Electronics Guide

What Counts as a Whisker

JEDEC standardizes the definition for test purposes: a whisker is a spontaneous columnar or cylindrical filament emanating from the surface, with a length of at least 10 micrometers and an aspect ratio (length divided by width) greater than two. Features shorter than 10 micrometers, or squat protrusions that fail the aspect-ratio test, are reported as nodules or hillocks rather than whiskers.

Whisker morphology varies widely. Common forms include:

• Straight filaments: The classic form, needle-like and often striated along their length, with the striations marking the incremental extrusion process.
• Kinked and bent whiskers: Filaments that change direction abruptly, reflecting changes in the stress state at the root during growth.
• Curved or hooked whiskers: Continuously curving filaments, sometimes closing on themselves.
• Nodules and hillocks: Squat eruptions that relieve stress without producing a long filament. They are not themselves a bridging hazard, but their presence indicates the same driving stress that produces whiskers.

Typical diameters are a few micrometers, with reported extremes above ten micrometers and, rarely, below one hundred nanometers. The NASA Goddard Space Flight Center whisker program, which maintains the most widely cited public photographic archive of whisker specimens, documents growth rates spanning roughly 0.03 to 9 millimeters per year, with enormous variability depending on plating chemistry, deposit thickness, substrate material, and storage conditions.

Distinguishing Whiskers from Dendrites

Whiskers are frequently confused with dendrites produced by electrochemical migration, and the distinction is the first thing a failure analyst must establish, because the two mechanisms have nothing in common beyond the resulting short.

• Driving force: A whisker grows from internal compressive stress in the deposit. A dendrite grows from an applied electric field acting on dissolved metal ions.
• Environment: Whiskers grow in dry air, in vacuum, and in sealed packages. Dendrites require an electrolyte, which in practice means condensed moisture plus ionic contamination.
• Direction: A whisker grows outward from its root in whatever direction the stress relief happens to favor, indifferent to the location of nearby conductors. A dendrite grows from cathode toward anode, tracking the field.
• Structure: A whisker is typically a single crystal of near-uniform cross-section. A dendrite is branched and tree-like.

The practical consequence is that dendrite failures are addressed by cleanliness and moisture control, whereas whisker failures are not. Cleaning a board more thoroughly does nothing for whiskers. Confusing the two mechanisms leads directly to a corrective action that cannot work. Related phenomena include zinc whiskers, which grow from electroplated zinc on raised-floor tiles and equipment enclosures and have caused repeated data center outages, and cadmium whiskers, which drove the original mid-century investigations.

Sources of Compressive Stress

• Intermetallic growth: The dominant source for tin plated directly over copper. Cu₆Sn₅ nucleates irregularly along tin grain boundaries at room temperature and occupies more volume than the copper and tin it consumes, wedging the surrounding tin grains into compression. This process needs no elevated temperature and continues throughout storage.
• Residual plating stress: Electrodeposits are rarely stress-free as plated. Bright tin, produced with organic brightening agents that refine the grain and codeposit organic material at grain boundaries, is the worst common case.
• Thermal expansion mismatch: Tin has a much larger coefficient of thermal expansion than the alloy 42 or copper alloy substrates it is plated on, so temperature excursions cycle the deposit through compressive states.
• Applied mechanical stress: Lead forming, connector insertion, clamping, board flexure, and probe marks all impose local compression. Whiskers on formed lead shoulders and on connector contact wipe areas are a recurring field observation.
• Oxidation and corrosion: Tin oxide and corrosion products formed at the surface occupy greater volume than the tin consumed, adding compression. This is the pathway through which humidity and atmospheric contaminants accelerate whisker growth even though moisture is not required for whiskers to appear.

Incubation and Statistical Behavior

Whiskers do not appear immediately. An incubation period, during which the compressive stress accumulates and the surface oxide is locally breached, precedes visible growth. Reported incubation ranges from days to years and depends on the same variables that govern growth rate.

This has an uncomfortable consequence for qualification: a finish that shows no whiskers on delivery, or after a few months of storage, has demonstrated very little about its behavior after a decade in the field. It also means that whisker behavior is statistical rather than deterministic. Whiskers nucleate on a small minority of grains, so nominally identical parts from the same lot can differ by an order of magnitude in worst-case length. Testing therefore reports the longest whisker found across a defined inspection area on a defined sample size, not an average, and the analyst must treat any single specimen as a lower bound on what the population can produce.

