Activated Carbon Adsorption: Isotherms, Kinetics and Bed Design

What is activated carbon adsorption, and how does it work?

Adsorption is the accumulation of a dissolved species (the adsorbate) on the surface of a solid (the adsorbent). It is distinct from absorption, where a species partitions into the bulk of a phase. On activated carbon, removal is dominated by physisorption — reversible physical binding through van der Waals and hydrophobic interactions, with a low enthalpy of adsorption (typically 5–40 kJ/mol). Chemisorption — the formation of chemical bonds with surface oxygen or nitrogen functional groups (40–400 kJ/mol) — plays a secondary role for polar and ionisable compounds.

Because physisorption is largely reversible and non-specific, activated carbon is a broad-spectrum polisher for taste-and-odour compounds, chlorinated solvents, pesticides, phenols, dechlorination, and a wide range of trace organic micropollutants — including the adsorption and filtration equipment used downstream of biological or membrane stages. The driving force is the concentration gradient between the bulk liquid and the equilibrium loading the carbon surface can support.

Why does pore structure and surface area matter?

Activation (steam or chemical) develops an enormous internal surface: a typical bituminous or coconut-shell GAC offers a BET surface area of 800–1,200 m²/g. That surface lives inside a hierarchy of pores, classified by IUPAC width:

• Micropores (< 2 nm) — provide almost all of the surface area and hold the bulk of small-molecule adsorption capacity.
• Mesopores (2–50 nm) — host larger molecules and act as transport pathways to the micropores.
• Macropores (> 50 nm) — the transport highways feeding the interior; they contribute little capacity but govern intraparticle diffusion.

The consequence for design is that a carbon must be matched to the target molecule: a compact taste-and-odour compound needs a highly microporous carbon, whereas natural organic matter (NOM) and larger micropollutants need adequate mesoporosity or they are excluded and diffusion-limited. Iodine number (roughly tracking micropore area) and molasses number (tracking mesopore area) are the usual bench indicators.

How do you describe adsorption equilibrium with an isotherm?

At a fixed temperature, the equilibrium loading q_e (mg adsorbate / g carbon) is a function of the residual liquid concentration C_e (mg/L). Two models dominate practice.

To fit either model, run a bottle-point isotherm: dose a series of carbon masses into fixed-volume aliquots of the water, equilibrate (often 3–7 days for GAC ground to powder), measure C_e, and back-calculate q_e from a mass balance. The Freundlich form usually gives the better fit for multi-component waters; a 1/n below 1 signals favourable adsorption.

Worked example: Freundlich isotherm to carbon dose and GAC bed life

Suppose bench testing on a phenolic industrial effluent yields a Freundlich fit of K_f = 40 (mg/g)(L/mg)^1/n and 1/n = 0.40. The influent is C₀ = 10 mg/L and the target is C_e = 0.1 mg/L.

Step 1 — PAC dose (single completely-mixed contact). The equilibrium loading in contact with the treated concentration is q_e = K_f·C_e^1/n = 40 × (0.1)^0.40 = 40 × 0.398 = 15.9 mg/g. A mass balance gives the carbon dose:

• Dose = (C₀ − C_e) / q_e = (10 − 0.1) / 15.9 = 0.623 g/L = 623 mg/L of PAC.

That high dose shows why PAC suits intermittent or low-flow duties, while a GAC bed is far more carbon-efficient because the carbon leaving service is in equilibrium with the influent, not the effluent.

Step 2 — GAC operating capacity. For a fixed bed run to exhaustion, the usable loading is q₀ = K_f·C₀^1/n = 40 × (10)^0.40 = 40 × 2.512 = 100.5 mg/g.

Step 3 — carbon usage rate (CUR). CUR = (C₀ − C_e) / q₀ = (10 − 0.1) / 100.5 = 0.0985 g/L = 98.5 g carbon per m³ treated.

Step 4 — bed life. Take a flow of 50 m³/h and an EBCT of 15 min. The bed volume is Q × EBCT = 50 × 0.25 = 12.5 m³; at a GAC bulk density of 450 kg/m³ the bed holds 5,625 kg. Volume treated to exhaustion = mass / CUR = 5,625 kg / 0.0985 kg·m⁻³ = 57,100 m³, i.e. 57,100 / 50 = 1,142 h ≈ 48 days between changeouts.

This is an upper bound: real service time is shorter because the mass-transfer zone leaks solute before the bed is saturated, and NOM competition erodes q₀. Treat it as the thermodynamic ceiling that a pilot column then discounts.

What sets the breakthrough curve — kinetics and the mass-transfer zone?

Equilibrium tells you capacity; kinetics tells you how a bed exhausts. Uptake proceeds through film diffusion across the liquid boundary layer, then intraparticle (pore and surface) diffusion into the micropores — usually the rate-limiting step. In a column this produces a finite mass-transfer zone (MTZ): a moving band of partially loaded carbon that travels down the bed. Effluent stays clean until the leading edge of the MTZ reaches the outlet (breakthrough), then the concentration climbs an S-shaped curve to exhaustion.

A sharper MTZ (faster kinetics, favourable isotherm) means the S-curve is steep and the bed is used efficiently; a broad MTZ wastes capacity. Two design parameters follow directly:

For robust polishing, deep beds or lead–lag pairs are used so the MTZ is always contained within the train even as the lead bed approaches exhaustion.

GAC or PAC — how do the two forms compare?

The same adsorption physics applies to granular and powdered carbon, but the contacting mode changes the economics and the operating envelope.

PAC shines when a taste-and-odour or spill event is transient — you dose only when needed and separate the spent carbon in a downstream clarifier or filter. GAC wins on lifecycle cost where the load is continuous, because reactivation recovers most of the carbon.

How do competitive adsorption and NOM fouling erode capacity?

Natural water is never a single solute. Background natural organic matter competes for the same pore surface and, being polydisperse, blocks micropore entrances — a slow preloading effect that steadily reduces the effective K_f for the target compound over weeks. In competitive systems the ideal adsorbed solution theory (IAST) predicts that a strongly adsorbing co-solute can displace a weaker one, so single-solute isotherms overstate field capacity, sometimes by an order of magnitude.

Practical mitigations are to place adsorption after upstream organics removal — coagulation, biological treatment or membrane pre-treatment such as ultrafiltration — so the carbon sees a lower NOM background, and to run realistic rapid small-scale column tests (RSSCT) that reproduce the preloading rather than a clean isotherm. Adsorption is a common polishing step for trace organics and PFAS and other emerging contaminants, where competition from NOM is often the binding design constraint.

When does thermal reactivation make sense?

Spent GAC can be restored by thermal reactivation: heating to 800–950 °C in a controlled steam atmosphere in a rotary or multiple-hearth furnace. Adsorbed organics are volatilised and pyrolysed, and controlled gasification reopens the pore network. Expect 5–15% carbon loss per cycle from attrition and burn-off, topped up with virgin make-up carbon. Reactivation typically costs a fraction of virgin carbon and is the reason continuous GAC duties beat single-use PAC on whole-life cost.

The decision hinges on carbon usage rate: high CUR (from strong loading or high influent concentration) favours reactivable GAC; very low usage or transient events favour dosed PAC. Where the spent carbon holds regulated contaminants, off-site reactivation with proper destruction of the desorbed load may be mandatory. Sizing the contactor and its backwash and carbon handling equipment together, and pairing the choice with upstream media filtration design, keeps the whole train efficient.

How to design a GAC adsorption contactor

1. Characterise the water. Measure target contaminant concentration, background NOM/TOC, temperature and flow profile. These set the adsorption duty and the competition the carbon will face.
2. Run isotherm and column tests. Fit a Freundlich isotherm from bottle-point data, then run an RSSCT or pilot column so preloading and mass-transfer are captured, not just clean equilibrium.
3. Select the carbon. Match pore structure to the molecule: microporous carbon (high iodine number) for small compounds, greater mesoporosity for larger micropollutants and NOM.
4. Set the EBCT and bed geometry. Choose an EBCT (typically 5–30 min) long enough to contain the mass-transfer zone, then fix bed depth and area from the design flow and a sensible linear velocity.
5. Estimate carbon usage and bed life. Compute q0 = Kf·C0^(1/n), the carbon usage rate and the service time, discounting for MTZ leakage and NOM preloading from the column data.
6. Plan changeout and reactivation. Adopt lead–lag beds for reliability and decide between thermal reactivation and single-use disposal based on carbon usage rate and any regulated loading.