Chemical Thermodynamics Interview Questions

The First Law states that energy cannot be created or destroyed, only converted from one form to another: dU = dQ - dW, where dU is internal energy change, dQ is heat added, and dW is work done by the system. For chemical engineering, it is the basis for energy balances around processes, sizing heat exchangers, calculating heating/cooling requirements, and analyzing power cycles. For open systems: dH = dQ - dWs (shaft work), making enthalpy the key property.

The Second Law states that entropy of an isolated system never decreases; for any spontaneous process, total entropy (system + surroundings) increases. It establishes the direction of natural processes and limits on energy conversion efficiency. Entropy (S) measures disorder or unavailable energy. Implications: heat flows from hot to cold spontaneously, 100% heat-to-work conversion is impossible, and minimum work required for separation equals the entropy of mixing. It defines the maximum theoretical efficiency of heat engines (Carnot efficiency).

Enthalpy (H) is defined as H = U + PV, combining internal energy with flow work (PV). For open systems at constant pressure, enthalpy change equals heat transferred: dH = dQ. This makes enthalpy the natural property for flow processes - heat exchangers, reactors, and separation equipment. Specific enthalpy (h) tables/charts are used for steam, refrigerants, and other process fluids. Enthalpy differences (delta_H) drive heating/cooling calculations and appear in energy balances around equipment.

The ideal gas law states PV = nRT, where P is pressure, V is volume, n is moles, R is universal gas constant (8.314 J/mol-K), and T is absolute temperature. It assumes molecules have no volume and no intermolecular forces. Valid when: pressure is low (typically <5-10 bar), temperature is high (well above critical), and gas is not near condensation. Deviations are quantified by compressibility factor Z = PV/nRT; ideal gas has Z=1. Real gases deviate, requiring equations of state for accurate calculations.

Gibbs free energy (G) is defined as G = H - TS, combining enthalpy and entropy effects. At constant temperature and pressure, the change in G determines spontaneity: dG < 0 indicates spontaneous process, dG = 0 indicates equilibrium, dG > 0 indicates non-spontaneous (requires work input). For chemical reactions: dG = dG_standard + RT*ln(Q), where Q is reaction quotient. At equilibrium, dG = 0, giving dG_standard = -RT*ln(K). Gibbs energy minimization is used to calculate chemical and phase equilibrium.

Vapor-liquid equilibrium (VLE) is the state where liquid and vapor phases coexist with no net mass transfer - the rates of evaporation and condensation are equal. At VLE, temperature, pressure, and chemical potential (or fugacity) of each component are equal in both phases. VLE data is essential for designing distillation columns, flash drums, evaporators, and any separation involving phase change. VLE behavior is described by Raoult's law (ideal) or activity coefficient models (non-ideal).

Fugacity (f) is an adjusted pressure that accounts for non-ideal behavior - it represents the escaping tendency of a substance. For an ideal gas, fugacity equals pressure; for real gases and liquids, fugacity = phi * P (phi is fugacity coefficient). At phase equilibrium, fugacities are equal: f_vapor = f_liquid. Using fugacity instead of pressure allows the same equilibrium equations to work for both ideal and non-ideal systems. Fugacity coefficient phi captures deviation from ideal gas behavior.

Activity coefficient (gamma) corrects for non-ideal liquid mixture behavior: a = gamma * x, where a is activity and x is mole fraction. For ideal solutions, gamma = 1. Non-ideal behavior arises from molecular size differences and interaction forces. Activity coefficients are used in modified Raoult's law: y*P = gamma*x*P_sat. Values can be >1 (positive deviation, less stable mixture) or <1 (negative deviation, more stable). Needed for polar, hydrogen-bonding, or dissimilar molecules. Calculated from models like NRTL, UNIQUAC, or Wilson.

