ALD of metal sulphide thin films has been known since the discovery of the technology in early 70s whereas ALD of metal selenide and telluride films has been limited because of a lack of precursors that would at the same time be safe and exhibit high reactivity as required in ALD. In this presentation we show that alkylsilanes of tellurium and selenium can be used as tellurium and selenium precursors in thermal ALD. Compounds with a general formula (R₃Si)₂Te and (R₃Si)₂Se react with various metal halides producing metal telluride and selenide thin films. Sb₂Te₃, GeTe and Ge₂Sb₂Te₅ films can be deposited by ALD at 90 °C using (Et₃Si)₂Te, SbCl₃ and GeCl₂·C₄H₈O₂ as precursors. All three precursors exhibit a typical saturative ALD growth behaviour. The Ge₂Sb₂Te₅ films show excellent conformality on a high aspect-ratio trench structure. Many other selenide and telluride films can be deposited by ALD using alkylsilanes of tellurium and selenium as precursors. Those deposited in this work include ZnTe, Bi₂Te₃, ZnSe, Bi₂Se₃, In₂Se₃ and Cu_xSe_y. The growth temperature has in some cases been 400 ^oC showing the thermal stability of these new Se and Te precursors. Growth rates of these binary chalcogenide films are typically between 0.5 and 1 Å/cycle. Other metal precursors than chlorides are also possible in the selenide and telluride depositions, as exemplified by the use of GeBr₂ and Sb(OEt)₃.

Metal ALD using thermal chemistry is limited and based on combustion reactions (Ru, Pt), organic or H₂ reduction (Cu, Pd) or fluorosilane elimination (W). Molybdenum (Mo) is a refractory metal that has applications in alloys, catalysis and electronics. Mo ALD can be achieved with fluorosilane elimination chemistry using MoF₆ and Si₂H₆ as reactants. This process is similar to W ALD using WF₆ and Si₂H₆ as reactants. This study reports Mo ALD using a quartz crystal microbalance (QCM) to monitor the growth of the Mo ALD films and Mo/W alloy films in a hot wall viscous flow reactor.

QCM studies showed that Mo ALD is self-limiting for both MoF₆ and Si₂H₆ reactants. MoF₆ produces a large mass gain and Si₂H₆ produces a small mass loss. A mass gain of 535 ng/cm² per cycle was observed at 120°C when both reactant exposures were in saturation.

Although long MoF₆ residence times were observed on the surface, the Mo ALD growth per cycle was independent of purge time. The Mo film growth reached a linear regime after a short nucleation period of only 3-4 cycles on Al₂O₃ ALD surfaces. X-ray reflectivity (XRR) experiments confirmed linear Mo ALD growth versus number of cycles. A growth per cycle of 6.4 Å/cycle was measured at 120°C. The average density of the Mo films was 8.7 g/cm³ and there was excellent agreement between the QCM and XRR experiments. The temperature dependence of the Mo ALD growth per cycle was investigated from 90 ºC to 150 ºC.

Atomic layer deposition (ALD) provides the unique and powerful capability to blend materials at the atomic scale for tuning and optimizing the properties of the resulting mixed layers. The ALD of mixed metal-oxide layers to maximize the charge storage capacity of dielectric materials is well known, but the ALD of mixed metal films is less well explored. The capability to synthesize mixed-metal layers with tunable physical and chemical properties could benefit numerous applications such as catalysis, hydrogen storage, corrosion resistance, and microelectronics, and the ALD of metal laminate nanostructures offers the possibility of core-shell structures and near surface alloys. In this study, we examine the ALD of platinum-iridium (Pt-Ir) mixed metal layers. The platinum ALD uses alternating exposures to (methylcyclopentadienyl) trimethylplatinum and oxygen while the iridium ALD uses alternating exposures to iridium(III) acetylacetonate and oxygen. The similar chemistries and process conditions for these pure metals facilitates ALD of the mixed metal layers. Furthermore, the tendency of Pt to form discrete nanoparticles on oxide supports makes this material attractive for catalytic application. We examined the Pt-Ir mixed metal ALD using in situ quartz crystal microbalance and quadrupole mass spectrometer measurements to investigate the growth mechanism for the pure and mixed materials as well as the effect of mixing on the metal nucleation and growth. In addition, ALD Pt-Ir films were prepared on planar substrates and examined with a variety of techniques to evaluate the thickness, morphology, crystal structure, and chemical composition of the films. These results demonstrate that the thickness and composition of the Pt-Ir films can be controlled precisely.

One of the least understood aspects of atomic layer deposition (ALD) is the initial stage of growth, which involves the first set of reactions between the thin film precursors and the substrate. As ALD growth is invariably conducted on foreign substrates, the surface must evolve from that representing the starting substrate, to that eventually representing the steady-state growth surface. Here we employ in situ x-ray photoelectron spectroscopy (XPS) to examine the initial stages of growth of TaN_x from the reactions of Ta[N(CH₃)₂]₅ and NH₃ on SiO₂, porous low-κ substrates, and both of these substrates modified with interfacial organic layers (IOLs). In this presentation, first, we will examine the effect of the density of the reactive adsorption sites (-OH groups) on SiO₂ and SiO₂ based porous low-κ substrates. Here we find that the saturation density (coverage) of Ta depends strongly on the initial density of –OH. Moreover, we find that there is a strong correlation between the amount of ligand loss in this first half-cycle, and the density of the reactive adsorption sites. For example, on porous low-κ substrates, reactions involving the loss of a single ligand and formation of –O-Ta[N(CH₃)₂]₄(a) species dominate. In contrast, on surfaces with a high density of –OH (SiO₂), ligand loss is much more significant. Second, another important feature that XPS can probe is the chemical state of the primary thin film constituents. One process observed in the first cycles of growth is the shift in the position of the N(1s) peak, from a binding energy of ca. 398.9 eV to 397.2 eV. The former binding energy corresponds well to that reported for –N(R)₂ species (398.6), while the latter is close to that reported for TaN (397.5) and Ta₃N₅ (396.9). As a final example, we will consider growth of TaN_x on SiO₂ and porous low-κ substrates modified with interfacial organic layers. We have found that adsorption of a branched polymer possessing a high density of –NH₂ species on the porous dielectric increases both the initial uptake of Ta[N(CH₃)₂]₅ in the first half cycle, and the growth rate per cycle. Here an important issue concerns the fate of the IOL—does the ALD thin film quickly and uniformly cover the IOL, or are there reactions between the ALD precursors and the IOL? We find that the fate of the IOL depends strongly on the density and spatial distribution of reactive groups in the organic layer. We will discuss the implications of these observations concerning the use of IOLs as nucleation promoters in ALD.