pubs.acs.org/est Article New Techniques for Assessing Critical Raw Material Aspects in Energy and Other Technologies Nick Martin,* Cristina Madrid-López, Gara Villalba-Méndez, and Laura Talens-Peiró Cite This: Environ. Sci. Technol. 2022, 56, 17236−17245 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information ABSTRACT: Transitioning to more sustainable energy technol- ogies is a vital step in the move toward reducing global greenhouse gas emissions. However, several physical constraints could hinder the implementation of these technologies, and many of the raw materials required to produce new infrastructure are scarce, nonrenewable, and nonsubstitutable. Various factors relating to material extraction and processing activities may also affect the security and sociopolitical aspects of future supply lines. Here, we introduce methods for quantifying three key indicators relating to raw material supplies for specific production processes: (1) overall supply risk, (2) environmental impacts from sourcing raw materials, and (3) environmental justice threats at sourcing locations. The use of the proposed methods is demonstrated via an exploratory case study examining projected electricity production scenarios within the European Union. Results suggest that renewable sources of electricity�particularly wind, solar, and geothermal technologies�are more likely to exacerbate supply risks and environmental issues than other technologies. Furthermore, projected expansions of wind and solar technologies mean that all three indicators appear likely to rise significantly systemwide by 2050. Ultimately, the methods represent a much-needed first attempt at providing practitioners with simple and robust approaches for integrating factors relating specifically to raw material supply into energy modeling and other applications. KEYWORDS: energy transition, critical raw materials, material supply, life cycle assessment, just transition, energy justice, energy modeling ■ INTRODUCTION Scientists and policymakers have now widely accepted the with supplying countries,19−21 and the issues of social justice and localized environmental damages that surround the need to reduce emissions of greenhouse gases (GHGs) at all increased demand for materials.22−24 All three aspects are scales.1 This is reflected in symbolic global initiatives like the likely to play a role in determining the speed and direction of Paris Agreement2 and in the many national, regional, and local the energy transition going forward. The European Commis- policies that are being formulated to address the issue. Within sion (EC) has begun to quantify supply risk for specific the rapidly evolving arena of energy and environmental policy, materials16 and now includes geographical concentration and the need to accelerate the adoption of more “sustainable” governance, import reliance, and responsible sourcing aspects sources of energy is viewed as one of the key pathways to as part of its triennial “Raw Materials Scoreboard” assess- ments.25 A handful of additional studies have also attempted to reducing emissions and achieving future targets.3 However, the concept of sustainability in energy systems is measure other aspects of material sourcing, particularly in evolving beyond the mere reduction of GHG emissions. relation to justice and conflict issues.26−28 However, these Among other things, the ongoing sourcing of the raw materials assessments generally only apply to individual materials. As and components required to implement new infrastructure such, despite a relative paucity of suitable data, a clear need for the quantification of raw material-related constraints relating to continues to gain policy focus4−8 and mainstream media individual technologies and processes is arising, particularly for attention,9−13 and several potential “roadblocks” have been identified. The range of issues triggered by the COVID-19 pandemic and war in Ukraine has further highlighted the Received: July 26, 2022 vulnerability of infrastructure development to supply chain Revised: November 13, 2022 disruptions.14,15 Accepted: November 14, 2022 A number of specific concerns have been raised in this Published: November 24, 2022 regard, mostly surrounding the available stocks of necessary materials,16−18 geopolitical and governance issues associated © 2022 The Authors. Published by American Chemical Society https://doi.org/10.1021/acs.est.2c05308 17236 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article Figure 1. Conceptual overview of methods used for deriving the three raw material indicators for the extraction and processing of 55 selected materials. It is noted that all three indicators relate solely to the activities involved in deriving and supplying a specific group of raw materials to a process and do not attempt to quantify all supply risk, environmental impacts, or environmental justice aspects relating to the entire life cycle of that process. those attempting to optimize systemwide transition pathways Union (EU) electricity network. A validation and sensitivity and minimize the exposure of these pathways to risk. analysis is also provided. Findings from the case study and To bridge this gap, we present here a series of methods� further aspects of the methods are then discussed alongside a developed for assessing energy system characteristics as part of final set of conclusions. the SENTINEL project29�that use raw material inventory information from life cycle assessment (LCA) databases alongside other data sources to generate three unique indicators specifically related to the supply of raw materials. ■ METHODOLOGY The proposed methods all use material requirement First, the risk of interruption to raw material supply channels is information provided by life cycle inventory (LCI) data as quantified by incorporating supply risk data published by the their foundation. An LCI represents one of the four phases EC.16 Two further indicators attempt to quantify the within a life cycle assessment (LCA).34 During this phase, all of possibility of localized issues occurring during the extraction the elementary material and energy flows that occur within a and processing of raw materials: the potential to exacerbate process are determined. This includes all sub-processes that local environmental conditions is estimated using ecosystem occur during the material extraction, processing and and human health data relating to individual materials from the manufacturing stages�and, if required, the product use and Ecoinvent LCA database,30,31 while the potential to reproduce disposal stages�within the entire life cycle of a process. The local environmental justice issues is quantified using data resulting breakdown includes listings of all inputs and outputs relating to sourcing countries within the Worldwide that occur for a range of different materials. Furthermore, it Governance Indicators (WGI) data set.32,33 Collectively, we will include specific items for the process in question�the believe these three indicators represent the