Molecular Sieve for Biogas Upgrading: 13X vs Activated Carbon for CO2 Removal
An engineering comparison of 13X zeolite and activated carbon for CO2/CH4 separation in PSA biogas units — with selectivity, methane slip, regeneration duty, and a 7-year TCO model.
By Aluminaworld Technical Team · Published June 30, 2026 · 12 min read
1. Why Biogas Upgrading Is a Real Engineering Problem
Raw biogas from an anaerobic digester or landfill is typically 55-65% methane, 30-40% carbon dioxide, trace hydrogen sulfide (50-3000 ppm depending on feedstock), water vapor at saturation, and various siloxanes and volatile organic compounds. To inject this into a natural gas grid, or to compress it as vehicle fuel, you have to remove the CO2 down to 1-3% and strip almost everything else out — without throwing away the methane you are trying to sell.
Pressure swing adsorption is the workhorse technology for plants in the 200-5000 Nm3/h range, the size band that covers the vast majority of farm, food-processing, and municipal-waste-to-energy projects. Two adsorbents dominate the engineering debate: 13X molecular sieve and activated carbon. They look superficially similar — both come as black or dark beads, both adsorb CO2 — but their pore structures, surface chemistry, and operating economics are very different.
This guide walks through the engineering side of that choice, with the numbers a process engineer or procurement manager actually needs to write a spec sheet or evaluate a bid. We focus on 13X (the type most commonly used in modern biogas PSA) versus standard bituminous-coal activated carbon (the workhorse of older and smaller units). For related reading on PSA oxygen and hydrogen units, see our guides on LiLSX for O2 concentrators and 5A vs 13X in hydrogen PSA.
2. How 13X and Activated Carbon Differ at the Molecular Level
13X is a sodium-form faujasite (FAU) zeolite with a uniform 9 Angstrom pore opening. Inside the crystal, CO2 (3.3 A kinetic diameter) and CH4 (3.8 A) both fit, but the electric field gradient from the extra-framework sodium cations and the aluminum tetrahedra in the framework creates a strong quadrupole interaction with CO2 that does not exist with the non-polar CH4 molecule. The result is a strongly preferential adsorption isotherm: at 1 bar and 25 C, 13X takes up roughly 5-7 mmol/g of CO2 versus 1-1.5 mmol/g of CH4, a working selectivity of 4-8.
Activated carbon is a disordered graphitic material with a wide distribution of pore sizes — micropores (under 2 nm) where most of the adsorption happens, mesopores (2-50 nm) for transport, and macropores (over 50 nm) for bulk flow. The surface is largely non-polar. Adsorption is driven by van der Waals dispersion forces plus weak induced-dipole interactions, with no electrostatic enhancement for CO2. At the same 1 bar and 25 C, a typical bituminous AC takes up 3-4 mmol/g of CO2 and 1.5-2.5 mmol/g of CH4, giving a working selectivity of 1.5-2.5 — meaningfully worse than 13X.
The other crucial difference is water. 13X adsorbs water strongly enough that the isotherm at high humidity is essentially vertical, which is why a 13X bed used for biogas must be kept dry or it becomes a water-removal bed first and a CO2-removal bed second. Activated carbon, with its largely hydrophobic surface, takes up much less water at the same relative humidity, but it still loses significant CO2 capacity to water loading at high RH. Both materials need a dryer upstream; the design target is feed gas below 50 ppmv water, ideally below 20.
| Property | 13X Molecular Sieve | Activated Carbon (Bituminous) |
|---|---|---|
| Pore structure | Uniform 9 A crystals | Disordered 0.5-50 nm distribution |
| CO2 capacity at 1 bar, 25 C | 5-7 mmol/g | 3-4 mmol/g |
| CH4 capacity at 1 bar, 25 C | 1-1.5 mmol/g | 1.5-2.5 mmol/g |
| CO2/CH4 working selectivity | 4-8 | 1.5-2.5 |
| Water tolerance | Poor — strong hydrophilic sites | Moderate — hydrophobic surface |
| H2S tolerance | Poor — irreversible poisoning | Moderate — some reversible capacity |
| Typical bulk density | 640-720 g/L | 450-550 g/L |
| Bed life in clean biogas | 3-5 years | 2-4 years |
3. Selectivity, Capacity, and Methane Slip — Side by Side
The single most important operating metric in a biogas PSA is methane slip — the percentage of feed methane that ends up in the waste gas stream rather than the product. Methane that leaves the bed with the desorbed CO2 is either flared (with CO2e implications), vented (a regulatory problem), or recycled back through the compressor (which costs electricity and capacity). A 13X system with a four-bed configuration typically lands at 2-4% methane slip. An activated carbon system on the same feed typically lands at 4-8%.
