Five-membered saturated nitrogenous heterocycles such as pyrrolidines are present in numerous biologically active compounds. Therefore, synthetic chemists are continuously interested in preparing and functionalising these heterocyclic compounds. Saturated five-membered N–heterocycles are significant not only for the preparation of pigments, drugs and pharmaceuticals, but also for the development of organic functional materials.

Palladium-catalysed carboamination of alkenes has become a useful and reliable method for the synthesis of a broad array of saturated nitrogen heterocycles (1). For example, N-acyl-protected pyrrolidines are synthesised stereoselectively from γ-(N-acylamino)alkenes by reacting them with aryl bromides under palladium catalysis. Moreover, the reaction occurs with high levels of enantioselectivity when the chiral ligand (S)-NMDPP is employed (Scheme I, Equation (i)) (2–4).

Scheme I.

In 2004, Wolfe and coworkers reported a palladium-catalysed coupling of γ-aminoalkenes with aryl bromides to yield 2-benzylpyrrolidines (Scheme I, Equation (ii)) (5). In addition to pyrrolidines being an interesting class of medicinally-relevant compounds (6), this carboamination method was demonstrated to involve a novel, intramolecular syn-aminopalladation step (3, 4).

In a related reaction, carboamination of aminoolefins with 4-bromoanisole proceeds with high diastereoselectivity leading to formation of valuable 2,5-cis-disubstituted pyrrolidines (Scheme I, Equation (iii)) (3, 4).

Six-membered heterocyclic compounds are widely abundant in pharmaceutical actives. Drugs containing saturated heterocycles such as substituted piperidines possess a wide range of pharmacological activities. For example, they are utilised to modulate angina pectoris, hypertension, diabetes, act as Ca²⁺ channel blockers, antitumour agents, and possess hepatoprotective properties. In addition, piperidine derivatives are frequently used as organocatalysts and organic bases in organic synthesis. In this section, the discussion will focus on applications of palladium catalysts to form substituted piperidines.

Larock, Weinreb and coworkers reported the synthesis of vinyl piperidines from N-tosyl aminoolefins and vinyl halides in the presence of a palladium catalyst (7). Nucleophilic attack of the allylpalladium intermediate affords N-sulfonamide–protected vinyl piperidines (Scheme II, Equation (iv)). These compounds could be used as building blocks and incorporated into lead molecules.

Scheme II.

Fluorinative cyclisation of aminoalkenes was carried out with palladium catalysis. Liu et al. reported an oxidative fluorocyclisation protocol of alkenes with a palladium catalyst (8, 9). As a result, various fluorinated piperidine derivatives are formed with high regioselectivity. Mechanistically, the reaction is very interesting and involves: (a) trans-aminopalladation of the alkene; (b) oxidation of the C(sp³)–palladium(II) intermediate to C(sp³)–palladium(IV); and (c) reductive elimination of C(sp³)–palladium(IV) intermediate. The final C–F bond is formed by reductive elimination following oxidative fluorination of the C–Pd bond by a combination of inorganic fluoride salt and oxidant (Scheme II, Equation (v)). Therefore, both the hypervalent iodine reagent and silver(I) fluoride are crucial for this transformation. Interestingly, the fluoropiperidine product is not observed when an N-aryl acrylamide is reacted under the AgF/ PhI(OPiv)₂ catalytic system. Instead, C–H bond activation of the solvent (acetonitrile) is seen, silver(I) fluoride acting as both a Lewis acid and a Brønsted base (10).

Michael et al. reported a hydroamination reaction of tethered aminoolefin substrates to access substituted piperidines (11). For example, an aminoolefin is converted to a methyl-substituted piperidine in the presence of a tridentate-ligated palladium catalyst and silver tetrafluoroborate (Scheme III, Equation (vi)). Presumably, this reaction proceeds through a nucleophilic attack of the amine on the palladium-activated olefin. The piperidine product is then released by protodemetalation.

Scheme III.

Next, the synthesis of piperidines via palladium-catalysed carboamination was carried out to: (a) examine and identify suitable reaction conditions for the transformation by screening various ligands on palladium; and (b) examine the diastereoselectivity of reactions that provide disubstituted piperidines (Scheme III, Equation (vii)) (11). Gratifyingly, the palladium-catalysed carboamination turned out to be successful when preparing 2,6-disubstituted piperidines. However, in most cases, only modest yields are obtained due to competing side reactions (Scheme III, Equation (viii)) (12).

Seven-membered heterocyclic compounds are important structural components found in numerous medicinal compounds. Because of their importance in pharmaceutical chemistry, seven-membered nitrogen-containing heterocycles are important molecules to consider. They are inherently non-aromatic and, therefore, embody useful non-flat scaffolds for drug discovery. For these reasons, many seven-membered nitrogen heterocycles are referred to as ‘privileged scaffolds’ in medicinal chemistry. In particular, azepanes and benzazepines have attracted much of chemists’ attention and their preparation is a topic of extensive studies.

However, seven-membered nitrogen heterocycles are relatively underexplored in medicinal chemistry, in particular when compared to their four-, five- and six-membered congeners. For example, it has been reported that among all US Food and Drug Administration (FDA) approved drugs only 33 of them possess seven- or eight-membered N –heterocycles. In contrast, the number of drugs which contain five- or six-membered rings are 250 and 379, respectively (13). The main reason for this is the scarcity of general and convenient synthetic protocols for the preparation of seven-membered nitrogen heterocycles. Most of the methods that have been developed for the construction of N–heterocycles lead mostly to five- or six-membered ring systems, while the synthesis of seven-membered and larger heterocyclic compounds still lags behind. Nevertheless, some efficient ring-forming protocols have been tailored for the construction of seven-membered rings. Many of these protocols are based on palladium-catalysed reactions.

For example, Nakamura and coworkers disclosed that an exo- methylene azepane derivative is formed in 84% yield by an intramolecular palladium-catalysed hydroamination of an amino-tethered methylenecyclopropane (Scheme IV, Equation (ix)) (14). Here, the key allylpalladium intermediate is produced via distal bond cleavage of the cyclopropane ring. Reductive elimination then furnishes the observed azepane derivative (15).

Scheme IV.

Buchwald et al. developed a palladium-catalysed C–N coupling reaction between aryl halides and amines. This reaction was extended to an intramolecular version which affords interesting benzazepine derivatives (Scheme IV, Equation (x)) (16).

Seven-membered ring-annulated indoles were also synthesised through palladium catalysis (17, 18). Lautens et al. reported a highly modular one-pot tandem reaction involving direct arylation of indoles (19). Interesting fused tricyclic indole derivatives were synthesised by reacting (bromoalkyl)indoles with phenyl iodide in the presence of a palladium catalyst and norbornene (Scheme V, Equation (xi)) (20–22). Importantly, different substituents such as amine, ester, OMe, Me, Cl, or NO₂ are tolerated under the reaction conditions without affecting its yield. However, only 38% yield was observed when a N-methyl tosyl substituent is present at the meta position of phenyl iodide, presumably due to unfavourable steric interactions.

Scheme V.

Stewart and coworkers synthesised a seven-membered benzazepine derivative with an exocyclic double bond by cyclising an allylamine-tethered aryl iodide through a palladium-catalysed 8-endo-trig process (Scheme V, Equation (xii)) (23).

In order to identify possible wear mechanisms for our alloys, we first sought to better understand the nature and role of the individual scratch. This was done by producing coupons in each alloy that could be scratched using a conical Rockwell C hardness tester with a diamond indenter under controlled loads. The samples were first ground plane-parallel on both sides and then polished on the side designated for testing, followed by scratching under both constant and increasing loads. A tape lift consisting of adhesive tape applied directly and uniformly to the scratch in order to embed and remove any spalled material allowed us to compare the susceptibility of the platinum and gold alloys to scoring damage. Tapes from the lift were subsequently examined with energy dispersive X-ray spectroscopy (EDX) to confirm composition of the metal chips as well as characterise the amount of chipping.

A key objective for our study is the simulation of typical human wear mechanisms as closely as possible. There are countless chemical environments and unique mechanical forces that jewellery items are subjected to during human wear, hence a standardised test that attempts to replicate such conditions can only be seen as an approximation of what actually happens in real-life conditions. Correlation with the anecdotal is therefore critical in terms of supporting experimental outcomes as representative of what may be experienced in the human population.

