Repeatable accuracy means more efficiency – the key metric for success.

One of the principal issues facing the solar energy industry is the efficiency of its products; a good photovoltaic (PV) cell is currently capable of harvesting around 18% of incident solar light. While this represents vast progress for the industry, there is still a lot of room for improvement, which is key to solar’s mid- and long-term success, as cell efficiency directly affects its most critical metric: cost per watt.

Over 90% of PV cells manufactured around the world are based on silicon wafers, onto which light harvesting capacity is built using diffusion techniques. On top of this, a grid of silver energy-collecting “fingers” and bus bars is printed, typically using silkscreen processes, that routes the energy off the cell.

Herein lies a problem. Given current industry practice is to print all these features on the light-gathering side, they effectively shadow the underlying silicon substrate, rendering that real estate incapable of doing what it was designed for.

One obvious route to increased cell efficiencies – the holy grail for the solar industry – is to reduce considerably the size of the energy-collecting features. A lot of effort has gone into taking grid lines to their current widths of around 100 to 120 µm, with aims in the immediate future to get these down to 50 to 60 µm. As anyone involved in electronics manufacture knows, this is eminently possible, but it is not without its complications, not least of which is that a smaller cross-section means higher electrical resistance. The solution is to give finer lines more height, but the physical properties of printing screens, and the release properties of standard silver screen-printing pastes, effectively put a maximum height on 50 to 60 µm lines of about 15 µm, which is insufficient. One of the ways around this is to print the lines twice over, in a print-on-print process that effectively doubles grid height, and therefore its current-carrying capacity.

As we know from our experience with this procedure for the semiconductor and biomedical sectors, repeatable accuracy is the key here. Primarily because without it, this degree of fine line work simply would be impossible. Consider, too, a high-definition screen print relies on an effective gasket between the substrate and underside of the print screen, so alignment must be perfect. This is particularly true of the second pass, where the landing area – formed by the first print – is so limited that a misalignment of even 10 µm can result in printing paste flooding out, ruining not only the print, but also the entire cell. Given the fact that printing is the last process in the cell manufacturing cycle, this can prove incredibly expensive.

Repeatable accuracy in print alignment is also fundamentally important for another efficiency-boosting solution being developed by all the leading solar cell manufacturers: Selective Emitter. This process is achieved by depositing extra n-type dopant in a pattern mirroring that of the collection grid. Thus, like print-on-print, two print patterns must be closely aligned with each other, in this case to within 10 to 12 µm. The added challenge here is that as the dopant, the first pattern to be deposited, is invisible, normal vision alignment systems cannot be used to align the subsequent silver collection grid pattern to it. To enable the accurate registration of layer upon layer, typically two small fiducials, etched at the outer extremes of the cell, are used, to which both deposition processes must be precisely aligned.

A further way to increase efficiency is to move the relatively wide bus bars from the front of the cell to the rear, connecting them to the collection grid by means of metal wrap through-holes, solar’s version of plated through-holes. Our work in this area shows that this, too, relies on high print alignment accuracy and repeatability.

Even without considering these developments, repeatable accuracy is essential to PV manufacture. The fact that a solar panel is made of numerous identical interconnected cells presupposes manufacturing processes that repeat the same task in exactly the same manner and to exactly the same parameters, time and time again.

As new technologies and thinner wafers come online, the mechanical alignment systems that have until now been perfectly adequate for PV manufacture will necessarily give way to the sort of vision-assisted alignment techniques that have served the demanding semiconductor industry so well for years.

Just as important, machine stability will become increasingly critical. Print machines are typically equipped with large mechanical parts such as work tables, print heads and handling systems that may go through extensive linear or rotational excursions thousands of times a day, in excess of the sub 3 sec. beat rate that is the solar industry standard. New developments like those described here will increase, and throughputs improve repeated alignment to within just a few microns. As this happens, it is essential such masses and their movements are minimized, as they may cause the machine to vibrate during printing, compromising print accuracy and quality, or equally damaging, progressively compromise print alignment accuracies over time.

Darren Brown is business development manager, alternative energies at DEK International (dek.com); dwbrown@dek.com. This column runs periodically.