Environmental Engineering Reference
In-Depth Information
facture and improved mass utilisation of poly-crystalline compared to mono-
crystalline material is counteracted by lower efficiencies since the numerous crys-
tal boundaries form recombination centres that reduce the diffusion length of the
minority carriers despite of passivation measures.
Since the mid 1960's studies have been conducted to directly manufacture sili-
con wafers in the form of bands or cast or sintered plates for photovoltaic pur-
poses, thus avoiding crystal growth or casting an ingot and posterior sawing.
Within the scope of these investigations more than 20 different processes have
been investigated and tested with regard to manufacturing technology /6-17/.
However, to date only the "Edge-defined Film-fed Growth (EFG-ribbon)" process
has been applied in practice for solar cell manufacturing successfully /6-18/. By
this process hexagon silicon tubes are obtained which are subsequently cut into
bands and plates by laser. Efficiencies of solar cells made of this kind of band
material amount to about 15 %.
For the actual photovoltaic cell manufacturing from silicon wafers only a few
process steps must be performed (Fig. 6.9). As base material, poly-crystalline or
mono-crystalline wafers, usually already p-doped, are used. First, chemical etch-
ing purifies the wafer surface. Subsequently, the p-n-junction is obtained by diffu-
sion of phosphorus within the material surface aiming at building-up an n-doped
surface layer (i.e. p-doping of the silicon wafer is overcompensated by diffusion
of phosphorus atoms up to a maximum depth of 0.2 to 0.5 µm). N-doping must
then be removed from the wafer edges by plasma etching. In addition, during the
diffusion of the phosphorus a phosphorus glass builds up at the silicon wafer sur-
face, which also has to be removed prior to the following process steps.
To remove the n-doping at the rear side of the wafer, first an aluminium coat-
ing is applied by silk-screen process printing, and dried prior to applying an addi-
tional rear side metallization. Subsequent sintering ensures that aluminium atoms
diffuse into the silicon wafer from the rear side and thereby overcompensate the
undesired n-doping at the rear side of the wafer. At the end of the day as p-n-
junction only the phosphorus-diffused layer at the front side of the wafer remains.
Subsequently, the front side contact is printed on the wafer in the form of a grid
and dried afterwards. After applying an anti-reflex coating to enhance light trap-
ping a last sintering step is performed prior to conclusive electrical gauging of the
photovoltaic cell.
Virtually all mono-crystalline and poly-crystalline silicon wafer manufacturers
apply the described technology as a standard technology. Cells manufactured as
mono-crystalline silicon wafers based on the Czochralski process allow
achievement of efficiencies from 14 to 18 %, while the efficiencies of poly-
crystalline wafers stretch from approximately 13 to 15.5 %. Thus, the described
process sequence is a compromise between a simple and cost-efficient process
design and wafer materials on the one hand, and minimised efficiency losses due
to simplicity, on the other.
Search WWH ::
Custom Search