Calculator D4

Membrane Separation Fundamentals: Solution-Diffusion Model and Permeability

Membrane separation works like a super-selective sieve: molecules dissolve into the membrane, move through it by diffusion, and pop out the other side — only certain ones get through.

Industry Applications
Natural gas sweetening, seawater desalination, hydrogen recovery from refinery purge gas, pharmaceutical solvent recycling, dairy whey concentration
Key Standards
ASTM D4580 (membrane permeability testing), ISO 4022 (RO membrane performance), SEMI F63 (ultrapure water membranes)
Typical Scale
RO plants: 10,000–500,000 m³/day; gas separation skids: 10–50,000 Nm³/h

⚠️ Why It Matters

1
Inaccurate permeability estimation
2
Under-designed membrane area
3
Failure to meet product purity specs
4
Excessive energy consumption for pressure or vacuum
5
Premature membrane fouling or replacement
6
Unprofitable plant operation

📘 Definition

The solution-diffusion model describes permeation through dense, non-porous membranes as a three-step process: (1) selective sorption (dissolution) of solute or solvent at the upstream interface, (2) Fickian diffusion driven by a concentration gradient across the membrane matrix, and (3) desorption at the downstream interface. Permeability (P) is the product of solubility (S) and diffusivity (D), quantifying the intrinsic transport capacity of a specific component in a given membrane–feed system.

🎨 Concept Diagram

Feed Side (High Concentration)Membrane MatrixPermeate Side (Low Concentration)SorptionDesorptionDiffusion

AI-generated illustration for visual understanding

💡 Engineering Insight

Permeability is not a fixed material constant — it’s a system property that shifts with feed composition, pressure history, and thermal cycling. A membrane rated at P = 100 GPU for H₂ at 25°C will lose >40% effective P after 100 h exposure to 5% H₂S due to sulfide-induced polymer relaxation. Always validate P under *actual* feed conditions — not just pure-component lab data.

📖 Detailed Explanation

At its core, the solution-diffusion model treats the membrane as a homogeneous, dense phase where separation arises not from pore size exclusion (like in microfiltration), but from differences in how readily molecules dissolve into and move through the polymer. Think of sugar dissolving in honey: smaller, more compatible molecules (e.g., helium) both 'like' the membrane material and slip between its chains easily — they have high S and D. Larger or polar-unmatched molecules (e.g., SO₂ in silicone rubber) may dissolve well but diffuse slowly, or vice versa.

Beyond ideal behavior, real systems exhibit coupling effects: co-permeation of water and ethanol in pervaporation swells the membrane, increasing D but decreasing α. Plasticization — where highly soluble gases (e.g., CO₂) act like molecular lubricants — causes time-dependent flux rise and selectivity collapse. This is why industrial models embed dual-mode sorption (Langmuir + Henry) and concentration-dependent D to capture nonlinearity.

At the frontier, advanced membranes decouple S and D via nanostructuring: MOF-polymer mixed matrix membranes introduce rigid, size-sieving pores that boost D without sacrificing S, while surface-grafted ionic liquids tune S via chemisorption. Yet even these rely on solution-diffusion as the governing framework — the model endures because it correctly identifies thermodynamic affinity and kinetic mobility as the two irreducible levers of separation.

🔄 Engineering Workflow

Step 1
Step 1: Define separation objective (target purity, recovery, throughput)
Step 2
Step 2: Screen membrane materials using ternary solubility parameters & group contribution methods (e.g., UNIFAC-FV)
Step 3
Step 3: Measure pure-component P and mixed-gas α under representative T, P, and composition
Step 4
Step 4: Model module performance (e.g., CSTR or plug-flow) incorporating concentration polarization & compaction
Step 5
Step 5: Size membrane area and staging configuration using economic optimization (CAPEX vs. OPEX tradeoff)
Step 6
Step 6: Integrate with upstream/downstream units (e.g., compressors, condensers, pretreatment)
Step 7
Step 7: Commission with stepwise pressure ramp and on-stream monitoring of flux decay & salt passage

📋 Decision Guide

Rock/Field Condition Recommended Design Action
High-pressure CO₂/CH₄ feed (>70 bar), acidic gas presence Use glassy polyimide membranes with crosslinking; apply acid gas pretreatment; limit recovery to <85% to avoid plasticization
Seawater desalination (35,000 ppm NaCl), ambient feed temp (25°C) Select thin-film composite (TFC) polyamide RO membrane; operate at 55–65 bar; maintain ΔP < 1.5 bar across elements to minimize compaction
Pharmaceutical solvent recovery (acetone/water, 90/10 wt%), low vapor pressure ratio Deploy pervaporation with hydrophilic PVA/PVAl composite membrane; use vacuum sweep (<50 mbar); pre-concentrate via distillation to <40% water

📊 Key Properties & Parameters

Permeability Coefficient (P)

10⁻¹² – 10⁻⁶ cm³(STP)·cm/(cm²·s·cmHg) for gases; 10⁻⁹ – 10⁻⁶ L/(m²·h·bar) for liquids

Product of solubility coefficient (S) and diffusion coefficient (D); measures steady-state flux per unit driving force (typically partial pressure difference).

⚡ Engineering Impact:

Directly determines required membrane area and system capital cost — errors >20% cascade into 30–50% overdesign or underperformance.

Selectivity (α)

2–100 for gas pairs (e.g., O₂/N₂), 10–1000 for water/salts in RO membranes

Ratio of permeability coefficients of two components (e.g., α_A/B = P_A / P_B), indicating intrinsic separation capability.

⚡ Engineering Impact:

Dictates minimum stage count and whether single-stage operation is feasible — low α forces multi-stage or hybrid (e.g., membrane + distillation) designs.

Diffusion Coefficient (D)

10⁻⁸ – 10⁻⁵ cm²/s for small gases in polymeric membranes (e.g., H₂: ~10⁻⁶ cm²/s in cellulose acetate)

Measure of molecular mobility within the polymer matrix, governed by free volume, chain flexibility, and penetrant size.

⚡ Engineering Impact:

Controls response time to feed fluctuations and limits maximum practical crossflow velocity — low D increases concentration polarization severity.

Solubility Coefficient (S)

10⁻⁴ – 10⁻¹ cm³(STP)/cm³·cmHg (gases); 10⁻³ – 10² g solute/g polymer (liquids, e.g., water in PDMS)

Equilibrium concentration of penetrant in the membrane phase per unit upstream partial pressure (gas) or activity (liquid).

⚡ Engineering Impact:

Dominates temperature dependence of P — high S often correlates with plasticization risk and loss of selectivity above Tg.

📐 Key Formulas

🎨 Technical Diagrams

Feed PhaseMembranePermeate PhaseSorptionDesorption
C₁C₂Diffusion PathFree Volume

📚 References