Understanding Diffusion in Silicon Crystals
A silicon crystal is essentially a fixed lattice of atoms. For the crystal to be doped through diffusion, the atoms of silicon and those of the foreign substance must be thermally excited to a level where they can freely move within the lattice. To facilitate this, the silicon wafers are subjected to high temperatures of approximately 1200 °C (2190 °F) in a diffusion furnace.
The term “diffusion” refers to the spontaneous spreading of a substance due to a concentration gradient. This article will explore the intricacies of diffusion, its impact on semiconductors, and the techniques employed to achieve effective doping.
During the diffusion process, the dopant, supplied through a carrier gas, permeates the wafer surface and gradually spreads within the silicon lattice. It’s worth noting that the dopants can be of two types: n-type or p-type. By introducing a 5-valent donor, such as phosphorus, free electrons are generated, leading to an n-type semiconductor. Conversely, a 3-valent acceptor like boron generates holes, resulting in a p-type semiconductor.
Despite the nomenclature, both n-type and p-type semiconductors maintain a balanced number of positive and negative charges. The level of doping directly influences the conductivity of the semiconductor, with higher doping concentrations leading to increased conductivity due to a higher carrier concentration.
Doping Techniques and Diffusion with an Oxide Mask
To control the diffusion and achieve precise doping patterns, a structured oxide layer serves as a mask. The oxide acts as a barrier, preventing dopants from passing through and protecting against the high temperatures involved in the process. The dopants only penetrate the wafer regions accessible through openings in the oxide layer. This controlled diffusion results in a concentration gradient of dopants in the silicon, known as “diffusion with an inexhaustible source.”
At the silicon surface, the dopant concentration is denoted as C(z = 0) due to the presence of the carrier gas. Over time, this concentration remains constant in the diffusion furnace. However, as time elapses, the concentration increases in all areas while decreasing with depth (z) within the silicon.
Due to the nature of isotropic diffusion, dopants tend to diffuse in all directions, including laterally underneath the masking oxide. This phenomenon, known as out-diffusion, causes an edge shift, leading to lateral expansion of the doped area beyond the dimensions of the oxide opening.
While thermal oxidation consumes the silicon surface, regions where oxidation follows doping steps experience a loss of doped areas. This loss primarily occurs at the silicon surface, which typically exhibits the highest dopant concentration.
Challenges of Doping through Diffusion and the Advent of Ion Implantation The aforementioned loss of doped areas at the silicon surface becomes a critical issue when multiple doping procedures with different masks are performed in series through diffusion. The repeated oxidation required for each mask consumes silicon, leading to increased step formation on the surface. Overcoming these challenges prompted the development of an alternative technique called ion implantation, which will be covered in the subsequent sections.
In conclusion, the process of diffusion plays a pivotal role in the doping of silicon crystals. By harnessing controlled thermal conditions and utilizing structured oxide masks, engineers can precisely introduce dopants into specific regions of the crystal lattice.