Education — Comprehensive Reference

Gas Laws & EOS — Theory, Physics & Engineering Science

In-depth educational content for undergraduate students, practising engineers, and research scholars. Covers the kinetic theory foundations, full mathematical derivations, thermodynamic consistency, phase equilibria, industrial applications, and known failure modes of real-gas equations of state.

KINETIC THEORYSTATISTICAL MECHANICS CUBIC EOS DERIVATIONFUGACITY & CHEMICAL POTENTIAL CORRESPONDING STATESVLE & FLASH DEPARTURE FUNCTIONSINDUSTRIAL APPLICATIONS

Kinetic Theory of Gases

Where the Ideal Gas Law Comes From — First Principles

The ideal gas law PV = nRT is not merely an empirical observation — it is a rigorous consequence of classical statistical mechanics applied to an idealised assembly of non-interacting point particles. Three foundational assumptions underpin it: (1) gas molecules occupy zero volume themselves, (2) there are no intermolecular forces between molecules except during perfectly elastic collisions, and (3) molecular kinetic energy is entirely translational and proportional to absolute temperature.

From these assumptions, the Maxwell-Boltzmann speed distribution follows directly, describing the probability that a molecule has a speed between v and v+dv. Integrating over this distribution gives measurable macroscopic quantities — the pressure, average kinetic energy, and heat capacity:

── Maxwell-Boltzmann speed distribution ────────── f(v) = 4π (M/2πRT)^(3/2) · v² · exp(−Mv²/2RT) Mean speed: v̄ = √(8RT/πM) [m/s] Most probable: v_p = √(2RT/M) RMS speed: v_rms = √(3RT/M) ── Pressure from momentum transfer to walls ──────── P = (1/3) · ρ · v_rms² = nRT/V ✓ ideal gas law ── Internal energy (translational only) ─────────── U = (3/2)nRT (monatomic — 3 translational DoF) U = (5/2)nRT (diatomic — 3 trans + 2 rotational) U = (6/2)nRT (polyatomic — 3 trans + 3 rot)

The idealised assumptions break down when molecules are close together — at high pressure and low temperature. Two physical effects cause real-gas deviations that the ideal law cannot capture:

Intermolecular attractions (van der Waals forces): at moderate separation, molecules attract each other. This pulls molecules slightly toward each other, reducing the effective pressure below the ideal prediction. Correction term: −a/V².
Finite molecular volume (excluded volume): molecules occupy real physical space, so the free volume available for motion is less than the total container volume V. Correction term: V → V−nb.

── When is ideal gas accurate? ───────────────────── SAFE Pr < 0.1 AND Tr > 2.0 → |Z−1| < 1% Air at 25°C, 1 bar: Z = 0.9994 (<0.1% error) Air at 25°C, 100 bar: Z = 0.9950 (0.5% error) CO₂ at 25°C, 50 bar: Z = 0.908 (9.2% error) CH₄ at −100°C, 100 bar: Z = 0.814 (19% error) NH₃ at 100°C, 100 bar: Z = 0.731 (27% error) → Never use ideal gas for refrigerants, LPG, or high-pressure process streams above ~20 bar.

💡 Van der Waals (1873) correctly identified both physical causes of deviation before quantum mechanics or intermolecular potential theory existed. His two-correction model is the conceptual ancestor of every cubic EOS in use today.

Corresponding States Principle

Why Tr and Pr Are Universal — The Pitzer Extension

Van der Waals showed in 1881 that if his EOS is written in reduced coordinates (Tr = T/Tc, Pr = P/Pc, Vr = V/Vc), the equation becomes the same for all substances — the law of corresponding states. This means that two gases with the same Tr and Pr will have the same compressibility factor Z, regardless of what the gases actually are. This universal chart was the primary engineering tool for gas property estimation until the 1970s.

Pitzer (1955) extended the principle by observing that simple spherical molecules (noble gases) obeyed van der Waals' law very accurately, but elongated and polar molecules deviated systematically. He introduced the acentric factor ω to quantify this non-sphericity — defined so that ω = 0 for perfectly spherical molecules:

── Pitzer acentric factor definition ─────────────── ω ≡ −log₁₀(Psat/Pc)|_{Tr=0.7} − 1 ── Physical meaning: departure from simple-fluid ─── Noble gases (Ar, Kr, Xe): ω ≈ 0.000 (perfectly spherical) Nitrogen N₂: ω = 0.037 Methane CH₄: ω = 0.011 Carbon dioxide CO₂: ω = 0.225 Ammonia NH₃: ω = 0.253 Water H₂O: ω = 0.345 n-Decane C₁₀H₂₂: ω = 0.492 (long chain) Higher ω → more elongated or polar molecule → stronger departure from simple-fluid behaviour.

