Failure Analysis of Capacitors and Inductors - EEE Parts Database | doEEEt.com (2024)

The general construction of Tantalum,Aluminumelectrolytic,Multi-layer Ceramics,Film, SupercapacitorsandCommon Mode ChokeandSurface Mount inductors are explained. Major failure modes and the mechanisms for each one of these are discussed. Different failure analysis approaches used for these components and the development of some of these techniques are described as well.

Introduction

There is an increasing use and reliance on electronics in daily life, from portable electronics to pacemakers, and high reliability of these systems is expected and demanded. Passive components might not be the “brains” of these electronic systems. However, failure of any of these could result in a partial to complete electronic system shutdown. These malfunctions lead to annoyance or can escalate to safety or even life-threatening situations. Failure Analysis (FA) of these components helps determine the root cause and improve the overall quality and reliability of the electronic systems.

Passive components can be broadly divided intoCapacitors(CAPS),Resistors, andInductors (INDS), each having drastically different functions and constructions. There are subcategories within each category, based on different materials and constructions used in each, as listed in Tables 1 and 2 [1]. This makes the FA of passive components a broad topic. The current work supplements an article in the previous edition of Microelectronics Failure Analysis Desk Reference [2]. The general construction of select components, typical failure modes, and relevant FA techniques are discussed.

Some common FA practices are briefly discussed here before delving into the specific FA approaches for different components. Examining and documenting the failed component in as-received condition, such as physical anomalies and damage, orientation on the board, condition of the surrounding parts, and so on, are crucial, as these provide invaluable information. The next important step is to understand, document, and confirm the reported failure mode (i.e., leakage or short, open circuit, etc.).

Next, non-destructive internal and external examinations should be performed by optical microscopy, X-ray, and C-mode Scanning Acoustic Microscopy (C-SAM). The next logical step is performing basic electrical characterization, discussed in individual sections. The reader is encouraged to refer to the specific components section below to ensure that no further damage occurs to the component due to any of the above analyses. Before performing any FA, it is critical to gather the history of the failed component, including lot information, board mounting and washing conditions, any testing performed on it, when and how it failed, what failure was detected and so on. Another consideration is ensuring that counterfeit or wrong component is not the cause of failure. Understanding the construction of the component plays a vital role in performing FA, as it varies within each category and across manufacturers (some of which are listed in references [1, 3-7]).

The type of failure, component construction and condition determine the best isolation approach. Removal with hot air pencil, cutting the leads or solder pads, cutting the board around the component, etc., are the common options for isolation. The goal is to remove the component with little or no damage.

Proper safety and handling practices and applicable codes should be followed. Improper handling of chemicals and/or electrical systems could cause bodily injury or even death.

Capacitors

A simple capacitor consists of a dielectric between two conductive materials. One way of having high capacitance in a small volume is to increase the dielectric-electrodes surface area in a given volume, which is done using different types of constructions, as listed in Table 1. Different types of CAPS, along with their constructions and failure modes, are discussed below.

Tantalum Capacitors

Tantalum CAPS (Ta-CAPS) typically consist of a porous Ta anode (for high surface area) with a Ta wire attached. The dielectric, amorphous Ta2O5, typically a few tens of nm thick, is electrochemically grown on all surfaces of the porous Ta anode [1]. Liquid electrolyte, MnO2, or conductive polymer are three commonly used counter electrodes, with the latter becoming the most commonly used. The positive connection is typically created by welding the Ta wire to a Lead Frame (LF).

The connection between the counter electrode (polymer or MnO2) and the negative LF is made using conductive pastes and/or solders for solid-state devices. Typically, the surface mount devices (SMD) are over-molded (Fig. 1) [1]. The wet and some through-hole devices are typically hermetically sealed in a metallic can or a ceramic container. An acid-based liquid solution is used as the negative electrode for wet Ta CAPS. Other through-hole devices are conformally coated or over-moulded.

Thanks to their benign failure mode and better electrical performance, conductive polymer Ta-CAPS are commonly used Ta-CAPS and the rest of the section is mainly focused on this type. However, most of what is discussed applies to other Ta-CAPS as well. Ta-CAPS can fail in high leakage/short, highESR(Equivalent Series Resistance) or open/lowcapacitancemode, with high leakage/short being the most common mode.

