5.2.2 Thin-Film Silicon PV Technology II

Let's look at a typical band diagram of such
double junction or also referred to as a tandem cell. On the left the electronic band diagram of
the amorphous silicon top cell is shown and on the right the electronic band diagram of the
nanocrystalline silicon bottom cell is shown. The blue-ish and green light is absorbed in
the top cell and excite the mobile electrons and holes. The red-ish light is absorbed in the bottom
cell exciting the charge carriers. Let's consider the two electron-hole pairs
excited in the top and bottom cell. The hole generated in the amorphous top cell
moves to the p-layer. The electron excited in the bottom cell drifts
to the n-layer. Both can be collected at the front and back
contacts. The electron excited in the top cell drifts
to the n-layer and the holes generated in the nanocrystalline bottom cell
drifts to the p-layer. Similar like in the III-V multi-junction, the
electrons and holes have to recombine at a recombination tunnel junction between the
n-layer of the top cell and the p-layer of the bottom cell. Often a very thin and defect-rich layer acts
like a tunnel recombination junction here.

Here these electrons and holes recombine. Again, the total current density is equal
to that of the junction with the lowest current density. It means that in an optimized multi-junction
cell all current densities in the various sub-cells have to be matched to achieve the
best spectral utilization. Let's look at the J-V curves of a single junction
amorphous silicon solar cell and that of a single junction nanocrystalline silicon solar cell. The high band gap amorphous silicon has a
high open-circuit voltage of let's say 0.9 V and a relative low short-circuit density
of 15 mA/cm^2, whereas the low band gap material of nanocrystalline silicon has a lower open-circuit voltage of 0.5 V and a higher short-circuit current density of 25 mA/cm^2.

How would the J-V curve of
the corresponding tandem cell look like? If we would make a multi-junction of both
components, we would get a cell with an open circuit voltage equal to the sum of the open-circuit voltages of the individual single cells. The resulting current density of the double
junction is lower than the currents in both bottom cells. The total current utilization of the tandem
cell is determined by the bottom cell, i.e. 25 mA/cm^2. Given the examples of the single junction
here, the best current matching of both cells would deliver 12.5 mA/cm^2. Here you see an example of thin-film silicon
based triple junction with an a-Si:H top cell, a-SiGe:H middle cell and a nc-Si:H bottom
cell as developed at United Solar. The colored part in the spectral power density
function represents the utilization of the solar spectrum by the individual cells. Blue represents the top cell, green represents
the middle cell and red represents the bottom cell.

In this figure you see the EQE of the three
p-i-n junctions of a record triple junction of United Solar. As you can see, the individual cells have
various overlaps, they don't look like the block functions as we have seen for the III-V
technologies. The light with wavelengths below 450 nm are
utilized by the top cell only, wavelengths at 550 nm are utilized by the top cell and
middle cell. Wavelengths at 650 nm are utilized by all
three junctions. Wavelengths above 900 nm are only utilized
by the bottom cell. Consequently, optimizing thin-film silicon
multi-junction solar cells is a complex interplay between the various thicknesses and light
trapping concepts used in the solar cell. In this slide you see the most studied and
developed thin-film silicon concepts on lab-scale. The record single junction amorphous silicon
solar cell developed by Oerlikon Solar has an efficiency of 10.1%. The best single junction nanocrystalline silicon
solar cell is 10.7% as obtained by EPFL Neuchâtel in Switzerland. The best result for a micromorph double junction,
or an amorphous nanocrystalline double junction is 12.3% obtained by Oerlikon Solar.

LG in Korea has the record for the a/nc/nc
triple junction with 13.4%. United Solar achieved an initial efficiency
of 16.3% for the triple junction based on amorphous silicon, amorphous silicon germanium
and nanocrystalline silicon. However, the stability of the amorphous alloys used
in this triple junction is a big issue. The hydrogenated amorphous silicon alloys
suffer from light-induced degradation and the stable efficiency drops below that of
13.4% as achieved by the a/nc/nc triple junction. The light-induced degradation, also referred
to as the Staebler-Wronski effect, is one of the biggest challenges for the thin-film
solar cells. The recombination of light-excited charge
carriers generates some metastable defects in the absorber layers.

More bulk defects means enhanced charge carrier
recombination in the bulk, which mainly affects the performance of the amorphous solar cells. The performance of amorphous solar cells relatively
decreases with 10-15% due to light-induced degradation, and therefore people talk about
stabilized efficiencies. If the SWE could be tackled, thin-film silicon
devices could achieve efficiencies well above 16%. Another aspect of thin-film silicon solar
cells is the current matching between the various solar cells.

Textured surfaces are being used to scatter
light into the various junctions to enhance the absorption path length, which allows to
use thinner absorber layers. This becomes more important for the bottom
cell as this nanocrystalline silicon film is the thickest layer in the device and it
has to absorb the most red-ish part of the spectrum. Secondly, intermediate reflector layers are
being used as a tool to manage the light management between the cells, like indicated by the blue layers. The top junction and bottom junctions are
separated with a low reflective index material. Due to the larger refractive index mismatch
between this intermediate reflective layer and the top cell, more light is reflected
back into the top cell. This allows the amorphous top cell to be made
thinner and makes the device less sensitive to light-induced degradation. Nowadays, doped nanocrystalline silicon oxide
layers are used as intermediate reflective layers. They have the bifunctionality of generating
the built-in field over the absorber layers and to act as transparent
intermediate reflective layers. Let's go to the Dimes Lab at the Delft University
of Technology. We will show how thin-film silicon solar cells
are made on lab-scale. Before deposition the samples have to be cleaned
in a so-called ultrasonic cleaning bath.