The reason is in the working selectivity numbers above. When the bed is regenerated, 13X lets go of CO2 more cleanly than CH4, so the desorbate is CO2-rich. Activated carbon releases more CH4 along with the CO2 because the isotherms are closer together, and that CH4 is lost on the next cycle unless you add a recycle compressor and a fifth "trim" bed. The 2-4% methane slip advantage of 13X translates, at a 1000 Nm3/h plant running 8000 hours per year, to roughly 1.6-3.2 million Nm3 per year of recovered methane, which at typical biomethane prices of 0.60-1.20 USD/Nm3 is 80,000-300,000 USD per year in additional revenue.
Capacity also matters because it drives bed size. 13X has roughly 1.5-2x the CO2 working capacity per kilogram in the operating pressure range of 4-7 barg. The flip side is that 13X is denser (640-720 g/L versus 450-550 g/L for AC), so the volumetric advantage is smaller, on the order of 1.2-1.4x. Most designers end up with 13X beds that are 70-90% the volume of the AC bed they would otherwise have specified, plus a smaller blower and vacuum pump because the cycle time is shorter.
| Operating Metric (1000 Nm3/h, 4-bed) | 13X | Activated Carbon |
|---|---|---|
| Product CH4 purity | 96-98% | 93-96% |
| Methane slip (vent loss) | 2-4% | 4-8% |
| CH4 recovery per year | 15.4-15.7 MM Nm3 | 14.7-15.4 MM Nm3 |
| Specific power | 0.20-0.30 kWh/Nm3 | 0.10-0.20 kWh/Nm3 |
| Bed volume relative | 1.0 (baseline) | 1.1-1.4x |
| Feed H2S requirement | < 1 ppm (strict) | < 50 ppm (tolerant) |
4. Regeneration: VPSA, TSA, and Energy per Nm3
Almost every modern biogas PSA runs as a vacuum pressure swing adsorption (VPSA) system, not a temperature swing (TSA) system. The reason is energy: TSA on 13X requires 250-300 C purge gas, and the heating/cooling duty for a 50-200 ton charge is enormous — typically 1.5-2.5 MJ per Nm3 of product, which dwarfs the compression energy. VPSA at 80-120 C gives up some working capacity but cuts regeneration energy by an order of magnitude, to roughly 0.05-0.15 MJ per Nm3.
The VPSA cycle on 13X looks like this: adsorb at 4-7 barg for 60-180 seconds, equalize with another bed to recover some CH4 from the void space, then desorb by pulling vacuum to 50-150 mbar absolute with mild heating to drive off residual CO2. The vacuum pump is a liquid-ring or rotary vane unit sized for 0.5-1.5 kW per Nm3/h of product. Activated carbon VPSA follows a similar shape but with shorter adsorption times (the bed saturates faster) and a slightly higher vacuum level because the CO2 isotherm has a longer tail.
For a 1000 Nm3/h biomethane plant, total specific power including the feed blower, vacuum pump, and auxiliaries lands at 0.20-0.30 kWh/Nm3 for 13X and 0.10-0.20 kWh/Nm3 for AC. At first glance AC looks like the winner on energy, but when you fold in the methane slip difference, the 13X system usually wins on total cost per Nm3 of sold methane. We'll quantify that in the TCO section below.
5. H2S, Siloxanes, and Moisture — the Poison Trio
If you take only one piece of advice from this guide, take this: the 13X bed is downstream of clean, dry, sweet gas. Period. The single largest cause of premature 13X failure in biogas service is feeding it H2S, water, or siloxanes above the design spec.