Before starting the actual destructive analysis, it is critical to document, as accurately as possible, the condition and the failure mode of the failed Ta-CAP in its as-received condition. The external and internal constructions of Ta-CAPS could be significantly different from one type of Ta-CAP to another (Fig. 3) and from one supplier to another [1, 3-7]. Consequently, a thorough understanding of the Ta-CAP’s construction is essential for a fruitful FA. Each manufacturer best understands their components and, hence, is best equipped for performing FA on them.

Recent developments in X-ray technology allow a better understanding of the internal construction of these Ta-CAPS. One such example is X-ray Computed Tomography (X-ray CT), which allows simultaneous virtual sectioning of a component from different directions. Fig. 4, for example, shows virtual sections of a polymer herm-sealed Ta-CAP from different directions, showing Ta anode alignment in the can and the presence of voids in the can-to-lid solder seal.

The next step in the FA process is to document the electrical characteristics of the Ta-CAP, such as capacitance, dissipation factor (DF) and ESR. It is noteworthy that DC leakage is purposely omitted here. Ta-CAPS, like some other CAPS, have self-healing capability. A Ta-CAP with high leakage conditions could heal once DC voltage and current are applied. Conversely, applying voltage and current can cause a thermal runaway condition, further destroying the Ta-CAP. Either of these scenarios is not helpful for FA.

It is therefore recommended to measure DC-resistance (DCR) instead of DC-leakage, which for a good Ta-CAP (depending on the type of CAP) will be in the megaohms range. The analyst should carefully limit the voltage and current the CAP exposes during DCR measurements, which could alter the CAP. In the absence of any external or internal anomalies detected by optical and X-ray examination, thermal imaging could help determine the location of the failure site.

The temperature gradient in a thermal image highlights the source of heat. A hot spot determined from thermal imaging indicates localized heating from a failure. Fig. 5 shows a bright red/white hot spot, highlighting the failure site near the negative end of the component. Thermal imaging, X-ray, and optical imaging help get to the fault plane, allowing detailed analysis of the fault site and the surrounding area. This combination is especially helpful for multiple anodes and stacks Ta-CAPS to determine how many and which anode(s) have high leakage or short condition. Caution should be exercised to limit the current and voltage applied to the Ta-CAP for thermal imaging, as it could cause additional damage.

A voltage applied should not exceed the application or rated voltage, whichever is lower. The current can be limited via the power supply or by adding a resistor to the circuit. Thermal imaging is a balancing act between limiting the voltage and current exposure and obtaining a thermal hot spot.

Examination of the fault site might not always allow the determination of the cause of the failure. Different causes of failures (Fig. 2) lead to similar results: damaged dielectric, thermal interaction between Ta, dielectric and counter electrode, etc. The thermal and mechanical damage resulting from a failure can destroy evidence as to what caused the failure.

This is where knowing the history and conditions the component was exposed to, be it board mounting, washing, storage, electrical and thermal exposures, etc., becomes very helpful in determining what failure mechanism/s might be at play. For example, most surface-mount Ta-CAPS are not hermetically sealed.

Extended exposure of these molded devices to harsh environments, such as high humidity and temperature, could lead to metal corrosion and ionic migration (e.g. Cu, Ag, Sn). This could create bridging on the positive end or compromise healing ability on the negative end. Significant work has been done to reduce/eliminate these ionic migrations [9]. Ionic migration might not resemble a typical dendritic structure because it occurs in different layers or interfaces between different materials. Significant moisture pickup can also cause internal counter-electrode swelling and/or oxidation, resulting in higher ESR.

Moisture pickup by the CAP can also increase capacitance by connecting areas not completely covered by solid electrolytes. It is worth mentioning that some of the Ta-CAPS are classified as Moisture Sensitivity Level 3 (MSL 3) by the manufacturers and should be treated as such. Improper handling could lead to immediate or latent failures. Storage of molded Ta-CAPS in a humid environment can lead to significant moisture pick up, resulting in a “popcorning effect” (swelling and cracking of mold epoxy due to gas pressure buildup) during reflow.

