Recently, the Washington Post published an article1 discussing Mark Jacobson's research2 positing that a strategy of 100% wind, water, and solar power can provide all the energy needs of the U.S. with no pollution or emissions.

Many on Twitter have been pushing back on this, including a thread from Brian Gitt with over 600 retweets and 1300 likes. I am skeptical of the first falsifiable claim in the thread (emphasis added):

Solar panels are manufactured using minerals, toxic chemicals, and fossil fuels. In fact, solar panels require 10 times the minerals to deliver the same quantity of energy as a natural gas plant. Quartz, copper, silver, zinc, aluminum, and other rare earth minerals are mined.

Is the bolded statement true?

  1. Patel, Kasha. "A year after Texas cold spell, study shows renewable energy could help prevent blackouts". The Washington Post. 20 February 2022.
  2. Mark Z. Jacobson, Anna-Katharina von Krauland, Stephen J. Coughlin, Frances C. Palmer, Miles M. Smith. Zero air pollution and zero carbon from all energy at low cost and without blackouts in variable weather throughout the U.S. with 100% wind-water-solar and storage. Renewable Energy, Volume 184. January 2022.
  • 3
    It's a pretty ambiguous figure as stated. I suppose it means to compare the "mineral" usage involved in the construction and maintenance of a gas plant throughout its entire life, as compared to that of a solar farm, divided by the lifetime energy production of each? Meanwhile, I'm not sure how sound it is to lump all minerals into one bucket when some are certainly going to be more costly or toxic than others.
    – Feryll
    Commented Feb 22, 2022 at 3:31
  • 48
    I do not see the point to the comparison. Once built the solar panel produces electricity. Once built a plant that turns natural gas into electricity just sits there unless nature gas is continually fed into it. Commented Feb 22, 2022 at 6:07
  • 10
    "Solar panels are manufactured using... fossil fuels" seems pretty circular. How are they comparing minerals? By weight? Is it useful to compare a kg of silicon to a kg of steel? Which solar panels and gas plants are we comparing? Is Brian Gitt someone who's opinion is significant?
    – Schwern
    Commented Feb 22, 2022 at 6:43
  • 32
    "Quartz, copper, silver, zinc, aluminum, and other rare earth minerals are mined." correct me if I'm wrong, but none of those minerals (quartz, copper, silver, zinc, aluminium) is a rare earth mineral, so the whole sentence is incorrect.
    – Marandil
    Commented Feb 22, 2022 at 16:14
  • 22
    The choice to define underground reserves of methane as "not a mineral" key to this argument: It's reasonable at a glance, since methane is a gas and we think of minerals as solid. However, arguably if we are trying to measure the resource usage of different power generation approaches, we should include it in some way. Commented Feb 22, 2022 at 19:39

3 Answers 3


Basically correct; however...

If by "solar panels" he means the classic photovoltaics which turn sunlight into electricity, the number is about 5 to 20 times more material depending on what type of panel and installation. If you count concentrated solar power, that number is more like 1 to 2 times more.

However, no context is given whether this is relevant to Mark Jacobson's conclusion about 100% renewable power. The claim is a small part of a broad argument shotgunning cherry-picked information at the readers to convince them that natural gas (or sometimes nuclear power) is better for humanity than solar power. That is not true.

Skipping natural gas extraction

LShaver found the author's original article with references: Solar’s dirty secrets: How solar power hurts people and the planet. Here's the expanded quote.

Solar panels are manufactured using minerals, toxic chemicals, and fossil fuels. In fact, solar panels require 10 times the minerals to deliver the same quantity of energy as a natural gas plant.

Taken in context, Mr. Gitt is arguing that solar panels are worse for people and the planet than natural gas... or is it nuclear power? He keeps switching.

He references Table 10.4 of the US Department of Energy's Quadrennial Technology Review 2015.

enter image description here

On first glance it seems like "10 times the minerals" is more like hundreds to thousands!

However, as an argument comparing natural gas vs solar power it is meaningless because it's not an apples-to-apples comparison. "Upstream energy collection" means all the extraction and mining associated with power production. Hydro, wind, solar, and geothermal don't have upstream costs; it comes to them, so it's easy to calculate the total cost. Natural gas does have upstream costs. By leaving out the material costs of natural gas extraction, processing, and transportation the argument has no meaning.

Using the complete energy production lifecycle

Cribbing from jkej's answer we can use the UNECE's Life Cycle Assessment of Electricity Generation Options from 2021 to get our answer.

Natural gas without carbon capture

enter image description here

Natural gas with carbon capture

enter image description here

We're interested in the first bars, expressed in equivalent resource depletion. Note that the majority of the mineral cost of a natural gas plant is in natural gas production which Mr. Gitt's source ignores.

Since Mr. Gitt is very concerned about climate change, we'll assume his dream natural gas generator is using carbon capture, so 0.314 mg Sb-Eq.

What is a "solar panel"?

Solar power comes in two major flavors: photovoltaics (PV), which turn sunlight into electricity, and concentrated solar power (CSP) which use sunlight to heat a fluid. Both use something which could be a "solar panel". Within those two major flavors are variations: panels mounted on the ground vs on a roof, what they're made out of, and for concentrators how they're concentrated. This affects their mineral cost and where that cost is derived.

PV, ground-mounted, poly-crystalline silicon

enter image description here

PV, roof-mounted, poly-crystalline silicon

enter image description here

PV, ground-mounted, copper-indium-gallium-selenide

enter image description here

PV, roof-mounted, copper-indium-gallium-selenide

enter image description here

CSP, parabolic trough

enter image description here

CSP, central tower

enter image description here

There's quite a spread from roof-mounted poly-crystalline silicon at 7.21 mg Sb-Eq to just 0.336 mg Sb-Eq for central tower concentrated solar. This tells us that "solar power" can use anywhere from 1 to 23 times more materials than natural gas. Or 1.4 to 30 times if Mr. Gitt doesn't care about how much carbon the natural gas plants are dumping into the atmosphere.

However, "compared to photovoltaics (PV), solar thermal, or concentrated solar power (CSP) technologies are a rather niche market" concentrated solar is a very small percentage of the market. It may be a larger part of the market in the future, but for now it's mostly PVs.

So it would be most fair to compare only the PVs with a spread of 1.66 to 7.21 giving us 5 to 21 times more materials compared to natural gas with carbon capture.

Solar panels can be recycled, their bases reused

The 5 to 21 times more number is a bit misleading once reuse and recycling are taken into account.

A natural gas plant could potentially be recycled; however, 70% of the material cost is in the extraction. Solar panels can be recycled, and the EU even mandates it and other countries could do the same.

The modules present on today’s market belong to two different categories, silicon and non-silicon based, which determine the recycling process to be used.

For silicon-based modules, aluminium frames and junction boxes are dismantled manually at the beginning of the process. The module is subsequently crushed and its several components are separated, allowing recovering up to 80% of the panel. Since a large quantity of these modules is composed of glass, it is not unusual for glass recyclers to be able to intervene in the recycling process.

Non-silicon based panels require the use of diverse recycling technologies. Cadmium telluride (CdTe) panels e.g. – a particularly common type – are first crushed into different fractions, much like non-silicon modules. But they also use chemical baths to separate the various semiconductor materials, allowing for the recovery of 95% such components. Recycling technologies for this type of panels have been widely increasing in recent years. For copper indium selenide (CIS) and Copper indium gallium (di)selenide (CIGS) photovoltaic modules similar chemical bath treatments apply.

Mr Gitt seems to disagree with the EU.

13/ Solar panels last only about 20 to 25 years. And they are difficult to recycle because they’re made with toxic chemicals.

In addition, much of the material cost of a solar panel is in building the platforms upon which they are installed. Once built, these platforms can be reused for new, upgraded solar panels while the old panels are recycled.

But is it better for climate change?

Mr. Gitt is arguing that "solar power hurts people and the planet" and that natural gas (or nuclear power, whichever fits his argument best at the moment) is a better alternative.

If we’re serious about tackling climate change, protecting the environment, and helping impoverished people around the world, we need to stop chasing fantasies about solar and wind energy.