Heat of reaction (delta_H_rxn) is the enthalpy change when reactants convert to products at specified conditions. Calculated from: delta_H_rxn = sum(n*H_products) - sum(n*H_reactants), using heats of formation or combustion. Standard conditions: 25C, 1 atm, species in standard states. Negative delta_H indicates exothermic (releases heat); positive indicates endothermic (absorbs heat). Temperature adjustment uses heat capacities: delta_H(T) = delta_H_298 + integral(delta_Cp*dT). Essential for reactor energy balances and safety analysis.

A P-T diagram shows phase stability regions (solid, liquid, gas) as functions of pressure and temperature. Key features: vapor pressure curve (liquid-gas boundary ending at critical point), sublimation curve (solid-gas boundary), fusion curve (solid-liquid boundary), triple point (all three phases coexist), and critical point (above which liquid and gas are indistinguishable). It helps determine phase at given conditions, boiling/condensation points at various pressures, and whether a substance can exist as liquid at given conditions.

Critical properties (Tc, Pc, Vc) define the critical point where liquid and vapor become identical. Critical temperature (Tc): above this, no liquid exists regardless of pressure. Critical pressure (Pc): pressure at critical point. Critical volume (Vc): molar volume at critical point. Importance: gases cannot be liquefied above Tc by pressure alone, critical properties are used in equations of state (reduced properties), corresponding states principle relies on critical properties, and supercritical fluids (T>Tc, P>Pc) have unique properties useful for extraction.

Heat capacity is the amount of heat required to raise the temperature of a substance by one degree. Cp (constant pressure): C = dH/dT, heat required at constant pressure - energy goes into both internal energy and PV work. Cv (constant volume): C = dU/dT, heat required at constant volume - all energy goes to internal energy. For gases: Cp - Cv = R (ideal gas), Cp/Cv = gamma (heat capacity ratio, ~1.4 for air). For liquids/solids, Cp approximately equals Cv. Cp is used in most process calculations (constant pressure operations).

Latent heat is the energy absorbed or released during phase change at constant temperature. Latent heat of vaporization: energy to convert liquid to vapor (e.g., water ~2257 kJ/kg at 100C). Latent heat of fusion: energy to convert solid to liquid. Latent heat of sublimation: solid to vapor directly. In process calculations: sizing evaporators and condensers, calculating steam requirements, evaluating refrigeration cycles, and drying operations. Latent heat varies with temperature, typically decreasing as temperature approaches critical point (zero at Tc).

The Clausius-Clapeyron equation relates vapor pressure to temperature: d(ln P_sat)/dT = delta_H_vap/(R*T^2), or integrated form: ln(P2/P1) = (delta_H_vap/R)*(1/T1 - 1/T2). It shows that vapor pressure increases exponentially with temperature. Uses: estimate vapor pressure at different temperatures from one data point and heat of vaporization, construct vapor pressure curves, and understand boiling point elevation and depression. Assumes constant heat of vaporization and ideal gas vapor - Antoine equation is more accurate for precise calculations.

At chemical equilibrium, forward and reverse reaction rates are equal, and Gibbs energy is minimized. The equilibrium constant K is related to standard Gibbs energy change: dG_standard = -RT*ln(K). Temperature dependence follows van't Hoff equation: d(ln K)/dT = delta_H_rxn/(R*T^2). Exothermic reactions: K decreases with temperature; endothermic: K increases. Pressure affects equilibrium position through Le Chatelier's principle but doesn't change K. Understanding these relationships guides reactor design and operating condition selection.

Van der Waals (1873): P = RT/(V-b) - a/V^2; first to include molecular volume (b) and attraction (a). Simplest but inaccurate for liquid densities. SRK (Soave-Redlich-Kwong, 1972): improves attraction term with temperature-dependent alpha function based on acentric factor. Better vapor pressures but poor liquid densities. Peng-Robinson (1976): modified SRK with different critical compressibility, better liquid density prediction (Zc=0.307). PR widely used in petroleum/gas industry. Both SRK and PR are cubic, solving for three roots (vapor, liquid, or single phase).

