Nothing Special   »   [go: up one dir, main page]

Chapter 1 - Fundamental To Microlectronic Fabrication

Download as pdf or txt
Download as pdf or txt
You are on page 1of 26

EMT 357/3

FUNDAMENTAL OF
MICROELECTRONIC FABRICATION

Chapter 1
Fundamental to Microelectronic Fabrication.
History and evolution of integrated circuits and
Overview of microelectronic fabrication.
Introduction

• What is Semiconductor Process Technology?


– The technology to produce Integrated Circuit (IC)
microchips
• IC chips are the backbone of the computer industry and have
spurred related technologies such as software and internet
• Every product of the information age is an offspring of IC
technology
• IC chips increasingly control functions in cars, TVs, VCRs,
cameras, mobile phones, toys, etc.
• The current technology is as a result of years of research and
development, taken many thousands of scientists, engineers
and technicians.
History
History of Semiconductor Devices
• First Transistor, Bell Lab 1947
• John Bardeen and Walter Brattain, Collector
demonstrated a solid state device
made from germanium. They
observed that when electrical
Emitter
signals were applied to contacts
on germanium, the output power
was larger than the input.
• These results were published in
1948.
• William Shockley, found out how
the bipolar transistor functioned
and published the theory in 1949.
• Three of them shared the Nobel
Prize in physics in 1956. Base
First Transistor and Its Inventors

They found amplification phenomenon when investigating Ge surface when putting


needles.
This is the Not Field Effect Transistor,
But Bipolar Transistor (another mechanism)
 Semiconductor industry developed rapidly and
germanium based transistor quickly replaced vacuum
tubes in electronics equipment due to:
• smaller size
• lower power consumption
• lower operating temperature
• quicker response time

 Single crystal silicon and germanium based devices


introduced in 1950 and 1952 respectively (better defect
control, hence higher yield).
Shockley left Bell Labs in 1956, to start his own
lab in San Francisco Bay, California. Nowadays
known as Silicon Valley. His lab has attracted
talented scientist such as Robert Noyce and
Gordon Moore.

Gordon Moore and Robert Noyce left Shockley


in 1957 to start Fairchild Semiconductor.
1958: 1st Integrated Circuit
Demonstrated functional IC, fabricated on germanium strip consists of:
• One transistor
• One capacitor
• 3 resistor
Jack S. Kilby

The first integrated ciruit chip (bar) made by Jack Kilby


1960: First MOSFET by D. Kahng and M. Atalla

Top View
First Silicon IC Chip by Robert Noyce, Fairchild Camera, 1961

Fairchild Semiconductor produced the 1st


commercial ICs in 1961. This IC consists of
only 4 transistors sold for USD 150 a piece.

In 1968, Robert Noyce cofounded Intel


Corp. with Andrew Groove and Gordon
Moore.

He built the first IC using silicon – si forms a


naturally occurring oxide – SiO2 can be
used as an insulating layer for MOSFET.

Monolithic IC – made from same wafer


Number of transistor has increased while at the same time
the size of transistor have decreased.

2 types of improvements
(1) Process improvement – fabrication with smaller sizes/dimentions.
(2) Structure improvement – new design for greater performance

Integrated Circuits (ICs): size of component (feature size)


and number of component.
Feature size & Wafer Size
Feature size : smallest dimension in the device

In less than 40 years, the minimum feature size of IC chips has shrunk dramatically, from about 50 microns in the 1960s
to 0.18 micron. By reducing the minimum feature size, one can make a smaller device. – more chips per wafer, or more
powerful chip with the same die size.
More components will be made on a chip and more chips made on wafer, and unit price for a transistor will continue
to decrease.
A Note on “Small”
Moore’s Law
In 1965, Gordon Moore, a co-founder of both Fairchild and Intel, described in his paper,
based on five data points available from Fairchild at that time:

‘with unit cost falling as the number of components per circuit rises, by
1975 economics may dictate squeezing as many as 65 000 components on a
single silicon chip’.

