Title: Template for SFR Presentations
1(No Transcript)
2Sensors Metrology
- SFR Workshop
- November 8, 2000
- E. S. Aydil, B. Dunn, K. Poolla,
- R. Smith and C. Spanos
- Berkeley, CA
3Sensor Milestones 2001
- September 30th, 2001
- Build and demonstrate Langmuir probe based
on-wafer ion flux probe array using external
electronics. (Aydil) - Design and build a single MEMS based retarding
field ion energy analyzer with external
electronics. (Poolla) - Design and fabricate first generation prototype
MEMS sensor array. Bench test using Joule
heating. (Smith) - Demonstrate cut-and-paste approach for membrane
arrays, LED arrays, and battery encapsulation.
(Cheung) - Develop thermally robust inorganic electrolyte.
Lid added to battery encapsulation scheme. (Dunn)
- Build Microplasma generating system. Test with
bulk optical components. (Poolla, Graves)
4Sensor Milestones 2002
- September 30th, 2002
- Build and demonstrate 8 on-wafer ion flux probe
array in industrial plasma etcher with external
electronics. (Aydil) - Demonstrate MEMS based ion energy analyzer in
plasma with external electronics. (Poolla) - Integrate the inorganic electrolyte into the
battery structure. Develop an in-situ lithium
formation process. (Dunn) - Build micro-optics for spectral analysis.
Complete the preliminary designs for integrated
MOES. (Poolla)
5Sensor Milestones 2003
- September 30th, 2003
- Integration of Si-based IC with sensor arrays.
Characterize and test integrated MEMS ion sensor
array. (Aydil, Poolla) - Battery operation between room temperature and
150C. Battery survivability to sensor soldering
operation. (Dunn) - Design and test integrated MOES. Calibration
studies, sensor characterization. (Poolla,
Graves)
6On-Wafer Ion Flux Sensors
- SFR Workshop
- November 8, 2000
- Berkeley, CA
- Tae Won Kim, Saurabh Ullal,
- Baosuo Zhou, and Eray Aydil
- University of California Santa Barbara
- Chemical Engineering Department
7Motivation and Goals
- Variation of ion bombardment flux and its spatial
distribution with plasma conditions is critical
to plasma etching. - Ion flux uniformity at the wafer determines the
uniformity of etching and etching profile
evolution. - There have been almost no measurements of the ion
flux or ion flux distribution across the wafer as
a function of both r and q in realistic etching
chemistry. - Design, build and demonstrate an on-wafer ion
flux analyzer with external electronics capable
of mapping J (r,q) on a wafer.
8On-Wafer Ion Flux Sensors Milestones
- September 30th, 2001
- Build and demonstrate Langmuir probe based
on-wafer ion flux probe array using external
electronics. - September 30th, 2002
- Build and demonstrate 8 on-wafer ion flux probe
array in industrial plasma etcher with external
electronics. - September 30th, 2003
- Integration of Si-based IC with sensor arrays.
Characterize and test integrated MEMS ion sensor
array. ( with Poolla)
9On-Wafer Ion Flux Probe Array
- 10 probes on 3 wafer
- Evaporated metal on PECVD SiO2 on Si wafer. Lines
insulated by PECVD SiO2 - External electronics based on National
Instruments SCXI platform - The array is scanned at a rate of 1000
Samples/sec (100 Samples/probe/sec) - Lab View Interface
10Ion Flux Uniformity Measurements in an ICP Reactor
50 W
100 W
200 W
- Ion Flux as a function of r and q over the whole
wafer is determined using Kriging extrapolation
between the probes. - Ion flux uniformity was measured in an
inductively coupled plasma reactor in Ar
discharge to demonstrate the probe operation. Qar
8 sccm, P 50 mTorr, Probe bias -70V.
11Plasma Instability J (r,q,t)
t 1.5 s
t 0 s
t 3.2 s
t 4.5 s
12On-Wafer Ion Flux Measurements in a Cl2 Discharge
in Lam TCP 9400 Reactor
- Goal extend the measurements to a commercial
reactor and realistic chemistry.
Measurement Probe (Biased _at_ -75V with respect to
reference)
Heavily Doped Si wafer (Reference)
- Ion Flux in Cl2 plasma increases as a function of
exposure time to Cl2 plasma until it finally
saturates. - Changes in chamber wall conditions is likely to
be responsible for the drift. - SF6 plasma clean resets the chamber back to
reproducible condition.
13Wall Probe
Plasma
Internal Reflection Crystal
Chamber Wall
From FTIR Spectrometer
To HgCdTe Detector
- IR radiation from a spectrometer is directed onto
one of the beveled edges of an internal
reflection crystal (IRC). The IR beam undergoes
multiple total internal reflections from the
crystals surface and emerges from the opposite
beveled edge. In this way, IR spectra of films
and species that are adsorbed on to the walls and
the IRC are recorded.
14Monitoring the Walls During Cl2/O2 Etching of Si
- SiO2 film is deposited on reactor walls from the
reaction of SiClx with O even in the absence of
O2 in the feed gas quartz window or walls can be
the source of Si and O. - Sensor is sensitive to even a few Å of oxide on
the walls. - Sufficiently long SF6/O2 plasma removes the oxide
film from the walls.
15Wall Cleaning/Conditioning Step Influences the
Ion Fluxes in the Subsequent Etching Steps
SF6 cleaning step
SF6O2 cleaning step
- Ion flux and its variation with time depends on
the wall conditioning step - If plasma reactor walls are cleaned/conditioned
with SF6O2 Ion Flux remains steady for a
longer time compared to conditioning with SF6
only.
16Relation Between the Ion Flux, Gas Phase
Composition and Wall Deposits
Ion Flux
Cl SiClx
SiO2 on the Walls
- Ion Flux monitored using ion flux probe
- SiClx and Cl concentrations monitored using
optical emission - Wall deposition monitored using the MTIR-FTIR
probe - Oxygen plasma oxidizes the surface of the wafer
and probe - Cl2 plasma (no bias power) etches the oxide layer
slowly compared to the Si. - Drift in Ion Flux is due to changing wall
conditions and plasma composition.
17Summary
- Designed and build an on-wafer ion flux probe
array with external electronics. - Demonstrated the use of the array for
- mapping ion flux uniformity in an Ar plasma.
- measuring spatiotemporal variation of the ion
flux in presence of a plasma instability. - Completed preliminary experiments in a commercial
reactor.
2002 and 2003 Goals
Build and demonstrate 8 on-wafer ion flux probe
array in industrial plasma etcher with external
electronics by 9/30/2002. Integration of Si-based
IC with sensor arrays. Characterize and test
integrated MEMS ion sensor array. 9/30/2003.
18Lithium Batteries for Powering Sensor Arrays
- SFR Workshop
- November 8, 2000
- Bruce Dunn
- UCLA
- Student contributors
- Nelson Chong, Jimmy Lim, Jeff Sakamoto
19Outline
- Background
- Status at the end of August, 2000
- Present Directions/Future Goals
(C4H5N)n (xlt0.25)
Energy density (Wh/kg)
Cathode material selection
20Operational and Dimensional Requirements
In order to provide on-board power of SMART
wafers, a low profile, thermally stable, high
energy density battery must be used.
Temperature capability 150C. Vacuum (10-2
torr). Operating voltage gt 2.5 V Discharge
current 2mA. Discharge time gt 10 minutes. Low
Profile 500?m or less. Area Less than 3 cm x 3
cm. Rechargeable 10 cycles.
21Status as of August, 2000
Microprocessor
Year 2 Milestone Lithium battery encapsulated
in in wafer well
LED
Thermistor
Battery
Voltage regulator
Thickness profile
22Status as of August, 2000
Key Features
Batteries exhibit good energy density and
cycling behavior Operation at elevated
temperature and under vacuum Epoxy
encapsulation system enables low profile
23Status as of August, 2000
Year 2 Milestone Evaluation of battery
robustness
Excellent cycling characteristics at room
temperature
24Status as of August, 2000
90
Alternating discharge at 85 C/vacuum and room
temperature/atmosphere.
80
70
Discharge
60
Temperature (oC)
Charge
50
40
2 mA discharge current to 2.5 volts.
30
20
0
5
10
15
20
25
30
35
Time (hrs)
T 85 ?C 10-2 torr
Poor cycling behavior after operation at
85C/10-2 torr
Cycle
25SFR Program for 2000 - 2001
- Increase operating temperature to 150oC
- Replace polymer electrolyte with
inorganic electrolyte - a) Sol-gel method
- b) Composite inorganic/organic electrolyte
Both approaches based on confining liquid
electrolyte in fine pore network SiO2 network
provides rigidity Liquid electrolyte gives Li
conductivity
Li liquid electrolyte
SiO2 network
26SFR Program for 2000 - 2001
First results with organic/inorganic system are
very promising
Battery fabricated with new electrolyte very
good discharge characteristics achieved
Fumed Silica R805 particles/aggregates
Electrolyte 1M Li Imide 0.5 cc PC 2.5 cc PEGdm 250
Current work Increasing SiO2 content to improve
temperature resistance
Li conductivity gt 10-3 S/cm Thixotropic
properties
27SFR Program for 2000 - 2001
Improve encapsulation by incorporating a
silicon lid
Lid attached
Encapsulation with low viscosity epoxy
Viscosity400 to 500 cps at room
temperature. Cure time18 to 24 hrs at room
temperature
Encapsulation with 5 minute epoxy Cure time 5
mins.
