C H A P T E R

4

Recycling Rare MetalsRobert U. Ayres1, Gara Villalba Mendez2, Laura Talens Peiro1

1INSEAD, Fontainebleau, France; 2Universitat Autonoma de Barcelona, Spain

4.1 INTRODUCTION

The application of material flow analysis(MFA) to trace metals throughout their life cy-cle is an exercise for quantifying the magnitudeof the uses and losses and identifying wherethey occur. Furthermore, by distinguishing be-tween losses that are dissipative versus lossesthat are recoverable, the MFA can help usassess the potential for recycling of the mate-rial. This is a very practical analysis for geolog-ically scarce metals used in new technologyproducts such as smartphones, wind turbines,and solar panels. These metals e.g. gallium,germanium, indium, tellurium, tantalum, andplatinum group metals (PGMs), are of specialconcern in terms of geological scarcity. Theyalso pose a problem for recycling due to thetrend toward miniaturization in their uses.Many of these metals are not found anywherein high concentrations but are distributed ascontaminants of other “attractor” metals towhich they are chemically similar. For thisreason, we call these metals “hitchhikers”when they accompany “attractors”. Forexample, molybdenum, rhenium, selenium, sil-ver, and tellurium are hitchhiker metals theproduction of which depends to a large extent

Handbook of Recycling

http://dx.doi.org/10.1016/B978-0-12-396459-5.00004-0 27

on the mining and smelting of copper (TalensPeiro et al., 2013). Losses occur initially in theproduction process that, for the case of thehitchhiker metals, includes the mining ofthe “attractor” ore, mineral processing, andfurther extraction, separation, and refining pro-cesses. A very simple calculation can be per-formed to estimate the theoretical amount ofthe hitchhiker metal that could be extractedbased on the quantity of the metal present inthe ore. Even though the potential metal pro-duction estimated this way does not considertechnological limitations and unavoidablelosses, it serves to identify which metals offeropportunities for improved resource manage-ment and which ones are already beingmanaged efficiently. We have performed thesecalculations for hitchhiker metals for whichone might expect revolutionary demand dueto their use in new technologies. We presentthe results in Table 4.1.

Metals for which potential mine productionis comparable to current (2010) output do nothave much margin for increasing recovery dur-ing the production phase. For example, cobaltfrom nickel ores is extracted efficiently (99%),whereas recovery from copper sulfide oressuch as carrolite is less efficient (8%) resulting

Copyright � 2014 Elsevier Inc. All rights reserved.

TABLE 4.1 Potential and Actual Production of Hitchhiker Metals in 2010

“Attractor” Metals or Mineral Ores “Hitchhiker” Metals

Name

Reserves

(106 t)

Production

(106 t) Name

Reserves

(103 t)

Current Mine

Production (t)

Potential Mine

Production (t)

Recovery

Efficiency (%)

Iron ore 87,000 2400 Rare earthoxidesa

110,000 54,000 4,114,280 1

Niobiuma 2900 63,000 89,140 71

Bauxite 28,000 211.00 Galliumb n.a. 106 10,550 1

Copper 630 16.20 Cobaltc 7300 31,000 408,800 8

Rheniumd 2.5 46 9370 <1

Molybdenumd 9800 133,000 281,050 47

Telluriume 22 475 1050 45

Seleniume 88 3250 4210 77

Zinc 250 12.00 Germaniume 0.45 84 597 14

Indiumf n.a. 574 1454 39

Galliume n.a. e 420 e

Nickel 76 1.55 Cobaltg 7300 44,000 44,600 99

PGMsd 66 11 17 63

Tin 5.2 0.26 Niobiumd 2900 2 373 1

Tantalumd 110 102 746 14

Sources: (a) Drew et al. (1991), (b) Jaskula (2011), (c) Shedd (2011), (d) Habashi (1997), (e) USGS (1973), (f) Tolcin (2011), (g) Berger et al. (2011).

4. RECYCLING RARE METALS28

in losses of up to 378,000 t of cobalt. The tablealso serves to point out a potential source forgallium from zinc (420 t) that is presently not be-ing exploited (Berger et al., 2011). On the otherhand, the PGMs selenium and molybdenumare currently being produced near their minepotential from current sources. These metalsare correspondingly important to recover andrecycle to meet future demand. For the rest ofthe metals listed in Table 4.1, a significant frac-tion of the potential is ending up as waste.

Similarly, the subsequent life cycle stageshave unnecessary losses and consequently op-portunities for recovery. In the next few pages,we illustrate how material flow analysis is usedto quantify the losses and recycling potential

I. RECYCLING I

during the whole life cycle of several criticalmetals: indium, europium, gallium, and plat-inum group metals (PGMs).

4.2 INDIUM

In our economy, we use the life cycle of in-dium to illustrate howmetal losses can be quan-tified from mining to end-of-life (EOL) based onconsumption data for 2010. Based on knowl-edge of the processes and end-uses of indium,we differentiate between dissipative and recov-erable fractions. The analysis is illustrated inFigure 4.1, which indicates the losses of indiumduring its life cycle in the present situation

N CONTEXT

FIGURE 4.1 Indium losses duringits life cycle (2010) in metric tonnes.