The number of components on a computer chip doubled every 12 month, while the price still same.
SCALING
100 1010
1.0E+10
Moore’s law and
Dennard’s scaling 10 9

CPU Transistor Count


1.0E+09
10 rules
10 8
Size (µm)

1.0E+08

10 7
1 1.0E+07

106
1.0E+06
130nm
0.1 105
90nm 65nm 1.0E+05
45nm 4
32nm 101.0E+04
22nm
0.01 3
10 1.0E+03
1970 1980 1990 2000 2010
• Dennard 2020
Scaling: as transistor shrinks it gets
The size and speed are related – the smaller – Faster
something is, the quicker it can be changed Year – Lower power (constant power density)
– Smaller/lighter
♦Thus, smaller transistors can switch at • Moore’s Law
higher speeds – Keep the cost/area about constant while shrinking
– More transistors/chip & lower cost/transistor
Moore’s law and
Dennard’s scaling rules

ENIAC 1946
 Lighter
 Cheaper
 Diverse functionality
 Longer battery life
 etc

W. Arden et al. More


than Moore, white paper
(ITRS)

Virtuous cycle of the semiconductor industry

Alexander Graham Bell ITRS executive summary:


1876 ‘a basic premise of the roadmap has been that continued
of electronics would reduce cost per function …… and
promote the market growth for integrated circuits’. 18
Scaling ?

2011
The switching speed of CMOS circuits is inversely
proportional to the drain current of the transistor:

In order to increase the circuit speed, IDsat has to be


increased
IDsat increase is achieved by reducing tox

MOSFET scaling improves Idsat, lower Cgate and reduces Vd to provide 30-40%
improvement per generation
IC Integration Scale
The scales of the integration level for the semiconductor industry.
Evolution of Industry (Intel)
Year Clock speed Num. of Technology
Transistor (Feature size)
4004 1971 108 kHz 2300 10µm
8086 1978 5 MHz 3x104 3µm

486 1989 25 MHz 1x106 1µm

Pentium IV 2000 1.5 GHz 4x107 180nm

Xeon 2007 > 3 GHz 8x107 45nm

45nm 32nm 22nm 11-14nm


(tri-gate) (future)
Limit of IC size??

• Size of an atom
What is the maximum wafer size??
No one knows for sure yet.
It is limited by many factors,
such as crystal pulling and wafer slicing
technologies, process equipment
development, and the demands of IC
fabrication.

Starting from 10 mm (2 inch) in the 1960s,


the wafer size has increased to 300 mm (12
inches).
Based on the history of semiconductor
technology, the silicon-based IC industry will
probably reach the final physical limit in the
2040s.
By that time, the wafer diameter could be
800 mm (32 inches).
Wafer size and Litho costs

Patterning process, which is the most fundamental IC manufacturing step used to transfer
The designed pattern to the wafer surface to form IC devices.
Refreshment 1

Make a list of 5 items in your everyday environment that you believe


contain IC chips..
Make a list of TWO DOZEN items in your everyday environment that you
believe contain IC chips. A PC and its peripherals are considered to be
one item.

Radar detector
Answering machine Radio
Alarm clock Satellite receiver/decoder
Automatic door Security systems
Automatic lights Garage door opener Smoke detector
ATM GPS Stereo system
Automobile: Hearing aid Digital watch Amplifier
Engine controller Inkjet & Laser Printers Digital scale CD player
Temp. control Light dimmer Digital thermometer Receiver
ABS Musical greeting cards Digital Thermostat Tape player
Electronic dash Keyboard synthesizer Electric guitar Stud sensor
Automotive tune-up equip. Keyless entry system Electronic door bell Telephone
Bar code scanner Laboratory instruments Electronic gas pump Traffic light controller
Battery charger Model airplanes Exercise machine TV & remote control
Calculator Microwave oven Fax machine Variable speed appliances
Camcorder Musical tuner Fish finder Blender
Carbon monoxide detector Pagers Drill
Cash register Personal computer Mixer
Cellular phone Personal planner/organizer Food processor
Copier Fan
Cordless phone Vending machines
Depth finder Video games
Workstations

You might also like