Updated: May 26, 2020
Do you want to know something crazy?
We get as much energy from the sun in 6 months than is embodied in all the non-renewable sources that ever existed on Earth put together (The World Counts, 2020). Then again, this fact should not come as a shock. Nearly all our energy sources are, in the most basic sense, embodied direct radiation from the sun; the notable exceptions being tides, geothermal energy and uranium. It is all-you-can-eat energy that is greener than any buffet will ever manage!
Trends in Solar Energy Consumption
Though we are getting better at making use of this vast resource, we still have quite a way to go. The following graph shows a breakdown of worldwide energy consumption by source in terawatt-hours. The watt is a unit of power which can be adapted to energy units by multiplying by time; the prefix terra signifying 1012, or a trillion. In 2017, solar accounted for a measly 1.73% of the total world electricity use (ibid). It is not all bad news as solar has shown rapid expansion over the past three decades- from 0.00 TWh in 1982 to 584.63 TWh in 2018. For reference, it took from 1983 to 2000 to go from <0.01 TWh to 1.13 TWh while the jump between just 2017-18 was around 131.11 TWh! This can be attributed to: (1) falling costs for solar, 100-fold decrease since 1976 (2) significant investment in solar technology, accounting for 47% of global finance in renewable energy in 2016 and (3) safety, especially relative to the death trap that is coal. I encourage you to use the link in the sources section to visit the good folks over at Our World in Data and learn more through some great graphs like this one:
So why does this matter? It matters because energy accounts for around 60 percent of total global greenhouse gas emissions (UN Sustainable Development Goals). It matters because we aim to “ensure universal access to affordable, reliable and modern energy services” (SDG 7.1). It matters because if we endeavour to “increase substantially the share of renewable energy in the global energy mix” (SDG 7.2), we will have to think about how to create, manage and scale these technologies appropriately. A couple of things must happen if we will do this properly. First, we must be mindful of the production processes behind these technologies. This will help us to hedge production risks (e.g. materials scarcity) but also ensure we are thinking of the end-of-service for these technologies from the very beginning. Otherwise, we may create more problems in the process of problem solving. A prototypical vicious cycle.
A Very Brief Overview of Silicon Cell Production Processes
Silicon cells are considered the first generation of solar cells and have at least 80% of the solar market share though I have seen estimates as high as 90%. Given their status as a poster child for green technology, you may be surprised to know that their production is very energy intensive. In fact, step number one is to combine sand and coal (gasp). Now you know why I cringe if somebody tries to talk in absolutes- even about solar. Due to the interconnectedness of our world and largely because research, commercialization and public policy seem to inhabit different planes of existence, nothing is ever quite as simple as it seems. It is a tall order to unravel all the interdependencies. This is precisely why we need more polymaths in general, but we especially need more in public office!
Subject the starting concoction of coal and sand to a toasty 1800°C (3272°F) and you get carbon monoxide and metallurgical grade silicon. This is 98% pure silicon. That sounds incredibly significant, but it is not nearly enough. For solar applications, the purity needs to be at least 99.999%. For the semiconductor industry, the purity threshold is 99.999999999% to produce computer chips so in some way, you can say the solar guys got off easy! Not that easy though because the metallurgical grade silicon is treated with hydrochloric acid and through additional purification steps, becomes high quality polysilicon*. The rods of polysilicon are then further processed into multicrystalline (AKA polycrystalline) or monocrystalline silicon*. The latter requires a more energy intensive production process. As the names of these types suggest, monocrystalline silicon is more uniform as it features only one crystal whereas the multicrystalline type is an amalgam of many crystals, often in different orientations. This has implications for the efficiency of the final cells as the boundaries between the distinct crystals can hamper the ability of the electrons to move out of the cell into the associated circuit. So, there is a bit of a trade-off here between the energy intensiveness of the production method and the efficiency of the cell.
The resulting bars of monocrystalline silicon and blocks of multicrystalline silicon are sawed into individual sheets, or wafers. This is to ensure the product is not too thick as to hamper the absorption of light. Though this step produces some waste of the product, the scraps can be recycled to use in earlier steps. The wafers are then further processed in a step-wise fashion through exposure to a variety of reagents: NaOH (for texturing and to mitigate surface contamination after cutting into wafers), isopropanol alcohol, phosphorus, chloroflurocarbons (abbreviated CFCs), silane (SiH4) and ammonia (NH3), Silver (as a paste), and Aluminum. These processing steps, though crucial to ensuring the proper functioning of the final cell will not be covered in any more detail here so that we can move on but I encourage you to have a look at PV Education in the sources for more information, photos and videos.
