The thermal oxidation of Cu interconnects, at 350°C in air, has been studied as a function of thickness of a CoWP capping layer. For thin CoWP layers (25 nm), a thick oxide layer (200 nm) is formed which is mainly composed of Cu20. For thick CoWP layers (50 nm), the oxide layer is much thinner (36 nm) and is mainly composed of CoO. For both CoWP thicknesses, depletion of the underlying Cu is often observed after oxidation and whisker growth is often observed on the surface. The results are consistent with an oxidation mechanism where metal is the dominant diffusing species. For thin CoWP layers, Cu diffuses more quickly to the surface than Co, and therefore mainly Cu oxides are formed. For thick CoWP layers, the Cu diffusion to the surface is greatly reduced, and as a result, mainly Co oxides are formed. These results indicate that CoWP is not a good barrier for thermal oxidation, so high temperature exposure to oxidizing ambients must be minimized during processing of integrated circuits where CoWP is used instead of a dielectric barrier.

Material interaction during integration of tungsten gate stack for 1 Gb DRAM was investigated by Transition Electron Microscopy (TEM), X-ray Diffraction analysis (XRD) and Auger Electron Spectroscopy (AES). During selective side-wall oxidation tungsten gate conductor undergoes a structural transformation. The transformation results in the reduction of tungsten crystal lattice spacing, re-crystallization of tungsten and/or growth of grains. During a highly selective oxidation process, a relatively small but noticeable amount of oxygen was incorporated into the tungsten layer. The incorporation of oxygen is attributed to the formation of a stable WO x (x<2) composite.

The effectiveness of a TiN capping layer to prevent the conversion of α-titantium to titanium nitride when annealed in a nitrogen ambient has been studied over the temperature range 300–700°C using in-situ high temperature diffraction and transmission electron microscopy. Over the time range of interest (four hours), no evidence of Ti reaction was observed at 300°C. At 450°C. nitrogen was found to diffuse into the Ti to form a Ti(N) solid solution. Above 500°C the titanium is transformed to a second phase: however this reaction follows two different kinetic paths, depending on the annealing temperature. Below 600°C. the reaction proceeds in two stages, with the first stage consisting of Ti(N) formation, and the second stage consisting of the conversion of the Ti(N) with a transformation mechanism characteristic of short range diffusion (grain edge nucleation). Above 600°C, a simple linear transformation rate is observed.

Interconnection metallization uses film stacks, often composed of thin (<10 nm) Ti, TiN, or Ti/TiN underlayer(s) with a thick (200–1000 nm) Al-alloy film deposited on top. The texture or preferred orientation in such film stacks has important implications for both processing and reliability. Earlier studies' have demonstrated the importance of the underlayers on Al texture; however, to date no systematic work has been done on the effect of processing conditions on underlayer texture. This study examines the effect of deposition parameters on the underlayer texture development as well as the effect of this underlayer texture on subsequently deposited Al-alloy films. Fiber plots were obtained for Ti <002> and <101> and Al <111> reflections for a series of 20 nm Ti/ 10 nm TiN/400 nm AlCu films using both a conventional Siemens D500 diffractometer with a pole figure attachment and a Siemens HI-STAR Area Detector system using Cu radiation from a rotating anode source. Because of overlap between the Al <111> and Ti <101> reflections, the Al was removed with a subtractive etch. In this way both the Al and underlayer film textures could be quantified. It was found that the Ti and Al-alloy film textures vary depending on the deposition temperature, deposition method and final film thickness. For example, an increase in the substrate temperature from 300° to 500°C caused the Ti film texture to change from <002> to <101>. Additionally, switching the TiN deposition process from physical vapor deposition (PVD) sputtering to chemical vapor deposition (CVD) in a Ti/TiN/AlCu film stack caused a degradation in the AlCu <111 > texture.

Material anisotropy implies that many film properties are affected by crystallographic orientation in the growth direction (out-of-plane texture) and / or in the plane of growth (in-plane texture). Physical vapor deposited (PVD) Ti and Al-alloy films deposited on silicon dioxide substrates typically exhibit strong fiber textures in the growth direction with little in-plane-texture observed. The strength of these fiber textures has been found to vary substantially depending on the details of the deposition process(es) and, to a lesser degree, on any post-deposition anneals. In this paper the role of the substrate surface roughness in defining film texture is reported. It was found that the substrate surface roughness determines the overlying film crystallographic orientation for Ti and Ti/AlCu films deposited on various oxides. Furthermore, it was found that the texture of the initial metal “seed” layer defines the texture in subsequently deposited films (texture inheritance). Modifications to the oxide surface which decrease the surface roughness lead to an improved crystallographic texture in Ti, AlCu, Ti/AlCu and Ti/TiN/AlCu films. Film orientation was determined from crystallographic pole figures measured using x-ray diffraction (XRD). The oxide surface roughness was measured using atomic force microscopy (AFM), transmission electron microscopy (TEM) and glancing incidence x-ray reflectivity (GIXR).

