Skip to content
Back to search results

UNIFAC

Introduced by
Aage Fredenslund, Russell Jones, John Prausnitz
Year
1975
Type
Activity coefficient model; predictive liquid-phase property method
Family
Thermodynamic Modeling
Font size

What does it do?

UNIFAC (Universal Functional-group Activity Coefficient) is a predictive model for liquid-phase activity coefficients of multicomponent mixtures.

Developed by Fredenslund, Jones, and Prausnitz in 1975, it decomposes molecules into functional groups and uses group interaction parameters to estimate non-ideal behavior. UNIFAC is revolutionary because it can predict phase equilibria for mixtures never experimentally measured, making it invaluable for process design and chemical engineering.

The intuition behind the method

Rather than measuring properties of every possible mixture experimentally, UNIFAC leverages a simple insight: the thermodynamic behavior of a molecule depends on its functional groups (CH3, OH, C=O, etc.), not on the specific molecule. By collecting interaction parameters between functional groups from available data, one can predict properties of any new mixture. The model separates activity coefficients into combinatorial (size and shape effects) and residual (interactions) contributions, each handled by proven submodels.

History

Fredenslund, Jones, and Prausnitz published UNIFAC in 1975, building on the UNIQUAC model and group contribution philosophy. Gmehling's modified UNIFAC (1993) improved parameters for polar compounds. UNIFAC-FV (Free Volume) extends the model to polymers. Today, variants like COSMO-SAC use more sophisticated quantum-chemical input.

When to use it

Use UNIFAC to predict liquid-phase properties of organic mixtures when experimental data are unavailable. It is ideal for preliminary process design, feasibility studies, and screening candidate solvents. Apply to nonpolar and slightly polar compounds. Avoid for hydrogen-bonding systems, aqueous solutions, polymers, and ionic liquids unless specialized UNIFAC variants are used (UNIFAC-FV, modified UNIFAC, COSMO-SAC).

Limitations and considerations

  • Limited to organic and weakly polar compounds; fails for water, alcohols, and strongly associating systems
  • Accuracy depends on group parameter quality; some rare functional groups have poorly fitted parameters
  • Cannot predict effects of temperature on group interactions well; requires variants like UNIFAC-T
  • Does not account for molecular size explicitly; may fail for very large molecules

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Predictive without experimental mixture data; saves time and resources
  • Mechanistic basis (group contributions); physically interpretable
  • Works for infinite dilution and concentrated solutions
  • Widely available; implemented in commercial process simulators

Applications

  • Designing separation columns (distillation, extraction) for novel solvent mixtures
  • Predicting mutual solubility of organic solvents for process design
  • Screening alternative solvents for green chemistry applications
  • Computing equilibrium compositions in multiphase systems without experimental data

How it works

How it works

1

Identify the molecular structures and decompose each molecule into its constituent functional groups.

2

Count the number of each functional group per molecule; look up group surface area (Q) and volume (R) parameters.

Formula
\text{Decompose molecule into functional groups}
3

Compute the combinatorial contribution to activity coefficients from group size and shape.

Formula
\ln \gamma_i = \ln \gamma_i^C + \ln \gamma_i^R
4

Calculate the residual contribution using group interaction parameters (A_mn) from the UNIFAC database; sum both contributions.

Formula
\gamma_i^R = Q_i \left( 1 - \ln \sum_k Q_k \Theta_k - \sum_k \frac{\Theta_k L_{mk}}{\sum_n \Theta_n L_{nk}} \right)
5

Use activity coefficients to predict vapor-liquid equilibria, flash calculations, and separation efficiency.

Common pitfalls

  • Applying UNIFAC to systems outside its domain (aqueous, hydrogen-bonding); use specialized models instead
  • Using outdated group parameters; modern versions (UNIFAC-FV, modified UNIFAC) have better parameters
  • Misidentifying functional groups; incorrect decomposition leads to large errors

Frequently asked questions

Why can't UNIFAC predict water behavior accurately?

Water has strong hydrogen bonding, which is not captured by UNIFAC's group interaction approach. Specialized models (NRTL-HB, UNIFAC-FV with hydrogen bonding terms) are needed for aqueous solutions.

How do I handle a functional group not in the UNIFAC database?

Use the closest chemically similar group or estimate parameters from similar groups. Alternatively, use more general models (COSMO-SAC) that derive activity coefficients from quantum-chemical calculations.

What is the difference between UNIFAC and modified UNIFAC?

Modified UNIFAC (Gmehling) has refined interaction parameters from additional experimental data, improving accuracy for polar and aromatic compounds. The model structure is the same.

Related methods

Sources

  1. Fredenslund, A., Jones, R. L., & Prausnitz, J. M. (1975). Group-contribution estimation of activity coefficients in nonideal liquid mixtures. AIChE Journal, 21(6), 1086-1099.

    Open DOI reference
    Copied.
  2. Prausnitz, J. M., Lichtenthaler, R. N., & de Azevedo, E. G. (2001). Molecular Thermodynamics of Fluid-Phase Equilibria (3rd ed.). Prentice Hall.

    ISBN: 978-0-13-977745-6
    Copied.
  3. Gmehling, J., Li, J., & Schiller, M. (1993). A modified UNIFAC model for the prediction of phase equilibria and excess enthalpies. Industrial & Engineering Chemistry Research, 32(1), 178-193.

    Open publication reference
    Copied.