[throwaway5752]

That must be news about how photovoltaics could not meet global energy needs. Last I'd read up, global capacity with realistic assumptions was 3 times current energy use. Also, it's already cheaper per kWh than nuclear (fully depreciated). Also, nuclear is non-renewable. It's still flipping the bird to future generations.

Perhaps you are conflating capacity with transmission and storage?

[ars]

It costs too much energy to make a solar cell. Solar thermal at scale uses less energy for what it returns.

[throwaway5752]

See my reply to you here: https://news.ycombinator.com/item?id=8248640

Also, I don't understand your response. 1.) what is "too much" energy. 2.) what's your source on solar thermal vs PV EROEI at scale? 3.) what's the relevance of the solar thermal vs PV comparison, since every building in the US can have solar on the roof but not solar thermal. 4.) what's the relevance of your response to what I said?

[ars]

> what is "too much" energy.

More energy than it costs to make them, or at least enough energy that it's better to build something else.

> what's your source on solar thermal vs PV EROEI at scale?

Solar thermal just needs a mirror (ideally aluminium rather than glass and silver) and the rest of the plant is the same as a regular power plant. A mirror costs less than fuel I'm sure.

Solar cells need ultra pure silicon which is very very expensive to make.

> what's the relevance of the solar thermal vs PV comparison, since every building in the US can have solar on the roof

What would be the point if the energy return is not there?

> 4.) what's the relevance of your response to what I said?

I said they could meet global energy needs, just that it's not worth it. You assumed for some odd reason that I said they could not, so I figured you were asking about a comparison.

[lutusp]

> Solar cells need ultra pure silicon which is very very expensive to make.

So sayeth the poster, sitting at a computer filed with the same kind of silicon out of which panels are made. The reason solar panels are so cheap to buy, is because they're cheap to build.

> What would be the point if the energy return is not there?

Everything except evidence that this is so. If solar panels didn't pay for themselves, people wouldn't buy them.

[ljf]

A reply to ars - a 1cm piece costs 200 dollars? Because I've got a solar pocket calculator here that I'm sure I picked up for £2 and I am pretty sure wasn't subsidised. Been doing its job for a very long time too...

[ars]

Obviously. I was saying the person I replied to was completely incorrect with his example of a computer.

It was a reply from absurdity.

[ars]

> So sayeth the poster, sitting at a computer filed with the same kind of silicon out of which panels are made.

And very expensive it is indeed. A piece about 1cm square costs $200 or more. Solar panels need square meters of the stuff.

> If solar panels didn't pay for themselves, people wouldn't buy them.

Exactly my argument. Thank you. People in fact don't buy them - except when they get subsidies that artificially reduce the price.

[lutusp]

> A piece about 1cm square costs $200 or more. Solar panels need square meters of the stuff.

Excuse me, but what planet are you posting from? Your claim above suggests that a monocrystalline silicon solar panel that's of average size (120 by 54 cm), 6480 sq. cm., should cost $200 per square centimeter, or 1.3 million dollars. In fact, such a panel costs $168.50:

http://www.aliexpress.com/item/200W-100W-X-2PCS-monocrystall...

> People in fact don't buy them - except when they get subsidies that artificially reduce the price.

All your claims have been falsified using readily available evidence.

[throwaway5752]

"A piece about 1cm square costs $200 or more. Solar panels need square meters of the stuff."

And that would be a real problem if solar panels producers had to have microprocessors etched into panels, had to pay the scientists to design the processors, and had to build the plants to scale that process to industrial levels, like Intel.

Happily, they just have to buy an ingot of the stuff and cut it with a saw into a 250 micrometer by 1 meter squared (or so) piece. I will leave it as an exercise to you to verify the spot price of solar grade silicon is less than $30/kg.

So lets say a panel takes 250 micrometers * 1 meter * 1 meter * 2g/cm^3 (density of silicon) * $30 / kg (assuming spot prices and not long term contracts or internal sourcing). That's 250 cm^3 * 2g/cm^3 * $30/1000g, or $15 for a whole panel of the stuff.

[power_rated]

Even with Subsidies and everything taken out, Wind and solar Power are now cheaper in Germany than other forms of energy.

[lutusp]

> It costs too much energy to make a solar cell.

Compared to what? Solar cells last a long time, and in sunny climates generate a lot of power, far more than enough to pay back the energy cost of their construction.

> Solar thermal at scale uses less energy for what it returns.

That may be true, but don't dismiss solar panels, in particular in remote locations where there's adequate sunlight.