4.2 INDIUM 29

(2010) and the potentially recoverable fractionmarked by a more sustainable practice repre-sented by the dotted line.

Indium is of special interest for several rea-sons. It is classified as a “critical metal” bynumerous reports because it is a scarce metalthat plays an important role in the solar energyindustry, in flat screens, and in other new tech-nology applications (US National ResearchCouncil and Committee on Critical Mineral Im-pacts of the US Economy, 2008; Buchert et al.,2009; European Commission, 2010a; ErdmannandGraedel, 2011). For many of these uses, thereare no substitutes that offer the same perfor-mance. Furthermore, given the miniaturizationtrend required by new technologies, indium isused in very small quantities that impede itsseparation, recovery, and recycling. Forexample, a 2005 cell phone screen contained aslittle as 5.5 mg of indium per phone (the metalspresently recovered from cell phones such asgold and silver represent 350 mg and 3.7 g,respectively) (Yoshioka et al., 1994; Hageluken,2007b). PC monitors contain between 79 and

I. RECYCLING I

82 mg of indium, whereas bigger liquid crystaldisplay (LCD) television screens use 260 mg(Buchert et al., 2009). However, there are somany of these products that these smallamounts of indium add up to a potentiallyimportant source for future supply. Withoutrecycling or implementing substitutes, the sus-tainability of these technologies is questionable.

Indium is most commonly recovered as a by-product of the zinc smelting process from zinc-sulfide ores (sphalerite). Other minor sourcesare from tin and lead refining processes, whichat the present time are not exploited. Thecontent of indium in zinc concentrate variesdepending on the geographic location of themine. For example, Rodier gives 0.027% (byweight) of indium content in zinc concentratefor Kidd Creek, Ontario, Canada; 0.010% inPolaris, Northwest Territories, Canada; and0.004% in Balmat, New York (Rodier, 1980).Based on the intermediate concentration of0.010%, an average zinc content of 55% in zincconcentrate, and a global production of eightmillion tonnes of zinc in 2010, we estimate that

N CONTEXT

4. RECYCLING RARE METALS30

the amount of indium that could have poten-tially been recovered from zinc-bearing ores in2010 is 1454 t. The USGS reports that 574 t of in-dium were produced globally for the same year(Tolcin, 2011). In other words, a third of the totalindium currently available from zinc refineriesis not recovered. This coincides with theamount Indium Corporation estimates is lost totailings or non-indium-capable refineries (abouttwo-thirds of all indium in zinc ores) (USDepartment of Energy, 2011). There is a marginfor improving recovery at this initial stage, butit is still not economically attractive to do so.

High purity indium is used to make com-pounds with other metals for the manufactureof products. The most important is indium-tin-oxide (ITO) (78% indium, 17.5% oxygen, and4.5% tin). Thanks to its electrical conductivityand optical transparency, ITO is used for trans-parent electrodes in products such as LCDs,TVs, and touch screens. Based on a previousstudy, we estimate that 425 t of indium wereused for the production of ITO in 2011 (TalensPeiro et al., 2013). ITO comes as a powder andis deposited as a thin, film coating on a substratevia a process called “sputtering” that is report-edly only 3% efficient.

Sputtering deposition techniques vary, butthe most common ones are electron-beam evap-oration and physical vapor evaporation.Approximately 3% of the indium ends up suc-cessfully on the substrate, although 70% is targetresidual material (recycled by reducing to in-dium metal), 20% is deposited onto the surfaceof tools and chamber walls, 5% is etched fromthe substrate, and 2% ends up in faulty panels(Yoshioka et al., 1994). These percentages arebased on a study from 2007, and it is possiblethat sputtering has become more efficient sincethen (Goonan, 2012). However, for lack ofupdated process data, based on these percent-ages we calculate that 12.8 t of indium areembodied in flat-panel display (FPD) products,297.5 t are recycled to make more ITO, and114.7 t are dissipated during the FPD production

I. RECYCLING I

process. These quantities are shown in the prod-uct manufacture stage of Figure 4.1. The upperdotted line represents a greater amount of in-dium available if the sputtering process resultedin half the indium losses (58 t instead of 114.7 t).The figure also shows the fraction of the indiumin ITO that is not successfully placed on the sub-strate and is recycled to make ITO.

Presently there is no indium recycling frompost-consumer products that employ LCDssuch as TVs, personal computers (PCs), mobilephones, and car navigation systems. The indiumis diluted and dissipated in waste managementschemes (recycling, landfill, etc.) at the EOL ofthe products. Sharp Corporation in Japan hasinvented a process for recovering ITO fromLCDs, which involves crushing the screensinto small chips and treating them in acid solu-tion. A large-scale implementation in the futurecould result in an important source of indium(Kawaguchi, 2006). Given the growth in con-sumption of touch screens and flat-panel TVs,as well as the increase in average screen size,this recoverable indium is crucial for future sup-ply. In Figure 4.1, the ideal EOL stage includesthe indium embodied in LCDs, whereas the pre-sent situation shows no recovery.

