title: How Activated Carbon Adsorbs: Mechanism, Pore Diffusion, Competition and Breakthrough description: A step-by-step engineering explanation of activated carbon adsorption — physisorption, pore diffusion, competitive adsorption, breakthrough and regeneration — with quantified data from Chinese and international sources and their confidence ratings.

How Activated Carbon Adsorbs: From Physisorption to Breakthrough

Direct answer: Activated carbon adsorbs pollutants in five linked steps — (1) physical adsorption onto pore surfaces by van der Waals forces, (2) diffusion of molecules through the pore network, (3) competition with water vapor and co-pollutants for the same sites, (4) progressive saturation of the bed expressed as a breakthrough curve, and (5) regeneration by supplying the energy adsorption released. The two numbers that govern engineering design — dynamic adsorption capacity and breakthrough time — are determined by the slowest of these steps under your specific conditions, not by the carbon's static surface area alone.

1. Step 1: Physical adsorption — the surface does the work

Activated carbon removes gas or liquid molecules primarily by physisorption: van der Waals/London dispersion forces bind molecules to the pore surface without changing their electronic structure. Binding energies are small (5–10 kJ/mol for gases on carbon), so the process is reversible — raising temperature or lowering pressure releases the adsorbate. This reversibility is both a strength (regeneration is possible) and a limit (capacity falls as temperature rises).

Three properties of the adsorbate dominate:

Boiling point / volatility — high-boiling, heavy molecules are adsorbed far more strongly than light, volatile ones; this is why activated carbon polishes VOCs easily but cannot hold methane or hydrogen at ambient temperature.
Molecular size — a molecule must fit the pore; iodine (≈0.27 nm) indicates micropore capacity, carbon tetrachloride (≈0.55 nm) probes slightly larger pores, methylene blue probes mesopores.
Polarity — polar molecules (alcohols, ketones) interact with oxygen-containing surface groups; the degree of surface oxidation changes capacity for polar species.

Chemisorption — a chemical reaction between adsorbate and surface (impregnated carbons removing H₂S, NH₃, mercury, or radioiodine) — is a different regime: stronger binding, more selective, and harder to regenerate. Standard VOC/water purification duty is dominated by physisorption; impregnated carbons are a special class (Impregnated Carbon).

2. Step 2: Pore diffusion — the transport network

Molecules travel in three stages: film diffusion across the external boundary layer → transport through macropores and mesopores (the "highways") → adsorption in micropores (<2 nm, the "storage"). IUPAC classification: micropore <2 nm, mesopore 2–50 nm, macropore >50 nm.

The engineering consequences:

Particle size sets kinetics. Powdered carbon (PAC, mostly <100 μm) has diffusion paths thousands of times shorter than granular carbon (0.5–4 mm), so PAC adsorbs far faster per unit mass — at the cost of being a one-shot material in water treatment.
Mesopores are the delivery network. A carbon with enormous micropore volume but few mesopores can show poor dynamic performance — molecules cannot reach the storage fast enough. This is why the mesopore volume matters for fast-cycling gas duty.
Temperature accelerates diffusion. Higher temperature shortens the time to saturation because diffusion coefficients rise — but simultaneously lowers equilibrium capacity (§4). Kinetics improves; thermodynamics worsens.

3. Step 3: Competitive adsorption — humidity is a competitor, not an atmosphere

Water vapor competes with pollutants for the same micropores. The severity depends on three factors, and oversimplified rules of thumb mislead:

Relative humidity (RH): activated carbon holds its toluene capacity roughly up to RH 50%, then capacity erodes as RH rises (2020 comparative study — High confidence). Quantified examples: a hydrophilic walnut-shell carbon lost 55.9% of its benzene capacity from RH 0%→90%, while a PDMS-hydrophobized version of the same carbon lost only 19.3% (2020, Sep. Purif. Technol.); a starch-based hierarchical carbon lost 46% of toluene capacity from RH 0%→80%, a hydrophobized composite 22% (Chem. Eng. J. 2024). An unmodified carbon was measured to take up 21.9 wt% water at equilibrium — a sense of how much pore volume water can occupy.
Concentration: at high VOC concentration (hundreds of ppm and above) humidity barely matters — the VOC competes successfully. At low concentration (single-digit ppm), humidity becomes the decisive factor. This is why coating/printing exhaust (low concentration, high flow) must address humidity explicitly, while solvent-recovery duty (high concentration) tolerates it.
Carbon hydrophobicity: coconut-shell and unoxidized coal carbons are relatively hydrophobic; oxygen-rich surfaces (chemical activation residues, oxidation treatments) are hydrophilic. Surface modification (silicone, PDMS, PDVB) is the documented anti-humidity solution — not "any coconut carbon".

A useful engineering takeaway: for humid, low-concentration VOC duty, prefer hydrophobic carbons or molecular-sieve adsorbents, or dehumidify upstream — see Zeolite vs Activated Carbon.

