Australia's renewable energy future

Solar photovoltaics: Power source for the future?

Tuesday, 7 October 2008

Professor Martin Green FAA
ARC Federation Fellow, Scientia Professor and
Research Director of the Photovoltaic Centre of Excellence
University of New South Wales

Martin Green and his group's contributions to photovoltaics include development of the world's highest efficiency silicon solar cells and commercialisation of several cell technologies. He is the author of six books on solar cells and numerous papers. He has received several international awards including the 2002 Right Livelihood Award, commonly known as the Alternative Nobel Prize, and the 2007 SolarWorld Einstein Award.

(Click on image for a larger version)

It is a real pleasure to be here tonight speaking on a topic that is so near and dear to me: solar photovoltaics. As Mike mentioned, I am with the Photovoltaics Centre of Excellence at the University of New South Wales. I am sure I do not have to explain to this audience what a photovoltaic module is. But, for any who have not had contact with the technology before, this is a typical commercial module shown here [indicating photovoltaic module on slide] and these are the individual cells made from silicon wafers [indicating wafers within module on slide]; they are pretty much the same as those used in microelectronics. When sunlight falls on this module, it just acts like a chemical battery. You get electrical power out of it for as long as the sun shines on it. Unlike a battery, it never wears out, as long as the sun is shining. You can do things like short circuit it and it is a lot friendlier than a conventional chemical battery. Silicon wafers like this are probably the purest form of material used in any commercial application, so historically the technology has been very expensive. But, as time has gone on, over the 30-odd years I have been involved with the technology, the costs have come down quite markedly and the expectation is that that trend will continue in the future for some of the reasons that I will outline tonight.

(Click on image for a larger version)

This type of application [indicating remote telecommunication system on slide] , which was actually pioneered by Telecom Australia, the forerunner of Telstra, has been economic for these solar cells, even though they have been expensive, for the last 30 years or so - where you are powering small electrical loads in areas that are remote from conventional power supplies. This wasn't a demonstration system; this was a purely commercial system that was installed in 1978 near Alice Springs. This system here [indicating large PV system adjacent the former Rancho Seco nuclear power plant on slide] was installed at about the same time and was a demonstration system. But this is the type of system that I believe will become very common in the future, where photovoltaics provide massive amounts of power to the electricity supply grid in parallel with conventional sources, such as the nuclear plant that is shown here.

(Click on image for a larger version)

This study by the German Advisory Council on Global Change resonates with my thinking about the possible future role of photovoltaics. This shows world primary energy use in exajoules. The 1600 [exajoules] here works out at 50 terawatts continuous – if that unit is more familiar to you – but at the moment we are using about 500 exajoules of energy per annum. As shown by the brown colours on this graph here, most of it is fossil fuel based. This is primary energy use worldwide.

This group [German Advisory Council on Global Change] was charged with the mission of looking at possible scenarios that we could evolve from this unsustainable dependence on fossil fuels to something that was sustainable in the long term. It was a broadly based group containing a range of skills. The view that they came to was that the most viable sustainable option was, in fact, this technology shown in yellow there, which is solar electricity production. What they saw in their scenario, I guess, was very encouraging to me, because that had been my view for some time – that photovoltaics was one of the few options that you could point at that could sustainably provide the world's energy and, if you looked at the others that might be possibilities, photovoltaics had some advantages. But it was also depressing because, when you looked at the type and rate of change that was required to enact this scenario, it seemed to be beyond the realm of reality. This was a scenario that stabilised carbon in the atmosphere at 450 parts per million, which was not substantially larger than the [present] 380, but the drastic changes that were required to enact this scenario brought home to me just the magnitude of the problem that we really are facing. In this scenario, which the group called their exemplary transition scenario, they saw solar electricity providing 25 per cent of primary energy by 2050 – not 25 per cent of electricity but 25 per cent of primary energy – and the vast majority, about 64 per cent, by the end of the century. I will try to give an idea of the issues that have to be faced if a scenario like that is to come about.

(Click on image for a larger version)

Incidentally, that is 0.02 per cent of the sunlight falling on the Earth's surface [indicating bar on slide], so there is no real resource constraint in enacting a scenario such as the Advisory Council developed.

(Click on image for a larger version)

The last five years has been a period when photovoltaics have changed from being a small-scale contributor to energy supply to being a more substantial one, and the next five years look like being a period when the technology could have an even more substantial impact. This indicates new electricity generating capacity that has been added over the years and I have shown photovoltaics in the yellow, nuclear in the red and wind in the blue. At the beginning of this period, photovoltaics were very small-scale technology; its capacity was not even equal to that of one conventional power plant being installed annually worldwide. Wind has been the renewable that's been installed in the largest capacity, as shown there [indicating on slide], and that trend has continued. Growth in wind energy is constrained at the moment only by the rate at which you can make new wind generators.

Over that period [5 years], photovoltaics have grown very rapidly. The year 2006 was a turning point: for the first time the capacity of new photovoltaics installed exceeded that of new nuclear plant. In 2007, the margin became even larger. The expectation of industry consultants – this [indicating on slide] is the expectation of how the industry is going to grow over the next five years – is of a massive explosion in photovoltaic growth. Some analysts believe that photovoltaics could well overtake wind as being the renewable installed in the largest capacity over that period [5 years to 2012].

