Is it time to move away from silicon-based solar? | Hacker News Reader

Is it time to move away from silicon-based solar?

arstechnica.com

56 pointsturing33 comments

PV "breakthroughts" are a dime dozen. Usually things quickly break apart due to these reasons:

1) The solar cell has to stable for 20 years in the field, not the 5 minutes it requires to collect the data for your publication.

2) You have to make 200 million cells per year with a yield >98%. Can you really do that?

3) Manufacturing cost on paper and in reality are two completely different things. Reducing material consumption is good, but it is not the only cost driver. Again: YIELD.

The first point is something that can be partially answered with a lot of additional research. Doable by a university, but often much more frustrating than the research they'd really want to do.

The only answer to the last two points is to try it on a large scale. And people did that for many technologies, see Nanosolar, Solyndra and many more. The problem is that so many failed, that currently nobody is willing to invest in new PV technologies.

Right now there is only silicon, CIGS thing film and CdTe thin film on the market. I doubt that a new technology is going to become relevant anytime soon as all of these technologies still have some room to breathe.

The article mentions, that 15 years passed without a new efficiency record in Si based solar cells. This is completely meaningless until that number is approached by mass manufactured cells.

64% of residential solar costs are the so-called soft costs of permitting and installation. PV's tech problems are smaller than its social/political problems. http://cleantechnica.com/2013/12/12/nrel-soft-costs-now-larg...

Yes, especially in the US. There is also a huge bureaucracy problem.

Even at the current cost of PV and the current electricity prices, residential PV generation amortises in areas where a lot of air conditioning is required. This basically applies to the entire south west of the US.

The problems rather seem to be ignorance and an unwilligness to see a house as a long term investment.

Also irrelevant because the price keeps going down. Who cares if it gets more efficient so long as it gets cheaper.

What about lower efficiency, but much cheaper panels? What is it that drives price up so much? Production capacity? Market forces?

That's exactly the direction some folks have gone.

http://en.wikipedia.org/wiki/Copper_indium_gallium_selenide_...

One of the startups, nanosolar, seems to have failed.

http://en.wikipedia.org/wiki/Nanosolar

You can actually pick up some nanosolar stuff on ebay right now.

http://www.ebay.com/itm/like/151190922913?lpid=82

I am not sure if there are any other companies still trying to make a go of this but last I checked -- a couple of years ago -- there were two or three others.

that's a damn shame about nanosolar -- they had a really cool process that made a lot of sense.

I was always under the assumption that Gallium Arsenide was the future after silicon.

What's the difference between GaAs and CIGS?

The article mentions Gallium Arsenide and that it's prohibitively expensive as a straight up replacement for Silicon. They're tinkering with the form factor to figure out the best way to focus more light on small bits of GaAs.

A number of people realized that they could lay down really thin films of CIGS materials and process the cells reel-to-reel. Imagine loading up a 10 ton roll of thin stainless steel in the morning and having 200 acres (or some crazy number) of solar cells by evening. If they had been able to sort the process technology out it would have been INCREDIBLE.

But of course the devil was in the details. This was a hardware based company so it's not terribly surprising that their R&D time went out past their funding. Look at how badly hardware Kickstarts do on average, blowing multiple "deadlines" because often-times hardware is more difficult than software. Not that it's impossible, but it's definitely unforgiving.

When it absolutely has to be exactly right the first time it's going to take a lot longer than you think, even once you account for the fact that its' going to take a lot longer than you think.

The price of non-silicon cells I think is largely dictated by the extra processing. I think the hardest thing you need to do for silicon is grow it as a single crystal, but even that isn't 100% required (but it will boost efficiency). And you've got abundant, cheap raw material.

Other types of cells have more complex growth and fabrication processes, often requiring a vacuum, tight stoichiometry/growth control, and expensive/finnicky/dangerous precursors.

One of the commenters on the article lives in rural Texas where it's too expensive to run electricity.

He lives comfortably on ~300 watts and recently bought a 100 watt panel for $130 shipped. His biggest expense? Batteries. Solar panels last, but batteries have to be constantly maintained and replaced.

Without a cheap, reliable storage medium, solar is useless or extremely expensive. All the more so at power plant scale. Maybe solar+hydro combination installations would work.

You mean in the far-from-norm situation you mentioned, is solar "useless" without storage. And even in that context, you're wrong. Many off-the-grid electricity consumers use diesel or fuel-oil based generators, there are already many solar systems that feed into such systems, backing off fuel usage when solar resources are available, then getting out of the way when it's not. No storage necessary.

In the "real-world", solar is connected to the grid, and the grid acts like the storage.

Storage becomes an actual concern when solar becomes a large fraction of the grid usage. Or in somewhat contrived situations such as an off-the-grid home using solar as the sole source of power.

Actually the grid has the same storage problem. On a small scale you can dump power into it and it is just a drop in the bucket. But on a large scale, you can't store power easily.

The most efficient alternative is to pump water up, then drop it through a turbine later. But world wide the total storage is only 3% of instantaneous generation capacity. So you can absorb/produce less than an hour's worth of electricity. Good for evening out a little fluctuation, but not a big one.

This is why natural gas is catching on. You can spin it up and down quickly, to accommodate the fact that wind and solar fluctuate quickly. By comparison both coal and nuclear require boiling a big tank of water. That's slow to heat up, and slow to cool down, so it doesn't adjust very fast as external power changes.

>This is why natural gas is catching on.

Fracking might have more to do with its sudden resurgence it than peaker plants, although granted it is probably the easiest power source to scale up and down.

Fracking makes it economical. But its ability to be a counterbalance to renewable is one of the reasons that environmentalists get behind it.