We demonstrate the use of a synchrotron radiation source for in situ x-ray diffraction analysis during rapid thermal annealing (RTA) of 0.35 μm Salicide (self-aligned silicide) and 0.4 μm Polycide (silicided polysilicon) TiSi2 Complementary Metal Oxide Semiconductor (CMOS) gate structures. It is shown that the transformation from the C49 to C54 phase of TiSi2 occurs at higher temperatures in submicron gate structures than in unpatterned blanket films. In addition, the C54 that forms in submicron structures is (040) oriented, while the C54 that forms in unpatterned Salicide films is randomly oriented. Although the preferred oreintation of the initial C49 phase was different in the Salicide and Polycide gate structures, the final orientation of the C54 phase formed was the same. An incomplete conversion of C49 into C54-TiSi2 during the RTA of Polycide gate structures was observed and is attributed to the retarding effects of phosphorus on the transition.

The effect of microstructure, including average grain size, grain size distribution, precipitate distribution and crystallographic texture, on the reliability of Al and Al-alloys is well documented. In this paper, the various x-ray methods available for measuring preferred orientation in blanket films are compared and contrasted in an effort to find the fastest, most accurate method to acquire crystallographic texture information. I(111)/I(200) ratios from Bragg-Brentano scans (θ/2θ), rocking curves, and complete pole figures (fiber plots) were compared for four Ti/AlCu films having greatly different crystallographic texture components. The results show that it matters how texture is measured in a thin film; only pole figures (fiber plots) are unambiguous. In addition, the local texture in a series of stress-voided 0.48 um wide Ti/AICuSi/TiN lines was measured using Backscattered Kikuchi Diffraction (BKD). Samples were chosen from two sets of identically prepared wafers (processed at different times) showing large differences in stress-voiding driven resistance versus time behavior. The more strongly textured (111 ) films had decreased stress-voiding lifetimes and tended to have smaller average grain sizes with slightly larger grain size distributions. This is in apparent contradiction with previous results in Al films, where improved stress-voiding and electromigration behavior were found in those films with the strongest (111 ) texture.

We have obtained theoretical stress-temperature curves for passivated Al lines undergoing thermal cycling. A finite element plane-strain cross-sectional analysis with a time-dependent constitutive property for Al, based on equations for discreteobstacle controlled plasticity, was performed. The parameters for this Al constitutive relation were obtained by fitting with experimentally obtained stress-temperature curves for Al blanket films on silicon. Theoretical results agree well with the x-ray diffraction experimental data of Besser et al.1 Using a time-dependent property for Al helps match the data better than a time-independent property. Theoretical stress-temperature curves were also obtained for the longitudinal, transverse, and normal stress components in aluminum lines for line-widths ranging from 0.5 to 10 µm. The hysteresis of the stress-temperature curve of Al gets less as the line-width gets smaller. All stress components in the Al line change substantially with linewidth for the same oxide thickness.

A sputtered aluminum-low copper (Cu concentration < 2wt.%Cu), multilayered, submicron, device interconnect metallurgy consisting of two TiAl3 layers (∼,0.1μm thick) under and over an Al-Cu alloy conductor (0.95μm thick) with either an Al-Cu or TiN cap layer (0.05μm thick) has been developed. These films were patterned by reactive ion etching, and showed both a low susceptibility to corrosion and a low resistivity. Electromigration lifetime data on Al, Al-Cu, and both dc magnetron sputtered and electron gun evaporated multilayered fine lines, fabricated and tested in the same laboratory, are included for a direct comparison. Outstanding electromigration behavior was measured for sputtered, multilayered, submicron films with copper concentrations between 0.12 – 2wt.%Cu. In contrast, electromigration lifetimes of evaporated, multilayered films were found to degrade rapidly at < 2wt.%Cu. This anomalous electromigration behavior was attributed to structural differences in the Ti-Al intermetallics formed. It is proposed that defects present in the Ti-Al superlattice phase cause decreased electromigration performance compared to films in which TiAl3 exclusively forms. Multilevel structures, consisting of CVD W interlevel vias, were also investigated and found to have significantly degraded electromigration performance compared to planar samples. This was attributed to geometric and material flux divergences at the via/conductor interfaces.

We have investigated the effect of an interfacial Ti layer on the mechanism of CoSi2 formation. The presence of this Ti intcrlayer shifts the completion of all the cobalt suicide reactions to higher temperatures, reduces the sensitivity of these reactions to the presence of the native oxide on silicon, and greatly modifies the morphology of the CoSi2/ < 100 > Si interface. The modification of the interface morphology arises from the formation of highly < 110 > oriented or < 100 > epitaxial CoSi2 and results in greater thermal stability of the disilicide.

There is a great deal of interest by the electronics Industry in understanding the reactions occurring at the interface between a solid metal and a liquid solder or braze. This is due to the complex nature of current microelectronics Packaging, where soldering or brazing operations often involve complex metallurgies and tightly controlled furnace profiles. However, to date only a few studies have been carried out on the reaction kinetics at a liquid metal-solid metal interface. This is due in part to the difficulty in carrying out such , experiments. In the past, two main techniques have been used to obtain solid-state kinetic data. The first, quantitative metallography, is slow and tedious to perform, and generally of limited accuracy. The second technique involves measuring the change of some property, such as electrical resistivity, that is a function of the concentration of one of the phases. The main disadvantages of these techniques are that absolute values of concentration are not obtained, and that the relationships between a property and constitution are rarely available. In addition, most physical properties are sensitive to factors other than constitution, and interference from these factors can often result in erroneous data.