Joule Heist search

references · updated 2026-09-25

Adsorption Cooling — Limitations & Mitigations

confidence: high volatility: warm verified: 2026-09-25fresh

Why adsorption cooling has low COP (a basic cycle's exergy efficiency is ~10 %, with ~77 % of exergy destruction in adsorber thermal coupling, compounded by small reservoir temperature differences and batch operation), how heat-rejection temperature, driving temperature and source fluctuation derate it in the field (up to −19 % COP under realistic fluctuating heat; COP 0.53 in desert heat vs 0.6–0.8 nominal), why sorbent durability — not COP — is the open question for MOFs, and the engineered mitigation stack: mass recovery (+24 % COP, +37.5 % SCP at 60 °C), heat recovery (+12.6 % COP but −10.8 % SCP), their combination (+37 % exergy efficiency), cycle-time optimisation, pore-window matching, and the single-adsorber Stratisorp cycle (heating COP 1.98).

Adsorption cooling’s low COP is structural, not a tuning failure: second-law analysis puts a basic cycle’s exergy efficiency at ~10 % and pins ~77 % of all destruction on adsorber thermal coupling, on top of small reservoir temperature differences and inherently batch-wise operation. This article quantifies the limitations — the derating that hot climates and fluctuating heat sources impose, the water burden of heat rejection, and the sorbent-durability gap — and catalogs the engineered mitigations with their measured payoffs. It is the counterpart to the optimistic numbers in Performance & Numbers.

Why COP is low (root cause)

Exergy analysis of the basic single-bed cycle (Entropy 2020) gives an exergy efficiency of only 0.1013 (10.1 %), and ~77 % of total exergy destruction is adsorber thermal-coupling loss — the irreversibility of heating and cooling a thermally massive, poorly conducting bed through its whole temperature swing every cycle. Two further structural causes compound it: the small temperature differences among the heat reservoirs (a 60–90 °C source, a 25–35 °C sink and a 7–15 °C evaporator leave little Carnot headroom), and batch-wise operation, which throws away the bed’s sensible heat at every switch-over. The Stratisorp authors describe the same thing from the other side: the dominant irreversibility of a two-adsorber machine is the driving-temperature difference across which heat is transferred. Because the loss is structural, no single fix is decisive — the gains stack, and they are largest at low driving temperatures (~60 °C), diminishing toward 90 °C.

The practical consequences are the standard criticisms: low COP and low specific cooling power drive up size and cost per ton; the water-based machine needs high vacuum tightness; the bed has large thermal mass; and few suppliers make capital cost high (see Commercial Adsorption Chillers).

The derating problem: temperatures and fluctuation

Three operating variables set what a machine actually delivers, and the field sources rank them.

Driving temperature. Sztekler & Mika’s two-bed silica-gel/water chiller went from COP 0.35 to 0.50 and SCP 27 to ~160 W·kg⁻¹ as driving water rose 55 → 85 °C: below ~60 °C the cycle barely works, and every degree of source temperature is worth more than any bed tweak. Under realistic fluctuating heat — emulated PEM fuel-cell waste heat at 70–76 °C with ±2.5–4 °C swings — Hua et al.’s built double-bed chiller lost up to ~7 % cooling capacity and ~19 % COP relative to steady-source ratings; the practical envelope is ±3 °C of drive fluctuation, and shaping the ramp in steps (70 → 73 → 76 °C) instead of a direct ramp recovers ~4.6 % COP — a control-level lever costing nothing. Solar and waste heat, the sources that justify adsorption in the first place, are exactly such fluctuating sources (Hua figures are abstract-level; flagged for full-text verification).