majority of key “foreground” system�alongside those for the broader issues in relation to raw material supply at present. industrial economy�the “background” system. Final material The methods enable composite values to be derived for requirements are given as the total mass of a material required individual technology types or, indeed, for any unit process to produce one “unit” of a process. Here, we use a small defined within LCA databases; higher scores highlight selection of the available LCI data�using the Ecoinvent processes that involve material sourcing from locations with database30�to perform a customized set of calculations higher inherent risks of supply interruption, with poor relating to the supply of a specific set of raw materials to a environmental impact characteristics, or where environmental given process. justice issues are potentially more likely to occur. Values could The methods use 55 of the raw materials identified as being be integrated into existing energy modeling applications to most important to the EU in accordance with the latest list of account for these aspects�for example, as in-built calculations so-called critical raw materials (CRMs) published by the EC. or soft-linked constraint parameters within integrated assess- The most recent investigation, from 2020,16 considers a group ment models or other energy system models�or be used as of 80 materials as potential CRM candidates, of which 44 were standalone indicators for assessing proposed energy system deemed critical using a standardized methodology35 based on configurations in other applications. Full descriptions of the economic importance and supply risk factors. The list includes methods and suggested data inputs for each indicator are the five platinum group metals, 10 heavy rare earth, and five provided. The approach is then operationalized via a case study light rare earth elements; holmium, thulium, lutetium, and involving current and projected scenarios for the European ytterbium are grouped as a single heavy rare earth entry. 17237 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article An attempt to align the 80 candidate materials from the required and available masses of materials such as silicon or 2020 EC study with the listings in the December 2020 version titanium can be up to 5 orders of magnitude higher than those of the Ecoinvent LCI database30 found that 30 of the 44 CRMs of rare earth materials. To overcome this bias, EU annual and 25 of the remaining 36 candidate materials are represented consumption levels4,16,37 were used as a “scaling” measure to in the database; full documentation is provided in the represent the relative magnitudes of the requirements for Supporting Information. While it is observed that 25 of the different materials in the EU. As such, each material 80 materials were found to have no suitable match in the LCI requirement value was first normalized by being divided by database, it is noted that 14 of these “missing” materials were the corresponding EU consumption rate. Accordingly, the categorized as “industrial and construction”�for example, proposed formula for calculating the net SR factor for a given aggregates, rocks, sand�or “biological and other”�for process is as follows example, rubber, cork, and wood�many of which are either i=1 too generic, not relevant, or too complex to quantify in LCI mi SR i SR process = listings. ci n Material requirement data for a given process�relating to the 55 selected materials�is then used alongside other data where, SRprocess = net supply risk factor for the process under for each material to create the three composite indicator study [yr/MJ], n = number of selected individual materials in values. That is, individual “scores” can be obtained for any the process under study, mi = mass of material i required by the process defined by an LCI. The three final scores then enable process under study [kg/MJ], SRi = supply risk factor of direct comparisons of raw material indicators to be made for material i [dimensionless], and ci = annual consumption level different processes. However, though the approach fundamen- in EU of material i [kg/yr]. tally provides scores for unit processes, the obtained scores can It is noted that, while the final value for the net SR factor is also be upscaled to provide composite scores for entire systems essentially dimensionless, the final units are actually the time of individual processes. For example, in an energy system, each frame of the consumption data divided by the unit that the indicator can be applied pro-rata according to the relative material intensity is based upon�in this case, the relatively contributions of each energy source to obtain composite scores meaningless years per megajoule. Although calculations could that allow complete system configurations, such as those also be undertaken using LCI data for processes based on derived from energy systems modeling, to be compared. In this different “functional units”�for example, megawatts of manner, the raw material characteristics of current and installed capacity or kilometers of travel�these would proposed energy systems can be analyzed for energy policy naturally return final values in different units. Though this and planning purposes. Furthermore, as the methods are demonstrates the flexibility of the method, it follows that one generically based on LCI definitions, it can equally be applied cannot directly compare final scores based on different to any process defined within existing LCI databases. Figure 1 functional units or consumption data. provides a final conceptual overview of the proposed approach Local Environmental Impacts. A second method was prior to the detailed descriptions of each method. developed to capture the potential for local environmental Supply Risk. The first method attempts to quantify the damages to occur during the extraction and processing of level of supply risk inherent to the sourcing of raw materials for primary materials for a given process. Here, we once again rely a given process. Alongside an indicator of economic on LCA data from the Ecoinvent database.30 However, in this importance, the EC uses a derived measure of supply risk instance, we follow the methodological guidance of Graedel et (SR) as one of the two key inputs in its own assessments of al.38 by utilizing LCIA endpoint indicators for the production CRM status.25,35 In essence, the EC’s SR factor quantifies the processes of individual materials. As in this study, dimension- potential risk of a disruption occurring within the supply chain less indicators are derived for both ecosystem quality and of materials by considering the current sourcing locations of human health for the production of a single kilogram of each supplies and the governance and trade attributes of those material in accordance with the ReCiPe Endpoint (H,A) locations. It is derived using the latest (2020) data for overall method.31 EU import reliance, a circularity indicator (the end-of-life The ecosystem quality indicator