Indium is also used to make alloys with basemetals to lower their melting point. The indium-based alloys are used as fusible alloys (neededfor devices such as fire sprinkler systems), hold-ing agents, and solders. Indium solders arepreferred in printed circuit board (PCB) assem-bly because indium is lead-free and has excel-lent wetting properties. Alloys of indium withprecious metals such as gold and palladiumare used in dental work due to their ductile na-ture. We estimate that 75 t of indium weredestined for alloys and solders in 2010, of which6 t were destined for PCBs (Talens Peiro et al.,2013). Some of these applications render the in-dium unrecoverable because it is used in suchsmall quantities. PCBs are presently beingrecycled to recover copper, gold, silver, andpalladium, but other critical metals such as

N CONTEXT

4.3 OTHER EXAMPLES OF RARE METALS 31

indium are not being recovered (Buchert et al.,2009). The use of indium in solders and alloyscan be considered dissipative at the EOL. The“ideal” scenario represented in Figure 4.1 doesnot include the 75 t of indium destined for alloysand solders as recoverable.

The solar energy industry is also an importantconsumer of indium. According to industry fig-ures, 57 t of indium were used in the formof copper-indium-gallium-diselenide (CIGS)employed as a semiconductor in photovoltaiccells in the form of thin-film photovoltaic mate-rial (Talens Peiro et al., 2013). ITO is used as ananode in the photovoltaic cells to increase thelight conversion efficiency because of its lighttrapping properties and high transparency.Other indium compounds, such as indium anti-monide (InSb), indium nitride (InN), and indiumphosphide, are also used as semiconductors inphotovoltaic cells and in other applicationssuch as infrared detectors and thermal imagingcameras. Photovoltaic cells have a lifetime of20e30 years, after which no recovery of indiumis presently taking place. Potentially the indiumpresent in the cells could be recovered in asimilar fashion as the FPDs, and this is repre-sented in the “ideal” scenario represented bythe dotted line in Figure 4.1.

Other semiconductor applications such aslaser diodes, fiber optic telecommunications, de-tectors, and light-emitting diodes (LED) wereresponsible for 17 t of indium. Indium in theform of indium gallium nitride (InGaN) isused as a light-emitting layer in blue and greenLEDs. Indium gallium arsenide (InGaAs) is apopular material in infrared detectors. We arenot aware of any recovery of the indium usedin these products, but potentially, this is a recov-erable fraction because the indium is not dissi-pated during use.

As is illustrated by Figure 4.1, less than 1% ofindium is presently recovered at EOL as hasbeen shown by other authors (Graedel et al.,2011). However, what the figure also shows arethe fractions of the indium produced that are

I. RECYCLING I

not dissipative and that could potentially berecovered. Recovering indium from LCDs andphotovoltaic cells could result in 384 t of indiumat EOL (minus unavoidable losses). If we as-sume a 50% improvement in the efficiency ofthe sputtering process, the recoverable indiumat EOL is further increased to 442 t. The usephase is not included in the diagram becausewe know of no losses of indium that occur dur-ing the use of the products described previously.

4.3 OTHER EXAMPLES OF RAREMETALS

Other metal life cycles can be analyzed in asimilar fashion. We next explain three examplesof scarce metals: europium, gallium, and plat-inum group metals (PGMs), all used in lowquantities, yet crucial for different technologies.

Europium is one of the 17 so-called rare earthmetals (REMs). Its major mineral source isbastnasite, a carbonate fluoride mineral foundin association with (some) iron mineral ores(such as magnetite, hematite, goethite, andlimonite). The content of bastnasite in ironores varies from 4e17%. Bastnasite containsabout 0.1e0.5% of europium, a very small con-centration compared to lanthanum, cerium,and neodymium, which are found in 42e50%,23e33%, and 12e20%, respectively (Guptaand Krishnamurthy, 1992). The commercialapplication of europium is based on its phos-phorescence. It is added to semiconductorsin trace amounts, either in the þ2 or þ3 oxida-tion state, to improve the emission of lightin various wavelengths. Divalent europium(Euþ2) tends to give blue phosphors, whereastrivalent europium (Euþ3) gives red phosphors.Both of them combined with terbium-basedphosphors yield white light (Gupta and Krish-namurthy, 2005).

In 2010, the entire global production of euro-pium (404 t) was used in phosphors for coldcathode fluorescent lamps (CCFLs), in LEDs

N CONTEXT

4. RECYCLING RARE METALS32

fitted in LCDs for background illumination, andin LED for light bulbs. Although most of themarket is now dominated by LCDs usingLEDs, to estimate the amount of europium stockin EOL LCDs, we need to know the amountscontained by each type of background lighting.The amount of europium in LCDs using CCFLsvaries from 8.10 mg for TV to 0.13 mg for note-book computers, whereas LCDs using LEDscontain an even lower amount: 0.09 mg forTVs compared to 0.03 mg for notebooks. Euro-pium is also used to give a reddish color inwarm white LEDs that have a correlated colortemperature lower than 3000 K. The averageweight of europium in warm white LEDs variesfrom 0.4 to 0.9 mg (Buchert et al., 2012). Consid-ering that losses of europium frommining to themanufacturing of phosphors are estimated to be25%, we estimate that the remaining 75% is allcontained in LCDs and LEDs.

