An oscillating heat pipe (OHP), also called a pulsating heat pipe, is a wickless, sealed tube that moves heat from one spot to another using the back-and-forth sloshing of liquid slugs and vapor bubbles inside a meandering channel. Unlike a conventional heat pipe that relies on a wick structure to return condensed fluid to the hot end, an OHP uses pressure differences between expanding and collapsing vapor plugs to drive fluid motion on its own, with no moving mechanical parts. The concept dates to a pair of patents filed in 1990 and 1993 by the Japanese inventor Akachi, who described a loop-type heat pipe in which a working fluid “circulates in a loop form in itself under its own vapor pressure.”1Elsevier / Applied Thermal Engineering. Innovations in pulsating heat pipes: From origins to future perspectives Since then, OHPs have attracted growing interest for cooling electronics, recovering waste heat in buildings, and even managing thermal loads in spacecraft.
How the Sloshing Actually Works
Picture a long, thin tube bent back and forth into a serpentine pattern, with one end touching something hot (the evaporator) and the other end touching something cool (the condenser). The tube is partially filled with a working fluid, and because the channel is narrow enough, surface tension breaks the fluid into alternating segments: liquid slugs and vapor plugs. When the evaporator heats a vapor plug, that plug’s pressure rises and pushes the neighboring liquid slugs toward the condenser. At the condenser, vapor plugs cool and shrink, pulling fluid back. The net effect is a self-sustaining oscillation that carries heat from hot to cold.
Researchers have modeled this motion as a kind of forced, damped vibration, where the driving force is the pressure difference between neighboring vapor plugs. One analysis showed that the oscillation’s amplitude and frequency don’t depend on where the liquid slug happens to sit when the device starts up, and that gravity can shift the phase of the oscillation by as much as 45 degrees without killing it.2International Journal of Heat and Mass Transfer. Flow and heat transfer of liquid plug and neighboring vapor slugs in a pulsating heat pipe That self-correcting quality is part of what makes OHPs appealing: the oscillation emerges naturally from the thermodynamics rather than requiring external pumps or control systems.
Why Channel Size Matters
For liquid slugs and vapor plugs to form in the first place, the channel has to be narrow enough that surface tension dominates over gravity. Engineers define a maximum hydraulic diameter beyond which the fluid would simply pool at the bottom of the tube rather than staying suspended as distinct slugs. In practice, most OHPs use channels somewhere between about 1 and 3 millimeters across, depending on the working fluid and the temperature range.
That said, the theoretical limit is not as rigid as textbooks sometimes imply. Researchers built a glass OHP with a 6-millimeter hydraulic diameter, far exceeding the conventional maximum for ethanol, and found that the device still functioned.3Journal of Heat Transfer. Visualization of an Oscillating Heat Pipe with Hydraulic Diameter Far Exceeding the Conventional Maximum Hydraulic Diameter The standard formula for the maximum diameter assumes a static balance between surface tension and gravity; once the fluid is already oscillating, the dynamic forces at play can maintain the slug-plug pattern in channels larger than the formula predicts. This is good news for manufacturing, because machining or extruding very tiny channels at scale is expensive and fiddly.
Closed-Loop Versus Closed-End Designs
OHPs come in two main flavors. In a closed-loop design, the serpentine tube’s two ends are connected to each other, forming a continuous circuit. In a closed-end design, each end of the tube is sealed off. The choice matters more than you might expect, especially when the device has to work in different orientations.
A comparative study of micro pulsating heat pipes found that the closed-end version was the more forgiving design at moderate turn counts. With 10 turns, the closed-end pipe achieved up to 2.5 times the effective thermal conductivity of the closed-loop version when mounted horizontally.4International Journal of Heat and Mass Transfer. Comparison of the thermal performances and flow characteristics between closed-loop and closed-end micro pulsating heat pipes The closed-loop pipe needed a full 20 turns before it could match that orientation-independent performance. However, at very low turn counts (5 turns), the closed-end pipe failed to start up at all, while the closed-loop version still worked in a vertical position. So neither design is universally better; the right choice depends on how many turns you can fit and whether the device might be tilted or flipped during use.
Fill Ratio and Heating Uniformity
How much liquid you pour into the tube before sealing it, expressed as a percentage of the total internal volume, is called the fill ratio or charging ratio. Too little fluid and the evaporator dries out. Too much and there isn’t enough vapor to generate pressure swings. The sweet spot typically falls somewhere between about 40% and 70%, but the exact optimum shifts depending on conditions.
One study that varied the fill ratio between 50% and 70% under different heating patterns found that the best fill ratio depended on how evenly the heat was applied. When the heat was uniform, 50% performed best. As the heating became more lopsided, higher fill ratios (60% and then 70%) gave better results, because the extra liquid helped maintain slug-plug flow in the portions of the evaporator receiving less heat.5Elsevier. Thermal performance characteristics of a pulsating heat pipe at various nonuniform heating conditions The practical takeaway is that you can’t just pick a single golden fill ratio from a table; you have to account for whether the heat source is a nice even spread or a concentrated hot spot.
What Happens Inside the Tube
High-speed camera studies have given researchers a close look at the flow patterns inside transparent OHPs. Three broad categories of flow show up: dispersed small bubbles, distinct vapor plugs separated by liquid slugs, and long vapor plugs that span multiple turns of the tube.6International Journal of Heat and Mass Transfer. High speed flow visualization of a closed loop pulsating heat pipe At low heat inputs, the motion is gentle and intermittent. As you turn up the heat, the oscillations grow larger and faster, and the short vapor plugs can merge into long ones that sweep liquid along the channel in bulk. Recent visualization work using magnetic nanofluids and cameras running at 100 frames per second has extended these observations to newer fluid formulations, confirming that the same general flow-pattern categories apply even when the fluid is loaded with nanoparticles.7Scientific Reports. Insights into two-phase flow dynamics in closed-loop pulsating heat pipes utilizing Fe3O4/water: experimental visualization study
Understanding these flow regimes isn’t just academic curiosity. When the evaporator gets too hot, the liquid slugs nearest the heat source can evaporate entirely, and the device enters a “dryout” regime where heat transfer collapses. A recent modeling framework captured three distinct operating regimes: a pre-startup phase where very little happens, steady oscillation where the device does its job, and dryout where it fails.8ASME Journal of Heat and Mass Transfer. Microscale Oscillating Heat Pipe Model for Prediction of Startup, Oscillation Dynamics, and Dryout Predicting where the boundary between useful oscillation and destructive dryout lies for a given design is one of the trickier challenges in the field.
Boosting Performance with Specialized Fluids
Water, ethanol, methanol, and acetone are the most common working fluids, but researchers have spent years trying to improve on them. Two strategies have gained the most traction: adding nanoparticles to the base fluid and using so-called self-rewetting fluids, which are mixtures whose surface tension increases with temperature rather than decreasing. That unusual property creates a surface-tension gradient that actively pulls liquid back toward the hottest zone, counteracting dryout.
An experiment combining both approaches, mixing nanoparticles into a self-rewetting fluid, found that the resulting “self-rewetting nanofluid” boosted an OHP’s heat transfer performance by about 16% compared to the self-rewetting fluid alone, and about 12% compared to a nanofluid without the self-rewetting additive.9International Journal of Heat and Mass Transfer. Experimental study on the heat transfer performance of an oscillating heat pipe with self-rewetting nanofluid A separate study using carbon nanotube-based self-rewetting nanofluids confirmed similar trends, with the self-rewetting fluid alone reaching up to a 22% improvement over pure water at higher heat inputs.10Case Studies in Thermal Engineering. Comparisons of thermal performances in a pulsating heat pipe by using a nanofluid and a self-rewetting nanofluid with carbon nanotubes These are meaningful gains for a device that already works reasonably well with plain water.
