The Microsystems for Bio-Imaging and Metrology Project designs, fabricates, and tests microelectromechanical systems (MEMS) for studying microscopic and nanoscopic magnetic phenomena. Project members are taking an approach based on chip-scale microsystems and nanosystems to advance instrumentation by improving sensitivity, portability, cost, and traceability to fundamental constants. The research has traditionally focused on the data storage, electronics, and communication industries and is currently exploring applications in medicine and bioengineering. Recent programs include magnetic manipulation and measurement of single molecules in microfluidic environments, engineered radio-frequency tags for magnetic resonance imaging (MRI) and microfluidics, single molecule enzymology, precision cantilevers for transfer standards and intrinsic force measurements, microfabrication of chip-scale atomic devices, and integration of chip-scale MRI microscopy systems.
Methods for measuring the properties and estimating the performance of nanomagnetic materials and devices are in the early development stage. Conventional magnetometers lack the necessary sensitivity for direct measurement of single nanoparticles or nanodevices. We target medical and security applications where we believe a significant impact could be made in the next five years in developing new measurement tools for basic research, rapid detection, assay, and diagnosis; early detection of diseases; and drug development and approval. We are developing new methods for magnetic measurements at the nanometer scale. Examples include cantilever magnetometers, where a sample is integrated with the sensor for maximum coupling, which hold promise for magnetic measurements at room temperature at the level of 1000 atoms. We develop: (1) methods for understanding the physical processes that give rise to the unique properties of nanomagnetic materials, (2) new instrumentation for nanometer scale MEMS magnetometry, and (3) new protocols for detecting the distinct signatures of magnetic nanoparticles.
The medical and security communities are acutely interested in harnessing the unique properties of nanomaterials, with nanomagnetism being one of the key physical characteristics of interest. We focus on the properties of nanoparticles dispersed in liquids or solids at the bulk and single-particle limits, with the ultimate goal of developing them for new imaging modalities. These include magnetic particle imaging (MPI) and MRI for medical applications, microfluidic devices that use magnetic particles for single cell and single molecule bioassay applications, radio-frequency tags for cell tracking in research animal models, and forensic identification of controlled substances.
The project’s technical strategy is to develop: (1) new, ultrasensitive magnetometers based on MEMS chips, MPI, and MRI, (2) integrated microsystems and new imaging modalities based on microscale and nanoscale magnetic particles, (3) quantitative cantilevers and calibration methods for quantifying the performance of atomic force microscopes, (4) magnetic microfluidic platforms for single-molecule studies, and (5) integrated microsystems based on atomic transitions.
Mems Magnetometry of Patterned Sub-Micrometer Dots — Sub-micrometer scale magnetic measurements have proven to be a challenge for conventional magnetometers, and new methods are being employed to probe magnetism on this scale. We will provide new magnetometers based on highly specialized MEMS chips fabricated at NIST. Such instruments will be inexpensive, since MEMS can be batch-fabricated in large quantities. In addition, large-scale magnetic wafer properties can be transferred to smaller MEMS magnetometers so that nanometer-scale measurements can be calibrated with reference to fundamental units. In particular, the focus is on developing torque and force magnetometers, magnetic-resonance spectrometers, and MRI microscopes on MEMS chips. Over the long term, we expect that this technology will lead to atomic-scale magnetic instrumentation for the measurement and visualization of fundamental magnetic phenomena.
Mems Based Magnetic Moment Standard Reference Materials — We are developing a method for defining the magnetic moment of a reference material based on a torsional resonator with a patterned magnetic film on its surface. Given accurate measurements of the magnitude of the applied field, the moment of inertia of the resonator, and the magnetic stiffening effect of the film on the resonance frequency of the resonator, an absolute measurement of the anisotropy energy and magnetic moment of the film can be made independently. With this approach, the anisotropy energy and the magnetic moment of the film are measured directly, eliminating the need for a detailed knowledge of the film’s saturation magnetization or sample volume. Calibrated moments as small as 10–6 ampere meter squared can be measured. In principle, all of the measurements needed for the reference material are traceable to an atomic clock frequency reference.