(Click on image for a larger version)

If we look at things on a logarithmic scale – I have shown cumulative capacity here versus time – we get an idea of how we are progressing, compared with the scenario that I outlined earlier [German Advisory Council on Global Change] , with this massive transition to solar electricity generation over this century. The red lines show the assumptions made in that scenario about the growth of wind and photovoltaics. The scenario tried to be realistic in assigning developmental paths to the different technologies based on land use, water use and all the other things that these technologies require. But, for wind and photovoltaics, the constraint was the rate at which the industry could grow. They believed that a 10-fold increase in capacity per decade was the maximum any large, growing industry could sustain. If we look at wind, we see that it is more or less following that scenario outlined. It is growing at a rate where its capacity is increasing 10 times every 10 years. In fact, last year was momentous in that wind supplied one per cent of the world's electricity, and that was a significant milestone to reach. Photovoltaics, much as I mentioned before, have been very small-scale installations, until quite recently. But the growth exceeded that forecast within the study. In the next five years, it is expected to depart even more significantly from that [forecast]. It could perhaps reach this similar milestone, this one per cent of the world's electricity, and not in 2022, as this study forecast, but 10 years earlier. So there is the possibility of actually cutting short some of these periods. This is what recent experience has shown. If we maintain this path here, we get to the scenario's 25 per cent of world energy generation. All you have to do is stay above that red line for that scenario to be fulfilled. So, on present experience, it seems as though the technology is ahead of schedule.

(Click on image for a larger version)

What is needed for this vision to come about? What really needs to be done? I think the main requirement is that the cost of the technology must continue to come down. That can come about in two ways. One is just through increased volume. Most of the reductions in the cost of the technology so far have occurred because there has been more of the photovoltaic product produced. So you have to increase volume, and I will talk about some of the ways that have been explored to increase the volume of the cells being produced with subsidies, feed-in tariffs and some financial engineering. I will talk about that a little later. The other thing that is required, I believe, is improved technology. That is the role that we see for our group: to continue developing technologies so that there is something in the pipeline to allow the technology to evolve from where we are now.

Up until now, most of the improvements have been incremental. You can imagine more revolutionary changes occurring by getting rid of the silicon wafers; as I mentioned, they are an expensive part of the present package. The efficiency of the panel – the ratio of the electrical power that you get out of the photovoltaic module to the amount of sunlight falling on it – is, I believe, a key parameter in trying to improve the long-term potential of the technology.

(Click on image for a larger version)

Photovoltaic differs a lot from some of the other energy generation sources, and you can use it in a myriad of ways. It is very modular and very self-contained and deployable, so there are [many] different ways in which you can use it. This [indicating on slide] just outlines a range of different applications. One that I mention here is called the Remote Area Power Supply (RAPS) system; 20 years ago, that was the only commercial application for the technology. There are other applications – in the developing world, in particular – where photovoltaics are already the lowest cost option, but it is just not affordable. For lighting and so on, studies have shown that photovoltaic systems are cheaper than kerosene, but just getting the financing to purchase such systems is the difficulty in the developing world – not being able to take advantage of the lower cost. However, in the Western World, the really large applications are where the electricity-generating source is connected to the electricity supply grid network and, even within that, there are different types of applications for photovoltaics. The most competitive is the one that I mentioned before where you have photovoltaic generators just generating electrical energy side by side with conventional large-scale generators. That is the most demanding application in that you need the lowest possible cost for photovoltaics. But there is a range of other applications where the economics are far less demanding and in [being able to address] some of these photovoltaics, I believe, they are unique amongst the renewables.

Looking at the residential use of photovoltaics, as with this house in Germany which is well equipped with photovoltaic panels [indicating on slide], you can see here that the economics are different because you are not competing with a coal-fired power station in terms of the cost that you have to reach for this application to be economic. You are competing not with the wholesale cost but the retail cost of electricity. As we will see, there is a big differential between those two. Photovoltaics is a little different from wind, for example, in that wind has to compete with other large generators connected to the grid whereas photovoltaics, instead of competing at the wholesale level, can compete at the retail level. So, even though electricity presently generated by photovoltaics is more expensive by a large margin than that by wind, there are still applications that it can tap into because of its modularity and deployability.

Then there are more technical applications within the grid network. Photovoltaics distributed within the grid can give you additional benefits additional to the electricity they generate. This [indicating on slide] is just a simple example showing that, if you install photovoltaics near a load centre that is growing, you can defer the need for upgrading of the transmission lines. So you get capital benefits as well as the energy benefits from the electricity that photovoltaics generate.

(Click on image for a larger version)

This looks at the present costs of photovoltaic generation. This study, which was done by the European Commission, shows the cost of photovoltaics from small rooftop systems and large ground mounted systems throughout Europe. The present costs range from 25 eurocents a unit of electricity in the Mediterranean regions going up to what I think is the largest at 75c, or about three times higher, in the Arctic latitudes. For the ground-mounted system, the costs are somewhat lower, going into the darker colours down here. I guess that is below the 25c level. It is perhaps 15c in the southern parts of Italy and Spain for the large ground mounted photovoltaic systems.

(Click on image for a larger version)

A lot of the discussion at the moment within Europe is about grid parity of photovoltaics. When I say 'grid parity', I mean grid parity at the retail level. So this slide just takes the figures from the previous graph and compares them with the retail price of electricity in various countries throughout Europe. Up this axis, you have euros per kilowatt-hour of various countries like the Mediterranean countries of Cyprus, Malta and Portugal. In some cases, such as Italy, the cost of electricity is quite high; already in some parts of Italy the costs of electricity generated by photovoltaics is more than competitive with the retail price of electricity. In other Mediterranean countries, you have lower photovoltaic costs, but electricity is quite cheap.