What Suppresses Growth

Lead is the one alloying addition with a long and unambiguous record of suppressing whiskers, which is precisely why the regulatory removal of lead reintroduced a risk the industry had largely forgotten. A few percent of lead by weight changes the grain structure and stress-relief behavior of the deposit enough to prevent whisker nucleation in the great majority of cases. Results for substitute alloying elements such as bismuth and silver are mixed and finish-specific: they reduce risk in some deposit and substrate combinations and not in others, so alloy claims warrant test data rather than assumption. Some of that data is encouraging. JEITA storage testing at 55 °C and 85 percent relative humidity held the longest whisker on parts finished with tin containing 2 percent bismuth below 50 micrometers, where pure tin controls exceeded 100 micrometers, and tin-silver finishes have performed well in the comparatively few studies published. The samples remain narrow, which is why the aerospace standards list these alloys as preferences to be demonstrated rather than mitigations to be credited.

Bridging Shorts and Intermittents

The primary mode is a filament bridging two conductors at different potentials. Because a whisker is only a few micrometers in diameter, its current-carrying capacity is small. NASA Goddard's experimental work indicates that more than 50 milliamperes can be required to fuse a whisker open, and GEIA-STD-0005-2 permits an analysis to assume fusing above 50 milliamperes, treating the event as an intermittent short of roughly 50 microseconds.

Two consequences follow. In circuits that can deliver well above the fusing current, whisker shorts are self-clearing and present as intermittent, non-reproducible faults, which are notoriously difficult to diagnose and are frequently written off as "no fault found." In low-voltage, high-impedance circuits, which include most sensor and signal-conditioning nodes, the available current never reaches the fusing threshold, and the short is stable and permanent.

Whiskers can also produce resistive rather than hard shorts. In the NASA Engineering and Safety Center's 2011 technical support work for the National Highway Traffic Safety Administration, destructive physical analysis of a failed accelerator pedal position sensor from a 2003 Toyota Camry found a tin whisker forming a 248-ohm resistive short between the two redundant sensor outputs, and whiskers were found in other assemblies of the same type that had not malfunctioned. The investigation concluded that whiskers were not shown to produce the large-throttle unintended acceleration events under study, but the case remains a well-documented example of whiskers reaching an automotive sensor in a consumer vehicle and altering a measured value rather than simply shorting a rail.

Metal Vapor Arc

The most severe mode occurs at low pressure. When a whisker fuses in a rarefied atmosphere, the vaporized tin forms a conductive plasma that can sustain an arc drawing current far beyond anything the solid filament could carry. NASA Goddard has demonstrated sustained arcs at pressures near 150 torr with source conditions of about 13 volts and 15 amperes or greater, and reports that such arcs can carry hundreds of amperes where the source can supply them. The arc does not stop when the whisker is consumed; it persists as long as the circuit sustains it, and the resulting damage is catastrophic rather than intermittent.

This mechanism is why whisker risk is treated as a hard design constraint in spacecraft, launch vehicles, and high-altitude equipment, where hardware spends its life at or transits through the pressure range in which metal vapor arcs are most readily sustained.

Foreign Object Debris and Other Effects

• Detached whiskers: A whisker broken free by handling or vibration becomes conductive foreign object debris that can lodge anywhere in an enclosure, including in a location unrelated to its origin. GEIA-STD-0005-2 directs that applications with optical or electromechanical requirements, and applications subject to extensive handling or repeated high-vibration exposure, be evaluated specifically for this risk.
• Mechanical interference: In relays, switches, micromachined devices, and optical assemblies, a whisker can obstruct motion or occlude an aperture without ever forming an electrical connection.
• High-frequency effects: At microwave frequencies a whisker can act as an unintended antenna or coupling structure, degrading isolation or introducing intermodulation, without producing a direct-current short.
• Non-bridging high-voltage failures: In high-voltage circuits a whisker distorts the local field and can initiate breakdown across a gap it does not physically span. GEIA-STD-0005-2 requires that circuit analyses for high-voltage applications address these non-bridging modes explicitly.

