Solvent-free ageing reactions of rare earth element oxides: From geomimetic synthesis of new metal-organic materials towards a simple, environmentally friendly separation of scandium

Article abstract of DOI:10.1039/d0gc00454e

The development of cleaner methodologies for the separation and processing of rare earth elements, including scandium and yttrium, is of high importance to materials science and industry. Here, we explore the use of simple, solvent-free and low-energy accelerated ageing reactions, inspired by geochemical processes of mineral weathering and neogenesis, to convert eight rare earth (RE) element oxides, specifically Sc, Y, La, Yb, Eu, Gd, Lu and Nd, into metal-organic oxalate derivatives. The reactions can readily be directed towards the formation of topologically distinct framework structures by introduction of additional oxalate ions in the form of simple alkaline metal or ammonium salts, providing not only simple, solvent-free access to novel metal-organic RE element architectures from readily accessible precursors, but also enabling a proof-of-principle of a simple route to separate scandium from at least seven other rare earth elements in a mixture of metal oxides. Specifically, the metal-organic structures obtained by ageing a mixture of scandium and other RE element oxides in presence of ammonium or sodium oxalates exhibit significant differences in aqueous solubility at boiling point of water, providing high scandium separation factors ranging from approximately 102-103, without requiring elaborate complexation reagents, toxic extractants, acidic media, or solvents other than water. This journal is

Full text of DOI:10.1039/d0gc00454e

View Article Online
View Journal
Green
Chemistry
Cutting-edge research for a greener sustainable future
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: I. Huskic, M.
Arhangelskis and T. Friši, Green Chem., 2020, DOI: 10.1039/D0GC00454E.
Volume 20
This is an Accepted Manuscript, which has been through the
Number
May 2018
Pages 1919-2160
9
7
Royal Society of Chemistry peer review process and has been
accepted for publication.
Green
Chemistry
Cutting-edge research for a greener sustainable future
rsc.li/greenchem
Accepted Manuscripts are published online shortly after acceptance,
before technical editing, formatting and proof reading. Using this free
service, authors can make their results available to the community, in
citable form, before we publish the edited article. We will replace this
Accepted Manuscript with the edited and formatted Advance Article as
soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes to the
text and/or graphics, which may alter content. The journal’s standard
Terms & Conditions and the Ethical guidelines still apply. In no event
shall the Royal Society of Chemistry be held responsible for any errors
or omissions in this Accepted Manuscript or any consequences arising
from the use of any information it contains.
ISSN 1463-9262
PAPER
Paul T. Anastas et al.
The Green ChemisTREE: 20 years after taking root with the 12
principles
rsc.li/greenchem

Please do not adjust margins
Page 1 of 14
Green Chemistry
View Article Online
DOI: 10.1039/D0GC00454E
ARTICLE
Solvent-free ageing reactions of rare earth element oxides: From
geomimetic synthesis of new metal-organic materials towards a
simple, environmentally friendly separation of scandium
Received 00th January 20xx,
Accepted 00th January 20xx
Igor Huskić,a Mihails Arhangelskis,a,b Tomislav Friščić*,a
DOI: 10.1039/x0xx00000x
The development of cleaner methodologies for the separation and processing of rare earth elements, including scandium
and yttrium, is of high importance to materials science and industry. Here, we explore the use of simple, solvent-free and
low-energy "accelerated ageing" reactions, inspired by geochemical processes of mineral weathering and neogenesis, to
convert eight rare earth (RE) element oxides, specifically Sc, Y, La, Yb, Eu, Gd, Lu and Nd, into metal-organic oxalate
derivatives. The reactions can readily be directed towards the formation of topologically distinct framework structures by
introduction of additional oxalate ions in the form of simple alkaline metal or ammonium salts, providing not only simple,
solvent-free access to novel metal-organic RE element architectures from readily accessible precursors, but also enabling a
proof-of-principle of a simple route to separate scandium from at least seven other rare earth elements in a mixture of
metal oxides. Specifically, the metal-organic structures obtained by ageing a mixture of scandium and other RE element
oxides in presence of ammonium or sodium oxalates exhibit significant differences in aqueous solubility at boiling point of
water, providing high scandium separation factors ranging from approximately 102-103, without requiring elaborate
complexation
reagents,
toxic
extractants,
acidic
media,
or
solvents
other
than
water.
industry. Its use in metal-organic frameworks (MOFs) has been
recently explored.21 Unlike other transition elements and REs, the
chemistry of scandium remains poorly explored and understood,
most likely due to its high price: on the global market the price of
Sc2O3 is above $4,000 per pound, with the price of pure metal being
even higher.22 Consequently, further investigation and use of this
poorly explored element requires the development of simple,
straightforward routes for its separation.
Introduction
Rare earth (RE) metals, including Sc, Y and the lanthanides, are
critical elements of high technological and strategic importance.1,2
The REs are central to the manufacture of high performance tools,
alloys, electronic components, catalysts and medicinal diagnostic
agents.3,4 Their isolation and production is a persistent challenge
because of their concomitancy, chemical similarity, and scarce
distribution of economically viable feedstocks.5–9 Common
processes for RE extraction and separation involve volatile organic
liquids, synthetically demanding custom ligands, and concentrated
mineral acids.10–14 Moreover, as REs are often a secondary product
in extraction of other elements (e.g. U, Al), their supply can strongly
depend on primary product market. This, combined with political
and economic instability in leading producer countries has led to an
increased interest in recycling of RE-containing components.15–17
Scandium is the lightest RE element and the lightest transition
metal. Its notable uses include acid catalysis, in the form of the
potent Lewis acid scandium triflate18,19 and the production of
durable, lightweight alloys.20 It is also a critical, irreplaceable
component of high-quality lighting elements in film and television
a. McGill University, Department of Chemistry, Montreal, Canada. E-mail:
tomislav.friscic@mcgill.ca
b. University of Warsaw, Faculty of Chemistry, Warsaw, Poland.
Electronic Supplementary Information (ESI) available: details of experimental
procedures, Rietveld refinement, additional PXRD, ICP analysis, thermal (TGA/DSC)
analysis, and infrared spectroscopy data. Crystallographic data for compounds 1-8
in CIF format (CCDC deposition numbers 1963550-1963557) See
DOI: 10.1039/x0xx00000x
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 2 of 14
ARTICLE
Journal Name
Figure 1. a) Flow schematic for a general process for isolation of RE elements from
phosphate minerals, adapted from reference 12. The ‘*’ sign denotes potential places
where the herein demonstrated routes for solvent-free transformation and separation
might be deployable. b) A flow schematic for the herein demonstrated solvent-free
transformation of RE oxides and separation of scandium.
Molycorp bastnesite, Rhone-Poulenc liquid-liquidVieweAxrtticrlaecOtniolinne,
Mintek-apatite processes),25,26,68,69 preseDntOinI:g10.1p0o3t9e/nDt0iaGlC0p0l4a5c4eEs
where the proposed approach could be implemented for early
scandium separation (Figure 1a) and simplification of subsequent
metal separation procedures by SX or IX cycles.
Currently used methods for the extraction of RE elements from
their ores are involved (Figure 1a) and, while details and steps are
often not fully disclosed by individual industries,23,24 many of them
employ oxalic acid (H2ox) and oxalates as precipitating agents once
the raw feedstocks have been digested by mineral acids (i.e. HCl,
HNO3, H2SO4). The precipitates are again converted to soluble forms
by calcination and/or acid treatment, and separated by solvent
extraction (SX) or multiple solvent-intense ion exchange (IX)
cycles.25–30
Experimental section
Materials used in the experiments were used as obtained,
without purification, except for Gd2O3, Nd2O3, La2O3 and Eu2O3
which were calcined at 800 ºC for 24 hours to ensure the
conversion of any potential hydroxides to the oxides. Oxalic
acid dihydrate (H2ox), potassium oxalate monohydrate (K2ox),
sodium oxalate (Na2ox) and ammonium oxalate dihydrate
((NH4)2ox) were obtained from Sigma Aldrich.
Brief milling of reaction mixtures, to ensure
homogenization prior to ageing, was performed on a Retsch
MM400 shaker mill for 30 seconds operating at 30 Hz, in a 15
mL volume Teflon® jar using a single 10 mm diameter zirconia
ball (weight ~3.6 g). Total mass of the components in the
milling jar was always ~200 mg.
Accelerated ageing was performed by placing the
homogenized mixture in a 20 mL glass vial which was then
placed in a sealed chamber at constant temperature (45 ºC)
and controlled relative humidity (RH) (see also Figure S1, ESI).
Relative humidity was controlled via a saturated solution of
potassium sulfate70 ensuring that the humidity never exceeded
98%, and was kept within 96%±0.5% at 45 ºC. The total mass of
the aged reaction mixtures was typically around 200 mg.
Powder X-ray Diffraction (PXRD) data was collected on a
PROTOXRD AXRD benchtop instrument equipped with a
Dectris Mythen2R 1D detector, using nickel-filtered CuKα (λ =
1.5418 Å) radiation, or the BRUKER D2 benchtop instrument
equipped with the Lynxeye 1D detector, using nickel-filtered
CuKα radiation. Conversion of file formats was performed in
the PowDLL program suite.71
Crystal Structures of Na3Scox3·6H2O (4), K3Scox3·4H2O (5),
K8Sc2ox7·10H2O (6) and K3Gdox3·3H2O (7) were determined by
single crystal X-ray diffraction, using a Bruker D8 ADVANCE X-
ray diffractometer equipped with a PHOTON100 detector and
graphite-monochromated MoKα (λ = 0.71073 Å) radiation. The
X-ray diffraction data sets were collected using the ω- and φ-
scan mode over a 2-range up to 54º. The structures were
solved by intrinsic phasing implemented in SHELXT72 and
refined using SHELXL.73 Structure refinement was performed
on F2 using all data. Hydrogen atoms participating in hydrogen
bonds were located from the electron density map whenever
data quality was sufficient. Otherwise, they were fixed to
appropriate distances and refined from the electron density
map.
Recently, solvent-free techniques such as mechanochemistry31–
44
or low-energy activated or accelerated ageing methods45–51 have
emerged as alternatives to solution chemistry, and our team has
proposed their potential use in metal production and recycling.
Such methods are simple and offer mild, environmentally non-
taxing means to transform and separate noble and critical
elements52,53 (Au, Pd, Pt, Ge) and base metals54 (Cu, Pb, Zn) in the
raw element or oxide form. Accelerated ageing, inspired by
geochemical mineral weathering (mineral neogenesis),55–58 provides
a minimalistic and more environmentally friendly route to the
synthesis of complex metal-organic materials from simple mineral-
like oxides or carbonates51,54 without requiring bulk solvents,
soluble and often corrosive metal nitrates, halides or sulfates, high-
temperature or mechanical agitation. Instead, accelerated ageing
chemistry relies on diffusion-controlled reactivity in static mixtures
of inorganic precursors with a suitable organic ligand.45,46,59
Here, we provide a systematic exploration of solid-state ageing
reactivity of eight RE oxides (Sc, Y, La, Yb, Eu, Gd, Lu, Nd) towards
oxalic acid (H2ox), an abundant naturally-occurring compound
already extensively used as a precipitating agent in extractive
metallurgy of these elements. Our results show facile, solvent-free
conversion of these RE oxides into corresponding oxalates through
accelerated ageing transformations, while the addition of
stoichiometric amounts of sodium, potassium or ammonium
oxalate as sources of additional oxalate ions also enabled the
production of previously not reported discrete or framework metal-
organic structures. On one hand, the results of this study provide a
potential model for the geological formation of RE-based organic
minerals in the absence of solvents, which is emerging as an area of
interest for understanding the mineral evolution and the carbon
cycle of Earth and other planetary systems.60–64 On the other, we
find that the metal-organic architectures obtained through ageing
reactions of RE oxides can exhibit significantly different dissolution
in boiling water, enabling a proof-of-concept for simple, mild
separation of scandium from a complex solid mixture with at least
seven other RE oxides. The separation factors from this simple
process (Figure 1b) range from ca. 90 to 2,200, comparable to
values encountered in conventional SX but without using
specialized complexation agents, organic solvents or toxic
extractants such as naphthenic acids or organophosphine oxides
(e.g. Cyanex or Cextrant range of extractants).65–67 Mixtures of RE
element oxalates and/or oxides often appear in popular
methodologies for RE element separation (e.g. the Baotou,
The crystal structures of Lu2ox3·6H2O (1), (NH4)Scox2·2H2O
(2), (NH4)Ybox2·H2O (3) and K3Euox3·3H2O (8) were determined
from PXRD data collected on a Bruker D8 Venture instrument
equipped with a CuKα non-monochromated X-ray source. The
measurements were performed using
geometry, with the sample material placed on a spinning plate.
The unit cell determination was achieved using DICVOL0674
a Bragg-Brentano
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 3 of 14
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
algorithm, accessed via graphical interface of the program days of ageing at 90 ºC and 96% relative humidiVtiye,waArt2ic0le0Onmlinge
DASH.75 The unit cell parameters were then refined using the sample of the mixture was suspended DinOI:d1i0s.t1i0ll3e9d/Dw0GaCte0r04a5n4dE
Pawley procedure76 and the extracted diffraction intensities briefly heated to boil. The liquid, mostly containing dissolved
were used for space group determination. Simulated annealing scandium species, was filtered and the solid residue was
(SA) structure solution and Rietveld refinement were then digested with 4% analytical grade HNO3. All samples were
performed in the program TOPAS Academic v.6.77 Diffraction diluted to concentrations between 50 ppb and 1 ppm,
peak shapes were modelled with a pseudo-Voigt function, appropriate for ICP-MS analysis.
while the background was described with 8th order Chebyshev
polynomial function. During the SA structure solution peak
shape, background terms and the unit cell parameters were
Results and Discussion
kept fixed, while the metal and water oxygen atoms were
allowed free movement. In addition, the geometry of the ox2-
Overview
To ensure homogeneity, solid-state reaction mixtures were
prepared by brief ball milling (30 seconds) of binary mixtures
of a RE oxide with H2ox, or of ternary mixtures containing an
oxide, H2ox and Na2ox, (NH4)2ox or K2ox. While no change was
evident in PXRD patterns of mixture immediately after milling,
PXRD analysis of these binary mixtures after several days
exposure to 45 ºC and 96% relative humidity (RH, achieved by
storing the samples in a sealed chamber in the presence of a
saturated solution of K2SO4) revealed the formation of metal
oxalates in all cases (Table 1, Figure 2a). Such behavior is
analogous to previously reported for ageing reactions of
transition metal oxides.26
fragments was defined using the rigid body approach, with
their positions and orientations freely refined during SA
calculation. The structural model with the lowest Rwp value
was selected for Rietveld refinement.78 At this point all the
previously fixed parameters (background, peak shape and zero
shift) were freely refined. In addition, atomic thermal motion
was modelled with a single isotropic Debye-Waller factor.
Structures have been deposited with the Cambridge
Structural Database, deposition codes 1963550-1963557.
Fourier-transform infrared attenuated total reflectance
(FTIR-ATR) measurements were done on a Bruker VERTEX 70
instrument with a PLATINUM diamond crystal ATR unit.
Although the reaction mixtures remained solid throughout
the ageing procedure, with consistency of moist sand, in some
cases after an extended period of time the product was
obtained in the form of crystals of sufficient size (up to ca. 0.2
mm) for structure determination via single crystal X-ray
diffraction. The formation of such macroscopic crystals in a
moist ageing system likely occurs through a process akin to
Ostwald ripening.79 Overall, the herein observed materials
Thermogravimetric and differential scanning calorimetry
(TGA and DSC) analysis was performed on a Mettler-Toledo
TGA DSC 1 Star system thermobalance using alumina crucibles
under an stream of nitrogen (50 ml min−1) and a heating rate
of 5 ºC min−1 from 25 ºC until 200 ºC and in a stream of air (50
ml min−1) and a heating rate of 10 °C min−1 from 200 ºC to 700
ºC Sample size was between 2 mg and 10 mg. Results were
processed with the Mettler STARe 9.01 software.
adopted
a limited number of metal-organic topologies,
Elemental analysis was performed by Thermo Fisher
Scientific iCAPQc inductively coupled plasma mass
illustrated in Figure 2. These include three-dimensional (3-D)
quartz (qtz) and Irish ladder (irl) topologies, 2-dimensional (2-
D) layered square grid (sql, or 4,4-connected) and honeycomb
(hcb, or 3,6-connected) topologies, one-dimensional (1-D)
zigzag chains and 0-dimensional (0-D, discrete) binuclear
clusters.
spectrometer (ICP-MS) equipped with
a CETAC ASX520
autosampler. In a typical example, an aged solid-state mixture
of rare earth oxides was combined with H2ox and either Na2ox
or (NH4)2ox in a respective 1:3:3 stoichiometric ratio. After 10
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 4 of 14
View Article Online
DOI: 10.1039/D0GC00454E
ARTICLE
Figure 2. a) Comparison of selected PXRD diffractograms of the mixtures resulting from the reaction between a rare earth oxide (RE2O3) and H2ox in the respective stoichiometric
ratio 1:3, with colour-coded simulated PXRD patterns of the resulting or structurally similar phases given below the corresponding experimental patterns and denoted by their
respective CSD codes. Graphical representations of the various frameworks observed to form in this work, with the underlying networks represented by grey spheres and yellow
rods: b) distorted hcb; c) one type of herein observed sql (sql-1); d) qtz; e) another type of herein observed sql (sql-2); f) irl; g) one type of herein observed 1-D chains (chain-1) h)
binuclear
discrete
clusters
and
i)
another
type
of
herein
observed
1-D
chains
(chain-2).
Notably, two types of structures with distorted sql-topology number of water guests (n) varies between 6 (in the case of Sc,
were observed, with planar and corrugated layers, and these Lu, and Yb) and 10.5 (in the case of Nd) per empirical formula
are herein designated as sql-1 and sql-2, respectively. unit, as seen from crystal structures and verified by TGA (see
Similarly, two types of zigzag chains, where the bridging ESI), and seems to depend on the size of the metal centre,
oxalate units are either coplanar or at a ~66º angle, herein with larger cations being able to accommodate more guest
designated chain-1 and chain-2, respectively, have also been molecules. As
a direct result of the cation size, the
observed (Figure 2). coordination environment of Sc, Lu and Yb consists of three
Typically, the products of reactions in binary mixtures ox2– species and two coordinated water molecules arranged in
containing only the RE oxide and H2ox were found to adopt the a square antiprism, while the coordination environment of Nd
empirical formula RE2(ox)3·nH2O, forming distorted hcb and Gd consists of three ox2– species and three coordinated
structures (Figure 2b) in which water molecules are water molecules arranged in a capped square antiprism. These
coordinated to the RE centre, as well as occupying the space coordination environments are consistent with those normally
within the hexagonal holes and between individual layers. The encountered in the stereochemistry of lanthanide
Please do not adjust margins

Page 5 of 14
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
complexes.80 Such hcb-structures have been observed to form stoichiometric mixture of RE2O3 and H2ox might be a result of
View Article Online
in geological environments as well, for example in the open local micro-inhomogeneities, as previouDslOyI:o1b0.s1e0r3v9e/Dd0iGnCa0g04e5in4gE
framework cerium-based mineral deveroite,61,81 suggesting formation of cocrystals86 and/or limited diffusion of H2ox in
that organic framework minerals might be a result of ageing- the solid environment. The possibility of forming yttrium
based weathering.
hydronium oxalate by ageing was additionally confirmed
The completion of ageing reactions was in each case through the addition of one equivalent of H2ox to a sample of
evaluated by complete disappearance of Bragg reflections of Y2ox3·6H2O, followed by ageing (Figure 3). As observed by
the oxide reactant, as was previously done for ageing or PXRD, this led to complete conversion into yttrium hydronium
mechanochemical reactions in which metal oxides convert into oxalate. For La2O3 as the reactant, the product PXRD pattern
diverse MOFs.82 The rates of ageing reactions differed strongly could not be related to any reported RE element oxalate
depending on the oxide, with Sc2O3 and Y2O3 undergoing structure and, unfortunately, attempts to index the PXRD
quantitative conversion within 4-5 days, and the lanthanide pattern of this product were unsuccessful.
oxides within one day. However, conducting the reactions at
90 ºC enabled quantitative conversion of all explored RE
oxides to corresponding oxalates in 1 day (see Table 2, also
ESI).
Reactions with H2ox only
The PXRD patterns of aged reaction mixtures involving
Sc2O3, Yb2O3 or Nd2O3 were found to match those simulated
for the reported structures of respective oxalates (Figure 2a).83
Similarly, ageing reactions of yttrium, lutetium, europium and
gadolinium oxides gave microcrystalline powder products
whose PXRD patterns matched those simulated for previously
published structures of oxalates of other RE metals (Table 1).
The crystal structure of Lu2ox3·6H2O (1) was here solved
from PXRD data and it was found to be isostructural to
Yb2ox3·6H2O (AOXYBH),83 validating the assumption that
materials with very similar PXRD patterns are likely to have
similar structures (see ESI, Table S2). Interestingly, Y2O3 and
Gd2O3 both gave mixtures of products with similar PXRD
Figure 3. Overlay of PXRD patterns demonstrating how the addition of one more
equivalent of H2ox into solid sample containing Y2ox3·6H2O leads to
spontaneous conversion into yttrium hydronium oxalate, based on the
Y(H2O)ox2- sql-framework. Starting mixture prepared by ageing of Y2O3 and H2ox
in a 1:3 stoichiometric ratio (see also Figure 2a), respectively: a) before addition
of 1 more equivalent of H2ox; b) after 1 day ageing in the presence of additional
a
patterns (Figure 2a). Based on the PXRD patterns, the mixtures H2ox and c) after 2 days ageing with additional H2ox; d) simulated PXRD patterns
for (H5O2)Yox2 ·H2O (CSD code YOXALT) and e) Gd2ox3·6H2O (CSD code VEXPUK).
in each case consisted of at least two distinct phases, one
being isostructural to gadolinium oxalate (CSD code VEXPUK)84
Reactions of three-component mixtures
and the other to yttrium hydronium oxalate (CSD code
YOXALT).85 The reported structure of the latter material,
yttrium hydronium oxalate, consists of sql-1 anionic layers
Following the successful demonstration of simple and
quantitative conversion of RE element oxides into
corresponding oxalates via low-energy ageing procedure, we
investigated the possibility of forming more complex oxalate-
based frameworks following an established strategy of
introducing simple oxalate salts that could change the ratio of
RE element cations and oxalate anions suitably for the
formation of open anions nets.54,87–89
-
(Figure 2c) of composition Y(H2O)(ox)2 , with the H5O2+ Zundel
cations as guests counterbalancing the anionic framework.85
The coordination environment around the Y3+ can best be
described as distorted capped square antiprism, consisting of
three bidentate ox2– ligands and three coordinated water
molecules.
The appearance of a hydronium oxalate material with a
stoichiometric RE:ox ratio of 1:2 from an initially 1:3
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 6 of 14
View Article Online
DOI: 10.1039/D0GC00454E
ARTICLE
Table 1. Overview of crystalline phases resulting from ageing reactions of RE element oxides with H2ox alone, and with mixtures of H2ox and (NH4)2ox, Na2ox or K2ox. Phases are
designated by the CSD code of the structure of the corresponding metal oxalates, or the CSD code for a metal oxalate structure that appears isostructural to the reaction product.
Shaded cells highlight novel phases that structurally resemble those corresponding to the given CSD code. Corresponding PXRD patterns are given in Figs. 2, 3, 5, 6 and in the ESI.
RE
Sc
Oxalic acid
AOXSCH
framework
NH4
framework
Na
framework
K
framework
hcb
2b
qtz
4
chain-1
5/6 mixture
chain-1/ binuclear
Y
VEXPUK / YOXALT
AOXYBH
hcb / sql-1
hcb
AYOXLH
3
sql-1
sql-1
Unknown
sql-2
irl
BOMLEU
SEYKOX
sql-1
sql-1
CIBZAO
CIBZAO
binuclear
binuclear
binuclear
Unknown
irl / chain-2
irl / chain-2
Yb
Lu
La
Nd
Eu
AOXYBH
1
Unknown
QUHYIB01
YEBKOG
YEBKOG
YEBKOG
BOMLEU
Unknown
Unknown
BOMLEUc
BOMLEUc
sql-1
CIBZAO
Unknown
NDOXALa
Unknown
hcb
Unknown
Unknown
sql-1
Unknown
DEXXOU/7d
DEXXOU/7d
DEXXOU/7d
NDOXALa
hcb
irl
Gd VEXPUK / YOXALT
hcb / sql-1
irl
sql-1
irl / chain-2
d
a PXRD also reveals unidentified phases (see ESI); b pattern resembles simulated one for the structure with CSD code DUJQUW; c material diffracted poorly; material converts to
structure 7 upon further ageing.
