Czochralski vs float zone: what is the difference?
Every silicon wafer starts as a single-crystal ingot, and the growth method sets the oxygen content, the purity, the resistivity range and the largest diameter available. For logic and memory the choice was settled long ago, but for power devices and detectors it still matters.
The short answer
The core difference is the crucible. The Czochralski (CZ) method pulls a single crystal from molten silicon held in a quartz crucible, so the crystal picks up oxygen from the crucible but can be grown at 300 mm diameter and is the standard for almost all wafers. The float zone (FZ) method melts a narrow zone of a polysilicon rod with an RF coil and needs no crucible, giving very low oxygen and very high resistivity, but it is limited to smaller diameters and costs more.
Side-by-side comparison
| Aspect | Czochralski (CZ) | Float zone (FZ) |
|---|---|---|
| Crucible | Quartz crucible holding the silicon melt | None; a molten zone is held between solid rod sections by surface tension |
| Starting material | Polysilicon chunks melted in the crucible, with dopant added to the melt | A high-purity polysilicon feed rod |
| Oxygen content | Contains interstitial oxygen dissolved from the crucible | Very low oxygen and carbon |
| Resistivity | Covers the range used by standard devices | Very high resistivity possible |
| Diameter | Up to 300 mm in volume production | Smaller diameters; 200 mm is the practical upper end |
| Mechanical strength | Oxygen pins dislocations and strengthens the wafer | Lower resistance to slip during thermal processing |
| Cost | Lower; mature high-volume process | Higher; needs high-grade feed rods and slower growth |
| Main uses | Logic, memory, image sensors and most other ICs | High-voltage power devices, radiation detectors, some RF substrates |
How CZ growth works and why it dominates
In the CZ method, polysilicon is melted in a quartz crucible, a seed crystal is dipped into the melt and slowly pulled upwards while seed and crucible rotate. Controlling pull rate, rotation and the thermal environment sets the diameter, the defect density, the oxygen concentration and, through the dopant added to the melt, the resistivity. The process scales to ingots for 300 mm wafers, and essentially all advanced logic and memory wafers are made this way.
The oxygen that dissolves from the crucible is not only a contaminant. During device processing it forms precipitates in the wafer bulk that trap metal impurities (intrinsic gettering), and dissolved oxygen makes the wafer more resistant to slip and warpage at high temperature.
How FZ growth works and what it offers
In the FZ method, a polysilicon rod is held vertically and a radio-frequency coil melts a narrow zone near a seed crystal. The zone is moved along the rod, and the silicon recrystallises as a single crystal behind it. Because the melt never touches a crucible, oxygen and metal contamination are very low, and impurities that prefer the liquid are swept along by the moving zone.
The result is silicon with long minority-carrier lifetime and very high resistivity. Uniform n-type doping of high-resistivity FZ silicon is often achieved by neutron transmutation doping, in which neutron irradiation converts a small fraction of silicon atoms into phosphorus. These properties suit high-voltage IGBTs, thyristors and diodes, and particle and radiation detectors. The limits are diameter, because the molten zone must be supported by surface tension alone, and cost.
MCZ: closing the gap for power devices
Magnetic-field-applied CZ (MCZ) applies a magnetic field to the melt to damp convection, which reduces the oxygen transported from the crucible into the crystal and improves resistivity uniformity. MCZ wafers bring CZ-type diameters and costs closer to FZ-type properties, and they are increasingly used for power devices that once required FZ material, especially at 300 mm.
The practical rule is therefore three-way: standard CZ for mainstream ICs, MCZ where low oxygen is needed at large diameter, and FZ where the highest resistivity and lowest oxygen are essential.
FAQ
Why are 300 mm wafers made by CZ?
In FZ growth the molten zone is held only by surface tension, which makes large diameters very difficult. CZ pulls the crystal from a melt supported by a crucible, so large, long ingots can be grown reliably, and all 300 mm production uses CZ or MCZ.
What are FZ wafers used for?
Applications that need very high resistivity, long carrier lifetime and low oxygen: high-voltage power devices such as IGBTs, thyristors and diodes, radiation and particle detectors, and some high-frequency device substrates.
Is oxygen in CZ silicon a problem?
It depends on the device. Oxygen helps gettering and mechanical strength, but it can form thermal donors that shift resistivity and can degrade some power and sensor devices. Growers control it through crucible rotation, pull conditions and magnetic fields (MCZ).
Which Japanese companies make CZ and FZ silicon?
Shin-Etsu Handotai and SUMCO, the main Japanese silicon wafer makers, both produce CZ wafers, including polished and epitaxial types for 300 mm.
Terms compared
- Czochralski Process (CZ Method)The standard method for growing single-crystal silicon ingots by slowly pulling a rotating seed crystal out of molten silicon in a quartz crucible.
- Float Zone Process (FZ Method)A crucible-free method that grows very pure, low-oxygen single-crystal silicon by moving a molten zone along a polysilicon rod.
- Magnetic Czochralski Method (MCZ)A Czochralski crystal growth method that applies a strong magnetic field to the silicon melt to suppress convection and control oxygen.
- Power SemiconductorSemiconductor devices built to switch, convert and rectify electric power at high voltage and current, such as power MOSFETs, IGBTs and diodes.
Japanese companies involved
- Shin-Etsu Quartz Products · High-purity quartz glass parts and crucibles
- Shin-Etsu Chemical · Silicon wafers, photoresists and photomask blanks
- SUMCO · Silicon wafers
- Mitsubishi Electric (power devices) · IGBT and SiC power modules, RF and optical devices
- Fuji Electric (power semiconductors) · IGBT modules and SiC power devices
- ROHM · SiC and silicon power devices, analog ICs
- Toshiba Electronic Devices & Storage · Power MOSFETs, IGBTs and discrete semiconductors
- Tokyo Electron (TEL) · Coat/develop, etch and deposition equipment
More wafers and substrates comparisons
Sources
Semiconductor comparisons
- All semiconductor comparisons
- SOI vs bulk silicon: what is the difference?
- Epitaxial wafer vs polished wafer: what is the difference?
- Fabless vs foundry: what is the difference?
- Tokyo Electron vs Applied Materials: what is the difference?
- EUV vs DUV lithography: what is the difference?
- High-NA EUV vs standard EUV: what is the difference?
- Single patterning vs multi-patterning: what is the difference?
Last updated: November 2, 2026