Product Classes

JESD201 sorts applications by service life and failure consequence rather than by industry:

• Class 1A: Consumer products with short product lifetimes, where whiskers are of minimal concern.
• Class 1: Industrial and consumer products with medium lifetimes.
• Class 2: Business-critical applications with long lifetimes and low tolerance for downtime, such as telecommunications infrastructure, high-end servers, and automotive electronics.
• Class 3: Mission-critical and life-critical applications, including military, aerospace, and medical. For this class the standard's position is that pure tin and high-tin finishes are not typically acceptable, which is why Class 3 is better read as a statement about finish selection than as a test one can pass.

Class 2 is the level at which most commercial component whisker qualification is performed and reported, and it is the level that GEIA-STD-0005-2 names when it lists successful JESD201 testing among the characteristics of lower-risk parts. Vendor whisker reports published for capacitors, inductors, and integrated circuit packages almost universally report against Class 2.

Exposure Conditions and Class 2 Criteria

The three exposures are complementary rather than redundant. Ambient storage represents the baseline room-temperature intermetallic-driven growth that proceeds on a shelf. Elevated temperature and humidity accelerates the oxidation and corrosion contribution. Temperature cycling exercises the thermal expansion mismatch between deposit and substrate. A finish that passes one exposure and fails another is common, and it tells the analyst which stress source dominates.

The elevated-humidity condition is one place where the revision matters. Revision A specifies 55 °C and 85 percent relative humidity, where the original release used a hotter and wetter condition, and published vendor reports sometimes run their own variants on top of that. Two whisker reports are comparable only when both name the same revision and the same exposure.

What Passing Does and Does Not Establish

JESD201 acceptance is a comparative statement about a finish under defined exposures, not a guarantee. Four thousand hours is under six months; service lives in the applications that care most about whiskers are measured in decades. Because incubation can exceed the test duration, and because whisker populations are dominated by rare outlier grains, a Class 2 pass reduces expected risk without bounding it. The standard itself states that its methodology may be insufficient for military, aerospace, and similar applications, and that additional requirements belong in the procurement documentation. That caveat is the hinge between the JEDEC framework and the aerospace framework described next.

Control Levels

The standard's central device is a set of control levels that a customer specifies and a supplier documents compliance against. Each level represents a different balance among three basic strategies: risk acceptance, risk mitigation, and tin avoidance.

• Level 1: Whisker risk is accepted. Intended for development models, test equipment, and other hardware that will not be fielded.
• Level 2A: Risk is managed primarily by acceptance, with limited use of design rules. Intended for lower-criticality applications; tin is permitted except where specifically restricted. If a control document says only "Level 2," 2A is the default.
• Level 2B: Risk is managed primarily by design rules, with tin avoidance secondary. Intended for non-critical boards and units, or well-redundant units, in systems with moderate to high failure consequences.
• Level 2C: Risk is managed primarily by tin avoidance, with design rules used in exceptional cases. Intended for critical boards with limited redundancy in systems with moderate to high failure consequences.
• Level 3: Control is by tin avoidance alone. Intended for units and boards where failure cannot be tolerated.

Levels are applied at the unit or board level, so a single system may carry several. The standard's informative annexes provide guidance on selecting levels, on the technical basis for each mitigation method, on inspection practice, on risk analysis, and on whiskers growing from bulk solder and solder joint fillets, a mechanism distinct from whiskers on plated finishes.

Accepted Mitigations

Mitigation requirements attach to Levels 2B and 2C, and the standard enumerates exactly what qualifies. The approved categories are hard potting or encapsulation; physical barriers; a tin-lead soldering process with validated complete coverage; conformal coating with validated coverage above a minimum gap; circuit and design analysis showing either that a bridging whisker would not affect the unit or that the sensitive areas carry a gap of at least one centimeter; lead-free tin parts with large gaps that have been installed with tin-lead solder and physically isolated from other lead-free tin surfaces; and any mitigation, or combination of mitigations, the customer approves. Level 3 needs no such list, because it admits no tin.

Two structural points in the standard deserve emphasis, because they are frequently misunderstood.

First, part-level practices do not by themselves satisfy the mitigation requirement. The standard is explicit that whiskers have been reported with virtually every part-level practice the industry has proposed, that part manufacturing and finishing lie outside the control of the aerospace supply chain, and that verifying those processes is difficult. A nickel underlayer or a post-plate anneal is a preference to be encouraged through the parts selection process, not a mitigation to be credited.