Specifically, we expected that ageing of three-component framework channels. The structure is similar to a previously
mixtures of an RE oxide with H2ox and M2ox, (where M=Na, reported scandium MOF (CSD code DUJQUW),21 however the
NH4, K) should yield known, and potentially also previously not unit cell dimensions differ significantly (between 5% and 10 %).
reported coordination architectures. While most such
Analysis by PXRD revealed that ageing of both Y2O3 and
reactions led to completion using 1:1:3 respective La2O3 leads to previously reported layered 2-D sql framework
a
stoichiometric ratio of the RE oxide, M2ox and H2ox, indicating structures with 9- (CSD AYOXLH,90 Figure 2c, sql-1) and 10-
the product composition MRE(ox)2, in some cases complete coordinated (CSD QUHYIB01,91 Figure 2e, sql-2) nodes,
conversion was achieved only using RE2O3, H2ox and M2ox in respectively.
respective stoichiometric ratios 1:3:3 and 1:4:3. Such reactant
Ageing of Yb2O3, (NH4)2ox and H2ox in a 1:1:3 respective
ratios indicate product compositions M3RE(ox)3 and stoichiometric ratio gave compound 3, a microcrystalline
M8RE2(ox)7, respectively.
material with a PXRD pattern strikingly similar to that of
previously reported (NH4)Y(ox)2·H2O (CSD code AYOXLH),
indicating isostructurality. The isostructurality and formation
Reactions in the presence of (NH4)2ox
Reactions with (NH4)2ox produced 2- or 3-D coordination of sql-topology framework (sql-1, Figure 2c) structure was
+
polymers with NH4 ions occupying the channels in the confirmed by structure solution from PXRD data followed by
structures. Ageing of Sc2O3 with (NH4)2ox and H2ox in a 1:1:3 Rietveld refinement (Figure 4b). The RE3+ metal nodes in all
respective stoichiometric ratio was shown to produce materials with sql-1 frameworks investigated herein are
compound 2, with composition (NH4)Scox2·2H2O, over a long coordinated by 4 ox2– in a capped square antiprismatic
period of time (>6 months). Again, as in case of simpler RE environment, often found in lanthanide complexes.80
element oxalates, the transformations could be greatly
.
accelerated by temperature increase: conducting the same
reaction at 90 ºC produced 2 within 2 days (see ESI).
Determination of the structure of 2 from PXRD data (Figure 4a,
Table S2) revealed an anionic open framework with a quartz
(qtz, Figure 2d) topology, consisting of eight-coordinate,
square antiprism Sc3+ nodes, each bridged by 4 ox2– ions,
+
arranged in a tetrahedral fashion, with NH4 cations in the
Please do not adjust margins

Page 7 of 14
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
1:1:3 stoichiometric ratio, respectively, with the simulated pattern for
View Article Online
DOI: 10.1039/D0GC00454E
Nd2(N2C3H12)ox4 (CSD YEBKOG).
During periodic monitoring of some of the three-
component ageing reactions, intermediate phases were
observed. An example is shown in Figure 6, corresponding to
the transformation in a mixture of Yb2O3, (NH4)2ox and H2ox in
a 1:1:3 respective stoichiometric ratio. The reaction mixture
rapidly (within 1 day) yields a still unknown microcrystalline
phase, which subsequently converts to 3 over several days.
Reactions in the presence of Na2ox
Ageing of Sc2O3 with Na2ox and H2ox in a respective 1:3:3 ratio
at 45 ºC over a period of several weeks produced compound 4,
of composition Na3Scox3·6H2O confirmed by TGA. Conducting
the same reaction at 90 ºC provided a moist powder, without
signs of deliquescence, that contained single crystals of 4 large
enough for structural analysis by single crystal X-ray
diffraction. The structure of 4 was found to consist of 1-D
anionic zig-zag chains (chain-1) exhibiting an oxalate-bridged
Sc3+–ox2––Sc3+ motif, with Na+ counterions and water
molecules occupying the space between the chains (Figure 2g,
Table S1), resulting in the overall empirical formula
Na3Scox3·6H2O. The Sc3+ metal nodes are coordinated by four
ox2– species, arranged in a square antiprismatic fashion. This
motif appears in all materials investigated herein exhibiting
the chain-1 structure.
Figure 4. Final Rietveld fits for structure solution from PXRD data for: a)
(NH4)Scox2·2H2O and b) (NH4)Ybox2·H2O. Measured pattern is shown in blue, simulated
in red and the residual is shown in grey. Vertical blue lines denote positions of expected
reflections, while vertical green lines in b) represent expected positions of the (NH4)2ox
impurity.
Reactions of Nd2O3, Eu2O3 or Gd2O3 with (NH4)2ox and H2ox in
respective 1:1:3 stoichiometric ratio gave products with
mutually very similar PXRD patterns. Whereas these PXRD
patterns did not match any reported Nd(III)-, Eu(III)- or Gd(III)-
Table 2. Time (in days) required for quantitative conversion of a mixture of RE2O3
and H2ox, with or without additional M2ox (M=NH4, Na or K), to the respective
complex oxalate product at 96% RH and 45 ºC. Some reaction mixtures exhibited
little to no appreciable change after extended time, which is denoted with ">".
+
based structures involving ox2- and NH4 , there is significant
similarity to the PXRD pattern of 1,2-propanediammonium
neodymium oxalate (CSD code YEBKOG) (Figure 5).92
Consequently, it is likely that these compounds adopt a similar
structure, which is a 3-D anionic open framework of irl
topology (Figure 2f), based on nine-coordinate lanthanide ions
bridged by ox2–. The RE3+ metal nodes in all irl frameworks
investigated herein are coordinated by 4 ox2– ligands and one
metal
H2ox
(NH4)2ox
Na2ox
K2ox
Sc
Y
4
5
4
10
1
1
>180
30
6
>60
>30
>180
>60
>60
6
14
1
14
7
7
7
Yb
Lu
La
Nd
Eu
Gd
3
14
14
3
water molecule adopting
a deformed capped square
antiprismatic coordination environment. It is reasonable to
assume that the RE3+ nodes in the structurally unknown
Nd(III)-, Eu(III)- or Gd(III)-based materials involving ox2- and
NH4+ would adopt the same coordination environment.
1
1
27
>60
7
7
14
Reactions of most RE oxides with Na2ox and H2ox in a
respective stoichiometric 1:1:3 ratio provide mutually
isostructural 2-D coordination polymers of sql-topology (sql-1)
with water molecules between the layers (Figure 2c),
quantified by TGA (see ESI).93,94 The exception to this are
materials containing Nd and La whose structures remain
unknown. Despite several attempts, the crystal structures of
sodium-containing complex oxalates of neodymium and
lanthanum have not been elucidated nor have they been
likened to other known structures based on PXRD patterns.
The ratio of Na to the RE element in these phases has been
determined to be 1:1, which was evidenced by the complete
disappearance of Bragg reflections of the starting materials
from the PXRD pattern of the mixture when reactions were
conducted in that ratio.
Figure 5. Comparison of PXRD patterns of the materials obtained by accelerated
ageing of a mixture of either Eu, Gd or Nd oxide with (NH4)2ox, H2ox and in a
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 7
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 8 of 14
ARTICLE
Journal Name
Reactions in the presence of K2ox
View Article Online
DOI: 10.1039/D0GC00454E
Ageing of Sc2O3, K2ox and H2ox in either the respective
stoichiometric ratios 1:1:3 or 1:3:3 produced mixtures of
phases. However, exploration of different stoichiometric
compositions revealed that ageing of the same starting
materials in the 1:4:3 respective stoichiometric ratio produces
only one phase. Serendipitously, single crystals of two
different phases (5 and 6) were isolated from the same solid
mixture and were used to establish their respective crystal
structures. Structural analysis revealed the composition
K3Scox3·4H2O for 5 and K8Sc2ox7·10H2O for 6 (see ESI, Table
S1). The structure of 5 resembles that of the corresponding
sodium analogue (4), as it consists of infinite chains (chain-1)
based on the repeating Sc3+–ox2-–Sc3+ motif, with K+ cations
and water molecules surrounding the chains (Figure 2g). In
contrast, 6 consists of 0-D discrete binuclear complexes
wherein two Sc3+ cations are bridged by a single ox2- anion and
are coordinated by additional three terminal ox2- groups in a
tetrahedral fashion that are further associated with K+ cations
and water molecules surrounding the complex (Figure 2h).
Ageing of other RE oxides with K2ox and H2ox produced
either a material consisting of binuclear 0-D (in the case of Y,
Yb, Lu and partially Sc) metal oxalate clusters95 (Figure 2h) or
3-D coordination polymers adopting the structural formula
KREox2 and the irl structural motif (Figure 2f), as concluded
from comparison of respective PXRD patterns (Table 1, ESI).96
Moreover, the KREox2 coordination polymers were found to
Figure 6. Comparison of selected PXRD diffractograms for the solid-state mixture
of Yb2O3, (NH4)2ox and H2ox in a respective stoichiometric ratio of 1:1:3,
monitored over a period of 60 days. Bragg reflections belonging to an unknown
convert to materials of general composition K3REox3 upon phase, at ≈12.5º and ≈15.5º 2 appear within 1 day and then gradually disappear
over a longer period of time, disappearing completely after more than 27 days.
further ageing with excess K2ox. The K3REox3 materials were
found to consist of anionic zig-zag chains (chain-2) based on
the RE3+–ox2––RE3+ motif, with K+ counterions and water
molecules occupying the space between them (Figure 2i). Each
RE3+ node in chain-2 structures was found to be coordinated
by four ox2– ligands and one water molecule, arranged in a
capped square antiprism. The structure of K3Gdox3·3H2O (7)
was been determined by single crystal X-ray diffraction using a
crystal obtained from the ageing experiment (see ESI, Table
S1), and the structure of the isostructural material
K3Euox3·3H2O (8) was determined from PXRD data using the
structural parameters of 7 as a starting point (see ESI, Table
S1).
The reaction rates for accelerated ageing of all herein
investigated ternary mixtures were found to differ
significantly, ranging from days to weeks, depending on both
the nature of the oxide and the oxalate salt used (Table 2).
Notably, Sc2O3 has shown to require longer times for full
conversion to the respective oxalate than most herein
Figure 7. Time-dependent change in the relative amounts of different phases in the
investigated oxides. In all cases, ageing of the REO-containing
mixtures at elevated temperatures greatly decreased the time
necessary for full conversion, without changing the reaction
outcome (see ESI). In the case of ageing Sc2O3 with (NH4)2ox
and H2ox, the time for full conversion of the oxide reactant
decreased from >60 days at 45 ºC to only 3 days by conducting
the process at 90 ºC.
reaction of Yb2O3 with three equivalents of H2ox and one equivalent of Na2ox. Lines
connecting data points are intended for easier visual identification and do not
represent fitted values.
During periodic monitoring of the reaction between REO,
H2ox and M2ox, the formation of MHox (the hydrogen oxalate
salt) was regularly observed. To ascertain whether MHox
played a role as an intermediate in the formation of the mixed
oxalate, a homogenous mixture of M2ox and pre-formed
lutetium oxalate was exposed to a humid environment at 45
8 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 9 of 14
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
ºC. The mixture containing lutetium oxalate as the starting analyses provided direct and quantitative insight into the
View Article Online
material converted to the corresponding mixed oxalate in an amounts of each RE both in the filtrateDaOnId: 10th.1e03s9o/Dli0dGrCe0s0id45u4eE.
approximately same time period as the mixtures starting from Specifically, analyses of the liquid extract and solid residue
pure oxide, possibly indicating that the formation of MHox show that this simple, non-optimized methodology in one step
does not play a vital part in conversion of REOs to mixed enriches scandium purity ca. five- or six-fold: from the initial
oxalates (see ESI). To achieve a better understanding of the 12.5 mol% to 68 mol% when using (NH4)2ox and 60 mol%
rate of chemical change from REO and simple oxalate salts to when using Na2ox.
more advanced metal-organic materials, a mixture of Yb2O3,
The analyses (Figure 8b,c, Tables 3 and 4) showed
Na2ox and H2ox in a respective stoichiometric ratio 1:1:3 was surprisingly efficient separation of scandium from other RE in
stored for 67 days at 45 ºC and 96% RH, with periodic the mixture. When using (NH4)2ox, the ratio (r) of Sc to Nd in
monitoring. The Rietveld analyses of the PXRD patterns aqueous filtrate was 152, and the ratio of Sc to Gd was 61. The
measured on the solid sample show that the reaction rate corresponding values for ratios of Sc to La and Eu were 98 and
decreases over time (Figure 7).97 Whether this is because the 79, respectively. The separation of Sc from Y, Yb and Lu was
reaction is diffusion limited, or there is a decrease in available poorer, with rSc/Y, rSc/Yb and rSc/Lu values 13, 6.6 and 5.3
surface area, or some other reason, remains unclear at this respectively. Analogous set of trials using an accelerated
point.
ageing mixture containing Na2ox gave comparable results
(Figure 8c, also Table 4).
Separation of scandium from a mixture of REOs
During attempts to grow single crystals by recrystallization
of (NH4)Scox2·2H2O it was observed that the material was
readily soluble in boiling water. Qualitative testing for
dissolution properties in water revealed that all mixed oxalates
containing potassium cations, as well as all herein prepared
scandium compounds, are readily soluble in boiling water.
Moreover, the mixed oxalates of yttrium, lutetium and
ytterbium containing sodium and ammonium cations exhibited
partial dissolution, which increased upon addition of solid
H2ox, Na2ox or (NH4)2ox, presumably due to the formation of
complex oxalate anions.98 In contrast, the obtained
compounds of gadolinium, neodymium, lanthanum and
europium were found to be insoluble in boiling water, even
upon addition of excess oxalates (Figure 8a).
These qualitative observations on solubility differences are
unexpected, considering that lanthanide oxalates of
conventional composition RE2ox3 are known to have very low
aqueous solubilities, with solubility products in the range of ca.
10-30 69,99
solubility differences, and investigate the possibility to
demonstrate proof-of-principle process for simple
.
This led us to conduct a quantitative analysis of the
a
a
separation of scandium from other RE metals that would start
from a mixture of oxides, without requiring mineral acids,
complex extractants or organic solvents. For this purpose a
model mixture was prepared by mixing equimolar amounts of
all eight herein investigated REOs (ca. 1 mmol each), along
with three equivalents of H2ox (ca. 24 mmol), to which three
equivalents of (NH4)2ox (ca. 24 mmol) were added. The
mixture was thoroughly mixed using a mortar and pestle. An
analogous mixture containing Na2ox instead of (NH4)2ox was
prepared as well. Each mixture was then aged at 90 ºC for six
days at 96% relative humidity, in order to convert the metal
oxides into the respective sodium or ammonium mixed
oxalates. A sample of the resulting solid product was then
suspended in 50 mL of boiling water, the undissolved solid
residue removed by filtration, the resulting solution diluted,
and metal content analyzed by ICP. The solid residue after
filtration was separately dissolved in aqueous HNO3, solution
diluted to 50 mL and subsequently also analyzed by ICP. These
Figure 8. a) Qualitative summary of aqueous solubility of phases resulting from ageing
of binary and ternary mixtures in this work. The distribution of different RE ions in
aqueous solution and in the solid residue based on ICP analysis of mixtures obtained by
ageing RE oxides with: b) (NH4)2ox and H2ox, and c) Na2ox and H2ox. The total amounts
of La, Nd and Eu determined in the solution and the solid residue appear lower than
that of Sc, Y, Yb and Lu due to incomplete digestion of the solid residue by 4% HNO3.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 9
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 10 of 14
ARTICLE
Journal Name
These results are comparable to and often exceed those of
previously reported and considerably more complex SX
systems that are based on popular extractants in organic
solvents or ionic liquids. For example, Zlobina et al. have used
trialkylbenzylammonium di(2-ethylhexyl)phosphate as the
extractant and obtained degrees of enrichment corresponding
to ratios of Sc to La, to Gd and Eu after extraction of 32.2, 23.0
and 26.8, respectively.100The values for Y and Yb were 4.4 and
2.1, respectively. All these numbers are three to four times
lower than those obtained herein. Similarly, Onghena et al.101
have recently used the ionic liquid betainium
bis(trifluoromethylsulfonyl)imide in extractive separation of Sc
from La, Nd, Y and several other REs, resulting in ratios of 7.1,
4.4 and 4.6 , respectively.
