Reliability Evaluation of PEMs Rejected During C-SAM - NASA ...

Reliability Evaluation of PEMs
Rejected During C-SAM
Examination
Alexander Teverovsky
Parts, Packaging, and Assembly Technologies
Office, Code 562, GSFC/ QSS Group, Inc.
Alexander.A.Teverovsky.1@gsfc.nasa.gov
CSME’04 1

Acknowledgements
This work was sponsored by the GSFC projects
and NASA Electronic Parts and Packaging (NEPP)
program.
The author would like to thank GSFC Parts
Engineers Bruce Meinhold and Gerry Kiernan for
providing samples for this work and initial C-SAM
images.
The author acknowledges his debt to engineers
and technicians of the Parts, Packaging, and
Assembly Technologies Office, Code 562, GSFC,
for help with testing, and to Darryl Lakins, Branch
Head of Code 562, for support of this work.
CSME’04 2

Purpose and Outline
Purpose:
To discuss effectiveness of C-SAM examinations during
screening of low-power devices in SOIC8 packages and
power devices in TO-220 and SOT-223 style packages.
Outline:
� General problems with C-SAM.
� Description of C-SAM rejects used for experiments.
� Reliability testing conditions and results.
� Conclusion.
CSME’04 3

Which Part to Choose?
Delaminations are due to stress relief in PEMs.
Typically rejected
Typically accepted
Before stress. After stress? Before stress. After stress?
This part might operate
reliably because
mechanical stresses have
been relieved and
delaminations are stable.
This part might fail if the
stresses have not been
relieved and delaminations
develop in critical wire bond
areas.
� Some types of delaminations might be not harmful.
� If delaminations are serious defects, is it possible to improve
reliability of a lot with 5% to 50% rejects by screening?
CSME’04 4

More Provocative Questions
Regarding Acoustic Microscopy
If no delaminations were observed by screening:
� Can they appear after soldering reflow, TC,
environmental stresses, operation, …?
� Can they develop reversibly at high/low temperatures
or during moisture sorption/desorption?
What if there is a sponge-like
structure at the interface?
How reliable is C-SAM data?
How often do delaminations self-heal?
How risky are different types of delaminations?
CSME’04 5

Is CSAM Testing Non-Destructive?
Each handling has the probability of ~0.1% of ESD
damaging the parts. This number might be larger for
low voltage microcircuits.
Requirements for the cleanliness of water are not
always enforced.
To screen large-quantity lots, the
parts are installed on a sticky
plate. This might contaminate
leads and mechanically damage
small parts.
Special holders to avoid damage
are not always available.
CSME’04 6

Description of Parts Used
Part Package
Die
coating
Total
QTY
Rejects,
%
Type of
delam.
PEM1 SOIC8 silicone 712 83% P, L
PEM2 SOIC8 silicone 469 20% P, L
PEM3 SOIC8 silicone 2143 5.1% P, L
PEM4 SOIC8 - 463 32.2% P, L
PEM5 SOIC8 silicone 297 18.9% P, L
FET1 D2pak - 995 14.6% P, L, D
FET2 D2pak - 200 28% P, L, D
FET3 SOT223 - 895 10.5% P, L
Delaminations:
P - top of paddle; L – leads at wire bonds; D – top of die.
CSME’04 7

Typical AM Images
(Linear Devices, SOIC8 Packages)
PEM1 PEM2
PEM2 had 2 WBs to the
paddle, which might
cause failures if broken.
PEM3 PEM4 PEM5
CSME’04 8

Typical AM images (Power FETs)
FETs 1 and 2 have
Al wire bonds.
Characteristics of
molding compound:
Tg = 165 o C
CTE1 = 21.5 ppm/ o C
FET3 has
Au wire bonds.
Characteristics of
molding compound:
Tg = 155 o C
CTE1 = 16.4 ppm/ o C
CSME’04 9

Testing of AM Rejects
Thirty samples of each part type rejected by C-SAM
screening were electrically tested, environmentally
stressed, and examined using acoustic microscopy:
� Before testing.
� After SMT simulation and preconditioning per JEDEC
JESD22-A113 (85 o C/85% RH/168 hrs +3 runs
through IR reflow chamber+ flux application and
rinsing).
� After 100, 300, and 1000 temperature cycles
between –55 and +125 o C (condition TC0).
� After 100 and 300 cycles between 0 and 180 o C
(condition TC1).
� After 100 and 300 cycles between 20 and 200 o C
(condition TC2).
CSME’04 10

Examples of AM Images Before and
After Stress Testing (Linear Devices)
PEM1
PEM4
PEM5
init after SMT after 1000 TC
The proportion
of delaminated
area
decreased
after SMT and
increased after
temperature
cycling.
CSME’04 11

Changes in Delaminations Due to
Stress Testing (Linear Devices)
Proportion of delaminations, %
Proportion of delaminations, %
80
60
40
20
0
100
80
60
40
20
0
PEM1
PEM3
PEM5
PEM4
init after SMT 1000 TC
init after SMT 1000 TC
Lead-to-MC
delaminations
Paddle-to-MC
delaminations
CSME’04 12

Examples of AM Images Before and
After Stress Testing (Power Devices)
FET 1die
FET 1 lead frame
FET 3
init after SMT after 1000 TC after HT TC
CSME’04 13

Changes in Delaminations Due to
Stress Testing (Power Devices)
delaminations, %
delaminations, %
60
40
20
0
100
80
60
40
20
0
Die surface Leads
init after SMT after 1000
TC
Top paddle
FET2
FET1
FET3
init after SMT after 1000
TC
after HT TC
after HT TC
CSME’04 14
delaminations, %
100
80
60
40
20
0
init after SMT after 1000
TC
FET2
FET1
FET3
after HT
TC
� No top-of-die delaminations
(TODD) in FET3.
� In FETs 1 and 2 TODD
increased and paddle
delaminations decreased
after SMT.

VOS, uV
Results of Electrical Testing
(Operational Amplifiers)
Offset voltage variation
during temperature cycling
200
100
0
-100
-200
PEM3
PEM4
PEM1
PEM5
init SMT 100 TC 300 TC 1000
TC
100 TC
H
100 TC
H2
300 TC
H2
Input bias current variation
during temperature cycling
1.E+06
1.E+05
1.E+04
1.E+03
1.E+02
1.E+01
1.E+00
init SMT 100
TC
CSME’04 15
IOSP, pA
PEM3
PEM4
PEM1
PEM5
300
TC
1000
TC
100
TC H
The parts had no failures and manifested only
minor parametric variations.
100
TC
H2
300
TC
H2

Results of Electrical Testing
(Voltage References)
Vout, V
2.501
2.5005
2.5
2.4995
2.499
Average output voltages
no load
load 10 mA
init SMT 100 TC300 TC 1000
TC
100 TC
H
100 TC300
TC
H2 H2
CSME’04 16
2.5
2.495
2.49
2.485
2.48
2.475
� Devices had minor parametric variations.
� One part failed at load conditions after 100 HT TC.
Vout load, V

VTH, V
IGP, A
4
3.5
3
2.5
2
1.5
1
Results of Electrical
Measurements (Power FET)
1.E-06
1.E-07
1.E-08
1.E-09
1.E-10
1.E-11
1.E-12
Average threshold voltage
FET1
FET2
FET3
Average gate current
init
after SMT
100 TC
300 TC
1000 TC
100 TC H
300 TC H
100 TC H
300 TC H
FET1
FET2
FET3
init
after SMT
100 TC
300 TC
1000 TC
100 TC H
300 TC H
100 TC H
300 TC H
� No failures after TC.
� No VTH variations.
� Significant increase
in leakage currents
after SMT simulation.
� All failed parts
recovered after TC.
� Some failed parts
had corrosion of Al
metallization after
HAST.
CSME’04 17

Results of Temperature Cycling of
Parts with Excessive Delaminations
Part SMT TC0 100 TC0 300
TC0 1k TC1 300 TC2 300
PEM1 0/30 0/30 0/30 0/30 0/15 0/15
PEM2 0/30 0/30 0/30 0/30 1/15 1/15
PEM3 0/30 0/30 0/30 0/30 0/15 0/15
PEM4 0/30 0/30 0/30 0/30 0/15 0/15
PEM5 0/30 0/30 0/30 0/30 0/15 0/15
FET1 0/30 0/30 0/30 0/30 0/15 0/15
FET2 0/30 0/30 0/30 0/30 0/15 0/15
FET3 0/30 0/30 0/30 0/30 0/15 0/15
PEM6* 0/30 0/30 0/30 0/30 - -
*Parts had excessive delaminations at finger-tips after HAST.
CSME’04 18

IG S P , A
Effect of Flux Application on
1.E-06
1.E-08
1.E-10
1.E-12
Power Devices
Effect of flux application.
init
after flux
applic ation
15 hrs RT
CSME’04 19
87 hrs RT
Test conditions:
Immersion into
KESTER 2331-ZX
water-soluble flux
for 10 seconds,
rinsing, and blowing
with dry air.
Excessive leakage
currents remained
after a few days of
storing in laboratory
conditions.

IG, A
Effect of Flux Application on
Power Devices, Cont’d
1.E-06
1.E-07
1.E-08
1.E-09
1.E-10
1.E-11
1.E-12
Failed samples.
in it S MT s imul.
+ CSAM
1 hr 125
oC
3 hr 125
oC
15 hr
100% RH
22 oC
10 sec in
deionized
wa te r
� Baking for a few
hours at 125 o C
mostly restored
leakage currents.
� Immersion in
water and storing
at 100%RH did
not increase IG.
� Effect of SMT
simulation is likely
due to good
wettability of flux.
� The type of flux used for SMT simulation might affect results of
qualification testing. JESD22-A113 might need a review.
� Flux used for soldering might be different from the one used for
the simulation, thus making results of qualification questionable.
CSME’04 20

IG, A
1.E-06
1.E-07
1.E-08
1.E-09
1.E-10
1.E-11
1.E-12
Can CSAM Screen Out Flux-
Related Failures?
paddle delaminations
Effect of immersion into flux
0 20 40 60 80
proportion of delamination, %
1.E-06
1.E-07
1.E-08
1.E-09
1.E-10
1.E-11
1.E-12
TODD
0 20 40 60 80
proportion of delamination, %
No correlation between TODD or paddle-to-MC delaminations
and leakage currents after flux immersion was observed.
CSME’04 21
IG, A

Conclusion
Parts with excessive delaminations at the paddle and leads
(at secondary bond locations) had no electrical failures after
1000 TC at –55 to +125 o C conditions.
The analyzed power devices in TO220-style packages are
prone to formation of top-of-die delaminations; however, no
wire bond fractures occurred even after multiple temperature
cycling.
Washable flux can penetrate to the die surface through
delaminations significantly increasing leakage currents and
causing corrosion of Al metallization.
No correlation between the proportion of delaminations and
flux-induced leakage currents were observed indicating a
failure of CSAM examination to screen out potential failures.
For the part types used, CSAM examination is not a valueadded
technique for screening; however, it is very useful for
qualification testing.
CSME’04 22