2015 NEPP Tasks Update for Ceramic and Tantalum Capacitors · 2015 NEPP Tasks Update for Ceramic...

2015 NEPP Tasks Update for Ceramic and Tantalum Capacitors

Alexander TeverovskyAS&D, Inc.

Work performed for Parts, Packaging, and Assembly Technologies Office,

NASA GSFC, Code [email protected]

NASA Electronic Parts and Packaging (NEPP) Program

Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov originally presented by Alexander Teverovsky at the NASA Electronic Parts and Packaging Program (NEPP) Electronics Technology Workshop (ETW), NASA Goddard Space Flight Center in Greenbelt, MD, June 23-26, 2015.

https://ntrs.nasa.gov/search.jsp?R=20170000475 2020-04-19T09:29:22+00:00Z

List of Acronyms

AF acceleration factor IM Infant mortalityBI burning-in JAXA Japan Aerospace Exploration Agency

BME base metal electrode MLCC multilayer ceramic capacitorDCL direct current leakage PHS polymer hermetically sealedESR Equivalent series resistance PME precious metal electrodeFB ferrite beads PV Prokopowicz-VaskasFR failure rate QA quality assurance

HALT highly accelerated life testing RVT random vibration testingHSD hot solder dip S&Q screening and qualificationHT High temperature VR rated voltage

HTS high temperature storage

2Deliverable to NASA Electronic Parts and Packaging (NEPP) Program to be published on nepp.nasa.gov originally presented by Alexander Teverovsky at the NASA Electronic Parts and Packaging Program (NEPP) Electronics Technology Workshop (ETW), NASA Goddard Space Flight Center in Greenbelt, MD, June 23-26, 2015.

Reasons for NEPP Tasks on Capacitors

Capacitors constitute the majority of elements in electronic systems.

New technologies and designs appear with increasing speed. There is a need for optimization of S&Q procedures and setting adequate requirements.

Physics behind degradation and failure processes needs better understanding.

Capacitors exhibit both, infant mortality and wear-out failures, and can be used as models to refine quality assurance approaches for variety of space components.


Outline Update on tantalum capacitors. Use of ferrite beads as surge current limiters. Polymer capacitors. Random vibration testing of advanced wet capacitors. Future work.

Update on ceramic capacitors. Effect of cracking on degradation of MLCCs at high

temperatures. Can we use automotive industry capacitors? Future work.


Ferrite Chip Beads as Surge Current Limiters

5

Contrary to inductors, FBs at high frequencies work like resistors and dissipate power in the form of heat.

NEPP report contains (https://nepp.nasa.gov/): Analysis of requirements of DLA DWG#03024 for hi-rel chips; Results of testing of 12 types of FB; Data on the specific features of FBs; Evaluation of the robustness of FB to soldering stresses; Behavior of FBs under multiple high current spikes. Recommendations for reliability assurance of tantalum

capacitors operating under surge current conditions. Conclusion: Due to decrease of impedance with frequency and current,

the effective resistance remains substantially below the value that is required to limit surge in tantalum capacitors (from 1 to 5 Ohm). Recommendations on current derating are available at

https://nepp.nasa.gov/.

05

1015202530354045

-10 10 30 50 70 90

curre

nt, A

time, us

Effect of ferrite beads on surge currents

15uF 10V at 22V surge

same with FB 03024-021

0.16 ohm

0.11 ohm

0

30

60

90

120

150

180

0 10 20 30

curre

nt s

pike

, Avoltage, V

BLM18PG181SNID, 0603, 180ohm, 90mohm, 1.5A

no FBSN1SN2SN3SN4SN5SN6failures


https://nepp.nasa.gov/


Polymer Tantalum CapacitorsA report on evaluation of PHS capacitors manufactured

per DLA LAM DWG#13030 (https://nepp.nasa.gov/):Literature review; analysis of requirements; characteristics, including thermal resistance; behavior of DCL under forward and reverse bias, recommendations.

Specific feature: operation of polymer capacitors requires certain amount of moisture in the case. What happens if cases dry out?

6

Variations of capacitance, ESR, and DCL with time of HTS

20

40

60

80

100

120

0 200 400 600 800 1000

capa

cita

nce,

uF

time, hr

PHS 100uF 60V during HTS at 150C

hermetic

non-hermetic

0

0.2

0.4

0.6

0.8

1

0 200 400 600 800 1000

ESR,

Ohm

time, hr

PHS 100uF 60V during HTS at 150C

hermetic

non-hermetic

1.E-08

1.E-07

1.E-06

0 200 400 600 800 1000

DC

L_10

00 s

ec, A

time, hr

PHS14 100uF 60V

pol SN7 pol SN8pol SN9 pol SN10pol SN11 dep SN7dep SN8 dep SN9dep SN10 dep SN11

non-hermetic

PHS can survive 1000 hr storage at 150°C without degradation. Non-hermetic parts degraded due to a substantial decrease in

capacitance and increase in ESR caused likely by increasing resistance of the polymer.



