Skip to main content

The Long Read: Carbon dioxide recycling


Instead of avoiding emissions of carbon dioxide entirely, should we convert this greenhouse gas into useful products? Our Long Read articles take a deep dive into a topic of scientific importance: this month, ESRF users explore the possibilities to recycle carbon dioxide.

  • Share

This article was first published in the June issue of the ESRFnews magazine.

Carbon dioxide was once the planet’s friend. It is the atmospheric ingredient for photosynthesis, the reaction that sustains all life – indeed, for millennia, nature has happily recycled several hundred gigatonnes of the gas every year. Then came the industrial revolution, the extensive burning of fossil fuels, and a significant tip in the carbon balance. Today, humans are responsible for adding more than 50 extra gigatonnes of CO2 annually, with potentially catastrophic consequences for the climate.

It is no wonder that CO2 is now the world’s number one enemy. Many governments have adopted policies to progressively go “net zero” and put an end to emissions of this greenhouse gas, in particular with the development of sources of renewable power. Yet the actual progress has not been fast enough. Most countries are still heavily reliant on non-renewable energy, either for primary power or to even out the irregularity of renewables. Battery-powered vehicles are still owned by a minority of road users. Fertiliser, fabrics, drugs, medical supplies, cosmetics, electronics, food packaging, steel and many other products still derive in critical ways from oil and natural gas. Arguably, our habits have not changed radically.

Against this uncomfortable backdrop, some scientists have been rethinking our relationship with CO2. Rather than regard it as a friend or enemy, these scientists have begun to think about the greenhouse gas as something in between – a kind of necessary evil. The idea would be to capture and then, with some renewable-powered chemistry, convert it into useful products, from chemical feedstocks for manufacturing to synthetic fuels. Known as “power to X” – the “X” being certain products that would previously have derived from oil and natural gas – the process itself cannot permanently reduce atmospheric CO2, but it can stop levels increasing further, and go a long way to help the fight against climate change.

“The truth is, achieving global net-zero emissions is a formidable challenge,” says Pieter Glatzel, the ESRF’s group head of electronic structure, magnetism and dynamics. “We need to reduce emissions, but we’re going too slow. People are beginning to accept that CO2 conversion has to be part of the solution.”



ESRF group head Pieter Glatzel assists with CO2 reduction studies at the ESRF’s ID26 beamline. Photo: ESRF/S.Candé

In fact, to some extent, it already is. In Patagonia last year, the German sports car manufacturer Porsche opened one of the first electrofuel, or “efuel”, pilot plants. Fed with electricity from a large wind turbine, the Haru Oni plant (pictured above) filters CO2 from the air while splitting water into its constituent hydrogen and oxygen via electrolysis. The oxygen is released back into the atmosphere, while the hydrogen is employed to chemically reduce CO2 into methanol. That methanol feedstock is then further processed to make ordinary petrol – about 130,000 litres a year currently, but rising to 550 million litres by 2027, if Porsche succeeds.

Porsche is not the only embracer of CO2 reduction for chemical synthesis. The multinational consumer-goods company Unilever has piloted the use of captured and reduced CO2 for the manufacture of surfactants for laundry detergent. Another multinational, Honeywell, has announced a plan to use it to make aviation fuel. But these are, at present, modest steps. For the process to operate at any kind of scale, scientists must develop catalysts that perform the reduction much more effectively.

Roham Dorakhan, a doctoral student at the University of Toronto in Canada, is working on this. Under the supervision of his group leader, Edward Sargent, he has been using X-ray absorption spectroscopy (XAS) at the ESRF’s ID26 beamline to analyse catalysts before and during CO2 electrolysis. In a very basic cell design, CO2 is dissolved in water in contact with a catalyst-covered cathode, separated with a permeable membrane from the anode. When a voltage is applied across the cell, the hydrogen from the water reacts on the cathode with the CO2 to form carbon monoxide (CO), hydroxide ions or hydrocarbons.

One of the problems with this basic set-up is that CO2 is able to react with the hydroxide ions in the electrolyte to form carbonate and bicarbonate products, which then cross the membrane to reform CO2 at the anode, reducing the reaction efficiency. To avoid this, Sargent’s group has designed a cascade cell that converts CO2 to CO, before converting CO to the desired hydrocarbons. The XAS at ID26 allowed the researchers to monitor what was going on in the catalyst, made of Ag–CuO.

“The high flux of the beamline allows us to analyse elements with concentrations as low as 1%, a level typically encountered in doped catalysts, where the effect of the dopant is being studied,” says Dorakhan. “Meanwhile, the beamline’s wide energy range broadens the number of elements accessible, especially those with higher atomic numbers.”

In experiments last year, the cell operated for 18 hours and produced acetate with an energy efficiency of 25%, twice as much as anyone had achieved before [1]. That bodes well for the sustainable production of acetic acid, an important feedstock for polymers, textiles, solvents and food additives, with a market size of $13bn.

Still, there are a wealth of other hydrocarbons needed by industry, and one of the major challenges of research into CO2 reduction is to improve the selectivity and stability of catalysts so that they produce exactly the right products for a long time. Formic acid is another important feedstock, for example, used in leather tanning, de-icing aircraft and extracting metals from ores. Among several potential catalysts for this chemical, such as lead and indium, bismuth has gained recent attention because it has relatively low toxicity and high abundance, but its active sites and structure during operation have been debated.

Experience required 

Ward van der Stam, a chemist at Utrecht University in the Netherlands, wanted to study the formation of bismuth active sites using synchrotron XAS, as it can probe structural features over multiple length scales.

“We applied to use ID26 mainly because of the X-ray flux and the energy range, but also because of their staff’s previous experience with electrocatalysis,” he says. “We wanted to look into the dynamics of the activation with high time resolution, so the high flux was useful. Also, the information from the in situ XAS measurements complemented and confirmed our in situ lab-based diffraction results.”

Last year, in their in situ experiment at ID26, van der Stam and colleagues were able to show that halides present in a bismuth oxyhalide precatalyst are able to guide the catalyst’s activation, with bromide in particular promoting the exposure of planar bismuth surfaces for more activity. Indeed, the bromide-activated bismuth achieved a formic acid selectivity of 90% at high current density of 150 mA/cm2 [2].

In future experiments, Van der Stam believes the combination of XAS with other X-ray techniques, such as X-ray diffraction, will be powerful in elucidating catalyst restructuring and stability in electrochemical environments. “The high-quality data provided by ESRF beamlines is crucial in this endeavour, due to the high brilliance and the possibility of combining experiments to draw as complete a picture as possible,” he says.

Several other ESRF beamlines are deeply involved in CO2 reduction research, and have been given a big boost by the EBS upgrade. At ID01, Marie-Ingrid Richard, a physicist based at CEA Grenoble, has been developing stress and strain mapping of individual catalytic nanoparticles via Bragg coherent diffraction at picometre resolution (see ESRFnews June 2023, p17). Coupled with theoretical simulations, these maps will help Richard and her colleagues to understand how CO2 molecules adsorb onto the surface of a catalyst in an alternative gas-phase approach, and to identify active sites. “Strain engineering has emerged as an effective tool to tune CO2 reduction selectivity,” she says.


ESRF visiting scientist Marie-Ingrid Richard has been using ID01 to map the stress and strain of individual catalytic nanoparticles for CO2 reduction. Photo: ESRF/S.Candé

Performance matters 

Meanwhile, there are the state-of-the-art X-ray absorption beamlines BM23 and ID24-DCM, which are able to probe the local atomic and electronic structure of metal centres that convert or store CO2. “This is done under operando conditions, at process-relevant timescales and using a wide range of complementary techniques,” says Kirill Lomachenko, the scientist in charge of ID24-DCM. “Also, both beamlines are equipped with the brand-new monochromators designed in-house at the ESRF, which offer excellent performance in terms of stability and data quality.”

As Jakub Drnec, a beamline scientist on ID31, explains, there are many aspects to improving CO2 reduction, beyond catalytic activity and selectivity. The catalytic durability, for example. “The catalyst can chemically degrade in the sense of chemical dissolution, or it can morph to a different form which is not as active or selective,” says Drnec. “On top of that, the catalyst can detach from its support and simply become electrochemically inactive, or wash out from the system. These processes are much less researched than activity or selectivity, but are essential to understand in order to make a reliable device.”

Recently, Drnec has helped ESRF users perform research on yet another aspect of CO2 reduction – cell stability. This is a particular problem for so-called zero-gap electrolysers, in which the space between electrodes is totally filled with a membrane electrode assembly (MEA). The set-up should be easier to scale, and be superior at preserving electric current, but suffers from flooding and salt precipitation at the cathode, where the new products are supposed to be synthesised.


ESRF scientist Jakub Drnec works on several aspects of CO2 reduction at beamline ID31. Photo: ESRF/S.Candé

A group led by Brian Seger at the Technical University of Denmark in Kongens Lyngby use a combination of wide-angle X-ray scattering and X-ray fluorescence operando to better understand why the problem occurs. “The ability to observe the movement of cations and water within an MEA during operation is truly extraordinary,” says Bjørt Joensen, one of the group members. “It’s a unique opportunity to gain insights into our devices.”

In results published this year, the researchers found that the flooding is linked to movement of cesium ions contained in the electrolyte near the anode, dragging water molecules from the anode and into the cathode, a phenomenon known as electro-osmosis. The knowledge should enable scientists to improve cell performance, by regulating the movement of cesium [3]. “We’re sure that continuous cooperation with the ID31 beamline will further enhance our understanding, and help us optimise the CO2 electrolysis system,” says Qiucheng Xu, another group member.

There is still a way to go until CO2 reduction becomes mainstream. MEA electrolysers currently operate at energy efficiencies of up to 34%, and for lifetimes of up to a few hundred hours, whereas industry requires minimum 50% efficiencies and thousand-hour operating times. Yet research is making fast progress. According to the Web of Science, an online platform that indexes most scientific literature, 10 years ago there were just a handful of papers published annually on the CO2 reduction reaction: today there are more than 1200. Moreover, together with the car manufacturer Toyota and several other industrial partners, the ESRF is a beneficiary of the Marie Skłodowska-Curie doctoral network “ECOMATES”, for the improvement of electrochemical CO2 conversion. “The ESRF–EBS is well-suited to be used for research into all the main issues,” says Drnec. 


Jon Cartwright

[1] C. Liu et al., Nat. Synth. 2, 448 (2023);
[2] S. Yang et al., Nat. Catal. 6 796 (2023);
[3] B. Joensen et al., Joule 8, 1 (2024);