Search

The use of a long throw sputter system for the deposition of refractory barriers is investigated by using the SIMBAD/SIMSPUD simulators. Findings indicate that both collimation and long throw systems produce a significant improvement in Ti film thickness uniformity. However, the long throw system has its disadvantage because of its asymmetrical film coverage over high aspect topography at the edge of the wafer.

Material movement resulting from the passage of an electrical current or the electromigration phenomenon continues to be a major reliability concern for the interconnects in very large scale integrated (VLSI) circuits. An experiment is conducted to investigate the temperature cycling effect on electromigration behavior under pulsed current conditions for metallization used in VLSI devices, the results of which are analyzed and discussed.

The one-dimensional electromigration boundary value problem under pulsed dc conditions is numerically examined by applying the transmission-line matrix modeling technique. A technique is suggested for easily transforming the pulsed stress into an equivalent dc stress that would cause electromigration damage at a similar rate. Using the fundamental diffusion-drift model, the vacancy buildup behavior under a pulsed dc stress can be described accurately by the dc stress scaled according to the duty factor of the current pulse.

Sputter deposition of titanium into trench and via topography has been studied under conditions of pressure as low as 0.3 mTorr (0.04 Pa) for long (30 cm) and short (9 cm) throw distances. Film resistivity and deposition rates at both distances have been measured for pressures from 0.3 mTorr up to 5 mTorr (0.67 Pa). Pinhole experiments have been used to examine the evolution of the angular distribution of sputtered flux as a function of pressure and distance. The experimental results show that by increasing the throw distance and reducing pressure, the bottom coverage in high aspect ratio topographical features may be increased to almost 90% but thickness nonuniformity under these conditions remains a significant problem due to the size of the target. We have tested for the presence of resputtering in the highly energetic, low collision regime which arises at low pressure and no resputtered material has been detected.

Transmission-line-matrix (TLM) method evaluates the grain boundary and interfacial diffusion problems in thin films, which is solved as a function of position and time. TLM treats complex structures, incorporates different boundary conditions, has an automatic time-stepping ability and uses a variable mesh. TLM is used in practical applications related to diffusion processes in thin films.

An absorbing filter that flattens the Gaussian intensity profile of a laser beam has been developed. The filter was fabricated by dc magnetron sputtering a thin tantalum film, by using a combination of substrate masking and motion. A model of the filter production has been developed that successfully predicts the film profile and performance. A typical filter produced by using these techniques results in a flat output to within [+ or -] 3% over a 2.2 mm with a transmission of 60% of the peak intensity of the Gaussian beam.

A quantitative analysis of time-dependent temperature distributions, as a result of high current pulse, in very large scale integrated (VLSI) metallization structures reveals that the thermal capacity of passivation materials can effectively inhibit the rate of temperature rise in the interconnects before the thermal equilibrium is achieved. The use of a passivation material, with thermal conductivity, increases instantaneous maximum current density to a high value.

The one-dimensional collisionless sheath model of electropositive plasmas is extended here to study the sheath structure of an electronegative plasma at steady state with the assumptions of cold positive ions and planar sheath configurations. Similar to the electropositive plasma sheath, a net charge layer is built within the sheath. However, the net charge density distribution within the sheath is sensitive to the plasma electronegativity (ratio of negative ion density [n.sub.-b] to the electron density [n.sub.eb]). It is most interesting that the sheath thickness varies nonmonotonically with the plasma electronegativity of the bulk plasma. Last, the effects of applied wall potential and temperature ratio of electrons to negative ions are discussed. Index Terms--Electronegative plasmas, sheath structure, sheath thickness.

Resonant glassy nanostrings have been employed for the detection of biomolecules. These devices offer high sensitivity and amenability to large array integration and multiplexed assays. Such a concept has however been impaired by the lack of stable and biocompatible linker chemistries. Diazonium salt reduction-induced aryl grafting is an aqueous-based process providing strong chemical adhesion. In this work, diazonium-based linker chemistry was performed for the first time on glassy nanostrings, which enabled the bio-functionalization of such devices. Large arrays of nanostrings with ultra-narrow widths down to 10 nm were fabricated employing electron beam lithography. Diazonium modification was first developed on SiCN surfaces and validated by X-ray photoelectron spectroscopy. Similarly modified nanostrings were then covalently functionalized with anti-rabbit IgG as a molecular probe. Specific enumeration of rabbit IgG was successfully performed through observation of downshifts of resonant frequencies. The specificity of this enumeration was confirmed through proper negative control experiments. Helium ion microscopy further verified the successful functionalization of nanostrings. [ABSTRACT FROM AUTHOR]

