Session A7 - MEMS, Nano-scale Physics Applications.
ORAL session, Monday morning, March 12
Room 609, Washington State Convention Center

[A7.001] Issues in Magnetic Resonance Force Microscopy at MilliKelvin Temperatures

Magnetic resonance force microscopy (MRFM) makes use of an ultrasensitive cantilever as a force detector capable of detecting attoNewton scale forces. In principle, this approach could allow for detecting the force from a single electron or nuclear spin. The force resolution is ultimately limited by the thermal mechanical noise in the cantilever, which scales in power with temperature. This fact provides a strong incentive to operate at the lowest practical temperature. Some of the challenges of performing MRFM in the milliKelvin regime will be discussed, in particular devising a detection scheme that results in minimal heating of the cantilever. We have made several improvements to a fiber-optic interferometer that allow operation with incident powers below 10 nW , or 20 times lower than previously used. Other issues in combining MRFM with a dilution refrigerator have been investigated, including adding a damped spring system to deal with external vibrations, and generating the GHz frequency magnetic fields necessary for magnetic resonance.

[A7.002] A Novel MEMS Pressure Sensor Integrated with an Optical Fiber

We present the design, fabrication, and characterization of an optically interrogated MEMS pressure sensor fabricated directly on an optical fiber. The sensor configuration involves anodically bonding of an ultra-thin (7 µm) piece of silicon onto the fiber end face over a cavity etched in the central portion of this end face. The silicon diaphragm and the cavity-fiber interface act as reflectors forming a Fabry-Perot interferometer. Final device diameter is thus the same as that of the optical fiber. We have employed both 200 and 400 m diameter multimode optical fibers. The micromachining procedure includes photolithographic patterning, wet etching of a cavity, and anodic bonding of a silicon diaphragm. A pressure sensor fabricated on an optical fiber has been tested displaying an approximately linear response to static pressure (0-80 psi) with a sensitivity of 0.1 mV/psi. This sensor is expected to find application in situations where small size is advantageous and where dense arrays may be useful.

[A7.003] Metastability and the Casimir Effect in Micromechanical Systems

Electrostatic and Casimir interactions limit the range of positional stability of electrostatically-actuated or capacitively-coupled mechanical devices. We investigate this range experimentally for a generic system consisting of a doubly-clamped Au suspended beam, capacitively-coupled to an adjacent stationary electrode. The mechanical properties of the beam, both in the linear and nonlinear regimes, are monitored as the attractive forces are increased to the point of instability. There "pull-in" occurs, resulting in permanent adhesion between the electrodes. This indicates that the free state of the system is merely metastable, and that the state of contact (after pull-in) has lower energy due to the strongly attractive Casimir interaction. We investigate, experimentally and theoretically, the position-dependent lifetimes of the free state (existing prior to pull-in). We find that the data cannot be accounted for by simple theory; the discrepancy may be reflective of internal structural instabilities within the metal electrodes.

[A7.004] MODELING AND VISUALIZATION OF MICROCRACKS DETECTION BY MAGNETIC TUNNEL JUNCTION DETECTORS

The presence of cracks, phase segregation, or even submicron-sized grain boundaries creates a disruption of the magnetic field response to an externally applied electrical current running through the material. We have considered disruptions of magnetic field in the external near-surface region caused by number of typical microcracks and flaws, of different dimensions and orientations, within the bulk of material. These disruptions can be mapped by an array of magnetic tunnel junctions that we are fabricating. To implement three-dimensional visualization, using a computer model of this array, magnetic "signatures" of flaws have been calculated using finite element analysis. The database of "signatures" thus generated allows fast recognition of faults and generation of their images in real time. Significant efforts have been made to provide an adequate three-dimensional visualization of the shape and distribution of microcracks, the magnetic field lines, and delineation of the surface of the faults in relationship to the component surface. Images can be viewed by using inexpensive stereo-ready graphics cards for PC's, at an ImmersaDesk, and at a portable CAVE-like system driven by clusters of PCs.

[A7.005] Quantum Mechanical Actuation of Microelectromechanical systems by the Casimir force

MicroElectroMechanical Systems (MEMS) have become vital in sensing and actuating applications. With further miniaturization, quantum effects may become significant in MEMS design and operation. The Casimir force, for example, arises from an alteration of the zero-point electromagnetic energy due to the boundary conditions imposed by two uncharged metallic surfaces. We demonstrate the Casimir effect in MEMS using a high sensitivity micromachined torque device. The device consists of a 3.5 um thick, 500 um square metallized polysilicon plate suspended on two of its opposite sides by thin torsional rods above two fixed electrodes. A 200 um metallized sphere is positioned close to one side of the torque sensor. Attraction between the sphere and the top plate results in a torque that rotates the plate about the torsional rods. Differential capacitive technique is used to detect the tilt angle with a 0.1 micro-radian resolution. The measured tilt angle is in agreement with theoretical calculations of the Casimir force taken into account the finite conductivity of the metal films.