28Summary and Future Work
- Accomplished Milestones for August, 2000
- Continued Improvements in Integrated Power
Source - a) Higher temperature operation/exposure
- Inorganic electrolyte (Sept. 2001)
- Integrate electrolyte into battery (Sept.
2002) - b) Improve battery fabrication/packaging
- Wafer lid (Sept. 2001)
- In-situ lithium formation (Sept. 2002)
29Microstructures for Temperature Uniformity
Mapping during PECVD
SFR Workshop November 8, 2000 Ribi Leung, Dwight
Howard, Scott D. Collins and Rosemary L.
Smith MicroInstruments and Systems Laboratory
(MISL) UC Davis
30Abstract
- The ever decreasing IC device geometry and
increasing substrate diameters requires high
degree of film thickness uniformity. PECVD rate
is a function of Plasma Chemistry and Substrate
(surface) Temperature. Uniformity depends on
spatial control of process parameters, including
plasma composition (gas flow rates and pressure),
plasma energy (power), and substrate surface
Temperature. Temperature uniformity is critical,
since deposition rate typically follows an
Arhenius dependence. The goal of this project is
to design, fabricate and test T mapping stuctures
for mapping surface Temperature during PECVD as
an aid in process and tool development.
31Milestones(1 year project)
- September 30th, 2001
- To design and fabricate MEMS sensor array to
record surface temperature variations. - Demonstrate in PECVD tool.
32Thin Film Temperature (T) sensor
Metal thin film bilayer resistors
Au/Cr Interdiffusion
- Au/Cr, Al/Au, Al/Cr
- function records accumulated time
- at temperature as increase in R
- mechanism interdiffusion and/or formation of
compound
Q1.13eV
A. Munitz, Y. Komem, The increase in the
electrical resistance of heat treated Au/Cr
films, Thin solid films, 71, 177-188 (1980).
33Metal Bilayer Resistor Pattern
500
µm
R0 670 ?
34Temperature Dependence of R
R/R0
RIE
PECVD
Al/Au/Cr
Au/Cr
Al/Cr
Temperature (C)
35Wafer T Mapping Demonstration
PolySilicon Etch Technics Parallel Plate RIE
- No Substrate Cooling
- fRF 160 kHz
- SF6 , 15 sccm, 150 mTorr
- RF power 200W
36Temperature vs Etch Uniformity
Temperature Map
Photograph of Wafer
hot
cool
Etch Incomplete, t 25 mins
37Temperature vs Etch Uniformity
Temperature Map
Photograph of Wafer
hot
cool
Etch Complete, t 31 mins
38Al/Cr for PECVD 2nd Phase Formation at T 290
R/R0
10
8
6
Al/Cr
4
Al/Au/Cr
2
1
Temperature, C
39Temperature Map PECVD Si3N4, 1200 Å
Al/Cr
Technics PECVD, Platen T 330 C, 10 min, SiH4
NH3
40MEMS Thermal Actuator (A)
d
8 R
Shield
R
d
Secondary Tip Motion
Arc 8 d
- Mechanical displacement with T
- Deflection recorded by masking of deposition by
shield. - Requires T structure gt T substrate.
- Calibration by Joule heating of legs with
injected current.
I
I
800µm
Primary Tip Motion
I
I
41Polysilicon Microhinge
R
MUMPS Chip Photo
d
8R
d
2 Layer PolySi Process
poly1 poly2 anchor
Key
1 mm
42MEMS Thermal Actuator (B)
Al/polySi Bimetal Actuator
PECVD
?T
43This years tasks
- Fabricate and Test PolySi/Al Bilayer Resistors
- Measure R vs Temperature for PolySi/Al
- Measure PolySi/Al Composite Film Stress vs. T
- Design and Fabricate Thermal Actuator
- Demonstrate MEMS thermal actuator in PECVD
44Spatially Resolved Heat Flux Sensor Array on a
Silicon Wafer for Plasma Etch Processes
SFR Workshop November 8, 2000 Mason Freed, Costas
Spanos, Kameshwar Poolla Berkeley, CA
45Motivation
- Plasma etch processes are highly sensitive to
wafer temperature, in terms of etch rate,
selectivity, and anisotropy - Heat delivered to the wafer has two principle
sources ion flux bombardment, and exothermic
chemical etch reactions - Very difficult to measure these two quantities,
spatially resolved, without wafer-mounted sensors - 2001 GOAL Design, build, test array of heat flux
sensors on a silicon wafer, with external
electronics.
46Methods for Constructing Heat Flux Sensors
- Simple, layered heat flux gauge
- Problem for semiconductor dimensions and
materials, ?T is very small
Incident heat flux (q? )
Dielectric, thermal conductivity ?
Temperature Sensors
t
47Possible Solution Thermopile
- Use series connection of many thermocouples to
amplify temperature difference, giving a
measurable output voltage.
- from Holmberg, Diller 1995
48Possible Solution Thermopile
- Benefits
- Sensitivity increases linearly with number of
thermocouples - Can use 100s or 1000s of them ? 1000X
amplification - Problems
- Sensor size is proportional to number of
thermocouples - Typical thermocouple materials are not part of
standard CMOS process ? cant easily combine with
electronics - CMOS thermocouples fabricated from n-poly /
p-poly are an order of magnitude less sensitive - Assumes no conduction along thermocouple leads
may not be a good assumption
49Possible Solution Gardon gauge
- Rotate the heat flow to travel laterally
instead of vertically ? increase the effective
dielectric thickness
Membrane Top View
?T
? depends on diameter squared!
D
Heat flow within thin dielectric membrane
Membrane Side View
Incident heat flux (q? )
Heat flow within membrane
Heat sink
Heat sink
w
?T
50Discrimination of Ion Flux / Etch Exothermicity
- Use two heat flux sensors, one with an exposed
layer of etched material (exposed in diagram)
and the other without this material (covered) - Place sensors into Wheatstone bridge arrangement
- ?
- ? etched material must be low conductivity to
avoid shorting the thermal path across the
membrane
51Proposed heat flux sensor geometry
- Add antenna to funnel heat through the center,
maximizing the temperature difference ?T
b
? now , a factor
10X higher
? now the conductivity of the top etched material
doesnt affect the operation of the sensor
522002 and 2003 Milestones
Demonstrate heat flux sensor in plasma etch
environment, with external electronics, by
9/30/2002. Design wireless heat flux sensor wafer
and demonstrate it in plasma etch environment, by
9/30/2003.
53Microplasma Optical Emission Spectrometer (MOES)
on a chip
SFR Workshop November 8, 2000 Michiel Krüger,
David Hsu, Scott Eitapence, K. Poolla, C.
Spanos, D. Graves, O. Solgaard Berkeley, CA
2001 GOAL to build a microplasma generating
system and test it with bulk optical components
by 9/30/2001.
54Motivation and background
- Motivation
- Precise detection of compounds near substrate
required during semiconductor manufacturing - Organic compounds, emitted during DUV, can coat
optics of stepper - Background
- Small atmospheric pressure glow discharges can be
used for species excitation. - Glow discharge optical emission spectroscopy has
long history in analytical chemistry
55Microplasma Optical Emission Spectrometer
- Basic idea
- OES from plasma reveals info about gas
composition in chamber - Interdisciplinary
- plasma physics and chemistry
- MEMS processing
- optics and metrology
- Inter-departmental
- chemistry
- electrical engineering
- mechanical engineering
56MOES (cont.)
- Generation of plasma with hollow cathode
- Generation of plasma possible if
0.05ltp.Dlt10Torr.cm - Smaller diameter (?75 mm) allows plasma
generation at atmospheric pressure! - This results in smaller sensor
- Many applications in (and outside!) IC processing
industry (for example in lithography)
D
57Schematic of initial MOES experimental
configuration
- Combination of
- Bulk optical optical
- components
- Microplasma chamber,
- fabricated in Si substrate
- Light emitted from
- discharge is captured by
- lens and collimated onto grating
- Diffracted light from grating is
- focused on detector array to record spectrum
58First experiments plasma in 200mm hole, 100Torr
N2 ambient
molybdenum
chip
mica dielectric
vacuum chamber
59Currently fabricated in UCB Microlab
- Relatively simple to make
- XeF2 etch to achieve required depth and undercut
- Very small diameters, i.e. high pressure, possible
plasma
cathode
anode
60Fabrication process and challenges
- Fabrication
- OES cavity defined by deep reactive ion
etching/XeF2 isotropic etch - anode/cathode defined on front and backside of
wafer (metal or doped Silicon) - Challenges
- Microplasma stability and contamination
- Device sensitivity
- Packaging of device
- Exploration of pulsed operation to make
autonomous power supply possible - Integration of micro discharges onto chips for
other applications
612002 and 2003 Milestones
- Build micro-optics for spectral analysis.
Complete the preliminary designs for integrated
MOES, by 9/30/2002. - Design and test integrated MOES. Calibration
studies, sensor characterization, by 9/30/2003.