Cadmium and iron removal from phosphoric acid using

commercial resins for purification purpose

Mohamed Taha, Ahmed Masoud, Yasser Khawassek, Ahmed Hussein, Hisham
Aly, Eric Guibal

To cite this version:

Mohamed Taha, Ahmed Masoud, Yasser Khawassek, Ahmed Hussein, Hisham Aly, et al.. Cadmium
and iron removal from phosphoric acid using commercial resins for purification purpose. Environmen-
tal Science and Pollution Research, 2020, 27, pp.31278-31288. �10.1007/s11356-020-09342-7�. �hal-
02769170�

HAL Id: hal-02769170

https://imt-mines-ales.hal.science/hal-02769170
Submitted on 20 Nov 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Cadmium and iron removal from phosphoric acid using commercial
resins for purification purpose
Mohamed H. Taha 1 & Ahmed M. Masoud 1 & Yasser M. Khawassek 1 & Ahmed E. M. Hussein 1 & Hisham F. Aly 2 &
Eric Guibal 3

Abstract
Three commercial resins bearing sulfonic, amino phosphonic, or phosphonic/sulfonic reactive groups have been tested for the removal
of iron and cadmium from phosphoric acid solutions. The sorption properties are compared for different experimental conditions such
as sorbent dosage (0.5–2.5 g L−1), phosphoric acid concentration (from bi-component solutions, 0.25–2 M), and metal concentrations
(i.e., in the range 0.27–2.7 mmol Cd L−1 and 0.54 mmol Fe L−1) with a special attention paid to the impact of the type of reactive
groups held on the resins. The sulfonic-based resin (MTC1600H) is more selective for Cd (against Fe), especially at high phosphoric
acid concentration and low sorbent dosage, while MTS9500 (aminophosphonic resin) is more selective for Fe removal (regardless of
acid concentration and sorbent dosage). Equilibrium is reached within 2–4 h. The resins can be ranked in terms of cumulative sorption
capacities according the series: MTC1600H > MTS9570 > MTS 9500. Sulfuric acid (0.5–1 M) can be efficiently used for the
desorption of both iron and cadmium for MTC1600H, while for MTS9570 (phosphonic/sulfonic resin) sulfuric acid correctly desorbs
Cd (above 96% at 1 M concentration), contrary to Fe (less than 30%). The aminophosphonic resin shows much poorer efficiency in
terms of desorption. The sulfonic resin (i.e., MTC1600H) shows much higher sorption capacity, better selectivity, comparable uptake
kinetics (about 2 h equilibrium time), and better metal desorption ability (higher than 98% with 1 M acid concentration, regardless of
the type of acid). This conclusion is confirmed by the comparison of removal properties in the treatment of different types of industrial
phosphoric acid solutions (crude, and pre-treated H3PO4 solutions). The three resins are inefficient for the treatment of crude
phosphoric acid, and activated charcoal pre-treatment (MTC1600H reduced cadmium content by 77%). However, MTC1600H allows
removing over 93% of Fe and Cd for H3PO4 pre-treated by TBP solvent extraction, while the others show much lower efficiencies (<
53%).

Keywords Phosphoric acid purification . Metal removal . Sorption isotherms . Uptake kinetics . Effect of reactive groups