Germany is an interesting case. Its electricity costs are quite high because there is not as much sunlight there as in Mediterranean countries; therefore, at present, energy generation costs are substantially higher than the retail price and so on. But over the coming decade it is expected that, as the cost of photovoltaics continues to decrease, starting from the southern bits of Europe and working up, the threshold below that latitude where photovoltaics would be economic compared to retail electricity will slowly move north through Europe. What is foreseen is a steadily expanding market for photovoltaics, as that demarcation line moves steadily north throughout Europe. So there is an enormous potential market at the retail level for photovoltaics.

(Click on image for a larger version)

I will talk now about some of the schemes that have been used to encourage the use of photovoltaics to try to get the production volumes up and to drive the costs down. That has been a common theme in the thinking about the commercial development of photovoltaics. The Japanese introduced a very effective subsidy program that I will talk a little bit about. It was really well organised. It started with the installation of lots of dummy photovoltaic systems on Rokko Island, which is the artificial island outside of Kobe, just to investigate all the issues involved with having distributed photovoltaic generators throughout the electricity supply network. This was done by the Central Research Institute of Electric Power Industry, which goes under the name of CRIEPI. They found that there was no fundamental problem with installing photovoltaic systems throughout the grid network.

(Click on image for a larger version)

They investigated a range of different scenarios [arising from PV in the grid network] and, as a result, the Japanese implemented the 'million roof program' [Million Solar Roof Initiative]. It was launched in the 1993-94 period, when world production of photovoltaic panels was at only about 60 megawatts, which is a very small amount by today's standard. Their aim was to install 100 times more photovoltaics over the decade up to 2010, equipping 70,000 roofs by 2000 and one million by 2010.

(Click on image for a larger version)

The way they set about doing this was by introducing a subsidy. This is a very complicated graph, but I will just spend a little bit of time going through it because the results are quite interesting. I have used some fairly weird units just to get everything on to the one graph. At the back there [indicating on slide] is the cost of the photovoltaic system within Japan in these very weird units of 100 kiloyens per three-kilowatt system. Three kilowatts is about the size of a system you require to provide a large part of the electricity for the average home. As you can see, the program was successful in driving the cost down, which was the intent. The yellow bar shows the subsidy per system that was paid under the Japanese government program. The difference between the yellow bar and the brown bar is what the householder actually paid. If you just follow these two curves down, you will see that, as the program developed, the homeowners who participated paid pretty much the same for the photovoltaics. So there was no disadvantage in getting in late as opposed to getting in early, but the subsidy per system went down quite dramatically. In this last period, 2004, you needed only five per cent of the system in price [as subsidy] – so a five per cent subsidy. It was massive at the start, when nearly 50 per cent of the system price was subsidised by the government; but, as the cost dropped, the subsidy dropped down. The green bar is really interesting. That is the number of systems installed and that just continued to grow. Even though the subsidy was dropping, the market was very vigorously growing and it seemed like this million-roof target would be reached. The final bar is the total amount of money paid in subsidies by the Japanese government. So it grew as the number of systems grew but, because the amount of the subsidy was decreasing, it dropped off in the final years of the program. Up to the end of this period here, over a quarter of a million Japanese homes have been equipped with photovoltaics. I calculate that as being about one in 100 Japanese homes that have been equipped with photovoltaics.

The Japanese government, I think under pressure from different groups, decided to drop the subsidy from five per cent to zero in 2006. The five per cent subsidy, as you can see, was not costing very much and the market was bubbling on, so it may have been more sensible to keep some type of subsidy in place. But the act of dropping the subsidy from a very small fraction of the total system cost to zero had a huge impact on growth. I did not get time to update this chart, but the market dropped back very markedly. This highlights one of the problems with a subsidy-based program, particularly when it is a government subsidy: you are very susceptible to political whims and so on. There was lobbying against the subsidy by various groups within Japan and the subsidy program was cut out. As a result, a program that had been so promising began to collapse. So subsidy programs involving governments have the problem – we have seen it here in Australia as well – of intermittency, which can undo the industry development and so on that occurs as a result of the subsidy program.

(Click on image for a larger version)

Germany has adopted the most successful renewable energy development program to date worldwide. This has been a remarkably effective program and it is based on a feed-in tariff. This slide demonstrates eurocents per kilowatt-hour that two technologies get under this German feed-in tariff scheme. The top one is photovoltaics. If you install a photovoltaic system in German in 2008, you are guaranteed 47c a kilowatt-hour of electricity that you produce over the next 20 years of the system's operation. That is locked in under German law. So you can go to your bank – banks there understand these schemes – and say, 'I want to buy a photovoltaic system; can I get a loan to allow me to purchase it?' The banks are familiar with this – and they say, 'Sure.' They know that this guaranteed income will support the repayments of the loan and the systems have a level of reliability and durability to make everyone confident that it is going to occur as planned.

The interesting thing about the German program is that each year the amount of subsidy that you get decreases by nine per cent. So it is 47c a year in 2008; in 2009 it is 43c or something like that. So it goes down nine per cent a year. So, if you install a system in 2009 rather than in 2008, you get 43c a unit of electricity produced. Then it drops down and, by 2020, you are down to the retail price of electricity in Germany, which is the 20 eurocents a unit that was mentioned before. So the scheme puts pressure on those that are benefiting from the subsidy, the people who are supplying these photovoltaic systems to the German consumers, to reduce their costs by nine per cent a year. They have built up a business that is expanding and doing well and, to maintain the vitality of that business, they need to make sure that their costs are nine per cent lower in the next year or something is going to change. It is a really clever scheme. That has been responsible almost entirely by itself for the massive growth in photovoltaics that we have seen worldwide. We saw it surpass nuclear in 2006, and all that is due just to this one scheme in Germany.

For different technologies, you get different guaranteed amounts. The other payment that I have shown here is for onshore wind, which is one of the cheaper renewable options. There you get something like 8c a unit now and, if the digression rate is less, you are down to 7c a unit in around 2020. But you are guaranteed that income from each output of electricity produced from wind generators. That has had a similarly enormous impact on the growth of the wind industry. At the moment, the wind industry is growing as fast as it physically can and is constrained only by the rate at which new generators can be built. That, again, is almost entirely as a result of this one German program.

(Click on image for a larger version)

How is that financed? This slide shows the retail or household price of electricity in Germany in eurocents per kilowatt-hour. From 1998 to 2007, there has been hardly any change in the price that a homeowner in Germany pays for their electricity. This is in 2007 dollars, so there has been no change in the real price that people pay for their electricity. You will see that the composition of the price is different. In about 2000, Germany reorganised its electrical power industry and a lot of excess, redundant capacity, particularly old capacity, was decommissioned. It became more efficient, particularly in the generation costs of electricity, this sort of reduced the dark areas [indicating generation cost component on graph], the generation costs of electricity. This is what I have been calling the wholesale price of electricity, which is the value of electricity at the output stage from a large coal-fired power station or something like that. On top of that, there are distribution costs, which are the costs of getting the electricity distributed around the grid network, which can be larger than the costs of generating it. They have stayed pretty constant. So what has changed to compensate for this lower generation cost? It has been the rather large increases in taxes that German householders pay on their electricity; so increasing taxes are paid on the electricity sold in Germany. The green bit, which is less than one eurocent per kilowatt-hour, is the cost of this scheme that I mentioned before. The scheme is financed by the additional cost – paying that 47c for a unit of photovoltaic electricity – being shared between all the customers of the electricity companies. So the government is involved only in setting up the legislation; it is the power companies that absorb the costs and pass them on to their other customers. This has been the cost of the Germany feed-in tariff scheme, which has revolutionised both the photovoltaic and the wind industries. It has been, I believe, a fairly small price to pay, particularly when other German taxes that are unrelated to renewable energy or global climate change have grown a lot more aggressively.

(Click on image for a larger version)

If we look at the pay-offs from that scheme, one of them is quite obvious. This shows the amount of gigawatt hours of energy that is generated by different sources within Germany. The blue shows the traditional hydro-electric generation, which has been the mainstream of renewable energy generation over the last century. The first tariff scheme was introduced in 1991 and, subsequently, there have been various developments of it. This one [indicating on slide] in 2000 was the one that really had the large impact. You can see the massive increase in this blue region, which is wind energy; and biomass is covered under this scheme as well. They are all within that 1c per kilowatt-hour. More recently, there are photovoltaics; this is electricity generated by photovoltaics. This 90,000 gigawatt-hours generated by renewables was about 14 per cent of German electricity generation last year. Now renewables provide 14 per cent of German electricity, and the percentage is obviously growing very quickly.

Australia's total electricity generation per year is about 200,000 gigawatt hours. This renewable generation within Germany would represent 40 to 50 per cent of Australian electricity generation, and Germany has a lot less wind and certainly a lot less solar than Australia has. It shows what these technologies are technically capable of.

(Click on image for a larger version)

This is employment within the renewable energy industry within Germany. This little box here [indicating on slide] shows the consolidated figures for 2004 – 2005 is missed for some reason – 2006 and 2007. But over a quarter of a million Germans now are employed in the renewable energy generation industries within Germany, and the number is growing quickly. Photovoltaics are the most rapidly growing employer at the moment. So there has been a massive increase in employment within the renewable energy sector within Germany as a result of this scheme.

(Click on image for a larger version)

We often hear that these schemes to promote renewable energies are going to send a country bankrupt. They are too expensive and we cannot afford them. What has happened in the case of Germany? Up until about 2002, Germany was known as the sick man of Europe. This little chart here demonstrates that graphically by showing GDP growth throughout Europe [1990 to 2002], with Germany notably badly performing. If we look at what has happened since then, we see that GDP in Germany has grown very rapidly. In fact, it has been one of the strongest of the major European economies in terms of GDP growth. If we look at unemployment, we see there has been this massive drop over the same period. It is not all due to solar, but a substantial fraction of it is due to solar and other renewables. So it has not sent Germany bankrupt.

(Click on image for a larger version)

What it has done is that it has simulated the more widespread adoption of these schemes. Obviously, they are a lot more successful than even the Japanese subsidy scheme and a lot of other schemes that have been tried in different countries around the world. A lot of countries throughout Europe have now adopted a similar feed-in tariff scheme. In Australia, in Queensland, South Australia, Victoria – and the ACT, of course – there is discussion about introducing similar feed-in tariffs schemes. There are some differences from the German scheme. In the Australian case, it is often the excess that you generate over your use that you get these special rebates for. That, I guess, differs from the German scheme, in that it does not make the proposition bankable; you do not have a firm income from renewable electricity generation. Can Australia afford a scheme like this? I think it can. Is it going to send us bankrupt? No.

