MEMS-Based Solid-State Airflow for Compact Thermal Management

As consumer electronics continue their rapid evolution, each new generation is defined by a familiar yet intensifying mandate: deliver more performance in less space. Today’s devices — whether smartphones, wearables, or portable computing platforms — are expected to be thinner, lighter, and more power-efficient, while simultaneously supporting increasingly demanding workloads such as high-speed connectivity, immersive media, and on-device artificial intelligence. This convergence of portability and performance has elevated thermal management from a secondary design consideration to a critical engineering challenge.

At the heart of the issue lies a fundamental constraint: heat. Compact, battery-powered devices must dissipate growing amounts of power within ever-shrinking form factors, all while maintaining safe and comfortable external temperatures for users. Industry standards such as IEC 60601-1 and 62368-1 permit surface temperatures up to 45°C for prolonged contact, yet in practice, manufacturers often target lower thresholds — typically below 42°C — to ensure user comfort and accommodate sensitivity in some populations. As a result, thermal limits at the device surface, rather than semiconductor junction temperatures, are increasingly dictating system performance, often triggering processor throttling before peak computational capability can be reached.

Traditional passive cooling approaches, including vapor chambers and graphite heat spreaders, have been widely adopted to address these constraints. By distributing heat over a larger surface area, these technologies reduce localized hot spots and moderate component temperatures. However, as power densities rise and device footprints continue to shrink, the effectiveness of heat spreading alone is reaching its practical limits. The challenge is particularly acute in emerging form factors such as smart glasses, where minimal volume and strict ergonomic requirements severely restrict thermal design options.

Even in larger personal devices such as smartphones, where more advanced thermal solutions are feasible, increasing computational loads — driven in part by on-device AI processing — are pushing existing cooling strategies to their limits. Similarly, solid-state drives and other high-performance components are frequently deployed in confined environments with limited airflow, relying on chassis conduction and passive spreading techniques that struggle to keep pace with escalating thermal demands.

These converging trends underscore a pivotal reality: as electronic devices become more compact and capable, innovative thermal management solutions will be essential to unlocking their full performance potential.

MEMS Cooling

The continued drive toward thinner, more powerful electronic systems has exposed a fundamental limitation in conventional thermal management: the lack of effective active cooling at miniature scales. Addressing this gap is a newly developed MEMS cooling device known as μCooling — a solid-state, MEMS-based airflow technology capable of generating net airflow with significant static pressure in an ultra-compact form factor.

Fabricated using an all-silicon, wafer-level manufacturing process, the MEMS cooling device enables high reliability, strong process traceability, and true scalability to mass production. Unlike traditional fans, the devices operate without macroscopic moving parts, resulting in virtually vibration-free and ultra-quiet performance — critical attributes for next-generation portable and wearable electronics.

Operation and Specifications

At the core of the device is a MEMS architecture built on silicon wafers. Thin silicon membranes are engineered into cantilevered vanes, which act as the fundamental actuation elements. A thin-film layer of lead zirconate titanate (PZT) is deposited onto these vanes to enable piezoelectric motion, as illustrated in Figure 1.

Each wafer contains arrays of these micro-scale cells, allowing designers to configure device size, shape, and airflow characteristics to match specific application requirements. This inherent modularity supports a wide range of geometries and integration strategies. For example, a 2×4 array of microcells takes about 7.25mm x 6.5mm on the wafer.

The vanes are actuated with ultrasonic drive signals, producing rapid oscillatory motion. Integrated venting within the device package enables pressure differentials, allowing the system to generate directional airflow. Although the airflow is produced in short pulses, the high actuation frequency results in a quasi-continuous flow profile.

A key advantage of the thin-film PZT approach is reduced power consumption and lower drive voltage. While traditional piezoelectric actuators often require 50–100 Vpp, these devices achieve full actuation at voltages below 20 Vpp, making them well-suited for low-power electronics.

