pressure-swing-adsorption

Pressure-swing adsorption (PSA)

Component Semi-native equipment
TRL Mars
Energy intensity
Required by
0
Requires
3

Separates and purifies gases at near-ambient temperature by cycling packed adsorbent beds between high pressure (adsorb) and low pressure (desorb), so the weakly-held component passes through pure. It is the simple, robust alternative to cryogenic distillation for oxygen concentration, hydrogen purification, and — critically — drying feed gas to protect cryogenic exchangers and catalysts. No phase change, no cryogenics; the cost is compression energy and the adsorbent.

Last reviewed: 2026-06-14

Governing equations

The adsorption isotherm: loading q rises with partial pressure p toward saturation q_m. The pressure dependence is the entire principle — adsorb at high p, desorb at low p. [1]

PSA trades product purity against recovery: pushing purity higher wastes more of the product to the purge. Cycle design tunes this balance for the duty. [1]

The pressure ratio between adsorb and desorb steps sets the working capacity (loading swing) per cycle — and the compression energy that dominates the PSA energy bill. [1]

Fast Skarstrom-type cycling lets small beds deliver continuous product by alternating roles — compact equipment, valve-driven, no large rotating machinery. [1]

Key constants & quantities

Symbol Value Units Conditions Description
O₂ purity (zeolite PSA) 90–95 vol% Oxygen purity from N₂-selective zeolite PSA — the Ar that co-elutes caps it near 95%; cryogenic distillation is needed for higher.[1]
H₂ purity (PSA) 99–99.999 vol% Hydrogen purification by PSA reaches very high purity — the standard industrial route to clean H₂ for synthesis and fuel cells.[1]
Dew point (drying PSA) -70 °C (or lower) Molecular-sieve drying drives gas dew point deep below freezing — essential pretreatment protecting cryogenic exchangers and expanders from ice plugging.[1]
Cycle time 30–600 s Adsorb/desorb cycle duration — fast cycling keeps beds small and output continuous.[1]
Specific energy (O₂ PSA) 0.3–0.6 kWh / kg O₂ Energy to concentrate oxygen by PSA, dominated by feed compression — competitive with cryogenic O₂ at small/medium scale.[1]

Operating envelope

ParameterRangeUnitsSource
Adsorption pressure 2 – 10 bar [1]
Desorption pressure 0.1 – 1 bar (vacuum-swing reaches lower) [1]
Temperature 0 – 60 °C (near-ambient) [1]
Cycle time 30 – 600 s [1]
Adsorbent life 3 – 10 years [1]

Mass balance

Basis: 1 kg O₂ concentrated from air-like feed (zeolite PSA, illustrative)

Inputs

Feed gas (compressed) 5 kg [1]
Electrical energy 0.5 kWh [1]
Adsorbent (periodic) 0.001 kg [1]
  • Feed gas (compressed): Mostly returned as N₂-rich purge; mass depends on feed composition.
  • Electrical energy: Dominated by feed compression; vacuum-swing adds vacuum-pump energy.
  • Adsorbent (periodic): Zeolite/CMS lasts years; small amortized make-up.

Outputs

Product gas (≥90% O₂) 1 kg [1]
Purge / reject (N₂-rich) 4 kg [1]
  • Purge / reject (N₂-rich): Vented or recovered as a nitrogen stream.
TRL · Earth
9/ 9
TRL · Mars
6/ 9
PSA is mature commodity technology (medical/industrial O₂, refinery H₂, gas drying) and adsorption-based gas systems have strong space heritage — ISS CO₂ removal (CDRA) and oxygen systems use molecular-sieve adsorption beds. Mars gaps are valve durability under dust and adsorbent supply, not the separation principle.[2]
Energy budget
0.5 kWhe / kg O₂ concentrated (zeolite PSA; energy dominated by feed compression) [1]

PSA energy is essentially the feed-compression energy — no cryogenics, no phase change. That makes it cheaper and far simpler than cryogenic distillation at small-to-medium scale, at the cost of lower ultimate purity (and no argon co-product).

Variants & trade-offs

Oxygen PSA (N₂-selective zeolite)

[1]

Zeolite adsorbs nitrogen preferentially, passing oxygen-enriched product — the medical/industrial oxygen concentrator scaled up.

