Electronic components have a wide range of failure modes. These can be classified in various ways, such as by time or cause. Failures can be caused by excess temperature, excess current or voltage, ionizing radiation, mechanical shock, stress or impact, and many other causes. In semiconductor devices, problems in the device package may cause failures due to contamination, mechanical stress of the device, or open or short circuits.
Failures most commonly occur near the beginning and near the ending of the lifetime of the parts, resulting in the bathtub curve graph of failure rates. Burn-in procedures are used to detect early failures. In semiconductor devices, parasitic structures, irrelevant for normal operation, become important in the context of failures; they can be both a source and protection against failure.
Applications such as aerospace systems, life support systems, telecommunications, railway signals, and computers use great numbers of individual electronic components. Analysis of the statistical properties of failures can give guidance in designs to establish a given level of reliability. For example, power-handling ability of a resistor may be greatly derated when applied in high-altitude aircraft to obtain adequate service life.A sudden fail-open fault can cause multiple secondary failures if it is fast and the circuit contains an inductance; this causes large voltage spikes, which may exceed 500 volts. A broken metallisation on a chip may thus cause secondary overvoltage damage.[1] Thermal runaway can cause sudden failures including melting, fire or explosions.
The majority of electronic parts failures are packaging-related. Packaging, as the barrier between electronic parts and the environment, is very susceptible to environmental factors. Thermal expansion produces mechanical stresses that may cause material fatigue, especially when the thermal expansion coefficients of the materials are different. Humidity and aggressive chemicals can cause corrosion of the packaging materials and leads, potentially breaking them and damaging the inside parts, leading to electrical failure. Exceeding the allowed environmental temperature range can cause overstressing of wire bonds, thus tearing the connections loose, cracking the semiconductor dies, or causing packaging cracks. Humidity and subsequent high temperature heating may also cause cracking, as may mechanical damage or shock.
During encapsulation, bonding wires can be severed, shorted, or touch the chip die, usually at the edge. Dies can crack due to mechanical overstress or thermal shock; defects introduced during processing, like scribing, can develop into fractures. Lead frames may contain excessive material or burrs, causing shorts. Ionic contaminants like alkali metals and halogens can migrate from the packaging materials to the semiconductor dies, causing corrosion or parameter deterioration. Glass-metal seals commonly fail by forming radial cracks that originate at the pin-glass interface and permeate outwards; other causes include a weak oxide layer on the interface and poor formation of a glass meniscus around the pin.
Various gases may be present in the package cavity, either as impurities trapped during manufacturing, outgassing of the materials used, or chemical reactions, as is when the packaging material gets overheated (the products are often ionic and facilitate corrosion with delayed failure). To detect this, helium is often in the inert atmosphere inside the packaging as a tracer gas to detect leaks during testing. Carbon dioxide and hydrogen may form from organic materials, moisture is outgassed by polymers and amine-cured epoxies outgas ammonia. Formation of cracks and intermetallic growth in die attachments may lead to formation of voids and delamination, impairing heat transfer from the chip die to the substrate and heatsink and causing a thermal failure. As some semiconductors like silicon and gallium arsenide are infrared-transparent, infrared microscopy can check the integrity of die bonding and under-die structures.
Red phosphorus, used as a charring-promoter flame retardant, facilitates silver migration when present in packaging. It is normally coated with aluminium hydroxide; if the coating is incomplete, the phosphorus particles oxidize to the highly hygroscopic phosphorus pentoxide, which reacts with moisture to phosphoric acid. This is a corrosive electrolyte that in the presence of electric fields facilitates dissolution and migration of silver, short-circuiting adjacent packaging pins, lead frame leads, tie bars, chip mount structures, and chip pads. The silver bridge may be interrupted by thermal expansion of the package; thus, disappearance of the shorting when the chip is heated and its reappearance after cooling is an indication of this problem. Delamination and thermal expansion may move the chip die relative to the packaging, deforming and possibly shorting or cracking the bonding wires.[1]
Electrical contacts exhibit ubiquitous contact resistance, the magnitude of which is governed by surface structure and the composition of surface layers.[2] Ideally contact resistance should be low and stable, however weak contact pressure, mechanical vibration, corrosion, and the formation of passivizing oxide layers and contacts can alter contact resistance significantly, leading to resistance heating and circuit failure.
Soldered joints can fail in many ways like electromigration and formation of brittle intermetallic layers. Some failures show only at extreme joint temperatures, hindering troubleshooting. Thermal expansion mismatch between the printed circuit board material and its packaging strains the part-to-board bonds; while leaded parts can absorb the strain by bending, leadless parts rely on the solder to absorb stresses. Thermal cycling may lead to fatigue cracking of the solder joints, especially with elastic solders; various approaches are used to mitigate such incidents. Loose particles, like bonding wire and weld flash, can form in the device cavity and migrate inside the packaging, causing often intermittent and shock-sensitive shorts. Corrosion may cause buildup of oxides and other nonconductive products on the contact surfaces. When closed, these then show unacceptably high resistance; they may also migrate and cause shorts. Tin whiskers can form on tin-coated metals like the internal side of the packagings; loose whiskers then can cause intermittent short circuits inside the packaging. Cables, in addition to the methods described above, may fail by fraying and fire damage.
Printed circuit boards (PCBs) are vulnerable to environmental influences; for example, the traces are corrosion-prone and may be improperly etched leaving partial shorts, while the vias may be insufficiently plated through or filled with solder. The traces may crack under mechanical loads, often resulting in unreliable PCB operation. Residues of solder flux may facilitate corrosion; those of other materials on PCBs can cause electrical leaks. Polar covalent compounds can attract moisture like antistatic agents, forming a thin layer of conductive moisture between the traces; ionic compounds like chlorides tend to facilitate corrosion. Alkali metal ions may migrate through plastic packaging and influence the functioning of semiconductors. Chlorinated hydrocarbon residues may hydrolyze and release corrosive chlorides; these are problems that occur after years. Polar molecules may dissipate high-frequency energy, causing parasitic dielectric losses.
Above the glass transition temperature of PCBs, the resin matrix softens and becomes susceptible contaminant diffusion. For example, polyglycols from the solder flux can enter the board and increase its humidity intake, with corresponding deterioration of dielectric and corrosion properties.[3] Multi-layer substrates using ceramics suffer from many of the same problems.
Conductive anodic filaments (CAFs) may grow within the boards along the fibers of the composite material. Metal is introduced to a vulnerable surface typically from plating the vias, then migrates in presence of ions, moisture, and electrical potential; drilling damage and poor glass-resin bonding promotes such failures. The formation of CAFs usually begins by poor glass-resin bonding; a layer of adsorbed moisture then provides a channel through which ions and corrosion products migrate. In presence of chloride ions, the precipitated material is atacamite; its semiconductive properties lead to increased current leakage, deteriorated dielectric strength, and short circuits between traces. Absorbed glycols from flux residues aggravate the problem. The difference in thermal expansion of the fibers and the matrix weakens the bond when the board is soldered; the lead-free solders which require higher soldering temperatures increase the occurrence of CAFs. Besides this, CAFs depend on absorbed humidity; below a certain threshold, they do not occur.[3] Delamination may occur to separate the board layers, cracking the vias and conductors to introduce pathways for corrosive contaminants and migration of conductive species.
Every time the contacts of an electromechanical relay or contactor are opened or closed, there is a certain amount of contact wear. An electric arc occurs between the contact points (electrodes) both during the transition from closed to open (break) or from open to closed (make). The arc caused during the contact break (break arc) is akin to arc welding, as the break arc is typically more energetic and more destructive.[4]
The heat and current of the electrical arc across the contacts creates specific cone & crater formations from metal migration. In addition to the physical contact damage, there appears also a coating of carbon and other matter. This degradation drastically limits the overall operating life of a relay or contactor to a range of perhaps 100,000 operations, a level representing 1% or less than the mechanical life expectancy of the same device.[5]
See also: Reliability (semiconductor).
Many failures result in generation of hot electrons. These are observable under an optical microscope, as they generate near-infrared photons detectable by a CCD camera. Latchups can be observed this way. If visible, the location of failure may present clues to the nature of the overstress. Liquid crystal coatings can be used for localization of faults: cholesteric liquid crystals are thermochromic and are used for visualisation of locations of heat production on the chips, while nematic liquid crystals respond to voltage and are used for visualising current leaks through oxide defects and of charge states on the chip surface (particularly logical states). Laser marking of plastic-encapsulated packages may damage the chip if glass spheres in the packaging line up and direct the laser to the chip.
Examples of semiconductor failures relating to semiconductor crystals include:
Vias are a common source of unwanted serial resistance on chips; defective vias show unacceptably high resistance and therefore increase propagation delays. As their resistivity drops with increasing temperature, degradation of the maximum operating frequency of the chip the other way is an indicator of such a fault. Mousebites are regions where metallization has a decreased width; such defects usually do not show during electrical testing but present a major reliability risk. Increased current density in the mousebite can aggravate electromigration problems; a large degree of voiding is needed to create a temperature-sensitive propagation delay.[7]
Sometimes, circuit tolerances can make erratic behaviour difficult to trace; for example, a weak driver transistor, a higher series resistance and the capacitance of the gate of the subsequent transistor may be within tolerance but can significantly increase signal propagation delay. These can manifest only at specific environmental conditions, high clock speeds, low power supply voltages, and sometimes specific circuit signal states; significant variations can occur on a single die.[7] Overstress-induced damage like ohmic shunts or a reduced transistor output current can increase such delays, leading to erratic behavior. As propagation delays depend heavily on supply voltage, tolerance-bound fluctuations of the latter can trigger such behavior.
Gallium arsenide monolithic microwave integrated circuits can have these failures:[8]
Metallisation failures are more common and serious causes of FET transistor degradation than material processes; amorphous materials have no grain boundaries, hindering interdiffusion and corrosion.[10] Examples of such failures include:
Most stress-related semiconductor failures are electrothermal in nature microscopically; locally increased temperatures can lead to immediate failure by melting or vaporising metallisation layers, melting the semiconductor or by changing structures. Diffusion and electromigration tend to be accelerated by high temperatures, shortening the lifetime of the device; damage to junctions not leading to immediate failure may manifest as altered current–voltage characteristics of the junctions. Electrical overstress failures can be classified as thermally-induced, electromigration-related and electric field-related failures; examples of such failures include:
See main article: Electrostatic discharge. Electrostatic discharge (ESD) is a subclass of electrical overstress and may cause immediate device failure, permanent parameter shifts and latent damage causing increased degradation rate. It has at least one of three components, localized heat generation, high current density and high electric field gradient; prolonged presence of currents of several amperes transfer energy to the device structure to cause damage. ESD in real circuits causes a damped wave with rapidly alternating polarity, the junctions stressed in the same manner; it has four basic mechanisms:[11]
Catastrophic ESD failure modes include:
A parametric failure only shifts the device parameters and may manifest in stress testing; sometimes, the degree of damage can lower over time. Latent ESD failure modes occur in a delayed fashion and include:
Catastrophic failures require the highest discharge voltages, are the easiest to test for and are rarest to occur. Parametric failures occur at intermediate discharge voltages and occur more often, with latent failures the most common. For each parametric failure, there are 4–10 latent ones.[12] Modern VLSI circuits are more ESD-sensitive, with smaller features, lower capacitance and higher voltage-to-charge ratio. Silicon deposition of the conductive layers makes them more conductive, reducing the ballast resistance that has a protective role.
The gate oxide of some MOSFETs can be damaged by 50 volts of potential, the gate isolated from the junction and potential accumulating on it causing extreme stress on the thin dielectric layer; stressed oxide can shatter and fail immediately. The gate oxide itself does not fail immediately but can be accelerated by stress induced leakage current, the oxide damage leading to a delayed failure after prolonged operation hours; on-chip capacitors using oxide or nitride dielectrics are also vulnerable. Smaller structures are more vulnerable because of their lower capacitance, meaning the same amount of charge carriers charges the capacitor to a higher voltage. All thin layers of dielectrics are vulnerable; hence, chips made by processes employing thicker oxide layers are less vulnerable.[13]
Current-induced failures are more common in bipolar junction devices, where Schottky and PN junctions are predominant. The high power of the discharge, above 5 kilowatts for less than a microsecond, can melt and vaporise materials. Thin-film resistors may have their value altered by a discharge path forming across them, or having part of the thin film vaporized; this can be problematic in precision applications where such values are critical.[14]
Newer CMOS output buffers using lightly doped silicide drains are more ESD sensitive; the N-channel driver usually suffers damage in the oxide layer or n+/p well junction. This is caused by current crowding during the snapback of the parasitic NPN transistor.[15] In P/NMOS totem-pole structures, the NMOS transistor is almost always the one damaged.[16] The structure of the junction influences its ESD sensitivity; corners and defects can lead to current crowding, reducing the damage threshold. Forward-biased junctions are less sensitive than reverse-biased ones because the Joule heat of forward-biased junctions is dissipated through a thicker layer of the material, as compared to the narrow depletion region in reverse-biased junction.[17]
Resistors can fail open or short, alongside their value changing under environmental conditions and outside performance limits. Examples of resistor failures include:
Potentiometers and trimmers are three-terminal electromechanical parts, containing a resistive path with an adjustable wiper contact. Along with the failure modes for normal resistors, mechanical wear on the wiper and the resistive layer, corrosion, surface contamination, and mechanical deformations may lead to intermittent path-wiper resistance changes, which are a problem with audio amplifiers. Many types are not perfectly sealed, with contaminants and moisture entering the part; an especially common contaminant is the solder flux. Mechanical deformations (like an impaired wiper-path contact) can occur by housing warpage during soldering or mechanical stress during mounting. Excess stress on leads can cause substrate cracking and open failure when the crack penetrates the resistive path.[18]
See main article: Capacitor. Capacitors are characterized by their capacitance, parasitic resistance in series and parallel, breakdown voltage and dissipation factor; both parasitic parameters are often frequency- and voltage-dependent. Structurally, capacitors consist of electrodes separated by a dielectric, connecting leads, and housing; deterioration of any of these may cause parameter shifts or failure. Shorted failures and leakage due to increase of parallel parasitic resistance are the most common failure modes of capacitors, followed by open failures. Some examples of capacitor failures include:
In addition to the problems listed above, electrolytic capacitors suffer from these failures:
See main article: Varistor.
Metal oxide varistors typically have lower resistance as they heat up; if connected directly across a power bus, for protection against voltage spikes, a varistor with a lowered trigger voltage can slide into catastrophic thermal runaway and sometimes a small explosion or fire.[23] To prevent this, the fault current is typically limited by a thermal fuse, circuit breaker, or other current limiting device.
Microelectromechanical systems suffer from various types of failures:
In order to reduce failures, a precise knowledge of bond strength quality measurement during product design and subsequent manufacture is of vital importance. The best place to start is with the failure mode. This is based on the assumption that there is a particular failure mode, or range of modes, that may occur within a product. It is therefore reasonable to assume that the bond test should replicate the mode, or modes of interest. However, exact replication is not always possible. The test load must be applied to some part of the sample and transferred through the sample to the bond. If this part of the sample is the only option and is weaker than the bond itself, the sample will fail before the bond.[25]