Thermodynamic Models For Real Gases That Break Ideal Assumptions
- 01. Why ideal gas models fail
- 02. Core modeling strategies for real gases
- 03. Virial equation of state
- 04. van der Waals equation
- 05. Redlich-Kwong and Soave-Redlich-Kwong
- 06. Peng-Robinson equation
- 07. Other real-gas equations of state
- 08. Quantifying deviation: compressibility factor and Boyle temperature
- 09. Illustrative comparison of real-gas models
- 10. Practical selection of a real-gas model
- 11. Historical evolution and modern context
The most widely used thermodynamic models for real gases replace the ideal gas law with corrected equations of state-such as the virial expansion, van der Waals, Redlich-Kwong, Soave-Redlich-Kwong, and Peng-Robinson-that explicitly account for finite molecular size and intermolecular forces, enabling accurate prediction of properties like compressibility, phase behavior, and critical phenomena at high pressures and low temperatures where the ideal model fails. These real-gas equations systematically "break" ideal assumptions by adding empirical or semi-theoretical parameters fitted to experimental data, and are standard tools in chemical, petroleum, and power-plant engineering for design and simulation of equipment such as compressors, pipelines, and liquefaction trains.
Why ideal gas models fail
Ideal gas models assume that molecules have negligible volume and do not interact, which means the pressure-volume-temperature (PVT) relationship can be written simply as $$pV = nRT$$ without correction terms. In practice, real gas behavior always deviates from this ideal law once measurements are precise enough, so engineers define a compressibility factor $$Z = pV/(nRT)$$ to quantify how far a gas departs from ideality under a given temperature and pressure.
Two conditions are particularly important for understanding why the ideal-gas approximation breaks down: high pressure, where the finite size of molecules makes their own volume comparable to the container volume, and low temperature, where attractive forces pull molecules together and alter the pressure. For common gases such as nitrogen or oxygen at room temperature and 1 bar, deviations from ideality are often under 0.1%, but at pressures of tens to hundreds of bar, $$Z$$ can differ from 1 by tens of percent, severely impacting calculations of work, enthalpy, and phase equilibrium if ideal models are used.
Core modeling strategies for real gases
The main strategies for modeling real gases thermodynamically are: virial expansions in powers of density or pressure, cubic equations of state like van der Waals and Peng-Robinson, and more complex multiparameter correlations for high-accuracy property databases. Each modeling approach trades mathematical simplicity against accuracy and range of validity, so industrial workflows often use simple cubic equations for process design and more detailed correlations only for critical safety or custody-transfer calculations.
- Virial equations of state expand the compressibility factor as $$Z = 1 + B(T)\rho + C(T)\rho^{2} + \dots$$ to connect macroscopic behavior to pairwise and many-body interactions between molecules.
- Cubic equations of state like van der Waals, Redlich-Kwong, Soave-Redlich-Kwong, and Peng-Robinson express pressure as rational functions of molar volume with a small number of empirical constants tuned to critical properties.
- Multiparameter formulations and reference equations, used in national and industrial standards, may employ dozens of fitted coefficients to match high-resolution PVT data over wide ranges of thermodynamic states.
In practice, process simulators typically allow users to choose among several equations of state, with default recommendations depending on fluid type-Peng-Robinson or Soave-Redlich-Kwong for hydrocarbon mixtures, virial-based models for low-density gases, and specialized formulations for refrigerants or polar mixtures. This flexibility reflects decades of research showing that no single thermodynamic model reliably covers all combinations of temperature, pressure, and composition encountered in modern energy and chemical systems.
Virial equation of state
The virial equation of state is conceptually the most direct way to "correct" the ideal gas law by expressing the compressibility factor as a power series: $$Z = 1 + B(T)\frac{p}{RT} + C(T)\left(\frac{p}{RT}\right)^{2} + \dots$$ or equivalently in density form, where $$B$$, $$C$$, etc. are virial coefficients that depend only on temperature. These virial coefficients have microscopic interpretations linked to intermolecular potentials-for example, the second virial coefficient encodes the balance between attractive and repulsive forces in molecular pair interactions.
Because the virial series converges only at relatively low densities, it is especially useful for accurate modeling of slightly nonideal gases at moderate pressures, such as in high-precision calorimetry or gas metrology near standard conditions. In 2000, thermodynamic studies of mixing "slightly imperfect" gases defined by a truncated virial expression $$Z = 1 + Bp/(RT)$$ were used to derive consistent expressions for excess Gibbs energy, enthalpy, and volume in binary mixtures, demonstrating how virial-based models can still underpin quantitative real gas thermodynamics for mixtures.
van der Waals equation
The van der Waals equation, proposed in 1873, was the first successful attempt to include both molecular size and intermolecular attraction in an analytic equation of state: $$\left(p + a/V_{m}^{2}\right)(V_{m} - b) = RT$$, where $$a$$ corrects for attractions and $$b$$ for excluded volume. This simple cubic equation qualitatively reproduces key features of real fluids, including the existence of a critical point and the S-shaped isotherms associated with liquid-vapor coexistence, a breakthrough that helped earn van der Waals the Nobel Prize in 1910.
