Substrate Selectivity of (^tBu-Allyl)Co(CO)₃ during Thermal Atomic Layer Deposition of Cobalt

Abstract

Tertbutylallylcobalttricarbonyl (tBu-AllylCo­(CO)3) is shown to have strong substrate selectivity during atomic layer deposition of metallic cobalt. The interaction of tBu-AllylCo­(CO)3 with SiO2 surfaces, where hydroxyl groups would normally provide more active reaction sites for nucleation during typical ALD processes, is thermodynamically disfavored, resulting in no chemical reaction on the surface at a deposition temperature of 140 °C. On the other hand, the precursor reacts strongly with H-terminated Si surfaces (H/Si), depositing ∼1 ML of cobalt after the first pulse by forming Si–Co metallic bonds. This remarkable substrate selectivity of tBu-AllylCo­(CO)3 is due to an ALD nucleation reaction process paralleling a catalytic hydrogenation, which requires a coreactant that acts as a hydrogen donor rather than a source of bare protons. The chemical specificity demonstrated in this work suggests a new paradigm for developing selective ALD precursors. Namely, selectivity can be achieved by tailoring the ligands in the coordination sphere to obtain structural analogues to reaction intermediates for catalytic transformations that exhibit the desired chemical discrimination.

Classification of processes for the atomic layer deposition of metals based on mechanistic information from density functional theory calculations

“…This is why the ALD of Co metal from Co(allyl) (CO) 3 and dimethylhydrazine (H 2 NNMe 2 ) is found to proceed selectively on a H-terminated Si surface rather than on OH-terminated SiO 2 . 24 DFT simulations show that the mechanism involves donation of a H atom (not H + ) from the substrate to Co and confirm that this nucleation reaction is thermodynamically favored on Si− −H and hindered on SiO 2 − −OH, reflecting the different H-donor capabilities of these surfaces.…”

Modeling Mechanism and Growth Reactions for New Nanofabrication Processes by Atomic Layer Deposition

“…[ 73 ] First principles simulations show that the mechanism involves donation of a H atom (not H + ) from the substrate to Co and confi rm that this nucleation reaction is thermodynamically favored on Si-H and hindered on SiO 2 -OH, refl ecting the different H-donor capabilities of these surfaces. [ 73 ] A quantum chemical study of the decomposition pathway of amidinate precursors for Co or Ni ALD shows that thermal and kinetic stability can be substantially enhanced by locating a CH 3 group at the β position, rather than H, which is consistent with experimental data. [ 74 ] Furthermore, chelation enhances the stability of the complex against β-H migration and decomposition.…”

Atomic-scale simulation of ALD chemistry

“…On a hydroxylated SiO 2 substrate, chemisorption of Cu 2 (amd) 2 is shown to proceed by elimination of amdH, followed by release of amidine during the H 2 pulse, while transfer of ligands to the substrate can lead to contamination with C. These conclusions are reached via the powerful combination of DFT and in situ IR spectroscopy [48]. Co films are found to nucleate better from Co( t Bu-allyl)(CO) 3 on H-terminated Si rather than on OH-terminated SiO 2 substrates and this is rationalised via DFT energetics that show the Si-H surface to be a superior donor of H to the allyl group via Co, analagous with hydrosilylation catalysis [45].…”