Wire bond Concern on Actel Devices with Au wire

1
Wire bond Concern on Actel Devices with Au wire
Raymond Kuang Manager, Packaging
Engineering May 08, 2002
2
Background

• An Actel device RT54SX16 with date code 9937 was
  found by GSFC having the following
• Strength of some wires pulled were below the 2.5
  Grams Mil-std minimum limit
• Parts exhibited intermetallic halo at ball
  periphery.
• Watermelon stripes were suggested to be
  contamination.
• GSFC inspected additional samples taken from 11
  Actel lots bonded with Au wires and result as
  follows
• All samples exhibited intermetallic halo around
  the ball periphery.
• Except on RT54SX16 with date code 9937, all
  sample parts passed the required minimum wire
  pull test limits.
• The bonding quality issue on RT54SX16 with date
  code 9937 was discovered in year 2000, resulting
  to PCN to customers with affected lots in July,
  2000.

3
Failure Analysis Low Wire Pull Strength on
9937 Lot
This issue was due to the use of insufficient
bonding power and force that resulted in immature
welds consisted of cluster of isolated
micro-welds. Since isolated micro-welds offers
more grain surfaces, and therefore more vacancies
defects, a rapid intermetallic diffusion
characterized by vacancies defects concentration
(voids) on Al-Au interface had occurred. The use
of insufficient bonding force and power is
evident on the shape of the ball see photos
below.
Al pad
Cross section image of RT54SX16 (9937 date code
lot)
Ball bonding on Pad
4

Critical Factor for Reliable Au Ball Bond
The only way to prevent rapid intermetallic
formation and growth from happening is to
assemble the part with very high quality wire
bonds which exhibits a uniform metallurgical
formation
Figure 2 Poor quality bond with Non-uniform
Intermetallic thickness
Figure 1 High quality bond with uniform
Intermetallic thickness
A bond with uniform intermetallic formation will
have stronger bond strength than a bond with
spotty intermetallic formation. Figure 1 above
shows an example of an ideal bond while Figure 2
shows a spotty bond that resulted to formation of
voids.
5
Actel Corrective Actions

• Equipment OptimizationUpgraded bonding machines
  with Ultra Sonic Generation (USG) delay and
  Gradual Power Application features. These
  allow for application of sufficient bonding
  force and power in a gradual schedule to avoid
  damaging pad metallization.
• Plasma Clean Prior to Wire bondHelps to expose
  a fresh, unoxidized bonding surface in order to
  achieve a stronger bond interface
• Units must be bonded within 3 hours after plasma
  clean.
• These enhancements are essential to ensure a
  uniform Au-Al intermetallic interface that is
  strong and stable being created during the wire
  bond process.

6
Bond pull Quality Analysis
RTSX Wire pull (1.0 mil wire)
Post Burn-In 100 wire pull
Assembly
Std Group B
7
Bond pull Quality Analysis
RTSXS Wire pull (1.0 mil wire)
Break mode of minimum reading from Std. Group B
and Post Burn-in 100 wire pull were wire
break at wire span.
8
Actel Preventive Actions

• Preventive Actions with Wire Bond Process Monitor
  --- Wire Pull Testing
• Enforcing extensive wire pull testing during
  assembly
• To detect any sign of process or material issue
  that may affect bonding quality.
• Require a minimum of 2 set up units prior to
  start of production
• Production wont start until set up units have
  passed pull test requirement.
• In-Process Monitor
• E-flow Devices 2 additional units must go
  through bond pull testing in every 2 hours of
  production.
• A minimum of 4 units per lot must be wire pulled
  if bonding process is completed within 2 hours.
• 100 of the wires must be pulled in each sample
  unit.
• Class B devices 2 additional units must be wire
  pulled at the completion of the production lot.
• 50 of the wires must be pulled from each sample
  unit.

9
Actel Preventive Actions .. Continue

• Preventive Actions with Wire Bond Process Monitor
  --- cont.
• Wire Pull Process Control limit
• A guard-banded minimum allowable bond pull
  strength during set-up and in-process monitor is
  4.5 grams.
• Pre-cap Source Inspection (CSI) is required for
  all E-flow production lots by Actel.
• Inspector will witness wire pull of 50 of the
  wires on 2 units randomly selected from
  production lot.
• Disposition of Non-Conforming Lots
• All production lots fail to meet minimum bond
  strength limit will be put-on hold for
  disposition from Actel
• Actel will evaluate if the failure modes are true
  or equipment and/or operator errors
• For equipment and/or operator errors, additional
  units must be tested and no failures allowed
• Failed lots will be down graded to commercial or
  scrapped.

10
Actel Preventive Actions .. Continue

• Preventive Actions with Wire Bond Process Monitor
  --- cont
• Post burn-in wire pull monitor
• 100 wire pull is perform on 2 units of the group
  B samples.
• Tracking wire pull data on assembly set-up and
  group B testing on a routine base.

11
Wire Bond Reliability Study

Parts from known bad lot (9937 d/c) and good lots
were subjected to Temperature Cycle Testing (TCT)
for 1,000 cycles from -65C to 150C per
Mil-Std-883, TM1010, Condition C. Wire pull test
data conducted before and after TCT are compared
as bellow

• The above result shows no significant degradation
  of bond pull strength after 1,000 Temperature
  Cycles.
• No ball lifting exhibited on units from known
  good lots.

12
Wire Bond Reliability Study .. Continue

• Wire pull test data on one unit exposed to 1,000
  hours of life test at 150oC are as follows
• Post Lid Seal Post 1K hours _at_150C
• Min. Strength 5.8 4.8
• Max. Strength 8.1 6.6
• Avg. Strength 6.72 5.78
• Sample size 100 235

• No significant degradation of bond pull strength,
  and no ball lifting observed.
• Picture on the right is taken after 1K hours life
  test.

13
Wire Bond Reliability Study .. Continue

• Wire pull test data on one unit exposed to 2,000
  hours of life test at 150oC are as follows
• Post lid Seal
  Post 2K hours _at_150C
• Min. Strength 5.2 3.4
• Max. Strength 12.7 9.5
• Avg. Strength 6.83 7.28
• Sample size 88 117

• Minimum reading is way above the mil-std 883
  minimum bond strength requirement after exposure
  of unit to 150C.

14
Engineering Analysis Intermetallic Mass
(Halo) at Ball Edge
Intermetallic mass
Intermetallic mass
RT54SX16 / 9937
RT54SX32S / 0049
RT54SX32S / 0049
Photo taken by Actel after 1000 hrs
exposure to 150C.
Photo taken by GSFC ( No 1000 hrs burn-in
test )
Intermetallic mass at ball edge
RT54SX16 / 9937
RT54SX32S / 0049

• Intermetallic mass can be seen also on units with
  good bond quality.
• Presence of intermetallic mass around the bond
  edge is normal and is not an indication of poor
  bond quality.

15
Engineering Analysis Intermetallic Mass
(Halo) at Ball Edge .. Cont.
Ceramic Capillary

Footprint of X-Y displacement of Au ball
Active contact area of Pad and Ball
16
Engineering Analysis Intermetallic Mass
(Halo) at Ball Edge .. Cont.

• Formation of Intermetallic Mass (Halo) at Ball
  Edge
• The intermetallic mass at the edge of the ball is
  a result of displaced intermetallic footprint by
  the scrubbing motion on both X and Y directions
  during the initial stage of bonding.
• The footprint of this Au-Al intermetallic mass
  will grow when the device is exposed to heat
  during the lid seal process.
• The size of the exposed intermetallic mass around
  the edge of the ball is also dependent to the
  final ball shape. If the ball is squashed, it
  may cover all intermetallic mass at the ball
  edge.
• Bonding force and power are less on area near the
  ball edge
• than at the bond center. As a result, void
  is normal to occur on area near the ball edge.
  This void, however, will not propagate toward the
  
• bond center and degrade the bond strength.

17
Engineering Analysis Dark (Watermelon)
Stripes on Ball Surface

• EDX analysis show no difference on the
  composition of dark and light spot at the ball
  surface.

Light spot
Dark spot
18
Engineering Analysis Dark (Watermelon) Stripes
on Ball Surface Cont.
Light spot EDX analysis at 5KeV
Dark spot EDX analysis at 5KeV
19
Engineering Analysis Dark (Watermelon) Stripes
on Ball Surface Cont.

• It is suggested to be an effect of reaction
  between the doping materials within the wire and
  the heat energy of the Electronic Flame Off (EFO)
  spark generated during free air ball formation.
• Wire Composition Analysis (Au wire vendor
  American Fine Wire)
• 1998 1999 Currently
• Au 99.99 99.99
• Ag 19 ppm 20 ppm
• Be 6 ppm 5.2 ppm
• Ca 1.5 ppm ---
• Fe --- 0.8 ppm
• The EDX analysis shows no evidence of organic
  contamination.

20
Summary

• The affected lots, assembled in 1999, with weak
  bond strength and formation of voids within the
  Al-Au intermetallic interface were due to
  insufficient power and force at the wire bond
  process.
• The intermetallic mass (Halo) around the edge of
  the ball is typical.
• The dark stripe at the ball surface is not an
  indication for contamination
• based on EDX analysis.
• Au-Al intermetallic formation will always exist
  in Au wire bonded devices.
• Optimized bonding process to achieve uniform
  intermetallic formation and high bond strength is
  essential to maintain device reliability.
• Actel maintains tight process control and
  stringent wire pull monitors to ensure bond
  reliability.
• Thanks !