The three parameters (Tc, Pc, ω) fully characterise a pure component in all cubic EOS. Tc and Pc locate the molecule on the universal reduced phase diagram; ω corrects the shape of the vapour-pressure curve. This is why every EOS input form asks for exactly these three parameters:

Gas	Tc (K)	Pc (bar)	ω	Experimental Zc
Helium He	5.20	2.27	−0.390	0.301
Hydrogen H₂	33.15	13.00	−0.219	0.305
Nitrogen N₂	126.2	33.98	0.037	0.290
CO₂	304.2	73.83	0.225	0.274
Water H₂O	647.1	220.6	0.345	0.229
Methane CH₄	190.6	46.10	0.011	0.286
Propane C₃H₈	369.8	42.48	0.152	0.281
Ammonia NH₃	405.6	113.5	0.253	0.242

⚠️ Helium and H₂ have negative ω — they are quantum fluids where zero-point energy is significant even at room temperature. Standard cubic EOS fail for He and cryogenic H₂. Special quantum-corrected critical parameters (Tc,eff, Pc,eff) must be used.

Historical Development

150 Years of Equations of State — Key Milestones

Year	EOS / Author	Key Innovation	Zc	Status today
1662–1834	Boyle, Charles, Gay-Lussac → Ideal Gas PV=nRT	First systematic gas laws; combined into ideal gas law by Clapeyron (1834)	—	Universal for dilute gases
1873	van der Waals	First cubic EOS. Introduced molecular volume (b) and attraction (a/V²). First to predict liquid-vapour phase transition and critical point from a single equation.	0.375	Teaching / historical
1901	Kammerlingh Onnes — Virial EOS	Rigorous statistical-mechanics expansion Z = 1 + B/V + C/V²… Second virial coefficient B has direct molecular interpretation.	—	Research / low density
1949	Redlich-Kwong (RK)	Replaced vdW T-independent attraction with T^(−0.5) dependence. Improved vapour phase accuracy significantly.	0.333	Superseded by SRK
1955	Pitzer — Acentric Factor	Introduced ω enabling 3-parameter corresponding states. Foundation of all modern cubic EOS.	—	Universal parameter
1972	Soave (SRK)	Replaced RK T^(−0.5) with α(Tr,ω) from vapour pressure data. Transformed accuracy of vapour-phase predictions. Industry standard for gas.	0.333	Industry standard
1976	Peng-Robinson (PR) ★	Modified denominator to Vm(Vm+b)+b(Vm−b) → Zc=0.307, dramatically improved liquid density. Gold standard for oil and gas.	0.307	Default EOS worldwide
1982	Peneloux Volume Correction	Additive volume translation c to correct PR/SRK liquid density without affecting VLE. Standard in LNG calculations.	—	Widely used extension
1996–2008	GERG-2004 / GERG-2008	Multifluid Helmholtz energy EOS for natural gas. ISO 20765 standard. Accuracy ±0.03–0.05%.	—	Custody transfer / metering
Ongoing	PC-SAFT, CPA, SAFT-VR	Molecular chain length and association terms. Required for polymers, water, HF, glycols, asphaltenes.	—	Specialised process simulation

💡 The 1976 Peng-Robinson paper is one of the most cited engineering papers ever written (10,000+ citations). It remains the default thermodynamic model in ASPEN Plus, HYSYS, PRO/II, and virtually every commercial process simulator.

Full Derivation — Peng-Robinson

From Physical Assumptions to the Cubic Polynomial in Z

All cubic EOS follow an identical structural template: a repulsion term describing excluded volume pressure, and an attraction term. The Peng-Robinson form differs from SRK only in the denominator of the attraction term — a seemingly small change that improves critical compressibility from Zc = 0.333 (SRK) to Zc = 0.307 (PR), matching experimental values of 0.23–0.29 far better.

── PR EOS: pressure-explicit form ───────────────── P = RT/(Vm−b) − a·α(T) / [Vm(Vm+b) + b(Vm−b)] ↑ repulsion ↑ attraction (PR denominator) ── PR parameters (from critical constraints) ──────── a₀ = 0.45724 · R²Tc²/Pc [Pa·m⁶/mol² — energy scale] b = 0.07780 · RTc/Pc [m³/mol — excluded volume] ── Soave α temperature correction ────────────────── κ = 0.37464 + 1.54226ω − 0.26992ω² α(T) = [1 + κ(1 − √Tr)]² α → 1 at T = Tc α → κ² as T → ∞ ── SRK for comparison ────────────────────────────── P_SRK = RT/(Vm−b) − a·α / [Vm(Vm+b)] m_SRK = 0.480 + 1.574ω − 0.176ω²