(Click on image for a larger version)

Other interesting things are happening as well. In the US, some financial engineering packages have been developed that are quite interesting. They offer another example of the way the industry might be able to grow. Power purchase agreements have been promoted by several companies around the US. Participants have often been major retailers, such as Macy's, who last year announced that they were going to solar power 26 stores in California. But the scheme works in that the company, such as Macy's, which has the solar system installed, just provides the roof space; they provide the site. The system is actually owned by a group of investors who are able to benefit from the tax credits and other financial opportunities that installing a system like this provide. There is a project developer involved as an intermediary, and the system supplier is the other party involved. So it is a complicated financial arrangement. But the benefit for a company like Macy's, for example, is that it has the publicity of having solar powered stores. The conversion to solar power is often done in conjunction with energy audits and efficiency improvements in the use of electricity throughout the company's operations as well, which is another important side benefit. Another benefit that these companies get is a guaranteed electricity price for the next 20 years. So they have their electricity price locked in, which is a hedge against inflation of electricity prices – which is very certainly going to happen within California. That is an important consideration for companies operating there. In addition, through the various financial instruments I mentioned, the investors are able to receive a return on the investment they make. Photovoltaic suppliers are happy because they get money from supplying the system, and then the project developer makes their commission on the whole system as well. It is a win-win-win situation. This is another potential way that the industry can grow through these financial engineering packages: providing a hedge against the likely escalation in electricity costs through having a generator whose costs are well understood and which is likely to be giving a reliable performance.

(Click on image for a larger version)

I have talked a lot about the industry and where it is and what the costs are; now I would like to talk a bit about the technology. As I have mentioned, present photovoltaic modules are based on silicon wafers, which I call first-generation technology. That is the technology that has been around for 30 years, and it has not really changed all that much over that 30-year period. Recently the industry reached a landmark in that it has now started to use more silicon than microelectronics. It used to be a small part of microelectronics, using the scrap from the microelectronics industry. But recently it grew to become larger than microelectronics, which started creating problems for the industry in that it could no longer rely on the microelectronics industry to supply its silicon. It caused a shortage of the hyper-pure silicon that these wafers require. That is one of the reasons for its growth being more constrained. Even though its growth has been very rapid over the last five years, it has been more constrained by the availability of silicon than it otherwise could have been. This is one of the reasons that a massive explosion is expected over the next five years where this silicon supply-shortage situation has been accommodated and the supply of silicon is not expected to constrain the growth of the industry in the same way.

Over the five-year period, the solar industry is also expected to become larger than the microelectronics industry in terms of its capital revenue, its revenue on sales. That is another important landmark. So it is starting to become big business. In trying to accommodate the silicon shortage, there have been technical issues like trying to make the wafers of silicon thinner. Companies have been very successful in making the wafers about half as thin as they were five years ago, just adjusting to the silicon supply shortage. The purification of the silicon that is required for these cells has been simplified. So a lot of simpler technology is now being used to produce the pure silicon, which still has to make its impact on the cost of the technology.

(Click on image for a larger version)

An important issue is higher conversion efficiency. It is there that my group is particularly well known internationally for its contribution to improving the efficiency of the photovoltaic converter. This is a history of efficiency evolution of silicon photovoltaics. The first cells were made in the 1940s and, in the 1950s, the same type of technology that led to the explosion in electronics led to big improvements. Our group got involved in the late seventies and, throughout the eighties and nineties, we progressively improved the performance of silicon cells; we improved their performance by about 50 per cent over that period. This is the group of researchers that made the first 20 per cent efficient cell – which was the four-minute mile of the photovoltaics area.

(Click on image for a larger version)

Some of this technology has been implemented commercially. This is the buried contact solar cell from our group that has surpassed one billion in sales.

(Click on image for a larger version)

We have also been successful in education. A number of my former Asian students have been very successful in adapting to the recent growth in the industry. Last year the number of solar cells produced in China exceeded those produced in any other country, and this was the first time that China had reached that status. If you look at the companies that were involved – these are the major Chinese companies [indicating specific companies on slide] – 70 per cent of the Chinese product was made by companies of which former students of our group are either chief technology officers or chief executive officers. So we have had a major impact in terms of educating people who have been able to take this industry forward.

(Click on image for a larger version)

This is just some of the more recent technology. Plenty of technology is being developed in Australia, and the ANU is particularly well known for its 'Sliver' cell.

(Click on image for a larger version)

I believe that the technology has to evolve beyond wafers to a second generation of technology where the active material is deposited uniformly over something like a glass sheet. That is called thin-film technology, because the converters are thin-film rather than their relying on silicon wafers. There are many advantages. An obvious one is a low material cost. You get rid of the wafers; they are replaced by a very thin layer. Instead of working with wafers as your unit of manufacturing, you are working with large glass sheets, and there are economies associated with that. The cells are not soldered together, as in a normal module. There are hundreds of cells within this module and they are just interconnected during manufacture. I think the aesthetics are better and I think there is potential for the device to be more rugged than the conventional modules. Even though conventional modules are extremely reliable, I think these can take the technology to a new level of reliability.

Many materials have been demonstrated to be suitable for depositing as these thin-films. Some of them are silicon materials. You can use silicon in various phases, either amorphous, microcrystalline or polycrystalline. There are different phases of silicon from which successful thin-film technologies have been developed [indicating on slide]. You can use other semiconductors as well. Chalcogenide semiconductors, which involve elements from Group VI of the Periodic Table, have been particularly well suited for photovoltaics. One example is cadmium telluride, which is a compound semiconductor. Silicon is just one element involved in photovoltaic material; cadmium telluride involves two – cadmium and tellurium. Then there is other technology such as CIGS, which involves a real potpourri of chemical elements: copper, indium, gallium, selenium and sulfur. So there is a whole mixture of technologies. Other organic based materials are also of considerable interest.

