Solar Panels: Another Exercise in Magical Thinking

I believe it’s high time that we put to rest the myth that solar panels are “sustainable”, “green” and “renewable” once and for all. They are none of this. Contrary to common sense, what photovoltaic panels generate is not electricity, but another round of “problems to solve”. Make no mistake, it is a fascinating technology, but there is a much better, simpler way to harness the power of the sun, one that doesn’t involve the pillaging of the entire planet.

Solar is the future, but not the way you are told.

---

I have to say, I’m baffled by the lack of technical understanding on display in the “renewables” and “electrification” space. Statistical data is thrown around about continuously improving EROEIs (Energy Return on Energy Invested) and falling costs like there is no tomorrow. These calculations, however, are based on a very-very limited understanding of how solar panels are manufactured, while completely dismissing a range of inputs essential for the creation of this magical technology.

I mean isn’t it magic, that you put a shiny black (or blue) slab of glass on your roof and it generates electricity out of thin air? After all we shouldn’t be surprised then, that now there are so many of them being installed in the hope of lowering electricity bills, that their continued deployment now threatens the very service they meant to make cheaper and more accessible. Apparently no one warned unsuspecting users that magic only works at small scale (usually in a sanctuary called a ‘lab’ and performed by magicians wearing white robes) and that a free lunch remains to be what it is: a pie in the sky. 

Some reckoning is in due order.

So, let’s start with the basics, shall we? First, let’s take a look at what these solar panels are made of. Going by their weight, the product’s heaviest component is the protective glass cover and the aluminum framing holding it all together. The essence of the technology, where magic happens — the set of silicon wafers glued to the backside of the glass — actually weighs less than 10% of the total weight of a panel. Now you just need to add some wiring to conduct the electricity away from the panel and you are all set! (OK. Almost.)

Here is where things get tricky. The manufacturing (and not the assembly) of all of these components is what takes a brutal amount of energy. In order to be melted, glass for example, needs to be heated to somewhere between 1500 to 1700°C (2700–3100°F), a temperature range wholly outside electric resistance heating and way above readings from Fukushima’s molten reactor cores. In other words: something only achievable by burning fossil fuels (mostly natural gas) and hydrogen. (As to why hydrogen is not the best idea, read my earlier post on the topic.) Melting and pouring glass into sheets is also not something you do on an on-off basis: it is a 24/7 operation. An abrupt loss of heat can easily lead to glass ‘freezing’ into the furnace and the onto other parts of the equipment, making it impossible to remove by means other than using dynamite and jackhammers.

Then comes aluminum: it is somewhat easier to melt and work with — once you’ve got a clean slab to manufacture sheets from — but making pure aluminum from its ore (bauxite) takes 17 kWh of energy per each kilogram of metal. Again, this is not something you do in an intermittent mode. Smelting is a sustained operation, one that is so energy hungry that most smelters usually have their own coal fired power plants, literally next door.

The raw materials (sand for glass and bauxite for aluminum) themselves are not coming free of charge either, of course. Both of these need to be mined and shoveled onto trucks by huge diesel powered machinery (no, batteries and hydrogen won’t do it here either), then transported to a factory, where the above mentioned melting and smelting happens. Well, as usual, no oil means no mining (at scale), and thus no raw materials for those oh-so-shiny magic panels on your roof. (By the way, the same goes to the glue holding the panel together: it is made from petroleum, just like many other chemicals and all the plastics we use in the industry.)

Silicon crystal being grown by the Czochralski method at Raytheon, 1956. The induction heating coil is visible, and the end of the crystal is just emerging from the melt. The technician is measuring the temperature with an optical pyrometer. (Source)

Now, a few words about the heart of every photovoltaic panel: the silicon wafer. Most panels produced nowadays are made from monocrystalline silicon, produced via the Czochralski process.

High-purity, semiconductor-grade silicon (only a few parts per million of impurities) is melted in a crucible at 1,425 °C (2,597 °F), usually made of quartz.

