How to Calculate Activated Carbon Dosage for Water Treatment (With Worked Examples)

By the YELI Technical Team · Updated August 2026 · 9 min read

Every plant asks the same question before its first order: how much activated carbon do we need? Suppliers quote prices per ton, but nobody quotes your dose. The number that matters is a ratio: milligrams of carbon per litre of water (mg/L) for powdered carbon, or tonnes of media per bed for granular carbon. This guide walks through the standard calculation methods — jar tests, isotherm estimates, and bed sizing — with worked examples you can adapt to your own numbers.

First, Know Which Carbon You Are Dosing

The calculation looks different for the two families of activated carbon:

• Powdered activated carbon (PAC) — dosed as a slurry in batches, then removed with the sludge. The question is mg/L per dose, and how often to redose.
• Granular activated carbon (GAC) — packed into a fixed bed that treats water continuously. The question is bed volume (from flow and contact time), then how many tonnes of media and how often to replace it.

If you are still choosing between them, our coconut vs coal comparison and the GAC bed design guide cover the trade-offs. The rest of this article assumes you know which one you are buying.

Method 1 — The Jar Test: The Only Ground Truth

No formula replaces a test on your actual water. The jar test is quick and gives a working capacity you can trust:

1. Fill six beakers with 1 L of the raw water to be treated.
2. Add carbon at increasing doses — for example 0, 5, 10, 20, 40 and 80 mg/L (0–80 mg for 1 L).
3. Mix rapidly for 1–2 minutes, then slowly for 20–30 minutes (typical contact time for PAC).
4. Filter a sample from each beaker and measure the target pollutant (colour, COD, TOC, pesticide, geosmin/MIB, etc.).
5. Plot residual concentration against dose and read off the dose that meets your limit.

Use your raw water, not distilled water. Natural organic matter and competing compounds consume carbon too, so a lab-water test will underestimate the dose by a wide margin.

Method 2 — Freundlich Isotherm: A First Estimate Without the Lab

Adsorption on activated carbon is usually described by the Freundlich isotherm:

q = K × C^1/n

• q = adsorption capacity at equilibrium (mg pollutant per g carbon)
• C = residual (equilibrium) concentration of the pollutant in water (mg/L)
• K and 1/n = carbon/compound-specific constants, usually supplied by the manufacturer or found in the literature for your pollutant

For a target residual C, this gives the expected working capacity q. The dose follows directly:

Dose (mg/L) = ΔC (mg/L) ÷ q (mg/g) × 1000

where ΔC is the concentration you actually need to remove (inlet minus target).

Worked example — geosmin in drinking water

Say your plant treats water with 50 ng/L geosmin and the target is 5 ng/L. ΔC = 45 ng/L = 0.000045 mg/L. For coconut GAC against geosmin, a typical K is around 60 (mg/g)(L/mg)^0.5 with 1/n = 0.5, giving q = 60 × (0.000005)^0.5 ≈ 0.13 mg/g — more than enough headroom, which is why T&O control doses are set by contact time rather than by capacity. The lesson: for trace contaminants, dose is rarely capacity-limited.

Worked example — organics removal with PAC

A lagoon treats water with 1.2 mg/L of a pesticide; the discharge limit is 0.01 mg/L. A jar test on this water gives a working capacity of 6 mg/g at the target residual.

• ΔC = 1.2 − 0.01 = 1.19 mg/L
• Dose = 1.19 ÷ 6 × 1000 = 198 mg/L
• For a 1,000 m³/day flow: 198 g/m³ = 198 kg per day, roughly 5.9 t/month

That is the calculation in one line. If your jar test says 6 mg/g on your water, use 6 — not the catalogue 20 mg/g, which is measured on ideal solutions.

Method 3 — Sizing a GAC Bed (Continuous Treatment)

For granular carbon, the dose question becomes a volume question. The two design inputs are flow rate and empty bed contact time (EBCT):

Bed volume (m³) = flow (m³/h) × EBCT (min) ÷ 60

Media weight (t) = bed volume (m³) × bulk density (t/m³)

• Coconut-shell GAC bulk density is typically 0.45–0.50 t/m³
• Drinking water T&O duty usually runs 7.5–15 min EBCT; PFAS duty often 15–20 min or more

Worked example — 100 m³/h at 10 minutes EBCT

• Bed volume = 100 × 10 ÷ 60 = 16.7 m³
• Media = 16.7 × 0.48 = ≈ 8 t per bed
• Two beds in parallel for continuous operation: ≈ 16 t installed, plus a spare charge

Method 4 — Annual Media Consumption: The Number Your Budget Needs

Install tonnes matter once; replacement tonnes matter every year. Two ways to estimate:

• PAC: annual usage ≈ average dose × annual treated volume, plus a safety factor of 1.2–1.5 for seasonal peaks.
• GAC: annual usage = installed media ÷ expected service life (typically 1–3 years for drinking water organics; see when to replace GAC). Add make-up for attrition losses, usually 3–7% per year for high-hardness coconut carbon.

Five Field Checks Before You Order

1. Test your water, not a synthetic one — background organics can cut working capacity by half or more.
2. Watch temperature — adsorption improves as temperature drops; size for your coldest month if the limit is a summer problem, or your worst case either way.
3. Check competing organics — if TOC or COD is high, expect a shorter life and plan for it in the contract.
4. Confirm contact time in the real system — short-circuiting through a bed reduces effective contact time below the design value.
5. Leave margin — nobody was ever fired for ordering 10% extra PAC or a spare GAC charge; running out mid-season is far more expensive.

The Bottom Line

Dose = removal required ÷ working capacity, measured on your water. For PAC, that gives mg/L and then annual tonnes. For GAC, translate the flow and target EBCT into bed volume, then into installed and replacement tonnes. Use the jar test as ground truth, the isotherm as a sanity check, and the field checks as your last line of defence. When you have your numbers, send them with your RFQ — we quote water-grade coconut GAC and PAC with full COA and can help sanity-check the sizing.