He attempts to do so by shotgunning every criticism he can find about solar power at the reader, but never does a total analysis.

If one were to do an honest evaluation of the material cost of natural gas vs solar power, one would also have to take into account the material of climate change; both its damage and our mitigation efforts.

Life Cycle Assessment of Electricity Generation Options does provide a "Climate Change Total" in Table 14.

  • Natural gas without carbon capture: 434g CO2/kWh
  • Natural gas with carbon capture: 128g CO2/kWh
  • Solar, Poly-Si, ground-mounted: 36.7g CO2/kWh
  • Solar, CdTe, ground-mounted: 11.9g CO2/kWh
  • Solar, CIGS, ground-mounted: 11.4g CO2/kWh

For all of them, this is almost entirely due to fossil fuels. For natural gas this is inherent to burning fossil fuels to generate electricity. But for solar this is due to the current energy mix used in the production and installation of solar panels and inverters.

...about half of greenhouse gas emissions can be attributed to silicon manufacturing (from primary production to solar-grade refining), while the reminder of emissions is split between the rest of the module, site preparation, and electrical equipment (inverters).

If we stick with natural gas, its impact on climate change gets worse as carbon continues to build up. If we switch to solar and decarbonize our energy production, the carbon necessary to produce solar panels will decrease and their impact on climate change will decrease, lowering their true cost to humanity in the long run.

If Mr. Gitt were serious about helping people, he would put this in his article. Instead he advises we invest in Liberty Oilfield Services, an oil and natural gas fracking company.

Solar panel material cost does need to be improved

The Quadrennial Technology Review has this to say about the "RDD&D opportunities in clean electric power technologies..."

Solar (photovoltaic and concentrating solar power): Reduce solar [photovoltaic] and [concentrated solar power] manufacturing and capital costs, reduce PV soft costs, improve grid integration—including with storage solutions, and identify and develop new PV materials and devices, particularly with abundant and environmentally-benign materials

So there are plenty of improvements to be made.

Natural gas is a mature technology, and, while it will improve, we don't expect radical changes - it's about as good as it's going to get. Solar power is an emerging technology still undergoing radical improvements in materials, cost, and efficiencies; the industry is still growing, and the power grid is still in the early stages of adapting. One must be careful when doing a flat comparison between mature and emerging technologies, especially using information which is seven years out of date.

Rare earth minerals are not rare

One of the rhetorical slights of hand used by people to argue against renewables is their use of "rare earth elements" with the tacit assumption that these are, well, rare and we're going to use them all up!

They aren't particularly rare, some are more abundant than copper. They can, however, be diffuse, making economically viable deposits "rare". If demand increases due to increased use in renewables, the price will go up and more deposits will become economically viable.

Natural gas plants also need rare earth minerals for their construction, particularly for their steel turbine blades.


Yes, mostly true

(although the relevance of this particular metric could be discussed)

Last year, the United Nations Economic Commission for Europe published a report titled Life Cycle Assessment of Electricity Generation Options. As the title suggests, the report features extensive life-cycle assessments of all the major electricity generation technologies used today. The technologies are evaluated on a number of different criteria, such as greenhouse gas emissions, ionizing radiation, human toxicity, land occupation, water use, material resource use.

The material resource use criterion is essentially what this question is asking about. Below are figure 45 and 46 from the report, showing the material requirements per MWh for the different electricity options. In the first figure, all resource requirements have been converted to antimony (Sb) equivalents of Abiotic resource depletion potential (ABP). In the second figure the requirements are given in grams for select materials.

Mineral and metal requirement

Material requirements

Although the exact number will depend on exactly which resource and what type of photovoltaic technology is compared, the factor 10 claimed in the statement in question seems to be a fairly good approximation of the typical difference between natural gas and photovoltaics. Several of the metals mentioned in the questioned statement (aluminium, copper, zinc) are shown explicitly in the second figure, and they seem to be high for photovoltaics (though it's hard to tell the exact proportion compared to natural gas).

What is included in the Life-cycle Assessments?