This time, due to the electric properties of pure metallurgical grade silicon, we can use radio-frequency or induction heating. This method still requires a very stable electricity supply though, responsible for the strong magnetic field required to heat up and organize silicon atoms. It is such a delicate process, that one tiny glitch can ruin the entire batch. (If you were wondering why solar panel wafers are not made by electricity from solar panels, then look no further.) This is not to mention the fact, that you need electricity on a truly industrial scale. Not a few kilowatthours here and there, but in the order of megawatts for 30 hours without the slightest interruption, to be able to pull an economically sized monocrystal, weighing several hundred kilograms. Like the one below:

No, this is not a space rocket. It is a silicon monocrystal, out of which the round slices are sown.

Now comes the slicing and dicing: effectively doubling the energy spent on each gram of silicon finding its way to a solar panel. Kerf loss alone (saw dust resulting from slicing operations) amounts to 30% of the total weight of the crystal tower above, not to mention the fact that you need to cut out rectangles from round wafers... The scraps literally weigh more than the actual solar cells themselves. All this activity, plus the polishing of slices takes and extra bout of energy, but sadly this is where EROEI calculations, like this photovoltaics report from the German Fraunhofer Institut, usually stop. Questions like: ‘How does the metallurgical grade silicon get to the manufacturing plant?’ are rarely asked. It just presumed to magically pop up in the warehouse at night, I guess. Just like the clean sheets of aluminum and glass.

What about mining, transportation, refining, smelting and melting then? What about the machines, trucks, ships, excavators, dumpers built solely for the purpose of mining high purity quartz (or SiO4, the raw material for silicon production)? What about the energy cost of getting rid of those four pesky oxygen atoms sticking to one silicon atom? What about glass manufacturing, transport and waste? What about aluminum production, bauxite mining and electrolysis for the frame? And the mining machines, trucks, ships, excavators, dumpers built solely for this purpose? At best, EROEI calculations presume these logistic factors to be a given, and calculate with the direct energy cost of assembling a panel plus the grossly underestimated direct energetic costs of their raw materials. What about the rest above? You guessed right: it’s all left out, and presumed to be just… there. Puff! Magic!

Mind you, there is nothing new or revolutionary in all this solar craze. We are using a manufacturing technology invented in 1915, applied to silicon monocrystal pulling in the 1950’s. Hoffman Electronics has created a 10% efficient commercial solar cell in 1959, and raised its efficiency to 14% just one year after. Sure, at first these panels were so outrageously expensive that only NASA could afford them... Still, they were there. We knew how to make them already, 64 years ago.

And as preposterous to believers of progress this may all sound, my father was a toddler still when solar panels were flying in space already.

Large scale commercial production of the technology only depended upon finding an abundant energy source to power the processes of smelting aluminum, melting glass and rising monocrystalline silicone. At scale. No ingenious inventions were needed. Since we are talking about a technology older then most of us, the last of low hanging fruits in manufacturing cost elimination were reaped 20 years ago, already. The only question was, what energy source to use? What country on Earth could power all these brutal energy hogs, not to mention providing skilled labor willing to work 12 hours a day, 6 days in a row…? Hmmm…

The answer was simpler than most thought: China. A country now responsible for burning 50% of all the coal mined on this planet. The high heat and stable electricity from the dirtiest of fossil fuels, combined with unfettered globalization is what made cheap solar panels possible. What we are seeing is pure economies of scale, combined with relatively cheap energy and abundant raw materials, all found in one place. Or rather, what used to be cheap energy and abundant raw materials…

---

But, but, but… But what about the unstoppable march of technology raising solar panel efficiency through the roof? Sure enough, performance is a key factor to a high energy return on investment. Efficiency gains have not come for free, however. The answer to the question how solar cells got so efficient recently lies in their material composition. Sorry to dispel this one too, but there is really nothing magical in this. (OK, just a little.) The solution was to add expensive to mine, corrosive and poisonous materials like Gallium to the silicon wafer; complemented by an increased level of complexity (new manufacturing methods, more layers, etc.) And while these metals make up only a minuscule amount of the total weight of a panel, without them we would be back into a rather disappointing efficiency range (practically half of today’s best in class). The story didn’t end with Gallium though. There are several other elements involved in making today’s top-notch multilayered (so called multijunction) super efficient panels reaching 45% efficiency. Take a look at this club sandwich for starters:

Composition of a multijunction cell. Ga = Gallium, Al = Aluminum, In = Indium, P = Phosphorous, As = Arsenic, Ge = Germanium. Now try and imagine recycling this… Source

These cells are not mass produced for a very good reason: cost and complexity. What makes up most of the sales nowadays are thus simple monocrystalline silicon panels with a typical efficiency of 15–18% (measured in sunlight converted into electricity with a best in class performance of 24–25%). However, these cells still require a rare metal called Germanium in the manufacturing process to reach such performance levels. Again, a finite mineral controlled by China, but that’s a story for a different day. Something, without which we would experience a significant drop in efficiency (not to mention copper wiring, the replacement of which is simply a non-starter). It’s not particularly hard to understand that solar panels produced by the millions (something which is only expected to ramp up as the “energy transition” unfolds) would quickly deplete existing Gallium and Germanium reserves, and thus would force us to return to a more simple material composition. It looks like that raising efficiency has hidden costs and comes with strings attached…

Sorry, no metal, no magic.

Wherever there is manufacturing though, there is decommissioning too. As these panels reach the end of their lifecycles, typically 25 years down the road, they will be scrapped. The question poses itself: what to do with all this waste? (Remember, the materials found in a panel are often poisonous to life, so it’s imperative that we get rid of them safely.)

there could still be a manageable 4 million tons of solar panels for scrap by 2030, but the amount could surge to over 200 million tons globally by 2050 as solar power deployment booms.

That is a considerable amount, to say the least... Then someone will surely recycle them! Right? Well, according to the article cited above: “the panel components with the highest value are aluminum, silver, copper, and polysilicon. Silver accounts for about 0.05% of the total weight but makes up 14% of the material value”. This already shows how tiny amounts of expensive metals can get the spotlight. However the same is not true for Germanium, Gallium, Arsenic and all the other additives used in wafer production. They are added literally in trace amounts to silicon: in the range of parts per million (not even a fraction of a percentage). No wonder that recycling companies are focusing on none of those, but on silver and aluminum instead. Found in much greater amounts, these metals can be extracted by using aggressive and highly poisonous solvents, after the panels were crushed to dust.

The process using these solvents retrieves more than 90% of the silver and aluminum in a period of 10 minutes. The silver recovered is high purity, which means that it can be reused in industrial settings.

Heck, this is not unlike mining a silver ore with 0.05% grade. Here though, instead of rocks coming from a mine nearby, we would have to deal with solar panels moved across the surface of the globe. So if you thought mining had an issue with moving ever larger amounts of rocks as ore quality slowly degrades, wait until solar panels need to be recycled… As the energy depletion (or rather the ‘big mad energy scramble’) unfolds, however, all this will become even more “challenging”…

Although at first glance glass seems to be recycled, the use of recovered glass is limited to less valuable products, with high transportation costs being an issue

Energy is the economy. Since you will always have to spend more energy on recycling due to higher transportation cost, then why bother…?

Yet another thing to consider is the degradation of materials with every round of recycling. While metals in theory can be recycled without a drop in quality, this is only valid for clean environments in a laboratory using 99.99% pure metals. Once you buy up scraps from all around the world, containing ‘who-knows-what-type’ of aluminum-alloys, ‘only-God-knows’ type of silicon wafers, and ‘no-clue’ branded glass, plus all the dirt and contamination two decades can leave behind, you can only mix a tiny portion of this curious stuff into brand new aluminum, glass, and silicon without degrading quality to the point of unreliability.

