Below are the questions asked during the event, along with their respective answers.


Q: What range of pore sizes are available?
A: It depends on the base material. The most common pore sizes for aluminum and copper are 5, 10, 20, and 40 Pores Per Inch (PPI). However, compressing the foam is a great way to increase the effective PPI (smaller pore size, but more “pores” in a given volume), while also increasing surface area and enhancing bulk thermal conductivity. Carbon and ceramic foams can go up to 100 PPI and can be compressed up to 10x, or ~1000 PPI. It’s important to note that the compressed PPI’s don’t create new “pores” but does effectively decrease the apparent pore size.

Q: How do you characterize the hydraulic diameter and surface area of the foam?
A: Historically pore size and ligament size were measured optically by discrete measurements under a microscope, and surface area has been measured using the BET method (which measures physical adsorption of gas molecules on a solid surface). However, advancements in Computerized Tomography (CT) technology have made it practical to scan (i.e. take x-ray images at incremental slices through the material) samples. Post-processing software can then create a 3D surface that can be interrogated with automated routines to extract rigorous and comprehensive statistics regarding pore size, ligament size, surface area, and more. Whether you are doing fundamental testing and correlating data to geometry, or you are doing quality control, CT scanning can be a very powerful tool.

Q: What are the mechanical characteristics of the foam. Is it strong enough to hold moderately high internal pressure?
A: Metal foam has many applications beyond heat transfer, including numerous structural applications, such as sandwich panels and impact absorbers. As a result, metal foam has been characterized in both compression and tension fairly well, and the foam has been proven in pressurized environments. The strength of the foam is a strong function of relative density (i.e. metal volume fraction). The foam can also be brazed into the housing and provides a strong thermal and structural connection to help handle high internal pressures.

Q: What is the process variability in the foam manufacturing process?
A: It depends on the application and the property of interest, both because of the quantity of data available and in terms of the actual variance. There is a lot more data on mechanical properties so it is easier to do statistical analysis on properties such as yield stress. A typical number for 95% confidence interval for yield stress is +/- 20%. Pressure drop variability is on the order of +/- 10%, although pressure drop and heat transfer data are less available and it can be hard to compare across different research institutions due to differences in experimental setup. At ERG we continue to build our own data sets, and we have yet to have any surprises (within about 5%) between predicted and as-built performance.

Q: If brazing or soldering to metal base plates, does capillary wicking of the molten metal plug the pores more than a few cells deep?
A: When using appropriate braze foil thicknesses and braze procedure, there is no risk of plugging any pores, and actually the surface tension effects help collect braze alloy around the base of the ligaments, improving the thermal and mechanical performance of the joint interface.

Q: What is the optimal density for 2-phase heat transfer?
A: Further research needs to be conducted on 2-phase cooling with foams in order to better answer that question. And even then the answer would likely depend on the specific cooling requirements, geometry, flow conditions, system-level requirements, etc. However a good starting place would be a 40 PPI foam with 8-10% relative density compressed by a factor of 2 (i.e. 16-20% final relative density).

Q: Current limitation of the geometry, like pore size and porosity?
A: It depends on the base material. For carbon and ceramic foams, 100 Pores Per Inch (PPI) are readily achievable. The most common pore sizes for aluminum and copper are 5, 10, 20, and 40 PPI, and porosity (i.e. 1-relative density) for uncompressed metal foam ranges from 0.85-0.97 (3% – 15% relative density). However compressing the foam is a great way to decrease the effective pore size, while decreasing porosity and increasing surface area. Compressing to final porosity of ~0.5 (relative density of 50%) is the practical limit for compressed foam (beyond that you start to close off pores and adversely affect the permeability of the foam). This can be achieved with an 8% relative density foam compressed by a factor of 6x.

Q: What is your insight on replacing fins with thin metal foam in natural convection application?
A: While most efforts using metal foam cooling have focused on high heat flux forced convection, at ERG we are currently developing various natural convection heat sinks for high-powered LED applications. Preliminary results are very promising and we expect to demonstrate significant reduction (~66% reduction in form factor) in overall heat sink volume.

Q: How does a foam heat exchanger compare to solid Cu heat exchangers in electronics?
A: In single-phase applications you can expect a thermal resistance that is about 2x lower when comparing foam heat exchangers to state-of-the-art finned heat exchangers. Lower thermal resistance is good and helps transfer more heat out of the system. There is insufficient data about the relative improvement in 2-phase cooling applications, though we hope that will change soon.

Q: Can metal foams be used in heat transfer?
A: Absolutely! Metal foam can be used wherever fins or microchannels are used to increase surface area and effective heat transfer.

Q: Any control on the shapes of the pores after compression?
A: There is limited control for pore shape during compression, although the typical resulting shape is well understood. Typically uniaxial compression is most practical, though a biaxial compression may be applied in a few incremental steps as a way to a control pore shape. However, note that many of a foam’s advantages in heat transfer (such as enhanced mixing and high film coefficient) benefit from the randomness and tortuosity of the compressed foam structure.

Q: What are the challenges of having composite metal foams with high thermal conductivity nanoparticles embedded for better heat transfer? Are these being investigated currently in industry research?
A: We are not aware of research in the area of embedding high thermal conductivity nanoparticles in the foam ligaments themselves. In general, it is hard to beat pure copper in terms of thermal conductivity. However, it is possible to have particles infused interstitially between pores (e.g. as nano-enhanced PCM), or as a surface coatings that create a nano-textured surface for enhanced heat transfer or for enhanced catalytic reaction rates. ERG has a long history of supporting research projects and is always interested in efforts to help push the current technology boundaries.

Q: Does dust accumulate inside foam after prolonged usage reducing air flow etc… and how do we clean that for bigger size foam structures?
A: Dust buildup and fouling is a very important but poorly understood phenomenon, not only for metal foams but for most compact heat exchanger cores in general. However, a 2016 paper by Hooman and Malayeri – “Metal foams as gas coolers for exhaust gas recirculation systems subjected to particulate fouling” ( shows very similar fouling rates for metal foam Exhaust Gas Recirculation (EGR) coolers vs. conventional corrugated ones. Furthermore, they found that fouling occurs predominantly on ligaments near the surface and not in the interior, and they demonstrate that simple brush cleaning is remarkably effective. The authors note that “compared to louvered or finned counterparts, metal foam samples investigated here are much easier to clean and reuse, which translates into significant savings and easier maintenance in the long run.”

Q: What is the range of thermal conductivity for an aluminum metal foam?
A: Uncompressed foam can achieve ~10W/m-K for aluminum foam and ~18 W/m-K for copper foam (per equation on slide 8, assuming 14% relative density for uncompressed foam). However compressing the foam is a great way to increase conduction area and consequently bulk thermal conductivity in a proportional way, and thermal conductivities of 35 W/m-K and 60 W/m-K can readily be achieved with compressed aluminum and copper foam respectively.

Q: Is wearing of PCM capsule possible?
A: In ERG’s experience with metal foam PCM heat exchangers we have not experienced any issues with wear due to thermal cycling.

Q: What are the radiation properties of metal foam?
A: The local radiation properties (specifically emissivity) are governed by the material and surface finish – an aluminum foam ligament without any coatings would have a comparable emissivity as etched or polished aluminum bar stock on a microscopic level. However, bulk emissivity would be effected by repeated reflections and scattering due to the complex surface of foam ligaments and pores, and that would lead to further attenuation. This would also depend on the length scales and overall thickness of the foam to some extent. Duocel® has successfully been used to block out light for gas chromatographs.

If the question is geared towards EMI shielding, note that the electrically conductive solid metal (e.g. aluminum or copper) ligaments and relatively small pore size (0.6-2.5mm) result in an effective Faraday cage. Shielding effectiveness above 80 dB has been demonstrated in the open literature for radio frequencies below 100 MHz, using a 0.5” layer of foam. And while there is limited research on the shielding effectiveness against ionizing radiation, low energy X-rays (e.g. 5 keV) have a very limited penetration depth and can be blocked as long as there is no line of sight to the sensitive electronic equipment. In addition, because Duocel® is open celled and permeable, it can be infused with secondary materials that can enhance radiation shielding while also serving a dual purpose in heat dissipation.

Q: How do you join the metal foam with solid metal?
A: There are numerous ways of joining foam to solid metal. For heat transfer applications the preferred methods are brazing (vacuum brazing or dip brazing) or soldering with solder paste (copper foam can be soldered readily, and aluminum foam can be nickel-plated to facilitate soldering). Lower-cost alternatives include various thermal adhesives, though they could increase thermal resistance and may not be as durable.

Q: Can these be used in a heat pipe?
A: The application of metal foam as a wick inside a heat pipe has not been investigated, though it is a very interesting proposition. Especially in applications with lager planar footprints, a Duocel® foam vapor chamber could provide significant manufacturability advantages over conventional vapor chambers or thermal spreaders with embedded heat pipes. The foam could also serve a multi-functional role in the construction – e.g. as a wick, structural core and EMI shield.

Q: Can the foam structure be formed within confined spaces?
A: Metal foam provides a lot of flexibility when it comes to fabrication into conformal shapes. It can be machined to net shape or near-net shape and assembled into a housing with just about any form factor with clearances or interference as necessary. Because of the solid ligament structure, Duocel® foam can also be compressed and formed inside a die without fracturing ligaments. Compressing by a factor of 8x is possible, and thin layers of foam can be formed around small radii of curvature (e.g. a 0.1” thick layer of foam can easily be wrapped around 0.25” diameter cylinder).

Q: What is the cost ratio between aluminum foam to aluminum plate?
A: The cost trade-off is usually a bit more complex because it involves a more system-level understanding. Generally, metal foam will have higher performance per volume so you can reduce the overall heat sink size and simplify manufacturing, reducing the overall product cost. Metal foams are often used when conventional heat sinks either don’t perform well enough, or they are too large for the system and the cost-benefit analysis trends towards a shift in technology. However, to give a general answer to your question on a per unit volume basis ($/cubic inch) the cost ratio is on the order of 1:10.