The wear testing performed consisted of three different tests. The first being an abrasion test that utilises a stone and sand media, the second a corrosion test in artificial human sweat and the third a polishing test employing a nutshell media. All media used were calibrated and laboratory grade. Cycles were done in sequence fashion with each of the first five cycles including abrasion, followed by corrosion, followed by polishing. Two subsequent cycles were performed that omitted abrasion and corrosion and only included polishing media. The total test duration amounted to 252.5 h.

Five cube-shaped and individually identified samples of each alloy were used for the testing as shown in Figure 2. Before and after each test in the sequence samples were weighed and characterised by optical microscopy and Vickers microhardness testing. Samples were cleaned in an ultrasonic bath with ethanol to assure any media that might be clinging to the surface was removed. The surfaces of select samples were also characterised by SEM.

The abrasion and polishing tests were based upon the European Industrial Standard DIN EN 12472. The apparatus consists of a motorised rotating drum (Figure 3) that is filled with either an abrasive blend of sand and stones (abrasion test) or nutshells (polishing test). According to the standard, the samples must be physically isolated from one another during testing in order to avoid mutual damage through sample-to-sample contact. Therefore, cubes were anchored along a nylon cord attached to both ends of the drum frame.

Fig. 3

The possible roles of corrosion and erosion corrosion, specifically in gold alloys that contain significant amounts of corrosion-prone base metal elements, were other areas we considered as possibly contributing to wear. The platinum alloys tested were pure platinum group metal (pgm) alloys that did not contain any base metals and are otherwise well-known for their high resistance to chemical corrosion. Therefore, while we did not expect this test to have any effect on pure pgm alloys we included them for the sake of completeness. The corrosion test was based upon the international standard ISO 3160-2. The test involves application of artificial human sweat to the test cubes followed by heating in a closed chamber at 40°C +/– 2°C for 24 h (Figure 4). This test was conducted for cycles one through five right after the abrasion test and prior to the polishing test. Table II gives the composition of the artificial sweat and Figure 4 shows the samples positioned in the chamber. Following the test, samples were cleaned in an ultrasonic bath of deionised water and documented by light optical microscopy.

Fig. 4

Tensile testing was performed in accordance with ISO 6892-1 and microhardness testing was done using a 100 g load (HV0.1) in accordance with DIN EN ISO 6507-1. Tensile properties for cast product were derived from the same casting processes as the test cubes and coupons with the exception of the gold-nickel alloys that were cast by the producer of these alloys. Details on tensile testing are described in (14).

Prior to testing, samples were documented by stereomicroscopy and light optical microscopy. Due to hand polishing the samples exhibit some deviation from the ideal shape as shown in the computer aided design (CAD) images. Selected samples were also documented to obtain details of the geometry, shape and surface condition (Figure 5). After the fourth and fifth cycles the surfaces of select samples were also investigated by SEM (Figure 6).

Fig. 5

Fig. 6

The cube dimensions were measured using a calibrated micrometre calliper. Mass was determined by an analytical balance with an accuracy of 10 μg. Density was determined with the same balance using the buoyancy method (Archimedes’ principle). The mass and volume losses were determined after the abrasion and polish tests and in order to compare the samples, mass loss was normalised with the sample surface area. Volume loss was calculated by dividing mass loss by density.

Vickers hardness of each sample was measured in the as-polished condition and after completion of each cycle (abrasion + corrosion + polish). One measurement was done on each side of the cubes with the exception of the side bearing the sample ID. Table III gives the average hardness value of each sample.

Through SEM analysis (Figures 7 and 8) we see the evidence that the depth of the scratch is impacted by the hardness of the alloy. As one might expect, the softer the alloy, the deeper the scratch and the more material is displaced. In the case of the soft alloy 950PtIr, the displaced material was concentrated at the edges and the tip of the scratch (Figure 8), which is typical for micro-ploughing. Local overload also resulted in cracking of the displaced material at the edge of the scratch that appears to be loosely connected. In comparison, the gold alloys showed not only cracking, but also significant chipping along the cracks. This was especially true for the 585AuNi, which has a stronger tendency for micro-cutting.

Fig. 7

Fig. 8

We noted that the alloys appeared to show different levels of porosity after polishing with the platinum alloys exhibiting low levels and the gold alloys exhibiting higher levels characterised as finely dispersed microshrinkage. From previous studies on the tensile properties of platinum alloys (13) it was established that the ductility values of elongation and reduction of area are significantly impacted by porosity levels. Therefore, increased chipping in the gold alloys may be not only a result of intrinsically lower ductility for these alloys, but also porosity-related decreases.

3.1.1 Tape Lift

Corrosion was qualitatively assessed by optical microscopy after each test. The presence of corrosion was most visible after the first cycle because the surface had less scratching from the wear tests than subsequent cycles. As expected for pure pgm alloys, both 950 platinum alloys (Table I) showed no visible changes following corrosion testing (Figure 10).

Fig. 10

The alloy that demonstrated the least amount of resistance to corrosion was the 585AuNi containing high amounts of nickel, copper and zinc (Figure 11). Following wear testing porosity was exposed to the surface, suggesting that corrosion was further promoted by microshrinkage pores that had been revealed. Such pores act as crevices where a concentration of corrodents is able to accelerate the corrosion process. This being the case, the casting quality level may be a contributor to reduced (or improved) wear resistance, particularly in alloys demonstrated to have low corrosion resistance such as the 585AuNi.

Fig. 11

The 750AuNi and both the 585AuPd and 750AuPd alloys did not exhibit visible corrosion after any of the five cycles. While higher corrosion resistance is expected with the greater noble metal content of these alloys, the potential effects of corrosion cannot be ruled out given their base metal content and the limited scope of our testing. Moreover, the corrosion testing performed was of a static nature, omitting the potential for an erosion corrosion dynamic that is likely present in human wear conditions. This topic is recommended for further testing to better understand the potential for effects on wear resistance in gold alloys.

The goal of this series of tests was quantitative determination of mass loss and volume loss through a combination of abrasion testing and polish testing. The total testing time can be segregated into abrasion time (sand + stone media) and polish time (nutshell media). Mass loss and volume loss were normalised with the surface area of the sample, allowing us to compare data from samples with a different geometry. The plotted values show the mass loss and volume loss per surface area of the sample. For simplicity, the terms ‘mass loss’ and ‘volume loss’ are used for normalised values in the text of this paper. Mass and volume loss were plotted against abrasion and polishing time and total wear time, respectively. The plots show the average loss of the five samples per alloy that were tested. This allowed for a segregation of data for the amount of wear measured in each of the different tests.

During the abrasion test portion of our assessment the mass and volume losses show a non-linear increase with increasing abrasion test time (Figure 12) in the beginning of the tests, which turns into a linear trend with increasing testing time. No remarkable difference between the alloys is observed and overall mass loss during abrasion testing is extremely small. The 585AuPd does show slightly higher wear compared to other alloys in this phase of the cycle, but mass loss was only 0.00216 g, or 0.08% of original mass.

Fig. 12

Volume loss was calculated by dividing mass loss by density. Due to the considerably different densities of the tested alloys three groups can be distinguished. The platinum alloys have a density of ca. 20 g cm⁻³; 750 gold alloys are at ca. 15 g cm⁻³; and 585 gold alloys are at 13–14 g cm⁻³. While mass loss is very similar for all alloys, the volume loss differs more due to these distinctly different density levels. The platinum alloys showed the lowest volume loss, followed by the 750 gold alloys and the 585 gold alloys. Total volume loss in the abrasion test was very low with a maximum value at only 0.0005 mm³, or 0.03% of the original volume.

For the polishing test the mass and volume loss rate (i.e., the mass and volume loss per unit of time) was comparable to the loss rate abrasion test. Mass loss was demonstrated to increase linearly with increasing polishing time. The platinum alloys again show the lowest mass loss with total mass loss after 244 h of combined testing at less than half that of the 750AuPd, which showed the highest mass loss in the group. The mass loss of the 585AuPd and the 585AuNi lies in between the two 750 gold alloys. The total mass loss after 244 h of testing was 0.013 g for the 950PtRu (lowest value) and 0.031 g for the 750AuPd (highest value). These are still very small amounts equal to 0.03% for the 950PtRu and 1.1% for the 750AuPd. However, when we consider volume losses these differences take on much greater significance. The volume loss of both 950 platinum alloys is a factor of three times lower compared to 750AuPd, and a factor of about two times lower compared to 585AuPd and both 750AuNi alloys.

