Skip to content

High Thermal Conductivity Diamond–Copper Composites Prepared via Hot Pressing with Tungsten–Coated Interfacial Layer Optimization

FieldValue
JournalMaterials
Published2025-08-19
AuthorsQiang Wang, Zhijie Ye, Lei Liu, Jie Bai, Yuning Zhao
DOI10.3390/ma18163882
Citations1

Hot-pressed diamond-copper composites reached 640 W/(m·K) at 50 vol.% diamond. The optimum used 400 µm particles, a 200 nm tungsten coating, and 2 h of annealing at 1100 °C. Its coating contained 2 wt.% W, 92 wt.% WC, and 6 wt.% W2C. The authors report the highest value among hot-pressed composites with no more than 50 vol.% diamond. Thermal diffusivity was measured at room temperature on 12.7 mm discs cut from 1.5 mm thick composite.

Tungsten coating and composite fabrication

Section titled “Tungsten coating and composite fabrication”

Each preparation began with 400 µm single-crystal diamond particles cleaned sequentially with nitric acid, sodium hydroxide, acetone, and anhydrous ethanol. Ultrasonic cleaning removed oils and impurities before drying. Pulsed magnetron sputtering deposited an average 200 nm tungsten coating in argon. A tilted rotating plate improved coverage of the particle surfaces.

Particles were annealed at 1100 °C under 5 × 10−4 Pa vacuum with a 10 °C/min heating rate. After activation and sensitization, chemical deposition in a rotary evaporator added copper to the metallized surfaces. Hot pressing produced a 30 mm diameter, 1.5 mm thick composite. The authors cut 12.7 mm discs and targeted a nominal 50 vol.% diamond fraction, ignoring coating volume.

Phase identification used XRD from 20 to 90 degrees 2θ. SEM examined (100) and (111) surfaces. Raman spectroscopy used a 532 nm source over 100 to 3200 cm−1, and FTIR measured nitrogen absorption. TEM examined extracted interfaces. Thermal diffusivity came from the LFA467 HyperFlash method at room temperature. Archimedes density and specific heat completed the thermal-conductivity calculation.

Process stepReported condition
Diamond particlesAverage diameter 400 µm
Surface cleaningNitric acid, sodium hydroxide, acetone, anhydrous ethanol, ultrasonic cleaning
Tungsten depositionPulsed magnetron sputtering in argon; average thickness 200 nm
Annealing1100 °C, 5 × 10−4 Pa, 10 °C/min
Copper depositionChemical deposition in a rotary evaporator after activation and sensitization
Hot-pressed sample30 mm diameter and 1.5 mm thickness
Thermal test12.7 mm diameter discs measured at room temperature

The initial sputtered coating contained 83% W and 17% WC. Heating during sputtering caused some carbide formation. One hour of annealing introduced metastable W2C and stable WC. By 4 h, XRD found no remaining W, although 6 h annealing still left some W2C. The 2 h optimum retained 2 wt.% W, 92 wt.% WC, and 6 wt.% W2C.

SEM showed a change from a continuous smooth layer to a porous structure during the first 1.5 h. Longer annealing refined the grains and reduced porosity, forming a denser carbide layer through 6 h. Grain growth was faster on the (100) face than on the (111) face. The authors attribute this difference to the higher surface energy of (100) and its better wettability and adhesion to copper.

Raman spectra tracked carbon conversion at 634 and 696 cm−1 and stronger D and G signals after annealing. The authors interpret this as partial sp3 to sp2 conversion at the diamond surface. They could not relate Raman peak ratios quantitatively to composite thermal conductivity.

Particle size, coating thickness, and diamond content

Section titled “Particle size, coating thickness, and diamond content”

Composites annealed for 2 h showed several processing effects. With a 50 nm coating, thermal conductivity rose from 344 W/(m·K) for 150 µm particles to 465 W/(m·K) for 400 µm particles. The authors link the increase to a better conduction path. At 150 µm, raising coating thickness to 200 nm increased conductivity from 344 to 395 W/(m·K).

With 400 µm particles and a 200 nm coating, conductivity ranged from 465 to 640 W/(m·K). The authors state that the coating-thickness response was unstable and may reflect higher interfacial thermal resistance. At 50 vol.% diamond, the observed range was 344 to 640 W/(m·K). Raising diamond content to 60 vol.% reduced conductivity to 607 W/(m·K). Poor particle contact was proposed as the cause.

Interface structure governs the measured conductivity

Section titled “Interface structure governs the measured conductivity”

The annealing series links bulk transport to interface structure. Thermal diffusivity peaked at 238.083 mm²/s after 1.5 h. Density peaked at 6.57 g/cm³ after 2 h, when diffusivity was 227.891 mm²/s and specific heat was 0.428 J/g/K. The 2 h sample reached 640 W/(m·K), while the 1.5 h sample had comparable final conductivity despite lower calculated interface conductance.

