[ben_w]

> How much coal will we need to burn in order to make all of those PV cells? How many mountain tops removed to get the raw materials?

Zero.

PV pays back it's own energy cost in a matter of months to single-digit years, even in the worst cases that's still enough to support the current exponential.

And the raw material are not found only in mountains, the main component by mass being — famously — what sand is made from.

[defrost]

Meanwhile, in real world, non-zero:

    Today, coal generates over 60% of the electricity used for global solar PV manufacturing, [...].

    This is largely because PV production is concentrated in China – mainly in the provinces of Xinjiang and Jiangsu where coal accounts for more than 75% of the annual power supply and benefits from favourable government tariffs.

that said:

    Continuous innovation led by China has halved the emissions intensity of solar PV manufacturing since 2011.

    This is the result of more efficient use of materials and energy – and greater low-carbon electricity production.

    Despite these improvements, absolute carbon dioxide (CO2) emissions from solar PV manufacturing have almost quadrupled worldwide since 2011 as production in China has expanded. 

and there's a bit of a bottleneck:

    Based on manufacturing capacity under construction, China’s share of global polysilicon, ingot and wafer production will soon reach almost 95%.

    Today, China’s Xinjiang province accounts for 40% global polysilicon manufacturing. Moreover, one out of every seven panels produced worldwide is manufactured by a single facility.

    This level of concentration in any global supply chain would represent a considerable vulnerability; solar PV is no exception. 

We're talking billions of tonnes of raw materials here to meet decadal global demands, and it simply isn't just sand (and remember that really good sand is a resource in demand also):

    Solar PV’s demand for critical minerals will increase rapidly in a pathway to net zero emissions.

    The production of many key minerals used in PV is highly concentrated, with China playing a dominant role. 

    Despite improvements in using materials more efficiently, the PV industry’s demand for minerals is set to expand significantly.

    In the IEA’s Roadmap to Net Zero Emissions by 2050, for instance, demand for silver for solar PV manufacturing in 2030 could exceed 30% of total global silver production in 2020 – up from about 10% today.

    This rapid growth, combined with long lead times for mining projects, increases the risk of supply and demand mismatches, which can lead to cost increases and supply shortages.

https://www.iea.org/reports/solar-pv-global-supply-chains/ex...

[ZeroGravitas]

> Despite these improvements, absolute carbon dioxide (CO2) emissions from solar PV manufacturing have almost quadrupled worldwide since 2011 as production in China has expanded.

That would be bad except:

a) production has increased by more than 10x in the same time period.

b) solar panels added to the energy mix pull down the average carbon and quickly pay back their manufacture

which means it's just a sensible investment in an incredibly low carbon and cheap energy source which has gotten even more incredibly low carbon and cheap over time.

[ben_w]

> coal accounts for more than 75% of the annual power supply and benefits from favourable government tariffs.

That's a choice, not a need.

The need is right by that:

> solar panels only need to operate for 4-8 months to offset their manufacturing emissions.

> We're talking billions of tonnes of raw materials here to meet decadal global demands, and it simply isn't just sand (and remember that really good sand is a resource in demand also):

1) Doing nothing leads to burning around 8 billion tons of coal per year just by itself.

PV, even when made from coal power, reduces that by a factor of 40-90.

And, as you do make and connect it, the fraction of power coming from coal constantly decreases anyway.

2) I said "main component by mass", not "just". My point stands.

3) You don't need "good quality" sand for PV. Crush some rocks if you like, silicates are everywhere.

[defrost]

> 1) Doing nothing leads to burning around 8 billion tons of coal per year just by itself.

Who's advocating doing nothing, is that something I said?

2) I said "main component by mass", not "just". My point stands.

You clearly stated "Zero". That's incorrect. The energy demands of mining are not insignificant by any means.

There are large amounts of material being mined, both sand, and silver, and others to support PV

3) You don't need "good quality" sand for PV. Crush some rocks if you like, silicates are everywhere.

You've not ever mined anything or worked in geology, have you?

[ben_w]

> Who's advocating doing nothing, is that something I said?

I mixed you up with the other poster, but your comment and theirs together very much pattern-matches to such a position, yes.

> You clearly stated "Zero". That's incorrect. The energy demands of mining are not insignificant by any means.

I said zero in the context of "how much coal and how many mountain tops are needed".

This remains correct.

Zero mountains need to be levelled, zero coal needs to be used.

And what do you mean by "insignificant"? Your own citation is saying 4-8 months to repay their own energy cost, for devices which last 25-30 years. I think 1.1-2.6% of their lifetime output counts as "insignificant" in proportional terms, even though that's a big number when you multiply 2 TW by 30 years to find out what it takes to scale to the current global electricity demand.

> You've not ever mined anything or worked in geology, have you?

Have you?

Silicon is the second most abundant element in earth's crust after oxygen.

The doping agents are less common, but also you need far less of them.

Again, no mountains need apply — even for the single most important element, the scale needed is a big hill, not even a small mountain.

[pfdietz]

It's my understanding that reduction of silica to metallurgical silicon, which is done in an arc furnace using carbon as the reductant:

SiO2 + 2C --> Si + 2 CO

is best done with charcoal, not coal, due to the porous microstructure of charcoal more effectively interacting with silicon monoxide vapor. So not only is coal not needed, it's not even the best feedstock for this process.