Microfabricated Radio Frequency Tags For MRI Tracking Of Single Cells — We are developing a new biological detection methods that would have application in two areas: (1) to replace conventional serial fl uorescencebased (and also more recently suggested optically probed micro-metallic barcode) techniques with a parallel detection scheme capable of simultaneously detecting multiple bio-agents without the need for optical access to the system, and (2) to improve the specificity of cell labeling strategies for MRI. Custom designed magnetic micro-particle tags that induce distinct magnetic resonance frequency shifts in the nuclear magnetic resonance (NMR) of protons in water surrounding the labeled biological agent are being microfabricated and tested. Magnetic detection would be based on the tag particle set being specifically tailored (in size, shape, and composition using microlithography) to give distinct NMR frequency shifts and resonance line shapes. Such a magnetic resonance system would enable noninvasive and inherently parallel detection both in vitro and in vivo.
Integrated Chip-Scale MRI Microscopes — We are developing methods for integration of DC and radio-frequency magnetic field sources and MEMS cantilever oscillators to perform magnetic resonance imaging and spectroscopy of small samples attached to the oscillators. The main goal is to push the technology to measure magnetic phenomena at the nanometer scale. Cantilevers operated at resonance have the potential for detecting single magnetic nanoparticle tags with 105 spins at room temperature, and there are prospects for single spin detection at low temperatures. Shrinking the magnetic subsystems of a magnetic resonance instrument into a chip-scale microsystem provides benefits in terms of reduced size, batch fabrication cost, and better performance via an increased detector bandwidth and sensitivity.
Calibrated Cantilevers For Force Transfer Standards At The Nanonewton Range — The ability to demonstrate traceable quantitative measurement of forces in the regime below 10 millinewtons is a tremendous challenge for a variety of manufacturers. We are developing metrology for use by primary, secondary, and industrial laboratories to support quantitative measurement of these forces. The goal is an instrument that incorporates NIST’s force and length metrology to traceably determine the spring constants of silicon reference cantilevers and torsion oscillators. Silicon reference cantilevers have been proposed by the Materials Science and Engineering Laboratory as Standard Reference Materials (SRMs) for calibrated adhesion tests. Calibrated piezoresistive cantilever load cells could ultimately serve as a standard reference material for the dissemination of force to atomic force microscopes (AFMs) and instrumented indentation machines. Torsion oscillators may be used as SRMs for magnetometers.
Cantilever Sensors For Intrinsic Force Measurements At The Femtonewton Range Of Single Molecules — Among the experiments made possible by the revolution in scanned probe microscopy, measurements of bond rupture (forces between atoms), binding rupture (forces between molecules), and molecular conformation changes (force induced structural changes of individual molecules) are perhaps the most remarkable. Such forces might then serve as intrinsic standards, effectively rendering this class of instruments self-calibrating. We are developing the needed femtonewton force sensors for the various applications. For instance, an ultrathin cantilever with an integral interferometric displacement sensor is appropriate for use in the binding rupture experiments since the unit must be compatible with aqueous biological environments. For the metal bond rupture experiments, a feedback-controlled torsion balance using either electromagnetic or electrostatic forces to provide the null compensation is more appropriate. These sensors will be calibrated in an ultrahigh vacuum, atomic force test and calibration apparatus in the Manufacturing Engineering Laboratory and the single molecule laser tweezers apparatus at JILA (jointly operated by the University of Colorado and NIST).
Magnetic Templates For Nanometer Scale Manipulation And Assembly Of Magnetic Particles And Devices — We are developing a nanometer scale assembly platform based on an array of “switchable” magnetic dots for manipulating magnetic components. In the current configuration, a magnetic force microscope (MFM) tip is used to distribute magnetic polystyrene beads in an array of patterned Permalloy (Ni-Fe) traps on a thin membrane as part of a microfluidic flow cell. We have demonstrated the ability to translate particles with nanometer precision and sort micrometer sized magnetic particles based on size differences in the array with the MFM. In addition, we are currently developing fluidic cells with the potential for large-scale integration based on “spin-valve” technology using magnetically balanced spin-valve structures that act as switchable permanent magnets with a ferromagnetic “on” or an antiferromagnetic “off” net magnetization state so that magnetic particles in the fluid cell can be electronically confined and released for transporting, sorting, or assembly applications. Spin-valve magnetic traps and a matrix addressable architecture similar to MRAM would then eliminate the need for a scanning MFM tip for translating particles. In principle, these techniques would be the basis for nanometer scale robotic assembly of components to achieve novel biological, chemical, electrical, or mechanical functionality at the single molecule level.
