control, and a simple nerve net can perform elementary tasks (jellyfishes
swim). Echinoderms have a central nerve ring with radial nerves (for
example, sea stars have central and radial nerves with nerve net). Planarians
have small brains that send information through two or more nerve trunks, as
illustrated in Figure 1.2.2.
Figure 1.2.2. Overview of invertebrate nervous systems
1.3. NANO- AND MICROELECTROMECHANICAL SYSTEMS
Through biosystems analogy, a great variety of man-made
electromechanical systems have been designed and made. To analyze, design,
develop, and deploy novel NEMS and MEMS, the designer must synthesize
advanced architectures, integrate the latest advances in nano- and microscale
actuators/sensors (transducers) and smart structures, integrated circuits (ICs)
and multiprocessors, materials and fabrications, structural design and
optimization, modeling and simulation, et cetera. It is evident that novel
optimized NEMS and MEMS architectures (with processors or
multiprocessors, memory hierarchies and multiple parallelism to guarantee
high-performance computing and decision making), new smart structures and
actuators/sensors, ICs and antennas, as well as other subsystems play a critical
role in advancing the research, developments, and implementation. In this book
we discuss optimized architectures, and the research in architecture
optimization will provide deep insights into how intelligent large-scale
integrated NEMS and MEMS can be synthesized.
Electromechanical systems, as shown in Figure 1.3.1, can be classified as
•
conventional electromechanical systems,
•
microelectromechanical systems (MEMS),
•
nanoelectromechanical systems (NEMS).
Nerve
Trunk
Brain
Ring
of Nerve
Radial Nerves
Nerve Net
cnidarian echinoderm planarian
© 2001 by CRC Press LLC
Figure 1.3.1. Classification of electromechanical systems
The operational principles and basic foundations of conventional
electromechanical systems and MEMS are the same, while NEMS are
studied using different concepts and theories. In fact, the designer applies the
classical Lagrangian and Newtonian mechanics as well as electromagnetics
(Maxwell’s equations) to study conventional electromechanical systems and
MEMS. In contrast, NEMS are studied using quantum theory and
nanoelectromechanical concepts. Figure 1.3.2 documents the fundamental
theories to study the processes and phenomena in conventional, micro, and
nanoelectromechanical systems.
Figure 1.3.2. Fundamental theories in electromechanical systems
Conventional
Electromechanical
Systems
Micro-
electromechanical
Systems
Nano-
electromechanical
Systems
Fundamental Theories:
Classical Mechanics
Electromagnetics
Fundamental Theories:
Quantum Theory
Nanoelectromechanics
Electromechanical
Systems
Electromechanical
Systems
Conventional
Electromechanical
Systems
Micro-
electromechanical
Systems
Nano-
electromechanical
Systems
© 2001 by CRC Press LLC
NEMS and MEMS integrate different structures, devices, and subsystems.
The research in integration and optimization (optimized architectures and
structural optimization) of these subsystems has not been instituted and
performed, and end-to-end (processors – networks – input/output subsystems –
ICs/antennas – actuators/sensors) performance and behavior must be studied.
Through this book we will study different NEMS and MEMS architectures, and
fundamental and applied theoretical concepts will be developed and
documented in order to design next generation of superior high-performance
NEMS and MEMS.
The large-scale NEMS and MEMS, which can integrate processor
(multiprocessor) and memories, high-performance networks and input-output
(IO) subsystems, are of far greater complexity than MEMS commonly used
today. In particular, the large-scale NEMS and MEMS can integrate:
•
thousands of nodes of high-performance actuators/sensors and smart
structures controlled by ICs and antennas;
•
high-performance processors or superscalar multiprocessors;
•
multi-level memory and storage hierarchies with different latencies
(thousands of secondary and tertiary storage devices supporting data
archives);
•
interconnected, distributed, heterogeneous databases;
•
high-performance communication networks (robust, adaptive intelligent
networks).
It must be emphasized that even the simplest nanosystems (for example,
pure actuator) usually cannot function alone. For example, at least the internal
or external source of energy is needed.
The complexity of large-scale NEMS and MEMS requires new
fundamental and applied research and developments, and there is a critical need
for coordination across a broad range of hardware and software. For example,
design of advanced nano- and microscale actuators/sensors and smart
structures, synthesis of optimized (balanced) architectures, development of new
programming languages and compilers, performance and debugging tools,
operating system and resource management, high-fidelity visualization and data
representation systems, design of high-performance networks, et cetera. New
algorithms and data structures, advanced system software and distributed access
to very large data archives, sophisticated data mining and visualization
techniques, as well as advanced data analysis are needed. In addition, advanced
processor and multiprocessors are needed to achieve sustained capability
required of functionally usable large-scale NEMS and MEMS.
The fundamental and applied research in NEMS and MEMS has been
dramatically affected by the emergence of high-performance computing.
