Macroscopic and microscopic insight into the mutual effects of europium(III) and phosphate on their interaction with graphene oxide

Abstract

Graphene oxide (GO) has been proved to be very efficient for radionuclide enrichment. However, its potential performance for binary sorbate systems containing both a radionuclide and phosphate receives little attention. In this study, the mutual effects of Eu(III) and phosphate on their co-removal processes by GO as a function of contact time, pH and temperature were investigated in detail. The results show that the presence of phosphate can enhance not only the removal amount of Eu(III) but also the removal kinetics of Eu(III) on GO surfaces, and vice versa. Both the external mass transfer and diffusion of solute inside GO are important for the Eu(III) and phosphate co-removal rate. Evaluation of the thermodynamic parameters (ΔG⁰ < 0, ΔH⁰ > 0 and ΔS⁰ > 0) indicates the co-removal process is endothermic and spontaneous. The results of X-ray photoelectron spectroscopy (XPS) and zeta potential (ZP) analyses indicate that the oxygen-containing functional groups on GO surfaces play an important role in the co-removal process. This study provides a quantitative investigation of Eu(III) and phosphate with GO, which can facilitate the assessment of GO as a potential adsorbent for Eu(III) and phosphate simultaneous cleanup.

---

1. Introduction

Radionuclide contamination generated from the development of the nuclear industry poses serious threats to aquatic habitats and living organisms due to its potential toxic and carcinogenic effects.^1,2 The radionuclides with high mobility can be accumulated through the food chain even at very low concentrations, ultimately becoming a part of animals and human beings.³ Over the past few decades, approximately 10.7 kg of lanthanides and actinides are present in each ton of used nuclear fuel.⁴ Therefore, the highly efficient elimination and long-term storage of the long-lived radionuclides from large volumes of nuclear wastewater have currently become an issue of great environmental concern. However, the high radioactivity and biological toxicity limit some lanthanides and actinides (such as ²³⁶Np, ²⁴¹Am, ²⁴³Am, ²⁴³Cm, ²⁴⁴Pu, etc.) for use for experimental investigation in laboratories.⁵ Europium (Eu(III)) has been used as a chemical analog for the trivalent lanthanides and actinides such as Am(III) and Cm(III).^6,7 Phosphate is deemed as a double-edged nutrient for both the growth of organisms and eutrophication of rivers and lakes in most ecosystems.⁸ Phosphate may be present in large amounts as a result of TBP (tributyl phosphate used in the PUREX process in reprocessing plants) degradation by radiolysis.⁹ Therefore, radionuclide and phosphate are likely to co-exist in the effluent of reprocessing plants. The development of a facile and efficient approach to simultaneously remove both Eu(III) and phosphate from nuclear wastewater is urgent and crucial.

Owing to the excellent water-solubility, large specific surface area (theoretical value of 2620 m² g⁻¹), and enriched oxygenated functional groups (such as carboxyl, epoxy and hydroxyl), graphene oxide (GO) has been reported of highly efficient sorption capacities for radionuclides.^10–12 Zhao et al.¹⁰ firstly applied GO to remove U(VI) from aqueous solutions, and found the sorption process was mainly dominated by inner-sphere surface complexation. Sun et al.¹¹ studied the interaction of Eu(III) with GO by extended X-ray absorption fine structure (EXAFS) spectroscopy, and found the sorption capacity of Eu(III) on GO was 1.15 mmol g⁻¹ at pH 6.0. Romanchuk et al.¹² reported the prepared GO using improved Hummers' method demonstrated high sorption affinity towards actinides (i.e., U(VI), Sr(II), Am(III) and Eu(III)). However, few studies focused on the application of GO as the potential adsorbent in the binary sorbate systems containing both radionuclides and phosphate.¹³ Phosphate can readily form complexes with many radionuclides and interact with scavengers, affecting their surface charge and overall reactivity.¹⁴ Both of these may potentially affect the sorption of Eu(III). Therefore, investigations that incorporate phosphate into the co-removal process may have more realistic implications for Eu(III) cleanup in nuclear wastewater than those without co-adsorbing anions.

The purposes of this study are: (1) to synthesize GO and apply it as the adsorbent for the co-removal of Eu(III) and phosphate; (2) to elaborate the effects of contact time, pH and temperature on the co-removal process; and (3) to investigate the possible sorption mechanisms of Eu(III) and phosphate on GO. This study provides new insights of GO as the potential adsorbent for Eu(III) and phosphate co-removal by adopting batch techniques, which can broaden the applicability of graphene derivatives in simultaneous radionuclide and oxyanion pollution cleanup.

2. Material and method

2.1. Chemicals and adsorbent preparation

Expandable graphite (<20 μm) was provided from Qingdao Tianhe Graphite Co., Ltd (China). The expandable graphite was used as the raw material instead of flake graphite to ensure more uniform oxidization.^15,16 The chemicals including H₂SO₄, NaNO₃, KMnO₄, H₂O₂, NaBH₄, NaH₂PO₄·2H₂O, NaCl, HCl and NaOH were purchased from Sinopharm Chemical Reagent Co., Ltd. All chemicals used in the experiments were of analytical grade and used directly without further purification. Milli-Q water (18.2 MΩ cm⁻¹) was used to prepare the solutions throughout the experiments.

The GO was synthesized via the chemical oxidation of expandable graphite by employing the modified Hummers' method.¹⁷ The stock GO suspension had a dark brown color and did not form coagulation even after 6 months of aging time.^10,18 The homogeneous dispersion property of GO in aqueous solutions will make the oxygen-containing functional groups on GO surfaces freely available to form strong surface complexes with Eu(III).

2.2. Characterization

The obtained GO samples were characterized by scanning electron microscopy (SEM), Fourier transform infrared (FT-IR) spectroscopy, Raman spectroscopy, and UV-vis absorbance spectroscopy techniques. The GO samples before and after the co-removal process of Eu(III) and phosphate were characterized by energy dispersive spectroscopy (EDS), X-ray photoelectron spectroscopy (XPS) and zeta potential (ZP) techniques. The SEM image was obtained using a JEOL JSM-6700F instrument operated at the beam energy of 15.0 kV. The FT-IR measurement was recorded in pressed KBr pellets (Aldrich, 99%, analytical reagent) by using a PerkinElmer Spectrum 100 system spectrometer at room temperature. The Raman spectrum was conducted using a LabRam HR Raman spectrometer with excitation at 514.5 nm for 10 s by Ar⁺ laser to avoid overheating of the GO. The absorbance of the GO sample in aqueous solution (∼25 mg L⁻¹) was characterized using UV-vis spectroscopy (UV-2550, PerkinElmer). The ZP values of the GO samples in the absence and presence of Eu(III) or/and phosphate were recorded as a function of pH using a Zetasizer Nano-ZS90 Instrument (Malvern Co., U.K.) at T = 298 K. The XPS spectra were conducted using a VG Scientific ESCALAB Mark II spectrometer equipped with two ultrahigh vacuum chambers. The peak energies of XPS spectra were corrected with C 1s peak at 284.6 eV as a reference.¹⁶ The deconvolution of C 1s and O 1s lines were performed using XPSPEAK41 program after subtraction of the background (Shirley baseline correction).

2.3. Batch sorption experiments

The sorption of Eu(III) and/or phosphate on GO were carried out by batch experiments at T = 293, 313 and 333 K. In order to achieve homogeneous dispersion, the stock suspension of GO was sonicated for 15 min in an ultrasonic bath before use. Appropriate amounts of NaCl (0.1 mol L⁻¹), Eu(III) (0.5 mmol L⁻¹) and/or phosphate (0.5 mmol P L⁻¹) stock solutions and the stock suspension of GO (100 mg L⁻¹) were added into a 10 mL polyethylene test tube to achieve the desired concentrations of different components. After adding the above components and controlling the final volume of suspension in the test tube to 6 mL, the initial concentrations of GO, NaCl, Eu(III) and phosphate in the test tube were 25 mg L⁻¹, 0.01 mol L⁻¹, 0.05 mmol L⁻¹ and 0.05 mmol P L⁻¹, respectively. The desired pH value of suspension was adjusted by adding negligible volumes of 0.1/0.01 mol L⁻¹ HCl or NaOH solution. Additionally, in order to find the difference between individual and co-existing systems, single Eu(III)/phosphate sorption as a function of contact time was also investigated at T = 293 K. After the suspension was shaken for 24 h, the solid and liquid phases were separated by the centrifugation at 9500 rpm for 30 min, and then the supernatant was filtered through a 0.22 μm membrane filter. The concentration of Eu(III) in the supernatant was measured by monitoring the Eu(III)–chlorophosphonazo complex by spectrophotometry at a wavelength of 632 nm based on GB/T 18116.2-2008. The concentration of phosphate in the supernatant was measured by monitoring the blue phosphate–molybdate complex at 700 nm based on GB/T 11893-1989. The co-existing phosphate will not affect the quantitative detection result of Eu(III), and vice versa (Fig. S1 in ESI†).

The sorption percentage (%), distribution coefficient (K_d), and the amounts of Eu(III) or phosphate adsorbed on GO (C_s) are calculated from the following equations, respectively:

where C₀ (mmol L⁻¹) is the initial concentration of Eu(III)/phosphate in suspension, C_e (mmol L⁻¹) is the equilibrium one in supernatant, C_s (mmol g⁻¹) is the concentration of ions adsorbed on solid phase, V (L) is the volume of suspension, and m (g) is the mass of adsorbent. All of the experimental data are the averages of triplicate determinations. The relative error values of data are less than 5%.

3. Results and discussion

3.1. Characterization

Fig. 1a shows the SEM image of the obtained GO samples. The wrinkled and aggregated ripples on GO surfaces are observed, which is consistent with the previous report.¹⁹ The oxygen-containing functional groups on GO surfaces are characterized by FT-IR technique (Fig. 1b). The absorption bands at 1077, 1624, 1725 and 3423 cm⁻¹ correspond to C–O bonds, sp² C=C bonds, C=O stretching vibrations and –OH stretching vibrations, respectively.¹⁹ The FT-IR spectrum indicates that a large amount of oxygen-containing functional groups (i.e., carbonyl, carboxyl, epoxy and hydroxyl) present on GO surfaces. In the Raman spectrum (Fig. 1c), the G band at ∼1580 cm⁻¹ is associated with the vibration of sp² carbon atoms in a graphitic 2D hexagonal lattice, and the D band at ∼1350 cm⁻¹ is related to the vibration of sp³ carbon atoms of defects. The D band is resulted from the structural defection created by the attachment of epoxy and hydroxyl functional groups on the carbon basal plane.¹⁶ The strong intensities of G and D bands are consistent with previous GO characterizations.^10,17 According to the UV-absorbance spectrum (Fig. 1d), the maximum peak at 227 nm corresponds to π–π stretching vibrations of aromatic C=C bonds.¹⁶ The high resolution scans for C 1s and O 1s are shown in Fig. 1e and f, respectively. The C 1s XPS spectrum of GO presents two separate peaks due to the high percentage of oxygenated functionalities. The deconvolution of C 1s spectrum indicates the presence of C–C at 284.8 eV, C–O groups at 287.1 eV, and –COOH groups at 288.1 eV, which can be comparable to the previous reports.^18,19 The deconvolution of O 1s spectrum exhibits two peaks with binding energies of 531.8 and 532.9 eV, which are assigned to C=O and C–O, respectively.¹⁸ From the C 1s and O 1s XPS spectra, the considerable degree oxygen-containing functional groups on GO surfaces are C–O, C=O and O=C–O, which have potentiality to form strong complexes with radionuclides.

Fig. 1 Characterizations of GO: (a) SEM image, (b) FT-IR spectrum, (c) Raman spectrum, (d) UV-vis absorbance spectrum and high resolution XPS spectra for (e) C 1s and (f) O 1s.

