Results tagged “solar”

When people hear I'm involved in the solar industry, they always ask whether they should put solar panels on their house. I don't actually know the answer to that -- I deal with the manufacturing side of things, which is far removed from what retail customers actually see. But a friend of a friend recently installed panels, so I jumped at the chance to find out more.

This particular installation used SolarBlend roofing, with panels by SunTech and installation by Eagle Roofing. It cost about $22,000 for a 4 KW installation. That's about $5.50 per watt, installed, before a 30% federal tax rebate. That's pretty good, and a substantial improvement from just a few years ago.

For the homeowner, though, the key question is how much electricity the array will generate, and how quickly the energy savings will pay back the cost. That's a hard one, because it depends on the climate. It's easy to see that western Washington gets less sunshine than southern Nevada, but even driving a few miles within western Washington will put you in a different microclimate. Fortunately, there's a very helpful resource, Gaisma, which merges astronomical and weather data to give solar insolation charts for many locations around the world. The installation I'm discussing here is located in Pahrump, NV.

Insolation is measured in kilowatt-hours per square meter, per day. That's the amount of light actually hitting a solar panel. Multiply by the area and conversion efficiency of the panel to get the amount of electricity generated:

Generated electricity = Insolation x Area x Conversion Efficiency

Suntech's Just Roof panels claim to produce 125 peak watts per square meter at 1000 Watts/square meter irradiance. So a 4KW installation will include 32 square meters of panels, and the panels are about 12.5% efficient. Plugging all of that into a spreadsheet, we get about 7150 kilowatt-hours per year from this installation. (Probably a bit less in practice, as panels are less efficient in hot weather.) That's probably more than an average household needs, especially if people aren't home during the sunniest part of the day. Which is why net metering -- the ability to sell power back to the grid -- is so important for residential solar installations. Let's assume that all the electricity generated by this array is either used on site or sold back to the grid.

The next step is to figure out how much the electricity is worth. That's difficult because many companies use tiered pricing: the more electricity you use, the more each incremental kilowatt costs. There is a push to implement time-sensitive pricing as well, reducing the cost for electricity use during off-peak hours. All of this is discussed in more detail here. For purposes of this discussion, I'm going to say the electricity generated is worth $0.15 per kilowatt-hour, for a total of $1072 per year, but that's just a back-of-the envelope calculation.

Without the 30% federal rebate, payback time for a $22,000 roof that generates $1072 worth of electricity per year is 20 years. With it, it's 14 years. The warranted panel life is 25 years. (This assumes that all of the $22,000 is for the array. Subtract any costs that would also be incurred by a conventional roof.)

For the sake of simplicity, I'm ignoring both the cost of money for the installation and the likely inflation in electricity costs over its useful life. I'm also ignoring any value that the installation adds to the overall value of the home. If we assume that electricity costs are going to go up over the next 20 years, then the combination of these effects should make a solar array more attractive, reducing the actual payback time.

Just for grins, I ran the same calculation for an installation in Bothell, WA. The relative lack of sun cuts the expected electricity generation to about 4850 kilowatt-hours per year. That's $727 per year at the same $0.15/kilowatt-hour rate.

Disclaimer: These values are estimates, and may not be applicable to any specific installation. If you are considering a solar roof, ask your installer to supply accurate cost and efficiency metrics.

Update: Michael Bluejay, author of the article on electricity costs linked above, emailed a link to the solar installation calculator on his site.

In last week's post on energy policy, I skipped over the effect of manufacturing efficiencies. If an information-intensive technology can be manufactured in a way that reaps large economies of scale, then the know-how used to create it can still be very cheap on a per unit basis.

Exhibit A for this effect is the integrated circuit industry. Each individual transistor is a small masterpiece of engineering. The intellectual property contained in a few square inches of silicon is enormous, but the industry spreads out the cost by making billions of them. Thin film solar manufacturers believe that manufacturing efficiency will allow them to reach a competitive price point. That would be very good news for me, my clients, and the world as a whole.

But still not necessarily for job creation.

The disconnect between information content and job creation is explored in more detail in this recent post at fivethirtyeight.com, analyzing the decline of manufacturing jobs in the US economy. It isn't due to "evil corporations shipping American manufacturing overseas." Until the start of the recent recession, US manufacturing output was at an all time high. Rather, it's due to massive improvements in productivity, which in this context means replacing humans with automation of various kinds.

If solar becomes cost competitive it will be through massive productivity improvements, achieved in part by automating every step that can be automated.