Flash calculation determines the amounts and compositions of vapor and liquid phases when a mixture is brought to specified T and P. For PT-flash: given feed composition z, T, P, find vapor fraction V/F, vapor composition y, and liquid composition x. Rachford-Rice equation: sum(z_i*(K_i-1)/(1+V/F*(K_i-1))) = 0, solved iteratively for V/F. K-values (y/x) from VLE correlation or equation of state. Material balance gives compositions. Used for separator design, flash drums, and any phase separation modeling.

NRTL (Non-Random Two-Liquid): uses local composition concept with non-randomness parameter alpha (typically 0.2-0.5). Has three parameters per binary pair, handles highly non-ideal systems including LLE, but alpha is often uncertain. UNIQUAC (Universal Quasi-Chemical): based on lattice theory with surface and volume fractions. Two parameters per binary pair, more theoretically based, foundation for UNIFAC predictive method. Both handle strongly non-ideal systems. NRTL often better for LLE; UNIQUAC better for polymer solutions. Selection depends on available data and system type.

UNIFAC (Universal Functional Activity Coefficient) is a predictive method for activity coefficients using group contributions - molecules are divided into functional groups with known interaction parameters. No experimental data needed for the specific system, only the group parameters. Useful when: no experimental VLE data available, screening many systems, and preliminary process design. Limitations: less accurate than fitted models (NRTL, UNIQUAC), poor for electrolytes and polymers, and group parameter availability varies. Modified UNIFAC (Dortmund) has improved accuracy.

Mixing rules combine pure component EOS parameters for mixtures. Classical van der Waals mixing rules: a_mix = sum_i*sum_j(x_i*x_j*a_ij), b_mix = sum(x_i*b_i), with a_ij = sqrt(a_i*a_j)*(1-k_ij). Binary interaction parameter k_ij is fitted to VLE data (typically 0-0.2). More sophisticated: Wong-Sandler rules (ensure correct second virial coefficient), Huron-Vidal (incorporate activity coefficient model at high pressure). Mixing rules are critical - same EOS with different mixing rules gives very different predictions. k_ij selection significantly impacts VLE accuracy.

A phase envelope (or phase diagram) for a mixture shows the two-phase region on a P-T plot. Bounded by: bubble point curve (liquid just starts vaporizing), dew point curve (vapor just starts condensing), and cricondentherm/cricondenbar (maximum temperature/pressure for two-phase). Construction: at each T, calculate bubble pressure; at each T, calculate dew pressure; plot the curves. Critical point is where bubble and dew curves meet. Used for: pipeline design (avoid two-phase), separator operating conditions, and understanding mixture phase behavior. Generated by process simulators using EOS.

At azeotrope: y_i = x_i for all components, meaning alpha = 1 (relative volatility equals unity). For binary with modified Raoult's law: gamma_1*P1_sat = gamma_2*P2_sat at azeotrope. Conditions favoring azeotrope formation: similar volatilities (P_sat ratio near 1) and strong non-ideality (gamma values significantly different from 1). To identify: plot y-x diagram (azeotrope where curve crosses diagonal), or solve gamma_1*P1_sat = gamma_2*P2_sat for composition. Can be minimum-boiling (positive deviation) or maximum-boiling (negative deviation, less common).

Joule-Thomson effect is temperature change during isenthalpic (constant enthalpy) throttling. Joule-Thomson coefficient: mu_JT = (dT/dP)_H. For most gases at normal conditions, mu_JT > 0 (cooling on expansion); for hydrogen and helium at room temperature, mu_JT < 0 (heating on expansion). Inversion temperature: where mu_JT = 0. Important in: refrigeration cycles (expansion cooling), natural gas processing (cooling can cause hydrate formation), cryogenic processes, and pressure letdown systems (temperature prediction for material selection).

