Aluminum capacitors are the workhorses of big
capacitors, they are small and cheap for their capacity and can be found in sizes from <1 uF to over 1 farad. They are commonly available to 450 volts working voltage, with a few to at least 600 volts, much
higher than other types of electrolytic capacitors. Aluminum electrolytic capacitors use a layer of aluminum oxide (with a K of about 8.5) grown on aluminum foil. The aluminum foil forms one electrode while
the other is a non-aqueous electrolyte (hence the name) in a thin paper separator, and another foil layer for the cathode. The seperator holds the electrolyte and keeps the cathode from rubbing on the dielectric
and damaging it. The classic electrolyte formulas date back to the 1930s and were usually a glycol or amine, in which a conductive salt is dissolved, plus 1-2% water. Many variations on this have been used
over the years, although glycol is still often used. A patent search will find any number of exotic electrolytes, but who knows how many of these are actually in production. Given the variations that
aluminum electrolytics are found in however, there are probably many propriety electrolytes in use. The electrolyte is specialized enough that many small pacific-rim cap makers buy their electrolyte (often from
Japan) rather than make it themselves. To make a non-polarized capacitor, the other foil electrode is also oxide coated.
The reason for the
electrolyte is that the surface of the oxide-coated foil electrode is etched to greatly increase the surface area. The ratio of capacitance with etched foil vs unetched foil is called "foil gain", and is typically
25- 100 for modern aluminum electrolytics, depending on voltage. The other foil electrode can't make intimate contact with the dielectric by itself so the electrolyte is used as an intermediary conductor.
Etching the foil is a science in its own right. High-voltage electrolytics require a different surface than low-voltage due to the different thickness in oxide coating required. There has been a
slow-but-steady improvement in etching that has improved volumetric efficiency over the years, roughly doubling over the last decade. As manufacturers phase in the improved foil, they will obsolete capacitor series that
use the old foil and replace them with smaller parts that use the new foil. However, the reduced-volume parts tend to have slightly higher ESRs than the equivalent older higher-volume parts.
Aluminum electrolytics have their problems, however; noise, high leakage, high temperature drift, high dielectric absorption, high inductance, high almost
everything bad. Low temperature is a problem for most aluminum capacitors. For most types, capacitance falls off rapidly below room temperature while dissipation factor can be ten times higher at -25C than
Most limitations can be traced to the electrolyte. At high temperature, the water can be lost to evaporation, and the capacitor (especially
the small sizes) may leak outright. At low temperatures, the conductance of the salts declines, raising the ESR, and the increase in the electrolyte“s surface tension can cause reduced contact with the
dielectric. The conductance of electrolytes generally has a very high temperature coefficient, +2%/C is typical, depending on size. The electrolyte is implicated in various reliability issues as well.
Because of aluminum capacitor's problems, other capacitor types compete for aluminum's turf in some applications. In small sizes, tantalum
capacitors are available, but at a higher cost. Ceramic capacitors are available into the 100s of uF and film capacitors are now available >1000 uF.
Aluminum capacitors are at their best doing simple jobs like line-frequency power supply filtering. If you want to use them for more demanding applications,
like switcher power supply filtering (low ESR), audio DC blocking (low-noise), or in high temperature environments (long life), you must use care in selecting one the of special-purpose aluminum capacitors. There
are many to choose from, however; apparently the basic technology gives capacitor designers a lot of room to maneuver. They include:
- Low noise for audio applications.
- Low leakage for RAM backup and timing applications. Vendors sometimes tout these as tantalum replacements
- Low equivalent series resistance (ESR) and/or low inductance (ESL) for switcher power supply filtering or for high ripple-current applications. There are even 4-lead types available.
- High-temperature types for improved life and reliability at elevated temperatures. Rated operating temperatures go to 130C for common capacitors, but at the cost of increased size. 150C types for
automotive applications have started to appear. Special types (rare, possibly no longer even in production) are available for operation to 200C.
- Low-temperature types, as low as -65C. Very high and very low temperature rating may not be found in the same capacitor.
- Non-polarized for speaker crossover networks and other audio applications.
- A variety of combination types such as high temperature + high reliability.
- Photo-flash capacitors to handle very high surge currents.