To convert the pressure-explicit EOS into a cubic polynomial solvable for molar volume, introduce the dimensionless compressibility factor Z = PVm/RT and dimensionless groups A and B:

── Dimensionless groups A and B ──────────────────── A = a·α·P / (R·T)² [dimensionless] B = b·P / (R·T) [dimensionless] Substitution: Vm = Z·RT/P into PR pressure form → ── PR cubic polynomial in Z ────────────────────── Z³ − (1−B)Z² + (A−3B²−2B)Z − (AB−B²−B³) = 0 ── SRK cubic polynomial in Z ───────────────────── Z³ − Z² + (A−B−B²)Z − AB = 0 ── van der Waals cubic in Z ────────────────────── Z³ − (1+B_vdw)Z² + A_vdw·Z − A_vdw·B_vdw = 0 A_vdw = aP/(RT)², B_vdw = bP/RT

The three different polynomials show why EOS choice matters: the coefficients of Z² and Z are different for each EOS. Even at the same A and B values, the roots will differ — PR gives a lower Zc because its Z² coefficient is (1−B), not 1 as in SRK.

Critical-Point Constraints — Where a₀ and b Come From

Deriving the Universal EOS Constants Analytically

The constants 0.45724 and 0.07780 in the PR EOS are not empirical — they are derived by imposing two thermodynamic constraints that define the critical point on any P-V isotherm: both the first and second partial derivatives of pressure with respect to volume vanish simultaneously:

── Critical point constraints ─────────────────────── (∂P/∂Vm)_{T_c} = 0 [P-V isotherm has inflection at Tc] (∂²P/∂Vm²)_{T_c} = 0 [critical isotherm curvature = 0] ── Solving for PR at T = Tc (α = 1) ──────────────── Two equations in two unknowns (a₀, b): From constraint 1: a₀ = b·RTc·(1+2√2·B_c − B_c²)/B_c From constraint 2: b = 0.07780·RTc/Pc Combined result: a₀ = 0.45724·R²Tc²/Pc ── Critical compressibility predicted ────────────── Z_c (PR) = P_c·V_c/(R·T_c) = 0.3074 ← PR prediction Z_c (SRK) = P_c·V_c/(R·T_c) = 0.3333 ← SRK prediction Z_c (exp) ≈ 0.23 – 0.29 ← experimental range PR is 3× closer to experiment → better liquid density, better near-critical behaviour, better two-phase results.

💡 The improvement in Zc from SRK (0.333) to PR (0.307) is the entire reason Peng and Robinson modified the denominator. The mathematical change is tiny — replacing Vm(Vm+b) with Vm(Vm+b)+b(Vm−b) — but the thermodynamic consequence is a 20% improvement in liquid density accuracy across most industrially relevant conditions.

Solving the Cubic — Roots and Phase State

Three Roots, Thermodynamic Stability, and the Equal-Area Rule

A cubic polynomial has either one or three real roots. For an EOS, the number and physical meaning of the roots depends entirely on where (T,P) falls relative to the two-phase envelope:

── Root count and physical meaning ────────────── 1 real root: single-phase fluid (Tr > 1, or superheated gas, compressed liquid) 3 real roots: inside two-phase region (Tr < 1, P between bubble/dew pressures) 2 real roots: near critical point (repeated root) results are sensitive to input precision ── When 3 roots exist ──────────────────────────── Largest Z → vapour-phase root (physical) Middle Z → thermodynamically UNSTABLE (discard) Smallest Z → liquid-phase root (physical) Middle root has (∂P/∂V)_T > 0 which violates the thermodynamic stability criterion. It is a mathematical artefact — it has no physical meaning whatsoever.

To determine which physical root applies (vapour, liquid, or a two-phase mixture), compute the fugacity coefficient φ for both vapour and liquid roots. At vapour-liquid equilibrium, the fugacities must be equal — this is the Maxwell equal-area construction expressed algebraically:

── VLE condition (Maxwell / fugacity) ─────────── f_vapour = f_liquid φ_V · P = φ_L · P → φ_V = φ_L ── Phase stability test ───────────────────────── If φ_V < φ_L: vapour root is stable → gas phase If φ_V > φ_L: liquid root is stable → liquid phase If φ_V = φ_L: two phases coexist → saturation ── Cardano discriminant ───────────────────────── Δ = 18pq − 4p³ + p²q² − 4q³ − 27r² Δ > 0 → three distinct real roots (two-phase region) Δ = 0 → repeated root (critical point) Δ < 0 → one real root (single phase)

⚠️ Never use the middle root — the (∂P/∂V)_T > 0 region is mechanically unstable and physically unrealisable. Selecting it will produce wrong density, wrong fugacity, wrong enthalpy, and wrong VLE predictions throughout your calculation.