Source: Energy Sage
The Externalities of Silicon Cell Production
You will notice that a lot of these materials are not exactly eco-friendly as it were. For example, you may recall chlorofluorocarbons were banned in 1996 due to their role in thinning the ozone layer (Wikipedia). One would hope all these substances are handled in ways that diminish their environmental impact as much as possible. But given the growth of the Solar PV industry and the silicon cell’s market share, these considerations cannot simply be ignored. Especially since transitioning economies tend spend a higher proportion of their GDP on energy expansion compared to their energetically developed counterparts (Our World in Data, 2020). Furthermore, the smuggling industry for some of these compounds, particularly CFCs, is still alive and well (Wikipedia) and this is obviously unregulated in any real sense so it is another risk factor we must address.
It is also worth noting that silicon is the second most abundant element in the Earth’s crust, so these production methods are not as sensitive to resource scarcity. The availability of silicon and the physical properties that make it amenable to such manipulation* can help at least partially explain the current market dominance. However, the laws of supply and demand still apply. The PV industry’s usage of silicon now far exceeds the integrated circuit industry which has “pushed up the price of raw crystalline silicon feed stock from under $20/kg to over $200/kg” (PV Education). The increase in materials prices may represent a further hurdle in the uptake of these technologies. Additionally, the energy intensity of the processes and the environmentally detrimental nature of some of the reagents has already been highlighted.
One key metric that we must also consider is energy payback time or the slightly fancier term, energy amortization time, the amount of time a solar plant must operate to offset the energy required for its construction. For silicon solar cells, the payback time is generally on the scale of 1.6-4.5 years and varies based on the production process used and the location of the installation. For example, for rooftop solar in cities, the payback time can be around 8-10 years due to less offset and higher construction costs in these areas. Considering silicon cells’ large market share and the growth of the PV industry in general, this means more energy has currently been spent making these cells than has been harnessed from them. To prevent us from grossly oversimplifying the matter, keep in mind that silicon solar cells have long lifetimes, around 25-35 years. Of course, this comes with the additional caveat that most solar manufacturers only provide a warranty of up to 25 years. Additionally, the cells themselves degrade at a rate of approximately 1-1.5% annually so by the time the warranty expires, the cells are running at about 80% of their original efficiency. Nonetheless, over the lifetime of a solar installation, it has made much more energy than was required to produce it. Thus, though the PV industry has so far been a net electricity consumer, we cannot categorically say the trend will continue in the future. For the record, let me establish that the purpose of this piece is to not bash the solar PV industry. I am a passionate advocate for sustainability and green technology but still recognize we cannot glaze over how the sausage is made, so to speak. So, what are the alternatives within the solar PV space?
Second Generation Solar Cells
The alternatives to silicon cells have been plagued by issues with resource toxicity, scarcity, scalability and efficiency. The second generation of solar cell technology is represented by three types of cells collectively referred to as thin film cells. These include Cadmium Telluride (CdTe), CIGS (copper-indium-gallium-selenide) and amorphous silicon. For the reasons highlighted above, I think there is a better way forward, but I think a brief discussion is warranted to reach my goal of a nuanced view of solar PV as promised.
Cadmium’s infamy was cemented by its inclusion in the UN Environmental program’s list of top 10 hazardous pollutants (Parsons & Dixon, Field Guild to the Elements, 2013). In Japan, there is a disease characterized by bone and joint pain called itai-itai, literally meaning ouch-ouch, and occurs due to prolonged exposure to Cadmium (ibid). What I find most interesting of all is that the detrimental health effects of Cadmium have been recognized since the 1960s due to litigation against dumping practices in the mining industry. Yet this knowledge seemed to have done little when it came to the initial phase of considering Cadmium-based cells as a potential for solar applications. Certainly, the physical attributes that make any of these elements amendable to solar PV should not be considered in isolation; that is, there is more to appropriately scaling a technology than simply assessing its production costs and resulting efficiency. This is another reason intelligent systems design and circular economy should be a priority! It is an undeniable fact that CdTe now displays efficiencies that rival silicon cells and their energy amortization time is on a scale of months rather than years. Additionally, as the production processes of these cells do not have high purity requirements to start with and they use lower quantity of materials, they are less energy intensive. However, the health hazards associated with both elements and the concerns around Tellurium supply being able to support demand in the solar industry (ibid) means that this does not seem like a viable option to me at all for the long term plan for solar.
