Hydrometallurgy 169 (2017) 306–313

Contents lists available at ScienceDirect

Hydrometallurgy

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

Solvent extraction of rare earths using a bifunctional ionic liquid. Part 1:

Interaction with acidic solutions
James E. Quinn a,⁎, Karin H. Soldenhoff a, Geoffrey W. Stevens b
a
Australian Nuclear Science & Technology Organisation, Australia
b
Department of Chemical and Biomolecular Engineering, The University of Melbourne, Victoria 3010, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Bifunctional ionic liquid extractants have been examined in recent years for the separation of rare earth elements
Received 9 August 2016 by solvent extraction. One such extractant is the quaternary ammonium phosphonate ionic liquid ‘R4N+EHEHP−’
Received in revised form 21 January 2017 formed from Aliquat 336 (trioctyl/decylmethylammonium chloride) and EHEHPA (2-ethylhexyl phosphonic acid
Accepted 12 February 2017
2-ethylhexyl ester, also known as P507 or PC88A).
Available online 14 February 2017
In this work, the uptake of hydrochloric acid by R4N+ EHEHP− is investigated, by studying the effect of pH, chlo-
ride, diluent and extractant concentration and composition on the extraction of acid. A fundamental understand-
ing of the interactions between R4N+ EHEHP− and acid is considered crucial in order to describe the distribution
of rare earths in such a system.
The results show that when dissolved in low concentration in a polar phase, the acid uptake by the ionic liquid as
a function of pH is indistinguishable from EHEHPA. However, when dissolved in toluene at process relevant con-
centrations (up to 0.5 M), the protonation reaction occurs over a broader pH range and shifts to lower pH when
compared with EHEHPA alone. Increasing the chloride concentration shifts the reaction to higher pH. The reac-
tion stoichiometry, 31P{1H} NMR spectra and observed trends are consistent with protonation of the phospho-
nate ion and simultaneous formation of R4N+ Cl−. It is suggested that ion-pairing between the quaternary
ammonium and phosphonate ions is responsible for the observed change in the protonation behaviour compared
with EHEHPA alone. Our results indicate that the combination of Aliquat 336 and EHEHPA behaves as a mixture of
R4N+ EHEHP− and R4N+ Cl− + EHEHPA, with the protonation dependent on the pH, and not as a different chem-
ical form. Hence, regardless of how the ionic liquid has been prepared, once it is exposed to acid containing so-
lutions, the acid-base behaviour of the ionic liquid is indistinguishable from a mixture of the two reagents.
Crown Copyright © 2017 Published by Elsevier B.V. All rights reserved.

1. Introduction
Industrial solvent extraction circuits used for the recovery and/or
separation of metal ions utilise an active component (the extractant)
dissolved in an inert and water immiscible organic solvent (the diluent).
The relatively high volatility of most commonly used diluents can be of
concern from an environmental standpoint. Room Temperature Ionic
Liquids (RTILs) are entirely ionic compounds that are liquid at or
below room temperature. They usually consist of bulky organic cations
(e.g. tetraalkylammonium, imidazolium etc.) and organic or inorganic
anions (e.g. hexafluorophosphate, tetrafluoro borate and chloride)
(Han and Armstrong, 2007). RTILs have received some attention in re-
cent years in a number of applications, including as a solvent in organic
synthesis, in catalysis and in separation processes (Marsh et al., 2004).
Apart from the interesting solvation chemistry often exhibited by this
class of compound, they have also been suggested as an example of a
so-called ‘green’ solvent. They are considered ‘green’ primarily due to
⁎ Corresponding author at: Locked Bag 2001, Kirrawee DC, NSW 2232, Australia.
E-mail address: jamesq@ansto.gov.au (J.E. Quinn).
their negligibly low volatility and generally high chemical stability
(Zhao et al., 2005).
RTILs have been examined for the solvent extraction of metal ions,
either as a replacement for conventional diluents or extractants, or
used neat as both the extractant and the diluent (Dietz, 2006). Most
RTILs, however, are considerably viscous at the near ambient tempera-
tures relevant to industrial solvent extraction circuits (Marsh et al.,
2004). Because of this, they would need to be dissolved into a suitable
organic diluent for industrial applications, in much the same way that
traditional extractants generally are. In fact, Aliquat 336 (trioctyl/
decylmethylammonium chloride) is one such ionic liquid that, as early
as the 1960s, has been examined or used industrially for metal extrac-
tions including Mo, V, Th, W, Zn and the rare earths (REs) (Ritcey,
2006). It is therefore important to note that there are no inherent envi-
ronmental improvements offered by diluted RTILs as a class for solvent
extraction, as a volatile diluent will still be required. Any environmental
benefits would be achieved through either the lower toxicity, higher
stability or lower aqueous solubility of the specific ionic liquid, or by a
change in process chemistry that results in a more efficient separation
and hence lower reagent consumption or production of waste solutions.

