Lead Free Electronics, Tin Whisker Risks, and Mitigation ...

Lead Free Electronics, Tin Whisker
Risks, and Mitigation Techniques

Fred W. Verdi
ACI Technologies, Inc.
One International Plaza, Suite 600
Philadelphia, PA 19113
3/11/09

Table of Contents

Introduction......................................................................................................................... 1
Metallurgical Theory .......................................................................................................... 4
Mechanical Stress Model.................................................................................................... 6
In 10, 20, and 30 Years, How Long Will These Whiskers Be? .......................................... 9
References......................................................................................................................... 11

Introduction

In 2006, the European Union (EU) passed the law called Restriction of Hazardous
Substances (RoHS) that restricted lead (Pb) to 0.1% or less, by weight, in any
homogeneous material or alloy (all surface finishes, solder joints, and coatings) in any
commercial electronic equipment offered for sale on the European market. Responding to
this law, Printed Wiring Board (PWB) and electronic component vendors changed the
surface finish for electronic components to pure tin (rather than the SnPb alloy they had
used for the last 60 years), and the PWB finish to immersion tin, immersion silver, or
immersion gold plating.

Unfortunately, both the electroplated pure tin now predominantly being coated on
components and the immersion tin now being applied to Pb-free Printed Wiring Boards
(PWBs) are susceptible to the electrically conductive tin whisker growths. Since the tin
whiskers are documented to conduct several milliamps of electrical current without
fusing (melting), they have caused short circuit failures of both military and commercial
electronics (e.g. satellites, radars, heart pacemakers, etc) that utilize tin coated parts (Pb-
free electronics). Tin whiskers can also vaporize at higher currents and voltages, such as
in power supplies, creating a tin plasma that can conduct many tens or hundreds of
amperes, thus causing catastrophic damage to the power supply.

Even though the military is not bound by the RoHS law, the military must use, in many
cases, the same types of tin-plated components demanded by the vastly larger
commercial electronics customer base. Commercial electronics now constitute
approximately 95-to-98 per cent of the worldwide electronics market, while the RoHS
exempt military applications represent the other 2 to 5%. Thus, tin-lead plated
components are becoming less and less available from the component and PWB vendors
as the vendors convert to pure tin plating. The military simply does not possess the
economic leverage to influence the type of surface finishes offered by the vendors.
Tin whisker growth has become a prime source of risk with the switch to pure tin plating
for lead-free (RoHS compliant) component terminals and circuit boards.
Tin whisker growth is not a new problem (figure 1). The Bell System documented the
growth of tin, cadmium, and zinc whiskers and subsequent short circuit failures, on
electroplated hardware in telephone switching equipment as early as 1946.(1) After
extensive study, Bell Laboratories recommended alloying the tin with lead, and the result
has been essentially whisker free electronic assemblies, using tin-lead alloy plating on
component leads and circuit boards and tin-lead solder attachment in assembly, for the
last 60 years.

Figure 1. Early photograph of spontaneously growing, tenth-inch long single crystal
tin whiskers from Bell Laboratories investigation of electrical short circuit failures
from tin plated details in telephone switching equipment in the 1940s.

From the 1940s until the last decade of the twentieth century, tin-lead electroplate and
tin-lead solder were routinely used for electronic assembly, with virtually no chance of
tin whisker failures because of the Pb content. The history since then begins to show the
growing menace and attempts at mitigation.

In 1993, M. E. McDowell of the United States Air Force (USAF) (2) outlined the method
used by the USAF in dispositioning tin (Sn) plated parts in inventory. No position was
taken relative to the prohibition of pure Sn usage (as previously recommended by Dunn
of the European Space Agency(3)). This would prove to be an unfortunate situation, as
later events were to show, relative to reliability failures on USAF electronic equipment.
Between 1990 and 2004, Brusse from NASA compiled a list of high reliability system
failures caused by metal whisker growth. Tin, Zinc, and Cadmium are all highly prone to
whisker growth. This list of failures in the time period from 1990 to 2004 counts scores
of well-documented whisker failures in electronic assemblies, from heart pacemakers to
NASA and commercial satellites, and many millions of dollars in monetary losses. A
typical tin whisker shorting issue is depicted in Figure 2 (After Brusse/NASA ref. 4).

Figure 2. Tin whiskers causing electrical shorts between connector pins after 10
years in service. (Photo courtesy of NASA, 2003)

This increased incidence of whisker failures has not gone unnoticed by the scientific
community.
In the first two years of the 21st century, there were more presentations/papers on Sn
whisker matters than in the prior 15 years. The advent of lead free electronics
manufacturing (in response to environmental legislation) undoubtedly accounts for much
of this renewed interest. Many commercial and military organizations have begun
consortia and R&D programs to define mitigation techniques that might be able to stem
the tide of whisker failures. As of this writing, no conclusive set of mitigation techniques
is proven, and so the only accepted mitigation techniques for tin whiskers is to
electroplate component terminals with a solderable metal other than pure tin. Immersion
Silver, Electroless Gold, Nickel-Palladium, all can be theoretically substituted for tin, but
at significantly higher cost than electroplated tin, so that electroplated tin is by far the
most common RoHS-compliant replacement for the former SnPb electroplate.

