Carl V. Thompson - US grants

Interconnects are playing an increasingly important role in limiting the speed of high performance integrated circuit chips and systems. This motivation has fostered research to use optical rather than electrical interconnects. The research in progress addresses the fabrication problems posed by efforts to use the standard silicon wafer with a gallium arsenide (GaAs) light emitter incorporated. The approach is experimental and investigates use of indium to aid in matching lattices of different materials.

This research consists of an experimental and theoretical investigation of grain growth in polycrystalline films on single crystal substrates. Reflected high energy and transmission electron diffraction, transmission electron microscopy, and thin film X-ray texture analysis are employed to characterize the film grain size and orientation. Metal-on-metal, metal-on-mica, and metal-on-alkali halide model systems are the principal materials under study. Experimental results on grain growth are compared to predictions from computer simulations and analytical theories. A primary goal is to determine the film/substrate interface energy as a function of relative film/substrate orientation by experimental methods and compare these results with atomistic computer simulations.

Misfit dislocations are thermodynamically stable in epitaxial films above a critical thickness, which depends mainly on the misfit magnitude between the film and substrate. The formation of dislocations does not occur immediately when the critical thickness is exceeded since kinetic factors are involved which are not adequately understood. This grant examines the role of misfit dislocations and strain in affecting behavior of magnetic films. The kinetics of dislocation nucleation in epitaxial thin films is pursued with an aim towards developing an in-situ method for monitoring misfit dislocation formation. The Magneto-optic Kerr (MOKE) effect is employed as a probe that is sensitive to the presence of small strains in magnetic films. This is used to characterize misfit strains and dislocations on epitaxial nickel/copper films. Other techniques employed are electron microscopy and secondary electron spin polarization analysis to determine the kinetics of dislocation formation. %%% The homogeneous strain of lattice mismatch and the highly non- uniform strains around dislocations have dramatic effects on magnetic properties. This research is directed towards developing a better understanding of the formation of misfit dislocations and their effects on magnetic films. An in-situ monitoring technique during film growth could result from the research.

The goal of this research is to understnd the fundamental features of ion induced surface reactions in focused ion beam microfabrication, and of the materials and surfaces that are produced by the resulting deposition or etching. This will be accomplished by bringing to bear expertise from chemical engineering, materials science, and microfabrication technology. In the case of the deposition, understanding of the basic process is expected to point to ways of eliminating unwanted carbon impurities, for example, by the addition of a second, appropriately reactive, precursor gas or an atomic beam. Eliminating or significantly reducing the impurities in the deposited material would be a major breakthrough. The reduced resistivity and increased density would immediately make focused ion beam device and circuit repair processes more effective and open the door to new applications of maskless, resistless, patterned conductor deposition by projection ion techniques and to in situ processing. %%% Ion induced deposition and ion assisted etching are new processes for material addition and removal. Since ion beams can be focused to extremely small dimensions (0.05 microns and below), this permits material manipulation with unprecedented resolution and flexibility and has spawned commercial applications in micro- repair of integrated circuits and in microsectioning for fault diagnosis. By developing a better understanding of the mechanisms of these new processes, and by extending them to new materials, the proposed research work aims to broaden the applications to the deposition of the original wiring of future ultrafast integrated circuits.

9710139 Thompson This is an experimental and theoretical research effort on the evolution of structure and the mechanical state of polycrystalline films during thin metal film deposition and during subsequent thermal processing. The goal of the research is to develop models and simulations which can be used to predict the course of grain size, grain orientation and stress evolution as a function of substrate, film and capping layer material properties, while also considering the deposition conditions and subsequent thermal history. The stresses in films deposited on substrates are monitored using wafer curvature measurements made while heating at controlled and variable rates and while annealing at different temperatures for different times. Film material, capping layers, film thickness, and film deposition temperature are among the varied parameters in these experiments. In situ TEM observations explore grain structure evolution in heated films on thin substrates which impose a strain in the films due to differential thermal expansion. Texture of the films is monitored by X-ray analyses. %%% As modeling and simulation capabilities evolve through comparison with experiments, new quantitative capabilities are developed for predicting grain structure, grain orientation, and stress evolution. This provides techniques for processing of films with properties that are optimized for specific engineering applications. Polycrystalline films are used in electronic, micromechanical, and magnetic devices. In these applications, the performance and reliability of polycrystalline films are strongly affected by their structure, specifically the distribution of grain sizes and crystallographic orientations. ***