In modern practice, the van der Waals model is rarely used for quantitative design, because its predictions can deviate substantially from experimental PVT data, especially near the critical region and for complex mixtures. However, the structure of the van der Waals equation remains foundational, and many later equations of state can be viewed as refined cubic forms that inherit its basic conceptual split between repulsive volume exclusion and attractive energy contributions.
Redlich-Kwong and Soave-Redlich-Kwong
The Redlich-Kwong equation, introduced in 1949, improves on van der Waals by modifying the temperature dependence of the attractive term, using a form like $$p = \frac{RT}{V_{m}-b} - \frac{a}{T^{1/2}V_{m}(V_{m}+b)}$$, which fits real-gas data more accurately at moderate and high temperatures. This Redlich-Kwong formulation became a workhorse in early computational thermodynamics because it balances improved realism with algebraic simplicity suitable for manual or low-power digital calculation.
In 1972, Soave introduced a modification-now known as the Soave-Redlich-Kwong (SRK) equation-that replaced the simple $$T^{-1/2}$$ dependence with an "alpha function" tuned to the acentric factor of each component, dramatically improving vapor-liquid equilibrium predictions for non-ideal hydrocarbons. The SRK model quickly became standard in petroleum and petrochemical engineering, and by the late 1970s was embedded in process design methods for refineries and gas processing plants, where it remains widely used for phase behavior calculations up to several hundred bar.
Peng-Robinson equation
The Peng-Robinson equation of state, published in 1976, further refined cubic EOS modeling by targeting both accurate liquid density and vapor pressure prediction, with a form $$p = \frac{RT}{V_{m}-b} - \frac{a\alpha(T)}{V_{m}(V_{m}+b) + b(V_{m}-b)}$$. This Peng-Robinson equation is particularly well-suited to natural gas and crude oil systems, where it can predict phase envelopes and critical points within a few percent of experimental values across wide ranges of pressure and temperature.
By the 1990s, surveys of industrial practice reported that more than 70% of large hydrocarbon process simulations in the oil and gas sector relied on either Peng-Robinson or SRK as the primary process EOS, reflecting their robustness and ease of implementation in numerical flash algorithms. Today, most commercial simulators default to Peng-Robinson for upstream gas processing and LNG design, while providing options to switch to SRK or specialized models when handling highly non-ideal or polar mixtures.
Other real-gas equations of state
Beyond the classic cubic EOS family, additional models like Dieterici, Berthelot, Beattie-Bridgeman, and Benedict-Webb-Rubin offer intermediate complexity with extra fitting parameters to improve accuracy for particular fluid classes. These extended equations remain important in niche applications and historical literature but are often superseded in modern software by more flexible multiparameter Helmholtz-energy formulations that can accommodate dozens of terms and reproducing experimental property surfaces within about 0.1%.
Reference-quality property standards for substances such as water, CO₂, and common refrigerants rely on highly tuned correlations that go far beyond cubic EOS, often developed under the auspices of national or international organizations and updated periodically as new high-precision data become available. In these cases, the governing equations are not just simple PVT relations but complete thermodynamic formulations from which all properties-entropy, enthalpy, sound speed, and more-are derived via exact thermodynamic identities.
Quantifying deviation: compressibility factor and Boyle temperature
A central diagnostic for real-gas behavior is the compressibility factor $$Z$$, defined by $$Z = pV/(nRT)$$, which equals 1 for a perfect ideal gas and differs from 1 whenever real-gas effects matter. Plots of $$Z$$ versus pressure at fixed temperature show how compressibility factor curves dip below 1 when attractive forces dominate at moderate pressures and rise above 1 at high pressures when repulsive volume exclusion becomes more important.
The Boyle temperature is defined as the temperature at which the second virial coefficient becomes zero, so that $$Z \approx 1$$ over a relatively wide pressure range. For many nonpolar gases, the Boyle temperature is several times higher than the critical temperature, which explains why gases near ambient conditions often appear "ideal" despite having significant attractive interactions that only become obvious at lower temperatures or higher pressures.