Numerical Root-Finding

How the Calculator Actually Solves Z — Cardano & Newton-Raphson

── Cardano's method for 3-root case (Δ>0) ──────── Depress: Z = t − (1−B)/3 Reduced cubic: t³ + pt + q = 0 p = A − 3B² − 2B − (1−B)²/3 q = 2(1−B)³/27 − (1−B)(A−3B²−2B)/3 + (AB−B²−B³) For Δ>0, three real roots via trigonometric form: t_k = 2√(−p/3) · cos[⅓·arccos(3q/(2p)·√(−3/p)) − 2πk/3] k = 0, 1, 2 → gives all three roots analytically

── Newton-Raphson iteration (1 real root) ────────── Start: Z₀ = 1.0 (gas guess) or Z₀ = B (liquid) f(Z) = Z³ − (1−B)Z² + (A−3B²−2B)Z − (AB−B²−B³) f'(Z) = 3Z² − 2(1−B)Z + (A−3B²−2B) Z_{n+1} = Z_n − f(Z_n)/f'(Z_n) Convergence: typically 3–6 iterations to |ΔZ| < 1×10⁻¹² ── Eigenvalue method (most general) ─────────────── Companion matrix of polynomial → all eigenvalues = all roots simultaneously, no starting-guess needed Used in production EOS libraries (REFPROP, TREND)

Chemical Potential and Fugacity

The Engineer's Thermodynamic Driving Force for Mass Transfer

The condition for equilibrium between any two phases is equal chemical potential μ in each phase. But chemical potential diverges logarithmically as pressure approaches zero, making it inconvenient for practical calculations. G.N. Lewis introduced fugacity f in 1901 as a "corrected pressure" with the same units as pressure that behaves identically to P for ideal gases but correctly captures non-ideal behaviour:

── Chemical potential definition ──────────────────── μ_i = μ°_i + RT·ln(f_i) ── Fugacity defined via departure from ideal ──────── ln(f/P) = ∫₀ᴾ (Z−1)/P dP [at constant T] f = φ · P [fugacity = coefficient × pressure] lim_{P→0} φ = 1 [ideal gas limit: f → P] ── Physical meaning of φ ────────────────────────── φ < 1 → attractive forces dominate → f < P φ > 1 → repulsive forces dominate → f > P φ = 1 → ideal gas → f = P

For cubic EOS, the fugacity integral ∫(Z−1)/P dP has a closed-form analytical result — different for each EOS based on the structure of the attractive-term denominator. The square-root factor √2 in the PR expression comes directly from the quadratic denominator Vm(Vm+b)+b(Vm−b), which requires completing the square during integration:

── PR fugacity coefficient (analytical) ───────────── ln φ = (Z−1) − ln(Z−B) − A/(2√2·B)·ln[(Z+(1+√2)B)/(Z+(1−√2)B)] ── SRK fugacity coefficient ────────────────────── ln φ = (Z−1) − ln(Z−B) − (A/B)·ln(1 + B/Z) ── Why √2 appears in PR ────────────────────────── ∫dV/[V(V+b)+b(V−b)] = ∫dV/[(V+b·(1−√2))(V+b·(1+√2))] ↑ partial fractions introduce √2 factors naturally

💡 The fugacity coefficient φ is the single most important quantity output from an EOS for engineering use. It governs VLE (phase splitting), absorption, extraction, reaction equilibrium, and membrane separations. The entire edifice of process simulation rests on computing φ accurately from an EOS.

Vapour-Liquid Equilibrium (VLE)

Flash Calculations, K-Ratios, and the Rachford-Rice Equation

A flash calculation determines how a feed mixture of composition z_i splits into vapour (mole fraction y_i) and liquid (mole fraction x_i) at specified temperature and pressure. The K-ratio Ki = yi/xi is the key variable, computed directly from the EOS fugacity coefficients for each phase:

── Equilibrium K-ratio from EOS ──────────────────── K_i = y_i / x_i = φ_L,i / φ_V,i ── Rachford-Rice equation (isothermal flash) ──────── g(ψ) = Σ_i z_i(K_i − 1) / [1 + ψ(K_i − 1)] = 0 z_i = feed mole fraction, ψ = vapour mole fraction Solution: ψ ∈ (0,1) for two-phase flash ── Successive substitution algorithm ─────────────── 1. Initialise K_i from Wilson equation (fast estimate): K_i = (Pc_i/P)·exp[5.373(1+ω_i)(1 − Tc_i/T)] 2. Solve Rachford-Rice for ψ (Brent / Regula Falsi) 3. Compute phase compositions: x_i = z_i / [1 + ψ(K_i − 1)] ; y_i = K_i · x_i 4. Call EOS → compute φ_L,i(x) and φ_V,i(y) 5. Update: K_i_new = φ_L,i / φ_V,i 6. Check: Σ_i (ln K_i_new − ln K_i_old)² < 1×10⁻¹⁰ ? YES → converged, NO → go to step 2