[A7.006] High-frequency Nanomechanical Structures in Silicon

Nanoelectromechanical systems (NEMS) are of interest from both scientific and technological standpoints. Such structures are being considered for use as chemical and biological sensors, force gauges and frequency filters. Small resonant structures also open avenues for mesoscopic studies of the mechanical properties of materials. One of the obstacles for practical applications are intrinsic losses which lower the mechanical quality factor of these devices. We study high resonant frequency structures with lateral dimensions as small as 50 nm, while trying to understand sources of dissipation on this size scale. We have recently reported the fabrication and excitation of single suspended wires with resonant frequencies as high as 380 MHz. We are currently working on more sensitive detection schemes to detect the motion of devices with resonant frequencies above 1 GHz. Silicon and silicon nitride mechanical resonators have been studied in our group for material effects on dissipation. We are also focusing on surface treatments and the effects of device geometry on dissipation. We are also studying the effects of various levels of doping in single-crystal silicon on dissipation and driving schemes, a study significant for industrial use in integration with electronic devices. The study of the dynamics of these structures has also uncovered some very important phenomena such as parametric amplification and tunability. The studies of clamping losses, mechanical isolation and induced stresses in materials are underway. Interaction with light is also of interest.

[A7.007] Low Temperature Mechanical Dissipation in Ultrathin Single Crystal Silicon Cantilevers

Ultrathin single-crystal silicon cantilevers have been used to demonstrate attonewton force resolution at 4K. The force resolution of these cantilevers is limited by thermomechanical noise, the analog of Johnson noise in a resistor. Measurements have indicated that the mechanical energy dissipation in ultrathin silicon cantilevers has a strong dependence on temperature. Initial measurements of the dissipation in an ultrathin silicon cantilever with areas of high levels of boron doping show a broad dissipation peak centered at about 135K and a sharp dissipation peak centered at 13K. Thus, experiments operated near these temperatures would suffer from reduced force resolution. We present results from an ongoing study to determine the physical nature of these dissipation peaks by investigating the dissipation as a function of the material and doping properties of the cantilevers.

[A7.008] Manipulation of magnetic particles using micro-electromagnets

We have fabricated micro-electromagnet devices (\muEMs) for controlling magnetic nano particles on a chip. The \muEMs consist of multiple lithographically patterned layers of \mum-scale Au wires[1] separated by transparent insulators on sapphire substrates. Magnetic field produced by these wires creates local field maximum on a plane where magnetic particles are guided. The \muEMs are designed to control the motion of Fe_3O_4 magnetic nano particles as the individual as well as a group motion of these particles is studied as the field is turned on. Actual devices fabricated include a single trap, an array of traps and a transport device. Characteristic size of devices ranges from 10 \mum to 50 \mum and the field gradient produced by these devices is ~ 100 T/m for less than 1 Amp of current. Its applications include study and control of magnetic particles as well as control and manipulation of biological organisms.

[1] M. Drndic et.al., Appl. Phys. Lett. 72, 2906 (1998).

This work was funded by ONR N00014-99-1-0347.

[A7.009] Micro-electromagnets for control of magnotactic bacteria

A microorganism called magnetotactic bacteria can orient and migrate along magnetic field lines due to their intracellular magnetic structure, the magnetosome, which is a chain of nano scale iron particles[1]. We have fabricated micro-electromagnet devices to control and manipulate these bacteria using their magnetic characteristics. Micro-electromagnet devices consist of layers of lithographically patterned micron scale wires that produce local magnetic field. Current densities up to 10^8 A/cm^2 and magnetic field gradient up to 100T/m were demonstrated in these devices and they have been used in atom optics to manipulate neutral atoms in vacuum[2]. The devices are designed to locally trap the bacteria, arrange them in an array and guide their motion in microscopic scale. By allowing the control of the motions of each bacterium, these devices are expected to aid study the bacteria’s behaviors.

[1] R. P. Blackmore, Science 190, 377(1975)

[2] M. Drndic et.al., Appl. Phys. Lett. 72, 2906 (1998)

This work was funded by ONR N00014-99-1-0347.

