The central thesis of this research is that VLSI interconnect design strategies should shift from using global wires that can support only a single binary transition during the latency of the line to global wires that can sustain multiple bits traveling simultaneously along the length of the line. It is shown in this thesis that such throughput-centric multibit transmission can be achieved by wave-pipelining the interconnects using repeaters. A holistic analysis of wave-pipelined interconnect circuits, along with the full-custom optimization of these circuits, is performed in this research. With the help of models and methodologies developed in this thesis, the design rules for repeater insertion are crafted to simultaneously optimize performance, power, and area of VLSI global interconnect networks through a simultaneous application of voltage scaling and wire sizing. A qualitative analysis of latency, throughput, signal integrity, power dissipation, and area is performed that compares the results of design optimizations in this work to those of conventional global interconnect circuits. The objective of this thesis is to study the circuit- and system-level opportunities of voltage scaling, wire sizing, and repeater insertion in wave-pipelined global interconnect networks that are implemented in deep submicron technologies.

Small form-factor packages with high integration density are driving the innovations in chip-to-package interconnections. Metallurgical interconnections have evolved from the conventional eutectic and lead-free solders to fine pitch copper pillars with lead-free solder cap. However, scaling down the bump pitch below 50-80µm and increasing the interconnect density with this approach creates a challenge in terms of accurate solder mask lithography and joint reliability with low stand-off heights. Going beyond the state of the art flip-chip interconnection technology to achieve ultra-fine bump pitch and high reliability requires a fundamentally- different approach towards highly functional and integrated systems. This research demonstrates a low-profile copper-to-copper interconnect material and process approach with less than 20µm total height using adhesive bonding at lower temperature than other state-of-the-art methods. The research focuses on: (1) exploring a novel solution for ultra-fine pitch (< 30µm) interconnections, (2) advanced materials and assembly process for copper-to-copper interconnections, and (3) design, fabrication and characterization of test vehicles for reliability and failure analysis of the interconnection. This research represents the first demonstration of ultra-fine pitch Cu-to-Cu interconnection below 200°C using non-conductive film (NCF) as an adhesive to achieve bonding between silicon die and organic substrate. The fabrication process optimization and characterization of copper bumps, NCF and build-up substrate was performed as a part of the study. The test vehicles were studied for mechanical reliability performance under unbiased highly accelerated stress test (U-HAST), high temperature storage (HTS) and thermal shock test (TST). This robust interconnect scheme was also shown to perform well with different die sizes, die thicknesses and with embedded dies. A simple and reliable, low-cost and low-temperature direct Cu-Cu bonding was demonstrated offering a potential solution for future flip chip packages as well as with chip-last embedded active devices in organic substrates.

For decades, advances in device scaling has proven to be critical in improving the performance and productivity of 2D systems. In this thesis, we explore how advances in technology have pushed functional integration to such a high-level that interconnection and packaging issues represent real barriers to further progress. While three-dimensional (3D) integration offers to be a potential contender to overcome the barriers of increased energy consumption due to interconnects and bandwidth limitations, there are certain challenges that must be overcome before systems can be successfully stacked. Cooling and power delivery are among these key challenges in the integration of high performance 3D ICs. To address these challenges, microchannel heat sinks for inter-stratum cooling and through-silicon vias (TSVs) for signaling and power delivery between stacked ICs were explored. Novel integration schemes to integrate these uidic and electrical interconnects in conventional CMOS processes were also explored. Compact physical modeling was utilized to understand the trade-offs involved in the integration of electrical and microfluidic interconnects in a 3D IC stack. These concepts were demonstrated experimentally by showing different CMOS compatible methods of fabricating microchannels and integration of high aspect ratio (~20:1) and high density (200,000/cm²) electrical TSVs in the fins of the microchannels for signaling and power delivery. A novel mesh process for bottom up plating of high aspect ratio TSVs is also shown in this work. Fluidic reliability measurements are shown to demonstrate the feasibility of this technology. This work also demonstrates the design and fabrication of a 3D testbed which consists of a 2 chip stack with microchannel cooling on each level. Preliminary testing of the stack along with interlayer electro-fluidic I/Os has also been demonstrated.

As the current electronic trend is toward integrating multiple functions in a single electronic device, there is a clear need for increasing integration density which is becoming more emphasized than in the past. To meet the industrial need and realize the new system-integration law [1], three-dimensional (3-D) integration is becoming necessary. 3-D integration of multiple functional IC chip/package modules requires co-simulation of the chip and the package to evaluate the performance of the system accurately. Due to large scale differences in the physical dimensions of chip-package structures, the chip-package co-simulation in time-domain using the conventional FDTD scheme is challenging because of Courant-Friedrich-Levy (CFL) condition that limits the time step. Laguerre-FDTD has been proposed to overcome the limitations on the time step. To enhance performance and applicability, SLeEC methodology [2] has been proposed based on the Laguerre-FDTD method. However, the SLeEC method still has limitations to solve practical 3-D integration problems. This dissertation proposes further improvements of the Laguerre-FDTD and SLeEC method to address practical problems in 3-D interconnects and 3-D integration. A method that increases the accuracy in the conversion of the solutions from Laguerre-domain to time-domain is demonstrated. A methodology that enables the Laguerre-FDTD simulation for any length of time, which was challenging in prior work, is proposed. Therefore, the analysis of the low-frequency response can be performed from the time-domain simulation for a long time period. An efficient method to analyze frequency-domain response using time-domain simulations is introduced. Finally, to model practical structures, it is crucial to model dispersive materials. A Laguerre-FDTD formulation for frequency-dependent dispersive materials is derived in this dissertation and has been implemented.

Thesis (M.S.)--State University of New York at Binghamton, Thomas J. Watson School of Engineering and Applied Science, Department of Mechanical Engineering, 2009. / Includes bibliographical references.

Copper interconnect requires liner materials that function as a diffusion barrier, a seed layer for electroplating, and an adhesion promoting layer. Ruthenium has been considered as a promising liner material, however it has been reported that Ru itself is not an effective Cu diffusion barrier due to its microstructure, which is polycrystalline with columnar grains. The screening study of Ru precursors revealed that all Ru films were polycrystalline with columnar structure, and, due to its strong 3D growth mode, a conformal and ultrathin Ru film was difficult to form, especially on high aspect ratio features. The microstructure of Ru films can be modified by incorporating P. Amorphous Ru(P) films are formed by chemical vapor deposition at 575 K using a single source precursor, cis-RuH₂(P(CH₃)₃)₄, or dual sources, Ru₃(CO)₁₂ and P(CH₃)₃ or P(C6H5)₃ The films contain Ru and P, which are in zero-valent states, and C as an impurity. Phosphorus dominantly affects the film microstructure, and incorporating > 13% P resulted in amorphous Ru(P) films. Metastable Ru(P) remains amorphous after annealing at 675 K for 3 hr, and starts recrystallization at ~775 K. The density of states analysis of the amorphous Ru(P) alloy illustrates metallic character of the films, and hybridization between Ru 4d and P 3p orbitals, which contributes to stabilizing the amorphous structure. Co-dosing P(CH)₃ with Ru₃(CO)₁₂ improves film step coverage, and the most conformal Ru(P) film is obtained with cis-RuH2(P(CH₃)₃)₄; a fully continuous 5 nm Ru(P) film is formed within 1 µm deep, 8:1 aspect ratio trenches. First principles density functional theory calculations illustrate degraded Cu/Ru adhesion by the presence of P at the interface, however, due to the strong Ru-Cu bonds, amorphous Ru(P) forms a stronger interface with Cu than Ta and TaN do. Cu diffusion studies at 575 K suggests improved barrier property of amorphous Ru(P) films over polycrystalline PVD Ru.

M. Tech. Metallurgical Engineering. / Aims at improving Cu interconnection problem by homogeneous distribution of ruthenium and tantalum in Cu matrix for excellent interconnection in electronics packaging. The aim will be achieved through the following objectives.Development of appropriate technology for homogenizing submicron metal powders with suitable methods for controlling grain growth during sintering. Study the mechanisms of synergistic incorporation of Ru, and Ta on improving copper interconnection properties. To investigate metallurgical interactions and phenomena occurring during sintering. To investigate specific property and behaviour advantages intrinsic due to the composites and material mix.

The international technology roadmap for semiconductors (ITRS) 2012 report foresees the use of porous dielectric materials with a low dielectric constant in conjunction with copper interconnects as a way to reduce the resistive-capacitive (RC) delay in microelectronic applications. However, the introduction of pores in the dielectric not only serves as stress raisers but also diminishes the structural strength of the material. The challenge therefore with the implementation of low-k dielectrics for high-performance flip-chip packages is to create a reliable die to organic substrate interconnect solution which induces low stresses on the die to prevent the cohesive cracking and the interfacial delamination of the dielectric material. Potential interconnect solutions that meet this challenge are MEMS-like compliant freestanding micro-structures. These structures are designed to work as spring-like elements which allow the free lateral and out-of-plane motion between the silicon die and the organic substrate under assembly conditions as well as under thermal or power cycling. Thus, the focus of this research is to design, fabricate, and characterize electrically and mechanically an innovative compliant interconnect approach that addresses these challenges. The proposed interconnect is scalable in dimensions and pitch, and consists multiple electrical paths which will provide redundancy against interconnect failure. The multi-path design employs parallel electrical paths which effectively split a larger cross-sectional area into several smaller areas making the overall design more compliant than otherwise. This research proposes wafer-level, high-yield, CMOS-compatible fabrication procedure using sequential photolithography and copper electroplating. The proposed interconnects are symmetric and are amenable to easy reflow assembly to substrates. The mechanical compliance of the fabricated structures is studied through nano-indentation, while the electrical characteristics are assessed through fabricated prototypes. The xvi thermo-mechanical reliability of compliant interconnects is also demonstrated. Lastly, the dimensional scalability of the interconnects is also demonstrated.

Multi-chip integration with emerging technologies such as a 3D IC stack or 2.5D interposer is primarily enabled by the off-chip interconnections. The I/O density, speed and bandwidth requirements for emerging mobile and high-performance systems are projected to drive the interconnection pitch to less than 20 microns by 2015. A new class of low-temperature, low-pressure, high-throughput, cost-effective and maufacturable technologies are needed to enable such fine-pitch interconnections. A range of interconnection technologies are being pursued to achieve these fine-pitch interconnections, most notably direct Cu-Cu interconnections and copper pillars with solder caps. Direct Cu-Cu bonding has been a target in the semiconductor industry due to the high electrical and thermal conductivity of copper, its high current-carrying capability and compatibility with CMOS BEOL processes. However, stringent coplanarity requirements and high temperature and high pressure bonding needed for assembly have been the major barriers for this technology. Copper-solder interconnection technology has therefore become the main workhouse for off-chip interconnections, and has recently been demonstrated at pitches as low as 40 microns. However, the current interconnection approaches using copper-solder structures are not scalable to finer feature sizes due to electromigration, and reliability issues arising with decreased solder content. Solid Liquid Inter-Diffusion (SLID) bonding is a promising solution to achieve ultra-fine-pitch and ultra-short interconnections with a copper-solder system, as it relies on the conversion of the entire solder volume into thermally-stable and highly electromigration-resistant intermetallics with no residual solder. Such a complete conversion of solders to stable intermetallics, however, relies on a long assembly time or a subsequent post-annealing process. To achieve pitches lower than 30 micron pitch, this research aims to study two ultra-short copper-solder interconnection approaches: (i) copper pillar and solder cap technology, and (ii) a novel technology which will enable interconnections with improved electrical performance by fast and complete conversion of solders to stable intermetallics (IMCs) using Solid Liquid Diffusion (SLID) bonding approach. SLID bonding, being a liquid state diffusion process, combined with a novel, alternate layered copper-solder bump structure, leads to higher diffusion rates and a much faster conversion of solder to IMCs. Moreover this assembly bonding is done at a much lower temperature and pressure as compared to that used for Cu-Cu interconnections. FEM was used to study the effect of various assembly and bump-design characteristics on the post-assembly stress distribution in the ultra-short copper-solder joints, and design guidelines were evolved based on these results. Test vehicles, based on these guidelines, were designed and fabricated at 50 and 100 micron pitch for experimental analysis. The bumping process was optimized, and the effect of current density on the solder composition, bump-height non-uniformity and surface morphology of the deposited solder were studied. Ultra-short interconnections formed using the copper pillar and solder cap technology were characterized. A novel multi-layered copper-solder stack was designed based on diffusion modeling to optimize the bump stack configuration for high-throughput conversion to stable Cu3Sn intermetallic. Following this modeling, a novel bumping process with alternating copper and tin plating layers to predesigned thicknesses was then developed to fabricate the interconnection structure. Alternate layers of copper and tin were electroplated on a blanket wafer, as a first demonstration of this stack-technology. Dies with copper-solder test structures were bonded using SLID bonding to validate the formation of stable intermetallics.

Materials, Processes, and Characterization of Extended Air-gaps for the Intra-level Interconnection of Integrated Circuits Seongho Park 157 pages Directed by Dr. Paul A. Kohl and Dr. Sue Ann Bidstrup Allen The integration of an air-gap as an ultra low dielectric constant material in an intra-metal dielectric region of interconnect structure in integrated circuits was investigated in terms of material properties of a thermally decomposable sacrificial polymer, fabrication processes and electrical performance. Extension of the air-gap into the inter-layer dielectric region reduces the interconnect capacitance. In order to enhance the hardness of a polymer for the better process reliabilities, a conventional norbornene-based sacrificial polymer was electron-beam irradiated. Although the hardness of the polymer increased, the thermal properties degraded. A new high modulus tetracyclododecene-based sacrificial polymer was characterized and compared to the norbornene-based polymer in terms of hardness, process reliability and thermal properties. The tetracyclododecene-based polymer was harder and showed better process reliability than the norbornene-based sacrificial polymer. Using the tetracyclododecene-based sacrificial polymer, a single layer Cu/air-gap and extended Cu/air-gap structures were fabricated. The effective dielectric constant of the air-gap and extended air-gap structures were 2.42 and 2.17, respectively. This meets the requirements for the 32 nm node. Moisture uptake of the extended Cu/air-gap structure increased the effective dielectric constant. The exposure of the structure to hexamethyldisilazane vapor removed the absorbed moisture and changed the structure hydrophobic, improving the integration reliability. The integration processes of the air-gap and the extended air-gap into a dual damascene Cu metallization process has been proposed compared to state-of-the-art integration approaches.