Items for Discussion
description
Transcript of Items for Discussion
![Page 1: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/1.jpg)
1
Items for Discussion
• Chip reliability & testing• Testing: who/where/what ???• GBTx radiation testing• GBTx SEU testing• Packaging– Low X0 options, lead free solder?
• lpGBTx plans and timescales
![Page 2: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/2.jpg)
2
Chip Reliability
• What is there to worry about?• Failure Mechanisms• Statistical analysis PoF• Collaboration on reliability testing
![Page 3: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/3.jpg)
3
Why worry?
• Traditionally failures in HEP not dominated by ASIC reliability– Connectors, solder, wire bonds, cracks in tracks
and vias, capacitors, power supplies– Non-ideal scaling in DSM processes• Aggressive designs target optimal performance• Voltage decreases insufficient to compensate density
increase higher T lower reliability.
![Page 4: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/4.jpg)
4
FA Webinar- Cheryl Tulkoff
![Page 5: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/5.jpg)
5
Physics of Failure (PoF)• Assumption of single dominant damage mechanism
can lead to wrong extrapolation of lifetimes from accelerated tests.
• PoF aims to understand different failure mechanisms– Fit model parameters to data for each damage mechanism– Combine results to predict reliability at operating
conditions– Health warning: competing models for some damage
mechanisms can give very different extrapolations to operating conditions.
![Page 6: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/6.jpg)
6
Time Dependent Dielectric Breakdown (TDDB)
• In DSM processes E fields over gate oxides ~ 5 MV/cm cf breakdown fields of > ~ 10 MV/cm.– Gradual degradation later failures
• Acceleration model– MTTF=A×10-βE exp(-Ea/kT)– Example fits look ok but activation
energy not constant? next slide– can’t fit to single failure mechanism!
Holes injected into oxide Stress Induced leakage currents by tunnelling breakdown
![Page 7: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/7.jpg)
7
TDDB Fits • Fits to Voltage (E field)
and T look ok but estimated value of Ea depends on E ?
MTTF vs E
MTTF vs 1/T
Fitted Ea not constant!
![Page 8: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/8.jpg)
8
Hot Carrier Injection (HCI)
• Non-ideal scaling larger E fields “hot” carriers can overcome barrier between Si and gate oxide– Trapped charges lead to changes in Vth and gm– Eventually lead to failure– t = c (Isub)-m
– T dependence because at low T electron mfp longer acquire more energy in E field impact ionization.
![Page 9: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/9.jpg)
9
HCI
• Example fits to threshold shifts.
• Typical fit values – m ~ 3
• Also need to consider T variation.
• Shift Min Vcc
![Page 10: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/10.jpg)
10
Electro-migration (EM)• High current densities, force exerted by electrons large enough
to cause diffusion of metal ions in the direction of the e flow.– Creates voids increases R thermal runaway open circuit– Excess build up of ions at the anode can give short circuit
• Very sensitive to material, doping, grain boundaries etc…• EM is thermally activated, T gradients flux divergences.• Best model
– Typical values for Al: Ea=0.6 eV and n ~ 2.
)/exp()( kTEjAMTTF an
e
![Page 11: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/11.jpg)
11
Other Mechanisms
• NBTI (Negative Bias Temperature Instability)– Degradation (Vth/Gm shift) occurring due to
negative biased BT (bias temperature) stress in PMOS FETs
• Stress migration– CTE mismatch can cause stress even with no
current.• Assembly & packaging
![Page 12: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/12.jpg)
12
Combining Failure Rates• Common method is just to assume exponential
distributions– Total failure rate: – But we know that failure distributions aren’t exponential !
• Failure distributions better modelled by Weibull or log-normal distributions.
• Finally we don’t actually want MTTF we need MTT01 (1% failure) or MTT110 (10% failure).– Need to combine distributions correctly from different failure
mechanisms.– Determine MTT0X numerically
i iTOTAL
![Page 13: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/13.jpg)
13
Compare Distributions• Compare exponential, Weibull and log-normal• Note Weibull and log normal totally different from
exponential for small x– This is just the region we are interested in!
0 1 2 3 4 5 60
0.1
0.2
0.3
0.4
0.5
0.6
exp(-x)Weibulllog normal
![Page 14: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/14.jpg)
14
Determining Model Parameters
• Brute force: Run ALT for matrix of different T and V and fit data to get model parameters.– Too many tests too slow/expensive.
• Smarter approach– High T/High V TDDB
• Vary T Ea, vary V exponent c– Low T/High V HCI
• Vary T Ea2, vary V g2
– High T/low V EM dominates• Vary T Ea3
![Page 15: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/15.jpg)
15
Next slides from Steve McMahon (RAL)
![Page 16: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/16.jpg)
16
![Page 17: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/17.jpg)
17
![Page 18: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/18.jpg)
18
Next Steps• Fix some parameters:
– Sample sizes– How many different test conditions can we run– Do tests have to continue until we reach 50% failures or can we
perform fits with fewer failures?• Toy MC study to optimise power of set of points in V &T
space.– Generate pseudo-data from the models
• Don’t need to assume unrealistic exponential distributions.• Failure probability given by
– Try to fit the pseudo-data to determine the model parameters.
))(1(1)( tptf ii
![Page 19: Items for Discussion](https://reader036.fdocuments.us/reader036/viewer/2022062815/56816976550346895de15e46/html5/thumbnails/19.jpg)
19
References
• Bernstein, Physics of Failure Based Handbook of Microelectronic Systems, RIAC.
• Srinivasaan et al, The impact of Technology Scaling on Lifetime Reliability, DSN-04.
• Semiconductor Reliability Handbook, www.renesas-electoronics.com