[A7.010] Quantum 1/f Noise in Resonant Tunneling Diodes

Resonant tunneling diodes consist of two potential barriers enclosing a quantum well. If the electron energy is close to the energy level in the well, resonance occurs and a peak I_P of the current occurs, for the voltage VP. If the voltage increases further, only a negligibly small non-resonant current trickle remains at the voltage V=VV. Scattering processes that reduce the energy of the carriers to a value close to eVP will always be present, generating a finite current minimum IV at VV. Between VP and VV there is a negative differential conductance G=-(IP-IV)/(VV-VP) on the I/V curve, that is used to generate oscillations. The 1/f noise in IV is given by the conventional quantum 1/f effect with (Dv/c)2=2eVV/m. This yields IV-2SIv(f) =2aA/f N. Here N is given by N =tIV/e, where t is the life time of the carriers. The quantum 1/f frequency fluctuations can be obtained from the formula Sdw/w =(1/4Q4)SdG/G ,which was derived in 1978. This yields Sdw/w =(1/4Q4)(4a/3p)(2eVV/mc2) for the fractional frequency fluctuation spectrum exhibited by the RTD if included in an RF circuit of quality factor Q.

[A7.011] Electrodynamical Quantum 1/f Noise of Elactromagnetic Helmholtz Resonators

Electromagnetic Helmholtz Resonators are oscillant systems that always include dissipative elements. If they are oscillating in a well defined mode, they are described by a simple harmonic oscillator equation with dissipative coefficient g and resonance frequency w2 = wo2 + g2. Differentiating this expression, and dividing by 2w2, we get S(dw/w)=(1/4Q4)S(dg/g). The physical quantum mechanical cross sections and process rates defining g must exhibit fundamental quantum 1/f noise given by S(dg/g) =S(ds/s)=S(dm/m)=L/f, with L=2a/pN. The spectral density of fractional fluctuations is thus the same for the conductivity s and electron mobility m in the resonator walls of area A enclosing the volume V. The coherent quantum 1/f formula was used for the quantum 1/f coefficient L, where a = e2/hc = 1/137 is the fine structure constant. Using N=Ad, no=N/V, and a well-known expression Q=kV/AD = w/2g; (k=1 is a geometrical factor) of the quality factor Q in terms of the penetration depth D, we get the spectral density of fractional fluctuations S(dw/w) =a/2pfkQ3Vno. With n=100GHz, Q=1E4, V=0.03 cm3, no=5E22 cm-3 we obtain S(dw/w) =1E-36/Hz.

[A7.012] Nanometer-Scale Scanning Sensors Fabricated Using Stencil Masks

In recent years, new forms of scanning probe microscopy have been developed which utilize small electrical devices that are scanned above samples to act as high-resolution sensors. Examples include scanning single electron transistors, scanning thermal microscopy, and gated scanning tunneling microscopy. All of these techniques are limited in resolution by the size of the sensor and/or the distance of the sensor to the sample. We have developed a new technique for fabricating scanning sensors at the 10-nm scale. Metal devices are deposited directly onto conventional atomic-force microscope (AFM) tips through a stencil mask. The stencil is a suspended Si_3N_4 membrane in which holes as small as 10 nm are fabricated using electron beam lithography. Operating as an AFM, the probe tip is positioned over the stencil and metal is deposited through the hole to form the desired device pattern. We present details of the apparatus construction as well as results of experiments where individual cobalt nanoparticles were deposited on tips for use in high-resolution magnetic force microscopy.

[A7.013] Surface Tension Flows in Bent Micro-Channels: An Energy Minimization Theory versus Experiment

We show how surface tension flows through micro-channels can be described effectively by the solution of a constrained energy minimization. The optimization cost is equal to the total potential energy and is comprised of a liquid/air and liquid/solid part: Ep = c1 LA + c2 LS. Here LA is the surface area of the liquid/air interface and c1 is the liquid/air surface tension energy per unit area. Similiarily, LS is the solid/liquid front area and c2 is the corresponding liquid/solid energy per unit area. The optimization constraints follow from the enclosed fluid volume and the micro-channel geometry. Using the calculus of variations we show how this problem can be reduced to a simple 2 variable, 1 free parameter (p = c2/c1) optimization problem which may be solved easily. This yields the interface shape for any enclosed fluid volume, hence it says how the surface tension front propagates through the micro-channel.

Next, the theory is compared with experimental flow through a micro-channel corner. The data is generated and filmed for (slow) flows through bent micro-channels (about 800um wide and 100um deep channels at <5ul per minute mass flows). We find good agreement between the measured and predicted front shapes for a variety of channel surfaces: both hydrophobic and hydrophyllic.

[A7.014] Quantum Aspects of Elastic Instability in Nanostructures

Current work in nanophysics at Dartmouth will be described. Both theoretical and experimental work will be discussed. I will also discuss my own work with double-well systems and the possibility of observing 'macroscopic' quantum phenomena using nanostructures.

Part A of program listing