Ionic Winds: A New Frontier for Air Cooling

AIR-COOLING IS THE oldest and, in many ways, the easiest method of cooling electronics, whether it be through fan-driven forced convection or simply natural convection. However, with the increasing speed of processors, shrinking of form factors, and expansion of device functionality, air-cooling has begun to find itself on the outside looking in as new cooling technologies are sought, and new developments in solid-state cooling, such as thermoelectrics, and liquid cooling, such as two-phase microchannels, have become more popular in the heat transfer research community. Yet, because it is cheap, easy-to-implement, and a “known quantity,” forced convection air-cooling remains a very attractive approach for thermal management. Further, since in most cases, particularly in portable electronic devices, the entire heat load is eventually dissipated to ambient air, it will always be a critical component of any cooling solution. However, because of the ongoing evolution of electronics and new technologies such as light emitting diodes, advances in air-cooling technologies must also keep pace.

Classically, forced convection is driven by the conversion of mechanical-to-fluid energy through a fan. However, these fans often come with power consumption and acoustic penalties, as well as issues associated with moving parts and challenges facing miniaturization. This is especially pertinent in the areas of portable electronics and light emitting diodes (LEDs). One promising new approach is to utilize electrical-to-fluid energy conversion, a so-called ionic wind, where bulk air motion is not generated by a rotating or flapping mechanical structure but by a gas discharge. Recent advancements have shown that ionic winds could be a new frontier of air-cooling.

History of Ionic Winds

An ionic wind (also called an electric wind or corona wind) is generated when a gas discharge is formed between two electrodes in ambient air (Figure 1). Gaseous ions formed in the discharge are accelerated by the electric field and undergo collisions with neutral gas molecules. This exchange of momentum, and the subsequent cascading effect, generates a bulk fluid motion called an ionic wind. The ionic wind phenomenon has been known for centuries dating all the way to Francis Hauksbee who first observed the “electric wind” phenomena in 1709, with such notable scientists such as Sir Isaac Newton, Benjamin Franklin, and Michael Faraday all recording their own observations of the effect [1]. In the modern era, ionic winds began to be considered for practical engineering applications with two seminal publications by Stuetzer [2] (1959) and Robinson [3] (1961), who developed some of the fundamental relations still used today. Though investigations for heat transfer applications enjoyed some attention in the 1960s and 1970s, it is only recently that it has become a viable technological option, with companies such as Tessera Technologies, Inc. [4] and Ventiva [5] exploring technologies for laptop and LED cooling.

Figure 1. (a) Schematic of positive mode point-to-plane corona discharge-generated ionic wind. (b) Schematic of dielectric barrier discharge ionic wind, also called a plasma actuator.

















There are two types of gas discharges typically used to generate an ionic wind – the direct current (DC) corona discharge and the alternating current (AC) dielectric barrier discharge – both of which are favored because of their inherent stability at atmospheric pressure. A corona discharge (Figure 1a) is formed when high potential is applied between a sharp (corona) and blunt (counter) electrode. The high degree of inhomogeneity in the electric field due to the asymmetric electrodes causes a partial breakdown of the air gap and a small plasma region to be formed around the sharp electrode. The remaining air gap then acts as a drift zone where the large ions drift from the plasma zone toward the counter electrode, and it is in this region where the ionic wind is generated. Corona discharges, and the subsequent ionic winds, can be generated in both positive mode (the corona electrode is the anode) and negative mode (the corona electrode is the cathode), as well as in pulsed modes.  An AC dielectric barrier discharge (Figure 1b) is formed between two electrodes operated at oscillating high potentials, where at least one of the electrodes is covered by a dielectric, insulating material. This insulating material collects surface charges that serve to saturate the potential across the electrode gap and extinguish the discharge, thus preventing sparks. By using an oscillating field, the discharge is essentially extinguished and reignited every half-cycle. By using asymmetric electrode geometry, an ionic wind wall jet is formed along the surface of the insulating material.

There are a few drawbacks to both of these methods.  Both require high potentials (> 1 kV) and generate ozone as a by-product of operation in atmospheric air. Further, because they are gas discharges, there is always the potential for sparking, which can be a safety issue; although with proper design and operation this can often be prevented with high certainty. In general, corona discharges have been favored for electronics cooling applications because the required DC electronics is often considered more practical for consumer products.

Recent Developments

Figure 2. (a) Schematics of two common electrode configurations for ionic wind duct blowers. (b) Schematic of using an ionic wind aligned with a flow to modulate the boundary layer. The ionic wind forms a wall jet that accelerates the flow to thin the boundary layer and increasing convection.















For electronics cooling and heat transfer, an ionic wind can be produced and used in different ways. Original studies investigated directing the ionic wind at the hot surface, which also served as the counter electrode (Figure 1a), such that the impinging wind would directly cool the surface [6]. However, recent developments have focused on two operating modes – using the ionic wind as a ducted blower to replace a fan (Figure 2a) or using it to modulate a pre-existing flow (Figure 2b). In blower mode, a corona electrode is aligned with the axis of a duct and either a permeable mesh or surface counter electrode is used. The ionic wind generated between the two electrodes is then directed down the duct. A significant amount of effort has focused on duct and electrode design to maximize the raw flow rate through the duct, including creating multiple blowers in series, such that flow rates as high as 70 slm (standard liters/minute) have been reported [7]. One new development that shows significant promise is using multiple downstream collecting electrodes to form a so-called assisted corona discharge ionic wind [8]. This elongates the drift zone so that flow is more effectively generated, but avoids the necessity for very high applied potentials because the primary collecting electrode acts as a gate electrode.