http://dx.doi.org/10.1016/j.hydromet.2017.02.008
0304-386X/Crown Copyright © 2017 Published by Elsevier B.V. All rights reserved.
 J.E. Quinn et al. / Hydrometallurgy 169 (2017) 306–313
Recently, bifunctional ionic liquid extractants (Bif-ILEs), also some-
times referred to as ‘binary extractants’ and ‘acid base coupled (ABC)
extractants’, have been suggested for rare earth separations (Belova et
al., 2009; Chen et al., 2013; Kalyakin and Kuz'min, 2000; Rout et al.,
2012; Sun et al., 2010a; Sun et al., 2010b; Yang et al., 2016; Yang et al.,
2012). The premise of a Bif-ILE is that both the cation and anion partic-
ipate in the extraction mechanism. One example of this is an ionic liquid
containing the quaternary ammonium cation from Aliquat 336 and the
phosphonate anion from EHEHPA (2-ethylhexyl phosphonic acid 2-
ethylhexyl mono ester, also known by trade names P507, PC 88A and
Ionquest 801). Several claims have been made concerning improved
separation factors achieved between certain rare earth ions (Chen et
al., 2013; Sun et al., 2010a; Yang et al., 2016). This Bif-ILE extractant,
henceforth referred to as ‘R4N+ EHEHP− ʼ, is the focus of the present
study.
Understanding the behaviour of the Bif-ILE extractant in equilibrium
with hydrochloric acid (HCl) solutions is a critical aspect of the study of
rare earth extractions. Chen et al. (2013) attributed the decrease in RE
distribution ratios with decreasing pH to competition for R4N+ EHEHP−
with the extraction of acid. They suggested that the mechanism of acid
extraction is most similar to that of solvating extractants such as tributyl
phosphate (TBP) or tri-n-octylphosphine oxide (TOPO) (Eq. (1)).
 
R4 Nþ EHEHP− þ HCl ⇌ ðR4 Nþ EHEHP− Þ⋅HCl ð1Þ
The extraction of acid by ABC extractants was summarised by Eyal
(1997). His analysis focussed on mixtures of amines with organic
acids, including results for quaternary ammonium ions. While a number
of different interactions were noted to be possible for these systems, the
dominant reaction for extraction of strong mineral acids by quaternary
ammonium ions at moderate acidities was one of conversion to the pro-
tonated acid/ammonium chloride forms as represented by Eq. (2)
(where HA denotes the organic acid).
  
