Improved Analysis of Gridline TLM Pattern Including Effect of Uncontacted Gridlines

Contact resistivity (ρ
_c
) for silicon solar cells is often measured using a pseudo-TLM pattern of equally spaced gridlines, which is cut from the cell itself. In this measurement, the uncontacted (floating) gridlines between two contacted gridlines are usually assumed to be disconnected from the silicon sheet. Analysis of such TLM data yields an approximate value of ρ
_c
, which is reasonably accurate for L
_T
/L > 1, but increasingly less accurate for L
_T
/L <; 1 (good contacts). In this article, the TLM analysis is extended to explicitly include the effect of uncontacted gridlines to provide a more accurate analysis of gridline TLM data. Three contact systems were measured and then analyzed by both the approximate and improved methods. In all cases, a commercial Ag paste was screen printed and fired through a dielectric coating (SiN
_x
/AlO
_x
or SiN
_x
) to contact a p
⁺
-poly (p-TOPCon) layer, an n
⁺
-poly (n-TOPCon) layer, or a p
⁺
-boron (p-diffused) layer. The improved TLM analysis gave ρ
_c
values of 3.51, 0.89, and 5.26 mΩ·cm
²
, respectively. Analysis of the n-TOPCon data is complicated by 2-D current flow across the tunneling oxide and into the n-substrate, with corresponding uncertainty. Expressions for converting ρ
_c
to a component of series resistance are also given. Additional calculations were carried out to determine ρ
_c
error with approximate TLM analysis. With assumed L
_T
/L values of 1, 0.5, and 0.25, the errors are 3.7%, 41%, and 280%, respectively.