Joseph P. Klesko

Substrate selectivity in the low temperature atomic layer deposition of cobalt metal films from bis(1,4-di- tert -butyl-1,3-diazadienyl)cobalt and formic acid

The initial stages of cobalt metal growth by atomic layer deposition are described using the precursors bis(1,4-di-tert-butyl-1,3-diazadienyl)cobalt and formic acid. Ruthenium, platinum, copper, Si(100), Si-H, SiO2, and carbon-doped oxide substrates were used with a growth temperature of 180 °C. On platinum and copper, plots of thickness versus number of growth cycles were linear between 25 and 250 cycles, with growth rates of 0.98 Å/cycle. By contrast, growth on ruthenium showed a delay of up to 250 cycles before a normal growth rate was obtained. No films were observed after 25 and 50 cycles. Between 100 and 150 cycles, a rapid growth rate of ∼1.6 Å/cycle was observed, which suggests that a chemical vapor deposition-like growth occurs until the ruthenium surface is covered with ∼10 nm of cobalt metal. Atomic force microscopy showed smooth, continuous cobalt metal films on platinum after 150 cycles, with an rms surface roughness of 0.6 nm. Films grown on copper gave rms surface roughnesses of 1.1-2.4 nm after 150 cycles. Films grown on ruthenium, platinum, and copper showed resistivities of <20 μΩ cm after 250 cycles and had values close to those of the uncoated substrates at ≤150 cycles. X-ray photoelectron spectroscopy of films grown with 150 cycles on a platinum substrate showed surface oxidation of the cobalt, with cobalt metal underneath. Analogous analysis of a film grown with 150 cycles on a copper substrate showed cobalt oxide throughout the film. No film growth was observed after 1000 cycles on Si(100), Si-H, and carbon-doped oxide substrates. Growth on thermal SiO2 substrates gave ∼35 nm thick layers of cobalt(ii) formate after ≥500 cycles. Inherently selective deposition of cobalt on metallic substrates over Si(100), Si-H, and carbon-doped oxide was observed from 160 °C to 200 °C. Particle deposition occurred on carbon-doped oxide substrates at 220 °C.

Unusual stoichiometry control in the atomic layer deposition of manganese borate films from manganese bis(tris(pyrazolyl)borate) and ozone

The atomic layer deposition (ALD) of films with the approximate compositions Mn3(BO3)2 and CoB2O4 is described using MnTp2 or CoTp2 [Tp = tris(pyrazolyl)borate] with ozone. The solid state decomposition temperatures of MnTp2 and CoTp2 are ∼370 and ∼340 °C, respectively. Preparative-scale sublimations of MnTp2 and CoTp2 at 210 °C/0.05 Torr afforded >99% recoveries with <0.1% nonvolatile residues. Self-limited ALD growth was demonstrated at 325 °C for MnTp2 or CoTp2 with ozone as the coreactant. The growth rate for the manganese borate process was 0.19 A/cycle within the ALD window of 300–350 °C. The growth rate for the cobalt borate process was 0.39–0.42 A/cycle at 325 °C. X-ray diffraction of the as-deposited films indicated that they were amorphous. Atomic force microscopy of 35–36 nm thick manganese borate films grown within the 300–350 °C ALD window showed root mean square surface roughnesses of 0.4–0.6 nm. Film stoichiometries were assessed by x-ray photoelectron spectroscopy and time of flight-elasti...

Low Temperature, Selective Atomic Layer Deposition of Nickel Metal Thin Films

We report the growth of nickel metal films by atomic layer deposition (ALD) employing bis(1,4-di- tert-butyl-1,3-diazadienyl)nickel and tert-butylamine as the precursors. A range of metal and insulating substrates were explored. An initial deposition study was carried out on platinum substrates. Deposition temperatures ranged from 160 to 220 °C. Saturation plots demonstrated self-limited growth for both precursors, with a growth rate of 0.60 Å/cycle. A plot of growth rate versus substrate temperature showed an ALD window from 180 to 195 °C. Crystalline nickel metal was observed by X-ray diffraction for a 60 nm thick film deposited at 180 °C. Films with thicknesses of 18 and 60 nm grown at 180 °C showed low root mean square roughnesses (<2.5% of thicknesses) by atomic force microscopy. X-ray photoelectron spectroscopies of 18 and 60 nm thick films deposited on platinum at 180 °C revealed ionizations consistent with nickel metal after sputtering with argon ions. The nickel content in the films was >97%, with low levels of carbon, nitrogen, and oxygen. Films deposited on ruthenium substrates displayed lower growth rates than those observed on platinum substrates. On copper substrates, discontinuous island growth was observed at ≤1000 cycles. Film growth was not observed on insulating substrates under any conditions. The new nickel metal ALD procedure gives inherently selective deposition on ruthenium and platinum from 160 to 220 °C.