Figure 13 shows the total mass and volume loss after all cycles of wear testing were completed. Since the absolute mass loss in the abrasive test was much lower than that in the polishing test, the abrasive test was omitted in the last two cycles of wear testing. The result in Figure 13 is very similar to that of Figure 14. Error bars indicate the results from the samples with the lowest and highest mass loss in one group of alloys, while the full symbols indicate the averaged mass loss of the five samples. The error bars confirm that the difference between the alloys remains significant. The mass loss curves demonstrate a linear trend that was fitted for select alloys. The slope indicates the mass loss per hour of wear, i.e., the rate of wear.

Fig. 13

Fig. 14

The assessment of surface quality focused on the rounding of corners and edges, which was qualitatively determined by stereo microscopy. Figure 5 demonstrates the samples with the lowest and highest volume loss, which are 950PtRu and 750AuPd respectively. Figures 5(a) and 5(c) were taken after completing the first two cycles of 10 h total wear testing. After 10 h very little difference can be detected in comparison with the as-polished condition of the samples. The mass loss after two cycles was only 0.0004 g, therefore this result is expected. Figures 5(b) and 5(d) demonstrate the sample surface after completing seven cycles. 950PtRu displays a very well-defined cube shape after the second cycle and only a very slight rounding of the corners following the seventh and final cycle. The absolute mass loss after the complete series of testing was 0.0131 g, or 0.3% for the 950PtRu.

All five of the 750AuPd samples displayed a less-defined cube shape in the as-polished condition as a result of hand polishing prior to testing. The surface also appears somewhat uneven (Figure 5). Nevertheless, a continuing deterioration of the cube geometry was demonstrated through testing. Following the second cycle edges and corners present with increased rounding, and this condition is even more pronounced after the seventh cycle, indicating mass loss had occurred during testing. Absolute mass loss for the 750AuPd after the completion of wear testing was 0.0307 g, or 1.1%.

The surface of select samples and conditions was captured by SEM imaging. Figure 6 depicts the samples with the lowest and highest volume loss after abrasion testing (Figures 6(a) and 6(c)) as well as subsequent corrosion and polish tests (Figures 6(b) and 6(d)). Following the abrasion test both sample surfaces are quite rough and exhibit deep dents and scratches. After the polish test both samples display a levelling of the topography of the sample. Notably, despite its lower hardness, (or perhaps because of it) 950PtRu exhibits a smoother surface finish compared to 750AuPd.

Tensile testing was performed to determine whether strength and ductility measures might play a role in mass loss. Table III shows the average results of tensile testing from four as-cast bars in each alloy. The hardness values are the average values that were measured on a set of five cube samples of each alloy.

We did not find any significant correlation with tensile properties or hardness and mass loss. As other studies showed before (2), it appears that high hardness is not an indicator for low mass or volume loss. However, the opposite also cannot be concluded. Rather, the situation appears to be more complex and depends upon the mechanism of mass loss during wear testing. The alloys exhibited very different hardness levels with one series of samples (585AuPd) showing a spread of more than 10%, indicating an inhomogeneous microstructure, due to porosity for example. Micropores were visible on the polished coupons of the 585 gold alloys (Figure 8).

It has been demonstrated in platinum alloys that the reduction of area value (ROA) is strongly reduced by microporosity (14). If this is the case, then the microstructure of the samples plays an important role on wear behaviour. Micropores along scratches will act as points of stress concentration and may cause the chips to break free. Increased levels of microporosity are likely to favour micro-chipping over micro-ploughing, suggesting increased mass loss due to metal chips. Further investigations will be necessary to prove such a hypothesis.

Industries face mounting challenges in the paradigm shift to a more circular economy. Research and development is increasingly focused on finding ways to turn waste into resources, recover energy and materials and make better use of resources extracted from the natural environment. At the same time industry and consumers seek to cause less harm in the form of pollution or CO₂ emissions. In this issue of the Johnson Matthey Technology Review, we look at current and future technologies that may be used by industries including energy, fuels, chemicals, pharmaceuticals and transport to create the products we need while meeting the United Nations (UN) 2030 goals for sustainable development (1): “development that meets the needs of the present without compromising the ability of future generations to meet their own needs” (2).

Continuing the theme from our previous issue (3), several articles present different approaches to future fuels and chemicals. These approaches include water electrolysis, utilisation of biomass and waste and CO₂ reduction.

Electrification will provide alternatives to fossil fuel use in many areas of industrial science and technology but some areas like long-haul aviation will likely continue to need liquid fuels. These fuels will be provided through one or more of the technologies being developed today. For example, a technology recently commercialised by Johnson Matthey and bp uses a Fischer-Tropsch process to create sustainable jet and diesel fuels from waste, biomass or existing CO₂ emissions. The challenges involved in achieving a commercially viable process at scale are explained.

The technoeconomics and life cycle assessment of producing sustainable commodity chemicals from waste biomass using aerobic fermentation at scale are explored in another Johnson Matthey collaboration, this time with the University of Nottingham and Northumbria University, UK. Rigorous process modelling has determined at what point the production of commodity chemicals from a lignin source will become commercially viable. The future of this promising technology looks bright, with the authors concluding that their platform has promise as a best-in-class technology for the production of a broad spectrum of renewable commodity chemicals.

Activated carbon can be produced and characterised from biomass waste for applications in environmental protection, clean energy and catalysis. The work is presented by Gebze Technical University, Turkey, in collaboration with Gasification Consultancy Ltd, UK. Waste biochar from the gasification of biomass is the feedstock, and removal of contaminants is key to its successful use.

Reducing CO₂ emissions from iron metallurgy will become increasingly important. The electrification of primary iron production in a carbon-free process is presented in a collaborative research article from National Technical University of Athens, Laboratory of Metallurgy and Mytilineos SA, Metallurgy Business Unit-Aluminium of Greece. The technology is demonstrated at an early stage with additional optimisation recommended by the authors. Catalytic hydrogenation of CO₂ to methane using power-to-gas combined with biomass gasification is another option to reduce the CO₂ emissions of the steel industry, presented by Montanuniversität Leoben, Johannes Kepler Universität Linz and K1-MET GmbH, Austria.

This journal has long championed sustainable technologies involving the precious metals. Metals are inherently recyclable and none more so than the platinum group metals (4). Today’s focus on electrification of transport and energy means that elements such as lithium, nickel, cobalt and manganese join their precious cousins as critical materials for the clean energy revolution. Clean and efficient extraction of these minerals from spent lithium-ion cathodes is an emerging area of study that will become increasingly important in the coming years when batteries begin to reach end-of-life. Recycling techniques need to be developed for the sustainable development of the lithium-ion batteries industry as discussed in this issue. Meanwhile life cycle assessment of the entire lithium-ion batteries production process from both primary ore and recycled material is provided in the output from an Innovate UK project involving Johnson Matthey and the Warwick Manufacturing Group, UK.

Energy efficiency will be a key enabler for a transition to a low carbon future. High technology industries like electronics, energy and medical applications require novel materials and processes. Cooling is a challenge, especially at the microscale. Nanofluids containing titania offer a potential solution and are investigated in this issue.

Conventional technologies will continue to be used alongside newer ones. To help define the next generation of emissions legislation to clean up the air in China, a portable emissions measurement system was used to investigate on-road tailpipe volatile organic compounds emissions in diesel trucks compliant with Euro III–V. The results with recommendations from the authors are presented in this issue.

It will become apparent from reading this issue that collaborations both within and between industry and academia are vital to progress. The research projects described here are just a selection. Many more advances can be expected in the coming years and decades as fruitful collaborations continue apace, with industry and academia working together to meet the challenges of the present and the future.

Growing environmental concerns have made it imperative to reduce global climate change and this has resulted in prolific development of various energy storage technologies for different applications ranging from portable electronic devices (PED) to electric vehicles (EVs) (1, 2). The most common chemical energy storage devices are batteries for applications requiring high energy density and electrochemical capacitors (ECs) for applications with high power density requirements (3–5). LICs which have the combined desirable properties of batteries (high energy density) and ECs (high power density) are increasingly being investigated as high-performance energy storage devices that have a significant role in the decarbonisation of the transport sector (6, 7).

While there are many promising negative electrode materials for LICs, the lithium titanium oxide (Li₄Ti₅O₁₂, LTO) based anode offers high stability towards charge-discharge cycles, faradaic efficiency and lower costs (8–10). As the envisaged use of the LTO based LIC is in hybrid and EVs to assist in decarbonising the transport sector, it becomes pertinent to conduct a LCA for the production of a LIC using primary ore minerals and make comparisons to a manufacturing process that relies on recycling end-of-life LIC. LCA is defined as a process to evaluate the environmental burdens associated with a product, process or activity by identifying and quantifying energy and materials used and wastes released to the environment (11). The assessment includes the entire life cycle of a product, process or activity, encompassing extracting and processing raw materials, manufacturing, transport and distribution, use, reuse, recycling and final disposal.

LCA facilitates informed decision making as comparative analysis of competing processes or products can be conducted based on environmental impact. At the early stages of R&D activities, LCA is an invaluable tool as it can inform process and material choices that support sustainability goals in addition to promoting innovation for designing products that are more amenable to recycling when they reach end-of-life (12–14). Increasingly, LCA is also being utilised to engage with stakeholders as an evolving green marketing tool through brand competitive differentiation on the basis of sustainability as well as regulatory compliance purposes (15–17). Besides the multifaceted benefits of LCA, its utilisation is not without limitations with uncertainties in inventory data, methodology and application of the weighting technique often being cited as major weaknesses of the approach (18, 19).

While the LCA methodology has been widely applied to energy storage systems this has mostly been for lithium-ion batteries (LIBs), with most studies having focussed on comparative analysis of LIBs to internal combustion engine (ICE) or sustainability of the different battery chemistries (20–22). There is a scarcity in the literature of LCA studies that have analysed production of energy storage devices using primary ore materials in comparison to manufacture of a similar product using recycled materials and specifically for LICs. This study objective is to take a comparative approach with the aim of utilising LCA to inform early phase R&D activities to improve the sustainability of the various process and reagents choices in the production of a LIC module.

The LCA study was conducted as part of the Advanced Lithium Ion Capacitors Electrodes (ALICE) project whose objective was to develop a 48 V LIC module for use in automotive, e-bus and materials handling equipment. The project consortium had industrial and academic partners for developing and scaling-up materials production including application of novel coating techniques to electrode structure to improve performance. The 48 V module built in the project was tested based on end user requirements and physics based numerical modelling applied at different stages of the project to interlink sophisticated layer structure characterisation results with cell performance.

2.1 Goal and Scope

2.2 Functional Unit and System Boundary

2.3 Methods and Databases

The elementary flows of material required to make the cells for a 48 V LIC module are based on pilot plant data. The recycled process data is based on a laboratory flowsheet that is simulated using a process model with appropriate scaling of model parameters from gram scale to a full-scale production plant. The main product and process stages for the primary (ore) and recycled materials manufacture of LIC are as follows: (a) anode powder material preparation; (b) anode preparation; (c) cathode preparation; (d) electrolyte preparation; (e) cell formation; and (f) recycling of LTO powder (recycled material process only).

The detailed breakdown of materials for the assemblies and product stages of the two LCA comparative projects is in Table I. The only difference between the two comparative studies is in the source of lithium carbonate and titania for making the anode LTO powder material. For the primary (ore) process, the information for the environmental footprint associated with lithium carbonate and titania is obtained from the ecoinvent 3.5 database based on ore extraction and salt formation environmental impact values. However, in the case of the LIC module made from recycled materials, lithium carbonate and titania are obtained from the recycling product stage and only have process environmental impact values associated with reagents and energy consumption demand to recycle the end-of-life LTO anodes.

The supplementary data which contains the flowcharts and inventory to produce a lithium ion capacitor module and the list of assumptions used in the study is located with the online version of this article.

3.1 LTO Powder Synthesis

3.2 Anode and Cathode Preparation

3.3 Electrolyte, Cell and Lithium-Ion Capacitor Module Formation

3.4 Modelling of LTO Recycling Process

While results for several environmental impact categories were available for analysis, for purposes of this study climate change (kilogram of CO₂ equivalent) and terrestrial acidification (kilogram of SO₂ equivalent) were analysed in greater detail for comparing the LIC module manufacture from primary ore materials against the recycled material process. The calculations are based on the ReCiPe Midpoint (H) with European Normalisation (24). The ReCiPe method was utilised because of its environmental relevance to the scope of the study, transparency and reproducibility. However, other methods which are also compatible with ISO standards could have been applied to the study.

Aluminium had the highest climate change and terrestrial acidification burdens to the extent of overshadowing contributions from other materials. To facilitate detailed analysis of environmental burdens of the other materials and processes, visual graphics of the results were plotted without the contribution from aluminium. Aluminium has established recycling processes but the decision if the quality of this recycled aluminium was of specifications sufficient for direct use in LIC manufacture was indeterminate and therefore the LCA credit process was not applied towards aluminium used. Figure 4 compares the climate change impact for making a 48 V LIC module using primary ore material and recycled LTO. Overall, utilising recycled LTO materials reduces the climate change impact by 12%. The order of decreasing climate change for the LIC module manufacture using primary (ore) materials is titania > lithium hexafluorophosphate > ethylene carbonate. For LIC module manufacture using recycled LTO, the order of decreasing climate change is lithium hexafluorophosphate > formic acid > ethylene carbonate. The highest contributor towards climate change for primary (ore) case is titania while for the recycled LTO it is the lithium hexafluorophosphate electrolyte. Lithium hexafluorophosphate and ethylene carbonate are both part of the electrolyte system and have significant contributions which are equal for LIC manufacture using either primary (ore) or recycled LTO materials. Therefore, significant reductions in climate change for LIC manufacture using recycled LTO can only be achieved by reducing the quantities of formic acid used. Table II shows the climate change impact over the various stages of manufacturing a 48 V LIC module. The anode preparation stage has the highest contribution towards climate change for the two comparative cases. However, using recycled LTO lowers the climate change impact by 21 kg_CO2eq compared to using primary (ore) during the anode preparation stage. The cathode preparation and cell formation stages have the same values as the two cases only differ in source of materials used the anode preparation stage.

Fig. 4

Comparative analysis of terrestrial acidification for producing a 48 V LIC module using primary (ore) and recycled LTO materials is shown in Figure 5. Usage of recycled LTO for the anode manufacture product stage results in 18% reduction in terrestrial acidification compared to using primary ore materials. The major contributors towards terrestrial acidification in decreasing order are lithium hexafluorophosphate > titania > ethylene carbonate for LIC module manufacture using primary (ore) materials. For LIC manufacture using recycled LTO, the major contributors towards terrestrial acidification in decreasing order are lithium hexafluorophosphate > formic acid > ethylene carbonate. Table III shows the terrestrial acidification associated with the various stages of manufacturing a 48 V LIC module for the two cases. The anode preparation stage has the highest contribution towards terrestrial acidification when the primary (ore) and recycled LTO material sources are compared. By utilising recycled LTO for the anode preparation process the terrestrial acidification impact is lowered by 0.21 kg_SO2eq compared to using primary (ore) materials.

Fig. 5

While there are environmental benefits from using recycled LTO, the existing recycling process flowsheet has a lot of optimisation opportunities especially regarding the quantities of formic acid used which have a significant contribution towards both climate and terrestrial acidification.

The application in early phase R&D activities is demonstrated in this section as applied to process development choices for the recycling stage. Before the hydrometallurgical treatment detailed in Section 3.4, the LTO has to be decoated (removed) from the aluminium foil to which it is bound. The widely used binder for coating the LTO and most active materials to current collectors is polyvinylidene fluoride (PVDF) because of its adhesive capabilities and electrochemical stability (25). The PVDF binder presents a challenge to the decoating process as it is only partially soluble in most common solvents. The common solvent for dissolving PVDF is NMP which is also used during the slurry coating process. However, NMP has high environmental and toxicity burdens which has resulted in stringent legislative restrictions of its usage (26).

To improve the sustainability metrics of the recycling process, several alternative solvents were investigated for their capabilities to remove the LTO from the aluminium foils. Acetone and polyethylene glycol (PEG) have similar properties to dipolar aprotic solvents like NMP and were identified as greener alternatives (27). Several other organic reagents such as acetic acid, formic acid, ethylene glycol and methanol were also screened as potential candidates for the process. Figure 6 shows images of LTO anodes after stirring in different reagents for 1 h at room temperature. From Figure 6, formic acid has higher technical performance compared to the other solvents as it removed all of the visible traces of carbon and LTO from the aluminium foil.