Second, the circuit and design analysis route focuses on effects, not probability. The analysis must show that a whisker bridging a gap would not compromise the unit; it is not required, and not intended, to estimate the likelihood that a whisker will grow or land in a particular place. This aligns the work with an existing failure modes and effects analysis or FMECA, and it sidesteps a probability estimate that the physics does not currently support.

Parts Selection Preferences

The standard lists the finish and process characteristics that aerospace customers generally prefer, while cautioning that none is risk-free. They include annealing or fusing at the part manufacturer close to the time of plating, hot-dipped rather than plated tin finishes, immersion tin for boards, low-profile parts and short-lead geometries that reduce the chance of a whisker bridging, a nickel underplate (provided it is not beneath bright tin), tin-silver finishes with roughly 1.5 to 4 percent silver, tin-bismuth with 2 to 4 percent bismuth, tin-lead with 1 to 3 percent lead, and successful passage of JESD201 testing at the Class 2 level.

Related Requirements Elsewhere

European space practice reaches the same conclusion through a blunter rule. ECSS-Q-ST-70-38C, the ECSS standard for high-reliability soldering of surface-mount and mixed technology, states that a tin finish with more than 97 percent tin purity shall not be used, which in effect requires at least 3 percent of an alloying element. The standard gives two reasons in its notes: whisker growth, and the low-temperature transformation of tin to grey tin powder. The requirement carries through unchanged into revision 1, issued in September 2017, and the standard points to dipping pure tin terminations in liquid solder to replace the tin with a tin-lead alloy. Programs that cannot avoid pure tin parts therefore specify refinishing, by hot solder dip or by replating, with the refinishing process itself qualified, because an incomplete strip can leave a whisker-prone layer beneath a compliant one.

Plating and Part Level

• Alloying with lead: The most effective measure and the least available, since lead-bearing finishes are restricted for most commercial parts. Where an exempt supply exists, a few percent of lead by weight remains the benchmark against which other measures are judged.
• Nickel underlayer: A nickel barrier between the copper substrate and the tin suppresses the copper-tin intermetallic that wedges into tin grain boundaries, attacking the dominant room-temperature stress source directly. Results are nonetheless mixed. Several studies report fewer and shorter whiskers on tin over nickel than on tin over copper; others report whiskers growing despite the barrier, with the difference apparently depending on the base material and on how the tin was deposited. GEIA-STD-0005-2 cites a recommended minimum nickel thickness of 0.5 micrometers for parts that will see lead-free reflow, and lists nickel underplate among its parts-selection preferences only where it does not sit beneath bright tin, whose intrinsic plating stress owes nothing to the substrate.
• Matte rather than bright tin: Matte deposits, plated without organic brighteners, have coarser grains and lower internal stress; iNEMI and JEDEC define matte tin by a carbon content of 0.005 to 0.05 percent and a grain size of 1 to 5 micrometers. Bright tin is the higher-risk of the two, but the benefit is easy to overstate. GEIA-STD-0005-2 declines to credit matte over bright tin as a lower-risk finish without supporting data specific to the part and process in question, on the grounds that the two labels are vendor-defined, seldom verified against the iNEMI definition, and inconsistently supported by test results. One study it cites found whiskers up to 2 millimeters long on matte tin plated over steel.
• Post-plate annealing: A bake soon after plating relieves stress and stabilizes the grain structure. The iNEMI recommendation is one hour at 150 °C within 24 hours of plating, accompanied by data showing that the anneal worked on that part and process. Timing matters, and so does the limit of the technique: the available evidence indicates that heat treatments lengthen the incubation period and delay onset rather than prevent whiskers permanently. GEIA-STD-0005-2 accordingly treats annealing as a parts-selection preference rather than a creditable mitigation, and warns that it can create false confidence over aerospace service lives.
• Reflow or hot dip: Melting and resolidifying the deposit removes the as-plated stress state entirely and produces a fundamentally different microstructure. Hot-dipped tin is generally preferred over plated tin in aerospace parts selection for this reason.
• Deposit thickness: Very thin deposits offer less material to feed a whisker but plate less uniformly, and very thick deposits accommodate more stress. Thickness alone is not a reliable control, and thickness recommendations vary by finish and substrate.