but also as ingredient for food, pharmacy, cosmetics, and elec-

Introduction
tronics. The manufacturing of phosphoric acid may be operated
through two main processes: (a) wet way or (b) thermal way.
Phosphoric acid is widely used in industry for applications such
The thermal processing usually limits the presence of metal
as fertilizer, metal surface cleaner (rust removal), and catalyst,
impurities, but the technique is highly expensive, nevertheless
wet processing is the most frequently used at industrial scale. Its
main drawback consists of the presence of relatively high con-
centrations of hazardous metals (such as uranium, cadmium,
mercury, arsenic, and other heavy metals) or other organic (hu-
mic acid, for example) and inorganic substances (such as sul-
* Eric Guibal fate) (El Zrelli et al. 2018). The requirements and regulations
[email protected] for phosphoric acid grade are of increasing complexity when
1
target application concerns fertilizers < food < pharmacy <
Nuclear Materials Authority, P.O. Box 530, El Maddi, Cairo, Egypt
electronics. The quote for the different grades of H3PO4 is log-
2
Egyptian Atomic Energy Authority, P.O. Box 13759, Cairo, Egypt ically following the same scale of purity requirements.
3
Polymers Composites and Hybrids (PCH), IMT Mines Ales, 6 For preparation of technical phosphoric acid, the wet pro-
avenue de Clavières, 30319 Alès cedex, France cess usually consists of the following (EFMA 2000): (a)
phosphate rock grinding, (b) sulfuric acid leaching, (c) filtra- Puromet MTC1600H bears sulfonic acid reactive groups,
tion (to remove calcium sulfate di-hydrate, or phosphogyp- aminophosphonic reactive groups are supported on Puromet
sum), and (d) concentrating step (produced diluted liquor is MTS9500, while Puromet MTS9570 is a bi-functional resin
concentrated by evaporation to reach 50% w/w P2O5 merchant constituted of phosphonic and sulfonic acid groups. The study
grade acid, MGA). Phosphogypsum residue that may contain compares the effects of these different reactive groups on the
substantial amounts of valuable metals, including rare earths, removal of cadmium and iron from phosphoric acid through
constitutes one of the most important by-products of the wet the evaluation of the influence of H3PO4 concentration, sor-
process. bent dosage, uptake kinetics and sorption isotherms from bi-
The origin of phosphate rock is thus of critical importance component synthetic solutions; metal desorption is also tested
since some areas are reputed for the presence of high concen- using acid solutions (HNO3, HCl, H2SO4). In a second part of
trations of these hazardous metals that decrease substantially the work, the resins are applied to the treatment of an industrial
the commercial value of produced phosphoric acid. In addi- phosphoric acid solution produced by Abu Zaabal Fertilizer
tion, the high concentration of these impurities affect nega- and Chemicals Company (AZFC) in Egypt. Metal recovery is
tively the filtration of phosphoric acid, and the solubility of the processed, as a polishing treatment, on solutions pre-treated
produced fertilizers (in the case of iron) (Abdel-Ghafar et al. with either activated carbon or solvent extraction (using tri-
2019) and have a harmful impact on the environment as well butyl phosphate extractant, TBP). TBP has been used for the
as the human health (in the case of cadmium) (Roberts 2014). purification of phosphoric acid for fluoride removal, metal
Several processes have been designed for the purification of recovery and acid separation (phosphoric acid vs. sulfuric ac-
phosphoric acid including precipitation (Abdel-Ghafar et al. id) (Lembrikov et al. 2004; Li et al. 2017a; Liu et al. 2016;
2019; Kouzbour et al. 2019), solvent extraction (Ahmed et al. Zuo et al. 2019).
2007; Ahmed et al. 2019; Chen et al. 2018; Li et al. 2017a; Li
et al. 2017b; Sanghani 2014; Ye and Li 2013), and extractant-
impregnated resins (Reyes et al. 2001). However, in the case Materials and methods
of solutions containing low levels of metal ions (below a few
hundred mg L−1), sorption processes may be interesting and Materials
competitive. Activated carbon and clays are frequently used
for the purification of phosphoric acid from organic impurities Selected resins were supplied by Purolite (King of Prussia,
(Nasr et al. 2005; Trabelsi and Tlili 2017) or metal ions (El- PA, USA). They were used as supplied. Table AM1 (see
Bayaa et al. 2011; Kussainova et al. 2019; Kussainova et al. Additional Material Section) summarizes the main properties
2016; Tjioe et al. 1988), including metal-loaded clays (Hamza of the resins (commercial data). For the study of synthetic
et al. 2016). However, sorbents can be also operated for metal solutions, phosphoric acid (analytical grade) was purchased
removal from crude phosphoric acid (Alexandratos and Zhu from Merck KGaA (Darmstadt, Germany). Other acids,
2016; El-Bayaa et al. 2011; Heres et al. 2018; Kouzbour et al. namely nitric, hydrochloric, and sulfuric acids (analytical
2019; Zhu and Alexandratos 2015). grade), used for desorption tests, were supplied by Sigma-
Metal recovery from phosphate rock processing is thus Aldrich (Saint-Louis, MO, USA). Cadmium (CdSO4) and
considered a way to limit environmental and health issues iron (FeSO4·7H2O) salts were obtained from Merck KGaA.
associated with application of metal-loaded phosphoric acid, Stock solutions of 1.0 g metal L−1 were prepared in Milli-Q
and to improve the cost of valorization of processed phospho- water (acidified by a few drops of nitric acid to prevent metal
ric acid. However, the treatment of both phosphoric acid and precipitation at long storage). Freshly prepared solutions for
phosphogypsum can also be regarded as an opportunity to experiments were obtained by dilution with phosphoric acid at
recover valuable and strategic metals (Khayambashi et al. the appropriate concentration prior to sorption tests. When
2016; Reddy and Kumar 2016; Virolainen et al. 2019). relevant, pH was measured using Cyber Scan pH 6000 pH-
Indeed, wastes and sub-products of industrial production and meter (Eutech Instruments, Nijkerk, Netherlands).
mining activities may be considered for valorization of sec-
ondary resources (Gomes et al. 2016; Islam et al. 2015). Sorption and desorption processes
Biopolymer composites (Abasiyan et al. 2019), clays (Lan
et al. 2019) biochars (Yin et al. 2019) have been successfully The sorption tests were carried out in batch reactor. Although
tested for cadmium sorption. the primary step in the design of a sorbent process is the study
In this work, a special attention is paid to the removal of of the pH effect on metal binding, in the current work, the
cadmium and iron from phosphoric acid solutions using three strategy focused on the development of preferred process
commercial resins (supplied by Purolite) bearing different re- based on unchanged pH (the only variation dealt with the
active groups. The three spherical resins are supported on effect of phosphoric acid dilution in the relevant section in-
macroporous polystyrene crosslinked with divinylbenzene. vestigating the effect of acid concentration). A fixed volume
(V, L) of solution containing 50 mg L−1 of both cadmium and Egypt (Yassien and Elameer 2014). El-Sebeaya plateau is
iron (C0, 0.44 mmol Cd L−1; 0.90 mmol Fe L−1) was mixed located between latitudes 25° 30′–26° 30′ and longitudes 32°
for 24 h with a fixed amount of resin (m, g) (agitation speed 30′–33° 30′ on both sides of the Nile Valley (Elmaadawy et al.
150 rpm). The concentration of phosphoric acid was varied 2015; NMA 2010). Abu Zaabal Fertilizer and Chemicals
between 0.25 M and 2.0 M. In most cases, the sorbent dosage Company (AZFC) produces phosphoric acid using the wet
(SD = m/V, g L−1) was set to 1.5 g L−1. When relevant, the di-hydrate process; phosphate rock is completely acidified
sorbent dosage was varied (in the range 0.5–2.5 g L−1 for the using concentrated sulfuric acid to produce di-hydrate phos-
study of the effect of sorbent dosage), as well as the initial phoric acid (~ 25% P2O5%) and insoluble di-hydrate calcium
concentration. For the study of sorption isotherms, the initial sulfate precipitate (phosphogypsum) (Shweikani et al. 2013),
concentration (C0) varied in the range 30–300 mg metal L−1, according Eq. 1:
0.27–2.7 mmol Cd L−1, and 0.53–5.3 mmol Fe L−1. The con-
centration ranges were selected to reach the saturation of the Ca10 F 2 ðPO4 Þ6 þ 10 H 2 SO4
resins and evaluate maximum sorption capacities (though þ 20 H 2 O →70−80°C 6 H 3 PO4 þ 10 CaSO4 2H 2 O
sorption is probably not competitive when metal concentra-
tion exceeds a few hundred mg L−1). For the uptake kinetics, þ 2 HF ð1Þ
the contact time was varied between 0 and 1440 min (1 day)
with samples collected at fixed contact times for building the Crude phosphoric acid supplied by AZFC was pre-
kinetic profiles. After filtration (filter membrane, pore size treated using two different processes: (a) sorption onto
1.2 μm), the residual metal concentration (Ceq) was analyzed activated carbon (case a), and (b) solvent extraction using
by ICP-AES (inductively coupled plasma atomic emission TBP (case b).
spectrometer, Activa M, Horiba-Jobin Yvon, Longjumeau, Activated charcoal (AC, Loba Chemie Pvt. Ltd., Mumbai,
France). The sorption efficiency was calculated as well as India) was mixed with crude phosphoric acid (≈ 22.5% w/w
the sorption capacity (qeq, mmol g−1) using the mass balance P2O5) in an agitated tank reactor for 24 h; the sorbent dosage
equation: qeq = (C0-Ceq) × V/m. was 5 g AC L−1 (NMA 2017). Filtrated pre-treated solution
For desorption investigation, similar experimental proce- was used for polishing treatment with ion-exchange resins.
dure was used. Metal loading was performed using the proce- For solvent extraction, tri-butyl phosphate (TBP, Loba
dure described above, and the effective amount of metal pres- Chemie Pvt. Ltd., Mumbai, India) was diluted in kerosene
ent on the resin was calculated from the mass balance equa- (extractant concentration: 2.55 M) and shaken with concen-
tion. Metal sorption was carried out by contact of the resins trated phosphoric acid (≈ 45% w/w P2O5) for 15 min at room
(SD 1.5 g L−1) with metal solutions containing both iron and temperature with a 4/1 org./aq. volume ratio. After phase sep-
cadmium (C0 50 mg L−1) in 0.25 M H3PO4 solution (pH0 1.5) aration, loaded organic phase was scrubbed with pure phos-
at room temperature for 24 h. Metal resins were washed with phoric acid (≈ 35% w/w P2O5) with a 1/1 org./aq. volume
demineralized water before processing the elution step. Bi- ratio, for 5 min. Finally, the scrubbed organic phase was
metal-loaded sorbents were mixed for 24 h with the eluent, stripped with double distilled water at room temperature, for
setting the sorbent dosage to 0.5 g L−1. After filtration, the 10 min; the org./aq. volume ratio was set to 1/1 (NMA 2018).
concentration in the eluate was analyzed to calculate the The eluted sample was used for polishing treatment with the
amount released, and to calculate the desorption efficiency. resin.
The modeling of uptake kinetics was performed using con- For the polishing treatment of pre-treated phosphoric acid
ventional equations: (a) pseudo-first order rate equation solutions, the solution was mixed with the resin for 24 h. The
(PFORE), (b) pseudo-second order rate equation (PSORE) sorbent dosage was 1.5 g L−1. After filtration, the residual
(Ho and McKay 1999; Spies and Wewers 2020; concentration of the metals was evaluated by ICP-AES, and
Wieszczycka et al. 2020), and (c) the Crank equation (associ- the mass balance equation was used for calculating the sorp-
ated with the resistance to intraparticle diffusion, assuming the tion efficiency and the sorption capacity. Fluoride concentra-
homogeneous diffusion of the sorbate at the surface of the tion was measured by a fluoride ion-selective electrode (Orion
sorbent, with constant surface diffusivity condition (Crank Star A215 pH/Conductivity Benchtop Meter, Thermo Fisher
1975)). The parameters were determined by non-linear regres- Scientific, Waltham, MA USA) coupled to an 8157BNUMD
sion analysis using the facilities of Mathematica ® software. Orion ROSS Ultra Triode pH/ATC electrode. The P2O5 con-
centration was determined using spectrophotometer-type
Application to treatment of industrial phosphoric acid Shimadzu UV 1208 by a specific colorimetric method (am-
solutions monium molybdate and ammonium metavanadate)
(Marczenko and Balcerzak 2000). The same procedure was
El-Sebaeya phosphate deposits of the Nile Valley are recog- applied to the treatment of crude phosphoric acid supplied by
nized among the most important sources of phosphate rock in AZFC.
Results and discussion the so-called soft acids (Pearson 1966), while Fe(III) is part of
hard acids and Fe(II) is assimilated to borderline acid. Based
Synthetic phosphoric acid solutions on the HSAB principle (hard and soft acid and base theory),
strong acids will preferentially react with strong bases (such as
Effect of phosphoric acid concentration N >> P; O >> S) (and reciprocal: weak acids with weak bases).
The introduction of sulfonic groups, which bring hydrophilic
The increasing concentration of phosphoric acid decreases properties, may affect the intraligand bonding, and, in turn, the
both the efficiency of cadmium and iron sorption, regardless affinity of the sorbent for target metal ions (Alexandratos and
of the resin (Fig. 1), and their relevant sorption capacities Smith 2004). Another important parameter may affect the af-
(Figure AM1, see Additional Material Section). However, finity of the resin for metals: the speciation of the metal ions in
the responses significantly differ for the systems resin/metal. complex solutions (including in phosphoric acid) may strong-
Table AM2 shows the relative variation of the sorption capac- ly influence the binding of metal species in solution onto re-
ities (compared with the qm at 0.25 M H3PO4 concentration): active groups (Figure AM2). Iron forms stable complexes
for cadmium sorption, MTC1600H sulfonic resin is less sen- with phosphate: the cationic complex FeH2PO4+ represents
sitive to acid concentration, while for iron removal most stable 91–96% of total Fe(II) in 0.25–2 M H3PO4 solutions (free
performances are obtained with MTS9500 aminophosphonic Fe2+ varies between 8.6 and 3.8%). On the opposite hand,
resin. On the opposite hand, high concentration of phosphoric cadmium is not forming stable species: in 0.25–2 M H3PO4
acid drastically reduces the sorption of cadmium on solutions, the fraction of CdHPO4 is negligible (less than
aminophosphonic resin and iron sorption is strongly depreci- 0.02%). Free Cd2+ species largely predominates (97.9–
ated by high concentrations for the sulfonic resin. The 99.4%), aqueous cadmium sulfate counts for 2% at low acidic
phosphonic/sulfonic resin (i.e., MTS9570) shows an interme- concentration and its fraction progressively decreases below
diary behavior; being sensitive to acid concentration for both 0.6% at 2 M H3PO4 concentration. The two metals form
cadmium and iron. predominating cationic species: monovalent FeH2PO4+ and
Alexandratos’ group documented the relative importance divalent free Cd2+ under selected experimental conditions.
of polarizability (Alexandratos and Zhu 2015), hydrogen The charge difference and the size of metal cations may influ-
bonding (Alexandratos and Zhu 2017; Zhu and ence the interaction mode with reactive groups (thermody-
Alexandratos 2011), and bi-functionality (Alexandratos and namics/equilibrium) but also the accessibility to reactive
Smith 2004) on the sorption performance of ligands (especial- groups (kinetics).
ly those bearing phosphonate and amidoxime groups) These different synergic or anti-synergic interactions/
(Alexandratos et al. 2016). The electronegativity of Cd (II) effects make complex the interpretation of the variations in
(i.e., χaq 2.660) is close to that of Fe (II) (i.e., χaq 2.636), while the affinity of the different resins for the two metals with
Fe(III) has a much higher χaq value (i.e., 3.835) (Li et al. increasing phosphoric acid concentration.
2012). Persson (Persson 2010) reported a Shannon radius for Figure AM3 shows the log-log plots of the distribution
hydrated cadmium (Cd (OH)62+) close to 0.95 Å, while iron ratios (D = qeq/Ceq, L g−1) with the concentration of phos-
hydrated radius was much lower (i.e., 0.78 Å for Fe (OH)62+, phoric acid for the different systems. Linear trends are
and 0.645 Å for Fe (OH)63+). Cadmium is classified among observed: the slope of this type of plot in usually associ-
ated, in ion-exchange processes, with the stoichiometry of
100 proton exchange. For Cd(II), the slopes vary between −
MTC1600H-Cd
MTS9500-Cd
0.87 (MTS9570 dual sulfonic/phosphonic resin) and −
80 MTS9570-Cd 1.44 (MTC1600H sulfonic); the aminophosphonic resin
Sorption efficiency (%)