Following continuing trends in nanofabrication, the near future may see the requirement to integrate and assemble devices and integrated circuits that are below the scale that conventional robotic pick-and-place systems can successfully accommodate. Presented here is a protein--ligand based approach to self-assembling micronscale components onto specific patterned locations on a substrate. Other than the benefits in scale, this integration method may be advantageous for its parallel nature, 3D capabilities, and the ability to integrate devices made from incompatible processing technologies into a single platform (heterogeneous integration). Five micrometer square silicon microtiles were fabricated as model devices for microscale integrated circuits. They were fabricated from a silicon-on-insulator substrate and released into solution by bath ultrasonication after the buried oxide layer underneath them was underetched. A silicon target substrate was also patterned with gold pads for the microtiles to assemble onto. Self-assembled monolayers were employed to functionalize both the microtiles and the gold pads with biotin and avidin, respectively. Due to the very strong protein--ligand binding between avidin and biotin, the functionalized microtiles in solution were able to attach onto the target gold pads with a high selectivity. In this demonstration, for 5 lm square microtiles assembling onto square gold pads of the same size, 2.0% of the gold pads were covered by the microtiles and a selectivity (microtiles assembling onto the gold pads as opposed to the silicon substrate) of 7.3:1 was achieved. [ABSTRACT FROM AUTHOR]

The reliable and complete filling of vias and trenches with an appropriate metal is an important process in the fabrication of microelectronic components. Due to its favorable electronic properties and its reliability, copper is a common choice for replacing aluminum as a metallization material. One metallization procedure of interest is the sputter reflow of copper. The sputter reflow process begins with depositing copper via traditional sputter technologies. The films are subsequently made to fill micrometer and submicrometer trenches and vias through surface diffusion enhanced by annealing [1]. This paper studies fundamental considerations such as deposition rates, underlayer material, and transport mechanisms through numerical simulation using the process simulators SIMSPUD and GROFILMS. The simulations are further used to study via filling using Cu reflow and alternative methods including in situ annealing, three-step deposition, and a four-stage deposition-chemical mechanical polishing procedure. Simulation results are presented as depictions of the film on the feature scale. A discussion of the algorithmic solutions to the three-dimensional problems associated with vias is also provided.

Exposure of polymethylmethacrylate (PMMA) during electron beam lithography (EBL) produces small polymer fragments that dissolve rapidly during the development process. The resist dissolution behavior varies greatly depending on the nature of the developer (solvent) and therefore influences the selection of the EBL parameters, such as dose (sensitivity). A molecular scale examination of the development process is necessary to elucidate the resist-developer interaction mechanisms. In this work, the authors investigate the interaction of short PMMA chains (containing up to 10 MM A units) with common developer components methyl isobutyl ketone (MIBK) and isopropyl alcohol (IPA). For this purpose, the authors conduct molecular dynamics simulations using the Accelrys Materials Studio package. The simulation results were used to characterize the mixtures in the spirit of the Flory-Huggins theory of polymers and also to extract the diffusivities. The authors found that the behavior of PMMA fragments differed considerably in MIBK as compared with IPA. PMMA fragments containing more than three monomers exhibit stronger attractive interaction with MIBK. For all fragment sizes simulated, the diffusivity of PMMA fragments is 60-160% higher in MIBK as well. Similarly, the authors observed differences in the gyration radii. The authors conclude that the kinetic factor seems to be more significant as compared to affinity factor when accounting for differences in exposure sensitivities due to developer selection. [ABSTRACT FROM AUTHOR]

An approach of using a current-controlled voltage source analogy to account for a general boundary condition of particle or thermal flux in transmission-line matrix (TLM) diffusion modelling has been developed. For mass diffusion, this boundary condition is expressed as D ∂C/∂x + αC = 0. where C is the concentration, D is the diffusion coefficient, and α is a parameter characterizing the out-diffusion mobility across the boundary of interest. Confirmation of the TLM numerical treatment through comparison with analytical solutions is presented. Besides the open-circuit and short-circuit boundaries, the matched-load boundary is also found to be a special case of the present boundary condition. This extension of the boundary treatment allows the highly flexible TLM method to be applied to a greater variety of diffusion problems. [ABSTRACT FROM AUTHOR]