At present, europium is rarely recycled (Duand Graedel, 2011; Graedel et al., 2011). Themost important obstacle is collection due to itsuse in very tiny amounts. As an example,recovering one ton of europium would requirethe collection of at least 1.3 trillion units ofwhite light LEDs. From a technological perspec-tive, the recovery of europium also requiresdetailed knowledge about the composition ofthe phosphors. For instance, phosphors gener-ally contain other rare earth elements such asyttrium, cerium, and lanthanum in the supportmatrix, not to mention terbium and gadoliniumas dopants. The combined use of rare earths inphosphors inhibits their recovery because theyall have very similar properties. Hence, the sep-aration of any one from the others requiresmany processing steps.

Although gallium is a relatively commonelement, it mainly occurs as a trace amount inbauxite and zinc ores such as sphalerite. At pre-sent, almost all gallium is obtained as a by-product of alumina production from bauxite,which contains an average of 60 ppm (Grayand Kramer, 2005). In 2010, the world

I. RECYCLING I

production of refined gallium was 161 t. Ofthat, 106 t were obtained as crude gallium andthe remaining 55 t from preconsumer recycling(Jaskula, 2010). In 2010, 65% of gallium outputwas obtained as a “hitchhiker” of aluminumsmelting. Table 4.1 estimates that the currentproduction of gallium represents only 3% ofthe total amount contained in bauxite. It alsoidentifies zinc ore as a potential source of gal-lium that is not being exploited at present.

About 66% of the gallium consumed in 2010was used in integrated circuits (ICs). Phosphorsaccounted for 18%, thin films 2%, and theremaining 14% of gallium consumption was forresearch and development, specialty alloys, andother applications (European Commission,2010b; Talens Peiro et al., 2013). The most impor-tant gallium compounds are arsenides, nitrides,and phosphides. Gallium arsenide (GaAs)and aluminum gallium arsenide (AsGaAl) canconvert electrical signals into optical signals athigh speed with low power consumption, andbetter resistance to radiation compared to othercompounds (Habashi, 1997). They are widelyused as substrates in the manufacture of semi-conductor components as transistors and ICsfor the electronics and telecommunications in-dustry. For example, a mobile phone containsfrom 0.3 to 1.5 mg of gallium, and fourth genera-tion smartphones use up to 10 times that amount(Talens Peiro and Villalba Mendez, 2011; Jaskula,2013). Gallium nitrides (GaN) are primarilyapplied in light-emitting diodes (LEDs) for thebacklighting of LCDs for TVs and notebooks.Buchert et al. estimates that an LCD TV contains4.90 mg and a notebook 1.60 mg of gallium(Buchert et al., 2012). Gallium phosphides andphosphides complex of gallium, aluminum andindium are also used for the production of opto-electronic components and ICs. Gallium is alsoused in thin-film solar cells as CIGS and in thetriple-junction cells: indium gallium phosphide(GaInP2) (Green and Emery, 2011).

More than 85% of gallium ends up in theoret-ically nondissipative uses: electronics,

N CONTEXT

4.3 OTHER EXAMPLES OF RARE METALS 33

phosphors, and thin film photovoltaic panels.The numerous steps during manufacturingand the high quality requirements for most ofthe electronic products using gallium lead tosubstantial processing losses. Typically only20e30% of the gallium is finally embodied inthe end-products (Koslov et al., 2003). Forinstance, in the production of GaAs substrates,30% of the initial gallium is embodied in thefinal product; the remaining 70% is lost duringsuch processes as etching and polishing. Eichlerestimates that nearly 35% of the losses can stillbe recovered in the different steps (Eichler,2012).

In CIGS manufacturing, gallium can bedeposited in the absorber cell layer by varioustechnologies: electron-beam (EB), electrochemi-cal deposition, and co-evaporation. The amountof gallium deposited for electron-beam and co-evaporation is only 20% of the total galliuminput. For electrochemical deposition, theamount deposited is 30% of the total input(Kamada et al., 2010). Based on the fact that a1 MW CIGS panel contains about 3.5e4 kg ofgallium, we estimate that 7.5e16.5 t of galliumare wasted during the manufacturing stage(Kalejs, 2009; Christmann et al., 2011). New de-velopments to minimize the loss of gallium dur-ing CIGS manufacturing are based on usinginkjet printing for deposition. This technologycould reduce the amount of gallium wasted by90% and thus reduce the cost of thin-film solarcells (Quick, 2011).

At present, gallium is only recovered fromnew scrap (preconsumer scrap) generated dur-ing the manufacturing of semiconductors,mainly from GaAs wafer waste, and from liquidphase deposition. There are several options forthe recovery of gallium from semiconductorwaste: thermal dissociation, oxidation with oxy-gen, nitriding in ammonia, and chlorinationwith chlorine gas (Koslov et al., 2003). Most ofthe difficulty in recycling gallium is due to theproblem of separation from the other compo-nents of contact alloys such as indium, tin,

I. RECYCLING I

germanium, lead, silver, copper, and gold. In2010, about 55 t of gallium were produced frompreconsumer scrap (Jaskula, 2013).