Table 1 in the report gives a summary of what what aspects of each of the technologies are included in the life-cycle assessments. Below I have extracted the rows for natural gas and for photovoltaics. The scope seems to be quite comprehensive for both of them. The main thing that is excluded is "potential recycling of dismantled equipment", but this is true for both technologies.

Table header Included for natural gas Included for PV


It could be argued that focusing too heavily on material resource use when comparing electricity generation technologies is somewhat misleading. Reading the full UNECE report will show that photovoltaics are much better than natural gas on other criteria. Still, it's important to remember that all electricity generation options have environmental impacts, and weighing them against each other is not always straightforward.

  • 2
    +1, nice find. Does the report talk at all about natural gas infrastructure, i.e. wells and pipelines to bring a MWh of gas to the plant?
    – LShaver
    Commented Feb 22, 2022 at 17:58
  • @LShaver I can't find specific mentions of material resource use for "wells" and "pipelines". But in figure 9 on page 22, they break down impacts on categories like "Natural gas production", "Natural gas transportation", "Power plant" and "Electricity production". I would imagine that construction of wells and pipelines are accounted for under production and transportation. Strangely there seems to be almost no impact from transportation, but I guess that pipeline construction might fall under production. Transportation is perhaps only road and sea transportation.
    – jkej
    Commented Feb 22, 2022 at 18:15
  • 1
    @LShaver I think table 1 on page 10 also indicates that they are including "all infrastructure construction, operation, and dismantling", "from extraction to combustion". The main thing that is excluded is "potential recycling of dismantled equipment", but that seems to be the same for all technologies.
    – jkej
    Commented Feb 22, 2022 at 18:25
  • @jkej good catch. Maybe it's reasonable to exclude recycling; only about 17% of tech products get recycled (in the EU). Not sure about infrastructure.
    – Clumsy cat
    Commented Feb 22, 2022 at 18:30
  • "Life cycle impacts" in Fig 9, 10, and 22-25 give the total picture. Excellent find.
    – Schwern
    Commented Feb 22, 2022 at 20:21

Yes, but not taking into account the full picture

As Schwern already pointed out, making a direct apples-to-apples comparison is near impossible.

However I recently encountered the following graph, which does seem to support the bolded statement: enter image description here

It's in French, but with gaz=gas, photovoltaique = solar panels, beton=concret, acier = steel and cuivre=copper it should be easy to understand. These aren't exactly rare metals but it can already give an indication of the difference.

It's essential to note that this only takes the cost of building the installation into account, so for gas the running and maintenance costs will be much higher. Adding on extra effects such as recycling will most likely further shift the picture.

So if the original tweet is talking mainly about construction costs, then it is most probably correct, but this might not be the most relevant aspect to discuss.

The plot is based on the book "Mineral Resources and Energy: Future Stakes in Energy Transition" which seems to dive really in depth on the topic if it interests you (unfortunately I don't have access to it, but you might). I don't know where I first saw the plot, but the version shown here comes from this reddit thread.

  • 6
    Can you give a source for your plot please? I know you said the book, but you clearly found it elsewhere, since you cannot access the book.
    – Clumsy cat
    Commented Feb 22, 2022 at 15:11
  • 1
    Good point. It's from a reddit thread, so I added the link. Commented Feb 22, 2022 at 15:29
  • 2
    I was a bit surprised at this, but if you consider 1MW solar to 1MW gas it's quite believable that there is at least an order of magnitude different in the quantity of materials required to produce a given amount of energy (in terms of power plant construction)
    – Dave
    Commented Feb 22, 2022 at 16:07
  • 4
    This is the actual answer. It could use more source material, sure. Upvoted. (Note to Schwern: Like it or not, the answer is pretty clearly "yes", even if the pot is calling the kettle black ;-).) Oh, and @Dave's finds make it very plausible. Commented Feb 22, 2022 at 17:24
  • 4
    @Peter-ReinstateMonica it's a plot with the default Excell formatting, which used a paywalled dataset, has ambiguous labeling, and a badly formatted citation. Without any idea what the methodology is, I can't get anything useful from it.
    – Clumsy cat
    Commented Feb 22, 2022 at 18:23

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .