A client provided several resistors to SEM Lab, Inc. for SEM/EDS analysis. The client reported seeing “yellowing” of the metallization after solder reflow on some components, most notably 0603 resistors. We examined both discolored (post reflow) and stock parts (control).
 

This is a BSE SEM image of the discolored resistors.

 

This is the elemental spectrum from a discolored termination.

 

Comparison of EDS spectra for discolored and stock parts. The problem appears to be related to silver.

 

Sungil Cho et al [1] studied oxidation kinetics for lead-free alloys and reported that silver modifies the oxidation rate of tin (i.e. promote the formation of tin oxides). It seems likely that the yellowing is related to thick oxide formation due to combination of factors including (1) excessive time-temperature exposure (2) silver availability from the silver-thick-film in the resistor and/or the silver in the SAC alloy.
 

Resistor manufacturers offer designs that are improved with respect to exposure of the silver-thick-film in the resistor to corrosion (usually by sulfur), e.g. Thick Film Chip Resistors, Military/Established Reliability MIL-PRF-55342 Qualified, Type RM. It was suggested that the client obtain some of the Type RM resistors and determine if they are also more immune to yellowing during reflow on this assembly.
 

[1] Sungil Cho et al, “The Oxidation of Lead-Free Sn Alloys by Electrochemical Reduction Analysis”, Journal of Metals, June 2005, pp. 50 – 52.
 

Check out SEM Lab, Inc.  to learn more.

A client provided some sample PCBs (1 bare, 1 assembled) that were exhibiting discoloration on the ENIG plated surface. The client asked us to characterize this defect and determine root cause.

This image shows what appears to be a "water spot" on an ENIG pad.
This image shows what appears to be a “water spot” on an ENIG pad.

 

This is a microsection of the same pad through the approximate center of the water spot.
This is a microsection of the same pad through the approximate center of the water spot.

 

This is a high magnification image of the ENIG plating.
This is a high magnification image of the ENIG plating.

The nickel (EN) deposit appears normal except at the interface with the gold (IG layer).  The defects (shown by the yellow arrows) are referred to as “IG spiking”, where the gold plating solution etches the Ni-P grain boundaries.  This creates small voids or pits under the IG plating that trap process chemistry and cause corrosion resulting in the stain on the surface.  Root cause = IG spiking.  A secondary cause is water droplets left on the board surface after cleaning.

Check out SEM Lab, Inc.  to learn more.

This is a fairly typical example of wear out failure of an aluminum electrolytic capacitor.

AE-Cap-1

This is an optical image of the capacitor that captures the part date code.

AE-Cap-2

This is an image of the top of the device showing the vent, which ruptured and leaked a small amount of electrolyte.

AE-Cap-3

This is an image of the capacitor after dissecting the can. The capacitor roll was charred indicating that the core reached excessive temperature.

AE-Cap-4

This is a plot of capacitance versus age for “passing” capacitors from the same population of capacitors. The trend line is an inverse exponential function indicative of the wear out process.

The analysis results suggest that the capacitor failure was caused by the normal wear out processes that are characteristic of aluminum electrolytic capacitors, i.e. capacitance decreases and electrolyte volume decreases (i.e. capacitor dries out) as the capacitors age and approach end-of-life.

Check out SEM Lab, Inc.  to learn more.

This is a somewhat classic case of solder joint fatigue failure, but with the added factor of gold embrittlement. The solder alloy is SN63, the package is a J-lead PMIC, and it is soldered to an alumina ceramic substrate.

solder fatigue 1

The elemental spectrum of the bulk solder joint suggested it contained ~ 3 wt% of gold, which is considered a threshold for gold embrittlement.

solder fatigue 2

The thermal fatigue fracture showed classic characteristics such as grain boundary separation and propagation through the bulk solder joint.

solder fatigue 3

The AuSn4 intermetallic compound is clearly visible in the solder joint microstructure at about 12% by area in the section. This harder phase in the Sn-Pb matrix likely accelerated the thermal fatigue failure.

The analysis results suggest that the thermal fatigue damage was likely due to the combination of (1) CTE mismatch between the J-lead package and the alumina substrate and (2) Au-Sn IMC in the solder joint microstructure at the threshold for gold embrittlement.

Check out SEM Lab, Inc.  to learn more.

LEDs tend to develop high levels of thermal expansion mismatch stress internally due to gross CTE mismatch between the lens polymer, the die, the lead-frame/cup, and the bond wires. In the case shown below, a silicone polymer was used to fill the cup perhaps to provide a stress buffer, but it did not appear to be effective.

Failure of IR LED 1

This view (photo above) is looking down through the lens at the die area. The damage shown is at the interface between the body material and the silicone polymer that fills the cup and encapsulates the device die.

Failure of IR LED 2

There was no evidence of damage at the wedge bond.

Failure of IR LED 3

This view (above) is a look through the bottom (lead) side of the device. There appeared to be some crazing at the rim of the cup.

Failure of IR LED 4

This image (above) was obtained at an intermediate stage in the microsectioning process. There is a fairly significant kink in the bond wire above the ball bond.

Failure of IR LED 5

The silicone encapsulant adhesively failed at the cup/silicone interface and the silicone/lens polymer interface.

Failure of IR LED 6

The image above shows a segment of silver plating that has separated from the underplating on the cup.

Failure of IR LED 7

Above is an optical image showing the location of the open at the kink, which is the high stress region of the gold bond wire.

Conclusion – The analysis results suggest that the LED failed due to an open bond wire just above the ball bond on the device die. The damage was likely caused by a thermal overstress event where the thermal expansion rate of the silicone die encapsulant and the peak temperature are the primary drivers. The kink in the bond wire and the adhesion strength between the silicone die encapsulant and the cup were also a likely contributing factors.

Check out SEM Lab, Inc.  to learn more.

Residues on PCBAs are unfortunately common place. SEM/EDS analysis helps to determine the nature of the residue.

residue on PCBA 1

The image above shows residue on the surface of the PCBA, but it is not clear what is the nature of the residue.

residue on PCBA 2

EDS analysis shows that the residue is likely tin bromide, suggesting corrosive bromide flux activator was left behind after assembly.

residue on PCBA 3

This is a BSE SEM image of some of the residue, which shows dendritic growth. The dendrites suggest that electrochemical migration (ECM) was a factor meaning ionic contamination, electric field, and moisture were likely at play. If the dendrites migrate between normally isolated signals (e.g. PWR & GND) then the circuit fails.

Check out SEM Lab, Inc.  to learn more.

A client provided two sensor assemblies for failure analysis of capacitor C4 on the internal PCBA.

MLCC Bending Fracture _1

This is an optical micrograph of the capacitor as mounted on the PCBA.

MLCC Bending Fracture _2

This is an optical image of an intermediate grinding stage during microsection preparation. A bending fracture was noted under one of the terminations.

MLCC Bending Fracture _3

This is a BSE SEM image of the capacitor microsection.

MLCC Bending Fracture _4

The fracture appeared to be due to flexure of the PCBA.

Check out SEM Lab, Inc.  to learn more.

Ionic contamination, moisture, and electrical bias can combine to create electro-chemical migration (ECM) shorts under an IC device as shown in this example.

ECM short 1

This is an optical image showing residue on the bottom surface of the FPGA package.

ECM short 2

This is a BSE SEM image of the same corner location. The residue bridges several signals at this corner.

ECM short 3

An EDS spectrum of the residue suggest that it contains chlorine and bromine. Part of the bromine signal likely originates from the brominated epoxy molding compound. The chlorine & perhaps some of the bromine is most likely residual solder flux activator.

The analysis results suggest that the most likely cause of failure was external corrosion and ECM due to halide contamination. The corrosion and ECM was likely exacerbated by elevated temperature and humidity conditions.

Check out SEM Lab, Inc.  to learn more.