Single Molecule Enzymology: Mechanics Of Replication — The structure of DNA and individual proteins is being examined in increasing detail; however our knowledge of the mechanics of their interactions is often limited to bulk experiments. We are developing a new single-molecule assay to assist in elucidating the mechanics of enzyme activity. The flexibility of this in-vitro assay will be used to determine the operation of individual enzymes, complementing existing genetic techniques. The ultimate goal of this project is to work towards a more complete understanding of the mechanics of DNA polymerization and how errors in this process can lead to a range of neurological diseases, eventually developing an in-vivo assay. To control and manipulate the DNA substrate we will be working with superparamagnetic beads in a microfluidic environment. By attaching a magnetic bead to one end and applying a magnetic field we can stretch out the DNA. This provides a static one-dimensional template along which enzyme activity can be easily observed for longer periods and correlated with the underlying base-pair sequence.
Mems Design, Fabrication, And Packaging For Chip-Scale Atomic Clocks And Magnetometers — Many important electronic devices (such as global positioning receivers, wireless receivers, portable magnetometers, and compact gyroscopes and accelerometers) would greatly improve if very small, highly accurate, low-power, and low-cost measurement references were available. These devices are typically very large (often laboratory-scale), often consume kilowatts of power or require cryogenic cooling, and are generally too expensive for widescale applications. The challenge is to shrink the size of the sensors and standards based on atomic properties from the laboratory scale (10 cubic meters) to the 1 cubic centimeter scale of a computer chip. While the chip-scale atomic clock (as an early example of a chip-scale atomic device) is only the size of a rice grain, it is a complex structure comprised of more than a dozen components; more advanced chip-scale atomic devices will likely be more complex. We are developing new methods for designing, fabricating, and assembling chip-scale atomic devices optimized for different applications. A key part of the chip-scale atomic device is the ultra-miniature atomic vapor cell containing the active sensing atoms. We are developing new technologies to design, fabricate, and assemble these crucial vapor cells into various chip-scale atomic device packages and applications.
Scanning electron micrographs of the double torsional oscillator geometry. (a) A typical chip containing 12 devices, all with double-side access. (b) A closer view of a double torsional oscillator with a 5 micrometer by 5 micrometer by 30 nanometer fi lm on the head. The illustration on the right side indicates the shape of the two main torsional modes of operation.
Custom cantilever assembly for submonolayer in-situ magnetometery of multilayer films (front and back views). The paddle is 1 millimeter on a side.
Graph of experimental magnetic interlayer exchange coupling measured by the micro-resonator magnetometer as a function of Cr layer thickness.
Video micrographs showing the rotation of a strand of magnetic particles while trapped at the edge of a single “on” state spin-valve element (highlighted by the white ellipses). The magnetic beads are 2.8 micrometers in diameter.
- Novel Fabrication of Micromechanical Oscillators with Nanoscale Sensitivity at Room Temperature — Over the past decade, several experimental methods have been developed to probe material properties on the micrometer and nanometer scales. Many of these novel methods employ the use of micromechanical cantilevers to achieve the desired sensitivity. Because these experiments are limited by the thermal noise of the cantilever itself, low temperatures often must be used, making the results less relevant to industrial applications. Thus, there is a strong demand for microcantilevers that are sensitive enough to obtain useful results at room temperature. A further complication arises when the experiments require that a micrometer-sized magnetic material be placed onto the cantilever. Doing this on an individual basis is not only time-consuming, but it also jeopardizes the uniformity and consistency of the results. Therefore, there is a clear need to develop a process to batch-fabricate ultrasensitive cantilevers with magnetic dots prealigned and deposited as part of the microfabrication process.