MTC1600H-Fe
MTS9500-Fe (MTS9500) shows a slope close to − 1.1. The stoichio-
MTS9570-Fe metric ratio is thus close to 1. This is not consistent with
60
the predominance of the divalent cations (free Cd2+). For
40 iron, the slope of the linear plots varies much more: −
1.99 for MTC1600H resin, − 0.79 for MTS9570, and
20 down to − 0.22 for MTS9500. The resins bearing two
reactive groups show much lower slopes and lower dis-
0 tribution ratios for iron sorption; especially at low phos-
0 0.5 1 1.5 2 2.5
phoric acid concentration, the differences tend to level off
H3 PO4 (M)
in 2 M H3PO4 solutions. For Cd(II), the resins can be
Fig. 1 Effect of phosphoric acid concentration on Cd and Fe removal
using selected commercial resins (bi-component solutions, C0 50 mg L−1,
clearly ranged according the series: MTC1600H
i.e., 0.445 mmol Cd L−1 and 0.895 mmol Fe L−1; contact time 24 h; (sulfonic) > MTS9570 (phosphonic/sulfonic) > MTS9500
sorbent dosage (SD) 1.5 g L−1; T 20 ± 1 °C) (aminophosphonic).
100
Figure AM4 plots the selectivity coefficient, SCCd/Fe (i.e.,
SCCd/Fe = DCd/DFe) as a function of phosphoric acid concen-
80

Sorption efficiency (%)

tration for the three resins. The SC coefficient is less than 0.15
for MTS9500 aminophosphonic resin: the sorbent has a 60
MTC1600H-Cd
MTS9500-Cd
marked preference for Fe, especially at low H3PO4 concentra- MTS9570-Cd
tion. On the opposite hand, MTC1600H sulfonic resin shows MTC1600H-Fe
40 MTS9500-Fe
a preference for Cd that increases with acid concentration. The MTS9570-Fe
dual phosphonic/sulfonic resin (MTS9570) cannot separate 20
the two metals, whatever the concentration of the acid: the
SC varies between 0.44 and 1. The closer the SC value to 1, 0
the more difficult the separation of the metal ions on selected 0 0.5 1 1.5 2 2.5
resins. The resin has little preference for Fe at low phosphoric Sorbent dosage (g L-1 )

concentration, while at higher concentration the two metals Fig. 2 Effect of sorbent dosage on Cd and Fe removal using selected
ions are almost equally sorbed. commercial resins (bi-component solutions, C 0 50 mg L −1 , i.e.,
0.445 mmol Cd L−1 and 0.895 mmol Fe L−1; [H3PO4] 0.25 M; contact
time 24 h; sorbent dosage (SD) 1.5 g L−1; T 20 ± 1 °C)
Effect of sorbent dosage

As expected, increasing sorbent dosage improves the sorption Uptake kinetics

of both Cd and Fe, regardless of the sorbent. It is noteworthy
that the curves are almost overlapped for MTC1600H for the Under selected experimental conditions (SD, 1.5 g L−1;
two metal ions, and the sorption efficiency is significantly [H3PO4], 0.25 M; C0, 0.445 mmol Cd L−1 and
higher than for the other systems (resin/metal) (Fig. 2). For a 0.895 mmol Fe L−1; pH0 1.45), the equilibrium is reached
sorbent dosage of 2 g L−1, both cadmium and iron are within 2–8 h, depending on the system (Fig. 3). Faster kinetics
completely removed from 0.25 M H3PO4 solution. For the are achieved within 2 h for MTC1600H sulfonic resin for both
other resins, even at 2.5 g L−1 sorbent dosage, it is not possible cadmium and iron (the curves are overlapped for the two
achieving the complete removal of the metals. Iron is more metals). The kinetic profiles for cadmium and iron uptake
efficiently sorbed than cadmium and MTS9570 phosphonic/ are almost overlapped in the case of MTS9570 resin. MTS
sulfonic resin shows higher performance than MTS9500 9500 aminophosphonic resin shows a different trend with
aminophosphonic resin. similar profiles for the two metals but the equilibrium concen-
Table AM3 shows that the sulfonic resin (MTC1600H) trations are substantially different. This is consistent with pre-
allows much higher distribution ratios for both cadmium and vious observations on the preference of the sorbent for iron vs.
iron than the alternative resins. Highest D values are reached cadmium.
for SD 2.5 g L−1 for Cd (D: up to 46.4 L g−1) and SD 2.5 g L−1 The uptake kinetics can be controlled by the resistance to
for Fe (D: up to 43.5 L g−1). At the highest SD, the D values film diffusion, to intraparticle diffusion and by the proper
for Cd are about 230 to 30 times higher for MTC1600H than reaction rate. The PFORE, PSORE, and RIDE (crank equa-
for MTS9500 and MTS9570, respectively, while for Fe, the tion) have been fitted to kinetic profiles (Fig. 3 and
improvements are about 27 and 14 times, respectively.
Figure AM5 shows the effect of sorbent dosage on the 1
selectivity coefficient SCCd/Fe. These results confirm that PFORE
MTS9570 resin is poorly selective (SCCd/Fe varies between 0.8
0.35 and 0.58) and that MTS9500 resin has a preference for MTC1600H-Cd
Fe over Cd (SCCd/Fe varies between 0.09 and 0.14); for these 0.6 MTS9500-Cd
C(t)/C0

two resins, the sorbent dosage has a limited effect. The sul- MTS9570-Cd
fonic resin (MTC1600H) shows a preference for Cd over Fe at 0.4 MTC1600H-Fe
low sorbent dosage (SCCd/Fe is close to 1.6), while at higher MTS9500-Fe
sorbent dosage (in the range 1.5–2.5 L g-1) the SC remains 0.2 MTS9570-Fe
close to 1.08: the two metal ions are equally sorbed by the
sulfonic resin. At high sorbent dosage, the full sorption of both 0
0 240 480 720 960 1200 1440
Cd and Fe tends to limit the selectivity effect.
Time (min)
Therefore, the selection of experimental conditions (type
Fig. 3 Uptake kinetics for Cd and Fe removal using selected commercial
of sorbent and sorbent dosage) depends on the target: purifi- resins—modeling with the PFORE (bi-component solutions C 0
cation (large spectrum of metals to be removed) or recovery 50 mg L−1, i.e., 0.445 mmol Cd L−1 and 0.895 mmol Fe L−1; [H3PO4]
(selectivity is preferred). 0.25 M; sorbent dosage (SD) 1.5 g L−1; T 20 ± 1 °C)
Figure AM6). Figure AM7 also shows the fit of PFORE and presence of sulfonic groups is supposed to improve the hydro-
PSORE to uptake kinetics with the representation of sorption philic behavior of the resin and therefore water (and metal
capacity (i.e., q(t)) as a function of contact time. The parame- ions) transfer. Other parameters probably influence this mass
ters of the models are summarized on Table 1 (PFORE), and transfer resistance.
Tables AM4 and AM5 (PSORE and RIDE, respectively). The The predominant species (Cd2+ and FeH2PO4+) have sub-
comparison of the fitted curves with experimental profiles, stantially different ionic sizes in solution: the complexation of
and the determination coefficients show that the PFORE fits iron with dihydrogen phosphate means a larger ionic size for
better experimental data. In addition, the values of qeq,1 (cal- predominant iron species compared to free divalent cadmium
culated sorption capacities for PFORE) are very close to the species. However, the variations of the rate parameters (k1 and
experimental values of the equilibrium sorption capacities: the De) are not very marked.
variation never exceeds 3.5%. The PFORE is usually associ-
ated to physical sorption mechanism such as ion-exchange.
The apparent rate coefficients (k1, min−1) are of the same order Sorption isotherms
of magnitude for cadmium and iron, while the equilibrium
sorption capacities are significantly higher for iron than for The sorption isotherms in bi-component solutions (0.25 M
cadmium; this is consistent with the difference in molar con- phosphoric acid) appear in Fig. 4. The curves are character-
centration for the two metals. ized by a steep initial section that shows the affinity of the
Although the RIDE fits less accurately the profiles, the resins for cadmium and iron. The curves tend to reach a pla-
general trends are respected: the resistance to intraparticle dif- teau for residual concentrations in the range 0.5–
fusion contributes to control uptake kinetics. The model al- 1.5 mmol L−1. Unexpectedly, above this limit range of con-
lows determining the effective diffusivity (De, m2 min−1) for centration the sorption capacity tends to decrease. This behav-
cadmium and iron with the different resins (Table AM5). The ior is sometimes explained, in multi-component solutions, by
diffusivity coefficients are of the same order for cadmium and a difference in the affinity of the sorbent for a solute, which is
iron for MTC1600H: 0.45 × 10−10 and 0.38 × 10−10 m2 min−1, displacing at high concentration another solute initially bound
respectively. The same conclusion can be deduced for onto the sorbent. However, Figure AM8 shows the plots of the
MTS9570: 3.22 × 10−10 and 3.01 × 10−10 m2 min−1, respec- cumulative sorption capacity against the cumulative concen-
tively. The diffusivity is about an order of magnitude greater tration of the residual metal ions. This figure clearly shows
for the phosphonic/sulfonic resin (MTS9570) compared with that the same trend is observed on the cumulative isotherm: at
the sulfonic resin (MTC1600H). The diffusivity of cadmium high residual metal concentration, the sorption capacity tends
in MTS9500 aminophosphonic resin is almost doubled com- to decrease. The replacement of solutes is not sufficient for
pared with iron (i.e., 4.23 × 10−10 and 2.23 × 10−10 m2 min−1, maintaining the total sorption capacity at a constant plateau.
respectively). The order of magnitude is comparable for the In order to rank the resins, the sorption capacities are com-
two resins bearing phosphonic reactive groups. The self- pared for Ceq = 1 mmol metal L−1. In the case of Cd(II), the
diffusivity in water for both Cd 2+ and Fe 2+ is 4.23 × resins are ranked according: MTC1600H (0.61 mmol Cd g−1)
10−8 m2 min−1 (3.62 × 10−8 m2 min−1 for Fe3+, for counter- > MTS9570 (0.25) > MTS9500 (0.063). The same classifica-
anions such as HPO42− and H2PO4−, the self-diffusivities are: tion is observed in the case of Fe(II) but with higher sorption
4.55 × 10−8 m2 min−1 and 5.75 × 10−8 m2 min−1, respectively) capacities: MTC1600H (0.93 mmol Fe g−1) > MTS9570
(Buffle et al. 2007). This means that the diffusivity coeffi- (0.65) > MTS9500 (0.51). The predominant species is a
cients for the metal ions in the resins are two or three orders monovalent species for iron (i.e., FeH2PO4+) while divalent
of magnitude lower that their self-diffusivity in water. This free cadmium predominates for cadmium. The stoichiometric
confirms that the resistance to intraparticle diffusion cannot ratio between metal species and reactive groups at the surface
be neglected, especially for MTC1600 sulfonic resin. The of the resins is more favorable for iron.