aggregates values for recycling input rate), a substitution index, and two versions of terrestrial acidification, terrestrial ecotoxicity, freshwater the Herfindahl−Hirschman Index (HHI)36�derived using ecotoxicity, marine ecotoxicity, freshwater eutrophication, Worldwide Governance Index (WGI)32 data�that reflect the agricultural land occupation, urban land occupation, natural locational concentration and governance issues for the land transformation, and climate change (on ecosystems). countries supplying the material at both EU and global levels. Analysis of the data suggests that this indicator is over- A complete listing of raw SR factor values for the materials whelmingly influenced by some combination of marine examined in the study is provided in the Supporting ecotoxicity, natural land transformation, and climate change Information. values for all materials examined. Meanwhile, the human health Here, a composite SR indicator for a particular process is indicator aggregates values for human toxicity, photochemical created by summing all SR factor values in proportion to the oxidant formation, particulate matter formation, ionizing amount of the corresponding material required (mass) to radiation, ozone depletion, and climate change (on human produce one unit of the final “product” defined by the LCI. health). In this case, the indicator is overwhelmingly influenced Initial attempts at deriving the indicator considered only these by human toxicity values for all materials. A simple average of two inputs. However, it was soon discovered that using “raw” the net ecosystem quality and human health indicators was values of material requirement placed a large bias on materials used as the final environmental impact (EI) value for each used in larger amounts; this tended to vastly overshadow the material. Although it is acknowledged that these impacts could significance of scarce materials used in much smaller amounts. occur anywhere along the supply chain of these raw materials, For example, although both are considered to be CRMs, the it is assumed here that a significant amount is directly related 17238 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article to the extraction and processing operations that occur near or +2.5�where negative scores denote less desirable conditions close to their source locations. A full listing of the processes, and positive scores denote more desirable conditions�values LCIA endpoint indicators, and final composite EI values for for each indicator and each country are first normalized to each material are provided in the Supporting Information. It is percentage scores according to the range observed across all noted that the values for gold and the three platinum group countries in that category. Accordingly, the proposed formula metals (PGMs)�palladium, platinum, and rhodium�are for calculating normalized composite WGI scores that equally orders of magnitude higher than most of the other materials weigh each indicator for each country is as follows tested. This can be traced primarily to extremely high impacts encountered specifically during the extraction and refinery g=1 vg , h min{vg , h}hz = 1 1 h operations relating to these metals.30 WGIh = p max{vg , h}hz = 1 min{vg , h}hz = 1 The values for individual materials are then used to create a p h h final indicator for a given process as follows where, WGIh = composite WGI indicator for country h [%], p i=1 mi EIi = number of individual indicator categories in WGI database, EI process = v g,h = value of indicator number g for country h ci n [dimensionless], z = number of individual countries in WGI database. where, EIprocess = net local environmental impacts score for the Composite EJ indicators for each material are calculated by process under study [yr/MJ], n = number of selected combining the WGI scores for each country and the individual materials in the process under study, mi = mass of percentage breakdown of global supply sources for each of material i required by the process under study [kg/MJ], EIi = the 80 candidate materials. However, as higher WGI scores local environmental impacts score for material i [dimension- reflect better environmental health characteristics, the values less], and ci = annual consumption level in EU of material i used are inverted by subtracting them from unity. Accordingly, [kg/yr]. the proposed formula for calculating the net environmental Local Environmental Justice. A third method adopts a justice indicator for a given material is as follows similar approach, this time attempting to determine how (un)just the sourcing of raw materials is likely to be for a given h= 1 process. Though perhaps less directly tangible than SR and EI, EJi = sh , i (1 WGIh) the environmental justice (EJ) indicator seeks to widely n embody a set of concepts that includes conflicts relating to the effects of pollution and the distribution of environmental where, EJi = local environmental justice score for material i risks.39 While the energy transition is widely predicted to [dimensionless], n = number of countries included in analysis, exacerbate such issues at the global scale,40,41 much of the sh,i = share of global supply of material i sourced from country existing discourse on “energy justice” is focused on the siting of h [%], WGIh = composite WGI indicator for country h [%]. new facilities and the extraction and mining of fuels42−47 or on As with the previous indicators, the composite EJ values for the embodied impacts caused by outsourcing energy, products, each material�as listed in the Supporting Information�can and services from other countries.48 In addition to this, a small then be used to create a final indicator for a given process, viz., number of previous studies have attempted to broadly address i=1 environmental justice issues in relation to the new infra- mi EJi EJprocess = structure required to implement the energy transition.26,49 n ci Meanwhile, a growing number of studies are attempting to quantify27,28 or catalogue50 justice-related issues specifically in where, EJprocess = local environmental justice score for the relation to resource extraction and processing. Moreover, the process under study [yr/MJ], n = number of selected burgeoning field of social life cycle assessment (sLCA) is individual materials in the process under study, mi = mass of beginning to address the impacts caused within these stages, material i required by the process under study [kg/MJ], EJi = including those used in energy production and in the new local environmental justice score for material i [dimensionless], renewable energy infrastructure in particular.23 Nevertheless, ci = annual consumption level in EU of material i [kg/yr]. to date, no studies have quantified justice elements in relation Possible Applications. Calculating values of the three to specific materials or processes. indicators for individual life cycle processes allows comparisons Here, we once again utilize information from an established of different technologies or sub-technologies