Control of the MEMS device is handled by a custom ASIC named ASTRA packaged in a compact 2.5mm square wafer-level chip-scale package (WLCSP) format. The ASIC provides voltage regulation, ultrasonic signal generation, and bidirectional airflow control. Operating from a 3V input and controlled via I²C, it also incorporates energy management functions for the MEMS structure. The basic operation is depicted in Figure 2, and current implementations support arrays up to 4×8 cells. The ASIC is paired with the MEMS device and can be mounted remotely.

Packaging and Integration

These devices are currently available in multiple configurations, including 1×2 (XMC-1200), 2×4 (XMC-2400), and 4×8 (XMC- 4800) arrays. The MEMS die is packaged on a laminate PCB substrate and enclosed with a nickel-plated copper lid. Airflow is facilitated through inlet holes in the PCB and precision outlet slots in the lid structure.

Both top-vented and side-vented configurations are available for the larger arrays, enabling flexible system integration. In forward operation, airflow enters through PCB inlets and exits through lid vents. Representative packaging options and specifications are shown in Figure 3.

Figure 3: Packaging and performance

Importantly, the platform is not restricted to these standard form factors. The MEMS architecture supports further miniaturization and can be integrated into system-on-chip (SoC) modules, opening opportunities for tightly embedded thermal solutions. The XMC-2400 series has a weight of 0.15g. The XMC-4800 series is on the order of 0.25g.

Performance Characteristics

Performance of the μCooling devices is best understood through pressure–flow (PQ) characteristics. Typical PQ curves for the XMC-2400 and XMC-4800 are presented in Figure 4. While airflow direction can be reversed dynamically, reverse operation introduces minor performance trade-offs.

In forward mode, the XMC-2400 delivers approximately 25 cc/s (0.06 CFM) with a maximum static pressure exceeding 1300 Pa. The larger XMC-4800 achieves 48 cc/s (0.1 CFM) with static pressure above 1100 Pa. These high-pressure characteristics distinguish μCooling from conventional fans, which typically prioritize high volumetric flow (CFM) but operate at low static pressure. For comparison, a 9mm x 9mm x 3mm axial fan (Sunon UB393-700B) shows performance of 19cc/s (0.041 CFM) with a maximum static pressure of 16.17 Pa.

In compact systems, where airflow pathways are highly restricted, system impedance increases significantly. Under these conditions, traditional fans lose efficiency, whereas μCooling’s high static pressure enables more effective heat removal. As a result, airflow velocity, pressure, and flow rate can be optimized for small-scale thermal environments.

Power dissipation for the XMC-2400 is on the order of 200mW for full flow operation including the ASIC. As flow is reduced, power dissipation also reduces. xMEMS is in process of full power characterization across the product line and will publish results later this year.

Application Scope

Due to their miniature size and silent operation, these devices are ideally suited for portable and wearable electronics. Target applications include smartphones, smart glasses, augmented and virtual reality headsets, wearable health devices, AI-enabled personal assistants, and audio systems.

Beyond consumer devices, the MEMS cooler is also applicable to embedded and component-level cooling challenges. High-density components such as solid-state drives (SSDs), DDR memory, and other thermally constrained subsystems can benefit from localized airflow in environments where conventional cooling is impractical.

The absence of vibration and acoustic noise makes MEMS cooling particularly valuable in noise-sensitive systems, enabling active cooling without compromising user experience or interfering with sensitive components.

Power Overhead (POH)

A simple relationship is used to estimate additional power capability when a cooling solution is applied. The method uses thermal resistance to calculate how much power can be added to the existing components. Equation 1. Power Overhead shows the method.

Where: POH = power overhead in watts

ΔT = change in temperature without cooling solution

θ_JX = thermal resistance with cooling solution

Pd_SOURCE = power dissipation used in original thermal resistance calculation

For example, in an experiment using smart glasses, the system has a thermal resistance of 69.9ºC/W without cooling. This is calculated by ΔT of 52.4ºC at a power dissipation of 0.75W. After the application of the cooling solution, the ΔT dropped to 33.3ºC and thermal resistance is reduced to 44.4˚C/W. Using the POH equation:

The result shows that a power overhead of 36% is realized with the cooling solution.

