What if you could make cold by squeezing metal? No compressor, no chemical refrigerant, no hiss of leaking gas. Just a bundle of metal tubes that get warm when you squeeze them and turn cold when you let go. That is the core idea behind elastocaloric cooling, a solid-state cooling system that uses stress in shape-memory alloys to pump heat.
It sounds like a lab trick, and for years it was. Researchers could feel the chill in a stretched nickel-titanium wire held between their fingers. The challenge was turning that flicker of cold into steady, useful cooling for a fridge or an air conditioner. In 2023, a team at the University of Maryland showed a practical path forward with a multimode prototype that ranks among the highest-performing caloric cooling systems ever reported, as described in their paper in Science. Since then, larger and colder prototypes have followed, and they point to a future where at least some cooling no longer depends on climate-damaging gases.
How Squeezing Metal Makes Cold
Elastocaloric cooling starts with a shape-memory alloy, most often nitinol, an alloy of nickel and titanium. At room temperature, nitinol sits in a crystal structure that can switch back and forth under stress. This reversible switch is called a martensitic transformation.
The cycle works in four steps:
- Squeeze or stretch the metal. The crystal structure snaps into a more ordered phase and releases latent heat, so the metal warms from its starting temperature to a higher temperature.
- Reject that heat to a sink. Water or another heat-transfer fluid carries the extra heat away, and the metal returns to its starting temperature while still under load.
- Release the stress. The crystal snaps back, absorbs heat, and the metal drops below its starting temperature.
- Absorb heat from what you want to cool. The cold metal pulls heat from indoor air, a fridge cabinet, or electronics, and then the cycle repeats.
Because the refrigerant itself stays solid the whole time, there is nothing to leak into the atmosphere. Engineers describe the process using the shape-memory effect of certain metals that produce a reversible temperature change under tensile or compressive load. Repeat the load and unload rhythm hundreds of times per minute, route fluid through valves at the right moment, and you have a heat pump.

Why Air Conditioning Needs Reinventing
Conventional cooling is an engineering triumph, yet it comes at a steep cost to the climate. Air conditioning, refrigeration, and other cooling technologies account for more than 20 percent of global electricity use, according to the Maryland team, while the hydrofluorocarbon refrigerants inside vapor-compression systems have a global warming potential thousands of times greater than carbon dioxide.
Even small leaks matter because those gases accumulate. The world is already phasing down high-GWP HFCs under the Kigali Amendment, and the European Union tightened its F-gas rules with Regulation 2024/573, which pushes toward a 95 percent cut in hydrofluorocarbon consumption from 2015 levels by 2030 and a full phase-out by 2050. Natural refrigerants like carbon dioxide, propane, and ammonia help, but they bring trade-offs around pressure, flammability, toxicity, and hot-climate efficiency.
Elastocaloric systems sidestep that dilemma. They use no vapor, no HFCs, and zero direct greenhouse gas emissions from the refrigerant itself. The remaining climate question is indirect emissions from the electricity that drives the actuator, which is why efficiency matters as much as eliminating leaks.
The 260-Watt Prototype That Proved It
The Maryland breakthrough came from pairing materials science with systems engineering. Materials professor Ichiro Takeuchi, mechanical engineering professors Reinhard Radermacher and Yunho Hwang, and their team built a device around fatigue-resistant nitinol tubes squeezed in compression rather than pulled in tension.
Compression matters because it lasts longer. Tension tends to open cracks, while compression keeps flaws closed. The team also designed a versatile fluid network that can switch between two operating modes by changing valve timing. One mode builds a wider temperature span through active regeneration, while the other delivers cooling more efficiently under heavy loads. Combining the best of both in one practical prototype was the key advance.
The result was 260 watts of useful cooling power and a maximum temperature span of 22.5 kelvin (equal to 22.5 degrees Celsius, or about 40.5 degrees Fahrenheit). That is enough for a compact wine fridge, and the group said commercial viability could follow within several years if efficiency and force requirements improve, scaling from small fridges to window units, whole-home systems, and commercial HVAC. Takeuchi put it simply when he called it a completely different, completely green, environmentally friendly cooling technology that relies on pushing and pulling pieces of metal to create cooling.
Inside the Metal: Nitinol, Fatigue, and Better Alloys
Commercial-grade nitinol remains the workhorse for almost every prototype published to date. It offers a useful latent heat of 10 to 15 joules per gram, equivalent to roughly 20 to 30 kelvin of direct temperature change, plus strong mechanical properties. Direct measurements on elastocaloric materials have shown temperature jumps as large as 31.5 kelvin for nickel-manganese-titanium (NiMnTi) and 38.5 kelvin for nickel-titanium (NiTi) under practical stress levels.
The catch is durability and efficiency. A practical cooler must survive millions of load cycles without cracking or losing its cooling effect. Reviews of the field put the target above one million cycles, with an ideal closer to 10 million stable cycles. Hysteresis, the wasted energy in each squeeze-release loop, also drags down efficiency in standard nitinol.
Three paths are emerging:
- Load in compression, not tension, to extend fatigue life.
- Move to low-hysteresis alloys such as TiNiCu or TiNiCo that need less mechanical work per cycle.
- Recover work between cycles, using the energy released on unloading to help drive the next compression.
