Tin Whisker Observations on Pure Tin-Plated Ceramic Chip Capacitors

AESF SUR/FIN Proceedings AESF

Jay Brusse, QSS Group, Inc., Greenbelt, MD

Tin whiskers pose a risk of intermittent short circuits in terrestrial applications and potentially catastrophic

metal vapor arcs in space (vacuum) environments. Research on whiskering and mitigation practices has

historically produced highly variable results. These variations and recurring observations of whisker-induced

failures are central to the high reliability community's reluctance to adopt lead-free finish alternatives,

especially pure tin, until clear evidence exists that the factors affecting whisker growth are understood and can

be controlled. This paper highlights the variability of previous claims using recent observations of greater than

200-µm-long whiskers on pure tin-plated ceramic chip capacitors as a classic example of the confounding

nature of this phenomenon.

For more information, contact:

Jay Brusse

Senior Components Engineer

QSS Group, Inc. at NASA Goddard Space Flight Center

Bldg. 22, Room 036

Greenbelt, MD 20771

[email protected].gov

This paper was presented at the Electronics Finishing II Session and published in the official proceedings of AESF SUR/FIN 2002,

June 24 - 27, sponsored by the American Electroplaters and Surface Finishers Society, Orlando, FL.

Posted with permission of the AESF.

AESF SUR/FIN Proceedings AESF

Introduction

Electroplated tin (Sn) and tin alloy finishes are used extensively in the electronic components industry to

enhance solderability and corrosion resistance of electrical conductors made from metals such as copper,

phosphor bronze and alloy 42 (iron-nickel alloy). Additional benefits of tin finishes include excellent control

and uniformity of plating thickness (especially critical in miniaturized, fine-pitched components), good

electrical conductivity, and non-toxicity

. "Bright" finishes (those using organic additives in the plating bath to

produce fine grain finishes) are also commonly used because they maintain an aesthetically pleasing shiny

surface. For all of these purposes, a wide assortment of pure tin and tin-lead (2% to 50% Pb) based

electroplated finishes have been used with much success

. Since the mid-1990's, the United States (U.S.)

military and other high reliability users have preferred tin-lead based finishes (minimum 3% Pb) over pure tin

due to documented failures resulting from whisker growths on tin-plated components. These failures will be

briefly described later.

Legislative pressures in recent years (particularly in Japan and Europe) have pushed the electronics industry to

consider methods of eliminating Pb from their products and manufacturing processes

. Generally speaking,

these rules are being developed in response to potential environmental and health hazards that may result from

the manufacturing and disposal of consumer products bearing Pb and other hazardous materials. Although the

U.S. is lagging Japan and Europe in introducing similar legislative restrictions, changes made by the U.S.

Environmental Protection Agency (EPA) in 2001 now require industries to formally report any manufacture,

use and/or disposal of Pb or Pb compounds in excess of 100 pounds per year

. Previously, the EPA reporting

limit was as high as 25,000 pounds per year. These EPA rule changes and the desire to remain competitive in a

world economy have prompted many U.S. electronic component manufacturers to explore alternatives to tin-

lead based plating systems.

Recently, significant attention has been given to pure tin plating (especially “matte” finishes) as a candidate to

replace widely used tin-lead finishes. With several factors in its favor, such as ease of converting existing tin-

lead plating systems, ease of manufacture and compatibility with existing assembly methods, in addition to

years of successful commercial use, pure tin plating is seen by many in the industry as a potentially simple and

cost effective alternative

Despite the benefits, there remains one major impediment to the simple adoption of pure tin for use in high

reliability applications: many pure tin electroplates develop potentially damaging growths known as tin

whiskers. Assured methods for predicting if or when such growths may occur do not currently exist.

Hazards of Tin Whiskers

Figure 1 shows examples of tin whiskers on various types of pure tin-plated electronic components. As the

name implies, tin whiskers are hair-like growths of near-perfect single-crystalline structures of tin that grow

outward from some electroplated tin surfaces. Typical tin whiskers have diameters of only a few microns (µm)

and have been reported to grow to lengths of several millimeters (mm), though lengths of less than 1-mm are

typically reported

1,6,7

. Their tiny dimensions make them extremely difficult to see unless proper illumination

and/or high power inspection techniques are used such as scanning electron microscopy (SEM). In addition,

troubleshooting of system problems may sometimes obscure whiskers as the root cause of failure if handling or

probing is used which results in removal of the offending whisker.

