In 2007, Q-Cells edged ahead of Sharp, and produced 389.2 MWp of monocrystalline and polycrystalline solar cells. This was an increase of 54% over 2006. But despite this, Q-Cells is not resting on its laurels.
For a start, its lead over Sharp is precarious. The two companies are neck-and-neck in terms of actual production volumes – and according to Navigant Consulting, Sharp retained a slight lead in terms of volume shipped: 363 MWp against Q-Cells' 344.1 MWp. Both have been building considerable new production capacity, with more due to come onstream in the near future, so the race is far from over.
But Q-Cells founder and CEO Anton Milner is less concerned with world domination than with cutting costs.
“Cost is the most important challenge for the solar industry today. As an industry, we already know how to realise cost savings of 40%-50%. Of these, 50% can be achieved through technological development, 25% through economies of scale and 25% via productivity gains,” he says.
The solar PV sector surged forward in 2007, attracting record levels of investment in technology, manufacturing and capacity installation (US$28.6 billion, according to New Energy Finance), but in general solar PV generation capacity thrives only where capital subsidies and feed-in-tariffs (FiTs) are there to support it. Even with oil prices at US$120/barrel, solar PV is still not competitive with conventional energy in the majority of markets, and is only just approaching grid parity where PV has been actively promoted and financially supported, as in California. In Germany, only the generous FiT makes PV economically viable.
However, this will change, believes Milner: “Grid parity is coming – you can't turn solar back any more,” he says. His goal is obviously to make this happen as quickly as possible. Wisely, though, he refuses to be drawn on just when parity will arrive, partly because this depends on which country you look at. In California, for example, PV at US$0.37c/kWh already competes with higher-tariff electricity at US$0.34c/kWh – and solar PV is also approaching grid parity in Italy and Southern Japan. Most other countries though still lag behind.
A question of cost
While electricity costs continue to rise worldwide, the solar PV industry's job is to keep cutting its costs so as to reach a crossing point as soon as possible. Technological advances and production scale-up all help, but persisting silicon shortages make cost-cutting an uphill struggle across the industry. It is notoriously hard to find out what silicon costs in the solar industry – cell manufacturers play their cards extremely close to their chest, and keep their supply contracts deliberately opaque, revealing volumes but rarely prices. But Milner mentions spot silicon prices as high as US$400/tonne, which he describes as “ridiculous”. When the supply bottleneck ease, he believes prices will fall to a more reasonable long-run price of around US$30-40/tonne.
“Silicon is a commodity and has a known production costs, but it will take a while for prices to fall this far. Solar PV silicon is supported by very strong demand and price elasticity – if supply increases, prices fall, in turn creating higher demand,” explains Milner, who is confident that when the bottleneck eases it will not lead to a structural oversupply of silicon. “But there will be short cycles of oversupply and adjustment, and this highlights the need for the PV industry to get costs down to make sure extra demand is there when it's needed,” he adds.
Capacity, capacity, everywhere
However, Milner does not expect to see supply and demand even begin to come back into balance until late 2009. The problem is that most cell manufacturers have dramatically increased their production capacity. Sharp has the largest capacity in the world (700 MWp), while Q-Cells expanded its capacity to 516 MWp, and is planning to add a further 400 MWp within the next year, including a new 160 MWp production line in Malaysia.
But unused production capacity pushes costs up – precisely what Q-Cells (and the rest of the industry) is trying to avoid. According to Navigant, capacity utilisation in 2007 fell to an all-time low of just 57%, with Sharp operating at 52% and Q-Cells at 67%, although Milner disputes the capacity figure. “In a fast-growing industry, nameplate capacity goes up as new production lines are built, but not necessarily ramped up immediately. Our capacity utilisation is about 70–80%; it doesn't make commercial sense to run a plant at low capacity, especially as the market gets more competitive,” he points out. During 2007, Q-Cells expanded capacity from 336 MWp to 516 MWp and produced 389 MWp, close to the group's average capacity over the year.
Silicon: a question of quality
Securing silicon, therefore, is vital. Industry-wide, average utilisation is being pulled down by manufacturers who have ramped up production without enough silicon supply to run at capacity. During 2007, Q-Cells has taken bold steps to secure silicon. It recently signed one of the largest polysilicon supply deals in the industry with Chinese wafer manufacturer LDK Solar, equivalent to 43,000 tonnes of silicon (enough to make 6 GWp), rising from 1000 tonnes in 2009 to 5000 tonnes a year by 2013). But it has also – and more controversially – signed two large contracts for metallurgical silicon, defying sceptics who question its purity and performance.
