Background introduction
With the advent of 5G communication and cloud computing, the continuously growing demand for high power presents unprecedented challenges in thermal management, creating bottlenecks in producing effective electronic devices. Research on non-oxide ceramic materials such as aluminum nitride (AlN) or silicon carbide (SiC), which are both electrical insulators but thermally conductive, has garnered significant interest in the electronics industry. This interest stems from the potential applications of these materials in efficient cooling devices, allowing for safe integration with electrically active components. Additionally, the compatibility between AlN and silicon due to their similar thermal expansion coefficients may eliminate long-term thermal mechanical fatigue, prompting researchers to incorporate AlN into microelectronic packaging or heat sinks/expanders for direct chip cooling. Other non-oxide ceramics such as SiC, BN, etc., also possess similar favorable properties.
Current manufacturing technologies enable the cost-effective mass production of high thermal conductivity AlN (≈180 W/mK, compared to commercially available SiC ≈125 W/mK or BN ≈45 W/mK), making it the preferred choice in recent research. The attractive characteristics of these ceramics have been utilized in the development of integrated circuit (IC) packaging or heat sinks using silicon-glass technology to reduce electrical-thermal effects. The application of AlN in ceramic matrices or polystyrene composite materials (as fillers or interface materials in electronic packaging) is also considered to enhance heat transfer by forming bridges in the matrix or through phonon transfer via thermal conductivity networks.
A snapshot of results
Recently, the team led by M. Megaridis at the University of Illinois at Chicago has made new progress in the research on aluminum nitride (AlN) ceramics. Aluminum nitride is a versatile ceramic with high thermal conductivity, high electrical resistivity, and a thermal expansion coefficient compatible with silicon, making it highly suitable for direct chip cooling applications in electronic products, implementation in wide-bandgap semiconductors, and for use in high-temperature heat exchangers.
Despite its numerous advantages, the application of AlN in liquid cooling applications is often hindered by surface degradation effects induced by the hydrolysis of the working fluid. This paper introduces a scalable yet highly tunable wetting engineering approach that allows for the effective realization of large AlN substrates in enhanced two-phase cooling of electronic products. This approach prevents AlN from being hydrolyzed by the working fluid, establishes control over surface roughness, while maintaining the overall integrity and material properties of the substrate.
This new method is demonstrated in spontaneous, pump-less, surface liquid transport, which is necessary if this ceramic is to play an indispensable role as a sealed, phase-change, coreless thermal management device component (e.g., vapor chambers or heat pipes), as these devices require rapid transport of working fluids within their multiphase interiors.
The novelty of this work lies in establishing a scalable approach to utilize and further enhance the performance of this non-oxide ceramic material for phase-change heat transfer sealing devices, thereby paving the way for the implementation of this intriguing material in the next generation of heat exchangers. The research findings are published under the title “Laser-Tuned Surface Wettability Modification and Incorporation of Aluminum Nitride (AlN) Ceramics in Thermal Management Devices” in “Advanced Functional Materials”.
Graphical guide
Figure 1.
A) Cleaning of commercially available bulk AlN substrates and imaging under scanning electron microscopy (inset) (Supporting Information Section S1: Material and Sample Cleaning).
(i) XRD (Cu K𝛼1) and (ii, iii) EDS spectra confirm the relative purity of the samples, with some oxygen and carbon present on the surface as contaminants.
(iv) Contact angle (CA) of non-attenuated samples with a pore-free 4.87 μL water droplet is 40°±0.7°.
B) Surface deep modification - shown in cross-sectional SEM micrographs, depicting laser ablation on AlN surfaces - through laser ablation (i), releasing nitrogen surfaces and forming aluminum oxide (ii, iii). The resulting surface is superhydrophilic, conducive to capillary core withdrawal (iv). EDS maps in A(ii) and B(ii) quantify the relative presence of each element (Al, N, O, C) on each surface in atomic percentage.
C) Submicron-thick layer of polytetrafluoroethylene (fluorinated amorphous polymer) confirmed through spin coating (as indicated by EDS spectra in (i)) to render the surface hydrophobic (ii), with a water droplet CA = 130°±0.5°, while allowing the surface to maintain its potential roughness (iii).
D) Wetting pattern (WP) of passivated body substrates demonstrates hope for achieving thermal management in condensation heat transfer and vapor chambers.
Figure 2.
A) AlN samples treated with different laser settings (samples i-viii, scale bar on the top row represents 100 μm). Different laser parameters result in different roughness characteristics.
B) To evaluate the liquid absorption capability of each surface, a 4.87 μL water droplet was once dripped onto 2 × 2 cm samples (second row) and visualized using a high-speed camera. The final wetting state of each surface is shown beneath their respective surface micrographs, with yellow arrows indicating the maximum extent of droplet spreading.
C) Polyethylene tetrafluoroethylene (PTFE) coatings render the surface hydrophobic: comparison of water droplets gently dropped on the untreated surface of AlN (CA≈40°±0.7°) with those on the untreated surface of PTFE-coated AlN (CA≈130±0.5°) of the same volume. Samples laser ablated were spin-coated with PTFE and cured on a hot plate, rendering them hydrophobic. (i - viii) Water droplets gently dropped on each surface are observed with an optical goniometer (scale 1 mm) showing contact angles. Sample ii excels in both droplet spreading (maximum when untreated) and repellency (highest CA when coated with PTFE) among all other samples.
D) (i) XPS studies of the chemical composition of each surface to understand the differences between untreated AlN, laser-ablated AlN (inset: high-resolution scan confirming the absence of nitrogen), and laser-ablated AlN with teflon coating. (ii) Grazing incidence XRD spectra showing the formation of aluminum oxide layer during the laser ablation process.
Figure 3. Laser ablation characterization of superhydrophilicity on Teflon-coated aluminum nitride samples for spontaneous pump-free fluid transport.
Figure 4.
A) Schematic representation of wetting patterned (WP) AlN substrates vertically mounted in an environmental chamber maintaining a constant relative humidity RH = 80% at different ambient temperatures, while samples are kept at 10°C via an additional cold plate.
B) Interlaced pattern composed of wavy hydrophobic (white) and superhydrophilic (black) domains enhances condensation heat transfer performance. Geometric features of the pattern are indicated in the schematic. Thermal resistances (RAlN, Rdrop, Rvapor, Rfilm) during the condensation process in FwC or DwC modes are marked in separate schematic on the right.
C) Experimental observations confirm droplet condensation on domains of polytetrafluoroethylene-coated AlN, while filmwise condensation occurs on laser-ablated (hydrophilic) portions.
D) Assessment of heat transfer performance of samples based on condensation mass collected per unit area and hour from each sample.
Figure 5.
A) Schematic side view of a hybrid coreless vapor chamber (wfvc), consisting of an AlN evaporator and a copper condenser separated by a fixed gap.
B) (Left) WFVC featuring surfaces with wetting patterns capable of enhancing condensate retention and condensate recirculation to the evaporator. (Right) Water (filling liquid) replenishment coupling mechanism from condenser mode to evaporator mode.
C) Thermal resistances of patternless WFVC (hydrophilic evaporator and hydrophobic condenser) and
D) Thermal resistances of WFVC with both evaporator and condenser patterns computed.