Desorption of La3+ and Ce3+ from Treated ‘Chert’ a Siliceous Byproduct of the Phosphate Mining Industry of Gafsa-Metlaoui Basin (Southwestern Tunisia)

  • Imen Bouchmila
  • Bochra Bejaoui Kefi
  • Radhia Souissi
  • Mohieddine Abdellaoui
Original Article


Chert, the most abundantby-product of the phosphate mining industry, collected from the Metlaoui Gafsa basin underwent purification treatments and was characterized by various analytical methods (XRD, XRF, SEM, FTIR, laser granulometry). The obtained treated chert is a siliceous phase SiO2 (~ 98%) mainly composed of opal-CT (trydimite phase and cristobalite) and traces of quartz. The chert morphology is granular. Treated chert has been successfully tested and proved to be an effective sorbent of rare earths. In this study, the desorption of La3+ and Ce3+ from ion-loaded chert was investigated and the reversibility of sorption reaction was verified. Several eluting agents at different concentrations were tested to desorb those lanthanides such as mineral acids (HCl and HNO3), chloride salt (CaCl2) and ultra pure water. Results show that, HCl and CaCl2 are efficient eluents and the adsorbed Ce3+and La3+ can be easily eluted by them. However, the highest desorption percentages were achieved using calcium chloride as eluting agent (Ce3+, 93.651%; La3+, 96.446%). Kinetics of desorption of La3+ and Ce3+ from treated chert were studied. The rate of desorption was initially fast in the first 30 min, but gradually declined with time. Among all tested kinetic models, the pseudo-second-order model fits the kinetic data best with a better correlation coefficient (R2 > 0.9) and the desorption isotherms fitted better to freundlich, which means that lanthanum and cerium release was dominated by multilayer desorption process from heterogeneous and independent surfaces.


Desorption Siliceous phase Lanthanum Cerium Kinetics Isotherms 

1 Introduction

Rare earths are a group of 17 chemically similar elements, including lanthanides (atomic numbers 57–71) and two other additional elements scandium (atomic number 21), and yttrium (atomic number 39) [1, 2]. They are grouped depending on the atomic number, in “light” rare earth elements (REE)—La, Ce, Pr, Nd, and “middle and heavy” REE—Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Y. Due to the physico-chemical similarity, rare earths are formed in the same minerals and behave as a single entity [2, 3].

Considering their exceptional and diverse properties (chemical, electrical, metallurgical, magnetic, optical, and catalytic properties) the demand on rare earths has increased in the last decades. They are strategically important and widely used in many high-technology industries, such as nuclear energy, metallurgy, medicine, chemical engineering, electronics, computer and agriculture…. [4].

Most rare earth elements are not as rare as the name of the group suggests. For example, lutetium (Lu), one of the least abundant of the group, is about 125 times more abundant than gold [1, 5]. Although relatively abundant in the earth’s crust, they tend to occur typically dispersed and in low concentrations. Therefore, an enrichment step combined with a matrix separation is often required.

Thus, several methods such as liquid−liquid extraction [6], chemical precipitation [7] and ion exchange [8] have been developed and employed to separate and to preconcentrate RREs. However, these conventional processes have some disadvantages, such as high consumption of reagent, low selectivity and high operational cost [4]. Thus, there is a great need for an inexpensive, rapid, easy and very efficient alternative process for the extraction and preconcentration of rare earths. Adsorption is a method that satisfies all these requirements [9]. Also and most importantly, it can be used for ions extraction and recovery even from low-concentration sources by means of relatively simple processes.

Many sorbent such activated carbon, bacterial cell walls, silica gel, resin, C18, polymer supports, granular hybrid, nanostructured materials have been investigated for the extraction and preconcentration of rare earths [10, 11, 12, 13, 14, 15, 16, 17].

Many tested sorbents have shown considerable rare earth adsorption capacity (activated carbon [10]; 4.13 mg g−1, crab shells [18], 3.23 mg g−1, multi-walled carbon nanotubes coated cellulose acetate membrane [17]; 28 and 23 mg g−1, SBA-15 mesoporous silicas functionalized with N-propyl salicylaldimine [18]; 5.1 mg g−1, SBA-15 mesoporous silicas functionalized with ethylenediaminepropylesalicylaldimine [18]; 15.6 mg g−1, bone powder [18]; 12.7 mg g−1, oxidized multiwalled carbon nanotubes [18]; 99.01 mg g−1). However, low cost industrial and mining by-products sorbent can be used as an alternative to more costly materials.

In recent years, adsorption by various low-cost adsorbents has become the major focus of numerous investigations [19]. Naturally occurring, industrial and mining by-products/waste materials are widely used due to their abundant availability and low cost.

The use of low-cost by-products as adsorbents might have a beneficial double impact on both society and environment. Indeed, it allows the recycling of these wastes that are currently discarded and their reuse as a low-cost processing and treatment tool. Both of these factors create added value for the waste materials [19].

In Tunisia, Metlaoui Gafsa basin is the largest phosphate mining center. But, during the process of phosphate production, many mining by-products such as chert alternating with the phosphate layers are thrown away [20, 21]. One of the main problems in the mining industry is the accumulation of waste. Therefore, it is necessary to find ways to reduce this accumulation [22]. Valorization of mining by-products in this basin is an answer to the problem.

Therefore, chert is tested in this work as potential sorbent for the solid phase extraction of rare earth elements from aqueous solutions.

For solid phase extraction and separation process of rare earth elements, adsorption method is used firstly and then to recover the adsorbed ions, desorption process is needed. Therefore, desorption kinetics and equilibrium are important in understanding desorption characteristics of the studied adsorbent [23].

In contrast to adsorption studies, desorption has been given a limited attention. Indeed, only a limited number of studies have investigated the desorption process of rare earth from siliceous adsorbents [24, 25, 26].

Usually, desorption study is conducted in conjunction with the adsorption study to determine the reversibility of this reaction. Desorption is the inverse process of adsorption by which the molecules previously adsorbed is “detached” from the adsorbent and is released into the solution. The adsorbed compounds may be recovered by washing the loaded adsorbent with an appropriate desorbing agent, leaching by means of chemical reagents, biological process or thermal treatment [27]. Leaching the sorbed ion by chemical reagents can be used with different types of eluents agents such as distilled water, HCl, HNO3, NaOH… [27].

Therefore, in the first part of this study, we focus on determining the most effective eluents to desorb La3+ and Ce3+ from chert and on the development and optimization of the desorption process. In the second part we determine the suitability of different kinetic models to describe the mechanism of desorption and release of these ions.

The present work aims to prepare a low cost and efficient sorbent from an industrial by-product which has no economic value for the extraction and recovery of rare earths. Allowing their recycling and their reuse as low-cost and effective alternative to more expensive sorbent.

2 Experimental

2.1 Reagents and Solutions

All chemicals were of analytical grade used without further purification. Solutions were prepared using ultrapure water (Milli-Q ultrapure water).

Nitric acid (HNO3, 65%) and hydrochloric acid (HCl, 37%) were from Scharlau Chemicals. Metal ion solutions are prepared from CaCl2.2H2O (Sigma-Aldrich) by dissolving appropriate amounts in ultra pure water.

Working standard solutions of REEs (Ce3+and La3+) were prepared on a daily basis from Ce(Cl)3·7H2O to La(Cl)3·7H2O obtained from Sigma-Aldrich by dissolving appropriate amounts in ultra pure water. Standard solutions for ICP-AES measurement (up to 4 mg mL−1 for the calibration curves) were freshly prepared before each set of measurements by appropriate dilution from the commercially available single element stock solutions (1000 mg mL−1) obtained from Sigma-Aldrich.

2.2 Sorbent Preparation

The sorbent used in this work is chert, a highly silified sedimentary rock [28] collected from Gafsa-Metlaoui basin in Tunisia.

The Gafsa-Metlaoui basin is located in the south-western part of Tunisia, between latitude 34° and 35° North and longitude 8° and 10° East (Fig. 1).
Fig. 1

Location of the Gafsa-Metlaoui bassin in the southwest of Tunisia [29]

The cherty levels in this basin are composed mainly of opal CT (cristobalite/tridymite) located generally at the base of the main phosphatic series between CVI and CVII phosphatic layers [28].

At first, the sample was ground into a fine powder. In order to remove any undesirable moisture the obtained powder was dried in an oven for 24 h at a temperature of 60 °C. Then successive purifications steps were applied to the material to remove impurities such as carbonates, iron oxides, clays, etc..

To remove carbonates (present in the forms of calcite and dolomite) and the iron oxides, an acid attack (HCl 12%) at room temperature was performed on the chert finely crushed.

To remove excess acid and iron chloride, the resulted material was washed using UP water. The suspension is then removed by centrifugation.

To eliminate clays, the decarbonated chert was washed by a volume of ultra-pure water. The clay suspension is then removed by centrifugation. This step was repeated several times. Finally, the obtained material is dried in an oven at 60 °C.

The adsorbent was then ready for the adsorption/desorption experiments.

2.3 Sorbent Characterization

X-ray diffraction (XRD) patterns were obtained using PANalytical, X’Pert PRO’ equipped with a copper anticathode (Cu Kα λ = 1.540598 Å). The acquisition conditions are: angular domain 3–100° in 2θ, step size 0.0330, scan step time 81.0102 s. Analysis of XRD patterns was performed by X’Pert HighScore Plus program connected to the ICDD-PDF2 data base.

Fourier-transform infrared spectroscopy type ‘FTIR VERTEX 70 ATR Bruker Diamant’ was used to identify the surface groups of these cherty materials. We proceeded with a resolution of 4 cm−1 and 40 scans. FTIR was recorded in the range 400–4000 cm−1.

X-ray diffraction and FTIR were used to gain qualitative information about the nature and the structure of cherty phase.

Magix PW2403 X-ray fluorescence (XRF) was employed to determine the overall chemical compositions of the raw ‘Chert’ and treated ‘Chert’.

