I think your comment is the first use of "breakthrough," to be fair. The article title and body refer to it as "breaking" the record. You are correct about the (vastly more expensive) multi-junction concentrating cells having higher efficiency, of course.
You're also correct that not all improvements make it to market. Without research and continuing efforts to improve, nothing makes it to market. What's the point of your second paragraph? The article makes it _very_ clear that this is not in production, and speaks specifically to concerns about whether the process is suitable for industrialization.
The point is that you see announcements like these at a rate of 4 to 5 every year and hardly any of those ever make it to market. I've been active in the renewable energy-scene off-and-on for two decades and I really can't count how many times there have been announcements like the one here that did not go anywhere. It makes me quite skeptical of announcements like these, they do not really convey much in terms of information other than to create an optimism about solar panel efficiency gains that is usually met with disappointment when someone actually costs out a system and realizes that over approximately a decade we tend to see gains on the order of 1 to 2%.
Hype simply doesn't help, at best this is 10 years away, worst you'll never hear from it again.
This paper is closer to industrialization than most and I expect production on scale in less than 10 years. It uses process steps that are already operating on an industrial scale. It's achieved on an industrial area wafer (compare with many academic records where cells are tiny, less than 10 cm^2). It's achieved on an industrial thickness wafer (some records use thinner wafers that are closer to the theoretical optimum, but those thinner wafers are too fragile to process industrially with current technology). It's just combining features of two cell concepts that are already manufactured on an industrial scale (interdigitated back contact cells and heterojunction cells). The core patents on heterojunction silicon cells expired just a few years ago so I expect more vanilla heterojunction and heterojunction-plus-other manufacturing going forward.
Well, I certainly hope you are right and I'm wrong. Still, the devil will be in the details and even with expired patents multi-junction cells are simply more expensive to produce. Keep in mind they only have to be ~24% more expensive to make to wipe out any efficiency gains and that's not a whole lot of margin. If this succeeds you might see a fraction of that passed on to the consumer and the cost of the bare panels long ago ceased to be the dominant factor in a regular (domestic) installation.
For large areas it might work out to be beneficiary earlier.
Heterojunction cells are an approach to reducing recombination losses near contacts in single junction silicon cells. Sanyo (now Panasonic) developed them in the 1990s and has shipped gigawatts of them. It's the same Panasonic cell technology that SolarCity (now Tesla) is expected to use in their New York factory. The vanilla heterojunction cell design is actually one of the simplest high-efficiency cells in terms of process flow. The main drawbacks are patents (until recently) and the limited heat tolerance of a-Si layers; standard screen printed silver metallization, fired at high temperatures, does not work for these cells.
I spent some time reading that, thank you very much for the pointers. Looks like you are right and this is in fact something that might just happen. Here's to hoping that it does.
You're also correct that not all improvements make it to market. Without research and continuing efforts to improve, nothing makes it to market. What's the point of your second paragraph? The article makes it _very_ clear that this is not in production, and speaks specifically to concerns about whether the process is suitable for industrialization.