H₂S Capacity on Activated Carbon: Why Figures Don't Compare

The same activated carbon — AirDep CKC, impregnated with 5 % potassium bicarbonate — tested at the same 50–200 ppm inlet H₂S and the same 30 °C, gave 3.91 mg H₂S per gram under anoxic conditions and 84.87 mg per gram with only 0.1 % oxygen added. One variable changed. The capacity moved by a factor of 22. That figure, from Goula and co-workers in Catalysts, is the reason a bare H₂S capacity number on a datasheet tells you almost nothing.

One carbon, one variable, a 22× difference. Everything to the right of the bars is what a capacity figure has to declare before it can be compared with anything.

The short version

• One carbon, one variable: 3.91 mg H₂S/g anoxic versus 84.87 mg/g with 0.1 % O₂, at 50–200 ppm and 30 °C (Goula et al., Catalysts 10:521).
• ASTM D6646 is titled an accelerated method: it runs 10,000 ppmv H₂S to a 50 ppm outlet endpoint, and ASTM itself states the method does not duplicate service conditions.
• Three different mechanisms produce “H₂S capacity” on carbon — physical adsorption, catalytic oxidation and caustic neutralisation — and they require opposite gas conditions.
• g/cc and g/g figures are not interchangeable. Gas-phase carbons sit around 0.4–0.6 g/cc, so the two bases differ by more than a factor of two.
• SURSORB CAG publishes an H₂S adsorption figure of 0.12 g/cc with no test method and no conditions stated. By the standard set out in this article, that figure is not comparable to anyone else’s, and we say so below.

What actually removes H₂S on activated carbon?

Three mechanisms, and they are not variants of one another. They need different gas conditions, they fail in different ways, and they carry different hazards.

Three mechanisms, three sets of required gas conditions. A capacity figure only means something once you know which of these produced it.

1. Physical adsorption on virgin carbon

Weak, and strongly condition-dependent. Under anoxic conditions carbons show roughly one-tenth the capacity seen under humid, oxygen-bearing conditions — reported by Ciahotný and co-workers, carried into the Goula review. In the odour-control design literature, plain virgin carbon is rated only for 1–10 ppm H₂S duty at 0.05–0.08 g H₂S per cc. That is a polishing duty, not bulk removal. Note what that band is, though: a volumetric design allowance, not a measured capacity. A laboratory figure taken at a fixed challenge concentration — the 1.5 mg H₂S/g reported for unimpregnated carbon at 3,000 ppm, for instance — and a design-literature g/cc band are not the same quantity, and converted onto a common basis they differ by orders of magnitude, which is this article’s whole thesis rather than an exception to it.

2. Catalytic oxidation on catalytic or re-agglomerated carbon

The overall reaction is H₂S + ½O₂ → S + H₂O, with elemental sulfur retained inside the pore structure. Goula’s review describes oxygen leading to formation of solid elemental sulfur adsorbed on the pores, with water released.

The critical point is what the reaction consumes. Calgon’s CENTAUR HSV product bulletin states that effective H₂S removal requires the gas stream to contain “at least an equivalent amount of oxygen” and relative humidity above 10 %. Its bulletin for ST 1X — which is a caustic/alkali-impregnated acid-gas carbon rather than a catalytic one — repeats that the stream must contain free oxygen and that humidity must exceed 10 %, then adds the constraint that most specifications miss: “condensation of water on the carbon will reduce its performance”. You need a water film on the pore surface, not liquid water flooding it. That a caustic-impregnated grade is specified with the same oxygen and humidity requirements as a catalytic one is not an anomaly but part of the problem: a single commercial product can span more than one mechanism, so the label on a datasheet does not reliably tell you which mechanism produced the capacity figure printed underneath it. A saturated, chilled gas arriving at a catalytic bed is not the same duty as a warm gas at 40 % RH, and it will not give the same number.

Calgon also publishes, on the same carbon, the requirement that RH stay below 50 % to maximise carbon utilisation for VOC control — the vendor stating in its own literature that one carbon cannot be optimally conditioned for both duties at once.

Catalytic carbons are rated in the design literature for 50–100 ppm duty at 0.25–0.35 g H₂S/cc — the highest of the three bands.

3. Chemical reaction on caustic or alkaline impregnated carbon

NaOH or KOH impregnation raises the surface pH, driving H₂S dissociation and neutralisation. Liu and Zuo report that relative humidity and oxygen both enhance capacity, that alkaline impregnation improves performance, and — a specification detail worth knowing — that NaOH-loaded carbon performs better catalytically than KOH-loaded carbon at the same content. Bandosz’s group established that the caustic route is not purely acid–base but mediates oxidation of H₂S on the surface.

Caustic-impregnated carbon is rated for 10–50 ppm duty at 0.10–0.15 g H₂S/cc. It also carries a hazard the other two do not, covered below.

Why ASTM D6646 numbers are not field numbers

ASTM D6646 is the method behind almost every g/cc figure you will see on a supplier datasheet. Its full title contains the word Accelerated, and that word is doing real work.

Real digester gas and sewer air run one to three orders of magnitude below 10,000 ppmv. Argonne’s fuel-quality review puts general anaerobic digester gas at 10–1,200 ppm and landfill gas typically below 100 ppm; dairy waste digesters are the high case at 500–3,000 ppm.

ASTM does not hide the gap. The standard states the method “does not duplicate conditions that an adsorber would encounter in practical service”, that the laboratory setup favours carbons with faster removal kinetics, and that the mass transfer zone in the 23 cm test column is “proportionally much larger” than in a typical industrial bed. In other words, D6646 is a ranking tool for carbons, not a sizing tool for beds.

The gap was significant enough that Bagreev, Kuang and Bandosz built a model specifically to bridge it — deriving parameters from high-concentration data and extrapolating to service conditions, validated at the North River Water Pollution Control Plant in New York City. If the accelerated number transferred cleanly, nobody would have needed the model.

One thing D6646 does not pin down in any publicly verifiable way is the relative humidity of the test stream — the standard describes the air only as humidified. A specific RH figure is widely quoted in trade material; we could not verify it against the standard, so we do not print one.

The seven things a capacity figure must state

If a supplier quotes an H₂S capacity and any of these is missing, the number cannot be compared with a competing number. Ask for all seven.

Item 7 causes more confusion in tender comparison than the other six combined. A carbon quoted at 0.14 g/cc and one quoted at 250 mg/g may be the same carbon, or may differ by a factor of two. You cannot tell without both densities.

What the published vendor numbers actually say

Three commercial figures — and only one of the three names its method:

Only General Carbon states the method it tested to. The two Calgon bulletins publish the capacity without naming one. Applying this article’s own argument consistently, that means a figure whose method is not printed beside it cannot be assumed comparable to one whose method is — including when both come from the same trade and sit in the same units. The 0.20 g/cc figure can be read against ASTM D6646-01; the 0.09 and 0.14 figures can only be read against whatever the bulletin does say, which is the number and nothing else.

Now set those beside the peer-reviewed figures at realistic concentrations: 3.91 to 84.87 mg/g on the same carbon at 50–200 ppm depending on oxygen; roughly 25 wt % sulfur capacity for iodine-impregnated carbons under 20 % O₂ and 15 g/m³ water; roughly 10 wt % for copper and potassium salts. Those are not worse or better numbers. They answer a different question, at a different concentration, against a different endpoint. Ranking them in one column is a specification error.

Our own figure fails this test, and we are fixing it

SURSORB CAG is a re-agglomerated coconut-shell pellet with no impregnant, published at CTC adsorption min 50 %, BET surface area 1000 m²/g, apparent density 520 ± 30 g/L, ball pan hardness min 96, crushing strength min 4 kg, ash max 8 % and moisture max 5 %. Its datasheet also states H₂S adsorption 0.12 g/cc.

That figure carries no test method and no test conditions. By the standard this article has just set out, it is not comparable to Calgon’s 0.09 or 0.14 g/cc, to General Carbon’s 0.20 g/cc, or to anything else. We are not going to argue otherwise, and you should not accept the same silence from any supplier. The figure is being requalified with method and conditions attached.

The other biogas-category grades publish clean numbers on everything they publish at all, but none of them publishes an H₂S capacity:

The PAG-10 loading figure is worth pausing on. Most suppliers of impregnated H₂S carbon will not state a loading percentage at all — the impregnation level is treated as proprietary. A published minimum of 10 % is a commitment you can hold a supplier to on incoming inspection. It is not, however, a capacity figure, and we are not going to present it as one. Likewise MAG: the published literature on magnesium-oxide-impregnated carbon for H₂S is thin and largely paywalled, so we publish the physical specification — including the high ash maximum of 15 % that follows from an inorganic impregnant — and no capacity we cannot condition-qualify.

The safety line most suppliers leave out

Calgon’s bulletin for ST 1X states that “The impregnants on this carbon have been shown to decrease the ignition temperature”, and that the product can “self-heat and ignite” when exposed to air. The trigger condition is low convective airflow rather than high process temperature, which makes shutdown, standby, storage and maintenance the dangerous states rather than normal operation. The UK Environment Agency’s Technical Guide 12 adds the line worth designing around — “carbon monoxide monitoring is a much better early indicator than temperature monitoring” — and the mechanism, the ignition-temperature method comparison and the documented incident record are set out in the impregnation matrix.

Can H₂S-loaded carbon be regenerated?

H₂S-loaded activated carbon regenerates only partially, and how far depends on which mechanism loaded it: catalytic carbon can be water-washed in place, thermal regeneration recovers at best about 30 % of original capacity, and caustic-impregnated carbon consumes its impregnant stoichiometrically and has no verified regeneration route at all.

• Catalytic carbon: Calgon states CENTAUR can be regenerated in place for continued H₂S removal by washing the carbon with water, which requires potable water availability on site. That is mechanistically consistent — the catalytic route produces water-soluble sulfur compounds.
• Thermal regeneration: Coppola and Papurello recovered at best a 30 % regeneration ratio, on CO₂ at 300 °C, with the full nitrogen-air-CO₂ comparison set out in regenerate or change out.
• Caustic-impregnated: a claim that this media “can be regenerated up to 10 times” circulates widely. We traced it to a single non-peer-reviewed source, found no vendor bulletin or peer-reviewed corroboration, and note that Calgon steers customers toward catalytic carbon specifically where regenerability is wanted. Treat it as unverified. Suracsh does not operate reactivation services in the United States, so we have no commercial position to defend either way.

Every grade named here publishes its full specification in HTML at suracsh.us, and every datasheet downloads without a form or an NDA — including the SURSORB CAG sheet carrying the figure this article criticises. The selection guide maps contaminant to grade to specification, and the seven conditions above are the ones to check any published figure against, ours included.

Sources

• Goula et al., Catalysts 10(5):521 (2020) — Removal of H₂S from industrial gases
• ASTM D6646-03(2022) — Accelerated Hydrogen Sulfide Breakthrough Capacity
• ANSI listing, ASTM D6646-03(2022)
• Bagreev, Kuang & Bandosz, Adsorption 11:461 (2005) — Predicting H₂S breakthrough at low concentrations
• Bandosz — A Role of Sodium Hydroxide in H₂S Adsorption/Oxidation on Caustic-Impregnated Activated Carbons
• Liu & Zuo, Water Air Soil Pollut 234:118 (2023)
• Calgon Carbon — CENTAUR HSV product bulletin
• Calgon Carbon — ST 1X product bulletin
• General Carbon — GC Sulfursorb-BB
• Ludwigson — Activated Carbon Odor Control Systems, PDH Academy (2023)
• Papadias, Ahmed & Kumar — ANL/CSE/FCT/FQ-2011-11, Fuel Quality Issues in Stationary Fuel Cell Systems
• UK Environment Agency / Organics Recycling — Technical Guide 12, Fire risk of activated carbon (2017)
• OSHA Hazard Information Bulletin, 30 July 1997 — Fire Hazard From Carbon Adsorption Deodorizing Systems
• EPA Chemical Safety Alert — Fire Hazard from Carbon Adsorption Deodorizing Systems
• Coppola & Papurello, Clean Technologies 1(1):40–57 (2019) — Activated carbon regeneration for H₂S removal

Questions