ceramic capacitor higher values?

[willbanks]

why aren't there ceramic capacitors that can go into the ranges of thousands of micro farads? why is it seemingly dominated by electrolytic capacitors?

[mariush]

It's just more efficient to use polymer or electrolytic or tantalum caps.

Ceramic capacitors are made from lots of layers stacked on top of each other: https://en.wikipedia.org/wiki/Ceramic_capacitor#/media/File:MLCC-Principle.svg
For bigger capacitance you'd need way more layers and more surface so they'd be big and tall
They'd be sensitive to damage from pressure or bending of pcb , then there's potential for microphonics (google it) and their ugly relationship with voltage (see https://www.maximintegrated.com/en/app-notes/index.mvp/id/5527  or http://www.murata.com/support/faqs/products/capacitor/mlcc/char/0005 )

[willbanks]

does the size of the electrodes matter? like the thickness?

[T3sl4co1l]

Part manufacturing, part cost.

MLCCs are laid down layer by layer.  More layers = more time.

You can simply use a wider chip, but that has manufacturing problems (yield?) and reliability problems (more body size = more vulnerable to cracking).  IPC doesn't recommend bigger than 1210 size caps; I'd be okay with 1812s, but nervous about larger sizes.  (Larger sizes are usable either in lower reliability or lower vibration applications, or with flexible terminations or leadframes.)

OTOH, electrolytics are very simple to build (foil is mechanically and chemically prepped, then wound into rolls and shoved inside a can).  Polymer caps, I think, have more steps and the electrolyte is kind of expensive (it's an organic polymer, but synthesizing special ones that happen to be usefully conductive isn't as straightforward as making polyethylene!).  Tantalum caps scale very easily (they're a slug of porous tantalum metal, coated with electrolyte; that slug can simply be as big as you like*), downside is, tantalum itself is kind of expensive to mine and isolate.

*Given limitations of reasonable ESR, that is.

Polymer caps are king right now, for large values and low impedances.  Take apart a laptop, video card, or any high power computer hardware along those lines (maybe cell phones and tablets too) and you'll find oodles of them.

Tim

[David Hess]

They do make much larger ceramic capacitors for high reliability and high operating life applications which are constructed by stacking multiple multilayer ceramic capacitors and adding end terminations which ameliorates problems with mounting stress.  If you have to ask the price, then you cannot afford them.

I think my old dual Pentium 3 Xeon server uses them on its voltage regulator modules but usually I see them in high reliability industrial, military, and aerospace applications.

[Lee Leduc]

High value ceramic caps are available but they're expensive. For example, Mouser sells a 100uF at 100 volts. The catalog part number is 661-KHD101E107M99C0B. The price is $46 each.

[Zero999]

High capacitance ceramic dialectics suffer from saturation: as the bias voltage increases, the capacitance decreases, so the energy density is a fraction of what you'd calculate using E = ¹/₂CV². Electrolytic capacitors, both solid and traditional, don't suffer from saturation, have comparable energy densities to ceramics and are cheaper.

[T3sl4co1l]

High capacitance ceramic dialectics suffer from saturation: as the bias voltage increases, the capacitance decreases, so the energy density is a fraction of what you'd calculate using E = ¹/₂CV². Electrolytic capacitors, both solid and traditional, don't suffer from saturation, have comparable energy densities to ceramics and are cheaper.

C0G types do not suffer from this, and therefore achieve about triple the energy density.

Unfortunately, C0G are quite a bit more expensive than type 2 ceramics already, and aren't available in anywhere nearly large enough at lower voltages.  (They're good for bypass and RF in mains-voltage applications though.)

Tim

[Zero999]

High capacitance ceramic dialectics suffer from saturation: as the bias voltage increases, the capacitance decreases, so the energy density is a fraction of what you'd calculate using E = ¹/₂CV². Electrolytic capacitors, both solid and traditional, don't suffer from saturation, have comparable energy densities to ceramics and are cheaper.

C0G types do not suffer from this, and therefore achieve about triple the energy density.

Even higher energy density than aluminium electrolytic capacitors? I haven't done any calculations, have you?

[T3sl4co1l]

Even higher energy density than aluminium electrolytic capacitors? I haven't done any calculations, have you?

I wouldn't make the claim if I didn't.

Looking at my notes, the 1/3 is relative to an ideal capacitor of the same size.

A real C0G part doesn't achieve the same density though, so the volumetric advantage is 1.6 times.

The exact comparison is a CL43B104KGINNNE (fitting a curve to the C(V) graph and summing over that to find total energy storage), to a C5750C0G2J104J280KC (assumed ideal).

Electrolytics are around 4 times more energy dense than C0G, but not as fast.  If you add a converter stage to effectively multiply the value of the C0G capacitor, you can achieve a smaller volume than electrolytics, for supply filtering applications.  Electrolytic cannot handle that much ripple (current or voltage), so they're stuck doing it by sheer numbers.  (This approach was used in the winning Little Box Inverter Challenge entry, but additionally using Ceralink type capacitors for higher average capacitance under bias conditions.)

Tim