3.2. Eu(III) sorption

3.2.1. Effect of temperature on Eu(III) sorption kinetics. The kinetic data of Eu(III) sorption on GO in the absence and presence of phosphate (0.05 mmol P L⁻¹) at different temperatures are shown in Fig. 2. In the absence of phosphate, the sorption process is rapid to reach equilibrium for Eu(III) within 2 h, which is consistent with the previous study.²⁰ In the presence of phosphate (the molar ratio of Eu(III) to phosphate is 1 : 1), the sorption process increases much more rapidly in the initial 30 min of contact time at the lowest temperature (T = 293 K), and then maintains a high level until the sorption process achieves equilibrium within 1 h. Fig. 2 also shows that the sorption capacity of GO for Eu(III) in the presence of phosphate is clearly influenced by temperature. The higher temperature, the more Eu(III) sorption, suggesting that higher temperature favors Eu(III) sorption in ternary system containing Eu(III), phosphate and GO. In general, the sorption process of Eu(III) is rapid and 2 h is enough to achieve sorption equilibrium. Based on the sorption kinetic data, a shaking time of 24 h is chosen in the following experiments to ensure that the sorption process could achieve complete equilibrium.

Fig. 2 Effect of contact time on Eu(III) retained by GO in the absence and presence of phosphate (0.05 mmol P L⁻¹) at different temperatures. C_{[Eu(III)]initial} = 0.05 mmol L⁻¹, pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl.

The migration process and overall sorption rate of the sorbate at water–solid interface are controlled by the surface characteristics and diffusion resistances of the solid particles.²¹ Therefore, the utilization of a suitable kinetic model can provide useful information in determining the underlying mechanism during the entire sorption process. The experimental kinetic data of Eu(III) sorption are simulated by the pseudo-first-order and pseudo-second-order models using the following equations, respectively:^22,23

ln(Q_m − Q_t) = ln Q_m − k′t (4)

where Q_t and Q_m (mmol g⁻¹) refer to the sorption capacity of sorbate at time t (h) and equilibrium sorption ability obtained from the kinetic model, respectively. k′ (h⁻¹) and k′′ (g mmol⁻¹ h⁻¹) are the sorption rate constants.

The fitting results of pseudo-first-order and pseudo-second-order models are shown in Fig. 3a and b, respectively, and the corresponding kinetic parameters are listed in Table 1. As shown in Table 1, the pseudo-second-order model (eqn (5)) simulates the kinetic data better than the pseudo-first-order model (eqn (4)) with higher correlation coefficients and the calculated equilibrium values (Q_mc) closer to the experimental ones (Q_me), which highlights the chemisorption-controlled mechanism.²⁴ The values of Q_mc (1.66 mmol g⁻¹) and k′′ (4.56 g mmol⁻¹ h⁻¹) in the presence of phosphate are much higher than the ones in the absence of phosphate, respectively, suggesting the presence of phosphate can enhance not only the removal amount of Eu(III) but also the removal kinetics of Eu(III) on GO surfaces.

Fig. 3 Tests of (a) pseudo-first-order model and (b) pseudo-second-order model on Eu(III) sorption retained by GO in the absence and presence of phosphate (0.05 mmol P L⁻¹) at different temperatures. C_{[Eu(III)]initial} = 0.05 mmol L⁻¹, pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl.

Table 1 Parameters of Eu(III) and phosphate simulated by different kinetic models as a function of contact time at different temperatures^a

Parameters	Eu(III)				Phosphate
	293^b K	293 K	313 K	333 K	293 K	313 K	333 K
a Qmc is the sorption capacity at equilibrium calculated from different kinetic models, Qme is the experimental value.b Parameters of single Eu(III) sorption simulated by different kinetic models.
Pseudo-first-order model
k′ (h^−1)	0.16	0.17	0.12	0.07	0.18	0.15	0.15
Qmc (mmol g^−1)	0.08	0.07	0.06	0.08	0.11	0.08	0.07
R^2	0.489	0.391	0.282	0.361	0.428	0.384	0.302
Pseudo-second-order model
k′′ (g mmol^−1 h^−1)	2.59	4.56	6.65	4.85	1.50	2.79	3.14
Qmc (mmol g^−1)	0.26	1.66	1.75	1.84	0.54	0.59	0.66
R^2	0.995	0.998	0.999	0.999	0.997	0.999	0.998
Qme (mmol g^−1)	0.27	1.67	1.78	1.88	0.56	0.61	0.67

---

3.2.2. Rate-limiting step determination of Eu(III) sorption. The kinetic studies of sorption are significant in the cases of following perspectives:²⁵

(1) The kinetic data can be used in determining the time required to achieve sorption equilibrium.

(2) Sorption rate calculated from the kinetic study can be used to deduce predictive models for batch experiments.

(3) Kinetic studies can be used in understanding the effects of variables on the sorption of solutes.

Generally, the sorption kinetics is controlled by the film diffusion step, particle diffusion step, and sequestration on the active sites via sorption, complexation or intraparticle precipitation.²⁶ The last step is supposed to be very rapid and can be recognized as negligible. For the purpose of identifying the slowest step in sorption process, it is necessary to differentiate film diffusion step from particle diffusion step. Sorption rate usually depends on parameters such as metallic ions concentration, structural properties of adsorbent (specific area, porosity, particle size, etc.), the properties of metallic ions (ionic radius, coordination numbers, and speciation), interaction between active sites of adsorbent and metallic ions.^25,26 If intraparticle diffusion plays the dominant role in sorption process, it should be well fitted by the relationship between Q_t and the square root of time (t^0.5) as follows:²⁵

where K_I (mmol min^−0.5) is the initial rate of intraparticle diffusion.

As shown in Fig. 4a, a nonlinear distribution of points with two distinct portions is observed. The double nature of the plots (curved and linear) obtained from Eu(III) sorption on GO indicates the existence of intraparticle diffusion. According to intraparticle diffusion model,²⁶ if there is a linear relationship between Q_t and t^0.5, it will indicate that intraparticle/pore diffusion is the rate-limiting step in sorption process. However, the plots shown in Fig. 4a contrast the prediction of intraparticle diffusion model, suggesting that intraparticle/pore diffusion may participate in the sorption process but is not the single rate-limiting step. The initial curved part of plots is attributed to film diffusion, the linear part, to intraparticle diffusion.²⁵

Fig. 4 (a) Intraparticle diffusion rate determination and (b) test of Urano and Tachikawa model on Eu(III) sorption retained by GO in the absence and presence of phosphate (0.05 mmol P L⁻¹) at different temperatures. C_{[Eu(III)]initial} = 0.05 mmol L⁻¹, pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl.

Because of the multistep nature of these plots, the linear portions are modeled by Urano and Tachikawa model, which is shown as follows:²⁷

where π (3.1416) is the constant of Pi, D_I (cm² min⁻¹) is the intraparticle diffusion coefficient based on the concentration in solids, d (cm) is the diameter of adsorbent.

According to Urano and Tachikawa model,²⁷ if the plots lie in a linear form and pass through origin, then the rate-limiting step of sorption process is internal diffusion, and vice versa. From Fig. 4b, it is obvious that only the plot of Eu(III) sorption on GO in the presence of phosphate at 313 K is linear but does not pass through the origin, suggesting that the rate-limiting step is film diffusion. Other plots are not linear, suggesting that either internal diffusion or film diffusion is not the rate-limiting step. It can be assumed that both the external mass transfer and the diffusion of solute inside GO are important for the single Eu(III) sorption rate and for Eu(III) and phosphate co-removal rate.

3.2.3. Effect of initial pH on Eu(III) sorption. The solution pH affects the species of Eu(III) and phosphate and the protonation degree of functional groups on GO surfaces, so pH is one of the most important factors controlling the co-removal process of Eu(III) and phosphate.⁵ Effect of pH on Eu(III) enrichment by GO in the absence and presence of phosphate (0.05 mmol P L⁻¹) was investigated in the pH range of 2.0–12.0. Fig. 5 shows that in the absence of phosphate, the sorption of Eu(III) on GO increases with increasing pH values (approximately lower than 6.0), and then maintains the high level, as expect for the radionuclide sequestrations by GO.^12,20 In the presence of phosphate, the sorption edge of Eu(III) shifts to lower pH values. The presence of phosphate does not change the shape of pH-dependent sorption edge of Eu(III) but moves it to lower pH values, suggesting that the presence of phosphate has a positive effect on Eu(III) sorption, which is consistent with the previous studies.^28,29 There are three possible reasons for the increased Eu(III) removal by phosphate: (1) Eu(III) can form complex with phosphate, which may have higher affinity to GO; (2) phosphate can be captured by the adsorbed Eu(III) and acts as the new active site for continued removal of the dissolved Eu(III) ions in solution; (3) Eu–P complex can precipitate from solution.

Fig. 5 Effect of pH on Eu(III) retained by GO in the absence and presence of phosphate (0.05 mmol P L⁻¹). C_{[Eu(III)]initial} = 0.05 mmol L⁻¹, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl and equilibrium time = 24 h.

3.2.4. Sorption isotherms of Eu(III) on GO. Sorption isotherms of Eu(III) retained by GO in the presence of phosphate (0.05 mmol P L⁻¹) at different temperatures are shown in Fig. 6. The experimental data are simulated by the Langmuir and Freundlich isotherm models to obtain a better understanding of the sorption mechanism. Based on the assumption that adsorbent is structurally homogeneous, and the sorption process occurs on identically and energetically equivalent sorption sites, the Langmuir model is widely used for the fitting of a monolayer sorption.³⁰ Unlike the Langmuir model, the Freundlich model is a semi-empirical equation based on the assumption that the sorption phenomenon occurs on heterogeneous surfaces.³¹ The Langmuir and Freundlich models are expressed as follows, respectively:

C_s = K_FC_eⁿ (9)

where C_s,max (mmol g⁻¹) is the maximum amount of ions per unit weight of adsorbent to form a complete monolayer coverage on the surface, K_L (L mmol⁻¹) is the Langmuir constant related to the energy of sorption, K_F (mmol¹⁻ⁿ Lⁿ g⁻¹) represents the sorption capacity when the equilibrium concentration of radionuclides equals to 1, and n represents the dependence degree of sorption with equilibrium concentration.

Fig. 6 Sorption isotherms of Eu(III) retained by GO in the presence of phosphate (0.05 mmol P L⁻¹) at different temperatures. Symbols denote experimental data, solid and dash lines correspond to the model fitting of Langmuir and Freundlich equations, respectively. pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl and equilibrium time = 24 h.

The parameters calculated from these two models are listed in Table 2. It is clear that the Langmuir model fits the sorption isotherms better than Freundlich model with higher correlation coefficients, suggesting surface sites are homogeneous and that monolayer sorption is occurring for Eu(III) sorption in the binary and ternary systems.³² Based on the Langmuir model, the C_s,max values for Eu(III) sorption are the highest at T = 333 K and the lowest at T = 293 K, suggesting that higher temperature is more conducive to Eu(III) sorption. In the Freundlich model, n is lower than 1, suggesting that a nonlinear sorption occurs on GO surfaces. These results indicate that Eu(III) sorption should be associated with chemical sorption rather than physical sorption, which is consistent with the result of kinetic study. As compared with the adsorbents previously reported in other literatures (Table 3), GO possesses higher Eu(III) sorption capacity in the presence of phosphate than most of the common adsorbents.^6,7,11,12,33

Table 2 Parameters of Eu(III) and phosphate sorption isotherms modeled by the Langmuir and Freundlich equations at different temperatures

Parameters	Eu(III)			Phosphate
	293 K	313 K	333 K	293 K	313 K	333 K
Langmuir isotherm fit
Cs,max (mmol g^−1)	2.51	2.93	3.38	0.73	0.86	1.09
KL (L mg^−1)	41.33	50.50	67.80	19.95	16.05	11.60
R^2	0.996	0.996	0.998	0.996	0.996	0.998
RL	0.93	0.98	0.93	0.95	0.95	0.93
Freundlich isotherm fit
n	0.39	0.37	0.34	0.34	0.40	0.51
KF (mmol^1−n L^n g^−1)	4.57	5.39	6.21	0.95	1.23	1.75
R^2	0.946	0.945	0.938	0.945	0.960	0.974

---

Table 3 Comparison of GO sorption capacities for Eu(III) and phosphate with adsorbents reported in other literatures

Adsorbent	Experimental condition	Cs,max (mmol g^−1)	Ref.
Eu(III)
Bentonite–PAAm composites	pH = 3.5, T = 293 K	0.28	6
GO@TiP nanocomposites	pH = 5.5, T = 298 K	0.42	7
Ordered mesoporous carbon	pH = 4.0, T = 298 K	0.91	33
GO	pH = 5.0, T = 298 K	0.76	12
GO	pH = 6.0, T = 298 K	1.15	11
GO	pH = 4.0, T = 293 K, in the presence of 0.05 mmol L^−1 phosphate	2.51	This work
Phosphate
Iron oxide tailings	pH = 6.6–6.8, T = 293 K	0.26	39
MgO/CaO-loaded bentonite	pH = 7.5, T = 298 K	0.37	40
Magnetic Fe–Zr binary oxide	pH = 4.0, T = 298 K	0.44	41
γ-Al2O3	pH = 5.5, T = 298 K, in the presence of 0.09 mmol L^−1 Cu(II)	0.14	29
Natroalunite microtubes	pH = 7.0, T = 298 K, in the presence of 0.2 mmol L^−1 Cd(II)	0.72	32
GO	pH = 4.0, T = 293 K, in the presence of 0.05 mmol L^−1 Eu(III)	0.73	This study

---

In order to examine whether the sorption is favorable or not, the essential characteristics of the Langmuir model can be expressed in terms of a dimensionless constant separation factor or equilibrium parameter, R_L, which is given by the following equation:³²

According to Hameed et al.,³⁴ the value of R_L indicates the sorption to be unfavorable (R_L > 1), linear (R_L = 1), favorable (0 < R_L < 1) or irreversible (R_L = 0). The relevant R_L parameters listed in Table 2 fall in the range of 0–1, suggesting that the sorption of Eu(III) on GO is favorable, and Eu(III) ions tend to remain binding on GO surfaces.³⁴

3.2.5. Thermodynamic study of Eu(III) sorption on GO. To evaluate the thermodynamic feasibility of the sorption process of Eu(III) in the presence of phosphate, three thermodynamic parameters were calculated from the sorption isotherms. The parameters including standard Gibbs free energy (ΔG⁰, kJ mol⁻¹), standard enthalpy change (ΔH⁰, kJ mol⁻¹) and standard entropy change (ΔS⁰, J mol⁻¹ K⁻¹) are useful in defining whether the sorption process is endothermic or exothermic, spontaneous or not. The standard Gibbs free energy (ΔG⁰) can be calculated from the equation as follows:

ΔG⁰ = −RT ln K⁰ (11)

where R (8.3145 J mol⁻¹ K⁻¹) is the ideal gas constant, T (K) is the absolute temperature in Kelvin. The sorption equilibrium constant (K⁰) can be calculated by plotting ln K_d versus C_e and extrapolating C_e to zero.

The standard enthalpy change (ΔH⁰) and standard entropy change (ΔS⁰) are calculated from the following equation:

where the slope and intercept of the plot of ln K⁰ versus 1/T are −ΔH⁰/R and ΔS⁰/R, respectively.

Linear plots of ln K_d versus C_e and ln K⁰ versus 1/T for the sorption of Eu(III) retained by GO in the presence of phosphate (0.05 mmol P L⁻¹) at different temperatures are shown in Fig. 7a and b, respectively. Thermodynamic parameters of Eu(III) sorption are listed in Table 4. The negative ΔG⁰ values (−8.81, −12.46 and −16.61 kJ mol⁻¹ at 293, 313 and 333 K, respectively) indicate that the co-removal process on GO surfaces is favorable and spontaneous. The positive ΔH⁰ value (48.22 kJ mol⁻¹) confirms the co-removal process to be endothermic, which is consistent with the results of kinetic and sorption isotherm studies. According to the previous studies,^35,36 the positive value of ΔS⁰ reflects the increased disorder at the water–solid interface with some structural changes in the adsorbent and sorbate during the sorption process, and an affinity of the adsorbent towards sorbate. In our case, the value of ΔS⁰ is 194.40 J mol⁻¹ K⁻¹, which reflects the high affinity of GO towards Eu(III) in the presence of phosphate, and also indicates an increase in the degree of freedom of the adsorbed species.

Fig. 7 (a) Linear plots of ln K_d versus C_e for the sorption of Eu(III) retained by GO in the presence of phosphate (0.05 mmol P L⁻¹) at different temperatures. (b) Linear plots of ln K⁰ versus 1/T for the sorption of Eu(III) retained by GO in the presence of phosphate (0.05 mmol P L⁻¹). pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl and equilibrium time = 24 h.

Table 4 Thermodynamic parameters of Eu(III) and phosphate at different temperatures in co-existing systems

Temperature	Eu(III)			Phosphate
	293 K	313 K	333 K	293 K	313 K	333 K
ΔG^0 (kJ mol^−1)	−8.81	−12.46	−16.61	−4.26	−6.03	−8.42
ΔH^0 (kJ mol^−1)		48.22			26.07
ΔS^0 (J mol^−1 K^−1)		194.40			103.21

---

3.3. Phosphate sorption

3.3.1. Effect of temperature on phosphate sorption kinetics. The kinetic data of phosphate sorption on GO in the absence and presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures are shown in Fig. 8. In the absence of Eu(III), no obvious sorption of phosphate is observed, suggesting low sorption capacity of GO for phosphate. In the presence of Eu(III), the sorption of phosphate increases rapidly at initial stages, and then maintains a high level after 2 h. Fig. 8 also shows that higher temperature favors phosphate sorption. Generally, a shaking time of 24 h is enough to achieve sorption equilibrium, and the presence of Eu(III) changes the sorption characteristic of GO for phosphate obviously.

Fig. 8 Effect of contact time on phosphate retained by GO in the absence and presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures. C_{[phosphate]initial} = 0.05 mmol P L⁻¹, pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl.

Two kinetic model simulations (i.e., pseudo-first-order and pseudo-second-order models) for phosphate sorption on GO in the presence of Eu(III) at different temperatures are shown in Fig. 9a and b, respectively, and the corresponding kinetic parameters are listed in Table 1. The correlation coefficients of the pseudo-second-order rate equation (eqn (5)) are close to 1, which is higher than those of the pseudo-first-order model (eqn (4)). Meanwhile, the Q_mc values of phosphate sorption calculated from the pseudo-second-order model are closer to Q_me. Both suggest that the pseudo-second-order model simulates the kinetic data better, highlighting the chemisorption-controlled mechanism.

Fig. 9 Tests of (a) pseudo-first-order model and (b) pseudo-second-order model on phosphate sorption retained by GO in the presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures. C_{[phosphate]initial} = 0.05 mmol P L⁻¹, pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl.

3.3.2. Rate-limiting step determination of phosphate sorption. Based on eqn (6), plots of Q_t versus t^0.5 for the sorption of phosphate on GO in the presence of Eu(III) (0.05 mmol L⁻¹) are shown in Fig. 10a. A nonlinear distribution of points with two distinct portions is observed, suggesting that intraparticle/pore diffusion is not the single rate-limiting step in phosphate sorption process. With the linear portions linearized by Urano and Tachikawa model, the plots of −log[1 − (Q_t/Q_m)²] versus t (Fig. 10b) are not linear and do not pass through origin, suggesting that the rate-limiting step of the co-sorption process is not attributed to either internal diffusion or film diffusion.

Fig. 10 (a) Intraparticle diffusion rate determination and (b) test of Urano and Tachikawa model on phosphate sorption retained by GO in the presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures. C_{[phosphate]initial} = 0.05 mmol P L⁻¹, pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl.

3.3.3. Effect of initial pH on phosphate sorption. Effect of pH on phosphate retained by GO in the absence and presence of Eu(III) (0.05 mmol L⁻¹) was investigated and shown in Fig. 11. In the absence of Eu(III), nearly no obvious sorption of phosphate is observed. However, the sorption characteristic of GO for phosphate in the presence of Eu(III) changes obviously. The presence of Eu(III) increases the electrostatic potential on GO surfaces, which has a positive effect on phosphate sorption. One can see that the optimum pH for the retention of phosphate appeared at pH 5.9–9.0. At pH < 5.9, the adsorbed amount of phosphate decreases with the decrease of pH values. This might be attributed to the decrease of H₂PO₄⁻ proportion with a higher affinity to Eu(III) or GO–Eu, and to the formation of H₃PO₄ reducing the coulombic attraction between phosphate and Eu(III) or GO–Eu.^37,38 At pH > 9.0, the sorption of phosphate on GO is not favored with further increasing pH. OH⁻ competes strongly with phosphate for the limited active sorption sites, which might contribute to the decrease of phosphate uptake under alkaline conditions.³²

Fig. 11 Effect of pH on phosphate retained by GO in the absence and presence of Eu(III) (0.05 mmol L⁻¹). C_{[phosphate]initial} = 0.05 mmol P L⁻¹, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl and equilibrium time = 24 h.

3.3.4. Sorption isotherms of phosphate on GO. Sorption isotherms of phosphate retained by GO in the presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures are shown in Fig. 12. One can see that the sorption isotherm of phosphate at 333 K is the highest, and that of phosphate at 293 K is the lowest, suggesting that higher temperature favors phosphate sorption. As can be seen from the correlation coefficients (Table 2), Langmuir model fits the sorption isotherms of phosphate better than Freundlich model. The results indicate that the sorption of phosphate on GO surfaces is monolayer coverage. The obtained R_L parameters of phosphate sorption isotherms are listed in Table 2, which indicates that the sorption of phosphate on GO is favorable. From Table 3, it is obvious that GO possesses higher phosphate sorption capacity in the presence of Eu(III) than common adsorbents.^{29,32,39–41}

Fig. 12 Sorption isotherms of phosphate retained by GO in the presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures. Symbols denote experimental data, solid and dash lines correspond to the model fitting of Langmuir and Freundlich equations, respectively. pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl and equilibrium time = 24 h.

3.3.5. Thermodynamic study of phosphate sorption on GO. Linear plots of ln K_d versus C_e and ln K⁰ versus 1/T for the sorption of phosphate retained by GO in the presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures are shown in Fig. 13a and b, respectively. Thermodynamic parameters of phosphate sorption are listed in Table 4. With the increasing temperature, the calculated ΔG⁰ values decrease, with values of −4.26, −6.03 and −8.42 kJ mol⁻¹ at 293, 313 and 333 K, respectively. The negative ΔG⁰ values indicate that the co-removal process is favorable and spontaneous. ΔH⁰ and ΔS⁰ are calculated to be 26.07 kJ mol⁻¹ and 103.21 J mol⁻¹ K⁻¹, respectively. The positive ΔH⁰ value further confirms the sorption of phosphate to be endothermic, which is consistent with the results of kinetic and sorption isotherm studies.

Fig. 13 (a) Linear plots of ln K_d versus C_e for the sorption of phosphate retained by GO in the presence of Eu(III) (0.05 mmol L⁻¹) at different temperatures. (b) Linear plots of ln K⁰ versus 1/T for the sorption of phosphate retained by GO in the presence of Eu(III) (0.05 mmol L⁻¹). pH = 4.0 ± 0.1, m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl and equilibrium time = 24 h.

3.4. Interaction mechanisms of Eu(III) and phosphate with GO

3.4.1. XPS analysis. After the interaction with Eu(III) and phosphate, obvious visual changes of the suspension resulted from GO coagulation are observed (Fig. S2a in ESI†). The SEM image of coagulated GO and EDS patterns of GO before and after the co-removal of Eu(III) and phosphate are shown in Fig. S2b–d in ESI,† respectively. As shown in Fig. S2c in ESI,† the essential elements existing on GO surfaces are C and O (inset), referring to the abundant oxygen-containing functional groups on GO surfaces. After the co-removal process, the element distributions vary significantly (Fig. S2d in ESI†). New signals of Na and Cl appear, and the atomic percentages of Eu(III) and phosphate are 1.12 and 0.3%, respectively. The carbon weight percentage remains consistent, while the weight percentage of oxygen decreases significantly after Eu(III) and phosphate co-removal. One can deduce that the adsorbed Eu(III) complex with parts of the ionized oxygen-containing functional groups on GO surfaces, which will change the atomic environment of oxygen and make these oxygen escape from EDS detection due to the low sensitivity of EDS technique. Accordingly, it can be deduced that the oxygen-containing functional groups participate in Eu(III) and phosphate co-removal process on GO surfaces.

In order to further investigate the interaction mechanism of Eu(III) and phosphate with GO, the obtained GO samples before and after the co-removal process (i.e., denoted as GO and GO + Eu + P, respectively) are characterized by XPS technique. Fig. 14a shows the XPS survey spectra of GO and GO + Eu + P. One can see that carbon and oxygen are the predominant elements on GO surfaces, and the peaks of Eu 3d and P 2p are observed in the survey spectrum of GO + Eu + P.²⁰ Fig. 14b shows the high resolution scans for O 1s of GO and GO + Eu + P. The shift of binding energy and the decrease of peak area of O 1s spectra are observed after the co-removal process, suggesting that Eu(III) and phosphate are chemically adsorbed on GO surfaces.²⁰ The high resolution scans for Eu 3d_5/2 and P 2p of GO + Eu + P are shown in Fig. 14c and d, respectively. Fig. 14e and f show the deconvolutions of C 1s and O 1s high resolution scans before and after the co-removal process. One can see that there are significant differences in C 1s spectra associated with oxygen-containing functional groups such as C–O in carbonyl (286.6 eV) and O–C=O in carboxylic (287.0 eV) groups between GO and GO + Eu + P (both in intensities and peak positions). The O 1s spectra of GO and GO + Eu + P differ from each other significantly both in the shape and maximum position. The XPS survey spectra and high resolution scans for C 1s and O 1s clearly demonstrate that the co-removal process of Eu(III) and phosphate is related to the oxygen-containing functional groups on GO surfaces.

Fig. 14 (a) XPS survey spectra and (b) high resolution scans for O 1s spectra of GO before and after the sorption of Eu(III) and phosphate, high resolution scans for (c) Eu 3d_5/2 and (d) P 2p, high resolution XPS spectra for (e) C 1s and (f) O 1s before and after the sorption of Eu(III) and phosphate.

3.4.2. Zeta potential of GO suspensions before and after sorption. The variations of the ZP value of GO before and after Eu(III) and/or phosphate sorption as a function of pH were measured and shown in Fig. 15. The results indicate that GO has negative ZP values over the whole tested pH range, and further decreases with increasing pH due to the ionization of the oxygen-containing function groups.¹⁸ After the addition of phosphate into the suspension, the ZP values of GO further shifts toward lower pH, making GO become more attractive to Eu(III). While, after the addition of Eu(III) into the suspension, the ZP values of GO shifts toward higher pH, making GO become more attractive to phosphate. Therefore, the co-removal process of Eu(III) and phosphate by GO is electrostatically favorable. Based on Visual MINTEQ calculation, in the absence of GO, saturation index of EuPO₄ is positive at pH > 2.0 for an aqueous system containing 0.05 mmol L⁻¹ Eu(III) and 0.05 mmol P L⁻¹ in contact with an atmosphere containing 0.00038 atm CO₂ (Fig. S3 in ESI†), suggesting the formation of EuPO₄ precipitate is thermodynamically favored.⁴² As our previous study reported when GO, Eu(III) and phosphate were added simultaneously, the sorption process occurred more quickly than precipitation, forming ternary surface complexes with Eu(III) as a bridge linking GO and phosphate.⁴³ This is why the presence of phosphate decreases the ZP values of GO–Eu (Fig. 15). We conclude that when the saturation index of EuPO₄ is not high at the low pH values, Eu(III) was adsorbed on GO firstly and then served as the sorption site for phosphate retention. When the saturation index of EuPO₄ is extremely high with the pH values further increasing, the formation of EuPO₄ precipitate will become more thermodynamically favorable. Therefore, the enhanced removal amounts of Eu(III) and phosphate on GO surfaces are caused by electrostatic attraction, ternary surface complexation, and surface precipitation.

Fig. 15 Zeta potentials of GO before and after Eu(III) and/or phosphate sorption versus pH. m/V = 25 mg L⁻¹, I = 0.01 mol L⁻¹ NaCl.

4. Conclusions

In this study, batch experiments including contact time, pH and temperature were employed to study the performance of GO in the co-removal of Eu(III) and phosphate. The sorption characteristics of GO for Eu(III) and phosphate can be affected obviously by the presence of the other one, and the co-removal process is highly dependent on pH and temperature. The sorption kinetics of Eu(III) and phosphate follow the pseudo-second-order model, and the sorption isotherms can be well fitted by the Langmuir isotherm model. The standard thermodynamic parameters indicate that the co-removal process of Eu(III) and phosphate on GO surfaces is endothermic and spontaneous. The sorption enhancement of Eu(III) in the presence of phosphate will minimize the potential risk of Eu(III) to water quality. Findings of this study highlight the potential utility of GO as the attractive adsorbent for Eu(III) and phosphate co-removal in nuclear wastewater treatment.

Acknowledgements

Financial supports from Anhui Provincial Natural Science Foundation (1608085QB44 and 1508085MB29), and the National Natural Science Foundation of China (21307135, 21225730, 41273134, 21377132, 91326202) are acknowledged.

References

1. Y. B. Sun, J. X. Li and X. K. Wang, Geochim. Cosmochim. Acta, 2014, 140, 621–643.
2. Y. B. Sun, Z. Y. Wu, X. X. Wang, C. C. Ding, W. C. Cheng, S. H. Yu and X. K. Wang, Environ. Sci. Technol., 2016, 50, 4459–4467.
3. T. Wen, X. L. Wu, M. C. Liu, Z. H. Xing, X. K. Wang and A. W. Xu, Dalton Trans., 2014, 43, 7464–7472.
4. J. A. Shusterman, H. E. Mason, J. Bowers, A. Bruchet, E. C. Uribe, A. B. Kersting and H. Nitsche, ACS Appl. Mater. Interfaces, 2015, 7, 20591–20599.
5. X. X. Wang, S. B. Yang, W. Q. Shi, J. X. Li, T. Hayat and X. K. Wang, Environ. Sci. Technol., 2015, 49, 11721–11728.
6. X. X. Wang, S. S. Lu, L. Chen, J. X. Li, S. Y. Dai and X. K. Wang, J. Radioanal. Nucl. Chem., 2015, 306, 497–505.
7. C. R. Li, Y. Huang and Z. Lin, J. Mater. Chem. A, 2014, 2, 14979–14985.
8. H. B. Yin, M. X. Han and W. Y. Tang, Chem. Eng. J., 2016, 285, 671–678.
9. Z. Y. Chen, Q. Jin, Z. J. Guo, G. Montavon and W. S. Wu, Chem. Eng. J., 2014, 256, 61–68.
10. G. X. Zhao, T. Wen, X. Yang, S. B. Yang, J. L. Liao, J. Hu, D. D. Shao and X. K. Wang, Dalton Trans., 2012, 41, 6182–6188.
11. Y. B. Sun, Q. Wang, C. L. Chen, X. L. Tan and X. K. Wang, Environ. Sci. Technol., 2012, 46, 6020–6027.
12. A. Y. Romanchuk, A. S. Slesarev, S. N. Kalmykov, D. V. Kosynkin and J. M. Tour, Phys. Chem. Chem. Phys., 2013, 15, 2321–2327.
13. S. Bachmaf, B. P. Friedrich and B. J. Merkel, Radiochim. Acta, 2008, 96, 359–366.
14. S. R. Wang, W. L. Yi, S. W. Yang, X. C. Jin, G. D. Wang and F. C. Wu, Appl. Geochem., 2011, 26, 286–292.
15. X. L. Li, X. R. Wang, L. Zhang, S. W. Lee and H. J. Dai, Science, 2008, 319, 1229–1232.
16. Y. B. Sun, S. B. Yang, C. C. Ding, Z. X. Jin and W. C. Cheng, RSC Adv., 2015, 5, 24886–24892.
17. G. X. Zhao, L. Jiang, Y. D. He, J. X. Li, H. L. Dong, X. K. Wang and W. P. Hu, Adv. Mater., 2011, 23, 3959–3963.
18. H. Xu, G. Li, J. Li, C. L. Chen and X. M. Ren, J. Mol. Liq., 2016, 213, 58–68.
19. Y. B. Sun, S. B. Yang, Y. Chen, C. C. Ding, W. C. Cheng and X. K. Wang, Environ. Sci. Technol., 2015, 49, 4255–4262.
20. C. C. Ding, W. C. Cheng, Y. B. Sun and X. K. Wang, Dalton Trans., 2014, 43, 3888–3896.
21. S. T. Yang, P. F. Zong, X. M. Ren, Q. Wang and X. K. Wang, ACS Appl. Mater. Interfaces, 2012, 4, 6891–6900.
22. Y. S. Ho and G. Mckay, Process Saf. Environ. Prot., 1998, 76, 332–340.
23. Y. S. Ho and G. Mckay, Process Biochem., 1999, 34, 451–465.
24. W. C. Song, X. X. Wang, Q. Wang, D. D. Shao and X. K. Wang, Phys. Chem. Chem. Phys., 2015, 17, 398–406.
25. V. C. T. Costodes, H. Fauduet, C. Porte and A. Delacroix, J. Hazard. Mater., 2003, 105, 121–142.
26. C. O. Ijagbemi, M. H. Baek and D. S. Kim, J. Hazard. Mater., 2009, 166, 538–546.
27. K. Urano and H. Tachikawa, Ind. Eng. Chem. Res., 1991, 30, 1897–1899.
28. P. N. Pathak and G. R. Choppin, Radiochim. Acta, 2007, 95, 267–273.
29. X. M. Ren, S. T. Yang, X. L. Tan, C. L. Chen, G. D. Sheng and X. K. Wang, J. Hazard. Mater., 2012, 237, 199–208.
30. I. Langmuir, J. Am. Chem. Soc., 1918, 40, 1361–1403.
31. H. M. A. Frendlich, J. Phys. Chem., 1906, 57, 385.
32. H. Xu, B. S. Zhu, X. M. Ren, D. D. Shao, X. L. Tan and C. L. Chen, J. Hazard. Mater., 2016, 314, 249–259.
33. D. Dayananda, V. R. Sarva, S. V. Prasad, J. Arunachalam and N. N. Ghosh, Chem. Eng. J., 2014, 248, 430–439.
34. B. H. Hameed, A. T. M. Din and A. L. Ahmad, J. Hazard. Mater., 2007, 141, 819–825.
35. S. B. Wang and Z. H. Zhu, J. Hazard. Mater., 2006, 136, 946–952.
36. V. S. Mane, I. D. Mall and V. C. Srivastava, Dyes Pigm., 2007, 73, 269–278.
37. K. A. Krishnan and A. Haridas, J. Hazard. Mater., 2008, 152, 527–535.
38. T. P. Moss, J. X. Wang, S. Jones, E. May, D. Olive, Z. R. Dai, M. Zavarin, A. B. Kersting, D. Y. Zhao and H. Nitsche, J. Mater. Chem. A, 2014, 2, 11209–11221.
39. L. Zeng, X. M. Li and J. D. Liu, Water Res., 2004, 38, 1318–1326.
40. J. R. Li, L. Zhu, J. F. Tang, K. Qin, G. Li and T. H. Wang, RSC Adv., 2016, 6, 23252–23259.
41. F. Long, J. L. Gong, G. M. Zeng, L. Chen, X. Y. Wang, J. H. Deng, Q. Y. Niu, H. Y. Zhang and X. R. Zhang, Chem. Eng. J., 2011, 171, 448–455.
42. I. Carabante, M. Grahn, A. Holmgren, J. Kumpiene and J. Hedlund, Environ. Sci. Technol., 2012, 46, 13152–13159.
43. X. M. Ren, Q. Y. Wu, H. Xu, D. D. Shao, X. L. Tan, W. Q. Shi, C. L. Chen, J. X. Li, T. Hayat and X. K. Wang, Environ. Sci. Technol., 2016 DOI:10.1021/acs.est.6b02934.

---

Footnote

† Electronic supplementary information (ESI) available: Working curves simulation (Fig. S1), SEM image and EDS patterns of GO (Fig. S2), and saturation index (Fig. S3). See DOI: 10.1039/c6ra18629g

---