First 30 years of MBE System Design

1
First 30 years of MBE System Design

• RUSNANOTECH Forum Moscow October 6th, 2009
• Pierre Bouchaib

2

• Riber started to design MBE systems in the mid of
  70s.
• Why ?
• Because Molecular Beam Epitaxy was the first
  definite application which need absolutely Clean
  Ultra High Vacuum ( a pressure range 1015 times
  lower than the pressure of this room) and Riber
  was mastering these pressures every day in its
  systems for analysis purpose.
• Because very soon we were in contact with the
  pioneers like Al CHO in this new method of
  atomic scale growth of semiconductors.
• As an European company we were not focused in
  only one geographical direction, but also Europe
  and Asia.
• We have listened to the pioneers of the
  techniques and translated their needs in reality
  thanks to an enthusiastic team of designers held
  by Claude Buisson, a master in UHV design.
• We have believed the MBE will be sooner or later
  a powerful mean to produce the sophisticated
  devices of the XX century end.
• Because MBE process is a non equilibrium process
  , this means we can force nature to produce
  improbable structures quite easily compared to
  other techniques

3
MBE 500 half MBE half Surface Analysis
4
EARLY STAGES OF MBE SYSTEM

• A big vacuum system with all available surface
  techniques to try to understand the MBE processes
• A process capability few cm2 of AsGa every week (
  no load lock)
• A not so good vacuum cleanliness ( lack of
  cryopanels and load lock )
• Quite difficult to use and maintain

5
Artistic (?) view of an early MBE research lab
6
The big changes of MBE systems at the 80s
endMBE 1000

• Loading chamber to save time and large cryopanels
  to get the best cleanliness
• Rotating substrate to get /- 1 uniformity in
  composition and all over 3 substrate
• Few experiments a week to several per day were
  then possible
• From low electron mobility to the huge ones (
  can be 7 106 cm2/Vs after few runs !! at PTB)
• Preparation chamber capability to
  clean the substrates (desox ..)

7
MBE 1000
8
The bright idea modularityMBE32 MODUTRAC

• With the introduction of the modular concept,
  Modutrac MBE It was then possible to offer to
  the research community an ideally adapted to
  each situation system at cost and delivery of a
  standard system
• Our motto was We make Your Dream a reality

9
MBE 32 P Modutrac with double chambers
10
MBE 32 P Modutrac with 15 separate chambers
11
Gas-source MBE
12
Compact 21 / Max. diameter 3"
13
In the 90 multi wafer platen systems were
designed

• Large number of effusion cells ( up to 12)
  needed for new structures with 4 group V and 2
  group V (AlGaInSbAsP plus 3 dopants)
• Large cell up to 20 liter of Arsenic , or 10 Kg
  of Gallium !
• From 13 x 2 to 4 x 8
• Full process automation

14
Epineat / Multi-2"
15
MBE 49 / Multi-4"
16
MBE 6000 / Multi-6"
17
MBE 7000 / Multi-8"
18
PRODUCTION SYSTEMS
Throughput 15 000 x 6
19
What are the production capabilities of such
systems ?Example for a p-HEMT

• They works 24H/day for a 12 to 15 months run
  without any stop !
• They need only few weeks of maintenance reload
  and calibration per run or more than 80 uptime
• They are fully automated
• They are environment safe (no lethal gas are
  used)

20
Some device structures

• p HEMT for RF application ( III V GaN )
• II VI deep IR detectors
• VCSELs - Quantum Dots lasers ( III V)
• High power diode Shottky diodes (GaN)
• Hall effect sensors
• QWIP QCL Terahertz lasers
• etc..

21
A World Leader Position
The largest number of customers in Research and
Production
Research 600 250 customers
Universities Research Institutes Research
center
The Whos who of the industry
Production 150 21 customers /31 - 68
22
The next future (already in b tests)

• MBE system for large wafer fulfilling the
  SemiStandards
• Full factory automation
• Multi modules arrangments
• The MPVD300

23
SEMI STANDARD MPVD300
24
MPVD 300 PROCESS MODULES SCHEME
Cluster tool
EFEM
25
CONCLUSION

• A very sophisticated laboratory tool has become a
  very powerful tool for fabrication of the most
  sophisticated devices in electronics and
  optoelectronics
• From few cm2 per week in 1975 the MBE tool can
  produce 1000 M2 per year without any stop over 12
  to 15 Months ! ( no other semiconductor process
  tool in any field can do it !)

26
CONCLUSION

• The famous physicist Al Cho told me several years
  ago
• With MBE we are only limited by our
  imagination
• In our company the main raw material we like to
  work on every day is
• the Future, the Innovation.
• Thank you.