Fig. 6

The decision-making process considered the environmental footprint of these reagents in addition to their technical performance for removing the LTO from the aluminium foils. Figures 7(a) and 7(b) show the contribution towards climate change and terrestrial acidification respectively based on using 1 kg of these reagents. NMP has the highest environmental burden for climate change and terrestrial acidification, while methanol has the least environmental footprint. However, methanol efficiency in decoating LTO was low and therefore a trade-off of technical performance, environmental impact and costs resulted in formic acid as the alternative reagent choice for the decoating process stage. The bulk purchase price for these reagents is in Table IV.

Fig. 7

While several product and process research activities are focussing on novel binders that are less toxic and low costs compared to PVDF (29), the electrodes bound with PVDF that have already been manufactured will still require a more environmentally sustainable process to recover the active material at their end-of-life. The approach applied in this case study of early phase R&D process development activities demonstrated sustainable choices in alternate reagent selection in alignment to the triple bottom line approach (30). The choice of formic acid when compared to NMP results in intersection of people, planet and profit (3Ps) requirements of sustainability. For this case study, formic acid had reduced environmental impact, toxicity and meets the profit criteria through high decoating technical efficiency at lower costs compared to NMP.

The LCA methodology was applied to quantitatively determine the environmental burdens associated with manufacturing a 48 V LIC module. The prospective LCA compared the environmental impact of manufacturing a LIC module using primary ore materials versus LIC manufacture using recycled materials from end-of-life LICs. The anode preparation stage is associated with most of the environmental burden for manufacturing the LIC module for both processes due to the source of precursors used in production of the active LTO material. Utilisation of LTO precursors from recycled end-of-life LICs reduced both climate change and terrestrial acidification environmental impact categories for the LIC module manufacture. However, the sustainability metrics of the recycled process route of production could potentially be improved further by optimised application of formic acid which is used in the process stage for separating the LTO from the aluminium current collector foils.

The application of the LCA methodology in early phase R&D activities was demonstrated for the process development reagent choice case study. The LTO decoating reagent decision-making process considered the environmental footprint, technical performance and costs. The decision to utilise formic acid as a decoating agent was a sustainable choice which balanced environmental, economic and social performance. For the demonstrated case study, the choice of formic acid as decoating reagent reduced climate change and terrestrial acidification, lowered human toxicity values and met the profit criteria through high separation efficiency at lower costs.

In the 21st century, there is a need to deal with threats such as energy scarcity and environmental deterioration. The worldwide usage of fossil fuels accounted for 84.7% of global energy consumption in 2018, which is equivalent to 11.7436 billion tonnes of oil (1–2). Global CO₂ emissions, especially from fossil fuels, will continue to grow rapidly. It is crucial to explore green and renewable energy systems, such as wind, tidal and solar energy, and energy storage such as batteries, to replace fossil fuels. LIBs, with excellent energy storage properties, safety and stability, are among the most promising clean and sustainable energy storage equipment. LIBs are widely used in zero-emission vehicles (mainly electric vehicles (EVs), and plug‐in hybrid electric vehicles (PHEVs)), computers and electronic communication devices (3–6). The newly emerging model for super-performance LIBs, assembled with specific complex three-dimensional (3D), porous and polyhedral geometric structures in cathode and anode materials, is perfectly suited to the scale-up requirements for zero-emission vehicles (3–5). Increasing demand for new energy vehicles contributes to the expansion of the LIBs market. It has been estimated that the world’s production of LIBs would increase by 520% from 2016 to 2020 (7), and about 50 million electric buses will run on the road by the end of 2027 (8). Therefore the number of expired LIBs, as major electronic wastes, will inevitably increase. In China, the weight of retired LIBs was predicted to reach 500,000 tonnes by the end of 2020 (9), and that of the European Union to reach 13,828 tonnes in 2020 (10). Since harmful substances may damage the environment and the metals contained in spent LIBs are precious resources, the recovery of retired LIBs is bound to gain considerable social, economic and environmental benefit.

The cathodes of retired LIBs are rich in valuable nonferrous metals such as lithium, nickel, cobalt and manganese, which are secondary resources worth recycling. Considering potential immense profit, researchers have been working hard to develop various technologies to recycle the metals in spent LIBs. Current recycling technologies mainly include pyrometallurgy and hydrometallurgy. Although pyrometallurgical recovery processes have the advantages of a short process, high efficiency and easy industrial application, high energy consumption and generation of toxic gases still limit their development (11–14). In contrast, hydrometallurgical recovery processes have attracted extensive attention due to low energy consumption, environmental friendliness and high recovery efficiency (15–19).

Leaching valuable elements by chemical reagents is the core of the hydrometallurgical recovery strategy. After leaching, valuable metals in the leachate are extracted by chemical precipitation, solvent extraction and ion-exchange (20–25). The conventional hydrometallurgical process for waste electrodes recovery can be illustrated as follows: (a) acid-reductant leaching and selective chemical precipitation; (b) sulfuric roasting, acid leaching and selective chemical precipitation; (c) mechanochemical activation leaching and selective chemical precipitation, as shown in Table I (26–34). However, sulfuric roasting or mechanochemical activation before leaching may complicate the recovery process and decrease overall leaching rate. Based on the principle: waste substance + waste substance → rebirth resources, Li et al. (35) developed an in situ recovery model for graphite, Li₂CO₃ and cobalt by oxygen-free roasting with an anode graphite + wet magnetic separation process, without pretreating or adding any other chemical reagents. However, this technique is not suitable for recovering complex electrode materials. Prabaharan et al. (36) investigated an electrochemical leaching system: lead as anode + electrode scraps as cathode + H₂SO₄ as leachate, in which the leaching rate of manganese, copper and cobalt exceeded 96% through adjusting pH value. Although it can perfectly achieve integrated recovery of valuable metals, the high electricity consumption and recovery cost limit the use of this method. Surprisingly, Gomaa et al. (37, 38) explored a new multifunctional recovery method to recycle ultratrace Co²⁺ (~3.05 × 10^–8 M and 4.7 × 10^–8 M respectively) with visible selective ion extraction-separation-detection. It can extract almost 100% Co²⁺ from leachate in only 10–15 min and the used ion-extractors after activation can be regenerated and reused in repeated adsorption-desorption processes. Following desorption by HCl eluting reagent, the overall recovery rates of cobalt from waste LIBs or printed circuit boards (PCBs) are 98% and 95.7% respectively.

NCM batteries consist of more valuable metals that are worth recycling compared with traditional LiCoO₂ and LiFePO₄ batteries. However, few reports have been made to systematically clarify recovery techniques for waste NCM materials. In order to avoid loss of valuable resources and risk of secondary pollution, it is urgent to construct a sustainable recycling model for valuable metals in cathodes of spent LIBs. This review aims to describe progress in hydrometallurgical recycling of cathode materials from spent NCM batteries. The hydrometallurgical recovery strategy of waste LIBs can be classified into three steps: (a) pretreatment or separation of active substances; (b) leaching or extracting the valuable metals from the active substances with appropriate solvents; (c) separation of valuable metals by selective extraction from leachate by different methods to obtain the metals or metallic compounds. The conventional process flow for recycling NCM materials from waste LIBs by hydrometallurgy is shown in Figure 1. The advantages, disadvantages, existing problems and current status of each treatment method are analysed. Furthermore, advanced recovery technologies, DESs extraction and crystal repair direct-regeneration/recovery technology are illustrated. The challenges and prospects for metals recovery from ternary cathode materials of used LIBs by hydrometallurgy are described. By comparing the advantages and disadvantages of different methods, it is expected that this information will contribute to exploring economic, green, sustainable, high-efficiency leaching, separation and regeneration recovery systems for closed-circuit recycling of LIBs.