Analysis and simulation of NEMS and MEMS have significant outcomes. The
problems in analysis, modeling, and simulation of large-scale NEMS and
MEMS that involves the complete molecular dynamics cannot be solved
because the classical quantum theory cannot be feasibly applied to complex
molecules or simplest nanostructures (1 nm cube of nanoactuator has thousands
© 2001 by CRC Press LLC
of molecules). There are a number of very challenging research problems in
which advanced theory and high-end computing are required to advance the
theory and engineering practice. The multidisciplinary fundamentals of
nanoelectromechanics must be developed to guarantee the possibility to
synthesize, analyze, and fabricate high-performance NEMS and MEMS with
desired (specified) performance characteristics. This will dramatically shorten
the time and cost of developments of NEMS and MEMS for medical and
biomedical, aerospace and automotive, electronic and manufacturing systems.
The importance of mathematical model developments and numerical
analysis has been emphasized. Numerical simulation enhances, but does not
substitute for fundamental research. Furthermore, meaningful and explicit
simulations should be based on reliable fundamental studies and must be
validated through experiments. However, it is evident that simulations lead to
understanding of performance of complex NEMS and MEMS (nano- and
microscale structures, devices, and sub-systems), reduce the time and cost of
deriving and leveraging the NEMS and MEMS technologies from concept to
device/system, and from device/system to market. Fundamental and applied
research is the core of the simulation, and focused efforts must be concentrated
on comprehensive modeling and advanced efficient computing.
To comprehensively study NEMS and MEMS, advanced modeling and
computational tools are required primarily for 3D+ (three-dimensional
geometry dynamics in time domain) data intensive modeling and simulations to
study the end-to-end dynamic behavior of actuators and sensors. The
mathematical models of NEMS, MEMS, and their components (structures,
devices, and subsystems) must be developed. These models (augmented with
efficient computational algorithms, terascale computers, and advanced
software) will play the major role to simulate the design of NEMS and MEMS
from virtual prototyping standpoints.
There are three broad categories of problems for which new algorithms
and computational methods are critical:
1.
Problems for which basic fundamental theories are developed, but the
complexity of solutions is beyond the range of current and near-future
computing technologies. For example, the conceptually straightforward
classical quantum mechanics and molecular dynamics cannot be applied
even for nanoscale actuators. In contrast, it will be illustrated that it is
possible to perform robust predictive simulations of molecular-scale
behavior for nano- and microscale actuators/sensors and smart structures
which might contain millions of molecules.
2.
Problems for which fundamental theories are not completely developed to
justify direct simulations, but can be advanced or developed by advanced
basic and numerical methods.
3.
Problems for which the developed advanced modeling and simulation
methods will produce major advances and will have a major impact. For
example, 3D+ transient end-to-end behavior of NEMS and MEMS.
For NEMS and MEMS, as well as for their devices and subsystems,
© 2001 by CRC Press LLC
high-fidelity modeling and massive computational simulations (mathematical
models designed with developed intelligent libraries and databases/archives,
intelligent experimental data manipulation and storage, data grouping and
correlation, visualization, data mining and interpretation) offer the promise of
developing and understanding the mechanisms, phenomena and processes in
order to improve efficiency and design novel high-performance NEMS and
MEMS. Predictive model-based simulations require terascale computing and an
unprecedented level of integration between engineering and science. These
modeling and simulations will lead to new fundamental results. To model and
simulate NEMS and MEMS, we augment modern quantum mechanics,
electromagnetics, and electromechanics at the nano- and microscale. In
particular, our goal is to develop the nanoelectromechanical theory.
One can perform the steady-state and dynamic analysis. While steady-state
analysis is important, and the structural optimization to comprehend the
actuators/sensors, smart structures, and antennas design can be performed,
NEMS and MEMS must be analyzed in the time domain. The long-standing
goal of nanoelectromechanics is to develop the basic fundamental conceptual
theory in order to determine and study the interactions between actuation and
sensing, computing and communication, signal processing and hierarchical data
storage (memories), and other processes and phenomena in NEMS and MEMS.
Using the concept of strong electromagnetic-electromechanical interactions, the
fundamental nanoelectromechanical theory will be developed and applied to
nanostructures and nanodevices, NEMS and MEMS to predict the performance
through analytical solutions and numerical simulations. Dynamic macromodels
of nodes can be developed, and single and groups of molecules can be studied.
It is critical to perform this research in order to determine a number of the
parameters to make accurate performance evaluation and to analyze the
phenomena performing simulations and comparing experimental, modeling and
simulation results.