Thermal Atomic Layer Etching of Silica and Alumina Thin Films Using Trimethylaluminum with Hydrogen Fluoride or Fluoroform

Thermal atomic layer etching (ALE) is an emerging technique that involves the sequential removal of monolayers of a film by alternating self-limiting reactions, some of which generate volatile products. Although traditional ALE processes rely on the use of plasma, several thermal ALE processes have recently been developed using hydrogen fluoride (HF) with precursors such as trimethylaluminum (TMA) or tin acetylacetonate. While HF is currently the most effective reagent for ALE, its potential hazards and corrosive nature have motivated searches for alternative chemicals. Herein, we investigate the feasibility of using fluoroform (CHF3) with TMA for the thermal ALE of SiO2 and Al2O3 surfaces and compare it to the established TMA/HF process. A fundamental mechanistic understanding is derived by combining in situ Fourier transform infrared spectroscopy, ex situ X-ray photoemission spectroscopy, ex situ low-energy ion scattering, and ex situ spectroscopic ellipsometry. Specifically, we determine the role of TMA, the dependence of the etch rate on precursor gas pressure, and the formation of a residual fluoride layer. Although CHF3 reacts with TMA-treated oxide surfaces, etching is hindered by the concurrent deposition of a fluorine-containing layer, which makes it unfavorable for etching. Moreover, since fluorine contamination can be deleterious to device performance and its presence in thin films is an inherent problem for established ALE processes using HF, we present a novel method to remove the residual fluorine accumulated during the ALE process by exposure to water vapor. XPS analysis herein reveals that an Al2O3 film etched using TMA/HF at 325 °C contains 25.4 at. % fluorine in the surface region. In situ exposure of this film to water vapor at 325 °C results in ∼90% removal of the fluorine. This simple approach for fluorine removal can easily be applied to ALE-treated films to mitigate contamination and retain surface stoichiometry.

Selective Growth of Interface Layers from Reactions of Sc(MeCp)2(Me2pz) with Oxide Substrates

The transformation of an oxide substrate by its reaction with a chemical precursor during atomic layer deposition (ALD) has not attracted much attention, as films are typically deposited on top of the oxide substrate. However, any modification to the substrate surface can impact the electrical and optical properties of the device. We demonstrate herein the ability of a precursor to react deep within an oxide substrate to form an interfacial layer that is distinct from both the substrate and deposited film. This phenomenon is studied using a scandium precursor, Sc(MeCp)2(Me2pz) (1, MeCp = methylcyclopentadienyl, Me2pz = 3,5-dimethylpyrazolate), and five oxide substrates (SiO2, ZnO, Al2O3, TiO2, and HfO2). In situ Fourier transform infrared (FTIR) spectroscopy shows that at moderate temperatures (∼150 °C) the pyrazolate group of 1 reacts with the surface hydroxyl groups of OH-terminated SiO2 substrates. However, at slightly higher temperatures (≥225 °C) typically used for the ALD of Sc2O3, there is a direct reaction between 1 and the SiO2 layer, in addition to chemisorption at the surface hydroxyl groups. This reaction is sustained by sequential exposures of 1 until an ∼2 nm thick passivating interface layer is formed, indicating that 1 reacts with oxygen derived from SiO2. A shift of the Si 2p core level position, measured by ex situ X-ray photoelectron spectroscopy, is consistent with the formation of a ScSi xO y layer. Similar observations are made following the exposure of a ZnO substrate to 1 at 275 °C. In contrast, Al2O3, TiO2, and HfO2 substrates remain resistant to reaction with 1 under similar conditions, except for a surface reaction occurring in the case of TiO2. These striking observations are attributed to the differences in the electrochemical potentials of the elements comprising the oxide substrates to that of scandium. Precursor 1 can react with SiO2 or ZnO substrates, since the constituent elements of these oxides have less-negative electrochemical potentials than do aluminum, titanium, and hafnium. Additionally, Sc2O3 and surface carbonates are deposited on all substrates by gas-phase reactions between 1 and residual water vapor in the reactor. The extent of gas-phase reactions contributing to film growth is governed by the relative pressure of water vapor in the presence of 1. These results suggest caution when using very reactive, oxophilic precursors such as 1 to avoid misinterpreting unconventional film deposition as that resulting from a standard ALD process.