US: SunShot director talks technology, funding and Chinese competition

Share

Do you believe that the SunShot initiative, which laid out the goal of driving photovoltaic installation costs down to $1 per watt, is on track?

There are two things I believe in: I absolutely believe in our goals, otherwise I wouldn’t be here, and I absolutely believe in Steve Chu, U.S. Secretary of Energy. The very reason I came here was because of Chu – this is an absolutely outstanding individual. And yes, we absolutely think we have a good chance of this happening over the time scales we have set ourselves.

What is the current status of this goal?

So if we look at the cost structure, depending on who you talk to and what market segment you talk to – we will use the utility as kind of the exemplar – you will see that module prices have been dropping like crazy. In some parts of the world where there are supporting policies, you can get modules as cheap as a USD$1.05 per watt. It is quite possible that by 2017, we are at a $1 to $1.20 kind of price structure at the utility-scale end, provided there are corresponding drops in the costs for balance of systems and power electronics. If you look at the whole pricing process right now, the module prices are being set by forces beyond DOE. It’s being set by global market forces and big Chinese companies, based on supply and demand and other subsidies that are being offered.

What about the goal of making solar as cost competitive as wholesale electricity rates? Is this also achievable?

Translated into utility language, there are two ways to look at the dollar-a-watt kind of a cost structure. The first way is to say, ‘Look, how much does it cost to install a watt of solar electricity?’ One could reply, ‘It’ll be a dollar a watt.’ Now, if you go and talk to the electricity people, they say, ‘Look, that doesn’t make sense, because you’re generating electricity over a 25-year period. You have to look at the whole levelized cost of electricity, the LCOE.’ A dollar-a-watt price structure translates to an LCOE of about five to six cents per kilowatt hour. That’s approximately the same as for fossil fuel, within a cent or so. That’s why you see that the goal of making solar cost competitive as wholesale electricity rates. By the way, what we mean by that is the lowest of the wholesale rates. The average value of the electricity is maybe around 12 cents in the U.S. The reason we set ourselves this goal of five cents is because it is really the lowest cost: We said, ‘Ok, we’re going to shoot for the sun – we might as well shoot for the best target.’

You are very confident that those goals will be achieved, but what happens if they’re not? Will more be invested in the solar industry?

Our time scale is until the end of the decade, so it’s still nine years away. Nine years is a long time. What we are doing on an almost daily basis is to evaluate the return on our investment, which is to say, we put in USD$100 million, what are the consequences? Is it creating jobs and deploying more solar? Those are the broad-level questions. Now, if those things are happening at a very rapid scale – let’s say, hypothetically, that we hit $1.20 in 2020 – ok, we don’t hit a dollar, but $1.20. In the process, we have deployed let’s say 200 gigawatts of solar and we have created five to 10 very successful U.S. companies. Everyone will say ‘Hey, this is a successful game.’ And if that has caught on, it is quite likely that we will walk away. We will not continue funding for the sake of funding. Now if, on the other hand, we think there are still parts of the technology landscapes that are untapped, we will continue to fund it. But that’s a decision that shouldn’t be made today. You have to see that nine years is a lifetime for many technologies.

Obviously an industry like solar is very reliant on funding. However, as has been widely documented, the U.S. economy has been suffering somewhat. How secure is such funding in the U.S. in light of this uncertainty?

I wish I could give a quantitative answer. There are forces beyond us: there are global forces, then there are forces within the U.S. – non-technological forces – and each one of them has to be evaluated. For sure, solar is a very hot area. It is definitely a focal point, it is a huge market and globally we think that at a dollar-a-watt kind of a price structure, the global market is going to be of the order of a few trillion dollars. Therefore, it is important for the U.S. to have a very strong strategic presence in that, and so we will be investing from DOE at the appropriate levels. How much, et cetera, is somewhat beyond our radar.

Do you really think, in the face of Asian competition, and the fact that most solar manufacturing is now taking place in Asia, that it is possible to make manufacturing profitable in other locations?

It’s a hair-raising thought, so let me take you though our thought process. Let’s go back to the beginning. Almost every one of the scientific innovations in the solar area has happened in the U.S. – multi-junction solar, all of these ideas on silicon, on CIGs, on cells – it’s all come out of DOE’s funding and it’s all coming out of labs, so there’s an incredible amount of innovation happening. And we’ve done quite a bit of analysis on this – approximately 75 percent of the early-stage innovations are being funded either by the government, like us, or though private equity. Now what has happened over the last several years is that debt financing has happened a lot more aggressively in China, for example. But, if you have automation in place, then it turns out that labor costs are not the primary driver. What is the primary driver is the fact that there is easy access to capital. In the solar business, the approximate rule of thumb is that a 100 megawatt plant costs about $100 million to fabricate. The question therefore is, ‘Should I give a start-up company $100 million?’ They’re not bankable yet, and they can’t go to a Bank of America or Deutsche Bank and say, ‘Look, we’re very credible, give us $100 million.’ They will probably give you the 100 million, but they will ask you for a 10 or 12 percent interest rate. And counting that in, the cost of doing business for them then becomes too big, which is why many of these people are moving to China or other places, where they are offered very low costs, low interest rates, or zero interest rate financing for capital, and all other kinds of rebates and incentives.

That has become the strategic difference here. The question, therefore, is whether one accepts this and says that this is not a game we’re playing. When the market is of the size of the order of a few trillion dollars, how does one come up with creative models, creative new mechanisms to do that? Our thinking is that indeed we can do it, because the intellectual property resides in this country anyway. And I can tell you our colleagues in Germany, in Europe, are thinking exactly the same way. As a matter of fact, we are even thinking of doing some joint projects with them. I have visited Brussels for the European Parliament meeting of the PV segment, Minh Le was at Hamburg and the sentiment is exactly the same – that people in Germany have deployed a lot more than we have than in the U.S. and they’re saying, ‘Hey, we don’t have a manufacturing facility – we need to have it.’The question is how does one do it? You need to have creativity.

What has been the biggest success of the SunShot initiative to date?

There have been many successes, but let me focus on one of our top ones, which has been very germane to our discussion. That has to do with taking a research idea and converting it to the early stages of the value proposition – what the people in the business circles would call the ‘valley of death’. This is where the university professors or lab researchers here have an idea and try to take it into the first stage of development. Here, in particular, we have run a very successful incubator program, which provides the framework to becoming a successful company. It is an investment which will pay off – I say this not as the SunShot director, but as a taxpayer. The primary goal of DOE’s SunShot Incubator Program is to advance the timeline and commercial potential of new manufacturing processes and products that could dramatically lower costs. Now we put in about $60 million in projects over the last five years, and the investment has leveraged on $1.3 billion of private equity. So it’s done a fantastic job of leveraging public money to draw significant private investment. This has had a stimulating effect on the economy and has been a very good success story for SunShot.

Another area of interest: within the U.S., about 40 percent of the cost structure of photovoltaics comes from things like permitting and inspection processes, which is very different in Germany. In Germany, that component is probably just 10 percent. Forty percent is a significant portion of the cost structure, so we said, ‘Hey, we need to do something to bring that down into a manageable level.’ If you look at all of the costs floating around the modules, they’re dropping significantly, while the cost of balance of systems remains high. In the SunShot program, we recently launched this very exciting initiative called the Rooftop Challenge, which goes about the U.S. and says, ‘You tell us what is your current cost for permitting and how long must you wait for a permit?’ In some cities, it’s USD$150 and two hours; in other areas of the country, it’s USD$2,500 and nine months! So we created the Solar Rooftop Challenge in which teams of jurisdictions will create common permitting practices that make sense in their communities and throughout their states. We think this challenge will not only decrease cost and red tape, but will also enable much more solar to be deployed on rooftops throughout the country.

Can you talk a bit about the cloud surveillance program taking place at Copper Mountain?

The objective is to measure two attributes of the solar irradiation. If you think about it, there are two things you want to know: what is the magnitude of the irradiation with respect to spatial coordinates; and what is the magnitude of the irradiation with respect to temporal coordinates. The reason this is so important is that if you’re a utility company, and you know that a cloud is going to pass over a big solar installation – if you have five minutes notice, the reaction that you would have is to fire up a gas turbine engine, which is boosted from the electricity that you’ve created. If you only had ten seconds of this, your reaction time is going to be much shorter, so you can’t have many options. Therefore, having a predictive algorithm based on experimental measurements of temporal and spatial variation of the solar irradiation is a big bonus. The Copper Mountain solar plant is where First Solar and other people have large scale installations, and so you can actually put sensor networks out on a real solar installation, where you measure the electricity being generated by the solar panel and you measure the irradiation and try to do one to one matching.

The end goal would be that once you have, for a given location, a full irradiation spectrum, you can develop a predictive algorithm, which can go to the utilities and say ‘Look, five minutes from now, we will have a lot of rain or five minutes from now your irradiation is going to drop by 50 percent so the output is going to change by a comparable amount.’ Therefore, it is very powerful in terms of enabling the markets or the utilities. Copper Mountain is an exemplar of how we’re working with a large-scale installation and doing R&D.

Do you think it will become a lot more common with large photovoltaic plants?

Absolutely. We are doing a lot of these right now. We’re doing what we call regional testing centers, which will have the ability to test solar panels at several megawatts. In addition to the solar panel testing where we can collect the data, etc., we’re putting in the sensor networks, where you can measure temporal and spatial irradiation. We’re working quite a bit with that, and I think it’s going to become pretty generic for large-scale plants. The other thing to keep in mind is this: let’s assume that you have a single solar panel – the cloud passes over it – gone, you get no electricity. But imagine you had a million solar panels, a huge installation, let’s say a 200 megawatt (MW) installation, while the cloud is walking over the left-most solar panel, the right solar panel is still creating electricity. So the size of your solar plant becomes an important factor. If you go from 10 MW to 50 to 100 to 500 to a gigawatt, the reaction to some changes in irradiation are very different. And this is something we are trying to create a predictive algorithm for.

By how much do you think the delivery of electricity could be affected?

It depends on the nature of the cloud. If you have a big black cloud, it can shut out a lot of the light. We have a lot of data that shows you actual measurements where you track the irradiation, and it goes from 1.4 kilowatts per meter squared and can go down to 300 watts per meter squared when you have a big thunder cloud. But how much the electricity changes depends on the size of the solar installation or the impact to the utility. At the end of the day, this is why we’re doing it. Because we want to know what impact solar has on the utility grid.

From the current projects which have been running where do you see the most cost-cutting potential?

We are a funding agency – we fund a full spectrum of R&D activities. We fund early-stage development, we fund deployment activities, we fund quite a bit of work on changing the market mindset and so with maybe all these activities, it’s very difficult to identify which has the most cost-cutting potential. I would say every one of our funding steps are aimed at reducing costs and removing market barriers so we would have a lot of difficulty in telling you if one is better than the other. When you look at our waterfall chart, the overall installed cost of PV has to come down by about 75 percent. Now, if you’re doing hardware science, like myself, then the costs of doing business is much higher, because you have to buy an MBE system or whatever. If you’re doing soft projects, you need a computer or you need databases and those kinds of things – your costs of doing business will be different. The importance of the activity is still at the same level. All of them are equally important.