Next, X-ray analysis in as-is form could provide valuable information for internal construction and solder connections. Once the CAP is electrically isolated by cutting the traces on the board, ESR should be measured again. Removal of the CAP from the board should be done with the least amount of thermal and mechanical stress on the CAP, as either one of these could change the connections.

Cutting the board around the CAP, with minimal damage, is one of the options. This allows thorough internal construction examination with X-ray analysis; in some cases, X-ray CT could be very helpful. De-soldering if necessary, should be done carefully to avoid overheating the CAP. Fig. 7 shows an example of a negative connection issue which could lead to high ESR. X-ray CT images (Fig. 7) show virtual sectioning of the same CAP from two different directions. Significant delamination and voids (red arrows) between the silvered Taanode and the conductive adhesive used for attaching it to the negative LF can be seen in this case.

Based on the X-ray analysis, the CAP could be prepared for detailed internal examination by sectioning it to a specific plane. Typically, for a single Ta anode CAP, sectioning it parallel to the Ta wire to the centre of it exposes all the interfaces. For a CAP with a significantly higher ESR, micro-probing of the sectioned part allows for further isolation of the problematic interfaces. Microprobing allows resistance measurements of interfaces in a Ta-CAP.

These measurements typically take the form of positive LF to Ta wire, negative LF to conductive adhesive (CA), counter electrode layers to CA, etc. This approach allows the determination of the interface/s of higher resistance, contributing to the overall ESR. SEM analysis can then provide a closer look at these specific interfaces to help determine the root cause. Caution must be exercised in sectioning and micro-probing because both steps can introduce artefacts. Potting and sectioning should be done to minimize smearing or layer separations, which could give misleading results.

Low capacitance and high DF are typically not common issues for Ta-CAPS, and interested readers can look up [8] for more details. In case of the wet Ta-CAPS breach in the hermetic seal, it can result in electrolyte leakage, which could cause low CAP and/or high leakage/short issue if it bridges between the positive wire and the can.

Al-tabs attached to both anode and cathode foils during the winding process provide electrical connections to the external terminals. Etched anode foil is used to gain significant surface area, hence higher capacitance in the same volume. The dielectric, aluminum oxide, is grown electrochemically over the etched surface before assembling the wound element. Fig. 9 shows examples of the internal construction of two different types of Al-CAPS from a manufacturer. Once the wound element is placed in the Al-can
and sealed (not hermetically), an insulative jacket/coating is typically applied to isolate the Al-can (negative terminal) from the environment. The rubber seal (not hermetically sealed) could allow slow electrolyte loss, eventually resulting in parametric failures.

Leakage of electrolytes because of poor workmanship (bad seal, vent, etc.), external damage to Al-can, seal or terminals, or exposure to high temperatures could also result in a similar problem. This can be detected by optical examination followed by any residual analysis, such as Fourier Transform Infra-Red (FTIR) spectroscopy. Bubble testing could be another possibility to check the seal of the can. Significant mechanical impact on the Al-can could result in foil damage and electrolyte leakage, leading to high leakage current.

High mechanical impact and/or vibration can also compromise the internal connections of the foils/Tab/terminals, which cause high leakage/short or open circuit conditions. Severe vibration can also result in similar internal or external damage to these CAPS, though it might not be easy to detect the effects of vibration. The failure analyst should examine if the parts were secured properly in their application environment (e.g., clamped).

High transient voltage can cause the breakdown of the Al-oxide dielectric, resulting in high leakage or short failure. Application circuit analysis is helpful in cases like these. Al-CAPS can fail in short/catastrophic breakdown, open or low capacitance mode. High leakage or short in these CAPS, like Ta-CAPS, also occurs due to dielectric compromise or by-pass of the active CAP. Fig. 10 lists some of the common causes for high leakage or short failure modes of Al-CAPS. The by-pass could occur internally or externally. In both cases, the bridging between positive and negative terminal/Alcan could occur due to corrosion or contamination.

The external bridging could occur on the board. Careful external examination of a failure in an as-received condition is critical for determining this. Entrapment of solution between Al-can and the insulative sleeve/coating (washing, condensation), either during manufacturing or post-board mount washing, can later cause the liquid to seep out and cause bridging. The presence of ionic species, typically from different components, the boards themselves, or the fluxes used, can make this liquid conductive. This can bridge and/or can cause electromigration of conductive species.