- Miniature sizes and special shapes (tall and skinny, short and fat).
- Axial leads, radial leads, screw terminals, tabs with holes for screws, quick-connect terminals and snap-in.
- SMD versions are now widely used in consumer electronics. Most SMD aluminum capacitors are just a modified version of the conventional aluminum can. True SMD aluminums are stating to appear
however. Some come in molded packages that might at first glance be mistaken for SMD tantalums.
http://www.evoxrifa.com/ Electrolytics with 150C rating.
Since the 1980s,
a solid "electrolyte" (actually a semiconducting polymer called 7,7,8,8-tetracyanoquinodimethane, or TCNQ) aluminum capacitor has been available from Sanyo under the OS-CON name. Solid electrolyte aluminum capacitors are now becoming widely available from many other manufacturers as well, some with TCNQ under the OS-CON name, some with other polymer electrolytes such as polypyrrol, and some with MnO
2 (see tantalum capacitors). Polymer electrolytes are two to five orders of magnitude more electrical conductive than traditional liquid electrolytes or MnO2. That gives these
capacitors better electrical properties, including much lower ESR for lower impedance in the 10 kHz to 10 MHz frequency range, high ripple current, and tantalum-like low-temperature performance.
Solid electrolyte capacitors have usually been considered "specialty" devices, but that has changed as users demand low ESRs in ever smaller packages. Solid electrolytes are only available in low voltages. I
have seen polymer types to 30 V. They are only available in moderate sizes at present, about 1 to 100 uF. Solid electrolyte capacitors first appeared in through-hole packages but their real popularity is now
in SMD. Because TCNQ based capacitors have such low ESRs, turn-on surge current is a potential problem. One manufacturer recommends no higher than 10 times the maximum ripple current, less for the smaller
Vishay for one has a line of remarkable MnO2
electrolytics. They claim extreme life at 125C, very low ESR and other good characteristics. They are only available at modest voltages however, up to 40V at 85C, derated to 25V at 125C. They are priced well above your average electrolytics.
Besides lower ESR, polymer aluminums have more reliable self-healing, and tend to fail open rather than short. A downside is that the ESR increases with
time due to trapped moisture (not true of all polymer electrolytes however). This is worse at higher temperatures. http://www.sanyovideo.com/ the original
Companies that advertise polymer-electrolyte electrolytic capacitors (tantalum or aluminum) include:
http://www.ic.nec.co.jp/ (or http://www.necel.com/home.nsf/Main?ReadForm&Passive+Components )
http://www.cornell-dubilier.com/surface.htm/ See new paper on much improved reliability parts.
Tantalum electrolytic capacitors are a step up from aluminum capacitors. They come in a
number of types with different advantages, but in general, they have smaller size, lower leakage, lower dissipation factor, lower ESR, more stable capacitance over temperature, and good service life. Tantalum
capacitors aren't made in the monster sizes that aluminum capacitors are, but are available to several hundred µF in common voltage ratings (to about 100 volts) and to several thousand uF at low voltage (6-10
The first tantalum capacitors were made from tantalum foil with a sulfuric acid electrolyte. Foil tantalums are made for use to at least
300 volts, and to 125C. They are made in both industrial and military styles but their usage is probably mostly military. Foil tantalums are now rare however, and seem be out of production.
The foil anode is a poor use of what is an expensive metal, and not in good supply. Most modern tantalum capacitors are what are called "slug"
capacitors. Tantalum powder is sintered into a porous yet strong slug, with a tantalum wire, which forms the anode of the capacitor. The many small particles produce a very high surface area. A variety
of particle shapes are used, depending on the operating voltage, and making them is a science in itself. A layer of tantalum pentoxide (Ta2O5
with a K of about 25) is grown over the particles for the dielectric. In one version of the capacitor, the electrolyte is gelled sulfuric acid. This is called a "wet-slug" and, as you can imagine, it must be well sealed. Wet-slug tantalums are the capacitor of choice for some applications, especially at very high temperatures (to 200C for some, often with derating). They are available for up to 900 volts operation. Construction details vary.
The vast majority of solid tantalums are the "dry-slug" or "solid tantalum"
capacitors. A layer of manganese dioxide (MnO2) is layered over the pentoxide followed by a layer of colloidal graphite and a layer of silver paint. There are minor
variations on this. With no liquid involved, they can be sealed with just an epoxy dip, although the better ones may have a molded body. The pentoxide layer is
prone to defects, and the key to the solid tantalum capacitor's reliability is that the MnO2 provides self healing.
If a flaw in the pentoxide layer develops, the leakage will cause localized heating in the MnO2 and convert it to Mn2O3
, a much less conductive oxide, sealing off the flaw. This mechanism is not perfect, however, and
failures due to dielectric flaws have been a traditional problem for solid tantalums. If the temperature of the
pentoxide gets high enough, about 500C, the pentoxide converts from its nonconducting amorphous form to its
conducting crystalline form and the capacitor goes up in flame. Solid tantalums aren't usually made with high
working voltages because the particle size limits the dielectric thickness that can be grown; 50 volts is usually
the upper limit. There are however, a very few solid tantalums with working voltages as high as 125 volts.
Most dry tantalums are rated to no more than 85C to 125C, but a few are rated for use to 150C with voltage
derating. Solid tantalum capacitors are commonly available in surface-mount packages, in molded bodies.
A weakness of slug tantalums is high-frequency performance. Depending on construction details, the
capacitance can fall to 50% in the 100-200 kHz region, compared to 1 kHz. Lower ESR parts can be made using larger particles (although at the expense of somewhat lower capacitance) and by various manufacturing
A rarely mentioned characteristic of solid tantalums is their rapid decline in leakage current. When both
aluminum and tantalum electrolytics are powered up, their leakage current starts high, but declines over time.
For aluminums, leakage take minutes to decline to a stable value. For tantalums, this occurs in seconds.
Even though "dry-slug" tantalums are not wet-chemistry devices, they are still polarized. Although many
sources claim that dry tantalums are very sensitive to reverse bias, they can actually to be very slow to fail
when reverse biased. In fact, failure of tantalums installed backwards may occur from seconds to more than a year after manufacture.
Because of the growing popularity of tantalum in miniature electronic equipment, tantalums (mainly SMD) have diverged into a number special types. They include:
- High temperature, to 150C with voltage derating.
- Low ESR using specially sized powders and multiple anodes.
- Miniature package sizes.
- Very low height. Standard SMD tantalums are 1.9 to >4 mm in height. Low profile are 1.2 or 1.5 mm.
- Fused, for power supply bypassing on digital boards.
- Tantalum capacitors with the semiconducting polymer "electrolyte" TCNQ have appeared. Sanyo, for one, makes them under the POSCAP name. See Sanyo OS-CON capacitors above. They tend to
have the same advantages as their aluminum cousins, including lower ESR and flatter temperature characteristics. Manufacturers also claim that they can fail without bursting into flame.
Companies that advertise wet-electrolyte tantalum capacitors include:
http://www.vishay.com/ The last know manufacturer, Vishay has apparently dropped wet-foil tantalums.
I think all Transitor does is tantalum. They are now a Vishay company.
http://www.kemet.com A recent addition.
Niobium (once called Columbium) has been under investigation for many years as a lower-cost
replacement for tantalum in electrolytic capacitors. The Russians were making very poor quality devices as
early as the 1950s. This research is paying off however as several companies are close to (or actually are) shipping niobium electrolytics in volume.
A much more common metal, niobium should not have the supply and cost problems seen by tantalum.
However, it has proven to be much more difficult to work with than tantalum, a major problem being to get leakage under control. Niobium, unlike tantalum, can form other oxides than the pentoxide, such as NbO andf
NbO2 which are conductive to some degree. The fewer the oxygen atoms, the higher the conductance.
Oxygen also will dissolve in niobium to some extent. At elevated temperature the nobium will start forming
lower oxides, increasing leakage. At first this kept the rated operating temperature to 85C. More recently
AVX as introduced 105C parts that are suitable for ROHS soldering temperatures (260C peak). Niobium“s dielectric (Nb2O5
) has a higher K than tantalum, about 40, but a lower breakdown voltage and a greater tendency for leakage requires that the film be thicker. The volumetric efficiency may turn out to be similar.
Niobium caps, like tantalum, require a voltage derating for applications that involve current surges. The fate of
niobium will probably depend most on user acceptance, which will rise and fall with the price of tantalum.
The next step seems to be to use niobium oxide, NbO, for the anode. NbO has advantages over Nb,
including lower weight, a much lower tendency to ignite (something niobium has in common with tantalum), and an improved ability to handle current surges. Only AVX is selling NbO caps to my knowledge.
Companies that advertise niobium electrolytic capacitors, all in SMD, include:
A Russian company, they seem to have little or no presence in the west.
http://www.epcos.com Has announced volume production.
http://www.vishay.com Vishay is shipping samples at least.
http://www.ic.nec.co.jp/compo/cap/english/topics/0107_e.html The only ones using a polymer electrolyte so far.
They areshipping niobium oxide caps. They also have good technical information.
Kemet has talked about niobium, but now seems less optimistic about its use in mainstream applications in the immediate future.
Sometimes called super caps, electrochemical caps, or Gold caps (an early Panasonic trade name), the
double-layer (DL) capacitor is a relative newcomer. I tend to think of DL capacitors as an odd form of electrolytic, but others regard it as a new class of capacitor. The classic DL capacitor does not use a
traditional dielectric, but rather uses air-gell carbon electrodes and a sulfuric acid electrolyte which forms a
vary thin "dielectric" layer. Many other systems are now appearing. Some DL capacitors are polarized, but
others are not. The basic carbon/sulfuric acid system is nonpolarized, so I am not sure why some of these caps are polarized.
More than a capacitor, less than a battery, to some extent the DL capacitor bridges the gap between
the two worlds. At least that“s what the manufacturers say. DL capacitor energy density (watt-hours/kg) is
much higher than conventional capacitors, almost as high as some batteries. Power density (watts/kg) is not as high as true capacitors, but much higher than batteries.
DL capacitors are widely used as replacements for small batteries such as for RAM backup, or to get
longer power bus holdup times when power failures occur. Self-discharge can be significant in some applications however. DL capacitors are also widely used for load leveling in small battery-powered systems
(mostly telecom), and there is a lot of research in this area. There are many applications where batteries give
out not because they are actually depleted, but because they can no longer handle current spikes. Future DL
caps are expected to be used for load leveling in larger and larger systems, even electric vehicles. DL caps are even touted for their environmental friendliness, they contain no lead or cadmium.
DL capacitors have generally been made from 0.047 F to 100 F (yes, farad), but much larger sizes are
now becoming available. However, they are usually available in very low voltage ratings. For board-level DL caps, most are about 2.5 volts,
while up to 6.3 volts is achieved by putting three cells in series in one package. Temperature range is limited; -25 to 70C is the most common, but a few are rated to 85C. ESR is usually
high; as low as <0.1 ohm for the really huge ones, but >100 ohms for the smallest sizes. Some newer types
do much better than this, with ESRs <0.1 ohms, even for the smaller sizes. DL capacitors are available in
several through-hole and coin packages. In spite of the limited temperature range, Tokin now has a 0.1 F/5.5 volt part designed for SMD soldering.
Like aluminum electrolytics, DL capacitors undergo an aging process. In general, capacitance gradually
drops, and ESR increases. This tends to level off after several thousand hours. Also like aluminums, leakage drops as DL caps age, leveling off after several 10s to 100s of hours.
The future of DL capacitors is considered so promising that there is extensive research into advanced
types that will overcome the shortcomings of existing devices. Some these are already reaching the market. http://www.skeletonnanolab.com/Supercap.html
. Companies that advertise double-layer capacitors include:
http://www.alumapro.com/cap.htm for 12 volt systems
http://www.elna-america.com/DLC-battery.htm Elna has the largest board-level DL caps I have seen
http://www.globar.com/maxcap.html also called cesiwid.com
http://www.chemi-con.co.jp/chemi-con/event/SWR98/double/double.html not sure these are available in USA
lots of good reference numbers
http://www.cap-xx.com/about/default.html includes technical articles
A common argument is over whether or not you can make a non-polarized capacitor by putting two
polarized electrolytics back-to-back. People have been doing this for years, with no problems to my knowledge.