We have developed a process to meet these demands. Careful consideration was given to the design of the oscillator shape, and finite-element modeling was used to study the resonant shapes and to make sure that the resonance frequencies were in the desired ranges for the specific applications. Our ultrathin (less than 1 micrometer) single-crystal silicon cantilevers with integrated magnetic structures are the first of their kind. They were fabricated with a novel high-yield process in which magnetic film patterning and deposition are combined with cantilever fabrication. These novel devices have been developed for use as cantilever magnetometers and as force sensors in nuclear magnetic resonance force microscopy. These two applications have achieved nanometer-scale resolution with these cantilevers. Current magnetic moment sensitivity achieved for the devices, when used as magnetometers, is 10–15 ampere meter squared at room temperature, which is more than a thousandfold improvement in sensitivity compared to conventional magnetometers. Finite element modeling was used to improve design parameters, ensure that the devices meet experimental demands, and correlate mode shape with observed results. The photolithographic fabrication process was optimized, yielding an average of 85 percent and alignment better than 1 micrometer. Post-fabrication focused ion-beam milling was used to further pattern the integrated magnetic structures when nanometer scale dimensions were required.
- MEMS In-Situ Magnetometers to Measure Interlayer Coupling in Fe/Cr/Fe Trilayers — The characterization of thin magnetic films, patterned recording media, and nanometer-scale magnetic devices has proven to be a challenge for conventional magnetometers at nanoscale and atomic dimensions. The limitation on the sensitivity of magnetic moment sensors used in these instruments is fundamentally understood by comparing the energy necessary to excite the sensor relative to the energy necessary to excite the specimen for measurement purposes. Conventional magnetometers are designed for measurements of large specimens and therefore have a relatively low signal-to-noise ratio for small specimens. Sensitivity can be improved tremendously by integrating the specimens with the measurement sensor using microfabrication methods. Further, ex-situ measurements outside of the growth chamber require sample transfer, so there is the potential for substantial oxidation of thin films and devices.
We have tailored an ultrasensitive magnetometer for the in-situ study of thin-film interface magnetism and interlayer magnetic exchange coupling. The magnetometer is based on a customized micro- resonator made with silicon MEMS fabrication techniques. The measured changes in magnetic moments using the MEMS magnetometer were compared to theoretical calculations and measurements made by more conventional methods. In particular, we followed a previous formalism that includes interface roughness (interdiffusion length of different atoms) for predicting the oscillation periods of interlayer magnetic exchange coupling of Fe/Cr/Fe thin film multilayers as a function of Cr spacer thickness. Our results indicate that the magnetometer is well suited for studying the subtle changes in the magnetic moment of multilayer films during fi lm growth. Specifically we determined the short and long period oscillation wavelength of the interlayer exchange coupling in the Cr layer to be 2.1 monolayers and 13.8 monolayers, respectively. These results illustrate the utility of MEMS magnetometers for studying ultrathin magnetic films and multilayer devices.
- Manipulation of Magnetic Particles by Patterned Arrays of Magnetic Spin-Valve Traps — To address some of the limitations of current single-molecule tweezers techniques, we developed a novel magnetic tweezers based on a chip-scale microfluidic platform that can trap, measure, manipulate and sort magnetic particles in an array. The platform consists of an array of magnetic spin-valve elements separated from the biological sample by an optically transparent thin membrane, effectively isolating the electronic or magnetic components from the fluid bead solution. The particles are confined by local magnetic field gradients in an array of magnetic spin-valve structures. The spin valves have a bistable ferromagnetic “on” and antiferromagnetic “off” net magnetization states in the absence of an externally applied magnetic field, which allows magnetic particles to be selectively confined or released for transport or sorting.
Single particles can be rotated by applying a global magnetic field, and thus, the platform may be potentially used as a magnetic molecular tweezer where particles are rotated about the x, y or z axes while attached to multiple points of a biological molecule such as the ends of a strand of DNA to impart torsional forces. Alternatively, spin-valve trap arrays can be adapted to a low power MRAM switching architecture for massively parallel particle sorting applications. Previous work with spin-valve trapping elements in terms of biological microfluidic applications focuses on their ability to detect the presence of magnetic particles as they attach to locations with specific biological antibodies. In contrast, the current platform incorporates spin-valve elements that can be switched between bistable states to provide a local magnetic field gradient sufficiently large to trap a magnetic particle that can be used not only for the purpose of investigating conformational dynamics of single biological molecules but also to capture and sort biological molecules for gene sequencing or bioassay applications.
