Joule Heist search

topics · updated 2026-10-27

Absorption Cooling

confidence: high volatility: warm verified: 2026-10-27fresh

The liquid-sorbent cousin of adsorption: a thermally-driven cycle using a liquid absorbent (lithium bromide/water for air conditioning, ammonia/water for refrigeration) and a small solution pump. Single-effect COP ~0.7 (0.73–0.87 on 100–120 °C geothermal water), double-effect ~1.2–1.4, driven by roughly 70–200 °C heat with a practical single-effect floor near 90 °C. More mature and higher-COP than adsorption, but with corrosion, crystallization (LiBr needs condenser water of at least ~20 °C), roughly double the heat rejection of an electric chiller, and (for ammonia) rectification and copper/brass incompatibility. On geothermal brine it beats an electric chiller only where electricity is expensive, the heat is otherwise wasted and the rejection sink is cheap or cold.

The mature, commercially dominant heat-driven cooling cycle — and the closest cousin of Adsorption Cooling. Where adsorption uses a solid sorbent in a batch cycle, absorption uses a liquid absorbent continuously circulated by a small solution pump. That one difference buys higher COP and continuous operation, at the cost of a liquid that can corrode, crystallize, or (for ammonia systems) need rectification. Absorption is the benchmark adsorption is usually measured against.

How it differs from adsorption

Same job — move heat from cold to hot using thermal energy instead of a compressor — but:

AbsorptionAdsorption
SorbentLiquid (LiBr or water/ammonia solution)Solid (silica gel, zeolite, salt composite)
OperationContinuous (solution pumped generator↔absorber)Batch / multi-bed switching
COP (single-effect)~0.7 (Román’s survey: 0.6–1.1 across single and double effect; double-effect ~1.2–1.4)~0.4–0.6 (Román: 0.4–0.6; catalog unit 0.53)
Drive temp70–200 °C (Yazaki water-fired 70–95 °C; Román’s survey 85–180 °C; geothermal models put the single-effect floor at 90 °C)50–100 °C (lower floor)
Moving partsOne small solution pumpNone (valves only)
Failure modesCorrosion (LiBr), crystallization, NCG, rectification (NH₃)Lower SCP, large footprint

Adsorption’s edge is the lower regeneration temperature and no moving/corrosive parts; absorption’s edge is higher COP and maturity. Many designs now combine them.

The two working pairs

  • Lithium bromide / water (LiBr = absorbent, water = refrigerant): the air-conditioning workhorse. Water is the refrigerant, so it only cools above 0 °C, runs under deep vacuum (El Haj Assad’s geothermal model: 7.4 kPa / 0.68 kPa, LiBr 56–62%), and risks LiBr crystallization at high concentration or low cooling-water temperature — a 0.5–1% concentration shift materially moves the crystallization point, so YORK specifies entering condenser water of at least ~20 °C (68 °F) where an electric chiller happily takes 13 °C. Response is slower than electric because of the solution’s thermal inertia. Single- and double-effect machines dominate large commercial chilled-water plants; Zhao et al. find LiCl/water within 1–2% of LiBr on 80–105 °C geothermal water, and that exergy efficiency improves with warmer chilled water (comfort cooling over low-temperature loads).
  • Water / ammonia (water = absorbent, ammonia = refrigerant): reaches below 0 °C for refrigeration and ice-making (Chena: −29 °C brine), runs at higher pressure (8–14 bar in Buehn’s design; 10–13 bar at 90–100 °C for water-ammonia), and needs a rectifier to purify the ammonia vapor of water. Ammonia (boiling point −33.5 °C, latent heat 1,371 kJ/kg) is compatible with carbon and stainless steel, aluminum and PTFE but not brass or copper. Favored for industrial refrigeration and off-grid solar ice makers.

Commercial machines

  • Yazaki water-fired LiBr/water chillers run on hot water at 70–95 °C — a notably low drive temperature, so more waste heat is usable per ton. Water-fired 5–100 tons; direct-fired (gas) 30–200 tons. Their CCHP case: a CHP plant delivers 80 useful energy units per 100 fuel vs 56 for separate power + boiler, by routing waste heat to the chiller, with space cooling the best waste-heat sink because of its ~14 h/day utilization. Maintenance is light (chemicals every 8,000 h, no purge pump or spray nozzles, a single hermetically-sealed solution pump; ~$1,000 every three years claimed vs competitors’ $10,000+/yr), and factory-charged packages start up in under half a day. Off-grid economics drive adoption where peak grid power exceeds 30 ¢/kWh or transmission costs up to $1M/km.
  • YORK / Johnson Controls frames the wider market: single-effect steam at COP 0.7, double-effect steam 1.4, direct gas-fired 1.2, against electric centrifugals at 6.5 — and concedes that “higher COP does not necessarily result in low operational cost,” while electric chillers remain “a sound choice for mission-critical applications such as data centers.”
  • Energy Concepts builds custom ammonia-water machines for low-temperature heat: Chena’s three-pressure double-lift ThermoChiller and the ISAAC solar ice maker (35 lb of ice/day from a 37 ft collector at ~$11,000).

These sit alongside the adsorption vendors in Commercial Adsorption Chillers; other LiBr benchmarks are tabulated in Performance & Numbers.

