Addressing HJT capex, opex in equipment supply


How would you describe HJT’s progress in the manufacturing landscape this year?

Heterojunction is slowly moving into the market, and a lot of people are looking into whether it is something that they can really use for the future. In particular for smaller startups it is quite nice because it doesn’t have all the value chains that a PERC line has, so it is much easier to get started. What we see is a small success in the market – however, the big topic is bringing down the cost. What manufacturers are saying is that with HJT, “I have better efficiency on the module, but how do I bring down the capex and bring down the opex?”

And how can you, as a production equipment supplier, contribute to that?

We looked at it from an LCOE point of view and put together a strategic roadmap about how to potentially overcome that. The first thing, reducing capex: How to do that in terms of manufacturing cost and scale. We worked on two programs. One was bringing the production of the machines into China – so German-engineered, made in China. Just recently, we successfully manufactured the first machines in our Chinese subsidiary.

The second topic was to increase the throughput of our PVD machine – the XEA|nova L. If we look at what we discussed three years ago, 6,000 wafers per hour is not good enough. It has to be scaled up to 8,000 or even up to 10,000 wafers – and that is what we did. Our engineering concept is finished and we are just now promoting the 8,000 wafers/hour tool, which can even be scaled up to 10,000. That is a good balance between the wet benches in the beginning and the PECVD and metallization line at the end.

And what about on the efficiency front?

We have shown that we have a real track record with a tier one customer, and have shown the superiority of TCO layers, processed on Von Ardenne PVD machines. This applies not only to technology, but in particular from the cost of ownership point of view.

In parallel is the topic of bringing the opex down further, for example with different target materials and increasing target utilization and things like that. That is really what drives the HJT market: Capex down, efficiency up, opex down, and that is the simple algorithm for the LCOE.

Are the machines the same size?

The XEA|nova L is a bit bigger than before, with a larger carrier. The wafer can be up to M4 format. It still fits in the space on the carrier. We use every millimeter available in the machine.

But how can you make this process three times as fast? It’s pretty incredible.

It is a good question – but if you challenge engineers, they become creative. In glass coating, we have established thin film machines for substrate widths up to 3.3 meters. By comparions, the XEA|nova L for PV applications is 1.6 meters in width. So the machine for the PV business is still around half of what we have done on the thin film side.

The reason why we did not apply even larger systems with 20,000 or more wafers per hour is the need to balance the PVD system in the overall PV production. You have the wet benches in the front and they have limitations, and you have PECVD in the middle and the metallization at the end.

How do these other process steps compare to, say, 8,000-wafer throughput?

The throughput of conventional PECVD is less than 4,000 wafers. So you need two or three PECVD and one PVD to match the HJT vacuum equipment. The upstream wet chemistry step can be done in one machine.

Does this bring down the capex?

Of course, the PECVD and PVD require more capex investment. Another advantage of a bigger machine is that you use more of the targets. The wider the magnetrons, the higher the overall utilization rate per wafer and the lower the opex.

What would you say is the outlook for HJT?

Honestly, now we are just looking to see whether there will be more investments into HJT. Let’s see what happens. We are ready to support the market – and let’s see if people can find the financing to move on with HJT.


This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: