A capacitor is fundamentally two conductive plates separated by an insulator. But knowing that abstract definition doesn't tell you why a 10µF multilayer ceramic capacitor (MLCC) costs $0.05 while a 10µF film capacitor costs $4.00, or why the MLCC will violently crack if you bend the PCB. To understand component selection, safe substitution, and catastrophic failure, you need to understand exactly how a capacitor is made. The manufacturing process dictates the parasitic effects, the temperature coefficients, and the physical limits of the part on your bench.
Inside the Shell: How Different Capacitors Are Actually Made
The physical construction of a capacitor determines its real-world behavior. Here is how the three most common bench types are manufactured.
Multilayer Ceramic Chip Capacitors (MLCC)
MLCCs are made through a process called tape casting. Ceramic powder (like barium titanate) is mixed with a binder and cast into ultra-thin sheets, sometimes less than 0.5µm thick. Internal electrodes are screen-printed onto these sheets using nickel or copper paste. Hundreds of these sheets are stacked, pressed, and cut into individual chips. The chips are then sintered in a kiln at temperatures exceeding 1000°C, burning off the binder and fusing the ceramic into a solid, brittle block. Finally, the ends are terminated with silver and plated with tin for soldering. This monolithic construction gives MLCCs incredibly low equivalent series resistance (ESR) but makes them highly susceptible to mechanical stress.
Aluminum Electrolytic Capacitors
To get high capacitance in a small volume, manufacturers need massive surface area. They achieve this by electrochemically etching high-purity aluminum foil, creating microscopic tunnels that increase the surface area by up to 100 times. This etched foil is anodized to form a thin layer of aluminum oxide, which acts as the dielectric. The anode foil is rolled together with a spacer paper and a cathode foil, soaked in a liquid electrolyte, and sealed inside an aluminum can. The liquid electrolyte actually serves as the true second electrode. Because it is a liquid, it slowly evaporates through the rubber end seal over the years, which is why these capacitors have a defined operational lifespan.
Film Capacitors (Polypropylene / PET)
Film capacitors start with a thin plastic film that is metallized in a vacuum chamber with a nanometer-thick layer of zinc or aluminum. This metallized film is either wound into a tight cylinder or stacked into a block, impregnated with oil or epoxy to exclude moisture, and enclosed in a plastic or metal shell. The extreme thinness of the metallization gives film capacitors a unique self-healing property: if a microscopic defect in the film causes an arc, the intense heat instantly vaporizes the surrounding metallization, clearing the short circuit and restoring functionality.
Decoding the Shell: Reading Capacitor Markings and Codes
Because of how they are made and sized, different capacitors use entirely different marking schemes. Here is what the markings mean when you pull a part from your bin.
MLCCs: The 3-Digit EIA Code
Small ceramic capacitors rarely have room for text. Instead, they use a three-digit picofarad code. The first two digits are the significant figures, and the third digit is the multiplier (number of zeros).
- 104: 10 followed by four zeros = 100,000 pF = 100 nF = 0.1 µF.
- 473: 47 followed by three zeros = 47,000 pF = 47 nF.
- 221: 22 followed by one zero = 220 pF.
You may also see a trailing letter for tolerance: J (±5%), K (±10%), or M (±20%).
Electrolytics: Explicit Values and Polarity
Electrolytics have enough surface area for explicit text. They will state the capacitance (e.g., 470µF) and the maximum DC working voltage (e.g., 25V). Crucially, they feature a prominent negative stripe on the side of the can, often accompanied by minus signs (-). The longer lead on a through-hole part is the positive anode.
Film Capacitors: Text and Legacy Codes
Modern film caps usually spell out the value (e.g., 0.1µF or 100nF) and voltage (e.g., 630VDC). Older or smaller film caps might use the same 3-digit EIA code as ceramics, or use a nanofarad notation like .047 (implying 0.047µF).
The Selection Matrix: Which Dielectric for Which Job?
Knowing how a capacitor is made tells you which one to use. You cannot blindly swap a 100nF capacitor for another 100nF capacitor if the dielectric material is wrong. According to Cornell Dubilier's Capacitor Application Guide, the dielectric dictates the temperature stability, losses, and parasitic behavior.
| Type / Dielectric | Construction | Tolerance & Tempco | Typical Use | Safe Substitution Rule |
|---|---|---|---|---|
| C0G / NP0 (Ceramic) | MLCC (Paraelectric) | ±5%, 0 ±30ppm/°C | RF filters, timing circuits, precision oscillators. | Sub only with other C0G. Never use X7R/Y5V here; temp drift will detune the circuit. |
| X7R / X5R (Ceramic) | MLCC (Ferroelectric) | ±15%, ±15% over temp range | Decoupling, bypass, general-purpose bulk storage. | Can sub higher voltage rating. Watch for DC bias derating (capacitance drops under voltage). |
| Y5V / Z5U (Ceramic) | MLCC (High-K Ferroelectric) | +22% / -82%, severe temp drift | Non-critical bulk bypass where space is tight. | Avoid entirely in modern designs. Sub with X7R of double the nominal value. |
| Aluminum Electrolytic | Etched foil, liquid electrolyte | ±20%, high temp drift, high ESR | Power supply bulk filtering, low-frequency coupling. | Sub with higher voltage/low-ESR variants. Can sub with polymer caps for longer life. |
| Polypropylene Film | Metallized plastic film, wound | ±5%, highly stable, very low loss | Audio crossovers, snubber circuits, high-voltage AC. | Sub only with identical film types. Never sub with ceramic for AC line or snubber use. |
Bench Autopsy: Failure Modes and Visual Symptoms
How a capacitor is made directly dictates how it dies. Here is what to look for when troubleshooting a blown board.
MLCC Flex Cracking (Short Circuit)
The Cause: Because MLCCs are sintered ceramic blocks, they are essentially tiny pieces of glass. If the PCB flexes during assembly, testing, or operation, the board bends but the capacitor does not. The stress concentrates at the rigid termination points.
Visual Symptom: A hairline fracture running diagonally from the bottom corner of the metal terminal up into the ceramic body. It is often invisible to the naked eye and requires a 10x loupe. Electrically, it presents as a dead short or a highly leaky path.
Electrolytic Venting and Drying (Open Circuit)
The Cause: Excessive ripple current, overvoltage, reverse polarity, or simply reaching end-of-life (10+ years) causes the liquid electrolyte to boil or evaporate.
Visual Symptom: The scored cross on the top of the can bulges outward or bursts open, leaving a crusty brown or black residue on the PCB. Electrically, the capacitance drops to near zero, and the ESR spikes massively, causing power supply ripple to skyrocket.
Film Capacitor Corona and Metallization Clearing
The Cause: High dv/dt (rapid voltage changes) or voltage spikes exceeding the partial discharge inception voltage cause internal micro-arcing.
Visual Symptom: The outer plastic shell may bulge or melt. If you cut open the potting, you will see black scorch marks where the metallization has vaporized to clear a fault. Electrically, the capacitance drops incrementally with each self-healing event until it falls out of spec.
War Story: When a 10µF MLCC Destroyed a Buck Converter
Understanding manufacturing quirks saves you from the most insidious bench traps. Here is a real-world scenario involving the hidden flaw of ferroelectric ceramics: DC bias derating.
The Setup: I was prototyping a 5V, 3A buck converter using a standard TPS5430 switching regulator. The datasheet called for a 10µF output capacitor to manage the 500kHz switching ripple. To save space and BOM cost, I grabbed a handful of cheap 10µF, 25V Y5V MLCCs from a bulk bin instead of the recommended X7R parts.
The Numbers: At 5V output, the converter requires a minimum of 8µF effective capacitance to maintain stability and keep output ripple under 30mV. The Y5V capacitor was rated at 10µF at 0V bias and 25°C.
The Outcome: When I applied a 2A load, the output voltage rippled wildly (over 200mV peak-to-peak). The converter's control loop went unstable, the inductor began audibly whining, and the chip rapidly hit thermal shutdown. I scoped the output and saw massive sawtooth ripple.
What Went Wrong: I ignored how the Y5V dielectric is made. High-K ferroelectric ceramics like Y5V (and to a lesser extent, X7R) suffer from severe DC bias derating. The alignment of the internal dipoles saturates under an applied DC electric field. At a mere 5V DC bias, a 25V-rated Y5V capacitor loses up to 85% of its nominal capacitance. My "10µF" capacitor was physically acting like a 1.5µF capacitor in the circuit. The control loop starved for bulk storage and oscillated.
The Fix: I replaced the Y5V part with a 22µF, 10V X7R MLCC. Even though X7R also loses about 40% of its capacitance at 5V bias, starting with 22µF gave me an effective ~13µF in-circuit, stabilizing the loop instantly. For mission-critical power rails, I now default to X7R ceramics or low-ESR polymer aluminum capacitors, completely banning Y5V/Z5U from my lab inventory.
Ultimately, a capacitor isn't just a schematic symbol; it is a physical object bound by the laws of materials science. When you know how the dielectric is cast, etched, or wound, you stop fighting the component and start designing with its actual physics in mind.
