Solar on water, robots, and 2-sided panels, oh my: solar tech’s near future

Back in 2008, when the US was deploying less than 300 megawatts a year of solar photovoltaic (PV) panels, and the boom in solar was but a gleam in President Barack Obama’s eye, it was very much an open question which solar technology might triumph in the end, or if any of them would triumph at all.

Most solar at the time used old-fashioned crystalline silicon panels, which have been around since the 1950s. But there were hopes for (and some extremely large investments in) multi-junction solar cells, thin-film solar panels, organic and carbon nanotube cells, concentrated solar plants, and a variety of other solar power-generation technologies.

Today, after a dizzying decade of solar growth — the US installed more than 10 gigawatts of PV in 2017, and five US states now generate 10 percent or more of their electricity from solar — most solar uses … old-fashioned crystalline silicon panels.

They’ve gotten better, of course, but more importantly, they have gotten cheap — more than 80 percent cheaper since 2009:


Thanks, Obama.
Wikimedia

Silicon PV dominates the market more than ever.

The point was driven home to me recently when GTM Research (which is in the process of becoming Wood MacKenzie) contacted me about a series of short reports on nascent solar technologies just breaking into markets.

I expected to hear about fancy new solar cells — nano-this, quantum-that. But that’s not what the solar market it is right now.

Rather than silicon alternatives, the new technologies nudging their way into commercialization complement silicon PV, enabling panels to deploy in new places (on lakes and reservoirs!) or reducing the cost of maintenance (panel-cleaning robots!).

At least for now, silicon PV has won the market race. The promising solar markets are for technologies that help it go farther.

There is some controversy over whether silicon PV can ever go far enough — whether it will be sufficient for the ambitious long-term decarbonization goals the US has set for itself. Some researchers worry it could plateau early and fall short.

But first, let’s take a look at a few of these PV-supporting technologies. There will be robots.

Solar on the water (fire in the sky)

In some growing solar markets, especially in East Asia, land is becoming a constraint. Much of the flat land is taken for agriculture. The land in and around cities is densely populated. Finding room for solar will only get more difficult.

One solution: Put it on the water! PV panels can be mounted on pontoons that float on freshwater lakes and reservoirs.

Floating PV has a few advantages: it can be built without traditional site preparation, pile driving, or fence and road construction; there’s less competition for the “land”; the panels stay cooler, which boosts efficiency; and floating systems can often be built closer to loads.


The Geumjeon floating solar power plant, in South Korea.

The Geumjeon floating solar power plant, in South Korea.
Seaflex

Floating PV has been around since 2007, but falling costs mean it’s just now beginning to take off in high-population-density, land-constrained markets like Japan, China, and South Korea. A total of 100 MW of floating PV was installed in 2017, with a 40 MW project completed in China and a 200 MW project announced for 2019 construction in Indonesia.

Just based on what’s currently in the pipeline, GTM expects floating PV to top 1.5 GW next year.


floating solar

GTM

And as PV grows, more markets will begin to find land constrained, leading to more opportunities for floating PV.

What would really crack open the floating PV market would be saltwater PV, i.e., “offshore solar” that could sit in bays or harbors, or offshore, adjacent to wind or oil platforms. For now, there are just a few startups and pilot projects in saltwater PV, but the potential might be big enough to attract investment in coming years.

Imagine, for instance, the potential of a floating PV system in New York Harbor: almost unconstrained in size, competing with nothing but shipping, and sitting right next to an enormous load. Juicy!

Autonomous robots that clean solar panels, because of course

For the operators of solar PV power plants, “soiling” — wind blowing dirt or sand up onto panels — can be a serious problem, leading to as much as $0.3/kWh of energy loss.

But getting them cleaned isn’t cheap either. “Most third-party module-washing companies charge around $0.25 per panel,” according to GTM’s report, which can add up to thousands of dollars per site visit. The efficiency loss of soiling versus the cost of cleaning is a delicate (and extremely site-specific) balance.

