Can Photovoltaics Provide our Electricity?

The difference between a scientific assessment and a politician’s statements is usually that the first has numbers attached to it, and that forces the analysis to come to some form of realism. You may have heard politicians say the answer to climate change is simple: solar energy. The sun, they say, has huge amounts of energy. That is true, but so what? We cannot simply pipe it to our homes and cars.

According to a recent article by Lennon et al in Nature Sustainability the International Technology Roadmap has estimated that to get photovoltaics to replace other forms of power it needs a peak output of 60 TW by 2050. Of course, one still needs a huge battery storage system because the sun does not shine at night, and domestic electricity peaks tend to be near dawn and dusk, not in the middle of the day, but let us put that aside for the moment. Let us concentrate on the material demands of generating it. If that does not add up, what follows is immaterial because we can’t use it, at least on the required scale.

First, consider copper. From Zhang et al. 2021(Energy and Environmental Science, 14: 5587) the auxiliary systems (cables, transformers, connections in modules) require 2,800 kg/MW,  which, to get to 60 TW, requires 168 million t. That is about 20% of the estimated global reserves. Similarly, the amount of silver would be about 90,000 t, which is about 16% of the estimated known world reserves, but three times the supply available now. The estimate for silver is that 1 TW would consume between 53 – 117% of current silver production. As can be seen, 60 TW will be a problem. Indium usage tends to be 50% higher than that of silver, and there are some indications it could be even higher. Global reserves of indium could be as low as 2.7% those of silver. The most optimistic estimate for bismuth usage is that 1 TW would consume 50% of the global bismuth supply. On top of that, you may ask why is the global supply so large? That is because these metals are currently used for other things as well as PV modules, and the other uses are increasing in sales volume. Thus the touch screens on your mobile phones rely on indium. Further, although more indium and bismuth are used in these PV modules, bismuth has only about 2/3 the global reserves of silver. We need more of these elements and there is much less available. The total resource level is not that great, and when we have mined those resources, what then? Anyone who says, “Recycle them,” should be asked how they propose to do that. Thus a given mobile phone has tiny amounts of indium, and of a large number of other elements. Separating them all will be extremely difficult, but when the known resources are gone, now what?

However, the problem does not stop there. It is one thing to have, say, silver sulphide dispersed through various rocks, and another to having silver in a form ready for use in a photovoltaic.

Not only that, but there is material not directly involved in electricity generation. Thus aluminium is used in mountings, frames, inverters and in many other energy technologies. Now refining aluminium is rather energy intensive. There are two main steps: refining bauxite into alumina, then electrolysing the alumina. A tonne of aluminium ingot requires about 63 GJ of energy to make. Just for photovoltaics we need an extra 486 Mt of aluminium, which requires 30.6 quadrillion Joules. This is a huge amount of energy, so a lot of fossil fuel will have to be burned with the corresponding effect on climate change. We can cut this back by using recycled aluminium, but the recycled aluminium is currently being used. Unless there is a surplus of recycled material or potentially recyclable material, recycling adds nothing because the uses it is taken from will have to use virgin material.

We can have substitution. Replacing aluminium with steel reduces the energy demand to make the metal, but increases the loss due to corrosion, and because it is heavier, increases transport greenhouse gas emissions. It is possible to reduce demands by making things lighter, but there is limited scope here because simple costs have led to most of these cherries already having been picked.

On top of that, we have ignored another elephant in the room. Silicon comes from silica, which is very inert. There is no shortage of silica and rocks are made of that bound to metal oxides. However, the making of silicon is very energy intensive. To make high grade silicon we need 1 – 1.5 GJ of energy per tonne of silicon. We need 13 t of silicon per MW, so 60 TW of energy requires 780 billion t of silicon, or a minimum of another 780 quadrillion J of energy. We shall make a lot of greenhouse gases making these collectors.