- Email: [email protected]
The impact of technological innovation on critical materials risk dynamics
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Sustainable Materials and Technologies journal homepage: www.elsevier.com/locate/susmat
The impact of technological innovation on critical materials risk dynamics Anthony Y. Ku, Johnathan Loudis, Steven J. Duclos
⁎
GE Global Research, 1 Research Circle, Niskayuna, NY 12309, United States
A R T I C L E I N F O
A B S T R A C T
Keywords: Criticality Helium Materials Rare earth elements Recycling Rhenium Substitution Supply chain Sustainability
As the technologies we use as a society have advanced, so have the materials used in these technologies. Some of these materials are exotic and highly specialized, making them particularly vulnerable to supply disruptions and supply disruptions particularly impactful. Such materials are designated as “critical” materials. Their level of criticality can be identified by accounting for a number of factors related to their supply risk and the extent to which a supply disruption would impact business operations or society at large. We highlight current methodologies used to assess materials criticality, how these assessments are used to reduce materials-related risk and to what extent there is room for improvement. Particularly, this paper reviews critical materials designations from the United States Department of Energy, the European Union, and the General Electric Company, and how they have changed over the period from 2008 to 2014. The changes suggest that the factors considered in criticality ratings have different natural time scales, and that criticality changes occur both due to supply-side risk mitigation as well as demand-side responses. Response options, whether on the supply or demand side, also span a range of time scales and the interaction between factors with different time scales can play a significant role in the dynamics. To date, many published analyses are snapshots in time. A detailed understanding of how risk profiles evolve remains an open question. The importance and impact of demand-side responses such as recycling, substitution and new technological development are discussed.
1. Introduction Modern society uses a wide range of raw materials and these materials go through cycles of surplus and shortage [1]. Efforts to anticipate materials shortages were renewed in 2008 with the National Resource Council (NRC) study, which introduced the concept of materials criticality [2]. Critical materials are designated as such because they are vulnerable to supply disruptions and such disruptions would have significant adverse impacts for businesses and society at large. The criticality of a given material or element is often considered along two dimensions, namely the level of supply risk and the impact a supply disruption would have. Criticality analyses use different factors that measure the exposure of a given material to each of these dimensions, and may also consider risks associated with a third dimension, environmental factors [3,4]. Factors are aggregated to create a score for each material along each dimension [5,6]. Virtually all published assessments are snapshots in time. Assessments can be a valuable tool in planning policy both at the industry and government levels [7,8]. They can also be used to identify and mitigate supply chain risks [9]. For example, the U. S. Department of Energy (DOE) established a Critical Materials Institute (CMI) to coordinate and provide strategic focus to
⁎
efforts to address critical materials relevant to the US energy infrastructure. In order to use these tools to inform policy, it is important to understand the drivers of materials criticality as well as how criticality changes over time. Generally, the criticality of a given material can change due to market responses, geopolitical factors, and technology development. Retrospective studies of market trends and geopolitics can provide insight into historical criticality trends [5,10]. Minerals and metal markets are difficult to accurately forecast for many of the same reasons as energy markets. To-date efforts can be categorized into two general approaches. Top-down approaches posit different scenarios for economic and political landscapes and then work out the materials market implications [6,11]. The logical consequences of particular narratives related to economic growth, government policy, or sociopolitical situations can be explored using sensitivity analyses and statistical tools. However, top-down approaches have difficulty capturing technological innovations that introduce non-linear changes in materials use patterns [12,13]. Bottom-up approaches attempt to reconstruct dynamics from different links in the supply chain. Agent-based dynamic material flow models can capture interactions by segmenting the links in the supply chain [10,14,15]. As with any modeling, both top-down
Corresponding author. E-mail address: [email protected] (S.J. Duclos).
https://doi.org/10.1016/j.susmat.2017.11.002 Received 1 May 2017; Received in revised form 29 September 2017; Accepted 18 November 2017 2214-9937/ © 2017 Published by Elsevier B.V.
Please cite this article as: Ku, A.Y., Sustainable Materials and Technologies (2017), https://doi.org/10.1016/j.susmat.2017.11.002
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Table 1 Comparison of critical materials weights used in computing composite supply risk score. Supply risk scoring
Yale [3]
DOE [6.7]
GE
BGS [22]
EU [8,24]
Materials (year of publication)
62 (2015)
14 (2010) 16 (2011)
33 (2008) 53 (2012)
52 (2011) 41 (2012)
41 (2010) 54 (2014)
Physical availability Reserves/Depletion time Companion production Recycling rate Production Producer concentration Producer stability Producer governance Producer policy Market factors Price volatility Substitutability Competing demand
33% 1/6 1/6
50% 2/5 1/10
22% 1/9 1/9
29% 1/7
67% 1/6 1/6 1/6 1/6
40% 1/5
33%
1/5
1/6 1/6 45% 1/9 1/6 1/6
10%
1/10
1/7 56% 2/7
Included Included
2/7 14% 1/7
Included
EU and GE differ and this will impact the ratings. Materials that GE considers critical to GE due to exposure from its manufacturing supply chain for a specific product line related to medical imaging may not be as critical to the DOE as it considers clean energy materials requirements a decade in the future. Exposure to the coking coal supply challenge may be less of a concern to both GE and the DOE, but for slightly different reasons. For instance, it is less related to the DOE's medium- to long-term energy goals, and it is not directly related to GE's most pressing supply concerns. This is an important point when considering dynamics because individual technological developments and their implementation would be expected to have a both larger and faster impact on criticality at the company level. Criticality assessments take into account two key dimensions associated with materials risks. The first dimension is supply risk, or how susceptible a material is to supply disruption. The other is the impact caused by a shortage. Typical assessments focus on specific materials of interest and assign a risk score to each material in each dimension. Scoring is sensitive to the interests of the organization performing the assessment: producers and users may rate supply risk differently, as might companies active in different industries. Materials are then plotted along the two axes and materials that score above a threshold are designated critical. Some analyses, such as the one performed by the Yale group also include an environmental impact axis in their assessments [3]. Scores for supply risk and impact of disruption are generally composite indices that take into account several factors. In considering dynamics, an appreciation of the different factors that contribute to criticality scoring allows one to consider how they change with time and to consider interaction effects. Tables 1 and 2 summarize how five organizations compute the supply risk and impact of disruption. Despite some differences in terminology and relative weighting of factors, most assessments focus on the same contributing factors. Four of the five methodologies compute risk score as the weighted sum of various factor scores, while the EU approach uses multiplicative formulae. Factors that contribute to supply risk include physical availability, production, and market factors. Physical availability is tied to global reserves, co-production, and recycling rates. Co-production refers to the fact that many elements are produced as by-products of a primary ore body [3,21]. Production factors include geographic concentration, geopolitical stability, and policy. Refining and distribution bottlenecks are also captured in production scores. This factor is especially important when considering materials that may have only a few capable refiners or suppliers. A third category includes market factors such as price volatility, the availability of substitutes, and competing demand between different end use industries. The competition between different industries is particularly
and bottom-up approaches rely on the quality of the assumptions used in the modeling. It is essential to recognize that changes in criticality are induced by both exogenous shocks to a supply chain or usage pattern, as well as endogenous market responses to such shocks. Technological innovation can introduce inflection points that can shift materials use patterns, which may increase or decrease the criticality level of a particular material or materials. For example, the maturation of nickel-base superalloys, coupled with a shortage of cobalt in the 1950′s lead to a shift towards the former [16]. Beginning in the late 1980s, a transition from halophosphate phosphors to rare earth-baed triphosphor blends gradually boosted demand for certain rare earth elements (REE) [17]. The microelectronics industry routinely introduces new materials-enabled devices. The implication is that increased usage of such exotic materials comes at the cost of increased criticality. There is also a bright side to technological development: advances in technology are not limited to initiating material uses, rather can also lead to decreased usage of critical materials. Progress in production, manufacturing and recycling methods can drastically reduce the demand for various materials [18,19]. There can also be interactions between technological innovation and supply. For example, a second shortage of cobalt in the 1970's was a contributing factor to the development and adoption of NdFeB magnets as a substitute for SmCo magnets [20]. One challenge that we will face with these criticality assessment tools is whether they can accurate anticipate crises early enough to allow meaningful action. This paper considers the dynamics of materials criticality with a special focus on the role and consequences of technological innovation. The discussion is organized into four parts. The first section introduces definitions of criticality and how criticality scores are assigned. The second part reviews assessments by the DOE, the European Union (EU), and the General Electric Company (GE) and how they changed over a period of one to four years. The third part considers how risk evolves. The final section makes some observations on how technological innovation responds to materials criticality, with an emphasis on the nature of substitutes and time tables around their development. Examples will be drawn from GE's experiences in responding to its materials criticality challenges and the recent literature on REEs, given the extensive public discussion of these materials. Technology-driven inflection points in the usage of various materials are briefly discussed. The paper concludes with some implications for the materials development community. 2. Definitions of criticality A first comment on criticality assessments is that they are dependent upon the scope of the review. The time scales and interests of the DOE, 2
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Table 2 Comparison of critical materials weights used in computing composite impact score Impact scoring: scope of study
Yale: global importance
DOE: importance to clean energy
GE: impact to operations
Importance Economic impact Usage % by population % of world's supply used Substitutability Performance Availability Environmental impact Market factors Cost pass through
50%
75%
50% 1/4
EU: economic importance 100%
1/2 50% 1/6 1/6 1/6
1/4 25% 1/4
25%
Included in supply risk
25% 1/4
3. Historical dynamics in criticality risk ratings
relevant when considering the effects of technological development on supply risk. The particular weightings reflect the concerns of the different interested parties. For example, GE's analysis emphasizes market factors, while the EU or DOE are more concerned with physical production related to national and international interests. The EU supply risk score (SR) includes factors from each of the three categories, and is computed by:
Assessments are conducted as snapshots in time, but changes in criticality scores between snapshots provide insight into the underlying risk dynamics. The three groups highlighted above – the DOE, EU, and GE – have published detailed updates over periods of up to four years. This section reviews changes in these assessments as a practical starting point for a discussion on dynamics. Others, such as the Yale group, are in the process of updating their original assessments. The United States Department of Defense (DOD) maintains a stockpile of materials for defense and issues annual reports to Congress on stockpiling needs [23]. However, these reports only summarize the final recommendations without detailed discussion of the methodology, uses, or dynamics. This makes the results interesting, but less useful for probing dynamics. We see a few common themes between these analyses. First, some new materials enter updated analyses and other leave depending on changing market conditions. Second, specific scores typically shift slightly, however, changes in the impact scores are generally more significant than those in the supply risk scores. This could be due to both technical and methodological reasons. Technical drivers include the fact that impact scores are primarily functions of technological development and market dynamics. Both of these factors can shift relatively more quickly than the primary determinants of supply risk (such as economic reserves, which change only infrequently with significant new ore deposit discoveries). We also note that materials with high composite criticality scores also attract attention aimed at mitigating such risk, which can amplify the rates of change.