Many efforts to recycle gallium focus onincreasing the collection of EOL products. InEurope, the PV Cycle association, which repre-sents more than 90% of the EU photovoltaicmarket, recently set up a voluntary collectionand recycling scheme. In 2010, PV Cyclecollected about 80 t of modules. In 2011, thisgrew to 1400 t and, in the first two quarters of2012, 2250 t (Neidlein, 2012). In the EU, photo-voltaic module take back and recycling becamemandatory in February 2014. Even though thinfilms represent only 5% of the current photovol-taic panels market, they may eventually domi-nate the future market because they have morepotential for cost reduction. Moss et al. esti-mates that thin films will have a market shareof 18% by 2020; thus, gallium demand wouldincrease to 14 t (Moss et al., 2011).

PGM is the term used for a group of sixmetals: platinum, palladium, rhodium, ruthe-nium, iridium, and osmium. They tend to occurtogether in primary and secondary mineral de-posits. In primary deposits, they are found inplatiniferrous ores in association with ironsulfides and sulfides of nickel, cobalt, and cop-per. Economical deposits mainly contain sperry-lite (PtAs2), cooperite (PtS), stibiopalladinite(Pd3Sb), laurite (RuS2), ferroplatinum (FeePt),polyxene (FeePteother platinummetals), osmi-ridium (OseIr), and iridium platinum (IrePt).PGMs can also be found in secondary depositsformed by weathering and washing of primarydeposits. The content of PGM varies dependingon the type of mineral deposit. In South Africa,the average platinum grade is about 8 g/t of ore,but it can reach 27g/t in deposits associatedwith nickel ore. In Russia, where platinum is ob-tained as a by-product of nickel and copper, theaverage grade is about 5 g/t. The average gradecan reach 15 g/t in sulfide deposits (Renner,1997). In 2010, the world production of PGMwas 689 t; 456 t were obtained from platinum

N CONTEXT

4. RECYCLING RARE METALS34

sulfides and arsenides, 222 t from recycling, and11 t were obtained as hitchhikers of nickel (Tal-ens Peiro et al., 2013). According to Table 4.1,the current production of PGM as a hitchhikerof nickel is near its potential level.

Among the PGM group, we can differentiatesubgroups based on production quantities. Plat-inum and palladium are a first group of metalsproduced in multiples of 100 t/year. Rhodiumand ruthenium are a second group producedin multiples of 10 t/year; iridium and osmiumcan be regarded as a third group, produced invery small quantities. In brief, rhodium, ruthe-nium, iridium, and osmium can be regardedas by-products (or “hitchhikers”) of platinumand palladium. In 2010, the global productionof PGMs was: 249 t of platinum, 360 t of palla-dium, 36 t of rhodium, 35 t of ruthenium, 7 t ofiridium, and 2 t of osmium.

PGMs are exceptional oxidation catalysts.They are electrically and thermally conductiveand offer a high oxidation resistance combinedwith extraordinary catalytic activity, chemicalinertness, high melting point, temperature sta-bility, and corrosion resistance (Lofersky,2010). Almost 50% of the demand for PGMs isfor use in automobile catalytic converters,14% is used for jewelry, 9% for investment,

FIGURE 4.2 Substance flow analysis of platinum in 2010 (i

I. RECYCLING I

4% for other electrical uses, 1% for catalysts inthe chemical industry, 1% for alloys with elec-trochemical applications, less than 1% for theglass industry, and the remaining 22% to otherunspecified uses (Lofersky, 2011). See Figure 4.2.Some PGMs are lost from their extraction in themanufacturing of end-products. Dissipativeuses of PGM include dental restorative mate-rials and in the glass industry for the develop-ment of flat-panel displays as LCDs, plasma(PDP), and as LEDs (Buchert et al., 2009). ThePGMs used as catalysts in the chemical andpetrochemical sectors are mostly close-looprecycled, and the amount dissipated duringuse is negligible. Platinum used in jewelry ac-cumulates over time, but some is recycled.Recoverable fractions of PGM include automo-bile catalytic converters and from electricaluses. Figure 4.2 shows platinum’s dissipative(regarded as lost), stock, and recyclingfractions.

As illustrated in Figure 4.2, almost 40% ofplatinum is lost in nonrecoverable uses, 18% isadded to the stock of end-products, and 23% isrecycled. Platinum is lost when it is used inmedical applications, in the glass industry, andfor other uses such as automotive sensors, thecoating of aircraft turbine blades, and spark

n metric tonnes).

N CONTEXT

4.3 OTHER EXAMPLES OF RARE METALS 35

plugs. It is also lost from the fraction of automo-bile exhaust converters and electrical equipmentnot collected and adequately recycled.

In 2010, automobile catalytic converters used99 t of platinum, 204 t of palladium, and 30 t ofrhodium (Butler, 2011). Catalytic converters arepart of the exhaust system designed to accel-erate the oxidation of carbon monoxide (CO),nitrogen oxides (NOx), and unburnt hydrocar-bons. They are composed by a ceramic or metalhoneycomb structure of channels coated withso-called “wash-coating” catalysts that arecomposed of a combination of PGMs and rareearth oxides (Borgwardt, 2001; Lifton, 2007).The best known combinations of PGMs are plat-inumerhodium (PteRh), platinumepalladium(PtePd), and a three-way combination of allthree metals (PtePdeRh). The first two combi-nations accelerate the complete oxidation of car-bon monoxide and hydrocarbons. The three-way catalyst also reduces nitrogen oxides topure nitrogen and oxygen. PGMs are fixed inthe wash-coat surface usually by impregnationor by coating with a solution of hexachloropla-tinic (IV) acid (H2PtCl6$6H2O), palladium chlo-ride (PdCl2), and rhodium chloride (RhCl3)salts. As the solvent evaporates, a dry layer ofthe PGM salts results in the surface of the hon-eycomb structure (Ravindra et al., 2004).

PGM losses during the wash-coat formationrange from about 2e6% as maximum (Patchettand Hunnekes, 2012). Taking into account thatcurrent catalytic converters contain an averageamount of about 2 g of PGM, about 4e12 mgof PGM are lost during manufacturing(Yentekakis et al., 2007). Based on the datashown in Figure 4.2, we estimate that between2 and 6 t of platinum were lost during wash-coat formation in 2010. PGMs are also lost dur-ing automobile engine operation due to therapid change of oxidative/reductive condi-tions, high temperature, and due to mechanicalabrasion. The average quantity of PGMsreleased into the environment is in the rangeof 65e180 ng/km (Ravindra et al., 2004).

I. RECYCLING I

A possible way to estimate the amount ofPGM lost during vehicle operations is toconsider the current stock of vehicles in serviceand the average distance traveled by them.

A more accurate estimate of the amount ofPGMs potentially recoverable can be compiledby counting deregistered vehicles. A case studyin Germany estimated that only 41% of cata-lytic converters from deregistered vehiclesreach dismantler facilities, and about 70% ofthe PGMs contained in those converters arerecovered. The remaining 59% of catalytic con-verters are not separated from the vehicle andare not recycled. Many (perhaps most) areexported to areas without adequate recyclinginfrastructure. In some cases, convertors arenot removed from car bodies before beingshredded, or they are mistakenly sorted intothe wrong fractions, from which separation isnot feasible (Hageluken, 2007a). In 2010, therewere 33.7 t of platinum, 41.2 t of palladium,and 7.3 t of rhodium recycled from automobilecatalytic converters (Lofersky, 2011).

In the electrical sector, PGMs are used foractive components such as transistors, ICs,thick-film hybrid circuits and semiconductormemories, and also in passive components,which include multilayer capacitors, thick-filmresistors, and conductors. The most importantPGM used in electronics is ruthenium in combi-nation with platinum, palladium, and iridiumin alloys. Ruthenium is used in small quantitiesto increase wear resistance for electric contacts,and in microelectronics in computer hard disks,multilayer ceramic capacitors, and in hybrid-ized ICs. At present, neither ruthenium noriridium is recycled (Lofersky, 2011).

In 2010, only platinum and palladium wererecycled from electrical and electronic waste.The amount of platinum recovered repre-sented 5% of the amount consumed, as shownin Figure 4.2. In 2010, there were 113.7 t ofpalladium recycled from electronic waste,whereas the global sales of palladium for elec-tronics were 43.9 t (Lofersky, 2011). These

N CONTEXT

4. RECYCLING RARE METALS36

figures show that the recycling system and thetechnology for recovering palladium are func-tioning. In fact, in 2010, recycled palladiumsupplied 44% of the world’s production.Platinum and palladium were also recycledfrom jewelry: 23.2 t of platinum and 2.5 t ofpalladium.

4.4 THE DISTANT FUTURE:GEORGESCU’S LAST LAUGH?

Georgescu-Roegen proclaimed that “mattermatters” and asserted that the “entropy law”is the “taproot” of economics (Georgescu-Roegen, 1971). His point was that dissipationof matter (i.e. chemical elements) due to theimpossibility of perfect recycling constitutes a“fourth law” of thermodynamics and an un-avoidable limit to economic growth. He waswrong about the fourth law and wrong aboutthe technical impossibility of recycling, givenan unlimited supply of useful energy (exergy)(Ayres, 1999). This technical error has somewhatdiscredited his entire thesis. However, thelong-term prospects for technologies that utilizegeologically scarce “hitchhiker” metals are verypoor from an economic perspective.

As noted earlier, the short- to medium-termsituation is that most rare metals are currentlyunderexploited in the sense that the potentialsupply from mining of the important primary(“carrier”) metals (iron, aluminum, copper,zinc, nickel, etc.) is larger than the actual output.In some cases, the potential supply is hundredsof times larger than the current consumption;therefore, there is no immediate need for recy-cling. But in other cases (e.g. cobalt and molyb-denum), current output is reasonably close tomaximum potential output, given the currentdemand for the associated carrier metal (e.g.copper). Of course, the potential supply of cop-per is also limited, although not (it seems) formany decades to come.