Exposure of parts during transit and/or at a customer’s location to chemicals can result in corrosion of the external terminals, resulting in leakage and/or ESR issues. Seepage of some chemicals into the Al-can through the seal (not hermetically sealed), especially chloride-containing chemicals, can then attack the aluminum oxide and create a leakage condition.

Once a high leakage or short condition is confirmed, an internal construction examination of the Al-CAP by X-ray CT could help reveal the fault site. After all the non-destructive analysis is performed, the wound element can be removed from the Alcan and unwound for further analysis. Analysis of leakage/short site by optical and/or SEM-EDS analysis could provide more insight into the cause (e.g. contamination, damage, etc.).

Multi-Layered Ceramics Capacitors (MLCC)

Unlike Ta and Al-electrolytic CAPS, MLCCs are non-polar devices and thus work with an electrical bias applied in either direction. The MLCCs are produced by stacking sheets of dielectric (e.g. BaTiO3, CaZrO3, etc.) layers on top of each other with electrode layers in between. The electrode layers are shorter in dimension than the dielectric layers to avoid exposure to the outer surface except from one side.

The stacking is done so that every other electrode is displaced to be exposed to one termination side, thus making two adjacent electrodes connected to opposite terminals. After firing (high-temperature sintering) to attain bonding between different layers, end terminations are applied to make an electrical connection with all the electrodes on each side. Two different constructions of MLCCs are shown in (Fig. 11): a regular design where every other electrode is connected to one terminal (Fig. 11a) and a higher reliability design which has floating electrodes, not connected to either terminal to provide safer failure mode (Fig. 11b). Apart from internal construction differences, MLCCs are available in a variety of different sizes and constructions (Fig. 12).

The contamination leading to external bridging could come from various sources, including the production process of MLCC, their storage and handling, board mounting and/or application. The presence of moisture and applied bias with surface contamination can create an ideal environment for the electromigration of metals such as Sn, Ag or Cu, which are used in the construction of these MLCCs and in the solders used for mounting and the printed circuit boards themselves.

The scanning electron microscopy image in (Fig. 13) shows the migration of Ag on the surface of a contaminated MLCC after high temperature and high humidity exposure with voltage applied. For an over-molded MLCC, if electromigration is suspected, the epoxy should be removed carefully, and both the epoxy and the MLCC surfaces should be examined for evidence of electromigration.

Once the MLCC has been analyzed externally, cleaning the surface and remeasuring IR can help confirm if the issue was caused by external bridging. For multi-chip devices, the problematic chip must be identified and isolated before doing any further analysis. The failure analyst needs to be careful not to introduce artefacts while doing this. If the external bridging is confirmed, no further FA is needed.

Another cause for low IR or short condition is cracks in the MLCC, especially flex cracking, which can result from stress on the rigid MLCC during board mounting and/or board handling. Cracking can expose new surfaces of two or more opposing electrodes nearby. Moisture or other conductive material can get to these exposed surfaces and can bridge the two opposing electrodes causing low IR or short conditions. The presence of moisture and bias could also result in electromigration in these cracks. The heat generated by a failure can further propagate the cracks, thus making it even worse. Flex cracking is one of the most common causes of failure and typically shows a crack from the edge of the termination at the board into the chip (Fig. 14). It is worth pointing out that chips are not always board mounted with the electrodes parallel to the board, especially for the near square chips (similar width and height), thus making detection of flex cracks somewhat tricky.

With a better understanding of flex cracking, manufacturers of MLCCs have developed flex crack mitigation designs. One such approach is the general improvement of the flexibility of the component. MLCCs that can withstand up to 5mm of flex are available [14]. Another improvement is the usage of flex terminations, which tear off from the ceramic body rather than flex cracking it, thus making MLCCs fail as an open circuit (safer failure mode) rather than low IR or short condition [14].

Delamination/ parallel cracks between the electrode and dielectric, typically a manufacturing fault, could grow with thermomechanical stress during the application. Consequently, these could jump across different electrode layers and cause cracking of the dielectric layer, thus exposing two opposing electrodes. Formation of a conductive path (as described above) between these opposing electrodes (moisture and/or ionic species) can lead to low IR or short conditions.