This project will develop a class of methods known as 'templated self-assembly' that control the growth and self-assembly of nanostructures on surfaces. This will enable formation of monodisperse nanoscale features in precise positions on a substrate. This work will be an enabling technology in the design of new devices that utilize the properties of quantum dots and other nanoscale objects, in which the control of the sizes and spatial positions of the features is paramount in optimizing performance. The objective is to use lithography to modulate substrate surfaces with features of periodicity of order 100 nm, to form templates for the growth and self-assembly of nanostructures. In this process, lithography is not used to form the nanostructures themselves, but instead is used to form a template that will 'seed' the formation of nanostructures in particular locations. The nanostructures will be considerably smaller in size than the period of the template. The goal of the project is to develop the templated self-assembly of arrays of nanoscale semiconductor and metal islands controlled by epitaxial strain, surface chemistry or topography. The island formation will be achieved using both the deposition from the vapor phase and by the spontaneous agglomeration of metastable coninuous sold films. This work will be carried out by an interdisciplinary team of researchers from MIT in collaboration with IBM and Sandia National Laboratories working with a group of students and a postdoctoral researcher. The outreach involves a communication effort designed to inform the general public about nanotechnology through development of a web site and other scientific communication avenues, including the involvement of undergraduate students as well as other activities such as school visits.
%%%
The research will be focussed on the templated self-assembly of arrays of nanoscale semiconductor and metal islands controlled by epitaxial strain, surface chemistry or topography. The resulting well-ordered nanoscale island arrays will have technological relevance in devices that include optically active structures involving plasmon wires, and patterned magnetic recording media. A range of other applications will also benefit from the methods developed in this proposal; for instance optical devices based on arrays of semiconductor quantum dots. The educational goals of this work are to contribute to the public understanding of nanotechnology and to the training of skilled researchers.
***

The grant explores a clear understanding of the atomic processes involved in film formation. New stress measurement capabilities allow studies of adatom-substrate interactions, as well as adatom-adatom interactions in the sub-monolayer regime. An objective of the proposed study is to correlate in-situ measurements of stress evolution during growth and during growth interruptions with ex-situ STM, ex-situ TEM, and in-situ TEM observations of structure evolution. Experiments involving both epitaxial and polycrystalline film growth will be conducted. Detailed effects of populations of adatoms will be simulated using molecular dynamics and ab-initio approaches. Simulation results will be compared with measured forces in epitaxial systems. Through these studies, the separate effects of changing adatom concentrations and cluster formation will be determined. Experimental studies of later stages of growth will allow separate characterization of the effects of cluster coarsening, growth, and coalescence. Measurements of stress evolution, when correlated with studies of structure evolution, will provide a powerful tool for in-situ real-time kinetic analyses of the various atomistic processes involved in the earliest stages of film formation.
%%%
The study will have applications in a wide array of micro and nano-devices and systems including microelectronic, nanomagnetic, and micro- and nanoelectromechanical systems where the properties, performance, and reliability of nanocrystalline metallic structures are strongly affected by their stress state and defect structures. Results from this and related studies will be disseminated through a variety of graduate and undergraduate courses within the parent institution, as well as through partnerships with other universities and through summer courses for professional engineers. The research will be carried out with collaborators at MIT and in Singapore, and will indirectly support the research activities of the research groups of the faculty involved. Research results will be disseminated through publications and presentations, as well as via the Internet, and will be incorporated in software tools that are used directly by engineers.
***