Table 3. Measured amounts of RE in aqueous solution (naq, mmol) and solid residue (nr,
DOI: 10.1039/D0GC00454E
mmol), after extracting the aged mixture of 8 RE oxides, H2ox and (NH4)2ox with boiling
water, along with corresponding concentrations of the aqueous solution (caq, mmol
dm-3), of the solution obtained by dissolving the solid residue in aqueous HNO3 (cr,
mmol dm-3), and the associated ratio of scandium to other REs in the aqueous solution
(rSc/RE), separation coefficient ()a and separation factor (Sc/RE).b
View Article Online
c
c
b
RE
Sc
Y
La
Nd
Eu
Gd
Yb
Lu
naq
nrd
caq
crd
a
Sc/RE
-
219
rSc/RE
-
13
98
152
79
61
6.6
5.3
2.45
0.187
0.138
2.31
2.14
1.96
2.26
2.40
1.87
1.840
49
3.74
2.76
46.2
42.8
39.2
45.2
48.0
37.4
36.8
17.77
0.081
0.012
0.008
0.014
0.017
0.197
0.251
0.0248
0.0161
0.0312
0.0398
0.368
0.496
0.322
0.624
0.796
7.36
1481
2221
1269
1045
90
0.462
9.24
71
An often reported separation metric is the separation
factor (), defined as the ratio of distribution coefficients (also
known as the separation coefficient) of two components. In
order to facilitate comparison to previously reported SX
systems, we approximate the separation coefficient in our
process by the ratio of the amount of a metal in the aqueous
solution to that in the solid residue () and, consequently,  as
the ratio of  values for two metals.
We were surprised to see that our simple, unoptimized
approach provided  values that are high compared to a
number of SX processes described in the primary and
secondary literature. Typical  values range between 10-1000
in order of magnitude,102 as illustrated in the work of Sun et al.
who explored scandium separation using the popular
extractant Cyanex 925 in combination with an ionic liquid,
yielding separation factors of Sc to Y, Eu, Yb and La of 34.3,
67.8, 86.7 and 172.8, respectively.11,103 An extraction
procedure used in a red mud pilot plant scandium extraction
process developed by Ochsenkühn-Petropulu104,105 is based on
the popular di(2-ethylhexyl)phosphoric acid extractant and
hexane as the organic phase, and exhibits Sc to Y and Sc to Yb
separation factors of ca. 175, and a Sc to La separation factor
of ca. 3,700.
Some of the highest -values for scandium separation from
REs were reported by Karve and Vaidya,106 who used Cyanex
302 extractant with xylenes as the liquid phase for separation
of Sc from La ( = 16,600), Nd ( = 1,180), Gd ( = 27,600) and
Yb ( = 3,840), with a poor separation from Y (highest reported
Sc:Y ratio 10). While poorer in separating Sc from La, Gd and
Yb, it is interesting to note that the herein reported simple
process still achieves a higher separation both for Nd and Y.
Indeed, considering the simplicity of the herein described
method, the herein obtained  values compare surprisingly
well to technologically much more advanced processes, such
as the pseudoliquid membrane design described by Yang and
co-workers,13 which exhibited  values for separating Sc from
Y, La, Nd, Eu, Gd, Yb and Lu of 769, 4945, 2413, 338, 21194,
1150 and 3262, respectively.
a) Separation coefficient  is defined as the ratio of amounts of metal in aqueous
solution and in solid residue after extraction: naq/nr; b) separation factor for Sc
with respect to a particular RE, Sc/RE, is defined as the ratio of  values for Sc and
a selected RE; c) established by ICP using a suitably diluted sample of aqueous
solution of aged mixture after extraction with boiling water; d) established by ICP
using a solution obtained by dissolving the solid residue after extraction with
boiling water in dilute aqueous HNO3.
It is worth also noting that the ratio of concentrations of Lu
and Nd in the liquid phase was 28:1 which makes this method
interesting for separation of not only Sc but potentially also
other RE, such as Y, Yb or Lu, from a mixture. Such very
promising results coming from an unoptimized separation
process warrant a more in-depth exploration of separation of
RE elements using other simple salts.
Table 4. Measured amounts of RE in aqueous solution (naq, mmol) and solid residue (nr,
mmol), after extracting the aged mixture of 8 RE oxides, H2ox and Na2ox with boiling
water, along with corresponding concentrations of the aqueous solution (caq, mmol
dm-3), of the solution obtained by dissolving the solid residue in aqueous HNO3 (cr,
mmol dm-3), and the associated ratio of scandium to other REs in the aqueous solution
(rSc/RE), separation coefficient ()a and scandium separation selectivity (Sc/RE).b
RE
Sc
Y
La
Nd
Eu
Gd
Yb
Lu
naq
nrd
0.0685
2.27
1.914
1.89
2.16
2.20
1.79
1.58
caq
crd
a
Sc/RE
-
357
426
822
rSc/RE
-
c
c
b
2.42
0.225
0.158
0.0816
0.067
0.0652
0.406
0.65
48.4
4.5
1.37
45.4
38.3
37.8
43.2
44
35.35
0.099
0.083
0.043
0.031
0.030
0.227
0.408
10.8
15.3
29.7
36.1
37.2
6.0
3.16
1.632
1.34
1.304
8.12
13.0
1140
1178
156
87
35.8
31.6
3.7
a) Separation coefficient  is defined as the ratio of amounts of metal in aqueous
solution and in solid residue after extraction: naq/nr; b) separation factor for Sc
with respect to a particular RE, Sc/RE, is defined as the ratio of  values for Sc and
a selected RE; c) established by ICP using a suitably diluted sample of aqueous
solution of aged mixture after extraction with boiling water; d) established by ICP
using a solution obtained by dissolving the solid residue after extraction with
boiling water in dilute aqueous HNO3.
Conclusions
In summary, we have demonstrated the ability to
spontaneously convert a range of rare earth element oxides,
including Sc, Y, La, Nd, Gd, Eu, Yb and Lu, into complex metal-
10 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 11 of 14
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
organic structures by solvent-free, low-energy ageing with Innovation (CFI) and McGill University William Dawson fund.
View Article Online
oxalic acid and its salts (Na, K and NH4) in moist air. These M.A. thanks the National Science CenDteOrI: o10f.1P03o9la/Dn0dGCfo00r4t5h4eE
systematic studies demonstrate accelerated ageing as a financial support via SONATA grant (2018/31/D/ST5/03619).
simple, mild and materials-efficient route to convert the inert
oxides into more tractable oxalates, including open framework
materials, as opposed to mechanochemical treatment that has
Notes and references
so far been unsuccessful in converting the oxides into metal-
organic structures. Accelerated ageing led to the discovery and
structural characterization of at least eight new phases, and
suggested a simple strategy to separate scandium from a
mixture of rare earth elements without special complexation
agents, organic solvents, or aggressive reagents. The herein
investigated solvent-free reactions have led to the structural
characterization of four new scandium metal-organic
1
2
3
4
5
S. M. Hayes and E. A. McCullough, Resour. Policy, 2018, 59,
192–199.
F. Wall Rare earth elements in: Critical Metals Handbook,
1st Edition, Ed. G. Gunn, John Wiley & Sons (2014)
D. Schlinkert, K. Gerald and V. Den Boogaart, Resour.
Policy, 2015, 46, 272–280.
N. Haque, A. Hughes, S. Lim and C. Vernon, Resources,
2014, 3, 614–635.
T. V Molchanova, I. D. Akimova, K. M. Smirnov, O. K.
Krylova and E. V Zharova, Russ. Metall., 2017, 2017, 170–
174.
compounds, providing
a significant contribution to the
understanding of this poorly explored element. The results are
also of potential significance in the context of geochemistry,
specifically the carbon cycle and mineralogy of the rare earth
elements, as they demonstrate the formation of known
mineral structures, such as cerium-based deveroite,107 under
conditions resembling mineral weathering in nature. In that
context, it is tempting to consider the diversity of phases that
were herein observed to form upon ageing of rare earth oxides
with oxalic acid and its salts as a library of potential, not yet
discovered mineral species.51 Finally, the herein explored
solvent-free ageing chemistry enabled a simple proof-of-
principle method for the separation of scandium from a
mixture with other rare earth element oxides, based simply on
different solid-state reactivity of the oxides and aqueous
solubility of the resulting metal-organic materials. The
proposed separation methodology uses exclusively oxalic acid
or its simple salts as the organic components of solid-state
reactivity. Precipitation with oxalic acid, ammonium oxalate
and/or sodium oxalate is a highly efficient way of separating
RE elements from other metals in solution, and is already
employed in a number of current mining and mineral
processing approaches. Consequently, the proposed
methodology should be highly compatible with existing
processes and enable pre-concentration of scandium and
simplify subsequent ion and/or or solvent exchange processes.
