- Overview
- Structure
- Properties on both advantages and limitations
- Applications
The process of IC manufacturing often requires hundreds of sequential steps,
each one of which could lead to yield loss. Consequently, maintaining product quality
in an IC manufacturing facility often requires the strict control of hundreds or
even thousands of process variables. The issues of high yield, high quality and low
cycle time are being addressed in part by the ongoing development of several
critical capabilities:
- Process Monitoring
- Process/Equipment Modeling
- Process Optimization
- Process Control
- Equipment Malfunction Diagnosis
- Parametric Yield Modeling
< MENU >
Process Monitoring
The most important step in semiconductor process modeling is the collection of
data. It is essential to gather a sufficient sample of representative data; or else it
is impossible to train a neural network or any other type of model. As an
example of a methodology for collecting real-time data for semiconductor
process modeling or process control, consider the project In-Situ Monitoring of
Reactive Ion Etching. This project describes an effort to design a real-time data
acquisition system for RIE as well as the development and implementation of a
prototype microsensor for monitoring of film characteristics during etching. This
sensor will facilitate in-situ characterization of etch rate on the wafer surface, a
significant contribution to the state of the art. The microsensor concept has
evolved further in the project A Smart Sensor for Monitoring Reactive Ion
Etching. Another project in the area of process monitoring is A Piezoelectric
Silicon Acoustic Sensor, in which the goal is to investigate the use of acoustic
signals for monitoring of plasma processes.
< Main Page >
Process Modeling
Several researchers have of late reported noteworthy successes in using neural nets
to model the behavior of several key fabrication processes. Here the ability of neural
nets to discover input/output relationships from limited data is very useful. At the
Georgia Tech Microelectronics Research Center (MiRC), back-propagation (BP) neural
networks have been used to model ion-assisted plasma etching, as well as a variety
of other processes widely used in semiconductor manufacturing.
The models are required by semiconductor manufacturers in order to predict process
behavior under an exhaustive set of operating conditions, and with a very high degree
of precision. However, plasma processing involves highly complex and dynamic
interactions between reactive particles in an electric field.
Due to this inherent complexity, approaches to plasma etch modeling which
preceded the advent of neural networks had met with limited success. Other
processes besides plasma etching have benefited from the neural network
approach. To name one, chemical vapor deposition processes are also nonlinear
and also have been modeled by these means to good effect. Researchers at the
MiRC have developed neural process models for plasma-enhanced chemical
vapor deposition of silicon dioxide films. Process models have also been
developed for metal-organic chemical vapor deposition(MOCVD) and molecular
beam epitaxy (MBE), two popular techniques for growing thin, high purity
epitaxial films.
< Main Page >
Process Optimization
A natural extension of the neural modeling of processes is using the models thus
obtained to optimize processes or generate recipes for them. This optimization is
designed to produce designated target output responses based on the functional
relationship between controllable input parameters and process responses supplied
by the neural models. Such a process optimization activities has been undertaken
in the projects Plasma Enhanced Chemical Vapor Deposition and Recipe Synthesis
for PECVD and Process Modeling and Recipe Synthesis of Integrated Capacitors.
< Main Page >
Process Control
Since they can be used to build predictive models from multivariate sensor data
generated by process monitors, neural nets have also been applied to the
control of semiconductor manufacturing processes. The Intelligent
Semiconductor Manufacturing Group is participating in process control efforts
through the projects Real Time Control and Endpoint Detection of Reactive Ion
Etching, Real-Time Control of Plasma Etching Using Integrated Sensors and
Adaptive Neural Networks, Modeling and Control of Variable Frequency
Microwave Processing, and Modeling and Control of Molecular Beam Epitaxy
using the RHEED Diffraction Pattern.
< Main Page >
Equipment and Process Diagnosis
The use of Neural networks for the diagnosis of semiconductor manufacturing
processes and equipment is relatively new. Toward this end, the Intelligent
Semiconductor Manufacturing group has adopted the strategy of using neural
networks in tandem with expert systems in the projects Automated Malfunction
Diagnosis Of A Reactive Ion Etcher, Real-Time Diagnosis of Reactive Ion Etch
Using Optical Emission Spectroscopy, and Modeling and Diagnosis of Excimer
Laser Ablation
< Main Page >
Parametric Yield Modeling
Historically, IC yield enhancement has focused on the reduction of manufacturing
anomalies such as particulate matter, mechanical damage, and crystalline defects.
Under this paradigm, circuit failure is primarily characterized internal opens or shorts,
so-called "catastrophic" defects which result in a loss of functionality.
However, degradation in IC performance can result from sources other than
catastrophic defects alone. In fact, even in a defect-free fabrication environment,
random variations in the manufacturing process itself will lead to products with
varying levels of performance. These manufacturing variations result from the
fluctuation of various physical parameters (i.e. - doping concentrations, oxide
thicknesses, line widths, etc.), which in turn manifest themselves first as variations in
IC device parameters (such as threshold voltages), and finally as variations in circuit
performance measures (such as bandwidth).
Accurate IC performance prediction requires precise characterization of these
variations. Tools are therefore necessary which are capable of modeling circuit
parametric performance based on device and manufacturing variation, rather than
simply relying on nominal parameter values.
Neural networks and various statistical methods are used as parametric yield
modeling tools in the projects Statistical Prediction of Integrated Circuit Performance,
Reliability Modeling and Parametric Yield Prediction of GaAs MQW Avalanche
Photodiodes, and Design and Optimization of Microwave Circuits and Systems Using
Artificial Intelligence.