Current advances and developments in modeling and simulation of
complex phenomena in NEMS and MEMS are increasingly dependent upon
new approaches to robustly map, compute, visualize, and validate the results
clarifying, correlating, defining, and describing the limits between the
numerical results and the qualitative-quantitative analytic analysis in order to
comprehend, understand, and grasp the basic features. Simulations of NEMS
and MEMS require terascale computing that will be available within a couple
of years. The computational limitations and inability to develop explicit
mathematical models (some nonlinear phenomena cannot be comprehended,
fitted, and precisely mapped) focus advanced studies on the basic research in
robust modeling and simulation under uncertainties. Robust modeling,
simulation, and design are critical to advance and foster the theoretical and
engineering enterprises. We focus our research on the development of the
nanoelectromechanical theory in order to model and simulate large-scale
NEMS and MEMS. At the subsystem level, for example, nano- and microscale
actuators and sensors will be modeled and analyzed in 3D+ (three-dimensional
© 2001 by CRC Press LLC
geometry dynamics in time domain) applying advanced numerical robust
methods and algorithms. Rigorous methods for quantifying uncertainties for
robust analysis should be developed. Uncertainties result due to the fact that it
is impossible to explicitly comprehend the complex interacted subsystems and
processes in NEMS and MEMS (actuators/sensors and smart structures,
antennas, digital and analog ICs, data movement, storage and management
across multilevel memory hierarchies, archives, networks and periphery),
structural and environmental changes, unmeasured and unmodeled phenomena,
et cetera.
To design NEMS and MEMS, we will develop analytical mathematical
models. There are a number of areas where the advances must be made in order
to realize the promises and benefits of modern theoretical developments
recently made. For example, to perform 3D+ modeling and data intensive
simulations of actuators/sensors and smart structures, we will use advanced
analytical and numerical methods and algorithms (novel methods and
algorithms in geometry and mesh generation, data assimilation, and dynamic
adaptive mesh refinement) as well as the computationally efficient and robust
M
ATLAB
environment. There are fundamental and computational problems that
have not been addressed, formulated and solved due to the complexity of large-
scale NEMS and MEMS (e.g., large-scale hybrid models, limited ability to
generate and visualize the massive amount of data, et cetera). Other problems
include nonlinearities and uncertainties which imply fundamental limits to
formulate, set up, and solve analysis and design problems. Therefore, one
should develop rigorous methods and algorithms for quantifying and modeling
uncertainties, 3D+ geometry and mesh generation techniques, as well as
methods for adaptive robust modeling and simulations under uncertainties. A
broad class of fundamental and applied problems ranging from fundamental
theories (quantum mechanics and electromagnetics, electromechanics and
thermodynamics, structural synthesis and optimization, optimized architecture
design and control, modeling and analysis, et cetera) and numerical computing
(to enable the major progress in design and virtual prototyping through the
large scale simulations, data intensive computing, and visualization) will be
addressed and thoroughly studied in this book. Due to the obvious limitations
and the scope of this book, a great number of problems and phenomena will not
be addressed and discussed (among them, fabrication and manufacturing,
chemistry and material science).
1.4.
APPLICATIONS OF NANO- AND
MICROELECTROMECHANICAL SYSTEMS
Depending upon the specifications and requirements, objectives and
applications, NEMS and MEMS must be designed. Usually, NEMS are faster
and simpler, more efficient and reliable, survivable and robust compared
with MEMS. However, due to the limited size and functional capabilities,
one might not attain the desired characteristics. For example, consider nano-
© 2001 by CRC Press LLC
and microscale actuators. The actuator size is determined by the force or
torque densities. That is, the size is determined by the force or torque
requirements and materials used. As one uses NEMS or MEMS as the logic
devices, the output electric signal (voltage or current) or electromagnetic
field (intensity or density) must have the specified value.
Although NEMS and MEMS have the common features, the differences
must be emphasized as well. Currently, the research and developments in
NEMS and molecular nanotechnology are primarily concentrated on design,
modeling, simulation, and fabrication of molecular-scale devices. In contrast,
MEMS are usually fabricated using other technologies, for example,
complementary metal oxide semiconductor (CMOS) and lithography. The
direct chip attaching technology was developed and widely deployed. Flip-chip
assembly replaces wire banding to connect ICs with micro- and nanoscale
actuators and sensors. The use of flip-chip technology allows one to eliminate
parasitic resistance, capacitance, and inductance. This results in improvements
of performance characteristics. In addition, flip-chip assembly offers
advantages in the implementation of advanced flexible packaging, improving
reliability and survivability, reduces weight and size, et cetera. The flip-chip
assembly involves attaching actuators and sensors directly to ICs. The actuators
and sensors are mounted face down with bumps on the pads that form electrical
and mechanical joints to the ICs substrate. The under-fill encapsulate is then
added between the chip surface and the flex circuit to achieve the high
reliability demanded. Figure 1.4.1 illustrates flip-chip MEMS.