Although the whisker prone, electroplated pure tin finish is commonly found in COTS
electronic hardware, it is showing up in military electronic assemblies at an alarming rate.
Because of this "inconvenient truth", as the tin whisker risk has been called, many
attempts at mitigation of these risks have been made. An example of a recent tin whisker
event, on the NASA Space Shuttle, is depicted in Figure 3.

Figure 3. Tin whiskers growing on NASA Space Shuttle avionics hardware and
documented in April 2006. (Photos courtesy of NASA, 2006)
It should be noted that these tin whiskers were not growing from tin plated electrical
components, since the Space Shuttle avionics manufacturing pre-dated RoHS by a
significant number of years, but from tin plated beryllium copper card guides, and may
have been growing for years. Today's lead free electronics would be expected to have far
more tin whisker susceptible components, since any tin plated (RoHS compliant)
electronic component could be a source of whiskers, as well as any tin plated mechanical
hardware.

Metallurgical Theory

Tin (Sn) metal displays the characteristic of growing "tin whiskers" from pure tin (most
actively on relatively thin, electrodeposited or immersion tin coatings), usually months or
years from the initial deposition of the tin. Tin whiskers are electrically conductive,
filamentary, single crystals of white (β) tin. These filaments of single crystal tin are
usually one to five microns in diameter, and a few microns up to several tens of
millimeters long that grow spontaneously from the tin coatings. Alloying additions of
several percent by weight of lead (Pb) essentially prevent these electrically conductive tin
whiskers from growing. Pb alloyed into the Sn was discovered to prevent the occurrence
of tin whiskers in electronic assemblies in the 1950s as the Bell Laboratories solution to
the problem of tin whiskers, which at the time were causing failures in AT&T telephone

switching equipment. The alloying of the tin with lead (Pb) has thus quietly averted
incalculable losses from short circuits in electronic equipment for the last 60 years.
Sixty years ago, the most commonly used surface finish for electronic components and
printed wiring boards was tin-lead (SnPb) solder. The composition of this solder was
most often "eutectic," or 63% tin (Sn) and 37% lead (Pb), both by weight. This
composition was ideal for the soldering process, while the Pb content suppressed
(virtually completely eliminated) tin whisker growth on the tin-lead (SnPb) coated
electronic hardware.
White Sn (β tin), the allotropic phase of tin constituting the tin whisker, has a body
centered tetragonal crystal structure. The other allotropic phase, grey tin (α tin), has a
diamond cubic structure to the crystal unit cell. The c/a ratio for the white tin (β Tin), is
less than one, meaning that the unit cell for the tetragonal white tin is longer on one side
than the other (rectangular in cross section). This configuration of a crystalline material
(i.e. not cubic) is usually an indication of anisotropic properties. For tin, the coefficient of
thermal expansion (CTE) and the self diffusion coefficient are higher in the "a" direction
(longer side of the unit cell) than in the "c" direction (shorter side of the unit cell). Other
metals that have such an anisotropic crystal structure (c/a ratio less than one), such as
Zinc and Cadmium with their hexagonal close packed (hcp) structures, also form
whiskers readily (figure 4). In fact, the first well-documented occurrence of whiskers was
on cadmium plated military hardware in 1948. Another common failure is for the zinc-
chromate coated steel supporting structures for raised floor laminar flow clean rooms to
grow spontaneous zinc whiskers over time, resulting in conductive particle contamination
of the laminar air flow in the clean room.

Figure 4. Similarity between the tin and zinc spontaneous whisker growths. Both are
conductive single crystals (From J. A. Brusse, NASA, 2003 reference 4)

Sn is known to have several active slip systems and direction along with an active
twinning system, (Table 1). Multiple growth directions for Sn whiskers have been found
during investigations of whisker growth.

Table 1. Slip Planes, Slip Directions Twin Planes and Whisker Growth Directions
for Sn (5).

Allotropic Slip Planes and Directions Twin Whisker
Phase Plane Growth
Directions
 – Tin (110)[001], (010)[100], {101}1/2[111], 331
(Body Centered {101}[101], {121}1/2[111],{121}[101] 110, 001, 103,
321
Tetragonal)

Investigations of high strain deformation of Sn in a "Deformation Processed Metal-Metal
Matrix Composite" (DMMC) shows Sn takes on a <001> texture. It’s assumed that this
texture is brought about by the Sn deforming in a plane strain condition (6). Several
authors investigating Sn whiskers have seen a similar texture in the Sn single crystals.
Many researchers have cited the <001> direction as a predominant growth direction for
tin whiskers.

Mechanical Stress Model

It is well documented that internal compressive stress is a driving force for the growth of
tin or zinc whiskers. This stress can come from the naturally occurring compressive stress
that occurs when tin is electrodeposited from a plating bath. The more organic
brightening additions in the plating bath, the higher the internal stress in the tin deposit
will be. The highest stressed deposits are obtained from baths that yield "bright" or
specular (mirror-like) deposits, often used for decorative electroplates. Somewhat less
prone to whiskers is the "matte’ or dull finished tin, which is usually lower stressed, but
still will grow (usually fewer and/or shorter) whiskers.

Other sources of internal stress in tin electroplate are intermetallic compounds formed
from the base metal onto which the tin is electroplated. For example, the tin is often
plated onto a copper component terminal lead or printed circuit trace. An intermetallic
compound, Sn6Cu5, or in some cases, Sn3Cu are readily formed, and cause internal
compressive stress in the tin layer as the intermetallic compound grows. This internal
stress can accelerate whisker growth. A mitigation method for this effect is to electro- or
electrolessly plate a nickel diffusion barrier between the tin and copper layers.

Hot-dipped or reflowed tin or zinc tends to grow fewer and shorter whiskers than
electrodeposited tin as well, once again so long as precautions are taken to avoid

compressive-stress-causing intermetallics (such as Cu6Sn5, common on the copper tin
plated copper lead frames used in electronics).
There are two areas of tin surface finish that pose a much more severe risk for the long
product life, high-reliability DoD (Department of Defense) applications than for the
relatively short product lifetime commercial electronics world. While product lifetimes
for military systems are routinely 30 to 40 years, there are few commercial products
expected to last as long as 10 years, with average product life expectancies nearer to 3
years, and commercial product warrantees 90 days or less.
The first area of concern is the tendency of tin whiskers to grow for extremely long
periods of time, long past the obsolescence (and replacement) timeframe of commercial
electronics. Figure 5 is an electron microscopic picture of a circuit board surface having
immersion tin plating (a common lead-free alternative finish) after 5 days of room
temperature storage at ACI Technologies. Note the small, 10 to 20 micron tin whiskers
that are present. These whiskers are too small to be a serious risk, even to fine pitch
circuitry. Figure 6 shows the same sample after two and a half years of storage (long time
for a commercial application, but a short time for a DoD application) showing some tin
whiskers that have grown to over 90 microns (long enough to electrically short closely
spaced modern electronic circuits). After 5 years, whiskers over 150 microns long (figure
7) were found on the same sample, implying continuous whisker growth over time.

Figure 5. Scanning Electron Micrograph of Tin Whiskers grown on immersion tin
coating after only 5 days from the tin plating of storage at ACI in October of 2003.

Figure 6. Scanning Electron Micrograph of same sample (different field, equivalent
area) stored at ambient at ACI, micrograph taken in 2006 (3 years after plating the
immersion tin).

Figure 7. Scanning Electron Micrograph of same sample (different field, equivalent
area) stored at ambient at ACI, micrograph taken in February 2009 (5.3 years after
plating the immersion tin)

In 10, 20, and 30 Years, How Long Will

These Whiskers Be?

This question represents the second of the two troublesome characteristics of tin
whiskers. Their growth rate, and methods that could be used to accelerate this rate, are
unpredictable. This makes the tin whisker phenomenon difficult to study. For example,
the only way to assure whiskers do not grow to excessive length in a given time is to
observe the whiskers for that length of time. For military applications, product lifetimes
of several decades, and similar (impractically long) whisker observation times, are
indicated. Note that some military product lifetimes as typified by the expected for the
F-35 Joint Strike Fighter could be as long as 40 years. For commercial applications, this
might be a practical span of several years, Many of the military and commercial
electronic shorting failures that have been caused by tin whiskers have occurred after
years or decades of use of the electronics in satellites, airplanes, one nuclear power plant,
and even the space shuttle, from the initial manufacture of the electronic hardware. These
tin whisker failures were impossible to predict when the electronics were manufactured.
If built today, the tin whisker failure prediction would be just as impossible as it was
then.

The literature contains abundant misinformation about tin whiskers. Many component
and/or PWB vendors have claimed their pure tin (RoHS compliant) coatings are immune
to whisker failures because of various "mitigation" techniques. Since none of these
vendors have inspected product for tin whiskers for 20, 30 or 40 years, any such claims
by vendors who have examined experimental product for less than two years are suspect.
In fact, many of the commonly held (and often cited by vendors) "mitigation techniques,"
such as nickel under plating of the tin, annealing of the tin plate, use of "matte" rather
than "bright" tin electroplate, have all been seen to allow tin whisker growth if observed
for a couple of years.