− −
R4 Nþ EHEHP þ Hþ þ Cl ⇌ R4 Nþ Cl þ HA ð2Þ
Rout et al. (2012) suggest that the reaction for the mixture of Aliquat
336 and the closely related DEHPA (di-2-ethylhexyl phosphonic acid) is
best represented by Eq. (2), although the researchers noted that the ex-
tent to which the reaction was occurring was not known. Belova and
Khol'kin (2015) have also examined the extraction of acids with this
type of extractant. They investigated the extraction of HCl by a mixture
of Cyanex 272 (bis-2,4,4-trimethylpentyl phosphinic acid) and Aliquat
336 and suggested that Eq. (2) was the more appropriate extraction
mechanism.
Clearly, in order to draw any meaningful conclusions from the study
of the extraction of RE with Bif-ILE, the interactions of the extractant
with HCl/NaCl solutions needs to be better understood. The question
that this study addresses is: In the context of RE extraction with R4N+ -
EHEHP− at extractant and acid concentrations relevant to industrial
processes, does the reagent act as an undissociated ionic liquid through-
out, or are the extractant properties reflective of the individual compo-
nents? Part 1 of this study focuses on the extractant/acid interactions
and Part 2 focuses on the extraction of rare earths. We chose to examine
HCl as this is the most commonly used acid for rare earth separations
(Liao et al., 2013).
2. Materials and methods
2.1. Materials
Aliquat 336 (methyltrioctyl/decylammonium chloride, N90% qua-
ternary ammonium content, BASF) was used without purification.
Ionquest 801 (2-ethylhexyl phosphonic acid 2-ethylhexyl mono ester,
EHEHPA) was supplied by SNF FloMin and also not purified. The diluent
307
in all experiments was toluene (Merck, N 99.9% purity), although pre-
liminary experiments revealed that aliphatic diluents would also be
suitable. Toluene was selected because the sodium salt of EHEHPA
does not form a third phase in this diluent, which simplifies compari-
sons between EHEHPA and R4N+ EHEHP−. The exception was one set
of results in Section 3.3, for which Shellsol 2046 (narrow cut kerosene
containing 17% aromatics, supplied by Shell) was used as diluent.
Shellsol 2046 was used in this test to determine whether there were
any significant differences when between toluene and a diluent more
appropriate for industrial applications. All other reagents were of ana-
lytical grade.
2.2. Synthesis of ionic liquid
The ionic liquid R4N+ EHEHP− was prepared using a modification of
the method outlined by Sun et al. (2010b). Aliquat 336 chloride (482 g,
0.96 mol) was made up to 2 L in isopropanol and added to 1 L of 0.96 M
KOH in isopropanol. A white precipitate of KCl formed immediately.
After mixing at 50 °C for 4 h, the precipitate was removed by centrifug-
ing using an Eppendorf Centrifuge 5810 at 3750 RPM, and the solution
was diluted to 5 L using deionised H2O. To this solution, 246 g of
EHEHPA (0.79 mol) was added and refluxed at 50 °C for 12 h. After
mixing was stopped, the upper layer was washed several times with
deionised water to remove excess R4N+ OH−, then the excess
isopropanol/water solvent was removed by using a Heidolph Hei-Vap
rotary evaporator (50 °C/10 mbar) to yield a viscous orange liquid.
2.3. Extraction of acid and chloride (Section 3.1)
A 0.25 M sample of the R4N+ EHEHP− was contacted with various
solutions of 1 M NaCl at different acidities (10 mL of each phase).
After the contact, the aqueous pH was measured, and the concentration
of acid extracted was determined by the difference between the feed
and equilibrium acidities. The chloride concentration in the organic
sample was determined by first back stripping the sample with 5 M ni-
tric acid to remove the extracted chloride. The chloride concentration
was then measured by titration with silver nitrate in the presence of po-
tassium chromate (Mohr's method). The chemical shift of the major res-
onance in the 31P{1H} NMR spectrum of each organic sample was
recorded. Spectra were acquired on a Bruker Avance DPX400 at
161.97 MHz. Samples were run neat with a coaxial insert containing
0.1 M H3PO4 in deuterated acetone for referencing and locking, as de-
scribed by Marie et al. (2012).
2.4. Potentiometric titrations
2.4.1. Polar phase (Section 3.2)
1.5 × 10−3 mol of R4N+ EHEHP− was added to 100 mL of 75% etha-
nol in water and titrated using 1.0 M HCl. pH measurements were per-
formed using a Metrohm 6.0232.100 probe and 631 meter. The
relationship between the pH as measured by the electrode and the con-
centration of acid in solution was determined by titration of a blank eth-
anol/water solution. The curves were then modelled as a simple acid
protonation reaction as described in the text with an allowance for the
acid required for the pH change. The same procedure was used for
EHEHPA, except that this extractant was titrated using 1.0 M NaOH
(to pH 11) followed by 1.0 M HCl back to the starting pH.
2.4.2. Two-phase titrations (Sections 3.3–3.5)
20 mL of organic phase was mixed in aqueous continuous mode
with an equal volume of aqueous solution, with the compositions of
both as described in the text. The ionic strength was fixed at 1.0 M
(NaCl/HCl) except when examining the effect of the total chloride con-
centration. Titrations were then conducted using 5.2 M HCl as the titrant
for protonation or 1.0 M NaOH for deprotonation. The pH in the aqueous
continuous dispersion was monitored directly using the same probe as
308 J.E. Quinn et al. / Hydrometallurgy 169 (2017) 306–313
described in Section 2.4.1, with the correlation between acidity and pH
determined by conducting an experiment containing 1.0 M NaCl/tolu-
ene only. The amount of acid or base consumed was then determined
by subtracting the amount of acid or base required for the pH change
(as determined from the blank experiment) from the known amount
of acid or base added. The % protonation was defined as:
nHþconsumed
% Protonation ¼  100, for titrations using
nEHEHPA or R4NþEHEHP−
HCl.
! phosphonic acid. However, in industrial solvent extraction circuits, the
nOH−consumed
% Protonation ¼ 1−  100, for titrations
nEHEHPA or R4NþEHEHP−
using NaOH.
3. Results/discussion
3.1. Extraction of acid
The extraction of chloride and acid by 0.25 M R4N+ EHEHP− is
shown in Fig. 1. The results show broad uptake of acid and chloride as
the equilibrium pH was decreased from 9 through to 0. The extractant
appears to be saturated at approximately a 1:1 ratio of extractant to
acid at around pH 1. The stoichiometry is therefore consistent with
the reactions as shown in either Eq. (1) or (2).
The 31P{1H} NMR results reveal that the chemical shift of the phos-
phorus resonance also changes considerably with the equilibrium pH
of the aqueous phase, with a smooth downfield shift from 21.4 ppm
(pH N 7.7) to 30.2 ppm (pH b 0.6). Phosphonic acids in general experi-
ence an upfield shift upon conversion to phosphonate salts (Quin and
Williams, 2004), consistent with our observations. The results therefore
provide good evidence that Eq. (2) is the more appropriate representa-
tion of the reactive acid extraction (i.e. conversion from quaternary am-
monium phosphonate salt to a mixture of phosphonic acid and
quaternary ammonium chloride salt).
3.2. Acid-base equilibrium in polar medium
The acid-base behaviour of R4N+ EHEHP− was compared with the
EHEHPA in a polar medium (75% ethanol in water). As can be seen in
Fig. 2, the titration curves for the two extractants are virtually identical,
with the only difference being that for R4N+ EHEHP− the phosphonic
acid is initially fully deprotonated, whereas for EHEHPA it is initially
fully protonated. The titration curves can be modelled as simple proton-
ation/deprotonation (Eqs. (3) and (4)) with pKa 4.4, in reasonable
agreement with the literature for this extractant in 75% ethanol/water
medium (pKa 4.85) (Kolarik, 2010). The modelled curves are shown
Fig. 1. Impact of pH on chloride/acid uptake and 31P{1H} NMR chemical shift of
R4N+ EHEHP−. (Aqueous phase: 1 M NaCl/HCl solutions. Organic phase: 0.25 M
R4N+ EHEHP− in toluene.)
in Fig. 2.
HA⇌Hþ þ A− ð3Þ
 