The potential dirt and dust particles are
removed. Since the solar cell device is only several
hundreds of nanometers up to a few microns thick, a dust particle on the substrate will
generate a shunt between the front and back contact in the final solar cell. Here we use a substrate that is coming from
the Japanese ASAHI glass company, and already has a TCO coating on it. The TCO coating is a fluorine-doped tin oxide
and is responsible for the hazy color. In this movie, we deposit a thin zinc oxide
layer on top of the ASAHI substrate to protect the tin oxide from the next processing steps. The substrate is mounted on a sample holder
and put into a load lock. A load lock is a chamber in which the substrate
is brought under low pressure before it's moved into the processing chamber.

This avoids the processing chamber to be contaminated
with various unwelcome atoms and molecules present in ambient air. During sputtering, the zinc oxide target is
bombarded using an ionized noble gas like argon. The generated aluminum zinc oxide species
are sputtered into the chamber and deposited onto the substrate. Here you see the sputtering plasma. After the sputtering processing step we have
a glass plate with a tin oxide with on top a very thin zinc oxide layer. Next the samples are mounted into a different
substrate holder. This substrate holder is used during the Plasma
Enhanced Chemical Vapor Deposition (PECVD) step to deposit the various thin silicon films. Again we use a load lock. The load lock allows access to the various
chambers. Every chamber is dedicated to deposit a type
of silicon layer: p-type silicon carbide, intrinsic amorphous
or nc-Si, and n-doped a-Si or n-doped nc-Si. Typical precursor gases in deposition plasma
are silane (SiH4) diluted with hydrogen (H2). An RF bias between electroplates
generates a plasma. The silane (SiH4) is dissociated and the silicon
contained in radicals and ions are deposited on a substrate. Adding diborane (B2H6) or phosphine (PH3)
to the plasma makes the films p-type or n-type.

Increasing the hydrogen content makes the
films nanocrystalline instead of amorphous. The samples are moved from chamber to chamber
through the load lock. After the various silicon deposition steps,
the sample with the thin-film silicon p-i-n junction on top looks like this. Next, the samples are covered by a mask. The mask determines the areas on the sample
where the metal contacts are deposited. The samples are mounted into a processing
chamber. Silver is used as source material and particles
of silver are put in a boat. The silver is evaporated
using an electron beam. The silver vapor deposits on the samples and
form the back metal contacts.

After taking the solar cell out of the e-beam
evaporation setup, we can see the final silicon solar cell on lab-scale. Every metal contact represents a solar cell. The configuration of solar modules of thin-film PV technologies are different from those based on crystalline silicon based wafers. Here we will show the concept of how the solar
cell in a thin-film PV technology are processed and interconnected. We will show it here for the thin-film silicon
PV technology. In the next blocks we will discuss other thin-film technologies like CIGS and CdTe.

Similar interconnection schemes are being
used for these technologies. The solar module and its interconnection is
processed in one process sequence. The substrate carrier of the module is a large
glass substrate. On the glass plates the front TCO is deposited. Using intense lasers, lines of TCO are removed. This process is called laser scribing and
determines the area of the solar cells. On top of the TCO the various silicon layers
are deposited making the PV active part. After the silicon process step, a second laser
scribing step is made. The metal back contact is deposited, after
which the last laser scribe step is used. The whole cell is finished by covering it
with an encapsulant material. In this interconnection scheme the metal back
contact is connected with the front zinc oxide contact of the next cell. A module consists out of long strips of solar
cells, which are interconnected. Here you see a picture of a micromorph tandem
module. You can see the various solar cells and the
laser scribes.

The open-circuit voltage of the module is
determined by the number of solar cell strips that are connected in series. Note, that shading effects on this type of
module is different from that of wafer based crystalline silicon solar cells. The best module efficiencies are in the order
of 11.0% as achieved by companies like Tokyo Elektron, Panasonic and Kaneka. Another advantage of thin-film PV technology
is that you have an option to deposit it on flexible substrates. Here I give an example of the HyET Solar company,
who develops a technology that is deposited on a temporary aluminum foil. The entire solar cell is processed on the
foil and encapsulated at the backside. Then the temporary substrate is etched away,
and the frontside is encapsulated. This results in a very low weight flexible
substrate, which can be integrated in curved roof top elements. The advantage of lightweight is that a professor
can lift it quite easily.

This technology can be installed on simple
roof top constructions that have a low maximum for the ballast weight. Compare this to glass based solar panels. These can be quite heavy. Secondly, if the product is integrated into
roof elements it saves significantly on the installation costs, which has become the largest
contributor to the non-modular costs of a PV system. At the moment only thin-film silicon technologies
have demonstrated flexible modules with reasonable efficiencies on lab-scale. This was the quick introduction into thin-film silicon PV technology. In view of time limitations I did not talk
about the thin-film silicon junction crystalline solar cells. These are based on thermal crystallization
of amorphous silicon deposited on non-wafer based substrates. On lab-scale close to 12% conversion efficiencies
have been achieved with this approach as well.

In the next block I will talk about
CIGS PV technology. This is the thin-film PV technology which has
the highest achieved efficiencies on both lab-scale and module level. See you in the next block..

You May Also Like