to be undertaken. dataset as a proxy indicator within the method. In this case, a For example, the indicators derived for a unit of heat or composite value has been derived for each material using electricity from non-renewable sources could be directly values taken directly from the Worldwide Governance compared with various renewable sources. Likewise, results Indicator (WGI) dataset,32,33 as used within the EC’s for different sub-technologies could be compared within a derivation of supply risk factor.16 The WGI provides values technology group such as wind turbines or solar photovoltaic by country across six categories: voice and accountability, (PV) panels. Moreover, while the present article focuses on political stability and absence of violence/terrorism, govern- energy-related applications of the methods, it could theoret- ment effectiveness, regulatory quality, the rule of law, and ically be applied to any process defined by an LCI. control of corruption. All six categories are thought to be At a wider scale, scores for entire systems can be generated generally associated with conditions that enable or reflect the by tallying the product of the indicator and the total energy potential for environmental justice issues to occur and, hence, generated by each technology to derive final systemwide are assumed to provide a suitable proxy for the potential values. This would enable, for example, the characteristics of occurrence of such issues. However, as the scores are provided current systems to be compared against multiple future on an arbitrary scale that typically ranges from around −2.5 to alternatives to inform policy decision-making. The proposed 17239 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article Figure 2. Results for supply risk factors and environmental impacts and environmental justice scores for all available processes, grouped by category. Mean values shown as colored triangles. formula for calculating aggregated scores over entire systems is scenario as a basis, all 51 regionally applicable electricity as follows production processes within the 2021 version of the Ecoinvent i=1 LCI database30 were collected and grouped. Values of the Isystem = Ei Ii three indicators were then derived for each individual process n on a per-MJ basis as displayed in Figure 2. A full listing of the inputs and results is supplied in the Supporting Information. where, Isystem = aggregated indicator score for the system under The results for the three indicators demonstrate a relatively study, n = number of selected individual processes in the clear pattern across all three methods. The mean results by system under study, Ei = total energy production derived from category suggest that risks and impacts are considerably lower technology, and iIi = indicator score of the process i. for lake and river hydropower and nuclear processes, reflecting ■ CASE STUDY: EU ELECTRICITY SUPPLY To demonstrate the value and functionality of the proposed their relative simplicity and lower reliance on CRMs. Values for the three fossil fuel sources�natural gas, petroleum, and solid fossil�are typically moderate, although natural gas scores methods, they are applied here to an exploratory case study are generally lower for SR and EJ. Notwithstanding this, major involving existing and projected electricity generation levels for variations are observed for natural gas in the EI category, the EU, by technology, according to the EC’s latest “reference where three of the 12 processes are significantly higher as a scenario.”51 Values are firstly derived for all available individual result of their high reliance on platinum and rhodium; all other LCI listings within the Ecoinvent database.30 Using mean natural gas processes are far more consistent with scores values for each technological group defined within the EC data, observed for the other two indicators. aggregated system values were then derived using values from Values for biomass sources also tend to be in this moderate the EC scenario data sets to determine predicted changes in range alongside offshore wind turbines, although the value for the three indicators under these assumptions. offshore wind is somewhat lower in the EI category. Individual and Grouped Scores by Technology. Using Conversely, the scores derived for onshore wind are the 11 technological categories defined within the reference approximately double these levels as a result of their elevated 17240 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article Figure 3. Normalized values of overall supply risk (SR) factors and environmental impacts (EI) and environmental justice (EJ) scores for the EU electricity system for the period 2005 to 2050 according to technological projections defined by the EU reference scenario.51 All values are relative to the dimensionless values of SR, EI, and EJ for 2005 of 67.8, 219.8, and 7.7, respectively. It is noted that the projections are based on the current values for SR, EI, and EJ for each material. As such, they do not purport to predict any future variations in these factors for different materials going forward or to reflect previous values. Rather they are used to broadly illustrate the vulnerability of forecast energy systems in the EU to raw material supply issues based on current estimates for each material. reliance on rare earth materials, predominantly in the that are significantly different from the findings for SR and EJ. permanent magnets used in certain generator mecha- Notably, despite the fact that both include data from the WGI nisms;52−54 the offshore turbine assessment within the database in their derivation, the “R-squared” (R2) value Ecoinvent data set assumes the use of hybrid approaches comparing SR and EJ at the material level was found to be that rely less on rare earth materials. Values for geothermal relatively low (0.15780). In any case, the common materials energy are high in the SR and EJ categories but are noticeably use amounts used in both calculations scale up these factors lower for EI and are only considered moderate. and provide similarity at the process level. The solar technologies group�which includes both PV Current and Projected Scores for the EU System. To panels and concentrated solar power (CSP) plants�is more demonstrate the application of the three indicators to real- extensive than other categories, reflecting the many different world scenarios, they are applied to projected values of gross approaches employed in the field. Values for different solar electricity generation, by source, from the EU reference technologies range from moderate to very high and more or scenario;51 observed and projected values for the 11 categories less cover an entire order of magnitude for each indicator. are provided at five-year intervals from 2005 to 2050, as listed Copper indium gallium selenide (CIS) cells represent the and illustrated in the Supporting Information. The raw data higher scores in all three indicators, largely based on a strong demonstrate that wind, solar, and biomass are the only need for gallium. technologies to have risen significantly since 2005 although The relatively consistent trend observed in the results this trend is not expected to continue for biomass. Although reflects the influence of using the same masses of material geothermal sources are expected to rise slightly after 2035, the (mi)�derived from LCI listings�across all three sets of utilization of onshore and offshore wind and solar technologies calculations. As such, each indicator can be seen to be, first and are projected to increase by factors of 2.9, 7.6, and 4.2, foremost, a reflection of the total amount of all key materials respectively, between 2020 and 2050. All other technologies required�relative to total consumption�per unit of output; a are seen to remain relatively stable going forward. However, in process that uses higher levels of key materials overall will the cases of petroleum and solid fossil fuels, levels are predicted always be more likely to obtain higher scores than those with to decrease by factors of 30.2 and 9.6, respectively. As such, lower material requirements. In this sense, while the overall wind, solar, petroleum, and solid fossil fuels are expected to trends are clear cut, the three inputs applied for each have the biggest influence on overall changes across all three material�SRi, EIi, and EJi�can be viewed as contributing factors. varying levels of additional “scaling” within each calculation. Values for individual technological categories are first Nevertheless, variability in these “scaling” inputs can still be calculated by multiplying the mean values for each influential and result in significant variations in indicator indicator�in yr/MJ�by the amount of energy reported for results, particularly where inputs are not well correlated for a that category�in MJ/yr�in the EU data. Final system values given set of materials. This issue is further investigated via a for the three indicators are then calculated by aggregating the series of regression analyses, provided in the Supporting scores for all 11 categories. Final values for the three indicators Information. Regression analyses on the three indicators at the at each interval�normalized to “base” levels in 2005�are material and process level revealed that the results appear to be shown in Figure 3; a full listing of these results is also provided suitably “unique” at the material level, particularly for EI values in the Supporting Information. 17241 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article The results indicate that the overall values of each indicator fragility of global markets to supply chain issues and made the are likely to increase dramatically under the forecast scenario. consequences of such disruptions more tangible in the minds Values of SR and EJ are both predicted to more or less double of many. Indeed, Russia currently produces 33 of the 44 between 2020 and 2050; the value of EI is expected to rise by materials identified as CRMs by the EC;16 for five of these� around 56%. A small downturn was noted in the SR and EJ palladium (40.0%), scandium (26.0%), titanium (22.0%), observations between 2015 and 2020, largely due to the platinum (12.9%), and rhodium (12.0%)�Russia supplies especially hurried withdrawal of solid fossil fuels and over 10% of current global supplies. petroleum�which have significantly higher per-MJ values for Meanwhile, China is a known producer of 39 of these 44 both SR and EJ than legacy non-renewable technologies like materials and is responsible for over 80% of current global natural gas and nuclear�and the slight reduction in the supplies of 16 such materials, including gallium, germanium, growth rate of the highest-ranking group, solar. Similarly, a and all light and heavy rare earth metals, all of which are drop in EI values between 2010 and 2020 is predominantly the important in the manufacture of wind turbines and solar PV result of a drop in the use of natural gas, which has the second- panels.4 Ongoing tensions between China and the west could highest per-MJ value for EI. A full set of results is provided in have very serious implications in this regard.52,55,56 For certain the Supporting Information. materials, increased levels of recycling could help offset strong It is noted that these calculations do not take future import reliances, although recycling activities would also need variations in SR, EI, and EJ inputs into account. Present-day to be undertaken at the local level to avoid further supply- values are assumed when, in reality, these values are likely to related issues relating to the importation of recycled materials. fluctuate over time because of geopolitical shifts, technical Either way, circularity principles look likely to become an advancements, changes in recycling practices, or discoveries of integral part of future raw material landscapes.57,58 Never- new reserves. Nevertheless, this example provides a simple theless, many CRMs are technically difficult to recover from demonstration of the potential issues that could result from waste streams25 and strong reliance on newly extracted transitioning to renewable technologies that will most likely materials looks set to continue for the foreseeable future. continue to rely on materials with higher risk factors. Collectively, these observations highlight the need to continue In the end, the key finding here is that the substantial rises in to monitor key materials and to assess the indicators that best electricity from wind and solar sources that are predicted by reflect the status of these materials over time. 2050 look likely to result in significant increases in the net In any case, while most discussions in this area concern the scores for all three of the examined indicators. Indeed, onshore locations of global reserves and the importance of maintaining wind and solar technologies are predicted to generate the adequate supply lines, localized environmental impacts during highest and third-highest amounts of electricity, respectively, the material extraction and processing stages, and aspects of by 2050 while also representing the third-highest and highest environmental justice that relate to these impacts, are per-MJ scores for each of the indicators. increasingly being considered. The methods introduced here Sensitivity Analysis. While using annual EU consumption represent a first attempt at addressing this gap. Furthermore, as values, ci, appears to be a logical way to normalize scores and the three methods are fundamentally based on listings of avoid issues of disproportionate weighing in the presented individual materials required to produce one “unit” of a given methods, it was deemed necessary to test the influence of these process, they could theoretically be applied to any process values on final scores. To do this, a simple sensitivity analysis defined by an LCI listing and could, theoretically, find use in was undertaken. To allow for uncertainties in the estimates of any number of applications inside and outside of