4. Temperature: equilibrium falls, kinetics rise

Two opposing effects, both real:

Capacity falls as temperature rises. Adsorption is exothermic; the adsorption enthalpy exceeds liquefaction enthalpy (ΔH_ads = ΔH_liq − RT·ln c). The Dubinin–Radushkevich equation, which predicts capacity for aromatic and chlorinated VOCs on activated carbon, contains temperature explicitly — capacity declines monotonically with T. There is no single universal "X% per °C" coefficient in the literature; the slope depends on the adsorbate-carbon pair and must come from measured isotherms. Treat any "per degree" rule of thumb you see as unsourced.
Kinetics improve as temperature rises. Pore diffusion accelerates, so the bed reaches saturation sooner.

Engineering red lines (China HJ 2026-2013, High confidence):

Exhaust temperature entering the adsorber should be below 40°C.
Bed temperature during the adsorption cycle must stay below 83°C; automatic alarm and cooling above that threshold — this is an anti-self-ignition limit.
Hot-air regeneration: below 120°C for activated carbon (fiber); below 200°C for molecular sieves; hot-air regeneration is prohibited for exhausts containing ketones and other easily ignitable species. Steam regeneration: below 140°C.
Heat release is real: a dry bed adsorbing high-concentration VOCs can warm measurably (adsorption heat); a wet bed warms less because water desorption absorbs heat. Design for the temperature excursion, not for isothermal operation.

5. Step 4: Breakthrough — how a bed actually exhausts

A fixed bed does not saturate uniformly. An adsorption front — the mass transfer zone (MTZ) — moves from inlet to outlet. The outlet concentration stays near zero until the front arrives; the moment outlet concentration reaches the threshold (an emission limit or a fraction of inlet concentration, e.g. 5%) is the breakthrough point, and the time to reach it is the breakthrough time.

The area above the breakthrough curve integrates to the maximum capacity; the quantity adsorbed up to breakthrough is the working capacity — what the bed delivers in service.
China's HJ 2026-2013 defines the engineering twin, dynamic adsorption capacity (动态吸附量): the average adsorbed mass per unit adsorbent when outlet concentration reaches the set value at constant temperature, pressure and flow (mg/g).
A steep S-curve means a narrow MTZ and high bed utilization; a shallow curve means a wide MTZ and wasted bed. MTZ widens with: larger particle size, higher velocity, lower inlet concentration, higher humidity. This is exactly why HJ 2026-2013 caps superficial velocities: granular beds below 0.60 m/s, activated-carbon fiber beds below 0.15 m/s, honeycomb below 1.20 m/s.
Pressure drop limits from the same standard: below 2.5 kPa for non-fiber beds, below 4 kPa for fiber beds.

For the full treatment including design parameters, see Adsorption Bed Design and Breakthrough Curves & Bed Sizing.

6. Step 5: Regeneration — paying back the energy

Desorption requires the energy adsorption released. Three routes:

Thermal swing (TSA): hot gas or steam reverses the equilibrium. Chinese regulatory ceilings: steam <140°C; hot air <120°C (carbon) / <200°C (zeolite).
Pressure swing (PSA/VSA): pressure reduction releases adsorbate.
Thermal reactivation (off-site or dedicated furnace): drying at ~105°C, desorption/decomposition at 500–900°C under inert atmosphere, then steam/CO₂ gasification of residues at ~800°C. Each cycle burns off 5–15 wt% of the bed.

Operationally, HJ 2026-2013 says replace the adsorbent when dynamic capacity falls to 80% of design value. See Regeneration for the full lifecycle economics.

7. What this means for specifiers

Ask for dynamic data, not just static surface area. BET and iodine describe the carbon; dynamic adsorption capacity (mg/g at your concentration, velocity, bed height) describes your system. The two can diverge by a large margin in humid, low-concentration duty.
State humidity and concentration together — they jointly determine whether you have a capacity problem. One without the other is unanswerable.
Design within the red lines: inlet <40°C, bed <83°C, and the velocity/pressure-drop caps of HJ 2026-2013 if the unit will be audited in China.
Anticipate the temperature excursion in high-concentration duty — including self-heating margins.
Prefer measured isotherms over "per-degree" folklore when extrapolating capacity to summer conditions.

Bottom line: capacity is a system property, not a material property. The carbon supplies the surface; your conditions — concentration, humidity, temperature, velocity, bed height — decide how much of that surface is actually used. The mechanism chain above is the checklist for finding where the loss is.

Data classification: mechanism statements are General technical knowledge; quantified humidity/temperature effects are Literature values with the cited studies (High confidence); HJ 2026-2013 figures are Regulatory/standard values (High). "X% per °C" folklore was searched for and NOT FOUND in any citable source — flagged deliberately. See [Data Classification(../methodology/data-classification.md) and [Sources & Evidence(../methodology/sources.md).