< Main Page >
Semiconductors are made up of individual atoms bonded together in a regular,
periodic structure to form an arrangement whereby each atom is surrounded by 8
electrons. An individual atom consists of a nucleus made up of a core of protons
(positively charged particles) and neutrons (particles having no charge) surrounded
by electrons. The number of electrons and protons is equal, such that the atom is
overall electrically neutral. The electrons occupy certain energy levels, based on the
number of electrons in the atom, which is different for each element in the periodic
table. The structure of a semiconductor is shown in the figure below.
Schematic representation of covalent bonds
in a silicon crystal lattice.
The atoms in a semiconductor are materials from either group IV of the periodic
table, or from a combination of group III and group V (called III-V semiconductors),
or of combinations from group II and group VI (called II-VI semiconductors). Silicon
is the most commonly used semiconductor material as it forms the basis for
integrated circuit (IC) chips and is the most mature technology and most solar cells
are also silicon based. A full periodic table is given in the page Periodic Table.
Several of the material properties of silicon are given in the page Silicon Material
Parameters. The bond structure of a semiconductor determines the material
properties of a semiconductor. One key effect is limit the energy levels which
the electrons can occupy and how they move about the crystal lattice. The
electrons surrounding each atom in a semiconductor are part of a covalent
bond. A covalent bond consists of two atoms "sharing" a single electron, such
that each atom is surrounded by 8 electrons.
The electrons in the covalent bond are held in place by this bond and hence
they are localised to region surrounding the atom. Since they cannot move or
change their energy, electrons in a bond are not considered "free" and cannot
participate in current flow, absorption or other physical processes of interest in
solar cells. However, only at absolute zero are all electrons in a bonded
arrangement. At elevated temperatures, the electron can gain enough energy
to escape from its bond, and if this happens, the electron is free to move about
the crystal lattice and participate in conduction. At room temperature, a
semiconductor has enough free electrons to allow it to conduct current, while
at, or close to absolute temperatures, a semiconductor behaves like an
insulator. The presence of the bond introduces two distinct energy states for
the electrons. The lowest energy position for the electron is to be in its bound
state. However, if the electron has enough thermal energy to break free of its
bond, then it becomes free.
The electron cannot attain energy values intermediate to these two levels; it is
either at a low energy position in the bond, or it has gained enough energy to
break free and therefore has a certain minimum energy. This minimum energy is
called the "band gap" of a semiconductor. The number and energy of the free
electrons is basic to the operation of electronic devices. The space left behind by
the electrons allows a covalent bond to move from one electron to another, thus
appearing to be a positive charge moving through the crystal lattice. This empty
space is commonly called a "hole", and is similar to an electron, but with a
positive charge.
Animation showing formation of "free" electrons
and holes when an electron can escape its bond.
< MENU >
All transistor types can be used as the building blocks of logic gates, which are
fundamental in the design of digital circuits. In digital circuits like
microprocessors, transistors act as on-off switches; in the MOSFET, for
instance, the voltage applied to the gate determines whether the switch is on or
off. Transistors used for analog circuits do not act as on-off switches; rather,
they respond to a continuous range of inputs with a continuous range of
outputs. Common analog circuits include amplifiers and oscillators.
Circuits that interface or translate between digital circuits and analog circuits are
known as mixed-signal circuits.
Power semiconductor devices are discrete devices or integrated circuits
intended for high current or high voltage applications. Power integrated circuits
combine IC technology with power semiconductor technology, these are
sometimes referred to as "smart" power devices. Several companies specialize
in manufacturing power semiconductors
< MENU >
A pure semiconductor such as silicon has uniform properties, with the same band
gap and lattice properties. A simple binary semiconductor such as GaAs or InP also
has uniform properties.
You can get other properties by adding one or two other elements from the same
groups of the periodic table to a compound semiconductor. This is equivalent to
blending different materials to get a new compound with intermediate
characteristics. For example, a mixture of 20% AlAs and 80% GaAs has band gap
energy and lattice spacing partway between pure AlAs and pure GaAs. The formula
for such a compound can be written Ga0.8Al0.2As or simply GaAlAs. (The order of
the two elements that replace each other-aluminum and gallium-is arbitrary; some
people write AlGaAs.)
Things get more complex if the compound semiconductor contains four elements.
Depending on the relative concentrations of the elements, such "quaternary"
semiconductors can have properties somewhere in a broad range.
For example, as shown in Figure 9-20, InGaAsP (more precisely InxGa1 xAs1 yPy) can
have lattice spacing and bandgap within an area defined by four possible binary
compounds: InAs, InP, GaAs, and GaP. The material's properties can vary within the
entire space because the In Ga and As P ratios can be adjusted independently. (The
odd shape of the area in the figure arises because pure GaP is an indirect band gap
material, unsuitable for lasers and somewhat different from the other binary
compounds. The dashed line indicates the indirect band gap region.) Plots like Figure
9-20 are very helpful in understanding semiconductor properties; they are commonly
drawn with lattice spacing on the bottom, but some have lattice spacing on the sides
and band gap on the bottom.
The band gaps of different layers in a semiconductor device do not have to match, but
the lattice spacing must be close to avoid flaws that degrade device performance.
This means that all compositions in a bulk InGaAsP structure must fall roughly on a
vertical line in Figure 9-20. The use of thin strained layers relaxes this constraint
somewhat, as mentioned earlier.
Semiconductor laser structures are grown on substrates of simple-to-produce binary
materials such as InP or GaAs, which account for most of the device volume. The
choice of substrate restricts composition of other layers to materials with similar
lattice spacing.
< MENU >
Diode
IC & Transistor
< MENU >
Reference :
http://users.ece.gatech.edu/~gsm/overview.html
http://pvcdrom.pveducation.org/SEMICON/SEMICON.HTM
http://www.classle.net/bookpage/semiconductor-and-pn-junctions
http://www.globalspec.com/reference/13700/160210/chapter-9-11-1-semiconductorproperties-and-composition
< MENU >