Heat-rejection (condenser / cooling-water) temperature. This is the variable a hot climate attacks. Hua’s measured sensitivity ranking is chilled-water temperature > hot-water temperature > cooling-water temperature, so a machine is most sensitive to how cold you ask it to go and least to its sink — but the sink is the variable you do not control on an August afternoon. The Madinah desert study (an absorption chiller on evacuated tubes, but the same thermally driven physics) reported COP 0.53 in extreme heat against a 0.6–0.8 nominal band (as cited in Hua). The condenser-derating curve carried from the June 2026 compile — 26 → 32 °C cooling water: −11.2 % COP, −28.9 % SCP; 32 → 24 °C: COP 0.44 → 0.66 and cooling power 3.04 → 10.50 kW (3.5×) — is not traceable to any raw source in this article’s set and should be treated as unverified until its origin is re-established; the single modelled datapoint in Performance & Numbers (COP 0.50 → 0.31 at a 42 °C condenser) is likewise registry-carried. The direction is not in doubt; the magnitudes are. Either way, SCP collapses faster than COP, which is the core argument for wet cooling towers over dry coolers in hot climates and for sizing on SCP, not COP.

Chilled-water temperature is the strongest lever of all (Hua), and the pore-engineered silica study shows why: the same sorbent’s SCP can differ by 60 % depending on whether the evaporator sits at 10 °C or 25 °C (below).

Stacked field derate. Datasheet COP is a best-case laboratory value; stacking hot-sink derating on fluctuating-source derating, realistic field performance plausibly lands 25–40 % below nameplate.

The water burden of heat rejection

The Madinah techno-economic study is the favourable case for solar-thermal sorption cooling — 32 % solar fraction, LCOC $0.028/kWh, 3.0-year simple payback, IRR 32 % — and its liabilities still bite. At COP ~0.5 a thermally driven chiller must reject its cooling load plus twice that in driving heat, i.e. ~3× its cooling capacity, against ~1.25× for a vapor-compression machine at COP 4: a structurally ~2.3× larger heat-rejection burden per kWh of cold, which in a wet tower is evaporated water, in precisely the water-scarce regions where solar cooling is pitched. (The prior compile’s figure of 827.78 m³/year of cooling-tower water for the Madinah system is not in the ingested summary and is carried forward unverified.) This interacts directly with the wet-tower-versus-dry-cooler argument above: the wet tower protects SCP; the dry cooler protects water.

Sorbent durability: the open question

Materials that win on COP do not necessarily survive chiller duty:

  • MOFs. A February 2026 Nature Communications paper states flatly that “most MOFs exhibit only moderate or poor hydrothermal stability,” and that even the “water-stable” champions (Al-fumarate, MIL-160, CAU-10, MOF-303) decompose slowly under repeated > 100 °C heating — while a chiller demands thousands to millions of cycles and stability papers typically demonstrate tens to hundreds. A ~70,000-cycle temperature-swing benchmark (Applied Thermal Engineering 2023) saw 5–10 % uptake loss in most sorbents but a 35–45 % collapse in MIL-160(Al) under some conditions. Source-vs-source tension: the ORNL review in Performance & Numbers shows MOFs consistently beating conventional adsorbents on COP; both are true — MOFs win on paper; cycle-life at chiller duty remains unproven, with MFM-300’s month-long soak and 20-cycle test the best counter-evidence so far. Part of the frontier is now routing around MOFs — the paper’s own amorphous molecular adsorbent (M-PyC: 30 wt% uptake, 90–120 °C regeneration, constant capacity over 100 cycles) is one example, salt confinement (Composite Salt Sorbents) another.
  • Zeolites are robust but need 200–300 °C for full regeneration and lose up to 30 % of uptake when operated in open air (Nature Communications 2026) — the regeneration-temperature problem Zeolite documents and the aluminophosphates work around.
  • Hygroscopic salts need > 175 °C for full regeneration and suffer deliquescence and leakage; stabiliser work continues (see Composite Salt Sorbents).

The mitigation stack (with numbers)

MitigationCOP effectSCP effectNotes
Mass recovery+24 % @ 60 °C+37.5 % @ 60 °C+18 % useful exergy; the cheapest gain and the one that matters most at high condensing temperature
Heat recovery+12.6 % @ 60 °C−10.8 % @ 60 °Ccuts external exergy loss ~50 % (exergy efficiency 10.1 → 23.2 %); a real COP-vs-SCP trade-off
Combined heat + mass recovery——exergy efficiency 0.1389 = +37.1 % over the basic cycle, +11.3 % over heat recovery alone; optimal heat-recovery time ~60 s
Cycle-time optimisationtunabletunablean optimum exists per condition; longer cycles raise COP but lower throughput (and distillate in desalination mode)
Multi-bed reallocation+25 %+20 % capacitythree-bed optimum, best when the heat source is limited — carried from the prior compile, not traceable to this article’s raw set
Coated / finned / metal-additive beds—up to +40 %; TPMS geometries a further ~+12–17 % vs finscut inert thermal mass; numbers documented in [[topics/adsorbent-bed-engineering
Stratisorp single-adsorber cycleheating COP 1.98, SCOP 2.09—stratified-storage internal heat recovery (below)
Pore-window matching—±60 % depending on chilled-water temperaturesorbent-side (below)

All exergy-derived gains (first three rows) are strongest at a 60 °C source and diminish by 90 °C — the mitigations matter most exactly where low-grade heat forces you to operate.

Advanced cycles

  • Sztekler & Mika (silica-gel/water two-bed, desalination mode) demonstrate the heat-recovery phase and the cycle-time trade-off experimentally on a 1.10 kW lab chiller: a dedicated heat-recovery phase within a four-phase cycle, COP reaching ~0.64 at a 700 s cycle (chilled 25 / cooling 30 / hot 80 °C) while distillate output peaks at a shorter 450 s cycle (2.32 kg·h⁻¹, 0.07 mS·cm⁻¹ from ~40 mS·cm⁻¹ feed) — you cannot maximise both at once. The desalination side is the consolation for low COP: adsorption desalination needs < 1.5 kWh·m⁻³ of electricity and yields water at ~$0.2/m³ with < 10 ppm TDS, against 3–4 kWh·m⁻³ for seawater RO.
  • Stratisorp (Schwamberger, Desai & Schmidt, KIT) attacks the dominant driving-temperature-difference irreversibility with a single adsorber coupled to a well-insulated stratified thermal-storage tank: heat released during adsorption is stored by temperature, at its own height in the tank and drawn back for desorption, giving the internal heat recovery of a multi-adsorber machine with one adsorber. Modelled with water/zeolite Li-Y (RUB04) and thermal oil at a 200 °C driving temperature, it reaches heating COP 1.98 at 5.4 kW (sink 34 °C, source 8 °C), can exceed 2.0, and scores a seasonal COP of 2.09 over the five VDI 4650 load points; second-law analysis confirms the driving-temperature irreversibility is strongly reduced versus an optimally switched two-adsorber cycle. The sensitivities are a design guide in themselves: doubling the adsorber’s U-value adds ~0.1 COP, halving its thermal mass adds ~0.1, and storage above ~300 kg buys nothing more.

Sorbent-side: match pore size to the operating window

Pore-engineered sorbents help only when matched to the operating relative-pressure window (Mohammed et al., npj Clean Water 2021). An engineered bimodal nanoporous silica (pore peaks 1.27 nm + 2.75 nm; BET 929 vs 621 m²·g⁻¹; pore volume 0.773 vs 0.33 cm³·g⁻¹; maximum uptake 0.53 vs 0.37 kg·kg⁻¹) beats conventional silica RD for high-evaporator-temperature cooling (SCP 0.82 vs 0.51 kW·kg⁻¹ at 25 °C chilled water, ~60 % higher) and for desalination (SDWP ~24.5 vs 15.1 kg·kg⁻¹·day⁻¹, GOR 0.77), and works on heat as low as 65 °C — but conventional silica RD wins for low-temperature air-conditioning (SCP ~245 vs ~150 W·kg⁻¹ at 10 °C chilled water), with the crossover near 17 °C. There is no universally best sorbent: pore structure must be tuned to the target chilled-water temperature, which is also why MOF headline uptakes do not translate directly into chiller SCP.

See Also

Sources