Solar & off-grid experiments

Solar ammonia refrigerators are a recurring DIY/appropriate-technology thread: Vanek & Vanek’s intermittent parabolic-trough ice maker (140 ft² collector, ~200 psi generation, ~3-hour charge) produced ~10 lb of ice per day/night cycle for ~$510; Buehn et al.’s senior-design vaccine cold-chain refrigerator (2–8 °C for 24–36 h in 35–50 °C ambients) used the same pair at 8–14 bar for a ~$251 build. Both are labelled “absorption” but use a calcium-chloride salt rather than water as the sorbent (forming CaCl₂·8NH₃, six of the eight ammonia molecules usable), chosen because salt does not evaporate into the refrigerant — so strictly they are solid-sorbent chemisorption cycles, the lineage of the Crosley IcyBall (1927); see Adsorption vs Absorption. They mirror the intermittent solar adsorption ice makers in Solar Adsorption Cooling.

Combined absorption–adsorption

Because the two cycles share the same evaporator/condenser hardware and the same waste-heat input, hybrids are being studied: Hassan et al. modeled a combined LiBr/water absorption + silica-gel/water adsorption system sharing the evaporator and condenser — in series at 85 °C it delivered 0.344 kW at COP 0.623, 58% more capacity and 27% higher COP than the adsorption stage alone, with much smoother chilled-water temperature (±0.9 vs ±1.8 °C). Román et al. simulated solar absorption (rated COP 0.70, generator capped at 116 °C) and adsorption (COP 0.53, capped at 95 °C) chillers managed by a modulating tempering valve that blends tank and return water to hold the generator inlet at its limit, cutting auxiliary-heater energy — most for the absorption unit. The combination aims to capture absorption’s COP and adsorption’s low-temperature reach in one machine.

Geothermal as the drive heat

Geothermal brine is the one waste-heat stream that arrives 24/7 without combustion, and absorption is the only sorption cycle that has ever run on it in the field. The record is thin and instructive — worked out in full in Geothermal-Driven Sorption Cooling:

  • Drive floor. Geothermal-specific models put the single-effect LiBr/water minimum at ≥ 90 °C, with COP 0.73 at 100 °C rising to ~0.87 at 120 °C and falling beyond; LiCl/water is within 1–2%. Once the heat exchanger is sized, brine mass flow barely affects COP — temperature and HX area are the levers. Crystallization limits tighten at high absorber UA (~1.9 kW/K at 90 °C, 2.5 at 110 °C). El Haj Assad’s configuration bottoms the chiller on a flash plant’s 100 °C reject brine.
  • Derating on real brine. Oregon Tech’s 1,095 kW LiBr chiller (1980) was rated for 116 °C; on 89 °C well water it made 526 kW — half nameplate — while pumping 37.8 L/s at only 11 °C ΔT (~4.6 gpm per ton). It was replaced by an electric centrifugal in 1998 for “low efficiency and high water usage” and generator-tube corrosion. Modern 70–95 °C-rated machines (the Yazaki class above) would derate less, but the flow-per-ton penalty of low-grade heat stands.
  • Cold sink beats hot source. Chena Hot Springs runs a three-pressure double-lift ammonia-water chiller to −29 °C brine on 74 °C spring water — only possible because it rejects to 4 °C river water. Output fell whenever the creek warmed to 7 °C. Note the pair-specific asymmetry: ammonia-water thrives on the coldest sink available, while LiBr machines need condenser water above ~20 °C.
  • Versus an electric chiller on the same plant’s electricity. Single-effect absorption beats a COP-6.5 electric centrifugal on energy cost only above ~$0.127/kWh (derived from YORK’s own 500-ton worked example at $4/1,000 lb steam), rejects ~2.4 kW per kW of cooling vs ~1.15 (roughly double the tower duty and water), and on 87 °C brine consumed ~5.6× the input energy of the electric baseline at $74/MWh of cold versus $42/MWh of heat (NREL, Tuttle OK). The niche is heat that would otherwise be reinjected, high flat utilization, expensive electricity, and cheap or cold rejection.

The micro-scale gap and market advantages

Commercial absorption machines sit at 100–300 tons (industrial) down to 5 tons (Yazaki’s smallest water-fired package). Below that — the 1–5 ton residential and small-commercial band where most cooling demand lives — no absorption product exists. An r26D internal assessment (2023) frames this as the central opportunity: the technology works now at scale; what’s missing is a right-sized machine.

Three market-side advantages stack in absorption’s favor for that segment:

  1. No licensed-technician gatekeeping. Vapor-compression service requires licensed technicians because of HFC refrigerants and 220 V power. Absorption runs on low-pressure water, a small pump, and fans — no controlled refrigerant, no high voltage. This removes the installer bottleneck that slows even heat-pump and mini-split adoption (where technicians often prefer conventional equipment).
  2. Grid resilience. An absorption chiller’s only electricity demand is a small pump and fans, so it can run on a generator during a power outage — precisely when cooling is most life-critical (wet-bulb emergencies). Vapor-compression AC drives peak grid load and fails exactly when demand peaks.
  3. Hotter climate = more capacity. Absorption performance improves with higher drive heat, so a warming climate that degrades vapor-compression COP (higher condensing temperature → lower lift) simultaneously improves the thermal economics of waste-heat or solar-driven absorption.

Open question — LiBr from lithium batteries. End-of-life lithium-ion batteries contain lithium compounds. The chemistry is trivial — recycled Li₂CO₃ or LiOH converts to LiBr in a one-step acid-base reaction — but the economics don’t work today: recycled lithium costs ~3× virgin lithium, and battery-grade Li₂CO₃/LiOH is more valuable returning to new batteries than diverting to lower-value LiBr. No published study evaluates this specific pathway. Viability could flip on lithium oversupply from recycling, battery chemistry shifts away from lithium (Na-ion), or sufficient micro-scale absorption demand. See research note.

See Also

Sources