(Click on image for a larger version)

Again, Australia has been heavily involved in the development of second generation technology. At present, this particular module here is being made under license to our university in Germany. Most of the thin-film panels are used in large grid-connected fields of photovoltaics modules because you need lower costs for the panel for that application.

(Click on image for a larger version)

An enormous amount of work is needed with organic cells. The hope is that one day we will develop a solar cell that can be rolled out like Gladwrap that will convert sunlight into electricity. The challenge with that is that polymeric materials are not necessarily all that resistant to the environmental hardships that they are going to experience in an application where you are trying to capture as much sunlight as possible. So durability is a real challenge and the initial efficiency is also a challenge at the moment. However, I think there are real applications in consumer products such as this [indicating a roll-out flexible panel on slide].

(Click on image for a larger version)

This shows the efficiency/cost issues in a slightly different format. Here we have efficiency versus cost of a panel like this, in dollars per square metre. As I mentioned before, a panel like this is quite expensive in that the silicon wafers are expensive and the encapsulation scheme that is used is quite sophisticated. At the moment, to manufacture a panel like this costs about US $300 a square meter. The efficiency of a panel – the ratio of electrical power output to solar in – is about 14 per cent, so it lies somewhere about here on this graph. These dotted lines show dollars per watt, which is the market metric. If you compare different technologies, you compare the dollars per watt that you pay. At the moment, to manufacture a panel like this costs about $2 per watt. The typical manufacturing point would lie somewhere around where I am directing the pointer. This bar here represents the efficiency limit on a panel like this. If you go to thermodynamics and calculate the limiting efficiency of a panel like this, you get 31 per cent as the answer.

(Click on image for a larger version)

I will briefly touch on why that is the case. It is a trade-off between two effects. These semiconductors, like silicon, have electrons in a band of states – they are fully occupied by the electrons here – then there is a gap in the available states and then there is another band of empty states up here [indicating on slide]. So sunlight absorption, such as by this green photon here, excites an electron from this fully-occupied band to the empty band, and then this is the electron that contributes the electrical current to the converter. This bandgap has a finite value. Some of the photons in sunlight, some of the wavelengths or colours of sunlight are not energetic enough to create that excitation event, so that just passes straight through. That is one loss mechanism. You could easily fix that by having this bandgap smaller, except for the second-loss mechanism, which is where a high-energy photon excites an electron from way down here to way up here, but the excess energy is very quickly lost; it loses it very quickly.

(Click on image for a larger version)

If you look at those two trade-offs, you get a maximum in the efficiency versus the bandgap within the material that is close to 30 per cent.

(Click on image for a larger version)

With the second generation technology, you have much lower costs per unit area, but the efficiency drops down a bit. However, you can still get some leverage in terms of the dollars per watt. Just about where I have the pointer now is where the most successful thin-film company internationally now is; it is about half the cost of the wafer technology.

(Click on image for a larger version)

The final bit of work I would like to talk about is our work on third generation photovoltaics. We have looked at the fundamental thermodynamics of solar conversion. If you just do a simple energy and entropy balance, it turns out that the limit of the efficiency you get is not 31 per cent but 74 per cent, which is very much higher.

(Click on image for a larger version)

We have started a line of work to try to develop a third generation of technology that has the cost advantages of the second generation, in terms of dollars per square meter, but breaks through this efficiency barrier, targeting much higher efficiency. I will not have time to talk about that, but I would just mention that is our research interest.

(Click on image for a larger version)

I will just finish by talking about another thing that is required for photovoltaics to be implemented on very large scales, and that is to address the issue of how you store the electricity for periods when the sun is not shining. To really address this issue, you have to look carefully at the way electricity is actually used. Demand for electricity is not constant upon an electrical grid network; it changes. Japan is quite an interesting example. This shows time of day and this is the recorded electricity use in Japan for various days. This is 31July 1975. In Japan, most of the electricity is used in the afternoon on summer days. August 1, 2002, is the peak in that particular year; that is around the time that sunlight is the strongest. Then, if you look at seasonal use, the demand peak for electricity is in summer afternoons. In winter – this is winter here – the demand is very much lower. So, in Japan, there is a big difference between the demand for electricity in summer and in winter. That has been the trend with all grids worldwide, with the demand for electricity slowly increasing to a summer peak due to the increasing use of air-conditioning and the decreasing use of electricity for heating.

If you install a photovoltaic system, one argument that you often hear is, 'But what happens if you install a photovoltaic system and the sun does not shine? You're going to need a conventional plant to back up the photovoltaic system because the sun might not shine; what will you do to supply your power?' But, basically, that argument is fallacious, and it is easy to see that. A conventional plant is not 100 per cent reliable, so you could say the same thing. If a conventional plant is only 95 per cent reliable, you are going to need another plant with the same capacity to back it up for the period during which it is not operating. So it is more of a probabilistic thing by which grid companies actually incorporate capacity on to their network.

