Separation and Purification Technology 194 (2018) 265–271

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

NdFeB magnet recycling: Dysprosium recovery by non-dispersive solvent T

extraction employing hollow fibre membrane contactor
⁎
Kartikey K. Yadav, Mallavarapu Anitha, D.K. Singh , V. Kain
Rare Earths Development Section, Material’s Processing & Corrosion Engineering Division, Materials Group, Bhabha Atomic Research Centre, Mumbai 400085, India

A R T I C L E I N F O A B S T R A C T

Keywords: Hollow Fibre Membrane (HFM) operation in non-dispersive solvent extraction (NDSX) mode has been suc-
Recycling cessfully employed for the separation of dysprosium from NdFeB magnetic scrap material using EHEHPA as an
NdFeB magnets extractant. Effect of various hydrodynamic parameters including aqueous phase acidity, extractant concentra-
Hollow fibre membrane tion, Dy concentration in feed, phase ratio and flow rate were investigated to optimize the condition for
Dysprosium
quantitative recovery of dysprosium from the leach liquor obtained by the dissolution of hard disk drive (HDD)
in nitric acid media. 0.5 M EHEHPA, 0.3 M HNO3 as aqueous feed phase, 100 ml/min flow rate and phase ratio of
1:1 were found to be optimum for both the cycle test runs. Taking the advantage of fast extraction kinetics of
dysprosium, two cycle HFM-NDSX approach was adopted to concentrate Dy from 20 to 83% in first cycle while
raising its purity > 97% in the second cycle. The overall process also yielded neodymium rich by-product for its
further purification.

1. Introduction
Rare earth elements are critical raw material for wide range of
modern age products ranging from catalysts, phosphors, magnet, lasers,
fibre optics and several more [1]. There is a great research interest
growing in rare-earth (RE) recycling due to increase in demand for REs
and associated complexity of REs separation from primary ores [2].
Neodymium Iron Boron magnet (NdFeB) is one of the major consumer
for rare earth elements which include Nd, Dy and Pr. NdFeB magnets
are best available magnets due to their superior energy product (with a
theoretical maximum of 512 kJ/m3) and fulfil almost 70% of perma-
nent magnetic material demand worldwide. NdFeB permanent magnet
contains about ∼32 wt% REEs (mainly 21–31 wt% (Nd/Pr), 0–10 wt%
Dy and small amounts of Gd and Tb) [3]. Dysprosium added to NdFeB
magnets plays a crucial role in preserving the performance of the
magnet at elevated temperature which is essential for high-temperature
applications [4]. Apart from NdFeB magnets, dysprosium also finds an
important place in the manufacturing of laser materials, nuclear re-
actors, hybrid cars and colour television tubes [5].
Du and Graedel [6] have recently published very interesting fact
about the global stock in use for four REs in NdFeB permanent magnets,
i.e., neodymium (Nd), praseodymium (Pr), terbium (Tb), and dyspro-
sium (Dy), which amount to a total of 97.0 kilotons (kt): 62.6 kt Nd,
15.7 kt Pr, 15.7 kt Dy, and 3.1 kt Tb. This vast resource of RE elements
could augment the existing supply as they are quantitatively almost
four times the 2007 annual production of the individual elements, if
recycled efficiently. RE recycling not only supplements existing demand
but can also reduce the geopolitical impact on economic aspects due to
supply risk. Less than 1% of the REEs were recycled by 2011, and de-
velopmental researches for new large-scale processes are scanty [7].
The scarcity of data on the current quantity of REE materials from
magnets in the waste streams and the fate of the magnets after shred-
ding has hampered the possible development of REE recycling processes
[8]. There are various metallurgical processes practised to separate and
recover the REEs in the NdFeB magnet scrap such as liquid metal ex-
traction [9], hydrometallurgical processing [10,11], and pyr-
ometallurgical slag extraction [4,12]. Hydrometallurgical, an important
route, involves leaching of magnetic scraps with acids at room tem-
perature, leading to dissolution of all the components in NdFeB mag-
nets. Subsequently rare earth can be precipitated as double sulphate salt
which could be easily converted to corresponding fluoride or oxide
[13]. The resultant oxide consists majority of rare earth elements
mostly Nd, Pr and Dy present in magnetic scrap. However subsequent
purification of dysprosium can enhance the value as well as importance
of the recycling process due to its invariably high cost and demand.
There are various conventional methods available for purification of
individual rare earth elements such as solvent extraction, ion exchange
and fractional precipitation [14]. Solvent extraction is the industrially
accepted method for purification of rare earth elements. In solvent
extraction process, organophosphorous acids based organic extractants

⁎
Corresponding author.
E-mail address: dksingh@barc.gov.in (D.K. Singh).

https://doi.org/10.1016/j.seppur.2017.11.025
Received 1 September 2017; Received in revised form 23 October 2017; Accepted 10 November 2017
Available online 13 November 2017
1383-5866/ © 2017 Elsevier B.V. All rights reserved.
K.K. Yadav et al.
are employed for separating REE, due to their solvation properties as
well as their chemical stability and low aqueous solubilities [15]. Acidic
organophosphorous extractants, such as 2-ethylhexyl 2-ethyhexylpho-
sphonic acid (EHEHPA also known as PC88A), di-2-ethyl hexyl phos-
phoric acid (D2EHPA) are widely used for the extraction and separation
of individual rare earths including dysprosium [16]. However, Dy re-
covery from NdFeB magnet scrap is difficult with solvent extraction
process owing to requirement of multiple stages and low Dy content in
feed source [17]. The present scenario consisting shortfall in supply of
critical rare earths elements has attracted a lot of research interest in
the field of rare earth separation by alternative methods i.e. membrane
based separation, solvent impregnated resins and ionic liquids from
waste and secondary sources [18–21].
Membrane assisted solvent extraction employing hollow fibre
membrane modules can provide suitable alternative to recover and
purify dysprosium from mixed rare earths feed [22]. Membrane based
techniques eliminates the disadvantages of conventional solvent ex-
traction process such as lean source, environmental hazard associated
with loss of organic extractant, loss of diluents in aqueous streams and
third phase formation [23]. In the case of hollow fibre membrane op-
eration mode, individual separation of rare earth elements is enhanced
in non-equilibrium conditions due to the continuous high driving forces
over extended periods of time compared to equilibrium limited con-
ventional solvent extraction [24]. Hollow fibre membrane modules
provide high surface area contact per unit volume resulting in high RE
extraction rate and can be employed in two modes namely supported
liquid membrane and non-dispersive solvent extraction (NDSX) mode.
In NDSX mode an organic extractant phase passing through one side of
the membrane wets the microporous hydrophobic membrane. A non-
wetting aqueous phase is passed through other side of the membrane at
a pressure higher than that of the organic phase, but lower than that
needed for the aqueous phase to displace the organic phase in the pores
of the membrane. The aqueous–organic (aqueous–membrane) interface
is essentially immobilized at the pore mouth of the hydrophobic
membrane support through which the solute mass transfer takes place.
Here, the membrane support does not function as a size selective sieve,
but it merely prevents the dispersion of one phase into the other.
Among them only few have explored hollow fibre membrane module
for rare earth separation [25]. However non dispersive solvent extrac-
tion mode which has promising potential for scale up among membrane
based separation methods has not yet been explored extensively.
Therefore it is desirable to develop an efficient separation method to
process the leach solution obtained from NdFeB magnet scrap material
to get high purity rare earth oxides.
Accordingly the objective of the present study is to explore the
HFM-NDSX under wide range of experimental variables to address the
complexity involved in Dy separation from NdFeB magnet scrap by
solvent extraction process by employing hollow fibre membrane
module in non dispersive mode to recover the same.
2. Experimental
2.1. Materials
2-Ethylhexyl 2-ethylhexyphosphonic acid (EHEHPA) shown in
Fig. 1 also known as EHEHPA having molecular weight of 306.4 and
Fig. 1. Structure of 2-Ethylhexyl 2-ethylhexyphosphonic acid.
266
Separation and Purification Technology 194 (2018) 265–271
Table 1
Characteristic properties of petrofin.
Characteristics Specifications
Density (at 15 °C) 0.75 ± 0.01
Purity 99%
Carbon distribution
C11 20–30%
C12 25–35%
C13 20–30%
C14 20–25%
Flash point 70 °C
purity of 95% has been received from Heavy Water Board and used as it
is. Experiments performed during the recovery of rare earth by em-
ploying hollow fibre membrane required organic extractant to be used
as carrier phase. The extractants were diluted to optimize hydro-
dynamic conditions of the process. Petroffin (heavy normal paraffin)
having specification shown in Table 1 has been used as diluent. Rare
earth (such as Dy (III), Nd(III), Pr(III) etc.) solutions were prepared by
dissolving their oxides (> 99% purity, received from Indian Rare Earth
Limited, India) in concentrated nitric acid and individual working so-
lution were prepared by the appropriate dilution of the stock solutions.
NdFeB magnetic scraps shown in Fig. 2 were recovered from end of life
Hard Disk Drive obtained from Computer Division, Bhabha Atomic
Research Centre, Mumbai. The magnets were first demagnetized by
heating well above the Curie temperature (30 min at 350 °C) and then
fed into a jaw crusher resulting in magnetically-neutral particles ready
for chemical processing. Subsequently the same was crushed in mortar
and pestle to powder form for efficient leaching of metal into corre-
sponding acids.
2.2. Non-dispersive solvent extraction studies
HFM contactor (Liqui Cell 2.5 × 8) used in the present work was
procured from Polypore, USA and it consists of hydrophobic micro-
porous polypropylene hollow fibres enclosed in a polypropylene shell of
2.5″ × 8″ dimension as shown in Fig. 3. The module contained about
10,000 fibres of 40% porosity and had an effective surface area of
1.4 m2 with other specifications, listed in Table 2. The aqueous and
organic phases (agitated continuously in the reservoir for uniform
concentration) were contacted in counter-current mode in the hollow
fibre module pores. The flow rates of aqueous and organic phases were
kept at 100 ml/min with the help of gear pumps (cole-palmer) through
the lumen side and shell side of the HFM module, respectively. The
polypropylene membrane is hydrophobic in nature so the organic phase
wets and even passes through the membrane fibres. To prevent the
dispersion of the organic phase into the aqueous feed phase a small
trans-membrane pressure in the lumen side is applied.
2.3. Sample analysis
The evaluation of rare earth elements extraction/stripping was
studied by monitoring its concentration in aqueous phase by using
Inductively Coupled Plasma-Atomic Emission Spectrophotometer, ICP-
AES (JY-Ultima 2). The experiments were performed thrice and the
mean values are reported. The errors in non dispersive solvent extrac-
tion experiments and metal ion concentration analysis were less than
4%.
3. Results and discussion
3.1. Batch extraction of Dy by EHEHPA
Separation of heavy rare earth (Dy) by Hollow fibre membrane
K.K. Yadav et al. Separation and Purification Technology 194 (2018) 265–271