I. RECYCLING I

With regard to standard industrial metalssuch as iron and aluminum, the ores or possibleores are plentiful, and most of the uses are notinherently dissipative. There are dissipativeuses of iron such as nails and small hardwareitems, as well as rust from structural beamsand iron oxide pigments. There are dissipativeuses of aluminum, especially aluminum foiland aluminum bottle caps. Hence iron andaluminum will be mined to some extent evenin the very distant future unless those “waste-ful” uses are also eliminated. Nickel, cobalt,and molybdenum will be recycled to a high de-gree because their metallurgical uses, such as instainless steel or refractory (heat resistant) al-loys, are such that collection and separationare not too difficult. Copper, zinc, and lead arevery easily recycled. In the case of copper, it iseasy to recycle copper roofing, copper tubingfor water, and copper from household wiring,motor and generator windings, and so forth.But copper for computer chips, brass cartridgesfor bullets, and copper chemicals are very diffi-cult or impossible to recycle. In the case of lead,old water pipes, radiation shielding, and lead-acid batteries are easily recycled, but lead-based solder is not, and lead bullets and leadshot for shotguns are quite dissipative. Zinchas few metallic uses except for some kinds ofhardware. Zinc coating for iron (“galvanizing”)is one of that metal’s biggest uses. Recovery ofzinc from galvanized scrap iron is possible butnot very profitable. Similarly, PGMs are easilyrecycled from industrial catalysts, or fuel cells,but it is getting very hard to recover PGMsfrom automobile exhaust systems unless theEOL vehicles are carefully dismantled prior toscrapping operations. However, in the verylong run, automotive vehicles with internalcombustion engines that burn gasoline or dieselmay be largely replaced by electric vehicles.

For the major metals, we may never eliminatethe need for some mining activity, but theresourcebase is large enough that no supply crisisis likely for centuries to come. Recycling will

N CONTEXT

REFERENCES 37

never be 100% effective, thanks to the SecondLaw, but fairly high recovery rates from demoli-tion scrap and scrap cars can be foreseen. Also,the value of secondary metals (especially goldand silver) is already keeping some mines open.Similarly, the increasing value of some veryscarce metals, such as indium or tellurium, mayeventually justify the investment in large-scale re-covery and in the recycling of certain electronicitems such as touch screens or PV panels.

Unfortunately, new IT products, such asiPhones and iPads, use more and more of thescarce metals in trace amounts. It is importantto ensure that this trend toward micro-miniaturization does not prevent us from recov-ering thesemetals. In the future, this will requiremuch higher pricesdsuch as those that we nowsee for gold and platinum. From a systemsperspective, this may have to be accomplishedby innovativemeans such as “renting” the scarcematerials. The implications of such schemes willrequire extensive study and analysis.

References

Ayres, R.U., 1999. The second law, the fourth law, recyclingand limits to growth. Ecological Economics 29 (3),473e483.

Berger, V.I., Singer, D.A., et al., 2011. Ni-Co Laterite Depositsof the World. Database and Grade and Tonnage Models.United States Geological Survey.

Borgwardt, R., 2001. Platinum, fuel cells and future US roadtransport. Transportation Research, Part D: Transportand Environment 6 (3), 199e207.

Buchert, M., Manhart, A., et al., 2012. Recycling Critical RawMaterials from Waste Electronic Equipment. SustainableInnovation and Technology Transfer Industrial SectorStudies. Oeko-Institut e.V.

Buchert, M., Schueler, D., et al., 2009. Critical Metals forFuture Sustainable Technologies and Their RecyclingPotential. Sustainable Innovation and TechnologyTransfer Industrial Sector Studies. UNEP.

Butler, J., 2011. Platinum 2011. PGM Market Review. John-son Matthey.

Christmann, P., Angel, J.M., et al., 2011. Panorama 2010 dumarche du gallium. Compagnie Europeenne d’Intelli-gence Strategique (CEIS), pp. 1e53.

I. RECYCLING I

Drew, L.J., Qingrun, M., et al., 1991. The geology of theBayan Obo iron-rare earth-niobium deposits, InnerMongolia, China. Materials Science Forum 13 (70e72),13e32.

Du, X., Graedel, T.E., 2011. Uncovering the global life cyclesof the rare earth elements. Scientific Reports 1, 145.

Eichler, S., 2012. Green gallium arsenide (GaAs) substratemanufacturing. In: The International Conference onCompound Semiconductor Manufacturing Technology,Boston, Massachusetts, USA.

Erdmann, L., Graedel, T.E., 2011. Criticality of non-fuelminerals: a review of major approaches and analyses.Environmental Science & Technology 45 (18),7620e7630.

European Commission, 2010a. Annex V to the Report ofthe Ad-hoc Working Group on Defining Critical RawMaterials. European Commission, Brussels, pp. 220.

European Commission, 2010b. Critical Raw Materials forthe European Union. European Commission.

Georgescu-Roegen, N., 1971. The Entropy Law and theEconomic Process. Harvard University Press, Cam-bridge, MA.

Goonan, T.G., 2012. Materials Flow of Indium in the UnitedStates in 2008 and 2009. US Geological Survey, Denver.

Graedel, T.E., Allwood, J., et al., 2011. What do we knowabout metal recycling rates? Journal of Industrial Ecol-ogy 15 (3), 355e366.

Gray, F., Kramer, D.A., 2005. Gallium and gallium com-pounds. Kirk-Othmer Encyclopedia of Chemical Tech-nology. Kirk-Othmer, John-Wiley & Sons.

Green, M.A., Emery, K., 2011. Solar cell efficiency tables(version 37). Progress in Photovoltaic Research andApplications 19 (1), 84e89.

Gupta, C.K., Krishnamurthy, N., 1992. "Extractive metal-lurgy of rare earth. International Materials Reviews 37(5), 197e248.

Gupta, C.K., Krishnamurthy, N., 2005. Extractive Metal-lurgy of Rare Earths. CRC Press.

Habashi, F. (Ed.), 1997. Handbook of Extractive Metallurgy.Wiley-VCH.

Hageluken, C., 2007a. Closing the loop e recycling auto-motive catalysts. Metall 61, 24e39.

Hageluken, C., 2007b. RO7 Metals Recovery from E-scrap ina Global Environment. OEWG, Basel. Umicore.

Jaskula, B.W., 2010. Mineral Commodity Summaries: Gal-lium. US Department of the Interior, Washington, DC.United States Geological Survey, pp. 28e59.

Jaskula, B.W., 2011. 2009 Minerals Yearbook: Gallium. USDepartment of the Interior, Washington, DC. UnitedStates Geological Survey I, pp. 1e21.

Jaskula, B.W., 2013. Minerals Commodity Summaries: Gal-lium. US Department of the Interior, Washington, DC.United States Geological Survey I, pp. 58e59.

N CONTEXT

4. RECYCLING RARE METALS38

Kalejs, J., 2009. NSF Workshop. Manufacturing Challengesfor PV in the 21st Century, Arlington, VA, USA.

Kamada, R., Shafarman, W.N., et al., 2010. Cu(In,Ga)Se2 filmformation from selenization of mixed metal/metal-selenide precursors. Solar Energy Materials and SolarCells 94 (3), 451e456.

Kawaguchi, Y., 2006. Coming full circle. Recycling Magazine.Koslov, S.A., Potolokov, N.A., et al., 2003. Preparation of

high-purity gallium from semiconductor fabricationwaste. Inorganic Materials 39 (12), 1257e1266.

Lifton, J., 2007. By-products III: Rhodium. Resource InvestorNews.

Lofersky, P.J., 2010. Mineral Commodity Summaries:Platinum-Group Metals. US Department of the Interior,Washington, DC. United States Geological Survey,pp. 120e121.

Lofersky, P.J., 2011. 2010 Minerals Yearbook: Platinum-Group Metals. US Department of the Interior, Washing-ton, DC. United States Geological Survey, pp. 1e12.

Moss, R.L., Tzimas, E., et al., 2011. Critical Metals in Stra-tegic Energy Technologies. Assessing Rare Metals asSupply-Chain Bottlenecks in Low-Carbon Energy Tech-nologies. JRC e Institute for Energy and Transport, pp.1e162.

Neidlein, H.C., 2012. PV module recycling mandatory for allEU members by Q1 2014. PV Magazine.

Patchett, J., Hunnekes, E.V., 2012. Gas Catalysts ComprisingPorous Wall Honeycombs. http://www.patentgenius.com. B. Corporation. US, Sample, D, pp. 1e19.

Quick, D., 2011. Researchers Cut Waste and Lower Cost of‘CIGS’ Solar Cells Using Inkjet Printing Technology.Gizmag.

Ravindra, K., Bencs, L., et al., 2004. Review: platinum groupelements in the environment and their health risk. Sci-ence of the Total Environment 318 (1), 1e43.

I. RECYCLING I

Renner, H., 1997. Platinum group metals. In: Habashi, F.(Ed.), Handbook of Extractive Metallurgy. Wiley-VCH(Section 25).

Rodier, D.D., 1980. Lead-zinc-tin ‘80. In: Cigan, J.M.,Mackay, T.S., O’keefe, T.J. (Eds.), Proceedings of WorldSymposium on metallurgy and environmental control.The Metallurgical Society of AIME, Las Vegas.

Shedd, K.B., 2011. Mineral Commodity Summaries: Cobalt.US Department of Interior, Washington, DC. UnitedStates Geological Survey, pp. 46e47.

Talens Peiro, L., Villalba Mendez G., 2011. Results from theWorkshop “Preliminary Assessment of MultifunctionalMobile Phones”. Brussels.

Talens Peiro, L., Villalba Mendez, G., et al., 2013. Materialflow analysis of scarce metals: sources, functions, end-uses and aspects for future supply. Environmental Sci-ence & Technology 47 (6), 2939e2947.

Tolcin, A.C., 2011. Mineral Commodity Summaries: Indium.US Department of Interior, Washington, DC. UnitedStates Geological Survey, pp. 74e75.

US Department of Energy, 2011. Critical Materials Strategy.US Department of Energy, pp. 1e196.

US National Research Council and Committee on CriticalMineral Impacts of the US Economy, 2008. Minerals,Critical Minerals and the US Economy. National Acad-emy Press, Washington, DC.

USGS, 1973. United States Mineral Resources. US Gov’tPrinting Office, Washington, DC.

Yentekakis, I.V., Konsolakis, M., et al., 2007. Novel electro-positively promoted monometallic (Pt-only) catalyticconverters for automotive pollution control. Topics inCatalysis 42e43, 393e397.

Yoshioka, K., Nakajima, T., et al., 1994. Sources of TotalFactor Productivity for Japanese Manufacturing In-dustry: 1962e1988. Keio University, Tokyo, Japan.

N CONTEXT