Departure functions express the difference between real fluid properties and ideal gas properties at the same T and P: H_real - H_ideal, S_real - S_ideal. Calculated from equation of state via integrals involving compressibility factor or volume. Usage: Real property = Ideal gas property + Departure + Reference state contribution. This approach separates ideal gas calculations (known analytically) from non-ideal corrections (from EOS). Enables consistent property calculations using EOS. Important for accurate enthalpy and entropy in high-pressure gas systems.

Steady-state energy balance: Q - W = sum(n_out*H_out) - sum(n_in*H_in), including heat of reaction via enthalpies of formation. Steps: calculate inlet enthalpy (sensible heat from reference), calculate outlet enthalpy including product species, heat of reaction appears as enthalpy difference. For exothermic: outlet enthalpy > inlet, Q negative (heat removal needed). For adiabatic: Q=0, solve for outlet temperature. Include: phase changes, mixing effects if significant, and temperature-dependent properties. Reference state consistency is critical.

LLE occurs when a liquid mixture separates into two liquid phases with different compositions. Condition: component fugacities equal in both phases: x_i^alpha*gamma_i^alpha = x_i^beta*gamma_i^beta. Requires strongly non-ideal behavior (gamma >> 1). Modeled using activity coefficient models (NRTL, UNIQUAC) with parameters fitted to LLE data - these may differ from VLE parameters. Type I systems have one pair of partially miscible components; Type II have two pairs. Important for extraction, product purification, and understanding mixture stability.

Excess properties quantify deviation from ideal solution behavior: G_E = G_real - G_ideal mixing. For ideal solution, G_E = 0. Activity coefficient is related to partial molar excess Gibbs energy: RT*ln(gamma_i) = G_E_bar_i. Total excess Gibbs energy: G_E = RT*sum(x_i*ln(gamma_i)). Excess enthalpy (H_E) gives heat of mixing; excess volume (V_E) gives volume change on mixing. Activity coefficient models (NRTL, UNIQUAC, Wilson) are essentially correlations for G_E. Measuring H_E and G_E provides data for model fitting.

Acentric factor (omega) characterizes the non-sphericity of molecules, defined as: omega = -log10(P_sat/Pc) - 1 at T_r = 0.7. Simple fluids (Ar, Kr, Xe) have omega near 0; complex molecules have higher values (e.g., n-decane ~0.49). Uses: three-parameter corresponding states correlations (improve over two-parameter), alpha function in SRK and PR equations of state, Lee-Kesler correlation for compressibility, and property estimation methods. Higher omega indicates molecules with more complex structure or stronger polar interactions.

Thermodynamic consistency tests verify that experimental VLE data satisfies fundamental relationships. Area test (Redlich-Kister): integral of ln(gamma_1/gamma_2) vs x should equal zero for consistent data. Point test: verify activity coefficients satisfy Gibbs-Duhem equation (sum of x_i*d_ln_gamma_i = 0 at constant T, P). Herrington test: modified area test with temperature correction. Inconsistent data may indicate experimental errors, impure components, or incomplete equilibration. Consistency checking is essential before using data for parameter fitting.

Selection guidelines: Light hydrocarbons/gases: use equation of state (PR, SRK). Polar/non-ideal liquids at low pressure: activity coefficient models (NRTL, UNIQUAC). High-pressure polar: combine activity coefficient with EOS (e.g., PSRK). Aqueous electrolytes: electrolyte-NRTL. Polymers: polymer-NRTL or PC-SAFT. Steam/water: specialized steam tables (NBS, IAPWS). Consider: component types, pressure range, presence of electrolytes, available data for parameter fitting. Validate against experimental data for key separations. Most simulators provide selection wizards based on system type.

Steam tables provide properties of water/steam at various conditions. Saturation tables: properties at vapor-liquid equilibrium (T_sat or P_sat given). Superheated tables: properties of vapor above saturation. Compressed liquid tables: subcooled liquid properties. Key properties: specific volume (v), internal energy (u), enthalpy (h), entropy (s). Interpolation for non-tabulated values. Usage: energy balances (using enthalpy), available energy calculations (using entropy), sizing (using volume). For mixtures, use quality (x = m_vapor/m_total): h_mix = h_f + x*h_fg.