The universally accepted driving force (perhaps not the only driving force) for tin
whisker growth is internal compressive stress in the tin coating. Such stress might be due
to the initial electroplating conditions used in the deposition of the tin (such as current
density, tin concentration in the electrolyte, organic "brighteners" used in the
electroplating process, mechanical scratches, or bending of the tin plated article, or even
the growing of intermetallic compounds such as copper-tin (Cu6Sn5) within the tin
coating. This intermetallic generation of internal stress is cited as potentially creating
compressive stress over very long times after the plating as copper in the base metal
substrate slowly diffuses into the tin coating.

It is commonly accepted that whisker growth follows a two part mechanism:

1. Diffusion of tin from anywhere in the tin plate to the site of the whisker, and
2. Incorporation of the tin atoms into the crystal lattice of the tin whisker.

Diffusion is encouraged by 1) temperature high or low enough to be a significant fraction
of the melting temperature at the high end, and absolute zero at the low end, and 2) a
stress gradient whereby the tin atom diffuses from a place of high free energy
(compressively stressed tin plated surface finish) to low free energy (a perfect crystal of β
tin) within the tin whisker.

Neither of the diffusion enhancing/impeding parameters can be readily mitigated. First,
temperatures at which electronics operate are rarely high or low enough to affect the
diffusion coefficient of tin (melting point 231 °C) appreciably.

Second, even though the internal compressive stress in the plated tin can be eliminated or
caused to become tensile at the initial deposition of the tin coating film, the long term
generation of intermetallic compounds in the tin coating on any of the electronic
component lead frame alloys can eventually, often over very extended times, re-generate
internal compressive stress in an initially stress-free tin finish coating. So, diffusion is not
something that can be controlled on a practical, long term basis.

The second half of the mechanism for whisker growth—incorporation of the tin atom into
the whisker crystal -- could possibly be prevented by coating the whisker, or the re-
crystallized grain of the tin plate that provides the first few layers of perfect crystal in the
whisker, with a conformal coating of a high elastic modulus material. If the elastic
modulus of the conformal coat is higher than the modulus of tin, then the incorporation of
the additional tin atoms may not be possible, because the elastic modulus of pure tin
would have to be exceeded to add additional tin atoms (and therefore whisker crystal
volume, to the whisker) thus "stretching the coating. Of course, the elastic modulus of the
standard conformal coating materials (acrylics, silicones, epoxies, Parylene, and
urethanes are all thousands of times too low, compared to the modulus of tin, to do this.

There is also a third approach possible to mitigate whisker risk, also the subject of an
MDA Phase II SBIR, and this would be not to attack the growth mechanism of the
whisker at all, but conformally coat the electronics with a tough viscoelastic conformal
coating that tents the growing tin whisker and causes Euler buckling of the whisker rather
than allowing the growing whisker to puncture the coating. Tin whiskers have been
shown to puncture and grow through all standard conformal coatings in a matter of
months or a few years.

The coating that is documented to exhibit higher elastic modulus than tin metal is called
ALD Cap (applied to the completed CCA (Circuit Card Assembly) as a conformal
coating by Atomic Layer Deposition). The ALD Cap material is a ceramic material with
a much higher elastic modulus than tin metal.

The viscoelastic conformal coating shown to cause Euler buckling of growing whiskers is
called Whisker Tough P1.

Several projects are underway and/or proposed by the EMPF and partner organizations to
identify valid long term mitigation and/or complete prevention of tin whisker growth on

RoHS compatible (Pb-free) electronic hardware, making use of COTS (Commercial Off
the Shelf) electronic components by the military a less risky proposition than it is now.

References

1) G. T. Galyon, "Annotated Tin Whisker Bibliography," IBM Server Group Chair,
NEMI Tin Whisker Modeling Project, July, 2003.

2) M. E. McDowell, "Tin Whiskers: A Case Study", Aerospace App. Conf., pp. 207-215,
1993.

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

4) J. A. Brusse, "A Discussion of the Significance of Metal Whisker Formation to the
High Reliability Community," NASA, November 2003.

5) Metals Handbook 9th Edition, ASM, Metals Park, OH.

6) Chen Xu, Chonglun Fan, Yun Zhang, and Joseph A. Abys, "Whisker Prevention"
Proceedings of the Technical Conference APEX, Jan. 19, 2002.

7) Yuki Fukuda, Tong Fang, Michael Pecht, and Michael Ostermand, "Effect of Heat
Treatment on Tin Whisker Growth" CALCE Electronic Products and Systems Center,
University of Maryland, College Park MD, 2004.

8) Dr. Henning Leidecker and Jay Brusse, "Tin Whiskers: A History of Documented
Electrical System Failures," NASA, April 2006.