Fig. 2. NdFeB magnets recovered from Hard Disk Drives with casing and without casing.

while subscripts (aq) and (org) depict the aqueous and organic phase
respectively. One mole dysprosium metal ion forms complex with 3 mol
of dimer of EHEHPA and liberate three moles of hydrogen ion.
[Dy (HA2)3](org ) [H+]3(aq)
K ex =
[Dy 3 +](aq) [H2 A2 ]3(org ) (2)
The distribution coefficient of the metal ion (D) is the ratio of Dy
(III) concentration in the organic phase to that of in the aqueous phase
as represented in Eq. (3).
[Dy (HA2)3](org )
D=
[Dy 3 +](aq) (3)
In order to select the aqueous acidity for HFM–NDSX test, pre-
liminary experiments on the effect of acidity on D value of dysprosium
extraction were performed in single contact mode (batch) by equili-
brating an aqueous solution containing fixed metal ion concentration
with varying nitric acid concentration (0.1–0.75 M) with 0.5 M
EHEHPA. Results are summarized in Table 3. It is evident with increase
in aqueous phase acidity from 0.1 to 0.75 M HNO3 distribution ratio for
Fig. 3. Hollow fibre membrane module. Dy extraction decreases from 53 to 0.3. The decrease in D value with
increase in acidity is due to protonation of EHEHPA given in Eq. (1).
The distribution data indicated that the extraction of Dy(III) can only be
Table 2
HFM module specification. performed at or below the acidity value of 0.3 M HNO3.

Parameter Specification 3.2. Non-dispersive solvent extraction

Fibre material Polypropylene
Number of fibres 9950 3.2.1. Effect of aqueous phase acidity on Dy extraction
Fibre internal diameter (µm) 240 In the case of solvent extraction by organophosphorus based acidic
Fibre outer diameter (µm) 300 extractants, aqueous phase acidity plays a key role in extraction as well
Fibre wall thickness (µm) 30 as stripping of metal ion. The nitric acid concentration (0.3–0.5 M
Effective pore size (µm) 0.03
HNO3) was varied to investigate the behaviour on Dy extraction with
Porosity (%) 40
Tortuosity 2.5 0.5 M EHEHPA using HFM in non dispersive solvent extraction mode
Effective fibre length (cm) 15 under recirculation at organic to aqueous phase ratio of 1:1. Results
Effective surface area (m2) 1.4 illustrated in Fig. 4 indicate that an increase in acidity from 0.1 to 0.5 M
HNO3 led to decrease in extraction of Dy(III) from 99.5 to 45.2%. The
decrease in extraction (transport) of Dy3+ from aqueous side to organic
module and operating the system in non dispersive solvent extraction
mode from end of life rare earth scrap material (which also consist Nd
Table 3
and Pr as major constituents in NdFeB magnets) requires understanding Distribution coefficient (D) for Dy(III) extraction
of two phase reaction mechanism for Dy extraction. EHEHPA an acidic by EHEHPA from different aqueous phase acidities:
extractant exists as a dimer in non polar diluents which extract dys- Organic phase: 0.5 M EHEHPA in petrofin, aqueous
prosium by cationic exchange mechanism from nitric acid media as per phase: 1089 mg/L Dy.

Eq. (1) [25,26]. [HNO3], M D

K ex
Dy(3aq+) + 3[H2 A2 ](org ) ⇐
⇒ [Dy (HA2)3](org ) + 3H(+aq) (1)
0.1 53.48
0.3 3.91
0.5 1.15
where H2A2 shows the dimer form of EHEHPA in non polar diluents, Kex 0.75 0.35
is the equilibrium constant for the chemical reaction shown in Eq. (1)

267
K.K. Yadav et al. Separation and Purification Technology 194 (2018) 265–271

Fig. 4. Effect of Aqueous phase acidity on the extraction of Dy(III) by 0.5 M EHEHPA; Fig. 5. Effect of organic phase flow rate on the percentage extraction of Dy(III) at fixed
aqueous phase: 1000 mg/L; flow rate: 100 mL/min, Phase ratio: 1:1. aqueous phase flow rate (100 ml/min); organic phase: 0.5 M EHEHPA; aqueous phase:
1000 mg/l of Dy(III) at 0.1 M HNO3; Phase ratio: 1:1.

phase with increase in aqueous acidity (0.1–0.5 M HNO3) leads to in-

crease in H+ ion concentration and equilibrium shown in chemical Eq.
(1) shifts in backward direction. Increase in acidity also affects the
equilibration time required from 30 to 60 min to achieve the quanti-
tative transport of Dy(III) from aqueous phase to organic phase. The
loaded organic phase of 0.5 M EHEHPA with Dy(III) resulted from this
experiment was quantitatively stripped by equilibrating the organic
phase with 3.5 M HNO3 in a separating funnel. The regenerated 0.5 M
EHEHPA was reused for the repeat experiments and results obtained
showed excellent reproducibility in terms of Dy(III) extraction and no
degradation of solvent system/membrane.

3.2.2. Effect of flow rate

In order to study the mechanism of mass transfer in NDSX mode of
operation in HFM and also to evaluate the rate determining step of total
mass transfer resistance, it is important to investigate the hydro-
dynamic characteristics of current system. The flow rates of chosen
system (RE-HNO3-EHEHPA-Petrofin) in both phases (feed and ex-
tractant) play a pivotal role in the transport of metal ion from the feed
side to the organic side. Flow rates of both organic and aqueous phases
in non-dispersive solvent extraction carried out in hollow fibre mem-
brane affects the contact time of organic and aqueous phase at the in-
terface. In the case of Dy(III) extraction with 0.5 M EHEHPA, effect of
organic phase flow rate was examined by varying them in the range of
50–200 mL/min, while keeping aqueous phase flow rate constant at
100 mL/min. Results shown in Fig. 5 indicate that though the extraction
of Dy(III) in organic phase was quicker at lower flow rate in the initial
stages of the experiment (< 10 min of run) it was independent of
contact time beyond 30 min. The result also led to the conclusion
that > 99.0% Dy(III) could be extracted in the organic phase in 35 min
irrespective of flow rates. Similar kind of results were also observed
when aqueous phase flow rate was varied between 50 and 200 mL/min
keeping organic phase flow rate at 100 mL/min. The minimal effect of
flow rates on Dy3+ extraction may be attributed to the availability of
excessive extractant (EHEHPA) (1:1 organic to aqueous phase ratio) on
organic side and low metal ion concentration in feed phase. Due to the
insignificant effect of aqueous and organic phase flow rates on equili-
brium extraction of Dy(III), all subsequent experiments were performed
at 100 mL/min flow rate of both the phases.
3.2.3. Effect of Dy(III) concentration
After investigating the effect of acidity and flow rates another im-
portant parameter the effect of Dy(III) concentration in aqueous solu-
tion was also monitored by carrying out the HFM test run in NDSX
Fig. 6. Effect of Dy(III) concentration on its extraction by 0.5 M EHEHPA organic phase
and 0.1 M HNO3 aqueous phase at flow rates of 100 mL/min for both the phases.
mode with fixed concentration of 0.5 M EHEHPA and nitric acid 0.3 M.
Fig. 6 summarizes the results with increase in Dy(III) concentration
from 250 to 2000 mg/L increased the time required for the quantitative
(99%) transport across the membrane. However, in all the cases > 98%
extraction of Dy(III) from aqueous phase was observed in 60 min even
with 2000 mg/L Dy (III) in the feed phase. The loaded organic phase of
0.5 M EHEHPA with Dy(III) resulted from this experiment was quanti-
tatively stripped by equilibrating the organic phase with 3.5 M HNO3 in
a separating funnel. The strippant was later analyzed for Dy(III) content
by ICP-AES and subsequently stripping results were used to ascertain
the mass balance of extraction by NDSX experiment. The observed
behaviour is in accordance with conventional solvent extraction process
where the percentage extraction of metal ion decreases with increase in
metal ion concentration in aqueous phase. The decrease in extraction is
explained in terms of the less anionic binding sites of free extractant in
organic phase due to continuous forward reaction leading to enhanced
H+ ions in the medium as shown in chemical Eq. (1).
3.2.4. Influence of extractant concentration on Dy(III) extraction
Organic extractant plays a significant role on extraction of metal ion

268
K.K. Yadav et al. Separation and Purification Technology 194 (2018) 265–271

Fig. 7. Effect of EHEHPA concentration on Dy extraction from 0.1 M HNO3 aqueous phase Fig. 8. Extraction of Dy(III) at different organic to aqueous phase volume ratios; organic
at flow rates of 100 mL/min for both the phases. phase: 0.5 M EHEHPA; aqueous phase: 1.0 g/L Dy(III) at 0.3 M HNO3; flow rate: 100 mL/
min.

from aqueous solution under comparable extraction conditions. In the

present investigation effect of EHEHPA concentration on percentage
transport of Dy(III) was studied in the extractant concentration range
from 0.2 M to 0.83 M. The results of percentage Dy(III) transport ob-
tained from varying concentration of EHEHPA are shown in Fig. 7. The
% transport of Dy(III) increased from 82 to 99 in 90 min of equilibra-
tion with increase in concentration of the organic phase from 0.2 M to
0.45 M, however, when EHEHPA concentration increased from 0.45 M
to 0.83 M, the increase in % extraction of Dy(III) was not obvious.
Within this concentration of EHEHPA range from 0.2 M to 0.45 M, the
availability of EHEHPA at the feed–membrane–organic interfaces in-
creased with the increasing of concentration of carrier. Beyond 0.5 M
EHEHPA the increase in metal ion extraction is independent of ex-
tractant concentration due to the excess of extractant present at the
interface [26]. When the EHEHPA concentrations were 0.2, 0.45, 0.6,
0.75 and 0.83 M, the percentage extraction values were 82.49, 99.60,
99.65, 99.81 and 99.80, respectively. The optimum concentration of
EHEHPA to perform further experiments was chosen to be 0.5 M
EHEHPA in petrofin.
3.2.5. Effect of phase ratio
In order to determine the loading of Dy(III) in the organic phase of
EHEHPA, effect of organic to aqueous phase ratio O/A was studied in
the range of 0.25–1.0. Experimental results depicted in Fig. 8 suggest
that the time required to achieve quantitative recovery of Dy(III) from
aqueous to organic phase increased with decrease in organic to aqueous
phase volume ratio. At an organic to aqueous phase ratio of 1, about 25
min of contact time was required for the quantitative extraction of Dy
(III) from a feed phase containing 1.0 g/L Dy(III) in 0.3 M HNO3.
However under the comparable experimental conditions, time required
for quantitative extraction (99%) of dysprosium increases from 80 to
130 min, when phase ratio decreased from 0.5 to 0.25. It is also ob-
served that ∼96% extraction of Dy(III) could be achieved in the case O/
A of 0.25 which is relatively less than the recovery obtained in phase
ratios of 0.5 or 1.
was prepared by demagnetizing the NdFeB magnet followed by
crushing and grinding in mortar and pestle and subsequent selective
leaching, precipitation and re-dissolution of rare earths described
elsewhere [13].
The feed solution prepared from magnetic scrap composed of RE
such as Dy, Nd and Pr in varying concentration depending upon the
type of source. Experiments were performed to evaluate the separation
of Dy(III) from other rare earths such as Nd and Pr, found in magnetic
scrap. The maximum amount of Dy found in rare earth magnet was
found to be approximately 20% of total rare earths and the composition
of resultant aqueous phase 200, 438, 385 mg/L of Dy(III), Nd(III) and Pr
(III) respectively in HNO3 media. To maximize the recovery of Dy,
aqueous feed having acidity of 0.3 M and 0.1 M respectively was pro-
cessed in HFM module employing 0.5 M EHEHPA of organic extractant.
Figs. 9 and 10 show that in both the feed acidities (0.1 and 0.3 M)
initially % extraction of Dy(III) in organic phase increased and rose up
to 70% and 95% respectively in 10 min of contact time, however
competing ions Nd and Pr were extracted < 3 and < 4% in 0.3 M
aqueous acidity and < 13 and < 11% in 0.1 M HNO3 media respec-
tively in same time. Experimental results depicted in Figs. 9 and 10
clearly indicate the possibility of separation as well as purifying Dy

3.2.6. Dysprosium recovery from NdFeB magnetic scraps

After optimization of experimental variables for quantitative re-
covery of Dy from aqueous nitric acid media using EHEHPA in HFM-
NDSX mode of operation, actual rare earth magnetic scrap (NdFeB)
obtained from end of life computer hard disk drives were evaluated to
recover valuable rare earth i.e. Dysprosium, Neodymium and Fig. 9. Extraction of Dy(III) in presence of competing rare earths ions; organic phase: 0.5
Praseodymium by employing the developed NDSX process in hollow M EHEHPA; aqueous phase: 220, 438, 385 mg/L of Dy(III), Nd(III) and Pr(III) respectively
in 0.3 M HNO3; flow rate: 100 mL/min.
fibre membrane module. The feed solution for the recovery experiments

269
K.K. Yadav et al.
Fig. 10. Extraction of Dy(III) in presence of competing rare earths ions; organic phase: 0.5
M EHEHPA; aqueous phase: 215, 416, 399 mg/L of Dy(III), Nd(III) and Pr(III) respectively
in 0.1 M HNO3; flow rate: 100 mL/min.
from the Nd and Pr mixture by employing hollow fibre membrane
contactor. Dysprosium purification possibilities from the mixture of Nd
and Pr by using EHEHPA in hollow fibre membrane were found to be
time dependent. This finding leads the way to recover high pure Dy
from the mixture of rare earths found in magnetic scrap material. After
complete equilibration (120 min of operation), % extraction of Dy, Nd
and Pr by 0.5 M EHEHPA from 0.3 M HNO3 feed in HFM contactor in
NDSX mode were found to be 99.88, 13 and 8% respectively. Whereas,
in the case of 0.1 M HNO3 media % extraction of Dy, Nd and Pr by HFM
module under similar condition was 99.8, 78.9 and 68.57 respectively.
A second cycle of non dispersive extraction in hollow fibre mem-
brane was introduced to further purify Dysprosium from 70% to > 90%
obtained from the above operation (first cycle) of NDSX in HFM module
of 10 min equilibration time. The loaded organic obtained from the first
cycle was stripped with 3.5 M HNO3 and this formed a feed for the
second cycle. Aqueous second cycle feed consisting 1059, 226 and
184 mg/L Dy, Nd and Pr in 0.1 M HNO3 was processed in HFM-NDSX
mode of operation using 0.5 M EHEHPA with flow rate of 100 mL/min
each. Results summarized in Fig. 11 suggested that the Dy separation
from Nd and Pr by HFM was due to kinetics of metal ion transport
across the membrane. A careful operation of the system yielded Dy
with > 97% purity and 94% recovery from its mixture with Nd and Pr
obtained from the NdFeB magnets of Hard Disk Drives.
Fig. 11. Purification of Dy(III) from mixed rare earths feed; organic phase: 0.5 M
EHEHPA; aqueous phase: 1059, 226, 184 mg/L of Dy(III), Nd(III) and Pr(III) respectively
in 0.1 M HNO3; flow rate: 100 mL/min.
270
Separation and Purification Technology 194 (2018) 265–271
Fig. 12. Stripping of Dy(III) from loaded 0.5 M EHEHPA phase; strip phase: 1.5 M HNO3;
flow rate:100 mL/min, phase ratio 1:1, organic phase 0.5 M EHEHPA having 3.3 g/L Dy
(III).
3.2.7. Stripping of Dy(III) from loaded EHEHPA by HFM
Experiments on rare earth stripping from loaded organic phase were
carried out to establish the viability of non-dispersive mode of opera-
tion by employing hollow fibre contactor to separate Dy from aqueous
phase in continuous mode operation. Furthermore, the effect of contact
time at a definite flow rate of strip phase i.e. 1.5 M HNO3 on the
transport of Dy(III) from loaded organic phase i.e. 0.5 M EHEHPA was
