{4.0} = 6.250 \] (All provided numbers: term_a = 0.04, term_b = 0.005, term_c = 6.25.)
Theoretical stages (Kremser equation) \[ N_{\text{theo}} = \frac{\ln(\text{term}_c)}{\ln(E)} = \frac{\ln(6.250)}{\ln(4.0)} = 1.322 \] (The result N_theo = 1.322 stages is obtained from the numerical calculation: ln(6.25)/ln(4.0) = 1.83258/1.38629 = 1.32193... rounded to three decimal places.)
Actual stages (applying tray efficiency) \[ N_{\text{actual}} = \frac{N_{\text{theo}}}{\eta} = \frac{1.322}{0.4} = 3.305 \]

Final Answer

The separation requires N_theo = 1.322 theoretical stages. With a tray efficiency of η = 0.4, the actual number of stages needed is N_actual = 3.305 stages, which rounds to 4 stages for a practical column design.

"Un projet n'est jamais trop grand s'il est bien conçu." — André Citroën

"La difficulté attire l'homme de caractère, car c'est en l'étreignant qu'il se réalise." — Charles de Gaulle