Influence of diamond surface termination on thermal boundary conductance between Al and diamond

Abstract

The effect of diamond surface treatment on the Thermal Boundary Conductance (TBC) between Al and diamond is investigated. The treatments consist in either of the following: exposition to a plasma of pure Ar, Ar:H and Ar:O, and HNO3:H2SO4 acid dip for various times. The surface of diamond after treatment is analyzed by X-ray Photoelectron Spectroscopy, revealing hydrogen termination for the as-received and Ar:H plasma treated samples, pure sp2 termination for Ar treated ones and oxygen (keton-like) termination for the other treatments. At ambient, all the specific treatments improve the TBC between Al and diamond from 23 ± 2 MW m–2 K–1 for the as-received to 65 ± 5, 125 ± 20, 150 ± 20, 180 ± 20 MW m–2 K–1 for the ones treated by Ar:H plasma, acid, pure Ar plasma, and Ar:O plasma with an evaporated Al layer on top, respectively. The effect of these treatments on temperature dependence are also observed and compared with the most common models available in the literature as well as experimental values in the same system. The results obtained show that the values measured for an Ar:O plasma treated diamond with Al sputtered on top stay consistently higher than the values existing in the literature over a temperature range from 78 to 290 K, probably due a lower sample surface roughness. Around ambient, the TBC values measured lay close to or even somewhat above the radiation limit, suggesting that inelastic or electronic processes may influence the transfer of heat at this metal/dielectric interface.

Qualitative link between work of adhesion and thermal conductance of metal/diamond interfaces

“…Experiment a. Experimental setup The experimental setup used in this paper is a coaxial two tints pump/probe experiment 34 and has been described in detail previously 19,35 .…”

“…TBC measurement For the measurement of TBC, datasets containing X/Y ratio over delay times from 0 to 4.02 ns were recorded. The obtained experimental data were fitted using the model first proposed by Cahill 36,37 with TBC between metal and diamond substrate as a free parameter 19,38 . The thermal model used did not include the TBC between the Al suscepting layer and the metal layer since the TBC values of metal-metal interfaces is typically one order of magnitude higher 12 than that of metal-dielectric ones and therefore it should not have a substantial influence on the measurement.…”

“…The substrates were then rinsed clean with acetone, ethanol and isopropanol. Though at this point the diamond surfaces are already H-terminated 19,39 , the surface condition of the diamonds was modified using 2 additional and different treatments.…”

“…the TBC is mainly governed by properties of the materials on either side of the interface, but several features of the interface itself can also have a substantial influence on the resulting performance. Examples of such features include: intermixing between the layer and substrate 11 , ion bombardment of the interface 12,13 , roughness 14,15 and chemical termination of the substrate [16][17][18][19] , all of which have an impact on the TBC. Recently, Losego et al showed that there is a link between the chemical bonding at an interface between gold and quartz and the TBC of this interface by adding Self Assembled Monolayers (SAMs) between these two materials 20 .…”

“…The metal/diamond interface has already been shown to exhibit a wide variety of TBC depending on the diamond surface termination 17,19 and orientation 32 . A similar trend has been reported on the Al/graphene interface 33 .…”