IC
SensorActuator
−
Actuator
Sensor
Figure 1.4.1. Flip-chip monolithic MEMS with actuators and sensors
The large-scale integrated MEMS (a single chip that can be mass-produced
using the complementary metal oxide semiconductor (CMOS),
photolithography, and other technologies at low cost) integrates:
•
N nodes of actuators/sensors, smart structures,
•
ICs and antennas,
•
processor and memories,
•
interconnection networks (communication busses),
•
input-output (IO) systems.
Different architectures can be synthesized, and this problem is discussed
© 2001 by CRC Press LLC
and covered in Chapter 2. One uses NEMS and MEMS to control complex
systems, processes, and phenomena. A high-level functional block diagram
of large-scale MEMS is illustrated in Figure 1.4.2.
Figure 1.4.2. High-level functional block diagram of large-scale MEMS
with rotational and translational actuators and sensors
Actuators are needed to actuate dynamic systems. Actuators respond to
command stimulus (control signals) and develop torque and force. There is a
great number of biological (e.g., human eye and locomotion system) and man-
made actuators. Biological actuators are based upon electromagnetic-
mechanical-chemical phenomena and processes. Man-made actuators
(electromagnetic, electric, hydraulic, thermo, and acoustic motors) are devices
that receive signals or stimulus (stress or pressure, thermo or acoustic, et cetera)
and respond with torque or force.
Consider the flight vehicles. The aircraft, spacecraft, missiles, and
interceptors are controlled by displacing the control surfaces as well as by
changing the control surface and wing geometry. For example, ailerons,
elevators, canards, flaps, rudders, stabilizers and tips of advanced aircraft can
be controlled by nano-, micro-, and miniscale actuators using the NEMS- and
Data
Acquisition
Sensors
Antennas
Amplifiers
ICs
VariablesMeasured
Actuators
Analysisand
Decision
System
Dynamic
Controller
Output
VariablesSystem
Criteria
Objectives
VariablesMEMS
SensorActuator
−
MEMS
SensorActuator
−
SensorActuator
−
IO
© 2001 by CRC Press LLC
MEMS-based smart actuator technology. This NEMS- and MEMS-based
smart actuator technology is uniquely suitable in the flight actuator
applications. Figure 1.4.3 illustrates the aircraft where translational and
rotational actuators are used to actuate the control surfaces, as well as to
change the wing and control surface geometry.
Figure 1.4.3. Aircraft with NEMS- and MEMS-based translational and
rotational flight actuators
Sensors are devices that receive and respond to signals or stimulus. For
example, the loads (which the aircraft experience during the flight),
vibrations, temperature, pressure, velocity, acceleration, noise, and radiation
can be measured by micro- and nanoscale sensors, see Figure 1.4.4. It should
be emphasized that there are many other sensors to measure the
electromagnetic interference and displacement, orientation and position,
voltages and currents in power electronic devices, et cetera.
ψφθ
,,
:AnglesEuler
ActuatorsFlight
BasedMEMSandNEMS
−−
SensorActuator
−
SensorActuator
−
GeometryWing
GeometrySurface
ntDisplacemeSurface
Control :
© 2001 by CRC Press LLC
Figure 1.4.4. Application of nano- and microscale sensors in aircraft
Usually, several conversion processes are involved to produce electric,
electromagnetic, or mechanical output sensor signals. The conversion of
energy is our particular interest. Using the energy-based analysis, the general
theoretical fundamentals will be thoroughly studied.
The major developments in NEMS and MEMS have been fabrication
technology driven, and the applied research has been performed mainly to
manufacture structures and devices, as well as to analyze some performance
characteristics. For example, mini- and microscale smart structures as well as
ICs have been studied in details, and feasible manufacturing technologies,
materials, and processes have been developed. Recently, carbon nanotubes
were discovered, and molecular wires and molecular transistors were built.
However, to our best knowledge, nanostructures and nanodevices, NEMS
and MEMS, have not been comprehensively studied at the nanoscale, and the
efforts to develop the fundamental theory have not been reported. In this
book, we will apply the quantum theory and charge density concept,
advanced electromechanics and Maxwell's equations, as well as other
cornerstone methods, to model nanostructures and nanodevices (ICs and
antennas, actuators and sensors, et cetera). In particular, the
nanoelectromechanical theory will be developed. A large variety of actuators
and sensors, antennas and ICs with different operating features are modeled
and simulated. To perform high-fidelity integrated 3D+ data intensive
modeling with post-processing and animation, the partial and ordinary
nonlinear differential equations are solved.
ψφθ
,,
:AnglesEuler
Radiation
Sensors
Noise
onAccelerati
Velocity
ressureP
Vibrations
Loads
eTemperatur
Flight Computer
SensorActuator
−
SensorActuator
−
© 2001 by CRC Press LLC
Không có nhận xét nào:
Đăng nhận xét