3–15Successive-substitution iterations (typical)

2–5Newton-Raphson iterations (near convergence)

1×10⁻¹⁰Convergence tolerance on K-ratio sum of squares

Departure Functions

Computing Enthalpy, Entropy, and Cp from a Single EOS

The real power of a cubic EOS is that it allows computation of all thermodynamic properties — not just Z and density — through the concept of departure functions. A departure function is the difference between the real-gas value of a property and its ideal-gas value at the same T and P. The departure is computed analytically from the EOS, then added to the ideal-gas contribution (from heat capacity data) to give the total property. This is how ASPEN, HYSYS, and REFPROP compute everything from one model:

── Enthalpy departure (PR EOS) ───────────────────── H − H^ig = RT(Z−1) + [a·α − T·d(a·α)/dT]/(2√2·b) · ln[(Z+(1+√2)B)/(Z+(1−√2)B)] where: d(a·α)/dT = a₀·κ·α / (Tc·√Tr) (analytical) ── Entropy departure (PR EOS) ─────────────────────── S − S^ig = R·ln(Z−B) − [d(a·α)/dT]/(2√2·b) · ln[(Z+(1+√2)B)/(Z+(1−√2)B)] − R·ln(P/P°) ── Heat capacity departure ────────────────────────── Cp − Cp^ig = Cv_dep + T·(∂P/∂T)²_V / (−(∂P/∂V)_T) − R (involves second derivatives of EOS) ── Joule-Thomson coefficient ──────────────────────── μ_JT = (∂T/∂P)_H = −(1/Cp)·(∂H/∂P)_T = −V·(T·(∂V/∂T)_P − V) / Cp [sign determines cooling/heating on throttle]

💡 Departure functions are why a single EOS can consistently predict Z, density, enthalpy, entropy, heat capacity, speed of sound, Joule-Thomson coefficient, and phase envelopes. Thermodynamic consistency is guaranteed because all properties are derived from the same Helmholtz energy function A(T,V).

Oil & Gas — Pipeline and Metering

Z-Factor in Custody Transfer, Compression, and Gas Processing

The natural gas industry is the single largest user of EOS calculations in the world. Every custody transfer metering station continuously computes Z-factor to convert measured volumetric flow at operating conditions to standard conditions. Regulatory frameworks (ISO 20765 / AGA-8) mandate use of GERG-2008 for high-accuracy applications, but Peng-Robinson remains standard for wellsite, pipeline, and separator calculations.

── Standard volume conversion via Z ──────────────── Q_sc = Q_act · (P_act/P_sc) · (T_sc/T_act) · (Z_sc/Z_act) Standard conditions: 15°C (288.15 K), 1.01325 bar Z_sc ≈ 1.000 (standard conditions) Z_act from PR or GERG-2008 at pipeline T and P ── Compressor polytropic head ─────────────────────── H_poly = Z_avg·R·T₁/(M) · n/(n−1) · [(P₂/P₁)^((n−1)/n) − 1] Z_avg directly scales required compressor power. 1% error in Z → ~1% error in motor sizing. At 10 MW compressor: 1% = 100 kW = significant cost.

Dew-point calculations are critical in gas gathering and processing — liquid dropout in pipelines causes slugging, corrosion, and metering errors. The hydrocarbon dew point is computed entirely from EOS-based VLE at operating P and T:

── Dew-point condition ────────────────────────────── At dew point: Σ_i (y_i / K_i) = 1.0 where K_i from PR EOS fugacity ── Phase envelope key points ──────────────────────── Bubble curve: Σ_i (x_i·K_i) = 1.0 (first vapour appears) Dew curve: Σ_i (y_i/K_i) = 1.0 (first liquid appears) Cricondentherm: max T where two phases exist Cricondenbar: max P where two phases exist ── Hydrate and water dew-point ────────────────────── Hydrate risk T from Hammerschmidt: T_h = f(P, gas composition) MEG / methanol injection shifts T_h below pipeline T. EOS models water content of gas for dehydrator design.

⚠️ For LNG and cryogenic processing (T < −100°C), standard PR with classical mixing rules is insufficiently accurate. Use PR with Peneloux volume correction or GERG-2008. For H₂S-rich or CO₂-rich acid gas, BIP (kij) tuning is mandatory.