Table 1 Modeling of Cd and Fe

uptake kinetics with the pseudo- Metal Cadmium Iron
first-order rate equation (bi-com-
ponent solutions) Resin MTC1600H MTS9500 MTS9570 MTC1600H MTS9500 MTS9570

qeq,exp (mmol g−1) 0.285 0.042 0.193 0.590 0.333 0.414

qeq,1 (mmol g−1) 0.284 0.041 0.189 0.582 0.331 0.408
k1 × 102 (min−1) 3.61 1.80 2.72 3.80 1.53 2.66
R2 0.995 0.971 0.990 0.994 0.985 0.995
1
MTC1600H-Cd MTC1600H-Fe conditions for metal desorption correspond to 1 M H2SO4
MTS9500-Cd MTS9500-Fe
0.8 MTS9570-Cd MTS9570-Fe
eluent with MTC1600H resin: metal is completely eluted.
For iron desorption, the resins shows more difficulties for
qeq (mmol g-1 )

0.6 releasing the metal. Iron is almost not desorbed from loaded
MTS9500 resin, whatever the concentration and the type of
0.4 acid: maximum desorption does not exceed 3%. For
MTS9570 resin, desorption is a little better but never exceeds
0.2 39%; HCl and H2SO4 solutions are much better than HNO3
solutions. These conditions are highly unfavorable: the pro-
0 gressive saturation of the resin with iron will progressively
0 1 2 3 4 5
poison and saturate the sorbents; this will limit the life cycle
Ceq (mmol L-1 )
of the material. The sulfonic resin (MTC1600H) shows much
Fig. 4 Sorption isotherms for Cd and Fe removal using selected
commercial resins (bi-component solutions, C0 30–300 mg L−1, i.e.,
more favorable desorption for iron. However, only sulfuric
0.27–2.7 mmol Cd L−1 and 0.534–5.41 mmol Fe L−1; [H3PO4] 0.25 M; acid at the highest concentration (i.e., 1 M) succeeds in achiev-
sorbent dosage (SD) 1.5 g L−1; T 20 ± 1 °C) ing the complete desorption of iron (i.e., desorption yield
reaches 98.8%).
The selectivity coefficient for cadmium over iron increases The difference in desorption of cadmium and iron may be
with total metal concentration for MTC1600H (around 2), associated to the probable effect of metal speciation.
while for MTS9570, it tends to decrease (the sorbent progres- Cadmium does not form complexes with phosphate and the
sively enhances the preferential binding of iron) resin binds the metal ion under the form of free cadmium or
(Figure AM9). MTS9500 resin maintains a preference for cadmium chloride cation. On the opposite hand, the predom-
iron, independently of total concentration. inance of iron phosphate species (strongly stable), bound to
Compared with alternative sorbents for the recovery of Cd the resins, makes iron desorption more complex.
(II), MTC16900H resin is of the same order of magnitude than
recently developed materials: Ha/Fe-Mn oxides-loaded bio- Industrial phosphoric acid solutions
char (i.e., 0.60 mmol Cd g−1) (Guo et al. 2019), amino-
functionalized core-shell magnetic mesoporous composites The complexity of the solution may strongly decrease sorption
(i.e., 0.46 mmol Cd g−1) (Tang et al. 2013) or higher than performances while compared with the levels of metal recov-
functionalized mineral resources (thiourea-zeolite-tuff, ery from synthetic solutions (Prelot et al. 2014). Therefore, the
0.103 mmol Cd g−1) (Barragan et al. 2017) or earthworm resins are tested for the direct treatment of crude phosphoric
manure-derived biochar (i.e., 0.27 mmol Cd g−1) (Wang acid (Table 3) and for pre-treated phosphoric acid solutions
et al. 2017). (Table 4). First, the comparison of the concentrations in the
initial solutions shows the significant impact of solvent extrac-
tion on the content of hazardous elements. As expected, acti-
Metal desorption vated charcoal is less efficient than solvent extraction for re-
moving heavy metals and fluoride (without cost analysis). The
Metal desorption is a critical parameter for the evaluation of grade of phosphoric acid is of the same order of magnitude
process feasibility at industrial scale. Table 2 reports the tests (around 21–23% w/w), regardless of the eventual pre-
performed on the desorption of cadmium and iron from metal- treatment.
loaded resins using three types of acid eluents (i.e., nitric acid,
hydrochloric acid, and sulfuric acid) at three levels of concen- Crude phosphoric acid
tration (i.e., 0.1, 0.5, and 1 M). The two metals have very
distinct behaviors in terms of desorption, and the resins show Crude phosphoric acid is loaded with very high levels of metal
different efficiencies for desorption. ions (but also probably some organic compounds, such as
For cadmium, the efficiency of desorption increases with humic and fulvic acids, that may interact with the speciation
the concentration of the acidic eluent (especially for of metal ions). This leads to strong saturation of the resins and
MTC1600H and MTS9570); in molar acidic solutions the the sorption efficiency is generally low. Table 3 shows that
desorption efficiency ranges between 86 and 100%. For fluoride concentration is insignificantly changed after contact
MTC1600H and 1 M acid concentration, the desorption yields with resins. Iron content is reduced by 27% and 18% using
range between 98 and 100%; this is little better than MTC1600H and MTS9570, while Fe concentration remains
MTS9570. For MTS9500, the concentration and the type of unchanged after treatment with MTS9500. For Cr (hard metal
acid hardly change the desorption performance for cadmium according HSAB principle), sorption efficiency varies be-
(varying between 36 and 52%). For cadmium, the best tween 19 and 25%. Manganese (initially present the
Table 2 Effect of type and
concentration of acids on metal Resin MTC1600H MTS9500 MTS9570
desorption efficiency (%) from
loaded resins (desorption Metal [Acid] (M) 1.0 0.5 0.1 1.0 0.5 0.1 1.0 0.5 0.1
experimental conditions: SD Cd Loading 31.9 mg Cd g−1 4.3 mg Cd g−1 22.1 mg Cd g−1
5 g L−1; contact time 24 h; T 20 ± HNO3 97.9 76.4 6.4 47.6 36.0 39.0 95.9 86.6 38.8
1 °C)
HCl 98.9 85.6 41.2 47.3 44.6 35.6 86.0 87.7 56.6
H2SO4 100 82.6 18.0 52.2 49.8 37.7 96.7 86.0 43.6
Fe Loading 32.8 mg Fe g−1 18.3 mg Fe g−1 23.9 mg Fe g−1
HNO3 49.0 21.1 9.7 0.6 2.3 2.5 1.5 1.8 6.0
HCl 86.3 64.4 9.6 2.3 1.5 2.3 24.0 38.9 4.2
H2SO4 98.8 82.8 22.0 2.5 2.9 2.3 28.8 29.2 10.0

concentration of 576 mg Mn L−1) is poorly removed from Crude phosphoric acid pre-treated with activated charcoal
phosphoric acid (less than 12%), whatever the resin. For Zn,
sorption efficiency varies between 21 (with MTC1600H) and The pre-treatment with activated charcoal is supposed to re-
30% (for both MTS9500 and MTS9570). For Pb (borderline move essentially organic contaminants. This pre-treatment
metal, according HSAB principle), MTC1600 sulfonic resin maintains high levels of metal ions (about 6.9 g Fe L−1,
shows poor efficiency; metal binding increases to 22% and 476 mg Mn L−1 or 241 mg Zn L−1, for example) and up to
26% for MTS 9500 and MTS 9570, respectively. This is with 7.5 g L−1 for F. There is a large excess of this metal ions
cadmium (soft acid, HSAB) that the largest differences are compared with cadmium concentration (i.e.,
observed; this can be also explained by much lower concen- 16.1 mg Cd L−1). This high salinity means that the resin
tration in phosphoric acid. Consistently with the data obtained may be rapidly saturated. This is confirmed by the residual
with synthetic solutions, for Cd the resins can be ranked ac- concentrations of target metals in the solution after being treat-
cording: MTC1600H (87%) >> MTS9570 (41%) >> ed with the resins (Table 4). Compared to initial metal levels,
MTS9500 (9%) (Table AM6). Under selected experimental the concentration abatements for Cr, Pb, Zn, and Mn are neg-
conditions, the residual concentrations of heavy metals remain ligible and the three resins have very comparable sorption
very high, making the acid inappropriate for food, pharmacy performances (Table AM7). For Fe and Cd, the resins show
or electronic applications. However, crude phosphoric acid the expected ranking: MTC1600H > MTS9570 > MTS9500.
after contact with MTC1600H resin becomes suitable for the However, even with the best resin (sulfonic-based
production of environmental friendly phosphate fertilizers: MTC1600H) the residual concentrations (especially for Fe)
cadmium concentration (about 7 mg kg−1 P2O5) is acceptable are not appropriate for high-added value applications.
with respect to the levels recommended by WHO, FAO, and
European Food Safety Authority (Roberts 2014; Ulrich 2019). Crude phosphoric acid pre-treated by TBP solvent extraction

The solvent extraction process substantially reduces residual

Table 3 Specifications of crude phosphoric acid before and after concentrations for the different metal ions (and fluoride that
contact with resins disappeared) (Table 4): lead and chromium are fully extracted
Compound Initial content MTC1600H MTS9500 MTS9570
by TBP and the levels of Fe, Zn and Mn are reduced by one
order of magnitude.
Concentration (%) After treatment by sorption on MTC1600H, the removals
P2O5 ≈ 23.2 ≈ 23.2 ≈ 23.2 ≈ 23.2 of Cd and Fe range between 93 and 96% (Table AM8).
Concentration (g L−1) Residual concentrations decrease to 0.2 mg Cd L−1 and
Fe 9.5 6.9 9.5 7.7 31.8 mg FeL−1 (Table 4). The combination of solvent extrac-
F 8.6 8.6 8.6 8.6 tion and sorption onto MTC1600H sulfonic resin drastically
Concentration (mg L−1) increases the purity of phosphoric acid. As expected, the other
Cr 120 96.2 90.5 90.3 resins are less efficient in cadmium and iron removal. For
Cd 16.7 2.2 14.7 9.9 MTS9570, the sorption efficiency ranges between 53 and
Pb 55.0 50.5 43.1 35.1 58%; residual concentration is rather limited for cadmium
Zn 271 215 189 190 (about 2 mg Cd L−1), but iron is still present at high concen-
Mn 576 506 520 539 tration (i.e., 203 mg Fe L−1). This means that the treatment is
not sufficient for applications requiring high purity. The
Table 4 Specifications of pre-treated phosphoric acid before/after contact with resins

Compound Pre-treatment

Activated charcoal (sorption) TBP (solvent extraction)

Initial content MTC1600H MTS9500 MTS9570 Initial content MTC1600H MTS9500 MTS9570

P2O5 ≈ 22.6% ≈ 22.6% ≈ 22.6% ≈ 22.6% ≈ 20.1% ≈ 20.1% ≈ 20.1% ≈ 20.1%

−1
Fe 6.9 g L 5.6 6.5 6.2 430 mg L−1 31.8 244 203
F 7.4 g L−1 7.4 7.4 7.4 – – – –
Cr 101 mg L−1 97.2 95.5 93.3 – – – –
Cd 16.1 mg L−1 2.0 14.4 9.3 4.8 mg L−1 0.2 4.2 2.0
Pb 45.0 mg L−1 44.5 40.2 37.6 – – – –
Zn 241 mg L−1 209 199 195 36.3 mg L−1 22.7 28.7 25.0
Mn 476 mg L−1 466 476 469 25.0 mg L−1 18.4 22.5 20.8

results are even worst for MTS9500: sorption efficiency varies sorption on activated charcoal and solvent extraction, using
between 13 and 43%, with residual concentrations as high as MTC1600H fixed-bed column as a polishing treatment. The
4.2 mg Cd L−1 and 244 mg Fe L−1. desorption could be also tested with alternative acid eluents
such as perchloric acid solutions (Kocjan 1999; Kushwaha
Recommendations et al. 2010; Montembault et al. 1999; Rivas et al. 2002); such
an eluent brings both acidic and oxidizing effects that could
The high levels of heavy metals of crude phosphoric acid contribute to enhanced desorption.
makes the resource inappropriate for manufacturing of fertil-
izers. The level of P2O5 remains unchanged after contact with
the three resins. The treatment of crude phosphoric acid with Conclusion
MTC1600H sulfonic resin (at the sorbent dosage of 2.5 g L−1)
significantly decreases the residual concentrations of heavy The purity of phosphoric acid controls its potential commer-
metals, including cadmium and iron. However, their levels cial uses. The removal of color and toxic metals is thus a
are not appropriate for food applications (for example). The critical step in the development of added-value products.
presence of other contaminants, not extracted by the resin, Due to the high concentration of iron and the toxicity of cad-
may declass treated phosphoric acid for fertilizer application mium, these two metals have been considered for purification
(and obviously other high added-value applications). The of phosphoric acid using three resins bearing different func-
treatment of crude phosphoric acid with activated charcoal tional groups (i.e., sulfonic, aminophosphonic, and bi-
allows removing some inorganic contaminants but the impact functional phosphonic/sulfonic groups). The study, performed
on heavy metals is rather limited. The combination of sorption on bi-component synthetic solutions, shows that the uptake
on activated charcoal and MTC1600H resin also meets the kinetics are comparable for the two metals (difficulty to play
requirements for fertilizer manufacturing but the grade is not on kinetics for separating cadmium and iron) with equilibrium
sufficient, at this stage, for food or pharmaceutical applica- time in the range 2–4 h. The uptake kinetics are fitted by the
tions. The “cleaning” of crude phosphoric acid with TBP sol- pseudo-first-order rate equation. The sorption efficiencies de-
vent extraction substantially remove metal impurities (and crease with increasing phosphoric acid concentration and de-
fluoride content); using MTC1600H sorption as polishing creasing resin dosage. The sulfonic-based resin (i.e.,
treatment allows reaching levels of contaminants much better MTC1600H) is much more efficient (higher sorption capaci-
than those required for fertilizer production. The optimization ties and higher selectivity for Cd over Fe) than alternative
of this treatment (playing with sorbent dosage, etc.) would resins. The selectivity for Cd over Fe increases with phospho-
probably allow reaching concentrations levels compatible ric acid concentration but decreases when the sorbent dosage
with food applications, providing the other contaminants to exceeds 1 g L−1. The superior properties of MTC1600H are
be also removed, such as sulfate, uranium, rare earth elements also demonstrated by the much higher desorption efficiencies
that may be also present in the crude phosphoric acid (depend- of Cd and Fe while using 1 M H2SO4 solutions: the desorption
ing on the origin of phosphate rocks). exceeds 99% contrary to other sorbents that show much lower
Future work could consist of testing the resins in dynamic desorption yield, especially for iron. Sulfonic acid resin ap-
systems for the treatment of phosphoric acid pre-treated by pears the most efficient for the removal of iron and cadmium.
The comparison of these three resins demonstrate the critical (VI) from phosphoric acid solutions: combining variables to opti-
mize ligand-ion communication. Solvent Extr Ion Exch 34:290–295
impact of the type of functional groups: sulfonic resin
Alexandratos SD, Zhu XP, Florent M, Sellin R (2016) Polymer-
(MTC1600H) > phosphonic/sulfonic bi-functional resin supported bifunctional amidoximes for the sorption of uranium from
(MTS 9570) >> aminophosphonic resin (MTS9500). seawater. Ind Eng Chem Res 55:4208–4216
The complex composition of crude industrial phosphoric Alexandratos SD, Zhu X (2017): The effect of hydrogen bonding in
enhancing the ionic affinities of immobilized monoprotic phosphate
acid solution (with huge concentrations of iron) does not allow
ligands. Materials 10
efficient metal removal by direct sorption on the resin; a pre- Barragan PP, Macedo MGM, Olguin MT (2017) Cadmium sorption by
treatment is required. Sorption on activated carbon does not sodium and thiourea-modified zeolite-rich tuffs. J Environ Sci 52:
allow reducing significantly iron concentration and a pre- 39–48
Buffle J, Zhang Z, Startchev K (2007) Metal flux and dynamic speciation
treatment with TBP solvent extraction appears more appropri-
at (bio)interfaces. Part 1: critical evaluation and compilation of phys-
ate for reducing residual concentrations to levels compatible icochemical parameters for complexes with simple ligands and
with the regulations for fertilizer uses. After solvent extraction fulvic/humic substances. Environ Sci Technol 41:7609–7620
pre-treatment, the comparison of sorption efficiencies for the Chen M, Li J, Jin Y, Luo J, Zhu X, Yu D (2018) Efficient solvent
extraction of phosphoric acid with dibutyl sulfoxide. J Chem
two metals and the tree resins confirms the conclusions raised
Technol Biotechnol 93:467–475
with synthetic solutions: MTC1600H resin allows removing Crank J (1975) The mathematics of diffusion. Oxford University Press,
about 93% of iron and 96% of cadmium, i.e., much higher Oxford, U.K., 414 pp
yields than for MTS9570 (53% and 58%, respectively) and EFMA (2000): Production of phosphoric acid. Best available techniques
even more for MTS9500 (43% and 13%, respectively). for pollution prevention and Controlin the European fertilizer indus-
try, 4. European fertilizer Manufacturers' Association, Brussels,
Combined with other treatments (for removal of other contam- Belgium, 48 pp
inants), the sorption of metal ions on MTC1600H offers prom- El-Bayaa AA, Badawy NA, Gamal AM, Zidan IH, Mowafy AR (2011)
ising perspectives for reaching the levels of decontamination Purification of wet process phosphoric acid by decreasing iron and
compatible with food applications. uranium using white silica sand. J Hazard Mater 190:324–329
El Zrelli R, Rabaoui L, Daghbouj N, Abda H, Castet S, Josse C, van Beek
P, Souhaut M, Michel S, Bejaoui N, Courjault-Rade P (2018)
Acknowledgments Authors thank IFE (Institut Français d’Egypte, Characterization of phosphate rock and phosphogypsum from
France), Ministère des Affaires Etrangères, Ministère de Gabes phosphate fertilizer factories (SE Tunisia): high mining po-
l’Enseignement Supérieur et de la Recherche (France) and STDF tential and implications for environmental protection. Environ Sci
(Science and Technology Development Fund, Egypt) for supporting the Pollut Res 25:14690–14702
bi-lateral collaboration between NMA and IMT-Mines Ales and the Elmaadawy KHG, Ezz El Din M, Khalid AM, Abouzeid AZ (2015)
funding of post-doctoral fellowship of Dr. Ahmed Masoud (at IMT- Mineral industry in Egypt– part II non metallic commodities –
Mines Ales). IAEA (Internal Atomic Energy Association) for funding phosphate rocks. J Min World Express 4:1–18
the visit of Dr. Taha Montaser to IMT – Mines Ales (Fellowship No. Gomes HI, Jones A, Rogerson M, Burke IT, Mayes WM (2016)
FS-EGY2013-1702837). Vanadium removal and recovery from bauxite residue leachates by
ion exchange. Environ Sci Pollut Res 23:23034–23042
Guo J, Yan C, Luo Z, Fang H, Hu S, Cao Y (2019) Synthesis of a novel
ternary HA/Fe-Mn oxides-loaded biochar composite and its appli-
References cation in cadmium (II) and arsenic(V) adsorption. J Environ Sci 85:
168–176
Hamza W, Chtara C, Benzina M (2016) Purification of industrial phos-
Abasiyan SMA, Dashbolaghi F, Mahdavinia GR (2019) Chitosan cross-
phoric acid (54%) using Fe-pillared bentonite. Environ Sci Pollut
linked with kappa-carrageenan to remove cadmium from water and
Res 23:15820–15831
soil systems. Environ Sci Pollut Res 26:26254–26264
Heres X, Blet V, Di Natale P, Ouaattou A, Mazouz H, Dhiba D, Cuer F
Abdel-Ghafar HM, Abdel-Aal EA, Ibrahim MAM, El-Shall H, Ismail
(2018) Selective extraction of rare earth elements from phosphoric
AK (2019) Purification of high iron wet-process phosphoric acid
acid by ion exchange resins. Metals 8
via oxalate precipitation method. Hydrometallurgy 184:1–8
Ho YS, McKay G (1999) Pseudo-second order model for sorption pro-
Ahmed H, Diamonta H, Chaker C, Abdelhamid R (2007) Purification of cesses. Process Biochem 34:451–465
wet process phosphoric acid by solvent extraction with TBP and Islam A, Zaidi N, Ahmad H, Kumar S (2015) Amine-functionalized
MIBK mixtures. Sep Purif Technol 55:212–216 m e s o p o r o u s p o l ym e r a s p o t e nt i a l s or b e n t f o r n i c k e l
Ahmed IM, Ammanoeil RN, Saad EA, Daoud JA (2019) Purification of preconcentration from electroplating wastewater. Environ Sci
crude phosphoric acid and leached apatite by solvent extraction with Pollut Res 22:7716–7725
CYANEX 923 in kerosene. Periodica Polytechnica-Chemical Khayambashi A, Wang X, Wei Y (2016) Solid phase extraction of ura-
Engineering 63:122–129 nium (VI) from phosphoric acid medium using macroporous silica-
Alexandratos SD, Smith SD (2004) Intraligand cooperation in metal-ion based D 2 EHPA-TOPO impregnated polymeric adsorbent.
binding by immobilized ligands: the effect of bifunctionality. J Appl Hydrometallurgy 164:90–96
Polym Sci 91:463–468 Kocjan R (1999) Retention of some metal ions and their separation on
Alexandratos SD, Zhu X (2015) The role of polarizability in determining silica gel modified with Acid Red 88. Mikrochim Acta 131:153–158
metal ion affinities in polymer-supported reagents: monoprotic Kouzbour S, Gourich B, Gros F, Vial C, Allam F, Stiriba Y (2019)
phosphates and the effect of hydrogen bonding. New J Chem 39: Comparative analysis of industrial processes for cadmium removal
5366–5373 from phosphoric acid: a review. Hydrometallurgy 188:222–247
Alexandratos SD, Zhu X (2016) Polymer-supported Kushwaha S, Sreedhar B, Padmaja P (2010) Sorption of phenyl mercury,
aminomethylphosphinate as a ligand with a high affinity for U methyl mercury, and inorganic mercury onto chitosan and barbital
immobilized chitosan: spectroscopic, potentiometric, kinetic, equi- Rivas BL, Pooley SA, Aceiton E (2002) Metal ion adsorption behavior of
librium, and selective desorption studies. J Chem Eng Data 55: poly 3-(dimethylamine) propyl acrylate and poly 3-(dimethylamine)
4691–4698 propylacrylate-co-acrylic acid resins. J Appl Polym Sci 84:1251–
Kussainova MZ, Pasa S, Zhumasilovich DU, Chernyakova RM, Atlan 1256
M, Temel H (2016) Comparative sorption capacity of Pb (II) and Cd Roberts TL (2014) Cadmium and phosphorous fertilizers: the issues and
(II) by natural zeolite in phosphoric acid medium. Desalin Water the science. Procedia Eng 83:52–59
Treat 57:12561–12571 Sanghani R (2014): Novel technique for purification of fertilizer phos-
Kussainova MZ, Chernyakova RM, Jussipbekov UZ, Pasa S (2019) phoric acid with simultaneous uranium extraction. In: Amalhay M
Structural investigation of raw clinoptilolite over the Pb2+ adsorp- (Editor), Symphos 2013 - 2nd International Symposium on
tion process from phosphoric acid. J Mol Struct 1184:49–58 Innovation and Technology in the Phosphate Industry. Procedia
Lan T, Li PF, Rehman FU, Li XL, Yang W, Guo SW (2019) Efficient Engineering, pp. 225-232
adsorption of Cd2+ from aqueous solution using metakaolin Shweikani R, Kousa M, Mizban F (2013) The use of phosphogypsum in
geopolymers. Environ Sci Pollut Res 26:33555–33567 Syrian cement industry: radiation dose to public. Ann Nucl Eng 54:
Lembrikov VM, Konyakhina LV, Volkova VV, Lobova MV, Pokidova 197–201
OV (2004) Interaction of tri-n-butyl phosphate, water, and phospho- Spies ARL, Wewers F (2020) Equilibrium, kinetics and thermodynamics
ric acid in purification of wet-process phosphoric acid. Russ J Appl studies of Cd sorption onto a dithizone-impregnated Amberchrom
Chem 77:1606–1608 CG-300m polymer resin. Arab J Chem 13:5050–5059
Li K, Li M, Xue D (2012) Solution-phase electronegativity scale: insight
Tang Y, Liang S, Wang J, Yu S, Wang Y (2013) Amino-functionalized
into the chemical behaviors of metal ions in solution. J Phys Chem A
core-shell magnetic mesoporous composite microspheres for Pb (II)
116:4192–4198
and Cd (II) removal. J Environ Sci 25:830–837
Li X, Li J, Jin Y, Chen M, Feng D, Guo Y (2017a) Wet process of
phosphoric acid purification by solvent extraction using tri-n-butyl Tjioe TT, Weij P, Wesselingh JA, Rosmalen GMV (1988) Removal of
phosphate and cyclohexanol mixtures. J Serb Chem Soc 82:579– cadmium anion exchange in a wet phosphoric acid process. Part 2:
592 equilibria, kinetics and regeneration. Solvent Extr. Ion Exch 6:505–
Li X, Li J, Luo J, Jin Y, Zou D (2017b) Purification of wet process 560
phosphoric acid by solvent extraction using cyclohexanol. Solvent Trabelsi W, Tlili A (2017) Phosphoric acid purification through different
Extr Res Dev - Jpn 24:23–35 raw and activated clay materials (southern Tunisia). J Afr Earth Sci
Liu C, Cao J, Shen W, Ren Y, Mu W, Ding X (2016) Liquid - liquid 129:647–658
equilibria of quaternary system phosphoric acid/sulfuric acid/water/ Ulrich AE (2019) Cadmium governance in Europe's phosphate fertilizers:
tri-n-butyl phosphate and the formation of extraction complex at not so fast? Sci Total Environ 650:541–545
303.2 K. Fluid Phase Equilib 408:190–195 Virolainen S, Repo E, Sainio T (2019) Recovering rare earth elements
Marczenko Z, Balcerzak M (2000) Chapter 37 - phosphorus. In: from phosphogypsum using a resin-in-leach process: selection of
Balcerzak M (ed) Marczenko Z. Elsevier, Analytical Spectroscopy resin, leaching agent, and eluent. Hydrometallurgy 189:105125
Library, pp 326–333 Wang ZH, Shen F, Shen DK, Jiang YH, Xiao R (2017) Immobilization of
Montembault V, Soutif JC, Brosse JC, Grote M (1999) Synthesis and Cu2+ and Cd2+ by earthworm manure derived biochar in acidic
complexing properties of resins containing aminocarboxylic acid as circumstance. J Environ Sci 53:293–300
functional groups from diethylenetriaminepentaacetic acid Wieszczycka K, Filipowiak K, Wojciechowska I, Aksamitowski P
bisanhydride and polyvinyl alcohols. React Funct Polym 39:253– (2020) Novel ionic liquid-modified polymers for highly effective
261 adsorption of heavy metals ions. Sep Purif Technol 236:10
Nasr B, Hedi B, Abdellatif G, Rodrigo MA (2005) Purification of wet- Yassien ZJ, Elameer ZAM (2014) Profit maximization for Sibaiya east
process phosphoric acid by hydrogen peroxide oxidation, activated open-pit phosphate mines by using spreadsheet software packages.
carbon adsorption and electrooxidation. Chem Eng Technol 28: Al-Azhar Univ Eng J 9:1–10
193–198 Ye C, Li J (2013) Wet process phosphoric acid purification by solvent
NMA (2010) Economic evaluation of U and REEs in Egyptian phospho- extraction using N-octanol and tributylphosphate mixtures. J Chem
rite ores. NMA, Cairo Technol Biotechnol 88:1715–1720
NMA (2017) Uranium extraction from dihydrate phosphoric acid. Yin GC, Bi LL, Song XW, Luo HY, Ji PP, Lin QT, Liu QJ, Tang GY
Nuclear Materials Authority, Cairo (2019) Adsorption of Cd (II) from aqueous solution by Pennisetum
NMA (2018) Purification of wet process phosphoric acid using solvent sp. straw biochars derived from different modification methods.
extraction technique. Nuclear Materials Authority, Cairo Environ Sci Pollut Res 26:7024–7032
Pearson RG (1966): Acids and bases. Science (New York, N.Y.) 151,
Zhu X, Alexandratos SD (2011) Effect of hydrogen-bonding in the de-
172-7
velopment of high-affinity metal ion complexants: polymer-bound
Persson I (2010) Hydrated metal ions in aqueous solution: how regular
phosphorylated cyclodextrin. J Appl Polym Sci 121:1137–1142
are their structures? Pure Appl. Chem. 82:1901–1917
Prelot B, Ayed I, Marchandeau F, Zajac J (2014) On the real performance Zhu X, Alexandratos SD (2015) Development of a new ion-exchange/
of cation exchange resins in wastewater treatment under conditions coordinating phosphate ligand for the sorption of U (VI) and triva-
of cation competition: the case of heavy metal pollution. Environ Sci lent ions from phosphoric acid solutions. Chem Eng Sci 127:126–
Pollut Res 21:9334–9343 132
Reddy BR, Kumar JR (2016) Rare earths extraction, separation, and Zuo Y, Chen Q, Li C, Kang C, Lei X (2019) Removal of fluorine from
recovery from phosphoric acid media. Solvent Extr. Ion Exch. 34: wet-process phosphoric acid using a solvent extraction technique
226–240 with tributyl phosphate and silicon oil. ACS Omega 4:11593–11601
Reyes LH, Medina IS, Mendoza RN, Vazquez JR, Rodriguez MA,
Guibal E (2001) Extraction of cadmium from phosphoric acid using Publisher’s note Springer Nature remains neutral with regard to jurisdic-
resins impregnated with organophosphorus extractants. Ind Eng tional claims in published maps and institutional affiliations.
Chem Res 40:1422–1433