The catch is that nanoparticle suspensions can settle out over time or clog narrow channels, and the optimal concentration of both nanoparticles and the self-rewetting additive (typically a small amount of an alcohol like n-butanol) is specific to the geometry and operating range. There’s no universal recipe yet.
Steering the Flow with Tesla Valves
In a closed-loop OHP, the oscillating fluid can theoretically settle into a net circulatory flow in one direction, which tends to improve performance. But left to its own devices, the flow direction can be random and intermittent. One clever fix borrows from Nikola Tesla’s 1920 patent for a check valve with no moving parts: a channel shape that offers low resistance in one direction and high resistance in the other.
Researchers integrated Tesla-type check valves into the adiabatic (middle) section of a flat-plate OHP and found that circulation in the desired direction was promoted, with the effect becoming stronger at higher heat inputs. The valve-equipped OHP consistently had lower thermal resistance than an identical device without valves.11Experimental Thermal and Fluid Science. Investigation of a flat-plate oscillating heat pipe with Tesla-type check valves Because Tesla valves have no moving parts and no seals to fail, they fit naturally inside a device whose main selling point is simplicity and reliability.
Surface Treatments That Lower Startup Temperature
The interior surface of the channel also affects performance. Making the evaporator section super-hydrophilic (strongly attracting water) helps liquid wet the surface and form a thin film that evaporates easily, while making the condenser section super-hydrophobic (strongly repelling water) encourages droplets to form and shed quickly, speeding up condensation.
A study on three-dimensional OHPs with this combination of surface treatments found that the startup temperature dropped by more than 22% in a vertical orientation and about 13.5% in a horizontal orientation compared to an untreated pipe. The effective thermal conductivity improved by roughly 15% vertically and a striking 62% horizontally.12International Journal of Thermal Sciences. Study on the thermal performance of three-dimensional oscillating heat pipe with super-hydrophobic/super-hydrophilic surface for thermal management application That horizontal improvement is especially interesting because OHPs generally perform worse in horizontal orientation, where gravity can’t assist the flow. Tailored surface coatings appear to partially compensate for the loss of gravitational assistance.
Cooling Server Chips and Power Electronics
One of the most active application areas for OHPs is cooling high-power computer chips. A dumbbell-shaped, three-dimensional flat-plate OHP designed for server CPU cooling demonstrated that it could handle 150 watts of heat input while reducing the chip junction temperature by about 49% compared to a plain aluminum plate of the same size. Scaled up into a full heat sink module, the same design managed 300 watts (a heat flux of roughly 41 W/cm²) while keeping the chip below 85 °C, a safe operating ceiling for most processors.13Energy Conversion and Management. A dumbbell-shaped 3D flat plate pulsating heat pipe module augmented by capillarity gradient for high-power server CPU onsite cooling
At the other end of the scale, researchers have explored embedding OHPs directly inside printed circuit boards to deal with localized hot spots from power transistors. A 44-channel open-loop design tested with dielectric working fluids (chosen for their low global warming potential) showed that the concept is viable for low-voltage power electronics applications where the board itself needs to be the heat spreader.14Applied Thermal Engineering. Thermal performance of a PCB embedded pulsating heat pipe for power electronics applications Embedding the heat pipe into the board eliminates the thermal interface layers that usually sit between a chip and a separate heat sink, which can account for a big chunk of the total thermal resistance.
Building Energy Recovery
OHPs aren’t limited to tiny electronics. A study explored using large-format oscillating heat pipes for heat recovery ventilation in commercial buildings. The idea is to use an OHP panel to transfer heat between outgoing stale air and incoming fresh air, reducing the energy the HVAC system needs to spend on heating or cooling. Modeling showed that such a system could pre-heat or pre-cool incoming air by more than 5 °C, recover upwards of 5 kilowatts of heat, and maintain a reasonable pressure drop below 200 pascals. In cities with continental climates like Chicago and Denver, the projected annual savings exceeded $2,500 per building.15Energy Reports. Energy and cost savings potential of oscillating heat pipes for waste heat recovery ventilation Compared to rotary wheel heat exchangers, an OHP-based system has no moving parts and no risk of cross-contaminating the air streams, which matters in hospitals and laboratories.