Fig. 1

The LIBs are mainly divided into five types depending on the composition of cathode materials: lithium iron phosphate (LiFePO₄), lithium cobalt oxide (LiCoO₂), lithium manganese oxide (LiMn₂O₄), lithium nickel oxide (LiNiO₂) and lithium NCM oxide (LiNi_x Co_y Mn_{1–x –y} O₂,0<x +y <1) (39–43). Of these, layered LiNi_x Co_y Mn_{1–x –y} O₂(NCM) materials are preferred because of their excellent cycle and rate performance. Currently, there are five types of NCM cathode materials that have been commercialised: NCM333 (LiNi_1/3Co_1/3Mn_1/3O₂), NCM424 (LiNi_0.4Co_0.2Mn_0.4O₂), NCM523 (LiNi_0.5Co_0.2Mn_0.3O₂), NCM622 (LiNi_0.6Co_0.2Mn_0.2O₂) and NCM811 (LiNi_0.8Co_0.1Mn_0.1O₂) (44). It is predicted that NCM batteries will account for nearly 41% of the world’s LIBs market in 2025 (42). The chemical characteristics and potential hazards of each component in ternary LIBs are summarised in Table II (45–48). NCM cathode materials are rich in lithium, nickel, cobalt, manganese and other strategic essential metals. The heavy metals in waste LIBs will pose a huge threat to human health and the environment (49). Furthermore, the scarcity and high cost of nonferrous metals such as cobalt make it imperative to recycle ternary cathode materials from retired LIBs. It is also reported that the global lithium resource reserves in 2018 are over 62 million tonnes, but China only accounts for 7%, about 4.5 million tonnes (50). As a core raw material for LIBs, the value of cobalt is as high as AUD$115,000 per tonne (51, 52). Efficient extraction of these valuable nonferrous metals from retired LIBs is of great significance for green and sustainable development of the batteries industry.

3.1 Recovery Process

3.2 Recovery Process of Hydrometallurgical Method

4.1 Regeneration of Nickel Cobalt Manganese from Leachate

Sustainable recovery and regeneration of NCM materials from leaching solution is the current main trend for recovery of exhausted LIBs. Common methods to synthesise cathode materials include chemical coprecipitation, high temperature solid phase, hydrothermal synthesis, sol-gel, microwave synthesis and electrostatic spinning method (95–100). The parameters of resynthesis methods to regenerate NCM cathode materials from leachate are summarised in Table VII (61, 101–107).

NCM cathode materials regenerated by a hydrometallurgical process hold superior crystal morphology and rate/cycle performance. Their electrochemical performance is comparable to that of commercial NCM materials (102, 104). The conventional hydrometallurgical regeneration process to prepare NCM materials is as follows: leaching liquid → coprecipitation → solid phase synthesis at high temperature, as shown in Figure 6. In the regeneration process, appropriate amounts of cobalt salts, manganese salts, nickel salts and lithium salts are added according to the stoichiometric composition of NCM materials.

Fig. 6

Regeneration flow of LiNi_1/3Co_1/3Mn_1/3O₂ cathode material from leaching liquid by coprecipitation

Li et al. (108) resynthesised NCM333 (R-NCM) materials from spent cathode materials leaching liquids by a one-step sol-gel method, and the related details and recycling mechanism are shown in Figures 7 and 8 (108). In the recovery system, H₂O₂ is used as reducing agent, and lactic acid plays the role of leaching agent in the leaching stage and chelating agent in the regeneration stage respectively. The mechanism of leaching stage can be shown using Equation (i):

(i)

Fig. 7

Resynthesis flow of LiNi_1/3Co_1/3Mn_1/3O₂ cathode material from leachate of spent LIBs by sol-gel method

Fig. 8

(a) Possible products and mechanism in the lactic acid leaching process; (b) XRD and SEM patterns of R-NCM and F-NCM samples; (c) electrochemical performances of R-NCM and F-NCM samples: charge/discharge profiles at 0.2 C; (d) cycling performances at 1 C; (e) rate performances at different currents and (f) Nyquist plots in the frequency range of 100 kHz to 0.01 Hz. Reprinted with permission from (108). Copyright (2017) American Chemical Society

Under optimal conditions (lactic acid = 1.5 M, s/d = 20 g l^–1, H₂O₂ = 0.5 vol%, temperature = 343 K, where s/d means solid/liquid (S/L) ratio) the leaching efficiency of the metals reached 98%. Electrochemical results proved that the R-NCM cathode material held excellent reversible discharge capacity of 138.2 mAh g^–1 while capacity retention reached 96% at 0.5 C after 100 cycles (105). Due to low charge-transfer resistance (R_ct) of R-NCM (58.78 Ω) compared with fresh NCM333 (F-NCM, R_ct = 70.02 Ω), which can provide higher lithium ion diffusion coefficient (D_Li⁺) and faster lithium-ion intercalation/deintercalation kinetic properties, the electrochemical performance, especially cycle capability, of R-NCM is self-evidently higher than that of F-NCM. It is indicated that the chelating agent lactic acid can efficiently recycle and resynthesise NCM materials by a sol-gel method with closed-loop recovery process.

4.2 Integrated High Value Utilisation of Electrode Scraps

4.3 Potential and Challenges of Mediate/Direct Regeneration Method

5.1 Extractive Methods Based on Deep Eutectic Solvents

Compared with traditional extractants, DESs have the advantages of being non-toxic, cheap, biodegradable and possessing extremely high dissolution and reduction capability for metal oxides (114, 115). However, there are few reports on recycling electronic waste using DESs. Tran et al. (116) first applied DESs to recycle cathode materials from waste LIBs. The adopted DESs were synthesised by choline chloride and ethylene glycol, and the possible synthesis reaction is shown in Equation (ii):

(ii)

The aluminium foil and binder PVDF were recovered respectively while valuable metals were extracted by DESs. The results showed that the leaching rate of cobalt in LiCoO₂ cathode material was as high as 99.4% at 180ºC, which was comparable to that of traditional leaching agents such as sulfuric acid (99.70%) (70) and formic acid (99.96%) (72). Cobalt in the DESs liquids can be recovered as Co₃O₄ by electrodeposition and calcination.

Wang et al. (117) developed a recycling process using DESs to detach active substances and aluminium foil of retired LIBs, as the recovery flow shown in Figure 10. The specific synthesis reaction of DESs is shown in Equation (iii):

(iii)

Fig. 10

The separation of cathode materials and aluminium foil from spent LIBs by using choline chloride-glycerol DESs

The DESs attacked the hydrogen atoms in PVDF molecular chain, forming unsaturated double bonds, which were further oxidised into hydroxyl and carbonyl to construct an unsaturated ketone structure on the molecular chain. Therefore, PVDF was forced to deactivate and dissolve, the active materials and aluminium foil were separated successfully. Hence, it can be indicated that DESs have great potential in regenerating cathode materials. High selectivity of valuable metals is the key to promote leaching performance of DESs in recycling spent LIBs.

5.2 Nickel Cobalt Manganese Particles Recovery

Despite high recovery rates of R-NCM from leaching solution, electrochemical performance is damaged to some extent due to inevitable contact with organic solvent during the treatment process. In the future, recovery technologies with low efficiency must be avoided. Sieber et al. (118) directly recovered NCM particles from used cathode scraps. The active material was detached from the aluminium foil by strong stirring in water to maintain the electrochemical performance and morphological characteristics of NCM particles. However, during the separation process, side reactions occurred when the NCM cathode contacted with water, resulting in a rapid increase in pH value and degradation of NCM particles followed by forming of Al(OH)₃ precipitate and LiAl(OH)₄. The main chemical reactions during separation stage are as follows (Equations (iv–vii)):

(iv)

(v)

(vi)

(vii)

Such side effects can be avoided by adjusting pH value with superior buffer solution, aiming to detach active materials from aluminium foil and hinder degradation of NMC particles. Under buffer solution control, the aluminium foil was successfully dissolved into solution while NCM particles were well-preserved.

The waste cathode materials in spent LIBs contain valuable metals, which can be recycled by a hydrometallurgical process which offers low energy consumption, environmentally friendly and excellent recovery efficiency. Based on the present analysis and summary of hydrometallurgical recycling processes, the following conclusions are drawn. The following flow is considered the most competitive process to recycle valuable metals from waste cathode materials at industrial scale: alkaline solution dissolution/calcination treatment → reductant (H₂O₂) → sulfuric acid leaching → coprecipitation → resynthesis at high temperature. The leaching capability of organic acid with reducibility is greater than that of inorganic acid leaching united with reductants. Bacterial leaching shows high selectivity for lithium, which can be selected to recycle high purity lithium-containing products. A combination of processes for recovering valuable metals may complement each other, which contributes to improve the quality of the recycled products. The DESs separation technology can achieve efficient overall recovery of valuable components from spent LIBs, which is conducive to sustainable development. The highly efficient direct-regeneration or direct-recovery of NCM materials will face huge challenges and potential. In the future, both will become research hotspots in the field of spent LIBs recovery.