Chemical Process Design

EOS in Distillation, Reactors, Heat Exchangers & Separations

Process simulation platforms (Aspen Plus, HYSYS, PRO/II, gPROMS, UniSim) use EOS as the thermodynamic backbone for every unit operation. Every distillation tray equilibrium, every heat exchanger Q-T curve, every reactor energy balance depends on EOS-computed H, S, Cp, and fugacities. The single most consequential modelling decision in process simulation is the choice of thermodynamic model:

PR or SRK: Non-polar hydrocarbons and light gases. Gas processing, refinery, cracker, olefin plants. Pressures to 200 bar.
PR with Peneloux correction: LNG, NGL fractionation, dense-phase CO₂ transport. Better liquid density.
CPA (Cubic Plus Association): Water, methanol, glycols (MEG, TEG), amines. Hydrogen bonding dominates — plain PR fails badly.
PC-SAFT: Polymers, heavy oils, asphaltenes, wax appearance. Chain-length effects are dominant.
eNRTL / MSA-EOS: Electrolyte systems (acid gas scrubbing, MEA amine treating, caustic wash).
GERG-2008: Custody transfer, pipeline simulations where ±0.05% accuracy on Z is required.

Application	Recommended EOS	Key concern
Dry natural gas pipeline	PR or GERG-2008	Z-factor accuracy
NGL fractionation	PR + Peneloux	Liquid density
LNG liquefaction	PR or GERG-2008	Liquid enthalpy / MHX design
Ammonia synthesis loop	PR or SRK	H₂, N₂ at 150–300 bar
CO₂ capture & compression	PR + kij tuning	CO₂–water interaction
Acid gas injection	PR with CO₂-H₂S kij	Dense-phase miscibility
Glycol dehydration (TEG)	CPA or NRTL-RK	Water activity
Amine gas treating (MEA)	eNRTL or Kent-Eisenberg	Electrolyte chemistry
Refrigeration cycle	PR or NIST REFPROP	Refrigerant Cp, enthalpy
Supercritical CO₂ EOR	PR with kij regression	MMP prediction

Mixing Rules and Binary Interaction Parameters

The Critical Importance of kij — Extending EOS to Mixtures

Pure-component EOS parameters must be combined for mixtures using mixing rules. The classical van der Waals one-fluid mixing rules are used by default, but the binary interaction parameter kij is an empirical correction for unlike-pair cross-attractions that differ from the geometric mean. Wrong kij values are the most common source of large EOS errors in industrial applications:

── van der Waals one-fluid mixing rules ──────────── a_mix = Σ_i Σ_j x_i·x_j·√(a_i·a_j)·(1 − k_ij) b_mix = Σ_i x_i·b_i k_ij = binary interaction parameter k_ij = 0 : identical-type non-polar pairs (CH₄-C₂H₆) k_ij = 0.10–0.15: CO₂-hydrocarbon (important!) k_ij = 0.05–0.10: H₂S-hydrocarbon k_ij = 0.10–0.20: N₂-hydrocarbon ── Wong-Sandler mixing rules (advanced) ───────────── Combines cubic EOS with Gibbs energy model (NRTL, UNIQUAC) Required for: water+organics, alcohols+hydrocarbons, acids a_WS/(b·RT) = Σ_i Σ_j x_i·x_j·(a_ij/(b_ij·RT)) / (1 − Gex/(RT·CW)) ── kij regression from VLE data ───────────────────── Minimise: F = Σ_k [(P_calc,k − P_exp,k)/P_exp,k]² Sources: DECHEMA VLE Databank, NIST ThermoData Engine, in-house lab measurements

💡 For CO₂-CH₄ systems at pipeline conditions, using kij=0 instead of kij=0.10 can introduce 3–8% error in the equilibrium K-ratio — meaning the dew-point temperature prediction is off by 5–15°C. This is operationally significant for pipeline design and condensate separation.

Limitations — Where Cubic EOS Fail

Critical Failure Modes Every Engineer Must Know

── Problematic fluid types ────────────────────────── Water H₂O: Strong H-bonding → up to 30% liquid density error, 10–20% VLE error with PR/SRK. Use CPA or PC-SAFT. Alcohols: Self-associating → similar errors to water. Methanol in gas streams needs CPA. HF acid: Extreme association (forms hexamers in vapour). Plain cubic EOS fundamentally fails. Amines (MEA/DEA): Association + ionisation → use eNRTL. Electrolytes: Ions present → Debye-Hückel or MSA needed. Polymers: Very high MW, chain effects → PC-SAFT. Helium / H₂: Quantum effects at cryogenic T → use quantum-corrected Tc_eff, Pc_eff.