Still, merely looking at the need for productivity improvements does undervalue the job creation opportunity. Even if the number of jobs per GW goes down -- which I think is not only likely, but necessary -- the total number of jobs will still go up as the market size increases. The current world solar market is about 6 GW per year, but the world's total electricity consumption is in excess of 17 trillion kWh. That's a whole lot of room for growth.

Unfortunately, that also leaves us back where we started. Promoting use of a product through subsidies is not usually a good way to make the manufacturing of that product more efficient. The government can improve manufacturing -- for example by investing in research and development through organizations like NIST and NREL -- or it can create jobs, but it isn't really good at doing both at once.

Disclaimer: Though my past and present clients may like this post better, my opinions are still mine alone.

Is the goal of energy policy to reduce the costs energy imposes on the US economy and/or the planet, or to create jobs?

As one of my favorite energy-oriented sites points out, the two goals don't necessarily have much to do with each other.

Thin film solar technology, for example, contains a lot of intellectual property. Lots of engineering expertise is required to build the deposition systems, optimize the processes, and keep the whole operation running. Add the skilled electricians who actually build solar farms and connect them to the grid, and you get a lot of high-skill jobs. You also get electricity that is substantially more expensive than that generated by plain old boring low-tech fossil fuel plants.

But if your goal is to reduce the world's (or America's) consumption of fossil fuels and generation of greenhouse gases, whatever alternative technology you pick needs to be as inexpensive as possible. Ideally, it should be simple enough for illiterate subsistence farmers to implement using locally available materials, perhaps with guidance from a handful of engineers.

Not that there's anything wrong with green jobs. I have one myself, since most of my current clients are in the solar space. And certainly the emergence of a renewable energy sector will benefit the communities where those jobs reside.

But the assumption that clean energy will both prevent climate change and revitalize the US manufacturing sector doesn't stand up to close examination. Energy is a commodity product, and as such will always tend to migrate to the lowest cost technologies and least expensive producers. The fewer high-skill manufacturing jobs an energy technology requires, the more likely it is to actually succeed in shifting the world's energy mix.

Disclaimer: My opinions are my own, and do not necessarily represent the views of any particular past or present client.

It's easy to achieve world record solar performance. You just have to define your niche properly. The following records were all announced recently:

• Most efficient solar cell: a five-junction concentrator device from the University of New South Wales. 43% efficient.
• Most efficient triple-junction cell: a concentrator cell from SpectroLab. 41.6% efficient.
• Most efficient screen-printed monocrystalline silicon production cell: more than 18% efficient, from Suniva. The "screen-printed" qualifier is important, as SunPower's non-screen-printed production cells top 20%.
• Most efficient multicrystalline silicon panels: 15.6% efficiency, achieved by Suntech Power. Though this is a lower number than the other records, it's actually pretty interesting. It's for a complete panel, not a cell, and beats Sandia's longstanding record.

Update: But wait, there's more! I missed this one...

• Most efficient flexible CdTe cell. 12.4%, achieved by a group in Switzerland. This one is important because it uses a low temperature process, compatible with roll-to-roll processing.

Solyndra recently received DOE loan guarantees to support construction of a new manufacturing facility. Good for them. The guarantees illustrate the challenges as well as the benefits of government support, though: the law authorizing them was passed in 2005, applications were due by the end of 2006, yet Solyndra is the first company to actually receive a guarantee. Inclusion of funds in the recent economic stimulus package no doubt helped the process along.

Solyndra has an interesting technology, with CIGS solar material deposited on a cylindrical substrate. The idea is to keep the same cross-sectional area facing the sun at all times, without the complexity and cost of tilted mounts or tracking motors. It makes sense, but unfortunately they haven't shared much information about cost or performance. Third party sources tell me that their installation is as simple as claimed, but are more skeptical about how cost effective the panels are once installed.

A new study from MIT compares energy and resource consumption of various industries on a pound-for-pound basis. Yep, energy consumption per pound of output for manhole covers versus integrated circuits. Not surprisingly, integrated circuits consume a whole lot more energy. A more useful metric might be energy consumption per dollar value, or perhaps a ratio of value in to value out.

While it's tempting to dismiss the study as obviously ridiculous, energy consumption is an important issue for energy-generating technologies such as solar power. The more energy it takes to make a solar panel, the longer it takes for solar panels to start reducing net fossil fuel consumption: the panel has to recover its own energy cost first. Unless the solar industry generates more power than it consumes, it's hard for it to claim to be "green" or "sustainable."

Greentech Media recently updated their list of solar industry startups. Not surprisingly, there are a lot of them. At least 200 who are willing to tell the world about their plans, goodness knows how many more in stealth mode.

Most new businesses fail, so many of these won't live long enough to grow up. But without entrepreneurs willing to take the chance, the Intels and Apples and Googles of tomorrow would never see the light of day.

I was going to write a political post here. I decided against it: the election is generating enough sound and fury on all sides without my help.

Instead, I'll link to an especially sensible proposal on funding for the political football known as the Renewable Electricity Production Tax Credit. What about a tax on non-renewable electricity? An extra penny per kilowatt hour would add only a dollar or so to the average bill, but would more than cover the cost of extending the tax credit indefinitely.

Unfortunately, the idea that more expensive energy might actually be a good thing is likely to be too toxic for either party to touch, especially in an election year. Sigh...

Conversion efficiency, open circuit voltage, and other solar cell performance parameters are critically important to solar cell manufacturers and their customers. These parameters define the financial model for an installation, from the size of the array needed for a given output power to the installation's likely generation revenue. They set a viability threshold that new technologies must cross. Yet measuring them accurately turns out to be surprisingly challenging.

Though the whole point of solar energy is that the sun is readily available, the sun is a terrible light source for accurate measurements of solar cell parameters. Clouds, haze, and the time of day and time of year all cause deviations from the often-quoted 1000 W/sq. meter solar irradiance value, and from the "standard" solar spectrum. Instead, companies use solar simulators.

No solar simulator will precisely match the sun's spectrum, though, which means the cell being tested might respond differently in actual use conditions. Correcting for spectral mismatch is especially complicated for organic solar cells (as discussed in an article in April's IEEE Spectrum -- free for IEEE members), but several people have warned me not to trust non-certified reports for inorganic thin film cells, either.

With the solar industry's rapid growth, it was only a matter of time before someone stepped up to fill the need. VLSI Standards has introduced an NREL-traceable Solar Reference Cell, using a monocrystalline silicon cell for calibration of solar simulators. The company also offers calibration services for certification of customer reference cells. (This product is so new that it isn't on VLSI's site yet. Contact the company for more information.)

I've added the Energy Outlook blog to my regular reading list. It's pretty dense, but a good window into issues like oil prices and alternative energy policy. I wouldn't say it's a must read for people interested in photovoltaics, but it's worth a look.

Yep, it's shameless plug time. One of the reasons I've been so quiet lately is that I've been working on a study of thin film photovoltaics for NanoMarkets. The study takes a comprehensive look at the various thin film photovoltaic technologies, and considers how they are likely to match and expand the universe of applications. I'm quite proud of it.

Nikkei BP reports that the first customer for Applied Materials' SunFab turnkey solar cell manufacturing line will be Moser Baer Photo Voltaic Ltd. The new plant, located just outside New Delhi, is part of an aggressive expansion plan that will take the company to 500 MW of capacity by 2010.

The article is part of a three part series on the growth of India's electronics industry. Like most Nikkei coverage, it's excellent.

(Once again, all references are to MRS paper numbers, abstracts for which can be found at the conference site.)

Today at the MRS Spring Meeting, John Robertson reported (paper A13.1) that his group at Cambridge University has achieved n-channel mobility of 450 cm²/V-sec in microcrystalline silicon TFTs, and 100 cm²/V-sec p-channel mobility. Both those values are very good, and that's a problem. Plenty of models exist to explain why the material's mobility might be bad, and those models break when the mobility is good. More research needed.

(Special thanks to Dr. Robertson for walking me through yesterday morning's session on graphene, too.)

Meanwhile, Yifei Huang and a Princeton University group demonstrated (paper A13.2) a self-aligned process for low temperature polysilicon TFTs. It uses nickel silicide source and drain regions, aligned using the gate structure. At low annealing temperatures, the nickel doesn't react with the gate and can simply be etched away. Results were among the best ever recorded for top gate TFTs.

In the solar cell sessions, Makoto Shimosawa described (paper A14.1) Fuji Electric's FWave flexible solar material. It laminates roll-to-roll amorphous silicon/amorphous SiGe tandem cells (deposited by PECVD on plastic) onto steel foil. Each 2 square meter sheet generates 92 watts at peak output and weighs just 16 kg (including the steel foil). The company is now ramping production to wider rolls, targeting production of 40 MW per year.

The a-Si/a-SiGe tandem cell may be on its way out, though, as Xixiang Xu's group at United Solar Ovonic reported (paper A14.2) better results with small area triple junction a-Si/nanocrystalline-Si/nc-SI cells. Scale-up to large areas and optimization of the nc-Si component cell are the next steps.

The conference's own coverage is definitely worth a look as well.