In the presence of a pre-existing bulk flow, an ionic wind can modulate the boundary layer, and, when aligned with the flow, it will accelerate the flow, thinning the boundary layer and enhancing convection. This area has received significant attention for drag reduction in aeronautics, often using dielectric barrier discharge plasma actuators [9]. Figure 3 shows infrared thermographic images of a heated glass plate when cooled by only natural convection, a low-speed pre-existing bulk flow (0.3 m/s), and a corona discharge ionic wind enhancing the bulk flow. For a constant heat flux of 4 W, when the bulk flow was superimposed with the ionic wind, an additional ~20 K of convective cooling was obtained, while the ionic wind consumed less than 100 mW of electrical power [10, 11]. Alternatively, the ionic wind can be directed perpendicular or opposed to the bulk flow such that the flow is distorted, to increase heat transfer [12]. In both of these scenarios, however, the use of an ionic wind is limited to laminar, low Reynolds number flows.  Because the ion-neutral interaction acts as a Coulomb body force, if the inherent inertia in the pre-existing flow is too great, the ionic wind is ineffective. An analogous situation occurs in mixed convection, when the Grashof number is small compared to the square of the Reynolds number such that the buoyant body force becomes negligible.

Figure 3. Infrared thermographic images of an ionic wind enhancing the cooling from a bulk flow (10, 11). The ionic wind enhanced cooling by nearly 20 K while consuming less than 100 mW of power. A schematic of the experimental set-up is shown in the inset. For these experiments, 4 W of heat was applied to the plate to steady state, and then the plate was cooled by a 0.3 m/s flow to steady state by approximately 15 K. An 15 μA (67 mW) ionic wind was then superimposed on the bulk flow to reduce the plate temperature an additional 20 K. The average heat transfer was increased by a factor of 2 when the ionic wind was applied.



















The Future and Ongoing Challenges

Overall, ionic winds are a promising technology for a wide variety of important, but niche applications.  Because of their poor electrical-to-fluid energy conversion, an efficiency that ranges from 1-2% [13], they will be hard pressed to compete with mechanical fans for large scale cooling applications such as servers. But because they are inherently silent, consume little power, and offer the potential to be scaled to smaller dimensions, they may find application in portable devices, ranging from laptops to smart phones or consumer LED products. However, there continue to be challenges. Future research must focus not only on improving flow generation and heat transfer, but also reducing the operating voltage, the degradation of electrodes over time, and minimizing ozone production, which are very real practical hurdles.  Still, because of their potential advantages the future is bright that ionic winds will soon find a home in the cooling technologies market.


[1] Robinson, M., 1962, “A History of the Ionic Wind,” American Journal of Physics, Vol. 30, pp. 366-372.

[2] Stuetzer, O., 1959, “Ion Drag Pressure Generation,” Journal of Applied Physics, Vol. 30, pp. 984-994.

[3] Robinson, M., 1961, “Movement of Air in the Electric Wind of the Corona Discharge,” Transactions of the American Institute of Electrical Engineers, Vol. 80, pp. 143-150.

[4] Jewell-Larsen, N.E., Ran, H., Zhang, Y., Schwiebert, M., Honer, K.A., Mamishev, A.V., 2009,  “Electrohydrodynamic (EHD) Cooled Laptop,” 25th IEEE Semiconductor Thermal Measurement and Management Symposium (Semi-Therm), pp. 261-266.

[5] Schlitz, D., Singhal, V., 2008, “An Electro-aerodynamic Solid State Fan and Cooling System,” 24th IEEE Semiconductor Thermal Measurement and Management Symposium (Semi-Therm), pp. 46-49.

[6] Owsenek, B. L., Seyed-Yagoobi, J., Page, R. H., 1995, “Experimental Investigation of Corona Wind Heat Transfer Enhancement with a Heated Horizontal Flat Plate,” ASME Journal of Heat Transfer, Vol. 117, pp. 309–315.

[7] Rickard, M., Dunn-Rankin, D., Weinberg, F., Carleton, F., 2006, “Maximising Ion-Driven Gas Flows,” Journal of Electrostatics, Vol. 64, pp. 368-376.

[8] Tirumala, R., Go, D.B., 2011, “The Multi-electrode Assisted Corona Discharge for Electrohydrodynamic Flow Generation in Narrow Channels,” IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 18, Issue 6, pp. 1854-1863.

[9] Corke, T.C., Enloe, C.L., Wilkinson, S.P., 2010, “Dielectric Barrier Discharge Plasma Actuators for Flow Control,” Annual Review of Fluid Mechanics, Vol., 42, 505-529.

[10] Go, D.B., Garimella, S.V., Fisher, T.S., Mongia, R.K., 2007, “Ionic Winds for Locally Enhanced Cooling,” Journal of Applied Physics, Vol. 102, art. no. 053302.

[11] Go, D.B., Maturana, R.A., Fisher, T.S., Garimella, S.V., 2008, “Enhancement of External Forced Convection by Ionic Wind,” International Journal of Heat and Mass Transfer, Vol. 51, pp. 6047-6053.

[12] Molki, M., Bhamidipati, K., 2004, “Enhancement of Convective Heat Transfer in the Developing Region of Circular Tubes Using Corona Wind,” International Journal of Heat and Mass Transfer, Vol. 47, pp. 4301-4314.

[13] Moreau, E., Touchard, G., 2008, “Enhancing the Mechanical Efficiency of Electric Wind in Corona Discharges,” Journal of Electrostatics, Vol. 66, pp. 39-44.