Bad Board Design

Design rules suggest that MLCCs should be placed away from board edges and panel break out zones for reasons illustrated in this example.

bad-board-design-1

The MLCC was placed far too close to the breakout feature during the design phase for this product.

bad-board-design-2

This is a higher magnification image of the flexure fracture that caused the MLCC to short circuit some time after the board was placed into service.

Check out SEM Lab, Inc.  to learn more.

Solder Joint Failure

This example is a SMT connector where the solder joint appeared suspect.

creep rupture 1

Below is a BSE SEM image of a microsection through the suspect solder joint and its neighbors. Note the coplanarity issues for these connector leads.

creep rupture 2

This appeared to be a creep rupture failure of the solder joint where the lead that failed was under stress that caused creep (time dependent plastic deformation) of the solder joint. The vertical displacement of the lead after the solder joint fractured is the key feature that suggests this was a creep rupture failure.

creep rupture 3

Check out SEM Lab, Inc.  to learn more.

Tombstoning of SMD Inductor

Tombstoning is a well known issue for SMT soldering and is usually caught at post solder visual inspection. However, at times the effect is a subtle lifting at one end of the component as was the case in this example of an SMD inductor. The failure was a high resistance in the associated signal net that wasn’t detected until final assembly and testing of the product.

Tombstone1

This is a BSE SEM image of a microsection of the device as soldered on the PCBA.

Tombstone2

The high resistance was caused by a failure of the original solder reflow process to wet the termination, which caused “tomb stoning” of the inductor.

Check out SEM Lab, Inc.  to learn more.

Silver Sulfide Crystals

Silver is a common material used to connect resistor terminals to the resistive element within the resistor. This example shows some unknown crystals growing out of the termination area.

Silver corrosion

These crystals are shown in the BSE SEM image below.

2000A_1_displayed

EDS analysis showed that these are silver sulfide crystals.

2000A_2_displayed

These crystals most likely formed as a result of gaseous sulfur corrosion of the silver thick film conductor near the termination of the resistors. Corrosion such as the type seen here is typically due to atmospheric sources of sulfur such as diesel fuel fumes or out-gassing from cardboard.

Check out SEM Lab, Inc.  to learn more.

IC Bond Pad Corrosion

Sometimes IC bond pad corrosion is caused by residual process chemistry from the die fabrication or packaging process as was the case in this example.

Screen Shot 2016-08-18 at 9.37.10 AM

The decapsulated IC showed corroded aluminum bond pads.

Screen Shot 2016-08-18 at 9.38.15 AM

EDS results showed a clear chlorine peak associated with the corrosion.

The corrosion caused signals to open circuit resulting in device failure.

Check out SEM Lab, Inc.  to learn more.

Crystal Frequency Shift

Crystal frequency shift

This is a BSE SEM image of a quartz crystal that reportedly shifted in frequency. The problem was apparently caused by tin-lead solder on the spring mounts that wicked out onto the silver contact metallization on the crystals, which would be expected to change the mechanical response and therefore the frequency of the crystal.

Check out SEM Lab, Inc.  to learn more.

Damage in BGA Device

BGAs are susceptible to damage from thermal-mechanical warpage stresses as well as mechanical bending stress.

BGA damage 1

The damage tends to be subtle and isn’t visible at low magnifications such as the image above of a microsectioned device.

BGA damage 2

At higher magnification a fracture became apparent in the upper layers of the die.

BGA damage 3

In the image above, the die attachment appears to have failed and there is a void in the molding compound due to a processing flaw at molding.

BGA damage 4

The bottom of the die separated from the molding compound and a fracture radiates away through the molding compound.

BGA damage 5

However, the PWB laminate under the BGA pads also showed fractures suggesting that the assembled PCBA was bent out of plane. It doesn’t take a great deal of bending to induce this type of damage, so the problem may not be apparent at the moment it is caused.

Check out SEM Lab, Inc.  to learn more.

Manufacturing flaws in ceramic capacitor

ceramic capacitor 1

This is an optical image of the capacitor showing the part number.

ceramic capacitor 2

This is the section after final polish. Note the lack of a solder fillet between the right side termination and the lead.

ceramic capacitor 3

This is a BSE SEM image of the section. These anomalies are parallel in the plane of the capacitor plates (i.e. knit line defects) raising a question about the “as-sintered” strength of the structure.

ceramic capacitor 4

This fracture shows characteristics of thermal shock damage, possibly during attachment of the device leads to the capacitor end caps.

ceramic capacitor 5

This fracture traverses capacitor plates of opposite polarity, which typically results in a shorted capacitor. The fracture appears to propagate from a knit line defect suggesting the root cause is related to the original firing of the multilayer ceramic element.

Check out SEM Lab, Inc.  to learn more.

Red Phosphorus in IC Molding Compound

Red Phosphorus

SEM/EDS is useful for detecting the use of red phosphorus in IC molding compounds.

At one time (~1990s), red phosphorus was introduced as an environmentally-friendly flame retardant in molding compounds for semiconductor devices.

This turned out to be a mistake.

A rash of electrical leakage related failures ensued.

There are now reports of red phosphorus use in wire insulation for cabling and power cords, suggesting that the industry may not have learned its lesson.

Check out SEM Lab, Inc.  to learn more.

ENIG corrosion problem

ENIG corrosion problem

Chlorine residue on this ENIG land caused corrosion of the underlying nickel and copper to form copper and nickel oxides on the gold surface.

Apparently the 3 – 7 micro-inches thick immersion gold layer is not always impervious to chloride migration.

Check out SEM Lab, Inc.  to learn more.

Failure Analysis Images and Discussion