(Click on image for a larger version)

This is a very interesting graph that shows the capacity credit that you can assign to photovoltaics, which is the fraction of the rating of the system that you can count on when it is needed, when the grid is demanding its maximum power. This is the capacity credit versus – something I talked about before – the ratio of the summer to winter peak. This is a US study, but this is unity for the ratio of summer to winter peak. As I said, most of the grids within the US are summer-peaking; they have their peak demand for electricity in summer, and the trend for that is to increase. But the capacity credit, once you get above that unity, increases from about 50 per cent of the rating of the photovoltaic system up to 80 or 90 per cent. So, because the summer peak is correlated with the availability of sunlight, a very high rating can be given to the capacity of the photovoltaic system.

(Click on image for a larger version)

I should finish there, but in this graph I have just mentioned various storage options that are being explored internationally. I think that is the other thing that is required for photovoltaics to be used on a really large scale: eventually, you will need some type of energy storage. That is an important area for research.

(Click on image for a larger version)

Just to finish: I believe that you need to fix the carbon problem at the source. To do that, you need a clean cost-effective electricity generation option. Photovoltaics, I believe, provide a solution, provided that you can increase the volumes and get the costs down dramatically. I believe that high-conversion efficiency is the key. I did not get to talk much about the third generation approach, but I believe that improving the efficiency of a thin-film product is the key to reaching the ultimate potential of the technology. So high-efficiency thin-film technology is the key to the post-2020 era. I did not talk too much about our work on third generation technology, but nanomaterials give you design flexibility in engineering new properties that allow some of these more sophisticated conversion approaches to be implemented. With that, I will bring my talk to an end. Thank you very much for your attention.

Discussion

Chair (Mike Dopita): Thank you very much. That is a lot of food for thought. I will throw the floor open to questions.

Question: One issue you didn't touch on was whether photovoltaics or solar thermal, using mirrors and the like, to generate solar electricity will be better in the long term for really big-scale solar generation of the sort that you had in one of the slides and of the sort that is being built at Mildura. Can you just tell us which of those two technologies is better for that big base load power, particularly considering that with solar thermal you can store heat which apparently is cheaper and more efficient than storing electricity produced by photovoltaics?

Martin Green: Yes, I think the big advantage of solar thermal is that you have this potential for cheap thermal energy storage. You need to store a lot more energy because it hasn't yet been converted into electricity; so you need something like three times the storage. But the storage per unit of energy is cheaper, at least potentially, if you store it as heat. I think that is the real advantage of solar thermal. In terms of the relative costs, it is a matter of a point in time. In the eighties, it was very clear that solar thermal was a much cheaper way of generating electricity than photovoltaics because the technology is basically there [already]. Mirrors, steam generators and so on: all the technology was developed already. But, with the costs of photovoltaics coming down, it is not that clear-cut any more. I think we are just reaching a transition period at present, where some of the costs of the photovoltaic systems that are being installed now are very competitive with those of the solar thermal systems.

There are other issues as well in that solar thermal systems have to focus the sunlight. You get more susceptibility to dust and so on whenever you have to focus the sunlight rather than just using a diffuse source of sunlight like the cells can. So differences in water consumption and issues like that are proving to be very important in deciding between technologies in installations that are going in now.

At the moment I think it is getting to a situation where it is fairly much line ball. My view is that, ultimately, photovoltaics can be made for the same types of costs as those for reflectors and so on that you need in a solar thermal system; they are the same types of costs as for architectural glass. That is all you need; you do not need the rest of the solar thermal system. My belief is that, in the long term, there is a very clear advantage to photovoltaics – but that is only if the technology keeps evolving and reaches its full potential. There is much less potential for evolution, I believe, in solar thermal technologies compared with photovoltaics.

Question: The storage –

Martin Green: But storage is the big advantage. That problem is already addressed within solar thermal systems.

Question: I am personally interested, as I am sure a number of other people are, in all the detail about the third generation systems that you have allied there. Do you have a source of that on your website? Is there a place that we can go to find out about that?

Martin Green: Yes, there is a lot of information there. A range of technologies have been suggested that can exceed that efficiency limit for a conventional solar converter and some of them have already been demonstrated. I ran out of time, so I did not get a chance to talk about them. The most developed is where you stack cells of different material on top of each other so that each cell converts a particular part of the solar spectrum that it can do most efficiently. But we are exploring a range of more adventurous technology than that. I believe that very little effort has been invested globally in investigating some of these options, and it could prove to be a very fruitful area for future research.

Chair (Mike Dopita): I will take advantage of my position as chair to ask a question. Given that the areas in Australia where you get the best return on solar energy are not those where people live, what do you have to say about long distance grid transmission and the associated problems?

Martin Green: Long distance grid transmission could be very important. One suggestion for addressing the issue of storage is just to have grids transmitting power globally. The sun is always shining somewhere, so you just shift the power from where it is sunny to where it is needed thus eliminating the need for storage. That has been suggested as an option. There are also issues like the one in Europe where, if you were to become solar powered, the energy intensity is high and the land area is not large; so there is a bit of a mismatch. But North Africa is already connected to Europe by long distance power lines and, as we speak, the strength of those connections is being increased. This concept of generating solar energy in North Africa and shipping it to Europe by long distance power lines, I think, is really quite technically feasible and probably going to happen.

Chair (Mike Dopita): High-voltage direct current.

Martin Green: Yes, high-voltage direct current is the technology of choice at the moment.

Question: Could you tell us what effect dust has on solar panels?

Martin Green: I briefly touched on that before. Generally, when you design a system using solar panels, you allow for dust accumulation in your design and the figure that you use is four per cent. You accumulate dust between windy periods or periods of rain and there might be a five or six per cent loss due to dust coverage; then the rain comes and washes it off and you are back to zero. Just as a design figure, you use four per cent. If you concentrate the sunlight, you become more sensitive to the dust. Systems such as solar thermal, which rely on focusing the sunlight, are about five times more sensitive to the dust than that, and will require some type of periodic cleaning. But with photovoltaics you generally do not worry about cleaning.

Question: Speaking as an architect, I am interested in the application of this technology. I am in the happy position of generating 110 per cent of my consumption, so I am very happy at the moment. However, the problem is that we are slapping panels and other technologies on our roofs. When are we going to get to the stage where we can integrate photovoltaics with the roof itself – in other words, get rid of the roof and use photovoltaics?

Martin Green: Yes. You can actually buy a product that is geared for that type of application now, so it is a market area that is being addressed. However, when you look at the actual demand for photovoltaics, most of it is in that sort of retrofit type of application. So, even though building-integrated photovoltaics make sense, the real demand is for retrofitted photovoltaic panels. The French have been particularly interested in 'building-integrated' [PV] and they offer a premium for building-integrated photovoltaics to try to enhance development of the types of products that you have mentioned. So there is a lot happening at the moment.

Question: In terms of climate change, the need to move to a low-carbon economy seems to be much more urgent than governments realise. Wobbling around the Garnaut report of a 10 or 25 per cent reduction in carbon emissions – we need to get to 70 or 80 per cent – do you think Australia at the moment has the capacity to make much deeper cuts in carbon emissions while at the same time improving employment prospects?

Martin Green: Yes. It is all a question, I guess, of competitiveness. If the world acts globally, it will be much easier to do something like that. But, at the moment, you may be placing yourself at a competitive disadvantage compared with countries that are still continuing to emit carbon. I think that is where the problem lies. But we have seen in the case of Germany that the aggressive adoption of renewables has not had any obvious [negative] impact upon the economy. In fact, the economy has boomed over the same period during which the country has been addressing those issues. So I think concerns about the negative economic impact of embracing some of these options are perhaps exaggerated.

Chair (Mike Dopita): I would go further and say that successive governments have been in the pocket of dirty coal. Dirty coal has the exports and the clout, whereas photovoltaics do not. That issue really has not been addressed successfully because we do not have the lobbying power that the coal industry has.

Question: I am particularly interested in the issue of storage for periods during which the sun does not shine. Have you examined the Snowy Mountains Scheme to see whether there are any pumps there that could pump the water uphill by day using solar energy? The water would run down to the same generators by night. Have you examined that at all? I think that would be the ideal battery for the whole East Coast.

Martin Green: Yes. I had to cut short my spiel on storage, but that sort of pumped hydro storage is the largest type of storage that is used internationally at the moment. I was going to explain how Japan has perhaps the largest amount of storage within their network at the moment, and that is to address that issue with the large summer peaks in demand. Japan has a large percentage nuclear plant which can be operated [by law in Japan] only at full capacity. They run at full capacity during the night, when there is very little demand for electricity. Japan uses that capacity to pump water uphill – a massive amount of that type of pumped hydro storage – to discharge it during the daytime. But the punch line is that the more photovoltaics you install in Japan the less storage you require, because it addresses the same peak that the pumped hydro storage presently addresses.

Question: I have a fairly prosaic question, but it has puzzled me. I have been through two fairly devastating hailstorms in my life where a huge amount of damage has been done to the two communities in which I have lived. How do these things stand up to that kind of attack?

Martin Green: Modules can be bought from reputable suppliers. In Australia we have a certification program, so all the certified suppliers are reputable. The modules have to undergo certification tests, one of which is hail impact. So the modules are designed to withstand the direct impact of hailstones of a certain diameter at their most critical points.

Question: Like those in the hailstorms that we just had in Canberra and in Sydney.

Martin Green: Sydney had a very large hailstorm in 1999 – I think it was in April or May. At that time, we had a large number of photovoltaic systems out and some were damaged. But in that hailstorm conventional roofing material was very seriously damaged; ceramic tiles were blasted apart and so on. I think the real design criterion is that it [PV] has to withstand the same type of hail impact as conventional roofing material. You would not really expect a solar panel to survive a hailstorm where people see their cars are getting written off and their roofs are getting busted apart, and insurance companies can pay for that as well.

Chair (Mike Dopita): I think we had better finish the questioning there. I am sure that Martin would be happy to answer any questions that people have afterwards. I would like to thank Martin for a fascinating talk. I do hope to see the Australian government produce feed-in tariffs, particularly in a unified national way. It is ironic, isn't it, that neither New South Wales nor Western Australia, the two states that have the highest insulation radiance and the best conditions for the production of solar energy, operate feed-in tariffs. Those feed-in tariffs that do operate around the country operate at incredibly different rates, ranging from 60c in the ACT to as little as 10c in, I think, Tasmania.

I think we have been shown here that it does not take a miracle; it doesn't take enormous effort to make solar photovoltaics and other alternative energy sources successful. It just needs a government will to do it. In the front row, my friend from the Greens party is snorting again. But we can do this with relatively small shifts of government policy. We need to get government to produce these small shifts, such as implementing the recommendations of the Cutler Innovation Report, for example, which suggests that the research and development be changed from a tax deduction to a tax credit. It is an amazingly simple change, but it is one which would basically help start-up enterprises and would stop all the money for research and development flowing to the big companies. So I think policy issues related to alternative energy are as important as technology. Martin has really shown how policy from an enlightened educated government that wants to produce an outcome doesn't have to spend an arm and a leg but it sure produces results. Martin, thank you for that.