the energy ci, an additional 20% was added to the annual consumption sphere. values for a group of 13 key materials, all of which are highly Results from the case study strongly suggest that renewable influential in determining indicator values while having annual technologies within the wind, solar, and geothermal categories consumption rates of less than 1,000 tonnes. As expected, all present higher SR, EI, and EJ values than other technologies, three indicators were shown to be sensitive to these changes, while fossil fuel technologies tend to present midrange values. with reductions of between 11.7 and 16.6% being observed. The higher scores for solar and wind energy present a However, very low standard deviations�between 1.0 and particular cause for concern in this regard, especially when 2.8%�were observed within the changes, suggesting that the coupled with the fact that both technologies are expected to method can maintain consistent delineation between processes play key roles in most predicted transition scenarios world- when uncertainties in inputs are experienced. Full details of the wide.51,59,60 While continuing to rely on fossil fuels would analysis are contained in the Supporting Information. result in lower scores in all three indicators, other ramifications ■ DISCUSSION Raw material supply is an ongoing concern in relation to the relating to these technologies�not least of which are far higher GHG emissions�mean that they are generally no longer considered viable future alternatives. Conversely, transition to renewable energy sources. Although we are although hydropower, biomass, and nuclear technologies also limited to present-day assumptions about material supply bring their own constraints and controversies, it is noted that characteristics, applying newly developed methods to EU their potential to introduce disturbances is among the lowest in system projections strongly suggests that the potential for all three metrics considered here. At any rate, it is hoped that environmental impacts and justice issues to occur during the the methods and findings presented will further highlight the extraction and processing phases of the identified set of key seriousness of raw material issues in energy transition processes materials looks likely to rise dramatically over the coming and the need to interrogate and balance these aspects when decades. Likewise, the overall risks associated with obtaining considering different technological options. these materials also look set to increase sharply based on Nonetheless, while these approaches are thought to current projections. Recent disruptive events such as the represent an original and valuable contribution to the field, COVID-19 pandemic and war in Ukraine have highlighted the several limitations are noted. First, they only consider the 17242 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article group of 80 materials identified as potentially critical by the Such assessments would be large undertakings and are well EC.16 As such, other key materials could potentially be beyond the boundaries of the current study. However, the neglected for certain processes, and aspects relating to the concept could provide a basis for future research. extraction and processing of fossil fuels and uranium� The supply of raw materials looks likely to remain a concern particularly in relation to localized environmental impacts as we attempt to implement greater levels of renewable energy and justice issues�are not included. Furthermore, 25 of the 80 and other strategic technologies going forward. As such, robust identified materials are not currently represented in the LCI methods for quantifying the constraints and other aspects databases. Again, though many of the omitted materials are not relating to raw material supply are vital to ensuring that considered vital, materials such as niobium, germanium, and decarbonization pathways are optimized at all levels. In this indium are known to be important in a number of key future sense, it is hoped that the introduced methods provide a technologies.4,6,49,61,62 Wider inclusion of materials in future valuable new contribution to the field of raw material supply at LCI data releases would provide more robust coverage in this large and a specific starting point for energy modeling and respect. related applications as we strive to optimize pathways toward Similarly, many key technologies are poorly represented in more sustainable energy systems. current LCI data sets, limiting deeper analyses or comparisons. For example, only one type of geothermal electricity and two types of solar CSP are represented, and listings for key renewable energy technologies such as biofuel production, ■ * ASSOCIATED CONTENT sı Supporting Information power-to-gas (P2G), power-to-liquids (P2L), hydrogen The Supporting Information is available free of charge at electrolysis, and most forms of electrical storage are almost https://pubs.acs.org/doi/10.1021/acs.est.2c05308. entirely absent in the current databases. Wind power, widely predicted to be a dominant player in most future energy Description and results of validation and sensitivity scenarios, is only represented by three onshore processes and analysis (PDF) one offshore process in the latest Ecoinvent database compared Full listings of data used in analysis (XLSX) to the 19 listings for solar technologies. Although data can sometimes be obtained from secondary sources,63 more complete listings of key technologies within universal databases such as Ecoinvent64 and GaBi65 would greatly improve the ability of practitioners to assess future energy systems. ■ AUTHOR INFORMATION Corresponding Author It is also important to address locational issues as they relate Nick Martin − Sostenipra Research Group, Institute of to the methods being presented. As the SR factors being used Environmental Science and Technology (ICTA-UAB), were specifically derived for EU supplies, they can strictly only Autonomous University of Barcelona, Barcelona, Catalunya be used for processes occurring within the EU. Naturally, local 08193, Spain; orcid.org/0000-0001-9023-9696; SR factors could be vastly different in certain countries, Email: nicholas.martin@uab.cat particularly in those that are dominant suppliers of particular materials or use different supply mixes. Conversely, the Authors calculations for EI and EJ are far more universal as they rely Cristina Madrid-López − Sostenipra Research Group, on global supply mix data or LCIA data where only a single Institute of Environmental Science and Technology (ICTA- global estimate is used. This highlights the fact that the SR UAB), Autonomous University of Barcelona, Barcelona, method intrinsically assumes that SR is the same whether Catalunya 08193, Spain materials are brought to the EU as raw and processed materials Gara Villalba-Méndez − Sostenipra Research Group, Institute or embedded within intermediate products; this is thought to of Environmental Science and Technology (ICTA-UAB), be an acceptable assumption