We are currently investigating other possible applications of
accelerated ageing in production of critical elements, such as
indium which is mainly obtained as a by-product of zinc mining and
processing,108 or for processing industrial waste products such as
the ‘red mud’.109
6
R. A. Schmitt, R. H. Smith and D. A. Olehy, Geochim.
Cosmochim. Acta, 1964, 28, 67–86.
7
M. A. R. Onal, C. R. Borra, M. Guo, B. Blanpain and T. Van
Gerven, J. Sustain. Metall., 2015, 1, 199–215.
C. R. Borra, B. Blanpain, Y. Pontikes, K. Binnemans and T.
Van Gerven, J. Sustain. Metall., 2016, 2, 365–386.
P. S. Arshi, E. Vahidi and F. Zhao, ACS Sustain. Chem. Eng.,
2018, 6, 3311–3320.
X. Zhao, M. Wong, C. Mao, T. X. Trieu, J. Zhang, P. Feng and
X. Bu, J. Am. Chem. Soc., 2014, 136, 12572–12575.
X. Sun, Y. Ji, L. Guo, J. Chen and D. Li, Sep. Purif. Technol.,
2011, 81, 25–30.
8
9
10
11
12
13
J. P. Faris and J. W. Warton, Anal. Chem., 1962, 34, 1077–
1080.
X. Yang, Z.-M. Gu and D.-X. Wang, J. Memb. Sci., 1995, 106,
131–145.
14
15
F. Habashi, J. Chem. Technol. Biotechnol., 1985, 35A, 5–14.
T. Dutta, K. Kim, M. Uchimiya, E. E. Kwon, B. Jeon, A. Deep
and S. Yun, Environ. Res., 2016, 150, 182–190.
K. Binnemans, P. Tom, B. Blanpain, T. Van Gerven, Y. Yang,
A. Walton and M. Buchert, J. Clean. Prod., 2013, 51, 1–22.
M. H. D. Bandara, J. W. Darcy, D. Apelian and M. H.
Emmert, Environ. Sci. Technol., 2014, 48, 6553–6560.
S. Kobayashi, Eur. J. Mineral., 1999, 15–27.
K. Ishihara, M. Kubota, H. Kurihara and H. Yamamoto, J.
Am. Chem. Soc., 1995, 117, 4413–4414.
A. F. Norman, P. B. Prangnell and R. S. M. C. Ewen, Acta
Mater., 1998, 46, 5715–5732.
J. Cepeda, S. Pérez-Yáñez, G. Beobide, O. Castillo, A. Luque,
P. A. Wright, S. Sneddon and S. E. Ashbrook, Cryst. Growth
Des., 2015, 15, 2352–2363.
16
17
18
19
20
21
Conflicts of interest
There are no conflicts to declare.
22
23
24
J. Gambogi, U.S. Geological Survey, Mineral Commodity
Summaries - Scandium, 2019, vol. 1.
C. K. Gupta and N. Krishnamurthy, Extractive Metallurgy of
Rare Earths, CRC Press, Boca Raton, 2005.
P. K. Parhi and S. S. Behera Processing technology for
extraction of scandium(III) from secondary sources – a
comprehensive approach, in: Critical and Rare Earth
Elements Recovery from Secondary Sources, Ed. A. Akcil, A.
Akcil, CRC Press, New York (2020)
Acknowledgements
We thank Prof. F. Schaeper and Dr. P. Daneshmand for help in
collecting single crystal X-ray diffraction data of 4. We
acknowledge the financial support of NSERC (RGPIN-2017-
06467, NSERC RGPAS 507837-17), Canada Foundation for
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 11
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 12 of 14
ARTICLE
Journal Name
25
W. Wang, Y. Pranolo and C. Y. Cheng, Hydrometallurgy,
2011, 108, 100–108.
F. Xie, T. An, D. Dreisinger and F. Doyle, Miner. Eng., 2014,
56, 10–28.
W. Wang and C. Y. Cheng, J. Chem. Technol. Biotechnol.,
2011, 86, 1237–1246.
B. Yagmurlu, W. Zhang, D. Avdibegovic, M. Regadío, R.
Koivula, C. Dittrich, K. Binnemans and B. Friedrich, ALTA
2018 Uranium-REE-Lithium Proc., 2018, 85–93.
D. Kołodyńska and Z. Hubicki, in Ion Exchange
Technologies, InTech, 2012, pp. 101–154.
F. Habashi, Can. Metall. Q., 2013, 52, 224–233.
I. Huskić, I. Halasz, T. Friščić and H. Vančik, Green Chem.,
2012, 14, 1597–1600.
A. M. Fidelli, B. Karadeniz, A. J. Howarth, I. Huskić, L. S.
Germann, I. Halasz, M. Etter, S.-Y. Moon, R. E. Dinnebier, V.
Stilinović, O. K. Farha, T. Friščić and K. Užarević, Chem.
Commun., 2018, 54, 6999–7002.
T. Friščić, C. Mottillo and H. M. Titi, Angew. Chem. Int. Ed.,
2019, 2–14.
N. R. Rightmire and T. P. Hanusa, Dalt. Trans., 2016, 45,
2352–2362.
D. Tan and F. García, Chem. Soc. Rev., 2019, 48, 2274–
2292.
E. Boldyreva, Chem. Soc. Rev., 2013, 42, 7719–7738.
D. Braga, S. L. Giaffreda, F. Grepioni, A. Pettersen, L. Maini,
M. Curzi and M. Polito, J. Chem. Soc. Dalt. Trans., 2006, 60,
12491249–12631263.
A. Beillard, X. Bantreil, T. X. Métro, J. Martinez and F.
Lamaty, Chem. Rev., 2019, 119, 7529–7609.
D. Hasa and W. Jones, Adv. Drug Deliv. Rev., 2017, 117,
147–161.
J. Andersen and J. Mack, Green Chem., 2018, 20, 1435–
1443.
49
50
51
52
53
I. Huskić, J. C. Christopherson, K. Užarević and T. Friščić,
View Article Online
Chem. Commun., 2016, 52, 5120–5D12O3I:.10.1039/D0GC00454E
26
27
28
M. Dud, O. V. Magdysyuk, D. Margetić and V. Štrukil, Green
Chem., 2016, 18, 2666–2674.
X. Feng, C. Jia, J. Wang, X. Cao, P. Tang and W. Yuan, Green
Chem., 2015, 17, 3740–3745.
J. L. Do, D. Tan and T. Friščić, Angew. Chemie - Int. Ed.,
2018, 57, 2667–2671.
M. Glavinović, M. Krause, L. Yang, J. A. Mcleod, L. Liu, K. M.
Baines, T. Friščić and J.-P. Lumb, Sci. Adv., 2017, 3,
e1700149.
29
30
31
54
55
F. Qi, R. S. Stein and T. Friščić, Green Chem., 2014, 16, 121–
132.
N. V. Chukanov, N. V. Zubkova, G. Möhn, I. V. Pekov, D. I.
Belakovskiy, K. V. Van, S. N. Britvin and D. Y.
Pushcharovsky, Mineral. Mag., 2018, 1–17.
J. Chen, H. P. Blume and L. Beyer, Catena, 2000, 39, 121–
146.
P. Adamo and P. Violante, Appl. Clay Sci., 2000, 16, 229–
256.
X. Lim, Nature, 2015, 524, 20–21.
C. Jia, J. Wang, X. Feng, Q. Lin and W. Yuan,
CrystEngComm, 2014, 16, 6552–6555.
I. Huskić and T. Friščić, Acta Crystallogr., 2015, A71, S58.
I. Huskić and T. Friščić, Acta Crystallogr. Sect. B, 2018, 74,
539–559.
R. M. Hazen, D. R. Hummer, G. Hystad, R. T. Downs and J. J.
Golden, Am. Mineral., 2016, 101, 889–906.
O. E. Piro and E. J. Baran, Crystallogr. Rev., 2018, 24, 149–
175.
C. L. F. Dazem, F. M. A. Noa, J. Nenwaa and L. Öhrström,
CrystEngComm, 2019, DOI: 10.1039/c9ce01187k.
