Transcript Wafer Level Reliability

Wafer Level Reliability
www.fairchildsemi.com
1
FSC Worldwide Quality & Reliability
Built-In Reliability Initiative
System
Assembly
PC Board
Assembly
Control Points
for Emphasis
Development
Manufacturing
End-Use
Customer
Component
Manufacturing
Raw Material
Supplier
(i.e. Leadframe)
2
Built-In Reliability Approach
Failure Mechanism Driven Reliability Characterization
Methodology
1. DEFINE WORST
CASE USE
ENVIRONMENT
• Temp, RH, # On/Off
Cycles
8. PERFORM
MATERIAL
ANALYSIS
• Dry vs. wet Tg, CTE1,
CTE2, Young’s Modulus
9. PERFORM
PKG / PROCESS
SIMULATIONS
• Thermal, electrical,
mechanical,
3. DEFINE
POTENTIAL
RELIABILITY
FAILURE
MECHANISMS
2. DEFINE QUALITY
AND RELIABILITY
EXPECTATIONS
• Max. ppm allowable,
product lifetime,
cumulative ppm
7. PERFORM
CONSTRUCTION
ANALYSIS
• Demonstrated mfg.
construction quality,
compliance to design rules
10. ALR / WLR
TEST
CHARACTERIZATION
• Corrosion,
electromigration,
intermetallic growth
6. DEFINE STRESS
LIMITATIONS
• Max. temperature, max.
current density
11. FINALIZE POR
(Process of Record)
• Gate Oxide QBD, Hot E
Gm degradation, bond
shear, DEA
3
4. DEFINE
RELEVANT
RELIABILITY
MODELS
• Arrhenius, Eyring,
Coffin-Manson
5. IDENTIFY
ACCELERATING
FACTORS
• Temperature, voltage,
current density, RH
Coffin Manson Reliability Model
TMCL Test Conditions
T (high)
T (low)
150
-65
C
C
TTF
Cycle/hr
n factor
500
2
3
Cycles
* Note: enter data inputs into appropriate grey shaded fields
Cycles / Day
Estimated Lifetime at Use Conditions (yrs)
1
2
3
4
5
20
1701.8
850.9
567.3
425.4
340.4
40
212.7
106.4
70.9
53.2
42.5
Temperature Delta (C)
60
80
63.0
26.6
31.5
13.3
21.0
8.9
15.8
6.6
12.6
5.3
4
100
13.6
6.8
4.5
3.4
2.7
120
7.9
3.9
2.6
2.0
1.6
Agenda
• What is Wafer Level Reliability (WLR)?
• Advantages over traditional Reliability
• Design for Reliability
• Develop for Reliability
• Manufacture for Reliability
5
WLR Definition and Advantages
• Reliability: The probability that a component will perform a specific
function under specific conditions for a specific period of time.
• Wafer Level Reliability uses mechanism—specific test structures under
accelerated conditions to assess the reliability impacts of process and tool
changes
• Advantages over Conventional Reliability
• Can be done at wafer level—doesn’t need packaging
• Accelerated conditions (i.e. voltage, current, temp) speeds-up testing
• Specific failure mechanisms can be analyzed individually
• Results are process based and can be applied to multiple products
• Reduced time to market for new products
6
Examples of WLR Applications
• HC Degradation:
• Splits: 3 spacers and 5 LDD
dose splits—15 splits; each
split needs 2 or 3 weeks HCS
in PL; total: 20 weeks or
more; new technology ~ 9
months to a year!
• Need WLR!!! Compare the
degradation rate between
PLR and WLR. Each split just
needs one hour stress, total
15 hours
• Payoff: 15 hours vs. 3000
hours; WLR: 200 times FAST
7
Reliability Objectives
• Allow more aggressive performance without increased
failure rate
• Eliminate reliability concerns prior to qualification
• Meet customers demands and expectations
• How long should product last (i.e. 2 years to 20 years)
• Expected failures in time (i.e.1 FIT =1 failure in 10^9 power on hrs.)
8
WLR Applications
• Infant Mortality:
Device failures due to built-in
defects. WLR stress can help to
determine the root cause. The WLR
stress: such as PIC, Dielectric BV,
Qbd test
• Wearout Failures:
The failure rate is increasing. The
failures are caused by mechanisms
such as electromigration, time
dependent dielectric breakdown,
and hot carrier degradation.
