MEAM Seminar Series Archive: Spring 2007

January 18
Jianmin Qu
"An electrochemomechanical theory of point defects in ionic solids"

Speaker: Jianmin Qu
Professor, G.W. Woodruff School of Mechanical Engineering
Georgia Institute of Technology

Charged defects diffuse through an ionic solid under electrochemical driving forces. Such a diffusion process can be affected by mechanical stresses in the solid. A deviation of defect concentration from its stoichiometric value during diffusion can cause volumetric strains in the solid. Such strains will result in mechanical stresses if the ionic solid is under mechanical constraint, or if the defect distribution is non uniform. We develop a framework to account for the coupling between mechanical stresses and diffusion of charged defects in ionic solids. The framework consists of a system of nonlinear differential/algebraic equations governing the defect concentrations, electrostatic potential and the mechanical stresses. As an example, the coupled theory is used to study the electrochemical and mechanical behavior of an electrolyte made of gadolinia-doped ceria in a planar solid oxide fuel cell. It is found that the interaction between mechanical stresses and electrochemical activities is rather significant, and cannot be neglected.

January 25
Chuan-Hua Chen
"Physicochemical hydrodynamics in micro and nano-systems"

Speaker: Chuan-Hua Chen
Teledyne Scientific Company
Assistant Professor, Duke University

Physicochemical hydrodynamics is an indispensable subject for many branches of micro and nanoengineering. In this talk, I will discuss the fundamental challenges of this multi-scale, multi-physical subject with complex dynamics, and show a few applications of physicochemical hydrodynamics in the development of micro and nanosystems.

  • To illustrate the multi-scale nature, I will present a lotus-mimicking superhydrophobic structure with carbon nanotubes deposited on micromachined texture, and demonstrate the first continuous dropwise condensation with engineered superhydrophobicity. The dynamics of such superhydrophobic condensation spans at least six orders of magnitude in both space and time.
  • To illustrate the multi-physical nature, I will present an electrohydrodynamic drop-and-place technique capable of deploying droplets at sub-micron resolution. This technique is enabled by the cone-jet transition with a unique capability of producing nanometer jets. The cone-jet configuration results from the interplay of hydrodynamic and capillary forces under electric fields.
  • To illustrate the complex dynamics, I will present an electrokinetic instability frequently encountered in bioanalytical microfluidics. I will discuss the multiphysical mechanism, analytical prediction, and experimental confirmation of convective and absolute instabilities. This electrokinetic instability exemplifies the complexity of physicochemical hydrodynamic systems at small scales.

I will close the talk with several areas of future research as well as potential applications. For instance, the superhydrophobic structures can be used to improve heat and water management in energy systems; and the electrohydrodynamic systems can be applied to enable nanosecond and nanometer resolutions for biomedical devices. Throughout the talk, I will show how fundamentals and applications can motivate and support each other, and how theory and experiment can be integrated for interdisciplinary research.

January 29
John Hutchinson
Tedori-Callinan Lecture: "Recent developments in thin film mechanics"

Learn more about the lecture here!

February 8
Jianping Fu, PhD Candidate
"Nanofluidic devices for rapid biomolecule analysis"

Speaker: Jianping Fu
PhD Candidate
Department of Mechanical Engineering
Massachusetts Institute of Technology

Direct analysis of biologically-relevant entities such as nucleic acids and proteins offers the potential to outperform conventional analysis techniques and diagnostic methods through enhancements in speed, accuracy, and sensitivity. Nanofluidic systems with critical dimensions comparable to the molecular scale open up new possibilities for direct observation, manipulation and analysis of biomolecules (single or ensemble), thus providing a novel basis for ultra-sensitive and high-resolution sensors and medical diagnostic systems. Inspired by this concept, we have developed a new class of nanofluidic filter devices and have implemented them as controllable molecular sieves for rapid analytical separation of various physiologically-relevant molecules such as proteins. In addition, we have conducted theoretical studies of molecular sieving process in the context of periodic free-energy landscapes created by the patterned nanofluidic filter arrays. The kinetic model constructed based upon the equilibrium partitioning theory and the Kramers rate theory properly describes the field-dependent sieving behavior, presenting notable progress beyond the existing equilibrium model in conventional gels. At the end of the talk, I will introduce a microfabricated anisotropic sieving structure consisting of a two-dimensional periodic nanofluidic filter array (anisotropic nanofilter array, ANA). The designed structural anisotropy in the ANA causes different-sized biomolecules to follow distinct migration trajectories, leading to efficient continuous-flow separation. Continuous-flow separation of dsDNA and proteins covering broad biological size ranges were achieved within a few minutes, thus demonstrating the potential of the ANA as a generic molecular sieving structure for an integrated biomolecule sample preparation and analysis system.