The surface morphology and micro-chemical composition of the sample were investigated using a scanning electron microscope (ESEM) type FEI Quanta 200.

Granulometric analyses of samples were determined by Lazer Granulometer type Masteriser SCIRRCCO 2000 with wet analysis using the Malvern Hydro 2000 dispersion unit. The sample has been analyzed in water suspension. In order to obtain good particle dispersions, a suitable ultrasonic sound time has been applied (120 s). In order to compare the granulometric parameters from different samples, mass medium diameter (MMD) has been calculated. D (v; 0.5) corresponds to the particle caliber for that 50% of the sample having a lower size and 50% of the sample having a higher size and D (v; 0.9) corresponds to the particle caliber for that 90% of the sample having a lower size and 10% of the sample having a higher size.

Analyzes of rare earths elements (REE) were performed by an inductively coupled plasma atomic emission spectrophotometer (ICP-AES) type HORIBA Jobin–Yvon. ICP-AES is the commonly used technique in the determination of trace REEs because of the capability for rapid multi-element detection over a wide concentration range with relatively low detection limits [30].

2.4 La3+ and Ce3+ Adsorption (Ions Loading) and Desorption (Ions Elution) Batch Experiments

Adsorption experiments were performed in order to load the sorbent with La3+ or Ce3+ by adding 0.4 g of treated ‘Chert’ into 200 ml of lanthanum (La) or cerium (Ce) salt solution. Initial concentrations of lanthanum and cerium are 37.60 mg L−1 and 37.40 mg L−1, respectively. These samples were then agitated using a magnetic stirrer for 3 h at room temperature, pH values were measured over the course of the reaction and values are between 4 and 5. This period of time was chosen to ensure that adsorption equilibrium was achieved since kinetics adsorption studies showed that the required time to reach equilibrium was approximately 30 min. The mass was filtered out and dried overnight at 80 °C.

The amount of La3+ or Ce3+ uptake was calculated based on the difference in rare earth ion concentration in the bulk of solution before and after adsorption by using the following equation [31]:
$$Q_{e} = \frac{{\left( {C_{0} - C_{e} } \right)V}}{m} \left( {{\text{mg g}}^{ - 1} } \right)$$

With Qe as ions uptake capacity (mg g−1) at equilibrium, Ce is ions concentration in solution (mg L−1) at equilibrium, C0 the initial concentration (mg L−1), V is the volume of the solution (L) and m the dry weight of the ‘Cherty’ material used (g).

Desorption experiments were performed using 0.1 g of ion loaded chert with known amount of lanthanide ions, immersed under stirring in 50 ml of eluting solution for 5 h.

Tested eluting agents are: ultra pure water, hydrochloric acid (0.5, 0.1, 0.3, 0.6, 2 and 4 M), nitric acid (0.1, 1, 2 and 4 M) and calcium chloride (0.01, 0.1, 0.5, 1 and 2 M).

To determine the instantaneous concentration Ct of desorbed ions, a volume of 2 mL was collected at defined times and then filtered and dosed by ICP-AES.

The amount of ions released from the loaded mass as a function of time was calculated from the following equation [32]:
$$Q_{\text{des}} = \frac{{(C_{t} V)}}{m} {\text{mg}}\;{\text{g}}^{ - 1}$$
The percent of desorption was obtained from the following equation [31]:
$$\% {\text{Desorption = }}\frac{\text{Amount of desorbed ions }}{\text{Amount of adsorbed ions}}$$

With Ct (mg L−1) is the concentration of La3+ or Ce3+ at time t and V is the volume of the eluting solution and m is the mass of the loaded ‘chert’ (g).

A kinetic study was carried out to examine the rate-controlling desorption mechanism of cerium and lanthanum by the treated adsorbent. In order to cognize the mechanism of desorption, it is necessary to determine the suitability of different desorption kinetic models. The most frequently used models for desorption include zero order, first order, pseudo first-order, pseudo second-order, parabolic diffusion, simple Elovich, two constant rate and simple forms were fitted to desorption data as follows [26, 27, 31, 33]:

Zero order model:
$$Q_{t} = Q_{0} - K_{0} t.$$
First order model:
$$Ln\left( {Q_{t} } \right) = Ln\left( {Q_{01} } \right) - K^{'}_{1} t.$$
Second order model:
$$\frac{1}{{Q_{t} }} = \left( {\frac{1}{{Q_{02} }}} \right) - K^{'}_{2} t.$$
Parabolic diffusion model:
$$Q_{t} = Q_{0p} + K_{p} t^{0.5} .$$
Two constant rate model:
$$Ln\left( {Q_{t} } \right) = Ln\left( a \right) + bLn\left( t \right).$$
Simple Elovich model:
$$Q_{t} = \left( {\frac{1}{{\beta_{s} }}} \right)Ln\left( {\alpha_{s} \beta_{s} } \right) + \left( {\frac{1}{{\beta_{s} }}} \right)Ln\left( t \right).$$
Pseudo-first order model:
$$Ln\left( {Q_{e1} - Q_{t} } \right) = Ln\left( {Q_{e1} } \right) - K_{1} t.$$
Pseudo-second order model:
$$\left( {\frac{t}{{Q_{t} }}} \right) = \left( {\frac{t}{{Q_{e2} }}} \right) + \left( {\frac{1}{{K_{2} Q_{e2}^{2} }}} \right).$$

With Qt, Q0, Q01, Q0p, Q01, Q02, Qe1 and Qe2 the amounts of rare earth ion (mg g−1) desorbed at any point in time t, at t = 0 and at equilibrium conditions, respectively. Respectively, K0 is the zero order rate constant (mg g−1min−1), K1 is the first order rate constant (min−1), K2 is the second order rate constant (g mg−1min−1 (mg g−1)−1), Kp is the diffusion rate constant (mg g−1)−0.5, a is the initial cerium or lanthanum desorption rate constant (mg g−1min−1)b, b is the cerium or lanthanum desorption rate coefficient (mg g−1)−1, αs is the initial cerium or lanthanum desorption rate (mg g−1min−1), βs is the cerium or lanthanum desorption constant (mg g−1)−1. K1 is the pseudo first-order desorption rate constant (min−1) and K2 (mg g−1min)−1 is the pseudo second-order desorption rate constant.

The most commonly used adsorption and desorption isotherms are Langmuir and Freundlich. The Langmuir equation describes only a monolayer surface coverage, adsorption/desorption takes place at independent sites, and adsorption sites are equivalent, whereas the Freundlich isotherm describes multilayer adsorption/desorption onto heterogeneous surfaces that provide adsorption sites of varying affinities with different energies [34].

Therefore, desorption data were fitted to the Freundlich and Langmuir isotherm models, as follows [35]:

Freundlich isotherm model:
$$\log \left( {Q_{e} } \right) = \log \left( {K_{f}^{\text{des}} } \right) + \frac{1}{{n_{des} }}\log \left( {C_{e} } \right).$$
Langmuir isotherm models:
$$\left( {\frac{{C_{e} }}{{Q_{e} }}} \right) = \left( {\frac{1}{{Q_{ \text{max} } b}}} \right) + \left( {\frac{{C_{e} }}{{Q_{ \text{max} } }}} \right)$$

Where Qe is the adsorption capacity at equilibrium (mg g−1), Ce is the equilibrium concentration (mg L−1), Qmax is the solid phase concentration corresponding to the complete monolayer coverage of desorption sites (mg g−1); b is the Langmuir desorption constant related to the free energy of desorption (L g−1), Kf des [(mg g−1)(L g−1)1/n] and ndes are Freundlich desorption constants.

3 Results and Discussion

3.1 Characterization of Chert

The chemical analysis by XRF of chert sample before and after purification is shown in Table 1. Silica is the main constituent of the samples. After purification steps, SiO2 content reached 97.29%. This purification allowed a complete elimination of carbonate and reduced the other oxides until 22.50%.
Table 1

Chemical analysis by XRF of raw chert and purified chert samples


Raw ‘Chert’ (wt%)

Purified ‘Chert’ (wt%)



























– content is lower than limit of quantification of XRF

The mineralogical structural characterisation using XRD (Fig. 2) revealed that both raw and purified cherts are mainly composed of opal CT (Cristobalite/Tridymite) as major mineral; quartz is present in lesser quantities. Further peaks are also observed in the raw material. They correspond to the hydroxyapatite and carbonate (the latter is observed under calcite and dolomite forms).
Fig. 2

XRD patterns of a raw chert and b purified chert

However, with the purified chert, we notice the disappearance of the crystalline phases of hydroxyappatite, calcite and dolomite. Yet, we also notice the appearance of clinoptilolite-Ca phase. Clinoptilolite appears to be present in the sample from the beginning and its identification has been possible due to the elimination of certain mineral phases during the purification performed on the raw chert.

The crystallographic properties of the silicate phases of the sample are shown in Table 2.
Table 2

Crystallographic characteristics of silicate phases

Crystallographic parameters




crystalline system




Space group




Number of space group




a (Å)




b (Å)




c (Å)




Calculated density (g cm−3)




Cell volume (106pm3)




Reference code




Fourier Transform Infrared spectroscopy (FT-IR) has been used to follow and to identify the specific chemical changes that occur in cherty material after purification steps. The spectral data in Fig. 3 show absorption bands of silicate which are generally observed between 400 and 1400 cm−1. Absorption bands at 447 and 791 are assigned to the lattice vibrations of tetrahedral Si–O and the wide bands centered at 1090 cm−1 should be due to Si–O–Si in-plane vibration (asymmetric stretching) [28, 36]. The IR spectra of all studied samples have at least one absorption band characteristics of cristobalite phase that appears at 619, 795, 1090 cm−1 and 1202 cm−1 [28]. The FT-IR spectrums, recorded before and after purification showed the disappearance of absorption bands characteristics of O–H at 1420 cm−1 and C–O bonds at 1420 cm−1, 874 cm−1 and 710 cm−1. Only in the spectrum of raw chert, there are peaks of hygroscopic water and calcite phase.
Fig. 3

