Understanding the Solar Cell Contacts With Atmospheric Screen-printed Copper

Although Cu is very close to Ag in conductivity, there are still some concerns with its high diffusivity into Si. There are two Cu ions of focus; interstitial
$(\mathbf{cu}_{\mathbf{i}}{}^{+})$
and substitutional
$(\mathbf{Cu}_{\mathbf{s}}{}^{+})$
, whereby, with regard to Si the fast-diffusing impurity is the interstitial,
$\mathbf{Cu_{i}}^{+}$
. An isolated
$\mathbf{Cu_{i}}^{+}$
acts as a shallow donor, it reacts with impurities and defects to alter the electrical properties of the material. However, Cu passivates shallow acceptors, forms pairs with various impurities, including itself, and precipitates at defects. Thus, these Cu precipitates become strong electron-hole recombination centers. With regard to n-type Si, the Cu impurity precipitates much easier than in p-type Si, provided that several of the
$\mathbf{Cu}_{\mathbf{i}}^{\boldsymbol{+,}}\mathbf{S}$
precipitate without trapping an electron. The diffusivity of species in the semiconductor generally depends on the time and temperature, thus, a diffusivity of
$\boldsymbol{7}\mathbf{x}\boldsymbol{10^{15}}\mathbf{cm}^{\boldsymbol{-3}}$
can be inferred for Cu at
$\boldsymbol{600}^{\circ}\mathbf{C}$
for 20 minutes. For the atmospheric screen-printed Cu solar cell contacts, the sintering is performed on a PERC wafer with a finger width of 83
$\boldsymbol{\mu} \mathbf{m}$
fired at a peak a temperature of
$\boldsymbol{593}^{\circ}\mathbf{C}$
at 325 ipm for approximately 2 seconds; thus, the measured diffusion coefficient would be different. More so, since the paste consists of glass frits and Cu powder, the glass must react first with the
$\mathbf{SiN}_{\mathbf{x}}$
to produce the molten glass which would then react with Cu. Since the reaction time is very short, the Cu will not have enough time to diffuse into the Si before cooling down and subsequent sequestration by the reformed glass. STEM will be used to understand the mechanisms which enable or disable the sequestration of Cu and the associated challenges will be discussed. Additionally, the solar cell electrical output parameters comparing the results of sequestered Cu on PERC Si wafers will be presented.