SR = S (1 − R) H where S is the substitutability (0 to 1, with 1 being least substitutable), R is the recycling rate (percentage of end-of-life supply) and HHI is the Herfindahl-Hirschman index for market concentration defined as H = ∑(σi2 ∙ WGIi) where σi and WGIi are the share of production and world governance indicator for country i [8]. Table 2 shows the factors for impact scoring from the five methodologies. They are grouped into three categories related to the importance of the material, the availability of substitutes, and other market factors. The EU economic importance score (EI) is given by:
EI =
BGS: no rating assigned
∑ (si vi ) GDP
where si and vi are the share of use and value contributed to each megasector, and GDP the overall gross domestic product of the EU. Depending on the methodology, substitutability appears as a factor in either supply risk or impact. The British Geological Survey (BGS) and EU consider it under supply risk, Yale and the DOE treat it as an impact factor, and GE includes it in both. From a supply risk perspective, substitution is a market mechanism that can shift demand. However, individual users utilize substitution to reduce risk for high impact materials. Since it is application-specific, substitutability has been difficult to quantify and to forecast.
3.1. U.S. Department of Energy assessments The DOE published an assessment in 2010 on materials important to the deployment of clean energy technologies within the US. A revision
Fig. 1. Comparison of short-term and medium-term DOE clean energy materials criticality assessments from 2010 and 2011 [6].
3
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
switched relative to the DOE assessment. Supply risk is on the y-axis and the Economic importance is on the x-axis. Critical materials from the 2014 assessment are in black (materials in both versions) and red (new materials added in 2014). Material designations from 2010 for selected materials are shown in blue. There were three types of changes. First, the supply risk changed due to production concentration increasing or decreasing. This can be seen in the vertical shifts in the platinum group metals (decreasing) and magnesite (increasing). Second, changes in the economic importance resulted in horizontal shifts, as seen in niobium (decreasing) and tungsten (increasing). Finally, there were four changes to the criticality list. Tantalum dropped off due to reduced supply risk concerns while borates, chromium, and magnesite were promoted due to increasing concerns around supply concentration. The EU supply risk emphasizes concentration risk, which was the primary driver of change over the three-year span between assessments. Similarly, growth or contraction in economic sectors is the primary driver of changes in the EI index. Technological deployment was not explicitly accounted for. Efforts to promote recycling and support the development of substitutes were supported, but three years is a short time to see significant impacts from new efforts. A final technical factor, relevant to the other analyses as well, is improving data quality. Since criticality assessments are still in their early stages, some of the changes in criticality score may also be due to improved data quality. The EU plans to refresh its analysis every 5 years, and initiated efforts in 2015 to improve the fidelity and rigor of their assessment methodology [8].
Fig. 2. Comparison of critical materials designations by the European Union from 2010 to 2014. The region shown corresponds to the upper right-hand (critical) portion of the EU criticality diagram; materials in the shaded region are considered critical. Blue dots correspond to materials from the 2010 analysis. Black (original) and red (new) dots correspond to materials in the 2014 study. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Adapted from references [8] and [24].
was published in 2011. The revised version expanded the list of elements from 14 to 16, while adjusting the results for changes in markets and technologies. The main changes were minor shifts in the position of several REEs, indium, and gallium. Indium was demoted from the near-term criticality watch list. The primary drivers were the improved competitiveness of silicon-based solar cells relative to thin-film solar technologies, an increase in the projected adoption rate for light emitting diodes (LED) for lighting, and the announcement of several rare earth mining projects leading to expectations of increased supply [7]. The trends are shown graphically in Fig. 1.
3.3. General Electric Company assessment GE performed a criticality assessment of its operations in 2008, and refreshed the analysis in 2012. The original assessment considered 33 elements, and the refresh expanded the scope to 53 elements. Elements were scored according to Supply risk and Impact of a disruption on company operations. Only raw materials directly purchased by GE were accounted for in the analysis. Materials used as manufacturing consumables (e.g., tungsten carbide tooling or inert gases for heat treatments) or embedded in product components purchased from original equipment manufacturers (OEMs) were not included. Fig. 3 shows company-level criticality diagrams from 2008 and 2012. Supply risk is plotted on the x-axis and Impact is plotted on the yaxis, consistent with the DOE plots. Each circle represents an individual element with the area of the circle being the raw material spend for that element. Most of the circles are unlabeled due to the proprietary nature of information pertaining to company operations. Rhenium, helium, and REEs are identified for illustrative purposes. At a glance, it can be seen that the supply risk of some of the rare earth elements relative to Re has increased. Similarly, the effects of mitigation efforts can be seen in the smaller spend on Re and the wider distribution of impacts among the rare earth family of elements. A closer look at changes in their criticality scoring over the intervening three years is informative. Rhenium added nickel-based superalloys improves the high temperature mechanical properties for turbine applications [25]. A price
3.2. European Union assessments The EU performed its first critical materials assessment in 2010 [8], focused on materials of importance to the European economy. Fortyone non-fuel, non-agriculture raw materials were scored on economic importance and supply risk, according to the sub-factors listed in Table 1. Fourteen materials were designated as critical materials including: antimony, beryllium, cobalt, fluorspar, gallium, germanium, graphite, indium, magnesium, niobium, platinum group metals (PGM), REE, tantalum, and tungsten. An update to the analysis was published in 2014 [24]. In the second assessment, the scope was expanded to include fifty-four materials, of which twenty were designated critical. Fig. 2 highlights significant changes in the “critical” region between the two assessments. As such the figure only illustrates the upper righthand (critical) portion of the EU criticality diagram. The axes are
Fig. 3. GE assessments – criticality diagrams from 2012 with helium, rhenium and REEs overlaid. Additional details beyond the general labels for rare earth elements, and those for He and Re, are not included due to their proprietary nature.
4
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Fig. 4. GE Lighting and Healthcare criticality diagrams from 2012.
development of the supply chains for these new applications. There were also some examples of substitution at a system level. One notable example was an announcement in 2011 that GE's 1.5 and 2.5 MW wind turbine product lines would use a doubly-fed induction generator (DFIG) architecture in lieu of direct drive generators that would use a significant amount of REE-based permanent magnets. Awareness of the supply challenges for rare earths continues to be high across the company.
spike occurred in 2008, due to a confluence of factors including limited supply, market concentration, and financial speculation. Given the importance of superalloys to the operations of GE's Aviation and Energy divisions, rhenium was rated one of the most critical elements to GE in 2008. Although it was among the most critical elements in 2012, the relative criticality of rhenium declined due to a combination of market responses and internal initiatives to reduce total demand and spend. Some of these internal efforts included the qualification of alloys with reduced rhenium content, improved utilization in manufacturing operations, and reclamation from end-of-life components. Improvements in manufacturing efficiency and adoption of relatively mature alternate alloys helped in the short-term. Longer term options, such as the development of superalloy compositions that require less rhenium and new materials such as ceramic matrix composites, also remain in play. Helium was added to the 2012 assessment. GE accounted for about 4% of global consumption in 2011, primarily for cooling of superconducting magnets used in magnetic resonance imaging (MRI) machines. Helium had a moderate-high supply risk score, and a high impact score. Supply risk resulted primarily from its co-production as byproduct of natural gas liquefaction and the high concentration of supply from a single source (about a third of global production was associated with the United States National Helium Reserve). The high potential impact of a disruption on GE operations was due to the lack of alternatives to helium in MRI magnet cooling and some manufacturing operations, and GE's non-trivial share of world consumption. Concerns over helium proved to be prescient as U.S. government policy debate over the future of the National Helium Reserve resulted in supply constraints in 2013. The impact of government action to extend the life of the Reserve, and internal efforts to reduce exposure are expected to show up in the next criticality assessment. GE's analysis assigned criticality ratings to each individual rare earth element. Although specific elements are not identified in Fig. 3, some insight can be gained by looking at the spreads in aggregate. The range in supply risk reflects the difference between light REEs, which are naturally found in greater abundance in most deposits, and heavy REEs. The spread in impact captures differences in use and substitutability across different product lines. Fig. 4 shows criticality diagrams for GE's Lighting and Healthcare divisions, which have heighted exposure to rare earths due to their heavy use of europium, terbium, and yttrium in phosphors for fluorescent lighting and scintillators for medical imaging. Helium appears in the GE Healthcare diagram as well. Over the period from 2008 to 2012, GE Lighting re-optimized its phosphor blends to improve manufacturing yield, evaluated the benefits of phosphor recycling, and engaged in R&D to develop alternate phosphor compositions with reduced rare earth content. These activities were complementary to efforts to improve the robustness of its supply chains, as a matter of course for business. In parallel, GE Healthcare took steps to improve the recycling of manufacturing scrap from scintillator production. Efforts were also made to identify and introduce substitutes. In one case, GE Aviation found that some heavy REEs have the potential to function interchangeably in some of their next-generation coatings [26]. This provided some flexibility in the
3.4. General trends and observations Despite different mathematical forms for computing supply risk and impact, the general trends across the DOE, EU and GE assessments are consistent. They agree on the continued criticality of REEs, particularly the heavy REEs, and that the biggest near-term drivers of change were economic cycles and the commercialization of technologies already under development. Since criticality analysis is still a relatively young field, the inclusion of new materials to the assessments can expand watch lists regardless of the underlying dynamics. Examples of this are coking coal, phosphate and silicon metal for the EU, and helium for GE. Major shifts in criticality appear to require more time, ostensibly due to the longer time frames associated with opening a new mine, or the development, validation and commercialization of entirely new technologies. The question of relevant time scales and their impact on risk dynamics is the focus of the remainder of this paper. 4. Criticality risk dynamics In practice, criticality assessments are used to screen a wide range of materials and identify a high risk short list for follow-up. The value of incorporating dynamics, even crudely, is that it offers the possibility of identifying materials with high potential to emerge as critical. A wide range of specific factors can drive dynamics and determine the rates of change. For example, geological availability is expected to change slower than price volatility. Technological innovation varies among end use sectors. An example of this concerns elements used in both microelectronics and superalloys such as tantalum, cobalt and some REEs. Consumer electronics product cycles are short compared to the time needed to develop and qualify superalloys for aviation applications. As a result, new applications in microelectronics could create unanticipated challenges for the aerospace industry because of the longer time horizons needed to develop alternatives. Since a supply chain can be disrupted by a bottleneck in any single link, the distribution of the factors that go into the supply risk and impact provides insight into root causes and possible risk mitigation strategies. An ever-present concern with mismatches in innovation timeframes is that, to a supplier, ultimately only the sales price matters. This could lead to significant amounts of a raw material going to a single application (or set of applications) having the highest marginal value causing a supply deficit for other applications. This innovation-timing mismatch can be a serious risk, particularly when industries with significantly different development cycles compete for the same materials. Future studies can 5
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
conservation measures triggered by near closure of the National Helium Reserve may reduce demand in a way that could offset lost production.
quantify these effects more rigorously for specific scenarios by computing response curves using publicly available assessment tools. Table 3 shows different factors and estimated ranges for the time scales for significant changes. Public information can be used to validate a number of these factors including market data (including pricing and product introduction cycle times in specific industry) and government statistics (such as political turnover or USG data on production and reserves). Since some indices used to estimate reserves and several production factors are updated annually, the possibility of data lags must be considered. Factors related to physical availability generally change over the course of years to decades, while production factors appear to change over years. Narrower ranges for specific elements can be estimated from mining and market data. Price volatility for commodity materials can be computed from market data; this may be more difficult for materials with smaller or more opaque markets. Time scales for substitution and changes across multiple industries can change are more industry-specific, with innovations requiring significant invention taking several decades to make it to market. These will occur against a backdrop of longer term trends related to physical availability and technological progress, and near-term perturbations that test the resilience of existing supply chains. The interplay between different supply risk factors can produce inflections in market equilibrium demand. One example of this is the introduction of LEDs into the lighting market. Driven by technology development and government policy towards energy efficiency, this transition is already underway and expected to continue through the end of the decade [6,30]. Due to the lower REE use per bulb of LEDs relative to fluorescent lighting, the adoption of LEDs is expected to reduce demand pressure. This transition is occurring at a time of heightened government awareness of rare earth criticality, and can work with efforts to identify material substitutes to reinforce a downward trend in the supply risk of certain REEs. A second example illustrates that opposing trends may also act to offset each other. Helium is primarily a by-product of certain natural gas fields and the increasing share of natural gas production from shale gas fields which are low in helium could have an adverse supply impact. At the same time,
5. Response dynamics and substitution The response to each materials crisis will depend on the exact nature of the disruption and use. However, response options can be grouped into five categories, shown in Fig. 5. Sourcing approaches focus on commercial arrangements. Manufacturing efficiency and recycling approaches aim to use available material more effectively in an effort to reduce net demand for raw material. Substitution, which typically requires technology development, seeks to replace or reduce the use of the material. Responses require different amounts of time to have an impact. The planning, permitting, construction, and start-up of a new mine can take over a decade. The recovery and reuse of manufacturing scrap can be implemented more quickly. This is particularly true in cases where the methods already exist and only require a price increase in raw material to make them economical. Substitution can occur in two ways. The first is a direct material replacement. The existence of adequate, but more costly, alternatives could hasten material substitution, but the resulting products may still require tooling or other upgrades at manufacturing facilities or qualification in regulated industries. The invention of a new material could add years or decades to this process. A second substitution option is a different system design that replaces the function of the critical material. Examples of this, discussed earlier, include the transition from fluorescent lighting to LEDs, and GE's decision to move from rare earth permanent magnet direct drive generators to induction architectures in wind turbines. While these actions reduce the risk associated with REEs, they simultaneously increase the risk for materials used in the substitutes. This feedback loop is one of the reasons that criticality diagrams needs to be periodically updated and the dynamics better understood. Additionally, as designers increasingly use more exotic materials it will be essential to incorporate a culture of designfor-sustainability so as to mitigate potential criticality issues during the development stages of new products and technologies.
Table 3 Estimated time scales for criticality risk factors. Supply risk scoring
Time scales
Examples
Physical availability Reserves/depletion time Companion production Recycling rate
10 to 30 years 5 to 20 years 1 to 10 years
Discovery of new reserves [27] Development of improved separation technologies Government policy; economics [18]
Production Producer concentration Producer stability Producer governance Producer policy
1 1 1 1
Structural changes in geographical distribution of production Infrastructure investment [28] Changes in political stability of producer nations Mining or trade policy [28]; Environmental regulations
Market factors Price volatility Substitutability Competing demand
1 to 3 years Up to 20 years Up to 20 years
Shifts in supply/demand balance; short-term disruptions in supply Technology innovation [28,29] Technology innovation [28,29]
Impact risk scoring
Time scales
Examples
Importance Economic impact Usage % by population % of world's supply used
1 to 10 years 5 to 10 years 1 to 10 years
Business cycles; introduction of new technology Structural economic shifts [28]; changes in technology [29] Changes in business strategy or models
Substitutability Performance Availability Environmental impact
Up to 20 years Up to 20 years Up to 20 years
Introduction of new technology Introduction of new technology Introduction of new technology
Market factors Cost pass through
< 3 years
Company responses to markets
to to to to
10 years 5 years 5 years 5 years
6
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Fig. 5. Responses to material supply disruption.
discussed herein offer the practitioner more tools to anticipate and mitigate these risks.
Shortages (and sometimes expectations of shortages) trigger multiple, simultaneous responses. Interactions between these overlapping responses can be an important driver of dynamics. Governments, producers and users of a given material can also coordinate their responses. However, cooperation is not guaranteed and disconnects between competing responses may make them less effective than they might otherwise be. Returning to the REE example, efforts to develop new mines outside of China have the potential to both increase and diversify supply. However, in the time it takes to open a mine, users of rare earths are pursuing substitutes to reduce their demand. Similar economic dynamics apply to recycling processes. High prices incentivize efforts to establish recycling programs, but easing of demand can reduce their economic feasibility. Government activity and public awareness have increased. Research in material substitutes, refining processes, manufacturing improvements, and recycling technologies is under way. Together, these responses have resulted in differentiation between rare earths. The EU, in its latest assessment, split their original rare earth designation into a higher risk group of heavy rare elements, and a lower risk group of light REEs. Efforts to understand and model the dynamics of criticality are still at an early stage. In particular, more detailed frameworks are needed to understand how fast an industry can adapt to perturbations in its material supply chains, how links in the material supply chain interact with each other, and the possible convolution of dynamic effects between multiple risk factors in criticality scoring. These efforts can build on work from dynamic materials flow analysis, which is already wrestling with the question of forecasts [31]. In particular, some of the statistical methods for modeling material flows may be appropriated for addressing technological innovation [32]. New metrics for how technology transitions from lab to commercial production may provide insight into how fast industry can adopt a new material.
Acknowledgments The authors would like to acknowledge helpful discussions with A Chowdhury, J Herzog, S Hung, F Johnson, D Lipkin, J Luo, W McLinko, L Quirk, and C Thom at GE, and T Graedel from Yale. References [1] T.E. Graedel, E.M. Harper, N.T. Nassar, B.K. Reck, On the materials basis for modern society, Proc. Natl. Acad. Sci. 112 (2013) 6295–6300. [2] National Research Council, Minerals, Critical Minerals, and the U.S economy: Committee on Critical Mineral Impacts of the U.S. Economy, The National Academies Press, 2008. [3] T.E. Graedel, E.M. Harper, N.T. Nassar, P. Nuss, B.K. Reck, Criticality of metals and metalloids, Proc. Natl. Acad. Sci. 112 (2015) 4257–4262. Available electronically at www.pnas.org/cgi/doi/10.1073/pnas.1500415112 (Accessed: May 5, 2015). [4] N.T. Nassar, R. Barr, M. Browning, Z. Diao, E. Friedlander, E.M. Harper, C. Henly, G. Kavlak, S. Kwatra, C. Jun, S. Warren, M. Yang, T.E. Graedel, Methodology of metal criticality determination: criticality of the geological copper family, Environ. Sci. Technol. 46 (2012) 1071–1078. [5] S. Gloser, L. Tercero Espinoza, C. Gandenberger, M. Faulstich, Raw material criticality in the context of classical risk assessment, Res. Policy 44 (2015) 35–46. [6] U.S. Department of Energy. Critical Materials Strategy: 2011 Technical Report, Washington DC. [7] U.S. Department of Energy. Critical Materials Strategy: 2010 Technical Report, Washington DC. [8] European Commission, Critical raw materials for the EU: 2010. Technical Report, Brussels, Belgium, Available electronically at http://ec.europa.eu/enterprise/ policies/raw-materials/critical/index_en.htm , Accessed date: 2 January 2015. [9] S.J. Duclos, J.P. Otto, D.G. Konitzer, Design in an era of constrained resources, Mech. Eng. 132 (2008) 36–40. [10] E. Muller, L.M. Hilty, R. Widmer, M. Schleup, M. Faulstich, Modeling metal stocks and flows: a review of dynamic material flow analysis methods, Environ. Sci. Technol. 48 (2014) 2102–2113. [11] E. Alonso, A.M. Sherman, T.J. Wallington, M.P. Everson, F.R. Field, R. Roth, R.E. Kirchain, Evaluating rare earth element availability: a case with revolutionary demand from clean technologies, Environ. Sci. Technol. 46 (2012) 3406–3414. [12] B. Buijs, H. Sievers, L.A. Tercero Espinoza, Limits to the critical raw materials approach, Proc. Inst. Civ. Eng. Waste Resour. Manag. 165 (2012) 201–208. [13] K. Oelich, D.A. Dawson, P. Purnell, C. Knoeri, R. Revell, J. Busch, J.K. Steinberger, Assessing the dynamic material criticality of infrastructure transitions: a case of low carbon electricity, Appl. Energy 123 (2014) 378–386. Available online at https:// doi.org/10.1016/j.apenergy.2014.01.052. [14] G. Gaustad, E. Olivetti, R.E. Kirchain, Toward sustainable material usage: evaluating the importance of market motivated agency in modeling material flows, Environ. Sci. Technol. 45 (2011) 4110–4117. [15] C. Knoeri, P.A. Wager, A. Stamp, H.-J. Althaus, M. Weif, Towards a dynamic assessment of raw materials criticality: linking agent-based demand with material flow supply modelling approaches, Sci. Total Environ. 461–462 (2013) 808–812. [16] C.T. Sims, A history of superalloy metallurgy for superalloy metallurgists, Superalloys (1984) 399–419. Available electronically at http://www.tms.org/ Superalloys/10.7449/1984/Superalloys_1984_399_419.pdf (Accessed: May 5, 2015).
6. Conclusions The materials use profile of modern society will continue to evolve. At any given time, criticality assessments can help identify important materials that are also vulnerable to supply disruptions. The dynamics play out over multiple time scales. An understanding of which materials are critical, the drivers and dynamics behind their criticality can inform efforts to develop new materials and technologies. Simply put, material shortages can motivate technology development and technology development can cause material shortages. In this context, recycling, substitution and technological development are all important avenues for innovation related to materials supply risk. Rather than simply rely on myopic price signals, materials criticality analyses like those
7
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
[17] A.M. Srivastava, T.J. Sommerer, Fluorescent lamp phosphors, InterfaceElectrochem. Soc. 7 (1998) 28–33. Available electronically at http://www. electrochem.org/dl/interface/sum/sum98/IF6-98-Page28-31.pdf , Accessed date: 5 May 2015. [18] Y. Houari, et al., A system dynamics model of tellurium availability for CdTe PV, Prog. Photovolt. Res. Appl. 22 (129) (2014). [19] T.E. Graedel, J. Allwood, J.P. Birat, M. Buchert, C. Hageluken, B.K. Reck, S.F. Sibley, G. Sonnemann, What do we know about metal recycling rates? (2011) USGS Staff - published research. Paper 596, Available electronically at http:// digitalcommon.unl.edu/usgsstaffpub/596 , Accessed date: 5 May 2015. [20] E. Alonso, J. Gregory, F. Field, R. Kirchain, Material availability and the supply chain: risks, effects, and responses, Environ. Sci. Technol. 41 (19) (2007) 6649–6656. [21] A.Y. Ku, S. Hung, Manage materials risk, Chem. Eng. Prog. (2014) 28–35. Available electronically at http://www.aiche.org/sites/default/files/cep/20140928.pdf , Accessed date: 3 May 2015. [22] British Geological Survey - Risk List, Available electronically at http://www.bgs.ac. uk/mineralsuk/statistics/risklist.html, (2012) , Accessed date: 8 April 2015. [23] Defense Logistics Agency, Strategic and critical materials 2015 report on stockpile requirements. Jan 2015, Available electronically at www.strategicmaterials.dla.mi/ Report%20Library/2015%20NDS%20Requirements%20Report.pdf , Accessed date: 5 May 2015. [24] European Commission, Report on critical raw materials for the EU. Technical
[25] [26]
[27] [28] [29]
[30]
[31] [32]
8
report, Brussels, Belgium, Available electronically at http://ec.europa.eu/ enterprise/policies/raw-materials/files/docs/crm-report-on-critical-raw-materials_ en.pdf , Accessed date: 2 January 2015. P. Caron, T. Khan, Evolution of Ni-based superalloys for single crystal gas turbine blade applications, Aerosp. Sci. Technol. 3 (1999) 513–523. A.Y. Ku, C. Dosch, T. Grossman, J. Herzog, A. Maricocchi, D. Polli, D.M. Lipkin, Addressing rare-earth element criticality: an example from the aviation industry, JOM 66 (11) (2014) 2355. P. Darling (Ed.), SME Mining Engineering Handbook, 3rd ed, Society for Mining, Metallurgy and Exploration, Inc, 2011. K.S. Gallagher, J.P. Holdren, A.D. Sagar, Energy-technology innovation, Annu. Rev. Environ. Resour. 31 (2006) 193–237. J. Shea, What do technology shocks do? in: B.S. Bernanke, J. Rotemberg (Eds.), NBER Macroeconomics Annual 1998, vol. 13, MIT Press, 1999Available electronically at http://www.nber.org/chapters/c11249 , Accessed date: 7 May 2015. Anthony Y. Ku, Anant A. Setlur, Johnathan Loudis, Impact of Light Emitting Diode Adoption on Rare Earth Element Use in Lighting: Implications for Yttrium, Europium, and Terbium Demand, The Electrochemical Society Interface24.4 (2015) 45–49. M.T. Melo, Statistical analysis of metal scrap generation: the case of aluminium in Germany, Resour. Conserv. Recycl. 26 (1999) 91–113. H. Rechberger, T.E. Graedel, The contemporary European copper cycle: statistical entropy analysis, Ecol. Econ. 42 (2002) 59–72.
