Solid state cooling explained: what a solid state fan, cooler, cooling system and solid state refrigeration each mean, and how the four approaches differ.
Ask four vendors for a solid state cooler and you will be shown four different objects. One is a silicon chip that vibrates a membrane at ultrasonic frequency. One is a thin film that ionises a shallow layer of air. One is a ceramic tile that pushes heat from one face to the other. One is a shape memory alloy loaded to several hundred megapascals. All four are sold under the same two words, and only two of them actually move heat. The rest move air.
That confusion is the reason this page exists. The technology is real and the category is growing, but the vocabulary has outrun the engineering, and buyers routinely specify one thing and receive another.
Solid state cooling is any thermal management method that moves heat or moves air using a field effect inside a solid material, with no rotating fan, no compressor and no pump. The field can be electrical, magnetic or mechanical. What every member of the family shares is the absence of a rotating assembly, bearings and the wear mechanisms that come with them.
This is the part most explainers skip. Each of these phrases is searched by a different kind of buyer, and each usually, though not always, means something specific.
The umbrella term for the whole category. It covers heat pumps and air movers alike, which is precisely why it is unhelpful in a specification document. Use it to name the field, then say which mechanism you mean.
Usually a discrete component you buy, mount and drive: a Peltier module, a micro air mover, an actuator with its driver board. In practice the phrase gets applied to thermoelectric modules and to microscale air movers interchangeably. If someone offers you a solid state cooler, the useful follow-up question is whether it lowers a surface below ambient temperature or simply increases convection at that surface.
A device that performs a fan's job, moving air across a hot surface, without a rotating impeller. Frore Systems and xMEMS both frame their products this way, and it is a fair description. xMEMS' XMC-2400, for example, measures 9.26 by 7.6 by 1.08 mm, weighs about 150 mg and delivers up to 39 cubic centimetres of air per second at up to 1,000 Pa of back pressure [1]. It is a genuinely impressive piece of miniature cooling hardware.
One nuance is worth knowing rather than arguing about. Devices in this class contain elements that move at microscale: xMEMS describes all mechanical operation as taking place at ultrasonic frequencies [1], and Frore's AirJet uses membranes that vibrate ultrasonically to produce pulsating jets [2]. Calling them solid state is legitimate, because there are no bearings and no rotating mass. It simply means their long-term failure mode is membrane fatigue rather than bearing wear. Our comparison of a DBD plasma actuator against a piezoelectric fan goes into that trade in detail.
The important point: a solid state fan does not make air colder. It moves air. Its floor is ambient temperature, exactly like a rotary fan's.
The assembly rather than the part: actuator or module, drive electronics, heat spreader or fin stack, sensing and control. People searching for solid state cooling systems are typically specifying at the enclosure level, not the component level, and the honest answer for every technology in this article is that the bare device is only part of the thermal budget.
Genuinely different, and this is the single most useful distinction on this page. Refrigeration means pumping heat against a temperature gradient so that one surface goes below ambient. Thermoelectric, magnetocaloric and elastocaloric devices do this. Solid state fans and ionic wind devices do not, and no amount of airflow will make them.
Colloquial shorthand for an electrohydrodynamic air mover, one that accelerates ions in an electric field and drags neutral air along with them. Note two things. First, vendors rarely use the phrase themselves: Ventiva, whose devices work on exactly this principle with an emitter and a collector electrode, calls its product an ionic cooling engine rather than an ionic fan [3]. Second, the same phrase is used in the consumer air purifier market for a completely different product. If you are sourcing for electronics, say electrohydrodynamic or ionic wind and the ambiguity disappears.
Everything above collapses into one question, and it is the question that decides whether a technology can solve your problem at all.
Air movers increase the rate at which a hot surface loses heat to the surrounding air. They can never take that surface below the ambient air temperature. Ionic wind devices, piezoelectric and MEMS air movers, and ordinary fans all sit here.
Heat pumps move thermal energy from a colder body to a hotter one using external work. They can take a surface below ambient. Peltier modules, magnetocaloric and elastocaloric devices sit here. The catch is conservation: the heat rejected on the hot side equals the heat absorbed on the cold side plus the electrical work consumed. Put a thermoelectric cooler inside a sealed box and the box gets hotter overall, even as the target component gets colder.
This is why the two families are not substitutes. If a laser diode must hold a set point below ambient, no solid state fan will do it. If a processor simply needs better convection in a 3 mm gap, a heat pump is the wrong tool and an expensive one.
| Approach | What it physically does | Representative published figures | Where it wins | Where it does not |
|---|---|---|---|---|
| Thermoelectric (Peltier) | Pumps heat. Below-ambient capable | Bismuth telluride modules reach ZT of roughly 1; review data puts thermoelectric COP at 0.38 to 0.45 against 2.6 to 3.0 for conventional systems [4] | Precise set-point control, small loads, no working fluid, mature supply chain | Low efficiency, large hot-side rejection load, solder fatigue under thermal cycling |
| Piezoelectric and MEMS air movers | Moves air. Ambient-limited | AirJet Mini: 2.8 mm thick [2], up to 5.25 W of heat removed at 1 W input, 21 dBA, 1,750 Pa back pressure [5]. xMEMS XMC-2400: 1.08 mm thick, 39 cc/s, 1,000 Pa [1] | High back pressure for the volume, reflowable packaging, shipping products today | Membrane fatigue as the wear mechanism, module-level heat budget per part |
| Electrohydrodynamic and ionic wind, including DBD plasma | Moves air. Ambient-limited | Ionic wind on a 0.3 m/s bulk flow doubled the average heat transfer coefficient and added around 20 K of cooling for under 100 mW [6]. Ventiva reports up to 100 W TDP at under 15 dBA [3] | Extremely thin and conformal, no mechanical wear element, very low power for local enhancement | 1 to 2 per cent electrical-to-fluid conversion efficiency [6], kilovolt drive, ozone management, metres per second not tens |
| Magnetocaloric and elastocaloric | Pumps heat. Below-ambient capable | An elastocaloric prototype reported a 25.4 K load-free span and system COP of 1.27 with work recovery [7]. Vapour compression sits near COP 3.6, and caloric systems need COP of four or higher to be worth switching to [8] | No refrigerant gases, high theoretical efficiency ceiling | Research and prototype stage, rare-earth magnets or 400 to 600 MPa stresses, fatigue over millions of cycles [8] |
Read the second column first. Two rows pump heat and two rows move air. That single split matters more than any performance number in the table.
YPlasma's Y-Flow thermal actuators sit in the third row. A dielectric barrier discharge actuator is two electrodes separated by a thin dielectric, one exposed to the air and one encapsulated. A kilovolt-class AC drive breaks down a shallow layer of air at the surface. The ions accelerate along the surface and collide with neutral molecules, dragging them into a wall jet of a few metres per second. The whole stack is roughly 200 micrometres thick and has no moving parts at any scale.
The discharge is non-thermal. The gas stays near ambient temperature, so the actuator does not meaningfully add heat to the surface it is cooling, which matters when the entire power budget for a local hotspot is a fraction of a watt.
The measured behaviour is specific and modest. Superimposing an ionic wind on an existing 0.3 m/s bulk flow roughly doubled the average heat transfer coefficient and produced about 20 K of additional cooling, drawing under 100 mW [6]. That is a boundary-layer result, not a bulk airflow result, and the distinction is the whole design philosophy: attack the thermal resistance where the air is already nearly stationary, rather than trying to push more air through the chassis.
The technology does not win everywhere, and pretending otherwise wastes engineering time.
Electrical-to-fluid conversion efficiency for corona-based ionic wind sits at around 1 to 2 per cent [6]. As a bulk air mover that is poor, and a rotary fan remains the correct choice when the problem needs high volumetric flow. Ionic wind produces metres per second, not tens of metres per second. Where a competing approach is genuinely stronger, it should be said plainly: Frore's and xMEMS' devices deliver far higher back pressure per module than an ionic wind film does, and they are shipping in commercial products now, which an early-stage technology cannot claim.
Beyond that: the drive is kilovolt-class AC, which is a design constraint on the power supply and on creepage and clearance. Operation in air generates some ozone, dependent on waveform, geometry and duty cycle, and manageable by design rather than eliminated, which we cover in ionic wind reliability and ozone. Dielectrics age under continuous electrical stress. And none of the air-moving technologies here, ours included, can take a surface below ambient.
For thermoelectric, the honest limit is efficiency: review data places typical COP well below one [4], so every watt of cooling costs more than a watt at the wall and adds to the hot-side load, which we examine against our own approach in plasma actuators versus thermoelectric coolers. For magnetocaloric and elastocaloric, the honest limit is maturity: these are laboratory and prototype cooling machines today, not components you specify into a product this year [8].
Work through it in this order. Does the surface need to go below ambient? If yes, you are shopping for solid state refrigeration, and thermoelectric is the only branch that is commercially mature. If no, you are shopping for an air mover, and the question becomes geometry: how much space you have, how much back pressure the flow path demands, and whether the enclosure is sealed. Thin, conformal and sealed favours the electrohydrodynamic branch. Higher back pressure through a duct favours MEMS. Plenty of space and no acoustic constraint still favours a fan, as our comparison with conventional fans sets out. Densely packed compute at the edge, where none of the space assumptions hold, is its own problem, discussed under Edge AI thermal management, and a broader vendor-by-vendor view is in the 2026 solid state cooling landscape.
A solid state fan is a device that moves air across a hot surface without a rotating impeller or bearings, typically using a vibrating microscale membrane or an electric field acting on ionised air. It replaces a fan's function, not a refrigerator's: it increases convection but cannot take a surface below the ambient air temperature.
For thin, sealed or acoustically sensitive designs, yes, because there is no rotating assembly to wear out and no impeller to fit. For raw volumetric airflow at low cost, no. A rotary fan still moves far more air per unit of electrical power, so the right answer depends on whether your constraint is space and reliability or bulk flow.
An ionic fan is an air mover that ionises air molecules in a strong electric field and accelerates them, dragging neutral air along to create a flow known as ionic wind. It has no moving parts at all. Note that the same phrase is also used for consumer air purifiers, which are a different product entirely.
It depends on which branch. Ionic wind used as a local boundary-layer enhancer can add meaningful cooling for under 100 mW [6], which is very low. Thermoelectric cooling uses considerably more power than the heat it moves, with typical COP well below one [4]. So solid state cooling is not automatically lower power, and the mechanism decides.
Solid state cooling is the umbrella term and includes devices that only move air. Solid state refrigeration is the subset that actually pumps heat against a temperature gradient, so it can take a surface below ambient. Peltier, magnetocaloric and elastocaloric devices refrigerate. Ionic wind and MEMS air movers do not.