Market Progress in Molybdenum-Copper Heat Sink Materials

Heat dissipation for power devices and optical modules increasingly involves molybdenum-copper. The idea is direct: molybdenum has a low coefficient of thermal expansion and copper conducts heat well, and the composite strikes a compromise between expansion and conduction, matching ceramic substrates and chip materials better than pure copper. Market interest is rising, but selection still has to come back to a specific ratio and process, otherwise material data is easily mistaken for device performance. Grades of molybdenum-copper are described differently by each supplier, so key metrics should be written into negotiations as test methods.

Why Molybdenum-Copper Rather Than Pure Copper

Pure copper conducts heat well but expands a great deal, and repeated thermal cycling can crack the solder layer and the ceramic. Molybdenum-copper adjusts the proportion of the two phases to bring the expansion coefficient closer to the chip side while retaining relatively high conductivity. The price is higher density, greater machining difficulty and a cost above pure copper. Whether that is worthwhile depends on the power density and reliability requirements of the device and on the overall thermal design of the package; where power density is modest, pure copper or aluminum silicon carbide also have their applications, and comparisons should be made under the same thermal design conditions.

The Ratio Sets the Performance Bias

As molybdenum content rises, expansion falls, conductivity falls and density rises; more copper does the opposite. Common ratios are graded by application, with packaging houses usually specifying a target expansion coefficient and material suppliers back-calculating the ratio. Note that the same ratio produced by different processes still performs differently, so grades alone are not a basis for comparison. Uniformity through the thickness matters as well, since local copper-rich zones shift the expansion. At incoming inspection, density and expansion alone will screen out most off-spec batches.

Thickness deviation is amplified in later plating and assembly, so multiple points should be measured across the tolerance band marked on the drawing at incoming inspection, with edge and centre recorded separately. At the small-batch sampling stage, two or three ratios can be trialled first and compared with measured device temperatures before the grade is fixed, which avoids locking in a specification at the first attempt. Thermal simulation and measurement often diverge, so leaving room for adjustment at the sample stage is the safer course.

Interface Bonding Is Critical

Molybdenum and copper do not dissolve into each other, so interface bonding depends on process control. Infiltration first sinters a molybdenum skeleton and then infiltrates copper melt; the alternative is roll bonding or powder compaction. If the interface contains pores or oxide films, it becomes a crack origin under thermal cycling, and metallographic examination should check whether the transition zone is continuous. Infiltration temperature and holding time determine whether the copper phase fills completely, and skeleton porosity should be specified as well. The bonding process is described under process technology.

Forming and Downstream Processing

Heat sinks are mostly flat or shaped blocks; roll bonding suits large thin sheet, while infiltration suits complex shapes. Nickel or gold plating follows to improve solder wetting, and coating uniformity and adhesion need control, since incomplete pretreatment makes the coating blister. Avoid overheating during cutting so that local copper phases do not melt or the interface crack. Shaped parts are usually grit-blasted or chemically pretreated before plating, and coverage uniformity at corners and edges should be checked separately. Product forms and machining capability are covered under products and specifications.

Inspection and Acceptance Points

• Density and relative density, to judge whether residual porosity remains;
• Coefficient of thermal expansion and thermal conductivity, checked against the target range;
• Interface metallography and ultrasonic inspection, to observe bond quality;
• Coating thickness, adhesion and solderability;
• Dimensional change and delamination after thermal cycling.

These items should be covered in full at first qualification, then sampled by batch in production; if the device will later go through multiple reflows, the number of thermal cycles should be set according to the actual process.