INVESTIGATION OF THE CURRENT TRANSPORT MECHANISMS IN FIRE THROUGH DIELECTRIC CONTACT (FTDC) TO SILICON SOLAR CELLS BY SPECTROSCOPIC ANALYSES
This thesis work investigates the current transport mechanisms at the Si/gridline interface of the screen-printed commercial silicon (Si) solar cell with lowly doped emitter. The use of lightly doped emitter is one way to improve the efficiency of a Si solar cell. In addition, developing the screen-printable Ag-Cu paste to reduce the cost of using 90-95% Ag powder is crucial to reach the cost-effective solar electricity. Hence, the contacts under investigation are, respectively, Ag/Si and Ag-Cu/Si, which are formed with screen-printable pastes consisting Ag powder (and, or Cu at some wt%), glass frits and organic binders. For screen-printable contacts formed on heavily doped emitter, the contact resistance at the Ag/Si is always low but because of the shadowing and surface recombination losses, the cell efficiency is low. In this work, the Ag/Si interface on lightly doped emitter was first studied to elucidate the understanding of the transport mechanisms at the interface and then was extended to the Ag-Cu/Si interface. Optical and electrical characterization of the Ag/Si interface was carried out after the contact formation at high temperature. For the Ag/Si interface on the lightly doped emitter, a very thick interface glass layer (IGL) was measured, which should show high contact resistance according to literature. However, in this case, the IGL was found to be conductive according the conductive AFM (c-AFM) I-V curve, which fitted a barrier height of only 0.1 eV. This low barrier height stems from the formation of the semimetal nano-alloys found in the glass layer revealed by Raman Spectrographs. These nano-alloys are low-bandgap compounds, PbTe and Ag2Te, and could be responsible for the high conductivity exhibited by the thick IGL. Thus, in the presence of these semimetal alloys, the contact behaves as ohmic as seen with the lithography and buried contacts which are pure metal-semiconductor contacts. For the Ag-Cu paste contacts, the SEM showed that the glass was first doped with Cu, which led to increased glass frit transition temperature (Tg). Thus, the high glass transition temperature impeded the uniform spreading of the molten glass and resulted in poor wetting and etching of the SiNx with the agglomerated Ag crystallites found in the Si. Further analysis with STEM showed that the IGL at the Ag-Cu/Si interface acted as an effective diffusion barrier layer to prevent Cu atoms from diffusing into the Si emitter, which is the primary requirement for applying Cu in the Si solar cells. More so, the c-AFM in conjunction with the SEM and STEM analyses revealed that the growth of Ag crystallites in the Si emitter is responsible for carrier conduction in the Ag-Cu contacts.For the Ag-Cu/Si contact fired at high temperature, Cu tended to dope the glass frit and was sequestered by the resulting oxide. The higher the ratio of Cu to Ag, the thicker the formed oxide. It can therefore, be concluded that part replacement of Ag with Cu may not be the way to go but it can be the use of Cu paste with 90-95% Cu powder as in the Ag paste, if the paste can be formulated such that Cu does not oxidize in the atmosphere. Thus, future work focuses on atmospheric Cu-paste. The preliminary study with the atmospheric Cu paste showed very promising results with the Cu particles being sequestered in the oxide so it does not diffuse into Si. Open circuit voltage of 628 mV and the ideality factor of 1.17 with Jo2 of 1.1*10^-8 A/cm^2 indicate that the p-n junction is not damaged by Cu diffusion.