── Problematic conditions ─────────────────────────── Near critical point (|Tr−1| < 0.05, |Pr−1| < 0.1): → All cubic EOS give divergent Cp, incorrect density. → Use crossover EOS (Kiselev) or scaled equations. Very low Tr (< 0.5): → Standard Soave-type α function becomes unreliable. → Use Mathias-Copeman or Twu α function. Very high Pr (> 20–30): → Excluded volume b underestimates molecular repulsion. → BACK, SAFT-VR, or PC-SAFT perform better. Solid phases: → Cubic EOS model gas + liquid only. → Solid-fluid equilibrium needs separate triple-point model. Strong association / solvation: → CPA (1 EOS parameter per associating site) or → PC-SAFT (chain + dispersion + association terms).

⚠️ Never use the ideal gas law for: pipeline gas above 20 bar, any refrigerant, liquefied gases, supercritical fluids, or any application where density accuracy better than ±5% is needed. Z can deviate from 1.0 by 5–30% in normal industrial conditions.

Complete Reference

Symbols, Terms, and Definitions — Gas EOS Engineering

Compressibility Factor Z

Z = PVm/RT [—]

Dimensionless measure of deviation from ideal-gas behaviour. Z=1 for ideal gas. Z<1 when attractive forces dominate (most gases at moderate conditions). Z>1 when repulsive forces dominate (high-pressure gases, H₂, He). The primary output of all EOS calculations; used to correct all flow and volume calculations for real-gas behaviour.

Critical Temperature Tc

K or °C

Temperature above which no amount of pressure can liquefy a gas. At T > Tc the fluid is supercritical — gas and liquid phases merge into a single fluid phase. The critical point (Tc, Pc) is the unique state where the densities of coexisting vapour and liquid become equal. All cubic EOS parameters are derived from Tc and Pc.

Critical Pressure Pc

bar or Pa

Minimum pressure needed to liquefy a gas at exactly Tc. With Tc, it defines reduced coordinates (Tr, Pr) for corresponding-states correlations. High-Pc fluids (H₂O: 220.6 bar; NH₃: 113.5 bar; CO₂: 73.8 bar) require much higher pressure to reach the critical point than low-Pc gases (N₂: 34.0 bar; CH₄: 46.1 bar).

Reduced Temperature Tr

Tr = T/Tc [—]

Dimensionless temperature normalised to Tc. The foundation of corresponding-states correlations — fluids at the same Tr and Pr have the same Z (for simple fluids). Tr < 1: subcritical (vapour and liquid can coexist). Tr > 1: supercritical (single fluid phase, no phase boundary). Most accurate EOS behaviour occurs for Tr = 0.8–2.0.

Acentric Factor ω

ω = −log₁₀(Psat/Pc)−1 at Tr=0.7

Pitzer's non-sphericity parameter. Defined so that ω = 0 for perfectly spherical monatomic molecules (Ar, Kr, Xe). Increases with molecular chain length and polarity. Enters SRK as m, enters PR as κ — controls the temperature dependence of the α function. Higher ω → stronger Tr-dependence of the attraction parameter a·α.

Fugacity f

f = φ·P [Pa or bar]

Lewis's "corrected pressure" — the thermodynamic driving force for mass transfer between phases and across membranes. At equilibrium, fugacity is equal in all coexisting phases for each component: f_i^V = f_i^L. Reduces to pressure for ideal gas (φ=1). Computed analytically from any cubic EOS via the departure integral. The basis for all VLE and reaction equilibrium calculations.

Fugacity Coefficient φ

φ = f/P [—]

Ratio of fugacity to pressure. φ=1 for ideal gas. φ<1 when attractive forces dominate (most gases below ~50 bar, Z<1). φ>1 when repulsive forces dominate (high-pressure gases). The VLE equilibrium criterion is φ_V = φ_L. Each cubic EOS has a unique closed-form analytical expression for φ derived from the departure integral.

Molar Volume Vm

m³/mol or L/mol

Volume per mole of substance: Vm = ZRT/P. Ideal gas at 25°C, 1 bar: Vm = 24.79 L/mol. Liquid propane at −40°C, 1 bar: Vm ≈ 0.088 L/mol (282× denser). The primary root of the cubic EOS polynomial. Three roots can exist in the two-phase region; the largest is the vapour root, the smallest is the liquid root.

EOS Parameters a and b

a [Pa·m⁶/mol²], b [m³/mol]

Van der Waals attraction (a) and excluded-volume (b) parameters. In all cubic EOS, a₀ ∝ R²Tc²/Pc (energy scale) and b ∝ RTc/Pc (size scale). Numerically different between EOS: PR uses a₀=0.45724·R²Tc²/Pc vs SRK's 0.42748; b=0.07780·RTc/Pc vs SRK's 0.08664. Both differences follow from enforcing the critical-point constraints for each specific EOS form.

Soave α Function α(Tr,ω)

α = [1 + κ(1−√Tr)]²

Temperature-correction function introduced by Soave (1972) to improve vapour-pressure predictions. Fitted to experimental vapour pressures via the acentric factor. α=1 at T=Tc; α increases as T decreases (stronger attraction at lower T). The key innovation over the original Redlich-Kwong EOS, transforming it from a research curiosity into an industrial standard. Mathias-Copeman and Twu α functions provide better accuracy at very low Tr.

Binary Interaction Parameter kij

kij ∈ [0, 0.2] typical

Empirical correction to the geometric-mean combining rule for unlike-pair cross-attraction a_ij = √(a_i·a_j)·(1−kij). kij=0 for similar non-polar pairs (CH₄-C₂H₆). kij=0.10–0.15 for CO₂-hydrocarbon; 0.05–0.10 for H₂S-hydrocarbon. Must be regressed from VLE data. The dominant source of EOS mixture error when improperly set. Not physically derivable from pure-component parameters alone.

Departure Function

(Property)_real − (Property)^ig

Difference between a real-gas property (H, S, Cp, etc.) and its ideal-gas value at the same T and P. Computed analytically from EOS integral expressions involving ∂P/∂T and ∂P/∂V. Added to ideal-gas contribution (from Cp° heat-capacity data) to give the total real-gas property. Guarantees thermodynamic consistency — all properties derived from the same Helmholtz energy A(T,V).

Phase Envelope (P-T diagram)

Bubble + dew loci

The locus of all bubble points (first vapour appears from liquid) and dew points (first liquid appears from vapour) on a P-T diagram, enclosing the two-phase region. Critical point is at the apex. Cricondentherm is the maximum temperature at which two phases can coexist; cricondenbar is the maximum pressure. Essential for separator design, pipeline sizing, and gas processing plant design.

Virial EOS

Z = 1 + B/Vm + C/Vm² + …

Series expansion in molar density derived rigorously from statistical mechanics. Second virial coefficient B represents two-body molecular interactions; C represents three-body interactions. Theoretically exact at low density. Truncated at B: excellent below ~10 bar for most gases. B can be calculated from intermolecular potential theory (Lennard-Jones), providing a direct link between molecular physics and thermodynamic properties.

GERG-2008 / ISO 20765

Multifluid Helmholtz EOS

Groupe Européen de Recherches Gazières multifluid EOS for natural gas. Uses residual Helmholtz energy formulation with 21 natural gas components. Accuracy: ±0.03–0.05% for Z-factor in pipeline range (270–450 K, 0–300 bar). ISO 20765 standard for custody transfer metering. Far more complex than PR or SRK but required where ±0.1% Z accuracy is legally specified for billing purposes.

Peneloux Volume Correction

Vm,corr = Vm,EOS − c

Simple additive volume translation (Peneloux, 1982) to correct the systematic liquid-density overprediction of SRK and PR. c is a component-specific constant regressed from saturated liquid density data. Does not affect vapour-liquid equilibrium (K-ratios, dew and bubble points) because fugacity expressions are invariant to volume translation. Standard in LNG and NGL calculations where liquid density accuracy to ±1% is required.

CPA — Cubic Plus Association

A = A_cubic + A_assoc

EOS combining a cubic (SRK) term with an association term from Wertheim's theory (1984). The association term explicitly models hydrogen-bonding between molecules such as water, alcohols, glycols (MEG, DEG, TEG), and amines. Each associating site is characterised by energy ε^A and volume κ^A. Required for glycol dehydration, amine treating, CO₂-water systems, and any application where hydrogen bonding significantly affects phase behaviour.

Corresponding States Principle

Z = Z(Tr, Pr, ω)

Van der Waals' 1881 theorem that all pure fluids obey the same EOS when expressed in reduced coordinates (Tr, Pr). Extended by Pitzer to three-parameter form using ω. Foundation of all generalised correlations (Lee-Kesler, Pitzer-Curl, AGA-8) and all modern cubic EOS. Fails for quantum fluids (H₂, He) and strongly associating fluids (water, HF) where molecule-specific interactions are dominant.

Gas Laws & EOS Calculator

Awaiting Calculation

P–V–T Isotherm Charts

All EOS — Same Conditions, Four Results

Click Compare All

Theory, Equations & Derivations

Gas Laws & EOS — Theory, Physics & Engineering Science