Thin-Film Silicon PV Technology I

Thin film PV technologies are usually cheaper at the cost price per Wp than crystalline silicon technology.  Thin film silicon solar cells can also be deposited on glass and flexible substrates.  There are 2 extreme phases of the usual lattice structure: amorphous (a-Si), and nanocrystalline, also known as microcrystalline (nc-Si).  The first types of alloys are the hydrogenated ones: a-Si:H, and nc-Si:H.  This means that some of the silicon atoms in the lattice have valence electrons passivated by hydrogen, whose quantity is about 5-15%.  The 2nd types of alloys are mixed with Ge: a-Si_xGe_1-x:H, and nc-Si_xGe_1-x:H.  The 3rd types are silicon carbides: a-Si_xC_1-x:H.  The 4th types are silicon oxides: nc-Si_xO_1-x:H.  These alloys can all be doped.

a-Si has a disordered lattice compared with c-Si, where on atomic length scales (short range order), atoms are still tetrahedrally coordinated, but with slight distortions to bond angles and lengths.  On larger length scales (long range order), there are volume deficiencies consisting of vacancies, multivacancies, and nanosized voids in the lattice, and the disorder becomes clear.  The surfaces of the volume deficiencies are passivated with hydrogen for a-Si:H, but not all valence electrons can make bonds with neighbouring atoms.  Hence, these dangling bonds act as defects.

nc-Si is heterogeneous, consisting of small grains of crystalline lattice, of a few tens of nm, embedded into clumps of a-Si:H (see diagram above for the phases of thin film silicon).  Fully crystalline nc-Si is slightly less crystalline than polysilicon due to more cracks and pores.  The best nc-Si bulk for solar cells has a crystalline volume fraction of about 60%.  The band gap of nc-Si is close to c-Si, while that of a-Si is larger due to the lack of crystallinity (see diagram below).

The lack of crystallinity also cause a-Si to have a direct band gap because the electron moment is poorly defined.  Hence, the absorption of a-Si:H is better for wavelengths below its band gap (see diagram below).

SRH recombination in a-Si:H is very high, so the diffusion lengths of charge carriers is only 100-300nm.  Hence, charge carriers in a thick absorber layer cannot depend on diffusion to move, and has to have a p-i-n junction, where the i is the intrinsic a-Si absorber film which is between thin layers of p- (about 10nm) and n-doped (about 20nm) a-Si (see diagram below for the solar cell in superstrate configuration).

The doped layers of a-Si create a built-in electric field over the intrinsic layer, thereby creating a broad electronic band diagram (see diagram below).  Light excited charge carriers will drift in the absorber layer due to the electric field.  There are no majority or minority charge carriers in the intrinsic absorber layer, but holes and electrons are the majority charge carriers in the p- and n-layers respectively.  The p- and n-layers cannot be thick due to the dominant diffusion transport mechanism and the low diffusion lengths.

For p-i-n junctions, the sequence of deposition is p-layer, i-layer, and n-layer.  The TCO (transparent conductive oxide) layer can be fluorine-doped tin oxide (FTO), aluminium-doped zinc oxide, boron-doped zinc oxide, hydrogen-doped indium oxide, or tin-doped indium oxide (ITO).  The films can be processed with sputtering, low pressure CVD, MOCVD, or atmospheric pressure CVD.  P-layers with higher band gap materials such as boron-doped silicon carbide or oxide layers can be used to improve the absorption of blue light.  The metal back reflector can be either aluminium or silver, which is more expensive, but has higher back reflection.

The best stabilised efficiencies of a single junction a-Si cell is 10.1%.  The best V_oc are about 1.0eV, which is much less than the average band gap of 1.75eV, so the band gap utilisation is quite low due to SRH recombination and the broad intrinsic layer.

When nc-Si is used as intrinsic absorber layer, the spectral utilisation is better when compared with a-Si due to the lower band gap of nc-Si.  The usual thickness of nc-Si is 1-3 microns in order to utilise the spectral part from 700-950nm.  The best V_oc are about 600mV, and the record efficiency for single junction nc-Si is 10.7%.

To improve the spectral utilisation, micromorph tandems are used.  It refers to a double junction with a-Si as the top cell due to its high band gap, and nc-Si (μc-Si in the diagram above) as the bottom cell.



Reference:
5.2.1 Thin-Film Silicon PV Technology I, Delft University of Technology, https://www.youtube.com/watch?v=dFL65RIeyF4