Maryland researchers estimated overall system efficiency could improve by a factor of six with more efficient actuators, and they are testing a copper-based material that shows a similar cooling effect under much smaller stress. That softer option would shrink motors, frames, and costs.
Metals are not the only option. Polymers such as natural rubber and engineered elastomers also warm when stretched and cool when released, because their long molecular chains line up under tension and relax on unloading. Lab elastomers with uniform chain lengths have shown cooling near 15 kelvin, and thermoplastic polyurethane fibers have produced about 6 kelvin of cooling at only 7.5 megapascals (MPa) of stress, far below the hundreds of MPa metals need. The trade-off is large stretch ratios and small cooling per volume, which is why polymers suit rotary or roller-driven concepts and soft, low-force devices rather than compact fridges.
Heating matters too. Run the same cycle in reverse against outdoor air and an elastocaloric device becomes a heat pump, offering both summer cooling and winter heating from one refrigerant-free core. That dual role is part of why European programs frame the technology as heating and cooling together, not cooling alone.
How Efficient Is It Really?
Efficiency in cooling is measured as coefficient of performance, or COP: useful cooling divided by work input. Higher is better.
At the material level, elastocaloric alloys look excellent, with reported material COP values above 30 in some analyses. At the system level, reality is tougher because motors, pumps, valves, and heat exchangers all consume energy. One solid-contact laboratory setup reported an electrical COP of 1.27 with 7 watts of cooling, while modeling of a newer regenerative concept projected a steady-state system COP of 5.8, still to be verified experimentally.
A 2019 multistage device from Saarland University and partners reached a 28 kelvin span, showing the stacking strategy works, though at modest cooling power. Later roller-driven and compact wire-based prototypes pushed spans near 9 to 25 kelvin with single-digit watts. Recent prototypes show rapid progress on both power and cold temperatures. A February 2025 device from the Hong Kong University of Science and Technology delivered 1,284 watts of cooling power and survived more than 500,000 initial cycles.
In January 2026, the same group reported a sub-zero elastocaloric freezer that reached minus 12 degrees Celsius from a 24-degree-Celsius heat sink with a COP of 3.4. A separate 3D-printed NiTi prototype produced 50 watts with a 20-degree-Celsius span over 31,500 cycles. The trend is clear: power is climbing, temperatures are falling, and lifetimes are stretching toward commercial needs.
Elastocaloric vs. Other Solid-State Cooling
Elastocaloric cooling belongs to a family of caloric technologies that move heat by changing a solid, not by boiling a gas. Each uses a different trigger:
| Technology | Trigger | Strength | Limitation |
|---|---|---|---|
| Elastocaloric | Mechanical stress | Large temperature lift, practical actuators, high power density | Fatigue, large forces, heat-transfer design |
| Magnetocaloric | Magnetic field | Decades of development, efficient materials | Costly magnets, bulky field sources |
| Electrocaloric | Electric field | Compact, fast response | Small volumes, high voltages |
| Barocaloric | Hydrostatic pressure | Works with soft materials including perovskites | Very high pressures, sealing challenges |
| Ionocaloric | Ion flow | Pumpable materials, dual heating and cooling | Early stage, scaling unproven |
Maryland’s team noted their eight-year-old elastocaloric program had already matched the performance level of magnetocaloric systems developed since the 1980s. Unlike magnetocaloric devices that need large magnets and electrocaloric devices that need large electric fields, elastocaloric devices can use practical mechanical actuators.
Optimization is also getting more systematic. In December 2025, Maryland researchers introduced a non-dimensional design metric called thermal utilization ratio (Z*) to better balance fluid flow, cycle timing, and cooling output across different elastocaloric system designs. Tools like this help engineers compare prototypes fairly instead of chasing a single headline number.
For readers following solid-state cooling, the contrast is useful. Harvard engineers built a barocaloric prototype from metal-halide perovskites that stays solid through both stages but needs about 2,900 psi, far above the 150 psi of a typical air conditioner. Ionocaloric cooling from Lawrence Berkeley National Laboratory takes yet another route, using ions to drive a solid-to-liquid change with a pumpable material. Elastocaloric stands out for using familiar metals and mechanical force.

Where You Will See It First
Do not expect a whole-home elastocaloric heat pump next year. Reviews still rank technical maturity below vapor compression for COP and practicality, with fatigue life, system COP, and heat-transfer efficiency as the critical bottlenecks.
The near-term path runs through compact, high-value uses where small size and no refrigerants matter most:
- Portable and compact refrigeration, such as wine fridges and hotel minibars.
- Electronics and battery cooling, where high power density pays off.
- Medical and laboratory cold boxes that must avoid refrigerant leaks.
- Zoned or supplemental home cooling paired with efficient envelopes, much like radiant Cold Tube panels that cut cooling energy by absorbing body heat directly.
Residential and automotive uses remain longer-term goals pending better materials and work-recovery drives. European projects such as Fraunhofer IPM systems and Slovenia’s E-CO-HEAT device are now pushing fatigue-resistant regenerative designs toward industry testing.
If those efforts hold, the first product you meet may look ordinary: a quiet cabinet that stays cold because metal inside is being squeezed and released, over and over, thousands of times a day. The vivid part is what is missing: no HFC charge, no leak detector, no end-of-life refrigerant recovery. Just physics, metal, and motion.
