Reducing Metallization Induced Recombination Losses in N-Type Solar Cells

VINODH CHANDRASEKARAN
Sean Lonsinger, Crystal Han | Heraeus Photovoltaics, USA

The PV industry continues to grow at a steady pace both in terms of technology and capacity. An important aspect of efficiency improvement is to reduce recombination losses to improve the open circuit voltage (V_oc) of the solar cell. Metal induced recombination is a significant component of the total recombination losses in the device. In p-type PERC cells, reduction in metal induced recombination losses is achieved by reducing the metal contact area on the rear.

An alternative approach that has seen steep progress and adoption by industry is “passivated contacts.” In one form of this approach, a thin tunnel oxide with a poly silicon layer is used as the passivation and contact layer whereby direct metal contact to the crystalline silicon substrate can be largely avoided.

This approach also has the benefit of being able to withstand high temperature processing steps including firing for contact formation, which is most common in the current PV industry. Passivating contacts (more widely known as TOPCon) can be used to improve efficiency in both n-type and p-type solar cells.

Metal induced recombination for n+Poly Si using screen printed and fired paste

Although laboratory cells using Poly-Si passivating contacts use evaporation, photolithography for contact formation, screen printing continues to be the preferred method for contact formation in industrial cells using this technology.

Significant improvements in metallization pastes have enabled the use of screen printing and firing to contact Poly-Si layers while preserving the passivation properties of this layer and inducing only very small recombination losses from metallization. J_0,met of <100 fA/cm2 using Heraeus pastes have been reported by A. Chaudhary [1] on 150nm poly-Si layers. In this paper, the effect of firing temperature on J_0,met was studied and analyzed using SEM imaging. They have reported that as the peak firing temperature is reduced, the J_0,met reduces quite significantly as shown in fig 1 and a value of ~69 fA/cm2 was measured with respectable contact resistivity.

SEM images in fig 2 reveal that as the firing temperature is reduced, the amount of poly Si etched by the screen printed paste is drastically reduced, which could be one of the reasons for the reduced J_0,met at lower firing temperatures.

This study clearly demonstrates the importance of lower firing temperatures to reduce the J_0,met and achieve the highest possible V_oc from poly-Si passivating contacts. In an n-type cell where n+ poly Si is used in the rear side of the solar cell, in order to achieve the full benefit in a finished solar cell device, the front side metallization paste used to contact the Boron emitter also needs to be able to perform optimally at the lower firing temperatures.

Boron emitter contact pastes optimized for lower temperature firing

Two development paste formulations referred to as 9380 M1 and 9380 M2, where the inorganic system was developed to support lower firing temperatures are shown in comparison to a reference paste, which is the incumbent, as tested on nPERT cells to clearly see the impact from front side paste only. In Fig 3, V_oc of cells using all three pastes perform quite similarly at the reference firing temperature. When the firing temperature is reduced to Ref-40˚C, the reference paste shows hardly any change in V_oc and drops when the temperature is reduced to Ref-55 ˚C. In contrast, the V_oc for 9380M1 increases by 2mV at the Ref-40˚C firing temperature. At even lower temperature, the V_oc is similar to that at Ref firing condition.

For 9380M2 paste, the V_oc continues to increase as a function of lower firing temperatures. With this paste, V_oc is improved by 5mV at Ref-55 ˚C.

In order for the cell performance to improve, we must ensure that the improvement in V_oc is not offset by a loss in other IV parameters, particularly FF. FF data for the three pastes as a function of firing temperature is shown in fig 4. The reference paste shows a small drop in FF at Ref-40˚C, but a significant drop in FF at even lower temperatures. 9380M2 has a trend of generally lower FF as the firing temperature is reduced, but the overall values for FF are still comparable to the Reference paste at it’s optimum firing condition.

On the other hand, 9380M1 shows an improvement in FF at Ref-40˚C firing temperature. Similar to the V_oc trend, the FF also drops at even lower temperatures, closer to that at Reference firing condition.

In order to understand the source of FF differences, series resistance and pseudo-FF data for the pastes at different firing conditions were analyzed. As it can be seen in fig 5, series resistance increases for all pastes as the firing temperature is reduced. However, it is interesting to note that the pseudo FF

increases for the development pastes 9380M1 and 9380M2 as the firing temperature is reduced. For 9380M1, again Ref-40˚C is the optimum and the peak pseudo FF achieved at this firing temperature. For 9380M2, Ref-55 ˚C is the optimum and the highest pseudo FF is achieved at this even lower firing temperature.

The trend of improving pseudo FF for the development pastes indicate that there is less damage induced by the pastes, which is also consistent with the improvement in V_oc observed for these pastes at lower firing temperatures.

The net result of efficiency is shown in fig 6 for the pastes under study as a function of firing temperature. We can see that, as the V_oc and FF are improved at Ref-40˚C for 9380M1, this results in a net improvement in efficiency, as expected. For 9380M2, while there is a loss in FF, the improvement in V_oc is quite significant that there is still a net increase in efficiency, particularly at a firing temperature that’s even lower than the optimum temperature for 9380M1. For both developmental pastes 9380M1 and 9380M2, the improvement in efficiency is ~0.15%. These pastes not only enable lower temperature firing but also provide the ability to optimize the processing and firing conditions for the Boron emitter contacting paste (SOL9380) that are also suitable for the back side passivated contact (SOL7200) to achieve the highest efficiency.

References: [1] “Screen Printed Ag Contacts for n-Type Poly Silicon Passivated Contacts”, Aditya Chaudhary, Jan Hoß, Jan Lossen, R.A.C.M.M. van Swaaij, Miro Zeman; Silicon PV 2019.