Hydrometallurgy 196 (2020) 105435

Contents lists available at ScienceDirect

Hydrometallurgy
journal homepage: www.elsevier.com/locate/hydromet

Recent progress in impurity removal during rare earth element processing: A T

review
W.D. Judgea, G. Azimia,b,
⁎

a
Department of Materials Science and Engineering, University of Toronto, 184 College street, Toronto, ON M5S3E4, Canada
b
Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College street, Toronto, ON M5S3E5, Canada

ARTICLE INFO ABSTRACT

Keywords: Supply of rare earth elements (REEs) critical to the development and maintenance of industrialized economies
Rare earth elements has become uncertain in recent years owing to the limited number of REE producers. In order to develop al­
Impurity removal ternative REE sources, new mines are being explored and end-of-life secondary sources have been targeted for
Solvent extraction REE recovery. However, the composition of ores and secondary sources often differ significantly from one source
Ion exchange
to another and hydrometallurgical practices are not rigorously established or matured for many alternative REE
Selective precipitation
sources. This presents a challenge in processing these materials while minimizing the entrainment of impurities.
This work presents a comprehensive and systematic review on recent progress in impurity removal during REE
processing. A large number of original research articles covering techniques including solvent extraction, ion
exchange, precipitation, and other emerging technologies have been reviewed in detail and critically assessed to
understand the role and behaviour of specific impurities in REE processing (including Al, Ca, Mg, Fe, Si, Th, U,
Ti, Zr, Hf, Cr, Mo, Mn, Co, Ni, Cu, Zn, Sn, Pb, and Bi). Control of the most troublesome impurities for industry
(Al, Fe, Th, U) depends on their concentration, redox ratio, and the degree of purity required. Most aluminum,
ferric iron, thorium, and uranium(IV) selectively precipitate as hydroxides prior to REEs. Ferrous iron does not
usually co-extract with REEs in their ion exchange or solvent extraction circuits. Precipitation of REE oxalates is
effective to partition REEs and uranyl (UO22+) species. To remove thorium and uranium to the highest degree
requires separate solvent extraction or ion exchange steps designed to specifically extract these impurities by
anion exchange.

1. Introduction significantly different than conventional ores and new hydro­

metallurgical practices must be established. While much research has
In the twenty-first century, rare earth elements (REEs) emerged as focused on cracking the ore or secondary source, less attention has been
critical materials in the development and maintenance of industrialized given to removal of impurities from the resulting processing streams. A
economies. During this time REEs were supplied by a limited number of comparison of impurities encountered during processing of different
mines/commercial entities that produced REEs in tonnage quantities. REE resources is depicted in Fig. 1. Since chemical beneficiation rarely
This enabled a rare earth crisis beginning in 2010 that sparked intense has total selectivity for the targeted REEs, subsequent processing
exploration and development of new sources for REEs (Gschneidner Jr., streams inevitably end up with some amounts of these impurities. Part
2011; Machacek and Fold, 2014; Sprecher et al., 2015); consequently, of the challenge is that the composition and concentration of impurity
the rare earth industry has changed significantly in recent years. Many elements may be distinct and unusual compared with the traditional
industrialized nations have been looking to develop new mines and rare earth leach liquors, and new procedures must be developed or
target end-of-life secondary sources to establish circular REE econo­ existing procedures must be modified to reject these impurities. The
mies. However, transitioning from promising mineral deposits or sec­ presence of impurities in REE-containing solutions has an enormous
ondary sources to marketable products requires research and develop­ impact on not only the final REE products, but also on the efficiency of
ment expertise to address the complex technological challenges around processing; some specific examples are provided in Table 1.
production, separation, and processing of REEs. In REE processing streams, impurities like aluminum and iron(III)
The chemistry of these new ores and secondary sources can be are detrimental to solvent extraction circuits. For example, in

⁎
Corresponding author at: Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College street, Toronto, ON M5S3E5, Canada.
E-mail address: g.azimi@utoronto.ca (G. Azimi).

https://doi.org/10.1016/j.hydromet.2020.105435
Received 5 February 2020; Received in revised form 27 May 2020; Accepted 27 July 2020
Available online 06 August 2020
0304-386X/ © 2020 Elsevier B.V. All rights reserved.
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Fig. 1. A comparison of the impurities encountered during REE processing from various resources.

Table 1
Impurities and their effects on subsequent REE processing and final REE products.
Impurity Examples of effects on REE processing Examples of effects on final REE products

Al • Accumulates
emulsions.
in solvent extraction circuits and forms • Reduces saturation magnetization and anisotropy field of NdFeB magnets.
• Reduces precipitation efficiency of REE oxalates.
• Impedes the formation of REE carbonate crystals.
Ca, Mg • Lowers REE recovery by REE co-precipitation with gypsum • Increases the basicity of rare earth polishing suspensions.
(Ca). • Decreases the ionic conductivity of REE oxygen sensors (i.e., YSZ).
• Component in the crud in solvent extraction circuits.
Fe • Accumulates in solvent extraction circuits and forms
emulsions (as Fe(III)).
• Produces unwanted colorization of glasses (e.g., La O -B O lenses)
2 3 2 3

• Reduces precipitation efficiency of REE oxalates.

Si • Forms gelatinous phases in aqueous streams and solvent • Reduces magnetic moment in NdFeB magnets.
extraction circuits. • Reduces magnetostriction in Terfenol-D.
• Makes clarification difficult.
Th, U • Some co-precipitation with REE double sulfates or oxalates • Bears some radioactivity, exposure usually strictly regulated.
(Th). • Forms undesirable intermetallics in some magnetocaloric materials (Th).
• Reduces REE loading in solvent extraction.
• Strict regulations for tailings disposal.
Ti, Zr, Hf, V, Nb, Ta, Cr, • Hydrolyzes and forms crud in solvent extraction circuits. • Reduces Curie temperature, anisotropy field, and remanence of NdFeB
Mo, W • Poisons solvent, extractants, and modifiers. magnets.
• Produces unwanted colorization of glasses.
• Poisons REE fluid cracking catalysts (V).
Mn, Co, Ni, Cu • Deposition and corrosion of downstream equipment
(especially Cu).
• Reduces
NdFeB.
Curie temperature (except Co), anisotropy field, and coercivity of

• Higher valence Mn can cause same problems as refractory • Lowers saturation magnetization of SmCo magnets (Cu).
metals. • Decreases Curie temperature of magnetocalorific materials (Mn in Gd).
• Reduces precipitation efficiency of REE oxalates (Cu). • Produces unwanted colorization of glasses.
Zn, Ga, Ge • Forms crud during solvent extraction (Zn). • Significant decrease in Curie temperature of SmCo magnets (Ga) and
• Co-extraction issues in highly acidic chloride environments magneticaloric materials (Ge in Gd).
(Ga). • Slight reduction in remanence of NdFeB magnets.
Sn, Pb, Bi • Forms crud during solvent extraction. • Traps Nd in intermetallics causing a large reduction in coercivity of NdFeB
• Reduces precipitation efficiency of REE oxalates. magnets.
• Concentrates in molten salt electrolytes during REE
electrowinning (especially Pb)

naphthenic acid circuits, aluminum and iron(III) are preferentially ex­ Mo, W) also tend to hydrolyze and form crud in solvent extraction
tracted before REEs, while in the scrubbing stage they are the first to be circuits (Ritcey, 1980; Naylor and Eccles, 1988). Frequently, the for­
stripped and work their way back down to the feeding stage (Zhang mation of crud is directly a result of the composition of the feed liquor
et al., 2016). As they progressively accumulate in the solvent extraction to solvent extraction circuits (Ritcey, 1980). Other elements, like
circuit, they eventually hydrolyze and form gelatinous hydroxides that thorium and uranium do not form much crud, but do load in the sol­
cause emulsions during operation that reduce contact area and greatly vent; thereby reducing recovery of REEs.
impede extraction (Ritcey, 1980). Silicon is another element forming Impurities also affect REE recovery and precipitation efficiencies. In
gelatinous precipitates which co-extracts in some newly explored REE the presence of sulfate, excess calcium in solution can precipitate as
deposits based on eudialyte (Davris et al., 2017). Other major compo­ gypsum with some co-precipitation of REEs; thus, reducing REE re­
nents of the crud formed in solvent extraction include calcium (as covery (Dutrizac, 2017). Similarly, the presence of thorium in solution
gypsum) (Ritcey, 1980). The refractory metals (Ti, Zr, Hf, V, Nb, Ta, Cr, interferes with the selective precipitation and co-precipitates with the

2
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Fig. 2. Trends in REE production from 2010 to present: (a) global production of REOs and (b) price of various REE products. Data collected from U.S. Geological
Survey, 2011-2020.

REE double sulfates or oxalates (Pilkington and Wylie, 1947; Verbaan polishing slurry and affects the polishing rate and surface finish in
et al., 2015). Although aluminum, iron, copper, and lead do not co- precise glass surface planarization. Impurities in REE polishing com­
precipitate with REE oxalates, they reduce the precipitation efficiency pounds can also contaminate the polished surfaces (Lucas et al., 2015).
by consuming the oxalic acid via complexation in solution before REEs Besides magnets and glasses, other advanced REE materials are af­
precipitation (Dahlberg et al., 2014; Xia and Griffith, 2018). During fected by impurity elements including magnetostrictive alloy Terfenol-
REE carbonate precipitation, aluminum is known to hydrolyze and form D (Kwon et al., 2006), gadolinium-based magnetocaloric materials
a colloidal hydroxide that interferes with the formation of REE carbo­ (Gschneidner Jr. et al., 2005; Franco et al., 2018), REE lasers, phos­
nate crystals (Zuo et al., 2014). Other impurities with comparatively phors, and battery materials (Krishnamurthy and Gupta, 2016). Some
positive reduction potentials (like Cu) can electrodeposit on equipment, refractory metals (especially V) poison REE fluid cracking catalysts
piping, valves, or in regeneration circuits and lead to galvanic corro­ (Wormsbecher et al., 1986). In any product or concentrate, the presence
sion. Impurities also affect downstream REE metal or alloy production of naturally occurring actinides (Th, U) is usually strictly regulated
by molten salt electrolysis in chloride or fluoride melts (Eisele and (Tiefenbach et al., 2005).
Bauer, 1974). Alkaline earths and lead gradually build up in the bath, Considering the ever increasing stringent REE material require­
altering its properties and spoiling the electrolyte over time. Phos­ ments and the recent significant changes in the REE industry, there is an
phorus impurity fed to molten salt cells enters REE metals or alloys and urgent need to review impurity removal during REE processing. Prior
gradually releases phosphine, a toxic and flammable gas. reviews are available on general hydrometallurgical processing of REEs
During downstream REE metal or alloy production, many impurities (Jha et al., 2016; Krishnamurthy and Gupta, 2016), REE processing by
present in the feed material are often quantitatively transferred to the solvent extraction (Xie et al., 2014; Zhang et al., 2015; Hidayah and
metallic phase (Gschneidner Jr. and Daane, 1988). Therefore, im­ Abidin, 2018), ion exchange and adsorption (Hidayah and Abidin,
purities present in liquors or concentrates often report to finished REE 2017). However, no prior review exists on impurity removal during
magnets where they have an especially pronounced effect. Small REE processing.
amounts of aluminum, silicon, refractory metals (Ti, Zr, Hf, V, Nb, Ta, In this work, a comprehensive and systematic literature review on
Cr, Mo, W), or the 1st d-block metals (Mn, Co, Ni, Cu) have measurable recent progress (since 2015) in impurity removal during REE processing
effects on the properties of NdFeB magnets, including the Curie tem­ is presented. Original research articles covering techniques including
perature, anisotropy field, and coercivity (Buschow, 1991; Fidler and solvent extraction, ion exchange and adsorption, precipitation, and
Schrefl, 1996; Shaaban, 2007; Salazar et al., 2016). Heavy metals (Sn, others have been reviewed in detail and critically assessed to under­
Pb, Bi) are especially troublesome in NdFeB magnets as these elements stand the role and behaviour of impurities in REE processing. Specific
sequester neodymium in intermetallics and cause a large reduction in challenges related to iron, actinide (Th, U), and other impurities (Al, Zr,
coercivity (Van Mens, 1986). In some cases, the negative effect of Si etc.) are discussed. A comparative summary is presented in the dis­
certain elements on the bulk properties of magnetic phases are accepted cussion section.
to bring about an overall better microstructure; however, this should be
a discretion that rests with the magnet manufacturers, not dictated by 2. Review of the REE landscape since 2015
the REE supply.
REE-based optical glasses of the La2O3-B2O3 system have stringent Since the REE market is comparatively small, it is easily disrupted
requirements for refractive index and transparency, requiring ex­ and much more volatile compared with other commodity metals. A
ceptionally pure lanthanum (Riker, 1981). Transition metals such as defining moment in the REE industry occurred in 2010 when the
chromium, cobalt, and nickel are tolerated < 1 ppm, while iron must Chinese government reduced their export quotas of REEs by 40% under
remain < 10 ppm. In these glasses, manganese and iron absorb ultra­ the pretense of environmental protection (Tse, 2011). As presented in
violet radiation, causing solarization of the glass (Riker, 1981). In X-ray Fig. 2, the Chinese controlled 97% of all rare earth oxide (REO) pro­
or radiological applications, these impurities lead to unwanted duction in the world at the time and brought about a severe rare earth
browning of the glass. Some REE compounds, predominantly CeO2, are crisis for the rest of the world (Gschneidner Jr., 2011). Chinese export
also used as glass or silicon wafer polishing compounds, whose superior quotas were eventually lifted in January 2015 (Sprecher et al., 2015).
action towards glasses involves a chemical-mechanical mechanism In the years following the rare earth crisis, many new REE projects
(Horrigan, 1981; Hedrick and Sinha, 1994; Borra et al., 2018). The were launched outside of China, but most of these remained in the
presence of basic (Ca, Mg) or acidic (Si) impurities alters the pH of the exploration or development stages (Machacek and Fold, 2014). Part of

3
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

the challenge in bringing alternative sources of REEs to market is the investigating the behaviour of impurities were selected for inclusion.
long-life cycle of mines and the changing REE landscape in the last 5 Relevant references within each article were also consulted and in­
years. Since 2015, market prices for REEs have continued to decline cluded where appropriate. Where multiple articles were available
(U.S. Geological Survey, 2011-2020), as shown in Fig. 2, further re­ reaching the same or similar conclusions, we opted to include the
ducing profit margins under which new REE projects must operate. seminal works.
Despite the adversity, the overall market share for REE producers
outside China has been growing, mostly in the USA and Australia. 4. Impurity removal during REE processing
In 2015, the USA produced 5900 t of REO at Molycorp's Mountain
Pass mine in California (U.S. Geological Survey, 2011-2020). Later that Impurity removal is an operation that can take place anywhere in
year, as REE prices dropped, Molycorp filed for bankruptcy, and the the REE processing flowsheet. In some cases, it begins with selective
mine ceased production (Zillman, June, 2015; Hals, January 8, 2016). leaching of an ore or secondary source to limit the co-extraction of
In 2017, the mine was acquired by a capital group, which operates the undesirable impurity elements. In other cases, impurity removal is
mine under the name MP Mine Operations LLC (Topf, June 16, 2017; conducted after leaching of REEs, before a mixed REE concentrate is
Egbaria, March 18, 2019). The Mountain Pass mine resumed operations produced. In another alternative, impurity removal can be performed
in 2018 where it produced 18,000 t of REO that year and subsequently after a mixed REE concentrate is dissolved or during individual REE
26,000 t of REO in 2019 (U.S. Geological Survey, 2011-2020). separation by solvent extraction or ion exchange. Sometimes impurity
Australia's market share in REO production has been steadily removal is conducted after a REE metal or alloy production to remove
growing since 2015. Lynas Corporation operates its Mount Weld mine residual interstitials by vacuum melting, distillation, or electrochemical
that produced 21,000 t of REO in 2019 (U.S. Geological Survey, 2011- techniques (Zhang et al., 2015). Impurity removal may also be con­
2020). Lynas Corporation had mined the REE containing ore and pro­ ducted during some or all of these stages and more. No prior review has
duced a rare earth concentrate on site which was then shipped to Ma­ focused on impurity removal during REE processing, and this study
laysia for processing and impurity removal (Ganguli and Cook, 2018). aims to fill this critical gap.
However, their processing plant faced continued heavy criticism in The present work focuses on removal of impurities present in the
Malaysia over its environmental impact (Westbrook et al., September gangue materials in ores or secondary sources because these are the
24, 2018). Recently, Lynas has announced that ore cracking and principal source of impurities in REE processing. Therefore, most of
leaching operations will move to Western Australia, while downstream these impurity removal processes are conducted in operations down­
processing and separation will remain in Malaysia (Lynas Corporation, stream from cracking/chemical beneficiation and involve hydro­
May 21, 2019). metallurgical treatment of REE-containing aqueous solutions. Here, the
The market share from other countries has also been increasing. In principle methods of impurity removal generally fall into one of the
2019, some amounts of REOs were produced in India (3000 t), Russia following techniques:
(2700 t), Thailand (1800 t), Vietnam (900 t), and Myanmar (Burma)
(22,000 t) (U.S. Geological Survey, 2011-2020). Some of these statistics 1. Purification techniques based on solvent extraction
may also include illegally mined REOs originating from China. 2. Purification techniques based on ion exchange or adsorption
The future appears optimistic as new REE projects come online and 3. Purification techniques based on selective precipitation
existing projects ramp up production outside of China. As of 2019,
China accounts for only 63% of the total REE production in the world. or
However, advancing new or existing projects outside of China, in the
face of declining market prices, requires the development of sophisti­ 4. Other novel purification techniques which are not listed above.
cated separation and impurity removal processes, which are reviewed
in the following sections. Fundamentally, these techniques are similar in that they all aim to
be selective. Impurity removal processes must be either selective to­
3. Methodology wards extracting REEs or selective towards extracting certain impurities
– but it cannot be both. As it will be discussed further, different tech­
The present review aims to systematically review recent progress in niques and types of selectivity are better suited for cases where im­
impurity removal during REE processing and to identify best practices purities are more concentrated with respect to REEs or vice versa.
for their removal. We considered REEs here to include scandium, yt­ A comprehensive examination of each impurity removal approach is
trium, and the lanthanides (Ln) lanthanum through lutetium, except presented under its own subsection in the following text. In total, 87
promethium which is radioactive and not encountered in primary or original research articles published since 2015 were reviewed in detail.
secondary REE sources. Impurities here cover most of the periodic table Impurity removal techniques employed by these articles are summar­
besides REEs, including alkali metals, alkaline earth metals, transition ized in Fig. 3a. Most publications since 2015 have focused on ion ex­
metals, post-transition metals, actinides, and metalloids. Phosphorus change or adsorption, followed by solvent extraction, selective pre­
and sulfur were the only non-metals considered in detail. Any method cipitation, and other techniques. The intensification of research towards
for impurity removal and any (primary or secondary) source of REEs ion exchange and adsorption may be driven by the higher relative pu­
were considered. Although emphasis is on impurities present in the rities and separation factors realized using this technique. Impurities
gangue materials in ores or secondary sources, both commercial and encountered during processing are outlined in Fig. 3b. Most impurities
synthetic REE feeds were considered. Including the latter was relevant encountered include iron, thorium, uranium, and aluminum, as these
because many articles using synthetic feeds covered the same impurities originate in many REE ores. However, research is also intensifying to­
present in the gangue materials of the commercial REE sources. wards treatment of impurities like cobalt, nickel, copper, and zinc,
A comprehensive literature search was conducted using a sub­ which are present in secondary REE sources (e.g., coatings for NdFeB
scription database (SciFinder, Chemical Abstracts Service: Columbus, magnets).
OH). Original research articles published since 2015 relating to REEs
and either solvent extraction, ion exchange, adsorption, precipitation, 4.1. Solvent extraction
impurity removal, or separation were identified by searching the da­
tabase. Rather than searching for specific impurities, each article Different types of solvent extraction operations are performed
meeting the above criteria was examined to ascertain whether the be­ during REE processing (Xie et al., 2014; Zhang et al., 2015). The first
haviour of any impurities was investigated or not. Only articles (primary) solvent extraction fulfills the need for separation and

4
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Fig. 3. Overview of original research papers on impurity removal during REE processing reviewed since 2015 (a) impurity removal techniques and (b) impurities
encountered.

concentration of REEs from impurities in leach solutions. This is also 1. Extractant (functions as a cation/anion exchanger or solvating ex­
used to transfer REEs from, for instance, sulfuric leachates to chloride tractant)
solutions for subsequent secondary solvent extraction, where individual 2. Diluent (secondary immiscible phase, usually organic)
REE separation is conducted. Co-extraction of impurities must be 3. Targeted elements in solution (usually REEs)
minimized in primary solvent extraction to ensure clean solutions for 4. Other elements present in solution (Fe, Al, Ca, etc.)
secondary operations. Behaviour of impurities in secondary solvent 5. Feed material (chemistry, concentration, synthetic or ore, etc.)
extraction is also important because they can build up and spoil the 6. Summary and comments (separation efficiencies, recoveries, chal­
operations. lenges, etc.)
Acidic extractants (cation exchangers) are the most common used
extractant for REE extraction from aqueous solution. Two classes of The results are summarized and presented in chronological order in
acidic extractants are used which include carboxylic (fatty) acids or Table 2. No previous review has focused on the behaviour of impurities
organophosphorus acids. The structure of some extractants is provided during solvent extraction of REEs. However, reviews are available on
in Appendix A. Cationic exchange of REEs between aqueous and or­ REE separation by solvent extraction (Xie et al., 2014; Zhang et al.,
ganic phases occurs according to reaction (1) (Xie et al., 2014): 2015; Hidayah and Abidin, 2018).
Recently, most research has focused on the use of organic diluents
REE 3 + + 3H2 A2 = REE (HA2 )3 + 3H+ (1)
and extractants, which function as cation exchangers towards REEs in
where A denotes an organic anion (cf. Appendix A) and the overbar aqueous solutions. Different feed materials containing REEs were used
indicates species in the organic phase. Notice that extractants tend to for solvent extraction techniques as summarized in Fig. 4. Most research
form dimers to reduce their polarity in non-polar organic solvents. After has focused on extraction of REEs from commercial materials, such as
physical separation of the aqueous and organic phases, REEs are ore leachates, rare earth concentrates, and leachates from secondary
stripped from the organic phase into a new aqueous phase by the re­ sources (e.g., permanent magnets), and less has focused on REE ex­
verse of reaction (1). traction from synthetic solutions. The majority of investigations con­
Besides acidic extractants, REE solvent extraction has been per­ sidered the extraction of REEs from chloride-based solutions, followed
formed with neutral (solvating) extractants and basic extractants (an­ by nitrate, sulfates, phosphates, and other solutions (Fig. 4). All these
ionic exchangers). The structure of some extractants are included in feed solutions have commercial significance, since many REE ores and
Appendix A. Chelating extractants are not usually used in REE proces­ secondary sources are leached with mineral acids.
sing because they do not perform as well as acidic extractants (Xie et al., Many new developments in solvent extraction processing have oc­
2014). Neutral and basic extraction of REEs are more effective in sulfate curred since 2015. Combining certain extractants has offered sy­
or nitrate solutions, where there is higher tendency for ion association. nergistic effects that improve extraction, selectivity, phase separation,
An example of neutral extraction of REEs in nitrate media by tributyl and stripping properties (Zhou et al., 2019). Moreover, the use of mi­
phosphate (TBP) is given in reaction (2): crofluidics has been shown to achieve enormously high separation
factors in extremely short times (Chen et al., 2017); however, the
REE 3 + + 3NO3 + 3TBP = REE (NO3 )3 (TBP )3 (2) throughput of microfluidic processes would pose a challenge for large
Basic extraction involves the exchange of anionic complexes be­ scale REE projects. Membrane-assisted solvent extraction is another
tween the aqueous and organic phases. An example of basic extraction promising development that can drastically reduce the amount of stages
of REEs in sulfate media by primary amines is given in reactions (3) and necessary for REE extraction (Kim et al., 2015). There is also growing
(4): interest in the concept of selectively stripping certain elements from the
organic phase which creates an additional partitioning step between
2RNH2 + H2 SO4 = (RNH3 )2 SO4 (3) REE and impurities (Ye et al., 2019). As well, the design and synthesis
of new extractants engineered for impurity removal is promising (Lu
2REE (SO4 )33 + 3(RNH3 ) 2 SO4 = 2(RNH3 )3 REE (SO4 )3 + 3SO42 (4) et al., 2016). Despite these new developments, the removal of im­
The REEs are stripped from neutral or basic extractants using mi­ purities, such as iron, thorium, uranium, and aluminum remains a
neral acids or solutions of REE complexing agents. challenge.
Original research articles pertaining to impurity removal by solvent
extraction since 2015 were compiled and systematically categorized 4.1.1. Behaviour of iron
according to: The behaviour of iron during REE processing varies with different

5
 Table 2
Summary of original research articles on impurity removal during REE processing using solvent extraction.a Separations were performed at room temperature (20–30 °C) unless otherwise indicated. Cases where all
naturally occurring lanthanides (La to Lu, except Pm) in solution are followed are denoted by Ln.
Investigators Extractant(s) Diluent(s) Targeted element Other element(s) in Feed material(s) Summary and comments
(s) in solution solution
W.D. Judge and G. Azimi

Wang et al. (2019) USTB-2, INET-3, Cyanex 272, n-octane Ln Al, Zn, Co, Cu, Na Synthetic (chloride) Developed a new cation exchange extractant with high
P507 1.08 mM REEs separation factors for LREEs/HREEs. In mixed solutions, Al and
1.62 mM impurities Zn co-extract with REEs, but Co and Cu remain in the raffinate.
Ye et al., (2019) P204 Kerosene Sc Fe, Al, Ca, Ti Bauxite residue leachate Achieved very high separation factors for Sc/Fe and Sc/Al, but
(phosphate) difficulties encountered with separating Sc and Ca. HCl was used
23.62 mg/L Sc to selectively strip Ca, followed by NaOH to recover Sc.
8900 mg/L impurities
Zhou et al., (2019) Cyanex 572, ODP Heptane Th Fe, REEs Industrial mineral waste Developed a high selectivity process for Th compared with Fe and
residue leachate (nitrate) REEs. Separation factors up to 7.6 for Th was achieved. Stripping
25 μM REEs was performed with HCl or HNO3.
1.41 μM Th, 936 μM Fe
25–45 °C
Amaral et al., (2018) Primene JM-T, Alamine 336 Dodecane with 5% (v/v) U, Th REEs, Al, Fe, Mn, Brazilian industrial residue Used two extractants to simultaneously extract nearly all Th and
tridecanol Ca, Si, Zn leachate (sulfate) U. Separation factors of 145 and 292 were achieved for Th/REE
9.1 g/L REEs and U/REE, respectively. Some Fe was co-extracted during
27.5 g/L impurities solvent extraction.
Battsengel et al., D2EHPA, TBP, TBPP, TOA, TOPO Kerosene Y, La, Ce, Pr, Nd, Ca, P, Si, Al, Fe, Sr, Mongolian apatite ore Preferentially extracted HREEs up to 93% with low co-extraction
(2018) Sm, Eu, Gd, Dy, Er S, Mg, Na, K, F leachate (sulfate) of LREEs and Ca. Some co-extraction of Al and Fe occurred, but
1.16 g/L REEs these remained in the organic phase after stripping with H2SO4.
0.691 g/L impurities
Hiskey and Copp, Primene JM-T Kerosene Y, Nd Al, Ca, Cd, Co, Cu, Arizona copper pregnant Sulfate concentration does not inhibit extraction. Stoichiometry
2018a) Fe, K, Mg, Mn, Na, leach solution (sulfate) shows that bi and tri sulfate complexes are responsible for REE
Ni, P, Zn 18–56 mg/L REEs extraction. They preconditioned organic phase by contacting

6
2.59–3.40 g/L impurities with 1 M H2SO4.
Hiskey and Copp, Primene JM-T Kerosene Y, Ln Al, Ca, Cd, Co, Cu, Arizona copper pregnant Analyzed extraction of REEs in terms of the sulfate complexes
2018b) Fe, K, Mg, Mn, Na, leach solution (sulfate) formed in the solution. Did not analyze extracted solutions for
Ni P, Zn 29.3 mg/L REEs non-REEs. Stripping was performed with 3 M HCl or 3 M HNO3.
26.0 g/L impurities
Innocenzi et al., Cyanex 272, Cyanex 572, Kerosene Y, La, Ce, Eu, Gd, Ca, Si, Na Fluorescent lamp leach The highest separation factors were achieved using D2EHPA and
(2018) D2EHPA Tb solution (sulfate) a solution pH < 1. No co-extraction of impurities was noted.
6290 mg/L REEs Stripping was performed with H2SO4.
1270 mg/L Ca
Dong et al., (2017) ODP, P507, TBP, Cyanex 272 Toluene Th Fe, Y, Ln Ion adsorption waste residue ODP showed the highest selectivity towards Th. Over 90% Th was
leachate (nitrate or chloride) extracted with virtually no REE or Fe co-extraction. Stripping was
19.9 mM REEs performed using HCl or H2SO4.
6.13 μM Th, 9.9 mM Fe
Gergoric et al., TODGA Solvent 70, Hexane, Gluene, Pr, Nd, Dy Na, Fe, Al, Co, Cu, Synthetic (nitrate) The type of diluent had a significant effect on the distribution
(2017a) Cyclohexanone, 1-octanol Mn, Ni, B 4.08 mM REEs ratio of elements. The best impurity removal was achieved with
21.49 mM Fe Solvent 70. Cyclohexane performed the poorest, co-extracting
NdFeB leachate (nitrate) both Al and B.
~1100 mg/L REEs
~2900 mg/L impurities
Gergoric et al., D2EHPA Solvent 70, hexane, octane, Nd, Dy, Pr, Gd Co, B NdFeB leachate (sulfate) No Fe was present in the magnet leachate. Hexane, octane, and
(2017b) cyclohexane, chloroform, 1- 15.7 mM REEs Solvent 70 were considered the best performing diluents. No
octanol, toluene 0.72 mM impurities measurable co-extraction of Co or B occurred. Stripping was
performed with HCl.
Huang et al., (2017) HEHEHP, HDEHP Hydrogenated kerosene (DT-100) Y, Ln Mg, Al, Fe, Ca Synthetic ore leachate Utilized a stepwise solvent extraction procedure to recover REEs.
(sulfate) HREEs could be selectively extracted by controlling the pH. Mg
1.95 g/L REEs co-extracted with LREEs. The Al or Fe did not accumulate with
4.24 g/L impurities continued cycling.
(continued on next page)
Hydrometallurgy 196 (2020) 105435
 Table 2 (continued)

Investigators Extractant(s) Diluent(s) Targeted element Other element(s) in Feed material(s) Summary and comments
(s) in solution solution

Peng et al., (2017) DMDDDGA, T2EHDGA, TODGA Kerosene U, Fe Nd Synthetic (chloride) TODGA showed the highest simultaneous extraction of Fe(III)
32 g/L Nd2O3 and U(VI). Introducing chloride ions via NaCl increased the
W.D. Judge and G. Azimi

1 g/L U, 1 g/L Fe distribution coefficient of U(VI).