FT-IR spectra of raw Chert and purified chert

SEM micrographie of the studied chert (Fig. 4) illustrate a granular morphology. Grains of different shape and size are observed. Bedded fossiliferous chert contains remains of siliceous organism as diatoms, radiolarians, and sponge spicules which are formed during recrystallization of siliceous oozes. The granular morphology may be the result of the transformation of bedded chert via the dissolution and precipitation mechanisms by hydrothermal waters where most of the silica dissolved is immediately reprecipitated leaving no trace of the fossil which confirm the chemogenic origin of chert [37]. EDX spectrum (Fig. 5), is in agreement with XRF analysis, and shows the existence of Si as major element with the presence of Mg, Al, K and Fe.
Fig. 4

SEM micrographs, in secondary electrons mode, of purified chert

Fig. 5

EDX spectrum of purified chert

Laser granulometry analyses present the relative quantities of particles having particular size. Granulometric distribution of cherty sample is illustrated in Fig. 6. Obtained results showed a granulometry lower than 100 μm and unimodal particle distribution between 1 and 100 μm. They showed also that chert contains 50% of particles having a diameter equal to or less than 8.61 μm and 90% having a diameter less than 19.70 μm (Table 3). Their greatest dispersion was founded around 9 μm.
Fig. 6

Particle size distribution for the purified chert

Table 3

Laser granulometry analyses


D10 (μm)

D50 (μm)

D90 (μm)





D(x) corresponds to the particle size for which x% of the population is below this value

3.2 Sorption of La3+ and Ce3+ on Purified Chert

After 3 h, the maximum uptake by purified chert was noted as 11.70 mg g−1 for Ce3+ and 11.35 mg g−1 for La3+. Generally, due to the similar electric configuration and coordination chemistry, adsorbed amounts of lanthanum and cerium are quite alike. However, small differences exist which arise from the “lanthanide contract” [37].

The REE’s serie is characterized by the increase in mass of the atomic nucleus, which causes a very regular decrease of the ionic radius from La to Lu (Ce  <  La), which results in a slight but predictable change in their chemical affinity [38]. Therefore, due to the higher affinity as compared to La3+, the amount of Ce3+ adsorbed by chert was always slightly higher.

3.3 Desorption of La3+ and Ce3+ with Ultra Pure Water

Figure 7 shows the desorption percentages of La3+ and Ce3+ using ultrapure water as eluting agent. Results show that ultra pure water did not have an impact on the amount of lanthanum and cerium adsorbed on chert. Indeed, only insignificant percentages of ion desorption (less than 3%) were obtained. We can also conclude that washing the La-loaded chert and the Ce-loaded chert has no effect. For this reason, only the necessary volume of ultra pure water for transferring the sorbent to the paper filter was used in the following experiments.
Fig. 7

Desorption of lanthanum and cerium with ulta pure water

3.4 Desorption of La3+ and Ce3+ with HCl, HNO3 and CaCl2

For efficient desorption and in order to attain a high percentage of ions (La3+ or Ce3+) desorption and recovery, optimization of eluent nature and concentration is necessary. Effective desorption eluents can be ions exchangers, complexing agents or contain competing ions [39]. Mineral acids such HCl and HNO3, the most commonly used desorption eluents, are considered to be protons exchanging agents [39]. Calcium is a competing cation, used as an eluent with ion exchange potential [39]. Ion exchange is the mechanism in which adsorbing ion take place of another ion already associated with the sorbent surface leading to their desorption [40].

For this purpose, HCl, HNO3 and CaCl2 at different concentrations were tested out with the loaded chert. Figures 8, 9 and 10 show the percentage variation of Ce3+ and La3+ desorption from the ion-loaded chert as a function of eluent concentration. From these figures, it is clearly observed that desorption of La3+ and Ce3+ depends on the nature and concentration of the eluent agent.
Fig. 8

Desorption of lanthanum and cerium with CaCl2

Fig. 9

Desorption of lanthanum and cerium with HCl

Fig. 10

Desorption of lanthanum and cerium with HNO3

As explained above, we assume that Ce3+ formed a stronger ion-mass bond with chert which leads to its better adsorption percentages as compared to lanthanum. So due to its higher affinity with chert, it was more difficult to desorb. Consequently, lanthanum showed the higher desorption behavior as compared to cerium under the same conditions and in all tested desorption media.

As shown in Fig. 8, calcium chloride is the most efficient eluent among all tested eluents. Indeed, 93.651% desorption of cerium was achieved at 0.01 mol L−1 CaCl2 and 96.446% desorption of lanthanum was achieved at 2 mol L−1 CaCl2. It was reported that adsorption/desorption behavior of ions is related to their physicochemical properties [41]. Diniz and Volesky [42] stated that substitution of lanthanides ions, characterized by a large ionic radius, usually involve large cations such as calcium. In the same way, Moldoveanu and Papangelakis [25] assume that cations with similar radius to those of lanthanides such as calcium are more successful in replacing them by meeting similar steric requirements on the surface. So, as it was observed in Fig. 8, we assume that when Ca2 + ions were added to the solution, they competitively substituted Ce3 + and La3 +, which led to the desorption of these lanthanides in the solution.