W. Le, S. Kuang, Z. Zhang, G. Wu, Y. Li, C. Liao and W. Liao,
Hydrometallurgy, 2018, 178, 54–59.
C. Liu, L. Chen, J. Chen, D. Zou, Y. Deng and D. Li, J. Rare
Earths, 2019, 37, 1002–1008.
32
56
57
33
34
35
58
59
60
61
36
37
62
63
64
65
66
67
68
69
38
39
40
41
J. L. Howard, Q. Cao and D. L. Browne, Chem. Sci., 2018, 9,
3080–3094.
42
43
G. W. Wang, Chem. Soc. Rev., 2013, 42, 7668–7700.
S. L. James, C. J. Adams, C. Bolm, D. Braga, P. Collier, T.
Frišcic, F. Grepioni, K. D. M. Harris, G. Hyett, W. Jones, A.
Krebs, J. Mack, L. Maini, A. G. Orpen, I. P. Parkin, W. C.
Shearouse, J. W. Steed and D. C. Waddell, Chem. Soc. Rev.,
2012, 41, 413–447.
J. L. Do and T. Friščić, ACS Cent. Sci., 2017, 3, 13–19.
(a) C. Mottillo, Y. Lu, M. H. Pham, M. J. Cliffe, T. O. Do and
T. Friščić, Green Chem., 2013, 15, 2121–2131; (b) C.
Mottillo and T. Friščić Molecules 2017, 22, 144.
D. Wu, C. Niu, D. Li and Y. Bai, J. Alloys Compd., 2004, 374,
442–446.
C. Liao, S. Wu, F. Cheng, S. Wang, Y. Liu, B. Zhang and C.
Yan, J. Rare Earths, 2013, 31, 331–336.
P. Josso, S. Roberts, D. A. H. Teagle, O. Pourret, R.
Herrington and C. Ponce de Leon Albarran, Miner. Eng.,
2018, 118, 106–121.
L. Greenspan, J. Res. Natl. Bur. Stand. Sect. A Phys. Chem.,
1977, 81A, 89.
N. Kourkoumelis, Powder Diffr., 2013, 28, 137–148.
G. M. Sheldrick, Acta Crystallogr. Sect. A Found.
Crystallogr., 2015, 71, 3–8.
G. M. Sheldrick, Acta Crystallogr. Sect. C Struct. Chem.,
2015, 71, 3–8.
A. Boultif and D. Louer, J. Appl. Cryst., 2004, 37, 724–731.
W. I. F. David, K. Shankland, J. van de Streek, E. Pidcock, W.
D. S. Motherwell and J. C. Cole, J. Appl. Cryst., 2006, 39,
910–915.
44
45
70
71
46
47
M. J. Cliffe, C. Mottillo, R. S. Stein, D. K. Bučar and T. Friščić, 72
Chem. Sci., 2012, 3, 2495–2500.
(a) D. Cinčić, I. Brekalo and B. Kaitner, Chem. Commun.,
2012, 48, 11683–11685; (b) S. Kaabel, R. S. Stein, M.
Fomitšenko, I. Järving, T. Friščić and R. Aav Angew. Chem.
Int. Ed. 2019, 58, 6230; (c) F. Hammerer, L. A. Loots, J.-L.
Do, P. D. Therien, C. W. Nickels, T. Friščić and K. Auclair
Angew. Chem. Int. Ed. 2018, 57, 2621-2624.
73
74
75
48
D. Braga, S. L. Giaffreda, F. Grepioni, M. R. Chierotti, R.
Gobetto, G. Palladino and M. Polito, CrystEngComm, 2007,
9, 879–881.
76
77
78
G. S. Pawley, J. Appl. Cryst., 1981, 14, 357–361.
A. Coelho, TOPAS Acad. 6, Coelho Softw., 2016.
H. M. Rietveld, J. Appl. Cryst., 1969, 2, 65–71.
12 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 13 of 14
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
79
80
81
P. W. Voorhes, J. Stat. Phys., 1985, 38, 231–252.
Mag., 2013, 77, 3019–3026.
View Article Online
J. C. G. Bünzli, J. Coord. Chem., 2014, 67, 3706–3733.
I. Huskić and T. Friščić, Philos. Trans. R. Soc. A, 2019, 377,
1–19.
108
109
A. M. Alfantazi and R. R. Moskalyk,DMOiIn:e1r0..1E0n3g9./,D200G0C30,01465,4E
687–694.
S. Reid, J. Tam, M. Yang and G. Azimi, Sci. Rep., 2017, 7, 1–
82
I. Halasz, T. Friščić, S. A. J. Kimber, K. Užarević, A. Puškarić,
C. Mottillo, P. Julien, V. Štrukil, V. Honkimäki and R. E.
Dinnebier, Faraday Discuss., 2014, 170, 203–221.
E. Hansson, Acta Chem. Scand., 1973, 27, 823–834.
L. Cañadillas-Delgado, J. Pasán, O. Fabelo, M. Hernández-
Molina, F. Lloret, M. Julve and C. Ruiz-Pérez, Inorg. Chem.,
2006, 45, 10585–10594.
9.
83
84
85
86
87
88
G. D. Brunton and C. K. Johnson, J. Chem. Phys., 2003, 62,
3797.
I. Huskić, J.-C. Christopherson, K. Užarević and T. Friščić,
Chem. Commun., 2016, 52, 5120–5123.
A. K. Cheetham and C. N. R. Rao, MRS Bull., 2005, 30, 93–
98.
S. Ayyappan, A. K. Cheetham, S. Natarajan and C. N. R. Rao,
Chem. Mater., 1998, 10, 3746–3755.
89
90
L. Öhrström and K. Larsson, Dalt. Trans., 2004, 347–353.
T. R. R. McDonald and J. M. Spink, Acta Crystallogr., 1967,
23, 944–949.
91
92
93
94
T. Bataille, M. Loue and J. Au, J. Solid State Chem., 2000,
95, 81–95.
A. Mohanu, C. Brouca-Cabarrecq and J. C. Ã. Trombe, J.
Solid State Chem., 2006, 179, 3–17.
T. Bataille and D. Louër, Acta Crystallogr. Sect. C Cryst.
Struct. Commun., 1999, 55, 1760–1762.
B. Chapelet-Arab, L. Duvieubourg, G. Nowogrocki, F.
Abraham and S. Grandjean, J. Solid State Chem., 2006, 179,
4029–4036.
95
96
97
98
I. A. Kahwa, F. R. Fronczek and J. Selbin, Inorganica Chim.
Acta, 1984, 82, 167–172.
Y. Liu and S. Li, Acta Crystallogr. Sect. E Struct. Reports
Online, 2007, 63, 312–313.
K. A. Connors, Chemical Kinetics: The Study of Reaction
Rates in Solution, VCH, New York, 1990.
C. E. Crouthamel and D. S. Martin, J. Am. Chem. Soc., 1951,
73, 569–573.
99
100
Y. Xiong, Appl. Geochemistry, 2011, 26, 1130–1137.
E. Zlobina, A. Ismailova and K. Tassibekov, MATEC Web
Conf., 2017, 96, 1–6.
101
102
103
104
105
B. Onghena and K. Binnemans, Ind. Eng. Chem. Res., 2015,
54, 1887–1898.
Y. Chen, H. Wang, Y. Pei and J. Wang, Sep. Purif. Technol.,
2017, 178, 261–268.
X. Sun, D. Wu, J. Chen and D. Li, J. Chem. Technol.
Biotechnol., 2007, 272, 267–272.
M. Ochsenkühn-Petropulu, T. Lyberopulu and G. Parissakis,
Anal. Chim. Acta, 1995, 315, 231–237.
M. T. Ochsenkühn-Petropoulou, K. S. Hatzilyberis, L. N.
Mendrinos and C. E. Salmas, Ind. Eng. Chem. Res., 2002, 41,
5794–5801.
106
107
M. Karve and B. Vaidya, Sep. Sci. Technol., 2008, 43, 1111–
1123.
A. Guastoni, F. Nestola, P. Gentile, F. Zorzi, M. Alvaro, A.
Lanza, L. Peruzzo, M. Schiazza and N. M. Casati, Mineral.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 13
Please do not adjust margins

Green Chemistry
Page 14 of 14
View Article Online
DOI: 10.1039/D0GC00454E
Table of Contents Entry
Solvent-free accelerated ageing converts metal oxides into metal oxalate frameworks, enabling
simple separation of scandium from other rare earth elements.