Recommendations for Use of PHS

7

PHS capacitors have lower weight and ESR compared to similar case size wet tantalum capacitors and their application in power lines can assure better filtering and lower ripple currents.

Polymer capacitors would mostly benefit low-temperature applications (below 0°C) or systems where a cold start-up is required. However, additional application-specific testing are required if the parts are to be used at T < -55°C.

Self-healing capability of PHS is much worse than wet capacitors and flaws in the dielectric that might be forgiven in wet capacitors might cause catastrophic failures in PHS. This requires a close attention to the results of S&Q, specifically, to measurements of leakage currents through the testing.


Random Vibration Testing

8

Report is available at https://nepp.nasa.gov/ Problems in assurance robustness of capacitors

under RVT have a long history. Larger anode size increases the stress during RVT. Existing requirements and practice: MIL-PRF-39006: 1.5hr in 3 directions; 30 min

monitoring every 0.5 msec “to determine intermittent open-circuiting or short-circuiting”.

Test techniques and failure criteria are not specified allowing different test labs to carry out testing differently, e.g limiting resistors from ohms to dozens of kohms, and failure criteria vary from 5% to 90% of VR.

Different set-ups have different sensitivity to short-circuiting. Different failure criteria cause inconsistency in test results. A single scintillation event is sufficient to cause lot failure.

Some test labs assume this level of spiking is acceptable



RVT: Step Stress Testing

9

0.0E+00

5.0E-06

1.0E-05

1.5E-05

2.0E-05

0 200 400 600 800 1000 1200

curre

nt, A

time, sec

680uF 50V Mfr.C at 34.02g rms

SN1

SN2

SN3

SN4

SN5

0.0E+00

5.0E-06

1.0E-05

1.5E-05

2.0E-05

0 200 400 600 800 1000 1200

curre

nt, A

time, sec

680uF 50V Mfr.C at 53.44g rms

SN1

SN2

SN3

SN4

SN5

Example of a part passing RVT at 34 g rms and failing at 53.44 g rms

0.0E+00

2.0E-06

4.0E-06

6.0E-06

8.0E-06

1.0E-05

0 200 400 600 800 1000 1200

curre

nt, A

time, sec

DWG#93026 470uF 75V Mfr.Aat 10.76g rms

SN1

SN2

SN3

SN4

0.0E+00

2.0E-06

4.0E-06

6.0E-06

8.0E-06

1.0E-05

0 200 400 600 800 1000 1200

curre

nt, A

time, sec

DWG#93026 470uF 75V Mfr.Aat 19.64g rms

Did this part fail at 10.76 g rms, at 19.64 g rms?


RVT: Post-testing Leakage Currents

10

Leakage currents were monitored with time after RVT.

0.0E+00

1.0E-05

2.0E-05

3.0E-05

4.0E-05

5.0E-05

0 200 400 600 800 1000 1200

curre

nt, A

time, sec

560uF 25V Mfr.A at 53.79 g rms

SN1SN2SN3SN4

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E+1 1.E+2 1.E+3 1.E+4 1.E+5

curre

nt, A

time, sec

560uF 25V Mfr.A at RT, 25V

initial

af ter RVT

Currents during RVT Currents after RVT

Spiking during RVT might not result in DCL failures after the testing.560 µF 25 V capacitors passed HALT after RVT at 53.8 g rms.Parts with excessive currents are recovering with time under bias.1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-3 1.E-2 1.E-1 1.E+0 1.E+1 1.E+2

curre

nt@

75V,

A

time, hr

470uF 75V Mfr.A after RVT

SN1SN2SN3SN4SN5

0.0E+00

2.0E-06

4.0E-06

6.0E-06

8.0E-06

1.0E-05

0 200 400 600 800 1000 1200

curre

nt, A

time, sec

470uF 75V Mfr.A at 34.02g rms


RVT: Recommendations

11

Different tests for different risk levels.

Each lot should be tested. Typical testing: 19.6 g rms , 6 samples. 15 min in each direction. DCL is monitored (10k, 0.1sec

sampling). Criterion I: Isp > 3DCL(5) = 3×I300

Criterion II: Q > Qcr

Criterion III: I300_RVT < 1.25×I300_init

Lots older than 5 years should be retested.


Future Work on Tantalum Capacitors

MnO2 chip capacitors. Rapid assessment of reliability acceleration factors. Degradation during long-term operation under reverse bias.

Advanced wet capacitors. Analysis of DCL(T, V, t), breakdown processes, gas generation, and

requirements for S&Q. Effect of HT storage on performance and reliability.

Polymer capacitors. Evaluation of chip tantalum capacitors and requirements for S&Q. Evaluation of new types of hermetically sealed capacitors.

Solid electrolyte super-capacitors for space application.


Life Testing of MLCCs with Cracks

13

Cracking does not affect IR at 125 °C but facilitates degradation of leakage currents.

In the presence of cracks, currents in BMEs start increasing after a few hours of testing, but stabilize with time.

Degradation in PMEs with cracks occurs at much higher levels of stress, and contrary to BMEs results in instantaneous short circuit failures (due to HT silver migration?).

Contrary to humid environments, at high temperatures, BMEs with cracks degrade faster than PMEs (degradation vs. catastrophic failures).

1.E-09

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E+2 1.E+3 1.E+4 1.E+5

curre

nt, A

time, sec

CDR 1825 with cracks at125C 200V

Mfr.V

Mfr.C

1.E-07

1.E-06

1.E-05

1.E-04

1.E+2 1.E+3 1.E+4 1.E+5

curre

nt, A

time, sec

BME 0.33uF 50V 1210 with cracks at 125C 100V

Mfr.C

Mfr.A

Typical variations of currents during HALT

Reference parts


Can we Use “AUTO” Capacitors?Benefits of using “auto” grade

capacitors are obvious.Adaptation of “auto” components

should start after ~5 years on the market.

For MLCCs we are ~10 years late.Major QA problems: Lead-free terminations (Sn whiskers). “Insufficient screening” (no BI) Lack of long-term reliability data.

Issues to discuss: Acceptable measures to mitigate

whiskering. Why do we need burning-in? What long-term testing tells us?

14

WhiskeringCan be mitigated by using Sn/Pb

solder, conformal coating, etc. JAXA uses HSD to replace Sn on

“auto” BMEs with Sn/Pb followed by additional screening.

0201, smartphones


Is Burning-In Necessary?QA wisdom: “Reliability should be designed into product and processes,

but not screened out by testing”. In practice, we require that parts for space applications go through BI. The purpose of BI is to remove IM failures from the lot.

15

)(1)()()(

BI

BIBI

tAFFtAFFtAFtFtF

×−×−×+

=

−−=

β

ηttF~

exp1)~(

Probability of failure after BI for tBI hrs

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+0 1.E+1 1.E+2 1.E+3 1.E+4 1.E+5 1.E+6

prob

abilit

y of

failu

re

time, hr

Effect of BI, eta=1E6 hr and AF=1000

b=0.2 no BIb=1 no BIb=3 no BIb=0.2 BI_240b=1 BI_240b=3 BI_240

BI is useless if lots do not have IM or their proportion is below a certain level.

BI reduces useful life for lots susceptible to wear-out. Burning-In might be not necessary.

β < 1 => IM failuresβ = 1 => random failuresβ > 1 => wear-out failures

1

)(−

=

β

ηηβλ tt

Weibull distribution determines type of failures


What Results of Life Testing Mean?Still, FR can be estimated if: There is no significant lot-to-lot variations.

(Verification of consistency of quality is built into MIL system. There is a greater portion of “trust/relationship” in “auto” industry.) Same mechanisms at life test and normal

conditions.(not always so, e.g. moisture, cracking/soldering) Failures are random (β = 1).

(probably never happens; instead: λ= const < FRspec). Accelerating factors are known.

(Is not true in most cases.)

16

tNAFn

×××

+=

52 1012

)22,(αχλ

Field failures are typically due to conditions that are not simulated by life testing.Without AF, comparison of life tests (PME vs. BME) is not correct.Emphasizing importance of 10 khr HALT can mislead

manufacturers. The focus should be on consistency of quality.

PME BME1000 hr 10,000 hr 1000 hr 10,000 hr

failures 0 1 0 1PV const. nV= 3, Ea= 0.8eV nV= 4, Ea= 1.1eV

AF 14,404 439,199FR 2.9E-04 6.4E-05 9.5E-06 2.1E-06

FR( in %/1000hr at 60% conf) at 50ºC, 0.5VR based on life testing of 22 samples

Life testing is typically a qualification (qualitative), not a reliability (quantitative) test.


Future Work on Ceramic Capacitors Cracking-related problems. Develop mechanical tests (board flex and strength) and assess their

effectiveness for quality assurance. Analysis of cracking on degradation and failures at high temperatures. Develop recommendations to mitigate risks of manual soldering/rework.

Comparative analysis of performance and reliability of BME and PME capacitors. Breakdown voltages, leakage currents and insulation resistance. Analysis of failures in BME capacitors with defects. Express testing to determine reliability acceleration factors for BME

capacitors. Guidelines for selecting “auto” MLCCs for different project levels.

Specifics of QA and attachment for small-size MLCCs. Analysis of requirements for stacking capacitors.


2015 NEPP Tasks Update for Ceramic and Tantalum Capacitors · 2015 NEPP Tasks Update for Ceramic...

Documents

Transcript of 2015 NEPP Tasks Update for Ceramic and Tantalum Capacitors · 2015 NEPP Tasks Update for Ceramic...