Environmental aspects

Raw materials and land

The main environmental credentials of PV are established beyond doubt: its important contribution to reducing carbon emissions; cleanliness and silence in operation; lack of spent fuel or waste; and general public accept­ability in terms of visual impact. We have already referred to such advan­tages at various points in this book. But there are further environmental considerations as PV accelerates into multi-gigawatt annual production – can Planet Earth provide the necessary quantities of raw materials, and is there enough land available for hundreds of millions of PV modules?

We start with the issue of raw materials. One point is clear: in so far as PV’s future is based on silicon solar cells, there is no problem. Silicon is one of the commonest elements in the Earth’s crust and, almost literally, as plentiful as sand on the beach. There is no future scenario in which it could become exhausted, and fortunately it is also essentially nontoxic. This is not to say that other materials involved in the manufacture of silicon PV modules are inexhaustible or problem – free, but silicon itself seems unassailable.

The situation is not so clearcut with the major new types of solar cell dis­cussed in Chapter 2 – principally copper indium-gallium diselenide (CIGS), and cadmium telluride (CdTe). However a report published in 2004 by the highly respected National Renewable Energy Laboratory (NREL)4, a facil­ity of the US Department of Energy (DOE), was generally reassuring. The scenario considered was a rise in annual PV sales in the USA to 20GWp by 2050, and the report estimated the requirements for ‘specialty’ materials needed to make the solar cells, additional materials to imbed them in PV modules, and ‘commodity’ materials for such balance-of-system (BOS) items as roof mountings and support structures. It then compared the amounts of the various materials with current global production levels and estimated the percentage annual growth required until 2050. It concluded that the above scenario would not create problems with materials availabil­ity, although the situation could change if growth proceeded much more quickly, or if world production were to reach 100 GWp/year.

Although the picture has changed somewhat since 2004, and will no doubt change again in the coming years, some of the report’s main findings remain helpful. They are summarised in Figure 6.13 using colour-coding to denote the degree of ‘supply constraint’ in the various materials needed

Cells Si CIGS CdTe	Modules	BOS
silicon	copper	cadmium	glass	copper
silver	indium	tellurium	aluminium	aluminium
	gallium		plastics	steel
selenium		concrete

Supply constraints:

I I none

I I slight

□ medium

Figure 6.12 Effectively inexhaustible: silicon for solar cells (EPIA/Solar World).

Figure 6.13 Supply constraints on major PV materials.

to satisfy the projected 20 GWp/year demand. As we might expect, silicon solar cells are given the all-clear, with no concern about either silicon itself or the silver used to screen-print cell interconnections. CIGS cells might be slightly constrained by shortages of gallium and selenium, and more so by shortages of indium; CdTe cells by a shortage of tellurium. Moving on to PV modules and BOS components, the only slight concern is over the large amounts of glass required – not because the raw materials (more sand!) would run out, but because global production capacity would have to rise substantially to meet PV ’s demands. In conclusion, the only real concerns are for the ‘specialty’ materials indium and tellurium, and to a lesser extent gallium and selenium.

The report wisely notes a number of uncertainties, and suggests strategies and developments that would mitigate shortages of key materials. It is, of course, unlikely that CIGS and CdTe cells will continue to be made exactly on today’s pattern – for example, the active layers may become much thinner. It is probable that other types of cell currently in the research phase, or entirely new ones not yet discovered, will be in volume production by 2050; and in any case the calculations were based on the unlikely assump­tion that all the required 20 GWp would be met by one of the existing thin – film PV technologies – not including any contribution from silicon!

Yet there are some significant, and ongoing, concerns, especially about indium and tellurium. In the past few years there have been scare stories in the press, notably about indium, and its price on world markets has fluctuated wildly. To a large extent the problem has arisen because of its use in liquid crystal display (LCD) monitors and TV screens, an applica­tion consuming up to 50% of world production that could not have been foreseen 20 years ago. This surely highlights the difficulties of predicting future availability of important materials that are constantly finding new applications, many of them far removed from PV, and fading away from old ones. Who knows what the next 20 years will bring, let alone the next 40?

A general point worth making is that the solar cell materials presently seen as potential bottlenecks are byproducts of major mining operations. Indium is a byproduct of zinc extraction, tellurium (and selenium) of copper extrac­tion. In the normal course of things, the amounts of these byproducts fluctu­ate in sympathy with production levels from the main mining operations. Interestingly, indium is not an especially rare element in the Earth’s crust; it is actually about three times more plentiful than silver, but only extracted at one-sixtieth the rate, emphasising the dependence of indium volumes on the scale of zinc mining. Experts make the point that increasing scarcity of a byproduct, inevitably reflected in the market price, tends to encourage more careful processing of the parent ore. It also stimulates recycling, which has recently been satisfying up to half of the demand for indium, and the search for alternative materials.

Gallium is another material considered in the NREL report because of its use in CIGS cells. In recent years gallium has branched out in the form of gallium arsenide (GaAs) cells for space vehicles and, potentially more important from the availability viewpoint, into high-concentration terres­trial PV modules (see Sections 2.4.3.1 and 3.4). GaAs is also in line for application in other fields including a future generation of very high-speed computer chips, replacing silicon. So gallium may be moving slightly higher on the list of supply constraints.

How does the overall situation compare with the scenario painted by the NREL report in 2004? The huge recent rise in global PV production, and the promise of further rapid growth to 2020 and beyond, certainly bring into focus the report’s caveat about global thin-film production exceeding 100GWp per year. Yet thin-film still accounts for less than 10% of global production, and much of the current surge is coming from new factories in China that manufacture wafer-based silicon PV modules. So it will be many years before thin-film manufacturing overtakes crystalline silicon, and even then it will certainly not be based on a single technology. Amid the rather confusing debates about specialised PV materials there are grounds for cautious optimism, especially given the ability of the PV community to innovate and adapt. And of course there are plenty of silicon enthusiasts who can afford to watch from the sidelines, ignoring all talk about scarcity of raw materials!

We now turn to the question of land use. This has already been mentioned in Section 1.5, where we suggested that an area of land 140 x 140 km, or 20000 km2, roughly three times the size ofLondon or Paris, would be suf­ficient to accommodate 1000 GWp of PV modules. It seems that by 2020, or soon after, we may be approaching this huge total, some 50 times greater than global installed capacity in 2009, assuming PV continues its present remarkable progress. But where would the land actually come from, and would we resent it?

If 20 000 km2 sounds like a large parcel of land, consider some even larger ones: the Sahara Desert is about 850 times bigger; the Australian Outback about 200 times; and the state of Arizona about 15 times. In the USA, cities and towns cover some 700000 km2 and in many countries wide tracts of land are set aside for military uses, airports, highways, fuel pipelines, and so on. In short, if the world’s PV is sensibly spread around among the

Figure 6.14 No need for extra land: a rooftop PV array at Munich Airport (EPIA/BP Solar).

world’s nations, the landscapes seen by the vast majority of people will be virtually unchanged from those they enjoy today.

Of course this is far from the whole story, because PV can be installed on buildings. There are vast numbers of existing homes, offices, public build­ings, factories, warehouses, airports, parking lots and railway stations with suitable roofs and fagades, and we may be sure the that tomorrow’s archi­tects will be even more aware of the possibilities. BIPV will undoubtedly provide a major part of PV’s future space requirements, leaving deserts and other unproductive land to supply most of the balance. Sunshine is every­where, high and low, city and country, and at fairly predictable levels. There is absolutely no need for PV to dominate with unsightly and unwel­come ‘blots on the landscape’.