High temperature stable passivating contacts

High temperature stable passivating contacts for silicon solar cells

Keywords: crystalline silicon, Passivation, high efficiency solar cells

Background

Photovoltaics has developed into a mature technology, but strengthening its share in the global energy production will require further development. A comparison of high-efficiency silicon solar cells and cells from mainstream production technology shows that the latter ones are limited by their low open-circuit voltages. In many cases this limitation can be related to recombination losses at the interfaces to the metallic contacts. Consequently, contacts with reduced interface recombination should be made compatible with manufacturing.
The HTPC group is devoted to the development of passivating contacts that are compatible with mainstream cell fabrication. This means that they should be able to tolerate process steps at high temperatures such as the one needed for sintering of the screen-printed metallization or the one needed for the formation of nickel-silicide which serves as template for electrochemically plated contacts.
The relation between open circuit voltage and surface passivation is illustrated by the four panels below for a p-type semiconductor which is the type of doping that is used in most crystalline silicon solar cells.

In darkness and thermal equilibrium, the material is characterized by a unique Fermi-level (panel a). Its position close to the valence band means that there is a large concentration of holes (majority carriers) whereas the large separation from the conduction band denotes a small concentration of electrons (minority carriers). Illumination injects additional pairs of electrons and holes (panel b). This non-equilibrium situation is generally described by the assumption of low-injection which means that the concentration of injected holes neglected with respect to the doping concentration. Thus, the Fermi-level of the holes remains unchanged. However, the concentration of injected electrons is much higher than the equilibrium concentration of minority carriers (electrons). This can be described by a second Fermi-level that applies only to the electrons. The splitting between these quasi-Fermi-levels is the maximum potential at which carriers can be extracted to the contacts and implies thus an upper limit for the open circuit voltage. Finally, recombination at interface states reduces the splitting of the quasi-Fermi-levels (panels c and d).
A key parameter for characterization is the (effective) lifetime of the injected minority carriers. It is not only measured easily, but it is also directly related to the concentration of minority carriers and thus to the quasi-Fermi-level splitting and the implied open circuit voltage.

Research highlights

Recombination at the silicon surface is suppressed either by blocking the recombination path, or by removing one type of the recombining charge carriers. The first concept is called chemical passivation and can be achieved by a monolayer of hydrogen that passivates all states in the bandgap. Alternatively, chemical passivation can be obtained with thin films amorphous silicon, silicon oxide, or silicon nitride. The second concept is called field-effect passivation and uses an electrostatic potential to repel one type of charge carriers – normally the minority carriers – from the surface.
Here, we investigate primarily silicon oxide in combination with doped layers. Since silicon oxide is an insulator it must be deposited thin enough for charge carriers to tunnel through. This limits the maximum allowed thickness of the oxide layer to ca. 1.5 nm. Subsequently, a doped layer is deposited and the whole stack is annealed. The annealing helps to “activate” the oxide passivation, it recrystallizes the deposited layer for better conductivity, and it may diffuse some of the dopants across the oxide and dope a small region of the wafer. Additionally, a wealth of other effects may happen, including changes of the residual defects in the bulk of the silicon wafer.

Passivating SiO_x😛 electron contacts

Solar cells based on p-type wafers are generally equipped with an n-type front contact. Since parasitic optical losses should be minimal at the front of the cell, we investigate passivating contacts on the basis of nanocrystalline silicon oxide doped with phosphorous (nc-SiO_x:P). The layers are deposited by plasma enhanced chemical vapour deposition (PE-CVD) from a mixture of hydrogen, silane, carbon dioxide, and phosphine. Depending on the flux of hydrogen, the film grows either fully amorphous or it nucleates small crystals of silicon which are embedded into a matrix of amorphous silicon oxide.
The left panel below shows concentration profiles of the phosphorous dopant that diffuses across the tunnelling oxide during annealing at different temperatures. The right panel illustrates that high values of the implied V_oc (determined from measurements of the effective lifetime) can be obtained above 850 °C.

Passivating SiC:B for hole contacts

A manuscript is in preparation, results will follow shortly. 

Key publications: