Persistent data nerd finds solar panels made in China produce 3x more carbon emissions than claimed!

A Lone And Obstinate Italian Data Crusader

The China-sized black hole at the heart of the world’s photovoltaic data might, in the context of the industry, seem obvious.

That didn’t make it any easier for Enrico Mariutti, an introspective but compulsive 37-year-old Italian from Rome, to convince others in the field there might be a problem. It was Mariutti who first made substantial efforts to flag the data discrepancies.

Like Greta Thunberg, Mariutti comes to the tale as an environmental obsessive passionate about facilitating the world’s transition from fossil fuels to cleaner forms of energy. Unlike Greta, Mariutti finished school and knows how to crunch through a data set. He holds a degree in geopolitics and global security, which, while unrelated to the field, has equipped him with enough quantitative skills to ensure he can recognize the difference between good and bad data.

Mariutti first noticed something wasn’t quite right with photovoltaic assessments about two years ago. He was preparing for an online renewables debate with Nicola Armaroli, a research director at the Italian Research Council. But being a data junkie, he decided to pour over the source material to try and figure out why. What he discovered unnerved him. The data didn’t reconcile.

“They [the data] showed how much solar photovoltaic systems used in terms of raw materials: silicon, aluminum, copper, glass, steel, and silver. Then I saw the carbon footprint. It just seemed way too small,” he told Environmental Progress.

According to his findings, the carbon intensity of solar panels manufactured in China and installed in European countries like Italy was off by an order of magnitude. An initial back-of-the-envelope calculation put it at between 170 and 250g of carbon dioxide per kilowatt hour (kWh), as opposed to the official estimate from the Intergovernmental Panel on Climate Change (IPCC) of 20-40g per kWh. Way off.

The scale of the IPCC’s undercount shocks once applied to the EU’s “clean” energy plans. Following Mariutti’s math, the esteemed scientific body underestimates the emissions from the EU’s solar installations built in 2022 alone by 5.4 to 7.6 million metric tons, equivalent to adding 3.4 to 4.8 million cars to the road.

By 2020, Mariutti felt compelled to make his findings public. He managed to publish an op-ed in Italy’s premier financial newspaper Il Sole. The piece argued it was wrong to describe an energy transition that depended on mineral-hungry tech, which “could double the exploitation of the earth’s resources within a few decades,” as a green revolution. It was a hit and went viral across Italian social media.

Enthused by what felt like a public mandate for his mission, Mariutti continued his research, firing off dozens of queries to the data compilers. Responses, however, were far from forthcoming. Until one day in November 2022, a leading Dutch renewables expert, Mariska de Wild-Scholten, replied.

Mariutti was delighted to have made an inroad. Wild-Scholten had been one of five key authors who had made significant contributions — she is named some 454 times — to the IEA report Life Cycle Inventories and Life Cycle Assessments of Photovoltaic Systems (2020) — a starting point for much government decision-making on net zero policy.

But what she told Mariutti was not reassuring.

De Wild-Scholten’s Secret Data Stash

In two emailed responses, Wild-Scholten said that, when determining the electricity consumption of silicon purification, which is used to make wafers, she rarely read scientific papers “because of low data quality, outdated data and or non-transparent data,” while saying little about her own preferred sources other than that it is based on surveys. Responses to other queries were no more reassuring.

Mariutti asked what she thought about 192 countries deciding their long-term energy strategy based on data that at the time reasonably underestimated the average carbon intensity of photovoltaic energy by one order of magnitude (40 vs. 250 gCO2/kWh). Her reply invited more questions than answers. “My experience is that nobody would like to pay for the data aggregation which is needed to come up with publicly available and free updates,” she said, adding that she was working on updating the public data “but only slowly.” Little to no indication was given for the source of her own data.

But, she said, she was happy to share the data she used to inform the 2020 IEA study — attaching it to the email — because it was now “outdated”. It was based on confidential individual company data, she said but did not specify the regional profile of those companies or any other aspects of their identity. She had kept it private only because she had not managed to get the studies funded.

By February 2023, Mariutti decided to self-publish his findings on his own website in a piece titled, “The dirty secret of the solar industry.” The piece made a bold claim: scientists were disingenuously using European data to model the carbon intensity of Chinese solar manufacturing. Was the goal here, he asked, to measure the carbon footprint of solar energy or merely to convince us that it’s green?

Read the entire thread here

With a little prod from Mariutti, it was picked up in May in a piece by Giovanni Brussato for the Milan-based weekly Panorama. Brussato drew on Mariutti’s claim that one needs only to look at the life cycle analysis of China’s glass industry by the China Development and Reform Commission to see if there is a data reconciliation problem.

According to Chinese sources, it noted, glass manufacturing – another critical input in solar production – bears a carbon footprint of only 0.68 kgCO2e/Kg despite an admitted 70 percent dependence on coal-fired energy. A comparable study by Western researchers into the UK’s glass industry, which is mostly powered by cleaner natural gas energy, based on data from Eurostat and Guardian Europe, assessed the industry as having a carbon footprint of 1.12 KgCO2e/kg. To compare, the IEA scores solar between 0.5 and 1 kgCO2e/kg and Ecoinvent with 1 kgCO2e/kg.

A major issue with solar data, according to Mariutti, is that data compilers have been slow to recognize the displacement of the industry to China. It wasn’t until 2016, long after much PV production had already moved east, that the transition appeared on data collectors’ radars. But even then, they depended on new estimates and models rather than data from the source.

“In 2014, they calculated the carbon intensity of PV energy as if the panels were made in Europe, with low-carbon energy,” Mariutti told Environmental Progress, referring to data compilers. “By 2016, calculations started to appear as if the panels were made in China, i.e., supposedly with carbon-intensive energy.”

However, whatever model was used, the resulting carbon intensity was always around 20 to 40 gCO2/kWh. “Had they done the math right, it would come out at around 80 to 106 gCO2/kWh, and that’s with important factors still left out,” claims Mariutti.

After the publication of the Panorama piece, Mariutti’s claims drew the reaction of Dr. Marco Raugei, a leading researcher of emissions from renewable technologies at Oxford Brookes University, embroiling both in an extended online spat.

“We all used Chinese electricity mixes for c-Si PV. And we still got results nowhere near as high as you imply one would. So something is clearly off in your back-of-the-envelope calculations,” Raugei tweeted in April this year. By way of example, he cited an influential paper from 2021 on the sustainability of PV systems by life-cycle analysts Enrica Leccisi and Vsili Fthenakis.

Mariutti had previously critiqued Leccisi and Fthenakis’ analysis in his self-published piece, noting that while the electrical input of solar was modeled according to a Chinese scenario, thermal input remained European. After Mariutti pointed out to Raugei that he had tried to contact Leccisi for comment on his findings without success, the conversation with Raugei went cold.

In further correspondence with Environmental Progress, Dr. Raugei stressed that in his research, he endeavored to use the closest possible approximations to Chinese data in order to create a realistic scenario.

The Missing Chinese Data

When scientists, academics, or researchers lack accurate data in the Western world, they usually work hard to fill the data void directly. Major efforts are undertaken, and huge sums are spent on sourcing ever more reliable and better data.

Not so, however, with the China data anomaly. A lack of transparency, language barriers, and a plethora of inaccessible institutions – alongside a general reluctance by researchers to unearth realities that might dispel existing assumptions – have led to an overreliance on models and inputs extrapolated from Western manufacturing processes.

The IPCC’s own estimate that solar’s carbon intensity is four times that of wind and nuclear, but 10 times less than gas and 20 times less than coal is derived from such assumptions.

Unsurprisingly, the authors of the IPCC’s sixth assessment report, casually referred to as AR6, base their life cycle assessments (LCA) of solar energy on studies that do not represent the current state of the industry. Of the four studies cited by the authors, two evaluate only European manufacturing of solar panels. The third model is a state-of-the-art Chinese-manufactured panel, the Upgraded Metallurgical Grade Silicon (UMG-Si), which is no longer in production. The fourth reviews 16 studies, all of which either model solar panels that are no longer in production, model panels that make up only a few percentage points of the global market, or employ Ecoinvent’s 1 or 2 inventories, which also use European electricity mixes.

Ecoinvent, the omnipresent database that is relied upon by policymakers and academics across the planet, as well as manufacturers, big and small, was founded by Dr. Rolf Frischknecht.

For over 20 years, his Swiss non-profit, funded at least in part by the Swiss government and the photovoltaic industry, has collected data on the environmental impact of renewable energies. Whether you’re modeling the low-carbon appeal of recycled plasticpackaging, automotive filters, or titanium powder, Ecoinvent is the likely source of thedata. A recently agreed collaboration on zero-carbon shipping with major players in that industry showcases the association’s still-growing influence.

Since the early 90s, the reputation of Dr. Frischknecht has grown in step with the renewables industry. Some 20 years ago, he began a collaboration with the IEA through the Photovoltaics Power Systems Programme (PVPS), a joint initiative from the IEA and the global PV industry to conduct research on solar and turn it into a global energy “cornerstone”.

Despite his careful stewardship of Ecoinvent, in 2021, Frischknecht quietly resigned from the body he had founded decades earlier. In his resignation letter he noted “irreconcilably different perceptions regarding materiality, reality, quality and accountability” of their latest data.

“There was a drastic shift from (appropriate) data to methodology,” Frischknecht wrote to Environmental Progress. Faced with a movement away from real-world data collection, discussion of what were the crucial data points, proper referencing, and extensive data quality checks, as he had explained in his resignation letter, Frischknecht felt obliged to move on. “During my career, I tried, and try, to be independent of direct, indirect, and subtle attempts to influence the modeling or the data,” he told Environmental Progress.

He then cast doubt on the quality of the Ecoinvent data, telling Environmental Progress, “The PV data in Ecoinvent is from 2011, and there is no data from Chinese information sources.” In email correspondence with Ecoinvent, Environmental Progress was able to confirm Frischknecht’s allegation.

Frischknecht now runs Treeze, a “young and experienced” life cycle assessment consultancy, which is “involved in large EU projects”. Treeze also receives funding from the Swiss Federal Office of Energy and collects life cycle data for the PVPS “Task 12” report on solar’s sustainability.

The total lack of Chinese input into Ecoinvent’s data, however, has not stopped the IEA from continuing to depend on the potentially outdated work of Frischknecht’s brainchild for their own estimates.

These revelations undermine the foundations of the sustainability industry, which bases a significant portion of its certifications on Ecoinvent’s data and promises businesses and governments that earning their certifications protects the planet.

The sector has ample reason to trust Ecoinvent without checking its data. The sustainability sector makes billions of dollars each year thanks to the scale of carbon reductions they claim to provide, and disclosing that it failed to deliver its most basic pledges threatens its business.

The IEA Perspective

After a number of lengthy exchanges with Environmental Progress, IEA representatives and collaborators could only offer evasive and not always entirely corroborating answers.

“It’s true that silicon purification is the most energy-intensive stage in PV production,” IEA’s senior energy analyst, Mr Heymi Bahar, told Environmental Progress over the phone. He also accepted that the data point was central to all calculations of solar’s carbon intensity. “But this data is collected anonymously and cannot be shared.”

Pressed by Environmental Progress, Bahar asked, “Why should the IEA share the names of the companies we get the data from?” To the point that verifying the accuracy of the data, given the huge implications for public expenditure, is important, he added, “Why? Does the fossil fuel industry offer any alternative?” implying there are no better options.

Environmental Progress asked whether the IEA would stand over the unverifiable data in its reports. “We are an umbrella organization,” said Bahar. “We do not oversee the work of all our collaborators. We do not always agree with their opinions.” On whether the IEA trusts that the data it publishes in collaboration is accurate? “Most of the time,” Bahar said.

Two members of the IEA’s solar industry collaboration, PVPS, offered slightly different takes. Daniel Mugnier, PVPS chairman, said that energy consumption data came from “anonymized interviews conducted by Task 12 experts” and that procuring further information would prove overly time-consuming and prevent PVPS from “protecting its sources.” One of those experts, analyst Izumi Kaizuka, told Environmental Progress the information came from a Chinese business association without offering further detail.

Environmental Progress put it to Mugnier and Kaizuka that the only primary data sources available to the PVPS may be the two mentioned by Frischknecht, those without noteworthy links to China: FirstSolar and TotalEnergies. Environmental Progress also asked to know the degree to which the PVPS had relied on assurances from de Wild-Scholten regarding the accuracy of their conclusions without actually being able to consult her proprietary data. At the time of publishing, no response was received.

Presuming the mystery data did come from a Chinese source, Environmental Progress asked Bahar whether this was trustworthy given the vested interest held by both Beijing and the industry it massively subsidizes. He replied that China was “part of the IEA family,” leaving him in no position to comment.

Environmental Progress reached out to Professor Angel de la Vega Navarro of the National Autonomous University of Mexico, who contributed to chapter six of AR6, which focuses on energy systems, and working group III on the mitigation of climate change.

He replied, saying he lacked the expertise to comment further but that the questions we were raising were important. Of the four experts he said were better suited to answer our queries, only one replied, Andrea Hahmann of the Technical University of Denmark: “Unfortunately, I am not knowledgeable enough to provide any comments on the information you have provided as it is outside my area of expertise.”

Environmental Progress contacted Garvin Heath, a contributor to the IPCC Fifth Assessment Report (AR5) and the IEA PVPS report, and received a non-response from the media relations department of his employer, the US National Renewable Energy Laboratory. “Unfortunately, due to current time constraints with our relevant research teams, we will be unable to participate in this.”

Ecoinvent, meanwhile, now headed by Dr. Nikolas Meyer, has not responded to requests for further comment on its lack of Chinese data.

Worryingly, Mariutti claims this is only the tip of the iceberg, with accurate modeling being missed on storage, grid upgrades, methane emissions and more. The IEA has admitted to Environmental Progress that its carbon footprint calculations do not account for three important factors in PV production: silicon mining; toxic panel waste, which promises to overwhelm recycling infrastructure; and something known as the albedo effect. This is when the highly reflective properties of dark-colored solar panels lead to an increase in the greenhouse effect.

According to the IEA, when taken into proper account, the first two factors alone could more than triple the “payback period” for panels, i.e., the length of time before they become carbon neutral after installation.

“Why is the IEA not being transparent about its sources and the gaps in the data?” asks Mariutti. “A hasty transition to solar and other renewables without cast-iron proof of the benefits, all the while handing control to China, could be a huge error.”

With critics like Mariutti being shut out of the debate, science, he says, “is behaving like a religion.” Patrons of the science are giddy bureaucrats charged with convincing taxpayers across the globe to hand over trillions in funding for the feted clean transition.

By seizing control of PV production and acclimatizing well-meaning analysts to the retention of basic data, China has won for itself the lion’s share of global subsidies. What little data may even exist on the sustainability of the PV industry is revealed to selected partners such as the IEA piecemeal and in a manner ensuring its unverifiability.

A picture emerges of an aspirational Western industry captured lock, stock, and barrel by secretive, coal-loving Beijing. It’s a worry for the West’s economic development, never mind energy security and climate action. If solar is anything to go by, the great transition seems less based on data than a mixture of blind faith and vested interests.


Guest Contributor

Self-Reliance Central publishes a variety of perspectives. Nothing written here is to be construed as representing the views of SRC. Reproduced with permission.