In other hand, as shown in Fig. 9, hydrochloric acid gave also good recovery percentages of lanthanum (77.688%, 2 M HCl) and cerium (46.237%, 4 M HCl) which means that the presence of protons (hydronium H3O+) plays a major role in desorption. Indeed, Stirk, and van Staden [39] suggested that one of the most crucial factor in the desorption process of ions is the pH, controlled by acid concentration, which influences the mechanism of proton exchange. In the same way, as Anastopoulos et al. [18] stated, REEs are in their ionic form with a positive charge at low pH values. At higher pH values REEs start to precipitate as hydroxide form. For this reason, the use of acids, such as HCl and HNO3 help to maintain the ionic form of REEs and to increase the concentration of protons. Consequently, the H+ can extract more easily the REEs, thus leading to their desorption by ion exchange mechanism.

So, based on Fig. 9, we notice that the higher the concentration of HCl, the lower the pH, the better the desorption percentages of Ce3+ and La3+.

However, as shown in Fig. 10, desorption capacity of HNO3 is apparently limited and even at high concentrations (2 and 4 mol.L−1), desorption percentages does not exceed 25%.

3.5 Effect of Agitation Rate

Kyzas confirmed [43] that agitation is an important parameter in sorption reactions, influencing the distribution of the solute in the bulk solution and the formation of the external boundary film. The rate of agitation reduced the boundary-layer resistance and increased the mobility of the system. The increase of agitation lowers the external mass transfer effect. For this reason, the influence of the agitation speed on desorption of La3+ and Ce3+ from chert was studied. The speed was varied from 0 to 600 rpm using a magnetic stirrer. The effect of the agitation rate on desorption is shown in Fig. 11. As shown in this figure, the agitation rate had no significant effect on the desorption of lanthanum and cerium from chert.
Fig. 11

Effect of agitation rate on desorption of lanthanum and cerium

3.6 Effect of Time

The required contact time between the eluent and the adsorbent vary, as for adsorption, depending on the nature of the compounds and the adsorbent. Figure 12 gives the desorption kinetics of (a) Lanthanum, (b) Cerium at optimum CaCl2, HCl and HNO3 concentrations. This Fig. 12 shows that all the desorption curves of La3+ and Ce3+ by the inorganic salt and the mineral acids could be divided in two distinct stages; a first rapid desorption stage followed by a slower one. The desorption rate of lanthanum and cerium from chert was initially fast in the first 30 min, but gradually decreased over time and then continued with a slower rate during the left time of the experiment.
Fig. 12

Desorption kinetics of a Lanthanum, b Cerium at optimum CaCl2, HCl and HNO3 concentrations

Kandpal et al. [44] attested that the rapid initial desorption rate is due to the release of ions from the water-soluble fraction and also from the lower binding energy adsorption sites. Therefore, these ions were held weakly on the sorbent surface. While, the slower desorption rate is due to the release of ions from relatively higher binding energy sites. Therefore these ions were closely bonded on the sorbent surface. So, we assume that La3+ and Ce3+ were adsorbed on chert surface by two types of sites and consequently there releases were achieved in two stages a first rapid desorption rate followed by a slow one.

3.7 Rare Earth Desorption Kinetics and Isotherms

The adsorption kinetics have been studied by many researchers, but the desorption kinetics studies remain limited [45]. So, data from desorption kinetics were processed to understand the mechanism of the desorption process in terms of order and speed limiting step. Therefore, eight kinetic models were investigated. A coefficient of determination (R2) was used as criteria for the best fit.

Desorption behaviors of La3+ and Ce3+ were somewhat alike in all different eluting media. Thus, only the desorption kinetics with 2 M CaCl2 and 0.01 M CaCl2, respectively were schematically shown in Fig. 13.
Fig. 13

Kinetics of Ce3+ and La3+ desorption by treated chert with CaCl2: a pseudo second order model, b pseudo first order model, c first order model d second order model, e parabolic diffusion model, f two constant rate model, g simple Elovich model

All these results clearly show that pseudo-second order yielded the best overall representation for the desorption of La3+ and Ce3+. Indeed the R2 values of the pseudo second-order model in all tested media were greater than 0.96 (Table 4), whereas the R2 values of all the other kinetic models were less than 0.80.
Table 4

Values of parameters obtained by pseudo-second order kinetic model


Parameters/eluting medium

R 2

Qe2 (mg g−1)

K2 [g (mg min)−1]


CaCl2/0.01 M




CaCl2/0.1 M




CaCl2/0.5 M




CaCl2/1 M




CaCl2/2 M




HCl/0.05 M




HCl/0.1 M




HCl/0.3 M




HCl/0.6 M




HCl/2 M




HCl/4 M




HNO3/0.1 M




HNO3/1 M




HNO3/2 M




HNO3/4 M





CaCl2/0.01 M




CaCl2/0.1 M




CaCl2/0.5 M




CaCl2/1 M




CaCl2/2 M




HCl/0.05 M




HCl/0.1 M




HCl/0.3 M




HCl/0.6 M




HCl/1 M




HCl/2 M




HCl/4 M




HNO3/0.1 M




HNO3/1 M




HNO3/2 M




HNO3/4 M




This result is in agreement with the results found by van der Watt and Waanders whom reported that the processes of adsorption and desorption of rare earths on siliceous clays follows the second order kinetic model [26].

However, Li et al. reported [24] that the desorption process of cerium from different soils of china were different. The Elovich equation was the best for cerium desorption in fluvo-aquic and black soil and the parabolic diffusion equation was the best model cerium desorption in the red and loess soil.

The final equilibrium concentrations (Qe2), rate constants (K2) and coefficients of determination (R2) for the pseudo-second order kinetic model are presented in Table 4.

Compared to Langmuir model, the Freundlich isotherm was fitted better to the desorption process as indicated by the R2 values (Table 5), suggesting that lanthanum and cerium release was dominated by multilayer desorption process from heterogeneous surfaces and adsorption/desorption sites provide a varying affinities with different energies [34].
Table 5

Values of parameters obtained by different isotherm models

Isotherm models












R 2







Qm (mg g−1)







B (L mg−1)








R 2







n des







K f des [(mg g−1) (L mg−1)1/n]







4 Conclusions

Chert underwent purification and treatments steps. The obtained material was characterized by XRD, XRF, SEM, FTIR and laser granulometry. The XRD, FTIR and XRF confirm that the prepared sorbent is a siliceous phase SiO2 (~ 98%) mainly composed of opal-CT (trydimite phase and cristobalite) and traces of quartz. The SEM shows that the chert morphology is granular.

In this study, batch desorption experiments for the solid phase extraction and recovery of lanthanum and cerium from aqueous solutions have been performed using treated chert as an adsorbent. The obtained results can be summarized as follows:
  • The complete recoveries of La3+ and Ce3+ are achieved by CaCl2 and the percentages of desorption approach their maximum values of, respectively 96.446% and 93.651%.

  • La3+ and Ce3+ desorption rate was initially fast and the desorption occurs in the first 30 min.

  • The equilibrium desorption behavior of these ions can be well described by the freundlich desorption isotherm.

  • All kinetics results indicate that pseudo second-order kinetic model was the best to describe La3+ and Ce3+ desorption from treated chert.

The high and rapid desorption rate obtained in this study display the good efficiency of chert as sorbent for the extraction and recovery of trace elements such as rare earths.

Since chert is naturally abundant by-product, low cost sorbent and can be used as an alternative to more costly materials, the process is low cost and can be used to extract and recover of REEs efficiently.


Compliance with Ethical Standards

Conflict of Interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.