February 15
Paulo E. Arratia
"Complex fluids in microfluidic systems"

Speaker: Paulo E. Arratia
Postdoctoral Researcher, Department of Physics & Astronomy
University of Pennsylvania

Polymeric solutions often possess complex rheological behavior such as shear thinning viscosity and viscoelasticity. In this talk, the effects of viscoelasticity and polymer molecular weight on filament thinning and breakup are investigated in a microchannel cross flow. When a viscous solution is stretched by an external immiscible fluid, a low polymer concentration (e.g. 100 ppm) strongly affects the breakup process, and this is compared to a Newtonian case of same shear viscosity. At late times, when viscoelastic stresses become important, polymer filaments show much slower evolution, different morphology featuring multiple connected drops, and different scaling with the ratio of flow rates. As the polymer molecular weight decreases, the filament thinning dynamics approach that of a Newtonian fluid. The filament thinning process can be described in terms of extensional viscosities of the two immiscible fluids, which for the polymeric solutions include strain hardening. These results show that microfluidic systems hold great promise in measuring the rheological properties of complex fluids, particularly those where sample volume is limited such as biological fluids.

February 22
Ongi Englander
"Localized synthesis, assembly and integration of one-dimensional nanostructures"

Speaker: Ongi Englander
University of California, Berkeley

The development of processes which enable the integration and assembly of nanostructures into systems and devices remains a significant challenge as we seek to fully realize the potential of nanotechnology. Current research goals are focused on developing high-yield, parallel and scalable manufacturing techniques for integration and assembly within the nanoscale as well as across length scales.

Here, I will present a technique for the direct integration of one-dimensional nanostructures with MEMS structures. In this approach, a top-down microfabrication process is followed by a bottom-up nanofabrication process. Specifically, simple MEMS structures are utilized as microscale platforms for the site-specific, bottom-up synthesis of one-dimensional nanostructures. Resistive heating of the MEMS structures is used to initiate and sustain the catalyst assisted synthesis of one-dimensional nanostructures. This approach confines the high temperature region permitting only localized nanostructure growth thus allowing the surroundings to remain at room temperature and enabling a CMOS compatible process. The nano-to-micro contact is established in-situ and therefore eliminates the need for any post-process contact formation steps.

We have applied this approach to realize the direct integration of both, silicon nanowires and carbon nanotubes with the MEMS structures. Furthermore, we have demonstrated self-assembled, two-terminal, micro-to-nano systems achieved as localized nanostructure growth links together multiple MEMS structures. We have also shown that the application of a localized electric-field during the synthesis and assembly process can improve nanostructure organization and alignment in the resulting system. Finally, the utilization of this self-assembled system in a proof-of-concept sensing application has been demonstrated.

February 26
Mark O. Robbins
"Contact and friction: Connecting atomic interactions to macroscopic behavior"

Speaker: Mark O. Robbins
Professor of Physics and Astronomy
Johns Hopkins University

Friction affects many aspects of everyday life and has played a central role in technology dating from the creation of fire when sticks were rubbed together to current efforts to make nanodevices with moving parts. The friction "laws" we teach today date from empirical relationships observed by da Vinci and Amontons centuries ago. However, the microscopic origins of these laws have remained unclear, because friction is a complex multiscale phenomenon that depends on both atomic interactions in contacts, and the macroscopic elastic and plastic deformation that determine the morphology and stress distribution within these contacts. New experimental and computational work has begun to unravel this complex interplay. The talk will begin with continuum studies of contact between elastic and plastic surfaces with self-affine surface roughness. We find a power law distribution of contact sizes, and a broad distribution of local pressures. The mean contact size is always comparable to the small scale cutoff in the roughness, which is often in the nanometer range. The talk will next describe tests of continuum contact mechanics in contacts whose width is a few to a few tens of molecular diameters. The results imply that the continuum analysis commonly applied to atomic force microscope measurements can give accurate bulk elastic moduli, but that contact areas and yield stresses may be off by a factor of two and shear moduli may be off by an order of magnitude. These same discrepancies are found between continuum theory and atomistic studies of self-affine surfaces using a new multiscale hybrid method to treat elastic deformations below the surface. The above results show that the area of contact between surfaces depends on many factors that do not influence measured friction forces. The talk will conclude by showing that a simple mechanism based on the presence of debris between surfaces naturally explains Amontons' laws and common exceptions to them.

March 1
Thomas Cubaud
"Dynamics of viscous threads in microfluidics"

Speaker: Thomas Cubaud
Postdoctoral Researcher, Department of Chemistry & Biochemistry
University of California, Los Angeles

The manipulation of viscous material at the microscale is a key challenge for implementing lab on chips with the ability to manage a variety of complex and reactive fluids. I will show some recent experimental results on two-fluid flows in microchannels. For large viscosity contrasts, the more viscous liquid can detach from the walls and form a viscous thread, which is lubricated by the less viscous liquid near the walls. This lubrication property allows for transporting viscous material in microsystems with moderate dissipation. Viscous threads display a wealth of behaviors depending upon flow geometries, fluid properties, and flow rates. In this talk, I will focus on two hydrodynamic instabilities between miscible liquids. First, the folding instability of a viscous thread that flows in a diverging microchannel is investigated. This periodic phenomenon provides a natural stretch-and-fold mechanism that can be applied to enhance mixing between liquids having different viscosities. Second, the shear-induced destabilization of a thin thread that flows off-center in a square microchannel is examined. This instability can cause the thread to fragment and form an array of individual viscous swirls, the miscible counterparts of droplets. Examples of flow stratifications will also be given with emphasis on “liquid walls” that allow for the deformation and the break-up of droplets and bubbles. This study shows new methods for controlling interfaces both between miscible and non-miscible fluids in microsystems.

March 15
Pramod Sangi Reddy, PhD Candidate
"Thermoelectric effects in metal-molecule junctions: Application to energy conversion"

Speaker: Pramod Sangi Reddy
PhD Candidate
University of California, Berkeley

Thermoelectric materials are used to build solid state energy conversion devices. However, their widespread use is limited by their efficiency and cost. We propose a new approach based on metal-organic molecule aggregates, which could potentially achieve high efficiencies and low cost. I will describe our experiments where we measured electrical conductance of single molecules and have also demonstrated - for the first time - thermoelectricity in single organic molecules. Finally, I will describe strategies that we are pursuing to tune the thermoelectric properties of single molecules and to build novel materials based on organic molecules.

March 22
Jennifer R. Lukes
"Thermal transport at nanostructure interfaces"

Speaker: Jennifer R. Lukes
William K. Gemmill Term Assistant Professor
Mechanical Engineering and Applied Mechanics, University of Pennsylvania

Thermal transport in nanotubes, nanowires, ultrathin films, and other structures with nanometer-scale characteristic dimensions differs dramatically from that in bulk materials. Nanocomposite ‘metamaterials’ formed of these nanostructure building blocks have recently gained much attention for their potential to exploit these unusual properties and achieve the extreme values of thermal conductivity demanded by many current and future applications. For example, ultrahigh thermal conductivity materials are required to meet the escalating heat sinking demands of next-generation high power density electronics, and ultralow thermal conductivity materials are required to make thermoelectric alternative energy systems competitive with conventional energy systems. A crucial consideration when building such ‘materials-by-design’ is the transport of thermal energy across the interfaces between the constituent nanostructures and their surroundings. Presented here is recent molecular dynamics simulation work that illustrates how interfaces can be tailored to modify thermal transport in such materials. Specifically, thermal transport between carbon nanotubes, around nanoparticles embedded in a host material, and within superlattices will be discussed. Key results observed include a four order of magnitude decrease in nanotube-nanotube thermal resistance as nanotube spacing decreases, a distortion of phonon scattering arising from host material crystal structure, and a clear thermal conductivity minimum in lattice matched superlattices. Additionally, other ongoing computational and experimental projects will be briefly discussed.

March 26
Youssef Marzouk
"Uncertainty and Bayesian inference in inverse problems"

Speaker: Youssef Marzouk
Sandia National Laboratories

Bayesian statistics provides a foundation for inference from noisy data and stochastic forward models, a natural mechanism for regularization in the form of prior information, and a quantitative assessment of uncertainty in the inferred results. Inverse problems—representing indirect estimation of model parameters, inputs, or structural components—can be fruitfully cast in this framework. Complex and computationally intensive forward models arising in physical applications, however, can render a Bayesian approach prohibitive. This difficulty is compounded by high-dimensional model spaces, as when the unknown is a spatiotemporal field.

We present new algorithmic developments for Bayesian inference in this context, showing strong connections with the forward propagation of uncertainty. In particular, we introduce a stochastic spectral formulation that dramatically accelerates the Bayesian solution of inverse problems via rapid evaluation of a surrogate posterior. We also pursue dimensionality reduction for the inference of spatiotemporal fields, using truncated spectral representations of Gaussian process priors. These new approaches are demonstrated on scalar transport problems arising in contaminant source inversion and in the inference of inhomogeneous material or transport properties. We also discuss extensions toward the inference of chemical models and reaction networks.

April 5
Igor Pivkin
"Multiscale modeling of biological flows"

Speaker: Igor Pivkin
Division of Applied Mathematics
Brown University

Petaflops computing can play a catalytic role in advancing systems biology at all scales. However, new approaches are required for simulations of the multiscale phenomena encountered in biological systems. In this talk, we will discuss two multiscale modeling approaches for continuum (PDE-based) and atomistic (dissipative particle dynamics) modeling with specific applications to the human vascular system. Examples of simulations of arterial flows and blood cells will be presented.

April 10
Vikranth Racherla, PhD Candidate
"Non-associated plastic flow and effects on macroscopic failure mechanisms"

Speaker: Vikranth Racherla
PhD Candidate
Advisor: John L. Bassani
Mechanical Engineering and Applied Mechanics, University of Pennsylvania

This dissertation focuses on the implications of non-associated flow on plastic deformation The physical basis of conventional macroscopic plasticity theories for crystalline solids is a simple slip mechanism, one that is controlled only by the shear stress on the slip plane in the direction of slip. This is somewhat, exceptional behavior since in all but FCC lattices the core structures of screw dislocations tend to be three dimensional (non-planar) and, as a result, slip depends upon so-called non-glide stresses. The effects of non-glide stresses result in non-associative flow behavior in both single and polycrystals. Frictional materials such as rock and sand are typically modeled as non-associated flow as well.

In this work the effects of non-glide stresses are shown to persist at macroscopic scales and strongly affect deformation behaviors. For example, the critical pressure at which cavitational instabilities occur and critical necking strains are significantly affected by non-associated flow. The structure of constitutive models is studied in detail, and new relations are proposed. In classical, rate-independent non-associated flow models uniqueness and stability of solutions to incremental boundary value problems can be lost even at small strains. However, in a corner theory, a deformation theory, or a rate-dependent theory we demonstrate that uniqueness and stability can be guaranteed at small deformations. Non-associated flow is shown to significantly alter strain localization in thin sheets (sheet necking), and this class of problems is studied in detail. The effects are most prominent near the plane strain loading state. To investigate the full three-dimensional nature of instabilities in sheet deformation, a finite element analysis is carried out for nearly rate-insensitive response using an implicit dynamics scheme. This led to the discovery of “strain bursts” as a consequence of non-associated flow, particularly for deformations near the plane strain loading state. Finally, preliminary investigations of effects of non-associated flow on crack tip fields (for plane strain conditions and Mode I loading) reveal that the crack tip fields are of the HRR type, and their amplitudes are slightly affected by non-associated flow.

April 16
Ulrich Hetmaniuk
"Eigenspace computations in linear structural dynamics"

Speaker: Ulrich Hetmaniuk
Post-Doctoral Researcher
Sandia National Laboratories

Eigenvalue analysis is common in many areas of engineering. For example, the knowledge of the eigenspectrum of a linear structure allows an analyst to decide whether an excitation frequency will be close to a resonance frequency, which could cause vibrations of large amplitude. The eigenpairs of a linear structure can also determine efficiently, in a linear superposition procedure, its transient or frequency response. For large-scale heterogeneous structures, where the finite element models reach ten millions or more degrees of freedom, researchers at Sandia National Labs frequently compute thousands of eigenmodes. For these analyses, efficient algorithms are critical. The talk will review recent work on eigenspace computations in linear structural dynamics and highlight several open challenges.