- Wafer-Level Fabrication and Filling of Cesium Vapor Cells for Chip-Scale Atomic Devices — The introduction of atomic vapor systems to MEMS has enabled a new class of hybrid microchips referred to as chip-scale atomic devices (CSADs) with integrated optical, radio-frequency, electronic, and quantum based atomic components. At the heart of a CSAD is a miniature cell containing alkali vapor. Suitable techniques for fabricating vapor cells must be developed such that they will not only be in a physical form suitable for integration into the rest of a MEMS assembly, but also in a manner that is amenable to true wafer- scale mass production. Whereas conventional cells are made by glass blowing, millimeter scale vapor cells are more easily fabricated by means of standard photolithography, silicon etching, and silicon-glass bonding techniques in a cleanroom. However, filling the cells with the highly volatile alkali metals and buffer gasses without exposure to air, and subsequent hermetic sealing of the cells, are formidable MEMS fabrication problems. The high temperature required for anodic bonding of silicon to glass is at odds with the low melting points of Cs and Rb. Thus, to date, cell fabrication requires technically complex methods that are limited to sequential rather than parallel processes to introduce the alkali metal into the cell and to subsequently seal the cell hermetically.
To overcome these disadvantages and simplify the cell-filling process, we have developed and tested a cesium batch-filling technique based on thin-film deposition and subsequent decomposition of cesium azide (CsN3). CsN3 is a solid at room temperature and is stable in air. It decomposes to produce pure cesium and nitrogen gas when heated to 450 degrees Celsius or through ultraviolet photolysis at room temperature. The azide method of cell filling eliminates the need for pipetting or microfluidic manipulation of liquid alkali metals since the Cs is liberated in situ after the cell has been hermetically sealed, thus simplifying the filling process and dramatically reducing the tooling cost.
Furthermore, the azide method, being based on thin-film deposition, is inherently wafer-level. Due to the lack of thermophysical and thermochemical data on CsN3, as well as the relative lack of published literature concerning CsN3 material properties, much of this work was exploratory in nature. While the photolysis of other metal azides has been studied in detail, the photolysis of CsN3 has not been previously demonstrated or studied before and as such the mechanisms for the photodissociation are just now coming to light.
- Chip-Scale Atomic Magnetometer — In collaboration with the Physics Laboratory, we have constructed, using MEMS techniques, a small low-power magnetic sensor based on alkali atoms. We used a coherent population trapping resonance to probe the interaction of the atoms’ magnetic moment with a magnetic field, and we detected changes in the magnetic flux density with a sensitivity of 50 picotesla per root hertz at 10 hertz. The magnetic sensor has a size of 12 cubic millimeters and dissipates 195 milliwatts of power. Further improvements in size, power dissipation, and magnetic field sensitivity are immediately foreseeable; such a device could provide a handheld battery-operated magnetometer with an atom shot-noise-limited sensitivity of 0.05 picotesla per root hertz.
Force Standard Reference Material for AFM Calibrations — We are designing a piezoresistive cantileves reducr force sensor that can serve as a force and/or stiffness SRM. The beam thickness in the gauge area waed to produce high sensitivity.
The trade-off for this sensitivity is a more complex artifact with the need to maintain additional critical dimensions. The numbered fiducial marks shown in the figure indicate points along the cantilever’s length where a calibrated stiffness will be determined. The functional dependence of stiffness on test location is more complicated in this design, but can be modeled assuming the gauge area acts as a cantilever hinge and the remainder of the beam is a rigid lever arm. The exact functional relationship will be examined using the electronic force balance in the Manufacturing Engineering Laboratory. As with the previously described stiffness SRMs, the goal is to settle on a fabrication strategy that yields a high degree of uniformity within a wafer. We hope to use resonance frequency, this time conveniently measured from the piezoresistor, with a bridge as a process monitor to check for mechanical uniformity within a wafer run. This sensor is currently in prototype production and testing results are not yet available.
U.S. Department of Commerce Silver Medal for the design, construction, and testing of the first operational chip-scale atomic clock and chipscale magnetometer based on atomic transitions, 2005 (John Moreland, Li-Anne Liew, EEEL; John Kitching, Peter Schwindt, Hugh Robinson, and Leo Hollberg, Physics Laboratory).