TEM of the 1 h sample showed bonded interfaces without obvious porosity or cracks, but incomplete reaction left large amorphous regions. The 2 h sample had a more uniform coating, a smaller amorphous region, and a denser interface. Its central layer became mainly ordered crystalline WC. Both samples showed W2C bonded directly to copper without large amorphous zones at that boundary. The authors related the 2 h structure to its higher conductivity.

Hasselman-Johnson and differential effective medium calculations placed interface conductance at 36 to 56 MW/(m²·K) for the 2 h condition. It fell to about 7 MW/(m²·K) after longer annealing. The authors relate the decline to graphitization and amorphous carbon, which lower interface transport.

AMM and extended DMM calculations gave 2.59 MW/(m²·K) for amorphous carbon against diamond and 29.78 MW/(m²·K) for amorphous carbon against W2C. The W2C/WC pair reached 11,400 MW/(m²·K) because of acoustic matching. The model calculation identifies diamond/W2C/WC/W2C/Cu as the best achievable sequence. The Cu-side W2C was proposed to regulate phonon modes and reduce scattering.

At 50 vol.% diamond, the 640 W/(m·K) composite remained below a reported W-coated hot-pressed value of 721 W/(m·K), but that sample contained 55 vol.% diamond. Literature hot-pressed peaks reached 685 W/(m·K) with Ti coating, 665 W/(m·K) with B coating, 615 W/(m·K) with Zr coating, and 593 W/(m·K) with Cr coating.

The paper also cites a 661 W/(m·K) hot-pressed W-coated composite and a 874 W/(m·K) vacuum-infiltrated composite. Spark plasma sintering and gas pressure infiltration produced composites above 800 W/(m·K). The authors position their result as competitive for hot pressing at no more than 50 vol.% diamond, not as the highest value across all processing methods.

Longer annealing caused diamond carbon to graphitize and increased low-conductivity amorphous phases. The authors expect better control of coating composition and interface microstructure to suppress this process. They also identify increased diamond fraction and reduced porosity as the main targets for future work.

Increasing diamond content above 50 vol.% carried a risk of cracks after hot pressing and reduced composite stability. The 60 vol.% trial reached 607 W/(m·K). This limits a simple strategy of adding more diamond.

Thermal conductivity was measured at room temperature. The reported particle conductivity was estimated from 147 ppm nitrogen rather than measured directly. The Raman peak ratio was not validated as a quantitative predictor of thermal conductivity. The method also gives no hot-press temperature, applied pressure, or hold time.

The measured conductivity exceeds the paper’s stated 400 W/(m·K) for copper and remains below the estimated 1717 W/(m·K) of the diamond particles. The composite therefore offers a practical balance between the matrix’s processability and diamond’s high conductivity. The authors identify high-power electronic cooling, thermal management systems, and thermoelectric conversion as possible applications.

For material selection, the controlled 50 vol.% structure is the demonstrated reference. Annealing time and carbide phase distribution matter as much as diamond content. Device qualification would require evidence for thermal cycling, elevated-temperature operation, and mechanical reliability, which the paper does not report.

View Original Abstract

Diamond-copper composites, due to their exceptional thermal conductivity, hold significant potential in the field of electronic device thermal management. Hot-press sintering is a promising fabrication technique with industrial application prospects; however, the thermal conductivity of composites prepared by this method has yet to reach optimal levels. In this study, tungsten was deposited on the surface of diamond particles by magnetron sputtering as an interfacial transition layer, and hot-press sintering was employed to fabricate the composites. The findings reveal that with prolonged annealing time, tungsten gradually transformed into W2C and WC, significantly enhancing interfacial bonding strength. When the diamond volume content was 50% and the interfacial coating consisted of 2 wt.% W, 92 wt.% WC, and 6 wt.% W2C, the composite exhibited a thermal conductivity of 640 W/(m·K), the highest value reported among hot-press sintered composites with diamond content below 50%. Additionally, the AMM (Acoustic Mismatch Model) and DMM (Diffusion Mismatch Model) models were utilized to calculate the interfacial thermal conductance between different phases, identifying the optimal interfacial structure as diamond/W2C/WC/W2C/Cu. This composite material shows potential for application in high-power electronic device cooling, thermal management systems, and thermoelectric conversion, providing a more efficient thermal dissipation solution for related devices.

  1. 2018 - Theoretical modelling for interface design and thermal conductivity prediction in diamond/Cu composites [Crossref]
  2. 2021 - Reinforcement size effect on thermal conductivity in Cu-B/diamond composite [Crossref]
  3. 2020 - High-Temperature Thermal Conductivity and Thermal Cycling Behavior of Cu-B/Diamond Composites [Crossref]
  4. 2012 - Preparation of Si-diamond-SiC composites by in-situ reactive sintering and their thermal properties [Crossref]
  5. 2018 - Combining Cr pre-coating and Cr alloying to improve the thermal conductivity of diamond particles reinforced Cu matrix composites [Crossref]
  6. 2018 - Enhanced thermal conductivity in Cu/diamond composites by tailoring the thickness of interfacial TiC layer [Crossref]