The first part will present state-of-the art iterative eigensolvers to compute the smallest eigenpairs for sparse symmetric matrices. Examples include the shift-invert Lanczos method and preconditioned eigensolvers.

We will illustrate their performance on practical engineering problems of over a million unknowns and when a large number of eigenpairs are to be computed. Alternatives in computing the eigenspace include component mode synthesis techniques and multigrid methods. We will comment on both approaches and present a new multigrid algorithm that uses only the algebraic information from the matrices.

Finally, we will introduce an explicit a posteriori error estimator for eigenvalue analysis of heterogeneous elastic structures. This estimator treats the cases of high order finite elements and discontinuous material coefficients and is equivalent to the error in the eigenvector, independently of the variation of the materials properties. We will assess the efficiency of this estimator with numerical experiments.

April 19
Joshua Lampe, PhD Candidate
"Interfacial characteristics of a gas bubble immersed in a surfactant and protein laden fluid: Experimentation and modeling"

Speaker: Joshua Lampe
PhD Candidate
Advisor: P.S. Ayyaswamy
Mechanical Engineering and Applied Mechanics, University of Pennsylvania

In the United States, approximately 650,000* open heart surgeries are completed each year. During cardiopulmonary bypass, an otherwise closed system, is opened, and joined with an extracorporeal fluid circuit, which will manage blood flow and oxygenation while the patient is in surgery. One of the major complications which can arise while undergoing cardiopulmonary bypass is the inadvertent introduction of an air bubble, or air embolism. The presence of an air bubble in the circulatory system often leads to one of several possible negative outcomes including stroke, heart attack, and death.

The physics of the gas-liquid interface play a significant role in determining the interactions between the bubble, the blood, and the vessel wall. Blood contains a variety of surface active molecules, which will adsorb to the bubble interface reducing the surface tension. Surface tension plays a significant role in the fluid mechanics of bubble motion. The balance between surface tension forces and viscous and inertial forces will dictate the bubble shape, and potential surface tension gradients will induce Marangoni flows, retarding bubble motion. Protein adsorption to the interface also plays a significant role in a variety of physiologic responses such as bubble adhesion, blood vessel occlusion, or the activation of thrombogenic or inflammatory pathways. Given the adverse effects linked to protein adsorption to the bubble interface, it is desirable to comprehensively understand the physics dominating protein adsorption as well as to investigate methods to mitigate or retard the protein adsorption process. In earlier studies, the presence of certain non-toxic surfactants has been shown to result in better outcomes and to reduce the adhesion force between a bubble and the blood vessel wall. Given that surfactant molecules are relatively small and that they experience a higher free energy gain through adsorption, it seems likely that surfactant molecules diffuse to the interface more rapidly and this may inhibit protein adsorption. The purpose of this investigation was to determine the adsorption dynamics of model blood borne proteins, bovine serum albumin and bovine fibrinogen, onto the surface of a gas bubble and investigate competition in surface adsorption between these proteins and the non-toxic surfactant Pluronic F-127™. Surface tension data, initial data from a new confocal microscopy surface concentration measurement technique and new adsorption models will be presented.

*Statistics from 2004 American Heart Association

May 3
Ani (Mong-ying) Hsieh, PhD Candidate
"Towards the development of robotic swarms for real world applications"

Speaker: Ani (Mong-ying) Hsieh
PhD Candidate
Advisor: Vijay Kumar
Mechanical Engineering and Applied Mechanics, University of Pennsylvania

