MadSci Network: Computer Science
Query:

Re: How does a smaller fabrication process allow faster processors?

Date: Thu Dec 2 08:35:09 1999
Posted By: Daniel Hadad, , N/A, Motorola, Inc.
Area of science: Computer Science
ID: 944131703.Cs
Message:

The micron size referred to is the critical dimension of the fundamental
transistors used in the processor.  This fundamental transistor is called a
MOSFET, which stands for Metal Oxide Semiconductor Field Effect Transistor. 
A MOSFET is a kind of switch, where the current between two terminals
called the source and drain are controlled by a voltage on the third
terminal, called the gate.  When the gate voltage is high enough, then the
transistor turns on.

A critical dimension of the MOSFET is its gate length, which is a factor in
how close the other two terminals (source and drain) are to each other. The
smaller the gate length, the closer these two terminals are to each other,
the stronger the electric field and thus the higher the current through the
MOSFET when it is "turned on".  When the electric field and current is
higher, then MOSFET takes less time to turn on.  

These MOSFETs are connected together in certain patterns to carry out
logical operations in the processor, like the boolean algebraic operations
of AND, OR, NOT, NAND, NOR, XAND, XOR, etc.  These blocks of MOSFETs that
perform boolean functions are called logic gates ("gate" is a word used in
many ways in electrical engineering).

The clock speed of a microprocessor is directly related to how fast these
logic gates operate.  The higher the current in the fundamental MOSFETs,
the lower the delay in switching, the faster the clock speed of the
processor.

Fabrication facilities are constantly striving to acheive lower and lower
gate lengths to get higher and higher microprocessor speeds.  Semiconductor
manufacturing uses a process called photolithography to pattern the devices
that end up in a part like a microprocessor.  This patterning is not unlike
the photographic developing process, where light is shined through a
negative and developed on paper.  Photolithography (which in greek means
"writing on stone with light";photo=light,litho=stone,graph=write) involves
shining ultraviolet light through a similar negative (called a mask) onto a
photosensitive layer added to silicon wafer. This photosensitive layer can
be selectively removed using an acid that only removes the parts that were
exposed to the UV light. Once this photosensitive layer (called
photoresist) is patterned, the underlying silicon can also be selectively
removed with acid that only eats away the silicon not covered by
photoresist.

The resolution limits of photolithography are continuously being pushed by
processes that tout 0.25um technology, or even 0.15um technology.
Photolithography is limited mainly by the wavelength of the light used,
diffraction effects that happen at such small dimensions, and the physics
behind the acid etching that takes place on both the photoresist and
silicon layers.

Needless to say, this is a key issue many semiconductor companies deal with
on a daily basis.

Some useful keywords for further research include: MOSFET, gate delay,
photolithography, photoresist, gate length, short channel effect, effective
channel length, clock speed, etch, silicon wafer.


Current Queue | Current Queue for Computer Science | Computer Science archives

Try the links in the MadSci Library for more information on Computer Science.



MadSci Home | Information | Search | Random Knowledge Generator | MadSci Archives | Mad Library | MAD Labs | MAD FAQs | Ask a ? | Join Us! | Help Support MadSci


MadSci Network, webadmin@www.madsci.org
© 1995-1999. All rights reserved.