Materials: LiX / NaX zeolite · Cycling valves · Feed compressor
  • Simple, robust ambient-temperature O₂ source — no cryogenics
  • Fast start/stop; good for intermittent power
  • Strong space heritage in adsorption O₂/CO₂ systems
  • Purity capped ~95% (Ar co-elutes); no argon product
  • Compression energy; valve wear is the maintenance item

When preferred: Medium-scale O₂ for life support, leaching, and combustion-free oxidant where ultra-purity isn't needed.

Hydrogen PSA (purification)

[1]

Multi-bed PSA adsorbs impurities (CO, CO₂, CH₄, N₂) and passes ultra-pure H₂ — the industrial standard for clean hydrogen.

Materials: Activated carbon + zeolite layered beds · Multi-bed valve manifold
  • Very high purity H₂ (>99.99%) for synthesis and fuel cells
  • Removes the very poisons (CO, CO₂) that kill catalysts and PEM cells
  • H₂ recovery <90% (rest lost to purge unless recycled)
  • Multi-bed complexity

When preferred: Polishing electrolytic/FT hydrogen for catalyst- and fuel-cell-grade purity.

Molecular-sieve dryer / TSA

[1]

Adsorbs water (and CO₂) to deep dew points; regenerated by pressure or temperature swing (TSA). The guard that protects cryogenics.

Materials: 3A/4A/13X molecular sieve · Regeneration heater (TSA)
  • Drives dew point to -70 °C or below — prevents ice plugging in cold boxes and expanders
  • Removes CO₂ ahead of cryogenic separation
  • TSA regeneration needs heat; cyclic capacity limited

When preferred: Mandatory pretreatment upstream of every cryogenic exchanger, expander, and air-separation column.

Vacuum-swing adsorption (VSA)

[1]

Desorbs under vacuum rather than just venting — lower energy for some duties, and Mars's near-vacuum ambient assists the vacuum step.

Materials: Adsorbent beds · Vacuum pump
  • Lower-pressure operation, often less energy than high-pressure PSA
  • Mars ambient near-vacuum aids the desorption step
  • Needs a reliable vacuum pump; larger beds

When preferred: O₂/CO₂ duties where the low-pressure swing and Mars ambient vacuum can be exploited.

Failure modes

Mode Cause Detection Mitigation
Adsorbent degradation / poisoning[1] Liquid water slugs, oil carryover, or strongly-held contaminants permanently reduce adsorbent capacity over time. Declining product purity/recovery; breakthrough earlier in the cycle. Upstream knockout/coalescing filters, guard layers, oil-free compression, spec-clean feed; replace adsorbent on schedule.
Cycle valve failure[3] PSA cycles its switching valves constantly (millions of cycles); a stuck or leaking valve breaks the cycle and crashes purity. Pressure-profile anomaly, purity drop, valve-position telemetry. High-cycle-rated valves, redundancy, condition monitoring, standardized replaceable valve modules.
Moisture breakthrough into cryogenics (safety/availability)[1] Dryer bed saturates or channels, letting water reach a downstream cold box where it freezes and plugs. Moisture analyzer on dryer outlet; rising cold-box pressure drop. Conservative dryer sizing with margin, dew-point monitoring interlocked to the cryo plant, lead-lag bed redundancy.
Bed fluidization / attrition[1] Excessive gas velocity lifts and grinds the adsorbent, generating fines that blind screens and reduce capacity. Pressure-drop rise, fines in product, capacity loss. Velocity control below fluidization, bed compression/hold-down, robust pellet selection.
Compressor dependency[4] PSA lives or dies by its feed compressor — the energy and the pressure swing both come from it. Tied to compressor health monitoring. Compressor redundancy/spares; see the compressor node's reliability practice.

Mars adjustments

The simple alternative to cryogenics[1]

Impact: Where the colony needs medium-purity O₂ or clean H₂ but not argon or ultra-purity, PSA delivers it at ambient temperature with far less complexity than a cryogenic plant — a robustness win for a maintenance-constrained settlement.

Mitigation: Use PSA for life-support O₂ and H₂ polishing; reserve cryogenic distillation for argon and propellant-grade liquefaction.