The continuing drop of the price-to-performance ratio of embedded systems, wireless networking, and high-speed processors in recent years has fueled the growing interest in multi-robot systems. As the number of robots increase, it is useful to think in terms of a swarming paradigm where capabilities are expressed by populations rather than super-capable individuals, as seen in the group dynamics of biological swarms. This is especially relevant in complex tasks where it is difficult to provide robots with specific instructions a priori due to incomplete knowledge of the environment, e.g. interplanetary exploration, or where mission criteria require capabilities that are varied in both quantity and difficulty, e.g. urban search and rescue. In this talk, I will present some of the challenges towards the development and deployment of actual robotic swarms for real world applications. Specifically, I consider the effects of communication on teams of robots and propose methods to maintain the integrity of the communication network. I will show how reactive controllers can be used to maintain end-to-end communication links and consider the differences between monitoring point-to-point signal strength versus data throughput. I will further demonstrate how these controllers, combined with more deliberative planning, can be used to coordinate a robot swarm to generate a two-dimensional geometric pattern while maintaining specified relative distance constraints. This is relevant for applications such as perimeter surveillance, environmental boundary tracking, and cooperative manipulation.

May 4
Itzhak Green
"On the contact mechanics of rough surfaces"

Speaker: Itzhak Green
Professor, Mechanical Engineering Department
Georgia Institute of Technology

Since in reality all engineering surfaces are rough to some degree, the modeling of contact between rough surfaces is very important. It leads to an improved understanding of the friction, wear, thermal and electrical conductance between real surfaces. When two rough surfaces are pressed together, only the peaks or asperities on the surface will be in contact, and thus they carry very high loads. Such high loads lead often to plasticity and purely elastic contact models of rough surfaces (such as the landmark work by Greenwood and Williamson (1966)) are not always adequate. This presentation will review existing contact models, address the issue of "hardness," highlight new findings in elasto-plastic spherical and cylindrical contact under normal loading (typical to RF switches and MEMS devices) including residual stresses under repeated contact, and discuss the challenges of modeling sliding junctions and surface modification.

June 6
Shamik Sen, PhD Candidate
"The interplay of cell tension, elasticity, and adhesion:  From the simple red cell membrane to differentiation and signaling muscle cells"

Speaker: Shamik Sen
PhD Candidate
Advisor: Dennis Discher
Mechanical Engineering and Applied Mechanics, University of Pennsylvania

Cell adhesion is central to cellular processes ranging from motility to differentiation and is increasingly understood to depend on an interplay of membrane and/or cell tension, and elasticity of the cell and/or substrate. The focus of this thesis has been to add experimental and computational insights into the role of these factors in basic biological processes ranging from passive adhesion of a simple cell membrane to contractility-modulated cell attachment in sensing, signaling and differentiation. Red blood cells simply consist of a membrane, but it is shown that passive adhesion induces tension that not only alters how it is probed by the AFM but also influence molecular attachment dynamics. For many if not most other tissue cells, inwardly directed tension is generated predominantly by the cell itself through myosin-based contractile mechanisms, and yet cell survival and development depends on sufficient adhesion to withstand self-detachment. Integrin based adhesions are central, but many cells also adhere to a matrix of basement membrane through a dystroglycan complex, which is missing or perturbed in several muscular dystrophies (MD). Integrin upregulation appears to be a natural compensatory process in such diseases, and it is also shown that upregulation of a number of additional focal adhesion and signaling proteins, especially paxillin, ultimately generates a stiffer, hypercontratile or hypertensive state in these diseased cells. Forced overexpression of paxillin in skeletal muscle cells leads to a similar phenotype and is accompanied by robust ERK/MAPK activation. Stretch activation of this same pathway is also observed in both normal and dystrophic muscle, but it is surprisingly absent under osmotically induced stress, which highlights signaling differences when stresses are imposed normal to the membrane or parallel to it (in stretch). Under normal stress, adhesive signaling diverges between normal and dystrophic myotubes, with enhanced recruitment of vinculin and paxillin observed in the latter. To gain some insight into the interplay between contractility and attachment with extension to elastic substrates, a finite element model is developed in which adherent cells are modeled as a prestressed and firmly attached hyperelastic solid. One key question addressed is the enhanced spreading behavior observed on soft matrices compared to stiff matrices, which ties into stiffness sensitive differentiation of muscle cells and stem cells. The model shows large deviations in mean interfacial strain between matrix and cells on soft matrices, indicating that cells feel on the length scale of their adhesions. The collection of results here ultimately adds to our understanding of the important interplay of cell or membrane tension with adhesion and its many downstream effects.