Predicting if or when a particular tin-plated product will form whiskers has perplexed researchers and users

alike for decades. Perhaps the most insidious traits of tin whiskers are the seemingly unpredictable nature of

AESF SUR/FIN Proceedings AESF

their growth initiation and subsequent growth rate. Although several researchers have reported whisker

incubation periods as short as a few days from the time of electroplating, others have documented incubation

periods that have lasted many years

8,9

. Growth rates as high as 9mm/year have been reported, but slower rates

are much more common

. These attributes of whisker growth are particularly challenging because of the

lengthy experiments (years) needed to establish meaningful results. Formally accepted test methods

(environments) to accelerate whisker growth have not been established. Various industry groups in the U.S.,

Japan and Europe are presently working to develop such effective screening practices

The primary concern with tin whiskers is that they are electrically conductive and can lead to short circuits in

electronic assemblies. It is tempting to expect these tiny conductive filaments will fuse open at the instant they

grow long enough to cause a short creating perhaps an unnoticeable "glitch" to the circuit performance.

However, documented experiences have shown tin whisker failures also include permanent low current (few

milliamps or less) shorts as well as catastrophic metal vapor arcs in vacuum capable of sustaining HUNDREDS

of AMPERES

. Table 1 categorizes the common failure modes attributed to tin whiskers while table 2 lists

some notable examples of tin whisker induced failures in medical, military and space applications.

In previous work performed at the NASA Goddard Space Flight Center (GSFC), it has been demonstrated that

tin whiskers (especially the longer, filament types) will bend in response to forces of electrostatic attraction (see

Figure 2)

. Therefore, a tin whisker growing from a surface at one electrical potential will be attracted towards

an adjacent conductor with a different potential. This behavior is highlighted to show that the probability of a

tin whisker shorting event is greater than a "random" occurrence dependent solely on the whisker's direction of

growth.

AESF SUR/FIN Proceedings AESF

A) Ceramic Chip Capacitor Termination

B) Electromagnetic Relay Terminal

C) Hybrid Microcircuit Lid D) Terminal Lug

E) Test Points F) DIP Microcircuit Lead

Figure 1: Examples of Electronic Components with Tin Whiskers (All are Pure Tin-Plated)

Table 1: Equipment Failure Modes Induced by Tin Whiskers

Failure Mode Description

Permanent Short Circuits

In low voltage, high impedance circuits there may be insufficient current

available to fuse the whisker open. Fusing currents between 30 mA and 75 mA

have been reported

8, 24

Transient Short Circuits

If the available current exceeds the fusing current of the whisker, the circuit may

only experience a transient glitch as the whisker fuses open.

Metal Vapor (Plasma) Arcing

In Vacuum

In vacuum (reduced atmospheric pressure) a much more destructive short circuit

phenomenon can occur. If currents of above a few amps are available and the

supply voltage is above approximately 18V, the whisker may vaporize creating a

plasma of tin ions that can conduct Hundreds of Amperes. An adequate supply

of tin from the surrounding plated surface can help to sustain the arc until the

available tin is consumed or the supply current is interrupted such as occurs

when a protective fuse or circuit breaker interrupts the current flow.

Debris/Contamination

Mechanical shock, vibration or handling may cause whiskers to break loose from

the plated surface. Once free to move about, these conductive particles may then

interfere with sensitive optical surfaces or the movement of

microelectromechanical systems (MEMS). This debris may also lead to short

circuits as noted above

AESF SUR/FIN Proceedings AESF

Before applying

50V to test

robe

After applying

50V to test

robe

Figure 2. Tin Whisker "Bending" in Response to Electrostatic Attraction.

(Note: Photo has Been Enhanced to Aid in Observation of the Tin Whisker which is Difficult to See in this Optical Photograph)

Table 2: Reported Field Problems Induced by Tin Whiskers

Application Problem

Medical

Heart Pacemaker

Class I Product Recall: Tin whisker shorts from pure tin-plated housing of

crystal causes complete loss of pacemaker output.

Military

F-15 Radar

Tin whisker from pure tin-plated hybrid microcircuit lids.

U.S. Missile Program

Tin whisker from pure tin-plated relays

U.S. Missile Program

28,29

Tin whisker growing from pure tin-plated TO-3 transistor can shorts collector to

case.

Phoenix Air to Air Missile

Tin whisker shorts inside hybrid microcircuit.

Patriot Missile II

Tin whiskers from pure tin-plated terminals

Space (Satellite)

GALAXY IV

32,33

Complete loss of satellite operations. Tin whisker short (metal vapor arc in

vacuum) from pure tin-plated relays

GALAXY VII

Complete loss of satellite operations. Tin whisker short (metal vapor arc in

vacuum) from pure tin-plated relays

SOLIDARIDAD I

Complete loss of satellite operations. Tin whisker short (metal vapor arc in

vacuum) from pure tin-plated relays

Additional Satellites

32,33

Three additional satellites of the same general design as above have lost one of

two redundant satellite control processors due to tin whisker shorts

Tin Whisker @

Ground Potential

AESF SUR/FIN Proceedings AESF

Compressive Stress Theory of Whisker Growth

Although tin whiskers were first reported in the 1940’s

, researchers still have not reached consensus to

explain the mechanism(s) that drives whisker formation. However, most experts concur that whisker formation

is the result of a mechanical stress-relief process. More precisely, it has been postulated that the development of

"compressive" stress in the tin layer provides the fundamental driving force for whisker growth

2,15,16

. The

following factors have been cited in the literature as potential contributors to compressive stress in the tin

layer

• Substrate Element Diffusion (e.g., copper, zinc) into Tin Layer

• Intermetallic Compound (IMC) Formation between Tin and Substrate (e.g., Cu

)

• Plating Chemistry (e.g., Organic additives termed “brighteners” may increase the "as-plated" residual stress)

• Plating Process Parameters (e.g., higher current density may produce high residual stress)

• Contamination (organic and inorganic)

• Substrate Stress (e.g., stamping, cold working, swaging)

• Environmental Stresses (temperature, humidity)

• Externally Applied Mechanical Stress (e.g., torquing of a fastener, scratches from handling)

Lee and Lee

have described one plausible model for whisker growth that is heavily reliant upon the processes

of diffusion and intermetallic compound (IMC) formation occurring between the tin layer and the underlying

substrate. In this model atoms with high diffusivity in tin (e.g., copper or zinc) migrate from the substrate

material (phosphor bronze in their study) into the tin layer preferentially along the tin grain boundaries. The

diffusing elements and the resulting IMCs (e.g., Cu

) occupy space within the tin crystallographic structure

producing a resultant compressive stress in the tin layer. As the stress increases with time, the thin, brittle tin

oxide layer that forms over the electroplated tin can rupture. Once the oxide layer has ruptured, tin grains

(whiskers) may then be extruded in a continuous fashion through the oxide layer as a means of relaxing

compressive stresses. As growth proceeds, the localized compressive stress tends to reduce until the stress level

is no longer sufficient to support further growth (equilibrium).

Based upon “compressive” stress theories such as this one, the majority of whisker mitigation practices being

explored today involve attempts to minimize or eliminate those factors which encourage the development of

compressive stresses within the tin electroplate. Multilayer ceramic capacitors (MLCCs) are one example of

tin-plated components where whisker mitigation practices used in their manufacture have been studied and

reported.

Multilayer Ceramic Capacitor Construction

Today, MLCCs are utilized in almost every electronics application. Annual production of MLCCs exceeds

several billion capacitors worldwide. MLCCs come in a wide range of capacitance/voltage ratings and package

configurations (leaded and surface mount). Surface mount chip MLCCs are by far the most popular package

configuration and are currently available with footprints as small as 0.5-mm x 0.25-mm (0.02-in. x 0.01-in.).

Figure 3 illustrates the typical construction of a ceramic chip capacitor designed for installation by soldering

methods.

AESF SUR/FIN Proceedings AESF

Figure 3. Typical Cutaway View of a Tin-Plated Multilayer Ceramic Chip Capacitor (Kemet Electronics Corp Data Sheet)

The capacitor element consists of layers of a ceramic-based dielectric material (e.g., barium titanate) with metal

electrodes (e.g., palladium-silver [Pd-Ag]) interleaved. The ceramic-electrode structure is sintered at high

temperature to form a homogenous block. Conductive terminations are applied to the ends of this ceramic

block to provide electrical contact points for this surface mountable device. The end termination structure

typically consists of three distinct layers as follows:

1. Silver Frit: Silver paste (with some glass frit mixed in) fired onto the ceramic to make electrical connection

to the electrode layers. The glass facilitates mechanical adhesion to the ceramic.

2. Nickel: Typically electroplated ~5-µm thick applied over the silver frit to minimize diffusion of silver into

the final solderable finish.

3. Tin or Tin-Lead Alloy: Typically electroplated ~5-µm to 10-µm thick over the nickel barrier layer to

provide a solderable final finish.

Note: For wire bonding or conductive adhesive (e.g., silver epoxy) attachment applications, MLCCs with gold or Pd-Ag final finishes

are also available instead of tin or tin-lead.

Previous Whisker Research on MLCCs

In the vast research conducted to date on tin whiskers, there are only a few studies that directly discuss tin

whisker growth in passive electronic components such as MLCCs

18-21

. These papers generally proclaim MLCCs

to be immune to tin whisker formation based on a number of characteristics of the MLCC termination structure.

Selcuker

and Endo

have categorized the following attributes of the typical MLCC construction and

manufacturing processes as factors that reduce their propensity to form whiskers:

1. "Large" (>5-µm) and well-polygonized tin grain structure

2. "Thick" tin-plated layer (~5 - 10-µm)

3. Electroplated Nickel barrier layer (at least 2-µm thick) limits silver diffusion into tin and itself has very low

diffusivity into tin.

4. Post-electroplate annealing processes help reduce residual stress in the tin layer and stimulate tin grain

growth

Recent research employing pure tin finishes over nickel barrier layers over copper substrates has postulated that

the nickel layer may also impart a "tensile" (rather than a compressive) stress to the tin layer

. If this is true,

then theoretically such tensile stresses tend to discourage whisker extrusion.

AESF SUR/FIN Proceedings AESF

In 1997 Endo

reported 18 years worth of whisker-free observation of MLCCs stored continuously at 50°C.

This storage condition was chosen because of previous whisker research citing 50°C as the optimal temperature

for whisker formation to occur.

In 2001, the author of this present work surveyed five major domestic manufacturers of MLCCs to solicit their

experience with tin whiskers

. All of those surveyed indicated they knew of no formal reports of tin whiskers

on any of their MLCC products.

Recent Experimental Observations of Tin Whiskers on MLCCs after Temperature Cycling

To illustrate the confounding nature of the tin whisker phenomenon, several recent examples of profuse tin

whisker formation (some approaching 250-µm long) on pure tin-plated MLCCs (Figure 4) are described. The

construction of these MLCCs is typical of the MLCC manufacturing industry and is comparable to MLCCs

previously studied. Despite having most of the "whisker reducing" attributes described in previous studies

18,19

it will be shown that these capacitors formed tin whiskers when exposed to environmental test conditions

different from those used in prior research.

Figure 4. Pure Tin-Plated MLCC from Manufacturer "A" with Tin Whiskers (Max. Whisker Length ~100-µm in this Image)

The following examples of tin whiskers on pure tin-plated MLCCs were collected from formal and informal

exchanges of information with Original electronic Equipment Manufacturers (OEMs) who reported their

observations to engineers at the NASA Goddard Space Flight Center (GSFC). As noted, some of the work has

been augmented by examination at GSFC and independent experiments performed by The Aerospace

Corporation

to confirm the effects of environment reported by other researchers. In all cases, these examples

are presented with due concern for the confidentiality of the OEMs involved.

Example #1: Pure Tin-Plated MLCCs from Manufacturer "A" in Hybrid Microcircuit

In 2000 an OEM ordered Pd-Ag terminated MLCCs from a major manufacturer (Manufacturer "A"). The

MLCC manufacturer mistakenly supplied pure tin-plated capacitors due to a logistical error. The OEM's intent

was to mount Pd-Ag terminated MLCCs onto gold plated mounting pads of a substrate inside a hermetically

AESF SUR/FIN Proceedings AESF

sealed hybrid microcircuit. The OEM's mounting process employs a widely used silver-loaded conductive

epoxy to accomplish the bond between the substrate pads and the MLCC terminations (Figure 5). Once

assembled, the hybrids are hermetically sealed in a Nitrogen atmosphere. These assembly practices are

commonly used in the manufacture of complex hybrid modules. However, the use of pure tin MLCCs with

conductive epoxy is not a recommended practice due to potential mechanical and electrical integrity concerns

when bonding epoxies to tin or tin-lead.

Figure 5: Optical View of Hybrid Package Substrate with Epoxy Mounted Pure Tin-Plated MLCCs

(MLCC dimensions 2.03 mm x 1.27 mm)

Example #1: Description of Experiment

In the present example, the OEM was fortunate to catch the MLCC manufacturer's mistaken shipment of pure

tin prior to issuing the MLCCs to the production floor. Rather than return the erroneous shipment for

replacement, the OEM opted to conduct an evaluation to determine if the pure tin MLCCs may still meet the

performance requirements of their application even when mounted using conductive epoxy. Therefore, the

OEM manufactured and sealed several representative hybrid packages utilizing the pure tin-plated MLCCs and

conductive epoxy mounting methods as described above. The hybrids were divided into two groups for the

following environmental test conditions designed to evaluate the electrical and mechanical integrity of this

bonding technique:

Condition 1: Thermal Cycle

-40°C to +90°C up to 500 cycles

Condition 2: High Temperature Storage

+90°C for 400 hours

The OEM removed representative hybrid packages from test at discrete intervals of time for inspection. The

inspection test protocol consisted of delidding the hybrid to enable optical and scanning electron microscope

(SEM) inspection of the capacitors, electrical conductivity measurements of the epoxy joints and mechanical

strength measurements of the bond via die shear testing.

Example #1: Experimental Results

One representative hybrid package was removed from test Condition 1 (temperature cycling) after 300+ cycles.

During visual inspection the user was completely surprised to find the entire termination surface of every

AESF SUR/FIN Proceedings AESF

MLCC in the package was covered with a dense forest of “moss-like” tin whiskers with some approaching

100-µm long (Figure 4).

Additional packages were inspected after 500 thermal cycles. The OEM reported whisker growths on the

MLCCs were longer than those observed at the 300-cycle inspection point. At this point in the evaluation the

OEM determined that the size and density of the tin whiskers observed were unacceptable for their densely

packaged circuitry.

Hybrid packages subjected to test Condition 2 (high temperature storage) were inspected after 400 hours of test.

Interestingly, there was no evidence of tin whiskers for specimens stored under this condition.

Incidentally, the user’s analysis of the bond integrity after Conditions 1 and 2 revealed less than desirable

adhesion strength and electrical conductivity. These results also show that it is not advisable to use conductive

epoxy mounting practices with pure tin-plated MLCCs. However, OEMs that do not have safeguards in place

to protect against incorrect shipment of pure tin-plated MLCCs when Pd-Ag MLCCs are ordered may be prone

to problems such as observed in Example #1.

Example #1: Effects of Additional Ambient Storage on MLCCs with Tin Whiskers

Component engineers at the NASA Goddard Space Flight Center (GSFC) learned of this OEM's dilemma

through a technical email discussion forum catering to topics in electronics assembly. Intrigued by the OEM's

experimental observations, GSFC engineers petitioned the user for representative MLCCs with tin whiskers to

further document the whisker formations and analyze the MLCC termination construction. One hybrid package

containing six epoxy mounted MLCCs was supplied for this analysis.

The OEM reported that the representative hybrid had been subjected to 200+ thermal cycles of test Condition 1.

As received by GSFC, the hybrid package had already been delidded by the OEM for optical inspection only

(i.e., the MLCCs were still epoxied to the substrate). Upon receipt of the hybrid, GSFC parts analysts performed

SEM inspection. Figure 6 shows the profuse density of growths (as high as 800 per mm

) and maximum

whisker lengths on the order of 100-µm thus confirming the OEM's previous reports.

Figure 6: NASA GSFC SEM Inspection of Pure Tin-Plated MLCC from Manufacturer "A" with Tin Whiskers

AESF SUR/FIN Proceedings AESF

The hybrid package with MLCCs still attached was placed in a protective enclosure to minimize exposure to

dust and handling contamination. Ongoing storage conditions consist of a room ambient environment which for

this GSFC facility is nominally 20°C - 25°C, 30% - 60% relative humidity.

After approximately six months of additional storage time under the GSFC ambient conditions, another SEM

inspection was conducted. Surprisingly, this inspection (Figure 7) revealed several whiskers had grown in

excess of 200-µm long (max. of ~250- µm). This finding suggests that the rapid whisker growth process

initiated by thermal cycling has continued despite the removal of the initiating environment.

Figure 7. NASA GSFC SEM Inspection of MLCCs after ~ 6 Months of Additional Ambient Storage After Original Thermal Cycle

Exposure.

Example #1: Cross Section Analysis of MLCCs with Tin Whiskers

During the initial receiving inspection at GSFC, one MLCC was mechanically removed from the hybrid

substrate (using a die shear tester) for cross section analysis. The purpose of this analysis was to document

possible construction differences between these whisker-laden MLCCs from the MLCCs described in previous

research on tin whiskers.

Figure 8 shows the cross section of the MLCC that was removed from the hybrid. SEM inspection and Electron

Dispersive Spectroscopy (EDS) were performed to measure the various termination layer thicknesses and

composition. The analysis showed the termination structure as follows:

1. Silver Frit: 17-µm thick

2. Nickel Barrier: 6.5-µm thick

3. Pure Tin: 6.5-µm thick

AESF SUR/FIN Proceedings AESF

Ceramic

Silver Frit

Nickel

Tin

Figure 8: Cross-Section of MLCC Indicating the Thickness and Composition of the Various Layers

The bulk ceramic dielectric was observed to be of a commonly used barium titanate formulation.

The cross-sectioned MLCC was next chemically etched to highlight the grain structure of the tin electroplate.

Though difficult to see in Figure 9, this analysis showed that the tin grains were polygonal structures ranging

from 5 to 15-µm in size.

Figure 9. Tin Etch of Cross Section of MLCC Termination Structure with Grain Boundaries Highlighted

The results of the cross section analysis show that the attributes of this MLCC are essentially identical to those

of the MLCCs reported to be whisker-free after 18 years in the earlier studies.

The most obvious difference between the experiments of the different researchers is the environmental test

conditions chosen for judging whisker propensity. The more recent OEM's experiments described in Example

#1 suggest that exposure to thermal cycle environments more actively excites the necessary stresses to spawn

whisker formation in devices with a construction like MLCCs. Coefficient of thermal expansion mismatches

among the various materials in the MLCC construction or possible IMC formation between nickel and tin are

suspected to be potential factors contributing to whisker formation in this study. However, the exact

mechanisms involved in these growths have not been determined.

Example #2: Thermal Cycle Evaluation of Pure Tin-Plated MLCCs from Manufacturers "B" and "C"

Having heard the reports of MLCCs with tin whiskers described by Example #1, The Aerospace Corporation

performed experiments designed to validate these previous observations. Despite the fact that the OEM

application would utilize reflow soldering of the pure tin MLCCs using Sn/Pb solder paste, the OEM was

µµ

AESF SUR/FIN Proceedings AESF

concerned that portions of the capacitor terminations may remain unreflowed or tin-rich as some soldering

processes do not guarantee complete wetting of the MLCC termination with Sn/Pb solder.

In a study to assess the risk of tin whisker formation, The Aerospace Corporation obtained pure tin-plated

MLCCs from two different manufacturers ("B" and "C"). The MLCCs from Manufacturer "B" are typical

commercial grade capacitors whose cross section is shown in Figure 10. Note that this lot of MLCCs contains a

nickel barrier layer ~ 6-µm thick with a pure tin final finish ~10-µm thick

Figure 10. Cross-Section of Pure Tin-Plated MLCC from Manufacturer "B"

("Typical" MLCC Construction with Nickel Barrier Layer)

The production lot of MLCCs from Manufacturer "C" was also of similar construction; however, this lot was

manufactured in accordance with U.S. Military specification requirements (MIL-PRF-55681).

To simulate the effects of reflow temperature exposure that might be expected from typical installation

methods, several parts from Manufacturer "B" were exposed to temperatures of 215°C for 5 seconds without

actual solder present. Additional samples from Manufacturer "B" and all of the samples from Manufacturer "C"

were not subjected to this simulated reflow.

All of the capacitors were then subjected to the same environmental test described by Condition #1 of Example

#1 (thermal cycle). Capacitors from Manufacturer "B" were examined after 500+ cycles and densely populated

regions of tin whiskers up to ~25-µm in length were observed (Figure 11) even for those parts exposed to

simulated reflow. The MLCCs from Manufacturer "C" were examined after 100 cycles and also showed

evidence of whiskers ~25 to 30-µm long (Figure 12).

Figure 11. Tin Whiskers on Pure Tin-Plated MLCCs from Manufacturer "B" After Thermal Cycle

AESF SUR/FIN Proceedings AESF

Figure 12. Tin Whiskers on Pure Tin-Plated MLCCs from Manufacturer "C" After Thermal Cycle

The results of these experiments confirmed most of the observations from Example #1. However, the

maximum length of the whiskers found in Example #1 were from 3 to 10 times longer than those observed in

Example #2. It will be interesting to see if the whiskers from Example #2 continue to grow under ambient

storage as reported in Example #1.

Example #3: Tin Whiskers on Pure Tin-Plated MLCCs from Manufacturer "B" After Vapor Phase

Installation and Thermal Cycle Testing

Many who have heard of the intriguing observations described in Examples #1 and #2 have suggested that these

capacitors would not have whiskered if they had been installed using soldering techniques for which tin-plated

MLCCs are designed. They cite references in the tin whisker literature that suggest the heat of the reflow

process will relax residual stresses in the plating thought to be key contributors to whisker formation. They also

cite references that suggest the introduction of Pb from the Sn-Pb solders commonly used to mount this type of

capacitor would cover most of, if not the entire, MLCC termination surface thus introducing an alloying

element (Pb) that is widely considered to have whisker inhibiting properties.

Circa 1997, a different OEM conducted experiments to determine the whiskering propensity of MLCCs after

vapor phase installation at 217°C using Sn63/Pb37 solder paste. Pure tin-plated MLCCs of four different chip

sizes all from Manufacturer "B" were assembled onto FR4 printed wiring boards using the OEMs standard

vapor phase process.

Some of the assembled boards were then subjected to thermal cycle testing from -55°C to +100°C for several

hundred cycles. SEM inspection was performed at discrete intervals to examine for evidence of tin whiskers.

Upon examination after 50 cycles, the two larger MLCC chip sizes showed evidence of tin whiskers (~10-µm

long) growing from the tin-rich areas which were not wetted by Sn/Pb solder during installation. Additional

cycling to 400 cycles showed evidence of tin whiskers approaching 30-µm long.

Example #4: Tin Whiskers on Pure Tin-Plated MLCCs from Ambient Storage

In 2000, the author heard anecdotal reports of pure tin-plated MLCCs from a fourth manufacturer

(Manufacturer "D") that showed "moss-like" growths on the terminations of parts taken from stock. The OEM

who made these observations reportedly had ordered Pd-Ag terminated capacitors for their hybrid epoxy mount

application, but was supplied with pure tin-plated MLCCs by mistake (similar to the scenario described by

Example #1). Unfortunately, formal documentation of this example has not been made available.

AESF SUR/FIN Proceedings AESF

Conclusions

The observations and circumstances surrounding tin whisker formation on multilayer ceramic capacitors

(MLCC) from four different manufacturers have been described. All of these observations provide a

contrasting view of previously held beliefs that MLCCs are immune to tin whisker formation.

1. Surface mount multilayer ceramic capacitors (MLCCs) with electroplated pure tin finishes are not immune

to tin whisker formation. Thermal cycling has been demonstrated to excite tin whisker formation in MLCCs

with some documented growths approaching 250-µm.

2. Electrical shorting due to tin whiskers remains a significant problem. The incidence of tin whisker induced

failures will increase in frequency with increased use of pure tin-plated electronic components and

assemblies unless significant discoveries are made regarding effective mitigation practices.

3. Even when prohibited by system design and procurement requirements, components manufactured with

pure tin finishes continue to appear in electronic equipment.

4. The factors that affect tin whisker formation are not fully understood. The author believes that there are

also interactions amongst the various factors involved in whisker growth that makes research of this

phenomenon difficult to control. Despite efforts to perform controlled experiments to judge whisker

propensity, many subsequent experiments show “other” factors that can change previously held beliefs.

Recommendations

There are literally hundreds of published papers that discuss general and specific aspects of the tin whisker

phenomenon. Reliance on simply a few of the more prominent publications on the subject could prove to be

dangerous to manufacturers and users of electronic components and assemblies requiring the utmost in long-

term reliable performance of their products. Despite the fact that this phenomenon has been known for more

than 50 years, all must be aware of its confounding and sometimes conflicting nature.

1. The electronics manufacturing industry must strive to achieve and verify through direct investigation a

consensus understanding of the fundamental mechanism(s) of whisker formation.

2. "Accelerated" test method(s) to judge a product or process's propensity to grow tin whiskers need to be

developed. Methods must be tailorable to assess a wide variety of electronic component constructions and

user application conditions including factors such as temperature, temperature change, humidity, and

pressure.

3. Experiences with tin (and other metal) whiskers need to be shared more openly in order to create an

environment conducive to resolution of this dilemma.

4. OEMs and electronic component manufacturers should adopt practices to assess and mitigate the risk of tin

whisker induced failures in their products.

Acknowledgments

The author would like to thank Mike Sampson, Dr. Henning Leidecker (NASA Goddard) and Jong Kadesch

(Orbital Sciences Corporation) for their support and contributions to this paper. In addition, the support of Dr.

Gary Ewell and Jocelyn Siplon (The Aerospace Corporation) is greatly acknowledged. Finally, the author

expresses his gratitude to Ingemar Hernefjord for valuable discussions.

AESF SUR/FIN Proceedings AESF

References:

1. P. Hinton, "Tin Plating,Tin-Nickel Electroplate And Tin Plating Over Nickel As Final Finishes On Copper", Proceedings of the

Technical Program, Surface Mount International , San Jose, CA September 10-12, 1996, pp 806-810

2. Y. Zhang, G. Breck, F. Humiec, K. Murski & J.A. Abys, "An Alternative Surface Finish for Tin/Lead Solders: Pure Tin", SMI'96

Proceedings, San Jose, CA, Sept. 1996

3. European Union, "Waste Electrical and Electronic Equipment (WEEE) and Restriction of Hazardous Substances",

http://www.lead-free.org/legislation/detail/soldertec-overview.html

U.S. Environmental Protection Agency Toxic Release Inventory (TRI) Program, "Lead and Lead Compounds; Lowering of

Reporting Thresholds; Community Right-to-Know Toxic Chemical Release Reporting; Final Rule. 66 Federal Register, 4499-

4547", January 17, 2001,

http://www.epa.gov/tri/lawsandregs/lead/reporting_pb.htm

5. S. Winkler & B. Hom, "A Look at the Past Reveals a Leadfree Drop-In Replacement", High Density Interconnect (HDI) On-Line,

March 2001.

6. P. Key, "Surface Morphology of Whisker Crystals of Tin, Zinc and Cadmium," IEEE Electronic Components Conference, pp.

155-157, May 1970.

7. R. Diehl & N. Cifaldi, "Eliminate Whisker Growth on Contacts by Using a Tin Alloy Plate", Insulation/Circuits, April 1976 pp.

37-39

8. B. Dunn, "A Laboratory Study of Tin Whisker Growth," European Space Agency (ESA) STR-223, pp. 1 - 50, September 1987.

9. B. Hampshire & L. Hymes, "Shaving Tin Whiskers", Circuits Assembly, pp. 50-53, September 2000

10. G. Stupian, "Tin Whiskers in Electronic Circuits," Aerospace Report No. TR-92(2925)-7, pp. 1 - 21, December 20, 1992.