Current washing options are manual (hoses, squeegees), semi-autonomous (a robot is placed on the end of each row in succession), or fully autonomous (there’s a cleaning robot on each row and it runs itself). Some autonomous solutions are waterless.

The fully autonomous market is still pretty tiny — it covers around 0.13 percent of the global solar market. One big problem: rain. Cleaning is only really a problem in areas with low rainfall and high soiling.


autonomous solar pv cleaning robots — the market

GTM

Nonetheless, GTM thinks that if robot vendors work more cooperatively with module vendors, costs come down a bit more, and some big incumbents get in the game, the market has a lot of headroom. It expects the total amount of robot-washed PV to grow from 1,905 MW today to 6,103 MW in 2022.

Solar all up in your buildings

“Building-integrated PV” (BIPV) is designed into materials so that it is part of initial building construction (or big retrofits). Theoretically, that covers all kinds of materials that people are trying to put PV into, from windows to awnings to concrete itself, some of which have small markets going. But GTM mainly focuses on solar roofing.

Solar roofing uses solar panels as shingles, replacing (rather than sitting on top of) a traditional roof. You may recall that Elon Musk made a big splash back in 2016 when he announced that Tesla would be getting into solar roofing. (Like many things in Musk’s life, that is not going very well.)

The GTM report (not yet published online) is skeptical about the solar roofing market.

It acknowledges that the advantages are substantial. “It is a pretty serious problem that many potential customers of residential PV don’t like how it looks,” Ben Gallagher, senior GTM analyst and lead author on the reports, told me. Solar tiles improve aesthetics.

And there are advantages to reducing the upfront “soft costs” of solar — customer acquisition, siting, engineering, permitting, etc. Theoretically, solar roofing avoids much of that by slipstreaming into the established construction process. It requires less specialized labor and fewer man-hours to install.

The “total addressable market,” GTM calculates, “encompasses 14 states totaling 15.49 GWs.” In almost all of those states, the demand just for re-roofing (or just for new roofing) exceeds current PV demand.


Tesla solar roof

Tesla’s solar roof tiles, installed.
Tesla

But it has never taken off. “From 2010 to 2017,” GTM reports, “BIPV has never accounted for more than 1% of annual residential rooftop installations.”

What’s the problem? To return to a theme: The price of “standard” rooftop solar has gotten so damn cheap that BIPV is having trouble getting a foothold. Some 13 companies that tried have either gotten out of the market or gone under.

And, though it acknowledges enormous uncertainty around the solar roofing market, GTM doesn’t expect any huge leap any time soon. Two things will spur the market a bit: Tesla reliably producing solar roofs and California’s new mandate that all new homes install solar. But even with those boosts, BIPV will be less than 1 percent of the residential rooftop market. (GTM also runs a high-end case, where the market catches on and big vendors get in, where BIPV reaches 5 percent.)

A few other PV-supporting technologies to watch (more robots!)

In addition to those above, here are a few more nascent markets growing in the silicon-PV ecosystem:

  • Drones: GTM has a report coming soon on unmanned aerial drones (ahem, flying robots) that can monitor PV farms, identifying nascent problems and helping with preventative maintenance.
  • Smart inverters: Another upcoming report looks at recent advances in inverters, the widgets that control PV at the module and string (i.e., string of panels) level. They’ve been getting smarter, with “more and more granular data from the string level,” Gallagher told me. That will continue improving the performance (and lowering the price) of PV plants.
  • Bifacial panels: Gallagher told me to keep an eye on panels with PV on both sides, which can capture the sunlight that reflects off the ground, boosting overall efficiency, according to manufacturers, up to 30 percent.

Prism Solar’s bifacial solar PV.

Prism Solar’s bifacial solar PV.
Prism Solar

You will note that, with the possible exception of BIPV (some solar roofing panels use thin-film PV), all the technologies described above are devoted to lowering the costs or boosting the deployment of silicon PV.

Is that all we can expect from now on? Is the cell-level solar race over?

The future of solar module technology is … cloudy

“At least in the near term,” Gallagher said, “the market for alternatives to silicon-based PV is not really there. But looking farther down the road, it’s hard to say.”

Right now, US consumers are paying almost twice as much as their international counterparts for residential solar PV, mainly because of soft costs. In coming years, Gallagher says, solar panels and inverters are going to get so cheap that the vast majority of solar costs are going to be soft costs. “The global spot price for a standard PV module right now is around $0.32 per watt,” Gallagher says, “in the next five years, we forecast it’s going to be closer to $0.18 to $0.19 per watt.”

That means, in the near-term, the most robust markets for new solar products are going to be for techniques and technologies that reduce the soft costs for silicon PV.

“I don’t know if we necessarily need perovskites or quantum dots or ‘disruptive technologies’ now,” Gallagher says. “We need to focus on deployment: making deployment easier, lowering lifetime costs, reducing the cost of capital. Those are the most important things for the PV industry.”

Investments in alternatives to silicon PV are, at least in the next five to 10 years, going to have a rough time getting anywhere.


The band gap of perovskites can be adjusted by changing their compositions to access different parts of the sun’s spectrum.

Perovskite research.
Dennis Schroeder/NREL

But that’s the market perspective. There’s also the broader public-policy perspective.

Though the issue remains controversial, many researchers warn that PV could be headed for a plateau — the top of the “S-curve” of development — well before it reaches the penetrations envisioned in deep-decarbonization scenarios.

The worry focuses on solar “value deflation.” Every new increment of solar makes all the other increments of solar worth a little less, because all solar produces energy at the same time, when the sun is out, so all solar competes with all other solar. To remain competitive, falling solar costs must chase a receding target. Solar must get cheaper faster than it loses value.

Energy storage can somewhat offset this problem, but no storage tech on the horizon can do nearly enough. The only way to do keep solar competitive in the long-term, they warn, is through new business models and intensive innovation, including in the chemistries and cells at the core of solar. (This argument is best summarized by energy researcher Varun Sivaram in his recent book Taming the Sun.)

Not everyone buys that broader argument, but whether you buy it or not, it seems obviously worthwhile to a) reform markets so they better reflect the time- and location-dependent value of solar, providing the necessary incentives to fight value deflation; and b) vastly increase the level of federal spending on energy research and innovation.

(People have been arguing for more federal energy research money and better research policy for decades. Despite all the good arguments, funding remains abysmally low and the Trump administration is working to slash it further. They remain good arguments, though. Research is good!)

It would not be wise to sit back and think that solar is done, that silicon is it and it’s all over but the soft costs. If only because decarbonization is such an important undertaking, it’s worth having a diverse set of tools to guard against contingencies.


BERKELEY, CA - AUGUST 23: A U.C. Berkeley Botanical Gardens visitor looks at the art installation ‘SOL Grotto’ that was made by using glass tubes that were to be used by the now defunct solar company Solyndra on August 23, 2012 in Berkeley, California. Be

Glass tubes, once used by Solyndra, now used in an art installation in the UC Berkeley Botanical Gardens. There’s a metaphor here somewhere.
Justin Sullivan/Getty Images

“There will be some sort of room” for other solar module technologies, Gallagher says, “I just can’t say when or what that will look like.” That’s more or less the state of expert knowledge.

So we should hedge our bets and keep developing alternative solar technologies, if only because we have no idea what’s going to happen or what we might end up needing.

But still. It’s worth pausing for a moment to appreciate how rapid and total silicon PV’s rise to dominance has been over the past 10 years. It bestrides solar markets, a hegemon, creating ancillary markets in its wake. And it’s just getting cheaper and cheaper.

Ten years ago, no one expected this. I wonder what we’ll fail to anticipate in the next 10 years.