Photovoltaic cells can be divided based on the materials they are made of and the structure of those materials ( Fig. 1 ).

Among silicon cells, the types shown in Fig. 2.

Commercially, silicon-based cells are the most common. Four types of cells used to build photovoltaic panels are produced: monocrystalline, polycrystalline, thin-film, and amorphous cells. A cross section through the structure of a monocrystalline solar cell is shown in Fig. 3. The cell structure consists of a front electrode (the one on which the light falls), an n-type layer about 2 \( \mu m \) thick, n-p junctions, a p-type layer, and a back electrode. The p-type layer is characterized by a thickness of 300 \( \mu m \), but the work reducing its thickness is ongoing and it will soon reach 160 \( \mu m \). The exciton produced by the absorption of the radiation quantum appears in the vicinity of the n-p junction and breaks into charges. The electric field at the p-n junction shifts the charges in different directions, electrons to the n-type semiconductor and holes to the p-type semiconductor. The charges separated at the junction result in a potential difference.

Photovoltaic cells built on monocrystals are carved from a single crystal of silicon, created by pulling a single crystal, usually by the Czochralski method. By adjusting the pulling speed and rotation of the single crystal, monocrystalline rods of a specific diameter can be obtained. The cell is made up of the following layers:

• the front electrode, on the side of the incident sunlight - it can be made of silver or ITO strips, or aluminium strips,
• an anti-reflective layer - it increases the amount of solar radiation reaching the cell,
• a passivation layer - it protects the semiconductor from changing properties over time,
• an n-type layer - up to 2 \( \mu m \) thick, it is usually etched to form a surface consisting of pyramids, which increases the optical path length of radiation quanta in the semiconductor, increasing the probability of radiation absorption,
• a p-n junction layer - intended to separate charges,
• a p-type layer - 180-300 \( \mu m \) thick,
• a back electrode (from the side opposite to incident sunlight) - made as a solid layer of aluminium (or silver), or in the form of strips, it may also be made of ITO, in the case of double-sided electrodes.

Silicon cell manufacturing technology is based on making suitable silicon wafers from a monocrystalline cylinder. An acceptor dopant is introduced into the material base during crystal growth, thus obtaining a p-type semiconductor. The grown cylindrical crystal is then cut with a laser into wafers 0.3 mm thick, in the shape of a square, hexagon or circle. Such wafers are p-type semiconductors, and the resulting wafers are polished to perfect purity and smoothness [1]. In single wafers, an n-type region is produced in a thin surface layer by diffusion, such as phosphor. At the junction of two types of semiconductors linked in this way, a p-n junction is formed. The surface is then textured. Cell wafers are typically 100-200 mm squares with a thickness of 200-300 \( \mu m \) (160 \( \mu m \) thickness is expected to be the future standard) [2]. In electronics, wafers with a diameter of 100-300 mm are used, and soon even those of 450 mm [3]. Silicon wafers thick up to 180 \( \mu m \) are currently used. The surface of crystalline silicon reflects incident sunlight (up to \( 40\% \)). To prevent this, a thin anti-reflection layer is applied to the wafer surface. Further fabrication involves applying current paths from thin strips of aluminium foil and protecting the entire cell from weathering with a special layer of EVA (Ethylene Vinyl Acid) organic film. Thanks to such hermetic structure, cells can work in year-round installations for over 25 years.

Polycrystalline silicon rods, on the other hand, are produced by the Siemens process (from 1953) with a purity of > \( 99.99999\% \). The disadvantage of this technology is the use of high temperatures. The basic material is silicon, ground and cast in a cuboidal form. Through controlled heating and cooling, the block crystallizes in one direction to produce inhomogeneous crystals several millimetres to several centimetres in size. The boundaries between the crystals represent defects that can degrade the efficiency of the photovoltaic cell. The polycrystal fabrication process forces a ribbon polycrystalline structure. A photovoltaic cell fabricated on polycrystalline silicon wafers can be compared to more monocrystalline cells connected in parallel.

Amorphous silicon a-Si is used in the production of photovoltaic cells, LCDs, or OLEDs. In a vacuum chamber, gases ( \( SiH_{4} \) with dopants) are decomposed in a glow discharge and an amorphous silicon layer is deposited on the substrate. With this technology, the production of a-Si is simpler, as well as energy and material efficient. Moreover, it allows obtaining cells with a large surface area (they are very cheap). The disadvantage is the low efficiency, up to \( 12\% \). The mechanism of photovoltaic cells is the same, regardless of the materials used to make them.
Fig. 4 shows cells made of (a) monocrystalline, (b) polycrystalline, and (c) amorphous silicon. The monocrystalline cell has truncated corners and a black color, on the polycrystalline one can clearly see the crystal areas, and for the amorphous the color is dark maroon to black.

Fig. 5 shows the main steps of the screen-printed solar cell manufacturing process. With more or less minor modifications, this process is now used by most photovoltaic cell manufacturers [4]. The main advantages of this 35-year-old photovoltaic technology are easy automation, reliability, good material utilization and high efficiency.

Here is a typical photovoltaic cell fabrication scheme using 180-210 \( \mu m \) thick silicon wafers, provided as an example. The Si wafer is doped with boron to obtain a p-type semiconductor with resistivity in the range of 0.5-6 \( \Omega \) cm. First, the wafer is cleaned of cutting impurities. The wafer thus prepared is etched in KOH to obtain a surface texture in the form of microscopic pyramids ( Fig. 5 ). Their size has to be optimized, too small leads to light reflection and too large makes it difficult to attach the electrodes, i.e. to receive the charges. Texturing can be done in several ways: alkaline etching, acid etching, plasma etching and mechanical etching. The next step is doping ( Fig. 5 ) usually with phosphorus. This is a process that requires high temperature. This creates an area of n-type semiconductor around the entire wafer, i.e. at the edges it needs to be removed or separated.

Titanium dioxide ( \( TiO_{2} \)) is applied to the surface on which light falls, and is used to form an antireflective coating because of its good antireflective properties, especially for encapsulated cells [5]. This process can be easily automated in a conveyor belt reactor. Other possibilities include, for example, screen printing of suitable pastes. Front electrode metallization is applied to such a prepared surface and is characterized by good adhesion at the silicon interface, low line width, good mechanical adhesion, solderability, and compatibility with encapsulating materials. Resistivity, price and availability make silver an ideal choice for a contact metal. A screen-printing method is used to make the back electrode using aluminium paste, forming a layer on the back surface of the cell [6]. The low eutectic temperature of the Al-Si system (577 \( _{}^{o}\textrm{C} \)) means that some silicon dissolves in Al and recrystallizes upon cooling after the firing step, forming a p-type layer. The characteristics of this layer (thickness, uniformity, reflectivity) depend on the amount of paste.
The finished cells are measured under STC conditions, i.e., a light spectrum consistent with that of Sun AM1.5 and power of 1000 \( \frac{W}{m^{2}} \) at a temperature of 25 \( _{}^{o}\textrm{C} \). Defective links are eliminated and the rest are passed on for further production.