6. 6
Patent Issued
1. A particle controller, comprising: an input port
configured to receive a particle stream; a
semiconductor cell comprising a cavity through which
at least a portion of the particles comprising the particle
stream is directed; and one or more electrodes
coupled to the cavity and configured to facilitate
creation of an electromagnetic field for directing the at
least portion of particles through the cavity; wherein the
cell is part of a set of semiconductor cells whose
cavities are aligned to form a tube through which the
at least portion of particles is directed
SAI confidential
8. 2D Single Chip Particle Accelerator
Cornell University
DARPA funded project 2011
Designed by MEMS Dept.
Single beam per chip
Energy of 30 Kev
Proves that an IC can accelerate a
particle beam at high energies.
Proves beam deflection~50 degrees
Demonstration of high acceleration
value
10. 10
Silicon Accelerator – how it works
Changing electric field
Accelerating region in IC
Drift tubes of equal
lengths
Frequency and phase
of each stage under
digital control of
I.C.’s
17. 17SAI confidential
Silicon Accelerator
Summary
Solid State Linear Particle Accelerator
Enabled by 3D SIC and TSV (through silicon via)
3D SIC per accelerating stage
Each 3D SIC contains
• Accelerating electrodes
• Drift tubes
• Electrostatic lens
• Digital and Timing controls
• Sensors
• Scanning electrodes
19. 19SAI confidential
Semi Manufacturing
In Crisis
Fab capital cost at 14 nanometers >$10 Billion
Next Generation Steppers (EUV) >$100M
Mask sets approaching $10M
FinFet transistors at 25 nm going to 10 nm
Wafer size increasing to 450mm
Consolidation of ~4 Major Fab Manufacturing at
14 nanometers
IC product volume threshold >millions of units
20. 20
Nanobeam Implanter
A Million Beams
Current Ion Implant NBI
Single Beam ~cm
Bream Array
Nanometer beams
Silicon Accelerator
SAI confidential
21. 21SAI confidential
Silicon Accelerators
Bandwidth
Massive Beam Array:
1 cm chip at 10 micron pitch ----1 Million beams
Electrostatic micron-sized lens ----Gigahertz scanning
Bandwidth is million beams times Gigahertz per beam
Embedded Digital processing & EDA database
Fully automated and robotic wafer handling
22. 22
Nanobeam Implanter
IC Doping Comparison
Current implant
Method
NBI
SAI confidential
Photo-lithography Digital-lithography
24. Maskless electron beam lithography has the potential to extend semiconductor
manufacturing to the sub-10 nm technology node. KLA-Tencor is currently developing
Reflective Electron Beam Lithography (REBL) for high-volume 10 nm logic (16 nm HP). This
paper reviews progress in the development of the REBL system towards its goal of 100 wph
throughput for High Volume Lithography (HVL) at the 2X and 1X nm nodes. In this paper we
introduce the Digital Pattern Generator (DPG) with integrated CMOS and MEMs lenslets that
was manufactured at TSMC and IMEC.
KLA
The lenslet consists of a densely packed array of 4µm
deep cylindrical holes with a 1.4 µm diameter and top
spacing of only 200nm. The electron beam entering the
lenslet holes is focused through a set of 4 ring electrodes.
The ring electrodes can be tuned to focus the
electron beams by applying static voltages up to 50V
on the ring electrodes. The bottom of each hole
consists of a small metal plate that can be switched by a
CMOS circuitry below, either reflecting or absorbing the
incoming electrons. In this way, the incoming electron
beam is split into 1 million smaller beamlets, a strategy
designed to enable higher throughput for the e-beam
writing process through parallelization.
Read more at: http://phys.org/news/2012-11-imec-
customized-lenslet-array-kla-tencor.html#jCp
25. 25SAI confidential
NBI
Market Opportunity
Worldwide Semiconductors market is >$300
Billion
Served by Equipment Market >$50 Billion
• Fab equipment: new and upgrades
• Back-end: Test and Assembly
Fab Segments impacted by NBI
Mask making
Photo-lithography: steppers
Resist: Track systems
Ion Implant
26. 26SAI confidential
NBI
Economic Potential
No Tooling cost: eliminates mask cost tooling
• Small production lots
• Low prototyping cost
• Customized even a few chip per wafer
Lower Fab Capital Cost
• Smaller Fabs economical: $millions versus
$billions
• Better clean room utilization: smaller footprint
• Lower Fab Inventory; less inventory risk
• Fewer Processing Steps: higher yields
27. 27SAI confidential
Performance Impact
Multiple processes simplified
• DRAM+Logic+Flash+Analog
Mixed Technologies Practical
• MEM's,LED,Laser,DLP
Improved Analog
• Wide materials selection for
– Resistors, super-capacitors
Transistor Structures
• different depth and doping across wafer
Advanced Technologies
• Graphene transistors, magnetoresistive RAM
28. 28SAI confidential
NBI
Summary
New Methodology of Semiconductor
Manufacturing
− Nanobeam Ion Implantation (NBI)
Million beam Silicon Accelerators
High bandwidth Digital Lithography
− Replaces Photo-lithography
Lower cost
Higher yields
High IC performance
31. 31
Fusion
Light elements—hydrogen—combine into helium
No radioactive by products—no meltdown possible
First discovered in 1930 using linear accelerator
Fission splits heavy elements---Uranium
---radioactive isotopes are by products
Research to develop Fusion Engine began in 1950
Fusion requires a hot dense compressed plasma
32. 32
Fusion
Light elements—hydrogen—combine into helium
No radioactive by products—no meltdown possible
First discovered in 1930 using linear accelerator
Fission splits heavy elements---Uranium
---radioactive isotopes are by products
Research to develop Fusion Engine began in 1950
Fusion requires a hot dense compressed plasma
33. 33SAI confidential
Fusion
Why fusion been so hard to achieve?
