Silicon, a semiconductor, has properties close to those of an insulator. Therefore, to enhance its electrical properties, impurities such as boron (B), phosphorus (P), and arsenic (As) are injected into the silicon to create n-type and p-type semiconductors. This process of injecting impurities is called doping.
Doping can be broadly divided into diffusion and implantation processes.
Types of Doping
Diffusion
The diffusion process largely proceeds in the following order: pre-deposition, activation, drive-in, and subsequent heat treatment. Because movement occurs due to differences in material concentration, it is influenced by concentration differences, temperature, and the time it takes for the material to move.
Doping via diffusion offers the advantage of allowing doping without damaging the Si crystal. However, it has the disadvantages of being difficult to control the doping concentration and the inability to precisely control the junction depth through which the impurity penetrates. Finally, because the doping direction is omnidirectional (isotropic), it is not suitable for fine patterns.
Implantation
An ion implanter is largely composed of an ion source that ionizes gases such as BF3, PH3, and AsH3 to extract the desired ions; a beam line that accelerates the ion source using a magnetic field; and an end station that injects the ions onto the substrate. Positive ions are used, and the desired ions are separated and used in the analyzer magnet of the mass spectrometer.
The ion implantation process produces precise junction depths depending on the applied energy of the ions.
Channeling
Since particles are injected into a single crystal structure, the ion junction depth varies depending on the direction of the crystal structure.
To prevent the above channeling phenomenon, an oxide film is laid on the surface during the ion implantation process to prevent the single crystal structure from being exposed (a), or the ion implantation is performed with the substrate slightly tilted (b). In addition, channeling is prevented through Pre-Amorphizing Implant (PAI), which implants Si or Ge ions into the substrate surface to make the surface amorphous before actually implanting ions (c).
When performing the ion implantation process using the tilting method, areas where ions do not enter may appear depending on the structure. This is called the shadow effect. The shadow effect can be mitigated by dividing the ion implantation amount by one-quarter and rotating the substrate by 0, 90, 180, or 270 degrees while performing the ion implantation.
Annealing
The ion implantation process damages the lattice structure of the substrate surface. Therefore, a heat treatment called annealing is performed after the doping process.
Heat treatment is performed at high temperatures (750-900 degrees Celsius) for more than 30 minutes, which can lead to excessively deep source/drain junctions. Therefore, shallow junctions are created by performing heat treatment at ultra-high temperatures (900-1100 degrees Celsius) for a short period of time, typically a few seconds. Therefore, rather than using hot-wall methods like furnaces, heat treatment is performed using cold-wall equipment like RTP (Rapid Thermal Process), which processes single wafers.
Diffusion VS ion implantation
| Diffusion | Ion implantation |
|---|---|
| – High temperature process / Hard mask such as SiO2 required – Isotropic dopant profile – Dopant concentration and bonding strength cannot be controlled independently. | – Low temperature / Processing possible with PR mask – Anisotropic dopant profile – Dopant concentration and bonding depth can be independently controlled |
References: Samsung Electronics
