Atomic Layer Deposition of Ruthenium Films Using Ruthenium Diketonates and O₂, H₂, or N₂O: The Role of Ruthenium Etching

Abstract

The evolution of the surface during the steps that comprise the atomic layer deposition (ALD) of ruthenium films on a nickel substrate using tris­(2,2,6,6-tetramethyl-3,5-heptanedionato)­ruthenium­(III) (Ru­(tmhd)3) and molecular oxygen was characterized using a combination of X-ray photoelectron (XPS) and reflection–absorption infrared (RAIRS) spectroscopies. The uptake of the Ru metalorganic precursor was determined to be activated, involving the average loss of two out of the three ligands (and the retention of the third, in molecular form, at the surface–vacuum interface), and self-limited, as required in ALD. The reaction of the resulting layer from that first half of the ALD cycle with O2 proved to be more complex: in addition to the desired removal of the carbon-containing material, the Ni substrate becomes oxidized, and some Ru is etched away in the form of the volatile RuO4 gas. By testing different combinations of exposures and temperatures it was determined that Ru film growth was possible, but tuning and optimizing the ALD or atomic layer etching (ALE) process conditions for maximum film growth or removal rate per cycle proved difficult. Replacing O2 with H2 was shown not to be viable either, as such an agent can reduce the nickel oxide formed after O2 treatments (in fact, Ru­(tmhd)3 can do this as well) but not remove the carbonaceous material deposited on the surface by the Ru precursor. The use of N2O, on the other hand, showed much promise, being capable of removing most of the surface carbon without affecting the Ru film or the Ni substrate.

The surface chemistry of the atomic layer deposition of metal thin films

“…We have recently advanced the idea of using nitrous oxide as a source of O• radicals: N 2 O is in some ways similar to O 3 , but it is more stable, and upon dissociation yields N 2 , an inert gas, as the byproduct. We have shown that, indeed, Ru metal films (using Ru(tmhd) 3 ) are difficult to grow with O 2 but can be developed using N 2 O (figure 10, right panel) [93]. However, when attempting to deposit Mn on a silicon dioxide substrate using MeCpMn(CO) 3 and N 2 O, a multicomponent film was found to form instead comprising of manganese silicate and manganese silicide phases [171].…”

“…The ALD reactor, shown in the left side of figure 2, is also kept under vacuum but can be pressurized to carry out the ALD steps; it includes a setup for RAIRS that can be used in situ during gas exposures. The sample is mounted on a transferring rod for fast exchange between these two chambers without exposure to the outside atmosphere, and also to cool down and/or resistively heat it to any temperature between 150 and 1100 K. We have used this type of equipment extensively already, for studies of both catalytic [89][90][91] and ALD [92][93][94] systems. Instruments designed with a similar philosophy have been recently reported by other research groups [95][96][97].…”

“…with a net result of no metal film growth. This is certainly an issue with ruthenium, which, upon reaction with O 2 or O 3 , forms RuO 4 and desorbs from the surface (figure 10, left panel) [93,170]; mild ALD conditions or milder oxidants are needed in such cases to stop the metal oxidation at lower oxidation states (RuO 2 ). We have recently advanced the idea of using nitrous oxide as a source of O• radicals: N 2 O is in some ways similar to O 3 , but it is more stable, and upon dissociation yields N 2 , an inert gas, as the byproduct.…”