Plasmas expands:
• No physical container possible: extremely hot
• Reaction time ~ plasma density
Hot Plasma loses energy:
• The electrons radiate light when hot
• Fusion energy must exceed loses
How to compress plasma at >100 million degrees?
35. 35
National Ignition Facility
World's most powerful Laser system: 192 Laser beams
~2 million joules at 500 Terra-watts
Inertial Confinement Fusion
Millimeter diameter hydrogen fuel pellet
Idea is to heat and compress fuel with laser beams
Fuel failed to ignite due to:
Poor beam uniformity, jitter, coupling inefficiencies
NIF funding for fusion ignition dropped
40. 40SAI confidential
Nanofusion
Millions of beams focused into nanometer region
– Uniform compression of hot plasma
– Sub-picoseconds timing reduces beam jitter to nanometer
– Ion energy of 100K electron volts = 160 Million degrees
– Density of plasma is sum of beam densities
Fuel is hydrogen and boron
– Converted to fast moving ions of helium
– Energy of helium ions re-converted into electricity
Nanofusion is a portable power source
– About the size of basketball
41. 41
A few Other Apps
SAI confidential
Nano Technology Cancer Therapy Holography
Data Archive Quantum ComputingInstrumentation
42. 42
Member Role History
Alok Mohan Executive Leadership NCR-VP
SCO-CEO
Sam Brown Technology strategy NCR—Microelectronics
Alpine Semi-CEO
Tom Brummet Business Development NCR---Microelectronics
Silego Semi -VP
Marketing
Marcelo Martinex IC Design Principal
Advanced Analog Design
Jonathan Wurtele Technical Adviser Berkeley
Professor of Physics
Senior Scientist LNL
Ed Pheil Technical Adviser General Dynamics
Nuclear Engineer
John Bryant Technical Adviser Atmel: VP Marketing
43. 43
Next Steps
Printed Circuit Board
Identify Semiconductor Partner
Expand Team
Release Analog IC
Release Digital IC
Nanobeam prototype
Release Development Kit
SAI confidential
44. 44
Confidence Factors
1. Manufacturing: Very High-Processes are In Production
2. Competition: No Direct Competitor at this Time
• Strong-Broad Patents-Trade Secrets
3. Engineering: Digital IC~Block Diagram complete. Analog
IC: critical circuits simulated. Need to Identify Partner
4. Theory of Operation: Proven in 2D chip
5. Market Entry: Acceptance of Development Tool-Intel's
Microprocessor Model
6. First Revenues:
• Now Partnership R&D Licenses
• Development Systems 18 Months
Notas del editor
Silicon Accelerators is the term for a new technology. It is also the name of the corporation. This presentation a conceptual view of technology, but we will touch on the business side. The Silicon Accelerator is a new invention, the patent was issued in 2011. Since that time, research has focused on engineering development and applications. The first prototypes are under development. A Silicon Accelerator controls charged particles. Charged particles include electrons, protons, sub- atomic particles, ions, and electrically charged molecules.
The presentation aims to provide answers to basic questions. The underlying science includes plasma physics, IC processes, electric field analysis, and classical mechanic Specific applications requires nuclear physics and quantum mechanics.
The popular image of particle accelerators is that of large scientific accelerators. The Large Hadron Collider is the world's most powerful synchrotron, made famous by the discovery of the Higgs particle. It is the most powerful accelerator in the world, hurling protons to 3.5 Teravolts. The particles build gain energy as they circle through the track. Powerful super-cooled magnet hold the particles within the track. The Stanford Linear Accelerator is the world's longest-- familiar to commuters passing over on the 280 freeway. The SLA accelerates electrons to Gigavolts along what has been called the world's straightest line. All particle accelerators are either a linear accelerator or a synchrotron. There are approximately 10,000 accelerators installed worldwide. The three major segments using accelerators are scientific, medical, and manufacturing. For the most part, the accelerators are customized and manufactured one at a time. The underlying technology of the installed accelerators is primarily vacuum tube amplifiers, discrete semiconductor components, and electromagnets. A relatively small commercial particle accelerator is the about the size of a refrigerator.
