Costs, Substitution, and Material Use: The Case of Rare Earth Magnets

Subscriber access provided by UNIV OF DURHAM

Article

Costs, substitution, and material use: The case of rare earth magnets Braeton Smith, and Roderick Eggert Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b05495 • Publication Date (Web): 02 Mar 2018 Downloaded from http://pubs.acs.org on March 3, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29

Environmental Science & Technology

2

Costs, substitution, and material use: The case of rare earth magnets

3

Braeton J. Smith*,1,2 and Roderick G. Eggert1,2

4 5 6

1

Division of Economics and Business, Colorado School of Mines, Golden, CO 80401, United States Critical Materials Institute, Ames, IA 50011, United States *Corresponding author; email: [email protected] 2

Dy/Tb in NdFeB (% weight)

1

9% 8% 7% 6% 5% 4% 3% 2% 1% 0% 2010 2016 120 ℃

7

2010 2016

2010 2016

150 ℃ 180 ℃ Maximum operating temperature

8 9 10 11 12 13 14 15 16 17 18 19 20

Abstract: Environmental technologies depend on raw materials, some of which are subject to volatile costs and availability concerns. One way to address these concerns is through substitution, of which there are many types. An important form of substitution in the short term is adopting an alternative production process, yielding a material with the same functional properties with less material input. In effect, technology substitutes for material. This study elucidates the role increased and uncertain material costs play in inducing different substitution types in the short to medium term. Specifically, this paper uses an expert survey to determine the relative importance of eight specific industry responses taken by magnet and wind turbine manufacturers in response to the 2010/2011 rare-earth price spike through 2016. Statistical tests show adopting an existing production process for magnets was the most important response, followed by cost passthrough, using an alternate magnet grade in a redesigned generator system, and using alternate systems altogether. The paper also provides specific findings for the magnet and wind turbine industries with respect to each substitution type.

21 22

Keywords: material substitution, material costs, technology adoption, rare earth elements, permanent magnets, clean energy technology, expert survey, nonparametric statistics

23

1

ACS Paragon Plus Environment


Page 2 of 29

24

1 Introduction

25

Modern technologies depend on raw materials. Less obvious is that some raw-material supply

26

chains are risky and some raw materials prices are especially volatile – discouraging

27

development and adoption of otherwise promising technologies. Risky supply chains and volatile

28

material costs are among the challenges to the widespread proliferation of clean energy

29

technologies.1,2 The ability to reduce the amount of certain raw materials in environmental

30

technologies through substitution would alleviate these concerns and increase the likelihood of

31

successfully meeting climate change and pollution goals.

32

The substitutability of material inputs is an important consideration for all cost-minimizing

33

firms, particularly when faced with supply constraints and unpredictable price fluctuations. The

34

ability to substitute affects one’s dependence on material markets and supply disruptions, and

35

thus, the ability to compete in a crowded marketplace. Accordingly, understanding

36

substitutability is paramount to addressing material criticality and energy system resilience

37

concerns3,4 as well as performing climate change policy and scenario analysis.5-8 Unfortunately,

38

material substitution is often poorly understood and difficult to quantify, largely due to a lack of

39

quality data at the detailed level. This paper explains how permanent magnet and wind turbine

40

manufacturers used substitution as a strategy to respond to the significant and rapid rare-earth

41

element price increases in 2010-2011.

42

There is some discussion in the literature about what induces substitution as well as what is

43

technically feasible. Graedel et al.9 and Nassar10 find that direct elemental substitution does not

44

exist for most elements in their major applications. Tilton11,12 and Schlabach13 identify several

45

more general substitution types from economic and technological perspectives, respectively, that 2


Page 3 of 29


46

can arise from many potential influences. Economists are especially interested in the role of

47

material costs, which have generally been shown to induce substitution in the long run for

48

highly-aggregated materials and end-uses.12,14 While this result has held for some disaggregated

49

studies of specific materials and end-uses with shorter time periods,15-18 it has not been

50

consistent.19-22 This leads Tilton23 to assert an “alternative” view of substitution that occurs

51

indirectly through technological innovations. High input costs and environmental policy are

52

among the key drivers of technological advances in environmental technologies.24-26

53

Most significant innovations happen in the long run, making manufacturers reliant on existing

54

technologies and small engineering refinements in the short run.27 Much attention has been given

55

to the optimal timing28,29 of technology adoption and the factors that induce it,30 yet the influence

56

of material costs has been largely overlooked. Still, changes in relative material costs have been

57

shown to induce substitution indirectly through the adoption of alternative existing production

58

technologies.21,22

59

The purpose of this paper is to elucidate the role increased and uncertain material costs play in

60

inducing substitution in the short term. Specifically, it examines the ways permanent magnet and

61

wind turbine manufacturers responded to the 2010/2011 rare-earth price spike through 2016. The

62

price spike, during which prices of certain elements increased by over 3,000 percent in a matter

63

of months, was a rare extreme event providing an uncommon opportunity to evaluate the

64

responsiveness of material users to severe price swings. Due to substantial data limitations for

65

quantities of specific elements used in specific applications over time, this study uses

66

quantitative and qualitative data from an expert survey (first described by Smith and Eggert31)

67

that asked experts to assess the degree to which companies implemented different strategies. It 3



Page 4 of 29

68

uses Smith and Eggert’s unique substitution “taxonomy” for permanent magnets in wind turbines

69

and substantially extends and interprets their findings for a different audience.

70

The contributions of this work lie in economics and industrial ecology. First, it provides an ex-

71

post case study (in an application with virtually no data) of the ways manufacturers respond over

72

a relatively short time period to rampantly increasing and uncertain material costs. Second, it

73

refines Tilton’s alternative view of substitution, arguing that high material costs can induce

74

substitution through the adoption of existing technology in the short run. Third, it argues that

75

being explicit about the nature of substitution communicates specific ideas with experts of

76

different disciplines. This study differs from previous case studies in this specific application,

77

which are mostly performed from engineering and industrial ecology perspectives,4,32-34 in its

78

historical emphasis on how high material costs induced actual changes in material and

79

technology use. Further, the Supporting Information (SI) provides specific and novel findings

80

about the way permanent magnet and wind turbine industries responded to the rare-earth price

81

spike.

82

1.1 Rare-earth Elements, Magnets, and Wind Turbines

83

Rare-earth elements (REEs) are essential ingredients in many modern technologies, such as

84

smart phones, televisions and monitors, hard drives, lighting, electric vehicles, and many others.

85

They comprise the group of 15 lanthanide metals and tend to occur in the same deposits as they

86

are chemically similar.35 REEs are divided into light REEs (LREEs), which have atomic

87

numbers 57 to 64, and heavy REEs (HREEs), which have atomic numbers 65 to 71.35

4


Page 5 of 29


88

Prior to 2012, China accounted for over 95 percent of global REE production.36 In mid-2010, the

89

announcement of a drastically reduced export quota for REEs caused panic worldwide among

90

end-users, governments, and other stakeholders, leading to rapid and extreme price increases (by

91

over 3,000 percent for some elements) over the following 12 months.31 The “price spike” lasted

92

through early 2012, when prices of most REEs fell substantially, though many remain well above

93

their pre-2010 levels.

94

One important use of REEs is in permanent magnets (PMs) used in wind turbines, electric

95

vehicles, and many other products. The four major PM types are ferrite, alnico, samarium-cobalt,

96

and neodymium-iron-boron (NdFeB). While NdFeB magnets, which contain the LREEs

97

neodymium (Nd) and praseodymium (Pr), are the most powerful magnets available at room

98

temperature, they demagnetize at high temperatures. Adding HREEs, usually dysprosium (Dy)

99

and occasionally terbium (Tb), improves temperature stability.37 NdFeB grades with higher

100

temperature resistance almost always have higher HREE contents. As NdFeB has roughly 32

101

percent total REE content by weight (most of which is Nd), magnets with more HREE contain

102

less Nd.31

103

Wind turbines contain an electric generator, traditionally a PM-free doubly-fed induction

104

generator (DFIG) coupled with a gearbox. Despite decades of use, DFIGs often break down,

105

requiring expensive maintenance. Direct drive generators, most of which use PMs, are more

106

efficient and require less maintenance.34 PM generators (PMGs) are especially useful in offshore

107

turbines, where the maintenance required in a DFIG configuration is prohibitive. PMG turbines

108

require substantial quantities of NdFeB material, often well over 1 ton per turbine.38 Despite their

5



Page 6 of 29

109

small market share, PMG use is expected to grow,39 especially as the first offshore windfarm in

110

the U.S. began supplying power with PMG turbines in 2016.40

111

2 Concepts and Methods

112

2.1 Responding to High Material Costs

113

There are several responses that firms might undertake when facing increasing or volatile

114

material costs. Broadly, these include either passing through or absorbing costs or pursuing

115

substitution of some type. Schlabach13 and Tilton11,12 define several substitution types from

116

technological and economic perspectives, respectively, ranging from narrow to broad. Although

117

they define them slightly differently with various subcases, there are five general substitution

118

types that reduce material inputs.

119

The first substitution type involves the direct substitution of certain materials for others, which

120

Tilton calls “material-for-material” substitution and includes Schlabach’s “physical” material

121

substitution. The second involves adopting an alternative production process that reduces the

122

quantity of one or more materials. Schlabach refers to this as “process” substitution, however,

123

Tilton’s “other-factors-for-material” substitution reflects the idea that it requires increases in

124

other inputs, such as labor or energy. The third involves altering the product itself to require less

125

material, either by reengineering its design (Schlabach) or reducing its quality (Tilton). The

126

fourth (and most general) involves meeting the product’s function in a different way, which

127

Schlabach and Tilton call “functional” and “inter-product” substitution, respectively, reflecting

128

the idea that an end-use is met by changing the mix of goods and services used to achieve it. The

129

last type involves an advance in technology (distinct from process and redesign substitutions) 6


Page 7 of 29


130

that reduces material requirements without increasing other inputs. Tilton calls this

131

“technological” substitution and includes Schlabach’s “quantitative” material substitution.

132

The subtle terminology differences for similar substitutions suggest that perceived disparities in

133

the ways economists and engineers view substitution are largely semantic. Accordingly, when

134

working with experts from different disciplines it is imperative to communicate effectively

135

without sacrificing precision. For disaggregated studies, then, it is necessary to explicitly define

136

substitution types along the relevant supply chain in a way that makes sense to both economists

137

and the experts.

138

Smith and Eggert31 define eight specific responses that PM and wind turbine manufacturers

139

could have had in response to the REE price spike, including six specific substitution types

140

resulting in the reduction of HREEs. Table 1 provides definitions and specific examples for each

141

response. The first two, cost pass-through and cost absorption, are not substitution at all. In both

142

cases, manufacturers essentially do nothing either because they can (low cost-share, high demand

143

for their product, etc.) or because they expect prices to fall. Focusing on PMs and wind turbines,

144

the six substitution types fall loosely into the categories above and range from specific (element-

145

for-element) to broad (system-for-system). Another potential response beyond the scope of this

146

study is the recycling and remanufacturing of magnets from end-of-life products, which would

147

increase material supply rather than reduce demand.41-43

148

Table 1: Response Categories modified from Smith and Eggert31

Response

Definition

Cost Pass-through

Pass increased material costs through to enduser by increasing sales price of final product

Cost Absorption Element-for-Element

Absorbing increased material costs by not increasing the sales price of final product Replace one element with another without

Example Increase magnet sales price by the same amount as materials cost increase Magnet sales price remains the same Substitute Tb for Dy in NdFeB

Other Literature --Material-for-material

7



Substitution

Process-for-Element Substitution

Grade-for-Grade Substitution Magnet-for-Magnet Substitution

significantly changing the properties of the magnet; does not include substitutions that result in a magnet with significantly different properties Reduce one or more elements by using an alternate production process to produce a magnet with the same properties; involves increasing other inputs, such as energy or labor Use a magnet (of the same type) with different properties in place of another; could result from system redesign or initial overspecification Use a different magnet type in place of the magnet in the original design in a reengineered system

Page 8 of 29

magnet of the same grade

subst.12; Physical material subst.13

Use grain boundary modification rather than powder metallurgy process to produce NdFeB magnet of the same grade

Technology-for-element subst.31; Other-factors-formaterial subst.12; Process subst.13; Reduced use41

Use NdFeB magnet with less temperature resistance in a direct drive generator

Quality-for-material subst.12; Design subst.13; Grade optimization4

Use ferrite instead of NdFeB magnets in a direct drive generator

Material-for-material subst.12; Design subst.13; Material subst.4 Functional subst.13,44; Interproduct subst.12; Invisible material subst.13; Technological subst.4

System-for-System Substitution

Meet the end-use in an entirely different way; most broad type of substitution

Use an induction generator with no PMs instead of a direct drive generator

Improved Manufacturing Efficiency

Using purchased materials more efficiently by reducing waste or reusing materials, in effect, substituting more efficient methods in place of materials

Reduce material waste by using thinner separating saws

Technological subst.12; Quantitative material subst.13

149

2.2 Survey Method

150

Expert consultation is a common approach for disaggregated material substitution studies.

151

Holmes45 argues for formal questionnaires performed via personal interviews with industry

152

experts when obtaining data for empirical investigations and that such data should be analyzed

153

through subjective and judgmental techniques. Eastin et al.15 and Eastin et al.16 also perform

154

formal surveys (via mail), asking experts to rate their level of (dis)agreement with several

155

statements about material use and the reasons behind any substitutions. They use t-tests to

156

analyze the results.

157

Other recent studies have also investigated HREE and PM substitution in wind turbines using

158

less-formal expert consultation.4,33,34 Smith and Eggert31 also provide an informal discussion of

159

REE substitution in PMs and wind turbines, based on several interviews pertaining to the

160

categories in Table 1. This paper offers a more formal approach and a significantly updated and

161

expanded discussion for a different audience. 8


Page 9 of 29


162

In the current study, a formal questionnaire was constructed to elicit expert views on the reaction

163

of PM and wind turbine manufacturers to the 2010/2011 REE price spike. The survey introduces

164

the project and explicitly defines the eight responses in Table 1 before asking experts to rate the

165

extent to which they observed each response on a four-point rating scale with the values 0 (“Not

166

at all”), 1 (“Somewhat”), 2 (“Moderate”), 3 (“Significant”), and “Unsure”. It then asks experts to

167

rate the extent each was driven by the price increases versus other factors. The remainder asks

168

specific qualitative and quantitative questions to better understand the answers to the initial

169

rating scale questions. These include before-and-after questions about the elemental

170

compositions of various NdFeB grades, which grades are used in wind turbines, the ability of

171

other magnet types to be used in turbines, and many others. The survey was emailed to

172

participants in advance and the questions were completed via phone interview.

173

As with any study involving surveys and interviews, there are biases that can affect the results.

174

Most notably, a possible lack of randomness exists in responses associated with consenting,

175

declining, and non-responsive experts. For this reason, requests were sent to many individuals

176

involved with different segments of the PM industry. A similar bias could arise from the choice

177

of whom to contact to participate in the study (convenience sampling). Interviewees were thus

178

asked to refer others who might partake in the study.

179

2.3 Profile of Experts

180

Experts were selected from the PM and wind turbine industries, academia, and U.S. national

181

laboratories. Most of the experts are materials scientists employed as magnet sales and

182

purchasing managers with many years of experience. The initial list included industry and

183

research partners with the Critical Materials Institute (an energy innovation hub funded by the 9



Page 10 of 29

184

U.S. Department of Energy) and contacts from industry conferences and was expanded by asking

185

each interview subject for additional references. The list was exhausted until all relevant

186

references were those who had already been interviewed, declined, or did not respond to three

187

email requests. Interviews were conducted over the phone and lasted between roughly 30

188

minutes and 2 hours, averaging about 1 hour.

189

There were 22 experts interviewed in total, 17 of whom were from the PM and electric motor

190

industries, 3 researchers from national laboratories, and 2 academic researchers. Of the industry

191

participants, 4 were magnetics industry consultants with previous experience in various segments

192

of the market, 6 were from sales departments of NdFeB manufacturing companies, 5 were from

193

NdFeB suppliers, and 2 were from research and development departments at major wind turbine

194

manufacturing companies. To ensure candid conversations with experts about their experiences

195

during the price spike and protect them from undue risk, their names and affiliations have been

196

kept confidential.

197

3 Results and Discussion

198

3.1 Relative Importance of Responses

199

Experts rated the extent to which PM and wind turbine manufacturers implemented the responses

200

defined in Table 1 between 2010 and 2016 on a four-point scale (0 to 3). Figure 1 shows the

201

medians and 90 percent confidence intervals of the distributions for each response. The experts

202

rated process-for-element substitution as the most significant response, followed by cost pass-

203

through,

grade-for-grade

substitution,

and

system-for-system

substitution.

Improved

10


Page 11 of 29


204

manufacturing efficiency, element-for-element substitution, and cost absorption all have medians

205

of 1, while magnet-for-magnet substitution happened the least or not at all.

Median and 90% confidence interval for response ratings 3

2

1

0 Process for Element

Cost Passthrough

System for System

Grade for Grade

Improved Element for Manufacturing Element Efficiency

Cost Aborption

Magnet for Magnet

206 207 208

Figure 1: Median and 90% confidence interval of ratings for responses to 2010/2011 rare-earth price increases;

209

Nonparametric statistical tests, such as the Sign test, provide several advantages over traditional

210

t-tests when dealing with discrete ordered (non-normal) data, particularly in cases with small

211

sample sizes.46 The Sign test examines whether two distributions have the same median and

212

makes no assumptions about their shapes, degrees of freedom, nor variance.46,47 Performing this

213

test for each pair of responses allows for the inference of a rank ordering. Figure 2 shows the

214

results from the paired Sign tests, where the shaded cells indicate statistically different responses

215

at different significance levels.

216

The experts consistently ranked process-for-element substitution as having occurred more than

217

every other response, although it is not statistically different from cost pass-through, grade-for-

218

grade, or system-for-system substitution, which make up a clear second grouping of responses at

219

the 0.10 significance level. Improved manufacturing efficiency is significantly different from

0 = “Not at all”, 1 = “Somewhat”, 2 = “Moderate”, 3 = “Significant”

11



Page 12 of 29

220

process-for-element substitution at the 0.05 significance level, placing it in its own group since it

221

also differs from cost absorption and magnet-for-magnet substitution. Element-for-element

222

substitution follows improved manufacturing efficiency and is significantly different from the

223

first four distributions at the 0.10 level. Last are cost absorption and magnet-for-magnet

224

substitution, which the experts clearly ranked as having occurred least. Further nonparametric

225

tests (with results provided in the SI) indicate the same ordering. Relative Ranking of Responses and Statistically Similar Distributions - Sign Test 90% Rank

Response

PE

1

Process-for-Element Subst. (PE)

2

Cost Pass-through (CPT)

2

Grade-for-Grade Subst. (GG)

2

System-for-System Subst. (SS)

5

Improved Manufacturing Efficiency (IME)

6

Element-for-Element Subst. (EE)

7

Cost Absorption (CA)

7

Magnet-for-Magnet Subst. (MM)

H0: The medians of each distribution are equal.

p > 0.10

CPT

GG

p < 0.10

SS

p < 0.05

IME

EE

CA

MM

p < 0.01

226

Figure 2: Results of pairwise Sign tests of response ratings for responses to 2010/2011 rare-earth price increases

227

While the sample size is small, it is large enough for the Sign tests to detect differences between

228

the most different responses at the 0.10 and 0.05 significance levels.47 Increasing the number of

229

experts would only improve the power of the tests, thereby increasing their ability to identify

230

differences between more similar responses.47,48 Thus, the differences depicted in Figure 2

231

would likely be even more significant, and significant differences would likely be estimated

232

between similar responses (such as between cost absorption and magnet-for-magnet

233

substitution).

12


Page 13 of 29


234

One- and two-sided Sign tests against discrete values indicate experts rated process-for-element

235

substitution as at least a moderate response (median >2) at the 0.10 significance level. Similarly,

236

there is insufficient evidence to reject that cost passthrough, grade-for-grade substitution, and

237

system-for-system substitution were moderate responses (=2). Improved manufacturing

238

efficiency is between somewhat and moderate (between 1 and 2, inclusive) at the 0.05

239

significance level. There is also insufficient evidence to reject that element-for-element

240

substitution, cost absorption, and magnet-for-magnet substitution are somewhat (=1), and

241

insufficient evidence to indicate that magnet-for-magnet substitution is less than somewhat.

242

It is important to note that it is possible, and indeed probable, that several responses occurred

243

simultaneously. One limitation of the relative rankings is that they are unable to test the degree to

244

which responses such as process-for-element and grade-for-grade substitution occurred at the

245

same time. To inform this process, the next section discusses specific findings related to the

246

timing and implementation of responses (with further details provided in the SI).

247

3.2 Timing and Technological Change

248

Substitution responses are not instantaneous and require varying amounts of time to be

249

implemented, often several years.4 The eight responses considered in this study are clearly

250

differentiated by time-scale as well as the amount of technological innovation required. Doing

251

nothing and passing costs through was the easiest and immediate response, but was not tenable

252

longer term. Substitutions involving the adoption of pre-existing processes or methods already

253

under development occurred rather quickly, whereas more significant improvements took longer.

254

Much of what happened between 2011 and 2016 was technological adoption and incremental

255

improvements rather than innovation—but not entirely. Since the price spike was relatively 13



Page 14 of 29

256

short-lived (lasting about 18 months), many experts asserted that there was simply not enough

257

time for certain responses to take place. Table 2 summarizes the key findings for each response

258

type, in order of ranking as discussed in Section 3.1. Most discussion of findings pertaining to

259

the PM industry are for the industry generally, except where noted for PMs in wind turbines

260

specifically.

261

Table 2: Summary of Findings

90% Rank

Response

Primary Industry

1

Process-forElement Subst.

PM

2

Cost Passthrough

PM

2

Grade-for-Grade Subst.

Wind

2

System-forSystem Subst.

Wind

5

Improved Manufacturing Efficiency

PM

6

Element-forElement Subst.

PM

7

Cost Absorption

PM

7

Magnet-forMagnet Subst.

Wind

Key Findings PM manufacturers able to reduce HREE content of NdFeB by 40-50% via grain boundary modification and dual alloy production processes. PM manufactures passed through high REE costs as much as possible; led to prominence of metal clauses in client contracts. Many turbine producers 1) lowered magnet grade specifications due to initial over-specification or 2) redesigned their generator systems to accommodate lower-grade magnets. Price spike 1) impeded adoption of PMG turbine systems for onshore applications and 2) contributed to adoption/innovation of new drive systems with less PM material. PM manufacturers implemented waste-reduction techniques such as near net-shape manufacturing and using thinner separation saws to reduce amount of swarf created in finishing process; allowed producers to improve material yield. Increase in Pr content of NdFeB through use of unseparated NdPr (didymium); limited substitution between HREEs (Dy and Tb). Limited ability of PM manufacturers to absorb costs; absorbed costs as mandated in existing metal clauses in client contracts. Not currently viable option for wind turbine manufacturers.

262

3.2.1 Immediate Response—No Substitution

263

The immediate response to the REE price spike by PM manufacturers (in general) was to pass

264

costs through as much as possible without making changes to the production process. Most

265

experts ascertained that there was little choice involved, since most producers were already

266

producing near the margin and Dy accounted for as much as 95 percent of the total magnet cost

267

in some cases. One expert estimated that prices of some NdFeB grades increased by over 600

268

percent between January 2010 and August 2011.31 Many experts noted that magnet producers 14


Page 15 of 29


269

absorbed costs as much as possible and passed through the remainder, however, absorbing costs

270

was unrealistic for most producers (especially those with smaller volume). Passing costs through

271

had disadvantages, particularly in the early stages of the price spike before customers were aware

272

of the REE supply situation. After several months, magnet producers were forced to consider

273

other options as they faced pressure to keep magnet prices low.

274

Longer term, magnet producers negotiated agreements within existing quarterly contracts called

275

“metal clauses”, which allowed the negotiated magnet sales price to vary for specified deviations

276

in material costs and exchange rates. Producers absorbed increased material costs for increases

277

less than the specified amount, but could pass additional costs over the specified amount to the

278

customer. Consequently, magnet producers both absorbed and passed costs through over the

279

duration of an existing contract.

280

3.2.2 Delayed Response—Changes in Technology

281

Changes in technology were behind significant reductions in the HREE content of both magnets

282

and wind turbines following the price spike. The extent to which the price spike influenced these

283

changes and the speed with which they were implemented varies by the level of technological

284

change.

285

3.2.2.1 Technological Adoption

286

Most of the substitutions that took place during and following the price spike were those for

287

which (approximate) solutions already existed. Both PM and wind turbine manufacturers

288

adopted (rather than innovated) alternative manufacturing processes and product specifications

289

fairly quickly. 15



Page 16 of 29

290

Process-for-element substitution was the most significant response to the REE price spike by PM

291

manufacturers due to two known production processes that reduced the HREE content of

292

magnets with high temperature resistance by 40 to 50 percent. The first process involves

293

improving the microstructure of the magnet by modifying the grain boundaries, namely through

294

the grain boundary diffusion (GBD) process, which allows manufacturers to microscopically

295

place the HREEs around individual grains rather than dispersing it throughout the magnet.49 The

296

second process is “dual-alloying” technology, which consists of using unseparated ferro-

297

dysprosium alloy (DyFe) to economize on Dy rather than inserting it separately.50 A third,

298

though less prevalent, process, called powder refinement technology, consists of using finer

299

grain sizes in the magnets.37

300

Most experts pointed to the adoption of the GBD process as the magnet industry’s main response

301

to the price spike, representing a major change from traditional sintering with the powder

302

metallurgy method. They noted that although the GBD process for HREEs was known prior to

303

the price spike (having been developed by Sumitomo Special Metals, now Hitachi Metals, in

304

2002)51, it is time-intensive and requires expensive equipment and labor, so it did not see

305

widespread adoption until material cost increases forced producers to economize on Dy.

306

According to several experts, the major Japanese producers hold grain boundary modification

307

patents, licensing their processes to a select group of companies worldwide.51-53 Despite the

308

relatively small number of licensed companies, one expert asserted that “more than 75 percent”

309

of the industry now uses the GBD process. Similarly, an expert from a major Chinese

310

manufacturer indicated that introducing the dual-alloying process allowed their company to

16


Page 17 of 29


311

“remove Dy from half their grades”, while another indicated that almost all companies now

312

purchase pre-alloy metals.

313

Despite consistently rating element-for-element substitution as less important, the experts

314

mentioned a few important developments. The most notable elemental substitution in NdFeB

315

magnets was the increased use of didymium (unseparated NdPr), which reduces the Nd content

316

while increasing the Pr content.41 Some companies used didymium prior to the price spike (circa

317

2007), but in 2011 the practice became widespread. The experts estimated that the Pr content

318

range in a typical magnet was 0 to 7 percent before the price spike and 3 to 8 percent after.

319

Figure 3 shows ranges for estimates of the REE content of NdFeB by element in 2010 and 2016.

REE content ranges of SH grade NdFeB, 2010 vs. 2016 35%

Percent by Weight

30% 25% 20% 15% 10% 5% 0% HREE

320

Pr

Nd

HREE

2010

Pr

Nd

2016

321 322

Figure 3: Nd, Pr, and HREE content range of SH grade NdFeB in the N40-N45 range, 2010 vs. 2016 (Note: horizontal lines represent maximum, median, and minimum; based on responses from 6 experts)

323

Substitution between HREEs is also possible to achieve the same temperature properties. Tb can

324

completely substitute for Dy and, according to several experts, is more effective.31 The experts

325

mentioned that some companies may have tried to use Tb in the short run, however, Tb prices

326

are kept high due to its use in phosphors.31,34 One expert noted that some NdFeB magnets 17



Page 18 of 29

327

contain both elements, estimating that Tb decreased relative to Dy as the total HREE content

328

decreased.

329

System-for-system substitution was the simplest solution for wind turbine manufacturers and

330

project planners, especially for onshore systems. More accurately, the price spike impeded the

331

widespread adoption of PMG systems since DFIGs already dominated the market.34,38 Several

332

experts mentioned industry decisions to pursue DFIG designs, such as General Electric’s

333

decision in 2012 to quit using PMGs and return to DFIG for their onshore turbine designs

334

(although magnet cost was a single factor of this decision).54 According to Adamas

335

Intelligence,39 total REE consumption for wind turbines fell in 2013, with the industry’s

336

consumption of Dy oxide falling by more than that for Nd and Pr.

337

3.2.2.2 Incremental Technological Improvement

338

Actual technological improvements also occurred in response to the REE price spike through

339

continued innovation, though they were less immediate, and, as several experts mentioned, might

340

have been more significant had prices remained high for a longer period. In economics,

341

technological (or “technical”) improvement strictly implies the ability to produce the same

342

output using less of an input without increasing other inputs, as with Tilton’s view of

343

technological substitution.23

344

Technical improvements via improved manufacturing efficiency at least somewhat occurred after

345

the price spike, though many experts pointed out that well-managed firms continuously improve

346

their processes regardless of material costs. Still, others thought the price spike hastened the

18


Page 19 of 29


347

speed of improvements. In the words of one expert, “everything that could be done was done” to

348

economize on expensive REE components.

349

There are two ways to use materials more efficiently: either reduce material waste, or reuse scrap

350

material from the manufacturing process. According to the experts, most improvements were

351

accomplished through waste reduction, though the degree of improvement varies by magnet size.

352

Using slimmer separating saws reduced waste for thinner magnets and, for larger magnets (such

353

as those used in wind turbines), manufacturing the magnet closer to its final shape reduced the

354

amount of grinding necessary. One expert noted that waste reduction techniques allowed

355

producers to reduce total purchased REE content in a typical magnet by about 5 percent. Another

356

stated that since 2011 Chinese companies reduced the separation saw size from about 1mm to

357

0.4mm, while Japanese companies pursued near net-shape manufacturing. Because large

358

magnets tend to be manufactured near their final dimensions, one expert noted that material yield

359

is over 95 percent. More significant improvements were made in smaller magnets, however, as

360

experts noted yields that were between 30 and 60 percent are now 70 to 85 percent with near net-

361

shape manufacturing. Reusing material in the production process is difficult and less common.

362

There was also some improvement for wind turbine manufacturers. Experts indicated that some

363

companies responded in the short run via grade-for-grade substitution where magnets were over-

364

specified (Sprecher et al.4 call this “grade optimization”). Once some of the larger manufacturers

365

realized their systems were not optimized for cost, they substituted magnets with reduced HREE

366

content. One expert also mentioned that some companies switched to less expensive, lower-

367

quality versions of the same magnet grade.

19



Page 20 of 29

368

3.2.2.3 Technological Innovation

369

Despite the brevity of the REE price spike, a small number of actual technological breakthroughs

370

reduced HREE use in wind turbines. One notable innovation occurred via grade-for-grade

371

substitution in 2014, about 3 years after the price spike. Many experts pointed to Siemens, which

372

developed a system to cool the PMG, allowing for magnets with reduced HREE contents.31,55

373

These efforts have led some to forecast the average NdFeB alloy in a PMG turbine will contain

374

just 3 percent HREE in 2020, down from an estimated 4.5 percent in 201439 (though this change

375

would likely occur due to the combination of responses discussed in this study). One expert also

376

mentioned that some companies experimented with magnet shapes to isolate different parts of

377

the magnet from demagnetization, which also allowed them to use lower-grade magnets.

378

There were also some longer-term innovations via system-for-system substitution. Experts

379

mentioned that manufacturers sought alternatives to PMGs and DFIGs for both onshore and

380

offshore turbines, but that most solutions were less efficient. These experts stated that

381

manufacturers have been returning to PMGs in recent years since REE prices have fallen. There

382

are also designs that substantially reduced the amount of REEs, such as Enercon’s low-speed

383

direct drive generator with no PMs34 and mid-speed “hybrid” generators by manufacturers like

384

Gamesa and Vestas, which contain smaller PMGs coupled with gearboxes.

385

While the price spike caused magnet-for-magnet substitution in other applications,4 the experts

386

ranked it as the least significant response for wind turbine manufacturers and most indicated it

387

was not feasible. The major advantage of NdFeB is its high energy product, implying it takes less

388

material to generate a comparable magnetic field. A system with less powerful magnets

389

necessarily uses more magnetic material, leading to larger (and more expensive) systems overall. 20


Page 21 of 29


390

One expert noted that switching to less powerful magnets was not worth investigating since it

391

would require years of additional testing and REE prices declined within a fairly short period.

392

3.3 Comparisons to Other Literature

393

Except for improved manufacturing efficiency and grade re-optimization, the experts indicated

394

most responses occurred directly because of the price spike. The price spike caused substitutions

395

fairly quickly via technological adoptions, providing a refinement to Tilton’s23 alternative view

396

of material substitution that occurs through technological improvement and innovation in the

397

long run. This is indeed the case for process-for-element substitution, where manufacturers

398

switched to two previously-known processes that were cost-prohibitive prior to the price spike.

399

Holmes similarly finds that technological adoption was important for substitutions in battery

400

electrodes21 and windows22 in the 1980s, but notes that the nature of final demand and technical

401

attributes of new materials were more significant drivers than material prices.

402

As both Holmes and Tilton point out, introducing new technology raises an important point

403

about the relationship between substitution and material prices. It is generally assumed that

404

demand is continuous and reversible, implying that if prices decline the quantity demanded

405

returns to its previous level. Changing production processes, however, can permanently shift the

406

demand curve inwards for a given material. The findings from this study indicate that demand

407

for HREEs in magnets is not continuous (material costs changed by thousands of percentage

408

points to induce a discrete 40 to 50 percent change), and possibly not reversible. REE prices have

409

since fallen and producers continue to use technology implemented during and following the

410

price spike. Were REE prices to fall substantially, there might be a return to previous technology.

21



Page 22 of 29

411

While this study differs in purpose and scope from Graedel et al.’s9 and Nassar’s10 work in

412

quantifying elemental substitution, it finds it is at least somewhat possible in NdFeB magnets.

413

Specifically, substitution of different HREEs is possible (though impractical since Tb is several

414

times more expensive than Dy), as is substitution of non-REE components of the magnet.56

415

Further, the ability to introduce unseparated metal (in the cases of both NdPr and DyFe) also

416

alters the elemental composition and lowers material costs. This study does, however,

417

corroborate Nassar’s assertion that substitution is more likely in other areas than at the elemental

418

level.

419

This study also corroborates Sprecher et al.’s4 claim that there is an important time dimension to

420

substitution. Namely, substitutions that involved existing solutions were implemented far more

421

quickly than those that involved greater effort. This is particularly noteworthy with the

422

widespread adoption of both the GBD and dual-alloy production processes over the five years

423

following the price shock. Responses which required more innovation either experienced a

424

longer delay or did not occur at all due to the relative brevity of the price spike.

425

Each of the responses in this study have different environmental implications. There are three

426

perspectives to consider when alleviating the dependence of wind turbines on HREEs: first,

427

improvement in the ability of electric power producers to curtail emissions from carbon dioxide

428

and other pollutants via offshore windfarms; second, a possible reduction in overall material

429

demand of wind turbines as PMGs weigh less with fewer moving parts;57 and third, a possible

430

reduction in harmful environmental impacts from primary REE production (in the case of

431

system-for-system substitution or if there is future material supply from magnet

22


Page 23 of 29


432

remanufacturing). The first two points are most relevant to this study, however, several other

433

studies have performed life cycle sustainability assessments of NdFeB magnets.58-60

434

While other studies4,33,34,61 have analyzed potential substitution strategies for PMs in wind

435

turbines and electric vehicles, they tend to be from engineering perspectives and are less

436

interested in the influence of material costs on firm behavior. This study is explicit in its focus on

437

how PM and turbine manufacturers actually responded to rampant material price increases. It

438

also provides general insight into how industries facing substantial increases or uncertainty in

439

material costs might respond in the shorter term. In the end, it appears the actual impact of the

440

price spike was improvements in the way materials and technology are used.

441

3.4 Limitations and Future Research

442

There are limitations to this study. Most notable are the relatively small number of observations

443

and the biases inherent in the survey interview. The conclusions are drawn from the collective

444

opinion and knowledge of industry experts, rather than actual price-quantity data (which does not

445

exist). Still, the method employed provides an advantage over statistical regression models in its

446

recognition of the actual ways manufacturers respond to material price increases, rather than

447

simply providing a numerical estimation of this relationship.

448

While the findings indicate demand for HREEs in PMs is price inelastic and non-zero (between 0

449

and -1), it is difficult to ascertain how much so. There was an observed increase in material costs

450

of about 3000 percent and an accompanying decrease in HREE content by many manufacturers

451

by 40 to 50 percent, yet it is unknown at what price level PM manufacturers would have initiated

452

this change. Demand is inelastic for any price increase over 100 percent (by definition), so the 23



Page 24 of 29

453

true estimate is likely closer to 0 than -1. It also appears demand shifts in discrete jumps as

454

production technology changes. Furthermore, demand for HREEs in wind turbines is highly

455

elastic at high prices, since low-HREE technologies already exists.

456

Finally, this study assumes responses to the REE price spike were in fact linked to material costs.

457

While reasonable, geopolitical factors (perhaps induced by the spike) arising from the

458

geographic concentration of REE production may have influenced manufacturers in certain

459

countries.

460

4 Associated Content

461

4.1 Supporting Information

462

Provides substantial additional information and figures pertaining to the eight responses in Table

463

1, summary statistics tables, and robustness checks.

464

5 Author Information

465

5.1 Corresponding Author

466

*E-mail: [email protected]

467

5.2 ORCID

468

Braeton Smith: 0000-0002-0667-6007

469

5.3 Notes

470

The authors declare no competing financial interest. 24


Page 25 of 29


471

6 Acknowledgment

472

The authors thank the experts who consented to be interviewed. This work is supported by the

473

Critical Materials Institute, an Energy Innovation Hub funded by the U.S. Department of Energy,

474

Office of Energy Efficiency and Renewable Energy, Advanced Manufacturing Office.

475

7 References

476 477 478 479

(1)

United States Department of Energy, 20% Wind Energy by 2030: Increasing Wind Energy’s Contribution to U.S. Electricity Supply; Office of Energy Efficiency and Renewable Energy, U.S. Department of Energy, 2008. https://www.nrel.gov/docs/fy08osti /41869.pdf (accessed Feb 27, 2018)

480 481 482

(2)

United States Department of Energy, Critical Materials Strategy; United States Department of Energy, 2011. https://energy.gov/sites/prod/files/DOE_CMS2011_FINAL_Full.pdf (accessed Feb 27, 2018).

483 484

(3)

Erdmann, L.; Graedel, T. E. Criticality of Non-Fuel Minerals: A Review of Major Approaches and Analyses. Environmental Science & Technology 2011, 45, 7620–7630.

485 486 487 488

(4)

Sprecher, B.; Daigo, I.; Spekkink, W.; Vos, M.; Kleijn, R.; Murakami, S.; Kramer, G. J. Novel Indicators for the Quantification of Resilience in Critical Material Supply Chains, with a 2010 Rare Earth Crisis Case Study. Environmental Science & Technology 2017, 51, 3860–3870.

489 490

(5)

Lu, Y.; Stern, D. I. Substitutability and the Cost of Climate Mitigation Policy. Environmental and Resource Economics 2016, 64, 81–107.

491 492

(6)

Geyer, R. Parametric Assessment of Climate Change Impacts of Automotive Material Substitution. Environmental Science & Technology 2008, 42, 6973–6979.

493 494

(7)

Beckman, J.; Hertel, T.; Tyner, W. Validating energy-oriented CGE models. Energy Economics 2011, 33, 799 – 806.

495 496 497 498

(8)

Jorgenson, D. W.; Goettle, R. J.; Wilcoxen, P. J.; Ho, M. S. The role of substitution in understanding the costs of climate change policy; Pew Center on Global Climate Change, 2000. https://www.c2es.org/document/the-role-of-substitution-in-understanding-the-costsof-climate-change-policy/ (accessed Feb 27, 2018).

499 500

(9)

Graedel, T. E.; Harper, E. M.; Nassar, N. T.; Reck, B. K. On the materials basis of modern society. Proceedings of the National Academy of Sciences 2015, 112, 6295–6300. 25



Page 26 of 29

501 502

(10) Nassar, N. T. Limitations to elemental substitution as exemplified by the platinum-group metals. Green Chemistry 2015, 17, 2226–2235.

503 504 505

(11) Tilton, J. E. In Material Substitution: Lessons from the Tin-Using Industries; Tilton, J. E., Ed.; Resources for the Future, Inc., 1983; Chapter Material Substitution: Lessons from the Tin-Using Industries, pp 1–11.

506

(12) Tilton, J. E. Substitution: Economics. Conservation & Recycling 1984, 7, 21 – 26.

507

(13) Schlabach, T. Substitution: Technology. Conservation & Recycling 1984, 7, 15 – 20.

508 509

(14) Slade, M. E. Recent advances in econometric estimation of materials substitution. Resources Policy 1981, 7, 103 – 109.

510 511

(15) Eastin, I. L.; Shook, S. R.; Simon, D. Softwood lumber substitution in the U.S. residential construction industry in 1994. Forest Products Journal 1999, 49, 21–27.

512 513

(16) Eastin, I. L.; Shook, S. R.; Fleishman, S. J. Material substitution in the U.S. residential construction industry, 1994 versus 1998. Forest Products Journal 2001, 51, 30–37.

514 515 516

(17) Shook, S. R.; Soria, J. A.; Nalle, D. J. Examination of North American softwood lumber species substitution using discrete choice preferences and disaggregated end-use markets. Canadian Journal of Forest Research 2007, 37, 2521–2540.

517 518

(18) Messner, F. Material substitution and path dependence: empirical evidence on the substitution of copper for aluminum. Ecological Economics 2002, 42, 259 – 271.

519 520

(19) Demler, F. R. In Material Substitution: Lessons from the Tin-Using Industries; Tilton, J. E., Ed.; Resources for the Future, Inc., 1983; Chapter Beverage Containers, pp 15–35.

521 522

(20) Canavan, P. D. In Material Substitution: Lessons from the Tin-Using Industries; Tilton, J. E., Ed.; Resources for the Future, Inc., 1983; Chapter Solder, pp 36–75.

523

(21) Holmes, M. Material substitution in battery electrodes. Resources Policy 1990, 16, 22 – 34.

524 525

(22) Holmes, M. Material substitution in the UK window industry, 1983-1987. Resources Policy 1990, 16, 128 – 142.

526 527

(23) Tilton, J. E. Material Substitution: The Role of New Technology. Technological Forecasting and Social Change 1991, 39, 127 – 144.

528 529

(24) Popp, D. Induced Innovation and Energy Prices. The American Economic Review 2002, 92, 160–180.

530 531

(25) Goulder, L. H.; Schneider, S. H. Induced technological change and the attractiveness of CO2 abatement policies. Resource and Energy Economics 1999, 21, 211 – 253.

532 533

(26) Schneider, S. H.; Goulder, L. H. Achieving low-cost emissions targets. Nature 1997, 389, 13–14. 26


Page 27 of 29


534 535

(27) Doraszelski, U. Innovations, improvements, and the optimal adoption of new technologies. Journal of Economic Dynamics and Control 2004, 28, 1461 – 1480.

536 537

(28) Balcer, Y.; Lippman, S. A. Technological expectations and adoption of improved technology. Journal of Economic Theory 1984, 34, 292 – 318.

538 539

(29) Farzin, Y.; Huisman, K.; Kort, P. Optimal timing of technology adoption. Journal of Economic Dynamics and Control 1998, 22, 779 – 799.

540 541 542

(30) Popp, D. Exploring Links Between Innovation and Diffusion: Adoption of NOX Control Technologies at US Coal-fired Power Plants. Environmental and Resource Economics 2010, 45, 319–352.

543 544

(31) Smith, B. J.; Eggert, R. G. Multifaceted Material Substitution: The Case of NdFeB Magnets, 2010–2015. JOM 2016, 68, 1964–1971.

545 546 547

(32) 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. Environmental Science & Technology 2015, 49, 6740–6750.

548 549 550 551 552

(33) Erdmann, L.; Bette, K.; Merino, J. M.; Velte, D. Roadmap for the Substitution of Critical Raw Materials in Electric Motors and Drives; Critical Raw Materials Innovation Network (CRM_InnoNet), 2015. http://www.isi.fraunhofer.de/isi-wAssets/docs/v/en/publikationen /CRM-InnoNet-Roadmap-for-CRM-substitution_Electric_Motors_And_Drives.pdf (accessed Feb 27, 2018).

553 554 555

(34) Pavel, C. C.; Lacal-Arantegui, R.; Marmier, A.; Schuler, D.; Tzimas, E.; Buchert, M.; Jenseit, W.; Blagoeva, D. Substitution strategies for reducing the use of rare earths in wind turbines. Resources Policy 2017, 52, 349 – 357.

556 557 558

(35) Gambogi, J. 2014 Minerals Yearbook; United States Geological Survey, 2014; Chapter Rare Earths, pp 60.1–60.16. https://minerals.usgs.gov/minerals/pubs/commodity /rare_earths/myb1-2014-raree.pdf (accessed Feb 27, 2018).

559 560 561

(36) Cordier, D. J. Mineral Commodity Summaries 2011; United States Geological Survey, 2011; Chapter Rare Earths, pp 128–129. https://minerals.usgs.gov/minerals/pubs /commodity/rare_earths/mcs-2011-raree.pdf (accessed Feb 27, 2018).

562 563

(37) Brown, D. N. Fabrication, Processing Technologies, and New Advances for RE-Fe-B Magnets. IEEE Transactions on Magnetics 2016, 52, 1–9.

564 565

(38) Lacal-Arantegui, R. Materials use in electricity generators in wind turbines – state-of-theart and future specifications. Journal of Cleaner Production 2015, 87, 275 – 283.

566 567

(39) Adamas Intelligence, Rare Earth Market Outlook: Supply, Demand, and Pricing from 2014-2020; Adamas Intelligence, 2014.

27



Page 28 of 29

568 569 570

(40) Schlossberg, T. America’s First Offshore Wind Farm Spins to Life. New York Times [Online], Dec 14, 2016. https://www.nytimes.com/2016/12/14/science/wind-power-blockisland.html (accessed Feb 27, 2018).

571 572

(41) Binnemans, K.; Jones, P. T. Rare Earths and the Balance Problem. Journal of Sustainable Metallurgy 2015, 1, 29–38.

573 574 575 576

(42) Yang, Y.; Walton, A.; Sheridan, R.; Güth, K.; Gauß, R.; Gutfleisch, O.; Buchert, M.; Steenari, B.-M.; Van Gerven, T.; Jones, P. T.; Binnemans, K. REE Recovery from End-ofLife NdFeB Permanent Magnet Scrap: A Critical Review. Journal of Sustainable Metallurgy 2017, 3, 122–149.

577 578 579

(43) Rademaker, J. H.; Kleijn, R.; Yang, Y. Recycling as a Strategy against Rare Earth Element Criticality: A Systemic Evaluation of the Potential Yield of NdFeB Magnet Recycling. Environmental Science & Technology 2013, 47, 10129–10136.

580 581

(44) Chynoweth, A. Electronic Materials: Functional Substitutions. Science 1976, 191, 724– 732.

582 583

(45) Holmes, M. Input prices and material substitution: Choice of methodology for a disaggregated study. Resources Policy 1988, 14, 112 – 120.

584 585

(46) Montgomery, D. C.; Runger, G. C. Applied Statistics and Probability for Engineers, Fourth Edition, 4th ed.; John Wiley & Sons, Inc., 2007.

586 587

(47) Dixon, W.; Mood, A. The statistical sign test. Journal of the American Statistical Association 1946, 41, 557–566.

588 589

(48) Stewart, W. M. A Note on the Power of the Sign Test. The Annals of Mathematical Statistics 1941, 12, 236–239.

590 591 592 593

(49) Rashidi, S.; Moon, T.; Huan, Z. K. S. Advances in Manufacture of Low or No Heavy Rare Earths NdFeB Magnets. Magnetics Business & Technology [Online], Summer 2016, 14– 16. http://www.nxtbook.com/nxtbooks/webcom/magnetics_2016summer/index.php#/14 (accessed Feb 27, 2018).

594 595 596

(50) Lin, C.; Guo, S.; Fu, W.; Chen, R.; Lee, D.; Yan, A. Dysprosium Diffusion Behavior and Microstructure Modification in Sintered Nd-Fe-B Magnets via Dual-Alloy Method. IEEE Transactions on Magnetics 2013, 49, 3233–3236.

597 598

(51) Okayama, K.; Ishigaki, N.; Okumura, S. Rare earth magnet and method for manufacturing the same. 2002; https://www.google.com/patents/US6491765, US Patent 6,491,765.

599 600

(52) Nakamura, H.; Hirota, K.; Shimao, M.; Minowa, T. Rare earth permanent magnet. 2009; https://encrypted.google.com/patents/US7488393, US Patent 7,488,393.

601 602 603

(53) Benecki, W. T. Hitachi Metals, Ltd. - The Magnet Industry Newsmaker. Magnetics Business & Technology [Online], Winter 2013, 6–10. http://www.nxtbook.com/nxtbooks /webcom/magnetics_2013winter/index.php#/6 (accessed Feb 27, 2018). 28


Page 29 of 29


604 605 606

(54) Quilter, J. After 10 years GE goes back to DFIGs. Wind Power Monthly [Online], Oct 9, 2012. http://www.windpowermonthly.com/article/1153928/10-years-ge-goes-back-dfigs (accessed Feb 27, 2018).

607 608 609 610

(55) Siemens, New Siemens D3 Wind Turbine for High Energy Yields at Low Wind Speeds [Press Release], Oct 29, 2014. http://www.siemens.ca/web/portal/en/Press-Archive/2014 /Documents/Press%20Release%20-%20SWT-3%203-130_Canwea_F_Oct29 _docx%20(3).pdf (accessed Feb 27, 2018).

611 612

(56) Kurniawan, M.; Keylin, V.; McHenry, M. E. Alloy substituents for cost reduction in soft magnetic materials. Journal of Materials Research 2015, 30, 1072–1077.

613 614

(57) Venas, C. Life cycle assessment of electric power generation by wind turbines containing rare earth magnets. M.Sc. thesis, Norwegian University of Science and Technology, 2015.

615 616 617 618

(58) Sprecher, B.; Xiao, Y.; Walton, A.; Speight, J.; Harris, R.; Kleijn, R.; Visser, G.; Kramer, G. J. Life Cycle Inventory of the Production of Rare Earths and the Subsequent Production of NdFeB Rare Earth Permanent Magnets. Environmental Science & Technology 2014, 48, 3951–3958.

619 620 621

(59) Jin, H.; Afiuny, P.; McIntyre, T.; Yih, Y.; Sutherland, J. W. Comparative Life Cycle Assessment of NdFeB Magnets: Virgin Production versus Magnet-to-Magnet Recycling. Procedia CIRP 2016, 48, 45 – 50, The 23rd CIRP Conference on Life Cycle Engineering.

622 623 624

(60) Wulf, C.; Zapp, P.; Schreiber, A.; Marx, J.; Schlor, H. Lessons Learned from a Life Cycle Sustainability Assessment of Rare Earth Permanent Magnets. Journal of Industrial Ecology 2017, 21, 1578–1590.

625 626 627

(61) Pavel, C. C.; Thiel, C.; Degreif, S.; Blagoeva, D.; Buchert, M.; Schuler, D.; Tzimas, E. Role of substitution in mitigating the supply pressure of rare earths in electric road transport applications. Sustainable Materials and Technologies 2017, 12, 62 – 72.

29


Costs, Substitution, and Material Use: The Case of Rare Earth Magnets

Recommend Documents