TECHNiCAL: Vapor-deposited metal films and structures generally evolve through the Volmer-Weber (VW) mechanism of crystal island nucleation, growth, and coalescence on substrate surfaces. Films formed via this mechanism do not stably wet their substrate, so film formation is a consequence of kinetic constraints that force the development of metastable or unstable films and particles. Consequently, kinetic processes dictate the course of all levels of structure evolution during film formation, and therefore define the final properties of the as-deposited structures. We plan to study to key aspects of structure evolution during VW growth: 1) pre-coalescence island evolution, and 2) post-coalescence surface structure evolution. In the latter case, we will include studies of evolution of surfaces of homoepitaxial films, to isolate effects of surface processes from effects associated with grain structures. In both regimes we will use in-situ stress measurements to probe both island-scale and atomic scale processes with both the measurement sensitivity and the temporal resolution required to characterize fast atomic-scale kinetic processes that operate during film formation conditions that are typical for engineering applications. We will also use in-situ electron diffraction to probe surface structure evolution, and ex-situ probe-based and electron-based microscopies to characterize morphology and crystallographic characteristics of quenched structures. We plan to investigate the effects of changes in deposition flux and substrate temperature on stress and structure evolution in both regimes. We will investigate effects of growth interruptions as well the temporal variations in the growth conditions (including island size focusing techniques). We also plan to investigate the effects of low-level surfactant coverage and of nano-scale lithographically defined substrate topography on the size, shape, spacing, orientation and ordering of island and surface features, and on the final properties of deposited structures. NON-TECHNICAL: Metallic thin films and nano-structures play critical roles in microelectronic, microelectromechanical, micromagnetic and microphotonic devices and systems, defining their performance and reliability. New envisioned applications also include metallic nano-particle arrays in new computing devices, energy harvesting, and biosensing applications. In all of these applications, stringent engineering control of the structure and properties is required. In all of the planned work, our goal will be to develop fundamental understandings that will have impact on engineering practices leading to improved control of metal films and nano-structures for applications. Two graduate research assistants will be directly supported by the program. In addition, the PI's research group of graduate and undergraduate students is now, and has historically been structured, to directly exploit new insights derived from NSF research in applications, especially in microelectronic and micromechanical devices and systems. Much of this research is directly supported by industry, and reported to industry. The research involves collaborations both inside and outside MIT, and results will be reported, as they have historically been, in both scientific and engineering venues. Research results and summaries are made available via the internet, and are included in MIT courses for undergraduate and graduate students, as well as professional engineers, the latter in the form of short courses at MIT and on-site in companies.

TECHNICAL SUMMARY Polycrystalline films are used in a wide range of micro- and nano-scale devices and systems in which their stress state and grain structure profoundly affect their performance, properties and reliability. Stress and structure evolution during film formation and subsequent processing are known to be strongly coupled. In-situ measurements of stress evolution have revealed a complex range of phenomenology, including evolution between tensile and compressive stress states, and apparently reversible stress changes during growth interruptions. Both ex-situ and in-situ characterization of grain structures and surface topography have also revealed complex processes that include grain growth and texture evolution during and after island coalescence, network formation during island coalescence, retention of deep trenches at grain boundaries, and changes of surface topography during and after film island coalescence. This phenomenology is not accounted for in current understandings of the linkage between stress and structure evolution. In this program, stress evolution will be studied during deposition of a range of materials under a range of conditions. Use of different materials and deposition temperatures will allow observations of behavior that are intermediate to those that have been the focus of prior studies. Use of materials with higher melting temperatures will also allow quenching of surface and grain structures for ex-situ characterization. Surface topography will be monitored during deposition, and during interruptions of depositions, using light scattering techniques. The angle of incidence of the atomic deposition flux will also be varied in order to controllably promote different levels of surface roughness. The size and spacing of the initial islands from which films are formed will also be varied using templated dewetting techniques. Through these studies, understandings will be developed that will allow application-specific engineering of structures and properties of polycrystalline thin films and the micro/nanostructures patterned from them.

NON-TECHNICAL SUMMARY The electrical devices and integrated circuits that power our cell phones, computers, and the Internet are created using very thin layers of metals and semiconductors that are formed on flat surfaces of materials like silicon. Extremely small patterns are then made in these films to create millions of devices that are connected to make circuits of devices. Sometimes the devices have parts that physically move as they work. Layers, or films, of different materials are usually made by essentially spraying atoms onto a flat surface. Once the atoms arrive on the surface, they move around to form small crystals and these crystals grow until they run into each other to form a continuous film. The way these crystals form and grow can vary tremendously, depending on how the atoms are sprayed and on which atoms are used. These variations strongly affect the properties of a film, including how easily electrons can move through them or how easily the moving parts can be moved. This makes it very difficult to make complex integrated circuits and limits what can be made. This research program focuses on understanding why these variations occur and on how they can be controlled, so that new devices and circuits can be made. To do this, the researchers will change the way the atoms are sprayed and measure how the properties and structure of films change, while they are being formed.