½A−  Hþ
Ka ¼ ð4Þ
½HA
These results show that when dissolved in a polar medium, the qua-
ternary ammonium ion has no effect on the acid/base chemistry of the
extractant is dissolved in a low polarity diluent and is typically more
concentrated than the 0.015 M solutions examined here. This is ex-
plored in Section 3.3.
3.3. Two-phase acid base equilibrium
A comparison of the two-phase acid-base behaviour of 0.25 M
EHEHPA and 0.25 M R4N+ EHEHP− in toluene with 1.0 M NaCl in the
aqueous phase is shown in Fig. 3. The results for EHEHPA show a similar
curve to that in Section 3.2, except that the reaction occurs at around
pH 7. The curve can be described as formation of the sodium phospho-
nate salt Na-A, allowing for formation of dimer H2A2 in the organic
phase (Eqs. (5)–(8)). The formation of associates such as NaA[HA]3 de-
scribed by other authors as occurring under low loading conditions (Fu
et al., 1990; Kojima et al., 1969; Komasawa et al., 1983), was not includ-
ed in this analysis, as under high loading the limiting species is Na-A.
The fitted parameter for logKex_Na is −4.4, assuming a literature value
for logK2 (in benzene) of 5.8 (Kolarik, 2010). Note that concentrations
have been used in this analysis and hereafter on the assumption that ac-
tivity coefficients are approximately constant at constant ionic strength.
 
2 HA ⇌ H2 A2 ð5Þ
h
 i
H2 A2
K2 ¼ h i2 ð6Þ
HA
 
HA þNaþ ⇌ NaA þ Hþ ð7Þ
hih i
NaA Hþ
Kex Na ¼h ih i ð8Þ
Naþ HA
The results for R4N+ EHEHP− are markedly different. The acid ex-
traction takes place at lower pH, and is considerably broader, occurring
over approximately 8 pH units. In toluene, the free quaternary ammoni-
um and phosphonate ions are not as well solvated as in the higher polar-
ity water/ethanol phase. Therefore, the association between the
quaternary ammonium and phosphonate ions is expected to be com-
paratively stronger (Cox, 2013). This explains why there is no detect-
able difference in the polar-phase titration curves between EHEHPA
and R4N+ EHEHP−, but a significant difference in the two-phase titra-
tions. The calculated protonation of the phosphonic acid as a function
of pH was the same regardless of the direction of titration, which indi-
cates that the reaction is fully reversible.
The curve for 0.25 M R4N+ EHEHP− using Shellsol 2046 as diluent is
quite consistent with the equivalent curve using toluene, but shifted by
approximately 0.2 pH units (results also provided in Fig. 3). Shellsol
2046 is a more attractive diluent for industrial applications due to the
lower aromatic content and cost, and these results suggest that the
acid-base properties of R4N+ EHEHP− in this primarily aliphatic diluent
are quite similar to toluene.
 J.E. Quinn et al. / Hydrometallurgy 169 (2017) 306–313 309

Fig. 2. Single phase potentiometric titration curve of EHEHPA and R4N+ EHEHP− in 75% ethanol/water. (1.45 mmol (EHEHPA) or 1.49 mmol (R4N+ EHEHP−) in 0.1 L 75% ethanol/water.
Titrant: 1.0 M HCl (Protonation) or 1.0 M NaOH (Deprotonation). Lines represent fitted curves using Eqs. (3) and (4).)

3.4. Comparison of R4N+ EHEHP− with mixtures of R4N+ Cl− and EHEHPA
The extraction of acid by 0.25 M R4N+ EHEHP− and various mixtures
of EHEHPA and R4N+ Cl− is shown in Fig. 4. The extraction of acid by
0.25 M EHEHPA +0.25 M R4N+ Cl− as a function of pH is identical to
that of the ionic liquid, with the only difference being the pH after first
equilibrating organic phases with neutral NaCl solution (1.7 for EHEHPA
+ R4N+ Cl− and 9.0 for R4N+ EHEHP−). This indicates that in the simple
mixture, the phosphonic acid is initially protonated, compared with
R4N+ EHEHP− which is fully deprotonated. These observations seemingly
contrast with the assertion by Chen et al. (2013) that the ionic liquid ex-
tractant is different from the simple mixture of EHEHPA/R4N+ Cl−. Once
the ionic liquid is equilibrated with an aqueous phase at a certain pH
(as would be the case in any solvent extraction circuit), it is indistinguish-
able, in an acid-base sense, from a mixture of R4N+ Cl− and EHEHPA at the
same pH.
Also shown in Fig. 4 is the effect of varying the ratios of R4N+ Cl− and
EHEHPA. When EHEHPA is in excess of R4N+ Cl−, the proportion of total
phosphonate groups in excess of the quaternary ammonium ions proton-
ates/deprotonates sharply at around pH 6–8, as noted for EHEHPA alone
(Section 3.3). The remaining phosphonate ions (i.e. the amount equal to
the quaternary ammonium ion concentration) undergo the broad conver-
sion at lower pH, consistent with results for R4N+ EHEHP−. In this case the
reaction can be viewed as a combination of two processes: protonation of
the sodium salt of EHEHPA at high pH (Eqs. (7)–(8)), and the broad pro-
tonation of the R4N+ EHEHP−. When the R4N+ Cl− concentration is
increased further to be in excess of EHEHPA, the protonation remains
broad and shifts to lower pH.
3.5. Impact of chloride and extractant concentration
The effect of chloride on the extraction of acid by 0.25 M R4N+ -
EHEHP− is shown in Fig. 5. The results show that extraction of acid is
favoured by increasing chloride concentrations (i.e. it is shifted to higher
pH values). This can be explained by an increased tendency for Aliquat
336 to form an ion pair with chloride rather than the alkyl phosphonate,
which subsequently allows the phosphonate to protonate at higher pH
as the chloride concentration is increased. The results shown in Fig. 6 il-
lustrate that the proportion of the acid protonated as a function of pH is
independent of the overall extractant concentration between 0.1 and
0.5 M R4N+ EHEHP−. The extractant is approximately 50% protonated
at pH 5.0 regardless of the extractant concentration. Both of these obser-
vations are consistent with Eyal's results for ABC extractants (Eyal,
1997).
3.6. Mechanism of acid extraction by R4N+ EHEHP−
The protonation and deprotonation of EHEHPA can be
described using Eqs. (5)–(8), as described in Section 3.3. This fit is
shown in Fig. 7a. The results from previous sections, and a review of
the available literature, suggests that the extraction of acid by R4N+ -
EHEHP− involves protonation of the phosphonate ion and destruction

Fig. 3. Calculated protonation of extractant from two phase potentiometric titrations of R4N+ EHEHP− and EHEHPA. (Aqueous phase: 1 M NaCl. Organic phase: 0.25 M R4N+ EHEHP− or
0.25 M EHEHPA in toluene or Shellsol 2046 as indicated. Titrant: 1.0 M NaOH or 5.2 M HCl. Protonation calculated from experimental results as described in Section 2.4.2. Solid line
represents the fitted curve for EHEHPA protonation/deprotonation.)
310 J.E. Quinn et al. / Hydrometallurgy 169 (2017) 306–313

Fig. 4. Calculated protonation of R4N+ EHEHP− and mixtures of R4N+ Cl− and EHEHPA of two-phase titrations of with simple mixtures of R4N+ Cl− and EHEHPA. (Aqueous phase: 1 M
NaCl. Organic phase: As indicated in figure, all prepared in toluene. Titrant: 1.0 M NaOH or 5.2 M HCl. Protonation calculated from experimental results as described in Section 2.4.2.)

of the R4N+ EHEHP− ion pair. The simplest extension to the model for pH can therefore be estimated from the measured chemical shift using
EHEHPA alone is to allow an additional reaction to describe the decom- Eq. (11):
position of R4N+ EHEHP− to the separate quaternary ammonium chlo-
ride and acid forms (Eqs. (9)–(10)) (Belova and Khol'kin, 2015; Eyal, δobs ðpHÞ−δb
%ProtonationðpHÞ ¼  100 ð11Þ
1997): δa −δb

   The measured NMR shifts shown in Fig. 1 were substituted into Eq.
 −
HA þ R4 Nþ Cl ⇌ R4 Nþ EHEHP− þ Hþ þ Cl ð9Þ (11) to estimate the degree of protonation at each pH value. These re-
sults were also plotted on Fig. 7b. The percentage protonation indicated

  h ih i by the 31P{1H} NMR resonances match reasonably well with the
−
R4 Nþ EHEHP− Hþ Cl modelled protonation using Eqs. (5)–(10), and is considerably sharper
KR4NA ¼ h i
h i ð10Þ than the observed acid extraction. It is therefore conceivable that
−
HA R4 Nþ Cl
there is another mechanism responsible for acid extraction in addition
to that described by Eq. (9); that is significant in the region between
The modelled curves using this approach, with the same constants about pH 4 and 9. One possibility is that the quaternary ammonium
for logKex_Na and logK2 for EHEHPA alone are shown in Fig. 7b. While ion could ion pair with a partially deprotonated dimer (R4N+HA− 2 ). In
this simple model correctly predicts the effect of adding R4N+ and Cl− this case there would be another reaction (Eqs. (12)–(13)) to describe
on the shift in pH of the protonation reaction, it does not explain the ob- the system along with the previous ones (Eqs. (5)–(10)).
served broadening of the protonation reaction in the presence of R4N+,
  
− −
which occurs over approximately 8 pH units. H2 A2 þ R4 Nþ Cl ⇌ R4 Nþ HA2 − þ Hþ þ Cl ð12Þ
If the 31P{1H} NMR shifts observed in Fig. 1 are indicative of the pro-
tonation of the phosphonic acid, then the protonation percentage can be  h ih i
−
estimated as a function of aqueous pH. Taking the signal at 30.2 ppm as R4 Nþ A− Hþ Cl
the resonance of the protonated phosphonic acid (δa) and 21.4 ppm as KR4NHA2 ¼ h ihi ð13Þ
−
H2 A2 R4 Nþ Cl
the deprotonated acid (δb), then the observed resonance of a partially
deprotonated acid (δobs) would be a weighted average of the signals
of the two fast exchanging species (Macomber, 1992). The proportion The broad curves evident in the acid uptake curves from previ-
of protonated species, and hence the degree of protonation at a given ous sections can be generated using this system of equations (Fig.

Fig. 5. Impact of aqueous chloride concentration on R4N+ EHEHP− protonation. (Aqueous phase: 0.5–3 M NaCl. Organic phase: 0.25 M R4N+ EHEHP− in toluene. Titrant: 1.0 M NaOH or
5.2 M HCl. Protonation calculated from experimental results as described in Section 2.4.2.)
 J.E. Quinn et al. / Hydrometallurgy 169 (2017) 306–313 311

Fig. 6. Impact of extractant concentration on protonation of R4N+ EHEHP− (Aqueous phase: 1 M NaCl. Organic phase: As indicated in figure, all prepared in toluene. Titrant: 5.2 M HCl.
Protonation calculated from experimental results as described in Section 2.4.2.)

7c). However the modelled curves show an additional inflexion
point (not evident in our experimental results) as the system is
predicted to convert from predominantly H 2 A 2 + R 4 N + Cl − to
R4N+HA− + −
2 and finally to R4 N EHEHP . In addition, it was not pos-
sible to obtain a good fit using the same constants when EHEHPA
was in excess of R4N+ Cl− (Fig. 7d).
As an alternative to the presence of the partially deprotonated dimer,
it is also possible that the ion pairing interactions between the various
charged organic species are more complex than those represented in
Eqs. (9)–(12). Nanometre scale structures such as micelles and fila-
ments have been observed in mixtures of ionic liquids and low polarity
molecular solvents, as the polar sections of the ionic associates orient
away from the low-polarity solvent. It is only at very low concentrations
of ionic liquids that single ion pairing is observed (Gomes et al., 2010).
Eyal (1997) also notes that organic aggregates are favoured for non-
polar diluents. The use of single, concentration based equilibrium con-
stants with no allowance for activity effects to account for quite complex
structural changes in the organic phase may simply not be valid for this

Fig. 7. Modelling of the extraction of acid by EHEHPA and R4N+ EHEHP−. (Experimental data and conditions as per Figs. 3 and 4. Model parameters: logKex_Na = −4.4, logK2 = 5.8,
logKR4NA = −1.4 (Figs. a, c and d) or 0.5 (Fig. b), logKR4NHA2 = −2.6. Protonation calculated from experimental results as described in Section 2.4.2.)
312 J.E. Quinn et al. / Hydrometallurgy 169 (2017) 306–313
reaction. More advanced analytical tools may be required to determine
the exact mechanism of acid extraction in the high pH range.
Regardless of the exact mechanism, the main implications of these
results are that:
1) The phosphonate ion is partially protonated in the pH region rele-
vant to industrial rare earth separation circuits (pH 1–5), and
2) There is a strong interaction between the phosphonate and quaterna-
ry ammonium ions when dissolved in a low-polarity organic phase.
3.7. Recycling of extractant
For an industrial rare earth separation circuit, recycling of the sol-
vent through the process is essential for favourable process economics.
In the conventional process, the extracted rare earths are removed from
EHEHPA (stripped) using HCl solution to reprotonate the solvent. The
stripped solvent is then partially ‘saponified’ (i.e. neutralised) using ei-
ther NaOH or ammonia prior to recycling back to the extraction stages
(Eq. (14)). There, the sodium or ammonium ion is exchanged for rare
earth cations (Liao et al., 2013).
 
þ −
HA þ NaOH ⇌ Na A þ H2 O
+ −
If R 4 N EHEHP was to be used for industrial RE separations,
then it is likely that the phosphonate ion would be protonated
after stripping with HCl. In order to reuse the extractant, it would
require deprotonation prior to extraction, in much the same way
as for EHEHPA. The primary difference between the two
extractants is that instead of exchanging Na + or NH + +
4 for H , for
+ −
R4N EHEHP the protonated phosphonic acid would deprotonate
and associate with the quaternary ammonium ion, releasing Cl −
into the aqueous phase (Eq. ((15)).
   London, I.M., Goode, J.R., Moldoveanu, G., Rayat, M.S. (Eds.), 52nd Conference of Met-
−
HA þ R4 Nþ Cl þ NaOH ⇌ R4 Nþ EHEHP− þ NaCl þ H2 O
To demonstrate that the protonation/deprotonation reaction is re-
versible, a 0.25 M solution of R4N+ EHEHP− in toluene was first acidified
using HCl, then deprotonated with NaOH to pH 10.5 and water washed.
The chloride content and 31P{1H} NMR shifts of the extractant before
and after contacting with HCl and NaOH (Table 1) confirm that the ex-
tractant can be cycled between the different forms by contacting with
acidic and alkaline aqueous phases. The extractant could therefore po-
tentially be used industrially in the same way as EHEHPA. One advan-
tage of R4N+ EHEHP− over EHEHPA is that the deprotonated form is
fully miscible with the organic phase in both the diluents examined in
this work (toluene and Shellsol 2046). This is not the case for the sodi-
um salt of EHEHPA, which forms a microemulsion third phase in ali-
phatic diluents (Zhengshui et al., 1995).
4. Conclusions:
The extraction of hydrochloric acid by a bifunctional ionic liquid ex-
tractant, containing the quaternary ammonium cation of Aliquat 336
and the alkyl phosphonate anion of EHEHPA (R4N+ EHEHP−), was
Table 1
Regeneration Experiment. (Aqueous phase: as indicated. Organic phase: 0.25 M R4N+ -
EHEHP− in toluene. Phase ratio: 1:1.)
31
P{1H} NMR chemical shift (ppm) [Cl−] in organic (M)
Step 1: H2O 21.4 0.0094
Step 2: 1 M HCl 30.3 0.27
Step 3: H2O + NaOH 21.4 0.021
to pH 10.5
Step 4: H2O 21.4 0.018
explored. When dissolved in low concentration in a polar solvent, the
acid/base properties of the extractant are indistinguishable from that
of EHEHPA. However, when dissolved in higher concentrations (0.1 to
0.5 M) in toluene, the quaternary ammonium ion interferes with the
protonation/deprotonation process, broadening the two phase titration
curves and shifting the reaction to lower pH. The observed results are
consistent with ion pairing between the quaternary ammonium and
phosphonate ions. Increasing chloride concentrations favour the pro-
tonated form, which can be viewed as competition between phospho-
nate and chloride ions for the available quaternary ammonium ion.
Our results indicate that the combination of Aliquat 336 and EHEHPA
behaves as a mixture of R4N+ EHEHP− and R4N+ Cl− + EHEHPA, with the
protonation dependent on the pH, and not as a different chemical form.
Hence, regardless of how the ionic liquid has been prepared, once it is ex-
posed to acid containing solutions, the acid-base behaviour of the ionic
liquid is indistinguishable from a mixture of the two reagents. The use
of R4N+ EHEHP− for rare earth separations and its performance in com-
parison with EHEHPA is explored in Part 2 of this study.
Acknowledgements
ð14Þ
The authors wish to acknowledge Nigel Lengkeek (Australian
Nuclear Science & Technology Organisation) for assistance with
NMR experiments.
References
Belova, V.V., Khol'kin, A.I., 2015. Extraction of mineral acids and lanthanum salts with bi-
nary extractants. Russ. J. Inorg. Chem. 60 (9), 1157–1162.
Belova, V.V., Voshkin, A.A., Kholkin, A.I., Payrtman, A.K., 2009. Solvent extraction of some
lanthanides from chloride and nitrate solutions by binary extractants. Hydrometal-
lurgy 97 (3–4), 198–203.
Chen, J., Guo, L., Deng, Y., Li, D., 2013. Application of [A336][P507]/[P204] on high selective
extraction and separation of rare earths (III) from mechanism to techniques. In:
ð15Þ allurgists (COM). Canadian Institute of Mining, Metallurgy and Petroleum, Montreal,
Canada, pp. 367–374.
Cox, B.G., 2013. Acids and Bases: Solvent Effects on Acid-Base Strength. Oxford University
Press, Oxford.
Dietz, M.L., 2006. Ionic liquids as extraction solvents: where do we stand? Sep. Sci.
Technol. 41 (10), 2047–2063.
Eyal, A.M., 1997. Acid extraction by acid-base-coupled extractants. In: Marinsky, J.A.,
Marcus, Y. (Eds.), Ion Exchange and Solvent Extraction: A Series of Advances. Marcel
Dekker, New York, pp. 31–89.
Fu, X., Hu, Z., Liu, Y., Golding, J.A., 1990. Extraction of sodium in bis (2, 4,4 -
trimethylpentyl) phosphinic acid “Cyanex 272”™: basic constants and extraction
equilibria. Solvent Extr. Ion Exc. 8 (4–5), 573–595.
Gomes, M.F.C., Canongia Lopes, J.N., Padua, A.A.H., 2010. Thermodynamics and Micro Het-
erogeneity of Ionic Liquids, Ionic Liquids (Topics in Current Chemistry). Springer-
Verlag, Berlin, Heidelberg.
Han, X., Armstrong, D.W., 2007. Ionic liquids in separations. Acc. Chem. Res. 40 (11),
1079–1086.
Kalyakin, S.N., Kuz'min, V.I., 2000. Binary extracting agents as new stationary phases for
the extraction chromatography of ionic compounds. Mendeleev Communications
10 (4), 151–152.
Kojima, I., Fukuta, J., Tanaka, M., 1969. Extraction of copper(II) and sodium with di-(2-
ethylhexyl)- phosphoric acid. J. Inorg. Nucl. Chem. 31 (6), 1815–1820.
Kolarik, Z., 2010. Review: dissociation, self-association, and partition of monoacidic or-
ganophosphorus extractants. Solvent Extr. Ion Exc. 28 (6), 707–763.
Komasawa, I., Otake, T., Hattori, I., 1983. Extraction of nickel and cobalt with 2-ethylhexyl-
phosphonic acid mono-2-ethylhexyl ester. J. Chem. Eng. Jpn 16 (3), 210–217.
Liao, C., Wu, S., Cheng, F., Wang, S., Liu, Y., Zhang, B., Yan, C., 2013. Clean separation tech-
nologies of rare earth resources in China. J. Rare Earths 31 (4), 331–336.
Macomber, R.S., 1992. An introduction to NMR titration for studying rapid reversible com-
plexation. J. Chem. Educ. 69 (5), 375.
Marie, C., Hiscox, B., Nash, K.L., 2012. Characterization of HDEHP-lanthanide complexes
formed in a non-polar organic phase using 31P NMR and ESI-MS. Dalton Trans. 41
(3), 1054–1064.
Marsh, K.N., Boxall, J.A., Lichtenthaler, R., 2004. Room temperature ionic liquids and their
mixtures—a review. Fluid Phase Equilib. 219 (1), 93–98.
Quin, L.D., Williams, A.J., 2004. Practical Interpretation of P-31 NMR Spectra and Comput-
er-Assisted Structure Verification. Advanced Chemistry Development, Toronto,
Canada.
Ritcey, G.M., 2006. Solvent Extraction, Principles and Applications to Process Metallurgy.
Vol. 2. G.M. Ritcey and Associates Inc., Ottawa.
Rout, A., Venkatesan, K.A., Srinivasan, T.G., Vasudeva Rao, P.R., 2012. Ionic liquid
extractants in molecular diluents: extraction behavior of europium (III) in
quarternary ammonium-based ionic liquids. Sep. Purif. Technol. 95, 26–31.
 J.E. Quinn et al. / Hydrometallurgy 169 (2017) 306–313
Sun, X., Ji, Y., Hu, F., He, B., Chen, J., Li, D., 2010a. The inner synergistic effect of bifunctional
ionic liquid extractant for solvent extraction. Talanta 81 (4–5), 1877–1883.
Sun, X., Ji, Y., Liu, Y., Chen, J., Li, D., 2010b. An engineering-purpose preparation strategy
for ammonium-type ionic liquid with high purity. AICHE J. 56 (4), 989–996.
Yang, H., Wang, W., Cui, H., Zhang, D., Liu, Y., Chen, J., 2012. Recovery of rare earth ele-
ments from simulated fluorescent powder using bifunctional ionic liquid extractants
(Bif-ILEs). J. Chem. Technol. Biotechnol. 87 (2), 198–205.
Yang, H., Chen, J., Cui, H., Wang, W., Chen, L., Liu, Y., 2016. Application of ionic liquid
extractants on rare earth green separation. In: Chen, J. (Ed.), Application of Ionic
313
Liquids on Rare Earth Green Separation and Utilization. Green Chemistry and Sustain-
able Technology. Springer-Verlag, Berlin Heidelberg.
Zhao, H., Xia, S., Ma, P., 2005. Use of ionic liquids as ‘green’ solvents for extractions.
J. Chem. Technol. Biotechnol. 80 (10), 1089–1096.
Zhengshui, H., Ying, P., Wanwu, M., Xun, F., 1995. Purification of organophosphorus acid
extractants. Solvent Extr. Ion Exc. 13 (5), 965–976.