Hydrogenated silicon nitride (SiNx:H) as surface passivation layer

The passivating properties of the layer were previously correlated to cell output properties like the open-circuit voltage VOC [1, 2]. Last decade more details about the passivating properties and mechanisms behind it have been elucidated. The bulk passivating properties of a-SiNx:H are related to the release of H from the layer, which diffuses into the Si wafer where it can passivate defects and impurities. The bulk passivating quality is related to the amount of H present in a-SiNx:H, the (thermal) energy available to break H from (mostly) N-H bonds in a-SiNx:H and the density of the a-SiNx:H layer itself. During the high-temperature firing step for contact formation, H is released from a-SiNx:H and increasingly so for higher temperatures and longer times [3, 4, 5]. The atomic H diffuses through the a-SiNx:H layer and Si wafer, and bonds to defects and most common impurities (for example Fe) it encounters, thereby passivating them. Logically, layers with a high N-H density will release more H than layers with hardly any N-H. Atomic H present in a-SiNx:H can easily form H2 in case these layers also have a low total atomic density, thus an open structure, thereby effectively reducing the concentration of atomic H in the layer. Because of the open structure the formed H2 can easily effuse to the ambient [3]. The resultant low atomic H concentration in the Si wafer has an unfavorable effect on VOC of the solar cell. For high Si-N bond density (and low N-H bond density), also corresponding to a high total atomic density, the number of atomic H released from the layers is too small to passivate most of the defects in Si, thereby effectively reducing VOC [1]. An optimum was found for layers with a Si-N bond density of 1.2-1.3·1023 cm-3 [1]. The overall result on solar cell’s efficiency significantly depends on the bulk quality of the Si wafer, i.e. the number of defects and impurities to be passivated [6