It guards every cryogenic process[1]

Impact: Molecular-sieve drying/CO₂-removal PSA is the mandatory pretreatment that keeps water and CO₂ out of cold boxes, expanders, and air-separation columns — the recurring "clean, dry feed" requirement those nodes all cite.

Mitigation: Place a TSA/PSA dryer ahead of every cryogenic unit; dew-point interlock to the cold plant.

Strong space heritage de-risks it[2]

Impact: ISS already runs adsorption beds for CO₂ removal (CDRA) and oxygen systems — PSA/TSA is among the most flight-proven separation methods, lowering Mars deployment risk versus a novel cryogenic plant.

Mitigation: Lean on ISS adsorption-system heritage for life-support gas processing.

Mars vacuum aids the desorption swing[1]

Impact: Vacuum-swing operation benefits from the near-vacuum ambient — the low-pressure desorption step has a head start, potentially cutting vacuum-pump energy.

Mitigation: Consider VSA configurations that exploit ambient pressure for the desorb step.

Adsorbent is a durable but import consumable[1]

Impact: Zeolites and carbon molecular sieves last years but are not yet locally made; they are a slow-turnover strategic spare.

Mitigation: Stock multi-year adsorbent inventory; protect beds from poisoning to maximize life; local zeolite synthesis as a long-term chemistry goal.

Alternatives & substitutes

distillation-column (cryogenic air separation)[5]

  • Higher ultimate purity; co-produces argon (welding gas) and nitrogen together
  • Full cryogenic plant — far more complex and energy-intensive at small scale

When preferred: Large scale, ultra-high purity, or when argon co-production is needed.

Membrane gas separation[6]

  • No moving parts, continuous, compact
  • Lower purity per stage; membrane is an import/fouling item

When preferred: Bulk pre-separation and modest-purity duties; ahead of PSA polishing.

Electrochemical separation (electrolysis / O₂ from CO₂)[7]

  • Direct production rather than separation (MOXIE-style O₂ from CO₂)
  • Different process entirely; not a drop-in for gas purification

When preferred: Producing O₂ from CO₂ rather than concentrating it from a mixture.

Requires

References

  1. Ruthven, D. M., Farooq, S., & Knaebel, K. S. (1994). Pressure Swing Adsorption. VCH Publishers. ISBN 978-1-56081-517-9. — PSA fundamentals: adsorption equilibria and kinetics, Skarstrom and advanced cycles, and design for gas separation (O₂/N₂, H₂ purification, drying).
  2. Knox, J. C. (2005). International Space Station Carbon Dioxide Removal Assembly Testing. 35th International Conference on Environmental Systems, SAE 2005-01-2864. doi:10.4271/2005-01-2864 — CDRA architecture, zeolite 13X/5A duty cycle, ISS performance history.
  3. Smith, P., & Zappe, R. W. (2004). Valve Selection Handbook, 5th Edition. Gulf Professional Publishing. ISBN 978-0-7506-7717-2. — Valve types, selection, sizing, and actuation: gate/globe/ball/check/control valves, leakage classes, and service-specific selection.
  4. Bloch, H. P. (2006). A Practical Guide to Compressor Technology, 2nd Edition. Wiley-Interscience. doi:10.1002/9780470117002 — Centrifugal and reciprocating compressor selection, performance maps, surge, sealing, and reliability practice.
  5. Timmerhaus, K. D., & Flynn, T. M. (1989). Cryogenic Process Engineering. Plenum Press. doi:10.1007/978-1-4684-8506-4 — Cryogenic cycle engineering: turbo-expanders, the Claude/Brayton cycles, air separation, and liquefaction plant design.
  6. Seader, J. D., Henley, E. J., & Roper, D. K. (2016). Separation Process Principles: With Applications Using Process Simulators, 4th Edition. Wiley. ISBN 978-1-119-23958-9. — Distillation, absorption, and extraction design: equilibrium stages, McCabe-Thiele and rigorous methods, packing and tray hydraulics.
  7. Hartvigsen, J. J., Elangovan, S., Frost, L., Larsen, D., Elwell, J., Bayless, A., & Stoots, C. (2017). Carbon Dioxide Electrolysis for Mars ISRU. ECS Transactions, 78(1), 2953-2966. doi:10.1149/07801.2953ecst — MOXIE precursor work — solid-oxide CO₂ electrolysis at Mars conditions.