Mechanism Investigation Of Etching In Alkaline Fluoride Solutions For Monocrystalline Silicon Solar Cells

Solar cells industry began to evolve for decades and crystalline silicon cells are the market mainstream, whose trend is high-efficiency and low-cost. Etching technology plays an important role in the increase of cell conversion efficiency whereas several questions such as initial growth of pyramid and etching anisotropy aren’t answered clearly. In order to understand etching mechanism thoroughly, we proposed NaOH/NH4F and NaOH/NH4F/Na2CO3etching solutions and made a theoretical and experimental investigation in detail by comparison with pure NaOH. Owing to experimental result such as morphology, etch rate, Raman spectra and so on, etching-condensation mechanism and its atomistic model are put forward. And alteration of morphology and etch rate as well as some phenomena concerning other etching solutions are clearly explained.Si(100) surface etched in pure NaOH solution is covered with pyramids that leads to a lowest reflectance. After adding NH4F to pure NaOH solution, the shape of hillocks is similar to pyramid or ridge and reflectance is in a higher range of15-16%, but etch rate decreases rapidly with NH4F concentration and increases a little with NaOH concentration. After another reagent Na2CO3is added into the solution containing NaOH and NH4F, the ridges become so long as to look like mountain chain whereas reflectance is close to the value obtained from pure NaOH solution(the lowest experimental value is12.85%in this paper) and the etch rate is a little lower than that in NaOH/NH4F system.For a thorough investigation, Raman spectra of different etching solutions and silicon wafer surfaces were measured. It was found that condensation of Si-OH groups at the wafer surfaces and in etching solutions varied obviously as the same tendency. The degree of condensation is highest in NaOH/NH4F/Na2CO3system and it is lowest in pure NaOH solution. Moreover, no Si-F bond is formed.Based on these experimental results, etching-condensation mechanism is proposed. It is believed that the silicate products affect the Si(100) surface morphologies and silicon oxides are the masks of all hillocks including pyramids, ridges and mountain chains. In alkaline fluoride solutions, SiHxFy(OH)z is the intermediate product of silicon etching, but it isn’t stable and should be eventually transformed into silicates. There are three paths of condensation:path1(attack of a nucleophilic deprotonated silanol on neutral silicate species), path2(condensation of Si-OH groups) and path3(polymerization of Si-F and Si-OH groups). The degree of condensation are lower by path1and path2, on the contrary, it is higher by path3. And path3should occur even if silicate concentration is very low. In pure NaOH solutions, silicon oxides are generated by path1and path2, consequently the degree of condensation is so low that the masks are smallest and pyramids are formed. In NaOH/NH4F system, three paths should take place and they result in the formation of bigger masks so that infinite ridges arise. CO32-ions only increase the pyramid population when they are added into pure NaOH solution, but in alkaline fluoride solution, they accelerate condensation in company with F" ions and an increase of condensation degree leads to the appearance of mountain chain.According to concentrations of H2O and OH-ions at the silicon surface and formation of oxide masks, three different surface morphologies have been explained by etching-condensation mechanism in this paper. If variation of etch rate resulting from surface morphologies is taken into account during the etching process, we can understand more phenomena such as softened pyramids loosing their tetragonal shape with defined triangular faces and effects of various additives.An atomistic model founded on etching-condensation mechanism is put forward systematically. It describes the formation of multifarious hillocks and reflects their difficulties in the growth process. A pyramid derives from an oxidized silicon atom and is easiest to come into being because of the minimum size of oxides. A long (111) hillock needs a mask composed of many tightest oxidized silicon atoms, so it is most difficult to emerge. For a long hillock with6planes, the density of those oxidized silicon atoms is lower than that of a long (111) hillock and it is possible to appear accordingly. The hillocks obtained in NaOH/NH4F system are similar to this shape. Furthermore,(110) planes should be formed if condensation occurs at (111) planes of a pyramid during etching process. On condition that oxides are located on the edge of the pyramid or formed (110) planes are too small to be ignored owing to a very small quantity of oxides produced at (111) planes, the angles between (111) and (100) planes should be slightly lower than the expected value54.735°. When oxides at (111) planes are so plenty that (110) planes are macroscopically visible, a pyramid with8planes should appear. By etching-condensation mechanism and its atomistic model, not only the question about initiator of pyramid has been answered but also lots of phenomena during etching process, such as alteration of morphology and etch rate, could be explained. It is an important supplement to theory of silicon etching and lay the foundation for developing low-cost and high-efficiency crystalline silicon solar cells.