C-mode Scanning Acoustic Microscopy (C-SAM) allows the determination of such delamination. High-reliability MLCCs are routinely scanned using C-SAM to inspect for any delamination. Once cracks or delaminations are detected, sectioning the part to the plane of interest could provide more insight into the root cause of the failure.

If it is not readily visible by optical examination, then Cu back-plating (a technique to decorate the active electrodes), thermal imaging, or microprobing can identify the culprit. Once the electrodes in question are identified, SEM could allow a deeper understanding of the cause of delamination.

Film Capacitors

Like MLCCs, film CPs are non-polar devices with a similar layered structure. In this case, the dielectric layer, polymer film typically polypropylene (others include polystyrene, polycarbonate, etc.) 2-20 μm thick, is metalized with a few nm thick Al and/or Zn layer (which acts as the electrodes). The film’s small “margin” is left unmetallized on one end.

Hundreds or even thousands of these layers are stacked or wound together, with every other layer being displaced to one side (Fig. 18) and the margin being on the opposing end. End termination is applied on each side, typically with the same metallization (Al or Zn) sprayed on it, followed by a solder layer, typically Sn. End termination makes the electrical connection to the thin metallization layer in each layer in the winding.

The external surface of the end termination provides a solderable region and physical support to the winding. Depending on the design, electrical terminals are attached on each end, and the whole assembly is put in a case filled with epoxy or polyurethane (PU) resin. Sometimes, this step is skipped, and the capacitor is used in naked/unpotted construction. A number of these capacitors could be joined together before encasing them to achieve the desired electrical properties.

Film CAPS are available in various constructions, sizes and designs (Fig. 19) for different applications. The failure analyst needs to understand these before starting the FA. The film CAPs are also capable of self-healing, preventing catastrophic failure. If there is a dielectric breakdown, the energy released by the arcing (plasma discharge) at the breakdown site evaporates the thin metal layer in the surrounding areas of the fault site. This so-called “clearing phenomenon” isolates the fault site thus restoring low leakage current with a small capacitance drop (Fig. 20).

Thin spots and holes in the film can provide leakage paths. Entrapment of any hard particles in the film winding can also create holes in the film, thus allowing opposing terminals to be in very close vicinity. Upon voltage application, this scenario can provide an electrical path resulting in high leakage or short.

Bridging between two opposing terminals could also result from the metallization of one film connecting to both terminals. This can happen during production, or localized heating in the CAP, or high-temperature exposure, especially during the board mounting process. An extreme example of this type of failure detected by an X-ray (Fig. 21), revealed solder reflowed into the “film winding” thus bridging the opposite metallization layers, resulting in a dead short.

Since the film-CAPS are made by winding the metallized film, it allows a unique FA approach for these, i.e. the film can be unwound after removing the casing and the end termination. This allows detailed visual examination of the foil in reflective and transmission modes for defects. The area of concern can be further analyzed using optical and/or electron microscopy techniques.

SuperCapacitors

Super CAPS are polar devices that differ greatly from regular CAPS in their operations and applications. These CAPS have much higher capacitance than the regular CAPS, with lower voltage limits. These CAPS bridge the gap between regular CAPS and batteries. Unlike regular CAPS, there is no dielectric in Super CAPS.

Rather, these consist of a conductive electrode and an electrolyte. Once the voltage is applied, the charge is stored electrostatically and/or electrochemically. Figure 22a shows the basic base cell of a super CAP, consisting of two opposing electrodes, typically conductive rubbers, with a mixture of activated carbon and dilute sulfuric acid paste between two conductive rubber discs. An organic separator isolates the positive and negative ends yet allows the charge to travel across.

Sealing material and curing of the rubbers seal the base cell. A desired number of these base cells are stacked on top of each other, and terminals are connected to obtain the desired product (Fig. 22b).

Unlike other capacitors, typically, super CAPS fail in high ESR or open mode. Most of these failures occur because of water evaporation from the electrolyte. Failure analysis involves external and internal examination, optical and X-ray analysis, and basic electrical testing. This is followed by decapsulation of the CAP to examine the individual cells. Cells are analyzed for any electrolyte leakage sites. The electrolyte leakage/evaporation could result from a lack of proper sealing during the manufacturing process or exposure to high temperature.

In some extreme cases, very fast heating to higher temperatures can result in pressure build-up and explosion of the part. Like Al-CAPS, Super CAPS are not hermetically sealed and have an end of life; therefore, knowing the part history is critical.

Inductors

Theinductor (IND) is an electrical component that can store magnetic energy. These are used mainly for power conversion like DC/DC converter and noise reduction applications. An IND is comprised of aconductivecoil andmagnetic material. The conductive coil is generally made of insulation-coated wire called “magnet wire or enamelled wire”, copper trace, or sintered silver paste. The representative magnetic materials are ferrites and iron-based metal alloys.

Based on the construction, INDS can be divided into two common types, Common Mode Choke Coil and Surface Mount Powder Choke Coil, (Fig. 23), both are discussed in the following.

Common Mode Choke Coil

Common Mode Choke Coil(CMCC) is used for common mode noise reduction in power lines. It is normally comprised of a toroidal core and wound wire. The core is ferrite, pressed metal powder, or laminated metal ribbon. Insulation-coated copper wire is used for winding. The ends of the wound wire are used as the connecting terminals.

CMCC’s electric performance depends on the permeability of the magnetic material, core size, and number of turns of the wire. The core has some conductivity and is encased to keep insulation between the wire and the core. The diameter of the wire is determined by its rated current. For high current ratings, sometimes two or three wires are wound simultaneously around the core instead of one thicker wire, which might be difficult to wind to the core. The material of the core is determined by required permeability and environmental conditions such as temperature, humidity, and mechanical stresses.

Surface Mount Inductors

Surface Mount Inductors (SMI) is used for power conversion circuits such as DC/DC converters. They are comprised of the preformed inner coil, which is molded over using a composite material (a mixture of a magnetic metal alloy’s powder and resin). The inner coil is made of round or squared wire with insulation coating. Insulation coating has two roles: insulating between adjacent wires and between the coil and the over-molded material. The manufacturing process involves setting up the inner coil in a die, backfilling the die with composite material and pressing it, curing the resin and plating solder on the wire ends, and finally folding the wire ends to form terminals.

Surface Mount Powder Choke Coil could fail to open, short or exhibit parametric losses. Improper conditions at pressing could damage the wire’s insulation by metal powder (part of composite) rubbing/pushing against it, consequently creating a short condition. This type of IND is attached to the board by two terminals, which also support its heavy body. Any damage by mechanical stress, especially vibration, can compromise either or both terminals, creating an open condition. When vibration frequency is consistent with IND’s natural resonance frequency, these terminals might break, resulting in an open mode failure.

Acknowledgements

The authors would like to thank KEMET colleagues for their help preparing and reviewing this chapter, especially A. Parker, B. Reeves, D.Hepp, P. Bryson, M. Fulton, Z. Dou, V.Andoralov, D. Adam, M. Wright, M. Michelazzi, D. Montanari, J. Chen, C. Fischer, C. MotaCaetano, A. Gurav, C. Riedl, J.Bultitude, O. Pirakaew, P. Khomwongthep, K. Oga and P.Lessner.

We would also like to thank J. Kaplan of Cornell Dubilier for his help.

References

[1] KEMET Electronics Corporation, www.kemet.com
[2] Ross, R., Microelectronics Failure Analysis Desk Reference 6th Edition, ASM International, (Ohio, 2011),pp. 111-120.
[3] www.avx.com
[4] www.cde.com
[5] www.murata.com
[6] www.vishay.com
[7] www.panasonic.com
[8] Qazi, J., “An Overview of Failure Analysis of Tantalum Capacitors”, Electronic Device Failure Analysis, Vol. 16, No. 2 (2014), pp.18-23.
[9] Ye. J, Stolarski, C., Yaun, M., “Conductive Polymer Based Tantalum Capacitors for Automotive Applications”, Tantalum International Study Center, 56th General Assembly Meeting, 2015.
[10]Freeman, Y., Hahn, R., Lessner, R., and Prymak, J., “Reliability and Critical Applications of Tantalum Capacitors”, Capacitors and Resistors Technology Symposium, Electronic Components, Assemblies and Materials Association (ECA), March 2007.
[11] Freeman, Y., “Effects of Thermal Oxide on Anodic Oxide of Ta”, Capacitors and Resistors Technology