Henry's Law: p_i = H_i * x_i for dilute solutions of gas in liquid. Henry's constant H depends on temperature and solvent. Determination: experimental measurement of gas solubility at known partial pressures, or correlation (e.g., from molecular properties). Temperature dependence: ln(H) = A + B/T (typically increases with T, meaning lower solubility). Applications: absorption column design, dissolved gas calculations, environmental partitioning, and VLE for dilute volatile components. Different conventions exist (pressure/mole fraction, molality-based), so units must be carefully tracked.

Compressibility factor Z = PV/nRT measures deviation from ideal gas. From EOS: rearrange to form cubic in Z (for cubic EOS like PR, SRK). PR equation: Z^3 - (1-B)Z^2 + (A-3B^2-2B)Z - (AB-B^2-B^3) = 0, where A = aP/R^2T^2, B = bP/RT. Solve cubic for Z: three roots possible (vapor Z largest, liquid Z smallest, middle root unstable). Use appropriate root for phase. For mixtures, use mixing rules to get a_mix and b_mix. Generalized correlations (Lee-Kesler) also give Z from reduced properties.

Methods include: Antoine equation: log(P) = A - B/(T+C), most common, limited temperature range per parameter set. Extended Antoine: additional terms for wider range. Lee-Kesler: generalized correlation using Tc, Pc, omega. Riedel: group contribution method. Wagner equation: high accuracy over wide range, used in databases. Selection: Antoine for quick estimates within range, Wagner for accuracy, Lee-Kesler when only critical properties known. Always verify against experimental data for critical applications. Process simulators use stored parameters.

Exergy is the maximum useful work obtainable when a system equilibrates with its environment (dead state). Exergy = enthalpy - T0*entropy (relative to dead state). Unlike energy, exergy is destroyed in irreversible processes. Exergy analysis identifies: thermodynamic inefficiencies (where exergy destruction occurs), improvement opportunities (reduce irreversibility), and relative performance of process alternatives. Applications: power plant efficiency analysis, heat integration assessment, and second-law efficiency evaluation. Exergy destruction is proportional to entropy generation times dead state temperature.

The phi-phi approach uses equations of state for both phases: y_i*phi_V*P = x_i*phi_L*P, where fugacity coefficients are calculated from EOS for vapor and liquid. Best for: high pressure systems, near-critical conditions, light hydrocarbons, and systems where both phases are well-described by the EOS. The gamma-phi approach uses activity coefficient for liquid and EOS for vapor: y_i*phi_V*P = x_i*gamma_i*P_sat*phi_sat. Best for: low to moderate pressure, polar/non-ideal liquids, and when reliable activity coefficient model parameters exist. Phi-phi is thermodynamically consistent at all conditions but requires good EOS parameters; gamma-phi handles liquid non-ideality better but fails near critical points.

PC-SAFT (Perturbed-Chain Statistical Associating Fluid Theory) is a molecular-based EOS that models molecules as chains of spherical segments. Parameters: segment number (m), segment diameter, dispersion energy, and association parameters for hydrogen bonding. Advantages over cubic EOS: fundamentally better for polymers and long-chain molecules, handles associating fluids (alcohols, acids, amines) more accurately, better liquid density predictions, and more physically meaningful parameters. Disadvantages: more complex, more parameters, slower computation, and fewer available parameter databases. Used in polymer processing, pharmaceutical applications, and complex chemical systems.

Thermodynamic model errors propagate significantly in distillation design. K-value errors affect: relative volatility (alpha) directly impacting required stages, minimum reflux calculation, and feed tray location. A 5% error in K-values can lead to 10-20% error in stage requirements. Azeotrope prediction errors can result in completely wrong separation schemes. Temperature profile errors affect heat duties and condenser/reboiler sizing. Mitigation strategies: validate model against experimental VLE data for key binaries, check azeotrope composition predictions, perform sensitivity analysis on K-values, and add design margin (10-15% extra stages). Regressing activity coefficient parameters to available data is essential for non-ideal systems.

Electrolyte solutions require special treatment due to ionic dissociation and long-range electrostatic interactions. Key considerations: chemical equilibria (dissociation, speciation), Debye-Huckel term for long-range ion-ion interactions, and short-range interactions (NRTL-type). Common models: Electrolyte-NRTL (e-NRTL) in Aspen, Pitzer model for aqueous systems, OLI for comprehensive speciation. Parameters needed: dissociation constants (temperature-dependent), ion-pair interaction parameters, and reference state (infinite dilution). Applications: acid gas treating (amine systems), chlor-alkali, and pharmaceutical crystallization. Mean ionic activity coefficient replaces component activity coefficient for ionic species.

Retrograde condensation occurs when a gas mixture forms liquid when pressure decreases (opposite of normal behavior). It happens in the retrograde region between cricondentherm and critical point on the phase envelope. Mechanism: at constant temperature, reducing pressure from single-phase gas crosses dew point curve, forming liquid that increases to a maximum then decreases. Impact on gas processing: liquid dropout in pipelines during pressure reduction, condensate accumulation in separators not designed for liquids, two-phase flow in unexpectedly designed single-phase systems, and composition variation through processing. Design considerations: avoid retrograde region or design for two-phase handling.

Two main approaches: stoichiometric method - write reactions explicitly, express compositions in terms of extent of reaction, solve K expressions simultaneously. Becomes intractable for many reactions/species. Non-stoichiometric (Gibbs minimization) - minimize total Gibbs energy subject to element balance constraints: min(sum n_i*mu_i) subject to atom balances. Solved using Lagrange multipliers; no need to specify reactions. Algorithms: RAND method (successive substitution), Newton-Raphson with line search. Handling: condensed phases included with appropriate activity, ideal gas or EOS for vapor. Process simulators use element-balance approach (RGibbs in Aspen). Provides compositions, phase distribution, and equilibrium temperature/pressure.

Parameter regression fits model parameters to minimize difference between calculated and experimental values. Objective function: sum of squared deviations in pressure, temperature, or composition, weighted by experimental uncertainty. Data types: isothermal VLE (T, P, x, y), isobaric VLE, infinite dilution activity coefficients, excess enthalpy, LLE tie lines. Procedure: select model, initialize parameters (from similar systems or UNIFAC prediction), minimize objective using nonlinear optimization (Levenberg-Marquardt), check convergence and parameter correlation. Validation: check consistency tests, predict data not used in regression, examine residual patterns. Tools: Aspen Properties regression, DECHEMA DDB, ThermoFit.

Near-critical challenges: large property fluctuations and long correlation lengths not captured by classical EOS, density differences between phases vanish creating convergence difficulties, classical EOS has wrong critical exponents (mean-field behavior), liquid and vapor roots of cubic EOS merge making phase identification difficult, and numerical instabilities in flash calculations. Mitigation: use specialized near-critical formulations (crossover models), implement robust phase stability analysis (tangent plane criterion), use continuation methods for flash calculations, apply regularization techniques, and avoid operating in near-critical region if possible. Some advanced EOS (PC-SAFT with crossover) handle critical region better.

SLE calculation: at equilibrium, component fugacity is equal in solid and liquid phases. For pure solid: x*gamma*exp(-delta_H_fus/R*(1/T - 1/T_m)) = 1, where delta_H_fus is heat of fusion, T_m is melting point. For solid solutions, include solid phase activity coefficient. Eutectic systems: each component crystallizes independently; eutectic point is where solubility curves intersect. Heat capacity difference affects temperature dependence. Applications: crystallization process design (yield, purity, operating conditions), freeze concentration, and pharmaceutical polymorphism. Complications: solid polymorphism, hydrate/solvate formation, and kinetic effects (supersaturation, metastable zones).