Q-Cells' decision to use metallurgical silicon is motivated not only by cost and feedstock security, but also by growth potential. “It's almost impossible to secure silicon in significant quantities even for 2010–2011 – and we want to grow more quickly than that,” says Milner.
The decision was not taken lightly. “We were sceptical at first, but when we ran tests on the material we were impressed. We found that efficiencies were basically the same as pure polysilicon. You can't use metallurgical silicon for everything – monos, for example – but it can address about 50% of the world market,” Milner explains.
Metallurgical silicon plants are generally cheaper to run and quicker to develop than polysilicon plants. However, impurity levels in the feedstock are higher, and this has to be very closely managed so that cell efficiency isn't compromised. “If efficiency is more than 0.5-0.7% lower than polysilicon, then the economics don't stack up. There's a threshold where the lost yield outweighs the production cost gains of cheaper silicon. And you've got to make sure you're within it,” says Milner, who believes that only a small handful of the 20 or so metallurgical silicon producers worldwide operate above this threshold.
Q-Cells is not just dipping its toe in the water. It has contracted to take half Norwegian company Elkem's initial output with an option to increase this (equivalent to 66,800 tonnes over 10 years, enough to produce 10 GWp of cells) and has more recently signed a supply contract with Timminco subsidiary Becancour Silicon, for 410 tonnes in 2008 and 3000 tonnes in 2009. A follow-up contract could see up to 6000 tonnes a year delivered between 2010 and 2013.
These are vast quantities of raw material – and already the first shipments have started to roll in. But Q-Cells does not intend to corner the market and turn a profit by selling it on to competitors. “We will use it all in our own cells. I estimate that metallurgical silicon will account for 10% of our production in 2008 and 40%-50% in 2009. We will not supply metallurgical silicon to other cell manufacturers – we need it all for ourselves,” says Milner.
Metallurgical silicon has pushed Q-Cells up the supply chain in to wafering, partly because this was the most cost-effective solution, but also to keep technical know-how within the Group. “We're already processing the metallurgical silicon into ingots, so by doing the wafering in-house, we can realise synergies in terms of fine-tuning the wafer to our cell production process and avoiding joint costs,” says Milner.
Q-Cells has also moved sideways into thin-film technologies. Three of its thin-film technologies are now in production and four new thin-film factories are under construction and will start manufacturing in 2008. Thin-film is expected to contribute appreciably to sales for the first time in 2008, and to account for 25% of Q-Cells' total production by 2010. The Group's target is to produce 2 GW of cells by 2010 – 1.5 GW of silicon cells and 400-600 MW of thin-film.
Q-Cells has interests in a range of technologies including micromorph silicon (Sontor), cadmium telluride (Calyxo), CIGS (Solibro), crystalline silicon on glass (CSG Solar) and amorphous silicon on plastic foil (Flexcell). But Milner says it is too early to pick winners. “The answer to which technology is going to win is many years away. There are many different technology platforms, and all have very strong potential and likelihood of getting generation costs below grid parity. So all of them have a raison d'etre commercially,” says Milner.
Thin-film, in turn, is leading Q-Cells downstream. In 2007, it set up a subsidiary – Q-Cells International – to develop solar PV projects. Its first project – 3 MW of thin-film plus normal cells – was installed at Talheim at the end of 2007, and will eventually reach 7 MWp. Q-Cells International (as its name suggests) is also assessing opportunities in four countries outside Germany. Because thin-film production involves fewer intermediate steps than traditional crystalline silicon, it puts the manufacturer closer to the end-user – and Q-Cells is keen to understand and cultivate this relationship.
“I believe that in a few years' time we won't be talking about PV, in the pure sense of producing PV products. We're moving increasingly towards talking about an energy generation form. Our move into projects is to get a foot in the door, to really understand the dynamics and develop volume possibilities in that market,” explains Milner. However, the venture is not purely experimental: Q-Cells International is expected to contribute €25m in 2008 and €50m in 2009.
All this might look as though Q-Cells has embarked on a strategy of horizontal and vertical integration – but Milner is quick to rebuff this suggestion.
“Our strategy has very little to do with integration. We are organised around a core business of cell manufacturing, with technical aspects relevant to this added on. I think horizontal and vertical integration is the wrong way to look at what we're doing. Everything is about building out the business systems around the cell, and taking as much cost out of the process as possible”.
Which is just as well – because otherwise Q-Cells might have to consider changing its name.