Assembly Level

• Tin-lead soldering with validated coverage: Soldering a pure tin lead with a tin-lead alloy dissolves and alloys the finish in the wetted region. This is a recognized mitigation where coverage is validated, and its weakness is precisely coverage: the shoulders and upper surfaces of a formed lead, and the interior surfaces of press-fit and separable contacts, are not wetted.
• Hot solder dip or replating: Refinishing components before assembly replaces the pure tin finish outright. It introduces its own risks, including thermal stress on the part, damage to package bodies, and the possibility of an incompletely stripped underlying layer.
• Conformal coating: Coating constrains growth and raises the voltage required to bridge, but whiskers penetrate thin regions and coating never reaches every surface on a populated board. GEIA-STD-0005-2 credits conformal coating only with validated coverage and only above a minimum gap, and it sets that gap by coating type and control level: parylene qualifies above roughly 150 micrometers at Level 2B and 250 micrometers at Level 2C, while other coatings require roughly 250 and 500 micrometers respectively. The logic behind the gap is mechanical rather than dielectric. The space must accommodate the coating and still leave enough free air that an advancing whisker buckles under its own compressive load before it spans the distance. Coating is risk reduction, not elimination.
• Potting and encapsulation: Hard potting immobilizes the surface completely and is credited as a standalone mitigation, at the cost of mass, thermal behavior, and the loss of any possibility of repair.
• Physical barriers: Insulating sleeves, spacers, and barrier films placed between whisker-prone surfaces and the conductors at risk.

Design Level

• Spacing: Increasing the gap between tin-plated surfaces and adjacent conductors is the most direct design control, because whisker length distributions fall off steeply. It is also the hardest to obtain on modern fine-pitch hardware, where lead-to-lead gaps are often smaller than the lengths whiskers routinely reach.
• Circuit tolerance: Designing nodes so that a momentary or resistive short is detected and tolerated rather than propagated. Current-limited supplies, fault detection, and voting redundancy all reduce the consequence of a bridging event.
• Orientation and geometry: Low-profile parts and short leads reduce the population of surfaces from which a whisker can reach a neighbor.
• Environment control: Reducing thermal cycling amplitude and humidity exposure lowers two of the stress sources, though neither eliminates the room-temperature intermetallic pathway.

Preserving the Evidence

Whiskers are destroyed by wiping, by ultrasonic cleaning, by conformal coating removal, and often by simply separating a connector. When a whisker short is suspected, the sequence must begin with electrical characterization and nondestructive imaging of the intact assembly, and any disassembly must be planned to expose the suspected region without disturbing it. Photographic documentation before every step is standard practice, because a whisker that is present in one image and absent in the next at least establishes that it existed.

Imaging

• Optical microscopy: Useful for survey work and for locating candidate sites, particularly with oblique or darkfield illumination, which makes filaments glint against the background. Resolution is inadequate for measuring short whiskers.
• Scanning electron microscopy: The primary tool. Whisker inspection per JESD22-A121 is performed at magnifications on the order of 3,000×. Low accelerating voltage limits charging on oxidized and coated surfaces. Because a whisker projects out of the plane, tilted imaging and stereo pairs are needed to recover true length rather than a foreshortened projection.
• Cross-sectioning: Sectioning through a whisker root reveals the deposit thickness, the intermetallic layer, and the grain structure that explain why the whisker grew, converting an observation into a root cause.

Composition and Finish Verification

A large fraction of whisker investigations turn on a simple question: what is the finish actually made of? Suppliers change plating chemistry without notice, and part markings do not record it.

• X-ray fluorescence: The standard nondestructive screen for finish composition and for detecting pure tin in incoming inspection. GEIA-STD-0005-2 directs that XRF measurement be consistent with JESD213, which addresses the accuracy and repeatability required when measuring lead content near the reporting thresholds.
• Energy-dispersive X-ray spectroscopy: Performed in the SEM, EDS confirms that a filament is tin rather than a fiber, a contaminant, or a dendrite, and identifies alloying elements in the deposit.
• Electron backscatter diffraction: Maps grain orientation and size in the deposit, which is how the relationship between grain structure and whisker nucleation is studied.

Supply Chain Detection

For programs operating at Level 2B and above, detection is a continuing obligation rather than a one-time qualification. The standard requires plans for monitoring materials on a sampling basis at Levels 2B and 2C, and lot-by-lot monitoring at Level 3, with requirements flowed down to lower-tier suppliers. In practice this means XRF screening of incoming parts, documented finish declarations, and change control on plating processes, because a part that was tin-lead at qualification can silently become pure tin at the next die-shrink or site transfer.