11. Private communications with members of a forum sponsored by the National Electronics Manufacturing Initiative (NEMI).

Forum established in 2001 to discuss tin whisker testing and fundamental growth mechanisms.

12. D. Van Westerhuyzen, P. Backes, J. Linder, S. Merrell & R. Poeschel, “Tin Whisker Induced Failure In Vacuum”, Proceedings

International Symposium for Testing & Failure Analysis, pp. 407-412, October 17, 1992.

13. J. Kadesch & J. Brusse, "The Continuing Dangers of Tin Whiskers and Attempts to Control Them with Conformal Coating",

NASA EEE Links Newsletter, July 2001, http://grizzly.gsfc.nasa.gov/eeelinks/July2001/

14. H. Cobb, "The Monthly Review", American Electroplaters Soc., Vol. 33, 1946. p. 28

15. B. Lee & D Lee, "Spontaneous Growth Mechanism of Tin Whiskers", Acta Mater., Vol. 46, No. 10, pp. 3701-3714, 1998

16. R. Schetty, "Minimization of Tin Whisker Formation for Lead Free Electronics Finishing", Proceedings IPC Works Conference,

Miami, USA, 2000.

17. J. Brusse, G. Ewell & J. Siplon, "Tin Whiskers: Attributes and Mitigation", Capacitor and Resistor Technology Symposium

(CARTS), March 26, 2002.

18. A. Selcuker, & M. Johnson, "Microstructural Characterization of Electrodeposited Tin Layer in Relation to Whisker Growth",

Proceedings Capacitor and Resistor Technology Symposium: USA, pp. 19 - 22, October 1990.

19. M. Endo, S. Higuchi, Y. Tokuda, & Y. Sakabe, "Elimination of Whisker Growth on Tin-Plated Electrodes", Proceedings of the

23rd International Symposium for Testing and Failure Analysis, pp. 305 - 311, October 27-31, 1997.

20. R. Kuhl & S. Mills, "Assuring Whisker-Free Components," Surface Mount Technology, Vol. 9, 1995, p. 48

21. G. Ewell & F. Moore, “Tin Whiskers and Passive Components: A Review of the Concerns”, Proceedings 18

Capacitor and

Resistor Technology Symposium, March 1998.

22. C. Xu, Y. Zhang, C. Fan, J. Abys, et al, "Understanding Whisker Phenomenon: Driving Force for the Whisker Formation",

Proceedings of the Technical Conference APEX, Jan. 19, 2002, pp. S06-2.1 to S06-2.6

23. J. Brusse, "Survey of MIL Ceramic Chip Capacitor Manufacturers Regarding Termination Finishes", memo May 23, 2001,

http://grizzly.gsfc.nasa.gov/whisker/ceramic_cap/ceramic_cap_term_finish_survey_for_www.pdf

24. Y. Hada, O. Morikawa & H. Togami, "Study of Tin Whiskers on Electromagnetic Relay Parts," 26th Annual National Relay

Conference, pp. 9.1 - 9.15, April 25-26, 1978.

25. B. Nordwall, "Air Force Links Radar Problems to Growth of Tin Whiskers", Aviation Week and Space Technology, June 20,

1986, pp. 65-70

26. Food and Drug Administration, "ITG #42: Tin Whiskers- Problems, Causes and Solutions",

http://www.fda.gov/ora/inspect_ref/itg/itg42.html, March 16, 1986

27. K Heutel & R. Vetter, "Problem Notification: Tin Whisker Growth in Electronic Assemblies", Feb. 19, 1988, internal memo

28. M. McDowell, "Tin Whiskers: A Case Study, (USAF)", Aerospace Applications Conference, pp. 207 -215, 1993.

29. J. Richardson, and B. Lasley, "Tin Whisker Initiated Vacuum Metal Arcing in Spacecraft Electronics," Proceedings 1992

Government Microcircuit Applications Conference, Vol. XVIII, pp. 119 - 122, November 10 - 12, 1992.

AESF SUR/FIN Proceedings AESF

30. L. Corbid, "Constraints on the Use of Tin-Plate in Miniature Electronic Circuits", Proceedings 3

International SAMPE

Electronics Conference, pp. 773-779, June 20-22, 1989.

31. Anoplate WWW Site: http://www.anoplate.com/news/pastnews/fall2000/tin.htm

32. R. Dore, "Launches of Hughes HS 601 Satellites Ready to Resume", Hughes Press Release, Aug. 1998

33. Satellite News Digest WWW Site, "HS601 Satellite Failures", http://www.sat-index.com/failures/