9
WLR Correlation
• If there is no statistical difference in
the performance of subsequent
product---subsequent product will
have the same reliability as the
qualification product.
• Otherwise the subsequent lot is
anomalous material needing further
engineering study.
• Build up the correlation between
WLR and PLR.
• Effects of assembly and packaging
• The figure for EM of 12-inch and 8inch wafers from a supplier.
10
WLR Objectives
• Design for Reliability
• Develop tools for and consult with Designers to maximize
performance of new products without impacting reliability
• Develop for Reliability
• Use mechanism-specific test structures under accelerated
conditions to eliminate reliability concerns prior to
qualification
• Manufacture for Reliability
• Use the same test structures to make tool and process
improvements and act as a line monitor
11
Design for
Reliability
12
Design for Reliability
• Provide the reliability data to designer
through the reliability calculator.
• Hot Carrier, Mobile ion, TDDB, EM
• Updated on an ongoing basis
• Assess reliability risk for new designs and
update the reliability design rules:
• Design needs to increase operating
voltage in gate oxide, current density in
metal, and operating voltage on caps
(such as HV poly sink caps).
• How to screen out the defect for gate
oxide extrinsic failure.
• Consulting available upon request
• Develop AC hot carrier stress Methodology
• Confirm Berkeley Model for HC
degradation in AC application.
13
Develop for
Reliability
14
Gate Oxide
Integrity
15
GOI Test Methodology
JRAMP
• J-Ramp test searches for the breakdown voltage
(Vbd) and then calculates the breakdown charge
(Qbd) of thin oxide capacitors.
JRAMP Test Sequence
• Makes timing measurements of all procedures
used during the ramp
• Checks the integrity of the structure to be stressed
(leakage current test at use voltage)
• Logarithmically ramps current (5 pS ramp rate)
• Records current, voltage, and time at each
step/interval
• Terminates the test when
- Voltage compliance is reached
- Max Qbd is reached
- Next current step (if allowed) exceeds max
current density
- Elapsed time exceeds pre-set max test time
- Percentage decrease in measured voltage
from previously measured voltage exceeds
specified ratio.
• Calculates outputs/results
16
Plasma Induced
Charging (PIC)
or
Antenna
17
Antenna Test Methodology
Linear Voltage Ramp
VRAMP
• V-Ramp test searches for the breakdown voltage (Vbd)
and then calculates the breakdown charge (Qbd) of thin
oxide capacitors.
VRAMP Test Sequence
• Makes timing measurements of all procedures used
during the ramp
• Checks the integrity of the structure to be stressed
(leakage current test at use voltage)
• Linearly ramps the voltage (5 pS ramp rate)
• Records voltage, current, and time at each
step/interval
• Terminates the test when
- Measured current exceeds breakdown current
- Max Qbd is reached
- Next current step (if allowed) exceeds max
current density
- Elapsed time exceeds pre-set max test time
- Shift in slope of the resultant I/V curve is
excessive or slope goes negative
• Calculates outputs/results
18
Mobile Ionics
19
Test Methodology
Test Method & Conditions
• Poly heater 250C
• Bias gate to ±100v
Effect of Mobile Ion under an Applied Field
Test Sequence
1. Initial threshold voltage is measured.
Vt#1
2. Stress #1: NMOS a negative bias is
applied. PMOS a positive bias is applied.
3. Vt#2 is measured.
4. Delta Vt between Vt#1 and Vt#2
5. Stress #2: NMOS a positive bias is
applied. PMOS a negative bias is applied.
6. Final threshold voltage is measured.
Vt#3
7. Delta Vt between Vt#2 and Vt#3.
20
Manufacture for
Reliability
21
Manufacture for Reliability
• Manufacture for Reliability is the application of WLR for monitoring and process changes.
• Process Qualification (PQPs)
• FMDRC (identify reliability concerns, possibly new structures or test methods)
• Define WLR & PLR (Wear-out) test plans to complete Stage IV (Qualification)
• Execute testing, analyze and publish results
• Define & execute test plan and publish results at Stage VI (Line Warming)
• Process Change Support (Change Review Board)
• Define failure mechanism risks and WLR test plan, analyze and publish results
• Examples: Gate pre-clean (SP-RCA0) and DUV Qualification Plan
•
Production Monitor
• Minimum one lot per week sampling across key technologies
• Monthly presentation of yield compliance for critical parameters by technology and
drive W3 improvement plans; initiated 2011 P01
22