Property routes define which methods and models calculate each thermodynamic property. Configuration levels: equation of state selection (base method), transport property correlations (viscosity, thermal conductivity), reaction chemistry (equilibrium constants), and special handling (electrolytes, polymers). Consistency requirements: same EOS should calculate all equilibrium-related properties, reference states must match between enthalpy and Gibbs energy, and mixing rules should be consistent with phase equilibrium model. Common errors: mismatched activity coefficient and Henry's law models, inconsistent reference states causing energy balance errors, and using wrong method for property type (e.g., ideal gas Cp for liquid). Always validate property predictions against data.

Phase stability analysis determines whether a mixture is single-phase or will split into multiple phases. Tangent plane criterion: a phase is stable if the tangent plane to Gibbs energy surface at the feed composition lies below all other points on the surface. Mathematically: minimize TPD = sum(y_i*(mu_i(y) - mu_i(feed))) over all trial compositions y. If minimum TPD < 0, phase is unstable and will split; trial composition gives approximate equilibrium composition. Implementation: multiple starting points needed (pure components, random), global optimization or successive substitution. Essential for: reliable flash calculations, identifying three-phase equilibrium, and avoiding false single-phase solutions.

Supercritical fluids (T > Tc, P > Pc) have properties between liquid and gas: liquid-like density and dissolving power, gas-like diffusivity and viscosity. Modeling approach: cubic EOS (PR, SRK) with appropriate parameters, including binary interaction parameters fitted to high-pressure data. Solubility modeling: use fugacity coefficient from EOS, often correlated with density. Key considerations: strong pressure sensitivity near critical, CO2 co-solvent effects, temperature crossover behavior (solubility may increase or decrease with T depending on P). Applications: supercritical CO2 extraction (caffeine, essential oils), supercritical water oxidation, and particle formation (RESS, SAS). Validate against high-pressure experimental data.

Pinch analysis uses thermodynamic principles to set targets for energy recovery. First law: energy balance determines total heating and cooling duties. Second law: heat flows from hot to cold, imposing temperature driving force constraint (delta_T_min). Composite curves: plot cumulative enthalpy vs. temperature for hot and cold streams; closest approach is the pinch. Above pinch: heat sink (needs external heating only); below pinch: heat source (needs external cooling only). Thermodynamic targets: minimum hot utility = gap at top of composite curves; minimum cold utility = gap at bottom. Grand composite curve shows cascade of heat between temperature intervals. Correct heat exchanger network design achieves thermodynamic targets.

Starting from fundamental relation: ln(phi) = integral from 0 to P of (Z-1)/P dP. For cubic EOS (PR as example): ln(phi) = (Z - 1) - ln(Z - B) - A/(2*sqrt(2)*B) * ln((Z + (1+sqrt(2))*B)/(Z + (1-sqrt(2))*B)), where A = aP/R^2T^2, B = bP/RT. For mixtures, use mixing rules for a_mix, b_mix and partial properties: ln(phi_i) = (b_i/b_mix)*(Z-1) - ln(Z-B_mix) - A_mix/(2*sqrt(2)*B_mix) * (2*sum(y_j*a_ij)/a_mix - b_i/b_mix) * ln((Z + (1+sqrt(2))*B_mix)/(Z + (1-sqrt(2))*B_mix)). This enables VLE calculation using EOS for both phases.

Irreversibility sources in chemical processes: heat transfer across finite delta_T (dominant in heat exchangers), mixing of streams at different compositions or temperatures, throttling (pressure reduction without work recovery), chemical reaction away from equilibrium, and friction/pressure drop. Identification methods: exergy analysis (exergy destruction = T0 * S_gen), pinch analysis for heat exchange, entropy generation calculation. Minimization strategies: reduce temperature driving forces (more heat transfer area), use counter-current instead of co-current contactors, recover work from pressure letdown (expanders/turbines), use reactive distillation to couple reaction with separation, and minimize mixing of dissimilar streams. Economic trade-off: reducing irreversibility requires capital investment; optimize total cost.