The CIGS cells face similar issues though their production processes are significantly more complicated than CdTe. Again because of my self-confessed bias towards these technologies and consideration of my readers’ time, I will not go into too many specifics here. If you are interested, I cannot recommend Introduction to Solar Cells via Coursera highly enough. The last of the group, amorphous silicon is not even an option in my view because of the issues with stability levels and the unremarkable efficiencies that would make scalability a real issue. With population projected to rise to a whopping 11.2 billion by the end of the century (Roser, 2019), we need to think critically about scalability balanced with resource scarcity and the availability of viable substitutes (as shown by Graedel et al, 2013, figure below). Though it has been said that these technologies have their place in the renewable portfolio, I would argue that rather than introducing such technologies that do not seem amenable to a circular economy due to their toxicity, we are far better off finding ways to improve the efficiencies of the third generation of solar cells which I will address next.
Third Generation Solar Cells
The third-generation solar cell classification encompasses many different cells with their own unique hurdles. This group includes organic cells, multi-junction cells, perovskite cells and quantum dot cells. I want to discuss the organic cells in a little more depth but for the time being, I will do a extremely basic survey of the others. Multi-junction cells are used on space probes and are made of Gallium Arsenide (GaAs). The multi-junction cells are the most efficient we currently have but the price tag is staggering – hence their use in space applications where emphasis on the longevity and efficiency of the instruments justifies the cost. Perovskite cells are a burgeoning technology which has shown significant efficiency gains in a short time, from around 3% in 2009 to a relatively recent record of 22%. Though the energy payback time is good, and they can be produced quickly, the main drawbacks of these include their lead content and some instability that can pose a challenge to their longevity. Admittedly, I know the least about the quantum dot cells though I am keen to improve my understanding of them if anybody wishes to recommend any resources! Although I will say the Wikipedia article on the subject is a good start. Mostly relevant to the present purpose is the observation that quantum dot solar cells also utilize heavy, toxic metals, something I have thus far been very vocal about. One thing I did find interesting is the inclusion of a company called UbiQD which is working on employing a non-toxic semiconductor to “provide semi-transparent windows that convert passive buildings into energy generation units, while simultaneously reducing the heat gain of the building” (Wikipedia). To me, this demonstrates yet another reason we must prioritize safe, sustainable materials – because doing so helps increase the scope of their usage. Imagine that!
The (General) Case for Organic Cells
As the old saying goes, saved the best for last. I will focus for the time being on a subset of these, the polymer solar cell. To avoid getting accused of having a limited view that does not account for the shortcomings of this technology, I will start with the disadvantages first. They are not as efficient such that it takes 4x as much area for a module of polymer cells to generate the same amount of power as silicon cells. Additionally, the lifetimes of the cells are not as good as some of the other alternatives. Finally, there are some concerns about the stability of the cell as organic molecules start to deteriorate when a light is shined on them. Let me temper that by saying recent years have shown improvements in both efficiency and lifetime so I suspect some more rigorous research & development, potentially funded in part from the money previously spent on less environmentally friendly technologies, would help to improve this situation.
The advantages of polymer cells include:
(1) production process involves, very crudely speaking, glorified ink-jet printing and other related printing methods so they can be manufactured at break-neck speeds not possible with other technologies
(2) flexibility of the cells which allows them to be used in more places,
(3) they do not feature scarce or toxic elements and
(4) can be installed very quickly because it is essentially like unfurling a giant roll and fixing it in place.
Perhaps my only qualm with the production process is the use of plastic, PET, substrates. However, with the 2016 discovery from Yoshida et al of Ideonella sakaiensis, a bacterium that breaks down PET (see sources), I can fancy a future where this could potentially be integrated into the end-of-life plan for such cells. Though I profess, I have no idea how the ink used in the process would play into this and would love to have that conversation!
The future of polymer cell processing is to get to an in-line production process where everything is completed on one machine such that you put raw materials in on one end, and get a fully finished product out on the other end with all the processes being done at the same time. Of course, this would greatly increase the production rates though the speed would be determined by the slowest step. It would also introduce risk in the form of sensitivity to faults at any one point. I would love to regale you with all the details of the production process but I fear that my decision to expound upon silicon cells leaves me less freedom to tell you more about the polymer cell lest I really start testing the limits of everyone’s attention. This of course, is a classic example of how expending resources due to the market share of a product means you have less to go around even if an underdog can potentially better serve your needs, albeit with a bit of work. Yes, I purposefully did that to illustrate a point. No, I have no regrets because I have already signed up for Organic Solar Cells – Theory and Practice, again through the Technical University of Denmark. I also have some ideas from an article in Nature Materials about organic metal ionic semi-conductors which perhaps could be useful in addressing some of the issues surrounding the efficiency and longevity of these solar cells. I do, however also have a duty to not mislead or misinform so I will continue to explore those options and further develop the ideas. So, do not despair because you know there will be another one of these just about organic solar cells in due course!
Thank you for making it this far! Although I tried my best to ensure this article was readable but not excessively over-simplified, I must acknowledge some technical details were purposefully left out in the interest of time and clarity. I would endeavour that the material provided was nonetheless illuminating (if you pardon the pun). I have demonstrated solar has the potential to satisfy our energy needs in theory, but I suspect it remains to be seen how that will evolve as the population increases. Therefore, it is imperative we diversify our renewables portfolio and supplement solar with other renewable technologies especially in places where solar is not the best option and/or until we find better means for battery storage. I have not been shy about my biases against toxic materials irrespective of efficiency gains; we simply cannot afford to continue acting as if human and environmental health are two lofty, fundamentally separate ideals. The other key point is that we cannot consider the scientific merits of any technology only in the lab but must unite research, commercialization and public policy in a coherent context. Silicon cells have provided us with a great starting point in the fight against climate change; I would dare to suggest the future is organic.
This article draws heavily from Introduction to Solar Cells, from the Technical University of Denmark via Coursera, delivered by Morten V. Madsen. I am immensely grateful for his sizeable contribution to my knowledge which has inspired me to relay this information in a way that is more accessible to a wider audience; perhaps those who due to time constraints or preconceived notions of interest or ability may not have considered the course. Any opinions provided are not representative of Coursera, DTU, course staff or publishing platforms.
Morten Vesterager Madsen (n.d) 'Introduction to Solar Cells', Coursera, Technical University of Denmark, retrieved 24 April 2020, < https://www.coursera.org/learn/solar-cells>.
The World Counts. (n.d.) Renewable Energy Resources. Retrieved 24 April 2020, from https://www.theworldcounts.com/challenges/climate-change/energy/renewable-energy-resources
Ritchie, Hannah & Roser,Max (2020) - "Energy". Published online at OurWorldInData.org. Retrieved from: https://ourworldindata.org/energy [Online Resource]
United Nations Sustainable Development Goals (n.d.) 7- Affordable and Clean Energy. https://www.un.org/sustainabledevelopment/energy/
Energy Sage (12/12/19). Monocrystalline and polycrystalline solar panels: what you need to know https://www.energysage.com/solar/101/monocrystalline-vs-polycrystalline-solar-panels/
Chloroflurocarbon (n.d.). In Wikipedia https://en.wikipedia.org/wiki/Chlorofluorocarbon
Honsberg, Christina & Bowden, Stuart (2019). Solar Cell Production Line. PV Education. https://www.pveducation.org/pvcdrom/solar-cell-production-line
Parsons, Paul & Dixon, Gail. (2013). The Periodic Table: A Field Guide to the Elements. Quercus.
Roser, Max (2019). Future Population Growth. Our World in Data. https://ourworldindata.org/future-population-growth
Graedel, T. E., Harper, E. M., Nassar, N. T., & Reck, B. K. (2015). On the materials basis of modern society. Proceedings of the National Academy of Sciences of the United States of America, 112(20), 6295–6300. https://www.pnas.org/content/pnas/112/20/6295.full.pdf
Quantum Dot Solar Cell (n.d.). In Wikipedia https://en.wikipedia.org/wiki/Quantum_dot_solar_cell
Yoshida, S. et al. A bacterium that degrades and assimilates poly(ethylene terephthalate). Science 351, 1196–1199 (2016). https://science.sciencemag.org/content/351/6278/1196?ijkey=9f4f5d0a7cc9c31f7a4c2ec32b2aad1f556782fc&keytype2=tf_ipsecsha