Case Studies Overview

This article examines three representative applications of the MEMS cooling technology:

1. Smart Glasses
2. Cell Phones
3. Solid-State Drives (SSDs)

Case Study: Smart Glasses

Thermal management in smart glasses presents a uniquely constrained design problem. Prior studies by Matsuhashi et al. [1] and Jung et al. [2] have explored passive strategies, focusing on materials and layout optimization. However, strict industrial design requirements — balancing aesthetics, ergonomics, and comfort — severely limit the available volume for thermal solutions.

Additionally, any cooling approach must operate silently and without vibration to avoid interference with integrated audio systems.

Attempts to miniaturize conventional fans have shown diminishing returns [3], [4]. For example, an 8mm × 8mm × 2mm axial fan produces only 0.006 CFM and 1.6 Pa of static pressure — insufficient for meaningful thermal management in confined spaces.

To evaluate the MEMS cooler, a representative smart glasses model was developed, including bridge and dual temple arms. A simulated heat source was integrated into one temple, with the MEMS cooler for active cooling, while the opposite temple served as a control. The model and experimental setup is shown in Figure 5 and Figure 6 respectively.

The heat source consists of a compact PCB assembly with an attached copper slug and resistor bank. An XMC-2400 unit is integrated within the temple arm, positioned near outlet vents.

Computational fluid dynamics (CFD) simulations were conducted under steady-state conditions (Ta = 23°C, PD = 0.864 W), with airflow parameters adjusted to match experimental hardware. Material properties were based on representative values for the system components.

Experimental validation was performed using thermistors and thermocouples to measure internal and external temperatures. As summarized in Table 1, simulation results show strong agreement with measured data for both active and inactive cooling states.

Table 1: Model/Measure Comparison

At a higher power dissipation of 1.5 W, results demonstrate significant thermal improvement. Internal temperatures were reduced by approximately 16°C, while external surface temperatures decreased by an average of 9°C. Temperature values are illustrated in Figure 7.

The use of MEMS cooling in typical smart glass designs produces lower internal and external temperatures. Up to 25% reduction in component temperatures and 17% reduction in external surface temperatures is achieved with a simple application. Power overhead is increased by up to 64%. Further refinements in vents, ducting, and placement can improve performance. The small size of the μCooling component allows placement in extremely small volumes and shows promise for longer operation time without discomfort to the user while allowing increased power handling capability.

Case Study: Compact Cell Phone

Cell phones are another technology segment where MEMS cooling can increase power handling and reduce or maintain temperatures for safe operation during human body contact. The small size is a feature that is lacking in forced air movers today and can be integrated into phone designs. Engineering solutions to cooling problems in cell phones have been addressed using spreaders and vapor chambers, but as noted earlier, these methods are reaching limits of performance.

Introducing active cooling into a cell phone design requires changes to the philosophy of a nearly hermetic system to allow inlet and outlet vents. Moving air around the inside of a sealed system is possible but does not cool the system significantly. The effect of internal-only air movement is a slightly reduced ΔT across the system.

To fully realize the benefit of forced convection vents are needed for cool air entry and hot air escape. This is an antithesis to conventional phone design because of the need for waterproof or near waterproof assemblies. Current cellphone design typically aims for IP68 ratings, where the unit can survive underwater for 30 minutes at a depth of 1 meter. Although this rating is considered water resistant, no existing phone design is completely waterproof, and most manufacturers will not warrantee any kind of water damage.

To address water ingress when applying active air cooling several approaches are considered. IP68 ingress protection can be achieved using IP6x rated mesh. However, this method needs careful study to allow enough airflow while reducing the risk. Water ingress protection materials, such as ePTFE, are popular and in use for acoustic vents for speakers but are not well suited for continuous flow. Other IP materials such as hydrophobic mesh are also used to slow water ingress.

Another way to mitigate damage to internal components in a phone is to not allow the air-cooling loop to penetrate the main internal assembly. This can be achieved by using concepts shown in Figure 8. In this configuration air is directed over the primary heating component using an isolated flow channel created with a small bead of adhesive around the desired path. As can be seen in the image, the flow channel is placed over the vapor chamber. Inlet and outlet vents are 5mm x 1.5mm slots, are confined to the flow channel, and do not have ingress into the primary volume of the phone, effectively maintaining a cooling path and isolating internal components.

Simulations of this scenario with a 4×8 array of cells (XMC-4800) show promise for reducing temperatures, allowing higher power dissipation or lower temperatures using the existing power dissipation.

Simulation results for this configuration are shown in Table 2 for off and on states of the MEMS cooler. Temperatures are shown for the internal Tmax, display and back cover. Results show >15% reduction of case and internal component temperature and an average of 40% power overhead. This is only one potential solution. With a bit of creativity, this approach can be designed into a wide range of compact products.

Case Study: SSD

Solid-state drives are the core support for most, if not all, computer technology. The space savings and access speed compared to a mechanical disk drive are clear. With the decreased size and continued capacity improvements, power density is increasing rapidly. The critical components in an SSD are the controller chip and the memory chips themselves. These components are typically mounted on a small form-factor PCB in a few standard sizes. They are thin and are found in laptops and other computer types. In desktop systems, large enclosure volume allows air cooling by using heat sinks mounted directly on the SSD modules. In more compact enclosures, the SSD may be placed where convenient for the design while considering thermal management using the enclosure case or localized spreaders. As performance and capacity increase, cooling must also increase, but in many cases, there is little room for additional cooling and no way to introduce forced convection cooling.

The concept in Figure 9 shows a typical SSD and a modified layout with μCooling installed. A 4×8 array of cells (XMC-4800) is installed with a 3D-printed manifold managing airflow.

A duct is added to direct airflow around the components. Some designs may omit the duct and direct airflow towards the component of interest. The duct increases the cooling by creating a flow channel focusing the μCooling airflow. Figure 10 shows the duct added to the assembly. The duct can be fashioned and attached using common EMI shield manufacturing techniques.

Figure 11 shows the cross-section profile and assembly detail. The thin assembly allows installation without major modifications to the original location. In this application, TIM is placed on top of the critical chips to allow conduction into existing thermal management structures. Air only flows around the periphery of the chips.

Results for this configuration are shown in Table 3. The result shows a 15% decrease in Tmax and 27% power overhead. This is a substantial savings in power and moves the maximum temperature within operating temperature to eliminate throttling.

Table 3: M.2 SSD Simulation Results

Toward Active Cooling in Compact Systems

Modern computing systems increasingly rely on active cooling solutions, including fans, heat exchangers, and liquid cooling. However, these approaches are difficult to scale down to compact mobile platforms such as smartphones and ultra-thin devices.

As processing demands grow — driven by applications such as gaming, video processing, and on-device AI — thermal constraints have led to widespread reliance on performance throttling. While passive techniques such as vapor chambers and heat pipes provide incremental improvements, they are often insufficient for sustained high-performance operation.

MEMS cooling offers a fundamentally new approach: a silicon-based, solid-state cooling solution designed specifically for the smallest form factors. By enabling high-pressure airflow in millimeter-scale packages, it addresses one of the most critical limitations in modern electronics design.

References

[1] 1. Grimes, R, Quin, D, Walsh, E, Davies, M, & Kunz, S. “A Theoretical and Experimental Investigation of the Scaling of Micro Fan Performance.” Proceedings of the ASME 2003 International Mechanical Engineering Congress and Exposition. Fluids Engineering. Washington, DC, USA. November 15-21, 2003. pp. 553-560. ASME.

[2] Soumyashree A Hiremath, M Sivapragasam, S Umesh and M D Deshpande, “Scaling the performance of a miniature axial flow fan.”, ASME International Conference on Nanochannels, 2003

[3] Velmurugan K, Abhay Lingayat, Chandramohan VP, Karthik Balasubramanian, Karunanidhi S, “A critical assessment on micro-blowers and pumps for different engineering applications, Sensors and Actuators A: Physical, Volume 365, 1 January 2024.

[4] Kays, William Morrow, and Alexander Louis London. “Compact heat exchangers.” (1983).