  1. 1.
    Pecht MG, Kaczmarek RE, Song X, Hazelwood DA, Kavetsky RA, Anand DK (2012) Rare earth materials: insights and concerns. Calce Epsc Press, MarylandGoogle Scholar
  2. 2.
    Krishnamurthy N, Gupta CK (2005) Extractive metallurgy of rare earths, 2nd edn. CRC Press, Boca RatonCrossRefGoogle Scholar
  3. 3.
    Office of Research and Development Cincinnati (2012) Rare earth elements: A review of production, processing, recycling, and associated environmental issues. Environmental Protection Agency, USAGoogle Scholar
  4. 4.
    Torab-Mostaedi M (2013) Chem Ind Chem Eng Q 19:79–88CrossRefGoogle Scholar
  5. 5.
    Kim P, Anderko A, Navrotsky A, Riman RE (2018) Minerals 8:106–130CrossRefGoogle Scholar
  6. 6.
    Rout A, Binnemans K (2014) Dalton T 34:3168–3195Google Scholar
  7. 7.
    Ahmed SH, Helaly OS, Abd El-Ghany MS (2014) Int J Chem Mol Nucl Mater Metall Eng 8:866–872Google Scholar
  8. 8.
    Garcıa Fern´andez R, Garcıa Alonso JI (2008) J Chromatogr A 1180:59–65CrossRefGoogle Scholar
  9. 9.
    Zhao F, Repo E, Meng Y, Wang X, Yin D, Sillanpa M (2016) J Colloid Interf Sci 465:215–224CrossRefGoogle Scholar
  10. 10.
    Farahmand E (2016) Geology 6:189–200Google Scholar
  11. 11.
    Takahashi Y, Châtellier X, Hattori KH, Kato K, Fortin D (2005) Chem Geol 219:53–67CrossRefGoogle Scholar
  12. 12.
    Marwani HM, Alsafrani AE (2013) J Anal Sci Technol 4:13–23CrossRefGoogle Scholar
  13. 13.
    Karadaş C, Derya K (2014) Water Air Soil Pollut 225:2192–2202CrossRefGoogle Scholar
  14. 14.
    Bahramifar N, Yamini Y (2005) Anal Chim Acta 540:325–332CrossRefGoogle Scholar
  15. 15.
    Roosen J, Binnemans K (2014) J Mater Chem A 2:1530–1540CrossRefGoogle Scholar
  16. 16.
    Roosen J, Spooren J, Binnemans K (2014) J Mater Chem A 2:19415–19426CrossRefGoogle Scholar
  17. 17.
    Zolfonoun E, Yousefi SR (2016) J Braz Chem Soc 27:2348–2353Google Scholar
  18. 18.
    Anastopoulos I, Bhatnagar A, Lima EC (2016) J Mol Liq 221:954–962CrossRefGoogle Scholar
  19. 19.
    Utomo HD, Salim MR (2009) Asian J Water Environ Pollut 6:73–80Google Scholar
  20. 20.
    API: Agency for the Promotion of Industry and Innovation in Tunisia (2009) Les perspectives de développement industriel du gouvernorat de Gafsa. API/CEPI, TunisiaGoogle Scholar
  21. 21.
    Tucker ME (2001) Sedimentary petrology: an introduction to the origin of sedimentary rocks, 3rd edn. Wiley-Blackwell, Hoboken, USAGoogle Scholar
  22. 22.
    Iakovleva E, Sillanpää M (2013) Environ Sci Pollut Res Int 20:7878–7899CrossRefGoogle Scholar
  23. 23.
    Bashiri H (2012) Phys Chem 2(5):80–85CrossRefGoogle Scholar
  24. 24.
    Li D, Huang S, Wang W, Peng A (2001) Chemosphere 44:663–669CrossRefGoogle Scholar
  25. 25.
    Moldoveanu AG, Papangelakis GV (2012) Hydrometallurgy 117–118:71–78CrossRefGoogle Scholar
  26. 26.
    van der Watt JG, Waanders FB (2012) J South Afr Inst Min Metall 112:281–285Google Scholar
  27. 27.
    Reyhanitabar A, Karimian N (2008) J Agric Environ Sci 4:287–293Google Scholar
  28. 28.
    Saidi R, Tlili A, Fourati A, Ammar N, Ounis A, Jamoussi F (2012) Mater Sci Eng 28:1–8Google Scholar
  29. 29.
    Tlili A, Saidi R, Fourati A, Ammar N, Jamoussi F (2012) Appl Clay Sci 62–63:47–57CrossRefGoogle Scholar
  30. 30.
    Liang P, Liu Y, Guo L (2005) Spectrochim Acta B 60:125–129CrossRefGoogle Scholar
  31. 31.
    Wankasi D, Horsfall MJ, Spiff AI (2005) Afr J Biotechnol 4:923–927Google Scholar
  32. 32.
    Al-Bayati RA, Ahmed SA (2011) Int J Chem 3:21–30Google Scholar
  33. 33.
    Shirvani M, Shariatmadari H, Kalbasi M (2007) Appl Clay Sci 37:175–184CrossRefGoogle Scholar
  34. 34.
    Ponou J, Dodbiba G, Anh JW, Fujita T (2016) J Environ Chem Eng 4:3761–3766CrossRefGoogle Scholar
  35. 35.
    Rani S, Sud D (2015) Plant Soil Environ 61:36–42CrossRefGoogle Scholar
  36. 36.
    Smith CB (1998) Infrared spectral interpretation: a systematic approach. CRC Press, Boca RatonGoogle Scholar
  37. 37.
    Hughes ID, Däne M, Ernst A, Hergert W, Lüders M, Poulter J, Staunton JB, Svane A, Szotek Z, Temmerman WM (2007) Nature 446:650–653CrossRefGoogle Scholar
  38. 38.
    Wen B, Shan XQ, Lin JM, Tang GG, Bai NB, Yuan DA (2002) Soil Sci Soc Am J 66:1198–1206CrossRefGoogle Scholar
  39. 39.
    Stirk AW, van Staden J (2002) Bot Mar 45:9–16CrossRefGoogle Scholar
  40. 40.
    Hubbe MA, Hasan SH, Ducoste JJ (2011) BioResources 6:2161–2287Google Scholar
  41. 41.
    Kang SY, Lee JU, Moon SH, Kim KW (2004) Chemosphere 56:141–147CrossRefGoogle Scholar
  42. 42.
    Diniz V, Volesky B (2006) Sep Purif Technol 50:71–76CrossRefGoogle Scholar
  43. 43.
    Kyzas ZG (2012) Materials 5:1826–1840CrossRefGoogle Scholar
  44. 44.
    Kandpal G, Srivastava PC, RAM B (2005) Water Air Soil Pollut 161:353–363CrossRefGoogle Scholar
  45. 45.
    Teng Y, Liu Z, Xu G, Zhang K (2017) Energies 10:115–128CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Imen Bouchmila
    • 1
  • Bochra Bejaoui Kefi
    • 1
    • 2
  • Radhia Souissi
    • 1
  • Mohieddine Abdellaoui
    • 1
  1. 1.Laboratoire des Matériaux UtilesInstitut National de Recherche et d’Analyse Physico-ChimiqueSidi ThabetTunisia
  2. 2.Faculté des Sciences de BizerteUniversité de CarthageBizerteTunisia

Personalised recommendations