The Czochralski process is a crystal-growth process used to produce a single large crystal. Today, the process has been largely adopted in the production of monocrystalline silicon. But it has other applications also. Other names of it are the Czochralski method and the Czochralski technique.
The Czochralski process is very effective in producing ingots. Ingots are a pure piece of material. In a single-crystal ingot (boule), the entire material is one single piece and has uniform crystal properties. They are the base materials in the semiconductor and photovoltaic industry.
The most common example of a single crystal material is monocrystalline silicon (mono c-Si). Silicon is a fundamental element in semiconductors. You can find them virtually in all integrated circuits, which are in your phones, computers, and other electronics.
Another major use of monocrystalline silicon is in the production of solar cells. Silicon wafers, which are sliced silicon ingots, are an indispensable part of solar cells.
We can also produce single crystals using the Bridgman–Stockbarger method.
The Czochralski process is named after the polish chemist Jan Czochralski. Born in Kcynia, Poland, in 1885, he remained one of the most influential scientists of the silicon age, but few knew about him.
After completing his education in metal chemistry, young Czochralski joined AEG, a German electrical equipment producer, in 1907. Working as an engineer at AEG in 1916, he accidentally discovered what we now call the Czochralski process.
The story goes as: One day in 1916, Czochalski was busy studying the crystallization of metals in the lab. A crucible containing molten tin was abandoned on his table to cool. Lost in his thoughts, he accidentally dipped his pen in the crucible instead of the inkpot. Realizing the mistake, Czochalski quickly pulled out the pen. And he saw something unusual. The nip of the pen had a long stripe of crystallized tin. And the discovery was made.
The discovery might be an accident, but the perception wasn’t.
Later on, Czochralski replaced the nib with a narrow capillary to initiate crystallization. He analyzed the crystal and found out it was a single crystal. In his experiments, Czochralski had produced millimeter wide crystals and 190 mm long. In 1918, he published the results in a German chemistry journal as a new method for measuring the crystallization rate.
More than 20 years after the discovery, his method was first used in 1941 to produce the first germanium solar cell and later silicon.
Many improvements have been accomplished in the process since the 1920s. However, basic science remains the same. Today, over 90% of the world’s silicon producers use the Czochralski process to produce single-crystal silicon. With advances in the process, we can grow a crystal as long as 2 m with a diameter of 400 mm that weighs over 450 kg.
Before moving forward, it will be useful to acquaint readers with some common terminologies.
- Crystal growth: It is an important process in crystallization in which newer structures unite in a specific pattern to form a crystal.
- Monocrystal: It is a crystal having unbroken crystal lattice throughout the material. Monocrystals have no grain boundaries and are free from defects. Other terms for it are single crystal and monocrystalline.
- Ingot: Ingot is a pure piece of material.
- Seed: A seed crystal is a tiny piece used to grow a larger crystal. It promotes the growth rate when brought in contact with the solution.
- Dopant: Dopants are foreign substances added to magnify the electrical properties of semiconductors.
Schematic of Czochralski process
The diagram below represents a simple industrial schematic of the Czochralski process. The description of the main elements in the diagram is as follows:
- Quartz crucible: A crucible made of quartz is a container in which silicon is melted. Quartz is a mineral and consists of SiO2. It has a melting point above 1650 °C.
- Melt: The melt is liquid silicon in the crucible.
- Heating coils: Heating coils are an induction heater used to heat silicon in the crucible. They generate heat by Joule heating—i.e., when a high current flows through a conductor, the resistance of the conductor produces heat.
- Pulling and rotating shaft: The pulling and rotating shaft is a rotating rod or wire used to lift the cylindrical monocrystalline silicon. In the figure, it is rotating anticlockwise.
- Crucible shaft: Another rotating shaft, it is affixed to the crucible. Its direction of rotation can be the reverse or the same as the pulling & rotating shaft. In the figure, it is rotating clockwise.
- Insulation: Insulation prevents heat from escaping the vessel.
- Radiation shield: Temperatures in the vessel are very high. To minimize energy losses and to avoid exposure to radiation, the radiation shield is used.
As mentioned earlier, the diagram is a simple schematic and does not include several minor details.
Working of Czochralski process
The working of the Czochralski process is very simple. Crushed high-quality polycrystals of silicon are placed in a quartz crucible. Dopant impurities, like boron, phosphorus, are also mixed with polycrystals in the right proportion. Boron will create p-type silicon; phosphorus will create n-type silicon.
Proper handling of the material is necessary. To assure the final product is a high purity, crushed silicon needs to be extra pure. And the vessel, especially the crucible, must be devoid of any unwanted particles.
The heat required to melt the materials is provided by induction coils. As the temperature of the system increases, polycrystals will start melting. The melting point of polycrystalline silicon is around 1414 °C. To have complete melting, we keep the temperatures inside the vessel slightly higher than the melting point, around 1420 to 1425 °C.
Keeping high temperatures for a while removes any tiny bubbles present in the melt. It decreases the chances of deflects during crystallization.
A seed with a known crystal orientation is lowered and partially dipped in the melt. The vertical motion of the seed is controlled by the shaft.
As the seed starts to melt, it is slowly pulled upward and at the same time rotated. The rotation of the seed and crucible increases homogeneity.
When the seed is lifted upward, it carries a small portion of the molten liquid with it due to surface tension. It causes cooling and crystallization of the liquid portion near to the seed.
The first pull is very crucible. It is such that the diameter of the growing crystal is decreased to a few mm. This method is the dash process, pioneered in 1959 by W.C. Dash. By reducing the initial diameter, the number of dislocations created in the crystal is minimized. A narrow path breaks the growth and movement of existing dislocations.
The reduced diameter region is the neck of the growing crystal. By regulating the temperature profile and pulling rate, the diameter of the crystal is gradually increased. In the process, the conical region, or shoulder, are formed—see the figure below.
Desired crystal growth
We continue to grow the crystal until the desired diameter is achieved. After the crystal has reached the desired diameter, the following growth proceeds at the constant diameter. Consequently, we will have a cylindrical ingot of a desired constant diameter.
However, there are many challenges to have a constant diameter ingot. Things become more complex as the crystal grows. Because of various parameters, like the temperature profile, the concentration of impurities, the dopant percentage, velocity fields, defects, changes with the melt depth. As a result, controlling the homogeneity of the crystal becomes complicated.
In our hands, we have five parameters to maintain the diameter and homogeneity of the monocrystal.
- Pulling rate of the crystal
- Rotation of the crystal
- Rotation of the crucible
- Heat supply to the crucible
- Applied magnetic field—Applying a magnetic field across the crucible does affect the convection flow of the melt. By studying its effects, we can suppress the undesirable convection in the melt arising from temperature fluctuations.
By manipulating the above parameters, we try to achieve a homogeneous ingot.
In practice, a produced ingot will always contain some degree of inhomogeneity. The concentration of impurities and dopants will increase from head to tail. The explanation of this is covered afterward in the article.
As we approach the tail of the crystal, we gradually increase the pulling rate of the crystal. The gradual increase in the pulling rate reduces the diameter of the crystal. Finally, the diameter becomes thin enough to detach itself from the melt. The end cone will be similar to the head cone.
The final stage is also crucial. Any recklessness can generate thermal shock due to the temperature gradient, which in turn produces dislocations. The produced dislocations can propagate to the entire crystal and damage it.
Argon gas provides an inert environment in the process. The gas is charged from beginning to end to avoid chemical reactions due to high temperatures. And it also carries silicon monoxide (SiO) and carbon monoxide (CO) gas produced during the process with it.
The vessel is also vacuumed before the start of the process to remove foreign contaminants, including oxygen.
SiO and CO
The quartz crucible used in the Czochralski process consists of SiO2. At high temperatures, it dissolves to some extent and gets mix with the melt. The melt is molten silicon (Si).
SiO2 + Si -> 2SiO
The formation of SiO increases the presence of oxygen in the melt. From the melt, a portion of SiO gets embed into the monocrystal, and some amount of it evaporates from the container.
The vessel contains several parts that have carbon in it, such as the graphite susceptor, pulling shaft, heater. Carbon impurities are also present in polysilicon. In a hot environment, carbon can potentially react with oxygen to form carbon monoxide and may dissolve into the melt and finally end up in the monocrystal.
Impurities are undesirable, and considerable presence degrades the performance of the end product. The degradation of the product is due to inhomogeneity. Impurities also interfere with the electrical and electronic properties of semiconductors.
One common impurity in silicon ingots is oxygen. Since the melt and crucible are in physical contact with each other, oxygen easily passes into the melt from the quartz crucible. The research says the presence of oxygen has both beneficial and detrimental effects on the crystal.
The detrimental aspects associated with oxygen is that it forms precipitates. Having oxygen precipitates at unwanted points hinders charge carriers. And consequently, it damages electrical properties. The positive side is the precipitation increases the mechanical strength of the crystal. An optimal concentration of oxygen can help to overcome the negative effects without compromising positive effects.
Besides oxygen, common foreign elements in silicon are aluminum, carbon, copper, iron, antimony, and arsenic. Boron and phosphorus are also present in silicon as a dopant.
Impurities in the crystal will always be present. The way to measure the distribution of foreign particles is through the segregation coefficient.
The segregation coefficient is the ratio of the concentration of an impurity in the solid to that in the melt. Both solid crystal and melt must be in thermodynamic equilibrium.
k0 = [CA]s ÷ [CA]l
Here, k0 is the segregation coefficient, [CA]s is the equilibrium concentration of impurity A in the solid, and [CA]l is the equilibrium concentration of impurity A in the liquid (or melt).
How does it affect the distribution of impurities?
When the value of the segregation coefficient is high, the impurity in the solid will be more than in the liquid. And when it is low, the liquid will have more impurity.
Most impurities have a low segregation coefficient. It means impurities prefer to stay in the melt than the crystal. It may seem good that the monocrystal will always be purer than the melt in the crucible. However, there are negative effects to that.
This prejudiced distribution of impurities causes inhomogeneity in the crystal. Let’s take the example of carbon, which is a common impurity.
In the beginning phase, carbon will have some concentration in molten silicon, which will be higher than the growing crystal. As the process advances, carbon will build up in the melt since it prefers to stay in the melt. In short, carbon will accumulate in the melt from start to end. In the final phase, the residue liquid will have the highest carbon concentration.
Because of the equilibrium, the same is reflected in the crystal. From start to end, the carbon concentration will increase in the crystal. It is true for all foreign elements that have a low segregation coefficient.
Thus, impurities always increase from head to tail in the crystal.
In the case of dopants—boron and phosphorus—the segregation coefficient is close to one. As a result, the distribution of boron and phosphorus is more homogenous. It is also one of the reasons why we choose boron and phosphorus as a common dopant.
Another important thing to note is the segregation coefficient, in practice, is not only a function of the concentration of impurities but also of the growth rate. When the growth rate of the crystal is very low, we can presume the system is in equilibrium. And the effect of the growth rate can be ignored. However, we need to account for the growth rate when it is not low.
The primary application of the Czochralski process is in the production of monocrystalline silicon. Silicon is a vital part of integrated circuits and solar panels. In the photovoltaic system, solar panels made of monocrystalline wafers give higher efficiency than polycrystalline.
Apart from silicon, the method is also used for manufacturing ingots of other elements. Some of them include germanium, gallium arsenide, palladium, gold, silver. Many gemstones and systematic crystals are also manufactured using the Czochralski process.