Rey et al., (2017) DEHCNPB Isooctane La, Nd, Eu, Dy, Yb Fe Synthetic (phosphate, nitrate) Utilized a surfactant with a chelating extractant to achieve
250 mg/L REEs synergistic effects. High selectivity over Fe(III) was found at
2500 mg/L Fe particular ratios of surfactant/extractant. Separation factors for
REE/Fe were up to 65.
Wang et al. (2017b) Cyanex 572, P507 n-heptane, 260# kerosene with Th Ln, Fe Synthetic (chloride) Increasing HCl concentration increased Th/REE separation
and without phase modifiers 0.021–0.224 M REEs factors. Modifying the diluent with isodecanol improved the
0.022–0.060 M Th separation factors of Th/REE. Lutetium was the most difficult
Low level radioactive waste REE to separate from Th.
residue leachate (chloride)
0.079–0.35 M REEs
5–10 g/L impurities
Zhao et al., (2017b) D2EHPA, HEHEHP Kerosene (sulfonated) La, Ce Al, P, Na, Fe, Si, K Spent fluid cracking catalyst D2EHPA was the superior extractant for REEs. Some co-
leachate (chloride) extraction of Al occurred with both extractants, most of which
2080 mg/L REEs stripped with the REEs upon addition of HCl.
13,300 mg/L impurities
Kolar et al. (2016) Cyanex 572 Shellsol D70 Y, Ln Fe, Al Commercial ore concentrate Performed a microfluidic solvent extraction process which had
leachate (chloride) fast extraction time (15 s) and high selectivity for HREEs. The Al
60 g/L REEs and Fe were precipitated by pH adjustment prior to solvent
0.1 g/L Fe extraction.
Lu et al., (2016) 2-ethylhexyl-N,N-di(2- n-heptane Th La, Pr, Nd, Sm, Gd, Synthetic (nitrate, chloride, Designed/synthesized a new phosphorodiamidate extractant.
ethylhexyl) phosphorodiamidate Yb and sulfate) Thorium extraction was highest in nitrate solution followed by
46.7 mM REEs chloride and sulfate. Some REEs co-extracted with Th. Stripping

7
0.192 mM Th Th was difficult and required 6.5 M H3PO4.
Kim et al. (2015) Cyanex 923, TODGA, TBP Isopar L Nd, Pr, Dy Fe, B, Co, Ga NdFeB leachate (nitrate and Utilized membrane assisted solvent extraction to extract REEs in
chloride) a single step. Impurities were better removed from nitrate feed
1000–2000 mg/L Nd compared with the chloride feed.
2500–6000 mg/L impurities
Mowafy and TEHDGA 75% (v/v) n-dodecane, 25% (v/v) Nd, Sm, Dy Fe, Ni, Cs Synthetic (chloride, nitrate, Significant Fe was co-extracted in the chloride system, while Ni
Mohamed (2015) n-octanol and sulfate) and Cs were not co-extracted under any conditions. Any co-
6 mM REEs extracted Fe was stripped with REEs.
6 mM impurities

a
Alamine 336 = tri-octyl/dodecyl amine; Cyanex 272 = di-(2,4,4′-trimethylpentyl) phosphinic acid; Cyanex 572 = proprietary mixture of phosphinic and phosphonic acid; Cyanex 923 = proprietary mixture of
trialkylphosphine oxides; D2EHPA = di-(2-ethylhexyl)phosphoric acid; DEHCNPB = butyl-N,N-di(2-ethylhexyl)carbomoylnonylphosphonate; DEHEHP = di(2-ethylhexyl) 2-ethylhexyl phosphonate; DMDDDGA = N,N′-
dimethyl-N,N′,-didodecyl diglycolamide; HDEHP = di-(2-ethylhexyl)phosphoric acid; HEHEHP = 2-ethylhexylphosphonic acid mono-2-ethylhexyl ester; INET-3 = (2,3-Dimethylbutyl) (2,4,4′-trimethylpentyl)phos­
phinic acid; Isopar L = synthetic isoparaffinic hydrocarbon solvent; ODP = n-octyl diphenyl phosphate; P204 = di-(2-ethylhexyl)phosphoric acid; P507 = 2-ethylhexylphosphoric acid mono-2-ethylhexyl ester; Primene
JM-T = tri-alkyl methylamine; Shellsol D70 = hydrocarbons C11-C14, n-alkanes, isoalkanes, cyclenes, ≤0.02 wt% aromatics; Solvent 70 = hydrocarbons C11–C14, ≤0.5 wt% aromatics; T2EHDGA = N,N,N′,N′-tetra(2-
ethylhexyl) diglycolamide; TEHDGA = N,N,N′,N′-tetra(2-ethylhexyl) diglycolamide; TBP = tributyl phosphate; TBPP = tributylphosphine; TOA = trioctylamine; TOPO = trioctylphosphine oxide; TODGA = tetraoctyl-
diglycolamide; UTSB-2 = (2,4-dimethylheptyl)(2,4,4′-trimethylpentyl)phosphinic acid.
Hydrometallurgy 196 (2020) 105435
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Fe3 + + H+ + 4Cl + S = HFeCl4 S (7)

where S is a solvating extractant such as TBP, Cyanex 921, or Cyanex
923. Peng et al., 2017 took advantage of this tendency to preferentially
extract iron(III) at high concentration (1 g/L) from REEs in chloride
solutions using solvating extractants. Of the three extractants studied,
TODGA performed the best for simultaneously extracting iron and ur­
anium, while minimizing REE co-extraction. Peng et al., 2017 also
found that impurity removal was enhanced by introducing chloride via
Fig. 4. Summary of feed type for solvent extraction articles in Table 2: (a) NaCl, further validating that chlorocomplexation occurs. Regarding
source of feed and (b) major anion in feed. chelating extractants, surfactants have been used to significantly reduce
iron(III) co-extraction (Rey et al., 2017).
extractants and solution chemistries. Although the ratio of Fe(III)/Fe(II) Basic extractants are usually used in sulfate or nitrate solutions.
was not followed in some investigations, iron(III) is usually the co-ex­ Hiskey and Copp, 2018a found that iron was always co-extracted with
tracting species (Zhang et al., 2016). Iron(II) does not usually extract as REEs using Primene JM-T in sulfate solutions. It was also found that
strongly as iron (III) or REEs with acidic extractants (Islam et al., 1979; iron reduced the loading of REEs in the organic phase. Iron probably
El-Nadi, 2017). competes with REEs as it has a tendency to form sulfate complexes.
Acidic organophosphorous extractants, typically di-(2-ethylhexyl) To summarize, an effective way to selectively extract REEs from iron
phosphoric acid (known as D2EHPA, DEHPA, HDEHP, or Chinese brand containing solutions is through acidic extractants like D2EHPA (Ye
P204), are widely used for REE extraction. Analyses of the literature et al., 2019) or HEHEHP (Huang et al., 2017) with appropriate con­
suggests that REEs are preferentially extracted over iron using acidic sideration of extractant concentration and O/A ratio. Solvating or basic
organophosphorous extractants (Ye et al., 2019; Battsengel et al., 2018; extractants are not effective for separating REEs and iron. Alternatively,
Huang et al., 2017; Wang et al., 2010). However, iron(III) may still be iron can be selectively extracted from chloride media using solvating
co-extracted when its concentration is high, extractant concentration is extractants. The effect of contact time was not considered in most in­
high, or when the phase ratio of organic to aqueous (O/A) is high. vestigations, but it is known that reducing contact time greatly sup­
Variations in ferric concentration, O/A, and extractant concentration presses ferric extraction (Ritcey and Lucas, 1969). In fact, commercial
reconcile the differing co-extractions of iron in recent investigations (Ye circuits extracting yttrium with D2HEPA and TBP have operated with
et al., 2019; Battsengel et al., 2018; Huang et al., 2017; Wang et al., only 15 s mixing time to minimize ferric extraction (Goode, 2013).
2010). Increasing O/A or extractant essentially means excess free li­
gands being available beyond what is required for REE extraction. Ex­ 4.1.2. Behaviour of thorium and uranium
traction of iron(III) by D2EHPA occurs according to reaction (5) (Islam Thorium and uranium are known to associate with REE miner­
et al., 1979): alization and are usually present in most commercial REE ore materials.
The usual approach to impurity removal by solvent extraction is to
Fe3 + + 3H2 A2 = Fe (HA2 )3 + 3H+ (5) perform a separate solvent extraction step specifically engineered to
which is the analog of reaction (1) involving extraction of REEs. Al­ selectively extract thorium or uranium while minimizing REE co-ex­
ternatively, hydrolyzed iron(III) may extract according to reaction (6) traction. Here, the higher tendency of thorium and uranium to form
(Islam et al., 1979): complexes compared with REEs is exploited to achieve selectivity. It is
generally more difficult to separate the actinides from heavier REEs like
Fe (OH )2 + + H2 A2 = Fe (OH ) A2 + 2H+ (6) lutetium (Wang et al., 2017b).
In most cases, basic extractants that function as anion exchangers
Evidence for competition between iron(III) and REEs is apparent
towards thorium and uranium (e.g., amines) are employed. Typically,
from the reduction in REE loading at high iron(III) concentrations
tertiary amines are more selective towards uranium, while primary
(Wang et al., 2010). It may seem advantageous to operate acidic or­
amines are more selective towards thorium in sulfate media. In sulfate
ganophosphoric circuits close to saturation to minimize iron(III) co-
solutions, extraction of thorium by primary amines such as Primene JM-
extraction and improve REE recovery, but this will increase solvent
T proceeds according to reaction (8) (Shimidt, 1971; Amaral and
viscosity and lead to problems with crud formation (Ritcey, 2004).
Morais, 2010; Amaral et al., 2018):
Besides organophosphorous compounds, carboxylic acids are also used
as acidic extractants for REEs. Naphthenic acid (a common carboxylic Th (SO4 )44 + 4RNH2 H+HSO4 = (RNH2 H+)4 Th (SO4 )44 + 4HSO4 (8)
acid extractant) preferentially extracts iron(III) over REEs (Zhang et al.,
2016). Extraction of uranium(VI) by tertiary amines in sulfate solutions
Although iron may be co-extracted with REEs, there is an oppor­ proceeds according to reaction (9) (Shimidt, 1971, Amaral and Morais,
tunity to partition that by selective stripping, i.e., the pH at which re­ 2010, Amaral et al., 2018):
action (5) or reaction (6) proceeds in reverse is often different from the
UO2 (SO4 )34 + 4R3 NH+HSO4 = (R3 NH+)4 UO2 (SO4 )34 + 4HSO4 (9)
pH at which reaction (1) proceeds in reverse. Sulfuric acid has been
used to selectively strip REEs from D2EHPA, while leaving iron in the Quaternary amines are also effective for selective extraction of ur­
organic phase (Battsengel et al., 2018). In naphthenic acid circuits, the anium(VI). Secondary amines, however, perform comparatively poor
reverse is true. Iron(III) is stripped prior to REEs, which poses problems for both thorium and uranium(VI) (Zhu et al., 2015). Although it is
in counter-current operations as iron(III) works its way to the feeding usually difficult to remove both thorium and uranium at the same time
stage (Zhang et al., 2016). in a single step, a combination of primary and tertiary amines is ef­
Solvating extractants are sometimes used for REE extraction, usually fective at simultaneously extracting them from REEs (Amaral et al.,
from sulfate or nitrate solutions. In nitrate solutions, TODGA, Cyanex 2018). The alternative approach has been to remove uranium(VI) by
923, and TBP have been used to extract REEs with minimal co-extrac­ tertiary or quaternary amines, followed by thorium removal using
tion of iron (Kim et al., 2015; Gergoric et al., 2017a). Difficulties with primary amines or acidic ogranophosphorous extractants (Zhu and
high iron co-extraction are encountered in chloride solutions, compared Cheng, 2011; Zhu et al., 2015).
with nitrate or sulfate (Kim et al., 2015; Mowafy and Mohamed, 2015). Solvating extractants are also effective at extracting thorium and
The reason is due to iron's tendency to form chlorocomplexes that are uranium from REEs in nitrate, chloride, and sulfate solutions. Using
required for extraction according to reaction (7) (Mishra et al., 2010): TBP with nitrate media, uranium(VI) is preferentially extracted over

8
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

thorium and thorium is preferentially extracted over REEs (Menzies and ligands are available from high extractant concentrations or O/A ratios.
Rigby, 1961; Dong et al., 2017). Cyanex 923 is another solvating ex­ Extraction of aluminum from sulfate solutions with D2EHPA is de­
tractant that has been effective at separating uranium(VI) and thorium scribed by reaction (13) (Mashimo et al., 1997):
from REEs. In nitrate solutions, selective extraction of thorium and
Al3 + + 2H2 A2 = AlA3 HA + 3H+ (13)
uranium(VI) occurs according to reactions (10) and (11), respectively
(Gupta et al., 2002): The REEs may be selectively stripped from co-extracted aluminum
on account of the different equilibrium constants for reactions (1) and
Th4 + + 4NO3 + 3R = Th (NO3 )4 3R (10)
(13) (Battsengel et al., 2018). Although this helps with initial separa­
tions, a build up of aluminum in the organic phase will eventually lead
UO22 + + 2NO3 + 2R = UO2 (NO3 ) 2 2R (11)
to crud formation.
where R is Cyanex 923. In sulfate solutions, Cyanex 923 is not effective Regarding naphthenic acid extractants (carboxylic acid), aluminum
for thorium extraction because of significant co-extraction of REEs (Jun is preferentially extracted and stripped ahead of REEs (Zhang et al.,
et al., 1998). Recently, another solvating extractant, ODP, has been 2016). Therefore, these extractants are not effective for extracting REE
demonstrated to selectively extract thorium with virtually no REE co- from aluminum-containing solutions. In counter-current operations,
extraction (Dong et al., 2017; Zhou et al., 2019). In nitrate media, aluminum works its way towards the feeding stage and results in crud
thorium is extracted according to reaction (12) (Dong et al., 2017): formation (Zhang et al., 2016). However, in an alternative process,
naphthenic acid extractant has been used to selectively extract alu­
Th4 + + 4NO3 + 2ODP = Th (NO3 )4 2ODP (12)
minum to less than 10 mg/L from REE processing solutions (Han et al.,
Separation factors with ODP are highest in nitrate medium com­ 2013).
pared with chloride and sulfate; a trend which follows the hydration Solvating extractants have recovered REEs from nitrate solutions
energies of anions (NO3− < Cl− < SO42−) (Dong et al., 2017). This is with little co-extraction of aluminum, provided the appropriate diluent
not surprising because it relates to the stability constant and ease of was used (Gergoric et al., 2017a). It was postulated that some diluents
metal-ligand complexes formation which are necessary for extraction. themselves are able to directly extract aluminum ions from the aqueous
With solvating extractants, the distribution coefficient of uranium(VI) phase. Moreover, dielectric properties of diluents are of particular im­
from chloride media is greatly enhanced by introducing chloride ions portance and may alter extraction mechanisms (Ritcey, 2004). The
via NaCl (Peng et al., 2017). behaviour of aluminum in chloride and sulfate media has not been
A different approach was taken by Lu et al., 2016 who designed and investigated because solvating extractants are not very effective at re­
synthesized a new phosphorodiamidate (solvating) extractant to selec­ covering REEs from these solutions (Battsengel et al., 2018).
tively remove thorium. A nitrogen element was incorporated into the Basic extractant Primene JM-T has been investigated for recovery of
phosphate structure to alter the electron density of the P]O group to REEs from sulfate solutions (Hiskey and Copp, 2018a). Aluminum was
target thorium. The highest extraction of thorium was in nitrate solu­ found to co-extract, likely because it formed similar sulfate complexes
tion followed by chloride and sulfate. Some REEs were co-extracted as REEs in solution. Competition between aluminum and REEs was
with thorium and stripping thorium from the organic phase was diffi­ apparent by the reduction in REE loading with co-extraction of alu­
cult and required 6.5 M phosphoric acid to recover 70% of thorium. minum (Hiskey and Copp, 2018a).
It is exceptionally difficult to selectively extract REEs from solutions
containing thorium and uranium owing to differences in charge density 4.1.4. Behaviour of base metals and other impurities
and electric field strength. The most common acidic extractants for Many base metals and other impurities hitherto not discussed ori­
REEs, i.e., acidic organophosphorous extractants and carboxylic acid ginate from gangue material in the ore or secondary sources and may be
extractants, extract thorium and uranium(VI) more strongly than REEs present in REE solutions. The behaviour of transition metals including
(Gupta and Singh, 2003; Zhu et al., 2015; Wang et al., 2017b). There­ manganese, cobalt, nickel, copper, and zinc must be considered owing
fore, it is feasible to selectively extract thorium or uranium from REEs to their presence in coatings or compounds in recycled NdFeB magnets.
using traditional acidic organophosphorous extractants like P507 or In most systems, none of the aforementioned elements were measurably
Cyanex 572 (Wang et al., 2017b). The better approach is to selectively co-extracted (Gergoric et al., 2017a; Gergoric et al., 2017b; Mowafy and
extract these impurities from REEs or to use a different impurity re­ Mohamed, 2015, and Kim et al., 2015). These experimental conditions
moval technique. cover a wide range of solution chemistries (sulfate, nitrate, chloride),
diluents, and extractants (acidic, solvating). One exception is the newly
4.1.3. Behaviour of aluminum developed cation exchange extractant (USTB-2) by Wang et al., 2019
Aluminum is a well-known troublesome impurity in REE processing which rejected most cobalt and copper, but tended to co-extract zinc.
by solvent extraction, as it accumulates and forms crud in solvent ex­ Owing to their larger charge density, REEs are always extracted
traction circuits. Therefore, it is important to understand the behaviour more strongly with acidic extractants than divalent transition metals
of aluminum impurities during REE processing. Aluminum behaves (El-Nadi, 2017). If excess free ligands are available by, perhaps, high
differently towards different extractants and in different types of solu­ extractant concentration or high O/A ratio, co-extraction of divalent
tions. metals is more probable. With acidic organophosphorous or carboxylic
Acidic organophosphorous extractants, such as D2EHPA, HEHEHP, acid extractants, divalent first row transition metals are extracted in the
and P507, used in chloride media tend to co-extract some aluminum order V < Cr < Mn < Fe < Co < Ni < Cu > Zn (El-Nadi, 2017).
(Zhao et al., 2017b; Wang et al., 2019). Stripping the organic phase Calcium does not usually co-extract with REEs during solvent ex­
with HCl simultaneously strips aluminum with REEs and contaminates traction with acidic extractants, except for the case of scandium op­
the final solution (Zhao et al., 2017b). Virtually no co-extraction of erations, perhaps because of their proximity on the periodic table. Ye
aluminum occurs in phosphate media with D2EHPA because of the et al., 2019 faced great difficulty with co-extraction of scandium and
formation of hydrophilic AlH2PO42+ions that are difficult to extract calcium from bauxite residue leachate. Rather than preventing the co-
from the aqueous phase (Ye et al., 2019). In sulfate media with acidic extraction, a two-stage stripping process was used to first selectively
organophosphorous extractants, aluminum has sometimes co-extracted strip calcium with HCl followed by the recovery of scandium using
with REEs and sometimes not (Huang et al., 2017; Battsengel et al., stripping with NaOH. If calcium co-extraction does occur, it may form
2018). The difference appears to result from dissimilar O/A ratios and crud in solvent extraction circuits or precipitate as gypsum. Magnesium
extractant concentrations. It appears REEs are preferentially extracted can co-extract with LREEs when it is present in high concentrations
over aluminum, but some aluminum co-extraction occurs if excess free (Huang et al., 2017). Selective stripping of magnesium would be an

9
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

interesting prospect to prevent contamination of the final REE product. where P represents the resin backbone (typically a polystyrene or
Alkali metals do not have any direct interference with solvent ex­ polyacrylic matrix) and R represents the anionic species of the resin
traction of REEs (Mowafy and Mohamed, 2015). With acidic ex­ (e.g., sulfonic groups –SO3−). Some common functional groups for ion
tractants, REEs are extracted much more strongly than alkali metal exchange are provided in Appendix B. The REEs are recovered by their
ions. Alkali metal salts are sometimes used with basic or solvating ex­ displacement or elution from the resin according to a reaction like the
tractants to increase the salting out effect and to help with complex reverse of reaction (14). Chelating agents, such as ethylenediaminete­
formation of REEs, thorium, or uranium. Many investigations listed in traacetic acid (EDTA), are used during elution to enhance REE se­
Table 2 took place in the presence of sodium or lithium, as salts of these paration because of varying stability constants of REE-chelates across
elements were used to adjust pH or the concentration of anionic species. the lanthanide series (Powell, 1979). Weak acid type resins work on the
Complications at high concentrations and in sulfate solutions could same principle; however, their functional groups do not ionize (dis­
include precipitation of rare earth double alkali sulphates or jarosite in sociate) to the same extent (e.g., carboxylic groups (–COO−). Less
the presence of iron. commonly, anion exchange resins have also been used for REE se­
Boron (as boric acid or borate) is a component of the leachate paration and the mechanisms behind REE anionic exchange are more
during recycling of NdFeB magnets. During solvent extraction of neo­ complex (Powell, 1979; Jha et al., 2016).
dymium, no co-extraction of boron occurs with acidic organopho­ These resins are most effective in salt solutions (e.g., lithium nitrate)
sphorous extractants (Gergoric et al., 2017b). However, with solvating as oppose to mineral acids (Marcus and Nelson, 1959).
extractants, boron was found to co-extract with neodymium, especially Adsorption is recently gaining attention for REE recovery because of
with cyclohexane or 1-octanol diluents (Gergoric et al., 2017a). Two its simplicity, low cost, and high efficiency for dilute solutions
explanations seem feasible; first, certain extractants may strongly as­ (Anastopoulos et al., 2016). Adsorption is a phenomenon closely related
sociate with borate leading to their transfer to the organic phase (si­ to ion exchange, but distinguished in that no ions are exchanged be­
milar to association of secondary amines with phosphate ions, Zhu tween the sorbent and the liquid phase. Ions or molecules from the
et al., 2015). Second, borate may have some intrinsic solubility in the liquid phase are bound to the surface of the sorbents by physical
diluents themselves (Gergoric et al., 2017a). Gallium is also present in (electrostatic) or chemical forces. Pure adsorption (i.e., without ion
some NdFeB leachates, but it did not pose a problem with solvating exchange) of REE cations from the solution may be simply described
extractants (Kim et al., 2015). according to reaction (15):
Silica is solubilized during leaching of fluorescent lamp powders;
REE 3 + = (REE 3+)ads (15)
however, it does not pose any issue with extraction of REEs (Innocenzi
et al., 2018). Perhaps in larger operations, silica would be more pro­ where ‘ads’ represents the adsorbed state. Thus, adsorption is char­
blematic because of crud formation. If this is the case, surfactants can acterized by the partition ratio of a species between the bulk liquid and
be used to reduce formation of crud from colloidal silica (Ritcey, 2004). the sorbent-liquid interface. Besides REE cations, other REE species and
Little information is available on the behaviour of refractory metals complexes can be adsorbed. Realistically, adsorption must be re­
(Ti, Zr, Hf, V, Nb, Ta, Cr, Mo, W) during solvent extraction of REEs. versible, so that REEs can be recovered by desorption according to the
However, with acidic extractants, it is usually the case that metal ions reverse of reaction (15). Adsorption and ion exchange are distinct
of higher charge and smaller size are extracted in favour of those metal processes; however, they seldom occur in their own right and many
ions of lower charge and larger size (El-Nadi, 2017). Therefore, re­ sorbents elicit a combination of these mechanisms (e.g., through sur­
fractory metal ions of high valence (i.e., +4) would probably co-extract face hydroxl groups).
before REEs and behave similar to thorium and uranium ions. Original research articles pertaining to impurity removal by ion
exchange and adsorption since 2015 were compiled and systematically
4.2. Ion exchange and adsorption categorized according to:

Ion exchange and adsorption are used less commonly than solvent 1. Matrix (solid phase)
extraction in industrial REE processing, but they fill an important gap. 2. Solvent or sorbent (responsible for extracting elements from solu­
Typically, solvent extraction produces REE salts or oxides from 99 to tion)
99.99% purity while ion exchange can produce substantially purer REE 3. Targeted elements in solution (usually REEs)
salts or oxides up to 99.9999% purity (Zhang et al., 2015). Therefore, 4. Other elements present in solution (Fe, Al, Ca, etc.)
ion exchange is frequently used for upgrading REE salts or oxides (e.g., 5. Feed material (chemistry, concentration, synthetic or ore, etc.)
upgrading 4 N REO to 6 N REO) and is the main method for producing 6. Summary and comments (separation efficiencies, recoveries, chal­
high purity or ultra-pure REE salts and oxides. lenges, etc.)
Several types of ion exchange or adsorption materials are available
for REE recovery from aqueous solutions. Polymeric resins are by far The results are summarized and presented in chronological order in
the most commonly used ion exchange materials in industry owing to Table 3. No previous review has focused on the behaviour of impurities
their low cost, stability, robustness, and cyclability. Additionally, a during ion exchange and adsorption of REEs. However, a review of the
wide array of functional groups is available to perform either cation or literature pertaining to REE separation by ion exchange (Hidayah and
anion exchange (see Appendix B). Functional groups are capable of ion Abidin, 2017) and adsorption (Anastopoulos et al., 2016) is available.
exchange only when they are ionized; thus, they are usually categorized Nearly all research has focused on the use of ion exchange materials
into strong/weak acid types (cation exchangers) or strong/weak base that function as cationic exchangers towards REEs in aqueous solutions.
types (anion exchangers). Besides polymeric resins, emerging ion ex­ A number of different materials containing REEs were used as the feed
change and adsorption materials include silica, microorganisms, bio­ material for ion exchange and adsorption techniques as summarized in
materials, or hybrid materials (Hidayah and Abidin, 2017). Fig. 5. The majority of investigators have studied the extraction of REEs
Strong acid type resins are the most effective type for REE recovery, from synthetic solutions with only about one third focusing on com­
especially polystyrene sulfonic acid resins (Lucas et al., 2015; Jha et al., mercial sources of REEs, such as ore leachates, REE concentrates, or
2016). Resins are loaded with cations, like NH4+ or H+, which ex­ leachates from secondary sources (e.g., fluorescent lamp phosphors).
change positions with REEs according to reaction (14) (Fitch and Most investigations considered the extraction of REEs from nitrate-
Russell, 1951): based solutions, followed by chlorides, and sulfates (Fig. 5). There were
also a significant body of research, which did not explicitly state the
REE 3 + + 3 P RNH4 = (P R)3 (REE ) + 3NH4+ (14) type of solution that was utilized (it is likely that a good majority of

10
 Table 3
Summary of original research articles on impurity removal during REE processing using ion exchange and adsorption.a Cases where all naturally occurring lanthanides (La to Lu, except Pm) in solution are followed are
denoted by Ln.
Investigators Ion exchange material(s) or sorbent(s) Targeted element(s) Other element Feed material(s) Summary and comments
in solution (s) in solution
W.D. Judge and G. Azimi

Abdel-Magied et al. Hierarchical porous zeolitic La, Sm, Dy Na, Mg, Ca, Ni Synthetic (type unknown) Adsorption capacity for REEs was 30–430 mg/g. High separation
(2019) imidazolate frameworks 36 mg/L REEs factors for impurities (> 260). Desorption performed with
36 mg/L impurities acetonitrile/methanol.
Arunraj et al. (2019) Yeast embedded in cellulose matrix Eu Y, La, Ce, Ca, Al Phosphor powder leachate (nitrate) The presence of impurities lowered the adsorption of Eu to 55%.
52 mg/L REEs (2 mg/L Eu) Desorption was performed with EDTA and the material was cycled 4
23 mg/L impurities times.
Avdibegović et al. Crystalline α‑zirconium phosphate Sc Ca, Al, Si, Fe, Ti, Bauxite residue leachate (chloride) Some co-extraction of most impurities occurred; however, a two-step
(2019) platelets Y, La, Ce, Nd, Dy 27.1 mg/L REEs (1.9 mg/L Sc) elution could recover Sc without any Fe or Al.
9550 mg/L impurities
Hamza et al. (2019) Magnetic resin U, REEs Al, Fe, Zn, Cu, Abu Mogherat ore leachate (sulfate) Enrichment factor reached close to 19. Desorption was performed
Ni, Co 184 mg/L REEs with H2SO4. Uranium and REEs were co-adsorbed and co-recovered.
417 mg/L impurities
Hu et al. (2019) Organo-functionalized, hierarchically Th REEs, Al, Fe, U, Synthetic (type unknown) Significant Al, Sc, Ti, Fe co-extracted with Th. Cycled 10 times with
porous silica monoliths Ca, Na, Ti 300 μg/L REEs no degradation. About 80–95% of Th could be extracted from
1200 μg/L impurities commercial samples.
Bauxite residue leachate (sulfate)
Khawassek et al. Dowex 50 × 8 resin Sc, Y, Ce, Eu Gd, Er, Al, Fe, Cd, Mn, Egyptian ore effluent (chloride) Adsorption capacity for REEs was 80 mg/g. The REEs were eluded
(2019) Yb Na, As, K, Pb, P, 300 mg/L REEs with HCl and precipitated with oxalic acid. The final product
S, Zn, Ti, Cu, Si, 4910 mg/L impurities contained some Ti, Si, Zr.
Ba, Cr
Li et al. (2019) Diglycolamide-grafted Fe3O4/ Sc, Y, Ln Na, K, Mg, Ca, Synthetic (nitrate) Achieved highly selective separation and easy separation of loaded
polydopamine nanomaterial Fe, Cu, Zn, Pb, 0.32 mg/L REEs hematite particles using a magnet. Elution was performed with
Cd, Co 7170 mg/L impurities EDTA and the material was cycled 5 times.

11
Liu et al. (2019) Diglycolamide polymer modified silica REEs K, Cd, Cu, Fe Synthetic (nitrate) At high concentrations, impurities decrease the adsorption capacity
150 mg/L REEs by about 20%. Iron had the strongest effect followed by Cu, Cd, K.
150–15,000 mg/L impurities
Lou et al. (2019) Acrylic acid functionalized metal Sc Cu, Zn, Mn, Co, Synthetic (chloride) Impurities inhibited the adsorption of Sc by up to 62%. The most
organic framework Al 20 mg/L Sc inhibiting impurity was Al followed by Mn, Zn, Co, and Cu.
100–200 mg/L impurities
Mondal et al. (2019) TEHDGA impregnated XAD-7 resin Sc, Y, Ln Si, Al, Fe, Ca, Coal fly ash leachate (nitrate) No major co-extraction of impurities occurred despite their high
Mg, Ti 56.3 mg/L REEs concentration. Desorption was performed with 0.01 M HNO3.
20,400 mg/L impurities
Ramasamy et al. CNT-reinforced silica composites Sc, Y, Ln Ca, Mg, Cr, Co, Acid mine drainage (sulfate) Studied both single and multi-walled CNT. The Pb, Cu, Cr, and Al
(2019) Ni, Cu, Fe, Zn, 2–5 mg/L REEs were co-extracted, but Ca, Co, Mg, Mn, Ni, Zn were not significantly
Mn, Na 274.2 mg/L impurities co-extracted.
Rodrigues et al. (2019) Cmp-based acrylamide polymer Gd Ni Synthetic (nitrate) Nickel tended to coextract more when the concentration of Gd was
0.19–12 mM Gd lower. At higher Gd concentrations, only 2% of Ni was extracted.
0.19–12 mM Ni
Xiong et al. (2019) Water-stable metal-organic Y, La, Ce, Eu, Lu Th Synthetic (chloride) The Th/REE separation factors reached 6–21. A 24 h equilibrium
framework material 5 mM REEs time was required. The Th adsorption was chemical in nature.
1 mM Th
Ang et al. (2018) Various chelating resins Th, U La, Ce, Gd, Yb Synthetic (sulfate) Bis-picolylamine resin and organophosphinic acid-impregnated resin
8 mM REEs were most for U, while iminodiacetic resins were most effective for
4 mM impurities Th.
Babu et al. (2018) Ethylenediaminetriacetic acid- La, Sm Ni, Co Synthetic (nitrate) Adsorption capacity for REEs was 50–60 mg/g. About 15–20% of Ni
functionalized activated carbon 1 mM REEs and Co were co-extracted with REEs. Elution was performed with
1 mM impurities HCl.
Hérès et al. (2018) Tulsion CH-93, Purolite S940, Sc, Y, Ln Al, Ca, Fe, Mg, Phosphate ore leachate (phosphate, Aminophosphonic and aminomethylphosphonic resins performed
Amberlite IRC-747, Lewatit TP-260, V, Zn sulfate) the best, extracting 20–80% of REEs. Iron was significantly extracted
Lewatit VP OC 1026, Monophos, and 51.8–79.9 mg/L REEs and was difficult to strip from the resin. The next most extracted
Diphonix resins 7.64–7.92 g/L impurities impurities were Al > V > Mg > Ca, Zn.
(continued on next page)
Hydrometallurgy 196 (2020) 105435
 Table 3 (continued)

Investigators Ion exchange material(s) or sorbent(s) Targeted element(s) Other element Feed material(s) Summary and comments
in solution (s) in solution

Li et al. (2018) Phenanthroline diamide U, Th Nd, Sm, Eu, Na Synthetic (nitrate, chloride) Adsorption capacities were 718 and 703 mg/g for U and Th,
functionalized graphene oxide 150 mg/L REEs respectively. Highest Th/REE, U/REE separation factors (⁓25)
W.D. Judge and G. Azimi

50 mg/L U, 50 mg/L Th occurred at low pH.

Liang et al. (2018) Polyacrylic acid grafted silica fume Dy Ca, K, Na Synthetic (type unknown) The Na and K had virtually no effect on Dy extraction even at the
200 mg/L Dy high concentrations. Calcium reduced the extraction of Dy up to
10–200 mM impurities 20%.
Reynier et al. (2018) Various commercial resins (total of 20 Th, U Fe, Y, Ln Canadian ore leachate (sulfate, nitrate) It was difficult to simultaneously remove both Th and U compared
investigated) 435 mg/L REEs with their individual removal. Nitric media allowed better extraction
⁓0.6 g/L Th, ⁓2 g/L U of U and Th compared with sulphate. Also performed column
1.53–3.20 g/L Fe experiments.
Rychkov et al. (2018) Purolite C160 REEs Ca, Fe, Al, Sr, Phosphogypsum (phosphate, sulfate) Used a resin-in-pulp process to simultaneously leach and extract
Na, K, P, Cu, Zn, 0.44 wt% REOs REEs in one step. Iron and calcium co-extracted with REEs and
As 25 wt% Ca followed them during desorption with ammonium nitrate.
< 3 wt% other impurities
Shinozaki et al. (2018) Polymeric adsorbents bearing Ln Fe, Cu, Zn, Al, Synthetic (chloride) Impurities did not co-extract except Fe which a few % adsorbed.
diglycolamic acid ligands Ca 1.4 mM REEs Desorption was performed with 2 M HCl and the material was cycled
500 mM impurities 5 times.
Tay et al. (2018) E. coli bacteria REEs Al, Ca, Cu, Fe, Synthetic (type unknown) The effect of impurities was minor below 10 mM, but sorption was
Ni 1.5 mM REEs reduced by 50% at 10 mM. The Al, Cu, and Fe had the strongest
23 mM impurities effect.
Wang et al. (2018b) Wrinkled mesoporous carbon Th Sc, Y, Ln, U, Zr Synthetic (nitrate) High selectivity for Th compared with U and REEs. Some co-
4.8 mg/L REEs extraction of Zr occurred. Stripping was performed with 0.1 M
0.3 mg/L Th, 0.3 mg/L U ammonium oxalate.
Yang et al. (2018) Functionalized hierarchically Th, U Y, Ce, Nd, Sm, Synthetic (nitrate, chloride) Simultaneous adsorption capacities reached 40 and 53 mg/g for U
mesoporous silica Eu, Yb, Na 300 mg/L REEs (VI) and Th(IV), respectively. Less than 3% of REEs were co-

12
50 mg/L Th, 50 mg/L U extracted.
Ang et al. (2017) Various anionic and cationic exchange Th, U La, Ce, Gd, Yb Synthetic (sulfate) Weak-base resins performed the best, extracting 78% of U and 68%
resins 8 mM REEs of Th with less than 5% REE co-extraction.
2 mM Th, 2 mM U
Ashour et al. (2017) Magnetite nanoparticles Y, La, Nd, Gd Mg, Ca, Ni Synthetic (nitrate) Good selectivity for REEs with only minor extraction of impurities.
functionalized with citric acid or L- 0.04–0.20 mM REEs Nickel co-extracted more than Ca and Mg. Elution was performed
cysteine 0.03–0.15 mM impurities with HNO3.
Park et al. (2017) Engineered bacteria Y, La, Ce, Pr, Nd, Al, Mg, Mn, Cu, Various tailings leachates (nitrate, Achieved high extraction and recovery of REEs with minimal co-
Sm, Gd, Tb, Dy, Er, Zn, Ga, Rb, Ag, chloride) extraction of Ca, Ba, Zn, Mg, Na, K, Mn, and Rb. Co-extraction or
Yb Cd, Fe, Sr, Ba, 0.98–61.6 μM REEs interference occurred with Al, Cu, Fe, Ga, and Pb. Desorption was
Pb, Th, U, Na, K 243–430,000 μM impurities performed using 5 mM sodium citrate.
Qi et al. (2017) Layered A2Sn3S7·1.25H2O Eu Al, Fe, Na Synthetic (chloride) The distribution coefficient of Eu was 44× that of Na, 4× that of Al,
(A = organic cation) 6 mg/L Eu and 12× that of Fe. Also performed column experiments. Eluted
501 mg/L impurities with KCl.
Ramasamy et al. PAN modified mesoporous silica/ Sc, Y, Ln Al, Fe, Au, Mg, Synthetic (type unknown) The Fe(III), Al, and Au were co-extracted, but Mg, Mn, and Zn were
(2017) chitosan Mn, Zn 25 mg/L REEs not significant. Adsorption of REEs became poor at high impurity
25–250 mg/L impurities concentration.
Qiu et al. (2017) ODP-IP impregnated XAD-7 resin Th Fe, Ln REE waste residue leachate (nitrate) Adsorption capacity for Th was up to 1.4 mg/g. In a mixture, the
0.189 mM REEs uptake of Th was 64% with no significant co-extraction of REEs or
0.0066 mM Th, 4.9 mM Fe Fe.
Smirnov et al. (2017) Tulsion CH 93, Purolite D 5041, Sc Th, Al, Fe, Ti, U Uranium leach liquor (sulfate) Thorium always tended to co-extract with Sc. Selective desorption
Lewatit TP 260, Purolite S 950 resins 0.78 mg/L Sc processes were developed to desorb 70% of Th with only 10% loss of
3580 mg/L impurities Sc.
Wang et al. (2017a) Polyethylenimine− acrylamide/SiO2 La, Nd, Eu, Dy, Yb Na, Mg, Al, Ca, Synthetic (chloride) All impurities were co-extracted; the order of decreasing co-
hybrid hydrogel Mn, Fe 55.8 μg/L REEs extraction was Al > Mg > Fe > Ca > Mn > Na. The REEs
678 mg/L impurities desorbed with Al, Mn, Fe.
Wilfong et al. (2017) Porous amine−epoxy networks La, Nd, Eu, Dy, Yb Na, Mg, Al, Ca, Synthetic (chloride) Nearly all Al and Fe co-extracted with REEs. The Na, Mg, and Ca did
Mn, Fe 57 μg/L REEs not co-extract while some Mn co-extracted. Some Al, Fe, Mn stripped
689 mg/L impurities with REEs.
Hydrometallurgy 196 (2020) 105435

(continued on next page)

 Table 3 (continued)

Investigators Ion exchange material(s) or sorbent(s) Targeted element(s) Other element Feed material(s) Summary and comments
in solution (s) in solution

Zhao et al. (2017a) Polyethylenimine-cross-linked La, Eu, Er Na, Ca, Fe Synthetic (nitrate) Each impurity was studied individually, all decreased the adsorption
cellulose nanocrystals 3 mM REEs capacity of REEs (Fe > Ca > Na). Regenerated sorbent using
W.D. Judge and G. Azimi

2 mM impurities HNO3.
Bonificio and Clarke Roseobacter sp. AzwK-3b Sc, Y, Ln Th Synthetic (nitrate) Utilized bacteria to perform adsorption. The LREEs desorbed by
(2016) 32 mg/L REEs decreasing pH, while HREEs remained bound, allowing separations
10 mg/L Th from low grade solution.
Dubey and Grandhi Nano maghemite Y Na, K, Ca, Mg, Synthetic (nitrate) Studied each impurity individually. The Na, K only weakly
(2016) Al 3.6 mg/L Y suppressed adsorption efficiency of REEs while Mg, Ca, Al had a
21.6 mg/L impurities larger impact.
Ogata et al. (2016b) Silica gel particles modified with Y, Ln Ca, Si, Al, Fe, Apatite leachate (sulfate) Performed column experiments. The REEs were retained longer than
diglycolamic acid groups Mn, Mg, Na, K, 153 mg/L REEs impurity elements and they were recovered without much
Ti, P 30,100 mg/L impurities impurities.
Ogata et al. (2016a) Silica gel adsorbents bearing Nd, Dy Fe, Zn, Cu Synthetic (chloride) Achieved selective extraction of Dy (> 90%) and Nd (> 80%), with
immobilized diglycolamic acid ligands 2 mM REEs low Fe extraction (< 30%). No co-extraction of Zn or Cu occurred.
3 mM impurities
Park et al. (2016) Caulobacter crescentus Y, La, Ce, Nd, Eu, Tb, Ca, Ni, Zn, Cu, Synthetic (chloride, nitrate) Genetically engineered bacteria to display lanthanide binding tags.
Dy, Yb Mn, Mg, Co, Al, 0.01–100 μM REE The Ca and Cu interfered with REE adsorption.
K 1–10,000 μM impurities
Roosen et al. (2016) Chitosan-silica particles Sc Na, Ca, Al, Fe, Greek bauxite residue leachate (nitrate) Focused largely on Sc/Fe separation. Achieved 80% Sc adsorption
functionalized with DTPA and EGTA Si, Ti, REEs 2 mg/L Sc with no Fe adsorption at pH 1.5. Performed column experiments.
3490 mg/L impurities Scandium break through curve occurred after all impurities.
Rychkov et al. (2016) Purolite C-100 resin Y, La, Ce, Dy Al, Fe, Ca, Th Synthetic (sulfate) The investigated strongly acidic cation exchange resin always co-
4 g/L REEs extracted impurities. The REE extraction did not strongly depend on
2.5 g/L impurities pH, but Al and Fe extraction increased with increasing pH. The
Russian uranium-barren leach solution process was useful for pre-concentration of REEs.

13
(sulfate)
36.7 mg/L REEs
2800 mg/L impurities
Smith et al. (2016) Carbon black derived from recycled Y, La, Ce, Nd, Sm Na, Ca, Zn, Mn Synthetic (chloride, nitrate) Achieved 60–90% extraction of REEs. The presence of Na and Ca
tires 100 mg/L REEs enhanced REE adsorption, whereas Zn and Mn reduced REE
200 mg/L impurities adsorption.
Van Nguyen et al. Synthesized resin containing glycol Sc La, Ce, Al Synthetic (chloride) Selectively extracted Sc with no co-extraction of Al. Some REEs were
(2016) amic acid 1070 mg/L REEs (20 mg/L Sc) co-extracted at higher pH. Elution was performed with 2 M HCl at
2000 mg/L Al 80 °C.
Zheng et al. (2016) Dual-template docking oriented ionic Nd Fe, Co, Dy, Tb, Synthetic (type unknown) The distribution coefficient for Nd was significantly greater than
imprinted mesoporous films Pr 50 mg/L Nd other REEs and Fe. In the NdFeB leachate, about 50% of Nd was
200 mg/L impurities recovered with minor co-extraction of Fe > Dy > Co. Stripping
NdFeB leachate (type unkown) was performed with HNO3. The material was reconditioned and
65.1 mg/L Nd cycled 5 times.
38.3 mg/L impurities
Ogata et al. (2015b) Silica gel particles modified with Ln Al, Ca, Cu, Fe, Synthetic (chloride) Performed column experiments with H2SO4 eluent. Impurities were
diglycol amic acid groups, Amberlite Zn 1.4 mM REEs co-extracted, but leaked after a few bed volumes, while REEs were
IR120 resin 0.5 M impurities retained. More Al and Fe co-extraction occurred with LREEs.
Ogata et al. (2015a) Silica gel particles modified with Nd, Dy Fe, Cu, Zn Synthetic (chloride, nitrate, sulfate) Solution chemistry and speciation have a major impact on
diglycol amic acid groups 2 mM REEs separations. Sulfate ions lower REE extraction. The Zn, Cu, and Fe
3 mM each ion co-extracted in acidic chloride environments (> 1 M HCl).

a
Amberlite IR120 = strong acidic cation exchang resin; Amberlite IRC-747 = aminophosphonic acid chelating resin; cmp = carbamoylmethylphosphonate; CNT = carbon nano tube; Diphonix = geminally
substituted diphosphonic acid, sulphonic and carboxylic acid cation exchange resin; DTPA = diethylenetriaminepentaacetic acid; EDTA = ethylenediaminetetraacetic acid; Dowex 50 × 8 = strong acidic cation
exchange resin; EGTA = ethyleneglycol tetraacetic acid; Lewatit TP260 = weak acidic cation exchange resin (aminophosphonic); Monophos = monophosphonic and sulphonic acid cation exchange resin; ODP-IP = n-
octyl diphenyl phosphate; PAN = 1-(2-pyridylazo)-2-naphthol; Purolite C100 = strong acidic cation exchange resin; Purolite C160 = sulfonic strong acid cation exchange resin; Purolite D5041 = phosphorus-containing
cation exchange resin; Lewatit VP OC 1026 = D2EHPA impregnated resin; Purolite S940 = aminophosphonic acid chelating resin; Purolite S950 = aminophosphonic acid chelating resin; TEHDGA = N,N,N´,N´-tetra(2-
ethylhexyl) diglycolamide; Tulsion CH-93 = aminomethyl phosphonic chelating resin; XAD-7 = weakly polar adsorbent resin.
Hydrometallurgy 196 (2020) 105435
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

to many cation exchange resins; thus, their tendencies towards iron are
also similar to cation exchange resins in which iron(III) usually co-ex­
tracts with REEs (Ogata et al., 2016a; Ogata et al., 2015b; Zheng et al.,
2016; Park et al., 2017; Ramasamy et al., 2017; Wang et al., 2017a, and
Wilfong et al., 2017; Tay et al., 2018). Direct competition between iron
(III) and REEs is apparent by the reduction in REE adsorption capacity
with iron co-extraction (Liu et al., 2019; Ramasamy et al., 2017; Zhao
et al., 2017a). Selective elution procedures have been developed for
zirconium phosphate platelets (Avdibegović et al., 2019) and functio­
Fig. 5. Summary of feed type for ion exchange and adsorption articles in
nalized silica (Ogata et al., 2016b; Ogata et al., 2015b) to increase REE
Table 3: (a) source of feed and (b) major anion in feed.
purity.
Chelating resins and materials functionalized with solvating ex­
these cases are dealing with nitrates because this type of standard so­ tractants have also been investigated for REE recovery from iron-con­
lution is readily available to most wet chemical labs). All these feed taining solutions. Mondal et al., 2019 used resin impregnated with a
materials have some potential commercial significance depending upon solvating extractant (TEHDGA), and reported high recoveries for REEs
the type of mineral acid used for leaching. from nitrate media with no significant co-extraction of iron despite its
A number of new developments in REE processing using ion ex­ high concentration. With regards to chelating resins, Hérès et al., 2018
change and adsorption have occurred since 2015. There is a growing found that iron was always co-extracted with REEs from phosphate
interest in the use of biomaterials, such as biopolymers and biochar for media and it significantly reduced the loading of REEs. It was possible
ion exchange or adsorption of REEs. These materials have the potential to recover REEs by selectively stripping the resin; however, iron was
for high adsorption capacities (e.g., 800 mg/g) with relatively low cost difficult to strip and accumulated in the resin; hence, reducing its ca­
and high availability (Jing et al., 2018). Bioadsorption of REEs using pacity. Usually basic resins have poor extraction capacity for both REEs
bacteria is another emerging area of research in REE separation and and iron (Reynier et al., 2018); thus, they are not suitable for REE/Fe
impurity removal (Park et al., 2016; Bonificio and Clarke, 2016; partitioning. In sulfate media, some loading of iron as Fe(SO4)n(3−2n) or
Hidayah and Abidin, 2017; Park et al., 2017; Tay et al., 2018). Most Fe(OH)(SO4)22− is possible on basic resins, but not to the extent re­
cells are naturally functionalized with phosphate or carboxylate groups quired for the selective separation of REE/Fe (Sole et al., 2018).
with high affinity towards REEs and can be stripped with weak organic
acids; thereby, functioning similar to conventional adsorbents. Bacteria
has also been bioengineered to display lanthanide binding tags that 4.2.2. Behaviour of thorium and uranium
enhance selectivity towards REEs (Park et al., 2016). Thorium and uranium associate with REE mineralization and tend
Legaria et al., 2017 developed a solid-phase magnetic extraction of to follow the REEs during processing, making their partitioning a
REEs using silica coated maghemite nanoparticles. The nanoparticles challenge in ion exchange or adsorption operations. Two approaches
had a very high surface area, which greatly improved the REE loading are used to separate thorium and uranium from REEs by ion exchange
with adsorption capacities up to 900 mg/g. After adsorption, the na­ or adsorption. In the first approach, REEs are selectively extracted while
noparticles were easily collected using a magnet and stripped with minimizing co-extraction of thorium and uranium, while in the second
ammonium sulfate. The results are promising for increasing REE ex­ approach, thorium and uranium are selectively extracted while mini­
traction and loading while mitigating difficulties associated with tra­ mizing co-extraction of REEs. Typically, better separations are possible
ditional (e.g., gravity) solid-liquid separation of small particles. This using the second approach; however, this approach increases proces­
technique has also been demonstrated using magnetite particles (Li sing costs because it adds another processing step.
et al., 2019). Even with these recent developments, challenges remain Using strong acid cation exchange resins to extract REEs from
with respect to the removal of impurities such as iron, thorium, ur­ thorium and uranium contiaing solutions is not usually very selective
anium, and aluminum. and inevitably results in co-extraction of thorium and uranium (Ang
et al., 2017; Smirnov et al., 2017). Cation exchange resins have higher
affinity for species with higher charge; thus, the adsorption of thorium
4.2.1. Behaviour of iron
(IV) and uranium(IV) cations will be favoured over REE cations, which
The behaviour of iron towards ion exchange and adsorption pro­
are favoured over uranyl (UO22+) ions. Complexation in solution can
cessing of REEs depends upon the type of ion exchange material or
adjust the affinity series, for instance, in sulfate solutions, thorium
sorbent used and the type of feed solution. Strong acid cation exchange
forms ThSO42+ ions that have lower affinity for cation exchange resins
resins are extensively used for REE recovery. Their success depends on
than REEs (Ang et al., 2017). However, the change in affinity is not
the availability of uncomplexed REE ions in solution; thus, they are
strong enough to perform selective separations.
frequently employed with chloride media. When iron ions are present
Basic anion exchange resins are much more effective at selectively
in the same solution, they are also frequently uncomplexed; thus, their
extracting thorium and uranium from REE solutions (Ang et al., 2017;
co-extraction with REEs virtually always occurs on acidic ion exchange
Reynier et al., 2018). These resins are most effective in sulfate or nitrate
resins (Ogata et al., 2015b; Rychkov et al., 2016; Hérès et al., 2018;
media in which thorium and uranium are known to form anionic
Rychkov et al., 2018). Most likely iron(III) competes more strongly with
complexes. For example, anion exchange of uranium(IV) from sulfate
REEs compared with iron(II) because of its similar charge density, size,
solutions proceeds according to reaction (16) (Merritt, 1971):
and hydration enthalpy to REE cations. Therefore, processing condi­
tions which increase the ratio of ferric to ferrous can be expected to UO2 (SO4 )34 + 4 P NR3 X = (P NR3 )4 UO2 (SO4 )3 + 4X (16)
correspondingly increase iron co-extraction. If co-extraction of iron
occurs, selective elution procedures can be used to recover pure REEs where X− represents the counter ion initially loaded on the resin. In
(Ogata et al., 2015b). sulfate solutions, uranium(VI) is extracted to the same extent using both
Besides polymeric resins, other cation exchange materials, including strong and weak base anion exchange resins; however, thorium is ex­
functionalized silica (Ogata et al., 2015a; Ogata et al., 2015a; Ogata tracted more when weak base resins with primary amine functional
et al., 2016a; Ogata et al., 2016b), bioogranisms (Tay et al., 2018; Park groups are used (Ang et al., 2017). Little co-extraction of REEs occurs
et al., 2017), and zirconium phosphate platelets (Avdibegović et al., under these conditions (< 5%). On the contrary, in nitrate solutions
2019) have been investigated for REE/Fe separations. Functional thorium is extracted more on strong base anion exchange resins
groups in these materials (e.g., phosphate or carboxylate) are common (Reynier et al., 2018). Here, the extraction of thorium proceeds through

14
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

its hexanitro complex according to reaction (17) (Ryan, 1960): synthesized with these functional groups. Synthesized resins were de­
monstrated to extract REEs from solutions with virtually no co-extrac­
Th (NO3 )62 + 2P NR3 X = (P NR3 ) 2 Th (NO3 )6 + 2X (17) tion of aluminum, despite its high concentrations (Van Nguyen et al.,
2016; Shinozaki et al., 2018). Mechanisms behind the selectivity have
However, using strong basic anion exchange resins in nitrate solu­
not been completely elucidated. A drawback, however, is that the ex­
tions also leads to more co-extraction of REEs (Reynier et al., 2018).
traction of REEs is a comparatively slow process, requiring up to 24 h to
Thorium and uranium also form anionic carbonate complexes that can
reach saturation (Van Nguyen et al., 2016).
be extracted on basic resins; however, this approach has not yet been
In addition to resins, extraction of REEs from aluminum-containing
investigated for treating REE solutions (Sole et al., 2018).
solutions has also been studied on other materials including poly­
Chelating and solvent-impregnated resins have also been used to
ethylenimine−acrylamide/SiO2 hybrid hydrogel (Wang et al., 2017a),
selectively extract thorium and uranium from REE solutions (Smirnov
PAN-modified mesoporous silica/chitosan (Ramasamy et al., 2017),
et al., 2017; Qiu et al., 2017; Ang et al., 2018; Reynier et al., 2018).
engineered bacteria (Park et al., 2016; Park et al., 2017), and layered
Using ODP impregnated resin, Qiu et al., 2017 extracted the majority of
A2Sn3S7•1.25H2O (A = organic cation) (Qi et al., 2017). In all cases,
thorium (64%) from nitrate solutions with little to no co-extraction of
aluminum was found to co-extract with REEs and usually reduced REE
REEs or iron. Resin impregnated with D2EHPA performed even better,
loading. Lack of selectivity is likely due to reliance of cation exchange
extracting 95% of thorium and 80% of uranium simultaneously
or adsorption mechanisms on the cation charge, size, and hydration
with < 1% co-extraction of REEs (Reynier et al., 2018). A number of
enthalpy, which are similar for REEs and aluminum.
other chelating resins investigated by Reynier et al., 2018 which did not
Less information is available regarding REE/Al selectivity on che­
perform as well, but the reason was not immediately apparent. A con­
lating, impregnated, or basic resins. A resin impregnated with a sol­
vincing argument explaining selectivity was proposed by Ang et al.,
vating extractant (TEHDGA) was found effective for extracting REEs
2018 after detailed investigation of seven resins. Interpreting extraction
from nitrate solutions (coal fly ash leachate) with virtually no co-ex­
with respect to coordination chemistry shows that functional ligands
traction of aluminum (Mondal et al., 2019). Using various chelating
with oxygen donor atoms lead to a high affinity towards thorium, while
resins, Hérès et al., 2018 observed aluminum always co-extracted with
softer donor atoms like nitrogen lead to a high affinity towards uranium
REEs from phosphate ore leachate and reduced REE loading. As well,
(VI) (Ang et al., 2018). Therefore, ligands with a combination of oxygen
there are no current practices to selectively extract aluminum from REE
and nitrogen, such as those in iminodiacetic and aminophosphonic re­
solutions by ion exchange or adsorption.
sins, are best for extracting both thorium and uranium(VI) simulta­
neously.
4.2.4. Behaviour of base metals and other impurities
Selective sorption of thorium and uranium has also been in­
A number of other impurities are usually present in REE processing
vestigated on other materials besides polymeric resins. Li et al., 2018
solutions from the gangue material present in the ore or secondary
investigated PDA-graphene oxide and found adsorption capacities
source. Transition metals other than iron are frequently used in coatings
above 700 mg/g for both uranium and thorium with Th/REE and U/
for NdFeB magnets and are present in the gangue material of electronic
REE separation factors up to 25. Li et al., 2018 also reported that in­
waste. The literature is mostly consistent that nickel, cobalt, and
creasing the ionic strength of the solution with alkali salts increased the
manganese do not significantly co-extract with REEs during ion ex­
extraction of uranium and thorium because the alkalis are preferentially
change or adsorption (Abdel-Magied et al., 2019; Hamza et al., 2019; Li
hydrated in solution; hence, they lower the solubility of uranium and
et al., 2019; Lou et al., 2019; Ramasamy et al., 2019; Rodrigues et al.,
thorium. Using wrinkled mesoporous carbon, Wang et al., 2018b
2019; Park et al., 2017; Ramasamy et al., 2017; Zheng et al., 2016) or
achieved a high selectivity for uranium over thorium and REEs with
co-extract only a minor amount (Babu et al., 2018; Ashour et al., 2017;
short equilibration times of only 5 min. Bonificio and Clarke, 2016 used
Wang et al., 2017a; Wilfong et al., 2017). This is expected because most
engineered bacteria to selectively extract REEs from low grade solutions
investigations considered cation exchange or adsorption materials
containing thorium. Separation of thorium and uranium from REEs has
which have higher affinity towards trivalent REE cations compared
also been investigated on functionalized porous silica monoliths (Hu
with divalent transition metal ions.
et al., 2019), metal organic frameworks (Xiong et al., 2019), and me­
Copper and zinc have been noted to inhibit adsorption of REEs (Lou
soporous silica (Yang et al., 2018).
et al., 2019; Smith et al., 2016; Roosen et al., 2016; Park et al., 2017)
and slightly co-extract in acidic chloride environments (Ogata et al.,
4.2.3. Behaviour of aluminum 2015a). Action of copper or zinc to inhibit REE adsorption was postu­
Similar to most REE cations, aluminum cations carry a trivalent lated to be due to the competition for the same binding sites as REEs
charge (Al(III)), which makes REE/Al separations challenging. If pre­ (Smith et al., 2016) or their binding with other sites in a manner that
sent in the feed solution, aluminum virtually always co-extracts with disrupts REE binding sites (Park et al., 2017). Copper(II) and zinc(II)
REEs during ion exchange or adsorption, but the extent of co-extraction are borderline Lewis acids; thus, their interference (co-extraction) could
depends on its concentration, solution chemistry, and the type of sor­ be expected to be more in cases where REE adsorption requires sharing
bent. Most REEs are extracted from the solution using strong acid cation or exchange of electrons (e.g., dissociative adsorption).
exchange resins, which are selective towards cation charge, size, and Calcium and magnesium do not usually co-extract with REEs on
hydration enthalpy. Therefore, the similarity in charge, size, and hy­ cation exchange or adsorption materials unless present in high con­
dration enthalpy between aluminum and REEs puts aluminum in con­ centrations. Li et al., 2019 reported the upper tolerance limit for cal­
tention with REEs for extraction with acidic cation exchange resins cium and magnesium to be 1.5 g/L each in REE containing solutions.
(Ogata et al., 2015b; Rychkov et al., 2016; Wilfong et al., 2017; Hérès Ramasamy et al., 2017 found that less than 2% of magnesium co-ex­
et al., 2018; Lou et al., 2019). There is a slight tendency for aluminum tracted with REEs even when magnesium was an order of magnitude
to co-extract more with LREEs on acidic resins (Ogata et al., 2015b). In more concentrated in solution than the REEs. Liang et al., 2018 ob­
column arrangements, REEs are usually retained longer than aluminum served calcium reduced REE loading by 5–20% as calcium concentra­
during elution; therefore, despite aluminum co-extraction, this impurity tion increased from 0.4 to 8.0 g/L and hypothesized calcium must be
may be separated by selective elution (Ogata et al., 2015b). competing with REEs for carboxyl sites (–COO−) at these high con­
Ligands deriving from glycol amic acid offer the highest selectivity centrations. Using strong acid cation exchange resins, Rychkov et al.,
between REEs and aluminum (Van Nguyen et al., 2016; Shinozaki et al., 2016 and Rychkov et al., 2018 found some calcium and/or magnesium
2018). This is an important finding since this is one of the few cation co-extraction and therefore Rychkov et al., 2016 opted to produce a
exchange materials to offer REE/Al selectivity and resins can be pre-concentrate of REEs containing 10% CaO and 3% MgO. On carbon

15
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

black, Smith et al., 2016 found that calcium improved the adsorption of although in this case, ammonium bicarbonate is the preferred pre­
REEs, but the mechanism was not investigated in detail. Calcium is also cipitant, according to reaction (21) (Lucas et al., 2015):
known to interfere with bioadsorption of REEs on bacteria (Park et al., REE2 SO4(aq) + 3NH4 HCO3(s) = REE2 (CO3 )3(s) + 3NH4 HSO4(aq) (21)
2016; Jiao et al., 2019). No clear model has been proposed to describe
co-extraction of calcium or magnesium. However, when calcium and Mixed REE carbonates produced contain about 5 wt% impurity
magnesium co-extraction reduces REE loading, it implies that they ei­ carbonates and are the feed material for REE separation plants in China
ther occupy REE adsorption sites or other adsorption sites that interfere (Lucas et al., 2015). Precipitation is also used for finishing after in­
with REE sites. In the case where calcium improved REE adsorption dividual REE separation by solvent extraction or ion exchange to pro­
(Smith et al., 2016), it was hypothesized that calcium helped improve duce REOs. A common practice is precipitation of REE oxalates using
wettability of the adsorbent; hence, facilitating the transfer of ions to oxalic acid followed by their calcination (900–1000 °C) to REOs ac­
the adsorbate surface. cording to reactions (22) and (23) (Lucas et al., 2015):
Refractory metals such as titanium, vanadium, and zirconium may
2REE(3aq+) + 3H2 C2 O4(aq) = REE2 (C2 O4 )3(s) + 6H+ (22)
slightly co-extract with REEs on strong acid cation exchange resins
(Hérès et al., 2018; Khawassek et al., 2019). This should be expected REE2 (C2 O4 )3(s) = REE2 O3(s) + 6CO2(g ) (23)
because of the dependence of the affinity of cations towards these resins
on the cation charge, size, and hydration enthalpy. Therefore, re­ Typically, oxalate precipitation yields precipitates with good crys­
fractory metal ions of high valence (e.g., > 3+) should be pre­ tallinity, but it is costly. Finishing may also be conducted by REE hy­
ferentially extracted over trivalent REEs. By the same principle, op­ droxide precipitation followed by calcination to REOs according to
erations specifically designed for thorium removal also usually co- reactions (24) and (25) (Lucas et al., 2015):
extract some titanium and zirconium (Wang et al., 2018b; Hu et al., REE(3aq+) + 3NH4 OH(aq) = REE (OH )3(s) + 3NH4(
+
aq) (24)
2019).
Alkali metals do not co-extract or interfere with REE extraction on 2REE (OH )3(s) = REE2 O3(s) + H2 O(g ) (25)
cation exchange or adsorption materials, unless they are so con­
centrated that salting out occurs (Liu et al., 2019; Liang et al., 2018). or by REE carbonate precipitation followed by calcination according to
One exception is REE extraction using bioorganisms, where sodium and reactions (26) and (27):
potassium may interfere with REE extraction (Jiao et al., 2019). Silica 2REE(3aq+) + 3NaHCO3(aq) + 3NaOH(aq)
may sometimes be solubilized in certain REE processing solutions. Al­
though it does not co-extract with REEs during cation exchange or = REE2 (CO3 )3(s) + 6Na(+aq) + 3H2 O(l) (26)
adsorption (Ogata et al., 2016b; Roosen et al., 2016; Mondal et al.,
REE2 (CO3 )3(s ) = REE2 O3(s) + 3CO2(g ) (27)
2019), silica may pose other problems, such as gel formation that in­
terfere with cation exchange or adsorption. Usually carbonate precipitates are less crystalline and require seed
Gold was found to co-extract more than 90% with REEs on PAN- crystals or surfactants. Therefore, we see precipitation is an important
modified mesoporous silica/chitosan (Ramasamy et al., 2017). Pla­ part of REE processing and may be conducted at various stages in the
tinum group metals have a tendency to form anionic complexes; thus, it REE flowsheet and different precipitates may be chosen. The following
is not clear if co-extraction of gold occurs via anion exchange/adsorp­ section examines the behaviour of different precipitation schemes to­
tion or cation exchange/adsorption. Either way this is an important wards impurity removal.
consideration if REEs are to be recovered from electronic waste lea­ Original research articles pertaining to impurity removal by selec­
chates. Ramasamy et al., 2019 also found that some co-extraction of tive precipitation since 2015 were compiled and systematically cate­
lead and chromium occurred on single and multiwalled carbon nano­ gorized according to:
tubes (CNT). Cadmium has been found to slightly reduce the adsorption
capacity of REEs (Liu et al., 2019). 1. Targeted elements in solution (usually REEs)
2. Other elements present in solution (Fe, Al, Ca, etc.)
4.3. Selective precipitation 3. Feed material (chemistry, synthetic or ore, etc.)
4. Precipitation scheme (compounds, pH adjustments, etc.)
Precipitation is a process vital to all REE flowsheets. It converts 5. Summary and comments (separation efficiencies, recoveries, chal­
aqueous REE species to solid products for separation, impurity removal, lenges, etc.)
sale, transport, or use in subsequent manufacturing of high value pro­
ducts. A variety of REE precipitants are available depending upon the The results are summarized and presented in chronological order in
objectives of the precipitation process, a comparison of REE solubility Table 4. No prior review has focused on the removal of impurities from
products is given in Appendix C. During REE processing, it is often REE processing solutions by selective precipitation. However, reviews
necessary to produce intermediate REE products for sale or transport to are available on general hydrometallurgical recovery of REEs (Jha
separation plants. For instance, bastnäsite and monazite concentrates et al., 2016; Krishnamurthy and Gupta, 2016).
are often cracked using hot sodium hydroxide (120–180 °C) to yield Considerably less research has been focused on impurity removal by
REE hydroxides according to reactions (18) and (19), respectively precipitation, as compared with solvent extraction and ion exchange/
(Lucas et al., 2015): adsorption. This is because the finite amount of possible precipitation
schemes available for REE processing solutions makes it challenging to
(REE ) FCO3(s) + 3NaOH(aq) = NaF(aq) + Na2 CO3(aq) + REE (OH )3(s)
make innovations in this area. Despite this fact, a number of high-
(18) quality investigations in impurity removal using precipitation have
been conducted recently. All of these have focused on industrial lea­
REE (PO4 ) (s) + 3NaOH(aq) = Na3 PO4(aq) + REE (OH )3(s ) (19)
chates, with most derived from the treatment of REE containing ores.
Another approach for treating bastnäsite and monazite concentrate About three quarters deal with sulfate leachates while others work with
is sulfuric acid baking to yield solid REE sulfates, which are converted chlorides or caustic leachates. One investigation was focused on lea­
to carbonates according to reaction (20) (Lucas et al., 2015): chates derived from secondary sources (NdFeB permanent magnets).
Regardless of the source material, most leachates were treated to re­
REE2 (SO4 )3(s) + 3Na2 CO3(aq) = REE2 (CO3 )3(s) + 3Na2 SO4(aq) (20)
move base metals, such as iron, calcium, and aluminum. Leachates from
A similar method is used after leaching ion adsorption clays, primary REE sources (i.e., ores) also need to be treated to remove

16
 Table 4
Summary of original research articles on impurity removal during REE processing using selective precipitation. Cases where all naturally occurring lanthanides (La to Lu, except Pm) in solution are followed are denoted
by Ln.
Investigators Targeted element Other element(s) in solution Feed material Precipitation scheme Summary and comments
(s) in solution
W.D. Judge and G. Azimi

El Afifi et al. REEs Th, Fe Egyptian monazite liquor 1a. Add potassium sulfate (0.23 M) and/or sodium sulfide Selectively precipitated about 99% of Fe and 90% of Th. Iodate can
(2019) (chloride) (0.23 M) also be added to precipitate 99% of Th and 99% of Fe.
1b. Add potassium iodide (0.155 M)
Güneş et al. (2019) La, Ce, Pr, Nd, Sm Th Eski¸sehir-Beylikova ore 1. add 12% (v/v) oxalic acid to precipitate Th A significant amount of REEs precipitated with Th. Oxalic acid was
leachate (chloride) not very selective to remove Th.
31,729 mg/L REEs
410 mg/L Th
Huang et al. (2019) Y, Ln Al, Ca, Mg Commercial REE 1. MgO slaking solution (80 °C) was used to precipitate Developed a non-ammonia enrichment process for extracting REEs
concentrate leachate mixed REE(OH)3 at pH 8.9 from ion adsorption type deposits. The REE precipitation efficiency
(sulfate) was 99.6%. Purity of the REE precipitate was 86 wt%. Too much
30 g/L REO MgO co-precipitated with REE hydroxides.
Hamza et al. REEs, U Al, Fe, Zn, Cu, Ni, Co, Mn, Ti Abu Mogherat ore leachate 1. (NH4)2S to precipitate ZnS Developed a series of selective precipitation processes to remove
(2019) (sulfate) 2. NH3/H2O to precipitate Fe(OH)3 at pH 4 impurities from the ore leachate. In some steps, acetic acid was
184 mg/L REE 3. NH3/H2O to precipitate Al(OH)3 at pH 5 added to the liquor (7 mL/L) to prevent re-dissolution of the
80 mg/L U 4. Oxalic acid to precipitate REEs at pH 1.5 precipitates. No significant losses of REEs or U occurred. The final
5. NaOH to precipitate Na2U3O7 at pH 9 REE product contained some Cu, Ni, Ti, and Co because these
elements co-precipitated with REE oxalates.
Silva et al. (2019) Sc, Y, Ln Ca, Mg, Mn, Fe, Al, Th, U, P, Na Ore leachate (sulfate) 1. CaCO3 to increase pH to 3.5 Steps 1 and 2 removed all Fe, P, and Th, and most of U and Al. Step
3.29 g/L REEs 2. Ca(OH)2 to increase pH to 5.0 3a or 3b were used to recover REE compounds as precipitates.
3a. Na2HPO4 to precipitate REE phosphates and Na/REE Higher REE recoveries were found using 3a, but these precipitates
double sulfates (20–70 °C) also contained higher levels of Ca and U. If Fe, Al, and Th are not
3b. Na2SO4 to precipitate Na/REE double sulfates (70 °C) removed in Steps 1 and 2, they will co-precipitate with REEs in Step
3a/3b. The final product contained < 100 mg U/kg. There was no

17
significant tendency towards individual REE selectivity.
Amer et al. (2018) La, Ce, Pr, Nd, Sm, Th, U, P, Fe, Ti, Si Rosetta monazite digest 1. Sodium phosphate crystallization at 60 °C Step 2 extracted 95% of REEs while U and Th were undetectable. The
Eu, Gd, Y (caustic soda) 2. Hydrous cake washed, selectively leached with LREE double sulfates were precipitated from Step 2 leachate.
ammonium sulfate (300 g/L, 120 °C) to extract LREEs, Uranium was precipitated from Step 3 leachate by adding 1 M
which were then precipitated as double sulfates ammonium hydroxide at pH 13. The HREEs tended to follow
3. Selective leaching of U with sodium carbonate/ towards Step 4.
bicarbonate.
4. Residue dissolved in 4 M HCl. Thorium precipitated with
oxalic acid at pH 0.35.
Borai et al. (2018) La, Ce, Nd, Y Mg, Al, Si, P, Cl, K, Ca, Ti, Mn, Egyptian monazite 1. Added concentrated HCl 1.2% (v/v) The HCl or NaCl additions in Step 1 were important to minimize REE
Cr, Fe, Zn, As, Sr, Zr, Pd, Sn, I, concentrates leachate 2. Diluted 1:14 to increase hydrolyses of REE and Th co-precipitation. Step 2 improved Th precipitation efficiency which
Hf, W, Ir, Pb, Th, U (sulfate) 3. Added 2.5% Na-pyrophosphate (4% v/v) reached 99% in Step 3. Final Th precipitate contained about 15%
4.9 g/L REEs REEs and 2.6% Fe(III). Also investigated optimizing the ratio of free
4.22 g/L Th acids. Boiling the solution increased the co-precipitation of Fe with
2.88 g/L U no effect on Th precipitation efficiency. The H2O2 additions had a
0.36 g/L Fe negative effect, causing oxidation of more Fe to Fe(III), which
interfered with Th precipitation.
da Silva et al. Sc, Y, Ln Ca, Mg, Mn, Fe, Al, Th, U, P, Na, Monazite leachate (sulfate) 1a. CaCO3 to precipitate Fe (pH 3.5) followed by Ca(OH)2 to Developed a selective precipitation scheme to purify monazite ore
(2018) Ti, Si 3.53 g/L REEs precipitate more Fe (pH 5.0) leachate. All Fe and P were precipitated through either Step 1a, 1b,
1.36 g/L Fe 1b. CaCO3 to pH 5 1c, or 1e. Aluminum was only precipitated through Step 1c or 1e.
6.9 mg/L Th 1c. Ca(OH)2 to pH 5 Optimized process achieved 98% removal of U, and 100% removal of
2.8 mg/L U 1d. MgO to pH 5 Fe, Th, P, Al. Final REE recovery was 85%.
2.57 g/L other impurities 1e. NaOH to pH 5
Önal et al. (2017) Nd, Pr, Dy Fe, B, Co, Ni, Cu, Ga, Al, Si, Mn NdFeB leachate (sulfate) 1. MnO2 to oxidize Fe(II) to Fe(III) Recovered REEs from leached magnet powders by selective
10.3 g/L REEs 2. Ca(OH)2 or MnO to precipitate Fe hydroxides (pH > 3) precipitation techniques. MnO was used to increase pH and
22.2 g/L Fe 3. Electrolysis for Mn/Co recovery precipitate Fe hydroxide. The remaining solution contained Mn and
2.5 g/L other impurities Co, which were recovered by electrolysis, but with low efficiency.
Using Ca(OH)2 to increase pH resulted in 23% REE losses.
(continued on next page)
Hydrometallurgy 196 (2020) 105435
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

thorium and uranium, which co-leach during chemical beneficiation of

loss of REEs occurred in Step 1 at higher BaCl2 concentrations. Step 2

precipitated all Th, but also a significant amount of REEs (especially
Recovered REEs from high phosphate material (25% P2O5). Some
REEs.

4.3.1. Behaviour of iron

The usual approach for selective precipitation of iron is the hydro­
HREEs). Step 4 was used to partition REEs from Fe. xide route in which the pH of acidic REE leach solutions is raised to
selectively precipitate ferric hydroxide according to reaction (28):
Fe(3aq+) + 3OH(aq) = Fe (OH )3(s) (28)
To have a successful hydroxide precipitation, the oxygen potential
of the solution should be high enough that most iron is present as ferric
ion (Qiu et al., 2012). Ferrous hydroxide cannot be selectively pre­
cipitated from REE solutions because hydrolyses of REE ions and pre­
Summary and comments

cipitation of REE hydroxides occur simultaneously (cf. Appendix C and

Rizkalla and Choppin, 1991). The choice of reagent for pH adjustment
also plays an important role on the efficiency and selectivity of the
process.
Silva et al., 2019 and da Silva et al., 2018 found that iron was re­
moved to undetectable levels from REE sulfate leach liquors by a two-
step process in which the pH was first increased to 3.5 by calcium
carbonate addition and then subsequently increased to 5.0 by adding
1. BaCl2 (1–20 g/L) to co-precipitate Th, U with BaSO4

4. Added oxalic acid to precipitate mixed REE oxalate

calcium hydroxide. However, this process also resulted in the co-pre­

cipitation of virtually all scandium in the solution. Iron may also be
2. Added NH3/H2O to pH 1.05 to precipitate Th

precipitated by increasing the pH with sodium hydroxide. Magnesium

oxide was not as efficient at removing iron as calcium hydroxide. A
possible explanation is the solubility of reaction products. In the sulfate
leach liquor, calcium forms sparingly soluble sulfates (gypsum) whereas
magnesium sulfates have higher solubility. It is possible that calcium
hydroxide enhances the removal of iron by its co-precipitation with
gypsum, whereas magnesium oxide yields little to no such precipitate.
Önal et al., 2017 found that increasing pH to above 3 using calcium
Precipitation scheme

hydroxide resulted in simultaneous REE losses as they co-precipitated

(separates Fe)
3. Added HCl

with gypsum. Therefore, the amount of calcium in solution should be

considered when developing an iron precipitation process. The beha­
viour of REEs during gypsum precipitation has been investigated in
detail by Dutrizac, 2017. Hamza et al., 2019 found that raising the pH
of the leach liquor to 4.0 with aqueous ammonia was successful in
precipitating iron hydroxide with little REE losses.
Other possibilities exist for selective precipitation besides the ferric
concentrates leachate
Egyptian monazite

hydroxide route. Silva et al., 2019 reported that it is possible to selec­

tively precipitate iron from REE containing solutions by the addition of
Feed material

sodium sulfate, but temperature must be kept low (~20 °C), so that
(sulfate)

sodium REE double sulfates do not precipitate. On the other hand, so­
dium REE double sulfates have been selectively precipitated from iron
in sulfuric acid NdFeB magnet leachates by adding sodium hydroxide
Na, Mg, Al, Si, P, Ca, Ti, Cr, Mn,
Fe, Ni, Zn, Zr, Hf, W, Pb, Th, U

(Lyman and Palmer, 1993). Precipitation of REE oxalates according to

Other element(s) in solution

reaction (22) can be conducted without co-precipitation of iron (Borai

et al., 2016). Likewise, REE carbonate precipitation according to reac­
tion (20) occurs with minimal co-precipitation of ferric oxide (Liu et al.,
2004; Rychkov et al., 2018). El Afifi et al., 2019 used inorganic sulfate,
sulfide, and iodide additions to selectively precipitate iron and thorium
from REE solutions, but the composition of the precipitate was not
characterized.

4.3.2. Behaviour of thorium and uranium

Targeted element

A few different approaches have been taken to separate thorium and

La, Ce, Nd, Gd
(s) in solution

uranium from REEs by selective precipitation. Thorium hydroxide is

more readily separated by selective precipitation compared with ur­
anium(VI) hydroxide (Zhu et al., 2015). A vital but sometimes over­
looked aspect of separation of uranium and REEs has been the im­
Table 4 (continued)

portance of the valence state of uranium (Silva et al., 2019). Uranium

Borai et al. (2016)

(IV) behaves similarly to thorium and can be readily separated from

Investigators

REEs by selective precipitation of its hydroxide. The corresponding

reactions are given below:
Th (4aq+) + 4OH(aq) = Th (OH )4(s) (29)

18
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

U(4aq+) + 4OH(aq) = U (OH )4(s) (30) scandium are to be recovered.

Silva et al., 2019 were able to reduce the aluminum content from
On the other hand, uranyl (UO22+)requires a different approach 4600 to 100 mg/L, by increasing the pH of REE sulfate leach liquors to
because its hydroxide precipitates at the pH close to REE hydroxides. 5.0. Similarly, Hamza et al., 2019 and Kolar et al., 2016 also found that
Silva et al., 2019 and da Silva et al., 2018 selectively precipitated a considerable amount of aluminum precipitated upon increasing pH to
thorium to undetectable levels in REE sulfate leach liquors by a two- 5.0 using aqueous ammonia or other bases. The REE oxalates can be
stage process in which pH was first increased to 3.5 by calcium car­ selectively precipitated from aluminum-containing solutions according
bonate addition, and then subsequently increased to 5.0 by adding to reaction (22), but excess oxalate is required because aluminum
calcium hydroxide. This process also resulted in a 65% reduction of consumes oxalic acid by forming complexes in the solution (Dahlberg
uranium in the solution (from 35to 12 mg/L). Most likely, tetravalent et al., 2014; Xia and Griffith, 2018). Carbonate precipitation of REEs
uranium is the species that precipitated according to reaction (30) upon has been conducted according to reaction (20) with low co-precipita­
this pH increase, while hexavalent uranium remained soluble (i.e., tion of aluminum (Liu et al., 2004; Rychkov et al., 2018).
UO22+). da Silva et al., 2018 also found that adjusting the pH with
magnesium oxide or sodium hydroxide was successful in selectively 4.3.4. Behaviour of base metals and other impurities
precipitating thorium and some uranium. Alternatively, additions of Many minor impurities exist in REE leachates and their behaviour
inorganic sulfate, sulfide, or iodide were successful at precipitating with regards to precipitation of REE compounds is important. In par­
thorium (El Afifi et al., 2019). The most effective reagent here was ticular, the tendency of some impurities to co-precipitate with REE
potassium iodide, which precipitated 99% of thorium. hydroxides, carbonates, or oxalates should be noted and is an important
Hamza et al., 2019 performed selective precipitation of REEs from consideration for REE producers. Although calcium and magnesium
uranium containing solutions by oxalic acid addition at pH 1.5. The hydroxides are more soluble than REE hydroxides, they sometimes co-
REE oxalates precipitated according to reaction (22), while uranium precipitate with REE precipitates and contaminate the final REE pro­
oxalates remained highly soluble. The form of uranium in the solution duct (Huang et al., 2019). According to Silva et al., 2019 and da Silva
in this case was predominantly uranyl. Silva et al., 2019 also found that et al., 2018, REE phosphates are selectively precipitated from sulfate
REEs can be selectively precipitated from solutions containing hex­ liquors containing calcium and magnesium when pH is increased to 5.0
avalent uranium using sodium sulfate at elevated temperatures (70 °C) with either calcium carbonate, calcium hydroxide, sodium hydroxide,
to form insoluble sodium REE double sulfates. However, this approach or magnesium oxide. To minimize calcium and magnesium co-pre­
is not effective for thorium containing solutions because thorium will cipitation with REE oxalates, oxalate precipitation should be conducted
co-precipitate with REE double sulfates (Kul et al., 2008; Zhu et al., with excess reagent at elevated temperatures (35% excess oxalic acid,
2015). Silva et al., 2019 also found disodium phosphate was not an 60 °C; Silva et al., 2019b). Calcium and magnesium do not co-pre­
effective precipitant for REE/U separations since uranium co-pre­ cipitate significantly with REE carbonates in reaction (20) (Liu et al.,
cipitated with REEs. 2004, Rychkov et al., 2018). In recycling of consumed acid from
According to Borai et al., 2016, aqueous ammonia is the preferred leaching of REEs from phosphogypsum, Walawalkar et al., 2016
reagent to precipitate thorium hydroxide according to reaction (29) showed calcium removal from solution is enhanced by seeding with
because it minimizes the co-precipitation of REEs. In a later work, Borai calcium sulfate anhydrite which rejects calcium from the solution and
et al., 2018 utilized the relatively low solubility of thorium phosphate enhances REE recovery.
to selectively precipitate thorium from REEs using sodium pyropho­ In precipitation of REE oxalates, minor impurities have been noted
sphate. In this process, the pyrophosphates of cerium(IV) and zirconium to co-precipitate. This includes titanium, silicon, zirconium (Khawassek
(IV) may also co-precipitated. Additions of chloride (using HCl or NaCl) et al., 2019); silicon (Amer et al., 2018); and manganese, cobalt, nickel,
were found to suppress the co-precipitation of REEs. and copper (Hamza et al., 2019). Hamza et al., 2019 found that zinc
Precipitation of REE oxalates should not be conducted when sulfide was selectively precipitated by adding ammonium sulfide. Borai
thorium is present in solution because thorium co-precipitates sig­ et al., 2018 found that a REE‑lead precipitate forms from monazite li­
nificantly and contaminates the REE oxalates (Amer et al., 2018; Güneş quors upon additions of sulfate or phosphate; however, the precipitated
et al., 2019). However, thorium oxalate can be selectively dissolved by compound was not identified.
washing with sodium carbonate or sodium bicarbonate solutions ac­ Radium is a radioactive decay product of thorium-232 and uranium-
cording to reaction (31) (Amer et al., 2013; Zhu et al., 2015): 238 and must be removed from some REE processing solutions. Similar
to most heavy alkaline earths, its sulfate is sparingly soluble and may be
Th (C2 O4 )2(s) + 4Na2 CO3(aq) + 2NaHCO3(aq)
co-precipitated with barium sulfate according to reaction (33) (Lucas
= Na6 Th (CO3 )5(aq) + 2Na2 C2 O4(aq) + H2 O(l) + CO2(g ) (31) et al., 2015):
This technique decreases the ratio of ThO2/REO to 0.007% with RaCl2(aq) + BaCl2(aq) + 2Na2 SO4(aq) = (Ba, Ra) SO4(s) + 4NaCl (aq) (33)
only 3% loss of REEs. Thorium tends to form anionic complexes with
Barium chloride and sodium sulfate are introduced to form and
carbonate which stabilize thorium in solution; therefore, in some pro­
precipitate barium sulfate, which is isomorphous with radium sulfate
cesses, selective precipitation of REEs from thorium is aided by the
and removes radium from the solution.
presence of sodium carbonate in solution (Zhu et al., 2015).
Heavy metals such as lead or zinc can be selectively precipitated as
sulfides (Lucas et al., 2015; Hamza et al., 2019). For example, pre­
4.3.3. Behaviour of aluminum cipitation of lead sulfide from chloride solutions can be conducted ac­
Mixed success has been found in selective precipitation of aluminum cording to reaction (34):
hydroxide from REE solutions. If aluminum is present in large quan­
tities, a significant amount can be removed by selective precipitation PbCl2(aq) + Na2 S(s ) = PbS(s ) + 2NaCl (aq) (34)
according to reaction (32): Sulfide precipitation is also effective from sulfate solutions (Lucas
Al(3aq+) + 3OH(aq) = Al (OH )3(s) et al., 2015).
(32)
Fluoride is present in some REE processing solutions and if large
However, final traces of aluminum remain in solution which co- amounts of iron are present, REE fluorides may be precipitated (Lyman
precipitate with REE hydroxides (Qiu et al., 2012). Yttrium and scan­ and Palmer, 1993). These precipitates cause issues because they are
dium hydroxides tend to co-precipitate with aluminum hydroxide; thus, gelatinous and difficult to filter. Therefore, magnesium oxide is used to
the hydroxide precipitation route should not be used when yttrium and fix fluorine in solution. Borate is present in some NdFeB leachates and

19
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

can be selectively precipitated as zinc borate (Lyman and Palmer, situ according to reaction (35):
1993). 1 11
Ce3 + (H2 O )8 + 2SeO32 + H3 O+ + O2 = Ce 4 + (SeO32 )2 (H2 O )4 + H2 O
4 2
4.4. Other methods
(35)

Besides solvent extraction, ion exchange or adsorption, and selec­ During crystallization of cerium selenite (Ce(SeO3)2), thorium im­
tive precipitation, other methods for impurity removal during REE purities are sequestered in crystals through their substitution with
processing are of interest to overcome shortcomings in existing tech­ cerium(IV) (i.e., (Ce,Th)(SeO3)2). In a single step, quantitative removal
nologies. Solvent extraction is well-suited for handling large volumes of of thorium occurred with thorium/REE separation factors on the order
dilute liquors and large tanks help even out fluctuations in REE and of 103–105. Cerium is a lower-value REE and usually present is rela­
impurity contents, helping with process control. However, solvent ex­ tively high proportions in primary REE sources; therefore, its loss may
tractants are expensive and inevitably degrade or are lost to the aqu­ not be much of a practical issue. It was proposed that the process can be
eous phase; thus, large volumes of usually flammable solvents are re­ extended to other f-block elements. Recently, another selective crys­
quired, large circuits are necessary, and many reagents/products are tallization technique was discovered by Yin et al., 2017 for efficient
not environmentally friendly. Ion exchange or adsorption are effective separation of individual REEs by selective borate crystallization. It
at producing high purity products; however, they have low yields, low seems the borate crystal system may be another promising avenue to
throughput, high cost, and often require environmentally toxic solu­ pursue for impurity removal by selective crystallization.
tions (Lucas et al., 2015). Precipitation is a simple technique, but can be
time consuming, requiring expensive solid-liquid separation tanks and 4.4.3. Supercritical fluid extraction
reagents (e.g., oxalic acid). Moreover, precipitates with poor crystal­ Zhang et al., 2018 and Yao et al., 2017 studied the supercritical fluid
linity are hard to clarify, and can have low yield and purity. Therefore, extraction of REEs from solid waste products comprising end-of-life
there is a need to develop novel methods for impurity removal during NdFeB magnets and NiMH battery anodes. They utilized carbon dioxide
REE processing that are simple, environmentally friendly, with high as the solvent to selectively extract REEs directly from the solid waste,
throughput, and producing pure REE products. while minimizing the co-extraction of impurities, such as iron, nickel,
Original research articles since 2015 pertaining to impurity removal cobalt, and aluminum. Some amount of tributyl phosphate was in­
by novel techniques not involving solvent extraction, ion exchange or cluded in order to chelate REEs and solubilize them in supercritical
adsorption, or selective precipitation, were compiled and systematically carbon dioxide. This research presents an important development that
categorized according to: REEs can be selectively extracted from solid feed materials without the
need for any pre-treatments such as acid leaching or roasting.
1. Type of method
2. Targeted elements in the feed (usually REEs) 4.4.4. Extraction utilizing ionic liquids
3. Other elements present in the feed (Fe, Al, Ca, etc.) The use of ionic liquids (IL) as extractants, diluents, or both is at­
4. Feed material (chemistry, synthetic or ore, etc.) tractive owing to their selectivity towards REEs, low vapour pressure,
5. Summary and comments (separation efficiencies, recoveries, chal­ thermal stability, and low flammability; therefore, there is a growing
lenges, etc.) interest towards them (Liu et al., 2012; Wang et al., 2017c). Generally,
three types of ILs are used for REE extraction. Either pure ILs (Vander
The results are summarized and presented in chronological order in Hoogerstraete et al., 2013; Larsson and Binnemans, 2017), IL with
Table 5. No prior review is available on the removal of impurities from conventional solvent extractants (Xiong et al., 2017), or functionalized
REE processing solutions by novel techniques. ILs (Davris et al., 2016). The structure of some ILs used for REE ex­
A number of promising developments have taken place regarding traction is provided in Appendix D. Extraction mechanisms in ILs and
separation and impurity removal in REE processing since 2015. Some of variants may involve cation exchange, anion exchange, or neutral ex­
the investigators listed in Table 5 may utilize concepts from conven­ traction (Wang et al., 2017c). Supported ILs and variants allow these
tional impurity removal techniques; however, their combination or materials to be used in solid-liquid separation and help increase contact
application is novel and brings important scientific contributions to area (Avdibegović et al., 2018). Many emerging ILs are highly selective
REE processing. Specifically, emerging ideas of selective dissolution or towards REEs and can be stripped without the use of strong acids.
leaching, selective crystallization, and supercritical fluid extraction for Davris et al., 2016 selectively leached REEs from bauxite residue
REE recovery are promising. using a functionalized IL, [Hbet][Tf2N]. Up to 85% of REEs were ex­
tracted with < 3% co-extraction of iron and minimal co-extraction of
4.4.1. Selective dissolution or leaching aluminum, calcium, and sodium. In [Hbet][Tf2N], REEs are extracted
Tunsu et al., 2016 developed a selective leaching process to extract on account of the solubility of their oxides according to reaction (36)
REEs from waste fluorescent lamp powder. First, the material was (Wang et al., 2017c):
treated using an iodine/potassium iodide solution to selectively leach
REE2 O3(s) + 6H+ = 2REE 3 + + 3H2 O (36)
mercury. Then, the product was leached in two steps using nitric acid.
First, with low concentration of nitric acid (1 M) and short time (1 h), The REEs were stripped from the ionic liquid using hydrochloric
impurities were selectively leached from the powder. Afterwards, REEs acid or water. Oxides of common troublesome impurities (e.g., iron
were selectively leached using a higher concentration of nitric acid oxide or aluminum oxide) are sparingly soluble in ILs so the use of ILs as
(2 M) and longer time (2 h). The process utilized different leaching selective leachants for REE oxides is very promising. The results of
kinetics for REEs and impurities to maximize REE extraction while Davris et al., 2016 are remarkable considering their bauxite residue
minimizing co-extraction of impurities. contained only 0.14 wt% REOs along with 43 wt% Fe2O3 and 25 wt%
Al2O3.
4.4.2. Selective crystallization Xiong et al., 2017 discovered a synergistic effect by combining IL
Wang et al., 2018a developed a selective selenite crystallization [P66614][Cl] and an organophosphorous solvent extractant, di(2-
strategy to remove thorium from REE solutions. The process depends on ethylhexyl) 2-ethylhexyl (DEHEHP), for the extraction of REE nitrates
the cerium in the feed to sequester thorium. Hydrothermal reaction of and rejection of impurities. Lithium nitrate, added for its salting-out
REE-containing feed and selenium dioxide was conducted at 230 °C for effect, significantly enhanced REE extraction according to reactions
3 days. During the process, cerium(III) was oxidized to cerium(IV) in (37) and (38):

20
 Table 5
Summary of original research articles on impurity removal during REE processing using novel techniques.a
Investigators Method Targeted metal Impurity element(s) in feed Feed material Summary and comments
element(s) in feed
W.D. Judge and G. Azimi

Liang et al. (2019) Selective cloud point extraction U Th, La, Nd, Yb Synthetic Cloud point extraction was performed using TX-114 and Cyanex 301
0.05 mM metal ions to selectively extract U(VI) in the presence of Th and REEs. Over 90%
0.1 M NaCl of U was removed with virtually no co-extraction of REEs.
Altaş et al. (2018) Selective dissolution Th La, Ce, Pr, Nd, Sm, Eu, Gd Synthetic mixed oxalate precipitate The mixed oxalate (Th with REEs) was selectively leached for Th in
Th:REE = 4:13.1 (by mass) ammonium oxalate solution. The best leaching was achieved at
pH = 6, up to 97% dissolution efficiency of Th was achieved while the
solution was mostly free of REEs.
Avdibegović et al. Supported ionic liquid phase Sc, Y, Nd, Dy Ca, Al, Fe, Si, Ti Synthetic (nitrate, chloride, sulfate) Although impurities were co-extracted, selective elution was
(2018) [Hbet-STFSI-PS-DVB] 4.4 mM REE performed on a column using H3PO4 and HNO3. Except Sc, the REEs
5.5 mM impurities eluded individually and were enriched with respect to starting
Bauxite residue leachate (sulfate) impurities.
14 mg/L REEs
10,800 mg/L impurities
Wang et al. (2018a) Selective selenite crystallization Th Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Dy, Synthetic (nitrate) Single step quantitative removal of Th from REE solutions was
Er, Yb 90 mM REEs performed by in-situ oxidation of Ce(III) to Ce(IV), which sequesters f-
10 mM Th block elements. Separation factors of up to 105 were achieved. The
0.1–0.5 M SeO2 process required heating to 230 °C for 3 days.
Whitty-Léveillé et al. Selective dissolution U, Th, LREE, HREE Fe, Si, Al Commercial monazite, bastnäsite, xenotime First leach was 0.5 h in 0.5 M NaHCO3 + 0.5 M Na2CO3 at 80 °C to
(2018) minerals; various Canadian REE-bearing extract 67% of U. Next was leach at 4.7 M HCl, 0.5 h, 20 °C to recover
ores 71% of LREE, 56% of HREE, 5% of Fe, 5% of Si.
Zhang et al. (2018) Aeriometallurgical (supercritical Nd, Pr, Dy Fe, B NdFeB permanent magnet Performed SCFE using CO2 and TBP/HNO3 to selectively extract REEs

21
fluid extraction, SCFE) 22.2 wt% REEs from solid NdFeB magnets. Achieved > 91% REE extraction with
only 62% Fe co-extraction in 1–2 h. Methanol addition was found to
enhance the separation factors for REEs compared with Fe.
Yao et al. (2017) Selective SCFE La, Ce, Pr, Nd Ni, Co, Al, Mn NiMH battery anode Conducted selective SCFE using CO2 and TBP/HNO3 to extract REEs
29.8 wt% REEs from solid NiMH battery anodes. Achieved up to 90% REE extraction
in 1–2 h. Extraction was enhanced by methanol additions. Also
examined an extraction mechanism using synthetic REOs.
Xiong et al. (2017) Ionic liquid-based synergistic Pr, Lu Al, Ca, Mg, Mn, Co, Ni, Cu, Ba, Li Synthetic (nitrate) A synergistic effect occurs with [P66614][Cl] and DEHEHP. The
extraction 150 mg/L REEs loading capacity for Pr was about 0.78 M and stripping was performed
1350 mg/L impurities with H2O. The distribution ratio for impurities was about 10 while
that for Lu was 110. No diluent was used.
Borai et al. (2016) Selective dissolution La, Ce, Nd, Gd Na, Mg, Al, Si, P, Ca, Ti, Cr, Mn, Egyptian monazite Monazite was washed with 0.1–1.0 M HCl at 60 °C with shaking for
Fe, Ni, Zn, Zr, Hf, W, Pb, Th, U 30 min. The process removed > 90% Na, 40% Ca, 30% Si while only
losing about 3% of REEs.
Davris et al. (2016) Selective leaching with ionic liquid Sc, Y, La, Ce, Pr, Nd, Fe, Al, Ca, Ti, Si, Na Bauxite residue Selectively leached REEs using HbetTf2N ionic liquid. About 70–85%
Sm, Eu, Gd, Dy, Er, Yb 0.14 wt% REOs of REE were extracted with < 3% Fe and little to no Al, Ca, Na. The
60–180 °C pulp density was 3–20% and stripping was performed with HCl or
water.
Tunsu et al. (2016) Selective leaching with aqueous Y, La, Ce, Eu, Gd, Tb Al, Ba, Ca, Fe, Mg, Mn, Hg, Na, Fluorescent lamp powder Selectively leached Hg using I2/KI solution, then selectively leached
solution Sb, Sr, B, Cd, Cr, Cu, K, Ni, Pb, Si, 18 wt% REEs REEs with nitric acid. Utilized the different leaching kinetics to
Sn, Ti, W, Zn, Zr selectively leach impurities, followed by a second leaching step to
selectively leach REEs. The process was followed by solvent extraction
processing.

a
Cyanex 301 = bis(2,4,4-trimethylpentyl)dithiophosphinate; HbetTf2N = betainium bis(trifluoromethylsulfonyl)imide; [Hbet-STFSI-PS-DVB] = betainium sulfonyl (trifluoromethane-sulfonylimide) poly(styrene-co-
divinylbenzene); NiMH = nickel metal hydride; [P66614][Cl] = tri(hexyl)tetradecylphosphonium chloride (Cyphos IL 101); SCFE = supercritical fluid extraction; TBP = tributyl phosphate. TX-114 = polyoxyethylene
(8) octylphenyl ether.
Hydrometallurgy 196 (2020) 105435
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

[P 66614][Cl](IL) + LiNO3(aq) = [P 66614][NO3 ](IL) + LiCl(aq) (37)

Highly effective (selective leaching or selenite

[P 66614][NO3 ](IL) + REE (NO3 )3(aq) = [P 66614] REE(NO3 )4(IL) (38)

At the same time, DEHEHP extracted REEs by neutral complexation

according to reaction (39):

Effective (SCFE or ionic liquid)

xDEHEHP(IL) + REE (NO3 )3(aq) = REE(NO3 )3 xDEHEHP(IL) (39)

Although there was no obvious interaction between DEHEHP and the

Highly effective (SCFE)

Effective (ionic liquid)

IL, there was a clear synergistic effect. With regards to impurity removal,
separation factors between lutetium and either magnesium, aluminum,
crystallization) calcium, manganese, cobalt, nickel, copper, zinc, and barium were be­
tween 12 and 14 despite the molar concentration of these impurities
being greater than lutetium. This is quite promising since REEs can be
Other

difficult to extract in high purity from nitrate solutions.

–

Can be effective (by sel. ppt. of Sn, Bi hydroxides –

Can be effective (by sel. ppt. of Fe(III) hydroxide)
Effective (sel. ppt. of REE phosphates or oxalates)

Can be effective (if [HCl] is high or fluoride ions

sel. SX = selective solvent extraction; sel. IX = selective ion exchange or adsorption; sel. ppt. = selective precipitation; SCFE = supercritical fluid extraction.
5. Discussion
Can be effective (by sel. ppt. of Th and U(IV)
Effective (by sel. ppt. of Al hydroxide or REE

Highly effective (by sel. ppt. of Ti, Zr, Hf

Can be effective (by sel. ppt. of sulfides)

As evident by the preceding sections, tremendous scientific efforts

have been recently devoted towards impurity removal during REE
processing. A comparative summary of all REE impurity removal
techniques and their performance with respect to certain impurities is
presented in Table 6. Established processes exist for removing most
Selective precipitation

impurities; however, their application must be considered on a case-by-

case basis for each REE flowsheet. The behaviour of specific impurities
is considered below and methods are briefly compared and contrasted
or Pb sulfide)
are present)

hydroxides)

hydroxides)

for their removal from REE processing solutions.

oxalates)

5.1. Removal of Iron

Can be effective (requires selective elution)

The ferrous and ferric ions behave quite differently in solution and
Can be effective (if Fe(III)/Fe(II) ratio is

Effective (may require selective elution)

their ratio in solution changes by each processing step or reaction with

dissolved gases, species in solution, and solids. Ferric iron co-extracts
Can be effective (sel. IX of Th, U
Highly effective (except with Sc)

with REEs in cation solvent extraction and ion exchange; thus, impurity
transfer can be reduced by minimizing the concentration of ferric iron
in the feed material (Peng et al., 2017; Zhang et al., 2016). Precipitation
of ferric hydroxide is a convenient means to remove ferric iron from
solution and is a usual procedure conducted prior to solvent extraction
recommended)
Ion exchange

or ion exchange (Kolar et al., 2016; Hamza et al., 2019). Prior to solvent
minimized)

extraction or ion exchange, the solution may be treated with metallic

Typical performance for impurity removal during REE processing

Effective

Effective
Effective

iron powder to ensure the reduction of ferric iron. Any ferrous iron in
Comparative summary of different impurity removal techniques in REE processing.a

solution will co-precipitate with REE hydroxides. There may be a pos­

sibility to selectively precipitate ferrous sulfide from REE-containing
Can be effective (requires selective stripping), but not

Can be effective (if Fe(III)/Fe(II) ratio is minimized)

solutions, but this approach has not been fully investigated. Other
Can be effective (sel. SX of Th, U recommended)

promising techniques for iron removal involve novel solvents with low
iron solubility such as supercritical carbon dioxide (Zhang et al., 2018)
or ionic liquids (Davris et al., 2016).
Can be effective (except Cr(III)), but not

Can be effective but not recommended

Highly effective (except with Sc)

5.2. Removal of thorium and uranium

Thorium and uranium typically co-extract during cation solvent

extraction or ion exchange separation of REEs and must be removed
Solvent extraction

prior to individual REE separations. Selective precipitation of thorium

Highly effective
recommended)

and uranium(IV) hydroxides is an effective means for separating these

recommended

impurities (Zhu et al., 2015; Silva et al., 2019). However, this procedure
Effective

cannot separate uranium(VI) (i.e., uranyl, UO2+) whose hydroxide

precipitates at higher pH along with REE hydroxides. To separate REEs
from uranyl, selective precipitation of REE oxalates is effective, but if
Ti, Zr, Hf, V, Nb, Ta, Cr, Mo,

any thorium or uranium(IV) remains in solution, they will co-pre­

cipitate (Altaş et al., 2018). The highest level of removal of thorium and
uranium is achieved by performing a separate solvent extraction or ion
Mn, Co, Ni, Cu, Zn

exchange operation specifically designed to selectively extract these

elements by anion exchange from REE-containing solutions (Reynier
Sn, Pb, Bi
Impurity

et al., 2018; Peng et al., 2017). In this approach, high separation factors
Ca, Mg
Table 6

W
Th, U

are realized for Th/REE or U/REE, but it is difficult to simultaneously

Fe
Al

Si

achieve high separation factors for both Th/REE and U/REE in a single

22
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

step (Reynier et al., 2018). It is always more difficult to separate limit. Most approaches to remove silica from solution involve additives
thorium and uranium from the HREEs, especially lutetium (Wang et al., which control or stabilize precipitates to avoid gel formation. Silica's
2017b). Owing to the lanthanide contraction, HREEs have relatively gel structure is destabilized under more acidic conditions and in the
smaller ionic radii (and higher charge density) that are close in value to presence of more electronegative anions (Cheremisina et al., 2019);
those of thorium and uranium(IV), so they can behave similar to these thus, many processes incorporate intensely acidic solutions (e.g., 8 M
impurities. Alternative techniques include selenite crystallization HCl) or addition of inorganic salts (e.g., NaF, AlCl3). The best approach,
(Wang et al., 2018a) and selective leaching (Altaş et al., 2018) as a of course, is to avoid the dissolution of silica in the first place. Because
promising proactive approach to thorium and uranium control. of silica's different solution chemistry, any soluble silica which remains
in solution does not usually co-extract with REEs in ion exchange or
5.3. Removal of aluminum solvent extraction (Ogata et al., 2016b).

Aluminum presents a major challenge to REE producers owing to its 5.6. Removal of manganese, cobalt, nickel, copper, and zinc
trivalence. Typically, aluminum co-extracts with REEs, at least to some
extent, in cation solvent extraction and ion exchange operations. Fourth period transition metal impurities not discussed previously
Despite this, there is still opportunity to separate aluminum by per­ (Mn, Co, Ni, Cu, Zn) do not usually co-extract with REEs during cation
forming stepwise selective stripping or selective elution (Battsengel solvent extraction or ion exchange (Gergoric et al., 2017a; Kim et al.,
et al., 2018; Ogata et al., 2016b). Although this is effective in both 2015; Li et al., 2019; Lou et al., 2019). In some instances, there may be
solvent extraction and ion exchange, it is not recommended for the co-extraction of manganese(III) with REEs (Wilfong et al., 2017). As
former because aluminum is a troublesome element that accumulates well, highly acidic chloride environments seem to lead to some co-ex­
and causes many problems in solvent extraction circuits. Selective tractions (Ogata et al., 2015a). In terms of selective precipitation, these
precipitation of aluminum hydroxide can be effective to remove the elements cannot be separated by hydroxide precipitation because they
majority of aluminum; however, the final traces will co-precipitate with precipitate close to the same pH as REE hydroxides. The best separation
REE hydroxides (Kolar et al., 2016). The best separation by precipita­ by precipitation is achieved by selective precipitation of the transition
tion is achieved by selective precipitation of REE oxalates; however, the metal sulfide upon addition of a sulfide source (e.g., Na2S). Another
presence of aluminum in solution requires excess oxalates (Dahlberg promising approach for NdFeB magnets is selective leaching with nitric
et al., 2014; Xia and Griffith, 2018). A promising alternative technique acid, which extracts REEs, while leaving nickel in a solid residue
involves the use of ionic liquids, which have low solubility of aluminum (Gergoric et al., 2017a).
complexes (Davris et al., 2016).
5.7. Removal of refractory metals
5.4. Removal of calcium and magnesium
Refractory metal impurities (Ti, Zr, Hf, V, Nb, Ta, Cr, Mo, W) are
Calcium and magnesium are not particularly problematic in REE becoming increasingly important as new REE minerals and secondary
processing. Neither element co-extracts much in usual REE cation sol­ sources bearing these elements become exploited. Contrary to their
vent extraction or ion exchange operations. There may be a slight title, they are not always chemically inert and participate in many
tendency for calcium and magnesium to follow the LREEs at high leaching operations, such as those for bauxite residue and zircon. Once
concentrations (Huang et al., 2017). The exception is REE operations liberated, many form higher valent oxyanionic species, which can be
specifically engineered for scandium recovery, where calcium and selectively precipitated from REE-containing solutions as oxyhydr­
magnesium have a high tendency to co-extract perhaps because of their oxides. This has been used in the case of titanium, zirconium, and
similar cation size (Ye et al., 2019). hafnium (Ye et al., 2019). Unfortunately, their behaviour in solvent
When calcium and magnesium are present in solution, they can co- extraction and ion exchange has not been fully elucidated, but owing to
precipitate slightly with REE hydroxides, carbonates, oxalates, oxysul­ their high valence, their cations can be expected to co-extract during
fates, or double sulfates, especially when calcium or magnesium are REE extraction by cation exchange (El-Nadi, 2017). In ion exchange,
present in high concentrations (Huang et al., 2019; Chi et al., 2003; Kul their co-extraction with REEs can be mitigated by a multi-stage elution
et al., 2008; Silva et al., 2019b). When using CaO or MgO to raise pH, scheme (Khawassek et al., 2019; Avdibegović et al., 2018; Smirnov
calcium or magnesium may become entrained in REE precipitates as et al., 2017). In the case of solvent extraction, it is not recommended to
CaO, MgO, or insoluble gypsum or Mg(OH)2. Alternatively, it is pos­ be an effective means to remove these impurities, as they are well
tulated that high levels of supersaturation lead to poor crystallization known to cause crud problems in circuits. Operations specifically de­
and high surface Gibbs free energy of REE precipitates that encourages signed for thorium removal can expect some co-extraction of refractory
absorption of foreign calcium or magnesium impurity ions which leads metals (Wang et al., 2018b; Hu et al., 2019).
to reduced surface energy (Huang et al., 2019).
The most effective separation by precipitation involves selective 5.8. Removal of tin, lead, and bismuth
precipitation of REE phosphates (Silva et al., 2019). In any operation
involving sulfate solutions, the level of calcium should be monitored to The heavy metals tin, lead, and bismuth are associated with some
avoid precipitation of gypsum (Dutrizac, 2017). If gypsum is inevitably hard rock mining operations of REEs. They do not usually co-extract
to precipitate, then the amount of co-precipitated REEs can be mini­ with REEs during cation solvent extraction or ion exchange and these
mized by using special additives (Sadri et al., 2018). are effective in separations. However, it is not recommended to feed
these elements to solvent extraction circuits as they are well known to
5.5. Removal of silicon form crud and interfere with the circuits (Ritcey, 1980). Tin and bis­
muth hydroxides are selectively precipitated from solution prior to REE
Solubilized silica is only an issue in REE processing under certain hydroxides. In the case of lead, its hydroxide co-precipitates at similar
circumstances. When its concentration exceeds about 0.1 g/L in solu­ pH to the REEs; therefore, hydroxide precipitation is not an effective
tion, a gelatinous silica phase forms that is impossible to filter. This means for lead removal. Similarly, lead has been noticed to co-pre­
usually occurs during leaching of certain minerals, like eudialyte cipitate with REE sulfates (Borai et al., 2018). The most effective means
(Davris et al., 2017), where silica is temporarily co-dissolved with other to selectively precipitate lead is as lead sulfide by addition of a sulfide
constituents during breakdown of the mineral, but it can happen in any source to solution (e.g., Na2S). Selective sulfide precipitation has also
operation in which the concentration of silica is beyond the solubility been used to remove zinc from REE solutions (Hamza et al., 2019).

23
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

5.9. Removal of other impurities

Although many industrially relevant impurities have already been

discussed, there are many other impurities that may be necessary to
treat in REE processing solutions. Boron (as borate or boric acid) and
gallium can be present in NdFeB magnet leachates, but do not usually
co-extract with REEs, except gallium in highly acidic chloride en­
vironments (Kim et al., 2015). Alkali metals do not cause any issues
with co-extraction but can cause salting out or form jarosite in the
presence of iron and sulfate. It is possible to wash some REE ores with
HCl to selectively remove > 90% of alkalis (Borai et al., 2016). Radium
can be selectively precipitated with barium sulfate through addition of
barium chloride and sodium sulfate to solution (Lucas et al., 2015).
The non-metals sometimes encountered in aqueous processing (e.g., P,
C, S, Cl, F) are not usually an issue since they are mostly burned off upon
firing of REE precipitates in air (Domingues et al., 1962). The behaviour
of other impurities not discussed can be inferred by their relationship to
any analogous elements already discussed or by some considerations
presented in the following section. Any amphoteric metal can be expected
to cause crud formation in solvent extraction circuits.
Fig. 7. Carbonate precipitation diagram for REEs and selected impurities at
25 °C.

6. Process development for impurity removal

Carbonate precipitation is frequently used to produce intermediate
Despite the seemingly endless combinations and possibilities of REE products. The carbonate precipitation diagram presented in Fig. 7
impurities present in REE- containing solutions, impurity removal can indicates metal carbonates precipitate in a comparatively narrower range
be predicted with a reasonable degree of accuracy by considering the compared with metal hydroxides. Therefore, the opportunity for selective
behaviour of individual components. That is, the reality of several in­ precipitation in the carbonate system is not as feasible and most industrial
terdependent equilibria involving, for example, hydroxide ions can be mixed REE carbonates contain 5 wt% impurity carbonates (Lucas et al.,
simplified by considering a single equilibrium established between 2015). According to Fig. 7, co-precipitation of UO2CO3, FeCO3, and
hydroxide ions and metal hydroxides. For instance, the ability of im­ PbCO3 is expected during REE carbonate precipitation. On the other
purity removal by selective precipitation can be readily inferred by the hand, REE carbonate precipitation occurs freely without contamination
hydroxide precipitation diagram presented in Fig. 6. This diagram was from ferric iron (Liu et al., 2004; Rychkov et al., 2018). Impurities not
constructed by the present authors with the most up-to-date data (cf. included in Fig. 7, such as aluminum, do not form carbonates, but rather
Appendix C) using the usual methods (Monhemius, 1977). The ordinate hydrolyze and precipitate as hydroxides (Zuo et al., 2014).
represents the thermodynamic equilibrium concentration of metal ion Certain troublesome impurities encountered during leaching of ion
in solution for a given pH represented on the abscissa. adsorption clays (mainly Cu, Pb) cannot be partitioned by hydroxide
Fig. 6 reveals that impurity hydroxides including ZrO(OH)2, TiO systems. Therefore, the sulfide precipitation diagram presented in Fig. 8
(OH)2, Fe(OH)3, Th(OH)4, and U(OH)4 are selectively precipitated to a was constructed. Fig. 8 reveals that impurity sulfides including Cu2S, CuS,
large degree, prior to REE precipitation. On the other hand, UO2(OH)2, Fe2S3, PbS, ZnS, NiS, CoS, and FeS selectively precipitate before most REE
Al(OH)3, and Fe(OH)2 do not selectively precipitate and will con­ sulfides. Thus, the sulfide system presents an opportunity to partition
taminate REE hydroxides, in line with the results of laboratory in­ some impurities that are difficult or impossible to remove otherwise.
vestigations (Silva et al., 2019). Thus, a higher ratio of U(IV)/U(VI) and Oxalate precipitation is a common finishing operation after individual
Fe(III)/Fe(II) in solution facilitates removal of uranium and iron to a REE separations; therefore, the oxalate precipitation diagram presented in
higher degree in the hydroxide system. Fig. 9 was constructed. Indeed, Fig. 9 reveals thorium and uranium(IV)
oxalates co-precipitate and contaminate REE oxalates, as observed

Fig. 6. Hydroxide precipitation diagram for REEs and selected impurities at

25 °C. Fig. 8. Sulfide precipitation diagram for REEs and selected impurities at 25 °C.
24
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Comparing speciation diagrams for REEs, thorium, uranium(IV),

and uranium(VI) reveals their dissimilar valency, which also corre­
sponds to dissimilarly charged complexes in solution. Here, thorium
and uranium(IV) tend to form cationic or neutral species at high sulfate
concentrations, while uranium(VI) tends to form neutral or anionic
species. Again, it is postulated that speciation partly explains why ca­
tionic ion exchange resins co-extract REEs, uranium, and thorium from
sulfate solutions without much selectivity, while anionic resins are
more preferential towards thorium and uranium(VI) (Ang et al., 2017;
Amaral et al., 2018). Furthermore, this is consistent with observations
of strong base anion exchange resins extracting uranium(VI) (which
more readily forms anionic complexes) over thorium (Ang et al., 2017).
Similarly, in solvent extraction, amine extractants (based on anion ex­
change) are more selective towards thorium and uranium(VI) over REEs
(Zhu and Cheng, 2011; Amaral et al., 2018). This also explains the
better extractions for nitrate media, in which, thorium and uranium
have a high tendency to form anionic species.
The preceding discussion was meant to illustrate the importance of
understanding the solution chemistry and partition mechanisms in the
development of impurity removal processes. Impurities which form si­
Fig. 9. Oxalate precipitation diagram for REEs and selected impurities at 25 °C.
milar complexes to REEs in solution may co-extract or alter the stability of
REE complexes and affect extraction (Li et al., 2018). In this respect, the
experimentally (Pilkington and Wylie, 1947; Amer et al., 2018; Güneş oxygen potential in solution is of great importance since it affects mul­
et al., 2019). However, REE oxalates can be selectively precipitated from tivalent metals and determines the redox ratio in solution (e.g., Fe(II)/Fe
uranium(VI) and iron owing to the higher solubility of these impurity (III)). The best opportunity for selective extraction of a particular species
oxalates, as corroborated by Hamza et al., 2019 and Borai et al., 2016. occurs when a peculiarity in its chemistry can be exploited, such as so­
Fig. 9 shows lead, nickel, calcium, and copper only co-precipitate with lubility constants (e.g., Silva et al., 2019), ion exchange/adsorbent ma­
REE oxalates when present at high concentrations or when REEs are re­ terial structure (e.g., Lou et al., 2019), or solvent extractants (e.g., Lu
moved to very low concentrations by large oxalate additions. Free oxalate et al., 2016).
is also consumed in solution by impurities such as iron and aluminum to
form soluble complexes (Dahlberg et al., 2014; Xia and Griffith, 2018). 7. Outlook and future work
Therefore, impurity removal by oxalate precipitation is a delicate op­
eration requiring knowledge of impurities present and careful con­ Although much progress has been made in impurity removal during
sideration of oxalate addition; however, Fig. 9 helps in this endeavor. REE processing, there remain many challenges as more REE sources
In solvent extraction and ion exchange, REEs or targeted impurity come online. Constantly evolving technological landscapes always re­
ions are usually extracted from chemically complex aqueous solutions quire increasingly pure REE products. However, there are not many
of high ionic strength. Here, a requisite for extraction from the aqueous studies on the effects of impurities in REE products. If the effect and
phase is that the hydration and complexation of the extracted ion in the tolerance of impurities could be better understood, there would be
aqueous phase must be broken or adjusted to allow its transfer to the opportunities for cost reduction and mutual benefits for REE producers
new phase (organic solvent or ion exchange resin). Therefore, solution and consumers.
thermodynamics and speciation play a critical role in determining ex­ Much research has focused on the recovery of REEs and impurity
traction. As an example, speciation diagrams for REEs, iron(II), iron removal from dilute solutions; however, more research is needed on
(III), aluminum, thorium, uranium(IV), and uranium(VI) presented in highly concentrated solutions as this is what tonnage REE producers work
Fig. 10 were constructed to illustrate the dependence of species on the with. Attention must be given to the chemistry of solutions; as well the
sulfate ion concentration. structure of ion exchange resins, adsorbents, and solvent extractants must
Speciation diagrams presented in Fig. 10 were constructed using be known to increase the selectivity for REEs. Design of new extractants
HYDRA and MEDUSA programs (Puigdomenech, 2006). Model solu­ specifically targeted for impurity removal is very promising.
tions were considered containing either 10 mM neodymium (1.44 g/L), In solvent extraction, even properties of the diluent, such as po­
10 mM lutetium (1.75 g/L), 30 mM iron (1.68 g/L), 70 mM aluminum larity, have been observed to have an impact on processing. Use of
(1.89 g/L), 1 mM thorium (0.23 g/L), or 1 mM uranium (0.24 g/L). phase modifiers to achieve faster separation and reduce emulsification
Ionic strength was set to 1.25 M, pH to 1.0, and temperature to 25 °C has not been investigated for many new REE systems. Similarly, the
while the sulfate ion concentration was varied from 0.01 to about 3.2 M effect of contact time has not been optimized for many systems but is
to simulate a commercial REE leach solution. For all elements, Fig. 10 another avenue for reducing impurity co-extraction.
shows a number of complexes form and decompose as the sulfate ion In ion exchange and adsorption, much research has focused on
concentration changes. achieving fast adsorption rates and high adsorption capacities.
Fig. 10 reveals that impurities which carry the same charge as REEs However, adsorption rates inevitably increase as more concentrated
and form analogous sulfate complexes are also the impurities industry has solutions are dealt with. For recovery of REEs, desorption is as critical
found the most problematic with regards to co-extraction, i.e., iron(III) as adsorption, but is often overlooked. Materials with fast adsorption
and aluminum (Zhang et al., 2016). It is postulated that the similar co­ rates and high adsorption capacities may also be difficult to strip; thus,
ordination environment of iron(III), aluminum, and REEs explains the more research is needed on desorption behaviours.
difficulty found in selectively separating REEs from iron(III) and alu­ Many emerging techniques for REE extraction and impurity removal
minum in sulfate solutions (Rychkov et al., 2016; Battsengel et al., 2018). are also promising. Conventional processes meeting the needs of in­
Thus, reducing iron(III) to iron(II) helps mitigate co-extraction because dustry are often environmentally harmful; hence, there is a need to
iron(II) carries a different charge and forms different complexes from develop new environmentally friendly processes that achieve the same
REEs. Lesser explored methods for iron(III) reduction include reduction technical requirements. Development of breakthrough processes can
through fine metallic iron powder, sulfurous acid, or electrolysis. have a profound influence on sustainability and open new REE

25
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Fig. 10. Speciation diagrams for sulfate leach solutions at pH 1, 1.25 M ionic strength, and 25 °C for (a) neodymium, (b) lutetium, (c) iron(III), (d) iron(II), (e)
aluminum, (f) thorium, (g) uranium(VI), and (h) uranium(IV).

26
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

resources that were previously inaccessible. The use of new solvents, REEs and impurities can usually be selectively eluted. Co-extraction of
like ionic liquids and supercritical fluids, are gaining traction owing to iron can be reduced by minimizing the ration of Fe(III)/Fe(II) in solu­
the potential of selective extraction of REEs with minimal impurities. tion. Anion exchange resins are highly effective at selectively extracting
uranium and thorium from REEs in solution.
8. Conclusions Selective precipitation can be utilized to perform effective separa­
tions of REEs and certain impurities. From acidic leach solutions, iron
A comprehensive literature review was conducted to summarize (III), thorium, and uranium(IV) hydroxides selectively precipitate prior
recent progress in impurity removal during REE processing. A total of to REEs when increasing pH. However, aluminum, iron(II), and ur­
87 original research articles covering the years 2015–2019 were re­ anium(VI) hydroxides co-precipitate with REE hydroxides and con­
viewed in detail and systematically categorized to help draw broad taminate the product. The REE oxalates can be selectively precipitated
conclusions about the behaviour of impurities during REE processing. from solutions containing aluminum, iron(II), and uranium(VI) with
Whenever possible, the behaviour of a number of impurities was fol­ some loss in precipitation efficiency as these impurities tie up oxalate in
lowed (including Al, Ca, Mg, Fe, Si, Th, U, Ti, Zr, Hf, Cr, Mo, Mn, Co, Ni, complexes.
Cu, Zn, Sn, Pb, and Bi). Impurity removal techniques including solvent
extraction, ion exchange and adsorption, and selective precipitation Declaration of competing interest
were covered as well as other emerging novel techniques.
Solvent extraction can be an effective technique for selectively se­
The authors declare that they have no known competing financial
parating REEs or certain impurities. Operations aimed at extracting
interests or personal relationships that could have appeared to influ­
REEs usually observe some co-extraction of aluminum, iron(III),
ence the work reported in this paper.
thorium, and uranium if they are present in solution. Although alu­
minum and iron(III) can be separated from REEs at the stripping stage,
their accumulation in circuits will lead to emulsification and crud for­ Acknowledgements
mation. Co-extraction of iron can be mitigated by maintaining a low
ratio of Fe(III)/Fe(II). Separating REEs from uranium and thorium The authors gratefully acknowledge funding through Natural
usually requires a separate solvent extraction step to selectively remove Resources Canada (NRCan) and the CanmetMINING REE R&D Program.
uranium and thorium with extractants utilizing anion exchange. Kind review of the manuscript and thoughtful suggestions from Ms.
Ion exchange and adsorption can be a very effective method for Eliza Ngai (CanmetMINING) are greatly appreciated.
selectively separating REEs or certain impurities and many promising
new materials have been developed in recent years. Since most REE ion Author statement
exchange operations are based on cationic exchange, some co-extrac­
tion of aluminum, iron(III), thorium, and uranium can be expected G.A. conceived and supervised the research. W.D.J. performed the
when these impurities are present in feed materials. This is not as much literature search, data analysis and interpretation. All authors con­
as a problem as in solvent extraction because in column arrangements, tributed to writing and revising the manuscript.

Appendix A. Structure of solvent extractants

Table A1
Structures and pKa values of common acidic organophosphorous extractants (after Zhang et al., 2015).

Brand name Full name and abbreviation(s) Chemical structure pKa

P204 Di-(2-ethylhexyl) phosphoric acid (D2EHPA, DEHPA, or HDEHP) 3.24

P507 2-Ethylhexyl phosphoric acid mono-2-ethylhexyl ester (HEHEHP) 4.51

P229 Di(2-ethylhexyl) phosphinic acid (PIA-8) 4.98

Cyanex 272 Di-(2,4,4-trimethylpentyl) phosphinic acid (HBTMPP) 6.37

(continued on next page)

27
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table A1 (continued)

Brand name Full name and abbreviation(s) Chemical structure pKa

Cyanex 302 Bis(2,4,4-trimethylpentyl) monothiophosphinic acid (HBTMPTP) 5.63

Cyanex 301 Bis(2,4,4-trimethylpentyl) dithiophosphinic acid (HBTMPDTP) 2.61

Table A2
Structures and pKa of some common carboxylic acid extractants (after Zhang et al., 2015).

Brand name Full name and abbreviation(s) Chemical structure pKa

Versatic 10 Neodecanoic acid 7.33

HA Naphthenic acid 6.91

CA-12 Sec-octylphenoxyl acetic acid 3.52

CA-100 Sec-nonylphenoxy acetic acid 3.53

28
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table A3
Structures of common neutral extractants (after Zhang et al., 2015).

Brand name Full name and abbreviation Chemical structure

TBP Tributylphosphate (TBP)

P350 Di-(1-methyl-heptyl) methyl phosphonate (DMHMP)

P503 Di-(2-ethylhexyl) 2-ethylhexyl phosphonate (DEHEHP)

Cyanex 921 Tri-n-octylphosphine oxide (TOPO)

Cyanex 923 Trialkylphosphine oxides (TRPO)

(continued on next page)

29
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table A3 (continued)

Brand name Full name and abbreviation Chemical structure

Cyanex 925 Trialkylphosphine oxides (TRPO)

ODP n-octyl diphenyl phosphate (ODP)

TODGA Tetraoctyl-diglycolamide (TODGA)

30
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table A4
Structures of common basic extractants (after Zhang et al., 2015).

Brand names Full name and abbreviation Chemical structure

N1923, Primene JM-T Primary amine (CnH2n+1)2CHNH2

n = 9–11
N235, Alamine 336 Trialkylamine (CnH2n+1)3N
n = 8–10
N263, Aliquat 336 Methyltrioctylamine chloride [(C8H17)3N+CH3]Cl−

Appendix B. Functional groups for ion exchange and adsorption materials

Table B1
Structure of common cation exchange resins (after Masram, 2013).

Functional group name and abbreviation Chemical structure on polystyrene resin (charged with H+)

Sulfonic acid or corresponding salt (–SO3−)

Carboxylic acid or corresponding salt (–COO−)

Phosphonic acid or corresponding salt (–PO3−)

Phosphinic acid or corresponding salt (–HPO2−)

31
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table B2
Structure of common anion exchange resins (after Masram, 2013).

Functional group name and abbreviation Chemical structure on polystyrene resin (charged with OH−)

Primary amine (–NH3+)

Secondary amine (–NH2R+)

Tertiary amine (–NHR1R2+)

Quaternary amine (–NR1R2R3+)

32
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Appendix C. Solubility products of REE and impurity precipitates

Table C1
Solubility products of REE hydroxides (HSC Chemistry 10.0).

Reaction log Ksp,

25 °C

Sc(OH)3(s) ⇌ Sc(aq)3+ + 3OH(aq)− −30.7

Y(OH)3(s) ⇌ Y(aq)3+ + 3OH(aq)− −31.8
La(OH)3(s) ⇌ La(aq)3+ + 3OH(aq)− −21.4
Ce(OH)3(s) ⇌ Ce(aq)3+ + 3OH(aq)− −23.7
Ce(OH)4(s) ⇌ Ce(aq)4+ + 4OH(aq)− −47.7
Pr(OH)3(s) ⇌ Pr(aq)3+ + 3OH(aq)− −25.8
Nd(OH)3(s) ⇌ Nd(aq)3+ + 3OH(aq)− −24.7
Sm(OH)3(s) ⇌ Sm(aq)3+ + 3OH(aq)− −24.0
Eu(OH)3(s) ⇌ Eu(aq)3+ + 3OH(aq)− −23.7
Gd(OH)3(s) ⇌ Gd(aq)3+ + 3OH(aq)− −22.4
Tb(OH)3(s) ⇌ Tb(aq)3+ + 3OH(aq)− −24.6
Dy(OH)3(s) ⇌ Dy(aq)3+ + 3OH(aq)− −27.2
Ho(OH)3(s) ⇌ Ho(aq)3+ + 3OH(aq)− −26.3
Er(OH)3(s) ⇌ Er(aq)3+ + 3OH(aq)− −27.1
Tm(OH)3(s) ⇌ Tm(aq)3+ + 3OH(aq)− −24.5
Yb(OH)3(s) ⇌ Yb(aq)3+ + 3OH(aq)− −25.4
Lu(OH)3(s) ⇌ Lu(aq)3+ + 3OH(aq)− −20.6

Table C2
Solubility products of REE carbonates (after Firsching and Mohammadzadei,
1986).

Reaction log Ksp,

25 °C

Sc2(CO3)3(s) ⇌ 2Sc(aq)3+ + 3CO3(aq)2− −35.8

Y2(CO3)3(s) ⇌ 2Y(aq)3+ + 3CO3(aq)2− −31.5
La2(CO3)3(s) ⇌ 2La(aq)3+ + 3CO3(aq)2− −29.9
Ce2(CO3)3(s) ⇌ 2Ce(aq)3+ + 3CO3(aq)2− −35.1
Pr2(CO3)3(s) ⇌ 2Pr(aq)3+ + 3CO3(aq)2− −33.2
Nd2(CO3)3(s) ⇌ 2Nd(aq)3+ + 3CO3(aq)2− −34.1
Sm2(CO3)3(s) ⇌ 2Sm(aq)3+ + 3CO3(aq)2− −34.4
Eu2(CO3)3(s) ⇌ 2Eu(aq)3+ + 3CO3(aq)2− −35.0
Gd2(CO3)3(s) ⇌ 2Gd(aq)3+ + 3CO3(aq)2− −35.5
Tb2(CO3)3(s) ⇌ 2Tb(aq)3+ + 3CO3(aq)2− −34.9
Dy2(CO3)3(s) ⇌ 2Dy(aq)3+ + 3CO3(aq)2− −34.0
Ho2(CO3)3(s) ⇌ 2Ho(aq)3+ + 3CO3(aq)2− −32.8
Er2(CO3)3(s) ⇌ 2Er(aq)3+ + 3CO3(aq)2− −28.3
Tm2(CO3)3(s) ⇌ 2Tm(aq)3+ + 3CO3(aq)2− −31.6
Yb2(CO3)3(s) ⇌ 2Yb(aq)3+ + 3CO3(aq)2− −31.7
Lu2(CO3)3(s) ⇌ 2Lu(aq)3+ + 3CO3(aq)2− −32.2

Table C3
Solubility products of REE sulfides (HSC Chemistry 10.0).

Reaction log Ksp,

25 °C

Sc2S3(s) ⇌ 2Sc(aq)3+ + 3S(aq)2− −42.8

Y2S3(s) ⇌ 2Y(aq)3+ + 3S(aq)2− −22.7
La2S3(s) ⇌ 2La(aq)3+ + 3S(aq)2− −16.5
Ce2S3(s) ⇌ 2Ce(aq)3+ + 3S(aq)2− −12.6
CeS4(s) ⇌ Ce(aq)4+ + 2S(aq)2− −47.6
Pr2S3(s) ⇌ 2Pr(aq)3+ + 3S(aq)2− −12.5
Nd2S3(s) ⇌ 2Nd(aq)3+ + 3S(aq)2− −9.86
Sm2S3(s) ⇌ 2Sm(aq)3+ + 3S(aq)2− −19.5
EuS(s) ⇌ Eu(aq)2+ + S(aq)2− +8.04
Gd2S3(s) ⇌ 2Gd(aq)3+ + 3S(aq)2− −17.5
Tb2S3(s) ⇌ 2Tb(aq)3+ + 3S(aq)2− −21.4
Dy2S3(s) ⇌ 2Dy(aq)3+ + 3S(aq)2− −21.9
Ho2S3(s) ⇌ 2Ho(aq)3+ + 3S(aq)2− −21.3
Er2S3(s) ⇌ 2Er(aq)3+ + 3S(aq)2− −23.9
Tm2S3(s) ⇌ 2Tm(aq)3+ + 3S(aq)2− −18.3
Yb2S3(s) ⇌ 2Yb(aq)3+ + 3S(aq)2− −22.8
Lu2S3(s) ⇌ 2Lu(aq)3+ + 3S(aq)2− −27.0

33
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table C4
Solubility products of REE oxalates (after Josso et al., 2018, except where noted).

Reaction log Ksp,

25 °C

Sc2(C2O4)3(s) ⇌ 2Sc(aq)3+ + 3C2O4(aq)2− −27.4a

Y2(C2O4)3(s) ⇌ 2Y(aq)3+ + 3C2O4(aq)2− −29.3
La2(C2O4)3(s) ⇌ 2La(aq)3+ + 3C2O4(aq)2− −29.2
Ce2(C2O4)3(s) ⇌ 2Ce(aq)3+ + 3C2O4(aq)2− −30.4
Pr2(C2O4)3(s) ⇌ 2Pr(aq)3+ + 3C2O4(aq)2− −30.9
Nd2(C2O4)3(s) ⇌ 2Nd(aq)3+ + 3C2O4(aq)2− −30.9
Sm2(C2O4)3(s) ⇌ 2Sm(aq)3+ + 3C2O4(aq)2− −31.4
Eu2(C2O4)3(s) ⇌ 2Eu(aq)3+ + 3C2O4(aq)2− −31.4
Gd2(C2O4)3(s) ⇌ 2Gd(aq)3+ + 3C2O4(aq)2− −31.4
Tb2(C2O4)3(s) ⇌ 2Tb(aq)3+ + 3C2O4(aq)2− −30.8
Dy2(C2O4)3(s) ⇌ 2Dy(aq)3+ + 3C2O4(aq)2− −30.7
Ho2(C2O4)3(s) ⇌ 2Ho(aq)3+ + 3C2O4(aq)2− −30.3
Er2(C2O4)3(s) ⇌ 2Er(aq)3+ + 3C2O4(aq)2− −30.1
Tm2(C2O4)3(s) ⇌ 2Tm(aq)3+ + 3C2O4(aq)2− −30.0
Yb2(C2O4)3(s) ⇌ 2Yb(aq)3+ + 3C2O4(aq)2− −30.0
Lu2(C2O4)3(s) ⇌ 2Lu(aq)3+ + 3C2O4(aq)2− −30.2
a
After Sokolov Sokolov, 1960.

Table C5
Solubility products of impurity hydroxides (HSC Chemistry 10.0, except where
noted).

Reaction log Ksp,

25 °C

Zr(OH)4(s) ⇌ Zr(aq)4+ + 4OH(aq)− −68.1

Ti(OH)4(s) ⇌ Ti(aq)4+ + 4OH(aq)− −63.3
Sn(OH)4(s) ⇌ Sn(aq)4+ + 4OH(aq)− −57.1
U(OH)4(s) ⇌ U(aq)4+ + 4OH(aq)− −55.0
Th(OH)4(s) ⇌ Th(aq)4+ + 4OH(aq)− −46.5
In(OH)3(s) ⇌ In(aq)3+ + 3OH(aq)− −44.2
Fe(OH)3(s) ⇌ Fe(aq)3+ + 3OH(aq)− −37.5
Bi(OH)3(s) ⇌ Bi(aq)3+ + 3OH(aq)− −36.2
ZrO(OH)2(s) ⇌ ZrO(aq)2+ + 2OH(aq)− −34.8
Ga(OH)3(s) ⇌ Ga(aq)3+ + 3OH(aq)− −34.6
Al(OH)3(s) ⇌ Al(aq)3+ + 3OH(aq)− −31.3
Cr(OH)3(s) ⇌ Cr(aq)3+ + 3OH(aq)− −31.0
TiO(OH)2(s) ⇌ TiO(aq)2+ + 2OH(aq)− −29.0a
Sn(OH)2(s) ⇌ Sn(aq)2+ + 2OH(aq)− −26.3
UO2(OH)2(s) ⇌ UO2(aq)2+ + 2OH(aq)− −23.5
Cu(OH)2(s) ⇌ Cu(aq)2+ + 2OH(aq)− −21.6
Be(OH)2(s) ⇌ Be(aq)2+ + 2OH(aq)− −21.4
Cr(OH)2(s) ⇌ Cr(aq)2+ + 2OH(aq)− −17.5
Zn(OH)2(s) ⇌ Zn(aq)2+ + 2OH(aq)− −16.5
Fe(OH)2(s) ⇌ Fe(aq)2+ + 2OH(aq)− −16.4
Ni(OH)2(s) ⇌ Ni(aq)2+ + 2OH(aq)− −15.3
Co(OH)2(s) ⇌ Co(aq)2+ + 2OH(aq)− −14.7
Pb(OH)2(s) ⇌ Pb(aq)2+ + 2OH(aq)− −14.7
Cd(OH)2(s) ⇌ Cd(aq)2+ + 2OH(aq)− −14.4
Mn(OH)2(s) ⇌ Mn(aq)2+ + 2OH(aq)− −13.2
Mg(OH)2(s) ⇌ Mg(aq)2+ + 2OH(aq)− −11.3
Ca(OH)2(s) ⇌ Ca(aq)2+ + 2OH(aq)− −5.41
Sr(OH)2(s) ⇌ Sr(aq)2+ + 2OH(aq)− −0.47
LiOH(s) ⇌ Li(aq)+ + OH(aq)− +1.98
Ba(OH)2(s) ⇌ Ba(aq)2+ + 2OH(aq)− +3.43
NaOH(s) ⇌ Na(aq)+ + OH(aq)− +6.93
KOH(s) ⇌ K(aq)+ + OH(aq)− +10.6
a
After Landis and Kaye, 1971.

34
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table C6
Solubility products of impurity carbonates (HSC Chemistry 10.0).

Reaction log Ksp,

25 °C

UO2CO3(s) ⇌ UO2(aq)2+ + CO3(aq)2− −19.2

PbCO3(s) ⇌ Pb(aq)2+ + CO3(aq)2− −12.9
Ag2CO3(s) ⇌ 2Ag(aq)+ + CO3(aq)2− −12.0
CdCO3(s) ⇌ Cd(aq)2+ + CO3(aq)2− −11.2
NiCO3(s) ⇌ Ni(aq)2+ + CO3(aq)2− −11.0
FeCO3(s) ⇌ Fe(aq)2+ + CO3(aq)2− −10.5
BeCO3(s) ⇌ Be(aq)2+ + CO3(aq)2− −10.3
ZnCO3(s) ⇌ Zn(aq)2+ + CO3(aq)2− −9.89
CuCO3(s) ⇌ Cu(aq)2+ + CO3(aq)2− −9.88
MnCO3(s) ⇌ Mn(aq)2+ + CO3(aq)2− −9.85
CoCO3(s) ⇌ Co(aq)2+ + CO3(aq)2− −9.35
SrCO3(s) ⇌ Sr(aq)2+ + CO3(aq)2− −9.34
BaCO3(s) ⇌ Ba(aq)2+ + CO3(aq)2− −8.75
CaCO3(s) ⇌ Ca(aq)2+ + CO3(aq)2− −8.33
MgCO3(s) ⇌ Mg(aq)2+ + CO3(aq)2− −5.10
Li2CO3(s) ⇌ 2Li(aq)+ + CO3(aq)2− −3.25
Na2CO3(s) ⇌ 2Na(aq)+ + CO3(aq)2− +0.80
K2CO3(s) ⇌ 2K(aq)+ + CO3(aq)2− +4.71

Table C7
Solubility products of impurity sulfides (HSC Chemistry 10.0).

Reaction log Ksp,

25 °C

Bi2S3(s) ⇌ 2Bi(aq)3+ + 3S(aq)2− −105

Fe2S3(s) ⇌ 2Fe(aq)3+ + 3S(aq)2− −88.3
In2S3(s) ⇌ 2In(aq)3+ + 3S(aq)2− −81.4
Ga2S3(s) ⇌ 2Ga(aq)3+ + 3S(aq)2− −77.0
SnS2(s) ⇌ Sn(aq)4+ + 2S(aq)2− −54.0
V2S3(s) ⇌ 2V(aq)3+ + 3S(aq)2− −51.1
Ag2S(s) ⇌ 2Ag(aq)+ + S(aq)2− −49.2
Cu2S(s) ⇌ 2Cu(aq)+ + S(aq)2− −48.1
TiS2(s) ⇌ Ti(aq)4+ + 2S(aq)2− −41.9
Cr2S3(s) ⇌ 2Cr(aq)3+ + 3S(aq)2− −39.7
ZrS2(s) ⇌ Zr(aq)4+ + 2S(aq)2− −37.3
CuS(s) ⇌ Cu(aq)2+ + S(aq)2− −35.9
HfS2(s) ⇌ Hf(aq)4+ + 2S(aq)2− −34.3
SnS(s) ⇌ Sn(aq)2+ + S(aq)2− −29.1
CdS(s) ⇌ Cd(aq)2+ + S(aq)2− −28.8
US2(s) ⇌ U(aq)4+ + 2S(aq)2− −28.3
PbS(s) ⇌ Pb(aq)2+ + S(aq)2− −28.1
ZnS(s) ⇌ Zn(aq)2+ + S(aq)2− −24.1
NiS(s) ⇌ Ni(aq)2+ + S(aq)2− −23.1
CoS(s) ⇌ Co(aq)2+ + S(aq)2− −22.3
FeS(s) ⇌ Fe(aq)2+ + S(aq)2− −19.0
ThS2(s) ⇌ Th(aq)4+ + 2S(aq)2− −15.3
MnS(s) ⇌ Mn(aq)2+ + S(aq)2− −13.5
CrS(s) ⇌ Cr(aq)2+ + S(aq)2− −10.9
CaS(s) ⇌ Ca(aq)2+ + S(aq)2− −1.83
BaS(s) ⇌ Ba(aq)2+ + S(aq)2− +2.25
SrS(s) ⇌ Sr(aq)2+ + S(aq)2− +2.65
MgS(s) ⇌ Mg(aq)2+ + S(aq)2− +4.91
BeS(s) ⇌ Be(aq)2+ + S(aq)2− +10.5
Li2S(s) ⇌ 2Li(aq)+ + S(aq)2− +11.8
Na2S(s) ⇌ 2Na(aq)+ + S(aq)2− +14.0
K2S(s) ⇌ 2K(aq)+ + S(aq)2− +20.4

35
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table C8
Solubility products of impurity oxalates (HSC Chemistry 10.0, except where noted).

Reaction log Ksp,

25 °C

Zr(OH)2C2O4(s) ⇌ Zr(aq)4+ + 2OH(aq)− + C2O4(aq)2− −42.1a

Th(C2O4)2(s) ⇌ Th(aq)4+ + 3C2O4(aq)2− −24.3b
U(C2O4)2 • 6H2O(s) ⇌ U(aq)4+ + 3C2O4(aq)2− + 6H2O(l) −21.4c
Hg2C2O4(s) ⇌ 2Hg(aq)2+ + C2O4(aq)2− −12.8d
Ag2C2O4(s) ⇌ 2Ag(aq)2+ + C2O4(aq)2− −11.3d
PbC2O4(s) ⇌ Pb(aq)2+ + C2O4(aq)2− −10.1d
NiC2O4 • 2H2O(s) ⇌ Ni(aq)2+ + C2O4(aq)2− + 2H2O(l) −9.96
FeC2O4 • 2H2O(s) ⇌ Fe(aq)2+ + C2O4(aq)2− + 2H2O(l) −9.50
CuC2O4(s) ⇌ Cu(aq)2+ + C2O4(aq)2− −9.35d
CaC2O4 • 3H2O(s) ⇌ Ca(aq)2+ + C2O4(aq)2− + 3H2O(l) −9.27
CaC2O4 • 2H2O(s) ⇌ Ca(aq)2+ + C2O4(aq)2− + 2H2O(l) −9.38
CaC2O4 • H2O(s) ⇌ Ca(aq)2+ + C2O4(aq)2− + H2O(l) −8.80
ZnC2O4 • 2H2O(s) ⇌ Zn(aq)2+ + C2O4(aq)2− + 2H2O(l) −8.83
CoC2O4 • 2H2O(s) ⇌ Co(aq)2+ + C2O4(aq)2− + 2H2O(l) −8.07
UO2C2O4 • 3H2O(s) ⇌ UO2(aq)2+ + C2O4(aq)2− + 3H2O(l) −7.89e
CdC2O4 • 3H2O(s) ⇌ Cd(aq)2+ + C2O4(aq)2− + 3H2O(l) −7.84d
MgC2O4 • 2H2O(s) ⇌ Mg(aq)2+ + C2O4(aq)2− + 2H2O(l) −6.77d
MnC2O4 • 2H2O(s) ⇌ Mn(aq)2+ + C2O4(aq)2− + 2H2O(l) −5.50d
BaC2O4 • 3.5H2O(s) ⇌ Ba(aq)2+ + C2O4(aq)2− + 3.5H2O(l) −3.96
BaC2O4 • 2H2O(s) ⇌ Ba(aq)2+ + C2O4(aq)2− + 2H2O(l) −3.84
BaC2O4 • 0.5H2O(s) ⇌ Ba(aq)2+ + C2O4(aq)2− + 0.5H2O(l) −3.65
Na2C2O4(s) ⇌ 2Na(aq)2+ + C2O4(aq)2− +4.92
K2C2O4(s) ⇌ 2K(aq)2+ + C2O4(aq)2− +1.06
a
After Kobayashi et al. (2009b).
b
After Kobayashi et al. (2009a).
c
After Zakharova and Moskvin (1960).
d
After Rumble (2019).
e
After Leturcq et al. (2008).

Appendix D. Structure of ionic liquids

Table D1
Structure of cations and anions of ionic liquids for REE extraction (after Wang et al., 2017c).

Name and abbreviation Chemical structure

[HBet]+

[P66614]+

[A336]+

[PF6] –
(continued on next page)

36
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Table D1 (continued)

Name and abbreviation Chemical structure

[Tf2N]−

References Chen, Z., Wang, W., Sang, F., Xu, J., Luo, G., Wang, Y., 2017. Fast extraction and en­
richment of rare earth elements from waste water via microfluidic-based hollow
droplet. Sep. Purif. Technol. 174, 352–361.
Abdel-Magied, A.F., Abdelhamid, H.N., Ashour, R.M., Zou, X., Forsberg, K., 2019. Cheremisina, O., Volkova, O., Litvinova, T., 2019. Influence of anion nature on acid
Hierarchical porous zeolitic imidazolate frameworks nanoparticles for efficient ad­ leaching of silicate minerals and solvent extraction of rare and rare-earth elements.
sorption of rare-earth elements. Microporous Mesoporous Mater. 278, 175–184. Geochemistry 1–7 (in press).
Altaş, Y., Tel, H., İnan, S., Sert, Ş., Çetinkaya, B., Sengül, S., Özkan, B., 2018. An ex­ Chi, R.A., Zhou, Z., Xu, Z., Hu, Y., Zhu, G., Xu, S., 2003. Solution-chemistry analysis of
perimental design approach for the separation of thorium from rare earth elements. ammonium bicarbonate consumption in rare-earth-element precipitation. Metall.
Hydrometallurgy 178, 97–105. Mater. Trans. B Process Metall. Mater. Process. Sci. 34 (5), 611–617.
Amaral, J.C.B.S., Morais, C.A., 2010. Thorium and uranium extraction from rare earth da Silva, R.G., de Morais, C.A., Teixeira, L.V., de Oliveira, É.D., 2018. Selective removal of
elements in monazite sulphuric acid liquor through solvent extraction. Miner. Eng. impurities from rare earth sulphuric liquor using different reagents. Miner. Eng. 127,
23, 498–503. 238–246.
Amaral, J.C., Sá, M.L., Morais, C.A., 2018. Recovery of uranium, thorium and rare earth Dahlberg, P., Noble, A., Pickarts, J., Rose, W., Jaacks, J., 2014. Bear Lodge Project
from industrial residues. Hydrometallurgy 181, 148–155. Canadian NI 43–101 Pre-Feasibility Study Report on the Reserves and Development
Amer, T.E., Abdella, W.M., Wahab, G.A., El-Sheikh, E.M., 2013. A suggested alternative of the Bull Hill Mine, Wyoming. Rare Element Resources Ltd., Lakewood, CO 516 pp.
procedure for processing of monazite mineral concentrate. Int. J. Miner. Process. 125, Davris, P., Balomenos, E., Panias, D., Paspaliaris, I., 2016. Selective leaching of rare earth
106–111. elements from bauxite residue (red mud), using a functionalized hydrophobic ionic
Amer, T., El-Sheikh, E., Gado, M., Abu-Khoziem, H., Zaki, S., 2018. Selective recovery of liquid. Hydrometallurgy 164, 125–135.
lanthanides, uranium and thorium from Rosetta monazite mineral concentrate. Sep. Davris, P., Stopic, S., Balomenos, E., Panias, D., Paspaliaris, I., Friedrich, B., 2017.
Sci. Technol. 53, 1522–1530. Leaching of rare earth elements from eudialyte concentrate by suppressing silica gel
Anastopoulos, I., Bhatnagar, A., Lima, E.C., 2016. Adsorption of rare earth metals: a re­ formation. Miner. Eng. 108, 115–122.
view of recent literature. J. Mol. Liq. 221, 954–962. Domingues, L.P., Wilfong, R.L., Furlong, L.R., 1962. Pyrolysis of five salts of yttrium,
Ang, K.L., Li, D., Nikoloski, A.N., 2017. The effectiveness of ion exchange resins in se­ lanthanum and cerium. In: U.S. Bureau of Mines Report of Investigations 6029. U.S.
parating uranium and thorium from rare earth elements in acidic aqueous sulfate Dept. of the Interior, Washington 19 pp.
media. Part 1. Anionic and cationic resins. Hydrometallurgy 174, 147–155. Dong, Y., Li, S., Su, X., Wang, Y., Shen, Y., Sun, X., 2017. Separation of thorium from rare
Ang, K.L., Li, D., Nikoloski, A.N., 2018. The effectiveness of ion exchange resins in se­ earths with high-performance diphenyl phosphate extractant. Hydrometallurgy 171,
parating uranium and thorium from rare earth elements in acidic aqueous sulfate 387–393.
media. Part 2. Chelating resins. Miner. Eng. 123, 8–15. Dubey, S.S., Grandhi, S., 2016. Sorption studies of yttrium (III) ions on nano maghemite.
Arunraj, B., Sathvika, T., Rajesh, V., Rajesh, N., 2019. Cellulose and saccharomyces J. Environ. Chem. Eng. 4 (4), 4719–4730.
cerevisiae embark to recover europium from phosphor powder. ACS Omega 4 (1), Dutrizac, J., 2017. The behaviour of the rare earth elements during gypsum (CaSO4·2H2O)
940–952. precipitation. Hydrometallurgy 174, 38–46.
Ashour, R.M., El-Sayed, R., Abdel-Magied, A.F., Abdel-Khalek, A.A., Ali, M., Forsberg, K., Egbaria, F., March 18, 2019. Value of Raw Mineral Materials Produced by U.S. Mines
Uheida, A., Muhammed, M., Dutta, J., 2017. Selective separation of rare earth ions Increases 3% in 2018. MetalMiner, Chicago, IL. www.agmetalminer.com.
from aqueous solution using functionalized magnetite nanoparticles: kinetic and Eisele, J.A., Bauer, D.J., 1974. Recovery and separation of rare-earth elements, barium,
thermodynamic studies. Chem. Eng. J. 327, 286–296. and strontium from bastnäsite with sulfuric acid. In: U.S. Bureau of Mines Report of
Avdibegović, D., Regadío, M., Binnemans, K., 2018. Efficient separation of rare earths Investigations 7990. U.S. Dept. of the Interior, Washington 10 pp.
recovered by a supported ionic liquid from bauxite residue leachate. RSC Adv. 8 (22), El Afifi, E., Borai, E., El-Din, A.S., 2019. New approaches for efficient removal of some
11886–11893. radionuclides and iron from rare earth liquor of monazite processing. Int. J. Environ.
Avdibegović, D., Zhang, W., Xu, J., Regadío, M., Koivula, R., Binnemans, K., 2019. Sci. Technol. 16 (12), 7735–7746.
Selective ion-exchange separation of scandium (III) over iron (III) by crystalline α- El-Nadi, Y.A., 2017. Solvent extraction and its applications on ore processing and re­
zirconium phosphate platelets under acidic conditions. Sep. Purif. Technol. 215, covery of metals: classical approach. Sep. Purif. Rev. 46 (3), 195–215.
81–90. Fidler, J., Schrefl, T., 1996. Overview of Nd–Fe–B magnets and coercivity. J. Appl. Phys.
Babu, C.M., Binnemans, K., Roosen, J., 2018. Ethylenediaminetriacetic acid-functiona­ 79 (8), 5029–5034.
lized activated carbon for the adsorption of rare earths from aqueous solutions. Ind. Firsching, F.H., Mohammadzadei, J., 1986. Solubility products of the rare-earth carbo­
Eng. Chem. Res. 57 (5), 1487–1497. nates. J. Chem. Eng. Data 31, 40–42.
Battsengel, A., Batnasan, A., Narankhuu, A., Haga, K., Watanabe, Y., Shibayama, A., Fitch, F.T., Russell, D.S., 1951. The separation of the rare earths by ion-exchange pro­
2018. Recovery of light and heavy rare earth elements from apatite ore using sul­ cedures. Can. J. Chem. 29 (5), 363–371.
phuric acid leaching, solvent extraction and precipitation. Hydrometallurgy 179, Franco, V., Blázquez, J., Ipus, J., Law, J., Moreno-Ramírez, L., Conde, A., 2018.
100–109. Magnetocaloric effect: from materials research to refrigeration devices. Prog. Mater.
Bonificio, W.D., Clarke, D.R., 2016. Rare-earth separation using bacteria. Environ. Sci. Sci. 93, 112–232.
Technol. Lett. 3 (4), 180–184. Ganguli, R., Cook, D.R., 2018. Rare earths: A review of the landscape. MRS Energy
Borai, E., El-Ghany, M.A., Ahmed, I., Hamed, M.M., El-Din, A.S., Aly, H., 2016. Modified Sustain. 5, 1–16.
acidic leaching for selective separation of thorium, phosphate and rare earth con­ Gergoric, M., Ekberg, C., Foreman Mark, R.S.T.J., Steenari, B., Retegan, T., 2017a.
centrates from Egyptian crude monazite. Int. J. Miner. Process. 149, 34–41. Characterization and leaching of neodymium magnet waste and solvent extraction of
Borai, E., Ahmed, I., El-Din, A.S., El-Ghany, M.A., 2018. Development of selective se­ the rare-earth elements using TODGA. J. Sustain. Metall. 3 (3), 638–645.
paration method for thorium and rare earth elements from monazite liquor. J. Gergoric, M., Ekberg, C., Steenari, B., Retegan, T., 2017b. Separation of heavy rare-earth
Radioanal. Nucl. 316 (2), 443–450. elements from light rare-earth elements via solvent extraction from a neodymium
Borra, C., Vlugt, T., Yang, Y., Offerman, S., 2018. Recovery of cerium from glass polishing magnet leachate and the effects of diluents. J. Sustain. Metall. 3 (3), 601–610.
waste: a critical review. Metals 8 (10), 801. Goode, J.R., 2013. Thorium and rare earth recovery in Canada: the first 30 years. Can.
Buschow, K., 1991. New developments in hard magnetic materials. Rep. Prog. Phys. 54 Metall. Q. 52 (3), 234–242.
(9), 1123–1214. Gschneidner Jr., K.A., 2011. The rare earth crisis–the supply/demand situation for