Operation in Microgravity
Because OHPs rely partly on pressure-driven oscillation rather than gravity-driven fluid return, they are candidates for space applications. Research on OHP behavior under microgravity conditions found that after startup, the system reached steady state in less than a second, with slug and bubble displacements settling into quasi-sinusoidal oscillations accompanied by small pressure fluctuations.16International Communications in Heat and Mass Transfer. Operational characteristics of oscillating heat pipes under micro-gravity condition The fact that oscillation sustains itself even without gravity is encouraging, though most microgravity testing so far has been limited to parabolic flight campaigns and drop towers, which provide only brief windows of weightlessness. Long-duration orbital testing remains relatively scarce.
3D Printing Opens New Geometries
Traditional OHPs are made by bending copper or aluminum tubing or machining channels into flat plates. Additive manufacturing (3D printing of metals) is starting to change what’s geometrically possible. A team used selective laser melting to fabricate a compact titanium alloy OHP measuring roughly 5 cm × 4 cm × 1.6 cm, containing four interconnected layers of circular mini-channels just 1.53 mm in diameter.17Additive Manufacturing. Additive manufacturing of heat exchangers: A case study on a multi-layered Ti–6Al–4V oscillating heat pipe A follow-up investigation of the same multi-layered design confirmed that it could be built as a sealed, hermetic unit with an integrated fill port, all printed in a single build.18International Journal of Heat and Mass Transfer. An investigation of a multi-layered oscillating heat pipe additively manufactured from Ti-6Al-4V powder
This matters because stacking channels in three dimensions dramatically increases the heat-transfer area you can pack into a given footprint. Titanium alloy is also far lighter than copper, which is appealing for aerospace. The tradeoff is that as-printed surface roughness can be high, and post-processing (polishing the inner walls of a 1.5-mm channel buried inside a solid block) is not trivial. Still, the approach demonstrates that OHP designs no longer have to conform to what a tube-bending machine can produce.
Flexible and Polymer-Based Designs
Copper and titanium aren’t the only options. Flat-plate OHPs made from flexible polymers have been tested to see whether bending them degrades performance, which matters for applications like wearable electronics or conformable thermal management on curved surfaces. The results were reassuring: bending the polymer OHP to various angles changed the startup heat input somewhat (ethanol-filled devices needed more heat to get going when bent more sharply), but once running, the overall thermal resistance settled to roughly the same asymptotic value regardless of the bend angle.19Applied Thermal Engineering. Performance of flat-plate, flexible polymeric pulsating heat pipes at different bending angles In other words, you can wrap one of these around a curved battery pack or mount it on a body-conforming surface without giving up much thermal performance.
Why Modeling Remains Difficult
Despite decades of experimental work, predicting how a specific OHP design will perform from first principles remains frustratingly hard. The flow involves simultaneous boiling, condensation, capillary effects, and oscillating two-phase motion in a geometry that can have dozens of turns, all interacting nonlinearly. Traditional computational fluid dynamics approaches struggle with the computational cost of tracking every liquid-vapor interface in real time across a full device.
Machine learning has emerged as a shortcut. A review of the literature found that neural networks, support vector machines, and other algorithms have been used to predict OHP thermal resistance, optimal fill ratios, and startup behavior with reasonable accuracy when trained on experimental data. But the review also flagged fundamental open questions: which model architectures are best suited to the task, and whether predictions trained on one lab’s experimental setup can extrapolate reliably to a different geometry or working fluid.20ASME Digital Collection. Machine Learning for Modeling Oscillating Heat Pipes: A Review For now, building and testing prototypes remains faster and more trustworthy than simulation for most practical design decisions, which is an unusual situation for a thermal technology that has been studied for over 30 years.