evaluated. Fig. 12 depicts the stripping of 0.5 M EHEHPA loaded with
3.3 g/L Dy(III) with 1.5 M HNO3 aqueous strip phase. It is evident from
the results that complete stripping could be achieved with 150 min of
equilibration through the pores of HFM. Lower phase ratio i.e. O/
A < 0.5 can be utilized to further concentrate the Dy product in strip
phase.
4. Conclusions
Extraction behaviour of rare earth elements (Dy, Nd and Pr) found
in NdFeB magnet, across the hollow fibre membrane pores with
EHEHPA as an extractant have been studied by operating HFM in non-
dispersive solvent extraction mode. The separation and purification of
Dy(III) by non dispersive mode of extraction was found to be affected
by process variables such as aqueous phase acidity, organic phase
concentration, metal ion concentration and presence of competing ions.
Successful stripping of Dysprosium from loaded organic phase in non-
dispersive mode of solvent extraction indicate the possibility of using
this method in continuous operation mode even at large scale for rare
earth separation. The recycling of the organic phase 0.5 M EHEHPA
resulted in excellent reproducibility of the extraction results.
Experimental findings revealed that under optimized conditions.
Dysprosium could be successfully separated from the scrap rare earth
magnetic material by employing NDSX method with product consisting
up to 97% purity and > 94% recovery. The results obtained with the
present study establish that hollow fibre membrane module not only
offers a better separation method but also have the potential to be up
scaled for rare earth recycling.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.seppur.2017.11.025.
K.K. Yadav et al.
References
[1] U.S. Department of Energy, Critical Materials Strategy. < http://energy.gov/sites/
prod/files/DOE_CMS2011_FINAL_Full.pdf > (accessed 09.03.2017).
[2] B.C. McLellan, G.D. Corder, S.H. Ali, Sustainability of rare earths—an overview of
the state of knowledge, Minerals 3 (2013) 304–307.
[3] Neodymium Magnet. < https://en.wikipedia.org/wiki/Neodymium_
magnet > (accessed 15.06.2017).
[4] Y. Yang, A. Walton, R. Sheridan, K. Güth, R. Gau, O. Gutfleisch, M. Buchert,
B.M. Steenari, T.V. Gerven, P.T. Jones, K. Binnemans, REE recovery from end-of-life
NdFeB permanent magnet scrap: a critical review, J. Sustain. Metall. 3 (2017)
122–149.
[5] E. Alonso, A.M. Sherman, T.J. Wallington, M.P. Everson, F.R. Field, R. Roth,
R.E. Kirchain, Evaluating rare earth element availability: a case with revolutionary
demand from clean technologies, Environ. Sci. Technol. 46 (2012) 3406–3414.
[6] X. Du, T.E. Graedel, Global in-use stocks of the rare earth elements: a first estimate,
Environ. Sci. Technol. 45 (9) (2011) 4096–4101.
[7] K. Binnemeans, P.T. Jones, B. Blanpain, T. Van Gerven, Y.X. Yang, A. Walton,
M. Buchert, Recycling of rare earths: a critical review, J. Clean. Prod. 51 (2013)
1–22.
[8] B. Sprecher, Y. Xiao, A. Walton, J. Speight, R. Harris, R. Kleijn, G. Visser,
G.J. Kramer, Life cycle inventory of the production of rare earths and the sub-
sequent production of NdFeB rare earth permanent magnets, Environ. Sci. Technol.
48 (7) (2014) 3951–3958.
[9] T.H. Okabe, O. Takeda, K. Fukuda, Y. Umetsu, Direct extraction and recovery of
neodymium metals from magnet scrap, Mater. Trans. 44 (2003) 798–801.
[10] J.W. Lyman, G.R. Palmer, Recycling of rare earths and iron from NdFeB magnet
scrap, High Temp. Mater. Process. 11 (1–4) (1993) 175–187.