"Without a cheap, reliable storage medium, solar is useless or extremely expensive. All the more so at power plant scale."

Really? Why is it more so at power plant scale? When you're connected to the grid you have the benefit of being attached to more users, which seems like it'd increase the probability that someone somewhere could use the power. Lots of power plants - e.g. anything fossil fuel based can simply burn less fuel - can reduce their output fairly simply, so until they all bottom out it seems you wouldn't have a problem getting the energy to someone who wants it.

Where this starts to fall apart is transmission losses; there is benefit to having the consumers and producers be closer together, in terms of resistance. It's possible that high-Tc superconductors may help in this space in the coming decades.

HVDC solves the line-loss issue with current technology. It's still expensive to install new transmission lines though.

Did you read the article? The panels already got cheaper, but the cost of installing and using them didn't, and are making up a bigger and bigger part of the total cost for PV systems. At some point, making the panels even cheaper doesn't make the cost of the energy so much cheaper:

SSSSSSSSSPPPPPPPPPPPPPPPPPPPPPPPPPPPP

SSSSSSSSSPPPPPPPPP

SSSSSSSSSPPPPPP

I did read the article. The problem is, panels got cheaper, but they are nowhere near cheap enough.

Note that installation costs can be dwarfed by panel costs if you are in a third world country, where wages are lower.

I agree about the rest of the infrastructure required. Batteries, charge controllers and inverters are very expensive still. More appliances need to use DC, so that we can use less inverters.

This is how you optimise software as well. Find the function that is the bottleneck and optimise it. After doing that, some other function will be the bottle neck. Rinse and repeat.

We seem to be in the first iteration of optimisation right now (from what I understand of the article).

Eventually you hit a wall - because the other costs mount up and the efficiency point has a lower bound it can't get past. Solar is stuck with only being productive at most < 50% of the day. It's like designing a super-efficient aircraft, but one that can only fly during the middle of the day. It has to be more than twice as efficient as a regular aircraft, and even then, airlines aren't likely to touch it. because the gains wouldn't be worth it. You have to have an order-of-magnitude increase in efficiency in not only the panels, but also in storage. But the kicker is that improvements in power storage efficiency would also benefit many other power generation methods.

> But that manufacturing innovation hasn't been matched on the basic research side; it's been over a decade since the last time anyone set a new efficiency record for silicon cells.

> The material in question is gallium arsenide, which can be fashioned into solar cells with efficiencies twice those of silicon

What about this breakthrough doubling the typical solar panel efficiency to 44.7%? Isn't it based on silicon solar panels, too?

http://www.ise.fraunhofer.de/en/press-and-media/press-releas...

No, anything in that range of efficient is a multi-junction cell. Multi-junction cells use a range of layered materials, each tuned to a narrow spectrum in which they can convert more efficiently than silicon, for instance.

The material used in silicon solar cells Galium Arsenide isn't an environmental friendly material. For us to have large quantity of solar power the materials used in the solar cells can't be dirty in themselves or the environmental advantage are lost.

Here are some alternative links

Organic solar cell http://www.heliatek.com/newscenter/latest_news/neuer-weltrek...

A question in my mind is if plants is the ultimate solar cell, cheap to produce, naturally converting solar energy into biomass, sugar and potentially diesel. There are also ecoli based solar conversion.

Boing Green diesel breakthrough http://www.energypost.eu/exclusive-report-boeing-reveals-big...

Ecoli biogasoline http://cleantechnica.com/2013/09/30/kaist-researchers-produc...

As an active user of silicon-based solar in self sustained home I would say that we need a more efficient and cheap battery technology. By cheap I mean cost/cycles, current lead acid batteries offer less than 1000 cycles considering 30% usage, improving number of cycles 10 fold and usage to 100% would be great.

Is he trying to violate the laws of thermodynamics?

  The front has to let sunlight in, but then keep photons from
  escaping. ..... This takes photons from a broad area and
  funnels them into the PV chip. The other end of the U acts
  like a reflective cap, making it very hard for a photon to
  escape from the chip without being reflected back into it.

So basically a funnel with a wide part and a narrow part? And since it's narrow it's less likely for a photon to enter?

That actually doesn't work - the intensity at the narrow end is higher, and the total number of photons going in each direction is exactly the same.

One way mirrors do not, and can not, exist.

(One thing that does work is having slanted walls and lots of reflections giving many opportunities for the photon to be absorbed. But if it doesn't it will inevitably escape again.)

Somebody has already explained it in the comments in the article:

"The light coming in the from sun is pretty directional, and a parabolic setup will focus it down to a handy point. The photons being emitted by the the chip are more diffuse and so most will be reflected back down to be re-captured."

Alta Devices has been in the thin-film GaAs space for years. They've attracted nine-figure investments and posted really spectacular efficiency of ~28% in GaAs thin films. It's all very exciting, but these discoveries are not untrodden ground.

http://en.m.wikipedia.org/wiki/File:PVeff(rev131204)a.jpg

http://www.greentechmedia.com/articles/read/Sources-Alta-Dev...

There's nothing wrong with silicon PV arrays, they're easy to manufacture in high volumes and low cost, they have reasonable efficiency levels and high durability. The problem facing greater adoption of solar power continues to be the storage problem. If people want to improve deployment of solar power then they should work on that problem first and foremost.

I'd think the greater natural abundance of silicon, coupled with the lower toxicity vs an arsenic compound, would make such a move unlikely in deployments that don't require maximum power output for a given area (satellites, for example).

there are nascent technologies for refining silicon into solar grade that are like 30-50% cheaper than what is currently used, they just need to be brought up to production scale. So, to answer the title- no

gallium arsenide costs a fortune, way more than twice of silicon

Reply on news.ycombinator.com