37
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

2010–2015. Mater. Matters 6 (2), 32–37. Kwon, Y., Kwon, O., Kim, M., Wee, D., 2006. Reaction of Terfenol-D melts with SiO2
Gschneidner Jr., K.A., Daane, A.H., 1988. Physical metallurgy. In: Gschneidner Jr.K.A., crucibles. Mater. Sci. Eng. B 129 (1–3), 18–21.
Eyring, L. (Eds.), Handbook on the Physics and Chemistry of Rare Earths, pp. Landis, V.J., Kaye, J.H., 1971. Radiochemistry of Titanium. National Academy of
409–484. Sciences–National Research Council, Oak Ridge, TN p.12.
Gschneidner Jr., K.A., Pecharsky, V., Tsokol, A., 2005. Recent developments in magne­ Larsson, K., Binnemans, K., 2017. Separation of rare earths by solvent extraction with an
tocaloric materials. Rep. Prog. Phys. 68 (6), 1479–1539. undiluted nitrate ionic liquid. J. Sustain. Metall. 3 (1), 73–78.
Güneş, H., Obuz, H.E., Alkan, M., 2019. Selective precipitation of Th and rare-earth Legaria, E.P., Rocha, J., Tai, C., Kessler, V.G., Seisenbaeva, G.A., 2017. Unusual seeding
elements from HCl leach liquor. In: Azimi, G., Hojong, K., Shafiq, A., Takanari, O., mechanism for enhanced performance in solid-phase magnetic extraction of rare
Neale, R., Neelameggham, A., Abdullahi, B. (Eds.), Rare Metal Technology 2019. earth elements. Sci. Rep. 7, 43740.
Springer, pp. 81–86. Leturcq, G., Costenoble, S., Grandjean, S., 2008. Uranyl oxalate solubility. In: Proceedings
Gupta, C.K., Singh, H., 2003. Uranium Resource Processing: Secondary Resource. from ATALANTE, Montpellier France. P3_16, 4 pp.
Springer-Verlag, Berlin. Li, F., Yang, Z., Weng, H., Chen, G., Lin, M., Zhao, C., 2018. High efficient separation of U
Gupta, C.K., Malik, B.P., Deep, A., 2002. Extraction of uranium, thorium and lanthanides (VI) and Th(IV) from rare earth elements in strong acidic solution by selective
using Cyanex-923: their separations and recovery from monazite. J. Radioanal. Nucl. sorption on phenanthroline diamide functionalized graphene oxide. Chem. Eng. J.
Chem. 251 (3), 451–456. 332, 340–350.
Hals, T., January 8, 2016. Rare Earth Miner Molycorp to Start Bankruptcy Sale of Li, F., Gong, A., Qiu, L., Zhang, W., Li, J., Liu, Z., 2019. Diglycolamide-grafted Fe3O4/
Business. Reuters, Toronto, ON. www.reuters.com. polydopamine nanomaterial as a novel magnetic adsorbent for preconcentration of
Hamza, M.F., El-Aassy, I.E., Guibal, E., 2019. Integrated treatment of tailing material for rare earth elements in water samples prior to inductively coupled plasma optical
the selective recovery of uranium, rare earth elements and heavy metals. Miner. Eng. emission spectrometry determination. Chem. Eng. J. 361, 1098–1109.
133, 138–148. Liang, T., Yan, C., Li, X., Zhou, S., Wang, H., 2018. Polyacrylic acid grafted silica fume as
Han, Q.Y., Liu, Z.Q., Yang, J.H., Li, J.F., 2013. A new technology for aluminum removal an excellent adsorbent for dysprosium (III) removal from industrial wastewater.
from rare earth feed liquid with naphthenic acid. Chin. Rare Earths 3, 017. Water Sci. Technol. 77 (6), 1570–1580.
Hedrick, J.B., Sinha, S.P., 1994. Cerium-based polishing compounds: discovery to man­ Liang, H., Chen, Q., Xu, C., Shen, X., 2019. Selective cloud point extraction of uranium
ufacture. J. Alloys Compd. 207, 377–382. from thorium and lanthanides using Cyanex 301 as extractant. Sep. Purif. Technol.
Hérès, X., Blet, V., Di Natale, P., Ouaattou, A., Mazouz, H., Dhiba, D., Cuer, F., 2018. 210, 835–842.
Selective extraction of rare earth elements from phosphoric acid by ion exchange Liu, Y., Liao, C., Zhang, B., Jia, J., Wang, S., Yan, C., 2004. Method of Separating and
resins. Metals 8 (9), 682. Purifying Cerium by Sulfuric Acid Treatment of Bastnaesite. Chinese Patent
Hidayah, N.N., Abidin, S.Z., 2017. The evolution of mineral processing in extraction of CN200410031179.3.
rare earth elements using solid-liquid extraction over liquid-liquid extraction: a re­ Liu, Y., Chen, J., Li, D., 2012. Application and perspective of ionic liquids on rare earths
view. Miner. Eng. 112, 103–113. green separation. Sep. Sci. Technol. 47 (2), 223–232.
Hidayah, N.N., Abidin, S.Z., 2018. The evolution of mineral processing in extraction of Liu, Z., Liu, Y., Gong, A., 2019. Preparation of diglycolamide polymer modified silica and
rare earth elements using liquid-liquid extraction: a review. Miner. Eng. 121, its application as adsorbent for rare earth ions. Des. Monomers Polym. 22 (1), 1–7.
146–157. Lou, Z., Xiao, X., Huang, M., Wang, Y., Xing, Z., Xiong, Y., 2019. Acrylic acid functio­
Hiskey, J.B., Copp, R.G., 2018a. Recovery of yttrium and neodymium from copper nalized metal-organic frameworks (MOFs) for Sc(III) selective adsorption. ACS Appl.
pregnant leach solutions by solvent extraction. Hydrometallurgy 177, 21–26. Mater. Interfaces 11 (12), 11772–11781.
Hiskey, J.B., Copp, R.G., 2018b. Solvent extraction of yttrium and rare earth elements Lu, Y., Wei, H., Zhang, Z., Li, Y., Wu, G., Liao, W., 2016. Selective extraction and se­
from copper pregnant leach solutions using Primene JM-T. Miner. Eng. 125, paration of thorium from rare earths by a phosphorodiamidate extractant.
265–270. Hydrometallurgy 163, 192–197.
Horrigan, R., 1981. Rare earth polishing compounds. In: Karl, A., Gschneidner, G.A. Lucas, J., Lucas, P., Le Mercier, T., Rollat, A., Davenport, W.G., 2015. Rare Earths:
(Eds.), Industrial Applications of Rare Earth Elements. American Chemical Society, Science, Technology, Production and Use. Elsevier.
Washington, pp. 95–100. Lyman, J.W., Palmer, G.R., 1993. Recycling of rare earths and iron from NdFeB magnet
Hu, Y., Giret, S., Meinusch, R., Han, J., Fontaine, F., Kleitz, F., Larivière, D., 2019. scrap. High Temp. Mater. Processes 11, 175–188.
Selective separation and preconcentration of Th(IV) using organo-functionalized, Lynas Corporation, May 21, 2019. Lynas 2025: Growing with the Market. Investor Day
hierarchically porous silica monoliths. J. Mater. Chem. A 7 (1), 289–302. Presentation. https://www.lynascorp.com/Pages/Reporting-centre-Presentations.
Huang, X., Dong, J., Wang, L., Feng, Z., Xue, Q., Meng, X., 2017. Selective recovery of rare aspx.
earth elements from ion-adsorption rare earth element ores by stepwise extraction Machacek, E., Fold, N., 2014. Alternative value chains for rare earths: the Anglo-deposit
with HEH (EHP) and HDEHP. Green Chem. 19 (5), 1345–1352. developers. Res. Policy 42, 53–64.
Huang, L., Gao, G., Wu, R., Zhang, Q., Lai, F., Xiao, Y., 2019. Non-ammonia enrichment of Marcus, Y., Nelson, F., 1959. Anion-exchange studies. XXV. The rare earths in nitrate
rare earth by magnesium oxide from rare earth leaching liquor in magnesium salt solutions. J. Phys. Chem. 63 (1), 77–79.
system. J. Rare Earths 37 (8), 886–894. Mashimo, M., Sato, H., Ueda, M., Komasawa, I., 1997. Extraction equilibria of aluminum
Innocenzi, V., Ippolito, N.M., Pietrelli, L., Centofanti, M., Piga, L., Vegliò, F., 2018. and beryllium from sulfate media by mixture of bis (2-ethylhexyl) phosphoric acid
Application of solvent extraction operation to recover rare earths from fluorescent and dinonylnaphthalene sulfonic acid. J. Chem. Eng. Jpn 30 (4), 706–711.
lamps. J. Clean. Prod. 172, 2840–2852. Masram, D.T., 2013. Polymer based ion exchange resin. In: Naushad, M., Al-Othman, Z.A.
Islam, F., Rahman, H., Ali, M., 1979. Solvent extraction separation study of Ti(IV) Fe(III) (Eds.), A Book on Ion Exchange, Adsorption and Solvent Extraction. Nova Science,
and Fe(II) from aqueous solutions with di-2-ethyl hexyl phosphoric acid in benzene. Hauppauge, NY 31pp.
J. Inorg. Nucl. Chem. 41 (2), 217–221. Menzies, I.A., Rigby, F., 1961. Separation of thorium from uranium and rare- earth ele­
Jha, M.K., Kumari, A., Panda, R., Kumar, J.R., Yoo, K., Lee, J.Y., 2016. Review on hy­ ments by solvent extraction with tri-n-butylphosphate-xylene. J. Appl. Chem. 11,
drometallurgical recovery of rare earth metals. Hydrometallurgy 165, 2–26. 104–113.
Jiao, Y., Brewer, A., Park, D.M., Chang, E., Lammers, L., Kou, T., Li, Y., 2019. Application Merritt, R.C., 1971. Extractive Metallurgy of Uranium. United States Atomic Energy
of biosorption for rare earth recovery from geothermal brines. In: Proceedings, 44th Commission.
Workshop on Geothermal Reservoir Engineering, Stanford CA, SGP-TR-214. Mishra, R.K., Rout, P.C., Sarangi, K., Nathsarma, K.C., 2010. A comparative study on
Jing, Y., Chen, Y., Wang, X., Nengzi, L., 2018. Adsorption of La3+, Ce3+ from aqueous extraction of Fe(III) from chloride leach liquor using TBP, Cyanex 921 and Cyanex
solutions by bayberry tannin grafted chitosan. IOP Conf. Ser.: Earth Environ. Sci. 153 923. Hydrometallurgy 104 (2), 298–303.
022017. Mondal, S., Ghar, A., Satpati, A., Sinharoy, P., Singh, D., Sharma, J., Sreenivas, T., Kain,
Josso, P., Roberts, S., Teagle, D.A., Pourret, O., Herrington, R., de Leon Albarran, C.P., V., 2019. Recovery of rare earth elements from coal fly ash using TEHDGA im­
2018. Extraction and separation of rare earth elements from hydrothermal metalli­ pregnated resin. Hydrometallurgy 185, 93–101.
ferous sediments. Miner. Eng. 118, 106–121. Monhemius, A., 1977. Precipitation diagrams for metal hydroxides, sulfides, arsenates
Jun, L., Zhenggui, W., Deqian, L., Gengxiang, M., Zucheng, J., 1998. Recovery of Ce(IV) and phosphates. Trans. Inst. Min. Metall. Sect. C 86, 202–206.
and Th(IV) from rare earths(III) with Cyanex 923. Hydrometallurgy 50, 77–87. Mowafy, E., Mohamed, D., 2015. Extraction and separation of Nd(III), Sm(III), Dy(III), Fe
Khawassek, Y.M., Eliwa, A.A., El Sayed, A.H., Omar, S.A., Abdel-Wahab, S.M., 2019. (III), Ni (II), and Cs(I) from concentrated chloride solutions with N,N,N′,N′-tetra(2-
Adsorption of rare earth elements by strong acid cation exchange resin thermo­ ethylhexyl) diglycolamide as new extractant. J. Rare Earths 33 (4), 432–438.
dynamics, characteristics and kinetics. SN Appl. Sci. 1 (1) 51. Naylor, A., Eccles, H., 1988. Tri-N-Butyl Phosphate-the Universal Solvent for the Nuclear
Kim, D., Powell, L.E., Delmau, L.H., Peterson, E.S., Herchenroeder, J., Bhave, R.R., 2015. Fuel Cycle. International Solvent Extraction Conference, Moscow, USSR, pp. 31–36.
Selective extraction of rare earth elements from permanent magnet scraps with Ogata, T., Narita, H., Tanaka, M., 2015a. Adsorption behavior of rare earth elements on
membrane solvent extraction. Environ. Sci. Technol. 49 (16), 9452–9459. silica gel modified with diglycol amic acid. Hydrometallurgy 152, 178–182.
Kobayashi, T., Sasaki, T., Takagi, I., Moriyama, H., 2009a. Solubility of thorium (IV) in Ogata, T., Narita, H., Tanaka, M., 2015b. Rapid and selective recovery of heavy rare
the presence of oxalic and malonic acids. J. Nucl. Sci. Technol. 46, 1085–1090. earths by using an adsorbent with diglycol amic acid group. Hydrometallurgy 155,
Kobayashi, T., Sasaki, T., Takagi, I., Moriyama, H., 2009b. Zirconium solubility in ternary 105–109.
aqueous system of Zr(IV)-OH-carboxylates. J. Nucl. Sci. Technol. 46, 142–148. Ogata, T., Narita, H., Tanaka, M., 2016a. Adsorption mechanism of rare earth elements by
Kolar, E., Catthoor, R.P., Kriel, F.H., Sedev, R., Middlemas, S., Klier, E., Hatch, G., Priest, adsorbents with diglycolamic acid ligands. Hydrometallurgy 163, 156–160.
C., 2016. Microfluidic solvent extraction of rare earth elements from a mixed oxide Ogata, T., Narita, H., Tanaka, M., Hoshino, M., Kon, Y., Watanabe, Y., 2016b. Selective
concentrate leach solution using Cyanex® 572. Chem. Eng. Sci. 148, 212–218. recovery of heavy rare earth elements from apatite with an adsorbent bearing im­
Krishnamurthy, N., Gupta, C.K., 2016. Extractive Metallurgy of Rare Earths. CRC Press, mobilized tridentate amido ligands. Sep. Purif. Technol. 159, 157–160.
Boca Raton. Önal, M.A.R., Borra, C.R., Guo, M., Blanpain, B., Gerven, T.V., 2017. Hydrometallurgical
Kul, M., Topkaya, Y., Karakaya, İ., 2008. Rare earth double sulfates from pre-con­ recycling of NdFeB magnets: complete leaching, iron removal and electrolysis. J. Rare
centrated bastnasite. Hydrometallurgy 93, 129–135. Earths 35 (6), 574–584.

38
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Park, D.M., Reed, D.W., Yung, M.C., Eslamimanesh, A., Lencka, M.M., Anderko, A., Fujita, Smirnov, A.L., Titova, S.M., Rychkov, V.N., Bunkov, G.M., Semenishchev, V.S., Kirillov,
Y., Riman, R.E., Navrotsky, A., Jiao, Y., 2016. Bioadsorption of rare earth elements E.V., Poponin, N.N., Svirsky, I.A., 2017. Study of scandium and thorium sorption
through cell surface display of lanthanide binding tags. Environ. Sci. Technol. 50 (5), from uranium leach liquors. J. Radioanal. Nucl. 312, 277–283.
2735–2742. Smith, Y.R., Bhattacharyya, D., Willhard, T., Misra, M., 2016. Adsorption of aqueous rare
Park, D.M., Brewer, A., Reed, D.W., Lammers, L.N., Jiao, Y., 2017. Recovery of rare earth earth elements using carbon black derived from recycled tires. Chem. Eng. J. 296,
elements from low-grade feedstock leachates using engineered bacteria. Environ. Sci. 102–111.
Technol. 51 (22), 13471–13480. Sokolov, 1960. Solubility of scandium oxalate. Tr. Khim. Khim. Tekhnol. 3, 594–597.
Peng, X., Li, Y., Ma, J., Cui, Y., Sun, G., 2017. Extraction and separation of uranium(VI) Sole, K.C., Mooiman, M.B., Hardwick, E., 2018. Ion exchange in hydrometallurgical
and Fe(III) from lanthanide chlorides with N,N,N′,N′-tetrasubstituted alkyl diglyco­ processing: an overview and selected applications. Sep. Purif. Rev. 47 (2), 159–178.
lamide extractants. J. Radioanal. Nucl. 313 (2), 327–332. Sprecher, B., Daigo, I., Murakami, S., Kleijn, R., Vos, M., Kramer, G.J., 2015. Framework
Pilkington, E., Wylie, A., 1947. Production of rare earth and thorium compounds from for resilience in material supply chains, with a case study from the 2010 rare earth
monazite. Part I. J. Soc. Chem. Ind. London 66 (11), 387–394. crisis. Environ. Sci. Technol. 49, 6740–6750.
Powell, J.E., 1979. Separation chemistry. In: Gschneidner Jr.K.A., Eyring, L. (Eds.), Tay, P.K.R., Manjula-Basavanna, A., Joshi, N.S., 2018. Repurposing bacterial extracellular
Handbook on the Chemistry and Physics of Rare Earths. Vol. 3. pp. 81–110. matrix for selective and differential abstraction of rare earth elements. Green Chem.
Puigdomenech, I., 2006. HYDRA (Hydrochemical Equilibrium-Constant Database) and 20, 3512–3520.
MEDUSA (Make Equilibrium Diagrams Using Sophisticated Algorithms) Programs. Tiefenbach, W., Hughes, M.G., Rodgerson, M.C., Chen, J., Theriault, M.B., Cuthill, C.,
Royal Institute of Technology, Stockholm. http://www.kemi.kth.se/medusa/. Stacy, R., 2005. Canadian Guidelines for the Management of Naturally Occurring
Qi, X., Du, K., Feng, M., Gao, Y., Huang, X., Kanatzidis, M.G., 2017. Layered A2Sn3S7·1.25 Radioactive Materials (NORM). Ottawa, ON, Health Canada 61pp.
H2O (A= organic cation) as efficient ion-exchanger for rare earth element recovery. Topf, A., June 16, 2017. Mountain Pass Sells for $20.5 Million. MINING.COM, Vancouver,
J. Am. Chem. Soc. 139 (12), 4314–4317. BC. www.mining.com.
Qiu, T.S., Wu, H.Q., Fang, X.H., Li, X.B., 2012. The current situation and trends of the Tse, P., 2011. China's rare-earth industry. In: U.S. Geological Survey, Open-File Report
technology for impurity removal of weathering crust ion-absorbed type rare earth 2011-1042. U.S. Department of the Interior, Reston, VA 11pp.
ores. Xitu (Chin. Rare Earths) 33 (4), 81–85. Tunsu, C., Petranikova, M., Ekberg, C., Retegan, T., 2016. A hydrometallurgical process
Qiu, S., Li, S., Dong, Y., Su, X., Wang, Y., Shen, Y., Sun, X., 2017. A high-performance for the recovery of rare earth elements from fluorescent lamp waste fractions. Sep.
impregnated resin for recovering thorium from radioactive rare earth waste residue. Purif. Technol. 161, 172–186.
J. Mol. Liq. 237, 380–386. U.S. Geological Survey, 2011-2020. Minerals Yearbooks. U.S. Department of the Interior,
Ramasamy, D.L., Wojtuś, A., Repo, E., Kalliola, S., Srivastava, V., Sillanpää, M., 2017. Reston, VA.
Ligand immobilized novel hybrid adsorbents for rare earth elements (REE) removal Van Mens, R., 1986. Ternary phase studies of Nd-Fe-X where X= C, Si, Pb, Sn. J. Magn.
from waste water: assessing the feasibility of using APTES functionalized silica in the Magn. Mater. 61 (1–2), 24–28.
hybridization process with chitosan. Chem. Eng. J. 330, 1370–1379. Van Nguyen, N., Iizuka, A., Shibata, E., Nakamura, T., 2016. Study of adsorption behavior
Ramasamy, D.L., Puhakka, V., Doshi, B., Iftekhar, S., Sillanpää, M., 2019. Fabrication of of a new synthesized resin containing glycol amic acid group for separation of
carbon nanotubes reinforced silica composites with improved rare earth elements scandium from aqueous solutions. Hydrometallurgy 165, 51–56.
adsorption performance. Chem. Eng. J. 365, 291–304. Vander Hoogerstraete, T., Wellens, S., Verachtert, K., Binnemans, K., 2013. Removal of
Rey, J., Atak, S., Dourdain, S., Arrachart, G., Berthon, L., Pellet-Rostaing, S., 2017. transition metals from rare earths by solvent extraction with an undiluted phos­
Synergistic extraction of rare earth elements from phosphoric acid medium using a phonium ionic liquid: separations relevant to rare-earth magnet recycling. Green
mixture of surfactant AOT and DEHCNPB. Solvent Extr. Ion Exch. 35 (5), 321–331. Chem. 15 (4), 919–927.
Reynier, N., Whitty-Léveillé, L., Laviolette, C., Courchesne, M., Fiset, J., Zinck, J., 2018. Verbaan, N., Bradley, K., Brown, J., Mackie, S., 2015. A review of hydrometallurgical
Radionuclide removal from ore and REE-bearing mineral by leaching and ion ex­ flowsheets considered in current REE projects. In: Symposium on Critical and
change separation. In: Davis, B.R., Moats, M.S., Wang, S., Gregurek, D., Kapusta, J., Strategic Materials 3. British Columbia Geological Survey Paperpp. 147–162.
Battle, T.P. ... Bhambhani, T. (Eds.), Extraction 2018. Springer, pp. 2807–2827. Walawalkar, M., Nichol, C.K., Azimi, G., 2016. An innovative process for the recovery of
Riker, L.W., 1981. The use of rare earths in glass compositions. In: Gschneidner, Karl A., consumed acid in rare-earth elements leaching from phosphogypsum. Ind. Eng.
G.A. (Eds.), Industrial Applications of Rare Earth Elements. American Chemical Chem. Res. 55 (48), 12309–12316.
Society, Washington, pp. 81–94. Wang, L., Long, Z., Huang, X., Yu, Y., Cui, D., Zhang, G., 2010. Recovery of rare earths
Ritcey, G., 1980. Crud in solvent extraction processing—a review of causes and treatment. from wet-process phosphoric acid. Hydrometallurgy 101, 41–47.
Hydrometallurgy 5, 97–107. Wang, Q., Wilfong, W.C., Kail, B.W., Yu, Y., Gray, M.L., 2017a. Novel poly­
Ritcey, G.M., 2004. Development of industrial solvent extraction processes. In: Solvent ethylenimine–acrylamide/SiO2 hybrid hydrogel sorbent for rare-earth-element re­
Extraction Principles and Practice, Revised and Expanded. CRC Press, pp. 293–349. cycling from aqueous sources. ACS Sustain. Chem. Eng. 5 (11), 10947–10958.
Ritcey, G.M., Lucas, B.H., 1969. Stagewise separation of uranium, thorium and the rare Wang, Y., Huang, C., Li, F., Dong, Y., Sun, X., 2017b. Process for the separation of thorium
earths by liquid-liquid extraction. In: AIME Annual Meeting. Mines Branch, Canadian and rare earth elements from radioactive waste residues using Cyanex® 572 as a new
Dept. of Energy, Mines, and Resources Technical Bulletin TB 113, 29 pp. extractant. Hydrometallurgy 169, 158–164.
Rizkalla, E.N., Choppin, G.R., 1991. Hydration and hydrolysis of lanthanides. In: Wang, K., Adidharma, H., Radosz, M., Wan, P., Xu, X., Russell, C.K., Tian, H., Fan, M., Yu,
Handbook on the Physics and Chemistry of Rare Earths. Vol. 15. pp. 393–442. J., 2017c. Recovery of rare earth elements with ionic liquids. Green Chem. 19 (19),
Rodrigues, D.G., Monge, S., Pellet-Rostaing, S., Dacheux, N., Bouyer, D., Faur, C., 2019. 4469–4493.
Sorption properties of carbamoylmethylphosphonated-based polymer combining Wang, Y., Lu, H., Dai, X., Duan, T., Bai, X., Cai, Y., Yin, X., Chen, L., Diwu, J., Du, S.,
both sorption and thermosensitive properties: new valuable hydrosoluble materials 2018a. Facile and efficient decontamination of thorium from rare earths based on
for rare earth elements sorption. Chem. Eng. J. 355, 871–880. selective selenite crystallization. Inorg. Chem. 57, 1880–1887.
Roosen, J., Van Roosendael, S., Borra, C.R., Van Gerven, T., Mullens, S., Binnemans, K., Wang, Z., Brown, A.T., Tan, K., Chabal, Y.J., Balkus Jr., K.J., 2018b. Selective extraction
2016. Recovery of scandium from leachates of Greek bauxite residue by adsorption of thorium from rare earth elements using wrinkled mesoporous carbon. J. Am.
on functionalized chitosan–silica hybrid materials. Green Chem. 18, 2005–2013. Chem. Soc. 140, 14735–14739.
Rumble, J.R. (Ed.), 2019. CRC Handbook of Chemistry and Physics. Wang, J., Liu, X., Fu, J., Xie, M., Huang, G., Wang, H., 2019. Novel extractant (2,4-
Ryan, J.L., 1960. Species involved in the anion-exchange absorption of quadrivalent ac­ dimethylheptyl)(2,4,4′-trimethylpentyl) phosphinic acid (USTB-2) for rare earths
tinide nitrates. J. Phys. Chem. 64 (10), 1375–1385. extraction and separation from chloride media. Sep. Purif. Technol. 209, 789–799.
Rychkov, V.N., Kirillov, E.V., Kirillov, S.V., Bunkov, G.M., Mashkovtsev, M.A., Botalov, Westbrook, T., Lee, L., Burton, M., Pullin, R., September 24, 2018. Rare earth Miner Lynas
M.S., Semenishchev, V.S., Volkovich, V.A., 2016. Selective ion exchange recovery of to Face Review in Malaysia. Media. Reuters, Toronto, ON. www.reuters.com.
rare earth elements from uranium mining solutions. AIP Conf. Proc. 1767, 020017. Whitty-Léveillé, L., Reynier, N., Larivière, D., 2018. Rapid and selective leaching of ac­
Rychkov, V.N., Kirillov, E.V., Kirillov, S.V., Semenishchev, V.S., Bunkov, G.M., Botalov, tinides and rare earth elements from rare earth-bearing minerals and ores.
M.S., Smyshlyaev, D.V., Malyshev, A.S., 2018. Recovery of rare earth elements from Hydrometallurgy 177, 187–196.
phosphogypsum. J. Clean. Prod. 196, 674–681. Wilfong, W.C., Kail, B.W., Bank, T.L., Howard, B.H., Gray, M.L., 2017. Recovering rare
Sadri, F., Kim, R., Yang, Z., Ghahreman, A., 2018. The effect of calcium sulfate crystal­ earth elements from aqueous solution with porous amine–epoxy networks. ACS Appl.
lization and the crystal modification on aqueous REE stability in Ca saturated REE- Mater. Interfaces 9 (21), 18283–18294.
Ca-SO4-H2O systems. Hydrometallurgy 182, 82–96. Wormsbecher, R.F., Peters, A.W., Maselli, J.M., 1986. Vanadium poisoning of cracking
Salazar, D., Martín-Cid, A., Madugundo, R., Garitaonandia, J., Barandiaran, J., catalysts: mechanism of poisoning and design of vanadium tolerant catalyst system. J.
Hadjipanayis, G., 2016. Effect of Nb and Cu on the crystallization behavior of under- Catal. 100 (1), 130–137.
stoichiometric Nd–Fe–B alloys. J. Phys. D 50, 015305. Xia, C., Griffith, W., 2018. Direct Oxalate Precipitation for Rare Earth Elements Recovery.
Shaaban, A., 2007. Effects of Zr substitution in NdFeB permanent magnets. AIP Conf. Patent WO2018195642A1.
Proc. 909, 63–67. Xie, F., Zhang, T.A., Dreisinger, D., Doyle, F., 2014. A critical review on solvent extraction
Shimidt, V.S., 1971. Amine Extraction. Keter Press, Jerusalem. of rare earths from aqueous solutions. Miner. Eng. 56, 10–28.
Shinozaki, T., Ogata, T., Kakinuma, R., Narita, H., Tokoro, C., Tanaka, M., 2018. Xiong, Y., Kuang, W., Zhao, J., Liu, H., 2017. Ionic liquid-based synergistic extraction of
Preparation of polymeric adsorbents bearing diglycolamic acid ligands for rare earth rare earths nitrates without diluent: typical ion-association mechanism. Sep. Purif.
elements. Ind. Eng. Chem. Res. 57, 11424–11430. Technol. 179, 349–356.
Silva, R.G., Morais, C.A., Oliveira, É.D., 2019. Selective precipitation of rare earth from Xiong, Y., Gao, Y., Guo, X., Wang, Y., Su, X., Sun, X., 2019. Water-stable metal–organic
non-purified and purified sulfate liquors using sodium sulfate and disodium hydrogen framework material with uncoordinated terpyridine site for selective Th(IV)/ln(III)
phosphate. Miner. Eng. 134, 402–416. separation. ACS Sustain. Chem. Eng. 7 (3), 3120–3126.
Silva, R.G., Morais, C.A., Teixeira, L.V., Oliveira, É.D., 2019b. Selective precipitation of Yang, Z., Chen, G., Weng, H., Shen, W., Huang, Z., Lin, M., 2018. Efficient and selective
high-quality rare earth oxalates or carbonates from a purified sulfuric liquor con­ separation of U(VI) and Th(IV) from rare earths using functionalized hierarchically
taining soluble impurities. Min. Metall. Explor. 36, 967–977. mesoporous silica. J. Mater. Sci. 53, 3398–3416.

39
W.D. Judge and G. Azimi Hydrometallurgy 196 (2020) 105435

Yao, Y., Farac, N.F., Azimi, G., 2017. Supercritical fluid extraction of rare earth elements highly efficient recovery of rare earth elements from water and a mechanism study.
from nickel metal hydride battery. ACS Sustain. Chem. Eng. 6, 1417–1426. Green Chem. 19, 4816–4828.
Ye, Q., Li, G., Deng, B., Luo, J., Rao, M., Peng, Z., Zhang, Y., Jiang, T., 2019. Solvent Zhao, Z., Qiu, Z., Yang, J., Lu, S., Cao, L., Zhang, W., Xu, Y., 2017b. Recovery of rare earth
extraction behavior of metal ions and selective separation Sc3+ in phosphoric acid elements from spent fluid catalytic cracking catalysts using leaching and solvent
medium using P204. Sep. Purif. Technol. 209, 175–181. extraction techniques. Hydrometallurgy 167, 183–188.
Yin, X., Wang, Y., Bai, X., Wang, Y., Chen, L., Xiao, C., Diwu, J., Du, S., Chai, Z., Albrecht- Zheng, X., Zhang, F., Liu, E., Xu, X., Yan, Y., 2016. Efficient recovery of neodymium in
Schmitt, T.E., 2017. Rare earth separations by selective borate crystallization. Nat. acidic system by free-standing dual-template docking oriented ionic imprinted me­
Commun. 8, 14438. soporous films. ACS Appl. Mater. Interfaces 9, 730–739.
Zakharova, F.A., Moskvin, A.I., 1960. Solubility product of uranium(IV) oxalate, com­ Zhou, H., Dong, Y., Wang, Y., Zhao, Z., Xiao, Y., Sun, X., 2019. Recovery of Th(IV) from
position and dissociation constants of U(IV) oxalate complexes in aqueous solutions. leaching solutions of rare earth residues using a synergistic solvent extraction system
Russ. J. Inorg. Chem. 5, 592–595. consisting of Cyanex 572 and n-octyl diphenyl phosphate (ODP). Hydrometallurgy
Zhang, Z., Jia, Q., Liao, W., 2015. Progress in the separation processes for rare earth 183, 186–192.
resources. In: Handbook on the Physics and Chemistry of Rare Earths. Vol. 48. Zhu, Z., Cheng, C.Y., 2011. A review of uranium solvent extraction: its present status and
Elsevier, pp. 287–376. future trends. In: ALTA 2011 Uranium Conference, pp. 1–22.
Zhang, J., Zhao, B., Schreiner, B., 2016. Separation Hydrometallurgy of Rare Earth Zhu, Z., Pranolo, Y., Cheng, C.Y., 2015. Separation of uranium and thorium from rare
Elements. Springer. earths for rare earth production – a review. Miner. Eng. 77, 185–196.
Zhang, J., Anawati, J., Yao, Y., Azimi, G., 2018. Aeriometallurgical extraction of rare Zillman, C., June 25, 2015. Molycorp is Filing for Chapter 11 Bankruptcy. Fortune
earth elements from a NdFeB magnet utilizing supercritical fluids. ACS Sustain. Magazine, New York, NY. www.fortune.com.
Chem. Eng. 6, 16713–16725. Zuo, X., Liu, Q., Deng, X., Liu, K., Chi, Y., 2014. Preparation Method of Crystalline Rare-
Zhao, F., Repo, E., Song, Y., Yin, D., Hammouda, S.B., Chen, L., Kalliola, S., Tang, J., Tam, Earth Carbonate. Patent CN104195332A.
K.C., Sillanpää, M., 2017a. Polyethylenimine-cross-linked cellulose nanocrystals for

40