---

After discovering that there is no viable way to recycle them at scale (forget even 90% rates) and understanding the predicament of running out of exotic materials to keep performance (and hence energy return high enough), here is an even bigger riddle to solve: how to manufacture these cells once fossil fuels are out of the picture…? How to power trucks and excavators bringing up bauxite and quartz from the mine…? How to provide mega- and gigawatts of stable 24/7 electricity to smelt aluminum and grow silicon monocrystals…? How to melt glass without natural gas…? How to ship the raw materials and panels halfway around the globe…? And please don’t get me started on fusion…

Let’s be honest — at least to ourselves — photovoltaic panels are not, and never were, a ‘viable technology’ (meaning something which is able to reproduce itself). Coal could be mined by coal at least until the early 1920's, but solar panels — so far at least — failed to produce new solar panels. As long as we have the luxury of having ample fossil fuel supplies providing the energy, and delivering the raw materials needed for the magic of solar panel manufacturing to happen, we will keep producing those panels, disregarding material limits or real energy inputs for that matter… and then?

There are no projects aiming to make solar panels by using “renewable” electricity and recycled materials alone for a very good reason: we haven’t got the slightest clue how to actually do it. Photovoltaic panel manufacturing requires high amounts of stable electricity, diesel and natural gas. The panels themselves are practically unrecyclable: all of their glass, along with a good portion of their metal content, will be always left behind. Arsenic leaching from broken panels will contaminate the soil for decades if not centuries to come. Solvents used in recycling (if tried) will do the same. What material do gets salvaged will be so contaminated (or of unknown composition) that massive portions of new materials needs to be mixed with it to make it suitable for new panel production. High-tech, high efficiency panels require exotic and rare metals, the supply of which is limited to only a few countries, not to mention the obvious fact, that one day we will use up all economically available reserves. If you were looking for a technology under the term ‘dead end’, then Dear Reader, look no further.

Having all this knowledge, it is much more appropriate to think of a solar panel as an energy carrier, like a uranium fuel rod. A consumable, which gives power based on what’s the weather like, but something which will eventually be destroyed in the process and have to be decommissioned. With its use, just like with uranium or fossil fuels, we are actively living up a set of finite material resources and turning them into dangerous waste — as opposed to tapping into an ‘infinite flow of energy’.

We are trading metals and fossil fuels for a little bit of extra energy, something so predictably intermittent that electric grids across the world are now having problems admitting more and more of it into the grid. As a result, there is now an oversupply of solar panels, filling up warehouses, waiting to be installed. Not a sign that this is the way forward, but who am I to tell?

Is there a better way to use solar energy then? Sure! For starters, we could just drop this crazy habit of turning finite materials into junk and destroying entire ecosystems in the process. Let trees, shrubs, plants, animals use the power from our central star and start healing the damage we have done. Since that is a tall order, and we need to do something before we can let that happen, why not use the Sun as a source of heat instead?

Why not use a drum painted black to heat water on your roof for starters? Or how about building a solar oven? Similarly, we could utilize large polished metal dishes to concentrate sunlight into a single spot, and convert it into mechanical work by using simple heat engines, like a Striling engine.

Sure enough, it won’t look as high tech as a solar panel (it’s much like something from a steam punk retro-future movie), but one could do an awful lot of things with it. For example, attaching a generator — from a car sitting idle due to a lack of affordable fuel — you could produce a stable 12 Volts of current. If built bigger, it can mill seeds for human consumption, lift water from a well, and do all sorts of useful work. Since the heat source is external, you can literally lay a fire underneath it, should the sun go down. 

Of course, the one above is but the simplest in design, with some engineering work however, one could devise much more efficient heat engines. These simple machines could easily serve as a basis for a new energy paradigm. Not as exuberant as the one we have now, but way better than having to watch the entire energy-economy relationship falling into ruins. No matter if done on a large or small scale, the core principle of this new paradigm based on solar heat remains the same though: instead of using exotic materials from far away lands, mined, carried, smelted and melted by fossil fuels, one could build such a device from scratch. Even at home. No quartz crucibles heated to 1400 degrees under a protective atmosphere is needed. No toxic tailings. No destroyed ecosystems. Just you, and your hands. And a lot of scrap left behind by this failing civilization, waiting to reborn as something really useful.

Until next time,

B