[11] S. Wellens, B. Thijs, K. Binnemans, An environmentally friendlier approach to hy-
drometallurgy: highly selective separation of cobalt from nickel by solvent extrac-
tion with undiluted phosphonium ionic liquids, Green. Chem. 14 (2012)
1657–1665.
[12] T. Saito, H. Sato, S. Ozawa, J. Yu, T. Motegi, The extraction of Nd from waste NdFeB
alloys by the glass slag method, J. Alloys Compd. 353 (2003) 189–193.
[13] J.W. Lyman, G.R. Palmer, Recycling of Neodymium Iron Boron Magnet Scrap,
United States Department of the Interior, Bureau of Mines, Report of Investigation
9481 (1993) 1–28.
[14] P. Enghag, Scandium, yttrium, lanthanum and the 14 lanthanides – rare earth
271
Separation and Purification Technology 194 (2018) 265–271
metals (REMs), Encyclopedia of the Elements, Wiley-Vch Verlag GmbH & Co, 2004.
[15] S.A. Ansari, P.K. Mohapatra, V.K. Manchanda, Recovery of actinides and lantha-
nides from high-level waste using hollow-fiber supported liquid membrane with
TODGA as the carrier, Ind. Eng. Chem. Res. 48 (2009) 8605–8612.
[16] T. Sato, Liquid-liquid extraction of rare-earth elements from aqueous acid solutions
by acid organophosphorus compounds, Hydrometallurgy 22 (1989) 121–140.
[17] H.S. Yoon, C.J. Kim, K.W. Chung, S.D. Kim, J.Y. Lee, J.R. Kumar, Solvent extraction,
separation and recovery of dysprosium (Dy) and neodymium (Nd) from aqueous
solutions: waste recycling strategies for permanent magnet processing,
Hydrometallurgy 165 (2016) 27–43.
[18] M. Gergoric, C. Ekberg, B.M. Steenari, T. Retegan, Separation of heavy rare-earth
elements from light rare-earth elements via solvent extraction from a neodymium
magnet leachate and the effects of diluents, J. Sustain. Metall. 1 (2017) 1–10.
[19] S. Riano, K. Binnemans, Extraction and separation of neodymium and dysprosium
from used NdFeB magnets: an application of ionic liquids in solvent extraction to-
wards the recycling of magnets, Green. Chem. 17 (2015) 2931–2942.
[20] K.K. Yadav, D.K. Singh, M. Anitha, L. Varshaney, H. Singh, Studies on separation of
rare earths from aqueous media by polyethersulfone beads containing D2EHPA as
extractant, Sep. Purif. Technol. 118 (2013) 350–358.
[21] F. Kubota, T. Kakoi, M. Goto, S. Furusaki, F. Nakashio, T. Hano, Permeation be-
havior of rare earth metals with a calix[4]arene carboxyl derivative in a hollow-
fiber membrane, J. Membr. Sci. 165 (2000) 149–158.
[22] L. Peia, L. Wang, W. Guo, N. Zhaoc, Stripping dispersion hollow fiber liquid
membrane containing PC-88A as carrier and HCl for transport behavior of trivalent
dysprosium, J. Membr. Sci. 378 (2011) 520–530.
[23] P.R. Danesi, Separation of metal species by supported liquid membranes, Sep. Sci.
Technol. 19 (1984–85) 857–894.
[24] D. Kim, L.E. Powell, L.H. Delmau, E.S. Peterson, J. Herchenroeder, R.R. Bhave, A
supported liquid membrane system for the selective recovery of rare earth elements
from neodymium-based permanent magnets, Sep. Sci. Technol. 51 (2016)
1716–1726.
[25] D. Kim, L.E. Powell, L.H. Delmau, E.S. Peterson, J. Herchenroeder, R.R. Bhave,
Selective extraction of rare earth elements from permanent magnet scraps with
membrane solvent extraction, Environ. Sci. Technol. 49 (2015) 9452–9459.
[26] Y. Huang, M. Tanaka, Solvent extraction equilibrium of dysprosium (III) from nitric
acid solutions with 2-ethylhexylphosphonic acid mono-2-ethylhexyl ester, T.
Nonferr. Metal Soc. 20 (4) (2010) 707–711.