Illustrative comparison of real-gas models
To make the conceptual differences concrete, consider a hypothetical comparison of how several real-gas models handle a simple nonpolar gas across a moderate pressure range at fixed temperature. The table below sketches plausible relative errors in compressibility factor predictions compared with experimental data, highlighting why some equations of state are preferred in modern engineering practice.
| Equation of state | Typical Z error at 50 bar | Typical Z error at 150 bar | Main application focus |
|---|---|---|---|
| Virial (truncated to B) | ~0.5% low | >5% low | Low-density gas metrology |
| van der Waals | 2-5% | 10-20% | Qualitative teaching |
| Redlich-Kwong | 1-3% | 5-10% | High-temperature gases |
| Soave-Redlich-Kwong | 1-2% | 3-8% | Hydrocarbons, VLE |
| Peng-Robinson | <1-2% | 3-6% | Natural gas, crude oils |
This kind of comparative data guides the selection of a working equation in real-world projects, with engineers often choosing the simplest model that keeps Z and key derived properties within an acceptable error band for the operating envelope of interest. Even when official property packages provide many options, experience and industry guidelines tend to converge on a small subset of models for each fluid class, improving reproducibility and safety in design.
Practical selection of a real-gas model
In practice, choosing a thermodynamic model for real gases is a workflow decision that balances fluid type, pressure-temperature range, property requirements, and available data. For example, a pipeline engineer designing a natural-gas transmission system at 80 bar might use Peng-Robinson for phase behavior and compressibility, while a metrology lab characterizing a calibration gas at 5 bar could use a virial-based standard for higher accuracy.
- Define the fluid system (pure gas, light hydrocarbon mixture, polar components, refrigerant, etc.).
- Map the operating envelope (expected ranges of temperature, pressure, and composition).
- Identify required properties (only Z, or also enthalpy, entropy, phase equilibrium, speed of sound, etc.).
- Select candidate models (e.g., SRK vs Peng-Robinson vs reference correlation) based on standards and prior practice.
- Validate against available experimental or literature data and adjust interaction parameters if necessary.
Many large energy companies maintain internal guidelines that prescribe which thermodynamic models are approved for particular categories of calculations, often requiring more sophisticated reference equations for custody transfer and safety-critical relief design while allowing simpler cubic EOS for preliminary process simulations. This layered approach mirrors academic best practice, where cubic EOS are taught as baseline tools and more detailed models are introduced when precision or complex phase behavior becomes important.
Historical evolution and modern context
The progression from the ideal gas law to virial expansions, van der Waals, Redlich-Kwong, and Peng-Robinson mirrors a century and a half of increasing experimental precision and computational capability. Early thermodynamic pioneers were constrained to simple analytic forms, whereas late-20th-century researchers could exploit numerical optimization and large datasets to fit many-parameter equations capable of closely reproducing entire PVT surfaces.
By the early 2000s, comprehensive assessments of thermodynamic models for design and analysis of power and refrigeration cycles concluded that no single model could dominate across all fluids and conditions, recommending a portfolio approach with explicit validation for each new application. As decarbonization efforts accelerate and new working fluids such as supercritical CO₂, alternative refrigerants, and hydrogen-rich mixtures become more important, the need for accurate real-gas modeling is only increasing, pushing ongoing development of next-generation equations of state and data-driven hybrid approaches.
What are the most common questions about Thermodynamic Models For Real Gases That Break Ideal Assumptions?
What are the main thermodynamic models for real gases?
The main thermodynamic models for real gases include virial equations of state, classical cubic equations such as van der Waals, Redlich-Kwong, Soave-Redlich-Kwong, and Peng-Robinson, and more complex multiparameter correlations used in reference property formulations. These model families collectively allow engineers and scientists to capture deviations from ideal behavior across a wide range of temperatures, pressures, and compositions, from near-ambient laboratory conditions to supercritical industrial processes.
When is the ideal gas law no longer accurate?
The ideal gas law becomes inaccurate at high pressures, where molecular volumes are no longer negligible, and at low temperatures, where intermolecular attractions significantly alter pressure and compressibility. Under such conditions, real gas effects can drive the compressibility factor Z away from 1 by several tens of percent, making ideal-gas-based predictions unreliable for engineering design, safety analysis, and energy efficiency calculations.
Which equation of state should I use in process simulation?
For hydrocarbon-dominated systems such as natural gas and crude oil, Peng-Robinson or Soave-Redlich-Kwong are commonly recommended defaults because they give good vapor-liquid equilibrium and density predictions over a broad operating range. For high-precision work with simple gases at low to moderate pressures, virial-based models or official reference equations are preferred, while specialized property packages are needed for strongly polar fluids, electrolytes, or refrigerant blends.
How do real-gas models handle mixtures?
Most real-gas equations of state handle mixtures by applying mixing rules to the pure-component parameters and introducing binary interaction coefficients that are fitted to experimental phase-equilibrium or PVT data. These mixing rules enable cubic EOS and related models to describe complex multi-component systems, but their accuracy depends strongly on the availability and quality of mixture data, especially for highly non-ideal or polar combinations.