Contents lists available at ScienceDirect
Sustainable Materials and Technologies journal homepage: www.elsevier.com/locate/susmat
The impact of technological innovation on critical materials risk dynamics Anthony Y. Ku, Johnathan Loudis, Steven J. Duclos
⁎
GE Global Research, 1 Research Circle, Niskayuna, NY 12309, United States
A R T I C L E I N F O
A B S T R A C T
Keywords: Criticality Helium Materials Rare earth elements Recycling Rhenium Substitution Supply chain Sustainability
As the technologies we use as a society have advanced, so have the materials used in these technologies. Some of these materials are exotic and highly specialized, making them particularly vulnerable to supply disruptions and supply disruptions particularly impactful. Such materials are designated as “critical” materials. Their level of criticality can be identified by accounting for a number of factors related to their supply risk and the extent to which a supply disruption would impact business operations or society at large. We highlight current methodologies used to assess materials criticality, how these assessments are used to reduce materials-related risk and to what extent there is room for improvement. Particularly, this paper reviews critical materials designations from the United States Department of Energy, the European Union, and the General Electric Company, and how they have changed over the period from 2008 to 2014. The changes suggest that the factors considered in criticality ratings have different natural time scales, and that criticality changes occur both due to supply-side risk mitigation as well as demand-side responses. Response options, whether on the supply or demand side, also span a range of time scales and the interaction between factors with different time scales can play a significant role in the dynamics. To date, many published analyses are snapshots in time. A detailed understanding of how risk profiles evolve remains an open question. The importance and impact of demand-side responses such as recycling, substitution and new technological development are discussed.
1. Introduction Modern society uses a wide range of raw materials and these materials go through cycles of surplus and shortage [1]. Efforts to anticipate materials shortages were renewed in 2008 with the National Resource Council (NRC) study, which introduced the concept of materials criticality [2]. Critical materials are designated as such because they are vulnerable to supply disruptions and such disruptions would have significant adverse impacts for businesses and society at large. The criticality of a given material or element is often considered along two dimensions, namely the level of supply risk and the impact a supply disruption would have. Criticality analyses use different factors that measure the exposure of a given material to each of these dimensions, and may also consider risks associated with a third dimension, environmental factors [3,4]. Factors are aggregated to create a score for each material along each dimension [5,6]. Virtually all published assessments are snapshots in time. Assessments can be a valuable tool in planning policy both at the industry and government levels [7,8]. They can also be used to identify and mitigate supply chain risks [9]. For example, the U. S. Department of Energy (DOE) established a Critical Materials Institute (CMI) to coordinate and provide strategic focus to
⁎
efforts to address critical materials relevant to the US energy infrastructure. In order to use these tools to inform policy, it is important to understand the drivers of materials criticality as well as how criticality changes over time. Generally, the criticality of a given material can change due to market responses, geopolitical factors, and technology development. Retrospective studies of market trends and geopolitics can provide insight into historical criticality trends [5,10]. Minerals and metal markets are difficult to accurately forecast for many of the same reasons as energy markets. To-date efforts can be categorized into two general approaches. Top-down approaches posit different scenarios for economic and political landscapes and then work out the materials market implications [6,11]. The logical consequences of particular narratives related to economic growth, government policy, or sociopolitical situations can be explored using sensitivity analyses and statistical tools. However, top-down approaches have difficulty capturing technological innovations that introduce non-linear changes in materials use patterns [12,13]. Bottom-up approaches attempt to reconstruct dynamics from different links in the supply chain. Agent-based dynamic material flow models can capture interactions by segmenting the links in the supply chain [10,14,15]. As with any modeling, both top-down
Corresponding author. E-mail address: [email protected] (S.J. Duclos).
https://doi.org/10.1016/j.susmat.2017.11.002 Received 1 May 2017; Received in revised form 29 September 2017; Accepted 18 November 2017 2214-9937/ © 2017 Published by Elsevier B.V.
Please cite this article as: Ku, A.Y., Sustainable Materials and Technologies (2017), https://doi.org/10.1016/j.susmat.2017.11.002
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Table 1 Comparison of critical materials weights used in computing composite supply risk score. Supply risk scoring
Yale [3]
DOE [6.7]
GE
BGS [22]
EU [8,24]
Materials (year of publication)
62 (2015)
14 (2010) 16 (2011)
33 (2008) 53 (2012)
52 (2011) 41 (2012)
41 (2010) 54 (2014)
Physical availability Reserves/Depletion time Companion production Recycling rate Production Producer concentration Producer stability Producer governance Producer policy Market factors Price volatility Substitutability Competing demand
33% 1/6 1/6
50% 2/5 1/10
22% 1/9 1/9
29% 1/7
67% 1/6 1/6 1/6 1/6
40% 1/5
33%
1/5
1/6 1/6 45% 1/9 1/6 1/6
10%
1/10
1/7 56% 2/7
Included Included
2/7 14% 1/7
Included
EU and GE differ and this will impact the ratings. Materials that GE considers critical to GE due to exposure from its manufacturing supply chain for a specific product line related to medical imaging may not be as critical to the DOE as it considers clean energy materials requirements a decade in the future. Exposure to the coking coal supply challenge may be less of a concern to both GE and the DOE, but for slightly different reasons. For instance, it is less related to the DOE's medium- to long-term energy goals, and it is not directly related to GE's most pressing supply concerns. This is an important point when considering dynamics because individual technological developments and their implementation would be expected to have a both larger and faster impact on criticality at the company level. Criticality assessments take into account two key dimensions associated with materials risks. The first dimension is supply risk, or how susceptible a material is to supply disruption. The other is the impact caused by a shortage. Typical assessments focus on specific materials of interest and assign a risk score to each material in each dimension. Scoring is sensitive to the interests of the organization performing the assessment: producers and users may rate supply risk differently, as might companies active in different industries. Materials are then plotted along the two axes and materials that score above a threshold are designated critical. Some analyses, such as the one performed by the Yale group also include an environmental impact axis in their assessments [3]. Scores for supply risk and impact of disruption are generally composite indices that take into account several factors. In considering dynamics, an appreciation of the different factors that contribute to criticality scoring allows one to consider how they change with time and to consider interaction effects. Tables 1 and 2 summarize how five organizations compute the supply risk and impact of disruption. Despite some differences in terminology and relative weighting of factors, most assessments focus on the same contributing factors. Four of the five methodologies compute risk score as the weighted sum of various factor scores, while the EU approach uses multiplicative formulae. Factors that contribute to supply risk include physical availability, production, and market factors. Physical availability is tied to global reserves, co-production, and recycling rates. Co-production refers to the fact that many elements are produced as by-products of a primary ore body [3,21]. Production factors include geographic concentration, geopolitical stability, and policy. Refining and distribution bottlenecks are also captured in production scores. This factor is especially important when considering materials that may have only a few capable refiners or suppliers. A third category includes market factors such as price volatility, the availability of substitutes, and competing demand between different end use industries. The competition between different industries is particularly
and bottom-up approaches rely on the quality of the assumptions used in the modeling. It is essential to recognize that changes in criticality are induced by both exogenous shocks to a supply chain or usage pattern, as well as endogenous market responses to such shocks. Technological innovation can introduce inflection points that can shift materials use patterns, which may increase or decrease the criticality level of a particular material or materials. For example, the maturation of nickel-base superalloys, coupled with a shortage of cobalt in the 1950′s lead to a shift towards the former [16]. Beginning in the late 1980s, a transition from halophosphate phosphors to rare earth-baed triphosphor blends gradually boosted demand for certain rare earth elements (REE) [17]. The microelectronics industry routinely introduces new materials-enabled devices. The implication is that increased usage of such exotic materials comes at the cost of increased criticality. There is also a bright side to technological development: advances in technology are not limited to initiating material uses, rather can also lead to decreased usage of critical materials. Progress in production, manufacturing and recycling methods can drastically reduce the demand for various materials [18,19]. There can also be interactions between technological innovation and supply. For example, a second shortage of cobalt in the 1970's was a contributing factor to the development and adoption of NdFeB magnets as a substitute for SmCo magnets [20]. One challenge that we will face with these criticality assessment tools is whether they can accurate anticipate crises early enough to allow meaningful action. This paper considers the dynamics of materials criticality with a special focus on the role and consequences of technological innovation. The discussion is organized into four parts. The first section introduces definitions of criticality and how criticality scores are assigned. The second part reviews assessments by the DOE, the European Union (EU), and the General Electric Company (GE) and how they changed over a period of one to four years. The third part considers how risk evolves. The final section makes some observations on how technological innovation responds to materials criticality, with an emphasis on the nature of substitutes and time tables around their development. Examples will be drawn from GE's experiences in responding to its materials criticality challenges and the recent literature on REEs, given the extensive public discussion of these materials. Technology-driven inflection points in the usage of various materials are briefly discussed. The paper concludes with some implications for the materials development community. 2. Definitions of criticality A first comment on criticality assessments is that they are dependent upon the scope of the review. The time scales and interests of the DOE, 2
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Table 2 Comparison of critical materials weights used in computing composite impact score Impact scoring: scope of study
Yale: global importance
DOE: importance to clean energy
GE: impact to operations
Importance Economic impact Usage % by population % of world's supply used Substitutability Performance Availability Environmental impact Market factors Cost pass through
50%
75%
50% 1/4
EU: economic importance 100%
1/2 50% 1/6 1/6 1/6
1/4 25% 1/4
25%
Included in supply risk
25% 1/4
3. Historical dynamics in criticality risk ratings
relevant when considering the effects of technological development on supply risk. The particular weightings reflect the concerns of the different interested parties. For example, GE's analysis emphasizes market factors, while the EU or DOE are more concerned with physical production related to national and international interests. The EU supply risk score (SR) includes factors from each of the three categories, and is computed by:
Assessments are conducted as snapshots in time, but changes in criticality scores between snapshots provide insight into the underlying risk dynamics. The three groups highlighted above – the DOE, EU, and GE – have published detailed updates over periods of up to four years. This section reviews changes in these assessments as a practical starting point for a discussion on dynamics. Others, such as the Yale group, are in the process of updating their original assessments. The United States Department of Defense (DOD) maintains a stockpile of materials for defense and issues annual reports to Congress on stockpiling needs [23]. However, these reports only summarize the final recommendations without detailed discussion of the methodology, uses, or dynamics. This makes the results interesting, but less useful for probing dynamics. We see a few common themes between these analyses. First, some new materials enter updated analyses and other leave depending on changing market conditions. Second, specific scores typically shift slightly, however, changes in the impact scores are generally more significant than those in the supply risk scores. This could be due to both technical and methodological reasons. Technical drivers include the fact that impact scores are primarily functions of technological development and market dynamics. Both of these factors can shift relatively more quickly than the primary determinants of supply risk (such as economic reserves, which change only infrequently with significant new ore deposit discoveries). We also note that materials with high composite criticality scores also attract attention aimed at mitigating such risk, which can amplify the rates of change.
SR = S (1 − R) H where S is the substitutability (0 to 1, with 1 being least substitutable), R is the recycling rate (percentage of end-of-life supply) and HHI is the Herfindahl-Hirschman index for market concentration defined as H = ∑(σi2 ∙ WGIi) where σi and WGIi are the share of production and world governance indicator for country i [8]. Table 2 shows the factors for impact scoring from the five methodologies. They are grouped into three categories related to the importance of the material, the availability of substitutes, and other market factors. The EU economic importance score (EI) is given by:
EI =
BGS: no rating assigned
∑ (si vi ) GDP
where si and vi are the share of use and value contributed to each megasector, and GDP the overall gross domestic product of the EU. Depending on the methodology, substitutability appears as a factor in either supply risk or impact. The British Geological Survey (BGS) and EU consider it under supply risk, Yale and the DOE treat it as an impact factor, and GE includes it in both. From a supply risk perspective, substitution is a market mechanism that can shift demand. However, individual users utilize substitution to reduce risk for high impact materials. Since it is application-specific, substitutability has been difficult to quantify and to forecast.
3.1. U.S. Department of Energy assessments The DOE published an assessment in 2010 on materials important to the deployment of clean energy technologies within the US. A revision
Fig. 1. Comparison of short-term and medium-term DOE clean energy materials criticality assessments from 2010 and 2011 [6].
3
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
switched relative to the DOE assessment. Supply risk is on the y-axis and the Economic importance is on the x-axis. Critical materials from the 2014 assessment are in black (materials in both versions) and red (new materials added in 2014). Material designations from 2010 for selected materials are shown in blue. There were three types of changes. First, the supply risk changed due to production concentration increasing or decreasing. This can be seen in the vertical shifts in the platinum group metals (decreasing) and magnesite (increasing). Second, changes in the economic importance resulted in horizontal shifts, as seen in niobium (decreasing) and tungsten (increasing). Finally, there were four changes to the criticality list. Tantalum dropped off due to reduced supply risk concerns while borates, chromium, and magnesite were promoted due to increasing concerns around supply concentration. The EU supply risk emphasizes concentration risk, which was the primary driver of change over the three-year span between assessments. Similarly, growth or contraction in economic sectors is the primary driver of changes in the EI index. Technological deployment was not explicitly accounted for. Efforts to promote recycling and support the development of substitutes were supported, but three years is a short time to see significant impacts from new efforts. A final technical factor, relevant to the other analyses as well, is improving data quality. Since criticality assessments are still in their early stages, some of the changes in criticality score may also be due to improved data quality. The EU plans to refresh its analysis every 5 years, and initiated efforts in 2015 to improve the fidelity and rigor of their assessment methodology [8].
Fig. 2. Comparison of critical materials designations by the European Union from 2010 to 2014. The region shown corresponds to the upper right-hand (critical) portion of the EU criticality diagram; materials in the shaded region are considered critical. Blue dots correspond to materials from the 2010 analysis. Black (original) and red (new) dots correspond to materials in the 2014 study. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Adapted from references [8] and [24].
was published in 2011. The revised version expanded the list of elements from 14 to 16, while adjusting the results for changes in markets and technologies. The main changes were minor shifts in the position of several REEs, indium, and gallium. Indium was demoted from the near-term criticality watch list. The primary drivers were the improved competitiveness of silicon-based solar cells relative to thin-film solar technologies, an increase in the projected adoption rate for light emitting diodes (LED) for lighting, and the announcement of several rare earth mining projects leading to expectations of increased supply [7]. The trends are shown graphically in Fig. 1.
3.3. General Electric Company assessment GE performed a criticality assessment of its operations in 2008, and refreshed the analysis in 2012. The original assessment considered 33 elements, and the refresh expanded the scope to 53 elements. Elements were scored according to Supply risk and Impact of a disruption on company operations. Only raw materials directly purchased by GE were accounted for in the analysis. Materials used as manufacturing consumables (e.g., tungsten carbide tooling or inert gases for heat treatments) or embedded in product components purchased from original equipment manufacturers (OEMs) were not included. Fig. 3 shows company-level criticality diagrams from 2008 and 2012. Supply risk is plotted on the x-axis and Impact is plotted on the yaxis, consistent with the DOE plots. Each circle represents an individual element with the area of the circle being the raw material spend for that element. Most of the circles are unlabeled due to the proprietary nature of information pertaining to company operations. Rhenium, helium, and REEs are identified for illustrative purposes. At a glance, it can be seen that the supply risk of some of the rare earth elements relative to Re has increased. Similarly, the effects of mitigation efforts can be seen in the smaller spend on Re and the wider distribution of impacts among the rare earth family of elements. A closer look at changes in their criticality scoring over the intervening three years is informative. Rhenium added nickel-based superalloys improves the high temperature mechanical properties for turbine applications [25]. A price
3.2. European Union assessments The EU performed its first critical materials assessment in 2010 [8], focused on materials of importance to the European economy. Fortyone non-fuel, non-agriculture raw materials were scored on economic importance and supply risk, according to the sub-factors listed in Table 1. Fourteen materials were designated as critical materials including: antimony, beryllium, cobalt, fluorspar, gallium, germanium, graphite, indium, magnesium, niobium, platinum group metals (PGM), REE, tantalum, and tungsten. An update to the analysis was published in 2014 [24]. In the second assessment, the scope was expanded to include fifty-four materials, of which twenty were designated critical. Fig. 2 highlights significant changes in the “critical” region between the two assessments. As such the figure only illustrates the upper righthand (critical) portion of the EU criticality diagram. The axes are
Fig. 3. GE assessments – criticality diagrams from 2012 with helium, rhenium and REEs overlaid. Additional details beyond the general labels for rare earth elements, and those for He and Re, are not included due to their proprietary nature.
4
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Fig. 4. GE Lighting and Healthcare criticality diagrams from 2012.
development of the supply chains for these new applications. There were also some examples of substitution at a system level. One notable example was an announcement in 2011 that GE's 1.5 and 2.5 MW wind turbine product lines would use a doubly-fed induction generator (DFIG) architecture in lieu of direct drive generators that would use a significant amount of REE-based permanent magnets. Awareness of the supply challenges for rare earths continues to be high across the company.
spike occurred in 2008, due to a confluence of factors including limited supply, market concentration, and financial speculation. Given the importance of superalloys to the operations of GE's Aviation and Energy divisions, rhenium was rated one of the most critical elements to GE in 2008. Although it was among the most critical elements in 2012, the relative criticality of rhenium declined due to a combination of market responses and internal initiatives to reduce total demand and spend. Some of these internal efforts included the qualification of alloys with reduced rhenium content, improved utilization in manufacturing operations, and reclamation from end-of-life components. Improvements in manufacturing efficiency and adoption of relatively mature alternate alloys helped in the short-term. Longer term options, such as the development of superalloy compositions that require less rhenium and new materials such as ceramic matrix composites, also remain in play. Helium was added to the 2012 assessment. GE accounted for about 4% of global consumption in 2011, primarily for cooling of superconducting magnets used in magnetic resonance imaging (MRI) machines. Helium had a moderate-high supply risk score, and a high impact score. Supply risk resulted primarily from its co-production as byproduct of natural gas liquefaction and the high concentration of supply from a single source (about a third of global production was associated with the United States National Helium Reserve). The high potential impact of a disruption on GE operations was due to the lack of alternatives to helium in MRI magnet cooling and some manufacturing operations, and GE's non-trivial share of world consumption. Concerns over helium proved to be prescient as U.S. government policy debate over the future of the National Helium Reserve resulted in supply constraints in 2013. The impact of government action to extend the life of the Reserve, and internal efforts to reduce exposure are expected to show up in the next criticality assessment. GE's analysis assigned criticality ratings to each individual rare earth element. Although specific elements are not identified in Fig. 3, some insight can be gained by looking at the spreads in aggregate. The range in supply risk reflects the difference between light REEs, which are naturally found in greater abundance in most deposits, and heavy REEs. The spread in impact captures differences in use and substitutability across different product lines. Fig. 4 shows criticality diagrams for GE's Lighting and Healthcare divisions, which have heighted exposure to rare earths due to their heavy use of europium, terbium, and yttrium in phosphors for fluorescent lighting and scintillators for medical imaging. Helium appears in the GE Healthcare diagram as well. Over the period from 2008 to 2012, GE Lighting re-optimized its phosphor blends to improve manufacturing yield, evaluated the benefits of phosphor recycling, and engaged in R&D to develop alternate phosphor compositions with reduced rare earth content. These activities were complementary to efforts to improve the robustness of its supply chains, as a matter of course for business. In parallel, GE Healthcare took steps to improve the recycling of manufacturing scrap from scintillator production. Efforts were also made to identify and introduce substitutes. In one case, GE Aviation found that some heavy REEs have the potential to function interchangeably in some of their next-generation coatings [26]. This provided some flexibility in the
3.4. General trends and observations Despite different mathematical forms for computing supply risk and impact, the general trends across the DOE, EU and GE assessments are consistent. They agree on the continued criticality of REEs, particularly the heavy REEs, and that the biggest near-term drivers of change were economic cycles and the commercialization of technologies already under development. Since criticality analysis is still a relatively young field, the inclusion of new materials to the assessments can expand watch lists regardless of the underlying dynamics. Examples of this are coking coal, phosphate and silicon metal for the EU, and helium for GE. Major shifts in criticality appear to require more time, ostensibly due to the longer time frames associated with opening a new mine, or the development, validation and commercialization of entirely new technologies. The question of relevant time scales and their impact on risk dynamics is the focus of the remainder of this paper. 4. Criticality risk dynamics In practice, criticality assessments are used to screen a wide range of materials and identify a high risk short list for follow-up. The value of incorporating dynamics, even crudely, is that it offers the possibility of identifying materials with high potential to emerge as critical. A wide range of specific factors can drive dynamics and determine the rates of change. For example, geological availability is expected to change slower than price volatility. Technological innovation varies among end use sectors. An example of this concerns elements used in both microelectronics and superalloys such as tantalum, cobalt and some REEs. Consumer electronics product cycles are short compared to the time needed to develop and qualify superalloys for aviation applications. As a result, new applications in microelectronics could create unanticipated challenges for the aerospace industry because of the longer time horizons needed to develop alternatives. Since a supply chain can be disrupted by a bottleneck in any single link, the distribution of the factors that go into the supply risk and impact provides insight into root causes and possible risk mitigation strategies. An ever-present concern with mismatches in innovation timeframes is that, to a supplier, ultimately only the sales price matters. This could lead to significant amounts of a raw material going to a single application (or set of applications) having the highest marginal value causing a supply deficit for other applications. This innovation-timing mismatch can be a serious risk, particularly when industries with significantly different development cycles compete for the same materials. Future studies can 5
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
conservation measures triggered by near closure of the National Helium Reserve may reduce demand in a way that could offset lost production.
quantify these effects more rigorously for specific scenarios by computing response curves using publicly available assessment tools. Table 3 shows different factors and estimated ranges for the time scales for significant changes. Public information can be used to validate a number of these factors including market data (including pricing and product introduction cycle times in specific industry) and government statistics (such as political turnover or USG data on production and reserves). Since some indices used to estimate reserves and several production factors are updated annually, the possibility of data lags must be considered. Factors related to physical availability generally change over the course of years to decades, while production factors appear to change over years. Narrower ranges for specific elements can be estimated from mining and market data. Price volatility for commodity materials can be computed from market data; this may be more difficult for materials with smaller or more opaque markets. Time scales for substitution and changes across multiple industries can change are more industry-specific, with innovations requiring significant invention taking several decades to make it to market. These will occur against a backdrop of longer term trends related to physical availability and technological progress, and near-term perturbations that test the resilience of existing supply chains. The interplay between different supply risk factors can produce inflections in market equilibrium demand. One example of this is the introduction of LEDs into the lighting market. Driven by technology development and government policy towards energy efficiency, this transition is already underway and expected to continue through the end of the decade [6,30]. Due to the lower REE use per bulb of LEDs relative to fluorescent lighting, the adoption of LEDs is expected to reduce demand pressure. This transition is occurring at a time of heightened government awareness of rare earth criticality, and can work with efforts to identify material substitutes to reinforce a downward trend in the supply risk of certain REEs. A second example illustrates that opposing trends may also act to offset each other. Helium is primarily a by-product of certain natural gas fields and the increasing share of natural gas production from shale gas fields which are low in helium could have an adverse supply impact. At the same time,
5. Response dynamics and substitution The response to each materials crisis will depend on the exact nature of the disruption and use. However, response options can be grouped into five categories, shown in Fig. 5. Sourcing approaches focus on commercial arrangements. Manufacturing efficiency and recycling approaches aim to use available material more effectively in an effort to reduce net demand for raw material. Substitution, which typically requires technology development, seeks to replace or reduce the use of the material. Responses require different amounts of time to have an impact. The planning, permitting, construction, and start-up of a new mine can take over a decade. The recovery and reuse of manufacturing scrap can be implemented more quickly. This is particularly true in cases where the methods already exist and only require a price increase in raw material to make them economical. Substitution can occur in two ways. The first is a direct material replacement. The existence of adequate, but more costly, alternatives could hasten material substitution, but the resulting products may still require tooling or other upgrades at manufacturing facilities or qualification in regulated industries. The invention of a new material could add years or decades to this process. A second substitution option is a different system design that replaces the function of the critical material. Examples of this, discussed earlier, include the transition from fluorescent lighting to LEDs, and GE's decision to move from rare earth permanent magnet direct drive generators to induction architectures in wind turbines. While these actions reduce the risk associated with REEs, they simultaneously increase the risk for materials used in the substitutes. This feedback loop is one of the reasons that criticality diagrams needs to be periodically updated and the dynamics better understood. Additionally, as designers increasingly use more exotic materials it will be essential to incorporate a culture of designfor-sustainability so as to mitigate potential criticality issues during the development stages of new products and technologies.
Table 3 Estimated time scales for criticality risk factors. Supply risk scoring
Time scales
Examples
Physical availability Reserves/depletion time Companion production Recycling rate
10 to 30 years 5 to 20 years 1 to 10 years
Discovery of new reserves [27] Development of improved separation technologies Government policy; economics [18]
Production Producer concentration Producer stability Producer governance Producer policy
1 1 1 1
Structural changes in geographical distribution of production Infrastructure investment [28] Changes in political stability of producer nations Mining or trade policy [28]; Environmental regulations
Market factors Price volatility Substitutability Competing demand
1 to 3 years Up to 20 years Up to 20 years
Shifts in supply/demand balance; short-term disruptions in supply Technology innovation [28,29] Technology innovation [28,29]
Impact risk scoring
Time scales
Examples
Importance Economic impact Usage % by population % of world's supply used
1 to 10 years 5 to 10 years 1 to 10 years
Business cycles; introduction of new technology Structural economic shifts [28]; changes in technology [29] Changes in business strategy or models
Substitutability Performance Availability Environmental impact
Up to 20 years Up to 20 years Up to 20 years
Introduction of new technology Introduction of new technology Introduction of new technology
Market factors Cost pass through
< 3 years
Company responses to markets
to to to to
10 years 5 years 5 years 5 years
6
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
Fig. 5. Responses to material supply disruption.
discussed herein offer the practitioner more tools to anticipate and mitigate these risks.
Shortages (and sometimes expectations of shortages) trigger multiple, simultaneous responses. Interactions between these overlapping responses can be an important driver of dynamics. Governments, producers and users of a given material can also coordinate their responses. However, cooperation is not guaranteed and disconnects between competing responses may make them less effective than they might otherwise be. Returning to the REE example, efforts to develop new mines outside of China have the potential to both increase and diversify supply. However, in the time it takes to open a mine, users of rare earths are pursuing substitutes to reduce their demand. Similar economic dynamics apply to recycling processes. High prices incentivize efforts to establish recycling programs, but easing of demand can reduce their economic feasibility. Government activity and public awareness have increased. Research in material substitutes, refining processes, manufacturing improvements, and recycling technologies is under way. Together, these responses have resulted in differentiation between rare earths. The EU, in its latest assessment, split their original rare earth designation into a higher risk group of heavy rare elements, and a lower risk group of light REEs. Efforts to understand and model the dynamics of criticality are still at an early stage. In particular, more detailed frameworks are needed to understand how fast an industry can adapt to perturbations in its material supply chains, how links in the material supply chain interact with each other, and the possible convolution of dynamic effects between multiple risk factors in criticality scoring. These efforts can build on work from dynamic materials flow analysis, which is already wrestling with the question of forecasts [31]. In particular, some of the statistical methods for modeling material flows may be appropriated for addressing technological innovation [32]. New metrics for how technology transitions from lab to commercial production may provide insight into how fast industry can adopt a new material.
Acknowledgments The authors would like to acknowledge helpful discussions with A Chowdhury, J Herzog, S Hung, F Johnson, D Lipkin, J Luo, W McLinko, L Quirk, and C Thom at GE, and T Graedel from Yale. References [1] T.E. Graedel, E.M. Harper, N.T. Nassar, B.K. Reck, On the materials basis for modern society, Proc. Natl. Acad. Sci. 112 (2013) 6295–6300. [2] National Research Council, Minerals, Critical Minerals, and the U.S economy: Committee on Critical Mineral Impacts of the U.S. Economy, The National Academies Press, 2008. [3] T.E. Graedel, E.M. Harper, N.T. Nassar, P. Nuss, B.K. Reck, Criticality of metals and metalloids, Proc. Natl. Acad. Sci. 112 (2015) 4257–4262. Available electronically at www.pnas.org/cgi/doi/10.1073/pnas.1500415112 (Accessed: May 5, 2015). [4] N.T. Nassar, R. Barr, M. Browning, Z. Diao, E. Friedlander, E.M. Harper, C. Henly, G. Kavlak, S. Kwatra, C. Jun, S. Warren, M. Yang, T.E. Graedel, Methodology of metal criticality determination: criticality of the geological copper family, Environ. Sci. Technol. 46 (2012) 1071–1078. [5] S. Gloser, L. Tercero Espinoza, C. Gandenberger, M. Faulstich, Raw material criticality in the context of classical risk assessment, Res. Policy 44 (2015) 35–46. [6] U.S. Department of Energy. Critical Materials Strategy: 2011 Technical Report, Washington DC. [7] U.S. Department of Energy. Critical Materials Strategy: 2010 Technical Report, Washington DC. [8] European Commission, Critical raw materials for the EU: 2010. Technical Report, Brussels, Belgium, Available electronically at http://ec.europa.eu/enterprise/ policies/raw-materials/critical/index_en.htm , Accessed date: 2 January 2015. [9] S.J. Duclos, J.P. Otto, D.G. Konitzer, Design in an era of constrained resources, Mech. Eng. 132 (2008) 36–40. [10] E. Muller, L.M. Hilty, R. Widmer, M. Schleup, M. Faulstich, Modeling metal stocks and flows: a review of dynamic material flow analysis methods, Environ. Sci. Technol. 48 (2014) 2102–2113. [11] E. Alonso, A.M. Sherman, T.J. Wallington, M.P. Everson, F.R. Field, R. Roth, R.E. Kirchain, Evaluating rare earth element availability: a case with revolutionary demand from clean technologies, Environ. Sci. Technol. 46 (2012) 3406–3414. [12] B. Buijs, H. Sievers, L.A. Tercero Espinoza, Limits to the critical raw materials approach, Proc. Inst. Civ. Eng. Waste Resour. Manag. 165 (2012) 201–208. [13] K. Oelich, D.A. Dawson, P. Purnell, C. Knoeri, R. Revell, J. Busch, J.K. Steinberger, Assessing the dynamic material criticality of infrastructure transitions: a case of low carbon electricity, Appl. Energy 123 (2014) 378–386. Available online at https:// doi.org/10.1016/j.apenergy.2014.01.052. [14] G. Gaustad, E. Olivetti, R.E. Kirchain, Toward sustainable material usage: evaluating the importance of market motivated agency in modeling material flows, Environ. Sci. Technol. 45 (2011) 4110–4117. [15] C. Knoeri, P.A. Wager, A. Stamp, H.-J. Althaus, M. Weif, Towards a dynamic assessment of raw materials criticality: linking agent-based demand with material flow supply modelling approaches, Sci. Total Environ. 461–462 (2013) 808–812. [16] C.T. Sims, A history of superalloy metallurgy for superalloy metallurgists, Superalloys (1984) 399–419. Available electronically at http://www.tms.org/ Superalloys/10.7449/1984/Superalloys_1984_399_419.pdf (Accessed: May 5, 2015).
6. Conclusions The materials use profile of modern society will continue to evolve. At any given time, criticality assessments can help identify important materials that are also vulnerable to supply disruptions. The dynamics play out over multiple time scales. An understanding of which materials are critical, the drivers and dynamics behind their criticality can inform efforts to develop new materials and technologies. Simply put, material shortages can motivate technology development and technology development can cause material shortages. In this context, recycling, substitution and technological development are all important avenues for innovation related to materials supply risk. Rather than simply rely on myopic price signals, materials criticality analyses like those
7
Sustainable Materials and Technologies xxx (xxxx) xxx–xxx
A.Y. Ku et al.
[17] A.M. Srivastava, T.J. Sommerer, Fluorescent lamp phosphors, InterfaceElectrochem. Soc. 7 (1998) 28–33. Available electronically at http://www. electrochem.org/dl/interface/sum/sum98/IF6-98-Page28-31.pdf , Accessed date: 5 May 2015. [18] Y. Houari, et al., A system dynamics model of tellurium availability for CdTe PV, Prog. Photovolt. Res. Appl. 22 (129) (2014). [19] T.E. Graedel, J. Allwood, J.P. Birat, M. Buchert, C. Hageluken, B.K. Reck, S.F. Sibley, G. Sonnemann, What do we know about metal recycling rates? (2011) USGS Staff - published research. Paper 596, Available electronically at http:// digitalcommon.unl.edu/usgsstaffpub/596 , Accessed date: 5 May 2015. [20] E. Alonso, J. Gregory, F. Field, R. Kirchain, Material availability and the supply chain: risks, effects, and responses, Environ. Sci. Technol. 41 (19) (2007) 6649–6656. [21] A.Y. Ku, S. Hung, Manage materials risk, Chem. Eng. Prog. (2014) 28–35. Available electronically at http://www.aiche.org/sites/default/files/cep/20140928.pdf , Accessed date: 3 May 2015. [22] British Geological Survey - Risk List, Available electronically at http://www.bgs.ac. uk/mineralsuk/statistics/risklist.html, (2012) , Accessed date: 8 April 2015. [23] Defense Logistics Agency, Strategic and critical materials 2015 report on stockpile requirements. Jan 2015, Available electronically at www.strategicmaterials.dla.mi/ Report%20Library/2015%20NDS%20Requirements%20Report.pdf , Accessed date: 5 May 2015. [24] European Commission, Report on critical raw materials for the EU. Technical
[25] [26]
[27] [28] [29]
[30]
[31] [32]
8
report, Brussels, Belgium, Available electronically at http://ec.europa.eu/ enterprise/policies/raw-materials/files/docs/crm-report-on-critical-raw-materials_ en.pdf , Accessed date: 2 January 2015. P. Caron, T. Khan, Evolution of Ni-based superalloys for single crystal gas turbine blade applications, Aerosp. Sci. Technol. 3 (1999) 513–523. A.Y. Ku, C. Dosch, T. Grossman, J. Herzog, A. Maricocchi, D. Polli, D.M. Lipkin, Addressing rare-earth element criticality: an example from the aviation industry, JOM 66 (11) (2014) 2355. P. Darling (Ed.), SME Mining Engineering Handbook, 3rd ed, Society for Mining, Metallurgy and Exploration, Inc, 2011. K.S. Gallagher, J.P. Holdren, A.D. Sagar, Energy-technology innovation, Annu. Rev. Environ. Resour. 31 (2006) 193–237. J. Shea, What do technology shocks do? in: B.S. Bernanke, J. Rotemberg (Eds.), NBER Macroeconomics Annual 1998, vol. 13, MIT Press, 1999Available electronically at http://www.nber.org/chapters/c11249 , Accessed date: 7 May 2015. Anthony Y. Ku, Anant A. Setlur, Johnathan Loudis, Impact of Light Emitting Diode Adoption on Rare Earth Element Use in Lighting: Implications for Yttrium, Europium, and Terbium Demand, The Electrochemical Society Interface24.4 (2015) 45–49. M.T. Melo, Statistical analysis of metal scrap generation: the case of aluminium in Germany, Resour. Conserv. Recycl. 26 (1999) 91–113. H. Rechberger, T.E. Graedel, The contemporary European copper cycle: statistical entropy analysis, Ecol. Econ. 42 (2002) 59–72.