In 2011 the European Commission presented “A Roadmap for Moving to a Competitive Low Carbon Economy in 2050” outlining the milestones, among them also the 83–87% CO₂ reduction of the industry sector (1, 2). Primary steel production via the blast furnace/basic oxygen furnace (BF/BOF) route or so-called integrated steel plant, is a predominant and well-established process, contributing 70.8% of the world’s 1807 million tonnes of crude steel production in 2018 (3). The reduction process of the iron ore to crude steel is linked to CO₂ emissions, resulting in 2016 for a total of 7% (160 million tonnes of CO_2eq) of EU-28’s greenhouse gas emissions (4). The energy efficiency potential of a modern integrated steel plant has already been exploited to a great extent through conventional process optimisations. It is, therefore, necessary to transfer steel production to climate-friendly processes through new and innovative approaches. Hydrogen-based direct reduction processes and electrolytic reduction methods are alternatives for the reduction of iron ore, but they require on the one hand huge amounts of renewable energy, for instance for green hydrogen production in water electrolysis, and cause on the other hand significant investment demand as the existing production infrastructure has to be replaced (5). Carbon capture and utilisation (CCU) processes are a second option which are near-term actionable, as they can be added to the existing infrastructure without a significant change in the steel production itself (6). The first step of a CCU process chain is the energy intensive separation of CO₂ from diluted exhaust or process gases. If these gases also contain carbon monoxide, as is the case in the steel industry (Table I), carbon monoxide is avoided of utilisation since the carbon capture processes selectively separate CO₂ (8). The separated CO₂ is then either biologically or catalytically converted to usable products (6).

The process gases in a steel plant, BFG, BOFG and COG, contain high shares of carbon monoxide and CO₂ (Table I). These low-calorific gases are currently utilised in an integrated steel plant as an energy carrier, i.e. in heating processes, and as fuel in the power plant. In Figure 1 the energy flows in an integrated steel plant are depicted. The main part of the BFG (white letter A in Figure 1) is directed to the enrichment process where it is mixed with BOFG (white letter B in Figure 1). The main share of the enriched gas fuels the power plant. COG (white letter C in Figure 1) is mainly used in internal processes and in the power plant as well. The power plant covers almost the total electricity demand of the steel mill. NG (white letter D in Figure 1) is used for heating in downstream processes, like in the hot strip mill, and in the power plant as well. The quantity of the process gases BFG, BOFG and COG covers up to 40% (9) of the steel plant’s energy demand, where the remaining part is provided by electrical energy and fossil fuels, like NG.

Fig. 1

The Sankey diagram of Figure 2 indicates the composition of the different byproduct gases and their internal use. It is obvious that a withdrawal of these low-calorific gases has to be compensated by the supply of other energy carriers, either synthetic natural gas (SNG) or external electricity, since the main part of the gases are used in the power plant downstream of the enrichment process. In the enrichment process, the byproduct gases are mixed and buffered in gasometers.

Fig. 2

However, due to their high carbon monoxide and CO₂ concentrations, the process gases may be conceived for further use as a carbon source for catalytic conversions. As summarised in the review from Frey et al. (10) the alternative utilisation of steel plant process gases for the production of ammonia, methanol or recovery of its derivatives has already been under examination since the early 1950s. The latest review from Uribe-Soto et al. (11) outlined three alternatives: (a) thermal use of the process gases (state-of-the-art); (b) recovery of the valuable compounds (hydrogen, methane and carbon monoxide) via different separation technologies; and (c) thermochemical synthesis to high-added value products (methanol, dimethyl ether, urea), with the focus on the latter. The consideration of their utilisation as a potential carbon source and coupling it with renewable energy within the context of power-to-X technology, has gained attention in the last years especially in Europe, resulting in various theoretical studies as well as research projects. The largest German steel producer, thyssenkrupp, is leading the “Carbon2Chem” project (12), where different scenarios for the synthesis of methanol (13, 14), ammonia or urea (15), as well as higher alcohols and polymers (16) from BFG, COG and BOFG are being investigated. In an ongoing research project the possibilities of converting BOFG and BFG into methanol and methane are explored under dynamic conditions (17, 18).

All studies referenced above utilise pure CO₂ which is separated in a first step from the process gases, and is subsequently converted with hydrogen in a catalytic synthesis. In this study, the catalytic conversion of BFG and BOFG to methane is investigated without a prior separation of CO₂. The process gases of the steel plant are only pre-cleaned upstream of the catalytic conversion by dust removal (i.e. by a venturi scrubber and a bag house filter) and separation of the sulfur compounds sulfur dioxide, carbonyl sulfide, mercaptans (i.e. in a series of two adsorbers in order to remove the catalyst poisons), see Figure 3. Consequently, the catalytic conversion is carried out with significant shares of nitrogen in the feed gas. Therefore, the following advantages arise:

Fig. 3

The aim of the present study is the assessment of different process chains for the direct utilisation of BFG and BOFG in catalytic methanation without a prior CO₂ separation. Since the conversion of CO₂ and carbon monoxide to methane requires renewable hydrogen, the hydrogen supply is ensured by a water electrolysis powered by renewable electricity (P2G plant) as well as by an additional biomass gasification plant (Figure 3). Therefore, a variety of possible implementation scenarios arise, and the following fundamental research questions have to be answered:

A withdrawal of process gases, particularly of the comparatively high calorific COG, would result in a shortage of internal energy supply in the integrated steel plant (Figure 1) which has to be substituted by NG or electric energy sourced externally. Therefore, in order to avoid significant changes of the existing steel production infrastructure, in this study COG was not considered, and BFG as well as BOFG are solely used as a carbon source for a potential utilisation process (Question 1(a)). Furthermore, it has been deliberately decided that the product gas from the methanation, nitrogen diluted SNG, substitute fossil NG currently used in the integrated steel plant, mainly for heating processes. Alternatively, it is used as reducing agent in the blast furnace, for example as substitute for pulverised coal injection (PCI). An injection into the NG grid is not possible since the required specifications are not met (Question 3). The technoeconomic and ecological questions (Questions 1(b), 2 and 4) have been treated by Rosenfeld et al. (19). Supporting laboratory experiments for biomass gasification have been published by Müller et al. (20). The focus of this study is on Questions 5 and 1(b).

Figure 3 provides an overview of the possible integration of a P2G plant (Figure 3(b)) as well as a dual fluidised biomass gasification plant (Figure 3(c)) into the integrated steel plant (Figure 3(a) in dashed lines). By integrating a P2G plant, renewable energy is used for the production of hydrogen by water electrolysis and subsequently for the catalytic methanation of the process gases BFG and BOFG. The combination with a dual fluidised biomass gasification (20–22) provides an additional biogenic hydrogen source. The biogenic CO₂ is vented to the atmosphere. Alternatively, it could be stored in a carbon capture and storage (CCS) process resulting in negative CO₂ emissions (bioenergy with carbon capture and storage (BECCS)) which is not further considered here (23). In addition, the oxygen from the water electrolysis has the potential for utilisation in steel production as well as in the biomass gasification process. The produced nitrogen diluted SNG is directly utilised in the steel plant as a substitute for NG in various processes, and PCI in the blast furnace.

To explore the integration potential, a number of different scenarios has been defined and three of them, supported by the experimental results at a laboratory catalytic methanation plant, will be presented in detail in the present work. The three chosen scenarios provide a good overview of the order of magnitude of the required renewable energy, as well as the resulting CO₂ reduction potential.

The scenarios with different integration variations were based on Austria’s biggest steel production sites. The integration of renewable energy by a P2G plant and a biomass gasification plant has been analysed by three extreme value scenarios and three constrained scenarios. The results are reported in (19). The three extreme value scenarios described a maximum utilisation of the process gases, either individually or in combination. The required hydrogen for the methanation was balanced, half from water electrolysis and half from biomass gasification. The constrained scenarios are realistic in the medium term. They are limited by the maximum plant size of the biomass gasification plant (100 MW_th), based on the current biomass fuel availability and already installed gasification capacity in Europe (21). The main cost influencing factor throughout all six scenarios is the energy supply cost, both for electricity and for biomass (19).

The aim of the aforementioned scenarios was the minimisation of, or complete substitution of, the integrated steel plant’s demand for fossil fuels like NG and PCI. The steel plant process gases (BFG and BOFG) were used as carbon source for the methanation. The three scenarios which are the basis for the considerations in this paper are:

For these three scenarios, the required amount of the renewable electricity, biomass as well as the withdrawal amount of the process gasses (BFG or BOFG) has been determined. The main evaluation criteria for all scenarios were set by the CO₂ reduction potential.

Dual fluidised bed gasification systems consist of two reactors, the gasification reactor (650°C) and the combustion reactor (900°C). In contrast to the conventional systems, the presented system uses the sorption enhanced reforming (SER) process. It allows selective transport of CO₂ between the gasification reactor and the combustion reactor, by the use of calcium oxide as bed material, resulting in a product gas with a high hydrogen (up to 75 vol%) and low CO₂ concentration. The hydrogen rich product gas of the biomass gasification substitute green hydrogen from the electrolysis, and thus reduces the demand of renewable electric power (22, 24). Additionally, when pure oxygen is used instead of air for the combustion (oxySER), an almost pure CO₂ stream can be obtained as an exhaust (flue) gas, suitable as biogenic CO₂ source (Table II). The data given in Table II are based on the gasification of wood chips. A thermal gasification power of 100 MW_th consumes 50,400 kg h^–1 wood chips from Austria as fuel, and produces 28,800 Nm³ h^–1 product gas with the composition according to Table II (21).

The experimental tests were performed at a laboratory test plant, which consists of three fixed-bed reactors (R1–R3) connected in series with the purpose of achieving a multi-stage fixed-bed methanation. A detailed description of the test plant can be found in Kirchbacher et al. (25) and Medved (26). The conversion of CO₂ and carbon monoxide was investigated for synthetic gas compositions of BFG and BOFG under different flow rates, variation of hydrogen surplus and presence of nitrogen, with the focus on achieving a complete CO_x conversion.

A commercial bulk catalyst with 20 wt% nickel load was used. The operating pressure was set to 4 bar, which coincided with the steel producer’s gas supply system. The reactor load was limited to gas hourly space velocity (GHSV) of 4000 h^–1 (GHSV = V̇_feedgas/V_catalyst) for the synthetic BFG and BOFG gas composition and added hydrogen. The temperature in the reactor was determined by multi-thermocouples (Figure 4).

Fig. 4

Seven measuring points in each reactor, five in the catalyst bed and one below and above the catalyst zone, gave an understanding of the axial temperature profile in the catalyst bed. The methanation gas composition for BFG and BOFG for the stoichiometric ratio with hydrogen according to Equations (i) and (ii) is listed in Table III.

(i)

(ii)

The experimental results obtained from the methanation of BFG and BOFG were used as support for the further analysis of the three selected scenarios. In the following, the experimental results and the scenarios are presented separately in the subsections.

5.1 Methanation

The methanation of process gases is a combination of CO₂ and carbon monoxide conversion according to Equations (i) and (ii).

A suitable parameter for the description of the stoichiometry is the ratio r_{H 2} of molar hydrogen flow and molar flows of CO and CO₂, respectively, in the feed gas given in the Equation (iii):

(iii)

r_{H 2} equals 1 for stoichiometric mixtures, r_{H 2} <1 for sub- and r_{H 2} >1 for over-stoichiometric mixtures, respectively.

Achieved CO_x conversion rates for each reactor (R1–R3), with variation of hydrogen surplus (r_{H 2} = 1; 1.02; 1.04; 1.05) with and without nitrogen for a synthetic BFG and BOFG feed gas compositions, can be seen in Figures 5 and 6. On the right y-axis, the mean reactor temperature represents the average of the measured catalyst bed temperatures, as well as the calculated heating values of the product gas in each reactor. For experiments without nitrogen in the feed gas, marked with -N₂, the H₂:CO_x ratio and the amount of the reactive gas in the experimental series remained the same, meaning that the GHSV was reduced to 3260–3280 h^–1 for BFG and 3680–3690 h^–1 for BOFG due to the absent inert gas flow.

Fig. 5

Methanation of BFG with and without nitrogen and hydrogen-surplus variation

Fig. 6

Methanation of BOFG with and without nitrogen and hydrogen-surplus variation

5.1.1 Methanation of Blast Furnace Gas

Complete CO_x conversions are achieved downstream of the third reactor with an over-stoichiometric ratio of 1.05, both with and without the nitrogen in the feed gas (Figure 5). The mean reactor temperatures are approximately 50°C lower with nitrogen present, which is a result of the additional heat capacity of the inert gas. Despite these lower temperatures, approximately 4% better conversions are reached in R1 for all ratios in the absence of nitrogen. The withdrawal of nitrogen from the feed gas resulted in lower GHSV, consequently prolonging the residence time in the reactor leading to slightly better CO_x conversions. The temperature decrease in R2 and R3 was expected, since the majority of the reactive gas converted in R1, resulting in lower release of the exothermic reaction heat. Therefore, nitrogen in the feed gas only has a significant influence on the heating value of the product gas. In the case of the product gas (R3) with nitrogen the heating values vary from 19.4–19.8 MJ m^–3 (r_{H 2} = 1.05–1), whereas without nitrogen the values almost double (36.0–37.9 MJ m^–3). Although with the higher hydrogen surplus better conversions are achieved, the unconverted hydrogen decreases the heating value of the product gas, due to its lower volumetric heating value compared to methane.

5.1.2 Methanation of Basic Oxygen Furnace Gas

As for BFG, similar test series were conducted for the methanation of BOFG. As shown in Figure 6, on account of lower nitrogen share in the feed gas (8.2%), no noticeable effect on the temperature and consequently conversion can be recognised. Furthermore, a complete CO_x conversion at 4% hydrogen surplus is achieved with or without nitrogen in the feed gas. When comparing the mean reactor temperatures and CO_x conversion in R1 of BOFG with the BFG test series, temperatures are 50–100°C higher and conversions 5–10% lower, respectively. This can be attributed to the higher carbon monoxide share in BOFG, resulting in higher reaction heat release. Therefore, the conversion in the first reactor (R1) is clearly thermodynamically limited. Since its lower share in BOFG compared to BFG, the influence of nitrogen on the heating value of the product gas is lower, and values vary from 27.2–28.8 MJ m^–3 (r_H2 = 1.05–1) with nitrogen and between 34.9–37.7 MJ m^–3 (r_H2 = 1.05–1) without nitrogen.

The product gas composition downstream of the methanation is given in Table IV. A complete conversion is achieved at 5% hydrogen surplus for BFG and at 4% hydrogen surplus for BOFG, where the unconverted hydrogen is a result of its over-stoichiometric addition.

Table IV

Product Gas Composition for the Methanation of BFG and BOFG

	Product gas molar fraction
	CH4	CO2	CO	N2	H2
BFG (rH2 = 1.05)	0.446	0	0	0.434	0.120
BOFG (rH2 = 1.04)	0.679	0	0	0.215	0.106

5.2 Results for the Selected Scenarios

In this study, three different scenarios for the implementation of a P2G plant and a biomass gasification in an integrated steel plant have been investigated. The aim was the quantification of the CO₂ emission reduction potential of steel production, avoiding significant modifications in the existing steel plant infrastructure. Furthermore, a carbon capture step shall be avoided as well, in order to improve the efficiency of the CCU process chain, resulting in a catalytic conversion of BFG and BOFG to methane in the presence of nitrogen.

Basic evaluation of the three chosen scenarios confirmed the possibility of CO₂ emission reductions between 0.81 and 4.6 million tonnes CO_2eq per year without considerable interference with existing steel production. The required plant sizes and the necessary fuel demand (renewable power and biomass, respectively) substantially exceed the current realistic possibilities of a P2G plant (electrolyser power 784–2877 MW_el) as well of a biomass gasification (100–3162 MW_th). Even for the scenarios realistic in the medium term, the amount of required renewable electricity beyond 700 MW_el cannot be provided in the foreseeable future. This underscores the need for new technologies for the production of CO₂-free hydrogen.

Experimental tests have shown that the methanation of BFG and BOFG is technically possible without separating the inert gas nitrogen, and thus saving the energy intensive carbon capture unit with the additional benefit of carbon monoxide utilisation contained in the process gases from the steel production. A complete conversion of CO_x was achieved with a 4–5% hydrogen surplus for both process gases, BFG and BOFG, with and without nitrogen, with three-stage methanation. The lower heating value enrichment of the unrefined BFG (up to 19.8 MJ m^–3) and BOFG (up to 28.8 MJ m^–3) via methanation without the necessity of nitrogen removal as lean product gas showed a utilisation potential in the integrated steel plant as a substitute for NG and PCI.

The first evaluation presented here provides a good overview on the order of magnitude of required renewable energy and biomass for the transition of the integrated steel plant towards renewable gas supply by adding a CCU process chain. Additionally, particularly for the utilisation of the product gases of catalytic methanation within the steel plant, intricate CO₂ separation is not required as has been shown by the experimental investigations.