NON-TECHNICAL ABSTRACT:
Metallic thin films are used in a wide range of devices and systems that have an impact on our everyday lives. They play critical roles in the integrated circuits used for computation and communication, microelectromechanical systems (MEMS) used for sensing and biomedical analyses, micromagnetic devices used for information storage, and microphotonic devices and systems used for information processing, communications and sensing. As these technologies advance, smaller and smaller metallic components are required. However, it has been found that when materials are made very small, their shape tends to evolve over time as they try to adopt spherical shapes, like droplets. This is limiting the development of new technologies, especially those involving metallic components. In this project, precisely controlled very small metallic structures are being made to study their evolution over time. These experimental studies are coupled with development of the theoretical models that are needed to explain this evolution. The goals of this project are to develop new techniques for making stable nano-scale metallic structures and for controlling shape evolution to make structures with complex shapes for new functions. This project involves students from two research groups, one focused on experiments and one focused on modeling. These students participate in meetings of both groups and will also extensively interact with senior and junior members of collaborating groups in the Technical University of Dresden, the University of Milano, and the University of New South Wales. Results from this project are included in courses at MIT and short courses for industry, as well as in massively open online courses.

TECHNICAL ABSTRACT:
Experimental and theoretical studies of solid state dewetting are being carried out to understand the effects of crystalline anisotropy on capillary-driven morphological evolution. Single crystal films have been lithographically patterned before heating to cause morphological evolution. It is found that this evolution is strongly affected by the crystallographic orientation of patterned features such as film edges. Edges were found to retract at orientation-dependent rates and either undergo pinch-off to leave behind sets of ligaments aligned in parallel with the retracting edge, or develop a fingering instability that leads to parallel ligaments aligned along the retraction direction. Ligaments are subject to a Rayleigh-like instability that leads to break-up into particles. This behavior is very reproducible, and leads to different intermediate structures that depend on the shape and orientation of the pre-patterned structures. In all cases, crystalline anisotropy strongly affects the observed phenomenology. In this project, systematic studies are underway in which films of different materials, thickness, and crystallographic texture are patterned with edges and other features within a range of in-plane crystallographic orientations. Kinetic studies of retraction of straight edges, rim pinch-off and fingering are underway. Different annealing ambients are being used to understand the role of surface structure and surface energy anisotropy on dewetting anisotropy. Morphological evolution in the systems under study occurs by capillarity-driven surface diffusion. 2D models for evolution in the case of isotropic surface energies are well developed, and 3D models based on phase filled approaches are emerging. However, the strongly anisotropic behavior that is observed in dewetting studies shows the need for 3D models that account for surface energy and diffusion anisotropy. As part of this project, the investigators are developing anisotropic 3D phase field models that will be tested by comparison with a wide range of experiments. It is anticipated that these basic studies will lead to an improved understanding of capillary-driven evolution of thin films and micro-/nano-structures. This will allow design of materials and systems with improved stability and enable the use of templated solid-state dewetting as a tool for generating complex structures with sub-lithographic feature sizes.

When electrons move through wires, they are scattered by vibrating atoms and wire imperfections. This scattering is the source of the electrical resistance and results in power consumption. For the wires used in modern electronics (interconnects) this is already a bottleneck to computing performance and is worsening as the wires (along with the transistors) are made smaller. The aim of this program is to make the wires so small (<< 10-nm in width and height) and so structurally perfect that quantum size effects arise and the electrons can travel in a ballistic fashion without scattering. This can result in orders of magnitude improvements in resistance and computing energy efficiency and enable a revolution in electronics. In addition to the societal benefit of improved computing, the program will support education and research at the undergraduate, graduate and post-doctoral levels at four institutions, Columbia University, Massachusetts Institute of Technology, Rensselaer Polytechnic Institute and the University of Central Florida. The outreach effort will include: (1) The Harlem Schools Partnership (HSP) for STEM education at CU, (2) MIT's Materials Day for industry outreach, (3) the Discovery Engineering program for high-school girls at RPI, and (4) the techCAMP "Future of Information" at UCF.

To achieve its goal of ballistic conduction in metallic nanowires, the project will include the preparation and atomic scale characterization of single crystal metallic films and lines as well as experimental measurement of electron transport behavior, with ruthenium as the metal of choice for the initial studies. The stability of the metallic lines will also be investigated, since this is critical to the reliability of interconnects in computing systems. A theoretical and computational physics modeling effort will aim to provide a quantitative understanding of ballistic transport in the size and defect limits, and serve to identify preferred metals for ballistic conductance of interconnects for future efforts. The computational models and codes developed will be made available on the Nanohub (https://nanohub.org/). The project will additionally provide a proof-of-principle demonstration of how the proposed metallic conductors can be fabricated for technology implementation by the semiconductor industry.