A silicon accelerator is a miniaturization of linear accelerators based on an advanced integrated circuit technology called 3D stacked IC. The 3D SIC process is based on making electrical connections from the front to the back of the chip using micron diameter holes. Through Silicon Via (TSV) VIA processing is recently entered volume production and is available from several of the major foundries, including TSMC and IBM. For a pitch of 10 microns, a centimeter chip can have a 1 million TSV. The area surrounding the TSV contain the transistors circuits of the Integrated Circuit. The full range of IC can be integrated: microprocessors, memories, logic, analog, etc. Ina Silicon Accelerator, the TSV, left open, form the pipes to contain charged particles. The charged particles can controlled and manipulated electric fields present in the metal layers of a proprietary Analog IC design.
Just advances in IC density enabled putting a computer on a chip---the Microprocessor, 3D SIC enabled the invention of the Silicon Accelerator. It takes it place as a fundamental invention, improving with time, aka “Moore's Law”, continuing to find new applications, and creating new market opportunities.
The invention is covered by broad patent coverage. There are 20 claims. As the first claim states, the patent covers any IC based particle accelerator.
We now present an overview of the internal operation of the silicon accelerator.
The breakthrough was in orienting the path of the particles perpendicular to the surface of the IC. IC are limited to in size by yield to a few square centimeters. Previous attempts to build IC based particle accelerators were severely constrained by IC size and processing. The novel architecture leads to precise control of matter at the nanometer range.
Electrically charged particles, enter the accelerator from left. Electric fields are switched on between electrodes when the particles are present. As the particles transverse the gap, they experience a force proportional the voltage between the electrodes. Rather than increase the lengths to compensate for the increased particle speed, digital logic adjust the timing of the fields. Variable timing allows the same 3D SIC to be used in all stages. Digital controls also accommodates particles of differing mass and charge.
In a 3D IC, Through Silicon Via (TSV) are holes which are etched from the front to the back side of a silicon chips. When the chips are stacked, the holes in the chips are aligned. The aligned holes become the pipes through which the particles can pass. Some of the TSV can be filled with conducting material to create electrical connections. The design can be optimized by combining digital processes and analog processes.
The accelerating electrodes are formed using the metal layers of the active layers. The electrodes are connected to high voltage transistors. The diameter of the pipes are 5 microns on a pitch of 10 microns. This image is not to scale as an actual IC contains 1 million pipes.
The beams are focused using electrostatic lens. The electrostatic lens are formed similar to the electrodes. This image is from a software simulator, Simon. Simon uses the Laplace equations to solve the electric field equations to high accuracy. The voltages on the lens elements can be adjusted under program control to accommodate a wide variety of particle types.
Each 3D SIC corresponds to a single accelerating stage. The slide shows two accelerating gaps separated by the drift tube. An electrostatic lens is positioned inside the drift tube.
Many stages can be concatenated to increase the power of the accelerator.
Its now projected that for the first time, the cost on a transistor basis may increase. The high development of advanced ICs limits innovation. Process complexity lengths production cycle time.
Placing a silicon accelerator in the beam of an ion implanter splits the single beam into a million separate beams.
A million beams provides the data bandwidth for a wafer throughput of 120 wafers an hour, competitive with the productivity of steppers
Openings in photo resist allow ions to penetrate the wafer. NBI directly implants transistor features without the need for photo-lithography.
Digital Lithography compared to photo-lithography is as film photography to digital photography. The nano sized beans from the silicon accelerator are modulated under the control of embedded digital processors.
The conversion to NBI maintains Moore's Law and innovation with low-cost new IC designs.
ICs made with NBI have greater performance, more features, lower cost.
In the core of stars, gravity give rise to extremely high density and temperatures necessary for fusion. Fusion converts matter to energy: E=MC 2 /
The National Ignition Facility is a $4 billion facility located in Livermore, Ca. The facility has been funded through the Department of Energy. The justification for the facility included research tied to the maintenance of the strategic stockpile.
The incoming plasma from the millions of particle beams create a plasma sphere collapsing into the target region. The collapsing plasma contains the hot plasma core region of a few hundred of nanometers.
The boron eleven contains 5 protons and 6 neutrons in its nucleus. The positive charge between the protons creates the repulsive Coulomb force. The kinetic energy of the fast moving particles can bring the proton within range of the attractive force of the nucleus, the Strong Force. The proton is captured by the nucleus. The addition of the proton is unstable and splits into three helium nucleus