Stable perovskite-silicon printable solar cells with 25% efficiency | Hacker News Reader

Stable perovskite-silicon printable solar cells with 25% efficiency

science.sciencemag.org

77 pointsbookofjoe13 comments

Metal halide perovskite cells have shown very high efficiencies before, but there have always been a load of problems with them that have made them unsuitable for large scale production. They are very sensitive to discrepancies in deposition thickness on the substrate. This means you can spin coat very small cells in a lab, but applications on larger substrates always gives a massive decrease in efficiency. I see this paper talks about cells with an area of 1cm^2, so presumably this problem is still not overcome.

I don’t know enough about this space to parse this article without lots of effort - what is the current state of the art for solar cell efficiency and how scalable is the manufacturing process? Is the printable part the significant advancement?

According to Wikipedia [1]:

"Perovskite materials, such as methylammonium lead halides and all-inorganic cesium lead halide, are cheap to produce and simple to manufacture."

[1] https://en.wikipedia.org/wiki/Perovskite_solar_cell

My rough understanding from skimming wikipedia articles is that silicon solar cells requires specially prepared wafers, not dissimilar to what you'd to make actual semiconductors devices, thing like CPUs and ram.

Preparing those silicon wafers is expensive, and perovskite avoids the need for them.

I can’t answer most of those questions, but here is a graph of efficiency by type and year: https://www.nrel.gov/pv/assets/images/efficiency-chart.png

NREL maintain the chart to look at when it comes to efficiencies for various processes. Perovskite cells are relatively new, not hugely efficient, but pretty cheap.

https://www.nrel.gov/pv/cell-efficiency.html

You can buy ~30%, and labs have demonstrated ~40%.

This could make multi-junction cells cheaper if they have solved some of the durability issues with perovskite.

What is the price at which you could buy ~30%? Ultimately, the useful parameter is LCOE.

Spectrolab XTJ Prime cells are 30.7% efficient (and fall to 26.7% over 15 years), but I suspect they are on the order of $10-100/W as they are space rated and the minimum order quantity is not kind. The LCOE of them is amazing if you are building something in orbit, not so much for things on the ground or air.

Similarly the Azur Space are roughly $200 for the 8x4cm cell, and they are also space rated.

What's interesting about the last ten years is previously the cells dominated the cost of panels. But they are so cheap that other costs, glass, aluminum, manufacturing, installation are significant. That starts pushing you back towards higher efficiency solar cells.

But yeah everything I've read about perovskite is it's got problems with moisture. There are reports from the field of 30 year old silicon based panels that 'work fine'. Since solar panels have high capital costs the more durable, the lower the depreciation rate the better.

Edit: Just a caution. A lot of work was done over the last 50-60 years on alternatives to Silicon for semiconductors and outside of a few niches[1] they've all failed.

[1] Examples: Gallium Arsenide for high performance RF. Silicon Carbide for high temp power transistors. Various materials for LED's.

Gallium Nitride for power supplies and >20GHz RF.

The current state of the art in watts per square cm (or meter) and dollars per watt are two very different things.

Commodity 156mm monocrystalline silicon cells can have ratings of 5.4W per cell now under STC (standard test conditions) meaning that a 72-cell panel can be rated at 390W. This is somewhere around a 20 to 22% cell efficiency.

Anything above 22% gets very costly, very quick.

There are all sorts of very expensive super high efficiency photovoltaic cell manufacturing methods. Even at 35% and above. Mostly of interest if you're building a telecom satellite and want the absolute best watts per square meter and watts per gram. Or if you get a solar cell manufacturer to donate some free cells to build an electric car to race across Australia.

Reply on news.ycombinator.com