Hydrogen sulfide. 13X adsorbs H2S very strongly and irreversibly in the temperature range of VPSA operation. 50 ppm H2S in the feed will halve the CO2 working capacity in 6-12 months. Industry standard is to drop H2S to below 1 ppm with a biological scrubber (cheap, common on agricultural digesters) or a sacrificial activated-carbon bed (more compact, used on landfill and food-waste plants). Activated carbon is more H2S-tolerant but eventually exhausts too, especially if the AC is also doing the CO2/CH4 split.
Siloxanes. These silicon-organic compounds come from deodorants, detergents, and silicone sealants in the waste stream. They adsorb on both 13X and AC and then, during regeneration, decompose to silica (SiO2) which deposits on the pore mouths and is permanent. The cure is upstream activated carbon polishing — AC has enough mesoporosity to hold siloxanes loosely so they can be steam-stripped off periodically, or it is simply replaced. Silica deposition on 13X is essentially end-of-life.
Moisture. Water loading on 13X displaces CO2 from the strongest sites and forces higher regeneration temperatures to recover. Wet feed also promotes hydrolysis damage to the zeolite framework, gradually reducing the surface area over years. The design target is feed below 50 ppmv water, ideally below 20, achieved with a chiller + knockout drum and a molecular sieve 3A or 4A guard bed upstream. See our 3A application guide for the guard bed design.
6. 7-Year TCO: 13X vs Activated Carbon for a 1000 Nm3/h Plant
Let's put numbers on the whole operating envelope. The reference plant is 1000 Nm3/h of upgraded biomethane, 8000 operating hours per year, 60% CH4 in feed, 13X or AC beds sized for 4-bar adsorption, 100 mbar vacuum regeneration, electricity at 0.10 USD/kWh, and biomethane at 0.80 USD/Nm3.
| Cost Item (7 years) | 13X | Activated Carbon |
|---|---|---|
| Initial sieve charge (60-80 t) | 90,000-120,000 | 50,000-70,000 |
| Replacement beds (year 3 and 5 for 13X; year 2, 4, 6 for AC) | 180,000-240,000 | 150,000-210,000 |
| Electricity at 0.20-0.30 vs 0.10-0.20 kWh/Nm3 | 1,120,000-1,680,000 | 560,000-1,120,000 |
| Lost revenue from methane slip (2-4% vs 4-8%) | 560,000-1,120,000 | 1,120,000-2,240,000 |
| 7-year total (mid-range estimate) | ≈ 2,400,000 | ≈ 2,650,000 |
The crossover is sensitive to electricity price and biomethane price, but in most geographies where electricity is above 0.08 USD/kWh and biomethane is above 0.50 USD/Nm3, 13X wins on 7-year TCO. Below those thresholds, AC becomes competitive. The real lesson is that the sieve itself is rarely the dominant cost line — methane slip and electricity are, and both are 13X-friendly.
7. Layered Bed Design: Putting AC on Top of 13X
The most robust biogas PSA design uses a layered bed: activated carbon on the top third of the vessel, 13X on the bottom two-thirds. The AC layer acts as an integrated guard bed for H2S, siloxanes, and higher hydrocarbons that slip through the upstream polishing, while the 13X does the actual CO2/CH4 separation. This configuration extends 13X life to 5-7 years and tolerates intermittent upsets in the upstream clean-up train.
Layered bed sizing is non-trivial. The AC layer must be thick enough to give a 30-60 second breakthrough time at peak H2S loading, which typically works out to 300-500 mm of 4x8 mesh AC. The 13X layer underneath is sized for the full CO2 throughput. The interface between the two materials is usually a perforated plate or a graded ceramic ball layer to prevent mixing during pressurization and depressurization cycles. Our engineering team can provide layered bed sizing for your specific feed composition — contact us with the H2S, siloxane, and moisture numbers from your gas chromatograph.
8. Specification Checklist for Procurement
When you request a quote, send your sieve supplier the following data points so they can match the right grade to your feed:
- Feed flow rate (Nm3/h) and operating hours per year
- Feed composition: CH4, CO2, H2S, H2O, siloxanes, NH3, VOC
- Adsorption pressure (barg) and regeneration vacuum level (mbar abs)
- Target product CH4 purity and methane recovery
- Bed geometry: diameter, height, number of beds
- Cycle time target (seconds per adsorption step)
- Layered or homogeneous bed
- Expected sieve life target and replacement cost tolerance
For biogas service, our typical recommendation is 13X in 8x12 or 4x8 mesh with Na2O content below 0.5% (low-Na2O grades have less fines generation in service) and a CO2 working capacity of at least 4.5 mmol/g at 1 bar and 25 C. We can also supply pre-blended layered bed charges with the AC and 13X sized to your specific feed. See also our related guides on activated alumina for the upstream moisture knock-out, and on regeneration best practices.
9. FAQ
Why is 13X the standard molecular sieve for biogas upgrading?
13X has a 9 A pore opening that admits both CO2 (3.3 A kinetic diameter) and CH4 (3.8 A), but the stronger quadrupole moment of CO2 plus the 5-7x higher polarizability of CO2 relative to CH4 produces a strongly preferential adsorption equilibrium. Typical CO2/CH4 selectivity is 4-8 at 1 bar and 25 C, which is enough to drive PSA to >97% methane purity in a 4-bed configuration.
How much methane slip does 13X PSA lose compared to activated carbon?
Well-tuned 13X PSA loses 2-4% methane in the tail gas, while activated carbon PSA typically loses 4-8% under the same conditions. For a 1000 Nm3/h biomethane plant, the 13X advantage is worth roughly 80,000-150,000 USD per year at current gas prices.
What is the typical regeneration temperature for 13X in biogas service?
VPSA regeneration uses adsorption at 0.2-0.5 barg and desorption at 50-150 mbar absolute, with mild heating to 80-120 C. TSA uses 250-300 C purge gas. Biogas plants almost universally use VPSA.
How long does 13X last in biogas service before replacement?
Typical service life is 3-5 years in clean biogas (H2S < 1 ppm, moisture < 50 ppmv). H2S is the primary poison: 50 ppm H2S can halve CO2 capacity in 6-12 months. Thermal regeneration in the field does not fully recover capacity.
Does biogas H2S need to be removed before the 13X bed?
Yes. Industry standard is biological or activated-carbon polishing to < 1 ppm H2S upstream of the 13X bed. Skipping polishing is the most common cause of premature 13X failure in biogas.
What moisture level is safe for 13X in biogas feed?
Feed should be < 50 ppmv water, ideally < 20 ppmv after a chiller and knockout drum.
Can 13X and activated carbon be combined in a layered bed?
Yes. A top layer of activated carbon captures H2S, siloxanes, and heavier hydrocarbons, while the 13X underneath does the CO2/CH4 separation. This can extend sieve life to 5-7 years at the cost of 50-100 mbar additional pressure drop.
How does pressure affect 13X CO2 capacity in biogas PSA?
CO2 loading on 13X increases roughly linearly with CO2 partial pressure up to about 5 bar, then plateaus. Most biogas plants adsorb at 4-7 barg.
What is the methane number impact of residual CO2 in product biomethane?
At 97% CH4 / 3% CO2, methane number is ~130. At 95% CH4 / 5% CO2, it drops to ~120. Pipeline spec in most EU markets is MN > 130.
What is the energy consumption of a 13X PSA biogas plant per Nm3 of biomethane?
Typical specific power is 0.20-0.30 kWh per Nm3 of upgraded biomethane. A 500 Nm3/h plant on 13X uses about 3,000-4,500 MWh per year of electricity.
10. Next Steps
For a 13X or layered-bed quote for your biogas plant, send us your feed flow, feed composition (especially H2S and moisture), target product CH4 purity, and adsorption pressure. We will return a sieve grade recommendation, charge weight, expected bed life, and a delivered CIF or FOB price.
Related reading on this site: LiLSX for medical O2 concentrators, 5A vs 13X in hydrogen PSA, 3A for moisture guard beds, regeneration best practices, and molecular sieve product page.