Professional Documents
Culture Documents
of Semiconductor Industry
In Malaysia
By
School of Microelectronic Engineering
Presentation Outline
The Evolution of Semiconductor Technology
Industry Structure
Technology Challenges & Trends
Semiconductor Industry in Malaysia
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.
IC Design: 1st IC
1st IC design by hand (Jack Kilby)
Currently, hundreds of designers work
on single product to design, validate
and lay outed will take several months
to complete with the help of CAD
tools.
Main considerations;
performance
die size
design time and cost
testability
Layout
Cross section
Physical Verification
Post Layout Simulation
Mask Design
Fabrication & Wafer Probing
Packaging, Assembly & Test
Hardware
SUN Workstation
IC Product Category
CMOS
BasedTechnology
Industry Structure
Semiconductor Manufacturing
A multi-dicipline processes, involved;
Circuit design
Manufacturing material
Clean room technology, processing, equipment
Wafer processing technology
Die testing
Chip packaging and final test
Design
Services
Mask Making
Wafer
Manufacturing
Wafer
Test
QFP
BGA
Assembly &
Final Test
SDIP
SSOP
Providers
st
IC
IC Manufacturers
Manufacturers MySem, Silterra, 1 Silicon, Infineon
Mask Shops
Fablite Model
Marketing/
Sales
(B2C)
Marketing/
Sales
(B2B)
Design/IP
Systems
Design/IP
Systems
Manufacturing
Manufacturing
Fabless Model
Marketing/
Sales
(B2B)
Foundry
partners
Design/IP
Systems
Manufacturing
In 1990 only 7 fabless companies existed ; Today more than 100 fabless companies exist
worldwide
Many companies such as Motorola, TI, Tosibha, LSI Logic plan to outsource > 50% of its
production
Second & third tier IDMs would be forced to adopt a pure fabless model or fablite strategy to
remain viable
Organic fabless growth fabless growth consistently outpaces overall industry
School of Microelectronic Engineering
Foundry
partners
Why Fabless?
Model allows necessary focus on
system/design level for success
Manage the risk related to the high cost of
building and maintaining a fab
Economies of scale/efficiency
Fabless companies are expected to account
for more than 60% of the total semiconductor
revenues by 2010
Fabless company funding sequentially
increased 62 percent year-over-year in 2004
The fabless sector has continuously achieved faster growth than the
overall industry.
Fabless
(In millions)
Source - FSA
School of Microelectronic Engineering
Moores Law
IC Integration Scale
EXAMPLE
Integration
Level
Cost
Speed
Power
Compactness
Functionality
School of Microelectronic Engineering
1975
1997
2003
10
100
0.25
0.08
30
150
600
50
200
300
60
80
40
85
95
40
85
95
90
99
99
Key Inferences
Continuously increasing
density of transistors on
single chip
More functionality on a
single chip
More yield at lower costs
Reduction of costs through
facility automation
Increase in the number of
chips per wafer
Description
Challenges in design with increased density (reduced line width)
Chips with increasing heterogeneity (integration)
Design
Material
Fabrication
Assembly and
Packaging
Description
Concurrent development of circuits and processes using test
element groups (TEGs) & simulation
Robust circuit and processes factoring manufacturing fluctuations
Defective product analysis and counter measure prototyping by
means of quick turnaround time(QTAT)
Declining number of masks leading to increase in number of chips
per wafer:
Reducing chip size
Expanding the area from which chips can be obtained
Manufacturing of increased wafer diameter of 300 mm
Lower cost materials and processes to meet new requirements
Reliability under thermal cycling (stress and moisture)
Compatibility with harsh environments (automotive)
Increasing reliability for soldering process
New materials for Opto packaging
Defect Reduction
Description
Need for new failure analysis as the traditional failure analysis is likely to
be inadequate due to:
Classification speed of defects
The number of defects that can be handled
Speed of chemical element analysis
UV defect inspection equipment for wafers failing at 130 nm node
Achieving low test costs and high test reliability
New test requirements for technology >100nm
Testing
www.itrs.net
10 year CAGR
between 10% & 20%
Worst ever
Semiconductor
industry downturn
witnessed in 2001-02
Industry witnessed a
ve growth rate of
around 30% during
the downturn
Revival of
semiconductor
industry in 2004
Growth Rate
2003
2004
2005
2006
2007
2008
Semiconductor Capital
Spending
29,661
44,763
50,767
43,058
35,693
39,872
Growth (%)
7.5
50.9
13.4
-15.2
-17.1
11.7
Capital Equipment
22,824
37,317
42,912
35,230
27,806
32,439
Growth (%)
10.3
63.5
15
-17.9
-21.1
16.7
16,742
27,364
31,144
25,848
20,598
23,176
Growth (%)
3.5
63.4
13.8
-17
-20.3
12.5
Packaging and
Assembly Equipment
3,060
4,994
5,114
3,602
2,949
3,988
Growth (%)
30.5
63.2
2.4
-29.6
-18.1
35.2
Automated Test
Equipment
3,021
4,960
6,655
5,780
4,260
5,275
Growth (%)
39.4
64.2
34.2
-13.1
-26.3
23.8
Source : MIDA
School of Microelectronic Engineering
Year
Output
1989
2,262
1990
2,565
1991
2,689
1992
3,121
1993
3,491
1994
3,410
1995
4,757
1996
5,237
1997
7,432
1998
8,951
1999
9,959
2000
16,373
2001
13,524
2002
15,036
2003
15,958
2004
18,228
Semiconductor Exports
MALAYSIA: Exports of Semiconductors
Year
1996
35.5
1997
40.8
1998
54.4
1999
65.4
2000
71.1
2001
60.5
2002
72.9
2003
85.1
2004
89.2
Source : MIDA
Logical
Design
EDA Tools
R&D
Design &
Development
Test
Equipment
Photo-resists
Masking
Fabrication
Product
Design
Lead
Frame
Dicing
Packaging
Physical
Design
Testing
Product
Manufacturing
Precision
Components
Fabrication
Equipment
EMS
Bonding
Chemical and Ultra
pure gases
Consumer
Electronics
Components
Substrate
Wafer
Processing
PC &
Peripherals
Plastic
Molding
End
Product
Industrial
Electronics
Automotive
Bio-medical
Others
Competence to be developed
Wafer Fabrication
Design
MIMOS/Universities
MNCs
Source: MIMOS Analysis based on IMP2 Report
1999
2000
2001
2002
24,612.1
29,681.8
36,411.6
45,232.3
51,132.1
0.5
0.7
5.0
10.0
25.0
92.3
155.8
1,365.4
3,392.4
9,587.3
Projected Indigenous
Design value (RM million)
12.3
20.8
182.1
452.3
1,278.3
Projected Malaysian
Semiconductor Demand
(RM million)
Targeted Indigenous
Product Content (%)
Assumptions
Value of design: 10% of product value
Profit margin: 25% of product value
School of Microelectronic Engineering
National Strategy
R&D Centre
Infrastructure
New materials
Methodology
D & M process
Expertise
Universities
Products
Industry
Technology
Growth
Challenges
Absence of a highly evolved R&D/design environment
Current factor conditions (Education system)
High costs of setting up niche research institution (s)
Close proximity of other technology centers like, Taiwan, South
Korea, Japan, Singapore, Israel, India
Conclusion
Semiconductor technology is a strategic knowledge in the ICT era
Semiconductor industry is the key to the countrys competitiveness
and growth.
Chip design and wafer fabrication are the KSF in securing the country
as a major semiconductor based component / product producer.