 We like lots of blue skies because blue skies means lots of sunlight but we also are going to need blue skies to think and produce new science because the future of solar is not settled. We haven't plateaued, there are fantastic opportunities and I want to talk about that this afternoon. As a backdrop conventional solar which means silicon solar cells has been doing very well. If you look at the rate of increase of installation of solar around the globe it's fantastic, it's growing exponentially. It's still only 1% or thereabouts of electricity but if it carries on exponentially it will be doing pretty well and it's growing at that rate because it is actually economical today in some parts of the world. And the real reason why we have to make solar work is because it's the only really abundant source of energy. That red blob is the land mass of the USA that would produce all the USA's energy. Just means saying goodbye to Kansas and you can have your views whether that's a good thing to do or not. Now silicon, I've said it's wonderful. By another measure it isn't. It takes a year or two for a silicon solar panel to generate the energy that was required to make it and put it in place. The year or two isn't bad. It's rather better than the payback time on nuclear generators. But if you compare that with green plants, here it's corn, a corn plant probably has to get its payback time measured in a couple of days every time it puts a green leaf out to get the energy to produce the next green leaf. Days, not years. That is a glimpse of what we should be able to do in the future. I want to go back to something near term and ask, can we make solar cells more efficient? Now here's an analogy. Photons are a bit like trees. They come in different heights or energies. And the homework for a standard solar cell is I want to chop the trees or photons to be the same height. And I want to optimise the total length of tree trunk and I have a trade-off. I can chuck away some trees that are too short and I chuck away a lot of the trees that are nice and tall. Yellow equals what we use in a typical solar cell today if it's silicon. Hugely wasteful, we throw two-thirds of the energy away. Let me make trees multi-coloured and then our photons. Photons, height is now energy in the photon. So we have ultraviolet, then we run through the colours of the rainbow and then down into infrared. And if we have a silicon solar cell, the infrared are rather colourless leaves there, sort of autumn or fall leaves. Silicon solar cells are more relaxed as you can see on the screen. Two-thirds of the energy or more is chucked away either because we don't absorb the photon or because the photons have more energy that is used in the solar cell and that's the waste heat. So lots of approaches may get a more effective use of the potential energy that we have in photons when they land on the cell. We're interested in splitting photons into two when they have enough energy that we can get two, which we could then feed through to a silicon solar cell underneath. So if we do that we can get almost half as much efficiency out of the solar panel if all that works. So what we want to do is to produce a coating that can sit on top of a standard silicon solar cell that will absorb the visible high energy photons, split them into two infrared photons and push them down into the silicon below. Some fundamental science that you will not find in a textbook on solar technology turns out to make that possible. There are some rather particular molecules which, when they absorb light, will create the initial quantum excitation splits into two in a 10 million millinth of a second. It's very efficient because it goes so fast and those two are down in the infrared and we're trying to work out how to coax those to become light emitting. That will make a difference near term. Long term, how do we really approach the materials efficiency that you have in green plants? Well we're not going to do it by slicing up crystals or bulls of silicon. We're going to have to go to material technologies that just use less material and actually printing, roll-to-roll printing, looks pretty good. If you look around at what's going on at research and demonstration level, there's some fantastic advances in getting relatively efficient still conventional solar cells most recently with late-night perovskites where when you've finished, you've just used much less stuff and that will in a few generations or a generation maybe make a big difference. I'll leave you with a slide which shows a solar installation in Africa. A combination of today's silicon solar cells, lithium-ion batteries and LED lights produces lighting and produces electricity for phone charging that is half the cost of the incumbent technology which is essentially burning kerosene for kerosene lamps. The developing world is going to embrace solar even if the developed world doesn't. It may well be that it's there that big changes happen, first of all.