in lieu of vastly more complex Autonomous University of Barcelona, Barcelona, Catalunya calculations. Likewise, owing to the complex array of 08193, Spain; Department of Chemical, Biological and components within most products and processes, it is assumed Environmental Engineering, Autonomous University of that using global data is suitable when assessing EI and EJ Barcelona, Barcelona, Catalunya 08193, Spain scores. Laura Talens-Peiró − Sostenipra Research Group, Institute of Nevertheless, higher levels of granularity in LCA data sets, Environmental Science and Technology (ICTA-UAB), particularly in relation to the locations in which sub-processes Autonomous University of Barcelona, Barcelona, Catalunya occur, would allow more complete assessments of intermediate 08193, Spain materials, components, and finished assemblies to be under- Complete contact information is available at: taken. In this regard, future studies could attempt more- https://pubs.acs.org/10.1021/acs.est.2c05308 detailed assessments involving sub-processes within overall processes. As many such sub-processes are likely to occur Author Contributions outside of the EU, SR factors would need to be calculated for N.M. and L.T-P. contributed the study conception and design; each material for different regional locations using a similar N.M. collected data; N.M. and L.T-P. performed data analysis approach to that used in the EU.16 For example, SR factors in and interpretation; N.M., L.T-P., C.M-L., and G.V-M. prepared China would be vastly different for materials they are currently the manuscript; and C.M-L. contributed in project admin- key suppliers of and a sub-process occurring in China would istration. All authors reviewed the results and approved the then need to use these inputs. The same is true for calculating final version of the manuscript. EJ scores in different territories, where specific supply mixes could theoretically be applied, and EI scores could use more Notes specific LCA processes for materials where regional data exist. The authors declare no competing financial interest. 17243 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article ■ ACKNOWLEDGMENTS The research is funded by the EU’s Horizon 2020 research and (21) Vlaskamp, M. C. The European Union and Natural Resources That Fund Armed Conflicts: Explaining the EU’s Policy Choice for Supply Chain Due-Diligence Requirements. Coop. Conflict 2019, 54, innovation program under the SENTINEL project (GA 407−425. 837089). (22) McLellan, B. C.Environmental Impacts of Mineral Sourcing ■ REFERENCES (1) Ripple, W. J.; Wolf, C.; Newsome, T. M.; Barnard, P.; Moomaw, and Their Impacts on Criticality. In The Material Basis of Energy Transitions; Bleicher, A.; Pehlken, A., Eds.; Academic Press: London, 2020; pp 109−120. (23) Fortier, M. O. P.; Teron, L.; Reames, T. G.; Munardy, D. T.; W. R. World Scientists’ Warning of a Climate Emergency. Bioscience Sullivan, B. M. Introduction to Evaluating Energy Justice across the 2019, 70, 8−12. Life Cycle: A Social Life Cycle Assessment Approach. Appl. Energy (2) United Nations. C.N.63.2016.TREATIES-XXVII.7.d Paris Agree- 2019, 236, 211−219. ment; New York, 2015. (24) Sovacool, B. K.; Martiskainen, M.; Hook, A.; Baker, L. (3) IRENA. World Energy Transitions Outlook: 1.5 °C Pathway; Decarbonization and Its Discontents: A Critical Energy Justice International Renewable Energy Agency: Abu Dhabi, 2021. Perspective on Four Low-Carbon Transitions. Clim. Change 2019, (4) Bobba, S.; Carrara, S.; Huisman, J.; Mathieux, F.; Pavel, C.Critical Raw Materials for Strategic Technologies and Sectors in the 155, 581−619. EU: A Foresight Study; Publications Office of the European Union: (25) European Commission. 3rd Raw Materials Scoreboard: Luxembourg, 2020. European Innovation Partnership on Raw Materials; Publications Office (5) Wellmer, F. W.; Buchholz, P.; Gutzmer, J.; Hagelüken, C.; of the European Union: Luxembourg, 2021. Herzig, P.; Littke, R.; Thauer, R. K.Raw Materials for Future Energy (26) Church, C.; Crawford, A.Conflict Minerals: The Fuels of Conflict Supply; Springer: Cham, 2019. in the Transition to a Low-Carbon Economy; International Institute for (6) Hund, K.; la Porta, D.; Fabregas, T. P.; Laing, T.; Drexhage, Sustainable Development: Winnipeg, 2018. J.Minerals for Climate Action: The Mineral Intensity of the Clean Energy (27) Lèbre, É .; Owen, J. R.; Corder, G. D.; Kemp, D.; Stringer, M.; Transition; Climate Smart Mining Initiative, World Bank Group: Valenta, R. K. Source Risks as Constraints to Future Metal Supply. Washington, DC, 2020. Environ. Sci. Technol. 2019, 53, 10571−10579. (7) The Material Basis of Energy Transitions; Bleicher, A.; Pehlken, A., (28) Lèbre, É .; Stringer, M.; Svobodova, K.; Owen, J. R.; Kemp, D.; Eds.; Academic Press: London, 2020. Cô te, C.; Arratia-Solar, A.; Valenta, R. K. The Social and (8) Nansai, K.; Nakajima, K.; Kagawa, S.; Kondo, Y.; Suh, S.; Environmental Complexities of Extracting Energy Transition Metals. Shigetomi, Y.; Oshita, Y. Global Flows of Critical Metals Necessary Nat. Commun. 2020, 11, No. 4823. for Low-Carbon Technologies: The Case of Neodymium, Cobalt, and (29) SENTINEL. Sustainable Energy Transitions Laboratory Platinum. Environ. Sci. Technol. 2014, 48, 1391−1400. (SENTINEL). https://sentinel.energy/ (accessed September 30, (9) Pattisson, P.; Firdaus, F.“Battery Arms Race”: How China Has 2022). Monopolised the Electric Vehicle IndustryThe Guardian, 2021. (30) Ecoinvent. Ecoinvent Version 3.8 (2021) Database. https://v38. (10) Searcey, D.; Forsythe, M.; Lipton, E.A Power Struggle over ecoquery.ecoinvent.org/Home/Index. Cobalt Rattles the Clean Energy RevolutionThe New York Times, (31) Huijbregts, M. A. J.; Steinmann, Z. J. N.; Elshout, P. M. F.; 2021. Stam, G.; Verones, F.; Vieira, M.; Zijp, M.; Hollander, A.; van Zelm, (11) Ambrose, J.Record Metals Boom May Threaten Transition to R. ReCiPe2016: A Harmonised Life Cycle Impact Assessment Green EnergyThe Observer, 2021. Method at Midpoint and Endpoint Level. Int. J. Life Cycle Assess. (12) Ewing, J.The World Wants Greenland’s Minerals, but 2017, 22, 138−147. Greenlanders Are WaryThe New York Times, 2021. (32) The World Bank. Worldwide Governance Indicators: World Bank (13) Glüsing, J.; Hage, S.; Jung, A.; Klawitter, N.; Schultz, S.Mining Data Catalog. http://info.worldbank.org/governance/wgi/ (accessed the Planet to Death: The Dirty Truth about Clean TechnologiesDer June 27, 2022). Spiegel, 2021. (33) Kaufmann, D.; Kraay, A.; Mastruzzi, M. The Worldwide (14) Hoang, A. T.; Nižetić, S.; Olcer, A. I.; Ong, H. C.; Chen, W. H.; Governance Indicators: Methodology and Analytical Issues. Hague J. Chong, C. T.; Thomas, S.; Bandh, S. A.; Nguyen, X. P. Impacts of Rule Law 2011, 3, 220−246. COVID-19 Pandemic on the Global Energy System and the Shift (34) ISO. ISO 14044:2006 Environmental Management - Life Cycle Progress to Renewable Energy: Opportunities, Challenges, and Policy Assessment - Requirements and Guidelines; International Organisation Implications. Energy Policy 2021, 154, No. 112322. for Standardisation: Geneva, 2006. (15) European Commission. Joint Research Centre (JRC), EC’s Raw (35) European Commission. Methodology for Establishing the EU List Materials Information System (RMIS) - RMIS Newsletter Special of Critical Raw Materials; Publications Office of the European Union: Edition: Focus on the Ukraine-Russia Crisis, 2022. Luxembourg, 2017. (16) European Commission. Study on the EU’s List of Critical Raw (36) Matsumoto, A.; Merlone, U.; Szidarovszky, F. Some Notes on Materials (2020) - Final Report; Publications Office of the European Applying the Herfindahl-Hirschman Index. Appl. Econ. Lett. 2012, 19, Union: Luxembourg, 2020. 181−184. (17) Calvo, G.; Valero, A. Strategic Mineral Resources: Availability (37) Eurostat. EU trade since 1988 by HS2,4,6 and CN8 [DS- and Future Estimations for the Renewable Energy Sector. Environ. 645593], Extra-EU28, IMPORT, QUANTITY_IN_100KG. https:// Dev. 2021, No. 100640. appsso.eurostat.ec.europa.eu/nui/show.do?dataset=DS- (18) Valero, A.; Valero, A.; Calvo, G.; Ortego, A.; Ascaso, S.; 645593&lang=en (accessed July 30, 2022). Palacios, J. L. Global Material Requirements for the Energy (38) Graedel, T. E.; Barr, R.; Chandler, C.; Chase, T.; Choi, J.; Transition. An Exergy Flow Analysis of Decarbonisation Pathways. Christoffersen, L.; Friedlander, E.; Henly, C.; Jun, C.; Nassar, N. T.; Energy 2018, 159, 1175−1184. Schechner, D.; Warren, S.; Yang, M.; Zhu, C. Methodology of Metal (19) Lee, J.; Bazilian, M.; Sovacool, B.; Hund, K.; Jowitt, S. M.; Criticality Determination. Environ. Sci. Technol. 2012, 46, 1063−1070. Nguyen, T. P.; Månberger, A.; Kah, M.; Greene, S.; Galeazzi, C.; (39) Martinez-Alier, J.The Environmentalism of the Poor: A Study of Awuah-Offei, K.; Moats, M.; Tilton, J.; Kukoda, S. Reviewing the Ecological Conflicts and Valuation; Edward Elgar: Cheltenham, 2002. Material and Metal Security of Low-Carbon Energy Transitions. (40) Marín, A.; Goya, D. Mining - The Dark Side of the Energy Renewable Sustainable Energy Rev. 2020, 124, No. 109789. Transition. Environ. Innovation Societal Transitions 2021, 41, 86−88. (20) European Commission. Raw Materials Scoreboard 2018: (41) Bainton, N.; Kemp, D.; Lèbre, E.; Owen, J. R.; Marston, G. The European Innovation Partnership on Raw Materials; Publications Office Energy-Extractives Nexus and the Just Transition. Sustainable Dev. of the European Union: Luxembourg, 2018. 2021, 29, 624−634. 17244 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245 Environmental Science & Technology pubs.acs.org/est Article (42) Ottinger, G. The Winds of Change: Environmental Justice in (64) Wernet, G.; Bauer, C.; Steubing, B.; Reinhard, J.; Moreno-Ruiz, Energy Transitions. Sci. Cult. 2013, 22, 222−229. E.; Weidema, B. The Ecoinvent Database Version 3 (Part I): (43) Sovacool, B. K.; Dworkin, M. H. Energy Justice: Conceptual Overview and Methodology. Int. J. Life Cycle Assess. 2016, 21, 1218− Insights and Practical Applications. Appl. Energy 2015, 142, 435−444. 1230. (44) Sovacool, B. K.; Burke, M.; Baker, L.; Kotikalapudi, C. K.; (65) Kupfer, T.; Baitz, M.; Colodel, C. M.; Kokborg, M.; Schöll, S.; Wlokas, H. New Frontiers and Conceptual Frameworks for Energy Rudolf, M.; Bos, U.; Bosch, F.; Gonzalez, M.; Schuller, O.; Hengstler, Justice. Energy Policy 2017, 105, 677−691. J.; Stoffregen, A.; Thylmann, D.GaBi Databases & Modeling Principles (45) Levenda, A. M.; Behrsin, I.; Disano, F. Renewable Energy for 2020; Sphera, 2020. Whom? A Global Systematic Review of the Environmental Justice Implications of Renewable Energy Technologies. Energy Res. Soc. Sci. 2021, 71, No. 101837. (46) Carley, S.; Konisky, D. M. The Justice and Equity Implications of the Clean Energy Transition. Nat. Energy 2020, 5, 569−577. (47) McCauley, D.; Heffron, R. Just Transition: Integrating Climate, Energy and Environmental Justice. Energy Policy 2018, 119, 1−7. (48) Akizu-Gardoki, O.; Wakiyama, T.; Wiedmann, T.; Bueno, G.; Arto, I.; Lenzen, M.; Lenzen, M.; Lopez-Guede, J. M. Hidden Energy Flow Indicator to Reflect the Outsourced Energy Requirements of Countries. J. Cleaner Prod. 2021, 278, No. 123827. (49) Dominish, E.; Teske, S.; Florin, N.Responsible Minerals Sourcing for Renewable Energy. Report Prepared for Earthworks by the Institute for Sustainable Futures; University of Technology Sydney: Sydney, 2019. (50) Martinez-Alier, J. Mapping Ecological Distribution Conflicts: The EJAtlas. Extr. Ind. Soc. 2021, 8, No. 100883. (51) European Commission. EU Reference Scenario 2020: Energy, Transport and GHG Emissions - Trends to 2050; Publications Office of the European Union: Luxembourg, 2021. (52) Rabe, W.; Kostka, G.; Stegen, K. S. China’s Supply of Critical Raw Materials: Risks for Europe’s Solar and Wind Industries? Energy Policy 2017, 101, 692−699. (53) Carrara, S.; Alves Dias, P.; Plazzotta, B.; Pavel, C.Raw Materials Demand for Wind and Solar PV Technologies in the Transition towards a Decarbonised Energy System, EUR 30095 EN; Publications Office of the European Union: Luxembourg, 2020. (54) Sprecher, B.; Daigo, I.; Murakami, S.; Kleijn, R.; Vos, M.; Kramer, G. J. Framework for Resilience in Material Supply Chains, with a Case Study from the 2010 Rare Earth Crisis. Environ. Sci. Technol. 2015, 49, 6740−6750. (55) Vakulchuk, R.; Overland, I.; Scholten, D. Renewable Energy and Geopolitics: A Review. Renewable Sustainable Energy Rev. 2020, 122, No. 109547. (56) Sattich, T.; Freeman, D.; Scholten, D.; Yan, S. Renewable Energy in EU-China Relations: Policy Interdependence and Its Geopolitical Implications. Energy Policy 2021, 156, No. 112456. (57) Gaustad, G.; Krystofik, M.; Bustamante, M.; Badami, K. Circular Economy Strategies for Mitigating Critical Material Supply Issues. Resour., Conserv. Recycl. 2018, 135, 24−33. (58) Babbitt, C. W.; Althaf, S.; Rios, F. C.; Bilec, M. M.; Graedel, T. E. The Role of Design in Circular Economy Solutions for Critical Materials. One Earth 2021, 4, 353−362. (59) IEA. World Energy Outlook 2021; International Energy Agency: Paris, 2021. (60) IRENA. Global Renewables Outlook: Energy Transformation 2050; International Renewable Energy Agency: Abu Dhabi, 2020. (61) Buchholz, P.; Brandenburg, T. Demand, Supply, and Price Trends for Mineral Raw Materials Relevant to the Renewable Energy Transition: Wind Energy, Solar Photovoltaic Energy, and Energy Storage. Chem. Ing. Tech. 2018, 90, 141−153. (62) Giurco, D.; Dominish, E.; Florin, N.; Watari, T.; McLellan, B.Requirements for Minerals and Metals for 100% Renewable Scenarios. In Achieving the Paris Climate Agreement goals: Global and regional 100% renewable energy scenarios with non-energy GHG pathways for +1.5 °C and +2°C; Teske, S., Ed.; Springer Open: Cham, 2019; pp 437−457. (63) Junne, T.; Simon, S.; Buchgeister, J.; Saiger, M.; Baumann, M.; Haase, M.; Wulf, C.; Naegler, T. Environmental Sustainability Assessment of Multi-Sectoral Energy Transformation Pathways: Methodological Approach and Case Study for Germany. Sustainability 2020, 12, No. 8225. 17245 https://doi.org/10.1021/acs.est.2c05308 Environ. Sci. Technol. 2022, 56, 17236−17245