Patent Granted: Making Solar Cells You Can Print

I got an email last Tuesday with the subject line “WO2013050373 - Status Update.” I nearly deleted it. It looked like spam. Turns out it was the notification that our patent had been granted. Two years of lab work, distilled into a PDF full of legal language that I barely recognise as related to anything I actually did.

But there it is. I’m a named inventor on a patent for dye-sensitised solar cells. And the story of how that happened is really the story of SolarPrint, which was one of the more intense experiences of my twenties.

What SolarPrint Was Trying to Do

The short version: print solar cells like you print newspapers.

The longer version: conventional silicon solar panels are expensive to manufacture. You need clean rooms, high temperatures, vacuum deposition. The raw materials aren’t that costly, but the process is. Third-generation solar technology was meant to change that. Dye-sensitised solar cells (DSSCs) use a thin layer of titanium dioxide coated with a photosensitive dye, sandwiched between two conductive glass plates with an electrolyte in between. The materials are cheap. The manufacturing process can, in theory, be done with printing equipment.

SolarPrint raised $3.2 million to try to make that theory into a product. I joined as one of the early employees, fresh out of my undergraduate degree, and found myself in a lab in Glasnevin trying to make solar cells that actually worked reliably.

The Reality of Early-Stage Hardware Startups

There’s a particular type of optimism that exists in a startup that’s raised money but hasn’t shipped a product yet. Everyone believes. The investors believe. The team believes. The university partners believe. The press definitely believes, because “Irish startup to revolutionise solar energy” is a great headline.

What nobody tells you is how much of early-stage hardware R&D is just repetition. I spent months varying the thickness of TiO2 layers by fractions of a micrometre, measuring the IV curves, recording the results, and doing it again. My lab notebook from that period is hundreds of pages of nearly identical entries. “Sample 247: 8.2 micron layer, N719 dye, 12h soak. Efficiency: 3.1%. Slight delamination on edge.”

The efficiency numbers were always the thing. Conventional silicon panels were hitting 15-20% efficiency. Academic DSSCs in perfect lab conditions might get 11-12%. Our cells were in the 3-5% range for most of my time there, which sounds terrible but was actually reasonable for a manufacturing-focused approach. The point wasn’t to beat silicon on efficiency. The point was to be cheap enough that lower efficiency didn’t matter.

(This is the bit that journalists never quite got right. Every article compared our efficiency to silicon and declared it poor. That’s like comparing a bicycle’s top speed to a car’s and concluding bicycles are useless. Different tool, different economics.)

The Patent Itself

Patent WO2013050373 covers a specific method for assembling dye-sensitised solar cells. I can’t go into enormous detail because, well, it’s a patent and the legal language is precise for a reason. But broadly: it relates to how you seal the electrolyte layer and control the internal spacing of the cell components during manufacturing.

The problem we were solving was consistency. You can make a brilliant DSSC in a lab, by hand, with tweezers and a steady hand. Making a thousand identical ones on a production line is a completely different challenge. The electrolyte layer needs to be uniform. The spacing between the electrodes needs to be exact. If it varies by even 20 micrometres across the cell, you get hot spots and efficiency drops.

Our contribution was a method that made this more controllable at scale. It’s not glamorous. It’s a manufacturing process patent, not a “we invented a new type of solar cell” patent. But that’s where the real problems were.

Being an Inventor at 23

I want to be honest about this: being named on a patent as a junior engineer is partly about being in the right place. I did real work. I ran experiments, collected data, and contributed ideas in the lab meetings where we figured out the approach. But the senior researchers did the heavy conceptual lifting. The patent system names everyone who contributed to the inventive step. I contributed. I’m proud of it. But I’m not going to pretend I was the primary brain behind it.

The patent process itself was bewildering. Patent attorneys speak a language that’s adjacent to English but not quite the same. I remember sitting in a meeting where the attorney explained that we needed to describe our invention in terms broad enough to be useful but narrow enough to be defensible, and thinking “this is the opposite of how I was taught to write in university.” Scientific papers want precision. Patents want strategic ambiguity within defined boundaries.

The whole process took about 18 months from initial filing to grant. It cost the company something in the region of €15,000-20,000, including attorney fees and filing costs across jurisdictions. For a startup with $3.2M in funding, that’s not trivial.

What Happened to SolarPrint

I’ll save the full story for another time. The short version is that the company didn’t make it. Third-gen solar technology hit a wall of manufacturing challenges that the entire industry struggled with, not just us. The timing wasn’t right. The funding runway wasn’t long enough.

But the technology was real. The cells worked. The patent is granted. And I learned more in eighteen months at SolarPrint than I did in four years of undergraduate study. About materials science, yes, but also about what it means to try to turn lab research into a product. The gap between “it works on the bench” and “it works at scale” is where most hardware startups go to die.

I’ve got a solar cell sample on my desk at home. It’s about 10cm square, slightly yellowed from the electrolyte degrading over time. It generates approximately nothing now. But it’s still my favourite thing I’ve ever made.